© 2003 ASM International. All Rights Reserved. Brazing (#06955G)
Brazing Second Edition
Mel M. Schwartz
ASM International® Materials Park, Ohio 44073-0002 www.asminternational.org
www.asminternational.org
© 2003 ASM International. All Rights Reserved. Brazing (#06955G)
www.asminternational.org
Copyright © 2003 by ASM International® All rights reserved No part of this book may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the written permission of the copyright owner. First printing, September 2003
Great care is taken in the compilation and production of this book, but it should be made clear that NO WARRANTIES, EXPRESS OR IMPLIED, INCLUDING, WITHOUT LIMITATION, WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE, ARE GIVEN IN CONNECTION WITH THIS PUBLICATION. Although this information is believed to be accurate by ASM, ASM cannot guarantee that favorable results will be obtained from the use of this publication alone. This publication is intended for use by persons having technical skill, at their sole discretion and risk. Since the conditions of product or material use are outside of ASM’s control, ASM assumes no liability or obligation in connection with any use of this information. No claim of any kind, whether as to products or information in this publication, and whether or not based on negligence, shall be greater in amount than the purchase price of this product or publication in respect of which damages are claimed. THE REMEDY HEREBY PROVIDED SHALL BE THE EXCLUSIVE AND SOLE REMEDY OF BUYER, AND IN NO EVENT SHALL EITHER PARTY BE LIABLE FOR SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES WHETHER OR NOT CAUSED BY OR RESULTING FROM THE NEGLIGENCE OF SUCH PARTY. As with any material, evaluation of the material under end-use conditions prior to specification is essential. Therefore, specific testing under actual conditions is recommended. Nothing contained in this book shall be construed as a grant of any right of manufacture, sale, use, or reproduction, in connection with any method, process, apparatus, product, composition, or system, whether or not covered by letters patent, copyright, or trademark, and nothing contained in this book shall be construed as a defense against any alleged infringement of letters patent, copyright, or trademark, or as a defense against liability for such infringement. Comments, criticisms, and suggestions are invited, and should be forwarded to ASM International. Prepared under the direction of the ASM International Technical Book Committee (2002–2003), Charles A. Parker, Chair. ASM International staff who worked on this project include Scott Henry, Assistant Director of Reference Publications; Bonnie Sanders, Manager of Production; and Nancy Hrivnak and Jill Kinson, Production Editors. Library of Congress Cataloging-in-Publication Data Schwartz, Mel M. Brazing / Mel M. Schwartz.—2nd ed. p. cm. Includes bibliographical references and index. ISBN 0-87170-784-5 1. Brazing. I. Title. TT267.S39 2003 671.5′6—dc21
2003051963 ISBN: 0-87170-784-5 SAN: 204-7586 ASM International® Materials Park, OH 44073-0002 www.asminternational.org Printed in the United States of America
© 2003 ASM International. All Rights Reserved. Brazing (#06955G)
www.asminternational.org
Contents Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii Chapter 1
Introduction to Brazing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 Brazing versus Soldering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 Historical Development of Brazing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 Advantages and Limitations of Brazing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Mechanics of Brazing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 Brazing versus Other Welding Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
Chapter 2
Brazing Fundamentals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 Adhesion, Wetting, Spreading, and Capillary Attraction . . . . . . . . . . . . . . . . . . . . . . . 7 Practical Experience, Work-Related Tips, and Problem-Solving . . . . . . . . . . . . . . . 12
Chapter 3
Elements of the Brazing Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 Filler-Metal Flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 Base-Metal Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 Filler-Metal Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 Surface Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 Joint Design and Clearance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 Temperature and Time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 Processes (Heat Sources) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 Torch Brazing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 Furnace Brazing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 Induction Brazing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 Controlled-Atmosphere Brazing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 New Induction Brazing Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 Resistance Brazing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 Carbon Resistance Brazing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 Direct Resistance Brazing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 Dip Brazing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 Infrared (Quartz) Brazing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 Exothermic Brazing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 Laser Brazing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 Braze Welding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53 Diffusion Brazing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 Microwave Brazing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
iii
© 2003 ASM International. All Rights Reserved. Brazing (#06955G)
www.asminternational.org
The Future of Braze Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 Practical Experience, Work-Related Tips, and Problem-Solving . . . . . . . . . . . . . . . 59 Chapter 4
Base Metals and Base-Metal Family Groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 Metallurgical Reactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 Base-Metal Family Groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70 Case Histories and Problem-Solving Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
Chapter 5
Brazing Filler Metals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177 Basic Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177 Filler-Metal Selection Criteria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179 Filler-Metal Types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181 Specialized Brazing Filler Metals and Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . 211 Filler-Metal Selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219 Filler-Metal Forms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219 Case Histories and Problem-Solving Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231
Chapter 6
Fluxes and Atmospheres . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243 Atmospheres . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243 Fluxes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 266 Case Histories and Problem-Solving Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281
Chapter 7
Fixturing, Tooling, Stopoffs, Parting Agents, Surface Preparation, Surface Cleaning, and Repair . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 289 Fixturing and Tooling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 289 Stopoff Materials and Parting Agents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 295 Surface Cleaning and Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 297 Surface Preparation for Specific Base Metals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 302 Clean Rooms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 307 Vacuum Brazing Cleaning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 307 Postbrazing Treatments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 308 Repair Techniques with Cleaning Agents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 308 Case Histories and Problem-Solving Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . 308
Chapter 8
Joint Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313 Types of Joints . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313 Joint Clearance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 315 Design for Assembly . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 319 Effects of Brazing Variables on Clearance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 322 Strength . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 327 Joint Design and Ceramics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 329 Case Histories and Problem-Solving Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . 336
Chapter 9
Evaluation and Quality Control of Brazed Joints . . . . . . . . . . . . . . . . . . . . . . . . 339 Design and Quality System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 339 Quality Standards for Brazing and Brazing Processes . . . . . . . . . . . . . . . . . . . . . . . 340 Types of Common Defects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 341 Brazing Process Planning and Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 341 Nondestructive Inspection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 342 Design Testing, Evaluation, and Feedback . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 344 Destructive Inspection and Testing Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 345 iv
© 2003 ASM International. All Rights Reserved. Brazing (#06955G)
www.asminternational.org
Chapter 10 Applications and Future Outlook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 347 Automation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 347 Fluxless Brazing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 351 Novel and Emerging Brazing Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 352 Future Outlook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 373 Case Histories and Problem-Solving Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . 376 Abbreviations and Symbols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 383 Subject Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 385 Filler Metal Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 403
v
© 2003 ASM International. All Rights Reserved. Brazing (#06955G)
www.asminternational.org
© 2003 ASM International. All Rights Reserved. Brazing (#06955G)
www.asminternational.org
Preface This updated and revised second edition of Brazing is intended to provide the reader with the information needed to braze materials that will be used in the 21st century. My goal has been to fuse experiences, basic understandings, theories, and practical information from the past with guidance about expected brazing applications and problem-solving considerations for the future and what it may hold for materials joining. It will be interesting to observe the continuing development and use of novel types of equipment and how engineers, scientists, and technology innovators will tackle the unknowns of this new millennium. In addition, this revised edition includes lessons learned on tooling, design, materials, atmospheres, processing, and equipment throughout the book. These examples should be helpful to the new brazing initiate as well as to more experienced brazing technologists and engineers. Several new and emerging topics are covered, including nanostructures and materials, microwave and laser brazing, more effective use of vacuum atmospheres, functionally gradient materials (FGM), and intermetallics, to name just a few. Coverage has been increased on beryllium alloys, aluminumlithium alloys, new titanium alloys, various composites (metal, ceramic, intermetallic), ceramic-tometal brazing, and ceramic-to-ceramic brazing. The enforcement of environmental regulations and the rising cost of metals such as copper and silver have forced manufacturers of brazing and soldering products to develop new fluxes and filler metals, while devising more efficient means of applying existing products. These trends have been reflected in the content of this new edition. Residual flux and its disposal have come under the scrutiny of regulatory agencies. To address this concern, a family of concentrated fluxes has come on the market that provides improved fluxing performance with smaller amounts being applied to the part. The quantities used can be reduced even more with automated dispensing, which places a small amount of flux exactly where it is required rather than the more expansive manual brushing of the flux. The increasing use of aluminum for parts has spawned the development of better aluminum filler metals and fluxes for brazing. The elimination of cadmium from braze filler metals is another area that has attracted attention in recent years. Cadmium has for years been used in certain brazing filler metals because of its characteristic as a temperature depressant, which facilitates melting and wetting of the filler metal. OSHA has established stringent regulations on allowable cadmium levels in the air. Some manufacturers have addressed this through the installation of sensitive filtration equipment and closing off the brazing area from other manufacturing operations. Another approach has been to utilize substitute cadmium-free braze filler metals now commercially available. These substitutes are similar in wetting characteristics to the cadmium types, but they do require more heat to produce the wetting. This factor should be taken into consideration when establishing the application of the heat. Other products that are considered are brazing powders and rods pre-coated with flux and equipment that automatically dispenses exact amounts of filler metals in paste form. vii
© 2003 ASM International. All Rights Reserved. Brazing (#06955G)
www.asminternational.org
Modeling can be used to increase reliability, repeatability, and efficiency of furnace operations. That is, the optimum cycle for a particular brazing process and workload is determined and a program is designed to allow the influencing parameters to be automatically compensated in real time, thereby achieving optimum operation. The basis for actual in-process control is modern computer methods and more sophisticated knowledge of the brazing process. The thinking and understanding of companies and their management must become global, and engineers must be able to reduce the technology development cycle for new materials and brazing systems from a norm of five to seven years to two to three years. One final area of concentrated development work is ceramic joining. Although the ability to fabricate ceramic materials that are reliably strong and tough is continuously advancing, the successful utilization of these materials will depend on the ability to assemble simple components into structures that will function effectively. While metal joining can be performed by starting at one end of the joint, ceramic joining requires the entire joint region to be brought into a reactive state simultaneously. Preliminary supplemental heating may be needed to prevent thermal shock during joining. A transfer of the experience gained in other material fields could prove helpful. Adaptation of physical and chemical deposition methods, surface engineering by laser beams, ion implantation, and chemical doping to improve surface reactivity and solid state adhesion are just a few of the techniques available for implementation. The development of generic joining technology for ceramic materials is vital and should be pursued in the context of metallurgical and ceramic sciences. Furthermore, it should be paralleled by an engineering effort to design and construct equipment specific to ceramic joining. Extending joining technology to ceramic materials will allow implementation of advanced technologies in circumstances where ceramics are the only materials having the requisite properties. Such a development provides an opportunity for the expertise of joining engineers plus venture capitalists to cash in on solving the underlying basic problems and thus augment the international competitiveness of U.S. products. If one tries to view the future and what it holds, woven through these fanciful visions are items that reflect trends observed in current reality. For instance, with the current pace of technological change, it can take less than six months for a state-of-the-art computer to become old technology. And software viruses can gobble up monthly production reports in seconds. But there is also good news. Within the next few years, micro-robots may crawl through intelligent joining systems, performing preventive maintenance in areas previously inaccessible. Solar power could become the cost-effective, environmentally safe way to fuel many manufacturing processes. Traditional methods of design and manufacture will give way to concurrent engineering (CE) strategies that enable improved communications between design and manufacturing. Automakers, aircraft manufacturers, and even job shops will design for manufacturability. These strategies will strongly influence development and deployment of advanced manufacturing technology well into the new millennium. However, with all the enthusiasm, enhancement, and advancements made by all the above industries, the basics in joining have not been altered. In spite of the move to automation, computer-controlled and monitored processes, and robots, the basic principles for all the various joining processes and fabrication techniques have not changed. You still need a heating media/source, material to be joined, a filler material to be added or preplaced in most cases, and a protective atmosphere/flux or vacuum in many situations. Over the past several years, many organizations have surveyed the changing face of science and engineering within the United States. In each case, the conclusions were the same: • Fewer engineering students will be available for industrial employment during the first decade of the 21st century.
viii
© 2003 ASM International. All Rights Reserved. Brazing (#06955G)
www.asminternational.org
• As a result of the aging work force and benefits of early retirement programs, the demands for qualified joining (brazing) engineers will increase especially with the retirement in 2010–2012 of the baby boomers. • The changing face of industry will result in the need for more technically trained individuals. Department of Labor statistics predict an increased need in the engineering work force of 165% for the ten-year period from 1999 to 2009. As indicated in a recent National Science Foundation report, “the educational system in the United states has always been hierarchical. For this reason, the progression (and attrition) of students from primary education through graduate programs is conceptualized as a pipeline.” The feeder to engineering education resides in middle school (grades 6 to 8). It is here that students begin to take an interest in career opportunities and the educational requirements needed to enter these professional careers. This is where they begin planning their high school curriculum, which will either prepare them for a college education or vocational endeavors. The real solution will take the cooperative effort from each portion of industry, i.e., professional societies, industry and educational organizations. Professional societies and educational foundations can assist in image-building and career planning. Finally, the secondary schools and universities must begin to view themselves as equal partners with industry and societies in course requirement development and articulation agreements. A knowledge of metallurgy is basic to an understanding of the brazing process. This does not mean that the brazing engineer, brazer, and the brazing inspector must become a metallurgist. But it does mean that engineering schools of the world increasingly must face up to the fact that in the future engineers must play a key part in management’s strategic planning group. The university system recognizes that tomorrow’s engineers will have to have enhanced capabilities due to three significant factors: increased product sophistication and variation; a global manufacturing environment; and a multitude of social and economic changes. The 21st century engineer or technologist will require a radically different education from that of his predecessors. The workplace of the future will need systems integrators, not individuals classified as metallurgists, brazing technologists, manufacturing engineers, quality engineers, or industrial engineers; individuals will need knowledge and talent in each of these fields. The engineer of the future must be fully conversant with modern materials applications. Those who are involved in developing industrial computer networks must also be familiar with the manufacturing processes themselves. Improvements in joining science and technology must keep pace with advances in materials science and technology, or else the benefits of these new materials will not be achieved in the marketplace. Mel Schwartz
ix
ASM International is the society for materials
engineers and scientists, a worldwide network dedicated to advancing industry, technology, and applications of metals and materials. ASM International, Materials Park, Ohio, USA www.asminternational.org This publication is copyright © ASM International®. All rights reserved. Publication title
Product code
Brazing
#06955G
To order products from ASM International: Online Visit www.asminternational.org/bookstore Telephone 1-800-336-5152 (US) or 1-440-338-5151 (Outside US) Fax 1-440-338-4634 Mail
Customer Service, ASM International 9639 Kinsman Rd, Materials Park, Ohio 44073-0002, USA
Email
[email protected]
American Technical Publishers Ltd. 27-29 Knowl Piece, Wilbury Way, Hitchin Hertfordshire SG4 0SX, In Europe United Kingdom Telephone: 01462 437933 (account holders), 01462 431525 (credit card)
www.ameritech.co.uk Neutrino Inc. In Japan Takahashi Bldg., 44-3 Fuda 1-chome, Chofu-Shi, Tokyo 182 Japan Telephone: 81 (0) 424 84 5550 Terms of Use. This publication is being made available in PDF format as a benefit to members and customers of ASM International. You may download and print a copy of this publication for your personal use only. Other use and distribution is prohibited without the express written permission of ASM International. No warranties, express or implied, including, without limitation, warranties of merchantability or fitness for a particular purpose, are given in connection with this publication. Although this information is believed to be accurate by ASM, ASM cannot guarantee that favorable results will be obtained from the use of this publication alone. This publication is intended for use by persons having technical skill, at their sole discretion and risk. Since the conditions of product or material use are outside of ASM's control, ASM assumes no liability or obligation in connection with any use of this information. As with any material, evaluation of the material under end-use conditions prior to specification is essential. Therefore, specific testing under actual conditions is recommended. Nothing contained in this publication shall be construed as a grant of any right of manufacture, sale, use, or reproduction, in connection with any method, process, apparatus, product, composition, or system, whether or not covered by letters patent, copyright, or trademark, and nothing contained in this publication shall be construed as a defense against any alleged infringement of letters patent, copyright, or trademark, or as a defense against liability for such infringement.
Brazing Second Edition Mel M. Schwartz, p1-5 DOI: 10.1361/brse2003p001
Copyright © 2003 ASM International® All rights reserved. www.asminternational.org
CHAPTER 1
Introduction to Brazing MANY PRODUCTS are assembled from two or more individual components that are often permanently joined to produce structurally sound assemblies. Joining methods include various fasteners, interference-type joints, adhesives, and, for the highest-integrity joints, the many techniques classified under welding. Brazing, although fundamentally different from fusion welding, is one such technique; soldering is another. The term brazing encompasses a group of welding processes that produce coalescence of materials by heating them to the brazing temperature in the presence of a filler metal having a liquidus above 450 °C (840 °F) and below the solidus of the base metal. (The liquidus, or melting point, is the lowest temperature at which a metal or an alloy is completely liquid, and the solidus is the highest temperature at which a metal or an alloy is completely solid.) The brazing filler metal is distributed between the closely fitted faying surfaces of the joint by capillary action. In the brazing process, the materials involved are heated to a temperature approximately 56 °C (100 °F) above the temperature of the filler material (usually a metal) that is being used. The filler metal turns liquid, covers all the mating surfaces, and creates an alloy bond with the faying surfaces. The filler metal can be preplaced, plated, or applied from an external source. If the filler metal comes from outside the mating surfaces, it requires capillary action to draw it between the faying surfaces (see Chapter 2, “Brazing Fundamentals”). When selecting a joining method, many factors, including requirements, must be considered, for example, service temperature and environment (corrosive, galvanic, etc.); service requirements for reliability, fatigue resistance,
impact resistance, and other cyclic conditions; and manufacturing considerations, such as cost, equipment required, and other assembly details. Brazing (usually conducted at 540 to 1620 °C, or 1000 to 2950 °F) is considered a hightemperature joining process compared to adhesive bonding, mechanical fastening, and soldering. Therefore, it cannot be used for plastics; however, it is used for metal-matrix composites and ceramic-matrix composites. Lap joints, which are the most common type of brazed joint, are usually as strong as or stronger than the base materials being joined as long as (a) the faying surfaces overlap for a distance equal to at least 3 times the thickness of the thinner of the two members being joined, and (b) the clearance between the two parts ( joint thickness) is kept to approximately 0.075 mm (0.003 in.) or less at braze temperature (Ref 1).
Brazing versus Soldering The basic distinction between brazing and soldering is that brazing is conducted at higher temperatures (soldering processes use filler metals having a liquidus not exceeding 450 °C, or 840 °F). The historical distinction between the processes has its origin in the earliest solders, which were based on tin, while brazes were based on copper-zinc alloys. Indeed, the word braze is a derivation of the Old English braes, meaning to cover with brass. On the other hand, the term solder is an adaptation of the Old French soudure, meaning to make solid. Brazing and soldering require the application of a number of scientific and engineering skills to produce joints of satisfactory quality and reliability. Brazing employs higher temperatures than soldering, but the fundamental concepts
2 / Brazing, Second Edition
are similar, particularly with respect to metallurgy and surface chemistry (Table 1.1). However, joint design, materials to be joined, fillermetal and flux selection, heating methods, and joint preparation can vary widely between the two processes. Economic considerations involving filler-metal and process technology are also varied, particularly in relation to automated techniques and inspection and testing. Brazing and soldering are performed in many industries, from exotic applications in the electronics and aerospace field to everyday plumbing applications. The type of metallurgical reaction between a filler and parent metal is sometimes used to differentiate soldering from brazing. Solders usually react to form intermetallic phases, that is, compounds of the constituent elements that have different atomic arrangements than the elements in solid form. By contrast, most brazes form solid solutions, which are mixtures of the constituents on an atomic scale. However, this distinction does not have universal validity. For example, silver-copper-phosphorus brazes react with steels to form the interfacial phase of Fe3P in a similar manner to the reaction of tin-base solders with iron or steels to form FeSn2. On the other hand, solid solutions form between silverlead solders and copper just as they do between the common silver-base brazes and copper. Soldering and brazing involve the same bonding mechanism, that is, reaction with the parent material, usually alloying, so as to form metallic bonds at the interface. In both situations, good wetting promotes the formation of fillets that serve to enhance the strength of the joints. Similar processing conditions are re-
quired, and the physical properties are comparable, provided that the same homologous temperature is used for the comparison.
Historical Development of Brazing Early metalworkers, stimulated by a desire to produce structures that were difficult or impossible to build using methods then in existence, realized that it was possible to fill the joint between two metal pieces with molten metal and allow it to solidify. These artisans soon learned by experience that, in order to achieve adherence, the metals to be joined and the filler metal had to be kept free of oxides and the filler metal had to have a lower melting point, and, furthermore, that a given filler metal would not necessarily adhere to all metals. From these basic requirements, brazing and soldering grew into crafts whose practitioners were well versed in what to do and what not to do in order to produce sound joints. Just as these joining techniques developed empirically, so did the lower-melting-point filler metals. Workers first used lead and tin solders as well as silver and copper-arsenic ores, which were readily available and had low melting points. Later, the alloy brass was developed and found to be more desirable for joining copper, silver, and steel structures, because it provided higher-strength joints and could withstand higher temperatures. Early silversmiths, probably wanting to produce white solder joints for aesthetic reasons, melted brass and silver together and found it to have an even lower melting point than brass,
Table 1.1 Comparison of soldering, brazing, and welding Process Parameter
Joint formed Filler-metal melt temperature, °C (°F) Base metal Fluxes used to protect and to assist in wetting of base-metal surfaces Typical heat sources Tendency to warp or burn Residual stresses
Soldering
Brazing
Welding
Mechanical <450 (<840)
Metallurgical >450 (>840)(a)
Metallurgical >450 (>840)(b)
Does not melt Required
Does not melt Optional
Optional
Soldering iron; ultrasonics; resistance; oven Atypical
Furnace; chemical reaction; induction; torch; infrared Atypical
...
(a) Less than melting point of base metal. (b) Less than or equal to melting point of base metal
...
...
Plasma; electron beam; tungsten and submerged arc; resistance; laser Potential distortion and warpage of base metal likely Likely around weld area
Chapter 1: Introduction to Brazing / 3
good adherence, and better corrosion resistance. Although innumerable combinations of silver, copper, and zinc subsequently evolved, primarily to meet melting-point requirements, these silver-brass, brass, and lead-tin alloys were essentially the only brazing and soldering filler materials available for generations.
Advantages and Limitations of Brazing Advantages. Brazing has many distinct advantages, including the following: • Economical fabrication of complex and multicomponent assemblies • Simple method to obtain extensive joint area or joint length • Joint temperature capability approaching that of base metal • Excellent stress distribution and heat-transfer properties • Ability to preserve protective metal coating or cladding • Ability to join cast materials to wrought metals • Ability to join nonmetals to metals • Ability to join metal thicknesses that vary widely in size • Ability to join dissimilar metals • Ability to join porous metal components • Ability to fabricate large assemblies in a stress-free condition • Ability to preserve special metallurgical characteristics of metals • Ability to join fiber- and dispersion-strengthened composites • Capability for precision production tolerance • Reproducible and reliable quality-control techniques Strong, uniform, leakproof joints can be made rapidly, inexpensively, and even simultaneously. Joints that are inaccessible and parts that may not be joinable at all by other methods often can be joined by brazing. Complicated assemblies comprising thick and thin sections, odd shapes, and differing wrought and cast alloys can be turned into integral components by a single trip through a brazing furnace or a dip pot. Metal as thin as 0.01 mm (0.0004 in.) and as thick as 150 mm (6 in.) can be brazed. Brazing allows for the substitution of complex castings (Fig. 1.1) (Ref 2) for plate, bar, or tubular shapes. The cost and times for manufac-
turing enter the picture here; one must consider the relative cost of material, availability of material, schedule to manufacture the assembly, and the production equipment. The brazing equipment could be torch, induction, resistance, or furnace. Figure 1.2 illustrates an assembly with several internal brazed joints that can be simultaneously joined in one production load in, for example, a furnace using fluxed joints and/or an inert or vacuum atmosphere where no flux is used. Figure 1.3 illustrates the joining of two copper plates whereby large surface areas are covered by the liquid filler metal. The criteria of strength and corrosion resistance were the key requirements that convinced the joining engineers to select brazing over several other joining techniques. Brazed joint strength is high. The nature of the interatomic (metallic) bond is such that even a simple joint, when properly designed and
Fig. 1.1
Use of brazing to enable replacement of complex castings with assemblies of basic components. The casting shown (top) needs to be faced, drilled, and tapped in three places. It is much easier to braze three threaded couplings/ tubes into a machined block (bottom). Source: Ref 2
4 / Brazing, Second Edition
made, has strength equal to or greater than that of the as-brazed parent metal. The natural shapes of brazing fillets are excellent. The meniscus surface formed by the fillet metal as it curves across corners and adjoining sections is ideally shaped to resist fatigue. (It should be noted that in brazed joints using eutectic-type filler metal, fillets often contain an excessive amount of brittle intermetallic compounds. In fact, fillets are 5 to 10 times thicker than the joint and thus have a much higher volume of the liquid phase from which these brittle phases crystallize. Therefore, crack nucleation often originates in fillets.) Complex shapes with
Fig. 1.2
Assembly in which several internal brazed joints are accomplished at the same time in one furnace brazing operation by the use of internal brazing performs. Source: Ref 2
greatly varied sections can be brazed with little distortion, and precise joining is comparatively simple. Unlike welding, in which the application of intense heat to small areas acts to move the parts out of alignment and introduces residual stresses, brazing involves fairly even heating, and thus part alignment is easier. Limitations. A brazed joint is not a homologous body but rather is heterogeneous, composed of different phases with differing physical and chemical properties. In the simplest case, it consists of the base-metal parts to be joined and the added filler metal. However, partial dissolution of the base metal, combined with diffusion processes, can change the composition and therefore the chemical and physical properties of the boundary zone formed at the interface between base metal and filler metal and often of the entire joint. Thus, in addition to the two different materials present in the simplest example given previously, a complicated transitional or even completely different zone must be considered. In determining the strength of such heterogeneous joints, the simplified concepts of elasticity and plasticity theory—valid for a homogeneous metallic body where imposed stresses are uniformly transmitted from one surface or space element to the adjacent ones—no longer apply. In a brazed joint formed of several materials with different characteristics of deformation resistance and deformation speed, the stresses caused by externally applied loads are nonuniformly distributed.
Mechanics of Brazing Brazing involves a limited dissolution or plastic deformation of the base metal (Ref 3). Brazing proceeds through four distinct steps:
Fig. 1.3
Brazing used to join large surface areas. Channels were grooved into these two copper plates prior to brazing to serve as internal cooling channels. Source: Ref 2
1. The assembly or the region of the parts to be joined is heated to a temperature of at least 450 °C (840 °F). 2. The assembled parts and brazing filler metal reach a temperature high enough to melt the filler metal (foil, wire, paste, platings, etc.) but not the parts. 3. The molten filler metal, held in the joint by surface tension, spreads into the joint and wets the base-metal surfaces. 4. The parts are cooled to solidify, or “freeze,” the filler metal, which is held in the joint by capillary attraction and anchors the parts
Chapter 1: Introduction to Brazing / 5
together by metallurgical reaction and atomic bonding.
Brazing versus Other Welding Processes The mere fact that brazing does not involve any substantial melting of the base metals offers several advantages over other welding processes. It is generally possible to maintain closer assembly tolerances and to produce a cosmetically neater joint without costly secondary operations. Even more important, however, is that brazing makes it possible to join dissimilar metals (or metals to ceramics) that, because of metallurgical incompatibilities, cannot be joined by traditional fusion welding processes. If the base metals do not have to be melted to be joined, it does not matter that they have widely different melting points. Therefore, steel can be brazed to copper as easily as to another steel. Brazing also generally produces less thermally induced distortion, or warping, than fusion welding. An entire part can be brought up to the same brazing temperature, thereby preventing the kind of localized heating that causes distortion in welding. Finally, and perhaps most important to the manufacturing engineer, brazing readily lends itself to mass-production techniques. It is relatively easy to automate, because the application of heat does not have to be localized, as in fusion welding, and the application of filler metal is less critical. In fact, given the proper clearance conditions and heat, a brazed joint tends to make itself and is not dependent on operator skill, as are most fusion welding processes. Automation is also simplified by the fact that there are many means of applying heat to the joint, including torches, furnaces, induction coils, electrical resistance, and dipping. Several joints in one assembly often can be produced in one multiple-braze operation during one heating cycle, further enhancing production automation.
As noted in Table 1.1, essentially no melting of the base metal occurs in brazing; however, the temperatures involved can affect the properties of the metals being joined. For example, base metals whose mechanical properties were obtained by cold working may soften or undergo grain growth if the brazing temperature is above their recrystallization temperatures. Mechanical properties obtained by heat treatment may be altered by the heat of brazing. On the other hand, materials in the annealed condition are usually not altered by brazing. As with other welding processes, brazing produces a heat-affected zone (HAZ) with a strongly altered microstructure due to intensive mutual mass transfer between base metal and filler metal. The width of this zone varies with the heating process used. In torch and induction brazing, for example, only a localized zone is heated; in furnace and dip brazing, the entire part is subjected to the brazing temperature. As a rule, the HAZ produced during brazing is wider and less sharply defined than those resulting from other fusion-related processes. ACKNOWLEDGMENT
Portions of this article are adapted from M.M. Schwartz, Introduction to Brazing and Soldering, Welding, Brazing, and Soldering, Volume 6, ASM Handbook, ASM International, 1993, pages 109–113.
REFERENCES
1. Brazing Handbook, 4th ed., American Welding Society, 1991 2. W.D. Kay, Ten Reasons to Choose Brazing, Weld. J., Sept 2000, p 33–35 3. E. Lieberman, Modern Soldering and Brazing Techniques, Business News Publications, 1988
Brazing Second Edition Mel M. Schwartz, p7-13 DOI: 10.1361/brse2003p007
Copyright © 2003 ASM International® All rights reserved. www.asminternational.org
CHAPTER 2
Brazing Fundamentals BRAZING does not involve any melting or plastic state of the base metal. Brazing comprises a group of joining processes in which coalescence is produced by heating to suitable temperatures above 450 °C (840 °F) and by using a ferrous and/or nonferrous filler metal that must have a liquidus temperature above 450 °C and below the solidus temperature(s) of the base metal(s). The filler metal is distributed between the closely fitted surfaces of the joint by capillary attraction. Brazing is distinguished from soldering in that soldering employs a filler metal having a liquidus below 450 °C. Brazing has four distinct characteristics: • The coalescence, joining, or uniting of an assembly of two or more parts into one structure is achieved by heating the assembly or the region of the parts to be joined to a temperature of 450 °C or above. • Assembled parts and filler metal are heated to a temperature high enough to melt the filler metal but not the parts. • The molten filler metal spreads into the joint and must wet the base-metal surfaces. • The parts are cooled to freeze the filler metal, which is held in the joint by capillary attraction and anchors the part together.
Adhesion, Wetting, Spreading, and Capillary Attraction Metals More than 195 years ago, Thomas Young (Ref 1) proposed treating the contact angle (θ) of a liquid as the result of the mechanical equilibrium of a drop resting on a plain, solid surface under the action of three surface tensions (Fig. 2.1). The surface tensions are γlv at the interface
of the liquid in equilibrium with its saturated vapor, γsl at the interface between the solid and the liquid, and γsv at the interface of the solid in equilibrium with the saturated vapor of the liquid. Hence: γsv = γlv cos θ + γsl
(Eq 1)
It is important to keep in mind that phases are supposed to be mutually in equilibrium. The designation γsv is a reminder that the solid surface near the liquid should have an equilibrium film of vapor due to the film pressure. Young’s equation has been used extensively in literature, which reflects its general acceptance. However, Eq 1 has never been verified experimentally. The problem is that surface tensions of solids are not easy to measure due to the inevitable presence of the interfacial tension between a solid and its liquid. More importantly, there is the difficulty that any tensile stresses existing in the surface of the solid would prevent the system from being in equilibrium. The surface tension at the solid-vapor interface (γsv) has a relationship with surface tension of a solid in vacuum (γs) as follows: γsv = γs – πe
(Eq 2)
where πe refers to the spreading pressure. Consequently, Young’s equation may be rewritten as: γs = γlv cos θ + γsl + πe
(Eq 3)
Because most of the solids have a negligible πe, particularly when the contact angle (θ) is greater than 10°, Young’s equation becomes: γs = γlv cos θ + γsl
(Eq 4)
A decrease of the contact angle causes an increase of the liquid drop surface area and thus increases the total liquid surface free energy.
Chapter 2: Brazing Fundamentals / 9
expansion. It can be shown thermodynamically that in the absence of a reaction, the driving force for wetting does not exceed γlv, resulting in a steady-state contact angle (Ref 3). The driving force with the contribution of the free energy of reaction in most cases exceeds the resisting force represented by γlv, because θ is 0° during spreading. A condition of an expanding drop during a reaction is defined as spreading. It can be seen that the free energy of a reaction in which the substrate is a passive participant does not contribute to the driving force for wetting; thus, spreading does not occur. The contact angle, however, adjusts to conform with the surface-energy changes of the liquid caused by composition changes due to the reaction. Example: Copper-Silver System. The equilibrium phase diagram for the copper-silver binary system (Fig. 2.2) can be used to illustrate examples of wetting and spreading (Ref 3). The system has a eutectic at 780 °C (1430 °F), with 72 wt% Ag. At 900 °C (1650 °F), the solid-solution limit is 5 wt% Ag in copper and 8 wt% Cu in silver. Several compositions are identified in the phase diagram by the letters A to D. When a drop of liquid C is placed on solid B at 900 °C (1650 °F), wetting occurs, with a contact angle of 11° and no chemical reaction, because the phases are in chemical equilibrium. This behav-
Fig. 2.2
ior corresponds to γsv > γlv, because, in a given system, the surface free energy of a liquid is less than that of a solid, due to its lack of long-range order. The liquid thus has the opportunity to rearrange its surface structure to a lower freeenergy state. However, when liquid C is placed on solid A, spreading occurs, because substrate A (as an active participant in the reaction) changes its surface composition toward B. The third equation in Fig. 2.1 applies in this case. Another example is that of liquid D on solid B. Liquid D is not in equilibrium with B and dissolves some of the substrate to change its composition to C. Even though a reaction occurs, there is no spreading, because B is a passive participant with no change in composition, even though it is being dissolved. However, with liquid D on solid A, spreading occurs, because both are active participants as they change to equilibrium compositions C and B, respectively. In both of the latter examples, liquid D is an active participant, because it dissolves some of the substrate to reach equilibrium compositions. It does not, however, contribute to spreading, which is controlled by the active participation of the substrate.
Ceramics Joining dissimilar materials invariably results in high interfacial energy; that is, the work of
Stable phase equilibrium diagram for the copper-silver system
10 / Brazing, Second Edition
adhesion is not sufficient to maintain the joint integrity. Most structural ceramic-metal interfaces are no exceptions; the liquid metal does not readily wet the ceramic surface. Therefore, the first challenge in metal-to-ceramic joining is to alter the interfacial thermodynamics to render the ceramic surface wettable. There are two approaches available today (Ref 4) to accomplish this task: metallization of the ceramic surface and reaction wetting. A thin layer of metal alloy is deposited on the ceramic by vapor deposition or sputtering. Brazing is then carried out by appropriate filler material or by simply melting the deposited layer (Ref 5). This two-step approach is not as enthusiastically embraced by industry as is direct reaction brazing. In reaction brazing, the filler metal is carefully chosen so as to facilitate compound formation at the interface. A small percentage of reactive metals, such as aluminum and titanium, are added to the otherwise inert base alloys (Ref 6, 7). The compounds that form are commonly spinels for the oxide ceramics and complex nitrides for the ceramic nitrides (Ref 8, 9). It is important to realize that wetting in such systems is time dependent. Successful bond formation relies on rapid transport of the reactive metal to the interface and a rapid rate of compound formation. Reaction wetting may not be the solution to all metal-ceramic joining
problems, because interfacial compound formation could create fragile layers (Ref 9), further complicating the development of good joint efficiency. The weakest among the adhesive strength at the ceramic-compound boundary, the cohesive strength of the compound, or the adhesive strength between compound and metal would determine the final joint strength. Table 2.1 (Ref 10) summarizes the various reaction products that have been identified in literature for the common metal-ceramic systems. Many researchers have concentrated their study on alumina surfaces, because it is one of the very few ceramics for which essential thermodynamic data are available. One of the first commercial applications of metal-to-ceramic brazed components is the turbocharger rotor (Ref 8), where a silicon nitride turbine blade is brazed to a stainless steel shaft. Other ceramics of interest to brazers are silicon carbide and zirconia. Ceramics exhibit very different thermal expansion behavior compared to metals; hence, considerable residual stress can build up during cooling. This thermal expansion mismatch more or less dictates the use of a ductile filler material. Most commercial brazing systems are therefore silver and copper base. The soft interlayer might not be sufficient to compensate for large differences in thermal expansion coeffi-
Table 2.1 Ceramic-metal interface formation and reaction products Bonding conditions System
Temperature, K
Load, MPa (ksi)
Al2O3(a)-Nb(a) Al2O3(b)-Nb(c) Al2O3(d)-Nb(a)
1925 1973 1973
20 (3) 10 (1.5) 6.4 (1.0)
0.1 2 1
Argon Vacuum 10–4 Pa Vacuum 10–3 Pa
(Al)Nb (O)Nb NbOx, (Al,O)Nb
Al2O3(b)-Nb(a) Al2O3(e)-Ti(a) SiC(b)-Ti(a) SiC(b)-Zr(a) SiC(a)-Al/Ti/Al(a) SiC(a)-Al(a) Si3N4(a)-Fe(a) SiO2-Al(a)
1973 1250 1773 1773 1273 1373 1683 875 ... 1270 1773
6.4 (1.0) ... 0.34 (0.05) 0.56 (0.08) 0.56 (0.08) 0 (0) 3000 (435) 10 (1.5) ... 20 (3) 20 (3)
1 60 1 1 1 1 1 10–4 24 0.25 0.25
Vacuum 10–3 Pa Vacuum 10–4 Pa Vacuum 10–3 Pa Vacuum 10–3 Pa Air Vacuum 10–2 Pa Argon 75N2/25H2 (H2/H2O = 102) H2/H2O = 2 × 105 H2/H2O = 105
1773 1473 1273 ...
20 (3) 1 (0.15) 0 (0) ...
0.25 0.25 0.6 8
H2/H2O = 5 × 105 H2/H2O = 2 × 103 H2 H2
NbOx, (Al,O)Nb Ti3Al, TiO, (AlO)Ti Ti3SiC2, Ti5Si(C), TiSi2 ZrSi, ZrC + (Si)Zr TiC, TiAl3Si Al4C, (Si)Al Fe3Si, (Si)Fe Not observed α-Al2O3, (Si)Al None None. After 6 h, (Al)Pt; after 1000 h, Pt3Al Pt3Al, (Al)Pt Pt3Si (Zr)Pt Pt3Zr, (Zr)Pt
Al2O3(b)-Cu(a) Al2O3(b)-Pt(a) Al2O3(b)-Pt(a) SiO2-Pt(a) ZrO2(a)-Pt
Time, h
Plane of interaction: (a) Polycrystal. (b) (0001). (c) (110). (d) (1010). (e) (1100). Source: Ref 10
Atmosphere
Reaction products
Chapter 2: Brazing Fundamentals / 11
cients (e.g., Si3N4 as compared to stainless steel). In such situations, laminated interlayers that provide a continuous gradient thermal expansion coefficient are used (Ref 6). Thermodynamic phenomena that occur at the interface can be studied in terms of the contact angle, θ, and the work of adhesion, W. These terms can be related to various surface or interfacial energies. The general case for a liquid metal in contact with a solid ceramic is shown in Eq 1, where a balance of surface tension forces results in the familiar Young’s equation. The Dupree equation is easily derived from Eq 1: W = γlv (1 + cos θ)
(Eq 7)
Attempts at understanding the nature of the force of adhesion across the interface have not been very successful. In 1965, researchers (Ref 11) rationalized, on the basis of the work of adhesion data for an alumina-metal interface, that the observed work of adhesion was the sum of two independent contributions arising from the van der Waals forces and a primary chemical bond. Predicting adhesion data in joining an alumina-metal interface is of great importance in many applications. The objective of proposed research to predict wettability and bond strength from measurable parameters and bridge the gap between a theoretical understanding and technology of observed work was undertaken (Ref 12). Researchers (Ref 13, 14) attempted to explain the entire work of adhesion across the metalceramic boundary in terms of physical forces using the dielectric principle. Such models are not of much use to the brazing industry, because most commercial metal-to-ceramic bonds are based on chemical bond formations.
Effects of Capillary Attraction and Wetting on Brazing Capillary attraction makes leak-tight joints a simple proposition for brazing. In a properly designed joint, the molten filler metal is normally drawn completely through the joint area without any voids or gaps, and brazed joints remain liquid- and gas-tight under heavy pressures, even when the joint is subjected to shock or vibrational types of loading. Capillary action results in the phenomenon where surface tension causes molten braze filler
metal to be drawn into the area that covers the parallel surfaces that are to be brazed. Capillarity is a result of surface tension between base metals(s), filler metal, flux, or atmosphere and the contact angle between base metal and filler metal. In actual practice, fillermetal flow characteristics are also influenced by dynamic considerations involving viscosity, vapor pressure, gravity, and metallurgical reactions between filler metal and base metal. As a matter of fact, present-day brazing practices have evolved as the result of an empirical approach to the phenomena of wetting and spreading, which are of prime importance in the formation of brazed joints. Classical, physical, and chemical principles led to equations governing the shape of liquid surfaces and the rate of filling a capillary gap in systems that do not interreact. However, the extension of theory to practical systems necessitates the consideration of a number of complicating factors, which often arise in everyday practice. A few of these factors include the condition of the solid surface as to the presence of oxide films and their effects on wetting and spreading, surface roughness, alloying between the filler metal and base metal and the extent to which this affects the thermodynamic properties of the liquid and solid surfaces, and the condition and properties of the brazing atmosphere. The factors that control the rate at which wetting, spreading, and capillary flow occur are of great practical, as well as theoretical, interest. Studies have indicated profound influences of various kinds of surface activation that cannot be explained in terms of surface energies or alterations in equilibrium contact angle (Ref 15, 16). Some of the most spectacular of these effects have been observed in systems in which a finite contact angle is thermodynamically unstable, because the solid-vapor surface energy exceeds the sum of the liquid-solid surface energies—that is, a system in which thermodynamics would predict complete spreading. In actual fact, spreading may or may not occur in this type of system, and the rate of spreading can be markedly dependent on surface chemistry, although the fundamental mechanisms of this dependence are not all clear. Wetting is, perhaps, best understood by example. If a solid is immersed in a liquid bath and wetting occurs, a thin, continuous layer of liquid adheres to the solid when it is removed from the liquid. Technically speaking, in the
12 / Brazing, Second Edition
wetting process, the force of adhesion between the solid and the liquid is greater than the cohesive force of the liquid. In practical terms, with respect to brazing, wetting implies that the liquid filler metal spreads on the solid base metal instead of balling up on its surface (Fig. 2.3). It has been demonstrated that wetting actually depends on a slight surface alloying of the base metal with the filler metal. A comprehensive theory of the wetting or spreading of liquids on solid surfaces is presented in Ref 17 and 18. It can be concluded that wetting is the ability of the molten filler metal to adhere to the surface of a metal in the solid state and, when cooled below its solidus temperature, to make a strong bond with that metal. Wetting is a function not only of the filler metal but also of the nature of the metal or metals to be joined. There is considerable evidence that in order to wet well, a molten metal must be capable of dissolving, or alloying with, some of the metal on which it flows. Wetting is only one important facet of the brazing process. A very important factor affecting wetting is the cleanliness of the surface to be wetted. Oxide layers inhibit wetting and spreading, as do grease, dirt, and other contaminants that prevent good contact between the filler metal and the base metal. One of the functions of a flux is to remove the oxide layer on the joint area and to expose clean base metal.
Fig. 2.3
Wetting and dewetting
Good wetting and spreading of the liquid filler metal on the base metal are necessary in brazing, because the mechanics of the process demand that the filler metal be brought smoothly, rapidly, and continuously to the joint opening. If the conditions within the capillary space of the joint do not promote good wetting, the filler metal is not drawn into the space by capillary attraction. It all boils down to the fact that, for successful joining of components by brazing, the filler metal selected must have a melting point above 450 °C (840 °F) and must also wet the base metal without melting it. Then, the joint must be designed so that the mating surfaces of the components are parallel and close enough together to cause capillary attraction.
Practical Experience, Work-Related Tips, and Problem Solving In order to braze tungsten carbide (WC) granules or diamonds to 1010 carbon steel wheels with BNi-2 filler metal, avoid the problem of filler metal sagging. Because the wheels must rest flat in the furnace, the diamonds or carbides are then on the vertical surface of the wheel diameter. Typical brazing takes place at 1040 °C (1900 °F) in a pure dry hydrogen atmosphere. There are a large number of variables that must be taken into consideration. It is much easier and more practical to use temperature as the controlling variable. Key variables that affect braze quality include: • The chemistry which is controlled by a specification and as a result the melting and flow characteristics of the filler metal are controlled by the chemistry. • Partial pressure of nitrogen will affect the melting characteristics of the filler metal. • A variation in the partial pressure of oxygen will affect the melting and flow characteristics of the filler. • The length of time in the oxidation range of 540 to 925 °C (1000 to 1700 °F) can also alter the melting and flow characteristics of the filler metal. • The heating rate, particularly at the high temperature where diffusion takes place, can alter the melting and flow characteristics.
Chapter 2: Brazing Fundamentals / 13
• The maximum brazing temperature is the best variable to control, because any one of the previously mentioned variables can change, requiring a change in the brazing temperature. A better braze filler metal for this type of application would be a very widemelting-range material and a filler metal of Cr-Ni-B-Si-Fe. This filler metal has a melting range of 970 to 1160 °C (1780 to 2120 °F). • Therefore, the large number of variables presents a problem, but considering and taking into account the various variables, control of the flow of the filler metal is feasible. REFERENCES
1. T. Young, Philos. Trans. R. Soc. (London) A, Vol 95, 1805, p 65 2. A.W. Adamson, Physical Chemistry of Surfaces, 4th ed., John Wiley & Sons, 1982, p 339 3. P.R. Sharps, A.P. Tomsia, and J.A. Pask, Wetting and Spreading in the Cu-Ag System, Acta Metall., Vol 29 (No. 7), 1981, p 855–865 4. M. Erg and A.W. Hennicke, Ceramics in Advanced Energy Technologies, A. Krockel et al., Ed., Dreidel Publishing, 1982, p 138 5. M.E. Twentyman and P. Hancock, in Surfaces and Interfaces in Ceramic and Ceramic-Metal Systems, Vol 14, Materials Science Research, J.A. Pask and A.G. Evans, Ed., Plenum Press, 1981, p 535 6. H. Mizuhara, Vacuum Brazing Ceramics to Metals, Adv. Mater. Process., Vol 131 (No. 2), Feb 1987, p 53–55 7. A.J. Moorhead and A. Keating, Direct Brazing of Ceramic for Advanced HeavyDuty Diesels, Weld. J., Oct 1986, p 117
8. R.E. Loehman, Interfacial Reactions in Ceramic-Metal Systems, Ceram. Bull., Vol 68 (No. 4), 1989, p 891 9. M.G. Nicholas and R.J. Lee, Joining Dissimilar Materials, Met. Mater., Vol 5 (No. 6), 1989, p 348 10. J.T. Klomp, in Ceramic Microstructures 86: Role of Interfaces, J.A. Pask and A.G. Evans, Ed., Plenum Press, 1988, p 307 11. J.E. McDonald and J.G. Eberhart, Adhesion in Aluminum Oxide-Metal Systems, Trans. AIME, Vol 233, 1965, p 512 12. G.R. Edwards and J.J. Moore, “Investigation of Brazing Alloys for Ceramic Substrates,” Research Proposal CSM 3264, Colorado School of Mines, Feb 1990, p 54–70 13. R.G. Barrera and C.B. Duke, Dielectric Continuum Theory of the Electronic Structure of Interfaces, Phys. Rev. B., Vol 13 (No. 10), 1976, p 4477 14. A.M. Stoneham and P.W. Tasker, in Ceramic Microstructures 86, Vol 21, Materials Science Research, J.A. Pask and A.G. Evans, Ed., Plenum Press, 1988, p 155 15. C.M. Adams, Jr., “Dynamics of Wetting in Brazing and Soldering,” Technical Report WAL TR 650/1, Army Materials Research Agency, Watertown Arsenal, Watertown, MA, July 1962 16. S. Weiss and C.M. Adams, Jr., The Promotion of Wetting, Weld. J., Vol 46 (No. 2), Feb 1967, p 49s–57s 17. W.D. Hawkins, Physical Chemistry of Surface Films, Reinhold, 1952, p 1–413 18. M.M. Schwartz, Fundamentals of Brazing, Welding, Brazing, and Soldering, Vol 6, ASM Handbook, ASM International, 1993, p 114–125
Brazing Second Edition Mel M. Schwartz, p15-62 DOI: 10.1361/brse2003p015
Copyright © 2003 ASM International® All rights reserved. www.asminternational.org
CHAPTER 3
Elements of the Brazing Process IN ORDER TO PRODUCE satisfactory brazed joints, a careful and intelligent appraisal of the following elements is required: • • • • • • •
Filler-metal flow Base-metal characteristics Filler-metal characteristics Surface preparation Joint design and clearance Temperature and time Rate and source of heating
This chapter reviews the important considerations related to each of these elements.
Filler-Metal Flow Wetting is only one important facet of the brazing process. If the molten filler metal does not flow into the joint, the effectiveness of the filler metal is greatly restricted. Flow is facilitated by capillary attraction, which in turn results from surface-energy effects (Ref 1, 2). It is therefore apparent that a high liquid surface tension, a low contact angle, and low viscosity are desirable for promoting filler-metal flow. Thus, a low contact angle, which implies wetting, is a necessary but not a sufficient condition for flow. Viscosity is also important. Filler metals with narrow freezing ranges that are close to the eutectic composition generally have lower viscosities than those with wide freezing ranges. Flow is the property of a filler metal that determines the distance it will travel away from its original position because of the action of capillary forces. In practice, reactions between the filler metal and the base metal are usually minimized by (a) selecting the proper filler metal, (b) keeping the brazing temperature as low as possible but high
enough to produce flow, (c) keeping the time at temperature short, and (d) cooling the brazed joint as quickly as possible without causing cracking or excessive distortion.
Base-Metal Characteristics The base metal has a prime effect on joint strength. A high-strength base metal produces joints of greater strength than those made with softer base metals (other factors being equal). When hardenable metals are brazed, joint strength becomes less predictable. This is because there are more complex metallurgical reactions involved between hardenable base metals and the filler metals. These reactions can cause changes in the base-metal hardenability and can create residual stresses (Ref 3). Also vitally important are the coefficients of thermal expansion (CTEs) where different materials make up the assembly and gaps can open or close as heating proceeds to the joining temperature. Also, during cooling after the filler metal has solidified, differences between the CTEs generate residual stresses that may produce distortion. The relative ease of brazing some base metals and other materials is shown in Table 3.1. There are several metallurgical phenomena that influence the behavior of brazed joints and, in some instances, necessitate special procedures. Included among these base-metal effects are (a) alloying; (b) carbide precipitation; (c) stress cracking; (d) hydrogen, sulfur, and phosphorus embrittlement; and (e) oxide stability. In addition to the base-metal effects mentioned previously and the normal mechanical requirements of the base metal in the brazement, the effect of the brazing cycle on the base metal
16 / Brazing, Second Edition
and the final joint strength must be considered. Base-metal alloys that are strengthened by cold working will be annealed and the joint strength reduced when the brazing-process temperature and time are in the annealing range of the base metal being processed. Hot-cold-worked heat resistant base metals also can be brazed; however, only the annealed physical properties will be available in the brazement. The brazing cycle by its very nature will usually anneal the coldworked base metal, unless the brazing temperature is very low and the time at elevated temperature is very short. It usually is not practical to cold work the base metal after the brazing operation. When it is essential to design a brazement having strength above the annealed strength of the base metal after the brazing operation, specifying a heat treatable base metal is necessary. The base metals can be of the oilquench type, the air-quench type that can be brazed and hardened in the same or separate operations, or the precipitation-hardening type in which the brazing cycle and solution treatment cycle may be combined. Parts can be hardened and then brazed with a low-temperature filler metal employing short times at elevated temperature in order to maintain the mechanical properties (prevent softening by annealing). The strength of the base metal has a profound effect on the strength of the brazed joint; thus, this property must be clearly kept in mind when designing the joint for specific properties. Also, some base metals are easier to braze than others. This is the case particularly for certain base metals joined by specific brazing processes. For example, reactive metals (titanium, beryllium, etc.) brazed in vacuum or in an inert atmosphere in a furnace are easier to braze than with a torch process.
Table 3.1 Relative ease of brazing various base materials Degree of difficulty
Impossible Difficult Fair
Easy
Materials
None Ti, Zr, Be, and their alloys; ceramics; graphite; glass; TiC Al, W, Mo, Ta, alloys with more than 5% metals forming refractory oxides, cast iron, WC Cu, Ni, Co, and their alloys; steels; precious metals
Filler-Metal Characteristics The second material involved in joint structures is the filler metal. The term brazing filler metal is essentially synonymous with the commonly employed term brazing alloy. Its selection is important but not for the reasons many engineers think. A specific filler metal cannot be chosen to produce a specific joint strength, which is unfortunate but true. Actually, brazing can provide strong joints with almost any good commercial filler metal if brazing methods and joint design are selected and applied correctly. Several characteristics that filler metals must possess or are desirable are: • Proper fluidity at the brazing temperature to ensure flow by capillary action and to provide full alloy distribution • Stability to avoid premature release of lowmelting-point elements in the filler metal • Ability to wet the base-metal joint surfaces • Low-volatilization characteristics of the alloying elements of the filler metal at the brazing temperature • Ability to alloy or combine with the parent metal to form an alloy with a higher melting temperature • Controllability of the washing or erosion between the filler metal and the parent metal within the limits required for the brazing operation • Depending on the service requirements, the ability to produce or avoid base-metal/fillermetal interactions One of the most broadly misunderstood facts relating to filler metals is that brazed-joint strength is completely unrelated directly to the melting method used. This fact is hard to accept, because it seems to contradict a long-established metallurgical truth with regard to the manufacture of steels and other constructional metals. The effect of melting practice on filler metals, however, is not the same as that of melting practice on metals during primary manufacturing. If constructional metals are produced by vacuum melting, for example, there is a definite relationship between the vacuum-melting practice and the final strength of the ingot, bar, or rolled sheet. That is not true with a filler metal, because the joint strength depends on such factors as joint design, state of stress, brazing temperature, amount of filler metal applied, loca-
Chapter 3: Elements of the Brazing Process / 17
tion and method of application, heating rate, holding time at the peak temperature, and many other considerations that make up what is termed brazing technique. The process by which filler metals penetrate and alloy with base metals during brazing is referred to as diffusion. In applications requiring strong joints for high-temperature, highstress service conditions (such as turbine rotor assemblies and jet engine components), it is generally good practice to specify a filler metal that diffuses readily and alloys with the base metal. When the assembly is constructed of extremely thin base metals (as in honeycomb structures and some heat exchangers), good practice generally calls for a filler metal with a low-diffusion characteristic relative to the base metal being used. Diffusion is an essential and normal part of the metallurgical process that contributes to good brazed joints. In choosing a filler metal, the first criterion is the working temperature. Very few filler metals possess distinct melting points. Filler metals in which the solidus and liquidus are close together do not usually exhibit a strong tendency to separate, and they are relatively fluid. They flow readily and should be used with small joint clearances. Other filler-metal selection criteria include corrosion resistance, such as oxidation and galvanic corrosion with other parts of the assembly and the service environment; color match to the base metal; electrical and thermal conductivity; joint-filling capacity; hardness and machinability; ductility and fracture toughness; and ability to form fillets. Additionally, the designer must consider the extent of alloying with the base metal. The improvement in mechanical properties of the joint and the increase in remelt temperature obtained by alloying could be beneficial. Structural changes in the interface layers of the base metals; aggressive, extensive alloying between the base metal and filler metals; formation of brittle intermetallic compounds; and erosion of components can be deleterious, and, at times, catastrophic. All such effects, beneficial or otherwise, vary greatly with the joint gap, temperature and time of brazing, and compositions of the base metals and the filler metal. Although the mechanical properties of filler metals in massive form can provide a guide to their suitability for use in different capillaryjoining applications, in general, designers can-
not work with the mechanical properties of assemblies brazed using different joint configurations or cycles of time and temperature. Finally, the placement of the filler metal is an important design consideration, not only because the joint must be accessible to the heating or filler-metal placement method chosen, but also because, in automatic heating setups, the filler metal must be retained in its location until molten. Filler metals are available in different forms, and selection may depend on which form is suitable for a particular joint design. The most common filler-metal forms are wire, powder, foil, flux paste, strip, shim, and nonfluxing paste. Preforms made from wire, strip, and foil can be used. Several general rules apply in filler-metal placement. Wherever possible, place the filler metal on the most slowly heated part of the assembly in order to ensure complete melting of the filler metal. Gravity may be used to assist filler-metal flow, particularly for those filler metals having wide ranges between their solidus and liquidus temperatures. Filler metals can be chosen to fill wide gaps or to flow through joint configurations where the gap can vary, for example, around a corner. Unless movement between the components being joined is unimportant or can be corrected manually, through self-jigging or by fixtures after the filler metal is molten, the filler metal should be placed outside the joint and allowed to flow through it and should not be placed between the joint members. To avoid forming a void, the filler alloy should not flow from two sides to a center point. If erosion of thin members is possible, the filler metal should be placed on the heavier sections, which heat up more slowly, so that flow proceeds toward the thin sections. Apart from suiting the placement method selected for the joint, the form chosen for the filler metal may be needed to gage accurately the amount applied—not just for economy and reproducibility but also to regulate and maintain joint properties and configuration.
Surface Preparation A clean and nearly oxide-free surface is imperative to ensure uniform quality and sound brazed joints. A sound joint may be obtained more readily if all grease, oil, wax, dirt, and nearly all oxides have been carefully removed
18 / Brazing, Second Edition
from the base and filler metals before brazing, because only then can uniform capillary attraction be obtained. Oils and greases should be removed before cleaning to remove oxides. Even after proper cleaning, all metal surfaces exposed to air will have a thin oxide film, which does not impede filler-metal flow during brazing. Brazing as soon as possible after the material has been cleaned is recommended. The length of time the cleaning remains effective depends on the metals involved, atmospheric conditions, storage and handling practices, and other factors. Cleaning is commonly divided into two major categories: chemical and mechanical. Chemical cleaning is the most effective means of removing all traces of oil or grease. Trichlorethylene and trisodium phosphate are the usual cleaning agents employed. Various types of oxides and scale that cannot be eliminated by these cleaners are removed by other chemical means. The selection of the chemical cleaning agent depends on the nature of the contaminant, the base metal, the surface condition, and the joint design. Regardless of the cleaning agent or the method used, it is important that all residue or surface film be removed from the cleaned parts by adequate rinsing to prevent the formation of other equally undesirable films on the faying surfaces. When faying surfaces of parts to be brazed are prepared by blasting techniques, several factors should be understood and considered. The purposes of blasting parts to be brazed are to remove any oxide film and to roughen the mating surfaces so that capillary attraction of the filler metal is increased. The blasting material must be clean and must be of a type that does not leave on the surfaces to be joined any deposit that restricts filler metal or impairs brazing. The particles of the blasting material should be fragmented rather than spherical, so that the blasted parts are lightly roughened rather than peened. The operation should be done so that delicate parts are not distorted or otherwise harmed. Vapor blasting and similar wet blasting methods require care because of possible surface contamination. Mechanical cleaning may be adequate, in which case the design must permit this during manufacture. In some cases where chemical cleaning is required, it may be followed by protective electroplating, necessitating access to the faying surface by the liquids involved.
Another technique in surface cleaning and protection is the use of solid and liquid brazing fluxes. At temperatures up to approximately 1000 °C (1830 °F), fluxes often provide the easiest method of producing or maintaining surface cleanliness, and, in such cases, the design must not only permit easy ingress of the flux but also allow the filler metal to wash it through the joint. Above 1000 °C, the flux residues can be difficult to remove, and surface cleaning by, for example, a furnace atmosphere is desirable, but the design must permit the gas to penetrate the joint. Apart from cleanliness and freedom from significant oxides, surface roughness is important in determining ease and evenness of flow of the filler metal. This varies with different manufacturing methods and may influence the engineer’s choice, or it may require access to a surface treatment/roughening process. Generally, a liquid that wets a smooth surface will wet a rough one even more. A rough surface will modify filler-metal flow from laminar to turbulent, prolong flow time, and increase the possibility of alloying and other interactions. Surfaces often are not truly planar, and, in some instances, surface roughening will improve the uniformity of the joint clearance. Conversely, the designer and engineer may require that filler metal not flow onto some surfaces. Stopoff materials often avoid this, but the design must permit easy application of the stopoff without danger of contaminating the surfaces to be joined. Self-fluxing filler metals in a suitably protective environment, such as vacuum, may provide the essential surface wetting. Examples include copper-phosphorus on copper and silver-copper-lithium on stainless steels; additional examples are given in Chapter 4, “Base Metals and Base-Metal Family Groups.”
Joint Design and Clearance A brazed joint is not a homogeneous body. Rather, it is a heterogeneous assembly that is composed of different materials with different mechanical, physical, and chemical properties. It is the simplest case; it consists of the basemetal parts to be joined and the added filler metal. Diffusion processes, however, change the composition and therefore the chemical and physical properties of the boundary zone formed at the interface between base metal and
Chapter 3: Elements of the Brazing Process / 19
filler metal. Thus, in addition to the two different materials present in the simplest example given previously, further dissimilar materials must be considered. Why should small clearances be used? The smaller the clearance, in situations where there is not extensive alloying and erosion, the easier it is for capillarity to distribute the filler metal throughout the joint area, and there is less likelihood that voids or shrinkage cavities will form as the filler metal solidifies. Small clearances and correspondingly thin filler-metal films make strong joints. The soundest joints are those in which 100% of the joint area is wetted and filled by the filler metal. They are at least as high in tensile strength as the filler metal itself, and often higher. If brazing clearances ranging from 0.03 to 0.08 mm (0.0012 to 0.003 in.) are designed, they are designed for the best capillary action and the greatest joint strength. Before the detailed design for a part to be brazed is made, the first decision is how and where the components are to be joined. Because brazing relies on capillary attraction, the design of a joint must provide an unobstructed and unbroken capillary path to enable flux, if used, to escape from it as well as to allow the filler metal to get into the joint. Where filler metal is added to a joint by hand, such as by feeding in a rod or wire, the joint entry must be visible and accessible. If preplaced rings or shims are used, the joint must be designed so that the preform can be placed in position easily and remains in place until molten. Some of the more important factors influencing joint design are the required strength and corrosion resistance, the necessary electrical and thermal conductivity, the materials to be joined, the mode of application of the filler metal, and postjoining inspection needs. The actual design itself should follow from a consideration of all the previously mentioned factors. Consideration also should be given to the ductility of the base metal, the stress conditions of the joint, and the relative movements of the two surfaces during joining, which may introduce problems and inaccuracies requiring careful consideration by the designer. The lack of concentricity of circular components may cause gaps to vary from excessive to over-tight. This can be avoided by designs and techniques such as knurling, splining, use of shims, and machining. Viscosity, surface tension, and specific gravity of the filler metal are not the only factors that
determine the gap-filling capability of a given filler metal. Many other considerations are involved, such as the tendency of the filler metal and parent materials to alloy with one another. Joint strength increases as joint gap decreases, down to a minimum. For stressed applications, optimal joint clearance may have to be designed inside the gap-filling range of the filler metal. Table 3.2 shows allowable joint clearances for various filler-metal systems. Other factors influencing optimal joint gap with a specific filler metal are joint length, brazing temperature, and base-metal reactions. It is important to remember that an assembly expands during heating and that the joint gap may either widen or close by the time the filler metal starts to melt and move. It is desirable to design the joint so that the solidifying filler metal is exposed to compressive rather than tensile stresses. This is much more important in brazing than in soldering, because brazing temperatures are higher, increasing the total expansion. With cylindrical joints, the component with the larger coefficient of expansion should, whenever possible, be on the outside. It is equally important to make sure that at the joining temperature the gap does not become impossibly wide. If the components in the assembly have to be reversed, design modifications can reduce stresses, but a sufficient gap must be provided to ensure that narrowing of the joint gap at the brazing temperature does not provide insufficient clearance for filler-metal flow. Finally, it is important to ensure that there is sufficient filler metal to absorb room-temperature tensile stresses in order to compensate for any reduction in joint gap.
Table 3.2 Preferred gaps for different brazing filler metals Brazing filler-metal system
Al-Si alloys(a) Mg alloys Cu Cu-P Cu-Zn Ag alloys Au alloys Ni-P alloys Ni-Cr alloys(b) Pd alloys
Joint clearance, mm (in.)
0.15–0.61 (0.006–0.024) 0.10–0.25 (0.004–0.010) 0.00–0.05 (0.000–0.002) 0.03–0.13 (0.001–0.005) 0.05–0.13 (0.002–0.005) 0.05–0.13 (0.002–0.005) 0.03–0.13 (0.001–0.005) 0.00–0.03 (0.000–0.001) 0.03–0.61 (0.001–0.024) 0.03–0.10 (0.001–0.004)
(a) If joint length is less than 6 mm (0.240 in.), gap is 0.12 to 0.75 mm (0.005 to 0.030 in.). If joint length exceeds 6 mm (0.240 in.), gap is 0.25 to 0.60 mm (0.010 to 0.024 in.). (b) Many different nickel brazing filler metals are available, and joint-gap requirements may vary greatly from one filler metal to another.
20 / Brazing, Second Edition
Joint clearance is probably one of the most significant factors in all types of brazing operations. Naturally, joint clearance receives special consideration when joints are designed at room temperature. Actually, joint clearance is not the same at all phases of brazing. It has one value before brazing, another value at the brazing temperature, and still another value after brazing, especially if there has been diffusion of the filler metal into the base metal. To avoid confusion, it has become general practice to specify joint clearance as the value at room temperature before brazing. The recommended joint clearances given in Table 3.2 are based on joints having members of similar metals and equal mass. When dissimilar metals and/or metals of widely differing masses are joined by brazing, special problems arise that necessitate more specialized selection among the various filler metals, and the joint clearance suitable for the job at hand must be carefully determined. Although there are many kinds of brazed joints, selection of joint type is not as complicated as it may seem, because butt and lap joints are the two fundamental types. All others, such as the scarf joint, are modifications of these two. The scarf joint is identical with the butt joint at one extreme of the scarf angle and approaches the lap joint at the other extreme of the scarf angle. Selection of joint type is influenced by the configuration of the parts as well as by stress requirements and other service requirements,
Fig. 3.1
such as electrical conductivity, pressure-tightness, and appearance. Also influential in selection of joint type are fabrication techniques, production quantities, method of feeding the filler metal, and other factors. Lap joints are generally preferred for brazing operations, particularly when it is important that the joints be at least as strong as the weaker member. For maximum strength, the lap-joint length should equal 3 times the thickness of the thinner member. Figure 3.1 shows good and bad butt-joint and lap-joint designs. Butt joints are usually used where strength requirements may not be critical or where the use of a lap joint would be objectionable. The lap joint shown in Fig. 3.1 is the easiest type of brazement that can be made. The reasonable amount of overlap for lap joints is 3 to 6 times the thickness (T) of the thinner of the two members being joined.
Temperature and Time The temperature of the filler metal naturally has an important effect on the wetting action, because the wetting and alloying action improves as the temperature increases. Of course, the temperature must be above the melting point of the filler metal and below the melting point of the parent metal. Within this range, a temperature generally is selected that gives the best filler-metal wetting and flow. Usually, the lowest satisfactory brazing temperatures are preferred to (a) economize on heat
Comparison of some good and bad ways to assemble lap and butt joints. BFM, brazing filler metal. T, thickness
Chapter 3: Elements of the Brazing Process / 21
energy required, (b) minimize the heat effect on the base metal (annealing, grain growth, or warpage, for example), (c) minimize basemetal/filler-metal interactions, and (d) increase the life of fixtures, jigs, or other tools. Higher brazing temperatures may be desirable to (a) use a higher-melting but more economical or otherwise superior filler metal; (b) combine annealing, stress relief, or heat treatment of the base metal with brazing; (c) permit subsequent processing at elevated temperatures; (d) promote base-metal interactions in order to modify the filler metal (this technique is usually used to increase the remelt temperature of the joint); (e) more effectively remove surface contaminants and oxides with vacuum brazing; and (f ) avoid stress cracking. The time at brazing temperature also affects the wetting action, particularly with respect to the distance the filler metal can creep. If the filler metal has a tendency to creep, the distance generally increases with time. The alloying action between the filler metal and parent metal is, of course, a function of both temperature and time. In general, for production work, both temperature and time are kept at a minimum consistent with good quality. In conclusion, the filler metal and brazing process must be selected with a true understanding of both the physical metallurgy of the base material and the interactions of the base material with the filler metal.
Processes (Heat Sources)
joining requires efficient transfer of heat from the heat source into the joint. The heat capacity and thermal conductivity of the assembly must be considered. A 0.025 mm (0.001 in.) diameter wire cannot be brazed to a large mass of copper with a small torch. The size of individual assemblies, the numbers required, and the rate of production necessary influence selection of the heating method. Many other factors must be considered before the choice is made. The rate of heating, thermal gradients, and cooling rates, both external and internal, vary tremendously with different methods of heating, and the effects of these on dimensional stability, distortion, and metallurgical structure must be considered. The joint can be heated in many ways, which are commonly categorized by the actual method of heating. There are six commonly used methods: • • • • • •
Torch brazing Furnace brazing Induction brazing Dip brazing Resistance brazing Infrared brazing Lesser known heating methods include:
• • • •
Laser brazing Exothermic brazing Weld brazing Microwave brazing
Torch Brazing
The heating methods available often place a constraint on the designer and engineer in selecting the best type of capillary joint. In principle, there are many methods of heating available for brazing (Table 3.3). Effective capillary
Manual torch brazing is the method most frequently used for repairs, one-of-a-kind brazing jobs, and short production runs as an alternative to fusion welding. Heating the assembly with a
Table 3.3 Characteristics of various heating methods Characteristics(a) Method
Torch (flame) Electrical resistance Induction Furnace (atmosphere) Furnace (vacuum) Dip (flux bath) Infrared (a) H, high; M, medium; L, low
Capital cost
Running cost
Basic output
Flux required
Versatility
Operator skill required
L/M M M/H M/H H L/M M
M/H M M M/H L M/H L
L M/H M/H H H L/M M
Yes Yes Y/N Y/N No Yes Y/N
H L M M M L L
Yes No No No No Yes No
22 / Brazing, Second Edition
gas (oxyfuel) flame, either a hand-held torch or machine-mounted burners, is the most common method for brazing. Automatic setups comprising several burners can produce several hundred assemblies per hour. The torch brazing technique is relatively simple and can be mastered by the mechanically adept in a short time. Those already experienced in torch welding and the brazing of other metals generally encounter no difficulty learning torch brazing. Flux is normally required with the process. An exception is the joining of pure copper base metal with phosphorus-bearing filler metal. The phosphorus joins with the oxides on the surface to promote wetting of the filler metal. Torch brazing is widely used because of its relatively low cost and portability. The flame is generated by the combustion of a combination of oxygen and a fuel gas. The same equipment can be used as with oxyfuel gas welding. The process lends itself to the use of low-melting filler metals, which have excellent flow characteristics. Torches. Torch brazing employs the same type of torch, controls, and gases used for gas cutting and welding. To braze, the operator uses a brazing torch tip. In the oxyfuel torch, fuel gas and oxygen flow through separate tubes and combine in a mixing chamber, then travel through the torch tip before igniting. The outer barrel of the handle is often knurled or embossed to give a good grip. The typical torch is 305 to 600 mm long (12 to 24 in.) and 1.38 to 2.76 kg (3 to 6 lb) in weight. Handles are of brass or aluminum and stainless steel; tips are of copper alloys. Adjusting the gas flow alters the blend of gases to give the desired flame. Equal-pressure mixers, used with fuel gas above 13.7 kPa (2 psi), receive fuel gas and oxygen at nearly equal pressures to produce a neutral flame. The lowpressure or injector-type torch, for low fuel-gas pressures, receives oxygen at 69 to 207 kPa (10 to 30 psi); fuel gas aspirates into the mixing chamber. Injector torches come in versions for specific gases and pressure ranges. Torch Tips. Brazers fit torches with tips larger in diameter than those used for welding. Tip style varies with fuel gas—brazing with acetylene requires a tip with a squared-off orifice face; orifices of tips for natural gas and propane are recessed (Ref 4).
Fuel Gases. Depending on the temperature and heat required, all commercial gas mixtures can be used to fuel the torch: oxyacetylene; oxyhydrogen; oxynatural gas; acetylene and air; hydrogen and air; propane, methane, and natural gas and air; and certain proprietary mixes. Flame temperature and cost are major factors in selecting a fuel gas. Oxyacetylene and oxynatural gases are the mixtures most often used commercially and are preferred in that order. The adjustment of the flame is very important. Generally, a slightly reducing flame is desirable. The oxyacetylene combination produces the highest temperature. The other gases are cooler, and their flames are less concentrated. Thus, they are easier to use and advantageous on lightgage material. Flame Characteristics. As the ratio of oxygen to fuel gas changes, the nature of the flame changes. The most common flame conditions
Fig. 3.2
Conditions for an oxyacetylene flame
Chapter 3: Elements of the Brazing Process / 23
used are reducing and neutral. Oxidizing flames are not recommended, because they oxidize the filler and base metals, and such oxidation impedes wetting and flow (Fig. 3.2). Carburizing Flame. When the ratio of oxygen to fuel gas is small, the flame is carburizing, and soot is produced. As oxygen is added to the fuel gas, the flame becomes luminous, and sooting disappears. As the oxygen content is increased, the luminous part of the flame becomes smaller and is centered near the torch tip. A blue zone, consisting of an excess of fuel gas, forms around the outside edge of the flame. This flame can be used for brazing. Reducing Flame. As the oxygen content is increased further, the luminous area becomes smaller and consists of an inner cone with a feathery trail extended out toward the flame end. This condition indicates a slight excess of fuel gas and is an excellent flame for brazing. Neutral Flame. When the oxygen addition reaches the ratio necessary for the fuel gas to be completely combusted, the feather that extended out from the bright inner cone disappears. This flame is used in brazing when an excess of carbon in the reducing flame is detrimental to the base metals or when maximum flame temperature is required. Oxidizing Flame. When the oxygen-to-fuelgas ratio exceeds that needed for complete combustion, the flame becomes oxidizing. The flame produces a hissing sound. The inner cone in the flame appears to be constricted. As stated previously, oxidizing flames are not recommended for brazing. Fuel Flame. When hydrocarbon fuel gases are burned without oxygen or are added through the torch, they typically produce a yellowish flame. Soot particles are usually present, because the oxygen in the air is not sufficient to support complete combustion. This flame is not useful in brazing. Apparatus. The torch-brazing operator also needs gas cylinders and the appropriate accessories, including regulators, valves, hoses, check valves, and flashback arrestors. For manual brazing, some operators prefer oxygen combustion for its high heating rate. Compressed air can be used, although flame temperature and heating rate are lower. In manual torch brazing, the filler metal is usually face fed in the form of wire or rod, or preplaced. In the latter case, care must be exercised in the placement of the filler metal and the
guidance of the torch to preclude premature melting of the filler metal. One way to prevent overheating is to use flux with a melting temperature not too far below that of the filler metal. The proper brazing temperature is indicated when the flux becomes liquid. To ensure uniform heating throughout the joint, which is very important, it may be advisable to use a multipletip torch or more than one torch. Mechanization. In automating the torch brazing process, fabricators use devices that move the workpiece or torches. Typically, the workpiece indexes—by conveyor or turntable—through one or more torch stations. At any station, brazing may be manual or automatic. Stations can be fitted with automated feeders of flux and filler metal or can position filler-metal rings at the joint. Single or multiple burners can be mounted at each station. Other stations can house cooling devices, preheating torches, and loading and unloading devices. Figure 3.3 shows a specially designed tenstation rotary brazing system that has enabled a major manufacturer of air-conditioning components to automate production of copper manifold assemblies. One operator loads the copper assemblies; all other functions are automatic. The new system completes 250 parts/h and features adjustable fixtures that allow the machine to accommodate 22 different part cycles. The key to the versatility of the system is the use of a paste filler as the brazing medium. Finely atomized filler metal, proper flux, and a creamy neutral binder are blended into one homogeneous mixture. After paste application, the assemblies move through three natural-gas-fired heating stations, where the copper-phosphorus alloy liquefies at 795 °C (1465 °F). The brazed parts are then gradually cooled via an air quench and a water quench before being manually unloaded by the operator.
Furnace Brazing The popularity of furnace brazing derives from the comparatively low cost of equipment, the adaptability of the furnace, and the minimal jigging required. With many brazing assemblies, the weight of the parts alone is sufficient to hold them together. With other configurations, one or two rectangular blocks of metal are all the fixturing needed.
24 / Brazing, Second Edition
Furnace brazing is a medium- to high-volume production process for self-fixtured assemblies with preplaced filler metal. Operators lay filler metal as a thin solid sheet or as a clad layer into the joint before inserting the assembly into the furnace. The furnace is then purged with a gaseous atmosphere or evacuated of air and heated to a temperature above the liquidus of the filler metal but less than the melting point of the base metals. The brazements are then cooled or quenched by appropriate methods to minimize distortion and produce the required properties in the filler and base materials. This cycle is designed to produce the required melting and solidification of the filler metal to join the components without melting or impairing the properties or shape of the base metals. The process joins parts that would otherwise distort from the localized heating of torch brazing. Compared to other brazing processes, capital equipment cost can be high, but it may be easily recovered by the high throughput possi-
Fig. 3.3
ble with batch processing in furnaces. Jointdesign skill is critical for furnace brazing, because brazing occurs without observation or intervention by an operator. There are two basic types of furnaces used for brazing: batch furnaces that process mediumproduction work lots and continuous conveyorfed furnaces that turn out the highest-volume production. Heating is typically by resistance, but furnaces that combine radiation and naturalgas convection heating can decrease heating time by 50%, resulting in a more uniform heating of the part, and can reduce energy costs. The parts should be self-jigging or fixtured and assembled, with filler metal preplaced near or in the joint. The preplaced filler metal may be in the form of wire, foil, powder, paste, slugs, or preformed shapes. Fluxing is used except when a reducing atmosphere, such as hydrogen, and either exothermic or endothermic combusted gas can be introduced into the furnace. In some instances, both flux and a reducing atmosphere
Automated brazing system using a special paste alloy to braze up to 250 parts/h. Source: Ref 5
Chapter 3: Elements of the Brazing Process / 25
may be necessary. Pure, dry, inert gases, such as helium and argon, are used to obtain special atmospheric properties. For brazing in a batch furnace, the operator loads all of the assemblies at one time into the furnace. Loading is from the top, side, or bottom of the furnace. In loading from the bottom, the furnace is raised above a pit-type work area; after loading, the furnace is lowered over the work area. With cycle times as short as 15 min, batch furnaces can produce several hundred brazements a day. The simplest type of batch furnace is the muffle type. These supply heat by gas combustion. For better quality, brazers employ electricresistance or radiant-tube furnaces or furnaces that combine resistance and combustion heating. Box-type furnaces load and unload from the side (Fig. 3.4). These connect the heating chamber and a water-encased cooling station, each with its own source of controlled atmosphere. A refractory-lined baffle door separates the chambers. Work moves through the furnace and cooling station in trays or baskets. Most batch furnaces are lined with refractory material. With the development of nickel brazing came the need for a furnace that could withstand high heat—the retort-bell-type combustion furnace. For brazing in a purified-hydrogen atmosphere, these comprise an inner container of a heat-resistant alloy sealed from outside air and the products of combustion. They are twopiece containers—an operator loads the workpiece on a base, and a cover lowers over the base. Sand or an O-ring seals the base-cover joint. Dry hydrogen enters the top of the cover; the container is purged and then lowered into a pit-type
Fig. 3.4
Electrically heated batch box-type brazing furnace
furnace. Hydrogen flows through the container during preheating, brazing, and cooling. When continuous-type furnaces are used, several different temperature zones may be used to provide the proper preheating, brazing, and cooling temperatures. The speed through a conveyor-type furnace must be controlled to provide the appropriate time at the brazing temperature. It is also necessary for the assembly to be properly supported so that it does not move while traveling on the belt. This may require special fixtures, but most often, brazements are designed to be self-supporting. Compared to batch-furnace brazing, part quality may be higher and manual labor greatly reduced. Typical production volume is 100 to 1000 kg (220 to 2200 lb) of brazed assemblies per hour. There are various types of continuous furnaces. For example, a newly designed and built furnace is a hydrogen-atmosphere hump meshbelt electric furnace rated for temperatures up to 1177 ± 11 °C (2150 ± 20 °F). The furnace has a nitrogen curtain located on the furnace entrance ramp that is acceptable by the Occupational Safety and Health Administration (OSHA) code to isolate explosive gases. The furnace has two heating zones, the first with 18 electric heating elements and the second with 12 heating elements. The hot zone measures 2.86 m (9.4 ft) in length, with a 0.9 m (3.0 ft) slow-cooling zone before the exit. The furnace walls consist of an internal wall of ceramic blanket, an intermediate wall of ceramic-fiber board, and an outer steel shell. The innovative design of the hydrogenatmosphere hump mesh-belt furnace system has produced parts such as those shown in Fig. 3.5.
26 / Brazing, Second Edition
Innovative designs such as this continue to provide insight into the advanced technology being made available to the thermal processing industry and display the potential for development of advanced materials processing systems (Ref 6). A second type of semicontinuous controlledatmosphere brazing (CAB) furnace is shown in Fig. 3.6. This type of furnace was designed for brazing aluminum heat exchangers. Additionally, this type of furnace and control system was designed to accommodate the production demands in the small-to-intermediate-sized brazing shops. By monitoring the temperature differential through the load and adjusting the fan speed (i.e., heating rate) and cycle time accordingly, the semicontinuous furnace system has the ability to successfully braze parts of various dimensions or those loaded in different configurations under a single set of parameters. Part-to-part temperatures can be maintained at the recom-
Fig. 3.5
mended temperature range of 600 to 620 °C (1110 to 1150 °F) for 3 to 5 min, with a temperature of ±3 °C (±5 °F). Although CAB brazing of aluminum heat exchangers is a mature process, significant improvements in both product yield and quality can be achieved through process modifications that include both furnace design and proper materials selection. A prototype CAB braze furnace design is shown in Fig. 3.7 and is currently envisioned for use in a typical aluminum heat-exchanger manufacturing plant. The “intelligence” of the furnace centers around a fuzzy logic controller and a CAB thermal model that are based on a personal computer or Unix software architecture. To meet the future braze quality requirements of aluminum heat-exchanger manufacturing, it will be necessary for furnace designers and suppliers to consider making CAB braze furnaces that are smarter, more adaptive, and highly costeffective. This can be done by using logic con-
Examples of aerospace brazed manifolds and tube assemblies. The part on the right is approximately 61 cm (24 in.) in length.
Gas fired burners Fan assembly
Atmosphere curtains
Fan assembly
Airflow baffle
Muffle
Atmosphere curtains
Air blast curtains
Muffle tension cylinder
Load area/dryoff oven
Fig. 3.6
Entrance purge chamber
Convection chamber
Fan assemblies
Water jacketed cooling chamber
Schematic of the semicontinuous controlled-atmosphere brazing furnace system. Source: Ref 7
Air blast chamber
Wire brush assembly
Chapter 3: Elements of the Brazing Process / 27
trollers that can self-adjust furnace conditions for optimal efficiency, using image systems to increase potential problem awareness, using new materials that can provide durability and are cost-effective, and by coming up with new designs that reduce floor space requirements without reducing the quality of the process. It should be noted that several furnace manufacturers have installed twin silica (woven tape) curtains at the entrance and exit ends of brazing furnaces and have reduced natural-gas consumption by approximately one-third (Fig. 3.8) (Ref 8). The curtains serve both as a barrier to and combustion chamber for escaping gases. An additional benefit arises from the fact that the product line speed can be increased during continuous furnace operation. The tape curtains used perform at temperatures varying from 760 to 1100 °C (1400 to 2000 °F). In this particular application, they lasted for an average of one week—very good service, considering that quite a number of other materials evaluated over the years could not withstand the high temperatures involved. The natural-gas savings are very significant. Although the silica-based tape was found to be a little more expensive than some of the other filament tapes, its ability to withstand the high
temperature and the added service life, coupled with the reduced energy costs, made it very attractive. The effective cooling of components in continuous nitrogen-hydrogen atmosphere furnaces depends on the heat-transfer efficiency of the furnace atmosphere. The cooling of parts in a continuous furnace is controlled by the convective heat transfer, which is greatly influenced by the composition of the nitrogen-hydrogen atmosphere. The heating of parts is dominated by radiant heat transfer from the furnace walls and is not greatly influenced by the composition of the atmosphere. The cooling capacity or the productivity of a continuous furnace can be manipulated simply by changing the concentration of hydrogen in the atmosphere. This provides heat treaters added flexibility to optimize their atmosphere requirements, depending on the production demands. A model has been developed based on heattransfer principles to determine the cooling capability of a continuous furnace. The model indicates that the cooling capability can be increased by increasing hydrogen concentration in a nitrogen-hydrogen atmosphere or by using a high-convection cooling system. The selection of a higher hydrogen concentration or a
4. Infrared imaging cameras
5. Innovative muffle & element designs 6. Innovative belt & fixture designs
3. Pre-braze convection 2. Optical part sensor
1. Intelligent fuzzy logic
Fig. 3.7
Proposed next-generation furnace for controlled-atmospheric brazing. Source: Ref 8
28 / Brazing, Second Edition
high-convection cooling system would, however, depend on the desired increases in cooling capability and on the overall economics (Ref 9). A large volume of furnace brazing is performed in vacuum, which prevents oxidation and often eliminates the need for flux. Vacuum brazing has found wide application in the aerospace and nuclear fields, where reactive metals are joined or where entrapped fluxes would be intolerable. If the vacuum is maintained by continuous pumping, it removes volatile constituents liberated during brazing. There are
several base metals and filler metals that can be harmed by brazing in a vacuum, because their low-boiling-point or high-vapor-pressure constituents cause part of the metal to be lost. Vacuum is a relatively economical method of providing an accurately controlled brazing atmosphere and is an effective means of screening the work to be brazed from oxidizing gases and other impurities. The vacuum pressures used for brazing generally range from 0.13 to 1.3 Pa (2 to 20 × 10–4 psi). This range corresponds to a gas that is several hundred times purer than the purest gas used for atmosphere brazing. Vacuum brazing does not allow as wide a choice of filler metal as does atmosphere brazing. Vacuum furnaces are invariably heated by electricity in any one of a number of forms. Vacuum brazing takes place in either a singleor double-pumped retort or in a batch-type furnace. A single-pumped retort is loaded with the brazement, evacuated, and heated externally by resistance or by gas or oil combustion. The maximum furnace temperature is 1150 °C (2100 °F), although most operate at 870 °C (1600 °F) or lower. For higher-temperature brazing, manufacturers can employ a doublewall retort in which the workload sits in an inner high-vacuum (1.33 Pa, or 20 × 10–4 psi, or lower) container that sits inside a rough-vacuum
Fig. 3.8
Brazing furnace using silica-based woven tape curtains to reduce natural-gas consumption
Fig. 3.9
Diagram of cold-wall vacuum furnace showing the various components
Chapter 3: Elements of the Brazing Process / 29
(1.33 to 13.33 × 102 Pa, or 0.19 to 1.9 psi) chamber. Thermal insulation and the electrical heating elements fit between the walls of the outer and inner retorts. Batch-type vacuum furnaces can be a hot- or cold-wall type; the cold-wall furnaces are more popular. The furnace diagram in Fig. 3.9 illustrates the major components of a simple batch-type fluxless vacuum furnace for aluminum brazing. Figure 3.10 illustrates the microstructure of a fluxless brazed aluminum joint, while Fig. 3.11 shows a graphic sequence of events for a batchtype furnace. The cold-wall-type furnace in Fig. 3.9 consists of a horizontally charged, water-cooled
Fig. 3.10
1200
+3
650 Door opens Braze
760 mm atmospheric pressure
10 +2
Mg evaporation
Soak
1000
10 –2
Vacuum cool Convection cool
800
Temperature ˚F
Chamber vacuum torr
Rough vacuum
10 –0
10 –1
540
Mg sublimation
10 +1
10 –3
425
Temperatures: 600
Hot start cycle repeat
Program profile 100 micron staging pressure
315
Furnace actual Work surface Work internals
400
205 Vacuum break
10 –4 200
95
10 –5
High vacuum
Cold start 10 –6
0
0 5
10
15
Time, minutes
Fig. 3.11
Sequence of events during vacuum furnace brazing cycle
20
25
Temperature ˚C
10
Microstructure of furnace-brazed joint after fluxless aluminum brazing
vacuum furnace chamber with an internal resistance-heated hot zone, reflective metallic heat shields, an adjustable power supply, a single set of vacuum pumps, a water recirculating system, and a control panel. The operation can be initiated by automatic or manual sequencing. Depending on the geometry of the workpiece and the desired production rate, the hot-zone size can vary from less than 0.28 m3 (10 ft3) up to 17 m3 (600 ft3) or more. The hot zone can be rectangular, cylindrical, or flat in shape. There may be singular or multiple temperature-control zones within the hot zone. The designs to be considered depend on the workpiece shape, size, desired production rate, and user preference. Batch furnace cycle times can vary from 8 min to longer than a day. Small, lightweight parts that would fit into a 0.028 m3 (1.0 ft3) hot zone can be heated and brazed in approximately 8 to 20 min. Cycle times for most aerospace, automotive, and electronic components vary from 10 to 45 min. Large components, such as cryogenic heat-exchanger cores, have cycle times that vary from several hours up to 24 to 36 h. In the latter case, heat-exchanger cores measuring 1.2 × 1.2 × 6.1 m (3.9 × 3.9 × 20.0 ft) and weighing in excess of 9075 kg (20,000 lb) are routinely brazed in vacuum. Cold-wall-furnace configurations include horizontal designs with front loading, vertical designs with top or bottom loading, or clamshell
30 / Brazing, Second Edition
designs with front loading. Heated by radiation, temperatures peak at 2200 °C (4000 °C), and vacuums reach 0.00013 Pa (20 × 10–6 psi). In addition to batch-type furnaces, there are semicontinuous furnaces with entrance and exit vacuum locks in series with single or multiple heating stations. In all of these variations, the basic brazing process is essentially the same. The parts to be brazed are exposed to high vacuum, and heat is applied in a uniform and controlled manner (Ref 10). Semicontinuous furnaces have found application in high-volume production situations, such as the automotive industry, where this type of operation is customary. Figure 3.12 illustrates a system comprised of three chambers in series. Each chamber section is roughly equivalent in size to a single batchtype furnace chamber. The entrance and exit lock chambers are isolated from the heating chamber and from the atmosphere by large, automated gate-valve-type doors that are water cooled. Vacuum-sealed shafts are used for driving and positioning these doors each time a load of parts is charged. A workload may be present in all three of the chamber sections. A workload in the entrance lock could be undergoing evacuation and preheating while the workload in the furnace section is heated and brazed in a high-vacuum environment. Simultaneously, vacuum is broken in the exit lock, the workload is discharged, and the exit lock is closed, sealed, and then re-evacuated. Although the furnace chamber is exposed to intermediate vacuum levels and gaseous contamination from incoming work when the entrance-lock furnace door is opened and the work is indexed, the workload is not exposed to intermediate vacuum during the critical brazing portion of the cycle.
Plan view
Mechanical pumps (3) Booster pumps (2/3) Diffusion pumps (1/3)
Fan cool Vacuum tight doors (4) Exit vacuum lock (optional) Heat zone
Fig. 3.12
Power supply Entrance vacuum lock with heating capability Control panel
Semicontinuous vacuum furnace with threechamber design
The appeal of this vacuum furnace arrangement stems, in part, from the fact that the furnace chamber is not exposed to atmospheric conditions, thereby eliminating moisture and gaseous adsorption The quality of the brazed work is comparable to that obtained in a conventional batch furnace that is properly designed for aluminum brazing under vacuum. In the brazing of automotive aluminum heat exchangers, the vacuum furnace is equipped with a sophisticated residual gas analyzer for monitoring key brazing factors, such as magnesium evolution and oxidizing gas gettering. There are a number of other reasons, based on process mechanisms, why brazing aluminum under vacuum is advantageous. Oxides form on the surface of aluminum when it is exposed to air, which restricts the flow of molten filler metal. When aluminum is heated, the differential expansion of nascent aluminum and its thin oxide causes cracks in the oxide layer. If oxygen is available, the aluminum oxide envelopes readily reform. In vacuum, however, magnesium (a getter) in the filler metal vaporizes as it reaches the liquidus temperature, and the magnesium vapors react with any traces of residual oxygen and that absorbed in components within the chamber, in accord with these most prevalent reactions: Mg + H2O 3 MgO + H2 2Mg + O2 3 2 MgO Mg + CO2 3 MgO + CO
Good operating practice allows the maintaining of broken oxide film, with resultant wetting by the fluid aluminum filler, to form fillets. The remaining platelets, on breaking of the oxide film, become minute oxide inclusions without any significant deleterious effect. Heating Elements and Radiation Shields. Elements must withstand low vapor pressures and high temperatures and must have a large heating surface for good heat transfer. The most commonly used materials for elements, and their maximum operating temperatures, are molybdenum, 1900 °C (3450 °F), or 1590 °C (2900 °F) in high vacuum; tungsten, 2480 °C (4500 °F); and tantalum, 2205 °C (4000 °F). Radiation shields (Fig. 3.13) must perform the dual function of containing the heat and protecting the rubber vacuum seals that are incorporated in the door of the furnace and are fitted around all power leads or other controls that pass
Chapter 3: Elements of the Brazing Process / 31
through the furnace wall. A less obvious consequence of effective insulation is that the cooling rate is severely retarded. Because of this, the overall time of a vacuum brazing operation is usually much longer than that of other batch brazing techniques; this results in expensive equipment being tied up for long periods in processing comparatively small workloads. It should be remembered, however, that most of the metals on which vacuum brazing excels are also costly, and that any process that permits material economies through fabrication and ensures the necessary joint properties is justified. Equipment developments that have reduced processing times include transfer mechanisms, multizone furnaces, and inert-gas quenching. Controls. Brazing furnaces must be fitted with automatic temperature-controlling-and-recording devices. Control cabinets house vacuum gages and temperature instruments and controllers. Microprocessor-based units monitor and control temperature and holding times for the complete brazing cycle through heating, brazing, cooling, and quenching. They control the atmosphere, signaling gas-flow timing, and pressures. The previously mentioned devices and systems are able to control these parameters: temperature, temperature ramp rate, vacuum leak check, vacuum level, cooling, cycle events, and soak times, along with the capability of logging
Fig. 3.13
Vacuum furnace hot zone showing radiation shields
the various heating cycles performed by the vacuum furnaces and logging of process alarms and cycle modifications that are new additions to brazing furnace operations. Every production vacuum furnace operation should have its own fully automatic processing programmer, complete with data-logging instrumentation; run at vacuum levels of 0.00013 Pa (20 × 10–6 psi) to temperatures of at least 1450 °C (2650 °F); quench up to 100 kPa (2 bars) gas pressure; and run in partial pressures of nitrogen, argon, hydrogen, and helium. In addition, all horizontal vacuum furnaces have rail-guided work loaders to maintain part integrity, complete with a full overhead crane system. Cooling Systems. Some furnace brazements, depending on the specific alloys involved, must cool rapidly to ensure good mechanical properties and corrosion resistance. To quench the workpiece after brazing, backfill inert gases flow into the furnace. The backfill gases used are argon, nitrogen (except if base materials contain titanium, beryllium, zirconium, or aluminum), helium, and hydrogen. There are two preferred methods that are used for back filling: • With an internal gas-quenching setup, inert gases flow into a plenum behind the insulated hot furnace zone. Baffles and nozzles direct
32 / Brazing, Second Edition
the gas toward the brazement, and hot gases recirculate through a heat exchanger that cools the gas and redirects it to the work. • In vacuum furnaces with turbocharged gas quenching, backfill gases flow into the furnace hot zone in an orbital pattern. An internal turbofan recirculates this gas around the work, and a heat exchanger keeps cool gas flowing. Atmospheres. Furnace brazing takes place in vacuum or in a controlled atmosphere of high-purity inert or reducing gas. Both reduce the amount of flux needed or, in some cases, eliminate the need for fluxes altogether, to prevent oxidation during heating. Accurate, uniform temperature control and consistent atmospheric protection during brazing and cooldown ensure sound joints. Vacuum brazing often is preferred for joining alloys that contain chromium, aluminum, or titanium and for dissimilar-metal combinations of titanium, zirconium, niobium, molybdenum, or tantalum. To avoid contamination, these alloys also can be brazed in inert-gas atmospheres, if the gases are of high purity. Selecting the best controlled atmosphere for furnace brazing depends on base and filler materials and the capability of the furnace to maintain atmosphere quality through the cycle. Controlled atmospheres, which prevent formation of oxides to ensure good filler-metal flow and wetting, can be active or reducing, inert or relatively inert. Compared to vacuum brazing, atmosphere brazing enables use of a wider selection of filler metals, because some fillers contain low-boiling-point or high-vapor-pressure metals that can evaporate during vacuum brazing. In controlled-atmosphere brazing, gases pumped into the furnace are generated from natural gas or gases based on dissociated ammonia, nitrogen, hydrogen, argon, or helium. The escalating cost of natural-gas derivatives has led to increased use of nitrogen or hydrogen-nitrogen blends. Gas selection depends on base-material properties. For example, some grades of copper that contain copper oxide should not be brazed in reducing hydrogen atmospheres, because steam can form as the brazing process reduces the dissolved copper oxide. The compositions of American Welding Society (AWS)-designated controlled atmospheres recommended for brazing cover a wide
range (Table 3.4). These data are not intended as a comprehensive tabulation of atmospheremetal combinations but rather as a general outline of some of the more widely used combinations. Hydrogen in atmospheres reduces most metal oxides at elevated temperatures but can cause hydrogen embrittlement in some materials. Carbon monoxide reduces oxides of iron, nickel, cobalt, and copper. It is toxic, however, and must be properly vented or burned off. Carbon dioxide, properly mixed with carbon monoxide, inhibits decarburization and maintains a stable atmosphere when brazing steels. When using CO2-CO atmospheres, furnaces must be airtight; otherwise, the CO2 content can rise. Nitrogen in the atmosphere displaces air from the furnace and acts as a carrier gas for other atmosphere gases. High concentrations of nitrogen in the atmosphere are avoided when filler metals susceptible to nitriding—chromium, molybdenum, titanium, and zirconium—are present.
Induction Brazing Resistance to the flow of electricity induced by coils placed around the workpiece provides the heat for the induction brazing process. Resistance heating melts preplaced filler metal. Power densities range from 0.5 to 1.5 kW/cm2. Some advantages of the process include selective heating of the workpiece for brazing only where needed, quick attainment of brazing temperature, and the use of a variety of fluxes. This process lends itself to high-volume production applications that can be controlled remotely. High-strength components can be induction brazed with little loss of strength because of the precise heating capabilities of the process. Heating is localized to the part surface or just below, which is an advantage when joining components where metallurgical changes cannot be tolerated and on parts that allow minimal or no distortion. Most induction brazing occurs in air. Inductor design limits the complexity of assemblies, often making furnace brazing a more viable alternative for complex assemblies of several joints at one time. The process requires close, accurate part fit-up. Initial equipment cost can be high. The process relies on high-frequency alternating currents flowing through the induction
Combusted fuel gas (decarburizing)
Combusted fuel gas, dried
Combusted fuel gas, dried (decarburizing) Dissociated ammonia
Cryogenic or purified N2 + H2
Cryogenic or purified N2 + H2 + CO Cryogenic or purified N2 Deoxygenated and dried hydrogen Heated volatile materials (inorganic vapors—zinc, cadmium, lithium, volatile fluorides Purified inert gas (e.g., helium, argon)
Purified inert gas + H2 (e.g., helium, argon) Vacuum above 266.6 Pa (2 torr)
Vacuum from 66.65 to 266.6 Pa (0.50 to 2 torr) Vacuum from 0.13 to 66.65 Pa (0.001 to 0.50 torr)
Vacuum of 0.13 Pa (0.001 torr) and lower
2
3
4
6A
6B
9A
10A
10C
...
...
...
...
...
...
...
–68 (–90) –59 (–74)
–29 (–20)
–68 (–90)
–54 (–65)
–40 (–40)
–40 (–40)
RT
RT
Maximum dew-point of incoming gas, °C(°F)
...
...
...
...
1–10
...
...
... 100
2–20
1–30
75
38–40
15–16
14–15
5–1
H2
...
...
...
...
...
...
...
100 ...
70–97
70–99
25
41–45
73–75
70–71
87
N2
...
...
...
...
...
...
...
... ...
1–10
...
...
17–19
10–11
9–10
5–1
CO
...
...
...
...
...
...
...
... ...
...
...
...
...
...
5–6
11–12
CO2
Composition of atmosphere, %
BNi, BAu, BAlSi, titanium alloys
BCu, BAg
BCu, BAg
BCuP, BAg
...
Same as 5
BAg
Same as 5 Same as 5
Same as 5
BAg(b), BCuP, RBCuZn(b), BCu, BNi Same as 5
Same as 2
Same as 2
BCu, BAg(b), RBCuZn, BCuP
BAg(b), BCuP, RBCuZn(b)
Filler metal
Heat- and corrosion-resisting steels, aluminum, titanium, zirconium, refactory metals
Carbon and low-alloy steels, copper
Low-carbon steels, copper
Copper
...
Same as 5 plus titanium, zirconium, hafnium
Same as 3 Same as 5 plus cobalt, chromium, tungsten alloys, and carbides(e) Brasses
Same as 4
Same as 3
Copper(c), brass(b), low-carbon steel, nickel, Monel, mediumcarbon steel(d) Same as 2 plus medium- and highcarbon steels, Monel, nickel alloys Same as 2 plus medium- and highcarbon steels Same as 1–4 plus alloys containing chromium(e)
Copper, brass
Base metals
...
...
...
...
Special purpose. Parts must be very clean and atmosphere must be pure ...
Special purpose. May be used in conjunction with 1–5 to avoid use of flux
... ...
...
...
...
Referred to commonly as exothermic-generated atmospheres Decarburizes. Referred to commonly as endothermicgenerated atmospheres Referred to commonly as endothermic-generated atmospheres Carburizes
Remarks
RT, room temperature (a) types 6, 7, and 9 include reduced pressures down to 266.6 Pa (2 torr). (b) Flux required in addition to atmosphere when alloys containing volatile components are used. (c) Copper ahould be fully deoxidized or oxygen-free. (d) Heating time should be minimized to avoid objectionable decarburization. (e) Flux must be used in addition to the atmosphere if appreciable quantities of aluminum, titanium, silicon, or beryllium are present.
10B
10
9
8
6C 7
5
Combusted fuel gas (low hydrogen)
Source(a)
1
Brazingatmosphere number
Table 3.4 Atmospheres for furnace brazing
Chapter 3: Elements of the Brazing Process / 33
34 / Brazing, Second Edition
coils to create an electromagnetic field around the workpiece. The opposing currents that are induced into the workpiece generate the heat for brazing. The filler metal can be preplaced and can be in the form of wire, strip, or powder. The preform should be in good contact with the joint. Joint clearances of 0.05 to 0.13 mm (0.002 to 0.005 in.) have proven to be well suited for production work, but clearances up to 0.20 mm (0.008 in.) have been acceptable. Joint design should take thermal expansion into consideration. Large shear areas in the brazed joint provide maximum strength. Types of Joints. The most common joints— as with other heating methods for brazing—are butt and lap. A lap joint would be used to join two tubes, particularly if the joint requires strength and pressure-tightness (such as in the case of pumping a liquid through the pipes). In this example, the lap joint is easier to apply and provides more surface area to bond the assembly. Irregular-shaped parts do not lend themselves to induction brazing very readily, due to the complexity of coil configuration and heating
Fig. 3.14
Induction brazing of a steel base to a cast iron nose
uniformity. Any new user would, therefore, do well to seek the advice of the equipment manufacturer. Regardless of the type of joint, a good repeatable fit is desired. In assembly, one must consider that assembly is done at room temperature while brazing is at a higher temperature and that the parts expand. In the brazing of a steel base to a cast iron nose, for example, the coefficients of expansion of the two materials (cast iron, 18.9 × 10–6/°C, or 10.5 × 10–6/°F; steel, 14.9 × 10–6/°C, or 8.3 × 10–6/°F) must be considered in assembly design and heating for brazing (Fig. 3.14, 3.15). At 760 °C (1400 °F), the change of the joint gap is 0.06 mm (0.002 in.), and the gap is reduced. Coils. Induction brazing requires a coil, induction generator, and fixturing. Coils are made of copper tubing 4.76 to 9.52 mm (0.19 to 0.37 in.) in diameter, round or flattened. Depending on joint configuration, coils may be of square- or rectangular-cross-section tubing and are typically custom designed for each application. Size, shape, contour, number of turns, and
Chapter 3: Elements of the Brazing Process / 35
turn spacing all affect the strength of the electromagnetic field and the heat pattern. Examples of various coil designs are shown in Fig. 3.16. Proper coil design is important to the effectiveness of the brazing process. The geometry of the joints and the conductivity of the base metal must be taken into consideration. The heating pattern should be such that the joint area is heated uniformly. The joint area should reach the brazing temperature first to allow proper flow of the filler metal.
Induction Generator. Generators are rated by output power and frequency. For example, units can supply 25 to 200 kW power at 10 Hz and 15 to 50 kW at 25 kHz. The power rating of the induction generator and its ability to control current in the inductor directly influence heating rate. A low-power generator decreases heating rate, which allows time for temperature in the brazing joint to equalize. Generators may be one of three types: motor generator, tube oscillator, or solid state. Solidstate types have all but replaced motor-generator machines. Each of these generators operates at different frequency ranges—solid state from a few to 100 kHz and tube oscillators at 150 to 450 kHz. While frequency is often not of great importance in brazing, because it affects only the rate of surface heating as opposed to depthof-joint heating, brazing thin sections with a high-frequency (450 kHz) tube generator will likely produce better results than with a low-frequency solid-state unit. Conversely, heating with a solid-state generator with frequencies in the 8 to 10 kHz range works best on heavy sections. For sizing the generator needed for a particular application, fabricators take into account required production rate, base materials and their ability to absorb power (Fig. 3.17), and radiation from the work. A large number of induction-brazing defects stem from improper selection of power level, resulting in excessive heating rate. Calculate power absorbed by the workpiece from this equation: P = WTC/0.95t
Fig. 3.15
Diagram of assembly of cast iron and carbon steel components for induction brazing
Tube
where P is absorbed power in kilowatts, W is kilograms of material heated by the induction coil and by thermal conduction away from the
Induction coil
Tubing
Split Induction coil
Flange Alloy ring
Alloy ring Tube Induction coil
Fig. 3.16
Plug
Induction coil designs for specific applications
Tubing
36 / Brazing, Second Edition
joint, T is temperature rise in °C; C is mean specific heat of the base material, and t is heating time in seconds required to meet production requirements. Because determining W can be difficult due to the temperature gradient that develops in the work, use the accompanying graph to size an induction generator. The graph plots absorbed kilowatt per pound (kilogram) of workpiece versus temperature. Knowing the brazing temperature and base material, find the corresponding kilowatt per hour, and multiply by desired production rate, in pounds per hour (kilogram per hour), to determine absorbed power. Fixturing for induction brazing can be simple, especially if the assembly is self-jigging. If not, tack welding often suffices. If fixtures are necessary, construct any portion of the fixture within approximately 50 mm (2 in.) of the inductor coil of heat-resistant nonmetallic materials, such as ceramics or quartz. Any metallic materials used should be nonmagnetic, such as austenitic stainless steel or aluminum. Good fixturing contributes to repeatability of brazed parts. Flux and Atmospheres. Flux is used to clean the part while heating and to reduce the formation of oxides during heating. It is important to first clean the parts and flux them immediately prior to brazing. If the parts are fluxed and the flux is allowed to dry, it might
Brazing temperature, °C 540
760
1205
980
0.11
0.25
0.09
0.20
er
0.15
an
St
pp
Co
mi
0.06
Tit
nu m
ee l
0.07
ium
Ni
0.08
ck el
0.10
0.05
0.10
0.04 0.03
0.05
0.02
Absorbed energy, KW/kg of base material
315
Alu
Absorbed energy, kW/lb of base material
0 95
flake off. Critical to the success of the process is how the flux is applied and the amount that is applied (neither too little nor too much). Sometimes, it is possible to eliminate flux by using a controlled atmosphere (hydrogen-nitrogen-argon), which prevents oxidation. In brazing copper (and copper alloys), use can be made of silver-copper-phosphorus alloys. These alloys do not require flux; the phosphorus acts as a flux.
Controlled-Atmosphere Brazing Controlled-atmosphere induction brazing combines the unique benefits of induction heating with the advantages of brazing in a protected atmosphere. Induction heating provides: • Highly localized heating. Heating of the entire part is not required, as in furnace brazing. • Rapid heating • Precise control for repetitive results • Brazing of sequential joints in a given assembly as a result of localized heating With the addition of a controlled atmosphere, the following benefits are provided: • Avoid oxidation and/or volatilization of hightemperature metals and refractory materials • Joints produced are dependable, corrosionand oxide-free, and strong • Production of complex electronic devices requiring sequential brazing • Useful in critical structural assemblies with inaccessible areas • Leaves smooth fillets • Highly suitable for applications where brazed joints are hard to clean • Eliminates use of flux where residual flux can be particularly corrosive, as in complex vacuum tubes and joints of aircraft components. Also eliminates cost of fluxing and subsequent cleaning from flux
0.01 0
0 200
600
1000
1400
1800 2200
0
Brazing temperature, °F
Fig. 3.17
Graph that can be used to size induction generators
Reference 11 contains numerous applications where, on a production basis, the benefits of induction heating have been combined with the advantage of a controlled atmosphere. The examples cited include many copper alloys as
Chapter 3: Elements of the Brazing Process / 37
well as steel and stainless steel assemblies. The applications include a megawatt magnetron used in airborne radar systems, klystron tube components, and shielded grid triodes in specialized vacuum tubes.
New Induction Brazing Systems Major advances have been made in recent years in generator and temperature-control design. Solid-state generators have replaced motor generators and vacuum tube generators operating with intermediate power levels at frequencies from 50 to 350 kHz. Improvements have been made in generator control design, allowing precise and instantaneous responses to changes from sensors. The proper-sized generator required for a given operation is best determined by laboratory trial. However, a preliminary estimate can be made by considering the power absorbed by the pieces to be joined, the amount of radiation lost from the heated piece, and the required production rate. The estimate accuracy depends on good coil design and the electrical match between the generator and the workload. The most common frequencies for future use are 10, 25 to 50, and 50 to 350 kHz with solidstate generators, and 250 to 450 kHz with tube equipment. The lower frequencies heat more deeply with more uniform heating of matching surfaces if one of the components has a large mass. Higher operating frequencies, such as 450 kHz, may be suitable for brazing nonferrous metals or for brazing steel to a nonferrous metal or a nonmagnetic (austenitic) steel. Under certain circumstances, higher frequencies provide more efficient heating, and tube equipment operating at 2.5 to 8 MHz may be necessary for very thin components. The availability of a “sandwich” braze using a clad brazing strip (i.e., copper clad with silver filler metal on both sides) can overcome a problem of thermal expansion and the stress effects that have occurred in the past with induction brazing applications. Stress concentration and high residual stresses in the joint become critical factors when the joint members are stronger than the filler metal or when differential contraction takes place with dissimilar materials. The difference in thermal expansion for tungsten carbide may be troublesome when the carbide is placed in tension, resulting in cracked carbide. In this
case, stresses are minimized due to plastic deformation in the low-yield-strength copper layer. The flexibility of induction brazing has been enhanced by a hand-held induction heating system. Because it is not permanently mounted, the portable, lightweight system enables the operator to use this technique on nonuniform structures. Furthermore, the capability of taking the coil to the work, rather than the work to the coil, extends the area of applicability of this heating method. Not only are more varied heating uses possible, but also the new degree of flexibility lends itself to computerized, automated systems, including cellular operations. The unit has a small operating head and control equipment. The head is 102 × 152 × 330 mm (4 × 6 × 13 in.) and weighs only 36 kg (79 lb). The system delivers up to 25 kW into a workpiece. Full power is obtainable up to 129 m (425 ft) from the solid-state power supply using dry, flexible cable. The system operator is completely safe, because arcing to the workpiece does not occur if the coil should come in contact with the work. Controls for the system are mounted in the lightweight housing and can be operated through a computer interface. Some typical data provide a means for statistical process control, recording time, frequency, power, volts, amps, status, and date. The system has been used successfully in both refrigeration and compressor manufacturing. Tube-in-Place Induction Brazing. The development of manned spacecraft, maneuverable deep-space probes, and supersonic and hypersonic aircraft has resulted in the need for highly reliable, leak-free, and lightweight fluidtransmission systems. The obvious answer for this need is a system without joints, constructed from thin-wall, high-strength tubing. The acceptance of permanent joining methods for aerospace fluid systems was slow. The first uses of brazed joints in tubing for aerospace applications were in the B-70 (Ref 12) and A-4 vehicles. The first totally brazed system put into service was the hypergolic fuel and oxidizer system in the Gemini spacecraft (Ref 13). Since 1964, there has been a general acceptance of brazed systems by the aerospace industry. Brazed systems have been employed in a series of spacecraft programs, from Apollo to Space Shuttle to Space Lab to International Space Station (Ref 14–16). The advantages of brazed systems over welding and threaded connectors for all the aforementioned vehicles include the following:
38 / Brazing, Second Edition
• Freedom from leaks • Compatibility with system fluids, operating environments, and production techniques • Repairability and replaceability of system components in the field (or on the launch pad, if necessary) • In-place assembly and joining to provide a basically stress-free system
Fig. 3.18
• Potential for designing small tools, using thinner-wall tubing and less drastic heataffected zones Joint Design. The joint design illustrated in Fig. 3.18 provides for capillary flow of the filler metal in two directions in the joint to produce a maximum surface area of sound bonding. This enhances strength and provides a reliable seal. The filler metal, which, in the majority of applications, has been 82Au-18Ni, melting at 950 °C (1740 °F), flows between the outside wall of the tubing and the inside wall of the fitting being brazed. The maximum temperature attained does not exceed this temperature by more than 110 °C (200 °F), remaining below 1060 °C (1940 °F). Brazing is performed by an automated process in an atmosphere of dry argon, and the completed braze is of high quality with a high degree of reliability. Materials. Table 3.5, which is based on many studies (Ref 13), shows which tubing materials
Joint design for tube-in-place induction brazing. Source: Ref 15, 17
Table 3.5 Recommended filler metals for brazing of fluid systems
347
321
AM355
17-4 PH
PH 15-4
AM350
21-6-9
CP titanium
Ti-6Al-4V
Ti-0.15Pd
CP niobium
6061 aluminum
304L corrosionresistant alloy 304 corrosionresistant alloy 347 corrosionresistant alloy 321 corrosionresistant alloy AM350 corrosionresistant alloy 17-7 PH corrosionresistant alloy PH 15-7 Mo corrosionresistant alloy 21-6-9 corrosionresistant alloy CP titanium Ti-6Al-4V Ti-3Al-2.5V Ti-0.15Pd CP niobium LB2-coated niobium CP molybdenum Silicide-coated Mo Beryllium
304
Tube material
304L
Fitting material
1–6
1,6
1,6
1,6
...
...
...
...
...
...
...
...
...
...
1,6
1,4–10
...
...
...
...
...
...
...
...
...
...
...
...
1,6
1,6
1,6
1,6
...
...
...
...
...
...
...
...
...
...
1,6
1,6
1,6
1,4–10
...
...
...
...
...
...
...
...
...
...
...
6
...
6
1,6
...
...
6
...
...
...
...
...
...
...
...
...
...
...
1,6
...
...
...
...
...
...
...
...
...
...
...
...
...
...
1,6
...
...
...
...
...
...
...
1
...
...
...
...
...
...
...
1
...
...
...
...
...
... ... ... ... ... ... ... ... ...
... ... ... ... ... ... ... ... ...
... ... ... ... ... ... ... ... ...
... ... ... ... 1,6 1,6 6 6 ...
... ... ... ... ... ... ... ... ...
... ... ... ... ... ... ... ... ...
... ... ... ... ... ... ... ... ...
... ... ... ... ... ... ... ... ...
... ... ... ... ... ... ... ... ...
11,12 ... ... ... ... ... ... ... ...
... 11–14 11 ... ... ... ... ... ...
... ... ... 11,12 ... ... ... ... ...
... ... ... ... 15 ... ... ... ...
... ... ... ... ... ... ... ... 16
Note: 1, 82Au-18Ni; 2, 72Au-22Ni-6Cr; 3, 60Au-20Ag-20Cu; 4, 99.5Ag-0.5Li; 5, 92Ag-7.8Cu-0.2Li; 6, 72Ag-27.8Cu-0.2Li; 7, 92.6Ni-7.0Cr-2.9B-0.6C (max); 8, 85.1 Ni-7.0Cr-5.0Si-2.9B; 9, 79Ni-11.5Cr-3.5Si-3.0B-3.5Fe-0.15C (max); 10, 72.2Ni-20Co-4.5Si-3.3B; 11, 95Ag-5Al; 12, 81Pd-14.4 Ag-4.6Si; 13, Ti-15Cu-15Ni; 14, Ti49Zr-4Be; 15, Beta titanium; 16, Al-12Si
Chapter 3: Elements of the Brazing Process / 39
are the most compatible with the various corrosive hypergolic propellants. The materials 304L and AM350 have been used for most applications (Ref 12, 14, 15). The majority of the various types of fittings, including unions, elbows, tee fittings, reducer fittings, crosses, and valves, have been made of 304L corrosion-resistant steel. Additional studies (Ref 13, 16), also represented in Table 3.5, evaluated filler metals and their compatibility with fuels and base metals. Tooling. In general, the tube-in-place system is a total tube-brazing system. It is restricted to induction heating methods, because these methods permit fluxless brazing and the capability of completing joints in areas of very limited access. There are two distinct types of tooling employed for fluxless brazing of tubular components: • Open-coil tooling, in which the inductor is an unshielded multiturn solenoid coil and the
Fig. 3.19
Open-coil setup and typical joints
atmosphere-control chamber is constructed separately from the coil (Fig. 3.19) • In-place tooling, in which the inductor comprises either a split solenoid coil or opposing pancake coils, which are integral with the atmosphere-control chamber Tooling for in-place brazing combines the elements of the open-coil method into an integral tool. This type of tooling is available in either a clamshell-type (Fig. 3.20) or a plierstype (Fig. 3.21) configuration. The coil itself is the most critical part of the tooling. Split-solenoid coils require power transfer through pin connections that present arcing problems and limit the tool to a method of cooling external to the coil. Shaped pancake coil inductors provide the capability for using the water-cooling system of the induction generator to cool the induction coils and permit the use of pliers-type tools. In general, pliers-type tools provide the smallest
40 / Brazing, Second Edition
envelopes and the least cumbersome cooling, power, and gas connections. Brazing Procedure. Brazing is accomplished in a very straightforward manner. The tubing
Fig. 3.20
Clamshell-type tooling for in-place brazing
Fig. 3.21
Pliers-type tooling and typical joints
details and fittings are positioned on the structure and clamped where necessary. Once the brazing sequence is established, the internal tubing details and fittings are positioned on the structure and clamped where necessary. Once the brazing sequence is established, the internal argon-gas protection is provided, and the fitting is carefully positioned. The pliers-type tool is then positioned around the joint, and the brazing cycle is initiated. The automatically controlled brazing cycle consists of an argon-preflow period of at least 10 s, a heating period of from 10 to 30 s, and a postheating argon-flow period of from 60 to 90 s. The brazed joint is visually inspected for proper filler-metal flow and external filleting. If it is acceptable, the next joint is brazed. If the joint is unacceptable, the operator can try to promote the required filler-metal flow by reheating the joint with an identical brazing cycle. Only two reheats are permitted, because
Chapter 3: Elements of the Brazing Process / 41
there is a 10% tolerance in the brazing cycle settings, and because, if filler-metal flow is not promoted by proper use of the tooling, the chances of promoting flow are negligible. Excessive reheating when reactive filler metals are used also increases the chance for perforation of the tube wall through base-metal erosion. Nondestructive Testing. Each brazed joint is inspected visually to determine whether it is within dimensional tolerances and whether the filler metal is visible in the fillet around the entire circumference of the tube at the fitting end. Any melting of the base metal is cause for rejection. One-hundred percent radiographic inspection is performed, and all joints are required to meet rigid specification acceptance standards. This requirement is imposed on all procedure-qualification test specimens and also on all production brazes made for installation, repair, or modification. A new-generation induction brazing machine that features automatic temperature control has been developed. This brazing tool head contains a fiber-optics sensing device that functions in combination with a photoelectric cell, which feeds a signal through an amplifier to a controller recorder for automatic temperature control of the induction brazing cycle. Such developments, as a result of the use of solid-state electronics in the electrical circuitry, will continue to improve, and eventually lead to automation of, these brazing techniques.
Resistance Brazing Resistance brazing is most applicable to relatively simple joints in metals that have high electrical conductivity. For resistance brazing, the workpiece, with filler metal preplaced, is part of an electric circuit. Brazing heat comes from either placing carbon electrodes in contact with the brazement to conduct heat into the work or by relying on resistance of the brazement to generate heat. Electrode resistance brazing is best suited to joining of high-conductivity materials, such as copper-base alloys. Applied to steels, which are poor conductors, electrodes can cause local overheating. Using this process, often only one part of an assembly is heated by the electrode, the other by conduction. This procedure is most effective when brazing dissimilar materials. During direct resistance heating of the work, current flows for such a short time and heating
is so localized that the temperature of the workpieces does not change except in the local joint area. Advantages of resistance brazing are that it localizes heat and is flameless, non-contaminating, fast, and easily controlled. Disadvantages are that at least one workpiece, preferably all, must conduct electricity; large work, anything weighing over 2.3 kg (5 lb) or cross sections larger than 33 cm2 (5 in.2) may require so much current as to make the process too slow to be practical; and the process does not easily heat joints of nonuniform cross section. In the usual application of resistance brazing, the heating current, which is normally alternating current, is passed through the joint itself. Equipment is the same as that used for resistance welding, and the pressure needed for establishing electrical contact across the joint is ordinarily applied through the electrodes. The electrode pressure also is the usual means for providing the tight fit needed for capillary behavior in the joint. The component parts are generally held between copper or carbon-graphite electrodes. The heat for resistance brazing can be generated mainly in the workpieces themselves, in the electrodes, or in both, depending on the electrical resistivity and dimensions. The flux used must be conductive. Normally, fluxes are insulators when cool and dry, but they may become conductive from the heat of brazing. The process is generally used for low-volume production in joining electrical contacts and related electrical elements. Resistance brazing requires a transformer, electrodes, and fixtures. Standard resistancewelding transformers suffice for brazing, with the required capacity depending on joint size, heating capacity of the base material, and desired brazing time. In high-volume production applications, spot-welding equipment may suffice to resistance braze, using a transformer, upper and lower electrodes, and process controls. For manual brazing, operators wield handheld tongs that clamp the workpiece and hold the electrodes. Transformer power is typically rated at 10 to 25 kVA. Electrodes are of standard resistance-welding-grade copper alloys or of carbon. Most common are Resistance Welding Manufacturers Association (RWMA) class 2, chromium copper, and class 14, refractory (molybdenum) metal electrodes. With higher electrical resistivity, they generate more heat than do RWMA electrodes and
42 / Brazing, Second Edition
perform well for brazing highly conductive materials. Electrodes of high electrical conductivity braze low-conductivity materials, such as steels. Filler metals used for resistance brazing need to melt at relatively low temperatures to avoid oxidation. Most common are BAg-1, 1A, 2, 7, 8, and 18 and BCuP-1, 2, and 5. Preplace a strip or shim on large flat joints, a ring on cylindrical joints, paste or powder on irregular surfaces, and a rod for feeding additional filler to supplement preplaced filler in large joint gaps (Table 3.6). In resistance brazing, more so than in other brazing processes, attention is given to selecting from the metallurgically compatible filler metals the one having the lowest brazing temperature, because in resistance brazing, it is necessary to keep the maximum local temperature reached by the work as low as possible while providing uniform heating of the abutting joint surfaces and the filler metal. Fluidity of the filler metal is not critical in most resistance brazing, because the filler metal is usually preplaced, and the bond area is relatively large.
Carbon Resistance Brazing This method is generally used for joining of heavier sections. When this technique is applied to metals of low conductivity, such as steels, local overheating is likely to occur at the points of contact. The extent of overheating can be minimized by shaping the electrodes to the contour of the work, by heating at a lower rate, or by using only lower-melting-point filler metals. Carbon resistance heating of copper-base alloys is much easier. The choice of filler metals is no longer restricted to silver-base filler metals, and filler metals that contain phosphorus and no flux can also be employed. When a flux is used, care must be taken to prevent the carbon from
Table 3.6 General types of filler metals usually selected for resistance brazing of various classes of work metals Work metal
Steels, stainless steels, heat-resisting alloys, copper, copper alloys, nickel alloys Aluminum alloys Copper and copper alloys
Brazing filler metal
Silver filler metals (BAg type) Al-Si filler metals Cu-P filler metals
Note: These types of filler metals all have relatively low brazing temperatures.
becoming impregnated and to avoid the formation of an insulating film between the electrodes and the work. It is not always necessary for the current to pass through both the component parts. Frequently, only one part is resistance heated, and the second part is heated by conduction. This approach is particularly effective when either dissimilar metals or components of different masses are being joined. The equipment used for carbon resistance brazing may take a variety of forms, depending on the application involved. Hand-operated portable tongs are favored for many applications, particularly for joining heavy assemblies or when a number of joints, at different angles and locations, are required on one main unit. Heavy electrical-control-and-generating equipment provides many such applications. The fluxless method is used extensively for making connections of terminals or assemblies to stranded or braided copper electrical conductors. Carbon electrodes used in carbon resistance brazing are of two general types: carbon graphite and electrographite (artificial graphite). These electrode materials are made by simultaneously heating and blending the finely divided raw materials with coal tar pitch, which serves as a binder. Carbon or graphite blocks or rods are most frequently used, because filler metals do not wet them as they do metals. Various grades of these materials are available with ranges of resistance and hardness that permit them to be adapted to various jobs. The carbon or graphite is easily worked to the desired shape. These parts, however, must be considered expendable, because they waste away quite rapidly. Carbon resistance brazing is a far more attractive process when applied to copper than when applied to steel.
Direct Resistance Brazing Direct resistance heating is based on the same principles as resistance spot and projection welding, and the same machines can be used for both processes. The current must always pass through both of the parts that are being joined, and the highest temperature occurs at the interface between the parts. In most instances, the current flow is of such a short duration, and
Chapter 3: Elements of the Brazing Process / 43
heating is so localized, that no detectable increase in temperature takes place either in the electrodes or in the mass of the component. It is not unusual for the operator to hold the parts while the joint is being formed. When two metals of different resistivities are being joined, one part heats more rapidly than the other, unless precautions are taken. A simple cure is to introduce the current in short bursts and to allow time between the pulses for the temperature to balance out. Alternatively, using electrodes of different compositions can counteract unbalanced heating. By having the low-resistance electrode against the high-resistance member, and vice versa, the heating potential on both sides of the component interface is balanced. Direct resistance heating differs from carbon resistance heating, and from all other methods of brazing, in certain important respects. As already mentioned, severe temperature gradients are promoted within the workpiece, and only small localized areas at the joint interface attain the required brazing temperature. The area of joining is, therefore, small and has little mechanical strength. Attempting to increase the area is unlikely to be successful, even if sufficient power is available. Varying current densities over the larger area produces inconsistent results and severe local overheating. Also, because the highest temperatures occur at the interface, the filler metal must be preplaced between the mating faces of the component parts. Filler metals in the form of shims usually are the logical first choice for use with direct resistance heating. Electrodes for direct resistance brazing are made of high-resistance electrical conductors, such as chromium copper, tungsten, or molybdenum rods or inserts, or even steel in some instances. Direct resistance brazing is extensively used in the manufacture of electrical and electronic devices, many of which are of an extremely delicate construction by comparison with the heavier electrical equipment for which carbon resistance heating is preferred. The termination of fine wire conductors is a typical example of the use of direct heating. Applications. The use of flux-free (inertgas) brazing has been introduced in new uses of direct eutectic resistance brazing. Examples include brazing steel ring-piece pipe joints using resistance or induction heating, brazing chromium-nickel alloy steel fuel lines using resistance heating, and brazing steel servosteering parts in automated production pro-
cesses. Advantages of flux-free brazing over traditional flame and protective-gas furnace processes include: • Lower energy costs • Brazing without using flux • Easy integration of heating equipment into automated production processes • Meeting tight tolerances in the brazing of bent pipes Figure 3.22 shows ring pieces made of Deutsche Industrie-Normen (DIN) 7642 are used as connection elements for pipes/tubes (e.g., oil and fuel lines) in automobiles and mechanical equipment. The ring piece typically is joined using high-temperature brazing in a protective-gas, continuous-type furnace. This requires fixing the joint position via spot welding and fluxing before brazing. Disadvantages of the process include total or partial loss of brazing paste and the inability to be sure whether enough brazing paste is available to completely fill the brazing gap. Thus, when brazed parts must meet higher safety standards, expensive filler-metal rings often are substituted for brazing paste. Protective-gas, local-heating methods are more effective when brazing fewer pieces and
Fig. 3.22
Brazing ring-piece pipe joints using direct electric-resistance heating
44 / Brazing, Second Edition
thin-walled parts. Thus, brazing of ring-piece pipe joints using resistance heating is more economical. The brazing device (Fig. 3.23) essentially consists of a protective-gas chamber, two part electrodes (one stationary and one mobile), and one in the height-adjustable pipe electrode. A height-adjustable device between the part electrodes picks up the ring piece. The ring piece is placed in the fixture together with a pipe previously fitted with an alloy ring. To start the automatic brazing process, a part electrode is driven against the ring piece, which closes the secondary circuit. At the proper temperature, the filler metal liquefies and flows through capillary action into the braze gap. One braze joint is produced every 41 s, and 28 s are required to bend the pipe and check the seal. After a 10 s hold under the protective gas, the chamber is opened, and the brazed piece is removed. A ring piece joined to a steel pipe using a copper-tin alloy ring and resistance heating is shown in Fig. 3.24. Figure 3.22 shows resistance brazing apparatus used to join ring-piece pipe joints for an 8 to 12 mm (0.3 to 0.5 in.) diameter pipe. Servosteering parts traditionally have been brazed in protective-gas, continuous-type furnaces with a low failure rate of brazed parts.
Fig. 3.23
Direct electric-resistance-heating brazing device
Failures that occur are generally due to a partial loss of filler metal while in the oven. Because the required annual quantity totals over 1 million pieces, the brazing production process must be automated. Mounting the filler metal automatically and monitoring at the brazing temperature can reduce the failure rate. Both local heating processes are suitable for this application. In automatic resistance-heating brazing, the pipes and the ring pieces are isolated, automatically mounted, and brazed under a technical nitrogen gas. The accuracy of applying brazing wire is automatically examined by video inspection. Short and long pieces are alternately brazed with ring pieces after they have been automatically upset and collected opposite the brazing station. An automatic wire-feed device feeds the filler metal. The wire feed is warmed and follows the natural action of the filler metal even before reaching the working temperature of the brazed joint. The alloy is fed out after reaching this temperature. As a result, the heated filler metal melts immediately on the second dispensing. This differs from traditional methods of applying cold wire, where the wire often is pushed past the brazing joint. Following brazing, assembled components are transported by conveyor belt to the testing
Fig. 3.24
Ring piece joined to a steel pipe using a coppertin alloy ring and resistance heating
Chapter 3: Elements of the Brazing Process / 45
station, where a camera system checks the concave fillet and heated pipe area for flaws. Heating of the pipe area and the quality of the concave fillet are automatically tested. Parts detected having defects in the concave fillet exceeding 0.12 mm (0.005 in.) are removed from the process using a special procedure. The brazed assembly is turned four times. Flaws are displayed as black picture points, and the sum of these is used as a threshold value. When the camera registers fewer than 60 black points (0.015 mm2, or 0.0002 in.2) in a region, the program approves the detail picture. If the program counts more than 50 black points, a failure report is given. A part that passes inspection continues to the bending station. Brazing in three brazing stations enables a joining time of 15 s. During this time, the brazed assembly is upset, brazed, tested, and bent. The brazing process in reconditioning shafts of mechanisms and machines is based on producing an additional repair component in the form of a strip; deposition on its surface of a pasty filler metal; joining the strip to the reconditioned surface of the component by step-seam heating with a high-intensity pulse current and a low voltage, with simultaneous application of pressure; and cooling the brazing zone (Ref 18–20). As a consequence of the resultant transition resistance (Fig. 3.25) of the strip/fillermetal/component type, during a current pulse, the main component and filler strip are locally heated, and the filler metal is melted. Under the effect of the applied compressive force, the molten filler metal wets the surfaces being joined and is compacted, and after completing heating, it solidifies into a weld and forms a permanent joint. At the moment of formation of bonds between solid and liquid metals, the atoms of the parent and filler metals in the filler metal (Fig. 3.26) are excited. Consequently, the reaction rate, in a general form, can be expressed by the equation (Ref 19, 20):
dx Q = (N0 – x) ν exp – dt kT
cal bonds, k is the Boltzmann constant, and T is absolute temperature. One of the elements of the technological process is the temperature conditions of brazing, which take into account the temperature
Fig. 3.25
Diagram of electric resistances. Rt Rt1, and Rt2 are a transition resistance between the , electrodes on the surface of the strip, strip filler metal, and the filler metal of the main component, respectively. RM1, RM2, and RM3 are the resistance of the metal of the strip, filler metal layer, and the parent metal, respectively. (1) Roller electrodes. (2) Steel strips. (3) Filler metal. (4) Reconditioned shaft. Source: Ref 20
where x is the number of atoms entering into a chemical bond, N0 is the number of contacting atoms on the surface of the parent and additional materials of the component/strips, ν is the frequency of inherent vibrations of the atoms, Q is the energy of activation of formation of chemi-
Fig. 3.26
Diagram of the structure of the brazed joint
46 / Brazing, Second Edition
conditions of operation of the components, the melting point of the main metals joint, and the brazing temperature of the joint. Analyzing these temperatures with requirements made for the strength-reserve factor, the selected brazing temperature range (and consequently, the filler metal) was selected at 850 to 1150 °C (1560 to 2100 °F). The nickel, copper-zinc, silver, and copper-nickel filler metals are used most extensively in this range. All these filler metals satisfy the main requirements on brazing steels; namely, they wet efficiently the brazed surfaces, have sufficient fluidity and fill brazing gaps, ensure high strength of the brazed joint, and have minimum deformation of the distance at working temperatures. The results show that the size of the heataffected zone (HAZ) does not exceed 0.5 mm (0.02 in.), and in depositing the second and subsequent layers, the HAZ spreads only to the thickness of the first layer and does not reach the parent metal. It is thus possible to increase greatly the endurance limit of the brazed joint in comparison with arc surfacing methods. The cooling rate greatly affects the structure and, consequently, the strength of the brazed joint. This is reflected mainly in the formation of the primary structure and the distribution of the components of the filler metal in the brazed joint. Therefore, the proposed technology of reconditioning dynamically loaded worn components by electric resistance brazing a multilayer coating is characterized by the hyperactivity and low energy requirement of the process, by formation of joints with a small HAZ, and by retaining the initial properties of the metal of the component with a high strength of the brazed joint (Ref 20).
Dip Brazing As the name implies, during dip brazing, the assembly is immersed in a heated bath of either molten metal or a flux bath of molten salt. There are two methods that are considered to be dip brazing methods. One employs a molten metal bath, the other, a chemical (flux) bath. The molten metal bath dip brazing technique, which is not widely used, could be described as the high-temperature equivalent of dip soldering. It is used for the manufacturing of electronics and similar very small components.
In both of these processes, the parts being joined are held together and immersed in a bath of molten bonding metal that flows into the joints when the parts reach a temperature approaching that of the bath. In the chemical bath process, molten flux is used rather than molten metal. A different salt (flux) is often required if either the parent metal or the filler metal is changed, which may not be the case with the molten metal bath. However, one filler metal, such as a silver-base alloy, may be satisfactory for joining many brasses and steels. Molten Metal Bath Dip Brazing. The molten metal bath process is simple, cheap, and effective but is limited to brazing of small assemblies, such as wire connections or metal strips. A crucible, usually made of graphite or ceramics, is heated externally to the required temperature to maintain the filler metal in fluid form. A cover of flux may be maintained over the molten filler metal. The size of the molten bath (crucible) and the heating method must be such that immersion of parts to be brazed does not lower the temperature of the bath below that necessary for brazing. Parts to be brazed must be clean and are often protected with flux prior to their introduction into the bath. One potential future use for this process is in providing an alternative to soldering for electrical connections on motors and generators. The demand for increased output and/or reduced weight can now be met only by running the soldering machines hotter, and the softening temperature of the soldered joints is often the limiting factor. Another likely use of molten metal bath dip brazing is for pretinning of small inserts of difficult-to-braze metals that are to be subsequently brazed to heavier parts. Molten Chemical (Flux) Bath Dip Brazing. Dip brazing in molten salt (flux) involves the immersion of an assembly in a bath of molten salt, which provides the heat and may also supply the fluxing action for brazing. The bath temperature is maintained above the liquidus of the filler metal but below the melting range of the base metal. The advantages (Ref 21) of molten chemical bath (salt bath) dip brazing are: • Time for heating is approximately one-fourth of that required in a controlled-atmosphere furnace. • A protruding joint can be selectively brazed by partly immersing the assembly.
Chapter 3: Elements of the Brazing Process / 47
• A cocoon of frozen salt forms instantly around the assembly when it is immersed in the molten salt, which usually prevents premature melting of the filler metal by providing temporary insulation. • Brazing can usually be combined with carburizing or hardening, without the necessity for a separate reheating operation. • More than one assembly or joint in an assembly can be brazed at the same time, because production is limited only by the size and heating capacity of the furnace. • The workpiece is protected from scaling or decarburization by a thin film of salt that adheres to the surface of the assembly when it is removed from the salt bath. • Removal of the salt film is accomplished by dissolving during quenching or washing operations. • Because the density of the molten salt supports a considerable portion of the weight of the workpiece, the assembly weighs less when immersed, which can reduce the likelihood of distortion during heating. Limitations (Ref 21) of the process are: • The process is not generally used for intermittent operation, being better suited for work that requires daily production. • Joints that do not protrude from the assembly cannot be selectively brazed by partial immersion; most or all of the assembly must be heated to the brazing temperature in order for such joints to be brazed.
Fig. 3.27
Externally heated furnaces. Source: Ref 17
• The workpieces must be completely dry, because the molten salt reacts violently with moisture, splattering and possibly even exploding. If moisture is present, all work requires preheating. • The shape of the part must be designed to avoid trapping of air or salt and to drain completely after removal from the salt bath. • The assemblies should not require large, complicated fixtures. • Part cleaning may be difficult. • Salt residues, especially chloride, can corrode assemblies. • Proper maintenance of a salt bath furnace is difficult and has special problems, such as solidification during power outages. • Salt vapors may present health hazards unless properly ventilated. Equipment. Some salt bath furnaces are externally heated by gas, oil, or electrical resistance; this type of furnace lends itself more readily to intermittent operation and is not widely used for high-volume production. On the other hand, furnaces that are internally heated by immersed or submerged electrodes are not well suited to intermittent operation; therefore, they are used for high-volume production. Figures 3.27 and 3.28 illustrate the principal types of furnaces used for salt bath dip brazing. The externally heated furnaces shown in Fig. 3.27 are usually gas fired or oil fired and, less frequently, are heated by means of electricalresistance elements. When electrical-resistance heating is used, pot failure may result in total destruction of the heating elements. Figure 3.28
48 / Brazing, Second Edition
shows two types of internally heated furnaces that are energized with alternating current supplied by a transformer. The molten salt is an electrical conductor, and heat is generated within the salt between the electrodes by resistance to the passage of current. By closely spacing the electrodes, an electromagnetic stirring action of the salt is obtained that assists in maintaining temperature uniformity and controlling the temperature to within ±3 °C (±5 °F) (Ref 22). Theory and Start-Up. Figure 3.29 illustrates the theory of electrodynamic circulation. Maxwell’s well-known law of electromotive force states that any conductor that carries current, when placed in a magnetic field, tends to move at right angles to the direction of the magnetic field and at right angles to the direction of
Fig. 3.28
Internally heated furnaces. Source: Ref 17
the current. Applying this rule to the salt bath, two electrodes (A and B) are immersed vertically and close together in molten salt (S). A current of high magnitude (3000 A, for example) is passed between them. Each electrode is surrounded by approximately circular lines of force (D). High-density current flows between the electrodes at all points below the surface of the bath. At any given point between the immersed electrodes, such as point P, the salt element (represented by the heavy line) acts as a conductor of current between A and B. Because the electromagnetic field between the two electrodes extends outward from the plane of the drawing at point P, the liquid salt, which is acting as a conductor of current, is forced downward between the electrodes, as indicated by the
Chapter 3: Elements of the Brazing Process / 49
black arrows. If the direction of current is reversed, the direction of the magnetic field also reverses, and the direction of movement is still downward (Ref 17, 22). Once started, a molten flux bath is usually idled at a temperature high enough to maintain the salts in the liquid state. Electrodes. The electrodes in general use are made of carbon, wrought nickel, or Inconel 600. Such electrodes are less prone to attack than copper and copper-bearing electrodes and cause far less bath contamination. Carbon electrodes are the least costly. A carbon electrode lasts three to six months. Carbon has half the electrical conductivity of nickel, and therefore, for a given current density, carbon electrodes must have twice the cross section of nickel electrodes for equal capacity. Nickel electrodes are more costly than carbon electrodes. They may or may not be water cooled, depending on current density. Nickel electrodes wear most rapidly at their air-flux interface, and over-the-top electrodes generally need to be replaced in three to six months. They have the advantage of being easily removed. They are simply unbolted and lifted out of place with a forklift or other powerlift equipment. Over-the-top electrodes develop considerably more magnetohydrodynamic (MHD) force and therefore produce considerably more MHD flux circulation than sidewall- and bottom-positioned electrodes. This is due to the greater length of these electrodes and their attendant magnetic field in the flux.
Fig. 3.29
Schematic illustration of electrodynamic circulation. A, B, electrodes; S, molten salt; D, lines of force; P, point. Source: Ref 17
Submerged electrodes last much longer. Their average life is one to two years, with some sets lasting three years. Bottom-placed electrodes are more effective in producing convection currents in the molten flux and are less likely to be accidentally struck by dipped parts. Fluxes. The choice of a flux for dip brazing is not as much a matter of joint quality as of other important, although secondary, considerations. The aluminum dip brazing bath normally consists of a chloride mixture with small amounts of aluminum or sodium fluoride. The fluoride acts as a reducing agent, reacting with any oxides in the bath. This necessitates an exhaust system to remove volatile, corrosive fumes. Lithium lowers the melting point of flux and is an important additive to the salt bath—in the form of lithium chloride (LiCl)—for brazing of aluminum alloys. The fluxes containing LiCl are stable and require little chemical adjustment during operation. Salt bath dip brazing with silver-base filler metals is performed in a bath of neutral salts (Ref 17). Usually, a small amount of brazing flux is applied to the joint with the preplaced filler metal. After the moisture has been dried from the salt in a preheating cycle, the work is immersed into the molten salt for heating and brazing. On the other hand, aluminum is brazed in a molten salt bath consisting of active salts that serve as brazing flux and make it unnecessary to apply additional fluxes to the work before immersion. Process Operations. The salt bath process (Ref 17) is essentially the same regardless of whether aluminum, magnesium, titanium, steel, or nickel is being brazed. The process consists of six distinct steps: deburring and filing of machined or ground edges, cleaning, assembly, preheating, immersion, and flux removal. Deburring and Filing of Machined or Ground Edges. Because the brazed joint is made by virtue of filler metal flowing into the joint as a result of capillary action or gravity, burrs are removed, because they could restrict fillermetal flow and cause an imperfect union. Cleaning. The joint surface must be free of grease, oil, paint, oxide, and scale that would prevent the filler metal from wetting the workpiece surfaces. Assembly. During the assembly of parts, the filler metal, depending on its form, is applied to
50 / Brazing, Second Edition
the joints. The details can be held in position by tack welding, spring-loaded fixtures, spring clamps, staking, self-locking joints, or spot welding. (For more information on tooling, see Chapter 7, “Fixturing, Tooling, Stopoffs, Parting Agents, Surface Preparation, Surface Cleaning, and Repair.”) Preheating of an assembly before brazing serves several purposes. If prefluxing is used, preheating dries the flux and vaporizes all moisture from the assembly and the fixture. (Even a slight amount of moisture can cause spattering in contact with molten salt.) Preheating decreases the temperature drop of the salt bath, thus reducing brazing time, and also minimizes the premature melting of externally placed filler metal. For assemblies consisting of both heavy and light sections, preheating reduces thermal gradients and subsequent distortion and improves the wetting action on the heavier parts as well. To be effective, the preheating temperature must be at least 55 °C (100 °F) lower than the melting temperature of the filler metal. If oven preheating is used, oxidation must be avoided by using temperatures below 480 °C (900 °F); otherwise, an inert atmosphere is desirable. Finally, the preheating operation drives off all moisture from the assembly and fixture, thereby removing any possibility of any explosion, which might occur if water were introduced into the molten flux. Immersion. General ranges of brazing temperatures are used with various salts, and specific brazing ranges are employed for the individual filler metals. The time in the molten salt bath differs from one job to another. When only thin-section parts are to be brazed, the holding time may be as short as 1 min (Ref 17, 22). If assemblies are permitted to remain in the flux longer than 3 min, strength is substantially lowered, due to gradual transformation of the parent material from the as-wrought to the as-cast condition of the joint at the brazed interface (Ref 17). After the workpieces have been in the bath for the required time, they are carefully lifted from the salt bath. A uniform motion is necessary during removal from the bath; jerky movements can cause the liquid filler metal to be displaced from the joint. Flux Removal. A certain amount of flux adheres to the assembly after brazing, and this must be drained off while the parts are hot. The assembly is allowed to cool and then is
immersed in hot running water to remove the frozen salt cocoon. The assembly is normally removed from the fixture and subsequently dipped in a series of acid-to-water rinses to complete the flux-removal process.
Infrared (Quartz) Brazing Infrared brazing may be considered to be a form of furnace brazing in which heat is supplied by light radiation below the visible red rays in the spectrum. The principal heating is done by the invisible radiation from high-intensity quartz incandescent lamps. Heat sources (lamps) capable of delivering up to 5000 W of radiant energy are commercially available. The lamps do not necessarily need to follow the contour of the part in order to heat it, but the heat input varies inversely as the square of the distance from the source. Suitable reflectors, and sometimes parabolic focusing features, can present a unique source of energy for brazing. Because this process resembles furnace brazing, the brazing techniques are similar. The major difference is the source of heat for brazing. The quartz lamps supply brazing heat, in conjunction with electronic controls (including amplified feedback), to provide continuous, precise control of temperature and time to fractions of degrees and seconds. Infrared brazing setups are generally not as fast as induction brazing, but the equipment is less expensive. Honeycomb panels have been successfully brazed using opposing banks of lamps, and spot brazing of smaller parts has been accomplished using parabolic mirrors. An example of joining of Ti-6Al-4V by a rapid infrared processing technique has been investigated at temperatures between 1010 and 1250 °C (1850 and 2280 °F) and times of up to 120 s. With this infrared technique, joining is typically completed in seconds in an ambient atmospheric pressure of argon. Researchers (Ref 23) studied the effect of joining temperature and time on the microstructure and strength of the joints. The filler metal used in this study was a Ti-15Cu-15Ni wt% alloy. Results showed that the joint shear strength reached 554 MPa (80 ksi) when processed at approximately 1100 °C (2010 °F) for 60 s. Microstructural examinations of the joint with both an optical microscope and a scanning electron microscope indi-
Chapter 3: Elements of the Brazing Process / 51
2010
2190
500
75
400
60
300
45
200 900
(a)
1000
1100
1200
1300
Joint shear strength, ksi
Joint shear strength, MPa
1830 600
Joining temperature, °C Joining temperature, °F 2010
2190
500
75
400
60
300
45
200 900
(b)
1000
1100
1200
1300
Joint shear strength, ksi
1830 600
Joining temperature, °C Joining temperature, °F 1830
2010
2190
600
500
75
400
60
300
45
200 900
(c)
Fig. 3.30
1000
1100
1200
1300
Joint shear strength, ksi
The use of exothermic reactions for industrial brazing applications has been limited, because little information has been made available concerning exothermic brazing materials and the economics of their use. Exothermic brazing usually requires preplaced filler metal, as compared with other processing methods currently in practice that use controllable exothermic reactions. Exothermic brazing is a process in which the heat required for melting and flow of a commercial filler metal is generally produced by a solidstate exothermic chemical reaction. An exothermic chemical reaction is defined as any reaction between two or more reactants in which heat is given off due to the free energy of the system. Nature has provided us with countless numbers of these reactions, but only the solid-state or nearly solid-state metal/metal oxide reactions are suitable in exothermic brazing units (Ref 17). Exothermic brazing employs simplified tooling and equipment. The process uses the reaction heat in bringing an adjoining or a nearby metal interface to a temperature at which preplaced filler metal melts and wets the metal interface surfaces. The filler metal can be any commercially available material having suitable flow temperatures. The only limitations may be the thickness of the metal that must be heated through and the effects of this heat, or of any previous heat treatment, on the metal properties. The first step is to select a heat source and a filler metal. Then, the parts are cleaned chemically
Joining temperature, °F
Joint shear strength, MPa
Exothermic Brazing
(and mechanically, if necessary), the exo-reactant mixture suspended in alcohol is painted on the outside of the parts to be brazed, and the completed assembly with preplaced filler metal
Joint shear strength, MPa
cate that perfect wetting exists between the filler metal and Ti-6Al-4V for most joints. The brazeaffected zone increased with increasing joining temperature and time but did not show a direct influence on the joint strength (Fig. 3.30). Meanwhile, the Ti-6Al-4V base alloy exhibited no noticeable microstructural changes due to the rapid processing cycle of the infrared heating process. Based on this study, it is concluded that, compared with traditional joining methods, the rapid infrared joining technique has the following advantages: fast heating, little energy consumption, easy operation, no need for vacuum, little metallurgical modification to the base metal, and low cost.
Joining temperature, °C
Joint shear strength of infrared-bonded Ti-6Al-4V specimens as a function of temperature for: (a) 30 s, (b) 60 s, and (c) 120 s
52 / Brazing, Second Edition
is wrapped in a Fiberfrax (Unifrax Corporation) blanket and heat-lamp dried. The assembly is inserted into a furnace containing an inert atmosphere, and the furnace temperature is increased to the point at which ignition of the heat source can occur. For example, if the boron-vanadium pentoxide system were selected, the ignition temperature would be approximately 480 to 510 °C (900 to 950 °F). Finally, the external reaction products are washed away with hot water, and the assembly is complete and ready for use. Exothermic joining has proved applicable in numerous potential areas. It has been used to braze honeycomb sandwich panels, to braze inserts into panels locally, to attach small parts to larger structures, to make tube connections for aircraft hydraulic lines, and is considered a field-repair mechanism for aircraft structural honeycomb panels. Exothermic joining appears to be capable of fulfilling difficult requirements in producing reliable and reproducible joints, and its techniques are extremely economical and simple when compared with those of heat-source systems that depend on solar-energy conversion. These systems, except when the rays of the sun are focused directly on the work area, require the conversion of solar energy to electricity and the associated storage and conversion of the electrical energy to heat. The effectiveness of
Fig. 3.31
exothermic joining in vacuum has also been demonstrated (Ref 17).
Laser Brazing Laser brazing has been used in a limited number of applications. For example, a laser has been used where a very small, localized area of heat was required, such as for brazing of small carbide tips on printer heads for electronic printers. A finite-element model (FEM) was developed for the thermal analysis of a stud-to-plate laser brazing joint (Fig. 3.31, 3.32), and the transient temperature fields were analyzed by using a three-dimensional model (Ref 24, 25). Temperature-dependent thermal properties, effect of latent heat, and the convection and radiative heat losses were considered. The brazing parts used were American Iron and Steel Institute (AISI) 304 stainless steel stud and aluminum Al 5052 plate, and the filler metal 88Al12Si was used (Fig. 3.32). The brazing was done in open air with the aid of a powder flux, and neither a prebrazing nor a postbrazing heat cycle was attempted to control the thermal history of the joint. The studs to be joined function as guideposts or shafts for rollers, and the major quality factors are the perpendicularity of the
Experimental setup of stud-to-plate laser brazing system. VCR, videocamera recorder; IR, infrared; CNC, computer numerical control
Chapter 3: Elements of the Brazing Process / 53
stud with the plate surface, and the joint strength. The CO2 laser beam strikes the workpieces to heat locally in an area more than twice the diameter of the stud. The thermal diffusivity of aluminum is over 10 times higher than that of stainless steel, thus requiring more energy input to the aluminum plate. The beam impingement area was approximately 7.5 mm (0.3 in.) in diameter centered on the stud. This enabled both sides to be heated simultaneously for the filler metal to spread and penetrate into the braze gap. The braze flux melts within a few tenths of a second and covers the beam impingement area with a liquid film, and the filler metal being covered with the flux melts at the preplaced location. The filler metal, once melted at the periphery of the stud, instantly forms a small drop and moves toward the center of the joint geometry, that is, the highest-temperature region. A pseudo-TM01 mode (the so-called doughnut mode) of the continuous-wave CO2 laser beam was used as a heat source, for which TM00 mode generated by beam oscillator was opti-
cally modulated using an axicon lens. Relocation of the filler metal during the brazing process, including its wetting and spreading, was examined by using a high-speed motion analyzer, and the results were incorporated in the FEM for defining the solution domain and boundary conditions. The numerical results were obtained for typical process parameters and were compared with experimental ones determined by using the infrared and thermocouple measurements. The joining of a stainless steel stud to an aluminum plate by laser brazing is a new technology that is still under development. For a laser brazing performed in a few seconds by localized surface heating, control of thermal conditions in the workpiece is critical, even more so for a dissimilar-metal braze joint. In future work, the FEM used here will be refined to more accurately simulate the filler-metal behavior and further developed to fulfill a coupled solution of the heat conduction and filler-metal flow problem.
Braze Welding
Fig. 3.32
Schematic diagram of stud-to-plate laser braze (top) and a cross section of workpieces (bottom). O.D., outside diameter; SS, stainless steel
Braze welding offers considerably greater possibilities for joining composite materials, which are used on an increasing scale in various areas of technology (from sports equipment to space systems). The use of composite materials enables strength, stiffness, and impact loading resistance to be increased and the weight of structures to be reduced. Metal-matrix composites are used extensively, and these materials have a number of advantages in comparison with other materials: higher working temperature, shear and transverse strength under off-center loading and compression, higher impact toughness, and so on. Of these, composite materials with an aluminum matrix, reinforced with fibers of steel, boron, and carbon, as well as materials hardened with SiC and Al2O3 particles are used extensively. The method combines heating with a passing current with compression of the bond zone. Braze welding is carried out in projection welding machines using aluminum filler metals. The best results have been obtained using filler metals of the Zn-Al-Cu system. Filler metals based on tin have shown satisfactory spreading over the surface of the parent metal, but the strength was too low. Filler metals based on aluminum have high strength, but their melting point is
54 / Brazing, Second Edition
close to the melting point of the matrix alloy. At these brazing temperatures, debonding of the composite has been detected in the tinning stage, and brittle phases also grew. When using a filler metal of the Zn-Al-Cu system, the required strength of the joint was obtained by selecting the size of the lap. The results show that for a composite 1.2 mm (0.05 in.) thick, the thickness of the lap was 12 to 15 thicknesses of the material. However, compression during brazing reduced this value. Technology was developed for braze welding sheet composite materials (Ref 26, 27) and pipes and rods made of fiber-reinforced composites. Specialized equipment was developed for braze welding pipes and rods that enabled the process to be carried out with volume compression of the bonding zone (Ref 28). Braze welding technology has also been used to join composites to aluminum alloys, such as pipes of composites with endpieces of AMg6 (aluminum-magnesium) alloy. This technology was applied in the fabrication of frame structures, for example, a bicycle frame made of aluminum-boron composite. The frame weighs 1.2 kg (2.6 lb) and the fork 0.4 kg (0.9 lb). In the frame, the endpieces and pipes made of the composite were joined by braze welding. The technology of braze welding dispersionhardened composites does not differ from that used for braze welding fiber-reinforced composites. However, with an increase of the content of the hardening phase, the electrical resistance of the composite increases, thus reducing the time required to reach the brazing temperature. At braze welding temperatures, the gas saturation of the composite has no effect on pore formation (Ref 28). Automotive manufacturers have recently planned to introduce a new model and to use the gas metal arc braze welding process to join the roof panel to the front and rear quarter panels of the auto body. At the prototype stage, it was soon realized that this welding process was not meeting quality requirements. Some of the problems encountered were excessive weld reinforcement, panel distortion from heat buildup, ejection of braze metal in the form of spatter, and pinhole porosity in the braze-metal surface. The plasma arc braze welding process was suggested, although very little background was available on the process. A specialized torchand-wire feed system has been developed that would have the ruggedness needed for the assembly line.
The process has proven to have many advantages. Wire feed speed can be set separately of the current flow, thereby allowing the panels to be joined with current levels less than 50 A, which eliminates distortion. The process is capable of depositing 1.5 kg/h (3.3 lb/h) of braze metal, and brazing times are short: 10 and 20 s, respectively, for each joint. Because the plasma arc does not need to heat an electrode wire to melting, its force is not as violent as a gas metal arc. Therefore, spatter is not a problem, and porosity has been reduced to very low levels. Overall, the plasma arc braze welding process has proven capable of producing the quality required on this application (Ref 29). Structure tests (local buckling) on different stiffened structures (panel) made by braze welding were carried out at the temperature level from 360 to 750 °C (680 to 1380 °F). The materials of the panels were the titanium alloy Ti 1100, the intermetallic alloy Ti-48Al-2Cr, and the powder-metallurgy-produced DS-aluminum 620. The aim of the project was to consolidate the knowledge about the theory of the superposition of thermal loading and pressure loading used on structures such as hypersonic planes (Fig. 3.33) (Ref 30).
Diffusion Brazing Diffusion brazing is a hybrid joining process that combines the features of liquid phase joining and diffusion bonding and has the beneficial features of both techniques (Ref 31). Diffusion brazing and its lower-temperature analog, diffusion soldering, use a molten filler metal to initially fill the joint clearance, but during the heating stage, the filler diffuses into the material of the components to form solid phases, raising the remelt temperature of the joint. The steps involved in making a diffusion-brazed (or diffusion-soldered) joint are shown in Fig. 3.34. This process provides the ready means to fill joints that are not perfectly smooth or flat (a feature of liquid phase joining) while offering greater flexibility with regard to service temperature. The process also provides the following consequential advantages: • Facilitating the achievement of exceptionally good joint filling in large area joints • Allowing edge spillage from the joints to be tightly controlled and kept to a minimum
Chapter 3: Elements of the Brazing Process / 55
• Attaining high thermal conductivity with copper, silver, and gold systems, because the joint produced is composed of primary metal An alloy system suitable for diffusion soldering or brazing should have the following characteristics: • Preferably be a binary alloy, to keep the joint design and joining process as simple as possible • Have a phase constitution that includes a relatively low-melting-point eutectic reaction to initiate the melting process • Have as few brittle intermetallic compounds as possible, which should all melt at temperatures below or comparable to the joining temperature. This reduces the establishment of
Fig. 3.33
Vacuum weld brazed joint
diffusion barriers that can impede the process and lead to the formation of brittle interlayers. • The terminal primary metal phase should possess a wide range of solid solubility of the other constituents. This minimizes the risk of intermetallic phases precipitating during cooling of the assembly from the processing temperature and provides a greater process tolerance to the amount of filler metal introduced into the joint. Examples of alloy systems that satisfy these conditions and lend themselves to viable diffusion-soldering and -brazing processes are silver-tin, gold-tin, and nickel-boron (Ref 31). A study (Ref 32) of copper-tin diffusion brazing was made to identify the crucial process parameters needed to optimize this joining
56 / Brazing, Second Edition
process. Key parameters that were identified in practical trials were the thickness of the tin layer and the loading applied to the joint during the brazing cycle. It was established that the tin layer thickness must be controlled to 2 µm, within a tolerance of –0.5 µm, to obtain strong joints. A compressive load of 4 MPa (0.6 ksi) is adequate, while one of 3 MPa (0.4 ksi) is too low. The precise joining temperature is less critical, provided that it is 680 °C (1260 °F) or higher, sufficient to destabilize the brittle Cu3Sn intermetallic compound. There is a risk of reprecipitating this intermetallic phase on cooling if the heating operation does not adequately disperse the tin into the copper layers, due to the diminishing solubility of tin in copper as the joined assembly is cooled down to room temperature. This fact helps to explain why the thickness of the tin layer is highly critical, in
contrast with the silver-tin diffusion-soldering process, where the solubility of tin in silver is essentially maintained constant as the temperature is reduced. With regard to the application of this diffusion-brazing process to plasma-facing components for nuclear reactors, the initial concern about the relatively long reaction time at elevated temperatures (required for diffusion brazing and the resulting promotion of interfacial copper-beryllium intermetallic phases) was not borne out in practice, as indicated by the high strengths of the joints obtained in berylliumstrengthened copper assemblies (up to –230 MPa, or –35 ksi), both when produced under a pressure of 150 MPa (22 ksi) in a hot isostatic press and 4 MPa (0.6 ksi) in a uniaxial press.
Microwave Brazing
Fig. 3.34
Schematic of steps to make a diffusion-soldered (or diffusion-brazed) joint
The process of microwave heating is fundamentally different from any other form of heating. An understanding of the theoretical and practical aspects of the technology has been achieved (Ref 33). The joining of ceramic materials using microwave radiation has received increasing attention. There are potential economic benefits in terms of reduced energy costs as well as physical benefits due to localized heating. However, there is still a considerable amount of work to be done. Several groups have reported successful attempts to bond ceramics using microwaves (Ref 34–37). All bonds have been produced in the solid state and can be compared with diffusion bonding. The general mechanism of bonding is also the same as diffusion bonding, that is, the generation of heat across an interface, leading to atomic migration. Even when the parent ceramic does not couple well to microwave energy, for example, Al2O3 and aluminosilicate-base ceramics, bonds can still be produced by seeding the interface. Materials such as SiC, which heats very quickly in a microwave field, have been placed at the interface to induce high localized temperatures and aid bonding. In Japan, a research program (Ref 37), similar to that described in Ref 35 and 36, used a 6 GHz source with a maximum available power of 3 kW and described microwave heating and joining of Al2O3 and Si3N4 ceramics. Using ceramic rods 3 mm (0.12 in.) in diameter and 100 mm (4 in.) in length, butt joints were pro-
Chapter 3: Elements of the Brazing Process / 57
duced with and without intermediate layers. The joining operation was conducted in air for Al2O3 and in nitrogen for Si3N4. The samples were heated at a rate of 40 °C/s (70 °F/s) to the joining temperature (typically 1400 to 1850 °C, or 2550 to 3360 °F), held for 3 min with a pressure of 0.6 MPa (0.09 ksi), and then cooled at 15 °C/s (25 °F/s). Joints were evaluated by fourpoint bend testing. It was found that the bend strength increased with the rise in joining temperature. Three grades of Al2O3 ceramic—92, 96, and 99% (the balance being intergranular glass)— were evaluated. An average joint strength of 420 MPa (61 ksi) was obtained for 92% Al2O3. This value is approximately equal to the strength of the parent material. In general, joint formation became easier as the glass content of the material increased. This led the researchers to suggest a model for the microwave joining of Al2O3, assuming the microstructure of Al2O3 ceramics consists of Al2O3 particles surrounded by an intergranular glassy phase. The model relies on the intergranular glass heating preferentially to the Al2O3 particles and melting. Such melting (or softening) occurs over a temperature range of 1400 to 1800 °C (2550 to 3270 °F). The Al2O3 is not melted at these temperatures; consequently, the bond is formed via the glassy phase, and, on cooling, the bond line is not observed. The model is supported by observations that 99% Al2O3, which normally does not produce a bond to a similar material, does so if an intermediate layer containing a relatively high fraction of glass is used. The general conclusions that may be drawn from the work worldwide are that ceramics may be joined using microwaves. In particular, the presence of a glass phase, or interlayer, appears important. This is not detrimental, bearing in mind that most engineering ceramics do contain secondary glassy phases. The results also indicate that the use of microwaves allows ceramics to be joined in very short times. Typical bonding times are 5 to 15 min, compared to conventional bonding durations of 30 min and greater. If the only effect of microwave radiation was to heat the material, then it would be expected that joining times would be independent of the method used. In more recent work (Ref 38), a microwaveheating technique was developed for making a braze joint between a tungsten carbide (WC) support and a surface layer of polycrystalline
diamond or, alternatively, between a WC support and a relatively thin WC backing layer with polycrystalline diamond on its working surface. The technique would be used to fabricate diamond-covered cutting tool bits. Such bits could be used, for example, to drill geothermal wells and would be improved versions of some of the diamond-covered bits now used to drill oil and gas wells. Whereas the braze joints of the oiland gas-well versions become weakened at temperatures of 700 °C (1290 °F), the braze joints of the improved drill bits would be designed to withstand hard-rock-drilling temperatures up to 900 °C (1650 °F). The major problem in fabricating the improved drill bits is to use higher-melting-temperature brazing materials and to heat the braze joints accordingly to effect brazing, without overheating the diamond. Overheating in this context means heating to a temperature of approximately 1200 °C (2190 °F), causing the diamond to become graphitized and thereby to lose resistance to wear. The basic idea of this technique is to use the selective heating characteristics of microwaves to develop the required brazing temperature without overheating the diamond. Selective heating would be possible, because the commercially fabricated diamond is a very good absorber of microwaves, while the proposed brazing materials would be moderateto-good absorbers. Experience teaches that the best capillary action and shear strengths in braze joints on diamond/WC tool bits are achieved with fillets of 0.08 to 0.8 mm (0.003 to 0.03 in.), and that braze interlayers should be thick enough (at least 0.02 mm, or 0.0008 in.) to relieve stresses caused by differential thermal expansion between diamond and WC. The brazing material must be able, at the brazing temperature, to wet or diffuse into both the diamond surface layer and the WC substrate or into the WC backing layer and WC substrate, as the case may be. In preparation for a typical fabrication process according to this technique (Fig. 3.35), a diamond disk 2 to 3 mm (0.08 to 0.12 in.) thick is placed on top of a braze interlayer 0.08 to 0.8 mm thick on top of a WC substrate. This assembly of components is mounted in a region of strong electric field in a microwave-processing chamber. A pyrometer is focused on the diamond surface layer; during the subsequent microwave heating, the output of the pyrometer is used to monitor the temperature of the diamond and is used as a feedback signal to control
58 / Brazing, Second Edition
the microwave power to achieve the desired brazing temperature. The dimensions of the braze interlayer, the components to be brazed, and the process tooling are chosen, along with the temperature-versus-time heating curve, to obtain the strongest possible braze joint with minimal residual stress from differential thermal expansion. The braze interlayer could consist of a foil of a filler metal. Alternatively, the braze interlayer could be made of a combustion-synthesis compound, in which case microwave heating would be used to ignite a combustion wavefront with temperatures of thousands of degrees. When this wavefront reached the interfaces with the adjacent backing and substrate layers, it would provide sufficient local heating to form the desired braze joint. In conclusion, the feasibility of joining ceramic materials using microwave heating is proven, using interfacial agents or seeding interlayers, such as glasses, and advantages are expected in cost-effectiveness and improved microstructure and properties.
The Future of Braze Processing Automation of the brazing process is the key to maximizing production and quality and minimizing costs. Typical machines that are currently available are capable of automatic brazing by torch, furnace, induction, infrared, and resistance processes. Rotary and shuttle machines automatically apply paste and filler metal, heat the parts, torch braze the assembly,
Fig. 3.35
cool the brazement, and remove some or most of the flux. For small parts, automatic unloading is common. Many engineers are reluctant to try to automate brazing operations for fear of upsetting a well-established process and causing quality problems because of a method change. Their attitude, in short, is, “If it’s working and producing parts, don’t change it; don’t upset it.” However, brazing processes, when properly understood, are just as adaptable to increased productivity through automation as any other process; on the other hand, of course, they are also subject to similar limitations. Regardless of how attractive the idea of automating a brazing operation may be, the first step usually is to check out the economics. If automation is economically justified, the next step is to investigate the mechanics of how to automate the operation. In summary, the following steps are necessary for successful automation: • Make an economic analysis to justify automation. • Break down the total operation into its several parts and determine which can be eliminated, which can be combined, and which can be carried on in conjunction with other operations. • Determine the required type of automated equipment (rotary indexing machine, constant rotary machine, in-line conveyor, inline indexing conveyor, shuttle machine, or racetrack-type conveyor). The types of parts to be brazed and the required production rate
Microwave heating with temperature feedback control makes it possible to braze polycrystalline diamond to an underlying layer of tungsten carbide without overheating the diamond.
Chapter 3: Elements of the Brazing Process / 59
normally determine what type of mechanized equipment is most feasible. • Apply specialized skills and knowledge to the various design and engineering problems involved in realizing the objectives of automation: mechanics (in the selection of a mechanical movement and also in the design of the part-holding fixture), selection and application of flux and filler metals, type and position of heat application, cooling, additional mechanics of automatic loading and unloading, and electrical control systems.
Practical Experience, Work-Related Tips, and Problem Solving Example 1: Torch Brazing of a 304L Stainless Steel Tube to a Fitting Joint. An attempt was made to braze a tube to a fitting joint that does not have suitable strength, and when it pulled out of the joint, the filler metal did not appear to flow the full length of the joint. The fillets were good, and it appears to have been brazed well before testing. The base metal was 304L stainless steel, and the filler metal was BAg-7 (a silver filler metal). A white flux was used to cover the parts, which were then assembled and torch brazed. The filler metal was fed from the outside of the joint. It appears that the brazement was not heated uniformly, and while the outer end of the fitting was heated up adequately, the back portion of the fitting did not reach a temperature sufficient to allow the filler metal to flow. The front end of the joint was considerably thinner and lighter and thus heated up quite rapidly. However, the back section was considerably heavier and more difficult to heat. The molten filler metal froze before it reached the back end of the joint because the metal was cooler, and the filler metal solidified in the joint. To solve this problem, one must control the torch so that heating occurs closer to the back end of the fitting to ensure that this portion of the joint is up to temperature before the filler metal is applied. A second technique is to supply a ring at the back of the prefluxed joint, then assemble the tube and heat the assembly. Using 304L base metal is significant since the L grade is low carbon and would not form carbide precipitation and, potentially, avoid corrosion problems. The selection of the BAg-7 filler metal, which has a brazing range of 652 to 760 °C
(1206 to 1400 °F), was significant. This filler metal is one of the lower melting materials that does not contain cadmium. To ensure good strength of the joint, the suggested overlap distance should be 3 to 4T, where T is the thickness of the thinner member (which in this case is the tube). With this overlap distance, the strength of the joint should be adequate for this type of fitting. Example 2: Furnace Brazing of a Leaded Steel Fitting to a Copper Heat Exchanger Shell. Brazing a leaded steel with a silver filler metal could create a problem. If the parts were freshly machined, the BAg-4 filler metal with standard white paste flux for silver filler metals should be adequate for brazing the heat exchanger components. A second filler metal that could be suitable is BAg-24, which also contains silver-copperzinc-nickel. This filler metal has a narrower melting range and thus is good for brazing tighter joints. One point to consider when silver brazing a steel fitting to a copper shell of a heat exchanger is the possibility of galvanic corrosion. Depending on the electrolyte, a steel part may show a higher tendency to corrode. Example 3: Brazing Parts Exposed to Synthetic Machining Oils. The newer synthetic oils that are used in machining bear watching. Even though some brazing problems have been traced to oils and lubricants, a definitive answer on this subject has not been formulated to date. Example 4: Application of Filler Metal for Continuous Brazing. In another situation, if one is brazing 304 stainless steel with copper paste filler metal using a continuous brazing furnace with an Inconel muffle through the hot zone, one must be alert to any holes developing in the bottom of the retort in the hot zone. When there is excess copper, it can ball up at the bottom of the part or drip off, and it can drip through and land on the retort. Metallurgically, nickel and copper are mutually soluble in all proportions. Since the Inconel retort is a very high nickel alloy, the copper that pools in the retort actually creates a coppernickel alloy. If the concentration of copper gets high enough to cause melting, the hole in the retort occurs. In order to prevent the above problem, the best way is to tighten all the surface fits so that the clearance on all joints is zero. The best joint strength is obtained with the press-fit or metalto-metal surface where the copper in the joint is
60 / Brazing, Second Edition
less than 0.03 mm (0.0012 in.) thick when metallographically inspected. Example 5: Brazing of 409 Stainless Steel. In brazing 409 stainless steel for automotive tubing and fittings, on some occasions the brazed parts come out bright and shiny and sometimes with different purple hues, which interferes with the flow of the filler metal. The brazing work is performed in a continuous furnace with a nitrogen/hydrogen atmosphere with a dewpoint of approximately –46 °C (–50 °F). Copper filler metal is used on some assemblies and nickel-base filler metals on others. To solve the problem and the cause of the variation in color on these brazed parts and why it interferes with the flow of the filler metal is to utilize a nickel plate with approximately 0.01 mm (0.0004 in.) of electrolytic nickel. To provide for easier brazing, one could change to a 409Cb alloy. The niobium (columbium) addition to the 409 also ties up carbon, and the niobium does not produce an oxide on the surface like titanium that interferes with brazing. Example 6: Brazing of Leaded Brass. Lead is more readily vaporized than zinc, and thus any open globules of lead from a machine cut will outgas the lead. There is then a high lead content in the dust taken out of the furnace. This dust becomes a hazardous material to work with and dispose of, and must be handled properly. Example 7: Atmosphere Control. A vacuum furnace operates at 0.0013 Pa (2 × 10–7 psi) at a brazing temperature of 1070 °C (1950 °F). How does the operator determine that the atmosphere in the furnace is acceptable? One easy way to test atmosphere quality as well as base metals of titanium, zirconium, niobium and tantalum, is to use a 0.13 mm (0.005 in.) titanium foil, which is put in the vacuum furnace with or without the load. To check the furnace conditions, use the titanium foil without additional parts. Be certain that the titanium foil rests on a dense piece of alumina or is supported by clean ceramic paper. With a very good atmosphere, the titanium foil will be the same color after the heating cycle as it was when it went in the furnace. Next, the test is to take the foil and bend it back on itself. If the atmosphere was not good, the foil will fracture. If the atmosphere is very good, the foil can be opened without fracturing. The residual gas analyzer (RGA) is an excellent addition to the vacuum furnace. It can pro-
vide a great deal of information as to the residual gases in the vacuum furnace. As is typical of the conditions that can exist in furnaces, one is able to braze titanium very readily at 0.13 Pa (2 × 10–5 psi) in a very clean furnace. Both titanium and zirconium foil can be used for the same application. However, the titanium foil is more readily available. Example 8: Magnesium Buildup in a Vacuum Furnace Used to Braze Aluminum. Several people and companies are brazing aluminum in a vacuum furnace and are having problems with magnesium buildup on the walls and doors between the charge and exit vestibules. The magnesium tenaciously sticks to the metal surfaces and is difficult to remove. There is a material stopoff that has been used in the brazing industry for more than 50 years to limit the flow of filler metal. It has also been found to work well in the salt bath furnaces used in the brazing of aluminum. There are many types of stopoffs available for various applications. Stopoff materials are very interesting and necessary materials because not much material is needed, but they do prevent filler metals from flowing into unwanted areas. (Additional information on stopoff materials is provided in Chapter 7.) Stopoffs are available in solvent-based and water-based versions. The water-based type is recommended for use inside vacuum aluminum brazing furnaces because solvent vapors are undesirable in this confined area. The stopoff is painted on the metal surfaces where the magnesium usually condenses. Stopoff materials can be obtained in a painttype consistency that can be applied by brush or roller to the inside of the furnace in areas where the magnesium is condensing. Apply the stopoff right over what is already there to allow future condensed magnesium to be readily removed. Example 9: High-Frequency Induction Brazing of Stainless Steel. This example involves high-frequency induction brazing of stainless steel parts with BNi-7 filler metal. The parts are inside a quartz tube and use a vacuum atmosphere of 0.013 Pa (2 × 10–6 psi). The braze has been successfully performed; however, some areas of the part reach only a low-red heat. These areas discolor, and organic solvents will not remove the discoloration. At the joint, where the brazing temperature was 1050 °C (1920 °F), the part is bright, shiny silver. The cement being used to hold the filler metal in
Chapter 3: Elements of the Brazing Process / 61
place is not causing this discoloration. Chromium-containing base metals such as 304L will start to oxidize at approximately 540 °C (1000 °F) and will continue to increase the oxide thickness up to 760 °C (1400 °F) before getting progressively lighter. Good brazing quality will increase. When a base metal such as 304L is hot rolled or forged and pickled, the chromium is depleted from the surface, leaving an iron-nickel alloy at the surface. This makes for a much easier braze and requires a less critical atmosphere than the freshly machined chromium-containing stainless steels. To completely eliminate discoloration where there is chromium at the surface, the entire part must be heated in an adequate protective atmosphere close to the brazing temperature, and then a specific type of stopoff is applied to eliminate the discoloration.
REFERENCES
1. R.G. Gilliland, Wetting of Beryllium by Various Pure Metals and Alloys, Weld. J., June 1964, p 248–258 2. D.R. Milner, A Survey of the Scientific Principles Related to Wetting and Spreading, Br. Weld. J., Vol 5, 1958, p 90–105 3. Welding, Brazing, and Soldering, Vol 6, Metals Handbook, 9th ed., American Society for Metals, 1983, p 956 4. Weld. Des. Fabr., June 1993, p 16 5. Weld. J., Oct 1989, p 55 6. P. Kosir and D.K. Patrick, Data Collection and Modern Controls for Hump Mesh Belt Electric Brazing Furnace, Ind. Heat., Oct 1999, p 105–109 7. J. Boswell, C. Field, et al., Production Versatility Increased in Semi-Continuous Controlled Atmosphere Brazing Furnace, Ind. Heat., May 1999, p 73–77 8. T.V. Evans, R.P. Johnson, et al., Next Generation Controlled Atmospheric Braze Processing of Aluminum Heat Exchangers, Ind. Heat., Jan 1997, p 28–30 9. D. Garg, C.E. Baukal, et al., Effective Cooling of Components, Ind. Heat., Feb 1997, p 43–47 10. L.L. Ashburn, Fluxless Vacuum Furnace Brazing of Aluminum Particularly Advantageous for More Critical Applications, Ind. Heat., Mar 1994, p 47–51
11. P.F. Gerbosi and J.F. Libsch, ControlledAtmosphere Brazing with Induction Heating, Weld. J., Oct 1989, p 32–37 12. Brazed Joints, Thin-Wall Tubing Streamline B-70 Plumbing, Space/Aeronaut., May 1960, p 75–90 13. “Brazed Fluid Transmitting Systems,” Gemini Project, Contract NAS9-170, Progress Report 11, McDonnell Douglas, 12 Feb 1963 14. E.C. Nezbeda, “Lunar Excursion Module Manufacturing,” National Aeronautic and Space Engineering and Manufacturing Meeting (Los Angeles, CA), 4–8 Oct 1965 15. W.E. Clautice, Induction Brazing at the Kennedy Space Center, Weld. J., Vol 53(No. 10), Oct 1974, p 612–622 16. Compatibility of Au-Cu-Ni Braze Alloy with NH3, NASA Tech. Briefs, Vol 3(No. 2), 1978, p 226–227 17. M.M. Schwartz, Brazing, ASM International, 1987 18. L.B. Roginskii et al., Reconditioning by Brazing Cast Iron Crankshafts, Svar. Proizvod., Vol 6, 1993, p 14–15 19. S.V. Lashko and N.V. Lashko, Brazing of Metals, Publ. Mashinostroenie, Moscow, 1988 20. S.S. Nurkhanov and L.B. Roginskii, Reconditioning and Hardening Rolls by Electric Resistance Brazing, Weld. Int., Vol 11(No. 1), 1997, p 75–77 21. Q.D. Mehrkam, Never Underestimate the Power of a Salt Bath Furnace, Weld. Des. Fabr., AJAX Reprint 184, Mar 1968 22. Aluminum Brazing Handbook, 1st ed., The Aluminum Association, Jan 1971 23. C.A. Blue and R.Y. Lin, Rapid Infrared Joining of Ti-6A1-4V, Proc. Adv. Mater., Vol 4, 1994, p 21–28 24. J.S. Park and J.M. Kim, Finite Element Modelling for Thermal Analysis of Studto-Plate Laser Brazing for a Dissimilar Metal Joint, NTIS Alert, 1 Aug 1997, p 3 25. J.-S. Park and S.-J. Na, Heat Transfer in a Stud-to-Plate Laser Braze Considering Filler Metal Movement, Weld. J., Apr 1998, p 155–163 26. V.F. Khorunov and V.S. Kutchuk-Yatsenko, “Brazing of Sheet Composite Materials with an Aluminum Matrix,” International Symposium (Montreal, Canada), 1990.
62 / Brazing, Second Edition
27. I.S. Duckno and V.S. Kutchuk-Yatsenko, A Method for Making Permanent Joints of Aluminium-Based Metallic Composite Materials to Aluminium Alloys and Ce ramics, Second Japan. International SAMPE Symposium, Society for the Advancement of Material and Process Engineering, 1991, p 758–765 28. I.V. Zvolinskii, V.S. Kutchuk-Yatsenko, et al., Special Features of Brazewelding Metal Matrix Composites, Weld. Int., Vol 9(No. 1), 1995, p 41–43 29. R.V. Hughes et al., Weld. Met. Fabr., Vol 63(No. 3), 1995, p 110–111 30. T. Fleischer, Thermal Loaded Metallic Structures for Hypersonic Planes, NTIS Alert, 15 June 1997, p 27 31. G. Humpston and D.M. Jacobson, Principles of Soldering and Brazing, ASM International, 1994, p 129 32. S.P.S. Sangha, D.M. Jacobson, and A.T. Peacock, Development of the Copper-Tin Diffusion-Brazing Process, Weld. J., Oct 1998, p 432–438
33. J.A. Fernie, “Introduction to Microwave Heating and Its Use for the Joining of Ceramics,” 457/1992, TWI, Aug 1992 34. T.T. Meek and R.D. Blake, CeramicCeramic Seals by Microwave Heating, J. Mater. Sci. Lett., Vol 5, 1986, p 270–274 35. D. Palaith, R. Silberglitt, and E.L. Libelo, “Microwave Joining of Ceramic Materials,” Second International Conf. on Ceramic Materials and Components for Engines, 14–17 Apr 1986 (LubeckTravemunde, W. Germany) 36. D. Palaith, R. Silberglitt, C.C.M. Wu, et al., Microwave Joining of Ceramics, Mat. Res. Soc. Symposium, Vol 124, 1988, p 255–266 37. H. Fukishima, T. Yamana, and M. Matsui, Microwave Heating of Ceramics and Its Application to Joining, Mat. Res. Soc. Symposium, Vol 124, 1988, p 267–272 38. M. Barmatz, H.W. Jackson, and R.P. Radtke, Microwave Brazing of Polycrystalline Diamond onto Drill Bits, NASA Tech. Briefs, Dec 1998
Brazing Second Edition Mel M. Schwartz, p63-176 DOI: 10.1361/brse2003p063
Copyright © 2003 ASM International® All rights reserved. www.asminternational.org
CHAPTER 4
Base Metals and Base-Metal Family Groups THIS CHAPTER BEGINS by describing the general metallurgical considerations related to the selection of different base metals. The remainder of the chapter is devoted to describing specific considerations related to the groups of alloys that are most commonly joined by brazing.
Metallurgical Reactions Some metals and alloys exhibit metallurgical phenomena that influence the behavior of brazed joints and base-metal properties and, in some cases, necessitate special procedures. These phenomena may be classified as: • Base-metal effects, including carbide precipitation • Hydrogen embrittlement • Heat-affected zone and oxide stability • Sulfur embrittlement • Filler-metal effects, such as vapor pressure • Base-metal/filler-metal interactions, including alloying • Phosphorus embrittlement • Stress cracking Other factors that cause interactions between base metals and filler metals include postbrazing thermal treatments, corrosion resistance, and dissimilar-metal combinations. The extent of interaction varies greatly, depending on compositions (base metal and filler metal) and the duration and extent of the thermal cycles in the processing. There is always some interaction, except when mutual insolubility permits practically no metallurgical interaction. In addition to the base-metal effects mentioned previously and the normal mechanical
requirements of the base metal in the brazement, the effect of the brazing cycle on the base metal and the final joint strength must be considered. Cold-worked-strengthened base metals are annealed and the joint strength reduced when the brazing process temperature and time are in the annealing range of the base metal being processed. Hot-cold-worked heat-resistant base metals can also be brazed; however, only the annealed physical properties are available in the final brazement. The brazing cycle, by its very nature, usually anneals the cold-worked base metal, unless the brazing temperature is very low and the time at heat is very short. It is not practical to cold work the base metal after the brazing operation. When a brazement must have strength after brazing that is above the annealed properties of the base metal, a heat treatable base metal should be selected. The base metal can be an oilquench type, an air-quench type that can be brazed and hardened in the same or a separate operation, or a precipitation-hardening type that can be brazed and solution treated in a combined cycle. Parts already hardened may be brazed with a low-temperature filler metal, using short times at temperature to maintain the mechanical properties. The strength of the base metal has a profound effect on the strength of the brazed joint; thus, this property must be clearly kept in mind when designing the joint for specific properties. Some base metals also are easier to braze than others, particularly by specific brazing processes. Carbide Precipitation. If stainless steels are heated to temperatures from 425 to 815 °C (800 to 1500 °F), the carbon in the base metal combines preferentially with chromium to form
64 / Brazing, Second Edition
chromium carbide, usually at the grain boundaries. This chromium depletion reduces the corrosion resistance of the stainless steel. This condition has been defined as sensitization by some investigators. In certain corrosive environments, the mechanical properties may be impaired, with little or no apparent surface attack. A short brazing cycle keeps the chromium carbide precipitation to a negligible level with normal types of stainless steels. When this is not possible, one of the special grades of stainless steel may have to be used if its corrosion resistance is to be preserved after brazing. Precipitated carbides in stainless steels may be redissolved by heat treating at 1010 to 1120 °C (1850 to 2050 °F), followed by rapid cooling. Another stabilizing treatment that disperses the unprecipitated chromium uniformly throughout the structure consists of heating to 870 °C (1600 °F) for 2 h, followed by furnace cooling to 540 °C (1000 °F) and subsequent air cooling. If the cooling from the brazing temperature is rapid, no appreciable amount of carbides is precipitated. Where this cannot be done due to mass and it is necessary to braze stainless steels for corrosive service, one of the stabilized compositions, such as type 347 or 321, or an extra-lowcarbon grade, such as 304L, should be used. There are several ways to prevent or minimize the deleterious effects of carbide precipitation. First, because the reaction is time dependent, keeping the brazing thermal cycle as short as possible can minimize carbide precipitation. With short cycle times, such as would result from torch or induction brazing of small parts, even the unstabilized grades can be brazed without severe losses in corrosion resistance. The susceptibility to carbide precipitation also depends on carbon content. Thus, type 304 is less susceptible than type 302, and the extralow-carbon grades, such as type 304L, are relatively insensitive to carbide precipitation. For critical applications, type 347, the niobium-stabilized grade, is recommended. It has good high-temperature strength and can be brazed without danger of impaired corrosion resistance. Type 321 is also a stabilized grade, but it has slightly lower general corrosion resistance than type 347 and is more difficult to braze, because titanium is used as the carbidestabilizing element. When high-melting-point filler metals are used, precipitated carbides can be redissolved by heat treatment after brazing.
Alternatively, corrosion resistance can be restored by diffusing chromium back into the depleted area around the carbide precipitates. Two hours at 870 °C is the recommended homogenizing heat treatment. Hydrogen Embrittlement. Hydrogen can also be a source of trouble. Because of its small atomic size, it is able to diffuse quite rapidly through many metals, and the rate of diffusion increases with temperature. When hydrogen diffuses into a metal that has not been completely deoxidized, it may reduce the oxide of the metal, if the temperature is high enough. Metallic sponge and water vapor are the end products of this reaction. Once hydrogen has diffused into the metal, several things can happen. If oxygen is present, the hydrogen may combine with it to produce water vapor. The water-vapor molecule, unlike the hydrogen molecule, is too large to diffuse out of the metal, and the high vapor pressures that develop can literally tear the metal apart by starting many fissures and blisters, mainly at the grain boundaries. The ultimate result is hydrogen embrittlement. It commonly occurs in copper and copper-base alloys that have not been deoxidized. Pressures developed for tough pitch copper have been calculated to be as high as 620 MPa (90 ksi). Electrolytic tough pitch copper, silver, and palladium, when they contain oxygen, are subject to hydrogen embrittlement if heated in the presence of hydrogen. If tough pitch copper is to be brazed without embrittlement, hydrogen must not be present in the heating atmosphere. A better practice is to use deoxidized copper or oxygen-free copper where brazing is to be performed. Oxygen-free copper, if improperly heated, may also be oxidized and become subject to hydrogen embrittlement. It is impractical to salvage hydrogen-embrittled copper. A recently completed study (Ref 1) examined several commercial filler metals containing zinc, cadmium, or phosphorus that were found to cause embrittlement by migration of (copper) oxide to the grain boundaries, causing void formation and rupture of grain boundaries. (Oxides do not migrate as such but rather dissolve in the grains. The oxygen diffuses to the grain boundaries, where it recombines, forming lessstressed particles.) The brazing was performed both by the conventional fluxed and fluxless methods without the presence of any source of hydrogen; however, this still resulted in the
Chapter 4: Base Metals and Base-Metal Family Groups / 65
same embrittlement. Therefore, it was concluded that the influence of flux is insignificant, because embrittlement also persisted in joints brazed without flux under argon. To narrow the possibilities of embrittlement, a 72Ag-28Cu eutectic filler metal was used to fill several joints, and there was no such embrittlement, even in the most drastic brazing conditions employed. In comparing this filler metal with the others, the only difference was in the composition. While BAg-8 contains silver as the only addition to copper, the other filler metals are a ternary or quaternary formulation, containing additions of zinc, cadmium, and phosphorus. The difference between silver and these other alloying additions is in their ability to reduce copper oxide; silver cannot act as a reducing agent. This indicates that embrittlement occurs as a result of the interaction of the other alloying elements with the copper base metal and not directly because of other factors, such as flux, atmosphere, and the time-temperature cycle. These parameters were identical for all the filler metals, and their influences were as expected; that is, the embrittlement was enhanced when more drastic conditions were employed. Furthermore, embrittlement occurred only in tough pitch copper (containing oxygen as Cu2O precipitates) but not in phosphorus-deoxidized copper, which is completely free of oxides. As a consequence, the coexistence of the additional alloying elements of zinc, cadmium, and phosphorus, together with oxygen in the substrate, is the necessary prerequisite for embrittlement. The results indicate that the responsible elements are those that are capable of reducing copper oxide by a mechanism analogous to the hydrogen embrittlement of tough pitch copper. It seems quite certain that the thermodynamic activity of zinc, cadmium, and phosphorus is sufficiently high to cause the following reaction to take place: CuO + (X) 3 Cu + XO
where X stands for one of these elements in the filler metal. The mechanism suggested by this study is embrittlement induced by the coexistence of copper oxides in the base material together with certain alloying additions in the filler metal, such as cadmium, zinc, or phosphorus, capable of reducing the copper oxides. Steel is especially prone to another mecha-
nism for hydrogen embrittlement. In this type, hydrogen diffuses into the steel as atomic hydrogen in the same manner as it diffuses into copper, but it tends to accumulate in small voids, such as those around nonmetallic inclusions and at grain boundaries. Water vapor is not formed, as in copper, but the hydrogen atoms combine to form hydrogen molecules, which are less mobile and remain trapped at the discontinuities and, as a result, increase the concentration of molecular hydrogen, increase the vapor pressure, and lower the ductility of steel when stressed. However, steel and other ferrous alloys may be salvaged by allowing the hydrogen to diffuse out by baking at slightly elevated temperatures (95 to 205 °C, or 200 to 400 °F) or by permitting the steel to stand for long periods of time until the ductility is regained. A third type of embrittlement can occur when hydrogen combines with the metal to form hydride. The hydride lowers the notch toughness and affects the strain rate of the metal. Titanium, zirconium, niobium, tantalum, and their alloys are subject to this form of hydrogen embrittlement. Ductility can be restored if proper thermal treatments are followed after brazing; however, an inert or vacuum atmosphere should be used for brazing to avoid any embrittlement. Most other metals and alloys whose oxides may be reduced by hydrogen contain an excess of deoxidizing elements and are not subject to hydrogen embrittlement. Heat-Affected Zone. The heating of base metals may cause changes in their properties, particularly if the metals are heated above their annealing temperatures. Base metals whose mechanical properties were obtained by cold working (hard tempers) may soften or undergo an increase in grain size if the brazing temperature is above the recrystallization temperature. Where mechanical properties are obtained by thermal treatment, they may be altered by the brazing operation. Materials in the annealed condition generally experience no appreciable change due to brazing. The width of the zone through which these changes may occur varies with the process used. If the heating is localized, as in torch or induction brazing, the effects are confined to a narrow zone. If the whole assembly is heated, as in furnace brazing, the entire assembly is affected. In general, the heat-affected zone produced during brazing is wider and less sharply defined than those resulting from other welding processes.
66 / Brazing, Second Edition
Oxide Stability and Formation. When clean metals are heated to brazing temperature, their surfaces may form metal oxides if the atmosphere around the part contains oxygen. Oxidized metal surfaces are usually difficult to wet with most filler metals. Fluxes and special atmospheres are designed to prevent oxide formation or to reduce at elevated temperature any oxidation that occurs during initial heating (Ref 2). Chromium, aluminum, titanium, silicon, magnesium, manganese, and beryllium all have oxides that are difficult to remove, and, therefore, these metals usually require special preparation (Ref 2). Fluoride-bearing fluxes can reduce some oxides; hydrogen gas of sufficient purity can reduce them above certain temperatures, and techniques such as vacuum brazing may have to be used. Ideally, oxide formation should be prevented by brazing in low-dewpoint or vacuum atmospheres. Sulfur Embrittlement. Nickel and certain alloys containing appreciable amounts of nickel, if heated in the presence of sulfur or compounds containing sulfur, may become embrittled. This occurs when a low-melting nickel sulfide is formed preferentially at the grain boundaries; this sulfide, being brittle and weak, cracks if subsequently stressed. Material so embrittled is usually scrapped, because the damage that has occurred cannot be salvaged. Nickel and nickel-copper alloys are most susceptible to this attack, whereas alloys containing chromium are less susceptible. It is important that alloys in which nickel is the major component be clean and free of sulfur-containing materials (such as oil, grease, paint, and drawing lubricants) prior to heating and that heating be done in relatively sulfur-free atmospheres. Vapor Pressure. Every metal is in equilibrium with its vapor pressure; some amount of the metal is present in the gaseous state. For most metals, at normal temperatures, this vapor pressure is so small as to be considered nonexistent. For vacuum-tube applications, some metals, such as zinc and cadmium, have relatively high vapor pressures, give off undesirable gases at normal brazing temperatures, and therefore cannot be permitted as constituents of the filler metal. Accordingly, special vacuum-tube-grade filler metals have become commercially available (see Chapter 5, “Brazing Filler Metals”), and special fluxes are used in some situations. Base-Metal/Filler-Metal Interactions. There are always some interactions between the filler metal and the base metal. Although some
of this interaction aids in wetting the base metal, other detrimental effects may occur. Such effects include: • Formation of brittle intermetallic compounds that lower joint strength • Diffusion of the filler metal into the base metal to produce color changes • Creation of a new alloy—with a higher melting point than that of the original filler metal—that chokes off the flow of the filler metal Researchers (Ref 3–6) have reported on their work whereby modified brazing processes for nickel-base materials were used to reduce the formation of brittle phases in the braze joint and also to speed the joining operation. The three modified brazing processes developed by the researchers included: • Brazing under defined load • High-speed brazing • Application of mechanical-excitation brazing Additives (silicon, boride) used to reduce the melting point of nickel-base materials brazed with nickel filler metals cause brittle phases, which exert a negative influence on the mechanical properties of the brazed joints. Diffusion annealing and subsequent aging merge the brittle phases in the braze joint, whereas the aging causes hardening of the base material. After this, the mechanical properties of the joint are comparable to those of the base material. However, applying this heat treatment may cause the formation of coarse grains. Brazing under defined load pressures differs from conventional brazing in that a previously defined load is set up quickly after the brazing temperature is reached. Brazing temperatures of 1150 and 1180 °C (2100 and 2150 °F) with temperature retention times of 1 and 10 min, respectively, have proved to be the best parameter combination. This type of pressure-brazed joint with a more homogeneous microstructure produces good strength properties. Brazing time is of little significance; in contrast, the brazing temperature is of great importance to the strength properties. However, strength values that can be attained without diffusion annealing are comparable with those of conventionally brazed specimens using cost-intensive and time-consuming heat treatments.
Chapter 4: Base Metals and Base-Metal Family Groups / 67
The formation of brittle phases can be influenced or even avoided by combining a considerable reduction of the brazing time with a simultaneous increase in the brazing pressure. The only useful method to achieve the required high gradient of temperature is a conductive heat treatment technique. Another method to prevent the formation of brittle phases is by the mechanical excitation of the components during the brazing process. While brazing, a transducer directly connected to one specimen transfers high-frequency energy into the brazing couple. Using ultrasonic vibrations of approximately 30 kHz (amplitude of approximately 2 µm in the longitudinal direction of the specimens), the accumulation of brittle phases is prevented. The wetting of the base metal with filler metal is improved, because the ultrasonic vibrations destroy any existing surface oxides. The superposition of mechanical excitation produces seams of a quality comparable with those joints heat treated (1100 °C, or 2010 °F, for 20 h) after conventional brazing. Alloying is one of the significant basemetal/filler-metal interactions that can determine the behavior of brazed joints. The extent of interaction varies greatly, depending on the compositions of the base metal and the filler metal and on thermal cycles. There is always some interaction, except where mutual insolubility permits practically none. The term alloying is a general term covering practically every aspect of interaction. Some of these aspects are as follows. First, the molten filler metal can dissolve the base metal. Second, constituents of the filler metal can diffuse into the base metal, either through the bulk of the grains or along the grain boundaries, or can penetrate the grain boundaries as a liquid. The results of such base-metal dissolution or filler-metal diffusion may be to raise or lower the liquidus or solidus temperature of the filler-metal layer, depending on composition and thermal cycle. Examples include nickel, cupronickel, or Monel joined with pure copper filler metal; enough dissolution and diffusion occur so that the solidus of the copper filler metal is increased and flow is terminated. This also means that the remelt temperature of the filler-metal layer is higher than its original solidus temperature. In brazing of ferrous-base high-temperature alloys with filler metals containing boron, grainboundary penetration of the base metal by a low-
melting complex can cause joint degradation. This effect is particularly damaging if the base metal is thin, as in the case of brazed honeycomb sandwich panels. Producers and experienced users of high-temperature filler metals should be consulted during the design of parts for which the use of these filler metals is anticipated. Formation of intermetallic compounds as a result of interactions between constituents of the base and filler metals can occur, and these compounds are usually brittle. Whether or not such compounds form depends on base-metal and filler-metal compositions, time, and temperature, and just because intermetallic compounds do form, it does not necessarily follow that the joint is so embrittled as to lose engineering utility. This depends on the nature of the specific compound, its quantity, and its distribution. Phosphorus Embrittlement. Phosphorus combines with many metals to form brittle compounds known as phosphides. For this reason, copper-phosphorus filler metals are not usually used with iron- or nickel-base alloys; however, two nickel-base filler metals—10 to 12 P, 0.10 maximum C, balance Ni; and 13 to 15 Cr, 9.7 to 10.5 P, 0.08 maximum C, balance Ni—have been used in some applications for brazing heat-resisting alloys. The first filler metal is extremely free flowing, exhibits a minimum amount of erosion with most nickel- and iron-base alloys, and is good for use in exothermic atmospheres. The second is used for brazing of honeycomb structures, thin-wall tube assemblies, and other structures that are used at high temperatures. Erosion can be controlled because of low solubility with iron- and nickelbase alloys, and this filler metal produces strong, leakproof joints with heat-resistant base metals at relatively low brazing temperatures. Furthermore, it is recommended for nuclear applications where boron cannot be used. Stress Cracking. There are many highstrength materials, such as stainless steels, nickel alloys, and copper-nickel alloys, that have a tendency to crack during brazing when in a highly stressed condition and in contact with molten filler metal. Materials with high annealing temperatures—and particularly those that are age hardenable—are susceptible to this phenomenon. Such cracking occurs almost instantaneously during the brazing operation and is usually readily visible, because the molten filler metal follows the crack and completely fills it. This process has been described as stress-corrosion cracking (SCC), where the molten filler
68 / Brazing, Second Edition
metal is considered to be the corrosion medium. Cracking of stressed steel in a caustic solution or stressed brass in an ammonia solution are widely known examples of SCC. Sufficient stress to cause stress cracking can be produced
by cold work prior to brazing or by an externally applied stress from mechanical or thermal sources during the brazing operation. When stress is encountered, its cause can usually be determined from a critical analysis of the
Thermal expansion curves for some common materials. These nomographs assume a case wherein α1 exceeds α2, so that scale value for (α1 – α2) is negative. Resultant values of ∆CD are therefore also negative, signifying that the joint gap is reduced on heating. Where (α2 – α1) is positive, values of ∆CD are read as positive, signifying enlargement of the joint gap on heating. Clearance to promote filler-metal flow must be provided at brazing temperature. D, nominal diameter of joint, mm (in.); ∆CD, change in clearance, mm (in.); ∆T, brazing temperature minus room temperature, °C (°F); α1, mean coefficient of thermal expansion, male member, mm/mm · °C (in./in. · °F); α2, mean coefficient of thermal expansion, female member, mm/mm · °C (in./in. · °F). Source: Ref 7, 8
Fig. 4.1
Chapter 4: Base Metals and Base-Metal Family Groups / 69
brazing procedure. The usual remedy is to remove the source of stress. Stress cracking has been eliminated by: • Using annealed-temper rather than hard-temper material • Annealing cold-worked parts prior to brazing • Removing the source of externally applied stress, such as improper fit of parts or jigs that exert stress on the parts
Fig. 4.1 (continued)
• Redesigning parts or revising joint design • Heating at a lower rate • Heating the fluxed and assembled parts in a torch brazing application to a temperature high enough to effect stress relief, cooling to the brazing temperature, and then hand feeding the filler metal The age-hardenable high-nickel alloys are very susceptible to SCC. These alloys should be
70 / Brazing, Second Edition
brazed in the annealed or solution-treated condition with a relatively high-melting filler metal (preferably above 870 °C, or 1600 °F) that has sufficient strength to withstand handling during the age-hardening treatment. Postbrazing Thermal Treatments. A postbrazing thermal treatment to improve mechanical properties in brazed assemblies is frequently desired. In ferrous alloys, this treatment entails quenching from an elevated temperature, followed by tempering at some lower temperature. In other alloys, such as beryllium copper, 17-7 PH, Inconel X, and some Monels, the treatment consists of heating to some intermediate temperature, followed by a controlled rate of cooling. When a thermal treatment is performed subsequent to brazing, it is important that the filler metal selected have sufficient strength at the thermal treatment temperatures to withstand the necessary handling. It is also important for the base metal, filler metal, and postbrazing thermal treatment to be compatible relative to temperatures in heating and cooling. Postbrazing thermal treatments may generate residual stresses in brazed joints and may result in lowered joint strength. Dissimilar-Metal Combinations. There are many dissimilar-metal combinations that may be brazed. In fact, brazing can often be used where metallurgical incompatibility precludes the use of other joining processes. One of the most important factors to consider in brazing dissimilar metals is rate of thermal expansion. If a metal having high thermal expansion surrounds a low-expansion metal, clearances that are satisfactory for promotion of capillary flow at room temperature are excessive at brazing temperature. Conversely, if the low-expansion metal surrounds the high-expansion metal, no clearance may exist at brazing temperature. For example, in brazing of a molybdenum plug in a copper block, the parts must be press fitted at room temperature. However, if a copper plug is to be brazed in a molybdenum block, a properly centered loose fit at room temperature is required (Ref 7–9). Figure 4.1 shows thermal expansion of copper, molybdenum, and other common materials. Nomographs have been developed that are useful for learning the actual changes in clearance in ring-and-plug joints between dissimilar metals (Fig. 4.2). The equation may be used in cases where more accuracy is important or where one of the variables is off the nomograph scale. In more complex joint configurations, it is
usually best to prepare preproduction samples to establish ideal clearances. A technique often used in brazing of material with different coefficients of expansion is sandwich brazing. A common application of this technique is the manufacture of carbide-tipped metal cutting tools. A relatively ductile metal is coated on each side with the filler metal, and the composite is used in the joint. This places a third material in the joint that creeps during cooling and reduces the stresses caused by differential contraction. In some variations of the technique, wire mesh is used in place of the foil. There are other factors that also must be considered for successful brazing of dissimilar metals. The filler metal must be compatible with both the base metals. Wide differences in basemetal melting points must be considered when choosing the filler metal (Ref 7, 8). Where corrosion or oxidation resistance is needed, the filler metal should have properties at least equal to the poorest of the two metals being brazed. In addition, under conditions of the application, galvanic couplers that may promote crevice corrosion should be avoided. Filler metals that form low-melting phases with the base metals are not recommended unless, as in some special cases, techniques are used to form the final filler metal in situ. The metallurgical reactions that occur during brazing or subsequent thermal treatments between the filler metal and the base metal are important. One example is brazing of aluminum to copper. The copper reacts with the aluminum to form a low-melting brittle compound. Such problems can be overcome by coating one of the base metals with a compatible metal. In the case of aluminum to copper, the copper can be coated with silver or a high-silver alloy and then the joints made with a standard aluminum filler metal.
Base-Metal Family Groups In the ensuing list of base-metal groups, it is apparent that prerequisites for ductile joint behavior include freedom from related hard phase bands in the brazed joint and a matching of elastic-plastic behavior in the brazed joint and base material. It thus follows that not only is the optimal structural development required for a joint capable of deformation, but also the elastic-plastic behavior of the whole system of filler metal and base metal must be considered
Chapter 4: Base Metals and Base-Metal Family Groups / 71
(Ref 10–12). This results in a ductile behavior in the brazed joint and an elastic and plastic behavior in the base metal. When the elastic and plastic properties of the brazed joint and base metal are well matched, this leads to the demand for filler metals whose mechanical properties in the joint are well matched to those of the base metals (Ref 13).
Aluminum and Aluminum Alloys The non-heat-treatable wrought aluminum alloys that are brazed most successfully are the ASTM 1xxx and 3xxx series (e.g., 1350, 1100, 3003, 3004, 3005) and low-magnesium alloys of the ASTM 5xxx series (e.g., 5005, 5050, 5052). Available filler metals melt below the solidus temperatures of all commercial wrought non-heat-treatable alloys. The non-heat-treatable wrought aluminum
Nomograph for equation ∆ C = D∆ T (α – α )
mm D (in.) 150
6
125
5
alloys have manganese or magnesium as major alloying elements. Some alloys contain only one alloying element; others contain several. The alloys with higher magnesium contents are more difficult to braze by the usual flux methods because of poor wetting by the filler metal and excessive penetration into the base metal. Filler metals are available that melt below the solidus temperatures of most commercial non-heattreatable wrought alloys. The heat-treatable wrought alloys most commonly brazed are the ASTM 6xxx series (6151, 6951, 6053, 6061, 6063). The ASTM 2xxx and 7xxx series of aluminum alloys are low melting and, therefore, not normally brazeable, with the exception of the 7072 and 7075 alloys. The rest of the wrought aluminum alloys can be strengthened by heat treatment and, to a much smaller degree, by cold working. They are
D
2
1
(α – α ) 2
1
6
6
10 mm/mm/°C (10 in./in./°F) 100
75
4
mm
3
∆ T °C
(°F) 1800
950 50
900
2
–0.050
–1.00
–0.040
–0.80 –0.60 –0.50 –0.40
1600
–0.30
750
ex li
ne
–11
–9 –0.015
1300
–8 –7 –6 –5
–8 –7 Solution to sample problem
1400
1
–9
–10 –0.020
–0.20
–6 –3
Index lin
e2
–0.10
–4
–5
–0.08
1 650 1200
20
–12 –0.030
–0.005 700
25
–16 –15 –14 –13
–0.010
1500 800 Ind
D (in.)
–1.20
1700
850
∆C
–2 –0.04
600 0.75
–4
–0.06
1100
–3 –0.03 –0.001
550 15
1000 0.50
500
–2
900
–1 Sample problem
10
Given: D = 50 mm ∆ T = 680°C –6 (α – α ) = –5X = –5 x 10 mm/mm/°C 2
7
Solution:
1
∆
C = –0.17 mm D
0.25
Fig. 4.2
Nomograph for determining changes in diametral clearance caused by heating of dissimilar-metal joints. Source: Ref 7, 8
72 / Brazing, Second Edition
called heat treatable alloys. The heat treatable alloys generally contain magnesium and silicon. These alloys, because of their higher total alloy content, begin to melt at lower temperatures than higher-purity aluminum alloys and consequently are generally brazed at lower temperatures than the non-heat-treatable alloys. The heat treatable alloys, which include the commonly used 6061 alloy, are thermally treated during fabrication or after brazing to attain their high mechanical properties. Both groups of alloys lose strength at increasingly rapid rates as they are heated. At their respective annealing temperatures, they exhibit distinctly different behavior. The non-heattreatable alloys soften and return immediately to the O temper. The heat treatable alloys must be held at annealing temperature for a minimum of 20 min before an appreciable portion of their temper disappears. The heat treatable alloys can be returned to temper by heat treatment and aging. The non-heat-treatable alloys must be cold worked to be tempered. Aluminum is brazed at temperatures between 555 and 645 °C (1030 and 1195 °F). Obviously, a certain amount of annealing cannot be avoided during brazing. Aluminum alloys that have solidus temperatures above 590 °C (1090 °F) are easily brazed with commercial binary aluminum-silicon filler metals. Stronger, lower-melting alloys can be brazed with proper attention to filler-metal selection and temperature control, but the brazing cycle must be short to minimize penetration by the molten filler metal. High-quality castings are no more difficult to braze than equally massive wrought alloys. Aluminum sand and permanent mold casting alloys with high solidus temperatures are brazeable; brazeable aluminum casting alloys include 443.0, 356.0, 406, 710, 711, and 850.0. Alloys 443.0, 356.0, and 406 are used for both sand and permanent mold casting. Alloy A710.0 is primarily a sand casting metal. Alloys 711 and 850.0 are used for permanent mold casting. Alloys 443.0, 356.0, and 710.0 are the casting alloys most frequently brazed. Die casting alloys are difficult to braze; the castings are not easily wetted by the molten filler metal and tend to blister when brought to brazing temperature because of their high gas content and entrapped lubricants. Not all aluminum alloys can be brazed. The high-strength wrought aluminum alloys and certain casting alloys contain high amounts of alloying ingredients. These alloys have lower
melting temperatures than those of commercially available filler metals. Some experimental lower-melting filler metals have been produced that can be used for successful brazing of the not-readily-brazeable alloys 2024, 2219, and 7075. Low-melting filler metals contain zinc, copper, or other metal with or without silicon alloyed with aluminum. Although some of these combinations have been used successfully, none has been developed with adequate corrosion resistance to be recommended for commercial use. Therefore, the 2xxx and 7xxx series of aluminum alloys are not normally brazeable, with the exception of 7072 (used as a cladding material only) and 7005. All commercial filler metals for brazing of aluminum alloys are aluminum-base alloys themselves. These filler metals are available as wire or shim stock. A convenient method of preplacing filler metal is to use brazing sheet, which is an aluminum alloy base metal that is coated with a filler metal. Brazing sheet is coated on one or both sides. Core alloys 3003 and 6951 (a heat treatable alloy) are generally used. A third method of applying filler metal is to use a paste mixture of flux and filler-metal powder. Common aluminum filler metals contain silicon as the melting-point depressant with or without additions of zinc, copper, and magnesium. Commercial filler metals for brazing of aluminum are aluminum-silicon alloys containing 7 to 12% Si. Lower melting points are attained, with some sacrifice in resistance to corrosion, by adding copper and zinc. Filler metals for vacuum brazing of aluminum usually contain magnesium. More information on filler metals is presented in Chapter 5, “Brazing Filler Metals.” The optimal brazing temperature range for an aluminum-base filler metal is not only determined by the melting range of the filler metal and the amount of molten filler metal needed to fill the joint but is also limited by the mutual solubility of the filler metal and the base metal being brazed. The brazing temperature ranges of some filler metals are related to those of some base metals. Aluminum can be brazed by most of the standard practices. Most aluminum brazing is done by the torch, dip, or furnace process. Furnace brazing may be done in air or controlled atmosphere, including vacuum. Other methods, including induction, infrared, and resistance brazing, are used for specific applications, usually of the readily brazeable aluminum alloys. Regard-
Chapter 4: Base Metals and Base-Metal Family Groups / 73
less of the process, the temperature must be closely controlled for successful brazing. With dip or furnace brazing, automatic proportioning temperature-control devices are available that can maintain the flux bath within ±3 °C (±5 °F) and the furnace atmosphere within ±6 °C (±11 °F) of the desired brazing temperature. In manual torch, induction, or resistance brazing, operator skill and judgment are used to maintain the required temperature range for brazing based on flux color and on melting and flow of the filler metal. Automated torch and induction brazing utilize paste filler metals and preforms or wire with dispensable fluxes. On the increase is the use of vacuum furnace brazing for aluminum fabrication. Because it is done without flux, the joints are free from the corrosion problems commonly associated with residual or entrapped flux. Moreover, brazed assemblies containing inaccessible recesses can be fabricated efficiently. Furnaces operating in the 0.0013 Pa (10–5 torr) range are used. The success of the operation depends on the use of magnesium vapor as a getter of oxygen on the aluminum surface and magnesium alloyed in the filler-metal (aluminum-silicon) coating. Most furnace brazing of aluminum is done in air, but some benefits can be gained by brazing in a controlled atmosphere; by controlling moisture in the furnace at low levels, less flux is needed to achieve brazing (Nocolok process; Nocolok is a trademark of Solvay Fluor GmbH). New types of brazing sheet designed for fluxless brazing require nitrogen or other inert atmospheres. Furnaces operating at 0.0013 Pa have been used, and the success has stemmed from the use of magnesium as vapor, which acts as a getter of oxide on the aluminum brazing surface. Work has shown that magnesium in the parent aluminum alloy and in the brazing sheet can produce satisfactory brazed joints. Aluminum alloys are sometimes brazed to ferrous materials; in such cases, a nonoxidizing atmosphere is used to protect the ferrous material. Even though dip brazing of aluminum is still a popular and successful process, the search for improved methods of brazing complex aluminum structures, such as heat exchangers, that eliminate or minimize flux removal and corrosive residue problems has continued for many years. Aluminum alloys of the 1xxx, 3xxx, 5xxx, 6xxx, and 7xxx series can be vacuum brazed using No. 7, 8, 13, or 14 brazing sheet that is clad with 4004 filler metal. When additional
filler metal is required, 4004 in wire or sheet form can be introduced. The joint designs used for brazing with flux can also be used for fluxless vacuum brazing. The diffusion brazing process (Ref 14), which uses diffusion between the base metal and the filler metal, has been tried for joining aluminum-silicon alloy castings. The diffusion brazing process with a copper or brass preform, described in Ref 14, can apply for all hypoeutectic, eutectic, and hypereutectic alloys of aluminum-silicon system castings; the minimum temperature where the braze interface showed a liquid phase structure was 530 °C (990 °F) for the copper preform and 510 °C (950 °F) for the brass preform. The shear strength of the diffusion-brazed joint was dependent on the chemical compositions of the base metal, the type of material for the preform, and the brazing temperature and time. The maximum strength of the diffused-brazed joint under optimal conditions was 130 to 150 MPa (19 to 22 ksi) for the base metal of both Al-7Si and Al-12Si alloy castings and 100 to 130 MPa (15 to 19 ksi) for the base metal of Al-20Si alloy casting. The effect of brazing time on the strength of the brazed joint varied depending on the type of preform. The strength of the joint with the copper preform was determined by the characteristics of the braze interface structure. With the brass preform, the strength was controlled by the growth of silicon grains in the braze interface structure. Figure 4.3 shows the effect of brazing temperature and time on the strength of the diffusion-brazed joints of various aluminum-silicon system alloy castings. The 8019 (Al-8.3Fe-4.0Ce) aluminum alloy, a high-performance alloy, was weld brazed by researchers (Ref 15), and the work indicated that no thermal degradation was observed in alloy 8019 after 450 °C (840 °F) for up to 4 h; however, clear degradation could be found at 500 °C (930 °F) after only 100 s. The joining temperature for alloy 8019 was kept below 450 °C. Weld brazing is an innovative new approach that combines brazing with welding to provide synergistic benefits. The results of preliminary studies on 6061-T6 alloys showed that the technique could improve tensile-shear, tensile-fracture, and fatigue strengths. The new generation of aluminum alloys containing lithium exhibit improved specific stiffness and strength compared with conventional aluminum-copper (2000-series) and Al-Zn-Mg
74 / Brazing, Second Edition
(7000-series) alloys. Such alloys are therefore of particular interest to the aerospace industry, where weight reduction is a major requirement. The compositions of some important alloys are given in Table 4.1. High-strength diffusion-bonded joints have been obtained in aluminum-lithium alloy sheet using both solid-state and liquid phase tech-
30
200 Insert metal: Cu
100
500 °C 530 °C 540 °C
15
0 200 Insert metal: Cu - 30% Zn
500 °C 530 °C 540 °C
15
100
0
30
7 % Si 12 % Si 20 % Si
Shear strength, ksi
Shear strength, MPa
7 % Si 12 % Si 20 % Si
0
5
20
10 15 Brazing time, min
Fig. 4.3
Effects of brazing time and temperature on the shear strength of diffusion-brazed joints in relation to the types of preforms and aluminum-silicon alloy castings used
niques (Ref 16). The strengths are much greater than values for adhesive-bonded joints, and in the solid state, the bond microstructure and corrosion resistance are expected to be similar to those of the base metal. The highest shear and peel strengths and minimum variability have been obtained for joints fabricated under very clean conditions in vacuum, and diffusion bonding followed by superplastic forming has been demonstrated for aluminum-lithium 8090 alloy (Ref 16). In another study (Ref 17), diffusion bonds were produced between sheets of an Al-Li-CuMg-Zr alloy using Al-4%Cu vapor-deposited metallic interlayers. The joining was performed at a bonding temperature of 530 °C (990 °F), a pressure of 4 to 5.5 MPa (0.6 to 0.8 ksi), times in the range of 10 to 30 min, and a bonding atmosphere with partial oxygen pressure lower than 2 × 10–3 Pa (3 × 10–7 psi). Microstructural changes were analyzed that occurred both in the parent alloy and in the bond interface after diffusion-bonding cycles and postbonding heat treatments. For previous bonding conditions, elimination of the continuous interfacial oxide layer is possible, and only discrete oxide particles (probably lithium-rich spinels) were detected in the bond interface. This oxide elimination is accelerated if the bonding surfaces are chemically cleaned with a commercial deoxidant. In spite of the stability and high resistance to migration of the bonding interface, it is possible to obtain zones of local recrystallization, especially in superplastic joined sheets. It is probably due to mechanisms of local dynamic recrystallization associated with highly deformed zones.
Table 4.1 Compositions and densities of some important aluminum-lithium alloys Composition, wt% Production route
Ingot
Melt spun Powder atomized Mechanically alloyed Source: Ref 16
Alloy
Li
Cu
Mg
Si
Fe
8090 (U.K.) 8091 (U.K.) 2090 (U.S.) 2091 (France) Weldalite 049 (U.S.) 1420 (Russia) 678 (U.S.) . . . (U.S.) Cospray (U.K.) IN 905 XL (U.S.)
2.1–2.7 2.4–2.8 1.9–2.6 1.7–2.3 1.3 1.5–2.6 3.22 4.1 4.33 1.5
1.0–1.6 1.8–2.2 2.4–3.0 1.8–2.5 4.5–6.3 ... 0.54 ... ... ...
0.6–1.3 0.5–1.2 0.25 1.1–1.9 0.4 4.0–7.0 3.07 1.0 ... 4.0
0.2 0.3 0.1 0.2 ... ... ... ... ... ...
0.3 0.5 0.12 0.3 ... ... ... ... ... ...
Zr
0.04–0.16 0.08–0.16 0.08–0.15 0.04–0.10 0.14 0.05–0.3 0.83 0.2 0.11 ...
Density, 103 kg/m3
2.53 2.54 2.60 2.58 2.73 2.50 ... ... 2.41 2.58
Chapter 4: Base Metals and Base-Metal Family Groups / 75
The thermal bonding cycles applied to the alloys do not cause serious surface oxidation problems, although surface lithium depletion occurs by sublimation. In addition, the grain size of the superplastic alloy is kept in the range necessary for superplastic behavior. Diffusion-bonding trials were carried out using the same alloy (AA8090), both in the nonsuperplastic (T6) and superplastic conditions. The liquid phase bonding in air of unreinforced and SiC-fiber-reinforced aluminum using interlayers of copper-silver alloy was investigated (Ref 18). The bond strengths were measured using a simple shear jig, and the associated microstructures were characterized by electron microscopy and electron probe microanalysis. A shear strength of 65 MPa (9.4 ksi) was achieved when bonding unreinforced aluminum at 510 °C (950 °F) using a 50 µm thick alloy interlayer and a pressure of 10 MPa (1.5 ksi) for 30 min; the bonded region covered ~85% of the area of the joint. With reinforced aluminum, a bonding pressure of 20 MPa (3 ksi) was required to achieve sufficient contact and a similar area of bond; after application of pressure for 30 min at 510 °C, shear strength of 54 MPa (8 ksi) was developed at the joint. A mechanism for bonding that reflects diffusion behavior and interphase reactions is shown in Fig. 4.4. A novel brazing technique (a simplified and cost-effective method) using an alloy powder mixture instead of a clad surface has been developed for brazing aluminum, copper, and brass (Ref 19, 20). In these applications, at least one of the aluminum components is clad with filler metal consisting of an aluminum-silicon alloy of neareutectic composition, such as AA4045, AA4047, or AA4343 (Table 4.2) (Ref 21). These alloys contain 9 to 13 wt% Si and are characterized by a melting temperature (in a narrow range near 577 °C, or 1071 °F) (Ref 22) considerably lower than that of the core alloy
(~660 °C, or 1220 °F). Joining is carried out in the presence of a noncorrosive flux, such as a fluoroaluminate salt (Ref 23, 24), to remove native surface oxide films from the contacting aluminum surfaces. Oxide removal enhances wetting by the molten aluminum-silicon eutectic alloy at the brazing temperature and eases liquid metal penetration of the joint. The present brazing technique uses the eutectic bonding approach described in earlier publications (Ref 25, 26) but avoids the need to coat the base-metal surface with an intimately adhering layer of the eutectic-forming metal by electroplating or vacuum deposition. In the present technique, at least one of the aluminum surfaces
Al
Al FeAlx
Ag - Al Cu - Al Cu - Ag
Cu - Ag
Cu - Al Ag - Al Al 50 µm
(a)
Al
(b)
Al
Al
Al - Cu - Ag liquid
Al - Cu - Ag liquid
Ag2Al
Ag2Al
Al
(c)
Al
(d)
Al
Al
Al - Cu - Ag liquid
Al - Cu - Ag liquid
Ag2Al
Ag2Al
Al
(e)
Al
(f)
Fig. 4.4
Proposed mechanism for aluminum bonded with copper-silver alloy interlayer. (a) Start of process, 490 °C (915 °F). (b) Solid-state interdiffusion, 0 to 5 min. (c) Formation of Al-Cu-Ag liquid, 5 min, 502 °C (935 °F). (d) Expulsion of liquid, 5 to 10 min, 510 °C (950 °F). (e) Further reaction, 5 to 15 min, 510 °C (950 °F). (f ) Isothermal solidification, >15 min, 510 °C (950 °F)
Table 4.2 Composition limits for selected aluminum alloys Content, wt% Aluminum alloy
1050 1100 3003 3102 4045 6061 X-800
Si
Fe
Cu
Mn
Mg
Zn
Cr
Ti
Others
0.25
0.4
0.05 0.05–0.2 0.05–0.2 0.1 0.3 0.15–0.4 0.31
0.05 0.05 .1.0–1.5 0.05–0.4 0.05 0.15 1.11
0.05 ... ... ... 0.05 0.8–1.2 0.27
0.05 0.1 0.1 0.3 0.1 0.25 ...
... ... ... ... ... 0.04–0.35 ...
0.03 ... ... 0.1 0.2 0.15 0.008
... 0.15 0.15 0.15 0.15 0.15 0.15
0.95 Si+Fe 0.6 0.4 9–11 0.4–0.8 0.05
0.7 0.7 0.8 0.7 0.19
76 / Brazing, Second Edition
is coated with a thin layer of a powder mix consisting of an element capable of forming a lowtemperature eutectic with aluminum (e.g., silicon, copper, germanium, zinc) and a flux capable of dissolving surface oxide films (Ref 27), as illustrated in Fig. 4.5(a). A commonly available noncorrosive flux (Ref 24) was used in the present work. This flux consists of a mixture of KAlF4 and K2AlF5 · H2O powders in a molar ratio of the respective salts of approximately 13 to 1 (Ref 28), with a particle dimension of the order of 1 µm. Brazing is carried out by heating the joint at approximately 600 °C (1110 °F) in nitrogen gas at near-atmospheric pressure for a few minutes. During temperature ramp-up, the flux melts at ~562 °C (1044 °F) and dissolves the surface oxide layers on the aluminum, as illustrated in Fig. 4.5(b). Sufficient flux must always be present to remove these oxides. Oxide dissolution must occur more rapidly than reoxidation of the aluminum surface, allowing the silicon particles to come into intimate contact with the bare metal. At this juncture, the large, elemental concentration gradients at the aluminum-silicon
Fig. 4.5
interface cause the aluminum and silicon to interdiffuse (Fig. 4.5c). At temperatures exceeding 577 °C (1071 °F), it is found that the silicon particles diffuse rapidly into the aluminum surface and generate in situ a layer of aluminum-silicon liquid alloy of near-eutectic composition (Fig. 4.5d). The filler metal penetrates the joint of interest by capillary action and forms a fillet, thus producing a metallurgical bond on cooling. Any unused filler metal remains on the aluminum surface to form a layer of aluminum-silicon alloy of near-eutectic composition (Fig. 4.5e). In this process, only one of the joined surfaces need be covered with the silicon-flux mix, because the molten flux spreads rapidly across the joint to remove surface oxide films from the mating surfaces. The quantity of filler material formed from one coated surface is generally sufficient to yield a good metallurgical bond. Because fillets are formed through capillary flow of the filler metal, brazing requires only minimal contact force at the joint interface (Ref 29). In recent years, the use of rapid-solidification powder metallurgy has made it possible to
Successive steps in the novel brazing process. (a) Deposition of a silicon-flux powder mix on the aluminum surface. The silicon particle dimensions range from ~1 to 100 µm; the flux particle dimensions do not exceed 1 µm. (b) Melting of the flux at 562 °C (1044 °F) and dissolution of surface oxide films. (c) At 562 °C
577 °C (1070 °F), rapid dissolution of silicon to form localized pools of filler metal of neareutectic composition, followed by coalescence of liquid metal pools. (e) End of filler-metal generation and solidification
Chapter 4: Base Metals and Base-Metal Family Groups / 77
develop a new family of aluminum alloys exhibiting unique properties. One of these materials, dispersion-strengthened aluminum, derives its high strength from nanoscale AlN particles embedded in an aluminum matrix. Unlike precipitate particles in conventional aluminum alloys, these ceramic particles resist dissolution and/or coarsening during thermal treatment; therefore, the material is capable of maintaining much of its strength at high temperatures (Ref 30–32). Due to these properties, dispersionstrengthened aluminum is expected to be well suited as construction material for high-temperature applications where weight reductions are of particular concern. Two different brazing filler metals, based on the aluminum-silicon system, were investigated with regard to the brazeability of dispersionstrengthened aluminum (Ref 33). One of the filler metals was a commercial type with eutectic aluminum-silicon composition, whereas the other one was especially made for the present investigation (Table 4.3). The alloy was made with an addition of 1% Mg, which has been reported in numerous papers to have a beneficial effect on the wetting behavior of the filler metals in vacuum (Ref 34–38). During brazing, magnesium evaporates and has a gettering effect on the residual gas contaminants in the furnace atmosphere. This reduces the oxidation of aluminum and simultaneously breaks up the surface oxide layer through precipitation of indigenous MgO crystallites. The results on the wetting behavior of dispersion-strengthened aluminum under conditions applicable to brazing showed that a eutectic aluminum-silicon filler metal completely wets the base metal both under high-vacuum conditions and in controlled argon atmospheres, provided that the partial pressure of oxygen is sufficiently low. The main problem appears to be the stability of the matrix grain structure. In general, the
process of grain erosion and coarsening can be controlled by restricting the supply of the filler metal so that only a small metal volume is exposed to erosion. In addition, there is a great potential for reducing the thermodynamic driving force of the erosion reaction by proper adjustments of the filler-metal composition and/ or the brazing temperature (Ref 33). Still, grainboundary liquidation may be a problem, which, in turn, may require additions of surface-active elements to the filler metal to control the wetting behavior (Ref 39).
Beryllium and Beryllium Alloys Joint design, filler-metal selection, and a choice of brazing cycle are complicated by the low ductility of beryllium as well as its reactivity with many of the filler metals used for brazing. The low ductility of beryllium is aggravated by the presence of structural discontinuities, such as surface scratches, notches, and asymmetrical stress patterns produced by single-lapped joints. Beryllium parts must be handled carefully, and joint configurations such as butt, scarf, step, and double-lapped joints should be considered when beryllium structures are designed. Because beryllium reacts with the constituents of most filler metals, brazing should be done under conditions that minimize the formation of intermetallic compounds: rapid heating and cooling cycles, low brazing temperatures, minimum time at brazing temperature, and minimum amounts of filler metal. However, these procedures are not always practical, because many filler metals do not wet or flow well on beryllium surfaces. As a result, filler metal frequently must be preplaced between the joint members in amounts sufficient to produce the brazed joint. Longer brazing times than desired may be needed to ensure wetting and flow of the
Table 4.3 Compositions of filler metals investigated for brazing of dispersion-strengthened aluminum Content, wt% Filler metal
No. 1 No. 2 Source: Ref 33
Si
Mg
Ti
Br
Mn
Zn
Cu
Fe
Al
11–13 11
0.03 1
... 0.2
... 0.04
0.07–0.2 ...
0.05 ...
0.05 ...
0.4 ...
bal bal
78 / Brazing, Second Edition
filler metal; under such conditions, the diffusion that occurs between the base metal and the filler metal may tend to adversely affect the joint properties. Brazing is the preferred method for metallurgically joining beryllium. From a stiffness/ weight standpoint, beryllium is unexcelled by any bulk metal or alloy in existence. However, due to its chemical and metallurgical reactivity, braze-joining techniques must be highly specialized, and suitable filler-metal systems and brazing temperature ranges include: • Zinc (427 to 454 °C, or 801 to 849 °F) • Aluminum-silicon (566 to 605 °C, or 1051 to 1120 °F) • Aluminum (645 to 655 °C, or 1195 to 1210 °F) • Silver-copper (640 to 904 °C, or 1185 to 1659 °F) • Silver (882 to 954 °C, or 1620 to 1749 °F) Strictly speaking, zinc, with a melting point of 420 °C (788 °F), does not quite meet the accepted definition for liquidus temperature for a filler metal (425 °C, or 800 °F). Nevertheless, it is generally accepted as the lowest-melting filler metal for brazing of beryllium. Using rather specialized techniques, zinc brazing of beryllium has found important applications. Its successful use was demonstrated in production joining of over 2000 fuselage longeron components for the beryllium spacer for an intercontinental ballistic missile. Brazing was selected as the joining process for the longeron assembly because the brazed design had the lowest cost, the least weight, and it minimized clearance problems in the assembly (Ref 40). Zinc was selected as the filler metal because it caused no detrimental reaction with the beryllium base material. The dip brazing of beryllium with zinc filler metal offered the following advantages: • Zinc has no undesirable reaction with beryllium. • The brazing temperature was low enough that no recrystallization effects took place in beryllium sheet. • Shear strength at room temperature was consistent and high (128 MPa, or 19 ksi). • Shear strength at room temperature decreased linearly to a value of approximately 17 MPa
(2.5 ksi) at 425 °C (800 °F), indicating moderate temperature applications. • Sophisticated equipment was not required. Aluminum and aluminum-silicon filler metals that contain 7 to 12% Si can provide highstrength brazed joints for service at temperatures up to 150 °C (300 °F). These filler metals have been used widely in high-strength wrought beryllium assemblies, because joining is performed well below the base-metal recrystallization temperature. Fluxless brazing requires stringent processing but has been used effectively. A significant advantage of aluminumbase versus silver-base filler metals is that metallurgical interaction is minimal. This is of prime concern when thin beryllium sections or foils are to be joined. When beryllium assemblies are brazed using the aluminum-silicon-type filler metal, by virtue of the low furnace brazing temperature (605 °C, or 1120 °F), the wrought sheet beryllium base properties do not deteriorate but maintain the desirable isotropic mechanical properties of beryllium. Silver and silver-base filler metals, such as Si-7Cu-0.2Li, find use particularly in structures exposed to elevated temperatures. An added advantage with these systems is that atmosphere brazing is used and may be performed in inertgas atmospheres or vacuum. The lithium is added to improve wettability. Silver interacts favorably with beryllium, in that hard or brittle intermetallics do not remain on cooling. Brazing temperatures vary, depending on the mass of the parts to be joined, joint precision, and pressure applied to the components. Both the silver-base and aluminum-silicon filler metals, however, exhibit poor capillary flow, and preplacement is recommended. The use of the 7% Cu-0.2%Li sterling silver filler metal markedly improves the ability of the filler metal to make well-defined fillets and to bridge gaps of several mils. Copper beryllides, which generally form at the faying surfaces, inhibit rapid penetration of silver into the beryllium components, and, hence, a more stable liquid phase is present at the brazing temperature. Properly executed beryllium brazements may be made to exhibit useful strengths at temperatures up to approximately 870 °C (1600 °F) (Ref 40, 41). Research on brazing beryllium to a CuCrZr vabotron element (Ref 42) found that beryllium forms brittle intermetallic compounds with most
Chapter 4: Base Metals and Base-Metal Family Groups / 79
other metals; however, beryllium and copper form a eutectic at ~850 °C (1560 °F), and beryllium recrystallizes above 730 °C (1350 °F). The alloy CuCrZr is precipitation hardenable, requiring annealing (>900 °C, or 1650 °F) and hardening at ~500 °C (930 °F) for optimal strength, while beryllium forms an oxide layer on its surface. The use of BAg-18 (Ag-30Cu-10Sn) as a filler metal was successful in joining the beryllium component at 650 to 680 °C (1200 to 1255 °F) for 10 to 45 min under a pressure of ~0.3 MPa (0.04 ksi). Also used was the Incusil active brazing alloy (ABA) filler metal (Wesgo Metals) with vacuum induction brazing at 720 °C (1330 °F) for 30 s to 3 min at a pressure of 0.4 MPa (0.06 ksi).
Copper and Copper Alloys The copper alloy base metals include copperzinc alloys (brass), copper-silicon alloys (silicon bronze), copper-aluminum alloys (aluminum bronze), copper-tin alloys (phosphor bronze), copper-nickel alloys, and several others. The previously mentioned alloys are commonly brazed with copper- and silver-base filler metals (Table 4.4). The brazeability of copper and its alloys is generally rated from good to excellent. With
some alloys, however, difficulties may be encountered. For example, some lead-containing alloys can form dross that interferes with wetting, and tin-containing alloys, if not stress relieved before brazing, may crack when subjected to rapid localized heating. Softening of the base metal frequently occurs during brazing, because many copper-base alloys derive their properties as a result of heat treatment at relatively low temperatures, cold work, or both. The degree of softening increases with higher temperatures and longer times of exposure to elevated temperatures. Softening of areas in proximity to the braze can be minimized by cooling the assembly, except for the area to be brazed, by immersion in water, packing with wet rags or wet asbestos, or otherwise providing a heat sink to keep the overall temperature of the part as low as possible. In all cases, brazing with a low-melting filler metal for a minimum time reduces softening. The importance of uniform and controlled heating cannot be overemphasized, especially when dealing with brasses, cold-worked phosphor bronzes, and cold-worked silicon bronzes. They are especially susceptible to cracking. Tough pitch coppers are subject to embrittlement when heated in reducing atmospheres containing hydrogen. Consequently, although tough pitch coppers are generally rated as hav-
Table 4.4 Filler metals and other parameters for brazing of copper alloys Base material
Coppers C10100 to 14200 High coppers C15000 to 19400 Red brasses C20500 to 24000 Yellow brasses C25000 to 29800 Leaded brasses C31000 to 38500 Tin brasses C40500 to 45500 Phosphor bronzes C50200 to 52900 Aluminum bronzes C60600 to 64200 Silicon bronzes C64700 to 66100 Copper nickel C70100 to 82000 Nickel silver C73200 to 79800
Atmosphere
Flux
BCuP-2, BCuP-5 BCuZn; BAg-1, -1a, -2, -5, -6, -18 BAg-8, -1
Filler metal
None 1, 2, 5
None FB3A, C, D, E, I, J FB3A
BAg-1, -1a, -2, -5, -6; BCuP-5, -3; BCuZn BCuP-3, -4, -5; BAg-1, -1a, -5, -6 BAg-1, -1a, -2, -7, -18; BCuP-5 BAg-1, -1a, -2, -5, -6; BCuP-5, -3 BAg-1, -1a, -2 BCuP-3, -5; BAg-5, -6 BAg-1, -1a, -2, -3
1, 2, 5
BAg-1, -1a, -2 BAg-1, -1a, -2, -5, -18; BCuP-3, -5 BAg-1, -1a, -2, -5, -6; BCuP-3, -5
...
Notes
Do not braze oxygen-containing alloys in hydrogen atmospheres.
3, 4, 5
FB3 A, C, D, E, I, J FB3A, C, E
3, 4, 5
FB3A, C, E
3, 4, 5
FB3A, C, E
None 1, 2, 5 4, 5
None FB3A, C, E FB4A
Stress relieve prior to brazing
4, 5
FB3A, C, E
1, 2, 5
FB3A, C, E
Clean surface well. Stress relieve prior to brazing Stress relieve prior to brazing
3, 4, 5
FB3A, C, E,
Stress relieve prior to brazing
Braze cycle should be short to avoid softening Short braze cycle. Stress relieve prior to brazing ...
Apply flux with furnace brazing
Note: For information about brazing copper alloys under Atmosphere, see Table 3.4; under fluxes, see Chapter 6, “Fluxes and Atmospheres.”
80 / Brazing, Second Edition
ing good to excellent brazeability, they should not be brazed in a furnace that contains hydrogen with a reducing potential, such as dissociated ammonia, or in an exothermic-based or endothermic-based atmosphere. Heating by open flame or by torch also may result in hydrogen diffusion and embrittlement. Phosphorus-deoxidized and oxygen-free coppers can be brazed without flux in hydrogencontaining atmospheres without risk of embrittlement, provided that self-fluxing filler metals (copper-phosphorus) are used. The use of flux is required, however, when silver-base filler metals that contain additives such as zinc, cadmium, or lithium are used to braze these coppers to each other or to copper alloys or other metals. The coppers, including those that contain small additions of silver, lead, tellurium, selenium, or sulfur (generally no more than 1%), are readily brazed with the self-fluxing copperphosphorus filler metals, but wetting action is improved when a flux is used and when a sliding motion between components is provided while the filler metal is molten. Another copper group includes those with enhanced mechanical properties due to the addition of small amounts of alloying elements. These precipitation-hardenable copper alloys contain beryllium, chromium, or zirconium and form oxide films that impede the flow of filler metal. To ensure proper wetting action of the joint surface by the filler metal, beryllium-copper alloys should be freshly machined or mechanically abraded before being brazed. Removal of beryllium oxide from joint surfaces requires the use of a high-fluoride-content flux. Brazing precipitation-hardenable coppers in the aged condition reduces their mechanical properties. The high-strength beryllium-copper alloys (2% Be) can be furnace brazed and simultaneously solution treated at 790 °C (1455 °F). The temperature is lowered to 760 °C (1400 °F) to solidify the filler metal, and then the assembly is rapidly quenched in water and aged at 315 to 345 °C (600 to 655 °F), which develops adequate hardness in the base material. The 72Ag28Cu eutectic filler metal is generally used with flux. When the sections to be brazed are thin and can be cooled very rapidly, solution heat treated beryllium-copper may be brazed in the temperature range of 620 to 650 °C (1150 to 1200 °F). Other beryllium-copper alloys (0.5% Be) can be silver brazed rapidly with filler metals containing 45 to 50% Ag, 15% Cu, 16% Zn, and 18 to 24% Cd.
The chromium-coppers are brazed with silver-base filler metal and fluoride flux and simultaneously solution treated at 900 to 1010 °C (1650 to 1850 °F) and cold worked; aging can take place as a subsequent operation to develop improved mechanical properties. This is not the case with zirconium coppers. They do not precipitation harden without the benefit of prior cold working, a sequence that is incompatible with brazing. In the absence of cold working, the strength of zirconium coppers is not improved by aging treatments. Brasses. The copper-zinc alloys (brasses) are produced with varying ratios of the two elements to provide desired properties and casting characteristics. Other elements occasionally are added to enhance particular mechanical or corrosion properties; these special brasses are identified by the added element—for instance, lead, which gives leaded brass. Additions of manganese, tin, iron, silicon, lead, and aluminum, either singly or collectively, rarely exceed 4%. Red brasses (low) contain up to 20% Zn and are readily brazed with a variety of filler metals. Flux is normally required for best results, especially when the zinc content is above 15%. All the brasses can be brazed with the silver and copper-phosphorus filler metals, and the highermelting-point (low-zinc) brasses can also be brazed with the copper-zinc filler metal. It is recommended that protective atmospheres be used in furnace brazing of brasses; however, even in a protective atmosphere, flux should be used to promote good wetting by the filler metal and to reduce zinc fuming. When heated above 400 °C (750 °F), brass tends to lose zinc by vaporization. This loss can be reduced by fluxing the parts during furnace brazing or by using an oxidizing flame during torch brazing. Brasses are also subject to cracking and should therefore be heated carefully and uniformly. Yellow brasses (high) contain 25 to 40% Zn and are readily brazed, but low-melting filler metals should be used to avoid dezincification of the base metal. Leaded brasses are formed when lead is added to red or yellow brass in amounts up to 5%. Alloys containing more than 5% Pb are usually not brazed. The lead forms dross on heating that can seriously impede wetting and the flow of the filler metal. Consequently, in brazing of leaded brasses, the use of a flux is mandatory to prevent dross formation in the joint area. Naval brass, leaded naval brass, and admiralty brass contain up to 1% Sn and may contain other
Chapter 4: Base Metals and Base-Metal Family Groups / 81
alloying elements such as lead, manganese, arsenic, nickel, and aluminum. Except for the aluminum-containing alloys, these brasses are readily brazed; they have greater resistance to thermal shock and are less susceptible to hot cracking than the high-lead brasses. For proper wetting, brasses that contain aluminum require a special flux. Bronzes. Alloys of copper and tin are properly termed tin bronzes. However, these coppertin alloys contain small amounts of phosphorus (up to 0.25%) added as a deoxidizer. The tin bronzes have come to be known commercially as phosphor bronzes. Although susceptible to hot cracking in the cold-worked condition, phosphor bronzes have good brazeability and are adaptable to brazing with any of the common filler metals that have melting temperatures lower than that of the base metal. To avoid cracking, phosphor bronzes should be stress relieved at 290 to 345 °C (555 to 655 °F) before brazing. After this stress relief or anneal, the parts should be supported in a stress-free condition during brazing, and slow heating cycles should be used to avoid thermal shock. When the tin content is high or when there are appreciable lead additives, adequate flux protection during brazing is necessary. All the phosphor bronzes can be brazed with the Ag-Cu-Zn-Cd and copper-phosphorus filler metals. The copper-zinc filler metals are appropriate for brazing the low-tin varieties. Aluminum bronzes are copper-aluminum alloys with high copper contents and 3 to 13% Al, with or without varying amounts of iron, nickel, manganese, and silicon. Aluminum bronzes are generally considered difficult to braze because of their aluminum content, which results in the formation of refractory aluminum oxide at brazing temperature in alloys containing more than 8% Al. However, alloys containing 8% Al or less are brazeable, provided that appropriate fluxes are used to dissolve the aluminum oxide. The oxide, which inhibits the flow of filler metal, cannot be reduced in dry hydrogen. Electroplating of copper on the surface to be brazed is one technique used to enhance wetting of the low-melting silver filler metal, while use of flux and a protective atmosphere in a furnace is a technique employed if unplated surfaces are used. Silicon bronzes that contain up to approximately 3.25% Si and are in a highly stressed condition are susceptible to hot shortness and
stress cracking by molten filler metal and should be well fixtured during heating and brazing to prevent excessive stressing. To avoid cracking, these alloys should be stress relieved at approximately 290 to 345 °C (555 to 655 °F) before brazing. For best brazing results, joint surfaces of copper-silicon alloys should be freshly machined, cleaned, and flux coated or copper plated before brazing to prevent the formation of refractory silicon oxide. Mechanical cleaning is recommended, or, for light oxide, the material can be pickled. Silver-bearing filler metals and flux generally are used. Other Copper Alloys. Commercially available copper nickels may contain from 5 to 40% Ni and are susceptible to both hot cracking and stress cracking by molten filler metal. To prevent cracking, copper-nickel alloys should be stress relieved before brazing. Stresses should not be introduced during brazing. The Ag-CuZn-Cd filler metals and flux are preferred for brazing the copper nickels. Nickel silvers, which are brasses (Cu-Zn-Ni) that contain up to approximately 20% Ni but do not contain silver, are highly susceptible to hot cracking and should be stress relieved at approximately 290 °C (555 °F) before being brazed. The nickel silvers can be brazed readily with the same procedures used for brazing brass. When copper-zinc filler metals are used, however, care is required because of the relatively high brazing temperatures. Brazing Processes for Copper Alloys. Typical brazing processes used for copper and its alloys include torch, furnace, and induction. The advantages of furnace brazing that are applicable to the joining of copper and copper alloys relate to the furnace as a source of heat and to the cooling chamber that is provided on conveyor-belt furnaces as a means for cooling assemblies from the brazing temperature to 150 °C (300 °F) or below. To a lesser degree, the protective atmospheres used most conveniently in furnace brazing—notably, the exothermic-based and endothermic-based atmospheres—constitute another advantage of brazing deoxidized coppers and copper alloys in a furnace. For copper alloys susceptible to hot cracking, control of cooling rate in the cooling chambers of brazing furnaces is as important as control of the rate of heating to the brazing temperature. Torch brazing of copper and its alloys is selected largely on the basis of feasibility and cost. Torch brazing is often used when workpiece limitations preclude the use of alternative
82 / Brazing, Second Edition
brazing processes. Low equipment cost is a major advantage in manual torch brazing, which is particularly useful for assemblies involving unequal masses. A brazer with moderate skill can adjust and apply the heating flame so that unequal masses are brought uniformly to brazing temperature. The brazer can also apply the heat selectively to the joints of assemblies involving both large and small areas. The size of brazing flames may range from those of extremely small torches the size of a hypodermic needle used for electronic leads up to those of large torches used in brazing assemblies weighing hundreds of kilograms. Precautions should be taken in torch brazing certain coppers and copper alloys. Where it is necessary to braze oxide-containing coppers, a reducing atmosphere in the flame must be avoided, because it can promote hydrogen embrittlement. For these coppers, a neutral or slightly oxidizing flame and a short brazing cycle are necessary. Brasses are subject to volatilization of zinc when overheated or when held too long at brazing temperature. Application of flux suppresses zinc volatilization. Alloys containing elements that readily form refractory oxides (aluminum, beryllium, chromium, and silicon) must be protected by flux and should not be exposed to an oxidizing flame. The efficiency of heating by induction varies directly with the electrical resistivity of the alloy. Brass, because it has higher electrical resistivity, can be heated more efficiently than copper; steel, which has even higher resistivity, can be heated more efficiently than brass. In terms of the high-frequency power input and the time required to heat 0.45 kg (1 lb) of metal in a joint assembly at a brazing temperature of 705 °C (1300 °F) and a power input (450 kHz) of 15 kW, steel required 16 s, whereas brass required 30 s and copper approximately 55 s. The main advantages of induction brazing of copper and its alloys over torch brazing are minimized warpage, reduced postbraze cleaning, and less required operator skill. Induction heating is suited for mass producing brazed assemblies, primarily because inductors can be designed that heat a line of assemblies as they are carried through the induction field by conveyor belt or turntable. One of the general limitations of induction brazing is the cost of induction heating equipment, which far exceeds the cost of torch brazing equipment and usually exceeds the cost of equipment for resistance brazing or
dip brazing in molten salt. Because the efficiency of heating copper and copper alloys by induction is generally low, costs to achieve given production rates are high. The low longterm operating and maintenance costs of induction units, however, may outweigh the higher initial costs. Other general limitations of induction brazing may relate to the size and shape of assemblies that can be brazed, the design of inductors (cooling coils used to convey heat to the assembly), and the requirements for matching impedances that apply to brazing of copper and copper alloys. In a recently completed study (Ref 43) on the use of advanced particulate-reinforced composites and dispersion-hardened materials as candidate materials to replace pure annealed copper for the electromagnetic windings of the compact ignition tokamak nuclear fusion reactor, it was found that direct brazing of such materials may lead to brittle joints and that indirect brazing may need to be employed for joining such materials. As a result, 0.25 vol% alumina dispersionhardened copper (ADHC) was brazed satisfactorily using induction brazing in an argon atmosphere using 72Ag-28Cu eutectic and pure silver electroplate as interlayer materials. Both interlayers completely wet ADHC. For both interlayers, the application of additional stress (20 MPa, or 3 ksi) on the joint improved the joint quality. By plating a 17.5 µm layer of copper on the ADHC prior to brazing, other researchers have obtained shear strengths as high as 223 MPa (32 ksi). Resistance brazing is often used in joining copper conductors, terminals, and other parts in lap joints for electrical connections where heating must be localized and closely controlled during brazing and where the brazed joint must have low electrical resistance. Generation of heat in the filler metal and nearly complete filling of the joint with a thin layer of the filler metal help meet all these objectives. Normally, the filler metal most frequently used is copperphosphorus, which is used without a flux— especially in brazing of copper. The success of resistance brazing for joining small copper electrical conductors to massive copper assemblies has led to the use of the process for attaching armature leads to commutator bars on large electric motors and generators. Joints made by this method provide a large conducting cross section that prevents signifi-
Chapter 4: Base Metals and Base-Metal Family Groups / 83
cant resistive heating at the connections in service. The joints are made at a much lower temperature than would be possible by resistance welding. The use of conventional resistance welding equipment and high-resistivity electrodes with specially contoured tips makes it possible to concentrate the heating at the joint, to keep the heating time to a minimum, and to obtain efficient handling and comparatively high production rates. Another use of portable resistance welding machines for resistance brazing is in attaching copper bus-bar terminals or similar strip connectors to large electrical equipment that cannot be brought to, or positioned for brazing in, a conventional fixed-position resistance welding machine. Electrical connections to such equipment can often be made more economically by resistance brazing than by mechanical means and are made more readily by resistance brazing than by arc welding. Resistance brazing done with portable resistance welding machines is a convenient and economical way of interconnecting large copper electrical bus bars or of attaching either large or small copper bus bars to motor generators, transformers, and other electrical equipment. The low melting temperature of the filler metal helps to avoid overheating and excessive annealing of the work, and the usual selection of self-fluxing copper-phosphorus filler metals avoids corrosion problems and the need for flux removal. In some applications, filler metal for resistance brazing can be provided in the form of a coating already present on one or both members to be joined, eliminating not only the use of flux but also the operation of placing filler metal at the joint. This is done when copper wire is joined by high-speed resistance brazing to copper terminals clad with copper-phosphorus filler metal. Salt bath furnaces have been used for brazing of copper and copper alloys with silver filler metals. The same neutral salts, operating temperatures, and brazing procedures used for steel are used for dip brazing of most copper alloys. Applications of dip brazing of copper alloys in molten salt include waveguides and waveguide hardware, flowmeter hardware, and capillary tube and bellows assemblies. Copper, Other Metals, and Composites. Copper-steel and steel sections of engine systems produced by high-temperature brazing are
large, two-layer structures consisting of a ribbed wall and a smooth external jacket. They have the form of rotating bodies of complicated geometrical form with cooling circuits operating at high pressures. Therefore, special requirements on strength, corrosion resistance, and efficiency at high and low temperatures are imposed on brazed joints. The design of sections requires the production of brazed joints in combinations of copper and creep-resisting alloys with steels. In producing brazed sections, the filler metals are represented mainly by specially developed copper-manganese alloys of the PM17 type and nickel-manganese alloys of the G70NKh, G40NKh, and so on types, produced in the form of strips and foils 0.1 to 0.3 mm (0.004 to 0.012 in.) thick. Their use ensures strength characteristics and corrosion resistance of brazed joints close to those of the parent material (Ref 44). High-temperature brazing of sections is carried out in vacuum compression systems (VCSs) using induction heating, with automatic control of the temperature-time parameters of brazing and rotation of the brazed section around the horizontal axis. Figure 4.6 shows the diagram of a VCS for brazing sections of engine systems in the horizontal position, with rotation and compression of the surfaces of components during heating for brazing. The system consists of a brazing chamber with an induction heater, a system for loading and unloading sections, a vacuum system, and a system for forced pressurizing and
Fig. 4.6 position
Diagram of a vacuum compression system for brazing sections of engine systems in the horizontal
84 / Brazing, Second Edition
cooling. Brazing of engine system sections is carried out at 1000 to 1200 °C (1830 to 2190 °F) (depending on the grade of the alloy used) and with the pressure of the shielding gas up to 490 kPa (71 psi). However, when producing brazed sections with a small cross section of the channels, the use of filler metals in the form of strips, even 0.005 mm (0.0002 in.) thick, often leads to clogging of the channels and unstable hydraulic resistance. In addition, the technology of depositing filler metals in the form of strips on the brazed surfaces by resistance welding is very complicated, laborious, and not always efficient. Therefore, it is more efficient to deposit metallic coatings, which are components of the filler metals, on the surfaces to be brazed. In brazing, as a result of contact interaction of the metal of the coating with the parent material, a liquid phase with a lower melting point than in the brazed material forms and plays the role of the filler metal. This brazing method, the so-called resistance-reactive brazing, is used widely for producing sections where, in brazing, it is necessary to ensure accurate amounts of the filler metal to obtain brazed joints of the required quality. Metallic coatings can be deposited by electroplating or thermovacuum spraying on the surfaces of the brazed components. When brazing copper-steel sections of the engine systems, the main component of the filler metal is manganese, which can also be deposited on the ribbed surface of the copper wall. The optimal conditions of brazing coppersteel sections through a manganese coating include: thickness of the deposited coating layer, 12 to 15 µm; brazing temperature, 1010 ± 10 °C (1850 ± 18 °F); holding time, 10 min; and residual pressure, 6.65 Pa (10–3 psi). The tensile strength of the brazed joints is 200 to 225 MPa (29 to 33 ksi). X-ray spectrum microanalysis of the distribution of elements in the brazed joint shows that the composition of the brazed joint is identical with that of brazed joints obtained when brazing copper-steel sections with copper-manganese filler metals in compact form (strips, foils). Steel sections of engine systems are brazed with alloys where the main components are nickel and manganese, which are deposited in layers on one of the components to be brazed. In heating for brazing, manganese and nickel and melted by resistance-reactive heating, and a liquid phase forms.
The optimal brazing conditions of steel sections using a nickel-manganese coating as the filler metal include: thickness of the layer of the nickel coating deposited on each surface to be brazed, 10 to 15 µm; thickness of the layer of the manganese coating deposited on one of the brazed surfaces, 25 to 30 µm; brazing temperature, 1200 ± 10 °C (2190 ± 18 °F); holding time, 10 min; and residual pressure, 6.65 Pa (10–3 psi). The tensile strength of the brazed joints is 580 to 700 MPa (84 to 102 ksi). Locomotives used in the channel tunnel connecting Britain and France have a rotor made of parallel copper bars held together by an end ring (Fig. 4.7). This ring, which is made of Cr-Cu-Zr material, is brazed at each end to the copper bars. This braze must withstand the stress of rotor speeds up to 6000 rpm. It was essential to develop a brazement that retains a high strength during the brazing process. Flame brazing was the method chosen, because it provided a uniform heat over and around irregular shapes. It was also thought to be better for odd joint configurations. The gas chosen was an air-acetylene mixture. This mixture gave the desired flame characteristics for wraparound, depth of heating, and rate of heating. The number of heat devices varied depending on the size of the rotor being brazed; one setup had 42 and another had 64. Uniformity of heating was attained by rotating the rotor assembly. The part of the rotor that fit into the end ring was designed with a series of rectangular cavities adjacent to the joint area where the filler
Fig. 4.7
Rotors made of copper bars
Chapter 4: Base Metals and Base-Metal Family Groups / 85
metal and flux could be placed. The filler metal was a silver alloy that melted at 620 °C (1150 °F). The foil form was cut into small segments and placed at the joints. The flux was hand brushed onto the braze area and filler metal. Four pilot burners were used to ignite the flames in four quadrants around the ring. Brazing time, on average, was 21/2 to 41/2 min, with the unit rotating at 1 rpm. The rotors were quenched in cold water after brazing. Hardness before brazing was 148 to 154 HV. After brazing, it was 135–143, which was within acceptable limits (Ref 45). The first commercially available brazed copper-brass car and truck radiators have made their debut in the global automotive market. The design innovations and technological advances of copper-brass (CuproBraze) radiators have shown that these radiators are 35 to 40% lower in weight compared to traditional (nonoptimized) copper-brass radiators, because they are designed and manufactured with far less materials in their fins and tubes. Also, the heavy lead-base solder traditionally used in copper-brass radiators is replaced with a lightweight filler metal. Copper-brass radiators can be designed and manufactured to suit the diverse cooling requirements of the automakers of the world. They can be produced using the same furnaces as employed to braze aluminum radiators. Other advantages of copper-brass radiators include the inherent superiority of copper over aluminum in thermal conductivity, internal corrosion resistance, durability, ease of fabrication, longer manufacturing tool life, and simpler joining techniques. In laboratory performance tests, copper-brass radiator prototypes have lasted 8000 h without failure. This equals more than 800,000 km (500,000 miles) of service (Ref 46). Chief among the technologies that provide advantages in the manufacture of copper-brass radiators are no-flux brazing and electrophoretic coating. Newly designed materials and fluxless brazing of radiator components develop superior strength in fin, tube, and header joints and overall strength of copper-brass radiators. They comprise thinner fin and tube material; brazed copper fins are 0.05 mm (0.002 in.) thick or less, while brazed brass tubes are 0.127 mm (0.005 in.) thick. A filler metal of a nontoxic, low-temperature alloy (based on the Cu-Ni-Sn-P system) has been successfully employed. These alloys have a melting temperature as
low as 600 °C (1110 °F) and are self-fluxing in an inert atmosphere, so no rinsing is required after the brazing operation. Reference 47 reports on an investigation into a composite carbon-fiber/copper matrix. This type of composite manifests a number of advantages originating from the properties of the component materials, such as thermal and electrical conductivity, which make it suitable for, for example, resistance welding electrodes and electrical contacts or dilatation bases in semiconductor power elements. In order to make full use of the advantages of this material, it is necessary to join it to ceramics and to metals and alloys used in engineering. The research work examined the joining of carbon-fiber/copper composite to molybdenum and to FeNi42 alloy. These two materials were selected in view of the potential future applications of the resulting joints in the electronics industry, namely, for power diodes and thyristors (joints involving molybdenum) and for electrical contact pieces (joints involving FeNi42). The filler metal used was 72.5Ag-19.5Cu5In-3Ti (wt%). The joints were made using the diffusion-bonding method in a diffusion-bonding press. In the bonding of the volumetric composite carbon fiber/copper to molybdenum, the following process parameters were used: temperature, 950 °C (1740 °F); time, 30 min; pressure, 50 MPa (7 ksi); and vacuum, 2.66 × 10–4 Pa (3 × 10–8 psi). The laminar composites, consisting of carbon fiber in Ag-Cu-In-Ti alloy, were bonded to FeNi42 alloy and to Al2O3 ceramics under the following conditions: temperature, 850 °C (1560 °F); pressure, 0.6 MPa (0.09 ksi); time, 5 min; and vacuum, 2.66 × 10–4 Pa (3 × 10–8 psi). The experiments showed that it was possible to produce joints displaying good strength properties of carbon-fiber/copper composites to molybdenum and of carbon-fiber/AgCu alloy composites to FeNi42 alloy. In the examination of the microstructure, of the linear distribution of elements and of x-ray diffraction patterns, the bond of the composite carbon-fiber/AgCu alloy to molybdenum is likely to be of the diffusion type, which results from diffusion of silver and titanium out of the composite as well as from diffusion of iron and nickel in the opposite direction. In the case of the composite carbonfiber/copper bonding to molybdenum, the joint was formed mainly as an effect of nickel diffusion into molybdenum.
86 / Brazing, Second Edition
tion is performed in a flux bath. Furnace brazing offers the greatest potential for magnesium combustion, because the entire structure is heated to the brazing temperature. However, because the M1A alloy, with its high solidus temperature (648 °C, or 1198 °F), is the only magnesium alloy suitable for furnace brazing, the risk of combustion is minimal, provided that the furnace temperature is controlled within the prescribed limits and that brazing time is the minimum necessary to achieve filler-metal flow. Electric or gas heating furnace equipment with automatic temperature controls capable of holding the furnace temperature within ±6 °C (±10 °F) of the brazing temperature for magnesium base metal should be used to minimize incipient melting of the base metal and to reduce the potential of a magnesium fire. The best results in furnace brazing are obtained if dry powdered flux is sprinkled along the joint. Brazing times depend somewhat on the thickness of the materials used and the amount of fixturing necessary to position the parts. Torch brazing is accomplished using a neutral oxyfuel or air-fuel gas flame. Because of the close proximity of the initial melting point (solidus temperature) of the magnesium alloy base metal and the flow point (liquidus temperature) of the filler metal, manual torch brazing is generally preferred, using the AZ125A magnesium filler metal. Natural gas is also well suited for torch brazing, because its relatively low flame temperature reduces the danger of overheating. Dip brazing of magnesium alloys is accomplished by immersing the assembly in a molten brazing flux and holding the parts and fixture at
Magnesium and Magnesium Alloys Brazing techniques similar to those used for aluminum are used for magnesium alloys. Furnace, torch, and dip brazing can be employed, although dip brazing is the most widely used. Magnesium alloys that are considered brazeable, and recommended filler metals, are listed in Table 4.5. Furnace and torch brazing experience has been limited to the M1A alloy. Dip brazing has been used successfully for AZ10A, AZ31B, K1A, M1A, and ZK21A alloys. Other magnesium alloys have not been brazed with present filler metals and techniques developed, due to their low solidus temperatures. The two filler metals AZ92A or BMg-1 (9% Al, 2% Zn, 0.1% Mn, 0.00005% Be, balance Mg) and AZ125A or BMg-2 (12% Al, 0.5% Zn, 0.00005% Be, balance Mg) are both suitable for torch, dip, or furnace brazing. Because the AZ125A filler metal has a lower melting range (Table 4.5), it is usually preferred in most brazing applications. The ease of ignition of magnesium depends to a large extent on the size and shape of the material as well as on the size or intensity of the source of ignition. Fabricated assemblies, ingots, and castings are difficult to ignite, because heat is conducted rapidly away from the source of ignition. The probability of combustion occurring during torch brazing of magnesium is extremely remote, because only the joint area, and not the entire structure, is heated, and because the fluxes used prevent burning in the heated area. The dip brazing process presents no combustion problems with magnesium, because the entire opera-
Table 4.5 Brazeable magnesium alloys and recommended brazing filler metals Solidus ASTM alloy designation
Liquidus
Brazing range
Suitable brazing filler metal
°C
°F
°C
°F
°C
°F
BMg-1 or AZ92A
BMg-2 or AZ125A
632 566 649 648 626
1170 1050 1200 1198 1159
643 627 650 650 642
1190 1160 1202 1202 1187
582–616 582–593 582–616 582–616 582–616
1080–1140 1080–1100 1080–1140 1080–1140 1080–1140
X ... X X X
X X X X X
830 770
599 566
1110 1050
604–616 582–610
1120–1140 1080–1130
... ...
... ...
Base metals AZ10A AZ31B K1A M1A ZK21A
Brazing filler metals AZ92A(a) AZ125A(b)
443 410
(a) Available as wire or rod. (b) Available as wire, rod, strip, or powder
Chapter 4: Base Metals and Base-Metal Family Groups / 87
the desired brazing temperature. The flux serves the dual function of heating and fluxing. The immersion time in the flux bath can vary from 30 to 45 s up to 1 to 3 min. Large assemblies with more metal mass and/or fixturing require the longer times. Because of the large volume of flux and uniform heating, more consistent results are achieved with dip brazing than with other brazing procedures. The corrosion resistance of brazed joints in magnesium alloys depends primarily on the thoroughness of flux removal and the adequacy of joint design to prevent flux entrapment. Because the filler metal is a magnesium-base alloy, the problem of galvanic corrosion is minimized. Protection of magnesium alloys from corrosion is provided by applying appropriate chromate and other coatings. Torch brazing has been used to join hydraulic lift floats, whereas dip brazing has been used to produce battery containers and microwave antennas.
Nickel, Cobalt, and Heat-Resistant Alloys Nickel and the high-nickel alloys may be divided into the following classes: • • • • •
Commercially pure nickel Nickel-copper alloys Nickel-chromium-iron alloys Nickel-chromium-molybdenum alloys Thorium-dispersed nickel alloys
Commercially pure nickel and nickel-copper alloys are used primarily for applications where corrosion resistance is important or where product purity must be maintained. The other categories of alloys also have good corrosion resistance in many media and, in addition, have high strength and oxidation resistance at elevated temperatures. To use these properties, it is important to determine the effects of service conditions on the brazed joint. Nickel and the high-nickel alloys are embrittled by sulfur and low-melting metals, such as zinc, lead, bismuth, and antimony. Nickel and nickel alloy parts should be thoroughly cleaned prior to brazing to ensure the absence of substances that may contain any of these elements. Sulfur and sulfur compounds must also be excluded from the brazing atmosphere. Nickel and its alloys are subject to stress cracking in the presence of molten filler metals.
Parts should, therefore, be annealed prior to brazing to remove residual stresses or be carefully stress relieved during the brazing cycle. In selection of a brazing process and a filler metal for a nickel-base alloy, the characteristics of the alloy must be carefully considered. Nickel alloys differ significantly not only in physical metallurgy (precipitation strengthened versus solid-solution strengthened) but also in process history (cast versus wrought). These characteristics can have a profound effect on brazeability. The filler metals normally used for ferrous metals are suitable for joining nickel and highnickel alloys. It also is important to consider any heat treatments that may be required for the base metal, because the brazed joint must withstand the temperatures involved. In corrosive environments, high-silver filler metals are preferred, while cadmium-free filler metals are chosen to avoid stress-corrosion cracking (SCC). Most high-nickel alloys are capable of being brazed with pure copper filler metal, which is similar to the brazing of carbon and low-alloy steel, except that the copper filler metal characteristically alloys to a greater extent with nickel than with iron. Alloying during brazing makes capillary flow difficult. The copper does not flow far before it has picked up enough nickel to raise its liquidus and reduce its fluidity. To eliminate this problem, the filler metal should be placed as close to the joint as possible, and there should be a sufficient reservoir to fill the joint. Secondly, the assembly should be heated as rapidly as practicable to the brazing temperature. Phosphorus combines with many metals to form brittle compounds known as phosphides. For this reason, the copper-phosphorus filler metals usually are not used with any iron- or nickel-base alloy; however, the BNi-6 and BNi7 filler metals are used for brazing nickel-base alloys. Nickel-base filler metals offer the greatest corrosion and oxidation resistance and elevated-temperature strength. Nickel-base filler metals, containing palladium or platinum, and gold-base, palladiumbase, and platinum-base filler metals have also been used successfully to braze high-nickel alloys. These filler metals generally have good wetting and flow characteristics. They have a low interaction rate with most nickel-base metals and are used to advantage in many special applications where joints are required to have good ductility, high strength, and good oxidation resistance. Filler metals consisting of base
88 / Brazing, Second Edition
material plus additions of silicon and boron have been used successfully for brazing the precipitation-strengthened alloys of nickel. When appreciable amounts of aluminum and titanium appear (greater than 1%) in the precipitation-strengthened nickel-base alloys, the oxides of aluminum and titanium are almost impossible to reduce in a controlled atmosphere (vacuum or hydrogen). Therefore, nickel plating or the use of a flux is necessary to obtain a surface that allows wetting by the filler metal. Most commercial brazing processes may be used on nickel and high-nickel alloys. The most common of these are torch, furnace, induction, and resistance brazing; salt bath dip brazing has limited application. Whereas the silver-base filler metals may be used in torch brazing, the copper and nickel filler metals are usually used in controlled-atmosphere brazing. Attempting to braze over the refractory oxides of titanium and aluminum that may be present on precipitation-hardenable nickel-base alloys must be avoided. Procedures to prevent or inhibit the formation of these oxides before and/or during brazing include special treatments of the surface to be joined, and brazing in a highly controlled atmosphere. Surface treatments include electrolytic nickel plating and reduction of the oxides to metallic form. As stated earlier, a typical practice is to nickel plate the joint surfaces of any alloy that contains aluminum and/or titanium. For vacuum brazing, when aluminum and titanium are present in trace amounts, plating 0.01 mm (0.0004 in.) thick is considered optional. Alloys with up to 4% Al and/or Ti require plating 0.015 mm (0.0006 in.) thick, whereas alloys with aluminum and/or titanium contents greater than 4% require plating 0.02 to 0.03 mm (0.0008 to 0.0012 in.) thick. When brazing is done in a pure dry hydrogen atmosphere, thicker plating (0.03 to 0.04 mm, or 0.0012 to 0.0016 in.) is desirable for alloys with greater than 4% Al and/or Ti contents. Brazing of base materials containing more than a few percent of aluminum, titanium, zirconium, or other elements that form very stable oxides requires vacuums of 0.13 Pa (2 × 10–5 psi) or lower. Vacuum furnaces for this type of brazing usually employ a diffusion pump backed by a mechanical pump. Consideration must be given to the effect of the brazing thermal cycle on the base metal. Filler metals that are suitable for brazing nickelbase alloys may require relatively high thermal cycles. This is particularly true for the filler-
metal alloy systems most frequently used in brazing nickel-base alloys—the Ni-Cr-Si and Ni-Cr-B systems. It is significant to point out the effects of thermal cycles on solid-solution-strengthened and precipitation-strengthened nickel-base alloys. An example of the former is Inconel 600, a high-temperature nickel-base alloy containing 15.5% Cr, 76% Ni-Co, and 8% Fe that resists oxidation up to 1175 °C (2150 °F) and is used in furnace parts and fixtures, heat exchangers, chemical-handling equipment, and turbine and reactor parts. This alloy may not be adversely affected by nickel-base filler-metal brazing temperatures of 1010 to 1230 °C (1850 to 2250 °F). An example of the latter is Inconel 718, a nickel-base superalloy containing 50 to 55% Ni, 17 to 21% Cr, 4.75 to 5.5% Nb and Ta, 1% Co, approximately 0.20% Fe, and greater than 2% Al and Ti. This alloy is oxidation resistant up to 980 °C (1800 °F) and is used in aircraft turbine parts, pumps, and rocket motors. This alloy may, however, display adverse property effects when exposed to brazing cycles higher than its normal solution heat treatment temperature. Inconel 718, for example, is solution heat treated at 955 °C (1750 °C) for optimal stressrupture life and ductility. Brazing temperatures of 1010 °C (1850 °F) or above result in grain growth and an attendant decrease in stress-rupture properties, which cannot be recovered by subsequent heat treatment. The applicability of the transient liquid phase (TLP) metallic bonding method for joining finegrained Inconel 718SPF superalloy sheets by inserting a nickel-phosphorus or a Ni-Cr-P amorphous interlayer has shown that a joint with uniform chemical composition could be obtained for the Inconel 718SPF superalloy with a nickel-phosphorus interlayer at 1100 °C (2010 °F) for 8 h. When a Ni-Cr-P interlayer was used under the same metallic bonding conditions, the concentrations of nickel, iron, and niobium in the bond region and in the base metal had a difference of more than 2 wt%. This means that longer bonding time was required to homogenize the chemical compositions of bonds with a Ni-Cr-P interlayer. The shear strength of the TLP bonds with a nickel-phosphorus interlayer indicated higher bond strengths than did that with the Ni-Cr-P interlayer under the same bonding conditions. The second phases, which existed at the grain boundaries, retarded grain growth in the bonding region. Homogenization during the TLP metallic bonding process was thus limited (Ref 48).
Chapter 4: Base Metals and Base-Metal Family Groups / 89
Consideration of base-metal property requirements for service enables selection of an appropriate filler metal. For lower melting temperatures (below 1040 °C, or 1900 °F), filler metals are available within the nickel-base filler-metal family and within other filler-metal systems (Tables 4.6, 4.7) (Ref 49–51). Brazing is one of the preferred methods for joining dispersion-strengthened nickel alloys that must function at elevated temperatures. High-strength brazements have been made with special nickel-base filler metals tested up to 1315 °C (2400 °F). Procedures have been developed for brazing of thoria-dispersed nickel and thoria-dispersed NiCr foils for high-temperature service using the candidate filler metals shown in Table 4.8 (Ref 49, 51). Heat-resistant alloys are suitable for use under moderate to high loading in the temperature range from 540 to 1100 °C (1000 to 2010 °F). These metals are complex austenitic alloys based on nickel, cobalt, or both. They have often been termed superalloys. Their greatest use is in the construction of gas turbine engines and hot airframe components. Heat-resistant alloys are generally brazed in hydrogen-atmosphere or high-vacuum furnaces with nickel-base or special filler metals. Because the brazing temperatures are high, the effect of the brazing thermal cycle on the base metals should be taken into account. The nonheat-treatable alloys suffer moderate strength losses due to grain growth during brazing. Coldworked alloys should not be brazed unless the severe loss in strength from annealing during brazing is considered in the design. Brazing of cobalt-base alloys is readily accomplished by the same techniques used for nickel-base alloys. Because most of the popular cobalt-base alloys do not contain appreciable amounts of aluminum or titanium, brazingatmosphere requirements are less stringent. These materials can be brazed in either a hydrogen atmosphere or a vacuum. Filler metals are usually nickel- or cobalt-base alloys or goldpalladium compositions. Silver or copper filler metals may not have sufficient strength and oxidation resistance in many high-temperature applications. Although cobalt-base alloys do not contain appreciable amounts of aluminum or titanium, an electroplate or flash of nickel is often used to promote better wetting of the filler metal. Nickel-base filler metals such as BNi-3 (Table 4.7) have been used successfully on cobalt-base alloys (Ref 51, 52) for honeycomb
structures. After brazing, a diffusion cycle was used to raise the brazed-joint remelt temperature to 1260 to 1315 °C (2300 to 2400 °F). Additionally, it may be desirable to heat treat the brazement after brazing to attain optimal base-metal properties. The BCo-1 filler metal (Table 4.7) appears to offer a good combination of strength, oxidation resistance, and remelt temperature for use on cobalt-base alloy foil (Haynes 25) (Ref 53). Cobalt alloys, much like nickel alloys, can be subject to liquid metal embrittlement or SCC when brazed under residual or dynamic stresses. This frequently is observed when silver or silver-copper filler metals are used. Liquid metal embrittlement of cobalt-base alloys by copper filler metals occurs with or without the applications of stress; therefore, copper filler metals should be avoided in brazing of cobalt-base alloys. Superalloys can be subdivided into two categories: conventional cast and wrought alloys, and powder metallurgy (P/M) products. Powder metallurgy products may be produced in conventional alloy compositions and oxide-dispersion-strengthened (ODS) alloys. Oxide-dispersion-strengthened alloys are P/M alloys that contain stable oxide evenly distributed throughout the matrix. The oxide does not go into the solution in the alloy even at the liquidus temperature of the matrix. However, the oxide is usually rejected from the matrix on melting of the matrix, which occurs during fusion welding, and cannot be redistributed in the matrix on solidification; therefore, these alloys are usually joined by brazing. There are two commercial alloy classes of ODS alloys: the dispersion-strengthened nickel-chromium and dispersion-strengthened nickel, discussed previously, and the mechanically alloyed (MA) alloys. An example is Inconel MA 754, a turbine vane material made from powder. The material has extremely good creep and oxidation resistance at temperatures up to 1050 °C (1920 °F) and is used in gas turbines (vanes, nozzle, burners, combustors), chemical plants (liners, heat exchangers, combustion chambers), and so on. They owe their oxidation resistance to a stable and tenacious oxide film, and their creep resistance to the combination of a large and directional grain structure plus a dispersion of submicron-sized particles consisting chiefly of yttrium oxide. There is no joining method that can preserve the coarse and elongated grain structure of the recrystallized parent metal across a joint line
90 / Brazing, Second Edition
Table 4.6 Commercially available nickel-base filler metals Solidus (melting point)
Chemical composition, % Ni
99.26 98.88 97.25 96.00 95.50 95.20 94.60 93.72 93.60 92.80 92.60 92.54 92.40 92.30 91.70 91.60 91.30 91.22 90.80 90.66 88.95 88.80 88.40 87.10 85.90 84.39 84.08 83.40 82.97 82.50 82.10 81.50 81.50 81.10 80.85 79.25 79.20 78.87 78.65 78.40 77.00 76.80 76.30 75.60 74.85 74.60 74.40 73.90 73.62 73.40 72.44 72.20 72.00 71.90 70.00 69.20 68.59 68.10 67.00 66.85 65.50 65.00 63.00 61.95 61.50 61.40 60.35 59.15 57.10
Liquidus (flow point)
Cr
C
P
Fe
B
Si
Co
Mn
Cu
Other
°C
°F
°C
°F
0.04 ... ... ... ... ... ... ... 2.20 2.30 2.70 ... 2.90 2.25 2.30 2.50 2.70 ... 3.00 0.20 5.00 3.80 5.60 5.00 5.60 9.00 5.50 8.00 6.50 11.50 7.00 11.40 15.00 11.50 10.00 11.00 10.00 13.30 13.00 13.00 15.00 15.20 7.00 13.50 ... ... 15.00 14.80 17.10 17.00 15.00 ... ... 15.00 16.00 17.50 18.00 10.00 20.00 19.00 ... ... ... 12.00 19.00 19.50 11.50 19.50 3.50
0.04 0.015 ... ... ... ... 0.03 0.03 ... ... ... 0.06 ... ... ... ... ... 0.03 ... 0.14 0.25 ... ... ... ... 0.35 0.12 ... 0.03 ... 0.03 ... ... ... ... 0.50 0.35 ... ... 0.70 ... ... ... 0.15 0.15 ... 0.10 0.70 ... ... 0.06 ... 0.03 0.60 ... 0.80 ... 0.40 ... 0.15 0.01 0.07 ... 0.55 ... ... 0.55 ... ...
0.30 1.10 ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... 4.00 ... ... 1.50 6.00 ... ... ... ... ... ... 7.50 ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ...
... ... ... ... ... ... ... 1.25 1.00 1.00 1.10 ... 1.30 1.00 1.20 1.10 1.20 1.25 1.30 1.50 1.00 1.70 2.40 0.60 2.40 2.10 2.50 0.40 3.00 ... 3.00 ... ... ... 0.25 3.50 4.00 ... ... 3.50 ... ... 3.00 3.50 ... ... 3.00 3.80 ... ... 4.00 ... ... 4.00 2.00 4.50 4.50 3.50 3.00 3.00 ... ... 5.00 3.50 4.50 ... 3.40 2.25 1.00
... ... 0.75 1.50 1.50 1.80 1.90 1.50 1.60 1.50 1.90 2.90 1.90 1.70 1.80 1.80 1.80 3.00 1.90 3.00 1.80 2.20 2.80 1.20 2.50 1.66 2.50 0.80 3.00 ... 2.90 0.30 3.50 0.40 0.50 2.25 2.10 0.23 0.35 2.90 ... ... 3.20 3.25 ... 3.15 3.00 3.00 0.08 0.10 3.50 3.30 3.60 3.50 3.50 3.50 ... 2.50 ... ... ... ... 2.00 2.50 ... ... 2.50 ... 0.90
... ... 2.00 2.50 3.00 3.00 3.50 3.50 2.60 2.40 2.80 4.50 2.80 2.75 3.00 3.00 3.00 4.50 3.00 4.50 3.00 3.50 3.20 2.10 3.60 2.50 3.80 1.40 4.50 6.00 5.00 6.80 ... 7.00 0.90 3.50 4.35 7.60 8.00 1.50 8.00 8.00 4.50 4.00 8.00 4.25 4.50 3.80 9.20 9.50 5.00 4.50 4.50 5.00 3.50 4.50 9.00 3.50 10.00 10.00 7.00 7.00 ... 3.50 9.80 9.60 3.20 9.50 2.50
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... 18.00 ... ... ... ... ... 20.00 20.00 ... ... ... ... ... ... 0.50 ... ... ... ... 5.20 ... 2.50 ... ...
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... 17.00 ... ... ... ... ... ... ... ... ... ... ... ... ... ... 0.50 23.00 23.00 ... ... ... 9.50 ... 9.50 35.00
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... 2.50 ... ... ... ... ... 4.50 5.00 ... ... ... ... ... ... ...
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... 6.00 W ... ... ... ... ... ... ... ... ... ... ... 2.50 Mo ... ... 12.00 W ... ... ... ... 30.00 Mo 16.00 W ... ... 16.00 W ... ...
1082 1038 1021 1010 1010 1004 982 982 954 954 954 982 954 954 954 954 954 954 954 982 971 954 954 954 954 1010 982 982 971 1082 1038 1082 1054 1082 954 971 971 1082 1082 1082 1082 1082 977 976 1024 982 971 977 1082 1082 982 982 982 977 1093 949 1082 971 1079 1082 982 982 1066 971 1082 1082 1010 1082 987
1980 1900 1870 1850 1850 1840 1800 1800 1750 1750 1750 1800 1750 1750 1750 1750 1750 1750 1750 1800 1780 1750 1750 1750 1750 1850 1800 1800 1780 1980 1900 1980 1930 1980 1750 1780 1780 1980 1980 1980 1980 1980 1790 1789 1875 1800 1780 1790 1980 1980 1800 1800 1800 1790 2000 1740 1980 1780 1975 1980 1800 1800 1950 1780 1980 1980 1850 1980 1800
1121 1204 1066 1066 1066 1066 1066 1066 1038 1038 1038 1024 1038 1038 1038 1038 1038 1024 1038 1038 1182 1038 1038 1038 1038 1066 1038 1049 999 1138 1065 1138 1054 1138 1038 1160 1149 1138 1138 1138 1138 1138 1038 1052 1065 1024 1077 1038 1138 1138 1024 1024 1024 1052 1149 1065 1138 1093 1135 1138 1010 1010 1138 1104 1138 1107 1121 1121 1066
2050 2200 1950 1950 1950 1950 1950 1950 1900 1900 1900 1875 1900 1900 1900 1900 1900 1875 1900 1900 2160 1900 1900 1900 1900 1950 1900 1920 1830 2080 1950 2080 1930 2080 1900 2120 2100 2080 2080 2080 2080 2080 1900 1925 1950 1875 1970 1900 2080 2080 1875 1875 1875 1925 2100 1950 2080 2000 2075 2080 1850 1850 2080 2020 2080 2025 2050 2050 1950
(continued)
Chapter 4: Base Metals and Base-Metal Family Groups / 91
Table 4.6 (continued) Solidus (melting point)
Chemical composition, %
Liquidus (flow point)
Ni
Cr
C
P
Fe
B
Si
Co
Mn
Cu
Other
°C
°F
°C
°F
50.20 50.10 36.00 17.00 17.00 16.00 12.00 9.00 9.00 9.00 5.00 2.00
4.00 ... ... 21.00 19.00 ... 18.00 ... ... ... ... 16.00
... ... ... 0.80 0.40 ... ... 0.03 ... 0.03 ... 0.20
... 3.50 ... ... ... ... 0.10 ... ... ... ... ...
... 30.00 ... ... ... ... 66.90 ... ... ... ... 81.80
0.80 ... ... 3.25 0.80 1.00 ... ... ... ... ... ...
... 11.00 ... 3.00 8.00 ... 1.00 ... ... ... ... ...
... ... ... 44.95 50.80 16.00 ... ... ... ... 10.00 ...
45.00 ... ... ... ... 67.00 2.00 23.50 38.50 37.00 22.00 ...
... ... 59.00 ... ... ... ... 67.50 52.50 52.00 63.00 ...
... 5.40 Mo 5.00 In 10.00 W 4.00 W ... ... ... ... ... ... ...
996 954 1038 1038 1107 1004 1121 910 880 871 943 1149
1825 1750 1900 1900 2025 1840 2050 1670 1615 1600 1730 2100
1079 1001 1204 1121 1149 1021 1163 932 927 927 957 1204
1975 1835 2200 2050 2100 1870 2125 1710 1700 1700 1755 2200
(Ref 54, 55). However, some sacrifice of the maximum properties may be allowable if such a joint were to be sited outside the area of maximum stress; alternatively, a joint parallel to the major stresses might be acceptable in certain cases. Although brazing introduces a change in composition across the joint line, and although it does melt some of the parent metal and thus causes agglomeration of the dispersoid, nevertheless, these effects can be minimized by a suitable choice of composition and by close control of brazing and postbrazing conditions. Furthermore, brazing does not introduce mechanical distortion, and it can be carried out at temperatures well below the recrystallization range. Inconel MA 754 is one of the easiest alloys to braze in the family of ODS alloys. Vacuum, hydrogen, or inert atmospheres can be used for brazing. Prebraze cleaning consists of grinding or machining the faying surfaces and washing with a solvent that evaporates without leaving a residue. Generally, brazing temperatures should not exceed 1315 °C (2400 °F) unless demanded by a specific application that has been well examined and tested. The filler metals for use with these ODS alloys usually are not classified (Table 4.8); in most cases, they have brazing temperatures in excess of 1230 °C (2250 °F). These include proprietary filler metals containing nickel, cobalt, gold, or palladium. A series of filler-metal evaluation tests (Ref 55) found that brazing fine-grained MA 754 at 1200 °C (2190 °F) by induction heating for less than 30 s with a thin β-free coating equivalent to BNi-5 gave joints of excellent appearance and without recrystallization.
A second nickel-base ODS alloy is Inconel MA 6000. Like Inconel MA 754, it is γ strengthened. Inconel MA 6000 has a solidus temperature of 1300 °C (2370 °F); therefore, the brazing temperature should be no higher than 1250 °C (2280 °F). Nickel and cobalt filler metals have been used to join this alloy. Other mechanically alloyed ODS alloys include iron-base creepresistant MA 956, which was also recently examined (Ref 56). Although it was possible to produce a reasonable brazed joint in MA 956 by means of sputter-coated brazes based on the nickel filler metal BNi-la (72.5Ni-16.5Cr-4Fe4Si-3B), an agglomeration of the dispersed oxide at the braze interface was unavoidable. Furthermore, as in the brazing of nickel-base ODS alloys, the presence of boron in the braze causes local recrystallization of the parent material. This has led to future work whereby diffusion bonding was to be examined as the only joining process that did not require either a foreign material or a melted or disrupted region in the joint. Diffusion bonding therefore offered the ability to form a bond in an iron-ODS alloy in which the massive grain structure continued unbroken across the bond. To decrease the completion time during TLP bonding, an iron-base filler metal, comprising MA 956 base metal containing 7% Si and 1% B (wt%), was developed to join MA 956 base material. Transient-liquid-phase-bonded joints free of microvoids and bond line intermetallic phases were obtained using a bonding temperature of 1563 K, a holding time of 2 to 16 ks, and an applied pressure of 7.0 MPa (1 ksi). The bond line region in the TLP-bonded MA 956 base metal had a bamboolike microstructure,
Cr
B
Cr
37.0–38.0 79.5–80.5 34.5–35.5 81.5–82.5 29.5–30.5
Fe
C
P
Ni
18.0–20.0 16.0–18.0
7.5–8.5
Si
bal bal bal ... ...
Cu
3.5–4.5
W
Pd
1.0
Fe
... ... ... ... 33.5–34.5
Composition, wt %
... ... 2.5–3.5 bal 35.5–36.5
Ni
0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02
S
C
P
0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05
Al
0.7–0.9 0.35–0.45 0.02
B
Composition, wt %
(a) If determined, cobalt is 0.1% maximum unless otherwise specified. Source: AWS A5.8
BCo-1
Si
2.75–3.50 4.0–5.0 4.0–5.0 0.6–0.9 0.02 2.75–3.50 4.0–5.0 4.0–5.0 0.06 0.02 2.75–3.50 4.0–5.0 2.5–3.5 0.06 0.02 2.75–3.50 4.0–5.0 0.5 0.06 0.02 1.5–2.2 3.0–4.0 1.5 0.06 0.02 0.03 9.75–10.50 . . . 0.10 0.02 ... ... ... 0.10 10.0–12.0 0.01 0.10 0.2 0.08 9.7–10.5 ... 6.0–8.0 ... 0.10 0.02 3.25–4.0 ... 1.5 0.06 0.02
Cobalt-base alloy filler metals
AWS classification
BAu-1 BAu-2 BAu-3 BAu-4 BAu-5
Au
13.0–15.0 13.0–15.0 6.0–8.0 ... ... 18.5–19.5 ... 13.0–15.0 ... 13.5–16.5
Precious metals
AWS classification
BNi-1 BNi-1a BNi-2 BNi-3 BNi-4 BNi-5 BNi-6 BNi-7 BNi-8 BNi-9
Nickel-base alloy filler metals(a)
AWS classification
Composition, wt %
Mn
Cu
0.02
S
0.15 0.15 0.15 0.15 0.15
Other elements total
0.05
Al
990 890 975 950 1135
°C
0.05
Ti
Zr
0.05
Zr
°F
bal bal bal bal bal bal bal bal bal bal
Ni
bal
Co
1815 1635 1785 1740 2075
0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05
Solidus
0.05 ... ... 0.05 ... ... 0.05 ... ... 0.05 ... ... 0.05 ... ... 0.05 ... ... 0.05 ... ... 0.05 0.04 ... 0.05 21.5–24.5 4.0–5.0 0.05 ... ...
Ti
Table 4.7 American Welding Society (AWS) brazing alloys for elevated-temperature service
0.50
°C
1120
°C
2050
°F
Solidus
Liquidus
1790 1790 1780 1800 1800 1975 1610 1630 1800 1930
°F
Solidus °C
975 975 970 980 980 1080 875 890 980 1055
1015 890 1030 950 1165
Other elements total
0.50 0.50 0.50 0.50 0.50 0.50 0.50 0.50 0.50 0.50
Other elements total
1790 1970 1830 1900 1950 2075 1610 1630 1850 1930
1150
2100
°F
Liquidus °C
1860 1635 1885 1740 2130
°F
975 1075 1000 1040 1065 1135 875 890 1010 1055
°F
Liquidus °C
1860–2000 1635–1850 1885–1995 1740–1840 2130–2250
°F
1150–1230
°C
2100–2250
°F
Brazing range
1015–1095 890–1010 1030–1090 950–1005 1165–1230
°C
Brazing range
1950–2200 1970–2200 1850–2150 1850–2150 1850–2150 2100–2200 1700–2000 1700–2000 1850–2000 1950–2200
°F
Brazing range
1065–1205 1075–1205 1010–1175 1010–1175 1010–1175 1150–1205 925–1095 925–1095 1010–1095 1065–1205
°C
92 / Brazing, Second Edition
Chapter 4: Base Metals and Base-Metal Family Groups / 93
which was considered to be peculiar to bonded ODS alloys. For this reason, the growth mechanism of the solid phase during isothermal solidification was evaluated. During tensile testing at 923 K, the joints TLP bonded at 1563 K for 2 to 16 ks fractured in the base-metal zone, and consequently, the mechanical properties of the joint region and the base metal were similar. The creep-rupture properties of the joint regions were close to the base-metal properties in the transverse direction. In order to produce complex and high-temperature-resistant components, diffusion bonding/brazing has been investigated. The joining of a new iron-base ODS alloy, PM2000, was developed (Ref 57). Oxide-dispersion-strengthened superalloys are very promising metallic materials for different high-temperature-exposed components. The ferritic ODS material is especially interesting and suitable for high-temperature-exposed stationary components in turbine and diesel engine constructions. The combination of excellent oxidation and corrosion resistance and the high melting point (1480 °C, or 2700 °F) of the FeCr-Al alloy with a Y2O3-dispersion-strengthened matrix permits higher operating temperatures for the iron-base ODS alloy than ever attainable with the nickel-base superalloys. In iron-base Fe-Cr-Al ODS alloys, the Y2O3 dispersoid does not coarsen and coagulate during the joining. At room temperature, a bending strength of approximately 70% of base-material strength is attainable. A secondary recrystallization sometimes occurs in the bonding zone. Generally, the bond interface is formed as a straight grain boundary, which can limit the strength of the bond. It is very important to develop high-quality electric power systems that have excellent
energy efficiency. With this in mind, electric power systems combining gas and steam turbines must be improved to increase the energy efficiency. A key requirement for land gas turbines is that higher inlet gas temperatures can be applied. Single-crystal base material is one of the most suitable candidate materials for highquality land gas turbines. Nickel-base singlecrystal superalloys possess the excellent mechanical properties at elevated temperatures (because the melting temperature is increased when grain-boundary strengthening elements are not present). The elevated-temperature properties of single-crystal base material deteriorate when grain boundaries are formed in the joint region. Therefore, it is vital that singlecrystal base material is bonded without creating such a problem. Bonding and the crystallization behavior of a nickel-base single-crystal superalloy base material, CMSX-2, was investigated during TLP bonding using MBF-80 insert metal (Ref 58). Joint strength was evaluated using tensile and creep-rupture testing at elevated temperature. The TLP bonding of CMSX-2 was carried out at 1373 to 1548 K for 0 to 19.6 ks in vacuum (Table 4.9). Some of the conclusions include: • The dissolution width of the base metal increased when the bonding temperature and holding time increased. Base-metal dissolution was consistent with a derivation based on the Nernst-Brunner theory. • Material grew epitaxially into the liquid phase from substrates during the isothermal solidification, and single-crystal properties were maintained in completed joints. • The elevated-temperature tensile properties of completed joints were almost identical to those of base metal. The creep-rupture
Table 4.8 Filler metals for brazing thoria-dispersed (TD) nickel and TD-NiCr alloys Brazing temperature Base metal
TD-Ni
TD-NiCr
Source: Ref 51
Brazing filler metal
°C
°F
Nominal composition
TD-20 TD-6 J8600 Ni-Pd Td-6 CM50 NX77 NSB
1302 1302 1177 1246 1302 1066–1121 1177–1191 1288
2375 2375 2150 2275 2375 1950–2050 2150–2175 2350
Ni-16Cr-25Mo-4Si-5W Ni-22Cr-17Mo-4Si-5W Ni-33Cr-25Pd-4Si Ni-60Pd Ni-22Cr-17Mo-4Si-5W Ni-3.5Si-1.9B Ni-5Cr-7Si-1B-1W-4Co Ni-2Si-0.8B
94 / Brazing, Second Edition
strength and rupture lives of completed joints were also almost identical to those of the base metal. Joint fracture always occurred in the base material (Fig. 4.8).
Molybdenum, Niobium, Tantalum, Tungsten, and Their Alloys The refractory metals have a history of use in applications where structural integrity at high temperatures is required. For example, the filaments for incandescent lamps are made from tungsten. This metal is used extensively in vacuum tubes for cathode structures. Similarly, molybdenum is used for grids and supporting members in vacuum tubes and for lead wires in incandescent lamps. In current applications of the refractory metals in the aircraft, space, nuclear, and electronics industries, the joints must be of significant dimensions, with properties that are adequate to meet exacting requirements of strength, ductility, corrosion resistance, and so on. Because of its limited effect on base-metal properties, brazing has been accepted widely as a method of joining assemblies fabricated from refractory metals. The characteristics that affect the brazeability of refractory metals include ductile-to-brittle transition behavior, recrystallization temperature, and reactivity with oxygen, nitrogen, hydrogen, and carbon. The refractory metals include those metals that have melting points in excess of 2205 °C (4000 °F), that is, niobium, iridium, molybdenum, tantalum, ruthenium, tungsten, osmium, and rhenium. Sometimes vanadium, hafnium, rhodium, and chromium are included in this group of metals. The most important refractory metals, from a structural standpoint, are molybdenum, tantalum, niobium, and tungsten. Niobium, molybdenum, tantalum, and tungsten have much in common; however, there are differences that have a bearing on the manner in which these metals are used and joined. All of the refractory metals have a body-centered cubic crystal structure, with very high melting
temperatures, high to very high densities, low specific heats, and low coefficients of thermal expansion. Mechanical strength and structural integrity at high temperatures are excellent. The mechanical properties of refractory metals are affected markedly by ductile-to-brittle transition behavior, recrystallization temperature, and reactions with carbon and various gases. These characteristics must be considered when brazing procedures are established. Refractory metals, because of their body-centered cubic structure, have a well-defined transition from ductile to brittle behavior. With these metals, a large decrease in energy absorption occurs over a narrow temperature range, and an associated change in the type of fracture from tough to brittle is also evident. This temperature range is not a fixed property of the metal, because it varies with strain rate, alloying addition, impurities, heat treatment, and method of fabrication. The transition-temperature ranges for the pure refractory metals are as follows: • • • •
Niobium, –200 to –75 °C (–330 to –105 °F) Molybdenum, 150 to 260 °C (300 to 500 °F) Tantalum, <–195 °C (–320 °F) Tungsten, 260 to 370 °C (500 to 700 °F)
Therefore, molybdenum and tungsten are brittle at room temperature and must be handled carefully to avoid damage. Also, these metals must be brazed in a stress-free condition. The strength and ductility of refractory metals are adversely affected by microstructural changes that occur when the recrystallization temperatures of these metals are exceeded. Recrystallization-temperature range also varies with alloying additions, interstitial content, fabrication method (including degree of cold working), and time at temperature. Recrystallization-temperature ranges for unalloyed refractory metals are: • Niobium, 985 to 1150 °C (1805 to 2100 °F) • Molybdenum, 1150 to 1200 °C (2100 to 2190 °F)
Table 4.9 Composition of materials used in brazing a nickel-base single-crystal superalloy Composition, mass%
Base metal: CMSX-2 Insert metal: MBF-80 Source: Ref 58
Ni
B
Cr
Co
Mo
W
Ti
Al
Ta
bal bal
... 3.7
8.0 15.5
4.6 ...
0.6 ...
8.0 ...
1.0 ...
5.6 ...
6.0 ...
Chapter 4: Base Metals and Base-Metal Family Groups / 95
• Tantalum, 1100 to 1400 °C (2010 to 2550 °F) • Tungsten, 1200 to 1650 °C (2190 to 3000 °F) Some applications permit brazing with filler metals that melt below the recrystallizationtemperature range. Other applications require the use of filler metals that melt above this temperature range. The joint also must be designed to accommodate the loss in mechanical properties associated with recrystallization. Research has improved the high-temperature mechanical properties of the refractory metals and has led to an increase in recrystallization temperatures through alloying. For example, titanium, zirconium, and hafnium can be used to strengthen molybdenum and increase its recrystallization temperature. The recrystallization temperature of unalloyed molybdenum is approximately 1150 to 1200 °C (2100 to 2190 °F). In contrast, the recrystallization temperature of Mo-0.5Ti-0.07Zr molybdenum alloy (TZM) is approximately 1480 °C (2700 °F), and the stressrupture strength of this alloy at 980 to 1095 °C (1800 to 2000 °F) is several times that of unalloyed molybdenum. The third consideration is reactions with gases and carbon. The environment in which refractory metals are brazed is determined by the reactivity of these metals with oxygen, hydrogen, and nitrogen and the effects of these elements on mechanical properties. Tantalum and niobium are embrittled by the presence of hydrogen at relatively low temperatures. In contrast, molybdenum and tungsten
Stress (σ), MPa (ksi)
1000 (145) 900 (131) 800 (116) 700 (102) 600 (87)
Base metal CMSX-2 Insert metal MBF-80
Joint 1523 x 1.8ks (2.3MPa)
500 (73) 400 (58) 300 (44)
200 (30) Base metal
100 (15) 23
24
25
26
27
28
29
30
31
Larson–Miller parameter, T(20+Log t)x10–3
Fig. 4.8
Creep-rupture strength of superalloy CMSX-2 joints. Source: Ref 58
can be brazed in a hydrogen atmosphere. Molybdenum, tantalum, and niobium are embrittled by nitrogen at high temperatures. All of the refractory metals form carbides in the presence of minute quantities of carbon and its compounds. Based on the various characteristics of refractory metals, the following principles are applicable to brazing of refractory metals: • If maximum joint strength is required, refractory metals must be brazed at temperatures below those at which recrystallization occurs. • However, brazing at much higher temperatures may be necessitated by service requirements, and some decrease in joint properties must be anticipated. • All of the refractory metals can be brazed in a vacuum or in an argon or helium atmosphere with a very low dewpoint. • Graphite fixturing should not be used to position refractory metal parts during brazing, because these metals readily form carbides. However, graphite tooling may be acceptable if coated with a refractory material. • Ceramics can also be used for fixturing, but care in their selection must be exercised. Some ceramics cannot be used in a vacuum because of outgassing, whereas others react with refractory metals at high temperatures. • Refractory metals with higher melting temperatures than the one being brazed also can be used for fixturing. The filler metals used for brazing of refractory metals must be selected on the basis of the service condition, the specific application, and their compatibility with the base metal and coating (if a coating is used) (Table 4.10). Molybdenum and Its Alloys. Copper- and silver-base filler metals can be used to braze molybdenum for low-temperature service. For high-temperature applications, molybdenum can be brazed with gold, palladium, and platinum filler metals, nickel-base filler metals, reactive metals, and refractory metals that melt at lower temperatures than molybdenum. Nickel-base filler metals have limited applicability for high-temperature service, because nickel and molybdenum form a low-melting eutectic at approximately 1315 °C (2400 °F). There are three basic limits on brazing of molybdenum and its alloys for use above 980 °C (1800 °F):
96 / Brazing, Second Edition
• Recrystallization of molybdenum • Formation of intermetallics between refractory metals and filler metals • Relative weakness of filler metals at elevated temperatures The formation of intermetallics between molybdenum and filler metals is detrimental to joint soundness, because the intermetallics become brittle and may fracture at relatively low loads when the joint is stressed. The relative weakness of filler metals at elevated temperatures limits the use of brazed molybdenum assemblies. Most of the nickelbase elevated-temperature filler metals melt between 980 and 1150 °C (1800 and 2100 °F), where the superior elevated-temperature strength of molybdenum begins to manifest itself. Two binary filler metals, V-35Nb and Ti30V, have been evaluated for use with the Mo0.5Ti molybdenum base metal. Tee-joints were brazed in a vacuum for 5 min at temperatures of 1650 °C (3000 °F) for the Ti-30V filler metal and 1870 °C (3400 °F) for the V-35Nb filler metal. The filler metals had excellent metallurgical compatibility with the molybdenum base metal, and minimum erosion of the base metal occurred during brazing. Brazements made on Mo-0.5Ti alloy with the diffusion-sink filler metal Ti-8.5Si, which melts at approximately 1330 °C (2425 °F), exhibited excellent filleting and wetting, joint ductility, and freedom from cracks. Specimens with molybdenum powder added to the filler-metal powder were also evaluated and brazed at 1400 °C (2550 °F) (Ref 59). Because of its excellent high-temperature properties and compatibility with certain environments, molybdenum is a prime candidate for use in isotopic power systems for components of nuclear reactors and chemical-processing
systems, and the best filler-metal base on overall performance has been Fe-15Mo-5Ge-4C-1B (Ref 60, 61). The molybdenum alloy TZM (0.5Ti-0.08ZrMo) has also been brazed successfully at 1400 °C (2550 °F), with molybdenum powder added to the Ti-8.5Si filler metal (Ref 59, 62); other filler metals also can be used (Ref 63). A Ti25Cr-13Ni filler metal with a brazing temperature of 1260 °C (2300 °F) has produced the highest remelt temperature on TZM. Remelt investigations conducted to develop vacuum brazements of molybdenum and tungsten that can be used in seal-joint applications up to 1600 °C (2910 °F) were completed with the following filler metals: • • • • • •
Ti-65 wire Vanadium wire MoB-50MoC powder mixture V-50Mo powder mixture Mo-15MoB2 powder mixture Mo-49V-15MoB2 powder mixture
Brazing temperatures ranged from 1625 to 2255 °C (2960 to 4095 °F). Molybdenum joints made with Ti-65V, pure vanadium wire, V50Mo, and MoB-50MoC were as strong, or stronger, than the base metal at elevated temperature. The most resistant joints tested were brazed with Ti-65V, V-50Mo, and MoB50MoC. Without further testing, it probably would be best to limit the maximum service temperature to 1395 °C (2545 °F). A more recently completed wettability study (Ref 64) found that: • Filler metals generally wet molybdenum significantly better than they wet TZM. This is seen in Tables 4.11 and 4.12. The wettability index shown for each filler metal indicates that indexes >0.05 are indicative of good per-
Table 4.10 Filler metals for brazing of graphite, ceramics, and refractory and reactive metals Chemical composition, %
Solidus (melting point)
Liquidus (flow point)
Ti
Zr
Be
Ni
Al
Cu
°C
°F
°C
°F
43 45 48 43 70 48
43 45 47 47 ... 48
2 2 5 5 ... 4
12 8 ... ... 15 ...
... ... ... 5 ... ...
... ... ... ... 15 ...
799 899 893 927 950 1000
1470 1650 1640 1700 1742 1832
816 899 904 927 1000 1050
1500 1650 1660 1700 1832 1922
Chapter 4: Base Metals and Base-Metal Family Groups / 97
formance during brazing, and wettability indexes >0.10 are indicative of excellent performance during brazing. • The decrease in wettability of the filler metals on TZM versus molybdenum is believed to be due to the presence of titanium in the oxide on the surface of TZM. • Some nickel-containing filler metals have been shown to embrittle both molybdenum and TZM. Liquid metal embrittlement is believed to be the mechanism by which this embrittlement occurs. Torches, controlled-atmosphere furnaces, vacuum furnaces, and induction and resistance heating equipment can be used to braze molybdenum with the variety of filler metals described previously. In a report (Ref 66) on the vacuum brazing of P/M molybdenum TZM alloy plate and drawn molybdenum tubing assembly for spacecraft radiator panels, the full-scale test panel successfully passed a battery of tests, including heat transport, thermal shock, and launchlike vibration testing. Brazing was successfully employed despite the recrystallization of both the tubing and the isogrid substrate panel. Platinum filler metal demonstrated its ability to obtain high-strength reliable joints.
Niobium and Its Alloys. Niobium is used mostly for nuclear and aerospace applications. One aerospace application required the development of manufacturing procedures for fabricating flat and curved niobium alloy sandwich heatshield panels (Ref 59). Titanium and the titanium-base alloy Ti-11Cr-13V-3Al (B120VCA) produced joints with excellent ductility and less base-metal reaction (Ref 67, 68). The flow characteristics of the B120VCA filler metal were more sluggish than those of pure titanium; as a result, the filler metal could be used to bridge wide joint clearances. The previously mentioned honeycomb sandwich panels were successfully brazed with pure titanium at 1815 °C (3300 °F). Heat-shield panels have been brazed with B120VCA filler metal at 1650 °C (3000 °F) in the same vacuum brazing environment. Simulated re-entry environment tests have shown that the heatshield panel could be used at temperatures up to 1315 °C (2400 °F). Room-temperature and elevated-temperature tests of the structural honeycomb panel indicated that it possessed useful properties up to 1260 °C (2300 °F) (Ref 59). Other developments investigated and evaluated the high-remelt-temperature brazing techniques based on diffusion-sink and reactive brazing concepts. The diffusion-sink technique involves either permitting the filler metal to
Table 4.11 Wettability index values for commercial filler metals used in brazing of molybdenum alloys Temperature, °C (°F) Filler metal
750 (1380)
800 (1470)
825 (1520)
850 (1560)
875 (1610)
900 (1650)
950 (1740)
975 (1790)
1000 (1830)
1025 (1880)
1050 (1920)
1075 (1970)
Palcusil 5 Palcusil 10 Palcusil 15 Palcusil 20 Palcusil 25 Nicusil 3 Nicusil 8 T50 T51 T52 Cusiltin 5 Cusiltin 10 Braze 580 Braze 630 Braze 655 Nioro Nicoro 80 Palniro 7 Gapasil 9
... ... ... ... ... ... ... ... ... ... ... ... 0.030 ... ... ... ... ... ...
... ... ... ... ... 0 ... ... 0 ... ... ... 0.037 ... ... ... ... ... ...
... ... ... ... ... ... ... ... ... ... ... 0.021 ... 0 0.061 ... ... ... ...
0.012 0.011 ... 0.015 ... 0.024 0.020 0.023 0.023 0.024 0.008 0.037 0.045 0.066 0.075 ... ... ... ...
... ... ... ... ... ... ... ... ... ... 0.012 0.031 ... ... 0.069 ... ... ... ...
0.034 0.030 0.044 0.037 ... 0.053 0.052 0.052 0.049 0.057 0.057 ... ... 0.080 ... ... ... ... 0.017
0.045 0.045 0.045 ... 0.102 ... 0.096 0.069 ... 0.075 ... ... ... ... ... ... 0.024 ... 0.061
... ... 0.152 ... 0.119 ... ... ... ... ... ... ... ... ... ... 0.158 ... ... ...
... ... ... ... 0.418 ... ... ... ... ... ... ... ... ... ... ... ... ... ...
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... 0.071 0.023 ...
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... 0.119 ...
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... 0.102 ...
Source: Ref 65
Source: Ref 65
580 Nicusil 3 T51 630 Cusiltin 655 Palcusil 5 Palcusil 10 Nicusil 8 T50 T52 Cusiltin 5 Palcusil 20 Ticusil Cusil Palcusil 15 Gapasil 9 Silcoro 60 Palcusil 25 Nioro Nicoro 80 Ticuni Silver Incuro 60 Silcoro 75 Palnicusil Altizirbe Palniro 7 Gold Palniro 1 Palniro 4 Palco
Filler metal
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ...
750 (1380)
0.007 0.008 0.007 0.015 ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ...
800 (1470)
... ... ... ... 0 0.009 ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ...
825 (1520)
0.009 0.009 0.009 0.018 0 0.009 0 0.007 0.008 0.014 0.016 0 ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ...
850 (1560)
0.018 ... ... ... 0 0.025 ... ... ... ... ... 0 0.009 0.024 ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ...
875 (1610)
... 0.015 0.010 0.016 0 ... 0.011 0.016 0.018 0.020 0.017 0 0.017 0.101 0 0.023 0 0 ... ... ... ... ... ... ... ... ... ... ... ... ... ...
900 (1650)
... ... 0.020 ... ... ... ... ... ... ... ... ... 0.034 ... 0.009 ... ... 0.020 ... ... ... ... ... ... ... ... ... ... ... ... ... ...
925 (1700)
... 0.022 0.018 ... ... ... 0.027 0.028 0.024 0.025 0.026 ... ... ... 0.009 0.039 0.025 0.023 0.096 0 0.020 0.093 ... ... ... ... ... ... ... ... ... ...
950 (1740)
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... 0.105 0.070 ... ... ... 0.008 0 0.089 0.034 ... ... ... ... ...
975 (1790)
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... 0.50 0.041 ... 0.204 0.225 0.029 ... ... 0.012 0 ... ... ... ... ... ... ...
1000 (1830)
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... 0.013 0 0.318 ... 0 ... ... ... ...
1025 (1880)
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... 0.049 ... 0 ... ... ... ... 0.107 ... ... ... ...
1050 (1920)
Temperature, °C (°F)
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... 0.013 ... ... ... ... 0.133 0.088 ... ... ...
1075 (1970)
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... 0.013 ... ... ... ... ... 0.107 ... ... ...
1100 (2010)
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... 0.037 ... ...
1125 (2060)
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... 0.053 ... ...
1150 (2100)
Table 4.12 Wettability index values for commercial filler metals used in brazing of molybdenum alloy TZM
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... 0.070 0.037 ...
1175 (2150)
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... 0.044 ...
1200 (2190)
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... 0.047 ...
1225 (2240)
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... 0.041
1275 (2330)
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... 0.048
1300 (2370)
98 / Brazing, Second Edition
Chapter 4: Base Metals and Base-Metal Family Groups / 99
react with the base metal under controlled conditions or adding base-metal powder to the filler-metal powder. One diffusion-sink filler metal is Ti-33Cr, requiring brazing temperatures of 1455 to 1480 °C (2650 to 2700 °F) (Ref 67). The other is a Ti-30V-4Be reactive filler metal that requires brazing temperatures of 1290 to 1315 °C (2350 to 2400 °F). Test results of Ti-33Cr brazed joints with D-36 niobium alloy indicate an increase in lap shear strength from approximately 17 MPa (2.5 ksi) to more than 31 MPa (4.5 ksi) at 1370 °C (2500 °F) and 7 MPa (1 ksi) at 1650 °C (3000 °F). Other niobium alloys—D43 (10W-1Zr-0.1CNb), Nb-752 (10W-2.5Zr-Nb), and C-129Y (10W-11Hf-0.07Y-Nb)—have been brazed successfully with two filler metals, B120VCA and Ti-8.5Si, whose brazing temperature was 1455 °C (2650 °F). Tantalum and Its Alloys. The filler metals for moderate-temperature applications of tantalum and its alloys are chosen on the basis of the intended application. Nickel-base filler metals (such as the Ni-Cr-Si filler metals) have been used in successful brazing of tantalum. These filler metals are satisfactory for service temperatures below 980 °C (1800 °F). Copper-gold alloys having less than 40% Au can also be used as filler metals, but gold in amounts between 40 and 90% tends to form brittle, age-hardening compounds. Copper-tin, gold-nickel, and copper-titanium filler metals also have been used in brazing tantalum and its alloys. Because most uses of tantalum and its alloys are in elevated-temperature applications (at or above 1650 °C, or 3000 °F), there is a lack of high-temperature filler metals for use with tantalum. Investigations have examined conventional brazing, reactive brazing, and diffusionsink brazing concepts for new tantalum filler metals. The reactive brazing concept uses a filler metal containing a strong melt-temperature depressant. The depressant is selected to react with the base material or powder additions to form a high-melting intermetallic compound during a postbraze diffusion treatment. As the depressant diffuses into the base material, the joint remelt temperature is increased. Successful application of this concept appears highly dependent on controlling the intermetallic compound reaction to form discrete particles. Diffusion-sink brazing with titanium and Ti30V filler metals, whose brazing temperatures range from 1675 to 1760 °C (3050 to 3200 °F), has produced remelt temperatures exceeding
2095 °C (3800 °F) in tee and lap joints. Diffusion-sink brazing with 33Zr-34Ti-33V filler metal with a brazing temperature of 1425 °C (2600 °F) has produced remelt temperatures exceeding 1760 °C (3200 °F). Therefore, these remelt temperatures indicate that service temperatures could be 1925 °C (3500 °F) for the titanium and Ti-30V filler metals and 1650 °C (3000 °F) for the 33Zr-34Ti-33V filler metal. Most of the filler metals currently available for tantalum are in powder form, which is difficult to work with at elevated temperatures. New powder filler metals, such as Hf-7Mo, Hf-40Ta, and Hf-19Ta-12.5Mo, have been used successfully but require further development and refinement. Special techniques are necessary to braze tantalum satisfactorily. All gases that have any reactivity must be removed. Oxygen and carbon monoxide must be eliminated. Tantalum forms oxides, carbides, nitrides, and hydrides very readily with these gases. A loss of ductility ensues. At high temperatures, tantalum should be protected from oxidation. One method is to electroplate the surfaces with copper or nickel, in which case it is necessary for the filler metal to be compatible with the plating. Controlledatmosphere and both hot- and cold-wall vacuum furnaces have been used for brazing tantalum and its alloys. Induction, resistance, and torch brazing are not recommended. Tungsten and Its Alloys. Tungsten has been successfully brazed in vacuum. Filler metals that have been used to braze tungsten as well as the other refractory metals are listed in Table 4.10. These filler metals and pure metals having liquidus temperatures ranging from 650 to 1925 °C (1200 to 3500 °F) are potentially useful for brazing. Tungsten can be brazed in much the same manner as molybdenum and its alloys, using many of the same filler metals. Brazing can be accomplished in a vacuum or in a dry argon, helium, or hydrogen atmosphere. To some extent, the selection of the brazing atmosphere depends on the filler metal used. For example, filler metals that contain elements with high vapor pressures at the brazing temperature cannot be used effectively in a high vacuum. Care must be exercised in handling and fixturing tungsten parts because of their inherent brittleness; these parts should be assembled in a stress-free condition. Contact between graphite fixtures and tungsten must be avoided to prevent the formation of brittle tungsten carbides.
100 / Brazing, Second Edition
Although nickel-base filler metals have been used successfully to braze tungsten, a reaction between nickel and tungsten that results in basemetal recrystallization can occur; this reaction can be minimized by short brazing cycles, minimum brazing temperatures, and the use of small quantities of filler metal. Data on brazing of tungsten are more limited than data on brazing of other refractory metals. Unalloyed tungsten has been vacuum brazed with two experimental filler metals, Nb-2.2B and Nb-20Ti. The lap shear strength of joints brazed with the Nb-2.2B filler metal was approximately 35 MPa (5 ksi) at 1650 °C (3000 °F) and 55.2 MPa (8 ksi) at 1370 °C (2500 °F). The strength of joints brazed with Nb-20Ti was somewhat lower, approximately 21 MPa (3 ksi) at both test temperatures. The flow of Nb-20Ti filler metal was more sluggish than that obtained with Nb-2.2B. Filler metals based on the platinum-boron and iridium-boron systems were developed to braze tungsten for service at 1925 °C (3500 °F). They contained up to 4.5% B and could be used for brazing below the recrystallization temperature of tungsten. Tungsten lap joints were brazed and diffusion treated in a vacuum at 1095 °C (2000 °F) for 3 h. This cycle resulted in the production of joints with remelt temperatures of approximately 2040 °C (3700 °F). A slight increase in joint remelt temperature was noted when tungsten powder was added to the filler metal. The highest remelt temperature of 2170 °C (3940 °F) was obtained when joints were brazed with Pt3.6B plus 11 wt% W powder (Ref 69). Studies have been conducted to develop and evaluate filler metals that could be used to braze tungsten for nuclear reactor service at 2500 °C (4530 °F) in hydrogen atmosphere. Butt joints have been brazed using a gas tungsten arc as the heat source and the filler metals W-25Os, W50Mo-3Re, and Mo-5Os. Furnace brazing (inert-gas or reducing atmosphere and vacuum), torch brazing, resistance brazing, or induction brazing can be used to braze tungsten. Furnace brazing operations usually are used when the parts are larger than those considered practical for the induction and resistance processes. Tungsten-copper electrode tips are materials made by P/M processes. A normal analysis of one of the materials is 80W-20Cu. These tips may be used for resistance welding electrodes that are usually brazed to the copper-chromium alloys, which constitute the main current-carry-
ing portion of the electrode. Any of the silvercopper and/or zinc series of filler metals are suitable for this application. Tungsten-copper tips usually are applied by induction or torch heating in a neutral or reducing atmosphere. The most suitable filler metal is 72Ag-28Cu. To hold parts in alignment, pins are incorporated in the design, and the filler metal is preplaced in the form of shim stock located between the mating parts. In brazing of resistance welding electrodes with pure tungsten tips, precoating of the tungsten surface aids uniform flow of the filler metal between the faying surfaces of the joint. Usually, one of the silver-base filler metals is used as a precoating. Without pretreatment, poor wetting of the tungsten by the filler metal often results, causing premature joint failure under the high pressures and temperatures of resistance welding.
Cast Iron Brazing of gray, ductile, and malleable cast irons differs from brazing of steel in two principal respects: special precleaning methods are necessary to remove graphite from the surface of the iron, and the brazing temperature is kept as low as feasible to avoid reduction in the hardness and strength of the iron. The processes used for brazing of cast irons are the same as those used for brazing of steel, which include furnace, torch, induction, and dip brazing. As with other metals, selection of the brazing process depends largely on the size and shape of the assembly, the quantity of assemblies to be brazed, and the equipment available. In recent years, brazing of ordinary gray cast irons using silver-base filler metal has become practical commercially; the development of a satisfactory surface pretreatment has opened up many design possibilities. Intricate forms can be built from simple castings or by joining castings to standard wrought forms, such as tubing and rolled shapes. Foundry work may be simplified by making intricate castings in several parts to reduce coring. Gray, ductile, and malleable irons commonly are brazed to steels or to other cast irons. In brazing of ductile or malleable irons, certain precautions are imperative. If ductile or malleable irons are heated above 760 °C (1400 °F), the metallurgical structure may be damaged; brazing thus should be done below this temperature. The effects of graphite, which is present in all gray, ductile, and malleable cast irons, are sig-
Chapter 4: Base Metals and Base-Metal Family Groups / 101
nificant. Although gray, ductile, and malleable irons all have lower brazeability than carbon or low-alloy steels, the three types of iron are not equal in brazeability. Malleable iron is generally considered the most brazeable of the three types of cast iron, largely because the total carbon content is somewhat lower (seldom over 2.70%), and, therefore, graphitization is lower. Brazeability is also enhanced because the graphite occurs in the form of approximately round nodules and thus is easier to remove or cover up (as by abrasive blasting). Also, malleable iron is lower in silicon than the other types of cast iron and thus is less graphitized, which makes it better suited for brazing. Ductile iron can have a composition nearly the same as gray iron, but the graphite particles are spheroidal rather than flake-shaped. The spheroidal shape is more favorable for brazing. Shot or grit blasting is effective in rolling metal over graphite particles that are exposed at the surface. Gray iron, which is characterized by large flakes of graphite, is the most difficult type of cast iron to braze. Until the development of electrolytic salt bath cleaning, brazing of gray iron was considered impractical. Abrasive blasting with steel shot or grit has proved reasonably successful for preparing surfaces of ductile and malleable iron castings but is seldom suitable for preparing surfaces of gray iron castings. Electrolytic treatment in a molten salt bath, alternately reducing and oxidizing, has been the most successful method for surface preparation and is applicable to all graphitic cast irons. Ordinary chemical cleaning methods, such as degreasing, detergent washing, and acid pickling, have the distinct disadvantage of not removing surface carbon, which interferes with bonding. Before any procedure for cleaning is adopted, tests should be made by cleaning samples of the iron intended for use in the castings to be brazed, fluxing the samples, and applying filler metal (preferably on a smooth, flat surface). The samples are then heated to the preestablished brazing temperature, cooled, and examined visually. If the samples show an indication that the filler metal has not uniformly wetted the testpiece, the surface is not sufficiently clean. Alkaline-based salts (sodium hydroxide) operating at 400 to 480 °C (750 to 900 °F) are extremely effective in removing surface oxides and sand from iron castings. At these temperatures, salt baths exhibit the required high chem-
ical activity. This action is further enhanced in the cleaning operations by the introduction of electrical energy. The effective use of these salts and the removal of their by-products to produce an excellent metallurgical surface has resulted in quality bonds and has permitted direct babbitting of cast iron bearings and silver brazing fittings to cast iron surfaces. With proper cleaning and preparation of the cast iron, it has been possible to produce a brazed bond between metal surfaces that exceeds the strength of the parent metal. With proper surface preparation, the usual variety of filler metals suitable for steel brazing is satisfactory for cast irons. In selection of the filler metal, the metallurgical effects of the brazing cycle on the base-metal matrix (hardening, softening, etc.) should be considered. Various filler metals in the silver-copper series are commonly used. Copper filler metal has been used successfully in furnace brazing. Some cast irons have low melting (solidus) temperatures, suggesting that care must be exercised to keep the brazing temperature below 1120 °C (2050 °F). Brass filler metals (copper-zinc series) frequently are used to braze cast irons, while the silver-base filler metals have been used to join copper to cast iron. Most brazing of cast iron is performed to join assemblies at lower cost than is possible by another process or to fabricate parts that are difficult to produce—one-piece castings, for example. In some applications, two or more cast iron components are brazed together; in other applications, one or more components of a brazed assembly are made of another metal— most often steel but also copper alloys. Localized heating methods, such as torch and induction, have the advantages of restricted heat effects and limited metallurgical phase changes in base-metal components. Furnace brazing may be desirable for production lots due to cost savings and/or consistency of results. Brazing with copper and nickel filler metals require atmosphere controls during heating, usually accomplished by furnace brazing. The use of dry hydrogen as a protective atmosphere promotes wetting by reducing oxides such as silica. Electric arc braze welding processes have successfully been used with bronze-brass filler metals on cast iron and require special preheating techniques to prevent cracking. Cast irons that contain pearlite or free carbide graphitize and decrease in strength at elevated temperatures. Because graphitization is a func-
102 / Brazing, Second Edition
tion of time and temperature, some experimentation is usually necessary to develop a brazing cycle that produces acceptable joints without excessive graphitization and decreases in strength. Two identical assemblies were heated in different plants for 1 h. In the first plant, the assemblies were heated to approximately 705 °C (1300 °F) with no significant decrease in strength. In the second plant, the assemblies were heated to approximately 790 °C (1450 °F), again without significant decrease in strength. In the latter application, however, heating was done by induction instead of in a furnace, and the heating time was much shorter than would have been required in a furnace. In some plants using furnace brazing, the furnace is operated at a considerably higher temperature than is desired for brazing (sometimes as high as 870 °C, or 1600 °F), but the time cycles used are so short that the assemblies never reach furnace temperature. For applications in which little or no decrease in strength of the cast iron can be tolerated, it is mandatory to use a filler metal with as low a flow-temperature range as possible, thus permitting a low brazing temperature, and to keep the time at brazing temperature to the minimum. The temperature required for wetting the base metal and for flow of filler metal having a melting range of 620 to 645 °C (1150 to 1195 °F) may vary from approximately 690 to 845 °C (1275 to 1555 °F), depending on the complexity of the joint design—especially the distance the filler metal must flow. Simple joints with short flow distances can be brazed at lower temperatures than more complex joints.
Low-Carbon, Low-Alloy, and Tool Steels Brazing of low-carbon and low-alloy steels is a highly developed, low-cost production process. Filler metals can be either manually or automatically applied or preplaced in the joint. The heat treatment considerations of low-alloy steels are factors in determining the specific filler metals and brazing temperatures to be used. The steels covered in this section include low-carbon (less than 0.30% C) and low-alloy steels. The low-alloy steels include the Society of Automotive Engineers/American Iron and Steel Institute (SAE/AISI) 23xx nickel steels, 31xx nickel-chromium steels, 41xx chromiummolybdenum steels, 43xx nickel-chromium-
molybdenum steels, and several other types containing less than 5% total alloy content. Leaded steels can be torch brazed using filler metal and flux combinations normally recommended for steel. Tests conducted on low-carbon free-machining steel containing 0.25 to 0.35% Pb show detrimental effects after torch brazing with silver-base filler metals; satisfactory joints have been made in these leaded steels using copper- and nickel-base filler metals. All the silver-base filler metals can be used for brazing low-carbon and low-alloy steels (Table 4.13). The silver-base filler metals containing nickel usually provide better wettability and are preferred for brazing low-alloy steels wherein joint strength is most important. Copper-base filler metals are used mainly for preplacement in controlled-atmosphere furnaces. Copper is also generally preferred because of its low cost and the high strength of the joints produced. Copper-zinc filler metals often are face fed into the joint but can also be preplaced for furnace and induction heating. The high solidus temperatures (1095 to 1150 °C, or 2000 to 2100 °F) necessary when copperbase filler metals are used often allow simultaneous brazing and heat treating of certain lowalloy steels. Nickel-base filler metals have been used for joining low-carbon and low-alloy steels when special joint requirements so dictate. Brazing usually is restricted to controlledatmosphere furnaces. Low-carbon and low-alloy steels can be brazed using virtually all known processes. Torch, furnace, and induction heating techniques are the most common. Filler metals in the form of continuous wire or strip can be automatically applied using electromechanical wire feeders; powder filler metals blended with flux and pasteforming ingredients are applied automatically with pressurized dispensing equipment. For torch brazing, the equipment would include standard oxyacetylene or similar torches. Furnaces can be of the batch or conveyor type, with or without atmosphere control; they can be electric, gas, or oil fired and should provide accurate temperature control. The principal advantage of furnace brazing over other brazing processes is that it permits the use of a variety of prepared protective atmospheres—notably, the rich exothermicbased, endothermic-based, and some prepared and commercial nitrogen-base atmospheres. These atmospheres are among the least expen-
Strip, rod, wire, powder Strip, rod, wire, powder Rod Rod
Strip, wire, powder Strip, wire, powder Strip, wire, powder Strip, wire, powder Strip, wire, powder Strip, wire, powder Strip, wire, powder Strip, wire, powder Strip, wire, powder Strip, wire, powder Strip, wire, powder Strip, wire, powder
Product form
... ... ... ...
45 50 35 30 50 40 45 50 56 30 25 40
Ag
59 48 58 58
15 15.5 26 27 15.5 30 30 34 22 38 35 30
Cu
40 41 38 39
16 16.5 21 23 15.5 28 25 16 17 32 26.5 28
Zn
... ... ... ...
... 10.0 0.5 ...
... ... ... ... 3.0 2.0 ... ... ... ... ... ...
Ni
0.6 ... 0.95 0.95
... ... ... ... ... ... ... ... 5.0 ... ... 2
Sn
Fe
... ... 0.7 0.7
... ... ... ... ... ... ... ... ... ... ... ...
Nominal composition, %
24 18 18 20 16 ... ... ... ... ... 13.5 ...
Cd
... ... 0.25 0.25
... ... ... ... ... ... ... ... ... ... ... ...
Mn
... 0.15 0.08 0.08
... ... ... ... ... ... ... ... ... ... ... ...
Si
... 0.025 ... ...
... ... ... ... ... ... ... ... ... ... ... ...
P
890 920 865 865
605 625 605 605 630 670 665 690 620 675 605 650
°C
1630 1690 1590 1590
1125 1160 1125 1125 1170 1240 1225 1270 1145 1250 1125 1200
°F
900 935 880 890
620 635 700 710 690 780 745 775 650 765 745 710
°C
1650 1715 1620 1630
1145 1175 1295 1310 1270 1435 1370 1425 1205 1410 1375 1310
°F
Solidus temperature Liquidus temperature
910–955 940–980 ... 815–925
620–760 635–760 700–845 710–845 690–815 780–900 745–845 775–870 650–760 765–870 745–860 710–845
°C
1670–1750 1720–1800 ... 1500–1700
1145–1400 1175–1400 1295–1550 1310–1550 1270–1500 1435–1650 1370–1550 1425–1600 1205–1400 1410–1600 1375–1575 1310–1550
°F
Brazing temperature
Note: AWS, American Welding Society (a) Classified for braze welding and brazing. (b) Classified for braze welding. Source: Abstracted from the mandatory and nonmandatory sections of AWS A5.7, AWS A5.8, and Ref 65
RBCuZn-A(a) RBCuZn-D(a) RCuZn -B(b) RCuZn-C(b)
Copper-zinc alloys
BAg-1 BAg-1a BAg-2 BAg-2a BAg-3 BAg-4 BAg-5 BAg-6 BAg-7 BAg-20 BAg-27 BAg-28
Silver alloys(a)
AWS classification
Table 4.13 Filler metals for torch brazing low-carbon and low-alloy steels
Chapter 4: Base Metals and Base-Metal Family Groups / 103
104 / Brazing, Second Edition
sive; they can be generated in the plant in large volume, or, in the case of commercial nitrogenbase atmospheres, they can be stored in liquid form outside the plant. They provide excellent protection against oxidation, and they can be prepared with any carbon potential in the range of approximately 0.2% to more than 1.0% C, depending on the atmosphere. This range of carbon potential is sufficient to accommodate all carbon and low-alloy steels, including those carburized before brazing. By selecting an atmosphere with a carbon potential that matches the carbon content of the work metal, brazing can be accomplished without carburizing or decarburizing the work metal. Because the protective atmospheres used for furnace brazing are sufficiently reducing to iron oxide, they usually eliminate the need for fluxes in brazing carbon steel with copper filler metal. An oxide-free surface normally promotes wetting of the workpiece by the molten filler metal. However, some low-alloy steels that contain a total of more than 2 or 3% Cr, Mn, Al, and Si form more stable surface oxides, and they require highly reducing atmospheres (such as dry hydrogen or dissociated ammonia), a flux, or nickel plating to obtain adequate wetting. Most of the limitations of furnace brazing are directly related to the high temperatures required for brazing of steels with copper filler metal. These temperatures exceed the average brazing temperature required for brazing with silver-base filler metals by 280 °C (500 °F) or more. They are high enough to cause grain coarsening in medium-carbon, high-carbon, and low-alloy steels; however, grain refinement can be obtained by subsequent heat treatment. Steels brazed with copper filler metal develop lower tensile and yield strengths and increased ductility as the brazing time or temperature, or both, are increased. These changes in properties are a result of decarburization in some types of atmospheres and alteration of grain size. Original grain size can be restored by subsequent heat treatment below the remelt temperature of the copper brazing filler metal. Loss of carbon through decarburization is generally unimportant in low-carbon steels. However, surface hardness of some low-alloy steels may be substantially lowered. Such loss of surface hardness can be very deleterious when thin-gage material is involved. For alloy steels, the filler metal should have a solidus well above any heat treating tempera-
ture to avoid damage to joints that are heat treated after brazing. In some cases, air-hardening steels can be brazed and then hardened by quenching from the brazing temperature. A filler metal with brazing temperature lower than the critical temperature of the steel can be used when no change in the metallurgical properties of the base metal is wanted. In torch brazing, a neutral or slightly reducing flame usually is preferred, because the filler metal is face fed into the prefluxed joint. Fluxcoated filler metal is often beneficial. As in all brazing processes, it is important to avoid overheating in brazing to prevent undesirable effects on the base metal, the filler metal, or the flux. Time at temperature is an important consideration, especially when the filler metal contains volatile elements such as zinc and cadmium. Excessive heat also might affect the integrity of the braze and reduce the mechanical properties of the joint. Automated torch and burner-type production equipment is available for high-production applications. These units usually use brazing fluxes and silver-base filler metals. In production applications, the filler metal (usually copper) is preplaced in, or adjacent to, the joint before the preassembled parts are charged into controlled-atmosphere batch- or conveyor-type furnaces for brazing. Induction heating for brazing is also advantageous in that maximum temperature rise is restricted to the immediate joint area by selective coil design and the use of an appropriate frequency for the induction-heating circuit. In joining certain hardenable low-alloy steels, it usually is desirable to use the lowermelting silver-base filler metals, which can be applied below the lower transformation temperature of the steel. Some local annealing, however, may occur. When postheat treatments are required, the higher-solidus filler metals are necessary to preclude the possibility of joint impairment. Low-alloy steels are sometimes brazed and heat treated simultaneously using filler metals of the copper-zinc and silver classifications. The solidus temperature of the filler metal chosen must be above the austenitizing temperature recommended for the base metal prior to quenching. In an application of this type, the joint is made at normal brazing temperature, removed from the heat source to permit a drop to the hardening temperature, and quenched. The procedure is satisfactory only for base metals compatible with rapid cooling.
Chapter 4: Base Metals and Base-Metal Family Groups / 105
Dip brazing has also been used to join carbon and low-alloy steels with silver-base, copperzinc, and other copper-base filler metals. The types of salts used in dip brazing of carbon and low-alloy steels are neutral chloride salts, neutral chloride salts plus a fluxing agent such as borax or cryolite, and carburizing and cyaniding salts, which are also fluxing-type salts. Types and compositions of brazing salts and temperatures used for brazing of carbon and low-alloy steels with various filler metals are given in Table 4.14. Neutral salts, so called because normally they do not add or subtract anything from the surface of the steel being treated, protect the surface from attack by oxygen in the air. Oxide on the workpiece, however, cannot be reduced by the salt, and a flux must generally be provided. The neutral salts are mildly oxidizing to steel when they are used at recommended austenitizing temperatures. The oxides produced by heating steel in molten salt are largely soluble; hence, the steel is scale-free after heating. The accumulation of oxide in the molten salt, however, progressively makes the salt more decarburizing, and for this reason, baths may require periodic rectification (Ref 70). Carburizing and cyaniding salts provide their own fluxing action. In addition, they supply carbon or carbon and nitrogen to the surface of the steel assembly as it is being brazed. Although silver-base filler metals have been used successfully, copper-zinc filler metal is generally preferred. Various applications involve the use of filler metals in brazing low-carbon and low-alloy steels in everyday production. These include
many components of such vehicles as automobiles, trucks, bicycles, motorcycles, snowmobiles, all-terrain vehicles, and the like. Other common brazements include window and door frames, ducts, tanks, containers of all types, perforated and expanded steel panels, steel partitions, and shelving. A great variety of tubular steel furniture is production brazed, and brazing is also used in production of cutting tools and industrial knives, hydraulic oil tanks, reservoirs, electronic chassis and supports, hand tools, honing appliances, certain instruments, and steel assemblies of all types. Tool Steels. High-carbon steels contain more than 0.45% C. High-carbon tool steels usually contain 0.60 to 1.40% C. In discussing the brazing of tool steels, it is convenient to group them in two broad classifications: carbon steels and high-speed tool steels. Brazing of high-carbon steels is best accomplished prior to or during the hardening operation. Hardening temperatures for carbon steels range from 760 to 820 °C (1400 to 1510 °F). Filler metals having brazing temperatures above 820 °C should be used. When brazing and hardening are done in one operation, the filler metal should have a solidus at or below the austenitizing temperature. Tempering and brazing can be combined for high-speed tool steels and high-carbon, highchromium alloy tool steels that have tempering temperatures in the range of 540 to 650 °C (1000 to 1200 °F). Filler metals with brazing temperatures in that range are used. The part is removed from the tempering furnace, brazed by localized heating methods, and then returned to the furnace for completion of the tempering cycle.
Table 4.14 Typical salts used for dip brazing of carbon and low-alloy steels with various brazing filler metals Brazing temperature range (a) Filler metal
BAg-1 through BAg-8, and BAg-18 RBCuZn-A
RBCuZn-D
Type of salt
Neutral Cyaniding-fluxing Neutral Neutral Fluxing Carburizing-fluxing (water soluble) Carburizing and self-fluxing Neutral Neutral Neutral
(a) Temperatures shown are those of the salt bath. Source: Ref 65
°C
°F
621–871 649–871 732–871 913–940 913–940 913–940 816–927 1038–1051 1093–1149 1093–1149
1150–1600 1200–1600 1350–1600 1675–1725 1675–1725 1675–1725 1500–1700 1900–1925 2000–2100 2000–2100
106 / Brazing, Second Edition
Localized heating for brazing may decrease the hardness of heat treated steels when the brazing temperature is above the tempering temperature of the steel. Except for thin sections, these steels must be quenched drastically during heat treatment to achieve optimal properties. Alloying elements may be added to carbon steels to impart special properties, such as reduced distortion on heat treatment, greater wear resistance and toughness, or better hightemperature properties. Such steels are referred to as alloy tool steels; they are known by various trade names and grades, and their properties are adequately covered in manufacturers’ published information and in various handbooks. The alloy steel in question should be studied carefully to determine its proper heat treating cycle, the kind of quench necessary (water, oil, or air), the best filler metal, and the proper technique for combining the heat treating and brazing operations to achieve maximum properties in service life. High-speed steels, although logically falling in the alloy steel group, are classified separately because their properties depend on relatively high percentages of such alloying elements as tungsten, molybdenum, chromium, and vanadium. Their carbon contents normally are much lower than those of carbon tool steels. Highspeed steels are widely used in industry as metal cutting tools. A common type known as 18-4-1 contains 18% W, 4% Cr, and 1% V. The choice of the filler metal to be used depends on the properties of the tool steel being brazed and the heat treatment required to develop its optimal properties. Practically all filler metals of the BAg, BCu, and RBCuZn classifications are used at various times. The best filler metal to use should be determined for the specific application. Torch, furnace, and induction heating are the three most commonly used processes in brazing tool steels. Available equipment is frequently the main factor in process selection.
Stainless Steels Stainless steels include a wide variety of ironbase alloys containing chromium that are used primarily for applications demanding heat or corrosion resistance. Tighter process controls are required than for brazing of carbon steels. These more rigorous requirements are imposed by the inherent chemical characteristics of
stainless steels and the generally more arduous service environments in which they are used. Success in the fabrication of stainless steel components by brazing depends on knowledge of the characteristics of the various types of stainless steels, and rigid adherence to certain items of process control required by these characteristics. All stainless steels are difficult to wet because of their high chromium contents. Brazing of these alloys is best accomplished in a purified (dry) hydrogen atmosphere or in a vacuum. Dewpoints of –50 °C (–60 °F) or lower must be maintained, because problems with wetting may arise following the formation of chromium oxide. In torch brazing of these base metals, fluxes are required to reduce any chromium oxides present. Austenitic nonhardenable stainless steels contain sufficient nickel or nickel plus manganese additions to stabilize austenite down to room temperature, causing these alloys to be nonmagnetic and nonhardenable by heat treatment. Stainless steels of this class possess the highest heat and corrosion resistance. They are designated as AISI 300- and 200-series alloys. One commonly used alloy is type 302, which contains approximately 18% Cr and 8% Ni. In the 200-series stainless steels, some of the nickel is replaced by manganese. For example, type 202, the parallel to type 302, contains 18% Cr, 5% Ni, and 9% Mn. The 300-series stainless steels are used widely for both torch- and furnace-brazed assemblies. These alloys have relatively high thermal expansion and low thermal conductivity. This combination of properties makes thermal distortion a major concern in furnace brazing of large or complex assemblies or assemblies in which dissimilar materials are brazed to stainless steel. In design of fixtures, heat shields, and thermal cycles, the requirement to provide uniform heating and cooling must be considered. In brazed assemblies for which corrosion resistance is important, precautions must be taken to avoid sensitization to intergranular corrosion. This problem was discussed earlier in this chapter, but briefly, it occurs when an unstabilized grade of austenitic stainless steel, such as type 302 or type 304, is held at temperatures in the range from 425 to 815 °C (800 to 1500 °F) or slowly cooled through this range. The excess carbon combines with chromium and precipi-
Chapter 4: Base Metals and Base-Metal Family Groups / 107
tates as chromium carbide along grain boundaries of the austenite. The region around the precipitate is depleted of chromium and thus becomes susceptible to corrosion. The 300-series stainless steels have been used widely for brazed components; the 200series alloys, which are only 25 years old, have not seen as much use. Due to their high manganese content, the 200-series stainless steels are more difficult to furnace braze in hydrogen atmospheres than the 300 series. Manganese forms an oxide that is not reduced easily by dry hydrogen at the furnace brazing temperatures normally used for stainless steels. It is important to start with thoroughly cleaned surfaces and to maintain a low dewpoint in the hydrogen atmosphere. All the chromium-nickel steels are subject to SCC in the presence of molten filler metals. This phenomenon occurs when the base metal is under stress—either residual or resulting from applied loads while the braze is being made. One form of applied load results from thermal gradients during brazing. The filler metal penetrates the base metal along the grain boundaries at the points of stress, producing a greatly weakened base metal. Best results, therefore, are obtained with stress-relieved material. This stress relief may be done prior to or during the brazing cycle. If it occurs during the cycle, it must be done below the solidus temperature of the filler metal (Ref 71, 72). Ferritic nonhardenable stainless steels are basically low-carbon alloys of iron and chromium in which sufficient chromium has been added to the iron to stabilize ferrite, the low-temperature phase in steels, over a wide temperature range. The more common AISI stainless grades in this category are types 405, 406, 430, and 446. Type 430 is a widely used grade that is particularly subject to a form of interface corrosion when brazed with some silver-base filler metals. This corrosion apparently is caused by electrochemical action whereby the bond between the base metal and the filler metal is destroyed. In many cases, this action has been found to occur in the presence of tapwater. It has been found that addition of small percentages of nickel to silver-base filler metal prevents interface corrosion of brazed joints in most stainless steels. The nickel-containing silver-base filler metal is not completely effective, however, with type 430, even though its use greatly reduces the rate of attack. For type 430, a special silver-base
filler metal has been developed (63Ag-28.5Cu6Sn-2.5Ni). The ferritic stainless steels (405, 406, and 430) cannot be hardened, and their grain structure cannot be refined by heat treatment. These alloys degrade in properties when brazed at temperatures above 980 °C (1800 °F), because of excessive grain growth. They lose ductility after long heating times between 340 and 600 °C (645 and 1110 °F). However, some of the ductility can be recovered by heating the brazement to approximately 790 °C (1455 °F) for a suitable time (Ref 71, 72). Martensitic hardenable stainless steels are Fe-C-Cr alloys of two basic types: the lowchromium, low-carbon grades (types 403, 410, and 416) and the high-chromium, high-carbon grades (types 440A, B, and C). These steels are closely related to the ferritic nonhardenable grades, but their alloy compositions are so balanced that they air harden on cooling from brazing; this hardening occurs at a temperature above their austenitizing temperature range. Therefore, they must be annealed after brazing or during the brazing operation. These steels are also subject to stress cracking with certain filler metals. The primary precaution in brazing components made of these alloys is that the brazing thermal cycle must be compatible with the required heat treatment. If high-temperature nickel-base filler metals are used, it is possible to reaustenitize the assembly after brazing. Although this procedure would increase costs, it may be desirable to develop optimal properties in critical components. The thermal expansion properties of martensitic stainless steels are relatively low—in the same range as those of ferritic alloys (Ref 73). Precipitation-hardening stainless steels are basically stainless steels with additions of one or more of the elements copper, molybdenum, aluminum, and titanium. Such alloying additions make it possible to strengthen the alloys by precipitation-hardening heat treatments. Some of the designations of the precipitationhardening stainless steels are 17-7PH, PH148Mo, PH13-8Mo, 15-5PH, PH15-7Mo, AM350, AM355, 17-4PH, and A-286. As in the case of the martensitic hardenable stainless steels, brazing thermal cycles used in joining these alloys must be compatible with their heat
108 / Brazing, Second Edition
treatments. Because the heat treatments vary rather widely, specific brazing procedures are required for each alloy. Precipitation-hardening alloys that contain aluminum or titanium are difficult to wet in the usual furnace brazing atmospheres. Nickel plating generally is used as a surface treatment before furnace brazing. Metallurgical Considerations in Brazing Stainless Steels. In brazing of stainless steels, base-metal inclusions and surface contaminants are even more deleterious than in brazing of carbon steels. Base-metal inclusions, such as oxides, sulfides, and nitrides, interfere with the flow of filler metal. Flow is also impeded by surface contaminants, which may include lubricants such as oil, graphite, molybdenum disulfide, and lead that are applied during machining, forming, and grinding, or by aluminum oxide particles produced by grit blasting or by grinding with aluminum oxide wheels or belts. Some filler metals in powder form contain or are mixed with an organic binder to form a paste that is subsequently applied to stainless steel for brazing; acrylics and other plastics are often used for this purpose. Although some binders form a soot residue, this residue does not usually interfere with filler-metal flow. The brazing characteristics of stainless steels can also be severely impaired by unsuitable fixturing materials, such as graphite, or by protective atmospheres with nitriding potentials. Carbon in graphite fixtures unites with hydrogen to form methane (CH4), which carburizes stainless steel and impairs its corrosion resistance. Dissociated ammonia, unless sufficiently dry and completely (100%) dissociated, nitrides stainless steel. Filler Metals for Brazing Stainless Steels. A wide variety of filler metals, including silver-base, nickel-base, gold-base, and copper, are commercially available for brazing stainless steel parts (Tables 4.15, 4.16). The factors to consider in selecting a filler metal for a particular application include the following: • Service conditions, including operating temperature, stresses, and environments • Heat treatment requirements if martensitic or precipitation-hardening steels are involved • Brazing process to be used • Cost • Special precautions, such as sensitization of unstabilized austenitic stainless steels at certain temperatures
Commercial filler metals are available that have copper, silver, nickel, cobalt, platinum, palladium, manganese, and gold as the base or as addition elements; these are grouped conveniently according to service temperature. The most widely used filler metals for brazing stainless steels are the silver-base family (BAg group). Filler metal BAg-3, which contains 3% Ni, is probably the silver-base filler metal selected most frequently, although several other silver-base filler metals can also be used successfully. Where improved corrosion resistance is needed, BAg-3 and BAg-4 are recommended. Silver-brazed joints cannot be used for hightemperature service; the recommended maximum service temperature is 370 °C (700 °F), which is the maximum temperature for BAg-13 and BAg-19 filler metals. Of the silver-base filler metals shown in Table 4.15, all except BAg-19 and possibly BAg-13 are used at brazing temperatures that fall within the effective sensitizing range for austenitic stainless steels (540 to 870 °C, or 1000 to 1600 °F). Chromium carbide precipitation occurs in the sensitizing temperature range, resulting in impairment of the corrosion resistance of the base metal. Carbide precipitation, however, depends on time as well as temperature, and exposure within the sensitizing temperature range for only a few minutes is unlikely to produce a significant amount of precipitate. Nevertheless, the lower melting temperatures of the silver-base filler metals prohibit re-solution treatment of the base metal after brazing, and if corrosion resistance in service is sufficiently critical, an extra-lowcarbon, titanium-stabilized or niobium-tantalum-stabilized austenitic stainless steel should be selected instead of a nonstabilized type. Ferritic and martensitic stainless steels that contain little or no nickel are susceptible to interface corrosion in plain water or moist atmospheres when brazed with nickel-free silver-base filler metals using a flux. Filler metals containing nickel (BAg-3) help to prevent interface corrosion. However, for complete protection, special filler metals containing nickel and tin should be used, and brazing should be done in a protective atmosphere without flux. The nickel-base filler metals usually rank second in frequency of use as filler metals for brazing of stainless steels. Nickel-base filler metals provide joints that have excellent corrosion resistance and high-temperature strength. These filler metals alloy with stainless steel,
BAg-1 BAg-1a BAg-2 BAg-2a BAg-3 BAg-4 BAg-5 BAg-6 BAg-7 BAg-8 BAg-8a BAg-9 BAg-10 BAg-13 BAg-13a BAg-18 BAg-19 BAg-20 BAg-21 BAg-22 BAg-23 BAg-24 BAg-25 BAg-26 BAg-27 BAg-28
Filler metal
44.0–46.0 49.0–51.0 34.0–36.0 29.0–31.0 49.0–51.0 39.0–41.0 44.0–46.0 49.0–51.0 55.0–57.0 71.0–73.0 71.0–73.0 64.0–66.0 69.0–71.0 53.0–55.0 55.0–57.0 59.0–61.0 92.0–93.0 29.0–31.0 62.0–64.0 48.0–50.0 84.0–86.0 49.0–51.0 19.0–21.0 24.0–26.0 24.0–26.0 39.0–41.0
Ag
14.0–16.0 14.5–16.5 25.0–27.0 26.0–28.0 14.5–16.5 29.0–31.0 29.0–31.0 33.0–35.0 21.0–23.0 bal bal 19.0–21.0 19.0–21.0 bal bal bal bal 37.0–39.0 27.5–29.5 15.0–17.0 ... 19.0–21.0 39.0–41.0 37.0–39.0 34.0–36.0 29.0–31.0
Cu
14.0–18.0 14.5–18.5 19.0–23.0 21.0–25.0 13.5–17.5 26.0–30.0 23.0–27.0 14.0–18.0 15.0–19.0 ... ... 13.0–17.0 8.0–12.0 4.0–6.0 ... ... ... 30.0–34.0 ... 21.0–25.0 ... 26.0–30.0 33.0–37.0 31.0–35.0 24.5–28.5 26.0–30.0
Zn
23.0–25.0 17.0–19.0 17.0–19.0 19.0–21.0 15.0–17.0 ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... 12.5–14.5 ...
Cd
... ... ... ... 2.5–3.5 1.5–2.5 ... ... ... ... ... ... ... 0.5–1.5 1.5–2.5 ... ... ... 2.0–3.0 4.0–5.0 ... 1.5–2.5 ... 1.5–2.5 ... ...
Ni
Composition, %
... ... ... ... ... ... ... ... 4.5–5.5 ... ... ... ... ... ... 9.5–10.5 ... ... 5.0–7.0 ... ... ... ... ... ... 1.5–2.5
Sn
... ... ... ... ... ... ... ... ... ... 0.25–0.50 ... ... ... ... ... 0.15–0.30 ... ... ... ... ... ... ... ... ...
Li
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... 7.0–8.0 bal ... 4.5–5.5 1.5–2.5 ... ...
Mn
0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15
Other elements (total)
607 627 607 607 632 671 677 688 618 779 766 671 690 713 771 601 760 677 690 682 960 660 738 707 607 649
°C
1125 1160 1125 1125 1170 1240 1250 1270 1145 1435 1410 1240 1275 1325 1420 1115 1400 1250 1275 1260 1760 1220 1360 1305 1125 1200
°F
Solidus temperature
618 635 701 710 688 779 743 774 651 779 766 713 738 857 893 713 885 766 801 699 971 707 790 801 746 710
°C
1145 1175 1295 1310 1270 1435 1370 1425 1205 1435 1410 1325 1360 1575 1640 1325 1635 1410 1475 1290 1780 1305 1455 1475 1375 1310
°F
Liquidus temperature
Table 4.15 Typical composition and properties of standard filler metals for brazing stainless steels (silver base)
618–760 635–760 701–843 710–843 688–816 779–899 743–843 774–871 651–760 779–899 766–871 713–843 738–843 857–969 871–982 713–843 877–982 766–871 801–899 699–830 971–1038 707–843 790–846 801–871 746–857 710–843
°C
1145–1400 1175–1400 1295–1550 1310–1550 1270–1500 1435–1650 1370–1550 1425–1600 1205–1400 1435–1650 1410–1600 1325–1550 1360–1550 1575–1775 1600–1800 1325–1550 1610–1800 1410–1600 1475–1650 1290–1525 1780–1900 1305–1550 1455–1555 1475–1600 1375–1575 1310–1550
°F
Brazing temperature range
Chapter 4: Base Metals and Base-Metal Family Groups / 109
Cr
Filler metal
BNi-1 BNi-1a BNi-2 BNi-3 BNi-4 BNi-5 BNi-6 BNi-7 BNi-8
13.0–15.0 13.0–15.0 6.0–8.0 ... ... 18.5–19.5 ... 13.0–15.0 ...
Nickel alloys
18.0–20.0
Cr
BCo-1
Cobalt alloys
Filler metal
BAu-1 BAu-2 BAu-3 BAu-4 BAu-5 BAu-6
Precious-metal alloys
Filler metal
BCu-1 BCu-1a BCu-2
Copper alloys
Filler metal
2.75–3.50 2.75–3.50 2.75–3.50 2.75–3.50 1.5–2.2 0.03 ... 0.01 ...
B
16.0–18.0
Ni
99.90 min 99.0 min 86.5 min
Cu
Au
Fe
3.5–4.5
W
4.0–5.0 4.0–5.0 4.0–5.0 4.0–5.0 4.0–5.0 2.5–3.5 4.0–5.0 0.5 3.0–4.0 1.5 9.75–10.50 ... ... ... 0.10 0.2 6.0–8.0 ...
Si
7.5–8.5
Si
Sn
... ... ...
37.0–38.0 79.5–80.5 34.5–35.5 81.5–82.5 29.5–30.5 69.5–70.5
... ... ...
Zn
B
P
0.7–0.9
bal bal bal ... ... ...
Cu
... ... ...
Mn
0.075 ... ...
P
... ... ... ... 33.5–34.5 7.5–8.5
Pd
0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02
S
0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05
Al
0.02
Composition, %
0.35–0.45
C
P
Composition, %
... ... ...
Ni
Composition, %
0.6–0.9 0.02 0.06 0.02 0.06 0.02 0.06 0.02 0.06 0.02 0.10 0.02 0.10 10.0–12.0 0.08 9.7–10.5 0.10 0.02
C
1.0
Fe
... ... ...
Fe
Composition, %
0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.04 0.05
Ti
0.05
S
Ni
0.01 ... ...
Al
... ... ... ... ... ... ... 0.04 21.5–24.5
Mn
0.05
Al
... ... 2.5–3.5 bal 35.5–36.5 21.5–22.5
0.02 ... ...
Pb
... ... ... ... ... ... ... ... 4.0–5.0
Cu
0.05
Ti
0.15 0.15 0.15 0.15 0.15 0.15
0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05
Zr
0.05
Zr
0.10 0.30 0.50
bal bal bal bal bal bal bal bal bal
Ni
bal
Co
Others (total)
Others (total)
... ... ...
Si
0.50 0.50 0.50 0.50 0.50 0.50 0.50 0.50 0.05
Others (total)
0.50
Others (total)
1980 1980 1980
°F
1815 1635 1785 1740 2075 1845
°F
2050
°F
977 977 971 982 982 1080 877 888 982
°C
1790 1790 1780 1800 1800 1975 1610 1630 1800
°F
Solidus temperature
1121
°C
Solidus temperature
990 890 974 949 1135 1007
°C
Solidus temperature
1082 1082 1082
°C
Solidus temperature
Table 4.16 Compositions and properties of standard filler metals for brazing steels (copper, gold, cobalt, and nickel base)
1980 1980 1980
°F
1860 1635 1885 1740 2130 1915
°F
2100
°F
1038 1077 999 1038 1066 1135 877 888 1010
°C
1900 1970 1830 1900 1950 2075 1610 1630 1850
°F
Liquidus temperature
1149
°C
Liquidus temperature
1016 890 1030 949 1166 1046
°C
Liquidus temperature
1082 1082 1082
°C
Liquidus temperature
2000–2100 2000–2100 2000–2100
°F
1860–2000 1635–1850 1885–1995 1740–1840 2130–2250 1915–2050
°F
2100–2250
°F
1066–1204 1077–1204 1010–1177 1010–1177 1010–1177 1149–1204 927–1093 927–1093 1010–1093
°C
1950–2200 1970–2200 1850–2150 1850–2150 1850–2150 2100–2200 1700–2200 1700–2000 1850–2000
°F
Brazing temperature range
1149–1232
°C
Brazing temperature range
1016–1093 890–1010 1030–1090 949–1004 1166–1232 1046–1121
°C
Brazing temperature range
1093–1149 1093–1149 1093–1149
°C
Brazing temperature range
Chapter 4: Base Metals and Base-Metal Family Groups / 111
however, and form phases with two undesirable characteristics: the phases are considerably less ductile than either the base metal or the filler metal even at elevated temperatures and thus are a potential source of rupture, and the alloys formed with stainless steel are higher-melting alloys that are likely to freeze and block further flow into the joint during brazing. Because of the relatively high brazing temperatures required for nickel-base filler metals, their use is generally restricted to furnace brazing in a controlled atmosphere (including vacuum), although there are occasional exceptions. The nickel-base filler metals normally are supplied as powders; however, they can be obtained as sintered or cast rods, preforms, plastic bonded sheet, plastic bonded wire, and tape. Many of these filler metals now are available as metallic foils produced by ultrarapid cooling from the molten state. The BNi filler metals commonly are used on stainless steels for oxidation resistance at temperatures up to 980 to 1095 °C (1800 to 2000 °F). Filler metals BNi-1, BNi-2, BNi-3, and BNi-4 (Table 4.16) tend to erode thin sheet metal because of their interaction with many base metals. Time at brazing temperature and quantity of filler metal should, therefore, be controlled carefully when these filler metals are used. Boron-free filler metals BNi-5, BNi-6, and BNi-7 are suitable for use in nuclear reactor components where boron cannot be tolerated because of its absorption of neutrons. Oxidation resistance is good up to 1095 °C (2000 °F). The BNi-5 material has the highest melting point of all the nickel-base filler metals. The gold-base filler metals (BAu group in Table 4.16) are sometimes used for brazing stainless steels, although their high cost restricts their use to specialized applications such as heat exchangers for manned space-flight vehicles. When a gold-base filler metal is used, alloying with the stainless steel base metal is minimized, and, as a result, joints exhibit good ductility (Ref 74). In addition to the gold family of filler metals, there are palladium and platinum filler metals, such as Au-Ni-Pd, Cu-Pt, Ag-Pd-Mn, and Pd-NiCr, which are useful for brazing heat- and corrosion-resistant components in the jet- and rocketpropulsion and nuclear-energy fields. The cobalt-base filler metal BCo-1 (Table 4.16) is very rarely used for brazing stainless steels, but it has been included in the table for its availability. Furnace Atmospheres for Brazing Stainless Steels. Stainless steels can be brazed with
any of the various brazing processes. A large volume of controlled-atmosphere brazing is being performed on stainless steels, and the success of this type of brazing can be attributed to the development of reliable atmosphere and vacuum furnaces. The protective atmospheres most often used in furnace brazing of stainless steels are dry hydrogen and dissociated ammonia. These atmospheres are effective in reducing oxides, protecting the base metal, and promoting the flow of filler metal. The low-cost exothermic atmospheres that are widely used in furnace brazing of low-carbon steels are not suitable for stainless steels. An inert gas such as argon, or vacuum, may be used to satisfy special requirements and provide protection in applications for which hydrogen or hydrogen-bearing gases are satisfactory. A dry hydrogen atmosphere is preferred for many stainless steel brazing applications. Hydrogen, the most strongly reducing of protective atmospheres, reduces chromium oxide and provides for excellent wetting by some filler metals without the need for flux. The principal disadvantages of hydrogen are high cost, difficulty in drying sufficiently, need for special furnace equipment, and danger involved in storing and handling hydrogen. Dissociated ammonia, when it is free of moisture and 100% dissociated, is a suitable atmosphere for brazing of stainless steel with some filler metals without the need for a flux. Dissociated ammonia is strongly reducing but less so than pure dry hydrogen. Consequently, although it promotes wetting by reducing chromium oxide on the surface of the stainless steel, dissociated ammonia may not be sufficiently reducing to promote the flow of some filler metals, such as copper oxide powders. Because of its high (75%) hydrogen content, dissociated ammonia forms explosive mixtures with air and must be handled with the same precautions as those required for handling hydrogen. Unless the atmosphere used in brazing of stainless steel is completely decomposed (100% dissociated), even minute amounts of raw ammonia (NH3) in the atmosphere nitride stainless steels, especially those containing little or no nickel. Finally, to avoid oxidation of base metal and filler metal, the dissociated ammonia atmosphere must be kept dry and pure while it is inside the furnace. Argon is occasionally used as a furnace atmosphere in brazing stainless steels to other
112 / Brazing, Second Edition
stainless steels or to reactive metals such as titanium. Argon has the advantage of being chemically inert in relation to all metals; thus, it is a useful protective atmosphere for metals that can combine with or absorb reactive atmospheres, such as hydrogen. An argon atmosphere has the disadvantage of being unable to reduce oxides; consequently, the surfaces of stainless steel components must be exceptionally clean and free of oxides when brazed in argon. The principal advantages of furnace brazing are high production rates and the means for using controlled protective atmospheres at controlled dewpoints, which often make it unnecessary to use a flux to obtain satisfactory wetting. In most furnace-brazing applications, both of these advantages are exploited. Occasionally, however, furnace brazing is selected solely on the basis of production rate, and brazing is performed without a protective atmosphere but with a suitable flux. The lower-melting filler metals are generally selected for brazing under these conditions. Vacuum brazing of many structural configurations made of austenitic stainless steels offers excellent heat and corrosion resistance for hightemperature-service applications. For stainless steels, the fundamentals of torch brazing, as well as its advantages and limitations, are basically the same as carbon steels. Because of the metallurgical characteristics of stainless steel and its requirements for corrosion resistance, however, best results are obtained when special consideration is given to the type of flame at the torch and to the filler-metal composition. To aid in reducing the oxide already present, as well as to prevent further oxidation of the work-metal surfaces, a strongly reducing flame should be used for torch brazing stainless steel to itself. A reducing flame is also satisfactory for brazing stainless steel to nickel alloys or carbon steels. In brazing stainless steel to copper alloys, however, some compromise is necessary. Although a slightly oxidizing flame is normally best for brazing copper, for brazing stainless steel to copper a slightly reducing flame usually provides a satisfactory compromise in flame adjustment. The silver-base filler metals that flow at relatively low temperatures are used almost exclusively for torch brazing of stainless steels. Filler metal BAg-3 is most often used, because it flows well in the temperature range from 705 to 760 °C (1300 to 1400 °F) and provides joints that have greater resistance to corrosion than
those brazed with filler metals such as BAg-1 or BAg-1a, although these filler metals have been used. Depending on the metallurgical and physical properties of particular stainless steels, their behavior in heating by electrical induction may differ considerably from that of carbon and lowalloy steels and from that of the more widely used nonferrous metals. In addition, depending on whether a stainless steel is magnetic or nonmagnetic at room temperature, the response of the steel to induction heating varies considerably. Differences in specific heat and electrical conductivity markedly affect response to heating by induction. Ferritic and martensitic (400-series) stainless steels are ferromagnetic at all temperatures up to the Curie temperature. The Curie temperature is the temperature of magnetic transformation below which a metal or alloy is magnetic and above which it is paramagnetic (Ref 72). Thus, given the same power input, these steels generally heat faster than austenitic stainless steels, which are nonmagnetic in the annealed condition. Stainless steels may be induction brazed in an air atmosphere, using a suitable flux, although for critical applications, induction brazing is sometimes done in a protective atmosphere or in vacuum, whereas in other applications, an inert gas such as argon may be used as a backing gas to minimize oxidation. Brazing of stainless steel by immersion of all or a portion of the assembly in molten salt offers essentially the same advantages that would apply to brazing similar assemblies of carbon steel. Similarly, the same limitations are applicable.
Maraging Steels An investigation (Ref 75) found that thin gold-nickel alloy (BAu-4, 57.5Au-42.5Ni, at.%) interlayer brazes between strong base metals were found to have very high ultimate tensile strengths as a result of the triaxial stress state within the braze (reducing the effective, or von Mises, stress) and the high strength of the filler metal. Maraging steel was joined with the previously mentioned filler metal in this study. Results of the study found that: • The transmission electron microscopy (TEM) of the gold-nickel interlayer brazes showed a very fine-grain (0.1 µm) two-phase
Chapter 4: Base Metals and Base-Metal Family Groups / 113
microstructure, with both phases coincidentally oriented. • The ultimate tensile strength of the goldnickel bonds was high, largely due to a high interlayer yield strength that resulted from the refined microstructure. • The yield strength and quasi-steady-state flow stress of the filler metal were found to be relatively strain-rate insensitive. These results explain the absence of timedependent, or delayed, tensile failures that have been observed in other braze interlayers. The fracture process appears to be one of cavity nucleation by microplasticity, followed by cavity coalescence or impingement once the required nuclei density is achieved. Substantial cavity growth does not appear to occur. This failure process contrasts with that of other interlayer metals in which the failure mechanism appears to be unstable cavity growth.
Other Stainless Steel Applications Another study (Ref 76) found that companies saved a considerable amount of money by salvaging porous stainless steel castings by using the cleaning potential of the furnace atmosphere, vacuum, and the free-flowing character of the filler metal to fill the pores by capillary action. These filler metals can also be used for building up surfaces, for example, to adjust the throat widths of gas turbine compressor stators made by precision casting or to replace metal badly machined from large stainless steel valve components used in North Sea oil submersible equipment. The nickel-base filler metals have corrosion resistance generally comparable with Inconel, while some of the filler metals have oxidation resistances that are satisfactory to 1100 °C (2010 °F). Additional research found and reported on hydrogen detector units comprising a 2.3 mm (0.09 in.) outside diameter by 1.2 mm (0.05 in.) wall thickness nickel tube helix brazed into a type 316 stainless steel pressure hub approximately 25 mm (1 in.) in diameter and 75 mm (3 in.) thick; this was an extreme case of difference in thermal mass of the components. Brazing was carried out with BNi-2 filler metal at 1020 °C (1870 °F) for 5 min. The joint had to be tested using β-dense neutron radiography as well as helium leak testing and airflow checks. Finally, the research discussed the many thousands of stainless steel/copper heat ex-
changers that have been brazed for domestic water heaters and the stainless steel tube thermometers for electric cookers and cooker hot plates. Knife handles have been brazed from stainless steel pressings, as have very much larger valve bonnets for industrial use. It is very significant that joining stainless steel assemblies for food-handling equipment requires strong ductile joints while preventing formation of pits or other imperfections that can trap bacteria and contaminate food. When developing an assembly process for a doughdivider head assembly used in a dough-forming machine, graphite-brazing fixtures were constructed for five designs of the divider. When installed, the ring-and-blade assembly divides dough into portions small enough to make dinner rolls and large enough to make bread loaves. The five designs of divider head range in number of parts from 6 to 36. Each design calls for a ring and band to hold several blades in place; all materials are of type 304 stainless steel. Nickelbase BNi-2 filler metal holds the joints. This filler metal has been used in many foodindustry applications where clean, smooth surfaces to deter bacteria formation and resist oxidation are important. Both the brazing process and the nickel-base filler metal used have been approved by the United States Department of Agriculture food safety and inspection service and the privately operated National Sanitation Foundation. Production begins by placing ring and blades in an aluminum fixture. Then, the band is placed around the assembly and tack welded to the ends of the outer blades. Operators then use gas tungsten arc welding to create joints between the ring and blades to hold them in place during furnace brazing. Filler metal is applied in paste form to each butt joint; a 36-part assembly comprises 50 such joints. Then, an operator removes the assembly from the fixture and places it in the specially designed graphite fixture for furnace brazing. Brazing the head takes 12 h in an 81.3 by 106.7 cm (32 by 42 in.) top-loading vacuum furnace. Heating, brazing, and slow-cooling cycles are strictly controlled to prevent distortion. The procedure holds the bottom of the part flat to within 0.203 mm (0.008 in.), and six assemblies are brazed per furnace run. Palladium-containing filler metals 30Pd60Cu-10Co, 30Pd-60Cu-10Ni and 30Pd-50Cu10Ni-10Co (all in wt%) were studied for use in brazing stainless steel SUS 316 L at high temper-
114 / Brazing, Second Edition
atures (Ref 77). Palladium-containing filler metals were selected because of their ductility, oxidation resistance, relatively high melting points, and lower cost than gold-base filler metals. Wettability, microstructures, and reactions were studied with experimental brazed joints between base and filler metal. Joint strengths, tensile strengths at room and elevated temperatures of 473 to 1073 K, and creep-rupture strengths at 673 to 1073 K were investigated. The joints brazed with the 30Pd-50Cu-10-Ni-10Co filler metal had superior tensile strengths at elevated temperature and creep-rupture strengths. The joints with this filler metal had tensile strengths of 327 to 429 MPa (47 to 62 ksi) at temperatures of 293 to 873 K. The solidus and liquidus of the previously mentioned filler metals are shown in Table 4.17 (Ref 77). Joints brazed with the 30Pd-60Cu-10Co filler metal had the highest tensile strength of the tested joints, and the mean tensile strength for joint clearances ranging from 0.07 to 0.15 mm (0.003 to 0.006 in.) was 536 MPa (78 ksi). Joints brazed with the 30Pd-60Cu-10Ni filler metal had the maximum tensile strength at 473 K, and joints brazed with the 30Pd-50Cu-10Ni10Co filler metal did at 673, 873, and 1073 K. Joints brazed with the 30Pd-60Cu-10Ni filler metal had the highest creep-rupture strength at 673 K, and joints brazed with the Ni-Cr-W-FeB-Si filler metal did at 873 and 1073 K (Ref 77). It was concluded that 30Pd-50Cu-10Ni-10Co filler metal is the most suitable filler metal for high-temperature brazing of SUS 316 L. A parallel research study (Ref 78) examined two types of palladium-containing filler metals, 40Pd-50Ni-10Co and 30Pd-50Ni-10Cu-10Co (all in wt%). Commercially available filler metals, palladium filler-metal braze 33Pd-64Ag3Mn, and nickel-base filler metal Ni-Co-CrMo-Fe-B were also used. The microstructures and strength properties of the brazed joints were studied as well as Table 4.17 Solidus and liquidus temperatures of filler metals studied for brazing of 316L stainless steel Solidus Filler metal
30Pd-60Cu-10Co 30Pd-60Cu-10Ni 30Pd-50Cu-10Ni-10Co Ni-Cr-W-Fe-B-Si Source: Ref 77
Liquidus
°C
°F
°C
°F
1090 1139 1130 1010
1994 2082 2066 1850
1100 1150 1160 1165
2012 2102 2120 2129
the newly developed palladium-containing filler metals 40Pd-50Ni-10Co, 30Pd-50Ni-10Cu10Co, and 33Pd-64Ag-3Mn. The joints brazed with the 30Pd-50Ni-10Cu10Co and 40Pd-50Ni-10Co filler metals showed the highest tensile strength of the tested filler metals, and the mean tensile strength of joint clearances ranging from 0.05 to 0.15 mm (0.002 to 0.006 in.) was 516 MPa (75 ksi). The 30Pd-50Ni-10Cu-10Co filler metal showed maximum tensile joint strengths at 873 K or below, and the 33Pd-64Ag-3Mn filler metal did at 1073 K. The brazed joints with the 30Pd-50Ni-10Cu10Co and 40Pd-50Ni-10Co filler metals had the highest creep-rupture strength with the time to failure of 1000 h or over at applied stresses of 425 MPa (62 ksi) at 673 K and of 125 MPa (18 ksi) at 873 K, respectively. The joints with the Ni-Co-Cr-Mo-Fe-B filler metal had good creeprupture strength with the time to failure of 1000 h at applied stress of 50 MPa (7.3 ksi) at 1073 K. The 30Pd-50Ni-10Cu-10Co filler metal for SUS 316 L was finally found to be the most suitable filler metal for high-temperature brazing, because this filler metal provided superior strength properties at room and high temperatures (Ref 79).
Titanium and Zirconium and Their Alloys Titanium and zirconium combine readily with oxygen and react to form brittle intermetallic compounds with many metals and with hydrogen and nitrogen. Parts must be cleaned before brazing and brazed immediately after cleaning. Titanium and Its Alloys. Silver and silverbase filler metals were used in early brazing of titanium, but brittle intermetallics were formed, and crevice corrosion resulted. Type 3003 aluminum foil as a filler metal joins thin, lightweight structures, such as complex honeycomb sandwich panels. Electroplating various elements on the base-metal faying surfaces lets them react in-situ with the titanium/titanium alloy base material during brazing to form a titanium alloy eutectic. That transient liquid phase flows well and forms fillets, then solidifies due to interdiffusion (Ref 69, 80). Other filler metals with high service capability and corrosion resistance include Ti-Zr-NiBe, Ti-Zr-Ni-Cu, and Ti-Ni-Cu filler metals. The best braze processing is obtained in high-
Chapter 4: Base Metals and Base-Metal Family Groups / 115
vacuum furnaces using closely controlled temperatures in the range of 900 to 955 °C (1650 to 1750 °F). Selection of filler metals for use in brazing of reactive metals is critical to avoid formation of undesirable intermetallic compounds. Selection of filler metals and brazing cycles that are compatible with the heat treatment required for (α-β) and β-titanium base metals may present some difficulty. Ideally, brazing should be conducted at temperatures from 55 to 83 °C (100 to 150 °F) below the β transus, because the ductility of (α-β) base metals may be impaired if this temperature is exceeded. The β transus can be exceeded when β-titanium base metals are brazed; however, if the brazing temperature is too high, base-metal ductility after heat treatment may be impaired by considerable interaction between the properties of the heat treatable titanium alloys and may be affected adversely by brazing, unless the assembly can be heat treated afterward. For example, the (α-β) titanium alloys must be solution treated, quenched, and aged to develop optimal properties. It is not easy to select a filler metal that permits brazing and solution treating in a single operation. Similarly, it is not always possible to quench a brazed assembly at the desired cooling rate, and certain configurations (e.g., honeycomb sandwich structures) do not lend themselves to rapid quenching without distortion. Brazing at the aging temperature is impractical, because few filler metals melt and flow at these temperatures. The possibility of galvanic corrosion must be considered when filler metals are selected for brazing titanium base metals. Although titanium is an active metal, its activity tends to decrease in an oxidizing environment, because the surface undergoes anodic polarization in a manner similar to that of aluminum. Thus, filler metals must be chosen carefully to avoid preferential corrosion of the brazed joint. When titanium is brazed, precautions must be taken to ensure that the brazing retort or chamber is free of contaminants from previous brazing operations. As a precaution against any existing contaminants in the brazing furnace, a loose cover of pure titanium foil (0.03 to 0.08 mm, or 0.001 to 0.003 in., thick) should be put over the workpieces. This acts as a getter of any remaining contaminants. Care should be exercised to ensure that the foil does not come into direct contact with the workpieces. The choice of materials to be used in fixtures
must be carefully considered. Nickel or materials containing high amounts of nickel generally should be avoided; nickel and titanium form a low-melting eutectic (28.4% Ni) at approximately 940 °C (1725 °F). Should the titanium workpieces contact fixtures or a retort made from nickel-base alloy, the parts may fuse together if the brazing temperature is in excess of 940 °C. If a fixture material containing a high nickel content (such as stainless steel) is used, it should be oxide coated. In most applications, coated graphite or carbon steel fixture materials are used. Filler metals initially used for brazing of titanium and its alloys were silver-base filler metals containing lithium, copper, aluminum, or tin (Ref 81–83). Most of these filler metals were used in low-temperature applications (540 to 595 °C, or 1000 to 1100 °F). Later developments produced a number of successful commercial filler metals, including Ag-Pd, Ti-Ni, Ti-Ni-Cu, Ti-Zr-Be (Ref 84), and Ti-Zr-Cu-Ni (Ref 85, 86). Additionally, these filler metals could be used at temperatures from 870 to 925 °C (1600 to 1700 °F). For joining applications requiring a high degree of corrosion resistance, the 48Ti48Zr-4Be and 43Ti-43Zr-12Ni-2Be filler metals are outstanding (Ref 84). A Ag-Pd-Ga filler metal (Ag-9Pd-9Ga), which flows at 900 to 915 °C (1650 to 1680 °F), is another excellent filler metal that fills large gaps (Ref 87). The following four filler metals are new developments now available for use in products (Ref 85): • • • •
Ti-Cu-Ni-2 (Ti-20Cu-20Ni) Ti-Zr-Cu-Ni-1 (Ti-38Zr-12Cu-12Ni) Ti-Zr-Cu-Ni-2 (Ti-38Zr-15Cu-15Ni) Ti-Zr-Cu-Ni-Pd (Ti-37Zr-12Cu-12Ni-2Pd)
Initial tensile and corrosion testing of joints with the latter filler metals shows that the joints exhibit nearly the same mechanical and chemical characteristics as Ti-6Al-4V. Brazing of titanium honeycomb sandwich structures using aluminum filler metal is an achievement stemming from supersonic transport materials and process technology (Ref 81). Such aircraft structures up to 7 m (23 ft) in length have been successfully fabricated using Al 3003 brazing foils as filler metal, which provide satisfactory strength up to approximately 315 °C (600 °F). High-strength, corrosionresistant Ti-Zr-Be and Ti-Zr-Ni-Be filler metals are recommended for applications from 540 to 595 °C (1000 to 1100 °F) (Ref 84).
116 / Brazing, Second Edition
Researchers (Ref 88) examined the effect of joint clearance as a critical parameter for maintaining joint integrity in brazed Ti-6Al-4V and commercially pure titanium (CPTi) materials. They determined that the mechanical properties of Ti-6Al-4V and CPTi joints brazed with Ti(Zr)Cu-Ni-(Pd) filler metals (Table 4.18) are strongly dependent on the microstructure of the brazing zone, especially with the formation of brittle intermetallic phases in the center of the brazing zone, which reduce the mechanical strength of the brazed joints. These phases form due to an exceeding of the maximum jointclearance size of the system. The formation of these phases can be avoided by minimizing the joint-clearance size of the joint to be brazed and by enhancement of brazing time (Ref 89). Tensile tests of the brazed Ti-6Al-4V and CPTi joints revealed that especially Ti-6Al-4V joints exhibit excellent mechanical properties, even comparable to those of the base metal, if the brazing zone is free of brittle intermetallic phases. Fatigue tests of the Ti-6Al-4V joints confirm the excellent mechanical properties. The fatigue properties of the Ti-6Al-4V joints approach that of the base metal at maximum stresses below 700 MPa (102 ksi). In contrast to the Ti-6Al-4V joints, fabrication of CPTi joints with mechanical properties comparable to those of the base metal was not possible. The reason can be seen in the much lower maximum joint clearance of the system CPTi/Ti-Zr-Cu-Ni with regard to the formation of intermetallic phases. Therefore, under the given circumstances (50 µm filler-metal foils), formation of intermetallic phases could not be avoided. This study also revealed that the induction brazing process is an interesting alternative to the vacuum furnace brazing process. Especially from the economic point of view, the induction brazing process under inert atmosphere had many advantages due to the short processing time. On the other hand, the short processing
time was responsible for much higher requirements on brazing joint clearance compared to the vacuum brazing process. However, fabrication of titanium joints with mechanical properties comparable to those of the base metal by induction heating were possible, when the maximum brazing joint clearance of the system was not exceeded. This required extremely small joint clearances of <30 µm in the case of Ti6Al-4V/Ti-Cu-Ni-Pd joints and <20 µm in the case of CPTi joints. Using an amorphous 25Ti-25Zr-50Cu filler metal, researchers (Ref 90) induction brazed Ti6Al-4V alloy using argon as a shielding gas. The brazing cycles were rather short; radio frequency inductor power was supplied for only 40 to 60 s. The tensile strength and ductility of the brazed joint were at the level of that of Ti-6Al4V base metal. It was also found that the postbrazing natural cooling rate of 12.5 mm (0.5 in.) diameter rod samples was sufficient to suppress the formation of undesirable brittle λ-Cu2TiZr Laves phase appearing in vacuum furnace brazing due to inherently low cooling rates. From the practical point of view, it was proven that induction brazing of compatible samples located in a simple closed chamber may be carried out as an effective and inexpensive process. Zirconium and Its Alloys. Zirconium-base metals of commercial importance are the pure metal and several alloys. The most commonly used zirconium alloy is Zircalloy, which contains small percentages of tin, iron, chromium, and nickel. These structural alloys are used for corrosion resistance in nuclear applications, especially in pressurized-water nuclear power reactors. Like titanium and beryllium, zirconium reacts readily with oxygen, hydrogen, and nitrogen and is embrittled. It also reacts with many metals and alloys to form intermetallic compounds. Therefore, as a result of this reactivity, zirconium must be brazed in a vacuum or in a
Table 4.18 Compositions and critical temperatures of filler metals evaluated for brazing commercially pure titanium and Ti-6Al-4V Chemical compositions, wt%
Solidus
Liquidus
Brazing temperature
Filler metal
Ti
Zr
Cu
Ni
Pd
°C
°F
°C
°F
°C
°F
TiCuNi TiCuNiPd TiZrCuNi
60 60 35
... ... 35
20 19 15
20 19 15
... 2 ...
923 923 832
1695 1695 1530
934 934 850
1715 1715 1560
950 950 870
1740 1740 1600
Source: Ref 88
Chapter 4: Base Metals and Base-Metal Family Groups / 117
dry atmosphere of argon or helium. Zirconium joint members must be cleaned carefully before brazing to remove oxides and other surface contaminants, and brazing should be done immediately after cleaning. Compared with the other reactive metals, very little research has been done to develop filler metals and brazing methods for joining zirconium and zirconium alloys (Ref 91). To a degree, this lack of research on brazing can be attributed to the availability of other joining methods. However, most commercial filler metals do not wet or flow well on zirconium base metals, nor are they metallurgically compatible with zirconium. In addition, many of these filler metals do not possess the corrosion resistance required in reactor environments. Research was conducted to develop filler metals for producing brazed joints in Zircalloy 2 that possessed good resistance to corrosion in pressurized water at 360 °C (680 °F). The data from these studies and from metallographic examinations of the brazed joints indicated that the following filler metals most nearly met the service requirements: Zr-5Be, Cu-20Pd-3In, Ni-20Pd10Si, Ni-30Ge-13Cr, and Ni-6P (Ref 91). Additional studies to develop improved filler metals for brazing Zircalloy base metals for use in water-cooled reactors were recently completed (Ref 92). Candidate zirconium-base and non-zirconium-base filler metals were formulated, used in vacuum brazing under pressures of 0.0013 to 0.00013 Pa (10–5 to 10–6 torr), and screened by wetting tests and corrosion tests in pressurized, high-temperature water (315 °C, or 600 °F). The following filler metals had acceptable corrosion resistance and mechanical strength: • Zr-50Ag (brazing temperature, 1520 °C, or 2770 °F) • Zr-29Mn (brazing temperature, 1380 °C, or 2515 °F) • Zr-25Sn (brazing temperature, 1730 °C, or 3145 °F) The Zr-5Be filler metal previously mentioned has been used extensively to braze zirconium base metals to themselves and to other metals (e.g., stainless steel). For example, zirconium sheet stock has been brazed with this filler metal using the following cycle: 10 min at 1005 °C (1840 °F), followed by 4 to 6 h at 800 °C (1470 °F). Because of its ability to wet ceramic surfaces, Zr-5Be has also been used to braze zirconium to uranium oxide and beryllium oxide.
As stated earlier, titanium and zirconium are both highly reactive metals. The brazing process, therefore, must not allow the joint surfaces to come in contact with air during heating. Induction brazing and furnace brazing in inertgas or vacuum atmospheres can be used successfully. Torch brazing of these base metals is difficult, requiring special precautions and techniques. Induction brazing of small, symmetrical parts is very effective, because the speed minimizes reactions between filler metal and base metal. For large, precise assemblies, furnace brazing is favored, because it allows uniformity of temperature throughout the heating and cooling cycle to be controlled readily. Titanium and zirconium assemblies frequently are brazed in high-vacuum, cold-wall furnaces.
Carbides and Cermets Carbides of the refractory metals tungsten, titanium, and tantalum, bonded with cobalt, have been used for years in cutting tools and dies. It is necessary to join these carbides to metal parts, particularly for cutting tools. Closely related materials called cermets have been developed and are ceramic particles bonded with various metals. Their high-temperature strengths are intermediate between those of the ceramic materials and the binder metals employed. Their greatest disadvantage is their brittleness. The term cermets generally describes titanium-base hardmetals. These cermets bridge the gap between carbides and ceramics. Cermets are a unique class of materials that could have excellent applications in areas such as saws and wear parts. They are harder and much more wear resistant than tungsten carbide. They do not have the impact resistance and toughness of tungsten carbide, so they are not recommended in highimpact applications such as sawmills or cabinet shops where material is hand fed. The cermets are composed entirely of titanium carbide (TiC) and titanium nitride (TiN). These are the same materials used to coat ordinary grades of carbide to make them more wear resistant. A cermet is not coated; instead, it is solid coating material. It is impractical to use coated carbides in saws, because the saws are reground and the coating would be lost in the first grinding. Typical carbide coatings are TiC and TiN. The XT3 grade is solid TiC and TiN, so there is no coating to wear off.
118 / Brazing, Second Edition
Cermets have been used successfully in clamped or indexable machining operations for decades. They have not been used in brazed applications because they have not been brazeable. There is now a method to successfully braze cermets to steel holders. Previous brazing attempts resulted in the material forming a glob in the center of the insert and an extremely weak bond. The new braze technology gives excellent wetting of the insert and excellent braze strength. A successful braze is one where the strength of the joint is in excess of the rupture strength of the material. The force to fracture the brazed cermet was approximately equivalent to the force required to fracture an unbrazed cermet. Additional research can improve the impact resistance of brazed cermet joints, as it has for brazed tungsten carbide joints. Cermets typically have a life of 3 to 5 times that of carbide. This is due to superior wear resistance. An XT3 grade of cermet has the approximate toughness of C7 carbide while being much harder. There is also a C50 grade of cermet that is much tougher. Cermets are now brazeable, with tensile strengths of up to 685 MPa (100 ksi), and can tolerate temperatures up to 750 °C (1380 °F). New developments permit cermets to be brazed with the same filler metals and for the same cost using the same equipment and techniques used for tungsten carbide. These newly developed processes use chemicals that are commonly available, allow for brazing using standard filler metals, and also use these at ambient temperatures. The process involves preparing the cermet body as follows: • Cleaning: A preferred method of cleaning is a cathodic electrocleaning process in an alkaline solution. • Surface etching: The part is treated in such a manner as to prepare the surface to bond with an intermediate material. Typically, this is a chemical bath and may or may not use electric current. The surface is roughened and chemically activated. The part is immersed in a solution for an etching effect. This process may be accelerated or enhanced by the use of electric current, higher temperature, or different combinations of chemicals. • Deposition of cobalt or other metal: The surface is plated with a metal or other material that bonds to the ceramic as well as forms a layer suitable for brazing with standard filler
metals. Typically, this is also a chemical bath and may or may not use electric current. It is beneficial if non-cyanide nitrogen is introduced so as to create intermediate metallic compounds. Cyanide compounds may be used, but they are unnecessary and significantly add to the cost in several ways. • Postcleaning and passivation: This treatment of the parts may be desirable. The parts are then brazed on saws using standard filler metals (such as BAg-24 and those similar to it) and are used successfully in sawing and other applications (Ref 93) (see Chapter 10, “Applications and Future Outlook”). Brazing carbides and cermets is generally more difficult than brazing metals. Torch, induction, or furnace brazing is used, often with a sandwich brazing technique; a layer of weak, ductile metal (pure nickel or pure copper) (Ref 94) is interposed between the carbide or cermet and a hard metal support. The cooling stresses cause the soft metal to deform instead of cracking the ceramic (Ref 95). Silver-base filler metals, copper-zinc filler metals, and copper are often used on carbide tools. Although it is possible to use any of the BAg-1 through BAg-7 filler metals, those containing nickel (BAg-3 and BAg-4) are generally recommended, because nickel improves wettability. The RBCuZn-D and BCu filler metals also have been used, particularly where a postbraze heat treatment is required. The BCu filler metal retains practically all of its strength up to a temperature of 540 °C (1000 °F); however, it requires a hydrogen atmosphere furnace for best brazing results. The 85Ag-15Mn and 85Cu15Mn filler metals are used where the brazed joint is subjected to elevated temperatures in service and also for wetting the titanium-base or chromium-base carbides. Additionally, the BNi filler metals with boron and a 60Pd-40Ni filler metal have been used successfully for brazing nickel- and cobalt-bonded cermets of tungsten carbide, titanium carbide, and niobium carbide. Tungsten-base carbides generally are readily wetted by the BAg and BCu filler metals. However, titanium-base carbides are somewhat more difficult to wet. Where it is necessary to mount TiCs by brazing, either the joint must be made in an inert or vacuum atmosphere or the surfaces must be specially treated by the carbide manufacturer, or by nickel plating. In each case, the filler metals just mentioned then wet the TiC surface. In specifying a filler metal for a given job,
Chapter 4: Base Metals and Base-Metal Family Groups / 119
first consideration should be given to the temperature range of the anticipated application. This and the other considerations of corrosion and mechanical properties dictate the alloy composition selected, the brazing temperature required, the equipment to be used, and the joining atmosphere. Generally speaking, the filler metals mentioned previously are considered for the majority of applications that require simple equipment, fluxes, and brazing temperatures below 980 °C (1800 °F). The carbides, in general, are not wetted as readily by filler metals as are most base metals; thus, it is preferable, when possible, to preplace shims in the joint rather than to face feed the filler metal in the form of wire. For larger surfaces, shims having a core of copper or nickel with a coating of the filler metal on both surfaces are frequently used. The core of the shim generally accounts for approximately 50% of the total thickness, with the filler metal providing another 25% on each side. Filler metals can provide a cushioning or a shock absorber effect that actually helps the carbide survive in use. This is generally accomplished in one of three ways. The first is through the use of a soft filler metal, such as 50% Ag with cadmium. This is a pure cushioning effect. The second way is through the use of manganese. Manganese is a hard material, but it has the property of absorbing impact. This is more similar to a shock absorber. The third method is the use of trimetal sandwiches. These are typically filler metal/copper/filler metal. During the brazing process, the copper anneals to a dead soft condition. The cadmium alloys work well, but the cadmium is a serious health risk; more importantly, it is a health risk that the government really targets. Cadmium fumes at low temperature escape from the filler metal as it is being heated. It is also fairly soluble and dissolves into the grinding coolants when the saws and tools are ground. Finally, cadmium is very readily detectable, so it is easy to find. Manganese makes a good alloy for filler metal. However, the manganese makes the alloys a bit brittle, which means it is harder to draw into wire or roll into ribbons. The trimetal sandwiches are laminated together, which makes them very expensive. In addition, they have to be cut and bent to make brazing preforms. The metal sandwiches have to be preplaced into the joint. Three practices that can lead to improvements are recommended (Ref 93). First, the use
of a high-silver filler metal with manganese increases the bond strength and greatly increases the impact resistance. It also helps prevent thermal stress fractures. Secondly, the development of a cleaning and surface-treatment process cleans the carbide completely and then selectively removes the cobalt, which leaves the surface rough, because it is all exposed tungsten carbide grains. Finally, this surface is extremely wettable and provides an excellent gripping surface for the filler metal while also promoting the intermetallic joining characteristics for successful bonding.
Ceramics The increasing use of ceramic-to-ceramic and ceramic-to-metal joints in industrial and developmental applications is due to the unique combination of properties of ceramic materials (Ref 96–100). The large use of ceramic-to-metal joints in vacuum tubes in the electronics industry stems from the following properties: • Ceramic tubes can be outgassed at higher temperatures than glass tubes. • Ceramic tubes can withstand higher temperatures than glass tubes of similar dimensions. • Ceramic tubes are mechanically stronger and less sensitive to thermal shock than glass tubes. • Ceramic components can be ground to the precise tolerances required for vacuum-tube construction. • Ceramic materials have very low electrical losses at high frequencies. Because of their inertness in many corrosive environments, ceramics are used as seals in fuel cells and other devices that convert chemical, nuclear, or thermionic energy to electricity. Ceramics are also used as friction materials for brakes, clutches, and other energy-absorbing devices; coatings for nuclear fuel particles; constituents in high-temperature adhesives; radomes used to enclose antennae; and ablative materials. Glass-to-metal seals have been made for many years in the vacuum-tube industry, and the experience thus obtained gives a general insight into the problem of fabricating ceramic-to-metal joints (Ref 98). More recently, the fabrication of refractory-tipped tools, vacuum tubes, and various experimental devices has added to the available knowledge and techniques. Glass-Ceramic Joining. There are two types of intermediate bonds that can be em-
120 / Brazing, Second Edition
ployed, and the selection of which to use is influenced by a number of practical features, including the ease of fabrication, the temperature-withstand capabilities required, whether or not the bond must possess insulating properties, and the cost. The bonds can be either metallic (e.g., a braze) or vitreous (glass or glass-ceramic). In this type of bond, the usual design configuration is that of a butt seal in which the glass-ceramic components are joined on their end faces to the metallic components, although other types of seal, such as sleeve or concentric seals, are possible. The glass-ceramic body components are initially prepared using glass-forming methods and are then heat treated. The butt type of seal usually requires that the end faces of the glassceramic components are machined flat and parallel, and other surfaces, which are very often inner and outer cylinder surfaces, are machined to tight tolerances to fit the metal parts. The metal parts themselves also require preparation before assembly. In butt seals, the sealing faces must be flat to ensure good mating with the glass-ceramic seal faces. The preparation given to the metal prior to sealing depends on the type of intermediate bond to be employed—a metallic braze or a vitreous bond. In the case of a braze metal bond, the metal seal parts are required to have smooth sealing surfaces and to be clean and free from oxidation and other defects. The type of braze normally employed with a glass-ceramic is an active metal braze. In the butt seal design, thin washers of titanium and copper-silver eutectic are
Fig. 4.9
Exploded view of the constituent parts for an active metal brazed glass-ceramic/metal seal. Source: Ref 101
assembled together with the major components, as illustrated in Fig. 4.9. The titanium is placed adjacent to the glass-ceramic, and the coppersilver eutectic adjacent to the metal, the whole being in suitable jigs to provide and maintain alignment. Alternatively, a specially developed Cu-Ag-Ti filler metal can be used (Ref 102). The firing is carried out in a vacuum in order to prevent any oxidation of metallic constituents, which would inhibit wetting and flow of the braze at a temperature of approximately 850 °C (1560 °F). The titanium serves to react with the glass-ceramic during this firing, metallizing the surface and permitting reaction with the braze, which in turn wets and bonds to the metal surface. This type of bond permits subsequent processing (e.g., vacuum bakeout) and operation to temperatures in the region of 800 °C (1470 °F). Higher temperature-withstand capability can be achieved by using more refractory brazes, such as those containing nickel. However, it will be appreciated that the limitation may become the glass-ceramic component itself, although glass-ceramics are available that withstand temperatures well in excess of 1000 °C (1830 °F). It should be noted that the more-refractory glass-ceramics generally have relatively low thermal expansion, and consequently, matched expansion systems are only possible with the lower-expansion metals and alloys, for example, molybdenum and tungsten. It should be borne in mind that work at elevated temperatures may require the presence of an inert atmosphere to prevent oxidation and deterioration of metal parts, including the braze. In the case of a vitreous bond (glass or glassceramic), the sealing faces of the metallic component are roughened by means of grit blasting or etching, and the components are preoxidized by heating in a suitable atmosphere. The metal sealing faces are then coated with the vitreous bond, which is applied in powder form via a suitable suspension, usually by spraying. The coating is allowed to dry and is then fired, often in a protective atmosphere to protect the metal from excessive oxidation, to form a continuous adherent layer on the metal sealing face. It is usually unnecessary to preglaze the glassceramic sealing faces, although this can be done in a similar manner, if it is considered desirable. The various components are assembled, as shown in Fig. 4.10, for a butt seal and are then fired to cause the bonding materials to complete reactions with the glass-ceramic and metal components and so form the desired bond. If a glass-
Chapter 4: Base Metals and Base-Metal Family Groups / 121
ceramic bond is employed, the firing schedule also causes the glass to crystallize. A glass bond permits subsequent processing and operation up to temperatures of approximately 500 °C (930 °F), depending on the glass used, and a glassceramic bond enables temperatures in excess of 800 °C (1470 °F) to be withstood, again depending on the glass-ceramic used. The use of vitreous bonds is advantageous where more than two components are required to be bonded together in such a manner that sequential bonding operations are required. It is necessary in these circumstances, while maintaining thermal-expansion-matching/settingtemperature criteria, that the second and subsequent glasses are used at successively lower temperatures so that the firing conditions do not disturb the preceding glass and thereby allow relative movement of the previously bonded components. Fortunately, glasses are available that permit this sequential operation to be achieved, particularly for higher thermal expansion systems. If necessary, a glass-ceramic or braze can be used to provide the initial, more refractory bond. Examples of seals produced by direct bonding methods, such as in-mold shaping and graphite jig shaping, include: • Glass-ceramic-molybdenum high-voltage feed through seal, which comprises a glassceramic of the ZnO-Al2O3-SiO2 type that is matched in expansion and bonded to the molybdenum parts to form a concentric design (Ref 103)
Metal Pre-glazed layer
Glass-ceramic
Fig. 4.10
Exploded view of the constituent parts for glass- or glass-ceramic-bonded glass-ceramic/metal seal. Source: Ref 101
• Glass-ceramic-metal leadthrough seals (e.g., in thermocouple terminations). In this type of seal assembly, the glass-ceramic is matched in expansion to the metal(s) of the leadthroughs. In a thermocouple termination, these are Chromel and Alumel (Hoskins Manufacturing Company). In other types of light-current leadthrough seals, the leadthrough metal is often a nickel-iron alloy, for example, 48%NiFe. In this type of seal, a higher-expansion metal or alloy, for example, mild steel or a stainless steel, is frequently used for the outer metal components. A fully hermetic bond is developed between the glass-ceramic and the outer metal, and, owing to the differential expansion, or more precisely contraction, of the two components, the outer joint is in compression, which increases the overall strength of the assembly and its ability to resist thermal shock. Clearly, the deliberate introduction of compressive stresses into the outer glassceramic influences the design of the seal assembly. Challenges in Brazing of Ceramics. Ceramic materials are inherently difficult to wet with conventional filler metals. Most of these filler metals merely ball up at the joint, and little or no wetting occurs. When bonding does occur, it can be either mechanical or chemical. The strength of a mechanical bond can be attributed to interlocking particles or penetration into surface pores and voids, whereas a chemical bond derives strength from material transfer between the filler metal and the base material. Discussions of bonding mechanisms can be found in the literature (Ref 101–109). Another basic problem in brazing of ceramics results from the differences in thermal expansion between the base material and the filler metal and, in the case of ceramic-to-metal joints, between the two base materials. In addition, ceramics are poor conductors of heat, which means that it takes them longer to reach equilibrium temperature than it does metals. Both of these factors may lead to cracking in the joint. Because ceramics generally have lower tensile and shear strengths, crack propagation occurs at lower stresses in ceramics than in metals. In addition, the low ductilities permit very little distribution of the stresses set up by stress raisers. Alumina, zirconia, beryllia, thoria, forsterite (Mg2SiO4), and silicon carbide and nitride are the leading ceramic materials that can be joined by brazing.
122 / Brazing, Second Edition
and that use temperatures exceeding 750 °C (1380 °F) (Ref 117). Further work (Ref 118) showed the development of strong, reliable joints containing ceramic components for applications in advanced heat engines. This work was focused on the joining of Si3N4 by brazing. The technique of vapor coating ceramics to circumvent wetting problems that was developed for brazing zirconia at low temperatures was applied to brazing Si3N4. An acceptable porosity-free joint was obtained with 50Au-25Ni-25Pd filler metal. The joints made with the 50Au-25Ni-25Pd (wt%) filler metal were brazed at 1250 °C (2280 °F) (approximately 120 °C, or 216 °F, above the normal brazing temperature for the Au-Ni-Pd filler metal). This work also showed that the 50Au-25Ni25Pd (wt%) filler metal is particularly well suited for brazing titanium-vapor-coated Si3N4. Figure 4.11 shows that for this material combination, high joint strength was maintained up to 700 °C (1290 °F) before dramatic loss at 800 °C (1470 °F). Investigations continue into the apparent dependence of joint strength on the filler-metal thickness, with the tendency for improvement with reduced braze-layer thickness (from 40 to 50 µm to 10 to 20 µm). Other efforts (Ref 119) to develop filler metals for Si3N4 components for the Advanced Turbine Technology Applications Project (ATTAP) engine have included the development of filler metals for ceramic-to-metal joints, including the investigation of the different alloy systems Au-Ni-Cr-Mo-Fe-Nb, Au-Ni-Cr-Fe, Au-Ni-Cu-Cr-Mo-Fe-Nb, and Au-Ni-Cu-CrFe. The effort was to increase the high-temper500 (72.5)
Joint strength, MPa (ksi)
If the ceramic is premetallized to facilitate wetting, copper, silver-copper, and gold-nickel filler metals can be used. Titanium or zirconium hydride can also be decomposed at the ceramicmetal interface to form an intimate bond. Nonmetallized ceramics have been brazed with silver-copper-clad or nickel-clad titanium wires and other useful titanium and zirconium filler metals, such as Ti-Zr-Be, Ti-Zr-V, Zr-VNb, Ti-V-Be, and Ti-V-Cr (Ref 65, 96, 101, 110–113). Silicon Nitride (Si3N4). One process for brazing Si3N4 with metallic alloys involves vapor coating the ceramic with a 1.0 µm thick layer of titanium before the brazing operation. The coating improves wetting of the Si3N4 surfaces to the extent that strong bonding between the solidified filler metal and the ceramic occurs. Braze joints of Si3N4 are made with AgCu, Au-Ni, and Au-Ni-Pd filler metals at temperatures of 790, 970, and 1130 °C (1455, 1780, and 2065 °F) (Ref 114). A study also showed that Si3N4 surfaces are not easily wet by common precious-metal-based filler metals, but that vapor coating the surfaces with titanium prior to brazing improved their wetting characteristics and permitted filler metals to adhere strongly to the ceramic. The ability of vapor coatings, particularly of titanium, to improve the wetting characteristics of oxide ceramics has been known for some time (Ref 115, 116). Using a metallic vapor coating rather than incorporating an active element such as titanium directly into the filler metal permitted readily available commercial filler metals to be used for joining. The Si3N4 ceramics have recently been joined by an oxynitride brazing method. It was useful for large-sized parts or ceramics with complex geometries in which the high-temperature mechanical properties of the original ceramic components must be maintained. Joined parts can withstand temperatures that are 300 to 400 °C (540 to 720 °F) higher than conventionally joined parts can tolerate. Joint strength is comparable to the 400 MPa (58 ksi) of the Si3N4 ceramic material. If joint strengths do not need to exceed 200 MPa (29 ksi), then these joints can be used at temperatures as high as 1250 °C (2280 °F). These large or geometrically complex parts are used in gas turbine and adiabatic diesel engines, magnetohydrodynamic generators, pumps, and numerous other applications that require high strength and corrosion resistance
400 (58)
300 (43.5)
200 (29)
100 (14.5)
0 0
200 (390)
400 (750)
600 (1110)
800 (1470)
Temperature, °C (°F)
Fig. 4.11
Variation of joint strength with test temperature for titanium-vapor-coated SN220 Si3N4 lap joints brazed with Au-25Ni-25Pd wt% filler metal. Source: Ref 118
Chapter 4: Base Metals and Base-Metal Family Groups / 123
ature strength with a minimum loss of other properties necessary for brazing. Of those alloys, Au-Ni-Cr-Mo-Fe-Nb alloys, which melt at 1050 to 1100 °C (1920 to 2010 °F), seemed to meet the processing requirements for high-temperature ceramic-metal joints. Preliminary results from mechanical testing indicated that these alloys had good ductilities and strengths. They wetted well on Si3N4 with various coating materials in the vicinity of 1150 °C (2100 °F). Researchers (Ref 120) conducted an evaluation of brazed Si3N4 joints, including their microstructure and mechanical properties. The study involved Cusil ABA (active brazing alloy) (Wesgo Metals), an Ag-35Cu-1.6Ti filler metal with a liquidus of 815 °C (1500 °F), which was used to produce Si3N4/Si3N4 (SN/ SN) joints with room-temperature strengths similar to that of the monolithic ceramic. Brazing was achieved by holding at 850 °C (1560 °F) for 30 min under vacuum, using 50 µm thick filler-metal foil. Silicon nitride/AISI 316 steel joints were also produced by brazing with Cusil ABA when interlayers were used. Molybdenum was more successful as an interlayer than copper or niobium; its coefficient of thermal expansion matched that of the ceramic, and it had less effect on the reacting route for the joint formation. Joint strengths were higher for smaller workpiece bonded areas. Failure initiated predominantly from sites at the interfaces between the braze and the ceramic or metal workpieces. An investigation (Ref 121) reported on a joining method of ceramics-to-ceramics developed with a target goal of retaining high-temperature flexural strength with more than 80% of that of base materials. To join the Si3N4 material, the system of Si3N4/yttrium oxide/alumina was selected for the base material. The sintered Si3N4 material was joined by using the powder mixture of the grain-boundary phase composition. The paste of the powder mixture of the grain-boundary phase was screen painted on the joining surfaces of the Si3N4 material. The joining was carried out by firing them in a nitrogen atmosphere and hot isostatic pressing (HIP). Optimizing the conditions, the strength of joined specimens resulted in 800 and 640 MPa (116 and 93 ksi) at room temperature and 1250 °C (2280 °F). Silicon nitride braze joints have reportedly been produced in a single step at elevated tem-
peratures without premetallizing at the University of Illinois, Chicago. Reactive brazing with a filler-metal composition of Co-10Ti was used to join Si3N4 components; the brazing process involved a 15 min hold at 1300 °C (2370 °F) under a high vacuum. Active metal brazing with Ag-Cu-Ti alloys is currently used in the automobile industry to fabricate Si3N4-metal components, and no initial metallization of the ceramic surface is required. Nonetheless, because silver- and copper-base filler metals cannot withstand service temperatures above 500 °C (930 °F), filler metals with higher melting points must be added to really utilize the high-temperature potential of Si3N4. However, during tests with Co-10Ti filler metal, the reaction layer that developed was not uniform. While thicker areas were stable, the thinner areas showed some degradation, and regardless of thickness, gaps were observed between the braze metal and the ceramic substrate. This was caused by the premature isothermal solidification of the braze as it was depleted of titanium. Brazing with Ni-13Cr-Hf was also tested but was only possible with the Si3N4 premetallized. Hafnium from the braze metal moved toward the existing reaction layer, while copper dissolved in the matrix. Silver moved from the existing reaction layer toward the braze metal but did not dissolve; it remained as islands of pure silver (Ref 122). In a study (Ref 123) conducted to improve the joining strength of Si3N4 ceramics, pure nickel and 80wt%Ni-20wt%Cr alloy laminated interlayers were used for joining. Thin, pure nickel layers in contact with the Si3N4 ceramic were used for the following reasons. The gaps between pure nickel and Si3N4 ceramics closed easily with small joining pressure in the early stage of the joining process, because the flow stress of pure nickel is smaller than that of the nickel-chromium alloy. Then, the nickel and Si3N4 joined. Furthermore, chromium diffuses in pure nickel layers toward the joining interfaces and forms chromium nitrides there. The decomposition of the nitrides is suppressed by their interception from the argon joining atmosphere because of the closed gaps. The researchers concluded that the positive effects of thin, pure nickel layers placed in contrast with Si3N4 on the joining strength of Si3N4ceramics with nickel and nickel-chromium alloy laminated interlayers were confirmed experimentally. This joining technique can also
124 / Brazing, Second Edition
be applied for the fabrication of high-strength metal-Si3N4 joints. A team of researchers (Ref 124) assessed the mechanical properties of Si3N4 braze joints after fabrication and following exposure to elevated temperatures. These tests were conducted at room temperature and using four-point bending. The filler metal used in the investigation was palladium-base 58.2Pd-38.8Ni-3.0Ti. The Si3N4 substrate was coated with a Ag-Cu-In-Ti filler metal prior to fabrication of the braze joint. Two different sample geometries were used in the fabrication of the braze joints. Weibull statistics were applied to the interpretation of the data, but the results contradict what is expected from monolithic fracture theories. Oxidation was found to be significant when the braze joints were exposed to 800 °C (1470 °F) in air, which reduced the joint strength. An interlayer consisting of titanium, silver, and copper (0.1 mm, or 0.004 in., thick titanium foil; 0.1 mm thick silver foil; and 0.5 mm, or 0.02 in., thick copper plate) was sandwiched between two such Si3N4 rods in the symmetrical arrangement of titanium/silver/copper/silver/titanium, with copper at the center, and then treated for 60 min at 1273 K and 2 MPa (0.3 ksi) to form a SN/SN joint. The previously mentioned procedure was used to investigate the fatigue behavior of brazed SN/SN joint specimens (Ref 125). Three-point fatigue testing was conducted under static and cyclic loading at temperatures ranging from 300 to 1038 K in air. It was found that the process of strength degradation owing to fatigue is controlled by the rate of either the cavity growth or the crack growth, depending on which one is dominant. At low temperatures, the fatigue process is mainly controlled by crack growth; on the other hand, formation of a small cavity and its growth seem to be predominant factors in controlling the fatigue process at high temperatures. Therefore, the triaxial tensile stress component is a very important factor in evaluating the fatigue life of joints at high temperatures. However, a fairly good estimation of the fatigue life is obtained using nominal bending stress in the present investigation. It is also found that the fatigue life can be estimated using the bending strength of joints. Researchers (Ref 126) brazed Si3N4 with nickel-brazed filler metals having the same nickel-chromium ratio as American Welding Society (AWS) BNi-5 (Ni-18Cr-19Si, at.%) but different silicon content. Joining experiments
were performed at 1220 °C (2230 °F) under a N2 partial pressure of 15 Pa (0.002 psi) for different times between 5 to 15 min. The highest roomtemperature, four-point bend strength of the joints was 115 MPa (17 ksi), whereas 220 MPa (32 ksi) was achieved when the joints were tested at 900 °C (1650 °F). The high strength of the experimental joints was attributed to the reduction in residual stresses and the formation of a CrN reaction layer at the ceramic/fillermetal interface. This study shows that chromium is the most reactive element in the Ni-Cr-Si filler metals used for the brazing experiments. SIALON. Oxynitrides were initially referred to by the acronym SIALON, which stands for the Si-Al-O-N system. Both terms, oxynitride and SIALON, are used extensively throughout the literature to refer to specifically Si-Al-O-N materials as well as compounds derived from silicon nitrides or oxynitrides by simultaneous replacement of silicon and nitrogen by aluminum and oxygen. The term has become a generic one applied to materials where the structure involves (Si, Al) (O2, N2) or (Si, M) (O2, N2) tetrahedral (Ref 109, 127, 128). Silicon Carbide (SiC). Joining of SiC to SiC was investigated for advanced heat engine applications (Ref 129). The SiC-SiC joints were produced by cosintering β-SiC green forms with and without the use of pressure (HIP and vacuum sintering). The joints attained tensile strengths equal to or greater than 138 MPa (20 ksi) at 1530 °C (2785 °F), and no glassy phases were used for the joint. A program was undertaken to evaluate active brazing applied to SiC, because there is no sufficient metallizing process available for this ceramic (Ref 130). The employment of metallic interlayers is one way of reducing thermally generated stress. Ductile metal, such as copper, can be employed as an interlayer as well to reduce thermal stress by plastic deformation of the interlayer. Several commercially available active filler metals were evaluated in a program (Table 4.19). Resultant strengths using copper are reflected in Fig. 4.12. Silicon carbide sintered with the addition of both boron carbide and carbon was solid-state diffusion joined by glass-encapsulated HIP. A strength of 450 MPa (65 ksi) was obtained both at room temperature and 1400 °C (2550 °F). The joining method for the green body was adopted to prepare a SiC nozzle ring model for gas turbine bench tests. The model was successfully operated at 1400 °C (2550 °F) (Ref 131).
Chapter 4: Base Metals and Base-Metal Family Groups / 125
The roles of titanium in active brazing of SiC have been studied extensively (Ref 132), while studies on the roles of silver and copper, which constitute the major parts of the active filler metals, have been overlooked. The effects of the relative contents of silver and copper in the filler metal on the interfacial reactions and bond strength were investigated in this study (Ref 132). The interfacial reactions can be divided into the decomposition reaction of SiC by the filler-metal melt and the interfacial reaction of titanium with SiC. Brazing by the Cu-5at.%Ti filler metal induced SiC to be decomposed, but the addition of silver to the filler metal suppressed the decomposition of SiC. Titanium carbide (TiC) and Ti5Si3 were produced from the interfacial reactions of titanium independent of the filler metals. However, their morphologies and formation mechanisms differ greatly, depending on the relative contents of silver and copper. The bond strength and fracture modes are also dependent on the relative contents of silver and copper. Good bond strength of 159 to 178 MPa (23 to 26 ksi) was obtained by brazing
with the Ag-5Ti (at.%) filler metal at 985 °C (1805 °F) for 600 s, and fracture initiated at the interface of the reaction product layer and propagated through SiC. Silicon-carbide-base ceramic materials have been joined by the reaction-forming technique (Ref 133) for high-temperature applications. This method is unique in terms of producing joints with tailorable microstructures. The formation of joints by this approach is attractive, because the thermomechanical properties of the joint interlayer can be tailored to be very close to those of the SiC-base materials. In addition, high-temperature fixturing is not needed to hold the parts at the infiltration temperature. A variety of SiC-base ceramics and fiber-reinforced composites have been joined using this approach (Ref 133–138). A flow diagram of the joining method is given in Fig. 4.13. The joining steps include the application of a carbonaceous mixture in the joint area and curing at 110 to 120 °C (230 to 250 °F) for 10 to 20 min. This step fastens the pieces together. Silicon or a silicon alloy in tape,
Table 4.19 Commercially available active filler metals evaluated for brazing of silicon carbide Chemical composition, wt% Filler metal
1. Ag-Cu-Ti 2. Ag-Cu-In-Ti 3. Ag-Cu-Ti 4. Ag-Ti 5. Ag-Cu-Ti 6. Ag-Cu-In-Ti
Solidus
Liquidus
Brazing temperature
Cu
In
Ti
Ag
°C
°F
°C
°F
°C
°F
27.5 23.5 26.5 ... 6 19.5
... 14.5 ... ... ... 5
2 1.25 3 4 3 3
70.5 60.75 70.5 96 91 72.5
780 605 803 960 875 732
1435 1120 1477 1760 1610 1350
795 715 857 960 917 811
1465 1320 1575 1760 1683 1492
840 760 950 1050 970 950
1545 1400 1740 1920 1780 1740
Source: Ref 130
Bending strength, MPa (ksi)
220 (32) –
Sample a
Bending strength SiC – SiC
Sample b
200 (29) – 180 (26) –
Sample c
160 (23) – 140 (20) – 120 (17.5) – 100 (15) – 80 (12) – 60 (9) – 40 (6) – 20 (3) – 0– 1
2
3
4
Braze filler metal No.
Fig. 4.12
Four-point bending strength of active-brazed SiC-SiC joints. Source: Ref 130
5
6
126 / Brazing, Second Edition
paste, or slurry form is applied in the joint region and heated to 1250 to 1425 °C (2280 to 2600 °F)—depending on the type of infiltrant— for 5 to 10 min. The molten silicon or silicon alloy reacts with carbon to form SiC, with controllable amounts of silicon and other phases as determined by the alloy composition. Joint thickness can be readily controlled in this process by controlling the properties of the carbonaceous paste and the applied fixturing force. This process has been applied to two commercially available SiC-base ceramics (Ref 134). The materials used were Cerastar reaction-bonded SiC (Cerastar RB-SiC) (The Carborundum Company) and Hexoloy-SA (sintered alpha-SiC) (Saint-Gobain Advanced Ceramics Corporation). The average room-temperature flexural strengths of the as-received and joined specimens were 157 ± 11 and 147 ± 10 MPa (23 ± 1.6 and 21 ± 1.5 ksi), respectively. These strengths increase at high temperatures, possibly due to healing of machining flaws. The flexural strengths of joined bars are comparable to those of as-received materials. The fracture origins appear to be inhomogeneities inside the parent material. The 23 and 1350 °C (73 and 2460 °F) flexural strength of as-received specimens of HexoloySA sintered SiC were 401.7 ± 26.2 and 402.1 ± 9.3 MPa (58.3 ± 3.8 and 58.3 ± 1.3 ksi), respectively. The joined specimens with 45 to 50 µm thick joints had strengths of 251 ± 13 and 267 ± 43.7 MPa (36 ± 1.9 and 39 ± 6.3 ksi), respec-
Apply carbonaceous mixture to joint area Cure at 110–120 °C for 10 to 20 min
Apply silicon or silicon alloy (paste, tape, or slurry) Heat at 1250–1425°C for 5 to 10 min
Strong and tough joints with tailorable properties
Fig. 4.13
Flow diagram for reaction forming method for joining silicon-carbide-base ceramics
tively. In the joined materials, fracture does initiate in the joint region. Inhomogeneous silicon distribution has been observed in certain areas of these joints. This joining method has been used to join a wide variety of SiC-base materials, including fiber-reinforced ceramic-matrix composites in different sizes and shapes, as shown in Fig. 4.14. It can also be used to join tubular components. There is the potential to extend this joining approach to the repair of SiC composite components in service (Ref 139). A pure aluminum foil between two blocks of SiC was fabricated by diffusion bonding for temperatures ranging from 500 to 600 °C (930 to 1110 °F) (Ref 140). An interfacial amorphous phase was observed at the aluminum-SiC interface for diffusion bonds performed at and above 586 °C (1087 °F). The formation of this phase was essential for producing a strong bond between aluminum and SiC. When the thickness of the amorphous phase was greater than 9 nm, another amorphous phase was observed to form adjacent to the original amorphous phase. The morphology of the two phases indicated that the two phases are formed by a solidstate reaction. The results from indentation and in situ hotstage TEM experiments demonstrated that the amorphous phase forms by a solid-state reaction rather than a mechanism involving the formation of a low-melting eutectic liquid. Finally, the formation of the interfacial glassy phases was explained using thermodynamic and kinetic considerations. It was postulated that the strong attraction between the aluminum, carbon, silicon, and oxygen atoms results in the formation of a glassy phase that has a higher thermodynamic stability than aluminum and SiC crystals with a thin layer of silica in between. Further, the formation of competing crystalline phases, such as Al4C3 and Al2O3, is kinetically suppressed, because the formation of these phases is composition-specific and hence requires longer time. The major conclusions of a study (Ref 141) can be summarized as follows: • Additions of silicon to gold lead to a strong decrease of the contact angle (from values much higher than 90° to values much lower than 90°) without any significant reactivity. Solidified drops adhere well on SiC substrates. • The good wetting and adhesion are due to the formation of a strong chemical bond between
Chapter 4: Base Metals and Base-Metal Family Groups / 127
the alloy and the ceramic, localized at the interface and interpreted as a covalent Si-SiC bond. • The wetting results are sensitive to oxygen, even if this element is present in the furnace at very low levels (PO2 ≈ 10–20 Pa). Dissolved oxygen improves the beneficial effect of silicon on wetting, and nearly zero contact angles have been reached in some cases. Thus, oxygen seems to be tension-active at metal/covalent ceramic interfaces, as it is at metal/ionic ceramic interfaces. Mullite and ZrO2-toughened mullite, which were reaction sintered from Al2O3 and ZrSiO2,
Fig. 4.14
were joined with silver-copper eutectic filler metals that contain titanium or zirconium as active elements. Although neither the fillermetal compositions nor the processing conditions were optimized, four-point bend strengths of joined bars were as high as 108 MPa (16 ksi). Joining reactions were studied by conducting separate sessile drop experiments with molten Ti-Ag-Cu and Zr-Ag-Cu filler metals on mullite for times and temperatures that were similar to those used for joining. Detailed compositional and microstructural analyses of those metalceramic interfaces revealed a complex reaction zone in which the oxide of the active element
Photographs showing components fabricated from joined silicon carbide subelements. Source: Ref 139
128 / Brazing, Second Edition
was a principal reaction product. Compared with the titanium-containing filler metals, those with zirconium were more refractory, less reactive with the mullite, and did not segregate as completely to the interface during heating (Ref 142). Aluminum Nitride (AlN). Researchers (Ref 143) investigated the brazing of AlN, which is a good ceramic substrate in high-power electronic applications, to copper using indium-base active filler metals. Compositions of filler metals were chosen as In-1wt%Ti (IT1), In-19wt%Ag2-wt%Ti (IAT2), In-15wt%Ti (IT15), and In52wt%Ag-20wt%Cu-3wt%Ti (ACIT3). Brazing operation was performed in vacuum at temperatures of 650 to 900°C (1200 to 1650 °F). The filler metals showed good wetting on AlN and led to a strong bond between AlN and the filler metal. From the microstructural analysis, no evidence of a reaction layer was clearly found at the interface under the experimental brazing conditions. The composition of the filler-metal layer changed into Cu9In4 phase due to the extensive dissolving of copper from the base metal. Bond strength, measured by the fourpoint bend test, was obtained as high as 225 to 295 N (23 to 30 kgf) for the Cu/AlN/Cu joint brazed with IT15 and ACIT3 filler metals and was shown to be nearly constant even when the temperature was varied within 700 to 800 °C (1260 to 1440 °F). Most of the fracture appeared to proceed through the interior of the AlN ceramic. Based on the experimental results, it is believed that a strong bonding between AlN and the filler metal can be achieved without the apparent forming of a titanium-rich reaction layer at the interface. Alumina (Al2O3). A major problem with brazing an oxide ceramic is the resistance to wetting caused by the oxides on the surface of the ceramic. A means of rectifying the problem is to apply pressure to the filler metal with sufficient force to counteract the repelling force of the oxides. A study was undertaken to assess the strength of the joints of metals brazed to Al2O3 with a copper filler metal. Results of tests indicated that the strength of the joint steadily increased as the pressure increased up to 5 MPa (0.7 ksi) and then leveled off with any additional pressure. Strength increased as the brazing temperature rose to 1100 °C (2010 °F) and then dropped with further increases in temperature. Length of holding time under pressure also affected strength. The amount of Al2O3 in the ceramic also played a role in joint strength, with those of 100% Al2O3
being only 50 to 60% as strong as those with 94% Al2O3. Titanium-containing filler metals were used to join both 94 and 99+% Al2O3 compositions: Filler metal composition, % Filler metal
Cusil ABA Incusil 10 ABA
Cu
Ag
Ti
In
Brazing temperature range, °C (°F)
27.5 27.0
70.5 62.25
2.0 1.25
... 9.5
820–860 (1510–1580) 770–800 (1420–1470)
Resulting tensile strengths of 76.8 to 109.8 MPa (11.1 to 15.9 ksi) compared favorably with conventional molybdenum-manganese metallizing of Al2O3 surfaces and subsequent brazing (Ref 144). The two different titanium-containing filler metals that were used were essentially copper-silver eutectic compositions containing small amounts (1 to 3 wt%) of titanium. The basic difference between the two filler metals was that one contained approximately 10 wt% In, whereas the second filler metal contained no indium. The test results led to two general conclusions concerning the strength data (Ref 144): • For a given ceramic, higher strengths are observed when brazing in a vacuum and when brazing with the indium-containing filler metal. • All sets of variables examined (filler metal, atmosphere, and ceramic composition) yielded comparable tensile strength to conventional metallizing/brazing techniques. In order to verify the previous conclusions concerning the filler metal as an active filler metal (ABA), a component was selected (Fig. 4.15) to see how well the component (94% Al2O3) performed when brazed with the indium-containing filler metal. The component
Copper contacts
94% alumina
Active braze alloy
Molybdenum/alumina cermet
Fig. 4.15 Ref 144, 145
Ceramic header used for testing titanium-containing filler metals for joining alumina. Source:
Chapter 4: Base Metals and Base-Metal Family Groups / 129
chosen was a 94 wt% Al2O3 ceramic header with two molybdenum-Al2O3 cermet electrical feedthroughs. Attached to the connector end of this header were two copper contacts, which were subsequently brazed to the molybdenumAl2O3 cermet surface. The results of the shear tests are contained in Table 4.20. The filler-metal preforms with 10% In yielded the highest shear strength. As a result, it was found that ABAs provide a simplified method of joining Al2O3 ceramics. Second, comparable tensile strengths can be obtained from ABAs as from conventional molybdenummanganese metallizing techniques. Active filler metal sealing results in the migration of titanium from the bulk braze to the ceramic (Al2O3) surface. Additionally, a filler metal, 25Cr-21V-54Ti, has been successfully used to join Al2O3 to itself without any metallizing coatings (molybdenummanganese or titanium hydride pretreatment) (Ref 145, 146). Table 4.21 lists some of the systems that have held out promise; however, only Ti-V-Cr and Ti-Zr-Ta have successfully wet
Al2O3. Compositions with at least 25% Cr readily wet and flow on Al2O3; however, less chromium results in limited or no flow. The ductility of the filler metals also depends on composition. Filler metals with less than 25% Cr are ductile and can easily be rolled into sheet. The success of the previously mentioned type of filler metals in wetting ceramics is through the employment of active metals, such as titanium and zirconium, as components of the filler metal. While the joining of ceramics (Al2O3) has been successful without coatings prior to brazing, other investigators have examined other metallic coatings in lieu of molybdenum-manganese processing and the titanium hydride treatment (Ref 147–149). Finally, in a comprehensive study of filler metals and several engineered ceramic materials for uncooled diesel engines, researchers (Ref 149, 150) evaluated two filler-metal systems (Cu-Ag-Ti, Table 4.22; and Cu-Au-Ti, Table 4.23), three types of Al2O3, and two types of ZrO2 (MgO-stabilized partially stabilized zirco-
Table 4.20 Braze test results in evaluation of titanium-containing filler metals for brazing of alumina Temperature Braze material
Incusil-10 ABA (0.05 mm, or 0.002 in.) Cusil ABA (0.05 mm, or 0.002 in.) Cusil ABA (0.03 mm, or 0.001 in.) Cusil ABA (0.03 mm, or 0.001 in.) No metallize/nickel/Cusil
Shear strength
Atmosphere
°C
°F
Time, min
N
lbf
Vacuum Vacuum Vacuum H2 H2
810 840 840 840 820
1490 1545 1545 1545 1510
5 5 5 5 6
525 489 142 187 534(a)
118 110 32 42 120(a)
(a) Specification requirement = 289 N (120 lbf). Source: Ref 144
Table 4.21 Ternary systems of filler metals evaluated for brazing of alumina Approximate brazing temperature Filler metal system
Ti-V-Cr Ti-Zr-Ta Ti-Zr-Ge Ti-Zr-Nb Ti-Zr-Cr Ti-Zr-B Ti-V-Nb Ti-V-Mo Source: Ref 146
Materials
°C
°F
Refractory metals
Graphite
Al2O3
1550–1650 1650–2100 1300–1600 1600–1700 1250–1450 1400–1600 1650 1650
2820–3000 3000–3810 2370–2910 2910–3090 2280–2640 2550–2910 3000 3000
x x x x x x x x
x x x x ... ... ... ...
x x ... ... ... ... ... ...
130 / Brazing, Second Edition
nia, or PSZ, and Y2O3-stabilized tetragonal zirconia polycrystal, or TZP). From the filler metals based on the silvercopper eutectic with additions of nickel and titanium, the best results were achieved with the Cu-26Ag-29Ti (wt%) composition. This filler metal produced wetting angles less than 30° on all the ceramics and flexural strengths at 400 °C (750 °F) of >165 MPa (24 ksi) for Al2O3 and >130 MPa (19 ksi) for a PSZ brazement. Of the gold-bearing filler metals, the Cu20Au-18Ti (wt%) had superior properties. Flexural strengths at 400 °C (750 °F) of brazements made with this filler metal range from 106 MPa (15 ksi) for one Al2O3 to 218 MPa (32 ksi) for another Al2O3. The room-temperature flexural strength of PSZ brazed with this material was 258 MPa (37 ksi). Finally, the fracture toughness of composite specimens of PSZ brazed with a Cu-27Ag-26Ti (wt%) filler metal averaged approximately the same as that of the bulk ceramic, 6 MPa·m1/2. Researchers (Ref 151) studied the wetting and spreading of copper-manganese filler metals on Al2O3 using a sessile drop technique. The con-
Table 4.22 Silver-copper filler metals evaluated for brazing of alumina and other ceramics Brazing range(a) Composition, wt%
Cu-27Ag-26Ti Cu-26Ag-29Ti Ag-38Cu-1Ni-4Ti Ag-37Cu-0.75Ni-7.25Ti Ag-35Cu-0.7Ni-10.3Ti
°C
°F
900–950 920–1000 950–1050 850–950 950–1050
1650–1740 1690–1830 1740–1920 1560–1740 1740–1920
(a) Brazing temperature depends to some degree on substrate material and on the extent of flow desired. Source: Ref 149
Table 4.23 Experimental filler metals with copper and gold as major elements evaluated for brazing of alumina and other ceramics Brazing range(a) Composition, wt%
Cu-14Au-4Ni-6.5Ti Cu-13Au-3.5Ni-14Ti Cu-22Au-10Ti Cu-20Au-18Ti Cu-18Au-26Ti Au-36Ni-11Ti Au-30Ni-21Ti
°C
°F
1090–1190 1050–1150 1350–1450 1050–1150 1050–1150 1050–1200 1150–1250
1995–2175 1920–2100 2460–2640 1920–2100 1920–2100 1920–2190 2100–2280
(a) Brazing temperature depends to some degree on substrate material and on the extent of flow desired. Source: Ref 149
tact angle, solid-liquid interfacial energy, and the work of adhesion values were calculated based on the experimentally measured spreading radius in the temperature range of 1100 to 1300 °C (2010 to 2370 °F). Using an in situ measurement technique, the spreading radius could be continuously monitored as a function of time. The results were compared to similar work on the copper-titanium/Al2O3 system. Similar spreading behavior was observed for both systems, and it is proposed that spreading occurs by the same interfacial reaction-product nucleation and growth mechanism. While the maximum work of adhesion was lower for copper-manganese filler metals on Al2O3 (1425 mJ/m2) relative to that for copper-titanium filler metals on Al2O3 (2800 mJ/m2), the copper-manganese/ Al2O3 adhesion values are sufficient for many direct-bonding applications. The formation of a thin reaction layer and the lack of brittle copper-manganese intermetallics provides the potential for better joint mechanical properties and easier processing for direct brazing of Al2O3, with copper-manganese filler metals replacing copper-titanium filler metals. Sandia National Laboratories sponsored a series of studies (Ref 152). The researchers found that titanium and zirconium dissolved in molten silver-copper filler metals wet AlN, Al2O3, and mullite and react with them. Contact angles were lower for titanium than for zirconium and for higher concentration of either metal in the silver-copper filler metal. Titanium reacts with AlN to give TiN0.7; zirconium reacts to give stoichiometric ZrN. Both reactions are consistent with thermodynamic predictions and with previous studies. Titanium reacts with Al2O3 to give a reaction layer containing oxygen and titanium in approximately a 0.4 to 0.6 ratio. That is consistent with the reaction 8 Ti + Al2O3 = 3 TiO0.5 + 2 Ti3Al, which is predicted by thermodynamic calculations. Zirconium reacts with Al2O3 to give a reaction layer that contains oxygen and zirconium close to the stoichiometric 2-to-1 ratio for ZrO2. That reaction is not predicted by simple redox thermodynamics. Additional driving force may be provided by formation of aluminum copper and Ag-AlCu-Zr phases observed in cooled specimens. In the case of mullite, the reaction product for the Ti-Ag-Cu filler metals appears to be a complex titanium-copper oxide instead of the simple ZrO2 that was found for reaction of Zr-Ag-Cu filler metals. Although interface reactions have not been conclusively identified, formation of
Chapter 4: Base Metals and Base-Metal Family Groups / 131
titanium and zirconium oxides from metal-mullite redox reactions seems likely. The grains with the Ti-Cu-O composition in the titaniumCusil specimens require further investigation. In an investigation (Ref 153), a new joining process for ceramics to ceramics and ceramics to metals, squeeze (SQ) brazing, was developed. This process uses squeeze casting; a brazing material is squeezed into the interface channel to be brazed and is solidified under a high pressure. This new process has several advantages: low cost, mass producibility, high interface strength, high reliability, no severe reaction, and so on. Severe reaction, which sometimes limits adoption of the active metal brazing for certain kinds of ceramics, for example, not only Al2O3silica but also zirconia, diamond, lead-zirconium titanate, and so on, is not a serious problem in the SQ brazing. Nonactive combinations such as Si3N4-aluminum could be tightly brazed by this process with a simple surface activation treatment (Tables 4.24, 4.25) (Ref 153). The simple calculation for the melt infiltration in the SQ brazing gave a guide to the critical pressure required for brazing. In the present case, the melt flow was assumed to be laminar. However, if the melt flow is turbulent, it effectively breaks the surface oxidation film and may cleanse the surface of ceramics by the jet formed preceding the moving melt front of a
Table 4.24 Summary of bending strength of alumina joints formed by conventional brazing and by squeeze (SQ) brazing Strength Brazing method
Atmosphere
Thickness of braze layer, µm
MPa
ksi
Conventional
Air Argon flow Air ...
... 10 10 300
0 49 325 228
0 7 47 33
SQ brazing
See text for description of the SQ brazing process. Source: Ref 153
Table 4.25 Summary of bending strength of silicon nitride joints brazed by the squeeze (SQ) method Material
As-received Si3N4 Preoxidation treated Si3N4 Source: Ref 153
Bending strength, MPa (ksi)
57 (8) 404 (58.5)
braze, as in the case of the explosion bonding process. Thus, the squeeze cast speed, which is determined by the relation between the ram speed and the braze-layer thickness, has a certain influence on the quality of the SQ-brazed interface. Examples used in the study were limited to the use of aluminum as a filler metal. However, other filler metals such as copper alloys, silvercopper alloys, nickel alloys, and gold alloys with active metals such as titanium, which have higher melting points than aluminum, could be used as filler metals under a certain condition control. Oxidation of active metals can be prevented only by adoption of an inert atmosphere in molten brazing materials. The SQ brazing itself can be performed in an air atmosphere, because the exposure of the active brazing materials to an oxidative exposure atmosphere is limited to a very short period. In addition, glass-ceramics can also be used as brazing materials in the present process, if a sufficiently low viscosity and high pressure are established. Thus, it is concluded that SQ brazing has great potential for brazing ceramics for a wide variety of applications. Researchers (Ref 154) evaluated the thermodynamic reaction products and layering in brazed Al2O3 joints. The investigation covered the joints formed by brazing Al2O3 to itself or to a titanium alloy (Ti-6Al-4V) with a Ag-Cu-Ti filler metal. In the brazing process, titanium in the filler metal reduces Al2O3 to form a series of reaction products that have a layered morphology. The formation of M6X-type compounds, Ti4Cu2O or Ti3Cu3O, at the interface was characteristic of these joints. The other reaction products also belong to the Ti-Cu-O system (with the reduced aluminum in solution), and hence, this subsystem was chosen to assess the thermodynamic stability of the joints. The Ti-Cu-O section was established experimentally at 945 °C (1730 °F), and activities of elements in three of the threephase regions were estimated based on the phase boundaries of the ternary section and available binary thermodynamic data. The estimated free energies of formation of the two M6X-type compounds, Ti4Cu2O and Ti3Cu3O, are –120 and –122 kcal/mol, respectively. The highly negative values for the free energies of formation suggest that these compounds are thermodynamically stable. The activity data were also used to generate activity diagrams for the Ti-Cu-O system. The layer sequences at the
132 / Brazing, Second Edition
Four-point bending strength, MPa (ksi)
joints satisfied the stability requirements based on the ternary section and the activity diagrams, indicating that even though the interfaces formed in a matter of minutes, they were at local thermodynamic equilibrium. The suitability of active brazing technology for joining TiC-strengthened Al2O3 (ATC) to itself and to stainless steel was evaluated (Ref 155). The main emphasis was put on the investigation of the microstructural and mechanical properties of the active brazed joints. Furthermore, the electrical properties were investigated by determining the specific electrical resistance of active-brazed Al2O3-TiC joints. For fabrication of the joints, two process technologies were applied: vacuum furnace brazing and induction brazing under shielding gas. Four-point bend tests revealed that some of the vacuum-brazed ATC joints reach bending strengths comparable to those of bulk ATC with the Ag-In-Ti filler metal (Fig. 4.16). Besides joining ATC to itself, active brazing also allows fabrication of ATC-metal joints. However, mechanical properties are generally poorer than those of ATC-ATC joints. It was also shown that the induction brazing process is an interesting alternative to the vacuum furnace brazing technique when ATC has to be joined to steels. When using the very ductile silver-base filler metals Ag-Ti and Ag-In-Ti, high-quality joints can be fabricated by induction heating. The main advantages of the induction brazing process are short processing time, cheap equipment, and low energy input into the parts. A group of researchers (Ref 156–158) examined the ceramic joining and bonding of Al2O3 via copper/niobium/copper interlayers.
395
400 (58) 350 (51)
345
352
Minimum Average Maximum
355 320
300 (44)
275
238
250 (36) 200 (29)
180 153
150
150 (22)
103
100 (15) 61
50 (7) 0 AgCulnTi 900 °C/10ft (1650 °F/3m)
Fig. 4.16
A method of ceramic-ceramic joining that exploits a multilayer interlayer designed to form a thin, potentially transient layer of liquid phase has been used to join Al2O3 to Al2O3. Microdesigned multilayer copper-niobium interlayers were used to achieve bonding at 1150 °C (2100 °F). Flexure strengths of as-bonded samples ranged from 119 to 255 MPa (17 to 37 ksi), with an average of ≈181 MPa (26 ksi). The ability to form strong ceramic-metal interfaces was also indicated by instances of ceramic failure. The impact of postbonding anneals of 10 h duration at 1000 °C (1832 °F) in gettered argon on room-temperature joint strength was assessed. High strengths (198 to 238 MPa, or 29 to 35 ksi) were obtained. The retention of strength following annealing in low-oxygen partial pressure argon differs from the behavior previously observed in copper-platinum-bonded Al2O3. The results indicate that, as in the case of platinum-Al2O3 bonding, high-strength joints between niobium and Al2O3 can be produced at significantly lower temperatures than those previously used for diffusion bonding by introducing a thick intervening layer of copper. Compared to copper/platinum/copper-interlayer-bonded Al2O3, fracture strengths were generally higher and less scattered. Sessile drop experiments comparing the wetting behavior of copper and niobium-saturated copper on Al2O3 indicate that niobium acts to decrease the copper contact angle, and this is thought to promote a more favorable strength distribution than obtained with copper-platinum-base interlayers. More generally, the results suggest that precoating substrates with a thin layer of a reactive metal may improve the degree of contact between the transient liquid and the substrate
CuSiAlTi 1050 °C/10ft (1920 °F/3m)
AuPdTi 1250 °C/10ft (2280 °F/3m)
PdNiTi 1250 °C/5ft (2280 °F/1.5m)
Four-point bend strength of vacuum-brazed TiC-strengthened alumina joints. Source: Ref 155
Chapter 4: Base Metals and Base-Metal Family Groups / 133
and thereby increase the average strength, narrow the strength distribution, or both. Researchers (Ref 159) compared titanium Al2O3 diffusion bonding and titanium active brazing. The product layer sequence in the reaction of solid titanium with Al2O3 and in the reaction of titanium dissolved in liquid silver with Al2O3 were compared. The phase diagram and the diffusion path in the Ti-Al-O system were used to describe the interface formation during brazing of Al2O3 with liquid Ag-4wt%Ti filler metal. During brazing, first Ti-Al(O) intermetallics were formed. When the filler metal was depleted from titanium, the formed Ti-Al(O) intermetallics decomposed into oxygen-saturated Ti(O) and Al, which in turn diffused through the Ti(O) and dissolved in the silver. Researchers (Ref 160, 161) examined the interfacial reaction between Al2O3 and copper-titanium filler metal during reactive metal brazing. The copper-titanium filler metals reduced Al2O3 to form TiO at the interface. Thermodynamically, reduction of Al2O3 is possible through the dissolution of the aluminum by the filler metal. At 1300 K, for example, interfacial reaction can proceed until the activity of aluminum reaches approximately 0.02 in Cu-20Ti (at.%) filler metal. With time, the TiO layer grew toward the center of the filler metal, following a parabolic rate law, at the cost of another complex oxide, presumably Ti3Cu3O, which formed next to the TiO. The activation energy of TiO growth was 208 kJ/mol (50 kcal/mol), which corresponds to the activation energy of oxygen diffusion in the TiO. Therefore, it appears likely that the growth of TiO is controlled by oxygen diffusion. Coated Al2O3 substrates have been joined with a nickel foil interlayer through solid-state diffusion bonding. Here, multilayer nickel and titanium coatings with a 25 µm total thickness were deposited on polycrystalline Al2O3 substrates through direct current (dc) magnetron sputtering. Thicknesses and number of layers were varied, and each coating was constructed to produce one of the three intermetallics in the nickel-titanium system: Ti2Ni, TiNi, or TiNi3. The coated substrates were joined by solid-state diffusion bonding at bonding pressures of 4 and 12 MPa (0.6 and 2 ksi). The two major challenges to general brazed ceramic-to-metal components are the lower melting temperature of most filler metals, which limits the service temperature, and the forma-
tion of deleterious intermetallic compounds that limit the mechanical properties. Previous research indicated that the nickel-titanium system would provide the needed reactive, highmelting-point alloy, with NiTi providing the most acceptable ductility. The composition and structure of the coatings and bonded specimen were evaluated by x-ray diffraction, scanning electron microscopy, and energy-dispersive microscopy. The highest shear strength values were 54 MPa (8 ksi) for a NiTi five-layer coating bonded at 12 MPa (2 ksi) for 300 min at 900 °C (1650 °F). Shear strengths of the Ti2Ni were generally lower, and, as expected, shear strength diminished with the formation of TiNi3. Zirconia (ZrO2). Filler-metal systems have been developed not only for joining Al2O3 but also for joining partially stabilized ZrO2. The contact angles of molten aluminum-copper filler metals and their wettability on CaOstabilized ZrO2 have been measured, and it was found that a ZrO2 joint brazed with Al-1.7Cu (wt%) filler metal has produced the maximum fracture strength of 105 MPa (15 ksi), compared to 52 MPa (7.5 ksi) for that brazed with pure aluminum at room temperature. This improved strength of ZrO2 is maintained at elevated temperatures up to 513 °C (955 °F). The joining strength of a ZrO2 joint brazed with an aluminum-copper filler metal is dominated by the mechanical properties of the aluminum-copper in addition to the wettability of the aluminumcopper filler metal against ZrO2 (Ref 162). Researchers (Ref 163) evaluated the effect of the interfacial structure on the joining strength of a ZrO2-ZrO2 joint with aluminum foil. An oxygen ionic conductivity of ZrO2 was applied to strengthen the ZrO2-ZrO2 joint brazed with aluminum. The interfacial structure was examined by conventional TEM and ultrahighvoltage electron microscopy, and the joining strength was evaluated by the four-point bending test. The results obtained in this work were as follows: • It can be expected that the ZrO2-ZrO2 joints brazed with aluminum can be strengthened by applying a dc voltage across the ZrO2 specimens. • The applied current, the applied quantity, and the initial thickness of the aluminum layer strongly affect the joining strength. • Interfacial structures differed dramatically,
134 / Brazing, Second Edition
•
•
• •
depending on whether a dc voltage was applied or not. When a dc voltage was not applied, ZrO2 was deoxidized at the aluminum-ZrO2 interface. (Al2O3)δ was formed at the aluminum side of the interface; Zr2Al3 was formed between (Al2O3)δ and ZrO2. When a dc voltage was applied, α-Al2O3 or (Al2O3)δ was formed at the aluminum side of the interface. Additionally, very fine grains were formed between Al2O3 and ZrO2. The grains of Zr2Al3, α-Al2O3, and (Al2O3)δ were relatively large in size. The interfacial structure greatly influenced the joining strength.
The active brazing of partially stabilized ZrO2 to ZrO2 was carried out using 57Ag-38Cu5Ti filler metal in a vacuum of 7 × 10–3 Pa (10–6 psi) (Ref 164). The effects of joining conditions, such as temperature (from 1073 to 1323 K) and holding time (from 0 to 60 min), on the joint strength and the interfacial reaction were investigated. The joint strength was evaluated by shear testing. It was shown that the brazing temperature and holding time had a strong influence on the joint strength; the latter was controlled mainly by the interfacial morphologies. A lower or higher temperature and a shorter or a longer holding time were disadvantageous to joint strength due to the thinner reaction layer or cracks present at the interface, the maximum joint strength being achieved at 1123 K for 30 min when the interface reaction layer thickness was approximately 4.4 µm. X-ray diffraction analyses revealed that the reaction products at the interface between ZrO2 and the filler metal were δ-TiO and γ-AgTi3, and a layer transition structure of ZrO2/δ-TiO/δ-TiO + γ-AgTi3/γAgTi3/Ag-Cu formed at the interface. A definite choice (Ref 165) can now be made of the most suitable interlayer material(s) for three diffusion-bonded/brazed material combinations (ZrO2-ZrO2, ZrO2-Si3N4, and ZrO2/ AISI 316). ZrO2/Interlayer/ZrO2. In the ZrO2/interlayer/ZrO2 combination, copper and nickel were the best interlayer materials, in view of their positive behavior during the bonding/brazing process and the relaxation of residual stresses during cooling. However, copper is relatively weak and has a low corrosion resistance, and, therefore, its use as an interlayer material was limited. This led to the conclusion that nickel was the most suited metal to join ZrO2 to
itself. Titanium and AISI 316 might give problems because of chemical reactions with ZrO2, and molybdenum is not suitable because of the relatively high residual-stress level in the ZrO2/molybdenum/ZrO2 combination. Similar arguments given for the ZrO2/interlayer/ZrO2 combination hold for the case of the ZrO2/interlayer/Si3N4 combination, as far as the choice of nickel as the most suited interlayer material is concerned. Copper may also be used as an interlayer material, but it is excluded because of its low strength and corrosion resistance. Titanium and AISI 316 are expected to give severe reactions with Si3N4 and are, therefore, not applicable as interlayer material. The use of molybdenum results in very high stresses, which make molybdenum also not suitable as an interlayer material. ZrO2/Interlayer/AISI 316. Joining ZrO2 to AISI 316 is expected to be possible using copper and nickel as interlayer materials. As shown previously, nickel has better mechanical and corrosion properties that make it favorable in applications under severe circumstances. Although the use of titanium results in relatively low residual stresses, its chemical behavior in combination with both ZrO2 and stainless steel is expected to be problematic, which hinders its use as an interlayer material. Application of the candidate interlayer materials molybdenum and AISI 316 (direct bonding) results in relatively high stress levels. These materials are, therefore, assumed to be less suited as an interlayer material in diffusion bonding/brazing ZrO2 to stainless steel AISI 316.
Special, Miscellaneous, and Significantly New Materials Diamond. Single-crystal diamond tools are well established in many applications involving the high-precision machining of nonferrous materials. While previously this was mainly a preserve of natural diamond, now synthetic monocrystal diamond products are assuming greater importance as a cutting tool material (Ref 166). For a long time, it was widely, but erroneously, considered that diamonds could not be brazed directly to a support. Normal practice has been simply to mount the diamond by sintering or to fix it in a bed of filler metal, with no chemical bond forming between the diamond and the filler metal. However, by using certain metals, such as titanium, zirconium, niobium, or
Chapter 4: Base Metals and Base-Metal Family Groups / 135
tantalum, diamond can also be wetted directly by the formation of a carbide. Nevertheless, at the high temperatures required for this operation, the components that are to be brazed react extremely sensitively to the presence of even the smallest quantities of oxygen. Above approximately 450 °C (840 °F), in an oxidizing atmosphere the carbon near the surface of the diamond reacts with the oxygen in the air to form carbon dioxide; the result is that nothing can adhere to the surface of the diamond. Direct brazing in open atmosphere is, therefore, no longer possible. Consequently, at these temperatures, brazing must be carried out under a high vacuum or an oxygen-free inertgas atmosphere. High-vacuum machines/furnaces are relatively expensive, requiring a substantial investment, and are not economical with small production batches, as would be the case for most diamond toolmakers. The development of a diamond brazing unit (2500 W output), with an inert-gas environment, makes the brazing of monocrystal diamond blanks (generally between 0.5 and 1.2 mm, or 0.02 and 0.05 in., thick) to WC or molybdenum supports substantially less complicated and more economical. Carbide and molybdenum are the preferred substrate materials for brazing diamond, because their coefficient of thermal expansion is relatively close to that of diamond, and therefore, no damage is likely to be caused as a result of thermally induced stresses. The inert gas, which consists of 95% Ar and 5% H2, is supplied to the braze box through a reducing valve in order to ensure the flow of gas required for this operation. The braze box itself is provided with means for receiving and discharging the gas and with a glass door for opening and closing the housing. The glass door controls the gas supply, in that by closing the protective door the gas is introduced automatically, and conversely, the supply is shut off when the door is opened. The temperature of more than 900 °C (1650 °F) required for brazing is generated by resistance heating, as shown in Fig. 4.17; that is, the workpiece, which is fixed between two electrodes, is heated by passing an electric current through the electrodes. With this system, it is possible to distribute heat uniformly when working with tool holders up to approximately 10 by 10 mm (0.4 by 0.4 in.) in size (width by height), although special designs may be supplied on request. The current to the electrodes is activated
and can be adjusted by either the foot pedal or a hand control. The machine is powered by a standard 220 V alternating current system. The titanium-activated filler metal used in the process is of particular importance. It is normally 0.4 mm (0.016 in.) in diameter and becomes eutectic at a temperature of more than 900 °C (1650 °F). Because clean and oxygen-free surfaces are essential for effective brazing, the parts to be brazed should first be cleaned by a system recommended for the cleaning of diamond. Then, the active filler metal is placed on the carbide or molybdenum support clamped between the electrodes. On closing the glass door, the inert gas flows into the unit. To ensure that the unit is entirely filled with gas, in other words, that all the oxygen has been expelled, the operator should wait for approximately 20 s before initiating resistance heating with either the foot pedal or the hand control. In order to prevent premature operation, the system works with a time delay of 18 s before the current starts to flow. The special filler metal is generally used in the form of a 0.5 mm (0.02 in.) strip per operation. This is sufficient for brazing an area of approximately 5 mm2 (0.008 in.2), which is adequate in most cases. The substrate (support) is first pretinned with the filler metal, and only then is the diamond placed on top of the substrate for brazing. The temperature is then carefully raised until the eutectic is reached, at which time the actual brazing process occurs.
(I) Fig. 4.17
The principle of resistance heating. I, electric current. Source: Ref 166
136 / Brazing, Second Edition
Tests show that with a braze layer 0.02 to 0.05 mm (0.0008 to 0.002 in.) thick between the molybdenum and diamond, the shear strength of the bond is of the order of 340 MPa (50 ksi). Composites. Silicon carbide-fiber-reinforced SiC-matrix composites (SiC-SiC) produced by chemical vapor infiltration have been developed for use in structural applications at temperatures approaching 1000 °C (1830 °F). These composites contain approximately 40vol% SiC fibers (Nicalon, Dow Corning Corporation) and are infiltrated to approximately 85% of the theoretical density with SiC. In order to fully realize the advantages of these materials, practical joining techniques are being developed. Successful joining methods will permit the design and fabrication of components with complex shapes and the integration of component parts into larger structures. Joints must possess acceptable mechanical properties and exhibit thermal and environmental stability comparable with the composite that is being joined. Studies have focused on joints produced using TiC+Ni and SiC+Si interlayers. The microstructures of joints have been characterized, and the results appear promising. In one approach, hot pressing elemental titanium, carbon, and nickel powders formed a TiC+Ni joint interlayer, and in a second approach, silicon infiltration of SiC and carbon powders, or carbon cloth, produced SiC+Si interlayers (Ref 167). Researchers (Ref 168) investigated the effects of brazing temperature on joint properties of continuous SiC fiber in reinforced-aluminum-alloy-matrix composite and the behavior of the metal-matrix composite (MMC) during the heating process. The following results were obtained: • There is a brazing temperature at which maximum joint strength is obtained in the brazing process. A brazing temperature lower than this causes the brazed joint to fail in shear at the brazed interface. A brazing temperature higher than this makes the brazed joint fail in the MMC. The matrix was Al-10Si-1Mg filler-metal foil. The maximum joint shear strength in this study was 77.7 MPa (11.3 ksi). The brazing condition was 580 °C (1075 °F) for 1800 s. • The MMC showed degradation in tensile strength at room temperature after heat treatment higher than 550 °C (1020 °F). • The MMC began to melt at a temperature of approximately 580 °C (1075 °F) and com-
pletely melted at approximately 600 °C (1110 °F). Voids were produced, owing to the solidification shrinkage of the matrix metal. • When the melting point of the matrix metal was close to the brazing temperature, the reaction between fibers and matrix metal occurred easily. As a result, accurate temperature control was needed in the brazing of the MMC using the aluminum alloy matrix (Ref 169). Researchers (Ref 170) investigated the properties and microstructure of brazed joints of SCS-6 SiC-fiber-reinforced β21S (Ti-15Mo2.7Nb-3Al-0.2Si, wt%) titanium-matrix composite (TMC). Brazed joint specimens were fabricated from TMC using two different forms of commercially available Ti-15Cu-15Ni filler metal. The brazed joint specimens were tested in air at room temperature and at 815 °C (1500 °F) using overlap tensile shear tests. Metallurgical and fractographic analyses were used to characterize the microstructure, filler-metal/ TMC interactions, and joint failure modes. The fractographic results indicated that TMC delamination is a dominant failure mode for this type of joint. At room temperature, the TMC brazed joint specimens failed by TMC delamination and TMC tensile failure, with the brazed joint remaining intact. Therefore, the performance of the brazed joint specimens at room temperature was limited by the interlaminar strength of the TMC and not by the braze strength. At 815 °C (1500 °F), the TMC brazed joint specimens exhibited a combination of delamination and braze shear failure. Thus, the high-temperature performance of the brazed joint specimens may be limited by both the TMC interlaminar properties and the strength of the braze (Ref 171). Concurrently, researchers (Ref 172–174) were evaluating rapid infrared joining (RIJ) of SC5-6/β21S TMCs by transient liquid phase (TLP). Rapid infrared joining has been shown not to have many of the problems associated with conventional joining methods. Two models were used to predict the joint evolution and fiber reaction-zone growth. A TMC, 16-ply SCS-6/β21S, has been successfully joined, with total processing times of approximately 2 min, using the RIJ technique. The process uses a 50 °C/s (90 °F/s) ramping rate, a 17 µm Ti-15Cu-15Ni (wt%) filler metal between the faying surfaces, a joining temperature of 1100 °C (2010 °F), and 120 s of time to join the composite material. Joint
Chapter 4: Base Metals and Base-Metal Family Groups / 137
shear strength testing of the rapid infrared joints at temperatures as high as 800 °C (1470 °F) has revealed no joint failures. Also, due to the rapid cooling of the process, no poststabilization of the matrix material is necessary to prevent the formation of a brittle omega phase during subsequent use of the TMC at intermediate temperatures, 270 to 430 °C (520 to 805 °F), for up to 20 h. Researchers (Ref 175) evaluated the brazeability of several aluminum-matrix composites. They found the following: • Wettability and spreading of an Al-12Si filler metal on aluminum alloys reinforced with Al2O3 particles depend on the temperature, type of matrix, and reinforcement fraction. An increase in reinforcement fraction decreases both wetting and spreading, which hinders the brazeability of the aluminum composites. • Composites with a matrix of 7005 aluminum alloy present higher wettability and spreading than those with a matrix of 6061 aluminum alloy because of the decrease in the melting temperature of the braze metal by zinc diffusion. Optimal wetting temperatures determined by drop formation test are 570 to 580 °C (1060 to 1075 °F) for the 7005-matrix composite and 580 to 590 °C (1075 to 1095 °F) for the 6061-matrix composite. • Long wetting times aid silicon penetration in the base matrix and the reduction of its melting point and can lead to the penetration of molten braze into the substrate, resulting in erosion. This problem is less significant for the 7005-matrix composite and can be reduced if shorter wetting times are used. • The flowability of filler metal into a capillary gap in real MMC joints (T or single-overlap joints) is appropriate in the case of 10 vol% reinforced alloys but is limited when 6061/Al2O3/20p is brazed, because of its low spread behavior. • When proper joint filling is obtained, the mechanical strength is controlled by the filler metal alone. Maximum shear strengths greater than 90% of the nominal BAlSi4 filler metal alone have been reached for the brazed 10% particle-reinforced composites, which are close to 50% of their strengths in the T6 condition. Using copper and silver as interlayers, pure Al-20%SiCp and 6061-20%SiCp produced successful bonds (Ref 176).
Metallic interlayers have also been used to bond 8090 (aluminum-lithium) MMC materials (Ref 177–179). Transient liquid phase diffusion bonding (TLPDB) involves placing a metal foil insert between the surfaces to be joined, to induce formation of a liquid region. The foils are typically copper, zinc, and silver for aluminum MMCs (Ref 180, 181). The process of eutectic bonding was employed using an interlayer of copper as a prepared filler metal (Ref 182). This type of bonding is similar to that of liquid phase diffusion bonding in that the joining is effected by a liquid phase that forms at a temperature below the melting point of the base materials. Brazing of 6061-50%B has been carried out using AlSi brazing foils (melting point of 580 °C, or 1075 °F) and a lower-melting-point alloy of Al-Cu-Zn with a melting point of 380 °C (715 °F) (Ref 183). A composite aluminumboron tape has been successfully brazed (and diffusion bonded), with use on the space shuttle (Ref 183). Joints apparently of good quality have also been produced in W6A.15A (606115%Al2O3p) MMCs using TLPDB (Ref 184). Foils of silver, copper, and BAlSi-4 were used as interlayers. The best consistency, in terms of joint shear strength (>175 MPa, or 25 ksi), was achieved using the silver at 575 °C (1070 °F) for 100 min and BAlSi-4 at 585 °C (1085 °F) for 20 min. However, the copper produced joints of higher shear strength, but results appeared more scattered. Liquid phase diffusion bonding has, with inserts of 2017 aluminum alloy copper and silver foils, been used to bond Al2O3-fiber-reinforced 6061 aluminum alloy. This produced bonds of higher strengths than bonds made using similar procedures but without interlayers (Ref 180, 185). The vacuum brazing of aluminum-boron composites involves temperatures of 520 to 620 °C (970 to 1150 °F), and because of this, it is necessary to protect the boron filaments by covering them with SiC to prevent aluminum-boron interaction and reduction of joint properties (Ref 186). Vacuum brazing of aluminum-Borsic composites (Borsic being boron fibers coated with SiC for protection) has led to the development of a hybrid titanium-clad aluminum-Borsic composite. The titanium foil acts as a diffusion barrier to prevent fiber-matrix degradation during brazing (Ref 187). Using either ER4047 (Al10–13%Si) filler metal or 250 µm 718 aluminum filler-metal foil, titanium-clad aluminum-Borsic
138 / Brazing, Second Edition
unidirectional composite sheets have displayed longitudinal and transverse tensile strengths 0.9 and 2.5 times, respectively, those of the uncladaluminum-Borsic MMCs (Ref 187). Aluminides. Trial tests have demonstrated that it is feasible to join the various titanium aluminides, which are ordered intermetallic alloys that are arousing considerable interest in the aerospace industry for use in gas turbine engines and high-performance military airframes. Like most materials of this type (ordered intermetallic alloys), they are very brittle at ambient temperature, although they exhibit good hot ductility. Intermetallic alloys generally have high strength and modulus, which are retained to very high temperatures, and often have good resistance to oxidation at elevated temperatures too. There has recently been much effort to improve the low-temperature ductility of these materials by suitable alloying as well as to develop improved high-temperature mechanical properties and oxidation resistance. This work has reached the point where two types of titanium-aluminide-base alloys can be seriously considered for use in aerospace applications. The two alloy systems of immediate interest are based on the intermetallic compounds Ti3Al (α2 aluminides) and TiAl (γ aluminides). The α2 alloys are made more ductile by alloying with β stabilizers, such as niobium, molybdenum, or vanadium, which improve the number of slip systems, refine the microstructure, and permit a small amount of β phase to be retained at low temperatures. Unfortunately, this is at the expense of the density of the alloy. The key to joining lies in controlling the phase transformation from the high-temperature β to the α2 phase. The TiAl-base γ alloys are less well developed than the α2 alloys, but their potential is greater because of their lower density and improved high-temperature performance. Again, β-stabilizing elements are added to improve the ductility, and the alloys are usually designed to give a predominantly γ microstructure, containing laths of α2. The Welding Institute has been very active in developing joining techniques for both α2 and γ alloys. Figure 4.18 shows a diffusion bond in an α2 alloy, from which it can be seen that a goodquality bond could be obtained at a temperature below the β transus, thus avoiding the complications of the β-to-α2 transformation. The com-
paratively high alloy content does not apparently compromise the ability of titanium to be joined by this process. Joining of titanium aluminide, TiAl, by a rapid infrared processing technique has been investigated at 1150 °C (2100 °F) using a titanium filler metal (Ti-15Ni-15Cu, wt%). The effects and results of joining and post-annealing on the microstructure and strength of the joint show that the joint shear strength is 220 MPa (32 ksi) when processed at 1150 °C (2100 °F) for 20 s and postannealed at 900 °C (1650 °F) for 2 h. Microstructural examinations of the joint with both optical microscope and scanning electron microscope indicate that good wetting exists between the filler metal and TiAl for most joints. The braze-affected zone thickness shows little increase with joining time and does not show a direct correlation with joint strength. Meanwhile, the base material exhibits no noticeable microstructural or mechanical property changes as a result of infrared processing (Ref 188). Brazing of NiAl alloys was conducted in a two-step operation in a vacuum chamber. In the first step, specimens were heated to 1520 °C (2770 °F) and held for 5 min. Due to local evaporation of aluminum, a thin layer of self-generated filler metal forms on the faying surfaces, producing a glazed appearance. The resultant nickel-rich alloy had a melting point below 1520 °C (2770 °F), and far below that of the NiAl base metal (1638 °C, or 2980 °F). In the second step, the glazed faying surfaces were placed in contact and brazed by heating at 1520 °C (2770 °F) for 15 min. Sound joints have been produced with grain growth across the braze interface. Therefore, at the completion of the brazing process, the joint is not detectable by
50µ
Fig. 4.18
A diffusion bond in Ti-24Al-11Nb alloy
Chapter 4: Base Metals and Base-Metal Family Groups / 139
metallography. Some homogenization is produced during brazing. A separate postbraze heat treatment may be required for complete chemical homogenization. No foreign material (in the form of filler metals, fluxes, or any other substance) is added to the base metal. The mechanical properties of the joint are expected to match the base metal. This concept may be applicable to other material systems in which one element can be evaporated from the surface to produce a layer of material with a lower melting point than the base metal (Ref 189). Powder Metallurgy (P/M) Materials. The most often used P/M filler metal is a patented Cu-Mn-Ni-Fe powder that can be admixed with flux and a lubricant for preforming into slugs/rings or dispensed as a paste (Ref 190). This material is suitable for use with iron, steel, low-alloy, or stainless P/M components. Typically, 3 to 4 g (0.1 to 0.14 oz) of filler metal is required for each square inch of surface area to be joined. The quantity is dependent on joint design, joint clearance, and component density. Closer joint clearance and higher component density favor lower amounts. Some applications requiring specific properties may require the use of a silver-or nickel-base filler metal for electrical, high-temperature, or corrosive environments. Uniform heating of the components along with sufficient dwell time, to equalize the temperature between the parts to be joined, is important for proper flow and joint fill. A component having a lower mass or thinner section size likely achieves the flow temperature before a larger mating part. This results in a greater capillary force toward the area having the highest temperature or heat energy and may cause poor joint fill. If this does occur, it is sometimes helpful to reposition the components on the furnace belt to assist in achieving uniform temperature between the parts. It may also help to increase the preheat temperature, so parts can equalize before reaching the liquidus point. For Ancorbraze 72, that is approximately 1065.6 °C (1950.1 °F). Most of the commonly used sintering atmospheres can be used for P/M brazing. To assist in oxide reduction, a flux is generally added to the filler metal. In some circumstances, an additional borate or fluoride-type flux may be swabbed onto the mating surfaces, particularly when brazing sintered parts, stainless steels grades, materials containing sulfur or manga-
nese sulfide, and when wetting large surface areas to be joined. It has also been determined that high CO2 percentages in the sintering atmosphere can oxidize the fluxing agents and reduce the flowability of the filler metal. By far, the most common difficulty experienced with furnace brazing P/M components is the lack of filler metal in the joint. This is generally associated with the following conditions: • Improper cleaning or oxide reduction • Excessive joint clearance • Low sintering temperature or nonuniform part temperature • Insufficient filler metal • Entrapped lubricant, flux, or gases • Movement of the mating parts before solidification Graphite. The joining problems associated with brazing of graphite are very similar to those encountered in brazing the ceramic materials previously discussed (Ref 191, 192). Wetting of graphite is more difficult than wetting of metals, and the differences in coefficients of thermal expansion between graphite and conventional structural materials are pronounced. It should be noted that carbon and graphite can also be brazed both to themselves and to metals. These materials vary widely in degree of crystallinity, in degree of orientation of the crystals, and in size, quantity, and distribution of porosity in the microstructure. These factors are strongly dependent on the starting materials and on processing and, in turn, they govern the physical and mechanical properties of this product. Carbons and graphites can be manufactured by several processes that yield materials with a wide range of crystalline perfection and properties. In the most widely used process, polycrystalline graphites are made from cokes produced as by-products from the manufacture of petroleum or from natural pitch sources (Ref 193). The wetting characteristics of all the carbons and graphites are strongly influenced by impurities, such as oxygen or water, that are either absorbed on the surface or absorbed in the bulk material. Moisture absorption always occurs to some extent, with levels as high as 0.25 wt%. Brazeability also depends on the size and distribution of pores, which can vary significantly from one grade to another. A major consideration in brazing of carbon and graphite is the effect of the coefficient of thermal expansion of these materials, which can
140 / Brazing, Second Edition
range from 2 to 8 × 10–6/°C (1 to 4 × 10–6/°F) between 25 and 1000 °C (75 and 1830 °F). Before carbon or graphite is brazed, the type and grade of the material must be established so as to ascertain its expansion characteristics. This information is also important for brazing of carbon or graphite to itself. Joint failure, particularly during thermal cycling, may occur if too great a difference exists between the coefficients of thermal expansion of the graphite and the filler metal. If the joint gap increases significantly on heating because of a large mismatch in coefficient, the filler metal may not be drawn into the joint by capillary flow. On the other hand, if the materials and joint design cause the gap to become too small, the filler metal may not be able to penetrate the joint. In conventional brazing of dissimilar materials, the material having the greater coefficient of expansion is made the outer member of the joint. Joint tolerances are used that do not allow the gap between the surfaces to become too great for capillary flow. Additional problems occur in brazing of dissimilar materials when one part of the joint is a carbon or graphite. Carbons and graphites have little or no ductility and are relatively weak under tensile loading. These adverse conditions are usually compensated for in graphite-tometal joints by brazing the graphite to a transition piece of a metal, such as molybdenum, tantalum, or zirconium, with a coefficient of expansion near that of the graphite. This transition piece can be subsequently brazed to a structural metal if required. This minimizes shear cracking in the graphite by transposing the stresses resulting from the large difference in thermal expansion to the metallic components. Thin sections of metals that deform easily when stressed, such as copper or nickel, have also been successfully used for brazing dissimilar metals. Metals that have strong tendencies to form carbides (titanium, zirconium, silicon, chromium) have been found to wet graphite when they are molten. A commercial filler metal frequently used for brazing of graphite is silver-copper-clad titanium wire. Graphite is also readily wetted by molybdenum disilicide, titanium, and zirconium. In recent years, the requirements of the aerospace and nuclear industries have resulted in the development of several additional filler metals. In general, these filler metals incorporate substantial quantities of carbide-forming ele-
ments. They include 48Ti-48Zr-4Be, 35Au35Ni-30Mo, 70Au-20Ni-10Ta, nickel-clad titanium, 54Ti-21V-25Cr, 43Ti-42Zr-15Ge, and 47Ti-48Zr-5Nb (Ref 192). Additionally, the filler metal 49Ti-49Cu-2Be has been recommended for brazing of graphite as well as oxide ceramics. The filler metals mentioned previously wet graphite and most metals well in either a vacuum or inert atmosphere (pure argon or helium) and span a fairly wide range in brazing temperatures, from 1000 °C (1830 °F) for 49Ti49Cu-2Be to 1350 °C (2460 °F) for 35Au-35Ni30Mo (Ref 193, 194). At least two commercially available filler metals wet carbon and graphite as well as a number of metals. The first has a composition of 68.8Ag-26.7Cu-4.5Ti, a solidus of 830 °C (1525 °F), and a liquidus of 850 °C (1560 °F). This filler metal is suitable for low-to-medium-temperature applications but appears to have only moderate oxidation resistance. The second commercially available filler metal for graphite brazing has a composition of 70Ti-15Cu-15Ni. With a somewhat higher melting range (solidus, 910 °C, or 1670 °F; liquidus, 960 °C, or 1760 °F), and by virtue of its greater titanium content, it has better oxidation resistance than the silver-base filler metal. Some workers in Europe and the former countries of the Soviet Union have reported successful joining of graphite to steel using a filler metal with a composition of 80Cu-10Ti10Sn at 1150 °C (2100 °F) (Ref 195). In other work, using a technique called diffusion brazing, a metallic interlayer was placed between the graphite components; the components were pressed together at a specific pressure and heated to the temperature of formation of a carbon-bearing melt or eutectic. On heating to higher temperatures, the melt dissociated with the precipitation of finely divided crystalline deposits of carbon that interacted with graphite base material to form a strong joint (Ref 196, 197). Researchers (Ref 198) have developed a procedure for brazing a special grade of graphite to a ferritic stainless steel for a seal in a rotary heat exchanger. It seems apparent that the selection of type 430 stainless steel was based at least partly on its lower coefficient of thermal expansion compared with that of a typical austenitic stainless steel. In addition, a joint geometry was developed that minimized the area of the brazed joint, thereby reducing thermally induced stresses to acceptable levels. Specimens of graphite brazed in a vacuum furnace to thin type
Chapter 4: Base Metals and Base-Metal Family Groups / 141
430 stainless steel sheet with either Ni-20Cr10Si or Ni-18Cr-8Si-9Ti filler metal at 1125 to 1175 °C (2060 to 2150 °F) performed well in tests at 650 °C (1200 °F).
Special Ceramic Joining Methods (Coatings) Ceramic and Ceramic-Metal Joining Methods. It is difficult to form joints by brazing ceramics and/or ceramics-to-metals, particularly oxide ceramics, because few commercially available filler metals can wet their surfaces. This problem is usually overcome by treating the ceramic surfaces to give them metallic-wettable characteristics. For example, metal coatings can be applied by vapor deposition or sputter ion plating. However, the most widely used metallization process for oxide ceramics is that in which mixtures of a glassy phase and a refractory metal are applied as paint by brushing or screen printing, as exemplified by the moly-manganese process. The prime purpose of the glassy phase is to bond the refractory metal to the ceramic oxide, while that of the refractory metal is to render the surface amenable to electroplating and ultimately, brazing to make a joint. Metallization of oxide ceramics followed by plating and brazing has been established for several decades. Therefore, the problems involved in its application are those of practice rather than principle. Several other techniques for joining ceramics have been reported in the literature (Ref 97, 113). The sintered metal powder technique is a widely used brazing method. This process requires several steps to produce a joint: • The firing of metal powder held in suspension on the ceramic • Plating or deposition of a thin copper or nickel film • Brazing by conventional methods to make the ceramic-to-metal joint The affinity of titanium and zirconium for ceramics is the principle behind the active-alloy process. Highly active titanium or zirconium can be made available at the ceramic-metal interface by hydride decomposition of a powder slurry on the ceramic surface. The reaction of the titanium or zirconium with the ceramic, or with additional metals placed at the interface,
forms an intimate bond. In some cases, the titanium or zirconium is merely painted on the ceramic surface and placed in contact with a suitable filler metal and the base metal to which the ceramic is to be joined. This process has an advantage over some others in that only one firing operation is required. The affinity of reactive and refractory metals for ceramics is also the basis for the direct brazing approach. In this case, active metals are incorporated as one or more of the constituents in the filler metal, which is placed at the joint as in conventional brazing. An interesting means of promoting flow of the filler metal involves vapor-deposited coatings of titanium, zirconium, or other metals on the ceramic substrate. Electron beam heating has provided a unique means for producing the metal vapor (Ref 96, 199–201). Active metal brazing (ABA) has been practiced since the 1940s, when titanium hydroxide in an organic solvent was coated onto ceramic seals that were then sandwiched around a silver (or gold or copper) filler metal and brazed together in vacuum. The trouble was that under high vacuum, the excess titanium in the filler metal formed a brittle joint, while under low vacuum, most of the titanium formed brittle oxides and nitrides, and little was left for the actual active brazing. One way around the problem was to clad titanium wire or foil with a filler metal. While that protects the titanium from the furnace atmosphere until the filler metal starts to melt so that more titanium is available for brazing, the filler metal itself is not ductile enough to yield to stresses caused by the different rates of thermal expansion for the ceramic and metal. Filler-metal manufacturers attacked the problem by developing a series of filler metals consisting of Cu-Ag-Ti and Cu-Ag-In-Ti. Similar to cladding, alloying protects the active elements (titanium) until the filler metal starts to melt. Because the titanium is protected, less is needed (1.25 to 2.0 wt%), and the resulting braze is less hard and ductile enough to form strong, reliable ceramic-to-ceramic and ceramic-to-metal hermetic bonds without prior metallization (Table 4.20, Fig. 4.15) (Ref 98, 101, 102, 147, 150, 202–205). The system works best when the ceramic surface is free of surface fractures. Researchers found it took more than twice as much energy to peel apart an Fe-Ni-Co (Kovar) strip brazed to a
142 / Brazing, Second Edition
lapped ceramic surface as from a ground ceramic surface. Joint strength for Al2O3 ceramic has been shown to be equal whether prepared and joined by the molybdenum-manganese braze system or the ABA system (Fig. 4.18) (Ref 101, 203–206). A unique ceramic-to-metal seal system was developed (Ref 200) based on a composite Al2O3 insulator and a 49%Ti-49%Cu-2%Be filler metal. Pulsed laser beam, furnace, and induction joining techniques were all successful in brazing the sensor electrodes, ceramic insulator, and end seals in a number of brazing sequences and operations. Pure metal coatings have also been evaluated. A niobium coating film has been electrolytically precipitated at a temperature of approximately 740 °C (1365 °F) out of a fluoride melt onto sintered Al2O3. A thin film of niobium produced by gas phase precipitation provided the electrical conductivity of the surface that is essential for electrolysis. Metallizing coatings can also be created completely by means of gas phase precipitation, for example, by the thermal dissociation of halides or organic metal compounds (chemical vapor deposition). Tungsten coatings, for example, can be produced from WCl6 in a helium atmosphere at a temperature of approximately 1000 to 1300 °C (1830 to 2370 °F). Brazed joints on this type of metal coating have created high-vacuum seals; at ambient temperature, they have a transverse tensile
Fig. 4.19
strength of between 40 and 100 MPa (6 and 15 ksi). In cases where high-temperature-resistant metals, such as tungsten or niobium, are involved, their thermal resistance depends on the behavior of the filler metal used. The steps seen in Fig. 4.19 initially start as a primary metallizing layer, which is the mixture of refractory metal and glass that is coated onto the surface of the ceramic. The most commonly used material is a “paint” containing molybdenum and manganese. An alternative choice of refractory metal is tungsten, which is used when gold is to be plated onto the surface without an intermediate layer of nickel. The paint is an intimate blend of a refractory metal (tungsten or molybdenum), manganese oxide when the metal is molybdenum, a glass frit, a carrying vehicle, and a solvent. Typical formulations are given in Ref 204, 207, and 208. In achieving the best possible metallizing, many parameters need to be considered, of which the most important are the application method and the process parameters, particularly the temperature, time, atmospheric conditions, and the thickness of the paint. There are several ways of applying the paint to the ceramic, ranging from simple manual dipping to fully automated machines. Among the more popular methods are spraying, screen printing, syringe or nozzle brushing, and transfer using a wheel or tape. The application method chosen is largely determined by the
Comparison of two methods (metallizing/brazing and active brazing) for joining Kovar to Al2O3. Source: Ref 9
Chapter 4: Base Metals and Base-Metal Family Groups / 143
shape of the article, the complexity of the metallizing pattern, and the quantity or quality desired. No matter which method is chosen, it is essential to ensure that the paint layer is of the correct thickness, usually at least 5 times the molybdenum or tungsten particle size, which is commonly approximately 5 µm. The thickness achieved in practice depends on both the paint and the method of application. One of the most important paint characteristics is its viscosity. Paint thicknesses of less than 5 µm can result in braze penetration during subsequent assembly operations, while thicknesses greater than 5 µm may cause cracks or blisters during drying that may result in leaks or peeling of the metallizing. During the firing cycle, the metallizing paint undergoes many changes. At an early stage, there is decomposition and elimination of organic materials in the paint. Later, and even more importantly, there is sintering of the refractory metal particles and diffusion of the glassy phase from the paint onto and into the ceramic, or vice versa. The bonding between the ceramic and the metallizing layer is produced by diffusion of a glassy phase from the paint, in the case of highpurity ceramics, or by diffusion of glassy binder phase from the ceramic into glass-free paint on the surface of a debased ceramic. The most commonly used filler metals do not wet the surface of the primary metallizing layers. To achieve wetting, a secondary metallizing layer of a metal, such as nickel, copper, gold, tin, or lead, is deposited by plating (electrolytic or electroless). The plating operation contributes as much to the success or failure of the metallizing structure as the primary layer. The process must be controlled throughout. The roles of the plated layer are to aid wetting of the refractory metal by the filler metal and to resist braze penetration into the primary layer. One key parameter that affects both these roles is the thickness of the plated layer. If it is too thin, the filler metal may dissolve it, but if it is too thick, blisters can form during the braze process that lead to very poor adherence and leaks. In general, a plating thickness of 2 to 4 µm is sufficient to avoid such problems. This molybdenum-manganese-silicate coating (Ref 148) is a particular coating that enables joints having a transverse tensile strength of between 70 and 120 MPa (10 and 17 ksi) to be effective. In contrast to the conventional molybdenum-manganese processes (Ref 209) that re-
quire SiO2 from the ceramic in order to create a bonding phase, the former method permits strongly bonded metallization of very pure Al2O3 ceramics (99.8% Al2O3) and other ceramic oxide materials such as ZrO2 (Ref 148). Considerable success is being achieved at present in the application of reactive metal brazing (Ref 210). The next few years, however, could see a considerable expansion in the market for reactive metal brazes if their performance and associated fabrication procedures are optimized. Current filler metals for ceramics employ titanium as their reactive component, but it should be possible and perhaps technically preferable to use other reagents for some applications. Thus, the capability of titanium to form intermetallic compounds with many workpiece metals can be detrimental, and replacement by other multivalent metals, such as chromium, that can form ceramic reaction products with wide ranges of stoichiometry could be advantageous—although substantial development work will be required to optimize and produce the new filler metals in commercially usable forms. Improving the usability of existing filler metals is also of importance, and developments could come from the application of the rapid solidification technique to brittle alloys to produce the thin sheets needed for stamping out preforms. Of equal importance with the development of materials are the production of joint design codes and the acquisition of engineering performance data. These, combined with the tailoring of filler metals for use with specific materials, should enhance the acceptability of brazing as a joining process for ceramics and make it possible to exploit it more efficiently in advanced engineering projects (Ref 211). Finally, aluminum nitride (AlN) is currently under investigation as a potential candidate for replacing Al2O3 as a substrate material for electronic circuit packaging. The requirements for such a material are that it can be metallized and joined to produce hermetic enclosures for semiconductor devices as well as its coupling with its expansion matching that of silicon. A technique for brazing AlN using a nonactive filler metal has been developed. The study found that active filler metals, namely titanium, wet the AlN, whereas nonactive ones are nonwetting. This method has been used to successfully join AlN to a low-expansion lead-frame alloy. The interfacial reactions led to the formation of titanium and zirconium nitrides, and these
144 / Brazing, Second Edition
interfacial reactions enable the ceramic surface to be wet by the molten filler metal in the development of AlN packages for microcircuit applications (Ref 211). The characteristics of the molybdenummanganese, active metal, and active hydride processes are summarized in the following: • The molybdenum-managanese process is a multistep sealing process in which the ceramic surface is metallized and plated with one or two metals before brazing can take place. The operations are conducted at a high temperature in a controlled atmosphere of H2; H2 firing may discolor some ceramics and produce conductive surfaces. Despite the number of steps required to produce a seal, the moly-manganese process can be automated quite readily, and minor deviations in the process variables can be tolerated (Ref 96, 101, 212). • The active hydride process is essentially a single-step process in which hydride reduction and brazing proceed simultaneously. Joining in a vacuum or in a controlled atmosphere of H2 or an inert gas is accomplished at relatively low temperatures, permitting a fast brazing cycle. This process is more difficult to automate than the moly-manganese process, particularly if the joints are produced in a vacuum. Careful control must be exercised in coating the ceramic with the hydride. Even though the process is considered a onestep process, the hydride process has been supplanted by the active metal process. • The active metal process may be a one-step operation similar to the active hydride process. Joining proceeds at high temperatures in a vacuum or in a controlled atmosphere; vacuum joining is not readily automated.
Dissimilar-Material Combinations Many dissimilar-metal combinations may be brazed, even those with metallurgical incompatibility that precludes welding. Important criteria to be considered start with differences in thermal expansion. If a metal with high thermal expansion surrounds a low-expansion metal, clearances at room temperature, which are satisfactory for capillary flow, are too great at brazing temperature. Conversely, if a low-expansion metal surrounds a high-expansion metal, no clearance may exist at brazing
temperature. For example, when brazing a molybdenum plug in a copper block, the parts must press fit at room temperature; if a copper plug is to be brazed in a molybdenum block, a properly centered loose fit at room temperature is required. In brazing tube-and-socket-type joints between dissimilar base metals, the tube should be the low-expansion metal and the socket the high-expansion metal. At brazing temperature, the clearance is maximum, and the capillary fills with filler metal. When the joint cools to room temperature, the brazed joint and the tube are in compression. For a tongue-in-groove joint, one should place the groove in the low-expansion material. The fit at room temperature should be designed to give capillary joint clearances on both sides of the tongue at brazing temperature. As noted previously, the joining of dissimilar metals has become increasingly important during the past two decades because of the service requirements for structures used in missiles and rockets, supersonic aircraft, nuclear equipment, marine systems, electronics, and chemical-processing equipment. Although certain dissimilar metals have been routinely joined for many years, the advent of space and nuclear requirements has produced a need for sophisticated methods of joining the new structural materials that have been developed for these demanding applications. These new alloys possess exceptional mechanical properties and resistance to corrosive media under extreme operating conditions. However, such alloys are frequently used only for sections of a structure where their specific properties are required; conventional alloys are used for the remainder of the structure for reasons of economy, weight, ease of fabrication, and so on. Thus, there is a need for procedures for producing reliable joints between dissimilar metals. The ability to design and fabricate such joints is essential to many segments of our industrial economy. The selection and use of dissimilar metals in structural applications is governed by the service requirements of the structure and by material and fabrication costs. For example, a relatively inexpensive grade of steel may be used in fabricating the shell of a vessel for the chemical industry for reasons of economy, whereas the corrosion requirements are satisfied by lining the vessel with thin-gage tantalum or titanium. Corrosion of a dissimilar-metal joint is inevitable if there is an electrolyte present. The
Chapter 4: Base Metals and Base-Metal Family Groups / 145
degree of corrosion that takes place is dependent on the type of electrolyte and the difference in electromotive potential between the dissimilar metals and/or materials. The problem of galvanic corrosion between dissimilar metals has been discussed (Ref 213). The magnitudes of the solution potentials of metals and intermediate compounds depend on the nature of the electrolyte and the difference in potential between the alloys being joined. Applications requiring the joining of dissimilar metals are discussed briefly: • The lunar module contains 26 pressure vessels in its descent and ascent stages (Ref 214). Depending on their functions, these vessels are fabricated from various alloys plus titanium. Coextruded titanium/stainless steel transition joints are used to connect the titanium pressure vessels to the stainless steel feed system. • Joints between beryllium and such metals as aluminum, stainless steel, and titanium are encountered in space-vehicle design, where beryllium is an attractive structural metal because of its low density, its stiffness under load, its resistance to damage by impact with meteors, and its high heat capacity. • Procedures are required for joining dissimilar metals in nuclear reactor construction. These applications range from the cladding of fuel elements with zirconium alloys to fabrication of dissimilar-metal piping joints. • Dissimilar-metal joints are encountered in aircraft hydraulic and ducting systems as well as in engine and airframe construction. Because significant quantities of titanium alloys are used in new aircraft, there are occasions to join titanium to other structural alloys to meet specific design requirements (Ref 212). In general, dissimilar metals are used in structures to provide: • High-temperature or low-temperature strength • Resistance to oxidation, corrosion, or wear • Resistance to radiation damage • Other required properties Also, the use of dissimilar metals is often attractive from the standpoint of cost. Joints between ferrous and nonferrous metals are of interest to industry because they combine the strength and toughness of steels with the
special properties—such as oxidation resistance, corrosion resistance, and so on—provided by nonferrous metals. Joining of ferrous to nonferrous metals is far more complicated than joining of dissimilar ferrous metals because of the wider variation in the physical, mechanical, and metallurgical properties of the metals being joined. The extent of these property differences is an excellent indication of the difficulty to be anticipated in joining such metals (Ref 215–218). In brazing steel to copper, the steel heats much more rapidly than the copper, unless provision is made to equalize the heating rates. In the practice of induction brazing, the inductor is coupled more closely to the copper than to the steel or has additional turns to that portion of the inductor heating the copper. Similar provision must be made in brazing carbon steel to brass or austenitic (nonmagnetic) stainless steel. Carbon steel heats faster than either of these materials, although the differential in heating rates is less than that between carbon steel and copper. The problems inherent in joining dissimilar nonferrous metals are similar to those encountered when ferrous and nonferrous metals are joined, because of the differences in the physical and metallurgical properties of the base metals. Some dissimilar nonferrous metals have been joined routinely for many years; others, such as aluminum to titanium, titanium to nickel, titanium to Al2O3, and aluminum to uranium, are new combinations that owe their existence to their applications in aerospace, electronic, and nuclear hardware (Ref 9, 15). Brazing and metallizing of ceramics to form joints with metals find many uses, especially in small-scale and electronic applications, for example, Al2O3 to Fe-Ni-Co alloy for vacuumtube production. The molybdenum or molybdenum-manganese metallization layer is painted onto a ceramic, allowing a subsequent braze layer to wet. Brazing temperatures are high, with 1580 °C (2875 °F) being typical. At this temperature, the glassy phase in Al2O3 begins to mix with the metallizing mixture of semisintered molybdenum or molybdenum-manganese. The literature discloses several theories relating to this mechanism, including capillary action (Ref 101, 219–221). Metal-to-Metal Joining with Coatings. A study was conducted (Ref 222) of the brazeability of nickel-base filler-metal foil for joining nickel base metal to mild steel base metal. Alloy 600 was selected, which was clad to a mild steel
146 / Brazing, Second Edition
base metal, SM41B. Four different nickel filler metals in foil form were tested. These foils were BNi-3, Ni-Si-B, BNi-1, and BNi-2. The brazeability of these filler metals was determined by the shear strength, the observation of the cross-sectional microstructure, the hardness distribution, and the elemental distribution. The following results were obtained: • The shear strength of the specimens made with BNi-1 and BNi-2 filler-metal foils containing chromium and iron was higher than clad specimens made with BNi-3 and Ni-SiB filler-metal foils. Under all brazing conditions, the location of the failure during the shear test occurred in the brazed joint. If the eutectic structure remained in the brazed joint, the break occurred in the eutectic structure. The shear strength was increased by decreasing the amount of cross-sectional area of the eutectic structure at the brazed joint. The shear strength was increased to maximum by increasing the diffusion brazing time, which eliminated the eutectic structure in the brazed joint. • Brazing temperatures were 1050, 1100, and 1150 °C (1920, 2010, and 2100 °F). Brazing times were 10, 20, 30, and 60 min for each type of foil. • The maximum shear strength observed on this type of specimen was approximately 310 MPa (45 ksi) for each of the four filler metals tested. The elemental distributions of nickel, chromium, and boron were concentrated in the eutectic structure. It appears that the migration of boron influenced the reduction of the eutectic structure. • With the BNi-3 filler metal, the eutectic structure was formed in the first minutes of brazing. As time at the brazing temperature was increased, boron diffused into the base metal, resulting in the elimination of the eutectic structure. Chromium and/or iron diffuses from the base metal into the brazed joint, thus assisting in increasing the joint strength with added time at brazing temperature. In a similar study (Ref 223), the same researchers evaluated Inconel 600 and SUS 304 stainless steel with the same filler-metal foils. The brazed joint was obtained for all of the brazing conditions in this study. The shear strength of the specimen increased with increasing brazing time, except at 1050 °C (1920 °F).
At 1050 °C, the shear strength of the specimen was not influenced by brazing time. In this case, the break of the specimen during the shear test occurred in the brazed layer. At 1250 °C (2280 °F), the value of 450 MPa (65 ksi) was obtained as the maximum shear strength in this study; the break of the specimen occurred in the base metal. The shear strength of the specimen increased with increasing brazing temperature and with increasing chromium content in the filler-metal foil to 7 mass%. Researchers (Ref 224) developed a new method of brazing aluminum to austenitic stainless steel, involving squeeze casting, that has been named cast brazing. Without melt flow control, strengths exceeding 40 MPa (5.8 ksi) were achieved for the aluminum/stainless steel joint. The presence of a porous sintered layer on the bond face of the stainless steel increased the strength of the joint to ~70 MPa (10 ksi). By controlling the flow of aluminum filler metal during casting, the formation of an intermetallic reaction layer at the interface was effectively suppressed, and the strength of the brazed interface was increased to a value exceeding the strength of the aluminum. On brazing aluminum to aluminum with a preliminary nickel coating on the surfaces to be joined, interfacial strengths exceeding 50 MPa (7.3 ksi) were attained. Before fracture, this type of joint exhibited an elongation of up to 10%. Researchers (Ref 225) furnace brazed X2CrNi18/9 stainless steel/aluminum joints at 600 °C (1110 °F) using an aluminum-silicon eutectic filler metal and found that the interfacial zone between the aluminum-rich braze joint and the stainless steel substrate features two intermetallic layers. The first is formed in the first instants of the process and has an overall composition roughly similar to that of the compound FeSiAl5. The second appears after a 10 min hold at the brazing temperature and features an overall composition that parallels the FeAl3 intermetallic. Both layers are more complex in structure than is suggested by these stochiometric relations, featuring several phases and microstructural gradients. Both layers grow after the first 10 min of the brazing cycle, in accordance with the parabolic rate law. By comparison with published data gathered on 1100 aluminum/304 stainless steel solid-state diffusion couples, it would appear that formation of the second (iron-aluminum) layer is delayed by
Chapter 4: Base Metals and Base-Metal Family Groups / 147
the presence of silicon in the braze, and that the rate of layer growth is controlled by diffusion through this second (iron-aluminum) layer once it has formed. The shear strength of the bonds peaks at 21 MPa (3 ksi) after a 10 min hold time at the brazing temperature. This peak is associated with growth of the second intermetallic layer into a continuous layer, which is thus shown to make the joint significantly fragile. A practical implication of this work is that strong joints between stainless steel and aluminum via furnace brazing can be produced using a eutectic aluminum-silicon filler metal, provided brazing times are kept sufficiently short to avoid formation of the second, more fragile (iron-aluminum) intermetallic layer. At a brazing temperature of 600 °C (1110 °F), brazing times should remain less than approximately 10 min. Care and process control are thus required for success in the process; however, it should be noted that 10 min represent a sufficient time window for viable industrial application of the furnace brazing process to join aluminum with stainless steel. Several titanium and/or titanium alloys brazed or diffusion bonded to various stainless steels in studies (Ref 226, 227) and development programs have proven successful. In one study (Ref 226), the following conclusions were made: • Brazed joints are complex in nature and require detailed characterization to determine the distribution of alloying elements along the brazed joint and the intermetallics formation. • Titanium/304 stainless steel brazed joints using filler metal No. 1 (Ag-28wt%Cu) showed invariably better shear strength than the ones made using filler metal No. 2 (Ag46wt%Cu). • To maximize shear strength of titanium/304 stainless steel brazed joints, the brazing time must be limited to a maximum of 15 min. • Increase in copper concentration resulted in an increase in titanium content in the titanium/304 stainless steel filler metal and enhanced the formation of intermetallic compounds between titanium and 304 stainless steel. • Intermetallic compounds are not limited to the filler metal. They can also form in the base metal next to the interface. As such, the
characterization of these compounds should extend into the immediate base metal next to the filler metal. • Not all intermetallics are necessarily hard and detrimental to the properties of the joint. For example, Ti2Cu was present in the filler metal but did not deteriorate the mechanical properties of the brazed joints. • To design an optimal filler metal, the extent of diffusion of alloying elements in the base metal and in the filler metal, the kinetics of intermetallic compounds formation, and the mechanical and physical properties of the intermetallic compounds must be considered. Researchers (Ref 227, 228) have conducted a variety of investigations joining Inconel 625, AISI 316L, and AISI 4130 low-alloy steel to commercially pure titanium and titanium 6242 alloy (Ti-6Al-2Mo-4Cr-2Mn) by diffusion bonding. Encouraging results were obtained by joining Nb-1Zn/Inconel 718 where joints that were pressure brazed using nickel-base metallicglass foil had shown the most promise (Ref 229). Direct-diffusion-welded joints and joints diffusion welded with nickel and iron interlayers were judged to be unsuitable because of cracking and porosity problems. More brazing development studies will be required when the need is defined for a Nb1Zr/Inconel 718 joint of a specific size and geometry. Items to be addressed would include the composition of the metallic-glass filler metal, the effect of pressure on braze joint composition changes resulting from filler-metal expulsion, the effect of pressure on joint soundness and diffusion kinetics, joint design, postbrazing heat treatment, and mechanical property determination. Graphite-, Composites-, Diamond-, Aluminides-to-Metals. Direct brazing of graphite to metal has been successfully carried out by a one-step brazing process using a copper-base active filler metal. Researchers (Ref 230) carried out double-brazed shear tests to assess the shear strengths of the brazed joints. High shear strengths were obtained in all cases examined. The shear strength of the graphite-to-copper joint was high (29 MPa, or 4.2 ksi). This high shear strength can be attributed to the fact that copper plastically deformed to accommodate the residual stress caused by a large thermal expansion mismatch between the graphite and
148 / Brazing, Second Edition
copper. In the case of graphite-to-molybdenum, where the thermal expansion mismatch is small, the shear stress was 8 MPa (1.2 ksi), which is lower than that for the graphite-to-copper joint. The accommodation by plastic deformation in molybdenum is limited, due to its high yield stress, resulting in high residual stress (Table 4.26). Heat radiators of a proposed type feature thermally conductive fibers protruding from metallic surfaces to provide increased heat-dissipation surface areas. Originally intended for use in the vacuum of outer space, these radiators may prove useful on earth in special industrial and scientific applications that involve dissipation of heat in vacuum or in relatively still air. Because these radiators do not rely on internally contained liquid or gaseous heat-exchange media, they are free of leaks and more reliable than are radiators that incorporate heat pipes. They are also lightweight and relatively inexpensive. Figure 4.20 illustrates a radiator of the proposed type, consisting of a spherical metal shell covered with a sparse, brushlike array of graphite fibers extending perpendicularly from the surface. The shell surrounds the source of heat. The inner ends of the fibers are attached to the shell by solder joints. The fibers increase the effective radiative surface area to approximately 100 times that of a flat-plate radiator of comparable size. Each fiber is paced approximately 10 fiber diameters from its nearest neighbors, so that shadowing of fibers by other fibers is small, and there is ample solid angle for each fiber to radiate directly to space. Each fiber is metallized at one end so that it may be soldered or brazed to the shell. The first layer to be deposited in the metallization process is one of silicon, which bonds to the graphite fiber by forming SiC on its surface. Next, a layer of copper containing a small amount of titanium or zirconium is deposited to provide a solderable or brazeable surface.
Researchers (Ref 231, 232) conducted an investigation into the transient liquid phase diffusion bonding (TLPDB) of continuous SiCfiber-reinforced Ti-6Al-4V composite to Ti6Al-4V alloy. They concluded: • A Ti-Cu-Zr amorphous filler metal was used in the bonding of Ti-6Al-4V/SiC composite to Ti-6Al-4V alloy plate via the TLPDB process. A joint strength of 850 MPa (123 ksi) was obtained, and this value corresponded to 90% of the tensile strength of Ti6Al-4V. • Isothermal solidification was almost completed after 1.8 ks at the bonding temperature of 1153 K. The bonding layer had a typical acicular microstructure composed of Ti2Cu and α titanium with dissolved zirconium. • Liquid phase diffusion bonding can reduce bonding pressure in comparison with solidstate diffusion bonding. The deformation ratio of Ti-6Al-4V alloy necessary to obtain sufficient joint strength of 850 MPa (123 ksi) for TLPDB was ~2%; contrarily, the deformation ratio necessary for solid-state diffusion bonding was greater than 5%. • Brittle products such as (Ti,Zr)5Si3 or (Ti,Zr)5Si4 were formed at the interface between fibers and the filler metal at 1153 K for a bonding time of 3.6 ks. These products, however, did not affect joint strength, because they were formed only at the end of fibers in very small amounts. Researchers (Ref 233) performed brazing experiments at 750 °C (1380 °F) for 2 h between Ag-Cu-In-Ti filler metal and SiCw/Al2O3. They found that the first clearly nonbraze layer consisted of an oxide layer of metallic composition 33Ti-31Al-22Cu-14Si. Areas adjacent to the SiC whiskers were of different composition. A thin, continuous layer on the Al2O3 portion of the composite appeared to be γ-TiO. The SiC
Table 4.26 Shear strengths of graphite brazed joints Material
Graphite-to-graphite Graphite-to-molybdenum Graphite-to-copper Copper-to-copper Molybdenum-to-molybdenum
Filler-metal thickness, µm
Number of tests
Average shear strength, MPa (ksi)
Standard deviation, MPa (ksi)
Comment
50 50 100 50 100 50 50
4 4 4 4 2 1 1
40.0 (5.8) 6.1 (0.9) 8.0 (1.2) 28.6 (4.15) 28.6 (4.15) >40 (5.8) >138 (20)
1.8 (0.26) 1.4 (0.20) 1.4 (0.20) 1.5 (0.22) 1.0 (0.15) ... ...
... ... ... ... ... Specimen bent Testing machine capacity limit
Chapter 4: Base Metals and Base-Metal Family Groups / 149
whiskers were preferentially consumed and underwent reductions in diameter of approximately 40%. Observed were knobby whisker morphologies that may be related to SiC stacking faults. The η-type phases detected near the silver-copper eutectic portion of the joint appeared to consist of Ti-Cu-Al-Si-O and Ti3Cu3O. If accurate, this suggests that joint strength may be partially controlled by the quantity and orientation of these faults. Researchers (Ref 234) fabricated a surface set of diamond tools by an active metal brazing process, using bronze (Cu-8.9Sn) powder and AISI 316L stainless steel powder mixed to various ratios as the filler metals. The diamond grits were brazed onto a steel substrate at 1050 °C (1920 °F) for 30 min in a dry hydrogen atmosphere. After brazing practice, an intermediate layer rich in chromium formed between the filler metal and diamond. A filler metal composed of 70 wt% bronze powder and 30 wt% stainless steel powder was found to be optimal in that the diamond grits were strongly impregnated in the filler metal by both mechanical and chemical types of holding. The diamond tools thus fabricated performed better than conventional nickel-plated diamond tools. In service, the filler metal wore at almost the same rate as the diamond grits, and no pullout of diamond
Radiated heat
Spherical shell
Source of heat Core
Graphite fibers
Fig. 4.20
Advanced radiator featuring radial graphite fibers to carry heat away from a spherical shell and radiate the heat into space
grits or peeling of the filler-metal layer took place. Researchers (Ref 235) evaluated the diffusion brazing of NiAl, which is a promising candidate material for high-temperature applications. However, NiAl suffers from poor low-temperature ductility and toughness. Thus, in many potential applications, NiAl has to be used locally, in combination with nickel-base alloys, rather than to form entire components. Hence, suitable technologies are required for NiAl-to-nickel-base-alloy joining. In view of the poor low-temperature ductility and strong Al2O3-forming tendency of NiAl, diffusion brazing seems to be the most suitable technology for joining NiAl to itself and to nickel-base alloys. The authors examined the diffusion brazing of NiAl to nickel using Ni-Si-B interlayers and determined: • The progression of the diffusion brazing at the NiAl substrate side of the joints was not incompatible with that predicted by standard diffusion brazing models. No qualitative changes were found with respect to the two brazing temperatures studied (1065 °C, or 1950 °F, and 1150 °C, or 2100 °F). • Marked deviations from conventional models of the diffusion brazing process were observed at the nickel side of the joints. A prominent layer of borides was formed in the substrate adjacent to the original solid-liquid interface when the brazing temperature was below the nickel-boron binary eutectic temperature. When the brazing temperature was taken above the nickel-boron binary eutectic temperature, substantial liquation of the nickel substrate took place. This type of boride precipitation was consistent with previous observations of substrate boride formation at the brazing temperature after complete melting of the interlayer in nickel/Ni-SiB/nickel joints. Another series of investigations (Ref 236) examined transient liquid phase (TLP) bonding for NiAl, which offers the advantages of tolerance of the strong Al2O3-forming tendency of NiAl and compatibility with the poor low-temperature ductility of NiAl. Techniques have been developed for the TLP bonding of NiAl-NiAl (Ref 237, 238). How-
150 / Brazing, Second Edition
ever, relatively high bonding temperatures are required for these procedures, ranging from approximately 1200 °C (2190 °F) (Ref 237) to above 1460 °C (2665 °F) (Ref 238). These high bonding temperatures impede application to the joining of NiAl-nickel-base superalloys (which typically have melting temperatures of approximately 1300 °C, or 2370 °F). Thus, alternative interlayer materials are required if TLP bonding is to be applied successfully to the joining of NiAl to nickel-base superalloys. Ceramic-to-Metal Joining (Ref 239–243). Some excellent work was performed in the active alloy sealing of an Al2O3 ceramic to a copper stud via a titanium-nickel eutectic metallization for use as a high-reliability transistor package. The titanium-nickel metallization was chosen in lieu of the more widely used molybdenum-manganese metallization because of batchto-batch variations with the latter technique. The process used 0.5 µm of titanium evaporated onto the ceramic, followed by 7 µm of nickel. This was followed by heating to 1000 °C (1830 °F) for 2 min in dry H2. This heating step bonded the titanium to the ceramic by forming a liquid phase of titanium-nickel eutectic between the bulk of the nickel layer and the ceramic. The bulk of the nickel layer remained solid, because the proportion of nickel was far higher than required for the 71.5Ti-28.5Ni eutectic composition with its melting point of 955 °C (1750 °F). Close control of this postevaporation heating was found necessary. With too short a time or temperatures below 955 °C (1750 °F), the reaction to form the eutectic was complete; a weak joint between the metallization and the ceramic resulted. An excessively long heating time resulted in titanium diffusing through to the nickel surface, reacting with impurities in the H2 and rendering the surface difficult to braze. After this postevaporation heating, the joining was completed by brazing with a 72Ag-28Cu filler metal at >780 °C (1435 °F). Other studies of brazing (Ref 244) include an investigation into the use of an amorphous filler metal between Al2O3 and Fe-Ni-Co alloy and the effect of brazing conditions on the resulting shear strength. This complements other work (Ref 245) in the use of amorphous filler metals between two ceramic components. These results showed that, when using a copper-titanium filler metal, the joint strength was dependent on the amount of joining between Al2O3 and the intermediary TiO2.
Another unique metallizing process has been developed and applied at temperatures as low as 850 °C (1560 °F). This makes it especially suitable for materials (such as magnesium/partially stabilized zirconia, or PSZ) that cannot withstand temperatures above 1000 °C (1830 °F). The filler metal consists of 95% Sn, with the balance containing carbide or carbonyl formers and other alloying elements. The process consists of the formation of a chemical bond with the surface of the ceramic that continues to wet the surface even when it becomes molten. While the basic filler metal melts at relatively low temperatures, it is capable of alloying with brazing metals during joining to form ternary alloys with high melting points. Coatings and Metallizing. Of course, significant improvements in the techniques of metallizing have been made, and several new procedures have been developed and evaluated. Also, extensive research on the reactions that occur when a ceramic surface is metallized has contributed to the effectiveness of metallizing. Metallizing procedures were originally developed to improve the wettability of ceramic surfaces by conventional low-temperature filler metals. Later, investigators found that some active metals and their alloys or compounds (e.g., titanium and zirconium) wet unmetallized ceramic surfaces under certain conditions. Although variations of the so-called active metal process have been used commercially to produce ceramic-to-metal seals, they have not been accepted to the extent characterized by the metallizing-brazing concept of joining these materials. To ensure the production of reliable ceramicto-metal seals, most metallized surfaces are coated with nickel, copper, or other metals. The metals are usually deposited by electroplating; in some cases, however, the coatings are produced by reducing oxides of the desired metal. These coatings perform several functions, depending on the method used to produce the ceramic-to-metal seal. If the joints are to be brazed with conventional silver- or copper-base filler metals, the coatings serve the following purposes: • A metallizing layer is composed of metals and residual oxides not completely reduced during sintering. Such a surface is not conducive to good wetting by the filler metal. Plating with nickel or copper eliminates the
Chapter 4: Base Metals and Base-Metal Family Groups / 151
adverse effects of the surface on the wetting and flow characteristics of the filler metal. • When the metals used for metallizing are not wet readily by low-temperature filler metals, plating provides the surface with a metal easily wet by such filler metals. • To a degree, the plated metal acts as a barrier to the penetration of the metallizing layer by the filler metal. Some filler metals react with the metals used for metallizing. If the reaction is allowed to proceed too long, the filler metal may penetrate the metallized coating and lift it away from the ceramic. Metallized coatings are usually plated with nickel to retard penetration and with copper to provide good wetting. There are numerous methods to coat the metal-ceramic surface: • Sintered metal powder (Ref 246–254) • Reactive or refractory metal salt (Ref 254–256) • Metal-glass powder (Ref 257–261) • Vapor deposition (Ref 250) However, with the previously mentioned techniques and the early investigations, joining ceramics to metals by the active metal or active hydride process has advanced significantly. The strengths of joints made by this process are as great as those obtained with joints made by the moly-manganese process. Some difficulty has been experienced in making seals by the active metal or active hydride process in dry H2. The dew-point of H2 must be extremely low to prevent oxidation of titanium. Producing ceramicto-metal seals in a vacuum is advantageous in that the parts are outgassed during brazing. The concept of fabricating ceramic-to-metal joints and seals by the active metal or active hydride process was first applied in the electronics industry. In recent years, however, these joining processes have found other uses to meet the need of high-temperature vacuumtight seals in the nuclear and aerospace industries (Ref 262–264). Other New Ceramic-Metal and FillerMetal Combinations. New active filler metals are constantly undergoing changes and modifications in composition to meet the everdemanding requirements to permit metals (Ref 265) to be joined to ceramics without the ceramic materials being metallized. Some of
these silver-base filler metals (Cusil and Incusil) are ductile and adaptable to brazing metals to such materials as Si3N4, PSZ, transformationtoughened Al2O3, and SiC as well as many other refractory materials (Tables 4.27, 4.28). These brazing processes are shown in Table 4.29. Researchers (Ref 266) recently concluded in an investigation that the hermetic seal of Kovar/iron/aluminum/Al2O3 joints by the aluminum-silicon interlayer method depends not only on the composition and properties of the compound layer but also on the thermal stress induced by thermal expansion mismatch, especially on the circumferential stress parallel to the joint interface and on the radial stress. In trying to determine the influence of copper, Kovar, molybdenum, and tungsten interlayers on the magnitude and distribution of thermal stresses and on the tensile strength of brazed Si3N4-steel joints, researchers (Ref 267) found that joints made using low-yield-strength/highductility interlayers, such as copper, have lower thermal stresses and higher strengths than those made using low-thermal-expansion/high-yieldstrength interlayers, such as molybdenum or tungsten. A composite interlayer comprising copper and tungsten produces the lowest thermal stresses during brazing. Increasing the thickness of the interlayer decreases the thermal stresses produced during brazing, because the rigid restraint effect due to the high-yieldstrength/high-elastic-modulus steel substrate is reduced (Ref 267–271). Other studies (Ref 272) found that the silvercopper eutectic with 5% Ti filler metal for ceramic-metal joints has inherently poor oxidation resistance. It is necessary to improve this filler metal by adding other elements. Current work has demonstrated that aluminum is a very effective element in enhancing the oxidation resistance of Ag-Cu-Ti-base filler metals by forming a layer of protective copper-Al2O3 film. Other efforts (Ref 273) found that the oxidation behavior of titanium-containing filler metals on PSZ and Al2O3 can be improved. The filler metals included Cu-41.1Ag-3.6Sn-7.2Ti, Ag44.4Cu-8.4Sn-0.9Ti, Ag-41.6Cu-9.7Sn-5.0Ti, and Ag-37.4Cu-10.8In-1.4Ti. In investigations and evaluation of palladium-base filler metals, selected because of their oxidation resistance, ductility, and relatively high melting points in brazing ceramics to metals for heat engine applications, researchers (Ref 274) studied the brazed joints between
152 / Brazing, Second Edition
Si3N4 and nickel. They found that the joints brazed with the low-palladium filler metals, 70Au-8Pd-22Ni, 93Au-5Pd-2Ni, and 82Au18Ni, had shear strengths of 75 to 105 MPa (11 to 15 ksi) from 20 to 500 °C (70 to 930 °F), while the joints brazed with the high-palladium filler metals, 60Pd-40Ni, 30Au-34Pd-36Ni, and 50Au-25Pd-25Ni, all had shear strengths near 0. Using copper filler metals Cu-5Cr, Cu-1Nb, Cu-3V, Cu-5Ti, and Cu-10Zr, researchers (Ref 275) studied the reactions of Si3N4 and refractory metals (tungsten, molybdenum, niobium, tantalum). Combinations of brazed SiC (Ref 276) and others (Ref 277, 278) and various metals for automotive and aerospace applications include two combinations of interest to, for example, the automotive industry, namely, partially stabilized tetragonal zirconia (PSTZ) to a spheroidal graphite (SG) cast iron, and reaction-bonded SiC (RB-SiC) to a 0.4% C steel: • Vacuum-brazed joints between PSTZ and SG cast iron were made with shear strengths of approximately 200 MPa (29 ksi). • As an alternative to a complex alloy filler metal, a simple silver-copper eutectic together with a sputtered titanium coating on the PSTZ
•
•
•
•
produced joints with shear strengths up to 135 MPa (20 ksi). Many low-stress failures resulted from excessive oxidation of titanium arising from too high an oxygen partial pressure in the vacuum furnace. The work emphasized the importance of furnace atmosphere when brazes containing titanium were used. Diffusion bonding PSTZ to SG iron at 850 °C (1560 °F) with titanium and copper foil interlayers has produced an apparently crack-free joint, but others made at different temperatures or with titanium and silver foils contained internal cracks. Vacuum-brazed joints made with silver-copper + 5% Ti between RB-SiC and a 0.4% C steel failed by lack of wetting of the carbide surface, which was attributed to an undesirable reaction between titanium and free silicon in the RB-SiC. Surface treatments can improve the processes.
As a result, it was found (Ref 279) that filler metals containing titanium can make strong joints between PSTZ and SG iron, provided that excessive oxidation of titanium is prevented. Using titanium demands standards of vacuum
Table 4.27 Filler metals for brazing ceramics Name
Copper, BCu-1(a) Nicoro, BAu-3(a) Cu-Au(1) Cu-Au(2), BAu-1(a) Cu-Au(3) Cocuman Cu-Au(4) Silver Ticuni Nicuman 23 Nicoro 80 Nicuman 37 Palcusil 15 Silcoro 75 Gapasil 9 Palcusil 10 Ticusil Silicoro 60 Palcusil 5 Nicusiltin 6(b) Nicusil 3 Cusil, BAg-8(a) Incusil 10 Incusil 15 Georo Au-Sn
Composition, %
100 Cu 62Cu, 35Au, 3Ni 65 Cu, 35 Au 62.5 Cu, 37.5 Au 60 Cu, 40 Au 58.5 Cu, 31.5 Mn, 10 Co 50 Cu, 50 Au 100 Ag 70Ti, 15 Cu, 15 Ni 67.5 Cu, 23.5 Mn, 9 Ni 81.5 Au, 16.5 Cu, 2 Ni 52.5 Cu, 38 Mn, 9.5 Ni 65 Ag, 20.3 Cu, 14.7 Pd 75 Au, 20 Cu, 5 Ag 82 Ag, 9 Ga, 9 Pd 58.5 Ag, 31.8 Cu, 9.7 Pd 68.8 Ag, 26.7 Cu, 4.5 Ti 60 Au, 20 Cu, 20 Ag 68.5 Ag, 26.8 Cu, 4.7 Pd 62.5 Ag, 29 Cu, 2.5 Ni, 6 Sn 71.15 Ag, 28.1 Cu, 0.75 Ni 72 Ag, 28 Cu 63 Ag, 27 Cu, 10 ln 61.5 Ag, 24 Cu, 14.5 ln 88 Au, 12 Ge 80 Au, 20 Sn
Liquidus, °C (°F)
Solidus, °C (°F)
3598 (1981) 3427 (1886) 3362 (1850) 3346 (1841) 3330 (1832) 3326 (1830) 3232 (1778) 3200 (1760) 3200 (1760) 3184 (1751) 3087 (1697) 3087 (1697) 3006 (1652) 2989 (1643) 2941 (1616) 2851 (1566) 2844 (1562) 2827 (1553) 2714 (1490) 2689 (1476) 2665 (1463) 2617 (1436) 2455 (1346) 2374 (1301) 1243 (673) 997 (536)
3598 (1981) 3330 (1832) 3297 (1814) 3281 (1805) 3265 (1796) 2993 (1645) 3184 (1751) 3200 (1760) 3038 (1670) 3087 (1697) 3038 (1670) 2941 (1616) 2844 (1562) 2957 (1625) 2827 (1553) 2760 (1515) 2779 (1526) 2795 (1535) 2705 (1485) 2327 (1275) 2617 (1436) 2617 (1436) 2309 (1265) 2131 (1166) 1243 (673) 997 (536)
(a) American Welding Society specification. (b) Aerospace Material Specification (AMS) 4774A. Source: Ref 263, 264
Chapter 4: Base Metals and Base-Metal Family Groups / 153
Table 4.28 Selection guide to filler metals for brazing metal-ceramic joints Tool steel
Nickel, cobalt alloys(a)
Copper Copper-gold Palcusil Silcoro
...
...
Ticuni Ticusil
Ticuni Ticusil
Ticuni Ticusil
Carbon and low-alloy steel
Stainless steel
Copper(b)
Nickel(b)
Titanium, zirconium alloys
Refractory metals
Ceramic, metallized
...
...
Ticuni Ticusil
Carbon(c)
Tungsten carbine
Copper-gold Cusil Incusil Silver
...
...
Ticusil
...
Ticuni Ticusil
Ticusil
Cocuman Gapasil Nicuman Ticusil
...
Ticusil Cocuman Copper Nicuman Silver
Ceramic metallized Copper Copper-gold Cusil Nicusil Palcusil Silver
Copper-gold Copper Cusil Copper-gold Georo Cusil Gold-tin Nicoro Incusil Nicusil Nicoro Palcusil Nicusil Silcoro Palcusil Silver
Carbon (b) Ticusil
Ticuni Ticusil
Ticuni Ticusil
Tungsten carbide Cocuman Copper Nicuman Nicusiltin
Cocuman Copper Nicoro Nicuman Nicusiltin
Cocuman Cocuman Nicoro Nicoro Nicuman Nicuman Nicusiltin Nicusiltin Palcusil Palcusil
Copper-gold Cocuman Gapasil Nicuman Copper Ticuni Nicusiltin Copper-gold Ticusil Nicoro Nicuman Nicusiltin
(a) Corrosion- and heat-resistant alloys. (b) Includes alloys. (c) Graphite and diamond. Source: Ref 9, 264
Table 4.29 Summary of methods for metallizing ceramics No.
1.
Metallized layer type
Suitable ceramics
2.
Silver, silverplatinum Silver, silver-platinum + fluxing glass
Hard glasses, most ceramics Hard glasses, most ceramics
3.
Silver + copper oxide
Most ceramics
4.
Molybdenum + silver
Most ceramics
5.
Molybdenum + glass
Most ceramics
6.
Nickel + glass
Most ceramics
7.
Copper, copper alloys
Most ceramics
8.
Tungsten or molybdenum (+ manganese or iron) Titanium or zirconium(c) (active metal joints) Molybdenum + titanium
Oxide ceramics, BeO, Al2O3 (debased type) Oxide and some nonoxide ceramics Pure alumina
9.
10.
Metallizing materials in finely divided form
Mixtures of PtCl4, Ag2O, Ag, and Pt Mixtures of PtCl4, Ag2O, Ag, and Pt + <20% soft glass or flux, e.g., lead borate Ag2O + <10% CuO, Cu2O Mo + ~10% Ag2O
Metallizing temperature, °C (°F), atmosphere
500–900 (930–1650) in air 500–900 (930–1650) in air
>940 (1725) in air
Additional layers required
None Ag, Ag-Pt
Ag, At-Pt
1300 (2370) Ni(b) in dry H2 1200–1300 Mo + 10–20% glass Ni(b) (2190–2370) in dry H2 1300 (2370) NiO + ~10% glass Metallized layer in dry H2 needs buffing 1100 (2010) in air + CuO (+ alloying oxides) None 900 (1650) in H2 1450–1650 W or Mo (+Mn or Ni(b) (2640–3000) Fe compounds, in wet H2 e.g., 80% Mo, 20% Mn) Ti and Zr or, more >1000 (1830) in None commonly, TiH4, ZrH4, inert atm. or other compounds Mo + TiN or TiC 1450–1900 Ni(b) (2640–3450) in wet H2
Suitable solders or brazes
Sn, Pb, or Pb solders(a) Sn, Pb, or Pb solders(a)
Sn, Pb, or Pb solders(a) Cu, Cu-Ag, Ag Cu, Cu-Ag, Ag
Solders, Cu-Ag Solders Cu-Ag, Au-Cu
Zr, Ti eutectic brazes Cu-Ag, etc.
(a) Excessive reaction between solder and metallized layer can be prevented by an additional electroplated layer of copper on the metallizing or by using a solder with a high silver content, e.g., Cu-Ag. (b) Nickel coating can be achieved by electroplating or by a second coating, as in No. 6, but without glass. (c) This process can be done by direct brazing in vacuum with an active metal braze, using optionally TiH4 or ZrH4 as fluxes to wet the ceramic, i.e., a one-stage process.
154 / Brazing, Second Edition
practice well above those for conventional vacuum brazing. The use of sputtered coatings of titanium on oxide-ceramic surfaces prior to vacuum brazing could be commercially viable. Also, the improvement of the wettability of RB-SiC by modifying the surface or the filler metal should be studied further. Finally, the treatment of cast iron surfaces to avoid excessive diffusion of carbon into the joint and to give better bonding with the filler metal should be explored (Ref 279). Finally, a research program (Ref 277) on the mechanism of joining graphite to a high-Al2O3content ceramic (up to 95%) found that the TiAg-Cu filler metal was very successful in brazing the combination. For the existence of carbon, the Ti-Ag-Cu was used to produce AlAg3 and Ti3Al in the physical phase of the sealed region, which was an essential factor in the possible forming of a gastight seal between graphite and the 95% Al2O3 ceramic. For the existence of AlAg3 and TiAl3, the use of Ti-Ag-Cu could be a significant advance to future research of sealing between 95% Al2O3 ceramics and oxygen-free copper (Ref 280), Kovar, beryllium, beryllium oxide (Ref 281), tungsten, molybdenum, and stainless steel (Ref 282). A number of studies have investigated the active brazing of Al2O3 to superalloys (Ref 283), to steel with copper interlayers (Ref 284), and to titanium (Ref 285). In one study (Ref 283), the brazing of Al2O3 to nickel-base superalloy (Inconel 600) and cobalt-base superalloy (UMCo-50) was performed. The Ag-Cu-Ti active filler metals with titanium contents of 2 and 8 wt% were employed in this study. However, experiments showed that the Ag-Cu-2wt%Ti resulted in poor bonding, which was not suitable for the brazing investigation. With an increase of the titanium content to 8 wt%, the specimens could be bonded. However, the bonding strengths were still too low. The four-point bending strengths in this case were only 43 to 54 MPa (6.2 to 7.8 ksi). An improvement of the bending strength was achieved by inserting a Kovar interlayer. In this case, the bending strength was dramatically increased to 240 and 226 MPa (34.8 and 32.8 ksi) for the Al2O3/UMCo-50 joint and the Al2O3/Inconel 600 joint, respectively. Researchers (Ref 284) carried out experiments to show that multilayer bonding by using thin and very plastic material interlayers was a
good solution to the problem of making Al2O3steel joints with large dimensions. Using this type of interlayer allowed for the relaxation of stresses generated during the bonding cycle. It was concluded that a copper-material interlayer was suitable for this purpose. Direct bonding using active products of the reduction reaction of CuO (Cu2O, O2) can also be adequately used in the process of bonding joints with a copper interlayer. Using this method, they were able to fix the relationship between the quality of the bond (homogeneous microstructure at the whole bonding surface) and a protective atmosphere. They showed that their method can be used in joining Al2O3 to copper and copper to steel in one thermal cycle. Such bonded joints have high mechanical strength and a homogenous microstructure. Compared to other joining methods, this process does not need to implement any special joining conditions using the following features: low bonding temperatures of ~1340 K, a protective atmosphere of nitrogen containing below 40 ppm oxygen, and a bonding time of ~90 min. Particular attention must be paid to two effects when joining Si3N4 to metallic materials: interfacial bonding and residual stress. A wide variety of joining methods that depend on the fact that Si3N4 reacts with many metallic elements to form well-bonded interfaces have been developed for bonding to Si3N4 metals. Several categories can be involved, and each process has its own advantages. However, much remains to be learned. Thus, further experimental evidence could allow a more detailed understanding of the stress-corrosion effects produced not only by water but by other corrosive and oxidizing environments as well. Small components (up to 15 mm, or 0.6 in., in diameter) can be joined using soft metal or laminated layers. How to join a large-sized Si3N4 component to a metal, however, is still a major problem, due to the severe residual stresses. Because most Si3N4 components are used at elevated temperatures, evaluation of high-temperature properties, such as strength, oxidation, and thermal stress, is required. It can be concluded (Ref 286) that strong and reliable joints can be obtained between Si3N4 and austenitic stainless steel by diffusion bonding when using thin, metallic interlayers. Best results (average shear strength of 95 MPa, or 13.8 ksi, with a Weibull modulus of 6.8) were obtained when using an Invar (Carpenter Tech-
Chapter 4: Base Metals and Base-Metal Family Groups / 155
nology Corporation) interlayer at a bonding pressure between 7 and 20 MPa (1 and 3 ksi), a bonding temperature between 1000 and 1100 °C (1830 and 2010 °F), and a bonding time between 90 and 1440 min: • During the diffusion bonding process, a reaction layer was formed at the interface between the ceramic and the interlayer. This reaction layer consisted of a porous zone anchored into the Si3N4 and a diffusion zone extending into the interlayer. The formation of the reaction layer was due to the diffusion of free nitrogen and free silicon generated by decomposition of the Si3N4 in contact with the interlayer material. • The mechanical strength of the joint depended strongly on the thermal stress developed in the vicinity of the interface due to the difference in shrinkage between the ceramic and the metal and was directly related to the thickness of the reaction layer. A researcher has been developing joining technology for ceramics and, in addition to the development of materials, for the expansion of the use of ceramic applications in the fields led by the ceramic turbine. The aim of the research is the application of these joining techniques to the joining of a Si3N4 turbine with a metal shaft designed to resist high temperatures (Ref 287). The research was aimed at the improvement of joining strength and of thermal-cycle characteristics at high temperatures. It was found that the formation of a high-strength interface and the reduction of thermal stress were necessary for a successful joining between ceramics and metals. The realization of high-strength joining of Si3N4 with metal was extremely difficult in the past, owing to the poor wettability of this ceramic with metal, in addition to its low coefficient of thermal expansion (CTE) (approximately 3 × 10–8K–1), with a large difference from the CTE of metal materials. The formation of an interface was based on the ion-plating method shown in Fig. 4.21. The researcher has established a new method of combining brazing with the multilayer metallization, with titanium, an active metal, as the innermost layer. The outer layers serve in preventing the oxidation of titanium and improving the wettability of the filler metal, thereby achieving a durable interface. The reduction of thermal stress has become possible with the application of a WC-base alloy and other mate-
rials with low CTE. Joining strength surpassing 500 MPa (73 ksi) at 400 °C (750 °F) has been realized through these improvements. The joint is extremely tough, and the strength does not deteriorate, even under an unprecedented exacting thermal cycle test with 1000 cycles between room temperature and 400 °C (750 °F). Other studies and investigations combining Si3N4, steels (Ref 288), superalloys, and molybdenum evaluated the role of titanium in the development of the reaction layer in braze joining Si3N4 to stainless steel using titanium-active copper-silver filler metals. This reaction layer formed as a result of titanium diffusing to the filler-metal/Si3N4 interface and reacting with the Si3N4 to form the intermetallics titanium nitride (TiN) and titanium silicide (Ti5Si3). This reaction layer allows wetting of the ceramic substrate by the molten filler metal (Ref 289). Two filler metals were used in this study, Ticusil (Wesgo Metals) (68.8Ag-26.7Cu-4.5Ti, wt%) and CB4 (70.5Ag-26.5Cu-3.0Ti, wt%). The joints were processed in vacuum at temperatures of 840 to 900 °C (1545 to 1650 °F) at various times. Bonding strength was affected by reaction-layer thickness in the absence of titanium-copper intermetallics in the filler-metal matrix. The reaction-layer thickness increased with temperature and time. Its growth rate obeys the parabolic relationship. Activation energies of 220.1 and 210.9 kJ/mol (53 and 51 kJ/cal) were calculated for growth of the reaction layer for the two filler metals used. These values were close to the activation energy of nitrogen in TiN (217.6 kJ/mol, or 52.3 kJ/cal). Researchers (Ref 290) examined the effect of pressure in brazing Si3N4 to AISI 5140 steel. Pressures (0 to 40 MPa, or 0 to 5.8 ksi) were applied to the joints of Si3N4 ceramic to 5140
Fig. 4.21
Tapered Si3N4 joined to stainless steel shaft.
156 / Brazing, Second Edition
steel during vacuum brazing with Ag-Cu-Ti active filler metal. Pressurization started at various temperatures (873, 973, and 1073 K) and ended at room temperature during cooling. Results showed that there is an optimal starting temperature to pressurize, at which the maximum room-temperature shear strength of the joint is obtained. Researchers (Ref 291) successfully fabricated joints of Si3N4 and molybdenum with a vanadium interlayer using a vacuum hot pressing facility. The optimal joining conditions for producing a joint with the highest shear strength were found to be as follows: a temperature of 1328 K, a mechanical pressure of 20 MPa (3 ksi), and a bonding time of 5.4 ks. The strength level was initially 118 MPa (17 ksi) at room temperature, and this level gradually decreased as the test temperature rose. At 973 K, the strength level was still 70 MPa (10 ksi). Observations of the interface by scanning electron microscopy and electron probe x-ray microanalysis revealed that a layer of reaction product V3Si formed at the Si3N4-vanadium interface. Researchers (Ref 292) carried out detailed observations on the metallurgical behavior of joint brazing of nickel-base alloy Inconel 600 to Si3N4 with Ag71Cu27Ti2 filler metal, with emphasis on the interface between the filler metal and the Inconel 600 and the effects of nickel, which was the predominant element in the base metal. Based on the experimental results, the mechanism of bonding Inconel 600 to the filler metal is attributed to the diffusion of silver and copper along the grain boundaries of the Inconel 600, which results in mechanical anchoring. The effects of nickel on the metallurgical behavior of filler metal are summed up as enhancing the separation of silver- and copper-rich liquid phases from the molten filler metal, combining titanium and decreasing its activity in the reaction with Si3N4 at the interface with ceramics, promoting the diffusion of silver and copper into Inconel 600, and facilitating the flow of filler metal over the surface of Inconel 600. Material systems designed for 650 and 950 °C (1200 and 1740 °F) applications were evaluated in terms of torsion, torsion fatigue, and thermal fatigue (Ref 293). Researchers selected Si3N4/nickel/Incoloy 909 as the 650 °C (1200 °F) system, while Si3N4/molybdenum/Inconel 718 was selected as the 950 °C (1740 °F) system. The Au-5Pd-2Ni filler metal was used in
both systems. A cylindrical lap geometry with an interlayer was selected for these joints. Room-temperature and 500 °C (930 °F) torsion strengths of the 650 °C (1200 °F) system were measured in the range of 30 to 100 N · m (22 to 74 lbf · ft) with a 2 cm2 (0.3 in.2) brazed area, while the strength at 650 °C was significantly lower (1.6 to 7.0 N · m, or 1.2 to 5.2 lbf · ft). This was attributed to a reduction in the shrink fit at 650 °C. The Si3N4/nickel/Incoloy 909 joints showed excellent room-temperature fatigue behavior. A similar trend was seen in the high-temperature strength of the Si3N4/molybdenum/Inconel 718 joints, which had lower strength than the Si3N4/nickel/Incoloy 900 joints due to the high CTE of Inconel 718. Transient liquid phase bonding is a method for joining high-temperature metal alloys at relatively low temperatures. The method uses a bonding interlayer of a composition similar to that of the articles to be bonded, which contains a small amount of a melting-point-lowering material. To form the bond, the metal/interlayer/metal assembly is heated to, and held at, a temperature above the melting point of the interlayer to obtain the TLP. During this process, the melting-point-lowering material diffuses into the surrounding materials, raising the melting point of the interlayer and leaving a solid bond. The TLP bonding principle has been extended to the joining of ceramics to metals and ceramics to ceramics. To form these bonds, multilayer interlayer structures are employed between the articles to be bonded. In the simplest case, the multilayer structure is a refractory metal sandwiched between two layers of a metal with a melting point lower than that of the refractory metal. Such a structure may be gold/niobium/gold. Heating of this structure results in gold diffusion into the refractory metal layer and the formation of the required solid bond. A more complicated system consists of two ceramic layers and, for instance, Si3N4 bonded together using an assembly that comprises layers of a refractory metal, gold, titanium, and gold. Here, an additional function of gold is to protect the other metals against oxidation. Researchers evaluated TLP-insert metal brazing of Al2O3 and AISI 304 stainless steel (Ref 294). This joining technique allows the continuous replenishment of the active solute that is consumed by the chemical reaction that
Chapter 4: Base Metals and Base-Metal Family Groups / 157
occurs at the ceramic/filler-metal interface. Replenishment is facilitated by employing a sandwich of filler metals comprising tin-base filler metal and amorphous Cu50Ti50 or NiCrB interlayers. During Al2O3/AISI 304 stainless steel bonding, the highest shear-strength properties were produced using a bonding temperature of 500 °C (930 °F). Thick reaction layers containing defects formed at the ceramic/fillermetal interface when higher bonding temperatures were applied. Bonding at temperatures above 500 °C (930 °F) also increased the tensile residual stress generated at the periphery of Al2O3/AISI 304 stainless steel joints. The shear strength of joints produced using NiCrB interlayers markedly increased following heat treatment at 200 °C (390 °F) for 1.5 h. Heat treatment had little influence on the shear strength of the joint produced using Cu50Ti50 interlayers. Researchers (Ref 295) also investigated TLP bonding of metal-matrix composite (MMC) and Al2O3 ceramic substrates and reached the following conclusions: • The completion time is much longer and the rate of movement of the solid-liquid interface is much slower in dissimilar MMC-Al2O3 joints, because copper can only diffuse into the aluminum-base composite. For this rea-
son, the particle-segregation tendency is markedly increased when these dissimilar substrates are TLP bonded. • During TLP bonding at 853 K, the highest joint shear-strength properties are produced using 5 µm thick copper foil. When the width of the particle-segregated layer at the joint interface exceeds 10 µm, this region acts as a site for preferential failure during mechanical testing. However, when a thin (<10 µm thick) particle-segregated layer is formed at the bond line joint, failure occurs in Al2O3 material immediately adjacent to the bond line. A joining pair of SiC and molybdenum was fabricated using vanadium foil as an insert material (Ref 296). The optimal fabricating conditions were a bonding temperature of 1198 K, a bonding pressure of 30 MPa (4 ksi), and a bonding time of 10.8 ks. The joint had high shear-strength levels of 150 MPa (22 ksi) at room temperature and 52 MPa (7.5 ksi) at 973 K. At the interface of SiC and vanadium, a layer of vanadium silicide, V3Si, was formed. Table 4.30 summarizes the use of interlayers in joining a variety of ceramics and metals. An investigation was conducted (Ref 298) to gain information about the capabilities of some filler metals for direct brazing of cubic boron
Table 4.30 Summary of interlayers for reducing thermal expansion mismatch in ceramic-metal joints Group(a)
I
II
III
IV
Ceramic/metal
Interlayer
Joining conditions
Strength, MPa (ksi)
Al2O3/type 321 steel TZP/type 316 steel MgO/steel Al2O3/steel SiC/Al(c) Al2O3/W (d) Al2O3/Fe (d) TiN/Mo(d) Al2O3/Fe (d) TZP/W(d) Al2O3/type 405 steel Al2O3/type 316 Si3N4/steel SiC/steel Sialon/steel Sialon/steel Si3N4/type 405 steel SiC/super alloy Si2N3/super alloy SiC/type 316 steel Si3N4/type 405
Al Cu Cu/metal foam BA03 Cu-C fiber Al2O3-W Al2O3-Fe TiN/Mo FeO-Fe TZP-W Nb/Mo Ti/Mo ... BA03/WC ... Type 304/WC steel Fe/W Ni/Kovar/Cu Ni/Kovar/Cu Ti/Mo Al/Invar (cracking in intermetallic compound)
873 K, 50 MPa (7.3 ksi), 30 min. 1273 K, 1 MPa (0.15 ksi), 4 h 1273–1473 K (b) 883 K, 10 MPa (1.5 ksi), 30 min 1043 K (Al was joined at 823 K) (b) 2125 K, 8 h 1473 K, 3 GPa (435 ksi), 30 min 1573 K, 3 GPa (435 ksi), 30 min 1473 K, 29 MPa (4.2 ksi), 1 h 1673 K, 1 h in H2 1673 K, 100 MPa (14.5 ksi), 30 min 1373 K, 9 MPa (1.3 ksi), 3 h ... 883 K, 2 MPa (0.3 ksi), 1 h ... 1373 K, 5 MPa (0.7 ksi), 1 h 1473 K, 10 MPa (1.5 ksi), 30 min 1323 K, 54 MPa (7.8 ksi), 2 h 1323 K, 54 MPa (7.8 ksi), 2 h 1083 K, 0 min(b) 1073 K, 0.15 MPa (0.015 ksi), 7 min(b)
70 (10.2) 52 (7.5) 33 (4.8) 30 (4.4) ... ... ... 80 (11.6) ... 200–400 (29–58) 500 (72.5) 70 (10.2) 200 (29) 150 (21.8) 300 (43.5) 150 (29) 60 (8.7) 100 (14.5) ... 50 (7.3) 60 (8.7)
(a) I, soft metal; II, composite; III, laminate (soft metal/low expansion and hard metal); IV, crack layer. (b) Brazing; the others are diffusion or eutectic joining. (c) Soft metal with carbon fiber. (d) Grading (cermet). TZP, tetragonal zirconia polycrystal. Source: Ref 297
158 / Brazing, Second Edition
nitride (CBN) grits to a steel substrate. It has been found that a nickel-chromium filler metal, known for effective diamond brazing, failed to show satisfactory wetting and bonding characteristics toward CBN under the same brazing conditions as that of diamond. The situation could not be improved either by increasing the weight percent of chromium or the brazing temperature. Also, prolongation of the brazing time did not show any significant change. It was further revealed that a Ag-Cu-Ti filler metal, which is well established for the brazing of diamond and ceramics, exhibited good wetting and bonding toward CBN at moderate temperature. Based on the thermodynamic properties of various materials, the characteristic properties of the elements, and the physical observations made during the investigation, it was suggested that transition elements such as titanium or zirconium are preferred to transition elements such as chromium as an activator to promote the wetting characteristics of the filler metal toward CBN, which is far more chemically stable than diamond. The ability of titanium to react with CBN was achieved, thus changing its surface chemistry and the wettability of the silver-copper system. The combination of ZrO2-nickel-Si3N4 (Ref 299) was successfully joined under the following conditions: temperature, 850 to 1300 °C (1560 to 2370 °F); time, 90 min; bonding pressure, 14 to 37 MPa (2.0 to 5.4 ksi); and interlayer thickness, 0.1 to 0.8 mm (0.004 to 0.03 in.). It appears that the shear strength of the ZrO2nickel-Si3N4 combination is maximal for bonding temperatures in the range of 1000 to 1100 °C (1830 to 2010 °F). It was found that two effects play a role in the bonding process. At low temperatures, diffusion-controlled interface formation dominates, yielding an exponential dependence of the shear strength on temperature, with an activation energy of approximately 110 kJ/mol (26 kJ/cal). At high temperatures, Si3N4 decomposes, forming free silicon and nitrogen at the Si3N4-nickel interface. The silicon diffuses into the nickel, whereas the nitrogen atoms recombine to form molecular nitrogen, which escapes to the surrounding and/or forms pores along the interface. Due to the solid solution of silicon in nickel, partial melting of the nickel can occur during the bonding process, which is detrimental for the bond strength. The bonding pressure and the interlayer thickness do not affect the bond strength. The characteristic strength of bonds made under optimal conditions is approx-
imately 57 MPa (8.3 ksi), whereas the Weibull modulus is approximately 1.6. Failure occurs almost always at the Si3N4-nickel interface, which therefore turns out to be the weakest link of the ZrO2-nickel-Si3N4 combination. The optimal bonding conditions obtained for this material combination were used to make a diffusion bond in a practical application. A Si3N4 ring was diffusion bonded to a cylindrical part of a gear pump consisting of ZrO2. It turns out that bonding was possible but that due to residual stresses in this large geometry, cracks occurred in the ZrO2. The situation could be improved by using interlayers with modified geometry. The same combination, except AISI 316 stainless steel (Ref 299), was substituted for Si3N4 and was successfully joined under the following conditions: temperature, 1000 to 1250 °C (1830 to 2280 °F); time, 90 to 360 min; bonding pressure, 2 to 20 MPa (0.3 to 3 ksi); and interlayer thickness, 0 to 1.2 mm (0 to 0.05 in.). It was found that direct bonding without an interlayer leads to unreliable bonds with low strength. Finite-element modeling calculations show that the level of residual stresses is rather high in the vicinity of the interface between ZrO2 and AISI 316. These stresses are considered to be the most important cause of the poor bond strength. A number of ZrO2/nickel/AISI 316 diffusion bonds were made with different interlayer thicknesses between 0.2 and 1.2 mm (0.008 and 0.05 in.). The bonds were found to have a relatively low strength with a large scatter (σ0 = 30.4 MPa, or 209.6 ksi), independent of the interlayer thickness. Calculations of the residual stresses in ZrO2/nickel/AISI 316 diffusion bonds showed that high residual stresses were present in the vicinity of the ZrO2-nickel interface. These stresses, in particular, the tensile stresses in the ZrO2, limit the strength of the bonds. A new brazing technique was developed (Ref 300) for joining graphite to itself or to metals such as molybdenum, tungsten, or copper with conventional filler metals. Essentially, it is impossible to braze graphite with copper filler metal (AWS BCu-1), because no wetting occurs. However, when a graphite base material is combined with an iron base metal in copper brazing, the iron base metal dissolves in molten copper. Simultaneously, the dissolved iron grows as part of a columnar Fe6–9 Cu-1.6C alloy phase at the graphite interface at a constant brazing temperature; that is,
Chapter 4: Base Metals and Base-Metal Family Groups / 159
the dissolution and deposit of base metal takes place. By placing an iron foil insert between both graphite base materials, therefore, the columnar phase is formed at both graphite faces and grows toward the iron foil during heating. As a result, both graphite base materials are united by the columnar phase through the iron foil. In the same way, a graphite-molybdenum or -tungsten joint can be produced. Moreover, when using BAu-1, which has a lower melting point than that of BCu-1, it is also possible to braze graphite to copper. The shear strength of a graphite-graphite joint with a 0.12 mm (0.005 in.) thick iron foil at room temperature was approximately 32 MPa (4.6 ksi). Further, the bending strength of the graphite-graphite and -copper joints at 600 °C (1110 °F), as measured using the four-point bending test, was 35 and 11 MPa (5.0 and 1.6 ksi), respectively. In addition, the technique can be applied to the brazing of AISI 304 stainless steel to highcarbon steel with BCu-1 where, normally, Cr23C6 and Cr7C3 layers are formed at the highcarbon steel/braze interface; these carbide layers result in the loss of mechanical properties of the joint. When inserting a 0.3 mm (0.01 in.) thick iron foil between AISI 304 and 1.04 carbon steel in order to prevent the formation of chromium carbides, the dissolution and deposit of base metal takes place on either side of the foil. On the 1.04 carbon steel side, the 1.04 carbon steel and the iron foil are united with a columnar Fe-7–9 Cu-0.8C alloy phase that comes from the 1.04 carbon steel, and, on the AISI 304 side, the AISI 304 and the iron foil are also united with a columnar Fe-12–13 Cr-6–9 Cu-4–5 Ni-0.2C alloy phase coming from the iron foil. Consequently, both base metals are united by each columnar phase through the foil, so the Charpy U-notch impact toughness of the joint increased markedly from 1.0 to 9.0 J (0.7 to 6.6 ft · lbf) at room temperature. Further, this uniting with each columnar phase exhibited a remarkable effect on the impact toughness of the joint in the range between room temperature and 700 °C (1290 °F). Researchers (Ref 301) reported on their work regarding the thermochemical aspects of multiphase diffusion during brazing of hard metal. During brazing of WC-cemented carbide parts (hard metal) on steel using pure copper as a filler metal, an intermetallic compound is formed at the hard-metal/braze interface, whereas pores
are generated in the border zone of the hard metal. As a result, the tensile strength of the joint is greatly reduced, with fracture occurring preferentially in the border area of the hard metal. The calculation of the molar Gibbs energies of the different phases and of some of the element chemical potentials, as well as the experimental determination of some potential gradients, were used to explain the different processes (phase formation and multiphase diffusion) taking place during brazing.
Case Histories and Problem-Solving Examples Example 1: Brazing Cast Irons to Dissimilar Metals. Joining dissimilar metals is not an easy problem to solve. Many firms have encountered problems in joining cast iron in a furnace exothermic atmosphere. The several kinds of cast irons include white, gray, malleable, and ductile. There are many applications in which it is desirable to braze gray, malleable, and ductile irons either to themselves or to dissimilar metals; however, white cast irons are seldom brazed. Brazing of ductile and malleable irons must not be performed above 760 °C (1400 °F), because the metallurgical structure may be damaged. In an exothermic atmosphere, type BAg-18 filler metal could be used. Gray cast iron presents the most problems in brazing. Wetting the surface is difficult, because there can be interference from graphite, sand, oxides, and so on. The primary method of cleaning is a proprietary molten salt bath employing a reversing direct current. Other methods of cleaning include searing with an oxidizing flame, grit blasting, and chemical cleaning. When heating in an oxidizing atmosphere, the silicon, if high, could oxidize, thus interfering with surface wetting of the filler metal. It has been suggested that there is a specially formulated copper filler metal that has improved the wetting of copper on gray cast iron. With the many formulations and mixtures of cast iron that are available, there appears to be no easy one-step method of copper furnace brazing gray cast iron. Example 2: Brazing Aluminum Bronze to Naval Brass. In another application, a firm wished to braze aluminum bronze to naval brass in pure dry hydrogen with a BAg-8 silver-copper eutectic filler metal. The results initially pro-
160 / Brazing, Second Edition
duced very poor braze joints, and the filler metal did not wet and flow well. The aluminum-bronze base metal had an aluminum content of 7 to 10%. Because aluminum oxidizes very readily, even in the driest of hydrogen, this was undoubtedly the major problem being encountered. Likewise, if the atmosphere is quite dry, some aluminum gases off from the surface and aluminizes the filler metal (either in wire or powder form). This also causes the filler metal not to melt properly and to leave more of a residue. One of the simplest methods for solving this problem is to use a grade of flux specifically formulated to handle aluminum-bronze base metals, that is, AWS brazing flux type No. 4. When running the part in hydrogen, only a very small amount of this flux is required, because the dry atmosphere protects it from oxidation, so it may be thinned down and used as a thinner liquid rather than a heavy paste. If it is desired not to have any flux in the joint area, then it is possible to plate the aluminum bronze with either nickel or copper. The thickness of the plating must be thick enough so that the aluminum does not diffuse up to the surface before the filler metal melts and flows. This, of course, is dependent on time and temperature; thus, the plating thickness can be varied, depending on these two variables. However, the higher the temperature and the longer the time to come up to heat, the thicker the plating that is required. It is suggested that 0.01 mm (0.0004 in.) thickness is a good starting place. If, on brazing, the surface of the plating discolors to a gray color and flow is not adequate, a thicker plating is required. In brazing naval brass in a hydrogen atmosphere, the vapor pressure of zinc is increased to the point that dezincification occurs. This is true whether the atmosphere is exothermic, pure nitrogen, pure hydrogen, nitrogen-hydrogen, or any of the gas atmospheres. Because the BAg-8 silver-copper eutectic filler metal is used, brazing is approximately 788 to 816 °C (1450 to 1500 °F). To reduce the amount of dezincification, it is better to use the BAg-18 filler metal, which has a lower melting point and thus lowers the amount of zinc vaporization that occurs. Brazing range on this filler metal starts at 718 °C (1324 °F), and it is suggested that 760 °C (1400 °F) is a suitable brazing temperature. If it is necessary to reduce the amount of dezincification further, the parts may be plated with cop-
per or nickel. Dezincification is noted as a change from the brass color to a white surface, or, depending on the furnace, by brown fumes coming out of the furnace. Some firms have been very successful in brazing cartridge brass, which has only 30% Zn, with the BAg-18 at 760 °C in a pure dry hydrogen atmosphere. Example 3: Brazing Stainless Steel to Titanium. How would one braze AISI type 304L stainless steel to titanium with nickel 200? Titanium picks up oxygen quite readily, thus the parts are normally brazed in a high vacuum of 0.13 Pa (2 × 10–5 psi) or better. In brazing titanium, it is essential that the furnace be leakfree and have a clean interior prior to brazing. It is good to run a test cycle with titanium foil to assure that the furnace is clean and leak-free prior to brazing. If the titanium foil, when bent, exhibits brittleness, the furnace is not adequately clean. When the foil comes out essentially the same color as it went in, and when it can be bent over on itself, ironed flat, and opened back up again without cracking, the furnace is adequately clean, and brazing can then be accomplished. Before brazing, the 304L, nickel 200, all of the fixtures, and any materials that will be in the furnace during the brazing operation (except the titanium and filler metal) should be put into the furnace and adequately outgassed. Many materials have included oxides and other elements that gas off and are picked up by the titanium. Precleaning ensures a better braze assembly. After the 304L is precleaned, plate it with approximately 0.01 mm (0.0004 in.) of electrolytic nickel. This plating provides a better surface for the filler metal and improves wetting and flow. If there is a question about the bond of the nickel plate, it is desirable to test the bond by putting the part into the vacuum furnace, taking it up to 982 °C (1800 °F) for 1 to 5 min, cooling it down, and then removing it from the furnace. If the nickel plate is blistered, the part should not be brazed but should be returned to the plater for stripping and replating. The blisters indicate that the cleaning was not sufficient for plating, whereas, if no blistering occurs, the nickel plating is now diffusion welded to the 304L and is ready for brazing. After all the parts, fixtures, and other materials to be put in the furnace are outgassed and cleaned, brazing may be accomplished. There are several different filler metals that may be used, but one that works well is a 68Ag27Cu-5Pd filler metal. This filler metal has a
Chapter 4: Base Metals and Base-Metal Family Groups / 161
solidus of 807 °C (1485 °F) and a liquidus of 810 °C (1490 °F). The brazing temperature is 816 to 843 °C (1501 to 1549 °F). When vacuum brazing with filler metal at a temperature of 816 to 843 °C, a partial pressure of argon is used to prevent evaporation of the silver. A partial pressure of 50 µm prevents the silver from vaporizing. Good, clean argon is used, because the titanium picks up any contaminants. For this reason, nitrogen atmosphere is not used to provide the partial pressure, because the titanium picks up the nitrogen, with deleterious effects. As a note of caution, it is wise to remember that titanium does alloy with most other metals and, if taken up in temperature too far, alloys with and melts down the stainless steel fixturing and other metal parts in the furnace. Thus, it should not be placed directly on metal parts if the temperature is to be taken to the 927 °C (170 °F) range and above. Because the braze in question is titanium directly to stainless steel and to nickel, in this case, it is essential that the brazing temperature is below the eutectic melting point of the titanium-nickel and titanium/stainless steel. Thus, the filler metal suggested previously accomplishes this. Example 4: Copper Brazing of Stainless Steel Inserts. A manufacturing firm recently attempted to copper braze 305- to 400-series stainless steel inserts. Sometimes, the copper wet the 400-series inserts, and sometimes, it flowed away from them and did not wet, covering the surface of the 300-series stainless steel. The problem was why the copper did not wet the 400-series stainless steel, even though once in a while it brazed satisfactorily. Why did the copper flow all over the 305 steel, but the 400series stainless steel did not wet? In dissecting the problem, there are several points. First, the most serious problem is discussed: the wetting of the 400-series stainless steel. In looking at the part, it is obvious that it was stamped from a sheet and thus normally has burrs at the edges. These burrs were removed, apparently by tumbling or vibratory polishing. This part has a dull, gray finish, which is quite common for parts that have been tumbled or vibratory polished. This color, and the fact that the copper does not wet the 400-series stainless steel, indicates there has been some material smeared on the surface of the part during the deburring operation. It is possible that a stone-
polishing media was used or some polishing media had previously been used for deburring parts of aluminum, magnesium, or other parts that contained elements that were not cleaned up by the atmosphere or that produced an oxide in the standard brazing atmosphere. The parts that are not wet by the copper obviously have a contaminated surface. The fact that some parts braze satisfactorily and the copper wets the 400series stainless steel indicates there is not an element, such as aluminum or possibly selenium, in the stainless steel that also causes problems with wetting of the surface. The best method is to contact the supplier of the 400-series stainless steel and to check the previous operations. If a deburring media was used, some contaminant may be the culprit. The problem with the copper flowing all over the surface of the 305 stainless steel is caused by the very thin 305 stainless steel heating up first and the 400 series heating up later. The filler metal flows toward the hottest surface. The top of the assembly, which gets up to heat first, tends to be wet with the copper. There are several ways to control the flow of copper. The first approach is to bring the part up to heat more uniformly. Heat shielding of the top of the parts assists in slowing down the rapid heating of the top end of the assembly. Heat shields are very effective ways of controlling the surface heating of parts where one section heats up much faster than another. A second approach is to lighten the fixture so the part heats up more uniformly. The third approach is to change the dewpoint of the atmosphere in the furnace. The copper then has less tendency to flow out on the surface and thus stays closer to the joint. This approach is rather difficult and takes special equipment. A fourth approach is to change the ratio of hydrogen to nitrogen, keeping the dewpoint constant. Example 5: Brazing of a Ceramic-Matrix Composite. Joining ceramic-matrix composite (CMC) materials has attracted considerable interest in recent years as a class of materials that can combine the strength and high-temperature properties of ceramics with improved damage tolerance. In many applications, it is necessary to join ceramics to metals, and brazing is a common means of achieving this. A Si3N4-fiber-reinforced cordierite glass CMC was brazed to titanium and stainless steel using silver-copper filler metal in argon and
162 / Brazing, Second Edition
four different interlayers: copper, nickel, tungsten, and MMC (SiC continuous fiber reinforced with aluminum). The CMC needs to be coated with titanium to induce wetting and bonding. Joints brazed normal to the CMC layup plane are generally much stronger than those brazed parallel to the lay-up plane. The highest joint strength measured for a CMC-titanium joint was 91.6 MPa (13.3 ksi) in shear using a 1.0 mm (0.04 in.) copper interlayer. The highest CMC/stainless steel joint strength was 106.1 MPa (15.4 ksi) using a 1.5 mm (0.06 in.) MMC interlayer. Defect-free joints are produced with all combinations except when tungsten interlayers are used, but some debonding is evident at the edges of other specimens where residual shear stresses are highest. In shear testing, all joints except the strongest failed within the ceramic composite. Brazing has been found to be an effective technique for joining ordinary structural metals to brittle, low-thermal-expansion refractory metals. Specifically, a brazing process has been established for joining copper or nickel flanges to the ends of vacuum-plasma-sprayed tungsten tubes and for joining stainless steel flanges to the ends of tubes made of an alloy of molybdenum with 40% Rh. In the original application, the tubes were furnace cartridge tubes 59.7 cm (23.5 in.) long and 25.4 mm (1 in.) in diameter, with a wall thickness of 0.76 mm (0.03 in.). The tubes were of test-tube configuration, and the flanges were joined to the open ends. In preparation for brazing, the tubes were cleaned of surface contaminants by placing them in a hydrogen furnace in the temperature range of ~1100 to 1150 °C (2010 to 2100 °F) for 15 min. That area on each tube that was to be joined to a flange was then pretinned by wrapping the area with a thin filler-metal foil and heating it to the melting temperature. After pretinning, each flange was machined to provide a braze-joint gap of 0.076 to 0.10 mm (0.003 to 0.004 in.). Each tube was then put together with its flange, and a small amount of filler metal was placed at the opening of the braze joint to provide sufficient material to fill the gap. A braze stopoff solution was applied to the surfaces surrounding the joint to prevent the filler metal from flowing onto these surfaces, thereby making it unnecessary to perform a postbraze cleanup. The tubes were then brazed, variously, in an inert or vacuum environment in
the vertical orientation, with the flanges at the lower ends. Example 6: Brazing Nickel Alloy Strip to Copper-Aluminum Bar. Is it possible to braze a 15 by 19 by 50.8 mm (0.6 by 0.75 by 2 in.) Hastelloy X strip onto a 12.7 by 19 by 50.8 mm (0.5 by 0.75 by 2 in.) long copper (0.6 to 0.9%)aluminum bar? Since it is not possible to silver braze Hastelloy X, and it must be brazed with a nickel filler metal at 1121.1 °C (2050 °F), what does one do? First, it should be pointed out that a copperaluminum bar cannot be brazed at 1121.1 °C because the melting point of this material is ~1085 °C (1985 °F); therefore, at 1121.1 °C, it is molten. Because the bar contains 0.6 to 0.9% Al and is to be furnace brazed with a nickel filler metal, it must be plated with either electrolytic nickel or copper. This will protect the aluminum from the atmosphere and will not allow it to oxidize and prevent brazing. The assembly may be brazed with a BNi-2 filler metal at 1037.8 °C (1900 °F), and BNi-7 or BNi-6, brazed at 1010 °C (1850 °F). To torch silver braze a copper-aluminum bar, it is first required to have the FB4-A flux to inhibit the formation of aluminum oxides that prevent wetting the copper. The FB4-A flux has been specifically designed for brazing aluminum-bronze base metals. The next consideration is the heating of the copper, preferably from the bottom with a ribbon burner, and uniformly applying the extra heat needed to the Hastelloy X, which is on the top of the bar. Example 7: Vacuum Brazing of Nickel Alloy Tubes to Titanium Alloy Fittings. A warm-gas fuel pump manifold consisting of 30 vacuum-brazed joints between Inconel 600 tubes and Ti-6Al-4V fittings was successfully joined. The components were part of a pistonpumped rocket propulsion system, and the braze joints had to withstand high-temperature and short-term exposure to hydrazine propellant. The filler metal used was Ag-9Pd-9Ga. A second assembly was a warm-gas pressurization regulator. It required the brazing of a 1.5 mm (0.06 in.) diameter Inconel 600 U-tube into a Ti-6Al-4V body. Because of the different configurations of the components joined, two types of vacuum furnaces were used. Both brazing systems incorporated liquid-nitrogen-trapped diffusion pumps backed by rotary mechanical pumps. The base pressure of both systems was on the order of
Chapter 4: Base Metals and Base-Metal Family Groups / 163
10–5 Pa (10–9 psi) with the brazes typically performed at approximately 10–3 Pa (10–7 psi). The system used to braze the manifold was a 10 kW, 450 kHz radio frequency (RF) induction furnace. Induction heating took place in an electrically conducting object, not necessarily magnetic, when the object was placed in a varying magnetic field. The joint was placed within (but not in contact with) a water-cooled copper coil in the vacuum enclosure, and heating was performed locally—only the joint got hot. During the braze operation, the joint could be moved vertically inside the furnace to selectively apply heat to different parts of the assembly. The advantage to RF heating was that, depending on its size, the entire assembly is not necessarily raised to the braze temperature each time a joint was brazed. This resulted in a reasonably quick turnaround time and avoided remelting previous joints each time a new joint was brazed. Where the RF induction furnace was inappropriate, because of the part configuration, a radiant heating furnace was used. This furnace used the radiant energy from a series of tungsten filaments to heat the part. Molybdenum shielding contained the heat in the brazing zone, which was located inside a double-wall, water-cooled stainless steel vacuum chamber. The temperature was monitored and controlled by using thermocouples located within the hot zone. Radiant heating was very uniform, because the entire hot zone was elevated to the braze temperature. The furnace was allowed to cool overnight under vacuum before venting to argon and removing the brazed assembly. The filler metal selected was specifically developed to braze titanium alloys to themselves and was composed of 82% Ag, 9% Pd, and 9% Ga. A 0.508 mm (0.020 in.) diameter wire was used in brazing. This filler metal is very ductile and has a brazing range between 895 and 950 °C (1645 and 1740 °F), which is below the beta transus of the titanium alloy. Work performed on brazed samples revealed the presence of a layer of palladium-gallium at the filler-metal/titanium interface, which appears to act as a diffusion barrier to silver, preventing the formation of titanium-silver intermetallics. Because of the high vapor pressure of silver and gallium, it is recommended using this filler metal in a system backfilled with dry helium or argon. However, brazing can be performed in a hard vacuum by minimizing the time at braze temperature.
Example 8: Brazing of Copper-Graphite Assemblies. Copper-graphite assemblies are used as targets for pulsed proton beams. The graphite used is paralytic grade and is engineered to have a lamellar structure, perpendicular to the proton beam, such that the heat generated within the target during operation flows vertically through the graphite to the copper base. Copper is used because it has a high thermal conductivity and is subsequently water cooled to maintain efficient heat removal. When the brazing of these materials was first attempted, the graphite sections all failed as a result of cracking at the sharp-edged corners. Initially, it was felt that procedures/equipment could be at fault. Examination of the failed bonds produced at this stage showed classical residual-stress cracks emanating from each corner of the graphite. This was due to both the large difference in CTE between graphite (0.1 × 10–6/°C, or 0.06 × 10–6/°F) and copper (18 × 10–6/°C, or 10 × 10–6/°F) and also the presence of sharp corners, which act as stress concentrators. The potential ways of resolving this problem are: • A change in filler metal to reduce the brazing temperature and hence reduce the thermal stresses induced during the heat treatment • Modification to the copper heat sink, either by changing the composition to reduce the CTE (by alloying with a low-CTE metal such as tungsten) or by a change of dimensions to reduce the cross-sectional thickness • Modification to the design of the graphite to remove the sharp edges, which act as local stress concentrators An active filler metal was selected containing titanium, which is capable of wetting onto both graphite and copper and is Ag-Cu-Ti. This filler metal brazes at 850 °C (1560 °F) and forms a TiC with the graphite. The combination of reduced brazing temperature and the formation of a slightly more elastic carbide phase may alleviate some of the cracking problems encountered. Experimentally, this one change in brazing procedure (and no change to either the graphite or copper) reduced the cracking in each component by 50%. Therefore, this method may be used in conjunction with the other design modifications. A further alteration made to the use of the filler metal was the use of stopoff compound
164 / Brazing, Second Edition
(based on titanium oxide) near the joint area. Previous joints appeared to have failed at a point of initiation close to areas of excess braze material. These areas of excess braze wetting form due to flow of the filler metal as it wets onto the two substrates. To inhibit this excessive braze wetting, stopoff solution was painted around the joint area, so that the filler metal could only wet onto the bond area itself. The hermeticity of the bond was further improved by fine polishing of the graphite joint area prior to bonding. In this way, all sharp edges were removed from the joint area, and a surface finish of 0.15 µm was achieved on the graphite. The roughness and surface finish can influence the strength of bonding, because a rough surface can prevent complete contact at an interface and can damage the ceramic (in this case, graphite) because of severe residual stresses caused by deep scratches or poor surface finish. Conversely, there is an effect of anchoring on a rough interface due to mechanical keying. In reality, it appears to be a combination of these two properties that influences the mechanical properties of the joint. Wherever possible, a fault-free surface is preferred in order to inhibit cracks manifesting from residual flaws or surface imperfections. A combination of the previously mentioned modifications allowed the production of sound graphite-copper bonds for proton targets. In summary, these were: • Use of a lower-temperature filler metal • Reduction of the copper thickness at the bond areas, and slotting to produce mechanical keying as well as chemical bonding • Polishing of the graphite to remove surface irregularities, and removal of sharp corners from the joint area
REFERENCES 1. Al-Li Alloy Bonding in the Limelight, Connect, Dec 1989, p 6 2. H. Schoer, Aluminum Brazing Alloys for Fluxless Brazing of Aluminum, VAW, Light Metals 1989, P.G. Campbell, Ed., Minerals, Metals and Materials Society, 1988, p 699–702 3. A. Rabinkin, New Applications for Rapidly Solidified Brazing Foils, Weld. J., Vol 68 (No. 10), 1989, p 39–46
4. K. Suganuma, Y. Miyamoto, and M. Koizumi, Joining of Ceramics and Metals, Annual Review of Materials Science, Vol 18, R.A. Huggins, Ed., 1988, p 47–73 5. M. Naka, M. Kubo, and L. Okamoto, Joining of Silicon Nitride with Al-Cu Alloys, J. Mater. Sci., Vol 22 (No. 12), 1987, p 4417–4421 6. H.D. Steffens, B. Wielage, and K. Biermann, Modified High Temperature Brazing Techniques for Nickel Based Materials, Ind. Heat., 17–18 May 1989 7. M. Ueki, M. Naka, and L. Okamoto, Joining and Wetting of CaO-Stabilized ZrO2 with Al-Cu Alloys, J. Mater. Sci., Vol 23 (No. 8), 1988, p 2983–2988 8. M. Ueki, M. Naka, and I. Okamoto, Wettability of Some Metals Against Zirconia Ceramics, J. Mater. Sci. Lett., Vol 5 (No. 12), 1986, p 1261 9. Y.R. Mahajan and S. Kirchhoff, Rapidly Solidified Structures of an Al-Zr-Gd Alloy, Scr. Metall., Vol 21 (No. 8), 1987, p 1125–1130 10. E. Lugscheider and K.-D. Partz, High Temperature Brazing of Stainless Steel with Nickel Base Filler Metals BNi-2, BNi-5 and BNi-7, Weld. J., Vol 62 (No. 6), 1983, p 160s–164s 11. E. Lugscheider, K.-D. Partz, and J. Kruger, Rechnerunterstutzte Gefugebeurteilung an HochtemperaturlotverbindungenProzessoptimerung zur Qualitatssicherung and Bauteilsicherheit (Prerequisite Variables for High Temperature Joining— Process Optimization Resulting in Good Quality Techniques and Control), DVSBerichte 75, Dusseldorf: Deutscher Verlag fur Schweisstechnik, 1989 12. E. Lugscheider, O. Knotek, and K. Klohn, Development of Nickel-Chromium-Silicon Base Filler Metals, Weld. J., Vol 57 (No. 10), 1978, p 319s–323s 13. E. Lugscheider and H. Krappitz, The Influence of Brazing Conditions on the Impact Strength of High-Temperature Brazed Joints, Weld. J., Vol 65 (No. 10), 1986, p 261s–266s 14. T. Osawa, Changes in the Interface Structure and Strength of Diffusion Brazed Joints of Al-Si System Alloy Castings, Weld. J., Vol 74 (No. 6), 1995, p 206–212 15. R.W. Messler, Jr., Joining Processes, JOM, Sept 1993, p 17
Chapter 4: Base Metals and Base-Metal Family Groups / 165
16. D.V. Dunford and P.G. Partridge, Diffusion Bonding of Al-Li Alloys, Mater. Sci. Technol., Vol 8, 1992, p 385–398 17. A. Ureña, J.M. Gomez De Salazar, J. Quiñones, S. Merino, and J.J. Martin, Diffusion Bonding of an Aluminum-Lithium Alloy (AA8090) Using Aluminum-Copper Alloy Interlayers—Part 1: Microstructure, J. Mater. Sci., Vol 31, 1996, p 807–818 18. R.S. Bushby and V.D. Scott, Liquid Phase Bonding of Aluminum and Aluminum/ Nicalon Composite Using Interlayers of Cu-Ag Alloy, Mater. Sci. Technol., Vol 11 (No. 7), 1995, p 643–649 19. R.S. Timsit and B.J. Janeway, A Novel Brazing Technique for Aluminum, Weld. J., June 1994, p 119–128 20. R.S. Timsit and B.J. Janeway, A Novel Brazing Technique for Aluminum and Other Metals, J. Mater. Res., Vol 8 (No. 11), 1993, p 2749–2752 21. P.E. Fortin, W.M. Kellerman, F.N. Smith, C.J. Rogers, and M.J. Wheeler, “Aluminum Materials and Processes for Automotive Heat Exchanger Applications,” Technical Paper Series 852228, Society of Automotive Engineers Int. Congr. (Detroit), 1985 22. T.B. Massalski, Ed., Binary Alloy Phase Diagrams, ASM International, 1990 23. W.E. Cooke, T.E. Wright, and J.A. Hirshfield, “Furnace Brazing of Aluminum with a Non-Corrosive Flux,” Technical Paper Series 780300, Society of Automotive Engineers Int. Congr. (Detroit), 1978 24. E.R. Wallace and E.W. Dewing, Joining of Metal Surfaces, U.S. Patent 3,951,328, 1976 25. J.T. Niemann and R.A. Garrett, Eutectic Bonding of Boron-Aluminum Structural Components, Weld. J., Vol 53, 1974, p 175–184 26. R.R. Wells, Microstructural Control of Thin-Film Diffusion-Brazed Titanium, Weld. J., Vol 55, 1976, p 20–27 27. R.S. Timsit, Method of Brazing Aluminum, U.S. Patent 5,100,048, 1992 28. W.T. Thompson and D.G.W. Goad, Some Thermodynamic Properties of K3AlF6KAlF4 Melts, Can. J. Chem., Vol 154, 1976, p 3342–3349 29. D.J. Field and N.I. Steward, “Mechanistic Aspects of the NOCOLOK Flux Brazing
30.
31.
32. 33.
34. 35. 36. 37. 38.
39.
40.
41.
42.
Process,” Society of Automotive Engineers Int. Congr. (Detroit), 1987 E. Hellum and M.J. Luton, Proc. ESA Symposium: Space Applications of Advanced Structural Materials, (Noordwijk, NL), ESTEC, 1990, p 301–305 M.J. Luton, C.S. Jayanth, M.M. Disko, S. Matras, and J. Vallone, Proc. MRS Symposium on Multicomponent Ultrafine Microstructures (Boston, MA), Materials Research Society, 1988 O. Susegg, E. Hellum, A. Olsen, and M.J. Luton, Philos. Mag. A., Vol 68, 1993, p 367–380 B. Bjørneklett, Ø. Grong, L. Anisdahl, E. Hellum, and V. Sonde, Brazing of Dispersion-Strengthened Aluminum, Weld. J., March 1996, p 57–62 B. McGurran, M.G. Nicholas, and A. Karlsson, Proc. Aluminum Technology (London, U.K.), 1986, p 528–536 J.C. Ambrose and M.G. Nicholas, Brazing and Soldering, Vol 11, 1986, p 34–38 J.R. Terrill, C.N. Cochran, J.J. Stockes, and W.E. Haupin, Weld. J., Vol 50 (No. 12), 1971, p 833 O.W. Swaney, D.E. Trace, and W.L. Winterbottom, Weld. J., Vol 65 (No. 5), 1986, p 49–57 B.N. Perevezentsev, N.M. Sokolova, and G.A. Telitsyna, Brazing Aluminium in Low Vacuum, Weld. Int., Vol 10 (No. 12), 1996, p 980–982 A.A. Suslov, Brazing Laminated Structures of Light Alloys Based on Aluminium and Titanium, Weld. Int., Vol 9 (No. 7), 1995, p 570–572 J.V. Long and G.D. Cremer, “Beryllium: Brazing and Welding,” SAE Paper 680651, Society of Automotive Engineers Aeronautic and Space Engineering and Manufacturing Meeting, 7–11 Oct 1968 (Los Angeles, CA) G.D. Cremer, J.R. Woodward, and L.A. Grant, “Beryllium Brazing Technology,” SAE Paper 670805, Society of Automotive Engineers Aeronautic and Space Engineering and Manufacturing Meeting, 2–6 Oct 1967 (Los Angeles, CA) G.R. Longhurst, “Beryllium Technology Workshop,” EGG-FSP-10017, CONF9111209, Proceedings of Idaho National Engineering Laboratory, Energy Dept., 20 Nov 1991 (Clearwater Beach, FL)
166 / Brazing, Second Edition
43. A.A. McFayden, R.R. Kapoor, and T.W. Eager, Effect of Second Phase Particles on Direct Brazing of Alumina Dispersion Hardened Copper, Weld. J., Vol 69 (No. 11), 1990, p 399s–407s 44. A.A. Suslov, Advanced Technology of Brazing Copper-Steel and Steel Sections, Weld. Int., Vol 9 (No. 6), 1995, p 473– 474 45. P. Saxty et al., Brazing of Traction Motors for Channel Tunnel and Network Southeast Trains, Weld. Met. Fabr., Vol 64 (No. 4), 1996 46. S.B. Lasday, Potential for Increased Thermal Processing Through Copper Alloy Developments, Ind. Heat., June 1998, p 14–16 47. K. Pietrzak, Joining of Carbon-FiberCopper Composite to Metals, J. Mater. Process. Technol., Vol 55 (No. 3–4), 1995, p 129–133 48. M.S. Yeh and T.H. Chuang, Transient Liquid Phase Metallic Bonding of an Inconel 718 SPF Superalloy, Weld. J., Dec 1997, p 517–521 49. W.D. Brentnall, A.R. Stetson, and A.G. Metcalfe, “Joining of Superalloy Foils for Hypersonic Vehicles,” Tech Report AFML TR-68-299, Contract F33(615)67-C-1211, Wright-Patterson AF Base— AFML and ASD, Oct 1968 50. M.M. Schwartz, Metals Joining Handbook, McGraw Hill, 1979 51. N. Kenyon and R.J. Hrubec, Brazing of a Dispersion Strengthened Nickel Base Alloy Made by Mechanical Alloying, Weld. J., Vol 53 (No. 4), 1974, p 145s–151s 52. M.M. Schwartz, “Brazed Honeycomb Structures,” WRC Interpretive Bulletin 182, Welding Research Council, April 1973 53. J. McCown, C. Wilks, M.M. Schwartz, and A. Norton, Final Report on Development of Manufacturing Methods and Processes for Fabricating Refractory Metal Components, ASD Project 7-937, AF33(657)-7276, Wright-Patterson AF Base—AFML and ASD, Sept 1963 54. Mach. Des., Vol 57 (No. 8), 18 Apr 1985 55. I.A. Bucklow, Joining a Ni-Based CreepResistant (ODS) Alloy by Brazing, The Welding Inst. IIW Meeting and Proceedings on Joining (Montreal, Canada), 1990, p 293–298
56. I.A. Bucklow, Diffusion Bonding an FeBased ODS Alloy, The Welding Inst., IIW Meeting and Proceedings on Joining (Montreal, Canada), 1990, p 299–304 57. H.D. Hedrich, H.G. Mayer, et al., “Joining of ODS-Superalloys,” ASME 91 GT 411, Gas Turbine and Aeroengine Congress and Exposition, June 1991 58. K. Nishimoto, K. Saida, D. Kim, and Y. Nakao, Transient Liquid Phase Bonding of Ni-Base Single Crystal Superalloy, CMSX-2, ISIJ Int., Vol 35 (No. 10), 1995, p 1298–1306 59. D.S. Duvall, W.A. Owczarski, and D.F. Paulonis, T.L.P.: A New Method for Joining Heat-Resisting Alloy, Weld. J., Vol 53 (No. 4), 1974, p 203–214 60. N.C. Cole, R.W. Gunkel, and J.W. Koger, Development of Corrosion Resistant Filler Metals for Brazing Molybdenum, Weld. J., Vol 52 (No. 10), 1973, p 466s–473s 61. A. Semeniuk and G.R. Brady, Properties of TZM and Nuclear Behavior of TZM Brazements, Weld. J., Vol 53 (No. 10), 1974 62. M.G. Nicholas, Active Metal Brazing, Trnas. J. Br. Ceram. Soc., Vol 85, 1986, p 144–146 63. R.G. Gilliland and G.M. Slaughter, The Development of Brazing Filler Metals for High Temperature Service, Weld. J., Vol 48 (No. 10), 1969, p 463s–468s 64. M.M. McDonald, D.L. Keller, C.R. Heiple, et al., Wettability of Brazing Filler Metals on Molybdenum and TZM, Weld. J., Vol 68 (No. 10), 1969, p 389s–395s 65. M.M. Schwartz, Brazing, ASM International, 1989 66. G.A. Robinson, C. Burns, and J.A. Shields, Jr., Ind. Heat., Aug 1993, p 39–44 67. M.M. Schwartz, “The Fabrication of Dissimilar Metal Joints Containing Reactive and Refractory Metals,” WRC Interpretive Bulletin 210, Welding Research Council, Oct 1975 68. “Advanced Joining Technology,” Report of the Committee of Advanced Joining Technology, Publication N MAB-387, Materials Advisory Board, 1982 69. G.M. Slaughter, W.J. Werner, R.G. Gilliland, et al., Recent Advances in Brazing, Fifth Not. SAMPE Tech. Conf., Vol 5,
Chapter 4: Base Metals and Base-Metal Family Groups / 167
70. 71. 72.
73.
74. 75.
76.
77.
78.
79.
80.
81.
9–11 Oct 1973 (Kiamesha Lake, NY), Society for the Advancement of Material and Process Engineering, p 115–123 Heat Treating, Vol 4, Metals Handbook, 9th ed., American Society for Metals, 1981, p 675 C.A. Zappfe, Stainless Steels, American Society for Metals, 1949 Properties and Selection of Metals, Vol 1, Metals Handbook, 8th ed., American Society for Metals, 1961; and Welding, Brazing, and Soldering, Vol 6, ASM Handbook, ASM International, 1993 C.F. Burrows, “Vacuum Brazing GasQuenching Process for the Manufacturing of Metal Materials, Components and Hardware,” final report, Contract DAAA 15- 68-C-0659, Department of the Army, Feb 1970 M.M. Schwartz, Application for GoldBase Brazing Alloys, Gold Bull., Vol 8 (No. 4), 1975 M.C. Tolle, M.E. Kassner, E. Cerri, and R.S. Rosen, Mechanical Behavior and Microstructure of Au-Ni Brazes, Metall. Mater. Trans. A, Vol 26 (No. 4), 1995, p 941–948 G.R. Bell, Vacuum Furnaces for High Strength Brazing and High Temperature Heat Treatment of Stainless Steel, Consultant, Wall Colmonoy Ltd., U.K., p 10–12 H. Izui and Y. Suezara, High Temperature Brazing for Stainless Steel SUS 316 L with Pd-Cu Brazing Filler Metals, Trans. Jpn. Weld. Soc., Vol 24 (No. 2), Oct 1993, p 29–34 Y. Hisamori, A. Nishiura, and M. Machida, Brazing of Hot End Components in the Stirling Engine, Trans. Jpn. Soc. Mech. Eng., Vol 56–523, 1990, p 457 (in Japanese) H. Izui and Y. Suezawa, Development of Pd-Ni Brazing Filler Metals for SUS 316 L Stainless Steel, Trans. Jpn. Weld. Soc., Vol 24 (No. 2), Oct 1993, p 35–40 R. Wysopal and E.R. Bangs, “The Importance of Braze Alloy Application in High Temperature Brazing,” presented at Second International AWS-WRC Brazing Conference and Colloquium, 52nd Annual American Welding Society Meeting, 26–30 April 1971 (San Francisco, CA) D.T. Lovell and S.D. Elrod, “Development and Evaluation of the Aluminum-
82.
83.
84.
85.
86.
87.
88.
89.
90.
Brazed Titanium System,” Vol 1, program summary, Final Report FAA-SS73-5-1, Federal Aviation Authority, May 1974 T. Bales et al., Weld-Brazing of Titanium, Fifth National SAMPE Tech. Conf., Vol 5, 9–11 Oct 1973 (Kiamesha Lake, NY), Society for the Advancement of Material and Process Engineering, p 481–500 R.R. Wells, Low-Temperature LargeArea Brazing of Damage-Tolerant Titanium Structures, Weld. J., Vol 54 (No. 10), Oct 1975, p 348s–356s; and Tech. Report AFML-TR-75-50, Contract F33(615)-73-C-5161, Wright-Patterson AFB-AFML, May 1975 C.E. Smeltzer and A.N. Hammer, “Titanium Braze System for High Temperature Applications,” AFML-TR-76-145, Contract F-33(615) 74-C-5118, Wright-Patterson AFB-AFML, Aug 1976 E. Lugscheider and L.M. Ruiz, Brazing of titanium, Paper 23, International Conf. on Advances in Joining and Cutting Processes, 30 Oct to 2 Nov 1989 (Harrogate, UK), Abington Publishers, p 414–423 E. Lugscheider and L. Martinez, “Brazing of Ti-Al6V4 with Al-, Ag-, Au- and TiBased Filler-Metals,” AWS Annual International Brazing Conf., 1987 (Chicago, IL); and DVS, Brazing, High Temperature Brazing and Diffusion Welding, Berichte 125, 19–20 Sept 1989 (Dusseldorf, Germany), p 123–126 N. Gamer and J. Richardson, “Investigation of Ductile Brazing Alloy Compositions for Use in Joining Titanium and Its Alloys,” Western Metals and Gold Corp. Tech. Report 1492, April 1971, p 1–32 E. Lugscheider and U. Broich, Mechanical Properties of High-Temperature Brazed Titanium Materials, Weld. J., May 1995, p 169–176 E. Lugscheider and L. Martinez, “Entwicklung Neuer Titanbasislote Zum Fügen Von TiAl6V4, (Development of New Ti-Base Filler Metals for Ti6Al4V)” DVS Berichte Bd 132, DVS-Verlag O. Botstein, A. Schwarzman, and A. Rabinkin, Induction Brazing of Ti-6Al4V Alloy with Amorphous 25Ti-25Zr50Cu Brazing Filler Metal, Mater. Sci. Eng. A, Vol 20 (No. 1), 15 Feb 1996, p 14–23
168 / Brazing, Second Edition
91. R.E. Beal and Z.P. Saperstein, Development of Brazing Filler Metals for Zircaloy, Weld. J., Vol 50 (No. 7), July 1971, p 275s–291s 92. I. Amato and M. Ravizza, Some Developments in Zircaloy Brazing Technology, Energ. Nucl., (Milan), Vol 16 (No. 2), 1969, p 35–39 93. T.J. Walz, “Using Cermets and Ceramics Instead of WC in Saws and Other Brazed Applications,” Northwest Research Institute, Inc., Carbide Processors, Inc., Feb 2001 94. K.A. Thorsen, Wetting of Cemented Carbides and Subsequent Reactions with Steel Using Cu as Brazing Filler Metal, Conference on Joining of Ceramics, Glass and Metal, 27–29 March 1985 (Bad Nauheim, FRG), Vol 1 (No. 2), p 45–57 95. K.A. Thorsen, Wetting of Cemented Carbides, Lyngby, DVS, Brazing, High Temperature Brazing and Diffusion Welding, Berichte 125, 19–20 Sept 1989 (Dusseldorf, Germany), p 66–71 96. M.M. Schwartz, Ceramic Joining, ASM International, 1990 97. C.F. Lewis, Putting Ceramics Together, Manuf. Eng., Feb 1988, p 31–34 98. A.P. Tomsia and J.A. Pask, Chemical Reactions and Adherence at Glass/Metal Interfaces: An Analysis, Dent. Mater., Vol 1 (No. 2), 1986, p 10–16 99. R.E. Loehman et al., Why Metals Adhere to Si3N4, J. Am. Ceram. Soc., Vol 67, 1988 100. M. Vilpas, “Joining of Ceramics for High-Temperature Applications,” NASA TT-20030, N87-29678, NTIS HC A03/MF A01, Oct 1987, translation of “Korkeissa Lampotiloissa Kaytettavien Keraamisten Materiaalien Liittaminen,” Report UTT-TIED-481, Tech. Res. Center of Finland, Aug 1985 101. M.G. Nicholas, Joining of Ceramics, Chapman and Hall, 1990 102. H. Mizuhara and E. Huebel, Joining Ceramic to Metal with Ductile Active Filler Metal, Weld. J., Vol 65 (No. 10), Oct 1986, p 43–51 103. G. Partridge and S.V. Phillips, New Developments in Glasses and Ceramics for the Electrical Industry, GEC J. Sci. Technol., Vol 45 (No. 1), 1978, p 11–20 104. S.S. Cole and G. Sommer, Glass-Migration Mechanism at Ceramic-to-Metal Seal
105. 106. 107.
108.
109.
110.
111. 112. 113.
114. 115. 116.
117.
118.
Adherence, J. Am. Ceram. Soc., Vol 44 (No. 6), 1961, p 265–271 R.W. Rice, Joining of Ceramics, Advances in Joining Technology, Brookhill Publishing Co., 1976, p 69–111 J.T. Klomp, Bonding of Metals to Ceramics and Glasses, Am. Ceram. Soc. Bull., Vol 51 (No. 9), 1972, p 683–688 M.E. Twentyman, Mechanism of Glass Migration in the Production of MetalCeramic Seals, J. Mater. Sci., Vol 10, 1975, p 765–776 R. Morrell, Joining to Other Components, Part I: An Introduction for the Engineer and Designer, Handbook of Properties of Technical and Engineering Ceramics, Her Majesty’s Stationery Office, London, 1985, p 267–278 R.L. Tallman, R.M. Neilson, Jr., J.C. Mittl, et al., “Joining Silicon NitrideBased Ceramics: A Technical Assessment,” ECG-SCM-6572, DE84-011356, Department of Energy Contract DEAC07-761D01570, EG&G Idaho, Inc., March 1984 H.E. Pattee, “Joining Ceramics to Metals and Other Materials,” Welding Research Council Interpretive Report 178, Nov 1972 C.W. Fox and G.M. Slaughter, Brazing of Ceramics, Weld. J., Vol 43 (No. 7), 1964, p 591–597 C.R. Weymueller, Braze Ceramics to Themselves and to Metals, Weld. Des. Fab., Aug 1987, p 45–48 R.E. Loehman, M.L. Mecartney, and D. Rowcliffe, “Silicon Nitride Joining,” AFOSR-TR-82-0304, Annual Report for 12/1/80 through 1/31/82, SRI International, Menlo Park, CA, Feb 1982 M. Santella, Brazing of Titanium-Vapor Coated Silicon Nitride, Adv. Ceram. Mater., Vol 3 (No. 5), 1988, p 457–462 S. Weiss and C.M. Adams, Jr., The Promotion of Wetting and Brazing, Weld. J., Vol 46, 1967, p 49s–57s E.F. Brush, Jr. and C.M. Adams, Jr., Vapor-Coated Surfaces for Brazing Ceramics, Weld. J., Vol 68, 1968, p 106s– 114s Ceramics Unit of the Materials Technology Group, Idaho National Engineering Laboratory, U.S. Dept. of Energy, 1998, private communication D.R. Johnson, Ceramic Technology Proj-
Chapter 4: Base Metals and Base-Metal Family Groups / 169
119.
120.
121.
122.
123.
124.
125.
126.
127. 128.
129.
130.
ect, Semiannual Progress Report April 1991-Sept 1992, Ceram-Metal Joints, Joining of Ceramics for Heat Engine Applications, M.L. Santella, Ed., Section 1.4.1, p 193 D.R. Johnson, Ceramic Technology Project, Semiannual Progress Report April 1991-Sept 1992, Analytical and Experimental Evaluation of Joining Si3N4 to Metal and SiC to Metal for Advanced Heat Engine Applications, S. Kang, Ed., p 198 S.D. Peteves´ and M.G. Nicholas, Evaluation of Brazed Silicon-Nitride Joints— Microstructure and Mechanical Properties, J. Am. Ceram. Soc., Vol 79 (No. 6), 1996, p 1555–1562 H. Takeuchi, Fine Ceramics: Trial and Accomplishment of Next-Generation R and D, Sci. Technol.: Jpn., JPRS-JST-94009, May 1994, p 27–31, 33–39 S. Trickey, V. Trehan, and J.E. Indacochea, Brazing of Si3N4 with Active Filler Metals, Univ. of Illinois at Chicago, AM&P, Jan 1999, p 18 M. Nakamura, I. Shigematsu, K. Kanayama, and Y. Hirai, Joining of Si3N4 Ceramics with Nickel and Ni-Cr Alloy Laminated Interlayers, J. Mater. Sci. Ltt., Vol 12 (No. 10), 1993, p 716–718 W. Tillmann, E. Lugscheider, K. Schlimbach, et al., Heat-Resistant Active Brazing of Silicon Nitride, Part 1: Mechanical Evaluation of Braze Joints, Weld. J., Aug 1997, p 300s–304s K. Sasaki, K. Motoie, and T. Kawasaki, Strength and Fatigue of Silicon Nitride/ Silicon Nitride and Alumina Joints, JSME Int. J. A, Vol 39 (No. 11), 1996, p 9–17 A.M. Hadian and R.A.L. Drew, Strength and Microstructure of Silicon Nitride Ceramics Brazed with Nickel-Chromium-Silicon Alloys, J. Am. Ceram. Soc., Vol 79 (No. 3), 1996, p 659–665 K.H. Jack, Review: SIALONS and Related Nitrogen Ceramics, J. Mater. Sci., Vol 11, 1976 R.E. Loehman, Preparation and Properties of Yttrium-Silicon-Aluminum OxyNitride Glasses, J. Am. Ceram. Soc., Vol 62, 1979, p 9–10 G. Rossi, “Joining of Non-Oxide Ceramics,” ASM 1988 World Materials Cong., ASM-ACS Joining Session (Chicago, IL), 1988 E. Lugscheider, M. Boretius, and R.
131.
132.
133.
134.
135.
136. 137.
138. 139.
140.
141.
142.
Lison, “Active Brazing of Silicon Carbide to Steel Using a Thermal-Stress Reducing Metallic Interlayer,” IIW-Doc. 1-860-88, presented at 19th AWS Braze and Solder Conf., American Welding Society, 1988 K. Koga, Fine Ceramics: Trial and Accomplishment of Next-Generation R and D; Sci. Technol.: Jpn., JPRS-JST-94009, May 1994, p 82 H.-K. Lee, S.-H. Hwang, and J.-Y. Lee, Effects of the Relative Contents on the Interfacial Reactions and Bond Strength in the Active Brazing of SiC, J. Mater. Sci., Vol 28 (No. 7), 1993, p 1765–1774 M. Singh, Affordable Fabrication and Joining of Fiber-Reinforced Ceramic Composites, Technology 2006, NASA, 1996 M. Singh, J.D. Kiser, and S.C. Farmer, Joining of Silicon Carbide-Based Ceramics by Reaction Forming Approach, Ceram. Eng. Sci. Proc., American Ceramic Society, Issue 2, 1998 M. Singh and J.D. Kiser, “Joining of Silicon Carbide-Based Ceramics by Reaction Forming Method,” NASA CP-10193, 5:1–10, presented at the Physics and Process Modeling and Other Propulsion R and T Conference, NASA, 1997 M. Singh, A Reaction Forming Method for the Joining of Silicon Carbide-Based Ceramics, Scr. Mater., Vol 33, 1998 M. Singh, Microstructure and Mechanical Properties of Reaction Formed Joints in Reaction Bonded Silicon Carbide Ceramics, J. Mater. Res., Vol 14, 1999 M. Singh, Joining of Sintered Silicon Carbide Ceramics for High Temperature Applications, Scr. Mater., Vol 34, 1999 M. Singh, Joining of Silicon CarbideBased Ceramic Materials for High Temperature Applications, Ind. Heat., Sept 1997, p 91–93 P.L. Ratnaparkli and J.M. Howe, Structure and Mechanism of Bonding at a Diffusion-Bonded Al/SiC Interface, Acta Metall. Mater., Vol 42 (No. 3), 1994, p 811–823 B. Drevet, S. Kalogeropoulou, and N. Eustathopoulos, Wettability and Interfacial Bonding in Au-Si/SiC System, Acta Metall. Mater., Vol 41 (No. 11), 1993, p 3119–3126 R.E. Loehman and A.P. Tomsia, Wetting and Joining of Mullite Ceramics by
170 / Brazing, Second Edition
143.
144.
145.
146.
147.
148.
149.
150.
151.
152. 153. 154.
Active-Metal Braze Alloys, J. Am. Ceram. Soc., Vol 77 (No. 1), 1994, p 271–274 D. Huh and D.-H. Kim, Joining of AlN to Cu Using In-Base Active Brazing Fillers, J. Mater. Res., Vol 12 (No. 4), April 1997, p 1048–1055 R.T. Cassidy et al., “Bonding and Fracture of Titanium-Containing Braze Alloys to Alumina,” Monsanto Research Corp., MLM-3431 (OP), MLM-3394, DE87002197, and DE87009195, U.S. Dept. of Energy Contract DE-ACOE76DP00053, Oct 1986, 1987 S. Liu, D.L. Olson, G.P. Martin, et al., Modeling of Brazing Processes That Use Coatings and Interlayers, Weld. J., Vol 70 (No. 8), 1991, p 207–215 D.A. Canonico, N.C. Cole, and G.M. Slaughter, Direct Brazing of Ceramics, Graphite and Refractory Metals, Weld. J., Vol 56 (No. 8), 1977, p 31–38 H.W. Grunling, High Temperature-Resistant Ceramic/Metal Joints, DVS Working Subcommittee UG30.1, Schweissen Schneiden, Feb 1973 A. Meyer, The Adhesion Mechanism of Molybdenum/Manganese Metallized Coatings on Corundum Ceramics, DKG Report 42, Vol 11, 1965, p 405–415, and Vol 12, 1965, p 14–16 A.J. Moorhead and H. Keating, Direct Brazing of Ceramics for Advanced Heavy-Duty Diesels, Weld. J., Vol 65 (No. 10), 1986, p 17–31 A.J. Moorhead and P.F. Becher, Development of a Test for Determining Fracture Toughness of Brazed Joints in Ceramic Materials, Weld. J., Vol 66 (No. 1), 1987, p 26–32 A. Meier, V. Gabriel, O.R. Chidambaram, and G.R. Edwards, The Wettability of Copper-Manganese Alloys on Alumina and Their Potential as Direct Brazing Filler Metals, Mater. Manuf. Process., Vol 10 (No. 4), 1995, p 625–641 A.P. Tomsia and R.E. Loehman, Mater. Manuf. Process., Vol 9 (No. 3), 1994, p 547–561 K. Suganuma, New Process for Brazing Ceramics Utilizing Squeeze Casting, J. Mater. Sci., Vol 26, 1991, p 6144–6150 G.P. Kelkar, K.E. Spear, and A.H. Carim, Thermodynamic Evaluation of Reaction
155.
156.
157.
158.
159.
160.
161.
162.
163.
164.
165.
Products and Layering in Brazed Alumina Joints, J. Mater. Res., Vol 9 (No. 9), 1994 E. Lugscheider, W. Bürger, and U. Broich, Development and Characterization of Joining Techniques for Dispersion-Strengthened Alumina, Weld. J., Sept 1997, p 349–355 M.L. Shalz, B.J. Dalgleish, A.P. Tomsia, and A.M. Glaeser, Ceramic Joining III: Bonding of Alumina via Cu/Nb/Cu Interlayers, J. Mater. Sci., Vol 29 (No. 14), 1994, p 3678–3690 M.L. Shalz, B.J. Dalgleish, A.P. Tomsia, and A.M. Glaeser, Ceramic Joining, Part I: Partial Transient Liquid Phase Bonding of Alumina via Cu/Pt Interlayers, J. Mater. Sci., Vol 28, 1993, p 1673–1684 M.L. Shalz, B.J. Dalgleish, A.P. Tomsia, and A.M. Glaeser, Ceramic Joining, Part II: Partial Transient Liquid Phase Bonding of Alumina via Cu/Ni/Cu Multilayer Interlayers, J. Mater. Sci., Vol 29, 1994, p 3200–3208 G.H.M. Gubbels, L.S.K. Heikinheimo, and J.T. Klomp, A Comparison Between Titanium Alumina Diffusion Bonding and Titanium Active Brazing, Z. Metallkd., Vol 85 (No. 12), 1994, p 828–832 K.-S. Bang and S. Liu, The Interfacial Reaction Between Alumina and Cu-Ti Braze Filler Metal During Reactive Metal Brazing, Weld. J., March 1994, p 54–60 J.G. Li, Wetting of Ceramic Materials by Liquid Silicon, Aluminum and Metallic Melts Containing Titanium and Other Reactive Elements—A Review, Ceram. Int., Vol 20 (No. 6), 1994, p 391–412 B.J. Dalgleish, M.C. Liu, and A.G. Evans, The Strength of Ceramics Bonded with Metals, Acta Metall., Vol 36 (No. 8), 1988, p 2029–2035 H. Takeda, K. Nogi, and K. Ogino, The Effect of the Interfacial Structure on the Joining Strength of ZrO2-ZrO2 Joint with Al Foil, Mater. Trans., JIM, Vol 35 (No. 1), 1994, p 46–50 H.Q. Hao, Y.L. Wang, Z.H. Jin, and X.T. Wang, Joining of Zirconia to Zirconia Using Ag-Cu-Ti Filler Metal, J. Mater. Process. Technol., Vol 52 (No. 2–4), 1995, p 238–247 R.H. Vegter, Diffusion Bonding of Zirconia, Delft University of Technology, The Netherlands, 1999
Chapter 4: Base Metals and Base-Metal Family Groups / 171
166. Ind. Diamond Rev., Vol 54 (No. 3), 1994, p 117–119 167. B.H. Rabin, Joining of Fiber-Reinforced SiC Composites by In-Situ Reaction Methods, Mater. Sci. Eng. A, Vol 130, 1990, p L1–L5 168. H. Morimoto, T. Tanaka, T. Saito, et al., Effects of Brazing Temperature on SiC Fiber Reinforced Aluminum Alloy Matrix Composites, Advances in Joining Newer Structural Materials, Proceedings of International Conf. on IIW, Pergamon Press, Inc., 1990, p 137 169. V.F. Khorunov, V.S. Kutchuk-Yatsenko, I.S. Dykhno, et al., Brazing of Sheet Composite Materials with Aluminum Matrix, Advances in Joining Newer Structural Materials, Proceedings of International Conf. on IIW, Pergamon Press, Inc., 1990, p 143 170. E.K. Hoffman, R.K. Bird, and D.L. Dicus, Brazed Joint Properties and Microstructure of SCS-6/β21 S Titanium Matrix Composites, Weld. J., Nov 1995, p 378–384 171. E.K. Hoffman, R.K. Bird, and D.L. Dicus, “Effect of Braze Processing on Microstructure and Mechanical Properties of SCS-6/Beta 21 S Titanium Matrix Composites,” AIAA-92-5017, AIAA Fourth International Aerospace Planes Conference, 1–4 Dec 1992 (Orlando, FL) 172. C.A. Blue, V.K. Sikka, and R.A. Blue, Infrared Transient-Liquid-Phase Joining of SCS-6/Beta-21S Titanium Matrix Composite, Metall. Mater. Trans. A, Vol 27 (No. 12), 1996, p 4011–4018 173. Infrared Joining of Titanium-Matrix Composites, High-Tech Materials Alert, Technical Insights, 15 April 1996, p 4–5 174. C.A. Blue, R.A. Blue, R.Y. Lin, J.-F. Lei, et al., Joining of SCS-B/Beta 21S Titanium Composite, Process. Adv. Mater., Vol 4 (No. 2), 1994, p 105–110 175. A. Ureña, J.M. Gomez De Salazar, M.D. Escalera, and M.I. Fernández, Study of the Brazeability of Aluminum Matrix Composites, Weld. J., Feb 1997, p 92–102 176. M.B.D. Ellis, Joining of Al Based Metal Matrix Composites—A Review, Mater. Manuf. Process., Vol 11 (No. 1), 1996, p 45–56 177. P.G. Partridge, M. Shepherd, and D.V.
178.
179.
180.
181. 182.
183.
184.
185.
186. 187.
188. 189.
190.
Dunford, RAE Technical Report 90024, Royal Aeronautical Establishment, 1990 P.G. Partridge and D.V. Dunford, The Role of Interlayers in Diffusion Bonded Joints in Metal Matrix Composites, J. Mater. Sci., Vol 26, 1991, p 2255–2258 P.G. Partridge, M. Shepard, and D.V. Dunford, Statistical Analysis of Particulate Interface Lengths in Diffusion Bonded Joints in a Metal Matrix Composite, J. Mater. Sci., Vol 26, 1991, p 4953– 4960 S.B. Dunkerton, Applications of Diffusion Bonding, Welding and Applied Developments for the Process Industries, Proceedings of the International Symposium, A.F. Gifford, Ed., 12–14 April 1988 (London), The Welding Institute, p 273–282 K. Sudhakar, Ph.D. dissertation, Ohio State University, 1991 E.M. Breinan and K.G. Kreider, Braze Bonding and Joining of Aluminum Boron Composites, Met. Eng. Q., Vol 9, 1969, p 192–202 C.S. Beyukian, Brazing of Refractory, Superalloy, and Composite Material for Space Shuttle Applications, North American Rockwell Corporation, 1972 R. Klehn and T.W. Eagar, Joining of 6061 Aluminum Matrix-Ceramic Particle Reinforced Composites, Weld. Res. Counc. Bull., No. 385, 1993 P.G. Partridge, Diffusion Bonding of Metals, Superplasticity, AGARD Lecture Series, No. 154, NATO AGARD (Advisory Group for Aerospace Research and Development), Paper 5, Aug 1987, p 1–23 M.M. Schwartz, Composite Materials Handbook, McGraw Hill, 1984, p 6.1– 6.45 D.M. Royster, R.R. McWithey, and T.T. Bales, “Fabrication and Evaluation of Brazed Titanium Clad Borsic-Aluminum Compression and Skin-Stringer Panels,” NASA TP-1573, NASA TP-1976, NASA Langley Research Center, 1980 C.A. Blue, R.A. Blue, and R.Y. Lin, Infrared Joining of Titanium Aluminide, Proc. Adv. Mater., Vol 4, 1994, p 141–147 T. Moore and J.M. Kalinowski, More About Brazing or Welding NiAl Without Exogeneous Filler, NASA Tech. Briefs, LEW-15671, June 1996, p 84 J.A. Hamill, Jr., What Are the Joining Processes, Materials and Techniques for
172 / Brazing, Second Edition
191.
192. 193. 194. 195.
196. 197.
198.
199.
200. 201.
202.
203.
Powder Metal Parts?, Weld. J., Feb 1993, p 37–45 H.E. Pattee, R.M. Evans, and R.E. Monroe, “Joining Ceramics and Graphite to Other Materials,” NASA SP-5052, National Aeronautics and Space Center, 1968 R.G. Donnelly and G.M. Slaughter, The Brazing of Graphite, Weld. J., Vol 41 (No. 5), 1962, p 461–469 Welding, Brazing, and Soldering, Vol 6, ASM Handbook, ASM International, 1993, p 1023 Super Ceramic-Steel Ceramic-Graphite Bonds, High-Tech. Mater. Alert, Vol 5 (No. 5), 1988, p 1–2 D.V. Kochetov et al., Investigation of Heat Conditions in the Brazing of Graphite to Steel, Weld. Prod., Vol 21 (No. 3), 1974, p 15–18 L.T. Anikin et al., The High Temperature Brazing of Graphite, Weld. Prod., Vol 24 (No. 1), 1977, p 39–41 L.T. Anikin et al., The High Temperature Brazing of Graphite Using an Aluminum Brazing Alloy, Weld. Prod., Vol 24 (No. 7), 1977, p 23–25 I. Amato, P. Cappelli, and P. Martinengo, Brazing of Special Grade Graphite to Metallic Substrates, Weld. J., Vol 53 (No. 10), 1974, p 623–628; and Fiat-DCR Laboratori Centrali, SMT-25, Orbassano, Italy, Feb 1974 Y. Iino and N. Taguchi, Interdiffusing Metals Layer Technique of Ceramic-Metal Bonding, J. Mater. Sci. Lett., Vol 7, 1988, p 981–982 A.J. Moorhead, Welding and Brazing of Film Probe Sensor Assemblies, Weld. J., Vol 62 (No. 10), 1983, p 17–27 R.R. Kapoor and T.W. Eagar, “Oxidation Behavior of Brazing Alloys for Metal/ Ceramic Joints,” Prog. Report 9-41-88, Dept. of Materials Science and Engineering, MIT Indust. Liaison, Cambridge, MA, 1988 H. Mizuhara and K. Mally, Ceramic-toMetal Joining with Active Brazing Filler Metal, Weld. J., Vol 64 (No. 10), 1985, p 27–32 S.D. Peteves, Ed., Designing Interfaces for Technological Applications: Ceramic-Ceramic, Ceramic-Metal Joining, Elsevier Science and Publishing Co., Inc., 1989
204. M.G. Nicholas and R.M. Crispin, “The Role of Titanium in Active Metal and Activated Brazing of Alumina,” Proceedings of Second International Colloquium on the Joining of Ceramics, Glass, and Metals, 27–29 March 1985 (Bad Nauheim, Germany), Deutsche Keramische Gesellschaft e.V., Bad Honnef, FRG 205. H. Mizuhara, Vacuum Brazing Ceramics to Metals, Adv. Mater. Process., Vol 131 (No. 2), 1987, p 53–55 206. Brazing Handbook, 4th ed., American Welding Society, 1991 207. L. Reed, W. Wade, S. Vogel, R. McRae, and C. Barnes, AD636950, Report 12, TR66-1, R/E 66-115, Metallurgical Research and Development for Ceramic Electron Devices (Eimac), Division of Varian, 1966 208. W.H. Kohl, Handbook of Materials and Techniques for Vacuum Devices, Reinhold, 1967, p 441 209. H.J. Nolte and R. Spurck, Metal-Ceramic Sealing with Manganese, Telev. Eng., Vol 1 (No. 11), 1950, p 14–16, 18, 39 210. M.G. Nicholas, Reactive Metal Brazing of Ceramics, Scand. J. Metall., Vol 20, 1991, p 157–164 211. M.G. Norton, J.M. Kajda, and B.C.H. Steele, Brazing of Aluminum Nitride Substrates, J. Mater. Res., Vol 5 (No. 10), 1990, p 2172–2176 212. M.M. Schwartz, The Fabrication of Dissimilar Metal Joints Containing Reactive and Refractory Metals, WRC Interpr. Bull., 210, Oct 1975 213. J.G. Young and A.A. Smith, Joining Dissimilar Metals, Weld. Met. Fabr., July 1959 214. D. Peckner, Joining Dissimilar Metals, Mater. Des. Eng., Vol 56 (No. 2), 1962, p 115–122 215. T. Minegishi, T. Sakurai, and S. Morozumi, Electric-Assisted and Field-Depressed Segregation of Reactive Metals to the Bonding Interface in Braze Alloy Joining, J. Mater. Sci., Vol 26 (No. 2), 1991, p 5473–5480 216. M.L. Santella, “Joining of Ceramics for Heat Engine Applications,” ORNL/TM11489, DE 91-001004, Semiannual Report 4/89 through 9/89, p 227–231 217. S. Kang, H. Kim, and J.H. Selverian, “Analytical and Experimental Evaluation of Joining Silicon Nitride to Metal and
Chapter 4: Base Metals and Base-Metal Family Groups / 173
218.
219.
220.
221.
222.
223.
224.
225.
226.
227.
228.
Silicon Carbide to Metal for Advanced Heat Engines Application,” ORNL/TM11489, DE 91-001004, Semiannual Report 4/89 through 9/89, p 232–247 A.T. Hopper, J. Ahmad, and A. Rosenfield, “Analytical and Experimental Evaluation of Joining Ceramic Oxides to Ceramic Oxides and Ceramic Oxides to Metal for Advanced Heat Engine Application,” ORNL/TM-11489, DE 91001004, Semiannual Report 4/89 through 9/89, p 248–267 A.G. Pincus, Mechanism of Ceramic-toMetal Adherence, Adherence of Molybdenum to Alumina-Ceramics, Ceram. Age, Vol 63 (No. 3), 1954, p 16–32 K. Suganuma, T. Okamoto, and M. Shimada, New Method for Solid-State Bonding Between Ceramics and Metals, Commun. Am. Ceram. Soc., July 1983, C-117– 118 J.F. Burgess and C.A. Neugebauer, Direct Bonding of Metals to Ceramics for Electronic Applications, Advances in Joining Technology, E. Burke et al., Ed., 1974 Y. Miyazawa and T. Ariga, A Study of the Brazeability of Nickel-Based Brazing Filler Metal Foil for Joining Nickel Base Metal to Mild Steel Base Metal, Weld. J., July 1993, p 294–300 Y. Miyazawa and T. Ariga, Brazing of Inconel 600 and SUS 304 Stainless Steel with Use of Rapidly Solidified NickelBase Brazing Foil, Mater. Trans., JIM, Vol 33 (No. 5), 1992, p 509–518 K. Suganuma and N. Kawakami, Novel Processing of Brazing Aluminium to Aluminium and to Austenitic Stainless Steel, Mater. Sci. Technol., Vol 9 (No. 4), 1993, p 349–356 M. Roulin, J.W. Luster, G. Karadeniz, and A. Mortensen, Strength and Structure of Furnace-Brazed Joints Between Aluminum and Stainless Steel, Weld. J., May 1999, p 151–155 P.R.C. Camargo, R.E. Trevisan, and S. Liu, Microstructural Characterization of Titanium to 304 Stainless Steel Brazed Joints, Weld. J., Dec 1993, p 537–543 B. Alemán, I. Gutiérrez, and J.J. Urcola, Interface Microstructures in Diffusion Bonding of Titanium Alloys to Stainless and Low Alloy Steels, Mater. Sci. Technol., Vol 9 (No. 8), 1993, p 633–641 B. Alemán, I. Gutiérrez, and J.J. Urcola,
229.
230.
231.
232.
233.
234.
235. 236.
237.
238. 239.
240.
Interface Microstructures in the Diffusion Bonding of a Titanium Alloy Ti 6242 to an Inconel 625, Metall. Mater. Trans. A, Vol 26 (No. 2), 1995, p 437–446 T.J. Moore, Preliminary Study on Pressure Brazing and Diffusion Welding of Nb-1Zr to Inconel 718, Weld. J., March 1990, p 98–102 T. Oyama and H. Mizuhara, “Direct Brazing of Graphite Using a Copper-Based Active Brazing Filler Metal,” International Institute of Welding Document IA1012-97, 1997 S. Fukumoto, A. Kasahara, A. Hirose, and K.F. Kobayashi, Transient Liquid Phase Diffusion Bonding of Continuous SiC Fibre Reinforced Ti-6Al-4V Composite to Ti-6Al-4V Alloy, Mater. Sci. Technol., Vol 10 (No. 9), 1994, p 807–812 A. Hirose, S. Fukumoto, and K.F. Kobayashi, Joining Processes for Structural Applications of Continuous Fiber Reinforced MMCs, Key Eng. Mater., Vol 104–107, 1995, p 853–872 E.-S. Park, J.J. Lannutti, and J.D. Cawley, Microstructure and Reactions of SiCWReinforced Alumina with Ag-Cu-In-Ti, J. Am. Ceram. Soc., Vol 78 (No. 1), 1995, p 15–20 S.M. Chen and S.T. Lin, Brazing Diamond Grits onto a Steel Substrate Using Copper-Alloys as the Filler Metals, J. Mater. Eng. Perform., Vol 5 (No. 6), 1996, p 761–766 S.V. Orel, L.C. Parous, and W.F. Gale, Diffusion Brazing of Nickel Aluminides, Weld. J., Jun 1995, p 319–324 W.F. Gale and Y. Guan, Transient Liquid-Phase Bonding in the NiAl/Cu/Ni System—A Microstructural Investigation, Metall. Mater. Trans. A, Vol 27 (No. 11), Nov 1996, p 3621–3629 M.J. Strum and G.A. Henshall, Advanced Joining Technologies for New Materials II, N.F. Fiore and J.O. Stiegler, Ed., AWS, Miami, FL, 1994, p 76–88 T.J. Moore and J.M. Kalinowsk, MRS Symp. Proc., 1993, Vol 288 G.R. Edwards and S. Liu, Fundamental Aspects of Metal-to-Ceramic Brazing, Final Report Oct 1994, Colorado School of Mines, MT-CWJR-094-037, Oct 1994 G. Wang and J.J. Lannutti, Chemical Thermodynamics as a Predictive Tool in the Reactive Metal Brazing of Ceramics,
174 / Brazing, Second Edition
241.
242.
243. 244.
245.
246. 247. 248.
249.
250.
251. 252.
253.
254.
Metall. Mater. Trans. A, Vol 26 (No. 6), Jun 1995, p 1499–1505 R.E. Loehman, Ceramic-Metal Reactions in Composites, Ceramic Joining, and Electronic Packaging, Scr. Metall. Mater., Vol 31 (No. 8), Oct 15 1994, p 965–970 J.M. Howe, Bonding, Structure, and Properties of Metal-Ceramic Interfaces: Part 1. Chemical Bonding, Chemical-Reaction, and Interfacial Structure, Inter. Mater. Rev., Vol 38 (No. 5), 1993, p 233–256 The Promise of Ceramics, Adv. Mater. Process., Vol 131 (No. 1), Jan 1987, p 44–50 M.E. Twentyman, The Use of Metallizing Paints Containing Glass or Other Inorganic Bonding Agents, J. Mater. Sci., Vol 10, 1975, p 791–798 M. Naka, K. Sampath, I. Okamoto, et al., Influence of Brazing Conditions on Shear Strength of Alumina-Kovar Joint Made with Amorphous Cu50 Ti50 Filler Metal, Trans. Jpn. Weld. Res. Inst., Dec 1983 H.J. Nolte, Metallized Ceramic, U.S. Patent 2,667,432, Jan 26, 1954 K. Ettre, Automatic Metallizing, Ceram. Age, Vol 81 (No. 6), 1965, p 57–60 M.E. Staumanis and A.W. Schlechten, Titanium Coatings on Metals and Ceramics, Metallurgia, Vol 10 (No. 10), 1956, p 901–909 R.J. Heritage and J.R. Balmer, Metallizing of Glass, Ceramic, and Plastic Surfaces, Metallurgia, Vol 47 (No. 4), 1953, p 171–174 L.H. LaForge, Application of Ceramic Sections in High-Powered Pulsed Llystrons, Am. Ceram. Soc. Bull., Vol 35 (No. 3), Mar 1956, p 117–122 W.M. Phillips, Metal-to-Ceramic Seals for Thermionic Converters, A Literature Survey, JPL/Report 32-1420, Nov 1969 R.F. Spurck et al., Use Metallizing Tape for High Quality Ceramic-to-Metal Seals, Ceram. Inc., Vol 79 (No. 3), 1962, p 88–91, 94 L.A. Tentarelli, J.M. White, and R.W. Buck, Low-Temperature Refractory Metal-to-Ceramic Seals, Final Rept. ECOM03734-F, Cont. DA-36-039-AMC-03734 (E), Sperry Rand Corp., Gainesville, FL, Apr 1966 P.F. Varadi and R. Dominiguez, Tungsten
255.
256.
257.
258. 259. 260.
261.
262.
263.
264.
265. 266.
267.
268.
Metallizing of Ceramics, Am. Ceram. Soc. Bull., Vol 45 (No. 9), 1966, p 789– 791 K.H. Kohl and P. Rice, Electronic Tubes for Critical Environments, Tech. Rept. TR-57-434, Cont 33(616)-3460, Menlo Park, CA, Stanford Res. Ctr., Mar 1985 R.A. Quinn and R.F. Karlak, “Method of Coating a Body with Titanium and Related Metals,” U.S. Patent 3,022,021, 20 Feb 1962 C. Lindquist, Here Are Materials and Techniques for Coating Ceramics, Ceram. Ind., Vol 69 (No. 1), 1957, p 85– 88 C. Lindquist, How to Apply Conductive Coatings, Ceram. Ind., Vol 69 (No. 2), 1957, p 101–103 C. Lindquist, How to Dry and Fire Conductive Coatings, Ceram. Ind., Vol 69 (No. 3), 1957, p 144–150 A. Sedenka, Studies of the Effect of Firing Temperature on the Adherence of Silver to Ceramics, Am. Ceram. Soc. Bull., Vol 38 (No. 4), 1959, p 139–141 W. Knecht, Application of Pressed Powder Technique for Production of Metal-toCeramic Seals, Ceram. Age, Vol 63 (No. 2), 1954, p 12–13 M. Nakanishi, Joining Ceramics to Metals, Part 1: Interfacial Reactions Between Ceramics and Metals, Weld. Int., Vol 10 (No. 10), 1996, p 765–770 M.G. Nicholas, D.A. Mortimer, L.M. Jones, et al., Some Observations on the Wetting and Bonding of Nitride Ceramics, J. Mater. Sci., Vol 25, 1990, p 2679– 2689 G.L. White and P.J. Oakley, Industrial Ceramics—A Survey of Materials, Applications, and Joining Processes, Members Report Summary—Report 302/1986, Welding Institute Research Bulletin, 9444.1/85/476.3, June 1986, p 1–24 Metals Data Book, Maruyoshi Press, 1993 T. Yamada, M. Harino, K. Yokoi, et al., Hermetic Seal of Ceramic and Metal Joints by an Al-Si Interlayer, Mater. J. Sci. Lett., Vol 10, 1991, p 807–809 Y. Zhou, F.H. Bao, J.L. Ren, et al., Interlayer Selection and Thermal Stresses in Brazed Si3N4-Steel Joints, Mater. Sci. Technol., Vol 7, 1991, p 863–868 R.E. Loehman, S.M. Johnson, and A.J.
Chapter 4: Base Metals and Base-Metal Family Groups / 175
269.
270.
271.
272.
273.
274.
275.
276.
277.
278.
279.
Moorhead, Structural Ceramics Joining, Ceram. Eng. Sci. Proc., Vol 10 (No. 11–12), 1989 R.E. Loehman, Joining Engineering Ceramics, Advances in Joining Newer Structural Materials, Proceedings of International Conf. on IIW, Pergamon Press, Inc., 1990, p 83–97 A. Levy, Thermal Residual Stresses in Ceramic-to-Metal Brazed Joints, J. Am. Ceram. Soc., Vol 74 (No. 9), 1991, p 2141–2147 B. Fanghan, R. Jialie, Z. Yunhong, et al., Properties of Vacuum Brazed Si3N4/Steel Joint Using Active Brazing Filler Metal, Advances in Joining Newer Structural Materials, Proceedings of International Conf. on IIW, Pergamon Press, Inc., 1990, p 117–122 A.P. Xian, Z.Y. Si, L.J. Zhou, et al., An Improvement of the Oxidation Resistance of Ag-Cu Eutectic-5 at.% Brazing Alloy for Metal/Ceramic Joints, Mater. Lett., Vol 12, 1991, p 84–88 A.J. Moorhead and H.-E. Kim, Oxidation Behaviour of Titanium-Containing Brazing Filler Metals, J. Mater. Sci., Vol 26 (No. 15), 1991, p 4067–4075 J.H. Selverian and S. Kang, Ceramic-toMetal Joints Brazed with Palladium Alloys, Weld. J., Vol 71 (No. 1), 1992, p 25–33 Y. Nakao, K. Nishimoto, and K. Saida, Bonding of Heat-Resisting Fine Ceramics to Metals, Recent Trends in Welding Science and Technology, TWR ‘89; Trends in Welding Research, Proceedings of Second International Conf. on, ASM International, 1990, p 535–539 J.R. McDermid and R.A. Drew, Brazing of Reaction-Bonded Silicon Carbide and Inconel 600 with an Iron-Based Alloy, J. Mater. Sci., Vol 25, 1990, p 4804–4809 C. Qiu and H. Xia, A Research on the Mechanism of Joining Graphite to Ceramics of 95% Al2O3 Advances in Joining Newer Structural Materials, Proceedings of International Conf. on IIW, Pergamon Press, Inc., 1990, p 129–135 A.-P. Xian, Z.-Y. Si, and Z.-J. Hou, Oxidation of Ag57Cu38Ti5 Brazing Filler Metal for Metal-Ceramic Joint, J. Mater. Sci. Lett., Vol 10, 1991, p 726–727 I.A. Bucklow, Ceramic/Metal Bonding: A Study of Joining Zirconia to Cast Iron,
280.
281.
282.
283.
284.
285.
286.
287.
288.
289.
290.
291.
and Silicon Carbide to Steel, Weld. Inst. Bull., 414/1990, Sept/Oct 1990 C. Qiu, Mechanism Research on Seal Between 95% Alumina Ceramics and Activated Metal of Oxygen-Free Copper, Proc. of Fourth Annual Meeting of Printed Circuit Electronics Soc. E-vacuum Assoc., 1982, p 155–156 C. Qiu, Technical Improvements on Weld of R-1 Beryllium and Beryllium Oxide Ceramics, Proc. of Experimental Exchange Meeting on Brazing Materials and Its Applications, I, Vol 8, 1984 C. Qiu, Advantages of Titanium and Its Alloy in Welding Application, Proc. of Fifth National Weld. Sym., Vol 8, AWS, 1989 W.P. Weng, H.W. Wu, and Y.H. Chai, Interfacial Characteristics for Active Brazing of Alumina to Superalloys, J. Adv. Mater., Vol 28 (No. 2), 1997, p 35–40 W. Wlosinski, W. Olesinska, and K. Pietrzak, Bonding of Alumina to Steel Using Copper Interlayer, J. Mater. Proc. Technol., Vol 56 (No. 1–4), 1996, p 190–199 G.H.M. Gubbels, “Aspects of Active Brazing,” Internal Report TPD-CTKRTP-92-065, TNO, Eindhoven, Netherlands, 1992 B.T.J. Stoop and G. Denouden, Diffusion Bonding of Silicon-Nitride to Austenitic Stainless-Steel with Metallic Interlayers, Metall. Mater. Trans. A, Vol 26 (No. 1), 1995, p 203–208 M. Miyake, Sumitomo Electric Silicon Nitride Joining Technology Development Program, Japan’s 21st Century, June 1993, p 61 A. Frisch, W.A. Kaysser, W. Zhang, and G. Petzow, Stress Relaxation with Bonding in Si3N4/MA6000 Joints by Reactive Interlayers, Acta Metall. Mater., Vol 40 supplement, 1992, p S361–S368 R. Xu and J.E. Indacochea, Reaction Layer Characterization of the Braze Joint of Silicon-Nitride to Stainless Steel, J. Mater. Eng. Perf., Vol 3 (No. 5), 1994, p 596–605 W.Y. Liu, S.W. Yao, and J.X. Qu, Brazing Si3N4 Ceramic to AISI-5140-Steel Under Pressure, J. Mater. Eng. Perf., Vol 5 (No. 2), 1996, p 209–212 Y. Ito, K. Kitamura, and M. Kanno, Joint
176 / Brazing, Second Edition
292.
293.
294.
295.
of Silicon Nitride and Molybdenum with Vanadium Foil, and Its High-Temperature Strength, J. Mater. Sci., Vol 28, 1993, p 5014–5018 J.H. Chen, G.Z. Wang, K. Nogi, M. Kamai, et al., The Metallurgical Behaviour During Brazing of Ni-Base Alloy Inconel 600 to Si3N4 with Ag71Cu27Ti2 Filler Metal, J. Mater. Sci., Vol 28, 1993, p 2933–2942 J.H. Selverian and S. Kang, Ceramic-toMetal Joints, Part II: Performance and Strength Prediction, Am. Ceram. Soc. Bull., Vol 71 (No. 10), 1992, p 1511– 1520 Y. Zhai, T.H. North, and J. Ren, Transient Liquid-Phase Insert Metal Bonding of Al2O3 and AISI 304 Stainless Steel, J. Mater. Sci., Vol 32, 1997, p 1399–1404 Y. Zhai, T.H. North, and J. Serrato Rodrigues, Transient Liquid-Phase Bonding of Alumina and Metal-Matrix Composite Base Materials, J. Mater. Sci., Vol 32 (No. 6), 1997, p 1393–1397
296. Y. Ito and T. Jinbo, Joining of Silicon Carbide and Molybdenum with Vanadium Foil, Mater. Trans., JIM, Vol 34 (No. 10), 1993, p 966–968 297. K. Suganuma, ISIJ Int. Ceram. Rev., Vol 30, 1990, p 1046–1058 298. A.K. Chattopadhyay and H.E. Hintermann, On Brazing of Cubic Boron Nitride Abrasive Crystals to Steel Substrate with Alloys Containing Cr or Ti, J. Mater. Sci., Vol 28, 1993, p 5887–5893 299. B. Derby, Zirconia/Metal Diffusion Bonds, Diffusion Bonding, R. Pearce, Ed., Cranfield, p 195 300. H. Ohmura, T. Yoshida, K. Kawashiri, and O. Yoshimoto, A Technique for Brazing Graphite/Graphite and Stainless Steel/High-Carbon Steel Joints, Weld. J., Oct 1994, p 249–256 301. M. Stöck and K. Hack, Thermochemical Aspects of Multiphase Diffusion During Brazing of Hard Metal, Z. Metallkd., Vol 84 (No. 11), 1993, p 759–766
Brazing Second Edition Mel M. Schwartz, p177-241 DOI: 10.1361/brse2003p177
Copyright © 2003 ASM International® All rights reserved. www.asminternational.org
CHAPTER 5
Brazing Filler Metals THIS CHAPTER REVIEWS key considerations related to brazing filler metals: their basic characteristics, selection criteria, types, and forms.
Basic Characteristics For satisfactory use in brazing, filler metals must possess certain basic characteristics. First, filler metals must have the ability to form brazed joints possessing suitable mechanical and physical properties for the intended application. This often means strength but may include ductility, toughness, electrical or thermal conductivity, temperature resistance, and stability. An extremely important physical property for intended filler metals is that they have a coefficient of thermal expansion that closely matches the substrates being joined or, where severe temperature gradients persist, bridges the difference in coefficients of thermal expansion between the two joint-element materials. This is so that thermal mismatch stresses across the joint do not cause failure by fracture. Second, the melting point or range of an intended filler metal must be compatible with the base materials being joined and any surface metallizations and must have sufficient fluidity at the brazing temperature to flow and distribute into properly prepared joints by capillary action. Suitable melting range means below the solidus of the base materials but as high as necessary to meet service operating-temperature requirements. Usually, a margin is required between these two temperatures in order to achieve adequate fluidity of the molten filler metal. Third, the composition of the intended filler metal must be sufficiently homogeneous and
stable that separation of constituents, known as liquation, does not occur under the brazing conditions to be encountered. Obviously, the intended filler-metal composition must also be chemically compatible with the substrates to avoid adverse reactions during brazing or by subsequent sacrificial (i.e., galvanic) corrosion. Fourth, intended filler metals must have the ability to wet the surfaces of the base materials being joined to form a continuous, sound, strong bond. Fifth, depending on requirements, intended filler metals must have the ability to produce or avoid reactions with the base materials. Usually, it is desirable to avoid such reactions, because brittle intermetallics may result, thus degrading joint properties. However, for socalled active-metal or reactive brazing, it is necessary for the filler metal and the substrates to react chemically in a particular way. Other characteristics include: • It must be capable of producing joints at temperatures at which the properties of the base materials are not degraded. For example, many work-hardened and precipitation-hardened alloys cannot withstand brazing temperatures without loss of their beneficial mechanical properties. The first type of hardening involves subjecting the alloy to mechanical deformation, such as rolling or hammering, when reasonably cold. As the temperature is raised, the deformation damage is removed by atomic rearrangement in the metal. Precipitation hardening is accomplished by creating a finely divided phase within the material, which can be thought of as akin to a composite material. The dispersed phase is precipitated by means of an appropriate heating schedule, and its strengthening effect is likewise degraded by exposure to high temperatures.
178 / Brazing, Second Edition
• It must wet the parent materials or a metallization applied to the parent materials in order to ensure good adhesion through the formation of metallic bonds. • It must not excessively erode parent metals or metallizations at the joint interface. The associated reactions, which must occur to form a metallic bond, should not result in the formation of either large proportions of brittle phases or significant concentrations of brittle phases along interfaces or other critical regions of the joint. Even ductile phases can have weak interfaces with solidified filler metals. It must not contain constituents or impurities that might embrittle or otherwise weaken the resulting joint. Likewise, the parent material must not contribute constituents or impurities to the filler metal that will have a similar effect (Ref 1–3). Besides being compatible with the parent material, the filler metal and the joining process used must be mutually suited. For example, filler metals containing zinc, lead, or other volatile constituents are not usually appropriate for furnace joining, especially when reduced pressures are involved. The degree of temperature uniformity that can be achieved over the joint area will have an influence in determining the minimum temperature difference that can be tolerated between the melting temperature of the filler metal and the melting or degradation temperature of the parent material. This consideration is particularly relevant to the joining of aluminum alloy components at approximately 600 °C (1110 °F) with the well-known Al-12Si near-eutectic filler metal that melts hardly more than 20 °C (36 °F) below this temperature. The condition of the surface of the parent material may affect its compatibility with the filler metal, especially when fluxes are not used. As an obvious example, an oxidized surface will be less readily wetted by a filler metal than will an atomically clean metal surface. This consideration often determines the acceptable shelf life of components prior to joining. In order to establish whether a particular parent metal (or nonmetal with a surface metallic coating) is compatible with a given filler metal, the appraisal must be carried out under conditions representative of those used in any practical implementation of the process. Parameters such as process time and temperature can be
critical in this regard. Storage shelf life of the filler metal and the components is another relevant factor but is often neglected during transfer of a process from the laboratory to the factory. The properties of the filler metal and the joint that it is used to make must also be compatible with the service requirements. These are likely to involve a combination of at least some of the following considerations: • The strength and ductility of the joint must meet certain minimum requirements over the range of service temperatures. • The design of the joint should not introduce stress concentrations in the assembly that might arise through solidification shrinkage or the formation of intermetallic phases. • The joint must be appropriate for the service environment in terms of corrosion and oxidation resistance and compatibility with vacuum, in accordance with functional requirements. • The filler metal must comply with statutory needs, such as health restrictions on lead and cadmium for certain culinary and medical applications. • Aesthetic requirements must sometimes be met—for example, color, color matching in jewelry and utensils, and the ability of joints to accept surface finishes such as paints, electroplatings, and so on. Good fillet formation is often demanded for aesthetic reasons and also as a criterion of acceptable joint quality. • Requisite thermal and electrical properties must be achieved. The simultaneous attainment of several of these desired characteristics is frequently achievable with common filler metals, provided that simple design guidelines and process conditions are satisfied. Melting and Fluidity. Pure crystalline metals and ceramics melt at a constant temperature, and molten metals, in particular, are generally very fluid. Alloys, on the other hand, whether metallic or ceramic, melt over a range of temperatures from the solidus to the liquidus for the particular composition and can have a fluidity that varies widely depending on the relative amounts of liquid and solid present at the brazing temperature. This “mushy” state always reduces the fluidity versus the fully liquid state. The wider the mushy range, the more sluggish the flow of the filler metal by capillary action.
Chapter 5: Brazing Filler Metals / 179
So, special care must be taken in selecting and employing alloy filler metals for brazing, to ensure proper fluidity. Most filler metals are designed to be more complex than simple binaries, which are often filler metals for the purpose of altering the liquidus-solidus ranges and phase proportions as much as for any other property. A filler metal will always have a liquidus that is below the melting point of the lowest-melting component of the filler metal and thus will always be suitable for brazing that component. Liquation. Because the compositions of the solid and liquid phases of filler metals differ due to of the distribution coefficient for the alloy solute, the proportion and composition of each phase will undergo gradual changes as the temperature increases from the solidus to the liquidus. If the portion that melts first is allowed to flow away from the remainder of the unmelted filler metal by capillary spreading, the remaining solid has a higher melting point than the original composition, never melts, and remains behind as a solid skull. This phenomenon is known as liquation. Obviously, such separation is undesirable. The tendency for liquation should be minimized in properly designed filler metals. This is accomplished by employing filler metals with narrow melting ranges and heating rapidly through the melting range during brazing. The optimal brazing temperature for a particular filler metal is usually approximately 10 to 90 °C (20 to 160 °F) above the liquidus of the filler metal. This superheat ensures flow without liquation. Wetting and Bonding. To be effective, a filler metal must alloy with the surface of the base material without undesirable degrees of diffusion into the base material, dilution by the base material, base-material erosion, or formation of brittle compounds at the interface. These effects are dependent on the mutual solubility between the filler metal and the base materials, the amount of filler metal present, and the temperature-and-time profile of the brazing cycle.
Filler-Metal Selection Criteria The following factors should be considered when selecting a filler metal, whether it is a metal or a ceramic. First, it should be compatible with the base material and the joint design. Compatibility with the base material means properly
matching chemical, mechanical, and physical properties. Compatibility with the joint design means proper mechanical properties for the type and magnitude of loading (e.g., static or fatigue; tension, shear, or peel). Second, the filler metal must be suitable for the planned service conditions for the brazed assembly, including service temperature, thermal cycling, life expectancy, stress loading, corrosive conditions, radiation stability, and vacuum operation (i.e., outgassing). Third, the filler metal must be selected based on the brazing temperatures required and must be acceptable to the components of the assembly and to the production environment. Low temperatures are usually preferred for economizing on heat energy, minimizing heat effects on the base material (e.g., annealing, grain growth, and warpage), and minimizing interactions (e.g., embrittlement by intermetallics). On the other hand, high temperatures are preferred to take advantage of higher-melting filler metals for their economy, to combine stress relief or heat treatment of the base material with the process of brazing, to promote interactions that will increase joint remelt temperatures, or to promote the removal of certain refractory oxides in vacuum or certain atmospheres. Finally, filler-metal selection depends on the method of heating to be used. Brazing filler metals with a narrow, that is, 30 °C (55 °F), melting range can be used with any heating method. Filler metals with wider melting ranges that are prone to liquation should be brought to brazing temperature quickly; therefore, processes with more intense heating are preferable.
Coatings Another technique used to preserve the surface of the prebrazed component and be free of oxides is to electrochemically metallize the joint surfaces prior to brazing. The process removes surface oxides and applies a thin, adherent coating to give a protected surface for brazing. An operator can direct a handheld stylus over the joint; the stylus holds a conductive fluid that electrochemically coats the work. No heat, no distortion, and no masking are involved. The possibility of using the new coating technologies in brazing opens many new avenues for heretofore difficult-to-join materials. By depositing one or more layers of filler metals on the joint surfaces, traditionally difficult-to-join materials can be brazed together. Barrier coat-
180 / Brazing, Second Edition
ings, coatings for dissolution-solidification, and reactive-metal coatings (interlayers) are three different groups of coatings. To control brazing, the optimal amount of coating applied must be determined, and the interfacial chemical reactions must be understood, with the characterization of the products that result in the filler metal. The brazing of advanced engineering materials, and especially dissimilar materials, is coupled to the effective use of coatings. The most obvious application is to use the coating process to establish a highly wettable surface for the liquid filler metal. The coating may also be used as a barrier when the filler metal is incompatible with the base material, to avoid diffusion of certain alloying elements and to prevent intermetallic formations. Recent approaches (Ref 4–6) to coating technology in brazing recognize that coatings are no longer just an intermediate passive barrier, but they take an active part in the brazing process. Mass transport, either from the filler metal into the base material or the dissolution of the coating into the filler metal, can promote physical changes, such as increasing the liquidus temperature of the filler metal, which significantly affects the brazing process. With proper selection of the coating material and filler metal, very consistent brazing processes can be developed. The control of the chemical composition and thickness of the coating layers are far more critical than that required in traditional filler metals. Application. There are several ways to apply coatings on parts to be brazed. They are: • Electrolytic: electroplating or electroless deposition • Thermal: dip coating, barrel coating, roller coating, flow melting, and so on • Surface modification: cladding, thermal spraying, plasma spraying, and so on • Physicochemical: sputtering deposition, vapor deposition, ion implantation, and so on Because both chemical homogeneity and thickness are important properties of a coating, the coating procedure must be well controlled. For components that require critical dimensional control, continuous sputter cleaning and vapor deposition processes are preferred because of the uniform adhesion obtained at the interfaces. Three different coating schemes are the primary methods used in brazing. They are barrier
coatings, reactive-metal coatings, and coatings for dissolution-solidification. Barrier Coatings. It has been common practice, when no known suitable filler metal is available for the materials to be joined, to use a coating that can be wetted by the liquid braze (Ref 7). The coating also serves as a barrier between the base metal and the braze during processing. A barrier layer must be dense, ductile, and free from defects such as voids. It must also show good adhesion and wetting toward both the base and filler metals. Copper-phosphorus filler metal had been used in the past to join ferrous alloys; however, intergranular penetration of phosphorus was observed to promote subsurface embrittlement of the joint. A thin undercoating of nickel reduced phosphorus diffusion and provided good wetting for the copper-phosphorus filler metal (Ref 7). Sometimes, the barrier coating scheme may require a double coating. Material for the first layer is selected for its compatibility with the base metal, and the second layer (overlayer or outer coating) is used to promote wetting with the liquid filler metals, as illustrated in Fig. 5.1. The primary layer is generally a transitionmetal solution containing polyvalent elements such as titanium, manganese, zirconium, tantalum, or molybdenum, where the polyvalent states increase the bonding tendency. The second layer is usually a thin layer of noble metal, which provides a surface that is oxide-free, or a very thin oxide layer, which readily decomposes or dissolves during brazing. Such oxygenfree surfaces have, in general, high surface energy and offer excellent wettability. Metals
Fig. 5.1
Schematic diagram showing the noble-metal and transition-metal barrier coatings inserted between the base metal and the filler metal. Source: Ref 7
Chapter 5: Brazing Filler Metals / 181
such as silver, gold, and copper serve this function, requiring little, if any, fluxing agent. The double-coating scheme will also allow for longer brazing times, because the thickness of the other coating can be adjusted. Reactive-Metal Coatings. The use of reactive metals in brazing has been shown to promote the formation of a thin interlayer between the filler metal and the base metal (Ref 8–11). This layer promotes adhesion between the faying surfaces of similar and dissimilar materials. Hence, the resulting strength of the braze is dependent on the nature of the product layer and its thickness. An example is the use of titanium in noble-metal-based filler metals for the joining of metals to ceramic materials. Titanium in the liquid filler metal reacts with the substrate to form a thin reaction layer. This product layer can be a complex oxide, such as Cu2Ti4O and Cu3Ti3O2, or an intermetallic compound, depending on whether the base material is an oxide ceramic or a metal. Certain types of complex oxides, for example, (Cr, Mn, Mg)O·(Cr, Mn, Al2O3, spinel), also promote adhesion between metal and oxides. The characterization and understanding of the crystal structure and microstructure of the product layers must be carried out so that adequate process control can be achieved (Ref 12, 13). Dissolution-Solidification Coatings. Some coatings are deposited for the purpose of dissolution and solidification. In contrast to the passive barrier coatings, these dissolution-solidification coatings are considered active because of their contribution to the filler-metal composition and its thickness. A good example is the transient liquid phase (TLP) bonding (Ref 14), which is a diffusion brazing process that combines the features of both brazing and diffusion welding (Ref 15–17). It uses as filler metal a thin interlayer or filler metal of specific composition and melting temperature. At the bonding temperature, the interlayer may melt, or a liquid may form by alloying between the interlayer metal and the base metal. The liquid, by capillary action, fills the joint clearance and contributes to the elimination of voids at the braze interface. While the joint members are held at the bonding temperature, diffusion of alloying elements occurs between the liquid and base metal. Isothermal solidification of the joint results because of the solute composition change in the braze. Maintaining the joint at the bonding temperature after solidification will promote further homogeniza-
tion of the chemical composition and microstructure. Solid-state diffusion of elements away from the interfacial region reduces the initially large chemical composition gradient, thus avoiding the formation of intermetallics at the braze. An element of high mobility, both in the liquid and solidified filler metal, will decrease the time for completion of the TLP process. Conventional TLP bonding is mostly applied to binary alloy systems that show some intermediate, low-temperature reactions, such as eutectic transformation. Another example is liquid interface diffusion, which is most appropriately used on titanium and titanium alloys, as well as activated diffusion bonding (Ref 2).
Filler-Metal Types There are literally hundreds of different fillermetal compositions available on the market. However, these fall into a relatively small number of alloy families, defined according to the major metal constituent. Only a few filler metals are based on eutectic systems. More commonly, the constituent elements form solid solutions, enabling a large number of brazes with different compositions to be developed from each alloy family. Within each alloy family, the differences in composition usually reflect only slight variations in user requirements. The filler metals will be treated on the basis of this classification, with attention devoted to delineating the principal features of each alloy system. For more detailed coverage of the available filler metals, the reader should consult reference publications (Ref 1–3) as well as data sheets and manuals supplied by manufacturers. The American Welding Society (AWS) has specifications for filler metals and lists eight categories of filler-metal types (Ref 1–3, 18, 19). These categories are described subsequently.
Aluminum-Silicon Filler Metals (Designated BAISi) Aluminum-silicon filler metals are used primarily for brazing aluminum and its alloys to themselves or various other metals and alloys. Because of the presence of oxides on aluminum and its alloys, these filler metals always require flux (Table 5.1). Joints to aluminum components tend to be more susceptible to corrosion than similar joints
9.3–10.7
11.0–13.0
9.0–11.0
9.0–11.0
11.0–13.0
9.0–11.0
BAlSi-4
BAlSi-5(d)
BAlSi-7
BAlSi-9
BAlSi-11(f )
0.25
0.25
0.25
0.30
0.30
3.3–4.7
0.25
Cu
1.0–2.0
0.10–0.5
1.0–2.0
0.05
0.10
0.15
...
Mg
0.20
0.20
0.20
0.10
0.20
0.20
0.20
Zn
0.10
0.10
0.10
0.05
0.15
0.15
0.10
Mn
0.8
0.8
0.8
0.8
0.8
0.8
0.8
Fe
559
562
559
577
577
521
577
°C
1038
1044
1038
1070
1070
970
1070
°F
Solidus temperature
596
582
596
591
582
585
613
°C
1105
1080
1105
1095
1080
1085
1135
°F
Liquidus temperature
582–604
582–604
588–604
588–604
582–604
571–604
599–621
°C
1080–1120
1080–1120
1090–1120
1090–1120
1080–1120
1060–1120
1110–1150
°F
Brazing temperature range
For furnace and dip brazing. Comes in sheet form and clad on core of 3003 or 6951(b) General purpose, for all processes. Limited flow General purpose, for all processes. More free-flowing than BAlSi-3. Aluminum-silicon eutectic. Comes as paste, rod For furnace and dip brazing at lower temperature than BAlSi-2. Comes as sheet or clad on 6951(b), paste For all processes, including vacuum brazing(e). Comes clad on 3003 or 6951(b) For vacuum brazing(e). Comes clad on 3003 core. Use to join fins of heat exchangers of 5xxx- and 6xxxv-series alloys Sheet clad on 3105 for vacuum service(e)
Comments
AWS, American Welding Society. (a) Principal alloying elements; balance aluminum. (b) This alloy can be solution heat treated and aged after brazing. (c) Contains 0.15% Cr. (d) Contains 0.20% Ti. (e) Solidus and liquidus temperature ranges vary when used in vacuum. (f) Contains 0.02–0.20% Bi
6.8–8.2
BAlSi-3(c)
Si
BAlSi-2
AWS classification
Composition(a), %
Table 5.1 Aluminum-silicon brazing filler metals
182 / Brazing, Second Edition
Chapter 5: Brazing Filler Metals / 183
between other common metals. This is because aluminum has an electrode potential that is more negative than most other metallic elements, and this can give rise to a galvanic corrosion problem. The necessary emphasis placed on corrosion and its prevention has resulted in the adoption of a few select aluminum- and zinc-base alloys as filler metals for joining aluminum alloy components. Aluminum-silicon binary alloy filler metals are available as wires, foils, paste, and clad sheets. The clad material comprises a sheet of a suitable aluminum engineering alloy, coated on one or both sides with filler metal. Each cladding typically constitutes up to 15% by thickness of the total thickness of the sheet. Clad materials are particularly suited for fluxless brazing processes; the joint interface between two mated components that have been roll clad will liquefy on heating above the solidus temperature, and this helps to displace the alumina layer and other surface films present. Lower-melting-point filler metals can be achieved by adding copper, magnesium, and zinc, which further depress the melting point of aluminum-silicon, due to the existence of ternary eutectic points in all three aluminum-silicon-X (X is copper, magnesium, or zinc) systems. Magnesium has the additional benefit of aiding in wetting. Aluminum alloy filler metals are perfectly satisfactory for joining metals such as stainless steels, molybdenum, and tungsten, provided that the different electrode potentials of these metals are recognized and appropriate measures are taken to minimize the risk of corrosion of the joint. Several problems are associated with the use of aluminum-base filler metals with aluminum alloy components, which represents their main area of application: • Their melting points are very close to those of most aluminum engineering alloys. For some of these alloys, the melting points of the available filler metals can actually exceed the solidus temperatures of the engineering alloys. For this reason, brazing cannot be used to join many of the wrought and casting alloys. • Aluminum has high solubility in these filler metals, resulting in extensive erosion of the parent metal. Consequently, thin-wall components with a thickness of less than approximately 0.5 mm (0.02 in.) cannot be joined easily by this means. The alloying increases
the melting point of the filler metal and tends to impede lateral flow and fillet formation by the molten braze. • They exhibit poor wetting on aluminum alloys when used without fluxes. This is a consequence of the high reactivity of aluminum with oxygen in the atmosphere and the refractory nature of the alumina that forms. • Corrosive aluminum fluxes produce residues that are water soluble and easily removed; they are colorless and inoffensive in appearance (but corrosive). The noncorrosive residues are typically heavy and white. They do not cause corrosion and are not removed. The noncorrosive fluxes do react with the base metal, but more flux is not needed; stronger fluxes are. Simply adding more flux insulates the part, causing more heat to be required. • Most engineering alloys of aluminum rely on precipitation hardening for their boosted mechanical properties (i.e., hardening through the presence of a finely divided second phase in the material). The temperatures required for brazing with the available filler metals are too high to be compatible with the heat treatment step that precedes precipitation of the second phase. There have been attempts to remedy this situation by developing new filler-metal compositions with substantially lower melting points. Recently, a series of new multicomponent aluminum filler metals containing silicon, copper, nickel, and (optionally) zinc as the principal additions has been developed for fluxless brazing processes in inert gas atmospheres at temperatures as low as 520 °C (970 °F) (Ref 20–22). Notwithstanding these difficulties, the fluxless brazing of aluminum has been successfully developed for the fabrication of radiators and heat exchangers of this metal and is now routinely employed for mass production in the automotive industry. The process, as established by VAW in Germany, is performed in either high vacuum or in a high-quality nitrogen atmosphere and uses an aluminum-silicon filler metal containing specified minor additions that promote spreading and fillet formation. Special cleaning procedures applied to component surfaces are an essential part of the process (Ref 1, 2, 23, 24). A new filler metal for brazing previously unbrazeable aluminum alloys was developed (Ref 21, 22). Most high-strength aluminum engineering
184 / Brazing, Second Edition
alloys cannot be joined by brazing, because they either degrade or melt at the temperature at which commercially available aluminum filler metals are used. Previous efforts to develop aluminum filler metals with a significantly reduced melting point have tended to be frustrated by poor mechanical properties of the alloys; corrosion of the joints; or the high cost, toxicity, or volatility of the constituent materials. The development and assessment of a new filler metal with a composition of 73Al-20Cu-2Ni-5Si (wt%), which has been designed to overcome these limitations, has been reported (Ref 21). A fluxless joining process was established for using the filler metal in inert gas and vacuum furnaces at temperatures down to 525 °C (980 °F). This produced well-filled joints with smooth, rounded fillets and minimal erosion of the base material. The braze could be prepared in the form of flexible foil either by chill-block meltspinning or by roll bonding layers of aluminumsilicon and copper-nickel alloys in an appropriate thickness ratio. The shear strengths of simple lap joints are comparable to those of base metals, depending on their specific composition and metallurgical condition. Joints made to engineering alloys of aluminum appeared to be endowed with an appreciable resistance to galvanic corrosion. The constituents of the braze process are low cost, and the filler metal can be readily produced as a foil suitable for preforms. Because of their high specific strength and satisfactory corrosion resistance, aluminum alloys belong to the group of fundamental structural materials in modern engineering. Their wide use has been made possible as a result of developing advanced methods of processing and producing permanent joints by welding or brazing. However, the application of brazing aluminum alloys has been limited because of the problems in removing the strong and chemically resistant oxide film. This film does not dissociate in the vacuum that can be achieved in brazing, does not dissolve in the parent material, and is not reduced by active gas media such as hydrogen or boron trifluoride used efficiently in brazing other metals and alloys. However, it was found (Ref 25, 26) that these problems can be overcome by using metallic coatings that themselves do not oxidize during heating in vacuum, and, when deposited, the oxide film is broken up and can be removed from the surface of the parent material. Coatings are deposited by electroplating and thermal vacuum spraying. With the former, the
removal of the oxide film from the surface takes place as a result of deposition of an intermediate layer (prior to electroplating) of a more electropositive metal than aluminum. With the latter method, direct contact of the metal of the coating with aluminum is ensured mainly by disrupting the oxide film by rapidly flying particles of the sprayed metal in glow discharge. The most promising method is to use metallic coatings in the form of individual components of the filler metal, which forms in contact melting of the deposited coatings with aluminum in heating for brazing. This mechanism of formation of a liquid filler metal is possible when the area of contact of the metal of the coating with the parent material is characterized, as a result of solution and diffusion processes, by the formation of an eutectic filler metal with a lower melting point than in the aluminum alloys being joined. This brazing method is referred to as contact-reactive brazing and is used widely for brazing steels, titanium, and aluminum alloys. With this method of brazing aluminum alloys, copper and silver are the most suitable as coating metals. They do not oxidize during heating in vacuum, and, with aluminum, they form relatively low-melting eutectic filler metals at a temperature of 548 and 558 °C (1018 and 1036 °F). The experimental results showed that in contact-reactive brazing aluminum alloys, the quality of brazed joints depended strongly on the coating method. The best results in regards to the strength of brazed joints, the spreading of the resultant filler metals, and the formation of brazed joints were obtained when depositing coatings by thermal vacuum spraying. With brazing with a coating deposited by electroplating, the resultant filler metal spread poorly over the surface of the parent material, and, consequently, the fillets of brazed joints were nonuniform and intermittent. The metallic coating deposited by thermal vacuum spraying protected the aluminum surface against oxidation during heating for vacuum brazing up to the moment of formation of the filler metal. When the filler metal forms, the oxide film does not manage to form, and even if it forms, it can easily be disrupted in spreading of the filler metal on the surface of the parent material as a result of its small thickness and density. This ensures direct contact of the brazed surfaces and the formation of brazed joints. The properties of the brazed joints were determined on lap specimens of several Russian
Chapter 5: Brazing Filler Metals / 185
aluminum alloys: (a) 1911 (28Al-72AgO, (b) AMts (45Al-27Ag-24Cu-4Si) (AMts alloy clad with aluminum with 8 to 9.5% Si), and (c) AMtsPS (67Al-33Cu). Copper and silver were deposited on the surface of the specimens by thermal vacuum spraying. Spraying experiments were carried out in vacuum equipment with specially developed devices for securing and moving the specimens during spraying. These devices made it possible to produce silver and copper coatings with a uniformity of 2 to 3 µm and a minimum consumption of metal. It was found (Ref 25) that when using AMtsPS alloy clad with silumin as the parent material, contact-reactive brazing through copper and silver coatings resulted in the formation, in the composition of brazed joints, of eutectic alloys with a lower silver content and a higher strength of brazed joints, in comparison with brazing with the copper coating. Contact-reactive brazing of aluminum alloys in vacuum greatly expands the technological possibilities of producing sections of aluminum alloys and markedly increases the quality and reliability of brazed joints. Researchers (Ref 27) investigated the influence of oxygen content on the brazeability of a powder aluminum filler metal and found that removal of fine particles with less than 30 µm diameter was the key to obtaining excellent powder aluminum filler metals.
Powder
The procedure followed by the researchers (Ref 27) is shown in Fig. 5.2. They found that: • The fillet formation test on a T-joint was effective for evaluating the brazeability of a powder filler metal in this brazing process. • Brazeability depended on the powder-production atmosphere that affects the oxygen content of powder. The use of powders with less oxygen content is effective for producing a sound fillet with low flux content. The use of inert powders atomized in a nonoxidizing atmosphere was preferable; the use of air powder was not recommended. • The oxygen content of the powder depended on the powder size. The decrease in powder size increased the oxygen content. The oxygen content increased remarkably with a decrease in powder size to less than 30 µm in diameter. Thus, removal of the fine particles with less than a 30 µm diameter was the key to maintaining high brazing quality with powder filler metal. • To obtain 100% fillet formability, the necessary flux mass of paste was 70 times the oxygen mass of powder filler metal under the present experimental condition. Other investigators (Ref 28) designed studies to investigate low-melting aluminum-base filler metals that had been alloyed with different elements to improve their flow and wetting proper-
Flux
Al-10%Si(BA4045,BAlSi-5) Atomizing gas: Air, N 2, Ar Atomizing atmosphere: Air, Ar Mean diameter: 2791µm (1.13.6 mil)
Mixing
Organic binder KAlF4-K3AlF6 (ALCAN NOCOLOK type) Solidus: 562 °C(1044 °F) Content: 110%vs. powder
Powder filler metal: 6367% Flux: 6.30.7% Organic binder: Bal.
Paste Brazing
Fig. 5.2
Atmosphere: Ar gas Temperature: 600 °C(1112 °F) Holding time: 300 s(5 min) Heating rate: 0.65 °C/s(1.2 °F/s)
Procedure to make aluminum powder brazing paste, and conditions for the brazing test. Nocolok (Alcan Aluminum Ltd.). Source: Ref 27
186 / Brazing, Second Edition
ties. The starting alloy for the base materials of Al-Cu-Zn-Mg and Al-Li was Al-35Ge-3Si, and for the alloys Al-Mg-Si and Al-Mg-Mn, it was the filler metal Al-20Ge-7.5Si. Researchers (Ref 29) selected an aluminum alloy from four different family groups, A5052, A6061, A7003, and A2017. In their findings and investigative developments with respect to A6061 and A7003, strengths nearly as high as the base material were obtained using Al-Ge-SiMg group filler metals, and brazed products using these filler metals have already been produced. However, there are limitations in brazing conditions (the necessity of holding and applying the load for a long time), and, in addition, filler metals are expensive. Evaluation of corrosion resistance in a variety of environments is yet to be performed. Further improvement will be required before it can be established as a technology with wide application. With respect to A2017, vacuum brazing has been regarded as impossible to apply, but it has become feasible with Al-Ge-Cr-Si-Mg group filler metals with low melting temperatures (Ref 30, 31). The development of two aluminum sheet filler metals with very high postbraze strength in combination with long-life corrosion performance was reported (Ref 32). Hogal-3571 was designed for vacuum brazing, with an ultimate tensile strength of 247 MPa (36 ksi) and elongation of 11.6%. Hogal-3572, for controlled atmosphere brazing, had an ultimate tensile strength of 212 MPa (31 ksi) and elongation of 15.1%. The magnesium and silicon form precipitates of Mg2Si during the brazing cycle, which results in higher postbraze strength. Copper provides solid-solution strengthening and also enhances the aging response. The long-life corrosion properties are caused by a diffusion layer formed during brazing, which is anodic to the inner core materials and acts as a sacrificial layer. Higher-strength alloys enable downgaging for lighter heat exchangers or the use of higher pressure for increased cooling efficiency. Potential applications include automotive heatexchanger parts, such as radiator tubes, headers, and side supports; evaporator plates; condenser tubes; and oil-cooler parts. The application of ceramics in technical systems often requires the joining of a ceramic to a metal or to itself. In recent years, there has been interest in using aluminum-base filler metals for joining oxide and nonoxide ceramics (Ref 33). Aluminum possesses the following favorable
properties: it is oxidation resistant, it wets a variety of ceramics, and it can relax residual stresses that result from the joining process. It has been shown that strong joints can be achieved in the systems Al-Al2O3, Al-SiC, Al-Si3N4, and aluminum/partially stabilized zirconia (PSZ) (Ref 34–37). The filler metals used in these studies were based on aluminum with magnesium, silicon, or copper additives. To join zirconium-base systems, one would expect that aluminum filler metals might offer some promise. It has been found that pure aluminum and aluminum-copper filler metals can wet calcia-PSZ surfaces, although temperatures in excess of 1100 °C (2030 °F) seem to be required (Ref 36, 37). Aluminum-zirconium alloys offer some interesting possibilities as a braze material. One would suspect that the zirconium addition to the aluminum may aid in wetting zirconium ceramics, and, in addition, these alloys can be dispersion strengthened (Ref 38, 39). In this preliminary study (Ref 39), the use of a two-phase filler-metal system for joining ceramics may offer some important opportunities. Control of the size, volume fraction, and distribution of the precipitate may allow the optimization of the joint ductility and strength. For the particular case of yttrium/tetragonal zirconia polycrystal (Y-TZP), it was found that aluminum-zirconium filler metals can be a useful braze material. The Y-TZP was joined with an Al-5.8Zr (wt%) filler metal at 900 °C (1650 °F) and above. Large precipitates of the intermetallic phase, Al3Zr, can aid in strengthening of the joint, especially if they are close to the interface. With decreasing layer thickness, the strengths increased, with values as high as 420 MPa (61 ksi).
Magnesium Filler Metals (Designated BMg) Magnesium filler metals are used for brazing magnesium and its alloys to themselves, and, because of the presence of oxides on magnesium and its alloys, they always require flux. Magnesium filler metal AZ92A (Mg-1) is used to join AZ10A, K1A, and M1A magnesium alloys by torch, dip, or furnace brazing processes. The other magnesium-base filler metal, AZ125A (BMg-2), with a lower melting range, is used for brazing AZ31B and ZE10A compositions (Table 4.5). Heating must be closely controlled to prevent melting of the base
Chapter 5: Brazing Filler Metals / 187
metal. Joint clearances of 0.10 to 0.25 mm (0.004 to 0.010 in.) are best for most applications. Corrosion resistance is good if the flux is completely removed after brazing. Brazed assemblies are generally suited for continuous service up to 120 °C (250 °F) or intermittent service at 150 °C (300 °F) subject to the usual limitations of the actual operating environment.
Copper (BCu) and Copper-Zinc (BCuZn) Filler Metals Copper and copper-zinc filler metals are widely used with both ferrous and nonferrous metals and alloys. They are, without question, the workhorse filler-metal system. As a group, they have limited corrosion resistance but are highly fluid. They may or may not require fluxes, depending on the base material being brazed. Cadmium-Containing Filler Metals. New cadmium standards were established in 1992 in revisions to Occupational Safety and Health Administration (OSHA) standard 29 CFR, parts 1910, 1915, and 1926; these new standards have made many users of cadmium-containing filler metals switch to alternate filler metals without cadmium. The toxicity of cadmium was established many years ago, and the trend since has been to switch to a non-cadmium-containing filler metal. The first cadmium-containing filler metal was patented in the early 1930s and is known today as BAg-1a of the AWS A5.8 specification. A tin-containing silver filler metal, known today as BAg-7, as well as a whole series of tincontaining filler metals have been developed. It has been well documented over the years that the addition of cadmium to the Ag-Cu-Zn ternary system acts as a temperature depressant, reduces the solidus-liquidus melt range, and
improves wetting (primarily on iron-base alloys). Each of these properties is a factor to consider when brazing, but possession of all three is seldom an absolute necessity in the majority of brazed assemblies. Non-Cadmium-Containing Filler Metals. The first step in selecting a non-cadmium-containing filler metal is to identify the characteristics or limitations of the assembly and process. For example, if the braze is restricted to an extended heating cycle because of the large mass of the components, then the preferred replacement selection is one with a lower melting temperature. If the fit-up or capillary spacing of the design requires a filler metal capable of filling a large joint clearance, then the replacement filler metal chosen should be of similar viscosity. If iron-base materials or carbides are being joined with a nickel-containing filler metal, the replacement filler metal should also contain nickel. Each assembly and process should be critiqued individually to establish the specific characteristics of the brazement before considering non-cadmium-containing filler metals. There are a considerable number of Ag-CuZn ternary filler metals from which to choose a potential alternative. A suggested starting point in the selection process is with a filler metal of similar silver content. Table 5.2 lists some of the more common AgCu-Zn filler metals. From this list, depending on the cadmium filler metal being used, a noncadmium filler metal can be selected with parameters best approaching that of the cadmium filler metal. The selection process will probably be influenced by the melt temperature and/or melt range and how it relates to present practice. Today, with technical improvements in heat sources, reaching brazing temperature is not a significant problem, whether torch, furnace, in-
Table 5.2 Silver-copper-zinc brazing filler metals Nominal composition, %
Solidus
Liquidus
Alloy
AWS A5.8 designation
Ag
Cu
Zn
°C
°F
°C
°F
A-1 A-2 A-3 A-4 A-5 A-6 A-7 A-8
... BAg-6 BAg-5 ... ... BAg-35 BAg-20 ...
60 50 45 40 40 35 30 20
25 34 30 36 30.5 32 38 45
15 16 25 24 29.5 33 32 35
675 675 660 670 675 680 675 715
1245 1250 1225 1235 1245 1260 1250 1315
730 775 745 770 725 755 765 815
1345 1425 1370 1415 1340 1390 1410 1500
AWS, American Welding Society. Source: Ref 40
188 / Brazing, Second Edition
duction, or any of the other methods of heating is used. Copper and Copper-Zinc-Tin. These filler metals are used to join ferrous and nonferrous metals. The corrosion resistance of the copperzinc alloy filler metals is generally inadequate for joining copper, silicon bronze, coppernickel alloys, or stainless steel. Typically, lap and butt joints are used with brazing processes. The practically pure copper filler metals are used to join ferrous metals, nickel-base alloys, and copper-nickel alloys. They are free-flowing and often used in furnace brazing with a combusted gas, hydrogen, or dissociated ammonia atmosphere without flux. Copper filler metals are available in wrought and powder forms (Ref 2). Research made it possible to understand, in general terms, the basic mechanisms involved in bond formation (Ref 41). At temperatures in excess of ~1000 °C (1830 °F) and in the moderate vacuum involved, it appears that the oxide film in the surface of stainless steel is no longer self-repairing and that when copper melts at 1083 °C (1981 °F), it is able to wet the metal surface. Heat treatments involving short times and relatively low temperatures produce essentially a brazed joint between the two surfaces, and if enough copper is present, the bond is void-free. At longer times and higher temperatures, diffusion of copper away from the bond line permits the asperities on the two surfaces to come into contact and allows the formation of diffusion bonds. The results of these initial experiments show that high-strength void-free bonds can be formed by vacuum brazing of stainless steels using copper filler metals. In Nitronic 40 (Armco Steel Corporation), brazed joints have been formed with strengths in excess of the yield strength of the parent metal, and even at liquid nitrogen temperatures, the excellent mechanical properties of the parent metal are only slightly degraded (Ref 41). Copper-zinc filler metals are used on steel, copper, copper alloys, nickel, nickel-base alloys, and stainless steel, where corrosion resistance is not a requirement. They are used with the torch, furnace, and induction brazing processes. Fluxing is required, and a borax/boric acid flux is commonly used. The copper-zinc filler metals are used extensively in braze welding of lowcarbon and low-alloy steels. The copper-zinc filler metals have melting and brazing temperatures higher than those of
silver filler metals. Overheating must be avoided, however, because of their high zinc contents. When these filler metals are overheated, zinc vaporizes (fumes), causing voids in the joint. Researchers (Ref 42) investigated the use of the vapor phase in the process of high-temperature brazing, which makes it possible to greatly expand its technological possibilities. Consequently, it is possible to deposit barrier and processing coatings and filler metals from the vapor phase, ensure activation of the brazed surfaces in atmospheres with low purity, remove easily evaporated components and prevent their evaporation from the filler metal, and produce filler metals by contact melting. The extensive use of copper-zinc filler metals for brazing components of steel and copper is determined by the relatively low melting point and high mechanical characteristics of the joint. However, as a result of evaporation of zinc during heating, the filler metal may melt incompletely, and defects may appear in the joint. An optimal variant of solving this problem is to introduce an additional amount of zinc. This makes it possible to prevent evaporation completely and to produce a filler metal during brazing as a result of contact-solid-phase melting of the initial blank of pure copper. The quality of brazed joints increases in this case. Methods have been developed for furnacecontainer brazing components of steel and dissimilar materials using two-component brasses with the vapor phase of zinc. Brazing with brasses containing alloying additions such as nickel, tin, iron, and manganese expands the range of brazed materials and increases the quality of joints. By introducing the alloying components into the content of the initial blank, it is possible to greatly increase the strength and other service parameters of the joint and to reduce the temperature to which the material is heated during brazing. Additions of aluminum, manganese, iron, and nickel improve the mechanical properties of brasses and their heat resistance. Brasses containing small additions of tin and silicon have higher processing properties and ensure a higher density and leaktightness of the joint (Ref 43). When alloying with nickel, the mechanical properties of brasses greatly improve. Copper-nickel-zinc filler metals containing 3 to 10% Ni are available (Ref 43). The filler metal with the composition 45 to 47% Cu, 43 to 45% Zn, and 10% Ni gives stronger joints than two-component brass (Ref 44).
Chapter 5: Brazing Filler Metals / 189
Taking into account the equilibrium diagram of alloys of the copper-zinc system and the equilibrium diagrams of the Cu-Ni-Zn, Cu-Sn-Zn, and Cu-Fe-Zn alloys, the recommended brazing temperatures with the given filler metals, and the temperature dependence of the tension of saturated zinc vapors, researchers proposed to restrict the temperature range of examining the saturation processes to 900 to 980 °C (1650 to 1800 °F). They concluded that: • Additions of nickel and tin have no marked effect on the saturation kinetics of brass with zinc from the vapor phase. • The microhardness of filler metals of the CuNi-Zn, Cu-Sn-Zn, and Cu-Fe-Zn systems, produced by saturation with zinc from the vapor phase, as well as the strength of the joints brazed with these filler metals greatly increase. • Multicomponent brasses can be used for vapor-gas brazing heavily loaded joints, such as compound hard-alloy tools operating under impact-loading conditions. In the development of the Ag-Cu-Zn-Sn quaternary system, it was discovered that tin performed similar to cadmium as a temperature depressant, contributed to a reduction in the solidus-liquidus range, and improved wetting (primarily with the iron-base alloys). Although tin acts similar to cadmium, it is not a direct replacement. The degree of effectiveness is slightly less, but in conjunction with a marginal increase in silver content, properties approaching those of the Ag-Cu-Zn-Cd filler metals can be achieved. Table 5.3 lists the more common tin-containing filler metals available today. At the top of the list is AWS A5.8 BAg-7, which has become the first choice of many when switching to non-
cadmium filler metal. This filler metal has a solidus temperature approaching those of the cadmium family and a liquidus temperature lower than most of the more common cadmiumcontaining filler metals. A review of Table 5.3 shows that as the silver content is lowered and the tin content altered, the melt temperature and range increase slightly. Combining Two Families. Although tin provides properties similar to cadmium, it should be noted that tin also has the effect of a slight reduction in ductility. The lowering in ductility is more apparent at elevated temperature. This is most noticeable when joint design places the filler metal under severe tension on cooling. Experience and tests show that within the Ag-Cu-Zn-Sn family, the ductility decreases with a decrease in silver content. If the joint design cannot be modified to place the filler metal under compression, and ductility is a factor, then a filler metal from the Ag-Cu-Zn family may be the proper choice. To assist in comparison of the two families, the more popular filler metals selected to replace cadmiumcontaining filler metals from each family are listed in Table 5.4. These are listed in descending order of silver content. As a further aid in the selection of an alternate filler metal, Table 5.5 lists comparative brazing properties of these filler metals. Before the selection of a replacement filler metal is made, each individual brazement should be analyzed and rated relative to the importance of each of the braze properties. The apparent economical selection will normally be the filler metal with the lowest silver content that still allows for cost-effective fabrication and yields a quality brazed joint. Silver-Copper-Zinc-Nickel. For the specialized application of brazing tungsten carbide tips to steel shanks for arduous service, alternate noncadmium, nickel-containing filler metals are available. These are listed in Table 5.6. The
Table 5.3 Silver-copper-zinc-tin brazing filler metals Nominal composition, %
Solidus
Liquidus
Alloy
AWS A5.8 designation
Ag
Cu
Zn
Sn
°C
°F
°C
°F
T-1 T-2 T-3 T-4 T-5 T-6
BAg-7 ... BAg-36 BAg-28 BAg-34 BAg-37
56 50 45 40 38 25
22 22 27 30 32 40
17 25 25 28 28 33
5 3 3 2 2 2
620 645 645 650 650 685
1145 1190 1195 1200 1200 1265
650 665 690 710 720 770
1205 1230 1270 1310 1330 1420
AWS, American Welding Society. Source: Ref 40
190 / Brazing, Second Edition
AWS A5.8 BAg-24 with a low liquidus temperature is the preferred choice. When improved strength is desired at elevated-temperature applications, BAg-4 should be considered. Present practice or procedures require very little change, if any, when selecting one of these alternate filler metals. To sum up the copper filler metals, see Table 5.7. These filler metals are grouped into copperphosphorus (“beecup”) types, copper, and copper-zincs. The BCuP types serve mostly for joining of copper-alloy base material. They work well with all brazing processes. On copper, they are self-fluxing; for other materials, they require a flux. These filler metals braze at temperature ranges that start below their liquidus. BCu, the pure-copper classifications, join steels, iron, nickel-base alloys, and coppernickel alloys by all brazing processes. The RBCuZn types, heavy in zinc, should be used only where outgassing or volatilization of the zinc can be tolerated. These filler metals offer only low resistance to corrosion. The “R” prefix indicates that these materials can also serve as filler rods for braze welding. Copper-Tin-Nickel-Phosphorus. The CuproBraze process for manufacturing copper automotive radiators has been shown to cut costs and increase productivity in more than a year of production, according to reports from the International Copper Association, which says that the process is cost-effective, fast, and easy to adopt (Ref 45). When the core is assembled, the completed radiator core is brazed as one unit, sealing the tubes at the same time. The assembly of the core requires either a fin-tip roller or a machine that applies filler-metal paste to the tubes. The core is then assembled and connected to the header plates. The tube-to-header joints are made
leakproof through an application of filler-metal slurry prior to brazing. The assembly is brazed at 620 to 650 °C (1150 to 1200 °F), approximately 300 °C (540 °F) below the melting temperature of copper. This results in less scrap than aluminum processes, which have only a 40 °C (70 °F) margin for error. With minor adjustments, a variety of heat exchangers of various sizes and masses can be fabricated. Manufacturers can braze different models in the oven simultaneously. One firm has 90 models in production. The process involves common vacuum and controlled atmosphere gas brazing furnaces and existing brazing equipment that is also suitable for aluminum. The CuproBraze technology is lead-free; the filler metal is 75Cu-15Sn-5Ni-5P. Copper-Manganese-Tin. A new copper filler-metal system has been developed as an alternative to the expensive silver-base filler metals with cadmium, which has been declared as a toxic material and is avoided in many brazing applications. Copper to mild steel (MS) and MS to itself have been successfully joined by a Cu-Mn-Sn system. Cu-11Mn-16Sn-1Ni produces high-tensile-strength joints of MS to MS at temperatures above 850 °C (1560 °F). Cu12Mn-19Sn and Cu-13Mn-20Sn-1Ce can be used to braze MS to MS at approximately 800 °C (1470 °F) with acceptable joint strength. Cu10Mn-30Sn is capable of joining copper to MS at approximately 750 °C (1380 °F). Although the addition of a small amount of nickel improves the MS-to-MS joint strength considerably, it also raises the effective brazing temperature. The addition of a small amount of cerium improves the flowability of the filler metal but lowers the joint strength, so a balance has to be sought. Copper-Titanium. Scientists at the Industrial Technology Center of Saga and the Osaka University Welding Research Institute (Ref 46)
Table 5.4 Silver-copper-zinc and silver-copper-zinc-tin brazing filler metals commonly used to replace cadmium-containing filler metals Nominal composition, %
Solidus
Liquidus
Alloy
AWS A5.8 designation
Ag
Cu
Zn
Sn
°C
°F
°C
°F
T-1 A-3 A-4 A-6 T-4 T-5
BAg-7 BAg-5 ... BAg-35 BAg-28 BAg-34
56 45 40 35 40 38
22 30 36 32 30 32
17 25 24 33 28 28
5 ... ... ... 2 2
620 660 670 680 650 650
1145 1225 1235 1260 1200 1200
650 740 770 755 710 720
1205 1370 1415 1390 1310 1330
AWS, American Welding Society. Source: Ref 40
Chapter 5: Brazing Filler Metals / 191
have jointly developed a filler metal for ceramic bonding with a bonding strength 60% greater than ordinary filler metals made of silver, copper, or titanium. The content of titanium in the new filler metal is reduced, which prevents brittleness and facilitates the joining of thin products such as foils. Their experiments found success by adding gold or zinc to a copper-titanium alloy, with the filler metal consisting of 85% Cu, 15% Ti, and a content of 10% Au and 10% Zn. With gold present, the bonding strength was increased by 60%, compared with available filler metals consisting of silver, copper, and titanium; with zinc present, the bonding strength was increased by 30%. Active metals reacting readily with ceramics have been used when producing filler metals for ceramic bonding, such as titanium. To produce a copper-titanium filler metal with high bonding strength, it was necessary to increase the content of titanium. For example, filler metal with 40% Ti and 60% Cu provided a bonding strength of 500 MPa (73 ksi) necessary for ceramic bonding. However, when the titanium ratio becomes larger, a reaction between copper and titanium creates an intermetallic compound that makes this new alloy brittle (Ref 46). Titanium added to active filler metal has been widely used as the filler metal for ceramics
Table 5.5 Comparison of brazing properties for filler metals used to replace cadmiumcontaining filler metals Alloy
AWS A5.8 designation
Temperature
Melting range
Flow
Wetting of ferrous metals
T-1 A-3 A-4 A-6 T-4 T-5
BAg-7 BAg-5 ... BAg-35 BAg-28 BAg-34
1 3 4,5 3,4 2 2
1 3,4 5 3 2 3
1 3 3,4 4 2 2
1 2,3 3,4 4,5 2 2
Rating scale: 1 = best selection. AWS, American Welding Society. Source: Ref 40
because of its beneficial effect on the wettability. However, a substantial amount of titanium is required in binary copper-titanium alloy to improve wettability (Ref 47, 48), and it causes excess interfacial reaction and formation of a brittle intermetallic compound, which may decrease the bond strength. The wettability of ternary Cu-Ti-X (where X is aluminum, silicon, or yttrium) filler metal is enhanced by an increase in titanium activity. In some cases, however, the surface energy of the melt, which depends strongly on the properties of alloying additions and their contents, also influences the wettability. Moreover, interaction of the alloying elements with titanium also greatly affects the titanium activity and, accordingly, the wettability and reactivity to Si3N4 (Ref 49). The addition of aluminum, which has a high interaction with titanium, decreases the titanium activity and hence increases the contact angle and decreases the reaction-layer thickness. The addition of silicon, which has a strong interaction with titanium, decreases the titanium activity greatly. Hence, the contact angle increases and the reaction-layer thickness decreases markedly, even with the addition of small amounts of silicon. Large amounts of silicon depress the melting point of the filler metal and reduce the surface energy of the melt, thus causing the contact angle to decrease significantly. The addition of yttrium, which has high reactivity with Si3N4 and low interaction with titanium, increases the titanium activity (or activity of reactive components) and accordingly decreases the contact angle. While the thickness of the reaction layer increases markedly when the yttrium content is less than 5 wt% (total content of yttrium plus titanium is less than 10 wt%), the reaction-layer thickness decreases sharply with the yttrium addition of more than 10 wt% (i.e., total content of yttrium plus titanium is more than 15 wt%). The addition of tin and silver, which have low interaction with titanium, also causes the tita-
Table 5.6 Noncadmium alloys for carbide brazing Composition, %
Solidus
Liquidus
Alloy
AWS A5.8 designation
Ag
Cu
Zn
Ni
°C
°F
°C
°F
C-1 C-2
BAg-24 BAg-4
50 40
20 30
28 28
2 2
660 660
1220 1220
710 780
1305 1435
AWS, American Welding Society. Source: Ref 40
bal
bal
bal
bal
BCuP-4
BCuP-5
BCuP-6
BCuP-7
4.8–5.2
1.8–2.2
14.5–15.5
5.8–6.2
4.8–5.2
...
...
... ...
...
...
... ...
...
Ag
AWS, American Welding Society
bal
56–60 46–50
RBCuZn-C RBCuZn-D
BCuP-3
56–60
RBCuZn-B
bal
57–61
RBCuZn-A
BCuP-2
99.0 86.5
BCu-1a BCu-2
bal
99.9
BCu-1
BCuP-1
Cu
AWS filler-metal designation
...
...
...
...
...
...
...
bal bal
bal
bal
... ...
...
Zn
...
...
...
...
...
...
...
0.8–1.1 ...
0.8–1.1
0.25–1.0
... ...
...
Sn
Table 5.7 Copper brazing filler metals
...
...
...
...
...
...
...
0.25–1.2 ...
0.25–1.2
...
... ...
...
Fe
...
... ...
...
Mn
...
...
...
...
...
...
...
0.01–0.50 ...
0.01–0.50
Composition, %
...
...
...
...
...
...
...
... 9.0–11
0.20–0.80
...
... ...
...
Ni
6.5–7.0
6.8–7.2
4.8–5.2
7.0–7.5
5.8–6.2
7.0–7.5
4.8–5.2
... 0.25
...
...
... ...
0.075
P
...
...
...
...
...
...
...
0.04–0.15 0.04–0.25
0.04–0.15
...
... ...
...
Si
645
645
645
645
645
710
710
865 920
865
890
1083 1083
1083
°C
1190
1190
1190
1190
1190
1310
1310
1590 1690
1590
1630
1981 1981
1981
°F
Solidus temperature
770
790
800
720
815
795
924
890 935
880
900
1083 1083
1083
°C
1420
1450
1475
1325
1500
1460
1695
1630 1715
1620
1650
1981 1981
1981
°F
Liquidus temperature
705–815
730–815
705–815
690–790
720–815
730–845
790–925
910–955 940–980
880–980
910–955
1095–1150 1095–1150
1095–1150
°C
1300–1500
1350–1500
1300–1500
1275–1450
1325–1500
1350–1550
1450–1700
1670–1750 1720–1800
1620–1800
1670–1750
2000–2100 2000–2100
2000–2100
°F
Brazing temperature range
Joins ferrous materials, nickel-base alloys, and copper-nickel alloys. Free-flowing Powder form of BCu-1 Copper oxide suspension form of BCu-1 For steels, copper and its alloys, nickel and its alloys, and stainless steel. All brazing processes. Requires flux For steels, cast irons, copper and its alloys, nickel and its alloys, and stainless steel. Silicon suppresses zinc fumes. ... For tungsten carbide, steel, and nickel alloys. Not suitable for furnace brazing Ductile and sluggish at brazing temperature; use for resistance brazing. Wide joint gaps, 0.08–0.13 mm (0.003–0.005 in.) Flows well. For gaps 0.03–0.08 mm (0.001–0.003 in.) Sluggish flow. For gaps 0.05–0.13 mm (0.002–0.005 in.) Flows well. For gaps 0.03–0.08 mm (0.001–0.003 in.) Sluggish flow. For gaps 0.05–0.13 mm (0.002–0.005 in.) For gaps 0.05–0.13 mm (0.002–0.005 in.). Sluggish at 730 °C (1350 °F); flows well at 815 °C (1500 °F) Slightly sluggish flow. For gaps 0.05–0.13 mm (0.002–0.005 in.)
Characteristics
Chapter 5: Brazing Filler Metals / 193
nium activity to increase and hence the contact angle and reaction-layer thickness to decrease. The bond strength of the copper-titanium filler metal is affected more by the reactionlayer morphology rather than by the wettability. The effect of alloying elements on the wettability, interfacial reaction, and bond strength was very different. The aluminum addition up to 10 wt% reduced the reaction-layer thickness and increased the strength remarkably, irrespective of the variation of wettability. Therefore, aluminum is regarded to be very effective for the control of interfacial reaction. Because the silicon addition reduces the reaction-layer thickness and greatly decreases the strength, silicon is considered to be ineffective as a reaction-control element. The yttrium addition tends to promote the interfacial reaction by titanium and decreases the contact angle and the bond strength. However, the yttrium addition of less than 5 wt% to the copper-titanium filler metal containing less than 5 wt% Ti is very effective for the marked improvement of wettability and strength (Ref 49). Copper-Tin. An 80Cu-20Sn filler metal, available as foil and paste, is capable of brazing ferrous and nonferrous base metals by torch brazing with flux assist and by furnace brazing in nitrogen, argon, or vacuum (0.13 Pa, or 2 × 10–5 psi). Copper-Phosphorus Filler Metals (Designated BCuP). Copper-phosphorus filler metals are good for brazing copper and copper alloys to themselves; however, they tend to liquate. What is the effect of phosphorus when alloyed with copper? The following are some of the major effects: • It lowers the melt temperature of copper (a temperature depressant). • It increases the fluidity of the copper when in the liquid state. • It acts as a deoxidant or a fluxing agent with copper. • It lowers the ductility of (embrittles) copper. Listed in Table 5.8 are the most common filler metals of the BCuP family used today. A comparison of the nominal silver and phosphorus content shows that while the silver varies from 0 to 18%, the phosphorus range is much narrower (5.0 to 7.25%). Control within the narrow phosphorus range is very critical during the manufacturing and application of these filler metals.
A major advantage of the copper-phosphorus filler metals is the self-fluxing characteristic when joining copper to copper. They may also be used with the addition of a paste flux on brass, bronze, and specialized applications on silver, tungsten, and molybdenum. The selffluxing characteristic in joining copper to copper is related to the phosphorus element possessing a greater affinity for oxygen than it does for the element copper. Also, the phosphorus will combine with the oxygen in the air and in the flame if an oxyfuel torch is used. In addition to oxidation of the phosphorus during application, the phosphorus will also diffuse or amalgamate with the surface of the parts being joined. Any loss in phosphorus, if it is by diffusion or oxidation, results in an alteration to the properties of the remaining liquid as it flows into the joint. In most cases, the change reflects a more viscous liquid with a higher melt temperature. In the application of the BCuP filler metals, whether it is wire, rod, paste, or preform, care must be taken to minimize any overoxidation of the phosphorus element. A slight change in the phosphorus content has a significant effect on the liquid filler-metal flow characteristics (Ref 50). These filler metals are primarily used to join copper and copper alloys. They have some limited use for joining silver, tungsten, and molybdenum. They should not be used on ferrous or nickel-base alloys nor on copper-nickel alloys with more than 10% Ni. These filler metals are suited for all brazing processes. Corrosion resistance is satisfactory, except where the joint is exposed to sulfurous atmospheres at elevated temperatures. Brazed assemblies can generally be subjected to continuous service temperatures up to 150 °C (300 °F). Short service at 200 °C (390 °F) may be permissible, depending on the operating environment. However, flux is recommended with all other metals, including copper alloys. Lap joints are recommended, but butt joints may be used if requirements are less stringent. Joint clearances should be 0.03 to 0.13 mm (0.001 to 0.005 in.). The range of clearance depends on the fluidity of the particular filler metal. A new family of copper-phosphorus and copper-tin filler metals has recently been developed by rapid solidification techniques (Ref 47). Both the 78Cu-10Ni-8P-4Sn and 77Cu-10Sn7P-6Ni filler metals are silver- and cadmiumfree, thus enhancing brazing safety and lowering
194 / Brazing, Second Edition
cost. They are potential replacements for Cu-15 Ag-5P (BCuP-5) and 45Ag-15Cu-16Zn-24Cd (BAg-1) in brazing copper and its alloys. Both filler metals are available as rod, foil, and powder; are self-fluxing; are suitable for both torch and furnace brazing; and have brazing temperatures similar to that of the BAg-1 filler metal and lower than that of the BCuP-5 filler metal. Copper-to-copper brazed joint mechanical properties—tensile, shear, and impact strengths—are either similar or superior to those of BCuP-5 and BAg-1 brazed joints. Specialized Copper Filler Metal (CopperSilver-Silicon-Tin). Scientists at the National Metallurgical Laboratory (Ref 51) developed a copper-base filler metal with suitable additives. Major applications have been envisioned in the electronics and vacuum tube industries. The developed filler metal with only half the quantity of silver is less expensive than the standard silver-copper eutectic filler metal and possesses comparable brazing characteristics. In the currently developed copper-base filler metal (Cu-36.20Ag-3.10Si-0.15Sn, wt%), the melting point, fluidity, and wetting properties are as good as in the conventional silver-copper eutectic filler metal (72Ag-28Cu). The new filler metal possesses higher strength and good workability and hence should be amenable to forming foils or wires industrially. The mechanical properties of the brazed joints (copper to copper) using the new filler metal are comparable to those of eutectic brazed joints. The corrosion properties of the brazed joints (copper to copper) using the new filler metal are much better than the eutectic brazed joint. Thus, the new filler metal, with 50% savings in silver,
should be a potential substitute for the conventional 72% Ag and 28% Cu eutectic filler metal.
Gold Filler Metals (Designated BAu) Very expensive gold filler metals are used for brazing parts in jewelry, electronic assemblies, and vacuum tubes, where volatile components are undesirable and where electrical conductivity must be high. They offer excellent corrosion resistance with iron, nickel, and cobalt alloys. Gold-Copper and Gold-Nickel-Palladium. Gold filler metals are commonly used on thin sections because of their low rate of interaction with the base metal (Ref 52–55). Applied under a protective atmosphere, goldcopper and gold-nickel-palladium filler metals need not be fluxed; however, for certain applications, a borax/boric acid flux may be used. The gold-base filler metals are generally suitable for continuous service at 425 °C (800 °F) and intermittent service at 540 °C (1000 °F), depending on the operating environment. Normally, joint clearances of 0.03 to 0.10 mm (0.001 to 0.004 in.) are used. Several of the gold filler metals, BAu-1, -2 and -3 (Table 4.7), have different temperatures and are very popular for use in step brazing. The significance of step brazing is that when two or more joints are to be brazed in sequence at points not widely separated on the same assembly, filler metals that differ in working temperature are selected, and the higher-melting filler metal is used first. Some overlap of brazing temperature ranges can be tolerated, depending on the work metal, the size of the parts, the proximity of the joints, the presence or absence of heat sinks, and the closeness of temperature control (Ref 56–58).
Table 5.8 The most commonly used copper-phosphorus (BCuP) brazing filler metals Content, % Alloy
A B C D E F G H I J K
AWS classification
BCuP-2 ... ... BCuP-6 ... BCuP-7 BCuP-3 BCuP-4 ... BCuP-5 ...
Solidus
Liquidus
Flow point
Ag
P
Cu
°C
°F
°C
°F
°C
°F
0 0 1.0 2.0 2.0 5.0 5.0 6.0 6.0 15.0 18.0
7.25 6.8 6.0 7.0 6.5 6.75 6.0 7.25 6.0 5.0 6.25
bal bal bal bal bal bal bal bal bal bal bal
710 710 645 645 645 645 645 645 645 645 645
1310 1310 1190 1190 1190 1190 1190 1190 1190 1190 1190
795 800 815 790 795 770 810 720 810 800 670
1460 1470 1495 1450 1460 1420 1490 1325 1485 1475 1235
730 730 720 705 730 705 720 690 720 705 660
1350 1350 1325 1300 1350 1300 1325 1275 1325 1300 1220
AWS, American Welding Society. Source: Ref 50
Chapter 5: Brazing Filler Metals / 195
Two recently developed gold-base filler metals should also be mentioned. These filler metals (19Au-7Ni-6Pd-25Mn-43Cu and 30Au10Ni-10Pd-16Mn-34Cu) have lower gold contents and therefore are far less expensive than the other gold-base filler metals listed in Table 4.7. They are produced in wire, foil, and powder forms and are used for brazing in the temperature range from 960 to 1010 °C (1760 to 1850 °F). They are much stronger than other high-temperature filler metals of their type, provide better gap-filling capabilities and excellent wetting, exhibit nonaggressive behavior, have good oxidation and salt spray resistance, and are not embrittled in hydrogen environments. These filler metals were developed for vertical tube-to-tube brazing in the Space Shuttle main engine flight nozzle, but they can be used for many nonspace applications. Other Gold Filler Metals. The Cu-Au-Ni filler metals are used for metal-ceramic brazes in electronic assemblies because of their good wetting characteristics and low vapor pressure. Researchers (Ref 59) studied the tensile creep properties of annealed 62Cu-35Au-3Ni filler metal over the temperature range of 250 to 750 °C (480 to 1380 °F). The minimum creep rate behavior of the 62Cu-35Au-3Ni filler metal was compared with those of the 74.2Cu-25.8Au filler metal and pure copper. This comparison indicated that the 62Cu-35Au-3Ni has considerably higher creep strength than pure copper. This fact suggests that the 62Cu-35Au-3Ni filler metal can be used in low-mismatch metal-to-ceramic braze joints, such as molybdenum to metallized alumina ceramic, with few problems. However, careful joint design may be essential for the use of this filler metal in high-thermal-mismatch metal-toceramic braze joints. Researchers (Ref 60) examined the intermediate-temperature joining of dissimilar metals. They joined duplex stainless steel and Ag-Ni-Ag laminates to copper with gold-germanium filler metal. They determined that they could bond copper to silver and nickel surfaces between 450 to 550 °C (840 to 1020 °F) in vacuum by an intermediate melting process. The selected filler metals were Au-12Ge and Au-18In. The Au-12Ge filler metal typically yielded better wetting results than the Au-18In filler metal. Nickel plating of duplex stainless steel facilitated wetting by both gold filler metals. The interfacial reaction product was either Ni3Ge for the Au-12Ge filler metal or Ni3In for the Au-
18In filler metal. Direct wetting to the stainless steel surface was not possible without the nickel-plated layer. Bulk filler-metal tensile data revealed that Au-18In has higher tensile strength and less ductility than Au-12Ge. The creep behavior of the Au-12Ge filler metal also suggests that it would be more effective in relaxing residual stresses generated during thermal processing under relatively low applied creep loads and temperatures. Because the joint design is comprised of dissimilar metals having different thermal expansion properties, the more ductile Au-12Ge filler metal was chosen for the fabrication of prototype assemblies. Finally, prototype test specimens, oxygenfree high-conductivity copper rings and Ag-NiAg- or nickel-plated duplex stainless steel wheels, were joined with Au-12Ge at 450 °C (840 °F). The specimens yielded a mean shear strength of 160 MPa (23 ksi). Failures occurred by plastically deforming the copper ring. The Au-12Ge filler metal offers an alternative, intermediate solution to higher-melting filler metals, which can affect base-metal properties. Researchers (Ref 61) conducted brazing studies on polycrystalline alumina specimens in order to understand some of the factors that determine the mechanical and microstructural behavior when using Au-Ni-Mo-V filler metals. The following conclusions can be drawn from the results of this work: • The Au-15.5Ni-0.7Mo-base filler metals containing 1, 2, or 3% V exhibit limited wetting on either 94 or 99.8% alumina substrates at temperatures of 1000 and 1020 °C (1830 and 1870 °F) for 5 min in a dry hydrogen atmosphere. All filler metals showed a preferential tendency to wet the glassy phase found in the 94% alumina specimens. • Specimens made of 94% alumina showed consistently high strength and acceptable hermeticity after being brazed with the three active-braze-alloy (ABA) filler metals at the two brazing temperatures. Tensile fractures generally occurred in the ceramic substrate (Table 5.9). • High-purity 99.8% alumina specimens yielded hermetic joints when brazed with the 2 and 3% V filler metals. However, their joint strength was generally lower than comparable 94% alumina brazed specimens. The 1% V-containing filler metal yielded good tensile values, intermediate to the 1 and 3% V filler
196 / Brazing, Second Edition
metals, but produced significantly more joint leaks. The 99.8% alumina specimens brazed with the 2% V filler metal yielded the poorest tensile properties. Only the 3% V composition yielded joint strengths approaching the 94% alumina values. The fracture paths in the 99.8% alumina specimens occurred across the metal-ceramic interface. • Differences in joint strength between the two types of alumina can be attributed to the presence of a greater amount of glassy grainboundary binder phase present in the 94% alumina ceramic. New filler metals, based on the Au-Ni-Cr-Fe system, were developed and tested at room temperature and at 650 °C (1200 °F) for ceramicmetal brazed joints in ceramic heat engines (Ref 62). The two filler-metal systems developed were approximately Au-33–35Ni-3–4.5Cr-1–2Fe-1– 2Mo (SK-1) and Au-34–36Ni-4–5.5Cr-2–3Fe (SK-2) (wt%). These filler metals showed superior wetting and atomic bonding characteristics as well as excellent ductility, compared to the other compositions studied. These two systems were able to satisfy the requirements of high-temperature performance at 650 °C (1200 °F) for ceramic-metal joints in ceramic heat-engine applications. In the design of new filler metals, the microstructural criterion of both solid-solution and dual-phase strengthening was employed for high-temperature filler metals. Also, an effort was made to balance the property requirements of the filler metals for the ease of manufacture and hightemperature properties. The new filler metals provided superior joint performance at high temperature, as compared to conventional solid-solution filler metals. The SK-1 filler metal resulted in a significant gain in the torsion strength at 650 °C (1200 °F) for PY6-nickel-Incoloy 909 (Special Metals Corp.) joints, with a moderate loss at room temperature Table 5.9 Gold-base active brazing filler metals Content, wt% Alloy
Baseline 1 2 3
Melting range
Au
Ni
Mo
V
°C
°F
82.0 82.8 81.8 80.9
18.0 15.6 15.7 15.5
... 0.7 0.7 0.7
... 1.0 1.8 2.9
955 949–958 940–960 953–958
1750 1740–1755 1725–1760 1747–1755
Source: Ref 61
compared with baseline Au-Pd-Ni joints. The average SK-1 joint strength was 35.8 N·m at 650 °C (1200 °F), whereas the strength at room temperature was 43.1 N·m. The high-temperature torsion strength was far better than the Au5Pd-2Ni filler metal system (1.6 to 7.7 N·m). The creep performance of these joints at 650 °C (1200 °F) was outstanding. The rupture life of the SK-1 braze joint exceeded 160 h at the 20.9 N·m torque level. Most rupture failures occurred at the interface between the braze and PY6. Excellent performance was obtained when the atomic bonded area exceeded 80%. In addition, excellent room-temperature mechanical fatigue properties, as well as thermal fatigue resistance, were noted for the SK-1 joint at a fatigue amplitude of 3.9 to 20.9 N·m. The significant improvement in high-temperature torsion strength and creep performance is ascribed to the dual-phase microstructure of the SK-1 filler metal. The microstructures of the joints made of the filler metals show two discrete gold- and nickel-rich phases, which have different melting points. A patent (Ref 63) was issued for a gold-base filler metal in paste form for use with high-performance ceramics. The use of such ceramics, with their favorable properties in respect to hardness, wear and corrosion resistance, and electrical resistivity in high-technology applications, depends very much on the ability to reliably join: • Simple shape components to form complex assemblies • Unit length of material to form large systems • Ceramic components to metals For such bonds, active-metal brazing has been developed. With this technique, the wetting and bonding of the braze material is improved by the presence of small amounts of highly reactive metals such as titanium or zirconium. Preferred materials contained in the filler metal are 85 to 90Au, 0.5 to 7Ni, 0.5 to 6V, 0.25 to 4Mo, and 0.3 to 5Cr (wt%). In addition, relatively large amounts of vanadium can be added to the basic gold-nickel filler metals, which nevertheless maintain good ductility. Ductility can be further enhanced by adding small amounts of molybdenum. The corrosion and oxidation resistance of the filler metals is improved by minor amounts of chromium.
Nickel Filler Metals (Designated BNi) Nickel filler metals are used for corrosion and heat resistance (up to 980 °C, or 1800 °F, con-
Chapter 5: Brazing Filler Metals / 197
tinuous; 1200 °C, or 2190 °F, short term). They are excellent in vacuum systems and are the usual choice for use with nickel-base alloys. Nickel-base filler metals are more extensive in composition range and properties than even the silver-base filler metals. Their primary merit is the ability to endure high-temperature service (above 1100 °C, or 2010 °F), even in moderately aggressive environments. These properties can be largely attributed to the corrosion resistance of elemental nickel and its relatively high melting point (1455 °C or 2650 °F). Pure nickel is not widely used as a filler metal because of its high melting point, except for certain specialist applications, such as the joining of molybdenum and tungsten components intended for subsequent operation at elevated temperatures. The melting point of nickel is depressed by alloying additions of an appreciable number of elements (e.g., boron, carbon, chromium, copper, manganese, phosphorus, sulfur, antimony, silicon, and zinc), many of which are introduced in combination to produce the variety of nickelbearing filler metals that are commercially available. Of these, phosphorus and boron are particularly effective in low concentrations at promoting the wetting characteristics of nickelbase filler metals. One of the most widely used nickel-base filler metals, BNi-1, has the composition Ni-3B0.7C-14Cr-5Fe-5Si and melts in the range of 977 to 1038 °C (1791 to 1900 °F) (Tables 4.6; 4.7, and 4.8). Nickel filler metals are generally used on 300- and 400-series stainless steels, nickel- and cobalt-base alloys, and even carbon steel, lowalloy steels, and copper, when specific properties are desired. They exhibit good corrosionand heat-resistance properties. They are normally applied as powders, pastes, rod, foil, or in the form of sheet or rope with plastic binders. Nickel filler metals have the very low vapor pressure needed in vacuum systems and vacuum tube applications at elevated temperatures. The filler metals that contain nickel as their principal alloying element are most important from the standpoint of high-temperature service. The number of nickel-base filler metals with significant differences in their compositions is confusing to one who must select a filler metal for a specific application (Tables 4.6, 5.10). However, these filler metals vary extensively in their physical, mechanical, and metallurgical properties. These variations are intentional, because it is necessary to provide filler
metals that meet varying requirements (Ref 64, 65). Depending on composition, the temperatures at which joints made with the filler metals listed in Table 5.10 are resistant to oxidation vary from approximately 540 to 1095 °C (1005 to 2005 °F). Some nickel-base filler metals are extremely fluid and are used to braze long, close-fitting joints where the filler metal is distributed by capillary attraction. Others are moderately sluggish in their flow properties and must be preplaced on the faying surfaces of the joint to ensure consistent brazing. Still others are extremely sluggish when they are molten; they are used to braze joints whose clearances cannot be controlled readily (e.g., joints such as those encountered in sheet metal assemblies). Certain nickel-base filler metals have been developed for nuclear applications that contain elements resulting in a low-capture cross section to thermal neutrons; others are suitable for exposure in high-temperature water or in liquid metals or their vapors. Nickel-base filler metals developed for fabricating honeycomb sandwich structures have limited reaction rates with thin metals used in such structures. Nickel-base filler metals have not only the attributes discussed previously but also disadvantages and limitations that must be recognized to ensure proper selection. Some react severely with heat-resistant structural alloys; if the brazing conditions promote prolonged reactions of this nature, erosion and/or penetration of thinmetal sections may occur. The diffusion of certain alloying elements, such as boron (and silicon to a lesser degree), into the grain boundaries of the base metal can result in the production of joints with poor mechanical properties. Also, some nickel-base filler metals produce joints that lack ductility. To a large degree, the unfavorable characteristics of nickel-base filler metal (and other filler metals used for high-temperature brazing) can be limited or minimized by proper control of brazing variables (Ref 66, 67). These nickel-base filler metals not only provide oxidation and corrosion resistance but also are suitable for subzero applications down to liquid helium temperatures (–270 °C, or –455 °F). The nine major classes of nickel filler metals are listed in Table 4.7. Improvements in nickel-base filler metals for specialized applications have also been made. For example, several filler metals containing phosphorus instead of boron as a melting-point depressant have been developed for applica-
B
4776
4777
4778
4779
4782
...
...
... ... ...
...
BNi-1a
BNi-2
BNi-3
BNi-4
BNi-5
BNi-6
BNi-7
BNi-8 BNi-9 BNi-10
BNi-11
4.0–5.0 4.0–5.0 2.5–3.5 0.5 1.5 ... ... 0.2 ... 1.5 3.5 3.5
Fe
...
B50TF94 B50TF207 B14Y10
...
...
B50A820, B50TF81, B14Y3
B50TF26, B50TF206
B50TF84, B50TF205
B50TF204
...
...
General Electric
4.0–5.0 4.0–5.0 4.0–5.0 4.0–5.0 3.0–4.0 9.75–10.50 ... 0.10 6.0–8.0 ... 3.5 3.5
Si
P
0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 ... ...
S
0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 ... ...
Al
...
... 54752-VIII 54752-III
...
54752-XI
54752-V, 94782
54752-X, 94779
54752-I, 94778
54752-II, 94777
54752-XIII
54752-IV
Allied Signal (EMS)
56623
... ... 56625
...
...
56630, 56635
...
56600
56610
71005
71000
Allison (EMS)
Specification cross reference
0.6–0.9 0.02 0.06 0.02 0.06 0.02 0.06 0.02 0.06 0.02 0.10 0.02 0.10 10.0–12.0 0.08 9.7–10.5 0.10 0.02 0.06 0.02 0.5 ... 0.4 ...
C
Composition(a), % Cu
9500/703
... 9500/719 ...
...
9500/707
9500/116
9500/700
9500/114
9500/97
9500/705
9500/103
0.05 0.05 0.05 0.50 0.05 0.05 0.05 0.05 0.05 0.05 ... ...
Zr
bal bal bal bal bal bal bal bal bal bal bal bal
Ni
0.50 0.50 0.50 0.50 0.50 0.50 0.50 0.50 0.50 0.50 (d) (e)
Other(b)
...
... 36962 693
...
36100
...
...
...
...
996
...
171
... 150 170
50
10
30
135
130
LM
LC/LCP
125
Pratt & Whitney Wall Colmonoy (PWA) Corp. (Nicrobraz)
... ... ... ... ... ... ... ... ... ... ... ... ... ... 0.04 ... 21.5–24.5 4.0–5.0 ... ... ... ... ... ...
Mn
Rolls Royce (MSRR)
0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 ... ...
Ti
1790 1790 1780 1800 1800 1975 1610 1630 1800 1930 1780 1780
°F
1035 1075 1000 1035 1065 1135 875 890 1010 1055 1105 1095
°C
°C
1950–2200 1970–2200 1850–2150 1850–2150 1850–2150 2100–2200 1700–2000 1700–2000 1850–2000 1950–2200 2020–2200 2100–2200
°F
Brazing temperature range
1065–1205 1075–1205 1010–1175 1010–1175 1010–1175 1150–1205 925–1095 925–1095 1010–1095 1065–1205 1105–1205 1150–1205
Applications
1900 1970 1830 1900 1950 2075 1610 1630 1850 1930 2020 2000
°F
Liquidus temperature
General-purpose material used for turbine blades, jet engine components General-purpose; tougher than Nicrobraz BNi-1. Recommended for aircraft, missile, food-processing equipment, nuclear reactor components outside the core area. Good for torch brazing General purpose, with narrow melting point and high flowability; for jet engine diffuser components General purpose; lower sensitivity to atmospheres, compared to BNi-2. Used for missile guidance system tubes, honeycomb assemblies Wide melting range; low diffusion. Used for thin honeycomb structures, food-processing pump impellers, crack repair in glass molds. Also recommended for torch brazing Boron-free; high temperature; suitable for nitrogen-base atmospheres. Recommended for honeycomb assemblies, heat exchangers, and catalytic convertors Very fluid; generally insensitive to atmospheres; for aircraft compressors or induction brazing on aircraft spark plugs Similar to Nicrobraz 10, with better high-temperature corrosion resistance; for nuclear reactor grids, radiators, honeycomb sandwich panels, boiler finned-tube assemblies Brazing of nickel alloys and stainless steels; honeycomb assemblies Used for diffusion brazing and nozzle assemblies Extra strength at higher temperatures. Use with base metals containing Co, W, and Mo. Recommended for rocket engine rotor turbine assemblies Similar to BNi-10, but for wide gaps. Used for thin-walled combustion chambers
975 975 970 980 980 1080 880 890 980 1055 970 970
°C
Solidus temperature
AWS, American Welding Society. AMS, Aerospace Material Specification. (a) Single values are maximum percentages, unless otherwise indicated. All BNi alloys have a limit of 0.10 Co and 0.005 Se. (b) Other elements, total. (c) Nominal composition. (d) Contains 16% W. (e) Contains 12% W. Source: AWS A5.8 and Wall Colmonoy Corporation
4775
BNi-1
2.75–3.50 2.75–3.50 2.75–3.50 2.75–3.50 1.5–2.2 0.03 ... 0.01 ... 3.25–4.0 2.5 2.5
AMS
13.0–15.0 13.0–15.0 6.0–8.0 ... ... 18.5–19.5 ... 13.0–15.0 ... 13.5–16.5 12 10
BNi-1 BNi-1a BNi-2 BNi-3 BNi-4 BNi-5 BNi-6 BNi-7 BNi-8 BNi-9 BNi-10 (c) BNi-11 (c)
AWS designation
Cr
AWS designation
Table 5.10 Nickel-base brazing filler metals
Chapter 5: Brazing Filler Metals / 199
tions in the nuclear industry. Among them are Ni-25Cr-10P, Ni-26.3Cr-5.1Si-3P, Ni-14.8Cr8Si-3P-3Fe, and Ni-20.3Cr-11.5Si-0.5P (Ref 64, 65, 68). The phosphorus-containing filler metals suffer from low ductility, because they form nickel phosphides. Other work has concentrated on developing filler metals for brazing mechanically alloyed (MA), high-temperature, nickelbase alloys that are oxide dispersion strengthened (Ref 65). Because of their excellent high-temperature strength and corrosion resistance, these MA nickel-base alloy structural materials, such as Incoloy MA-356, which would be serviceable at temperatures up to 1200 °C (2190 °F), are potential candidates for various components for future gas turbine engines. The most promising brazing system has been Ni-CrPd with tungsten, which has improved mechanical properties by solid-solution strengthening. Another new filler metal developed especially to wet and flow at 1080 °C (1975 °F) on superalloys with combined aluminum and titanium contents of 1 to 3% (such as René 41) has a composition of 6% Si, 4.3% Cu, 0.079% Misch metal, and the balance nickel (no boron). Three experimental filler metals were evaluated (Ref 66) for suitability in industrial applications. The demand for boron-free, low-melting, nickel-base filler metals led to the development of three high-temperature filler metals with low amounts of phosphorus and iron. The evidence of the industrial suitability of the proposed new filler metals was made with stainless steel American Iron and Steel Institute (AISI) 321 using a combined test method. The three silicide-phosphide high-temperature filler metals were developed by modifying ternary Ni-Cr-Si alloys with phosphorus and iron. Filler metal 1 was developed by the addition of 0.5% P to the alloy 20.4Cr-11.6Si-bal Ni, which is almost the ternary eutectic point of the Ni-Cr-Si system. Hence, a melting range was reached between 1044 and 1060 °C (1911 and 1940 °F). Filler metal 2 was developed by alloying 3% P to the ternary alloy 27.1Cr-5.2Si-bal Ni, producing an alloy with a melting range of 952 to 1155 °C (1746 to 2110 °F). Filler metal 2 had a high chromium content, which can be seen in the wide melting range and the high liquidus temperature. Filler metal 3 was developed from the ternary alloy 15.8Cr-8.5Si-bal Ni by alloying it with 3%
P and 3% Fe. It has a solidus of 996 °C (1825 °F) and a liquidus of 1058 °C (1936 °F). The data from the tests showed that only through the use of the two combined investigation methods was it possible to come to a final conclusion about the suitability of a fillermetal/base-metal combination. Filler metal 1 showed the best match between metallographic investigation and tensile strength test. With brazed butt-joint tensile specimens, the filler metal reached the basemetal tensile strengths at a brazing temperature of 1070 °C (1960 °F) and a brazing time of 10 min. This particular filler metal might be used in a similar manner as the commercial filler metal BNi-5, where an 80 °C (145 °F) lower brazing temperature is possible. Filler metal 2 was not suitable for industrial applications. At tight joint clearances and with brazing temperatures of 1190 °C (2175 °F), satisfactory tensile strengths could not be reached. This filler metal was not an acceptable alternative to commercial filler metals or the other two experimental filler metals. Filler metal 3 can be used at brazing temperatures above 1070 °C (1960 °F). High tensile strengths were reached at temperatures of 1090 °C (1995 °F), making it a satisfactory alternative for industrial use. The minimum brazing temperature of 1090 °C (1995 °F) is approximately 60 °C (110 °F) lower than the brazing temperature of BNi-5. Researchers (Ref 67) found significant changes in oxygen contents of nickel-base filler metals during heating in simulated vacuum brazing cycles. They concluded: • The oxygen content of the nickel filler metals just below their solidus temperature is, in general, dependent on the partial pressure of oxygen in the brazing furnace and not on the initial oxygen content of the filler metal. • The boron filler metals, such as BNi-2 and BNi-3, drop significantly in oxygen content prior to reaching their solidus temperature when processed in a clean vacuum furnace operating in the vicinity of 1.3 to 0.13 Pa (2 × 10–4 to 2 × 10–5 psi). • The hydrated cement with nickel filler metal added to form a paste exhibits an increase of oxygen in the powder at lower test temperatures. At higher test temperatures, but below their solidus temperature, oxygen content drops substantially below the oxygen levels of the initial filler metal when run in a clean
200 / Brazing, Second Edition
furnace. Thus, the oxygen content of the original powder would have no significant effect on the braze results. Results are therefore dependent on the partial pressure of oxygen in the furnace. • The BNi-5 filler metal run in a reasonably clean production vacuum furnace prior to special cleanup exhibits an increase in oxygen at lower temperature, but prior to reaching the solidus, there is a drop in oxygen content to below its initial oxygen content. Because silicon dioxide reduction occurs at a much higher temperature than the chromium oxide, it was expected that a higher temperature would be required to show a significant drop in oxygen content. • Small additions of aluminum, magnesium, and other deoxidizing agents added to the filler metals will improve their wetting and flow characteristics, as indicated by previous tests.
Cobalt Filler Metals (Designated BCo) The cobalt filler metals are used for their high-temperature properties and their compatibility with cobalt-base metals. Brazing in a high-quality atmosphere or diffusion brazing gives optimal results. Special high-temperature fluxes are available for torch brazing. BCo-1 filler metal (Table 4.7), similar to the nickelbase filler metals, receives attention when designers need high-temperature resistance, especially for jet engine parts and honeycomb sandwich structures of cobalt-base alloys.
Palladium-Base Filler Metals Palladium is the major constituent of a series of filler metals that contain copper, nickel, and silver. These filler metals possess many of the beneficial properties of the gold-bearing filler metals but are less expensive. They confer on joints:
• High oxidation resistance at elevated temperatures, especially in the case of the palladium-nickel filler metals • Good corrosion resistance, although not as good as the gold-bearing filler metals • Low vapor pressure at typical brazing temperatures, comparable to that of the gold filler metals • A consistently narrow melting range, in most cases no more than 25 to 50 °C (80 to 120 °F) A representative range of commercial palladium-bearing filler metals is listed in Tables 4.11, 4.12, 4.17, and 5.11. These filler metals find application in refractory metal structures where cost is more critical than with certain aerospace uses, which can justify the gold-bearing filler metals. Platinum is superior to palladium in terms of its chemical inertness and thus finds occasional use in filler metals, but the applications are restricted by the high cost of this metal and by its relatively poor mechanical workability (Ref 1).
Silver Filler Metals (Designated BAg) Expensive but popular, silver filler metals are used for brazing most ferrous and nonferrous metals and alloys, except aluminum and magnesium, with all methods of heating. They offer excellent flow. Silver-Copper-Zinc. These filler metals may be preplaced in the joint or fed into the joint area after heating. Lap joints are generally used with joint clearances of 0.05 to 0.13 mm (0.002 to 0.005 in.) when mineral-type fluxes are used and up to 0.05 mm (0.002 in.) when gas-phase fluxes (atmospheres) are used. However, butt joints may be used if the service requirements are less stringent. Fluxes are generally required, but fluxless brazing with filler metals free of cadmium and zinc can be done on most metals in an Table 5.11 Palladium-bearing filler metals Composition, wt%
• Good mechanical integrity and freedom from brittle intermetallics. This is a consequence of the fact that palladium forms solid solutions with most common metals. • Enhanced mechanical strength at elevated temperatures. In this respect, they tend to be superior to the family of gold filler metals that do not contain palladium or other platinumgroup metals.
Melting range
Pd
Ag
Cu
Ni
Other
°C
°F
65 60 54 25 21 15 5
... ... ... 54 ... 65 68.5
... ... ... 21 ... 20 26.5
... 40 36 ... 48 ... ...
35Co ... 10Cr ... 31Mn ... ...
1230–1235 1237 1232–1260 900–950 1120 850–900 805–810
2245–2255 2259 2250–2300 1650–1740 2050 1560–1650 1480–1490
Source: Ref 1
Chapter 5: Brazing Filler Metals / 201
inert or reducing atmosphere (such as dry hydrogen, dry argon, vacuum, and combusted fuel gas) (Ref 12). Copper forms filler metals with iron, cobalt, and nickel much more readily than silver does. Also, copper wets many of these metals and their alloys satisfactorily, whereas silver does not. Consequently, the wettability of silver-copper filler metals decreases as the silver content increases in brazing of steels, stainless steels, nickel-chromium alloys, and other metals. Thus, a high-silver-content filler metal does not wet steel well when brazing is done in air with a flux. When brazing in certain protective atmospheres without flux, silver-copper filler metals will wet and flow freely on most steels at the proper temperature (Table 5.12). The addition of cadmium to Ag-Cu-Zn filler metals dramatically lowers their melting and flow temperatures. Cadmium also increases the fluidity and wetting action of the filler metal on a variety of base metals. Cadmium-bearing filler metals should be used with caution. If they are improperly used and subjected to overheating, cadmium oxide fumes can be generated. Cadmium oxide fumes are a health hazard, and excessive inhalation of these fumes must be avoided. Because cadmium-bearing filler metals are not intended for fluxless brazing, an appropriate flux should always be used with these filler metals when brazing in either air or furnace atmospheres. There are several special noble-metal filler metals that should also be mentioned (Ref 23). Among these are the vacuum-grade filler metals, which are made to high purities and are virtually free of high-vapor-pressure elements. The vacuum-grade brazing filler metals are used in brazing parts for vacuum tubes and electronic circuits that require durability in demanding applications. Made to be spatter-free, most vacuum-grade filler metals come in two grades. Grade 1 contains zinc and cadmium to a maximum of 0.001%; grade 2 contains 0.002 maximum zinc and cadmium. These filler metals are also low in lead, phosphorus, carbon, mercury, antimony, potassium, sodium, lithium, titanium, sulfur, cesium, rubidium, selenium, tellurium, strontium, and calcium, which are elements that have vapor pressures greater than 0.00013 Pa (2 × 10–8 psi) at 500 °C (930 °F). The cadmium-free filler metals, BAg-22, -24, -25, and -26, are listed in Table 5.12. Due to federal regulations and the requirements placed on health and safety, the use of cadmium-free filler
metals as required is extremely important. The addition of small and carefully controlled amounts of one or more of the elements tin, nickel, and manganese to filler metals of the AgCu-Zn ternary alloy system provides a series of useful cadmium-free filler metals. These newly developed and commercially available filler metals have particular potential for production brazing applications in the refrigeration, shipbuilding, cutlery, automobile, and rock-drilling tool industries. The availability of these materials provides industry with a new range of filler metals that are both technically and commercially attractive and that enable companies who wish to do so to avoid the use of cadmium in their production brazing operations. Zinc is commonly used to lower the melting and flow temperatures of copper-silver filler metals. Zinc is by far the most helpful wetting agent for joining alloys based on iron, cobalt, or nickel. Alone, or in combination with cadmium or tin, zinc produces filler metals that wet the iron-group metals but do not alloy with them to any appreciable depth. Tin has a low vapor pressure at normal brazing temperatures. It is used in silver-base filler metals in place of zinc or cadmium when volatile constituents are objectionable, such as when brazing is done without flux in atmosphere or vacuum furnaces, or when the brazed assemblies will be used in high vacuum at elevated temperatures. Tin additions to silver-copper filler metals result in wide melting ranges. Filler metals containing zinc wet ferrous metals more effectively than those containing tin, and where zinc is tolerable, it is preferred over tin. Generally, as the combined zinc and cadmium content is increased beyond 40%, the ductility of the filler metal decreases. This fact puts a practical limit on how much the flow temperatures of silver-base filler metals can be lowered. Stellites, cemented carbides, and other molybdenum- and tungsten-rich refractory alloys are difficult to wet with Ag-Cu-Zn filler metals. Manganese, nickel, and (infrequently) cobalt are often added as wetting agents in filler metals used for joining these materials (Ref 68). An important characteristic of silver-base filler metals containing small additions of nickel is improved resistance to corrosion under certain conditions. They are particularly recommended where joints in stainless steel are to be exposed to saltwater corrosion.
Ag
44.0–46.0
49.0–51.0
34.0–36.0
29.0–31.0 49.0–51.0
39.0–41.0 44.0–46.0
49.0–51.0
55.0–57.0
71.0–73.0
71.0–73.0
64.0–66.0
69.0–71.0 53.0–55.0
55.0–57.0
59.0–61.0
92.0–93.0
29.0–31.0
Brazing filler metal
BAg-1
BAg-1a
BAg-2
BAg-2a BAg-3
BAg-4 BAg-5
BAg-6
BAg-7
BAg-8
BAg-8a
BAg-9
BAg-10 BAg-13
BAg-13a
BAg-18
BAg-19
BAg-20
37.0–39.0
bal
bal
bal
19.0–21.0 bal
19.0–21.0
bal
bal
21.0–23.0
33.0–35.0
29.0–31.0 29.0–31.0
26.0–28.0 14.5–16.5
25.0–27.0
14.5–16.5
14.0–16.0
Cu
30.0–34.0
...
...
...
8.0–12.0 4.0–6.0
13.0–17.0
...
...
15.0–19.0
14.0–18.0
26.0–30.0 23.0–27.0
21.0–25.0 13.5–17.5
19.0–23.0
14.5–18.5
14.0–18.0
Zn
...
...
...
Ni
...
...
...
...
... ...
...
...
...
...
...
... ...
...
...
...
1.5–2.5
... 0.5–1.5
...
...
...
...
...
1.5–2.5 ...
19.0–21.0 ... 15.0–17.0 2.5–3.5
17.0–19.0
17.0–19.0
23.0–25.0
Cd
Composition,%
Table 5.12 Silver-base brazing filler metals
...
...
9.5–10.5
...
... ...
...
...
...
4.5–5.5
...
... ...
... ...
...
...
...
Sn
...
0.15–0.30
...
...
... ...
...
0.25–0.50
...
...
...
... ...
... ...
...
...
...
Li
...
...
...
...
... ...
...
...
...
...
...
... ...
... ...
...
...
...
Mn
680
760
600
770
690 720
670
770
780
620
690
670 680
610 630
610
630
610
°C
1260
1400
1110
1420
1275 1325
1240
1420
1435
1150
1275
1240 1260
1130 1170
1130
1170
1130
°F
Solidus temperature
(continued)
0.15
0.15
0.15
0.15
0.15 0.15
0.15
0.15
0.15
0.15
0.15
0.15 0.15
0.15 0.15
0.15
0.15
0.15
Other elements total
765
890
720
895
740 860
720
765
780
650
775
780 745
710 690
700
635
620
°C
1410
1635
1325
1640
1360 1580
1325
1410
1435
1200
1425
1435 1370
1310 1275
1290
1175
1150
°F
Liquidus temperature
765–870
875–980
720–845
870–980
740–845 860–970
720–845
765–870
780–900
650–760
775–870
780–900 745–845
710–845 690–815
700–845
635–760
620–760
°C
Comments
1150–1400 Lowest brazing temperature of the BAg series. Flows well into tight joints. All brazing processes Contains Cd 1175–1400 Narrow melting range; more fluid than BAg-1; more costly 1290–1550 Broader melting range for wide joint gaps. Heat rapidly 1310–1550 Similar to BAg-2, contains less silver 1275–1500 BAg-1a with Ni added for corrosion resistance, wetting of tungsten carbide. Use for carbide tool brazing 1435–1650 Similar to BAg-3, but Cd-free 1370–1550 Cd-free. Use for food equipment, brass brazements 1425–1600 Similar to BAg-1 and -2, but wider melting range for wide joints 1200–1400 Similar to BAg-1, but Cd-free. Use for Ni alloys, for color match 1435–1650 Ag-Cu eutectic; flows well. Use on Cu alloys, stainless steel, carbon steel 1410–1600 For furnace brazing in protective atmosphere. Use on stainless steels. Li promotes wetting. 1325–1550 Joins sterling silver, matches its color. Use for step brazing 1360–1550 Use for step brazing 1580–1780 Low zinc; suitable for furnace brazing. For service up to 370 °C (700 °F) 1600–1800 Similar to BAg-13, but no zinc; use for furnace brazing 1325–1550 Similar to BAg-8. Sn promotes wetting of stainless steel, Ni alloys, carbon steel. Use for fluxless brazing, step brazing with BAg-8 1610–1800 Same as BAg-8, but for higher brazing temperatures, where heat treatment accompanies brazing 1410–1600 Good wetting and flow characteristics. Cd-free. Economical
°F
Brazing temperature range
Ag
62.0–64.0
48.0–50.0 84.0–86.0
49.0–51.0
24.0–26.0
24.0–26.0
39.0–41.0
24.0–26.0 37.0–39.0 34–36
44–46 24–26
Brazing filler metal
BAg-21
BAg-22 BAg-23
BAg-24
BAg-26
BAg-27
BAg-28
BAg-33 BAg-34 BAg-35
BAg-36 BAg-37
26–28 39–41
29.0–31.0 31.0–33.0 31–33
29.0–31.0
34.0–36.0
37.0–39.0
19.0–21.0
15.0–17.0 ...
27.5–29.5
Cu
Table 5.12 (continued)
23–27 31–35
26.5–28.5 26.0–30.0 31–35
26.0–30.0
24.5–28.5
31.0–35.0
26.0–30.0
21.0–25.0 ...
...
Zn
... ...
16.5–18.5 ... ...
...
12.5–14.5
...
...
... ...
...
Cd
... ...
... ... ...
...
...
1.5–2.5
1.5–2.5
4.0–5.0 ...
2.0–3.0
Ni
Composition,%
2.5–3.5 1.5–2.5
... 1.5–2.5 ...
1.5–2.5
...
...
...
... ...
5.0–7.0
Sn
... ...
... ... ...
...
...
...
...
... ...
...
Li
... ...
... ... ...
...
...
1.5–2.5
...
7.0–8.0 bal
...
Mn
0.15 0.15
0.15 0.15 0.15
0.15
0.15
0.15
0.15
0.15 0.15
0.15
Other elements total
645 690
610 650 685
650
610
710
660
680 960
690
°C
1195 1275
1130 1200 1265
1200
1130
1310
1220
1260 1760
1275
°F
Solidus temperature
675 780
680 720 755
710
745
800
705
700 970
800
°C
1250 1435
1260 1325 1390
1310
1370
1475
1305
1290 1780
1475
°F
Liquidus temperature °F
Comments
1475–1650 Zn-Cd-free; for furnace brazing. Use on 300- and 400-series stainless steels; Ni alloys. No flux over 1010 °C (1850 °F). Sluggish flow 700–830 1290–1525 Use as BAg-3. Cd-free 970–1040 1780–1900 Torch or atmosphere furnace brazing. Hightemperature brazing. Free-flowing. For stainless steel, Ni alloys, high-temperature applications 710–845 1310–1550 Low-melting; free-flowing; Cd-free. For 300-series stainless steels, food equipment, carbide tool inserts 800–870 1475–1600 Low silver; economical. Cd-free. Use for carbide, stainless steel brazing 745–860 1370–1580 A low-silver version of BAg-2. Contains Cd. Limited ductility joints 710–845 1310–1550 Lower brazing temperature, narrower melting range than other Cd-free fillers 680–760 1260–1400 Low brazing temperature. High Zn and Cd 720–845 1325–1550 Cd-free. Similar to BAg-2 and -2a 755–840 1390–1545 Cd-free. General-purpose production brazing 675–815 1250–1495 Cd-free. Low brazing temperature 780–885 1435–1625 Cd-free; low silver; economical. Limitedductility joints
800–900
°C
Brazing temperature range
204 / Brazing, Second Edition
When stainless steels and other alloys that form refractory oxides are to be brazed in reducing or inert atmospheres without flux, silverbase filler metals containing lithium as the wetting agent are quite effective. The heat of formation of lithium oxide is very high; consequently, lithium is capable of reducing the adherent oxides on the base metal. The resultant lithium oxide is readily displaced by the filler metal. Lithium-bearing filler metals are advantageously used in very pure dry hydrogen or inert atmospheres. Continuous service temperatures for silverbase filler metals range up to 205 °C (400 °F), with intermittent service up to 315 °C (600 °F), adjusted for the actual operating environment. Silver-Copper plus Palladium. A recently completed study (Ref 69) reported that a 65Ag20Cu-15Pd (Palcusil-15, Wesgo Metals) and two other silver-base filler metals, 92Ag-7.8Cu0.2Li (BAg-19) and 60Ag-30Cu-10Sn (BAg18), were tested in edgewise compression in beryllium panels and carried loads up to 649 °C (1200 °F). Silver-Copper plus Manganese and Gallium. Researchers (Ref 70) developed a lowsilver filler metal for brazing corrosion-resistant steels. To reduce the silver content of the filler metal and retain the high ductility of the brazing joints in the corrosion-resistant steels at cryogenic temperatures, the standard silver brazing filler metals have been replaced by a face-centered cubic metal, that is, copper, whereas the components forming wide ranges of solid solutions with copper have been represented by manganese, silver, and gallium. The optimal percentage chemical composition of the filler metal was 5Ag, 10Mn, 5Ga, balance Cu. The filler metal was manufactured in the form of 0.1 to 2.0 mm (0.004 to 0.08 in.) thick strips and 0.5 to 4.0 mm (0.02 to 0.16 in.) diameter wires. The filler metal spread satisfactorily on the corrosion-resistant steels heated in vacuum and in an air furnace. The results of the mechanical tests on the filler metal showed that in the cast condition, the filler metal was sufficiently strong, ductile, and had high impact toughness, both at room temperature and at –20 °C (–4 °F). Silver-Copper-Hafnium. Due to the increasing applications of nonoxide ceramics in technical structures, new active filler metals are being developed for joining ceramics either to other ceramics or to metals. In commercially dominant active filler metals, titanium as the reactive agent is added to the base filler metal, par-
ticularly to the silver-copper eutectic. Apart from titanium and other reactive elements, hafnium additives are also known for promoting the wetting of ceramics by conventional filler metals. Researchers (Ref 71) investigated the hafnium-added silver-copper active filler metals and examined their brazing properties for joining SiC and Si3N4 ceramics to themselves and to steel. Active filler metals in a system of Ag-Cu(In)-Hf, with hafnium contents ranging from 2 to 5 wt% and a eutectic silver-copper composition, have shown that the filler metals have a melting behavior similar to titanium containing silver-copper filler metals. Wettability tests conducted on SiC and Si3N4 proved that they exhibit good wetting properties at elevated temperatures (>1000 °C, or 1830 °F). Brazing test joints at different brazing conditions showed that hafnium-containing filler metals are suitable for joining SiC/Si3N4 ceramics to themselves as well as for joining them to steel. The quality of the joint strongly depends on the quality of the vacuum because of the high reactivity of hafnium, even to traces of atmospheric gases. Regarding the mechanical properties of joints brazed with hafnium-containing filler metals, four-point bending test specimens showed strong joints with bending strengths above 149 MPa (22 ksi) in the case of Si3N4. The SiC ceramics display weaker mechanical properties due to their thicker reaction zone (Ref 71). Silver-Copper-Titanium (Ref 2, 3, 69, 72). Researchers conducted a series of experiments (Ref 72) to gain information about the wettability of AlN, BN, Si3N4, and two SiAlON (syalon) ceramics by using aluminum, coppertitanium filler metals, and a Ag-28Cu-2Ti filler metal. Wetting by aluminum and the Ag-28Cu2Ti filler metal was usually good. Both wetting and nonwetting filler metals containing titanium reacted to form TiN, and the achievement of wettability was associated with a certain degree of hypostoichiometry. While aluminum should also have reacted, no clear evidence was obtained. In supplementary experiments, it was found that bonds formed by brazing with aluminum at 1000 °C (1830 °F) could have shear strengths as great as 60 MPa (9 ksi). Although the experimental work was preliminary in nature, it suggested that good brazing systems could be developed. Researchers (Ref 73) joined Si3N4, using a 57Ag-38Cu-5Ti filler metal, and reported that
Chapter 5: Brazing Filler Metals / 205
various temperatures for brazing had an effect on the layer interface structure. Hot-pressed Si3N4 was joined using 57Ag-38Cu-5Ti filler metal in a vacuum, and the maximum bend strength of the joint measured by the four-point bend method was 490 MPa (71 ksi) when brazing at 880 °C (1615 °F) for 5 min. It is important that there is sufficient reaction during brazing between the ceramic and the active element in the filler metal, characterized by forming TiN with an appropriate thickness. On the other hand, insufficient interface reaction will decrease the joining strength. This work shows that one can expect to obtain a high joining strength using silver-copper filler metal with a small quantity of titanium as an active element, without forming a brittle dispersed phase such as silver-titanium or copper-titanium compound after brazing. From the previously mentioned results, in order to obtain high joining strength, the following points are suggested when one designs a filler metal for joining ceramic to ceramic or ceramic to metal: • The selected active element should have a high free energy of reaction with the ceramic and a strong segregation on the ceramicmetal interface. • The lower limit of active-element content in the filler metal should ensure good wetting of the ceramic by the filler metal, while the upper limit should be such that there is no brittle dispersed phase in the filler metal. • The brazing temperature and time should be sufficient to ensure interface reaction of the filler metal with the ceramic. A 27.5Cu-2Ti-62Ag filler metal (Cusil ABA, Wesgo Metals), an active filler metal, has been successfully used to join hot-pressed Si3N4 and Incoloy 909, a low-expansion superalloy (Ref 74). Brazing was carried out in vacuum (approximately 10–3 Pa, or 10–7 psi) at 950 °C (1740 °F) for 20 min. A 40Ag-5Ti-55Cu filler metal was successfully used to join alumina to Ti-6Al-4V and 3% Y/-partially stabilized zirconia (PSZ) to Ti-6Al4V, and brazing was performed by heating for 5 min in a vacuum furnace (2.6 mPa, or 3.8 × 10–7 psi) at 870 °C (1600 °F) (Ref 75). The rather good mechanical behavior reported can be related to the very similar thermal expansion coefficients of Ti-6Al-4V and zirconia. Another reason for these results can be found in the fact that the PSZ that was used was partially trans-
formed into monolithic zirconia, with such a transformation accommodating the stresses that developed during the brazing procedure. The zirconia/Ti-6Al-4V brazements exhibit interfacial phases that are very similar to those observed in the alumina/Ti-6Al-4V brazements. However, the tensile strength of joints using zirconia ceramics is much higher (150 ± 50 MPa, or 22 ± 7 ksi). The same filler metal was also used to join alumina to Ti-6Al-4V by active brazing. The high reactivity of the active filler metal formed a continuous and sinuous layer identified as Cu2(Ti, Al)4O at the braze-alumina interface. The presence of this layer was found to be beneficial to the bonding between the ceramic and the metal (Ref 76). Alumina joints of high integrity were produced with a 56Ag-36Cu-6Sn-2Ti (wt%) experimental filler metal by vacuum brazing at 900 °C (1650 °F) for 20 min. It was found by transmission electron microscope that the formation of a Ti3Cu3O-phase reaction layer in conjunction with a TiO layer can provide a more gradual transition in chemical bonding between the alumina and the silver-copper metallic fillermetal phases than TiO will alone. The Ti3Cu3Ophase layer may also provide a more gradual transition in physical properties and help to minimize the effect that local strains, which develop from thermal expansion coefficient mismatches, can have on adhesion (Ref 77–79). Successful application of new high-performance cutting materials can only be achieved if adequate joining techniques are provided. Brazing has been proved to be a promising approach for new advanced materials, such as cemented carbides with low binder concentration, or Si3N4. The AgCuTi active filler metal has been employed as a filler metal for both Si3N4 and cemented carbides. Cemented carbides have also been successfully brazed by filler metals (see Table 5.13). In order to reduce thermally induced stresses in ceramic materials, various interlayer materials were examined (Ref 80). Microstructural and mechanical analyses revealed that the joint formation as well as the interfacial interactions were of great significance for the joint quality. Although the use of interlayers is supposed to be an effective approach to reduce thermally induced stresses in bimaterial joints, the influence of microstructural effects within the joint cannot be neglected. The correlation of experimental results with finite-element calculations revealed that
206 / Brazing, Second Edition
there are discrepancies. These can be attributed to the fact that finite-element analyses do not take into account metallurgical effects.
4170 °F). Most of the work up to the present has involved metallizing. In order to obtain wetting of the joint surfaces by the filler metal, the area of the joint on the ceramic side must first be provided with a firmly bonded metal coating—the metallized film. These films can be applied by numerous techniques, including electrolytic precipitation, gas-phase precipitation, thermal and plasma spraying, ionic plating, and electron- and laser-beam coating. At temperatures up to approximately 1200 °C (2190 °F), brazed
Other Combinations Other groups of special filler metals include those that have been developed for joining refractory metals and their alloys (Table 5.14), graphite, and ceramics to themselves and to metals—especially those with brazing temperatures ranging from 1040 to 2300 °C (1905 to
Table 5.13 Commercially available filler metals for joining cemented carbides Filler metal
Ag-16Cu-23Zn-7.5Mn-4.5Ni Ag-26Cu-6In-2Ni-2Mn Cu-39.8Zn-0.2Si Cu-12Mn-2Ni Cu-3Co-10Mn
Solidus temperature
Liquidus temperature
Brazing temperature
°C
°F
°C
°F
°C
°F
685 730 890 970 980
1265 1345 1635 1780 1795
705 780 900 990 1030
1300 1435 1650 1815 1885
690 770 900 990 1020
1275 1420 1650 1815 1870
Source: Ref 80
Table 5.14 Filler metals for brazing of refractory metals Liquidus temperature Filler metal(a)
Liquidus temperature
°C
°F
2416 2996 960 1082 1454 1816 2127 1774 2299 2049
4380 5425 1760 1980 2650 3300 3860 3225 4170 3720
Ag-Cu-Zn-Cd-Mo Ag-Cu-Zn-Mo Ag-Cu-Mo Ag-Mn
618–702 718–788 779 971
1145–1295 1325–1450 1435 1780
Ni-Cr-B Ni-Cr-Fe-Si-C Ni-Cr-Mo-Mn-Si Ni-Ti Ni-Cr-Mo-Fe-W Ni-Cu Ni-Cr-Fe Ni-Cr-Si
1066 1066 1149 1288 1304 1349 1427 1121
1950 1950 2100 2350 2380 2460 2600 2050
Nb Ta Ag Cu Ni Ti Pd-Mo Pt-Mo Pt-30W Pt-50Rh
Filler metal(a)
°C
°F
Mn-Ni-Co
1021
1870
Co-Cr-Si-Ni Co-Cr-W-Ni Mo-Ru Mo-B Cu-Mn Nb-Ni
1899 1427 1899 1899 871 1191
3450 2600 3450 3450 1600 2175
Pd-Ag-Mo Pd-Al Pd-Ni Pd-Cu Pd-Ag Pd-Fe Au-Cu Au-Ni Au-Ni-Cr Ta-Ti-Zr
1316 1177 1204 1204 1316 1316 885 949 1038 2093
2400 2150 2200 2200 2400 2400 1625 1740 1900 3800
1649 1482 1427 999 1049 1249 1816–1927 1760–1843
3000 2700 2600 1830 1920 2280 3300–3500 3200–3350
Ti-V-Cr-Al Ti-Cr Ti-Si Ti-Zr-Be(b) Zr-Nb-Be(b) Ti-V-Be(b) Ta-V-Nb(b) Ta-V-Ti(b)
(a) Not all the filler metals listed are commercially available. (b) The liquidus temperature (and therefore the brazing temperature) depends on the specific composition.
Chapter 5: Brazing Filler Metals / 207
joints can be used on metallized ceramics. In higher temperature ranges, alloying between the filler metal and the metallizing film can influence the adhesive strength of the joint (Ref 81). Metallization of the surface of the ceramic is unnecessary when the newly developed ceramic filler metals are used. These are oxide mixtures, which, in the molten state, wet both the surface of the metal and also that of the ceramic base. Additionally, there are active filler metals that contain interfacially active components that reduce the interfacial energy between the ceramic and the molten filler metal to such a low level that wetting of the ceramic takes place. The best-known interfacial active elements are titanium and zirconium. Even small concentrations of a few percent in copper or silver are sufficient to produce excellent wetting of alumina and other oxides. Characteristic of the active metals (titanium and zirconium) is their great affinity for oxygen, which permits them to react with the ceramic oxides and form their own oxide phases. It is probably this reaction that makes possible the high degree of adhesion between the components of the joint but that, under certain conditions, can also lead to weakening of the joint. Due to the high degree of affinity for oxygen involved, brazing of such joints can be carried out only under conditions of high vacuum or in a dry, high-purity, inert gas atmosphere (Ref 81). Three new active silver-base filler metals have been developed that permit brazing of metal parts to high-alumina and other structural ceramics (such as wear-resistant, heat-resistant, and similar parts) without metallizing of the ceramic material. These new ductile filler metals are adaptable for brazing of metals to such materials as Si3N4, PSZ, transformation-toughened aluminas, and SiC, as well as many other refractory materials. Foil, wire, and preforms of these filler metals are available. Researchers (Ref 82) initiated a development study of tin-containing filler metals to replace the silver-base filler metals containing cadmium. The problem with cadmium-containing filler metals is the toxic fumes they generate during the brazing operation, which can be injurious to health, or even fatal, if inhaled (Ref 83). Accumulation of sufficient data on the toxic effects of cadmium resulted in the 1970 introduction of a threshold limit value for cadmium in most European countries (Ref 84, 85). This has generally
resulted in a decision to switch to cadmium-free filler metals rather than installing elaborate fume extraction systems (Ref 85, 86). In industry, it has been difficult to get a direct replacement filler metal equal in physical properties and intrinsic cost. The Ag-Cu-Zn filler metals are available but have silver contents as high as 55% to maintain a solidus temperature of 630 °C (1165 °F), which makes these filler metals very expensive. Tin has been added to some cadmium-free filler metals to maintain a low working-temperature range. A typical filler metal contains 60Ag-30Cu-10Sn. It can be used in fluxless controlled atmosphere brazing and vacuum brazing of ferrous and nonferrous alloys. It has a wide working-temperature range (600 to 720 °C, or 1110 to 1330 °F), which aids in filling the joint. Tin additions improve the wetting characteristics of ferrous alloys over that obtained with binary silver-copper filler metals (Ref 84, 85). The most effective composition in the range contains 55Ag-21Cu-22Zn-2Sn. This filler metal has a working temperature of 630 to 660 °C (1165 to 1220 °F) and has successfully replaced Ag-Cu-Zn-Cd filler metals in some cases. It is excellent for brazing cutlery, jewelry, hollowwares, and so on (Ref 87). The efficient tin-containing filler metals do not include toxic cadmium, but they have a large amount of silver, which is expensive and influenced by price fluctuations. These two major factors, that is, the high cost of silver and the toxicity of cadmium fumes, initiated the study for alternative filler metals for brazing copper, mild steel (MS), stainless steel, and so on within the temperature range of 600 to 850 °C (1110 to 1560 °F). The initial results of the study determined the suitability of Cu/15–40Sn/5–15Mn filler metals as universal filler metals for joining copper-copper, copper-MS, and MS-MS at 750 to 850 °C (1380 to 1560 °F). This series of filler metals can replace expensive and toxic Ag-Cu-Zn-Cdbase filler metals. The working temperatures of these filler metals are slightly higher than silverbase filler metals, but this has to be weighed out against its softening effect on copper joints, the economics, and the toxicity of Ag-Cu-Zn-Cd filler metals. Another set of filler metals, Cu-Sn-P and CuSn-P-Ni, can be used successfully to join copper and copper-base filler metals. It has been shown
208 / Brazing, Second Edition
experimentally that it is not advisable to use these filler metals for making copper-MS or MS-MS joints. In other experimental work, it was shown that a Cu-30Sn-12Bi filler metal could be used for joining copper-copper and copper-MS at 750 to 800 °C (1380 to 1470 °F). It also appears possible to develop suitable Cu-Sn-Ni filler metals for brazing stainless steel (Ref 82). Researchers (Ref 88) examined the metallurgical characteristics of Si3N4 braze joints fabricated for service at elevated temperatures. Filler metals containing palladium, platinum, copper, nickel, and silver were investigated. Most filler metals were arc melted, and then, differential thermal analysis was performed to determine the liquidus and solidus temperatures. Wetting tests were employed as selection criteria. The Si3N4 substrate was premetallized at a lower temperature with an AgCuInTi filler metal prior to brazing at elevated temperatures. The reaction layer developed during premetallizing remained stable at the higher brazing temperature, controlling the Si3N4 decomposition. Other braze joints were fabricated without premetallizing, using a 90Co-10Ti meltspun foil. They concluded: • The Si3N4 decomposes when brazed in 0.13 Pa (2 × 10–5 psi) of vacuum at 1250 °C (2280 °F). By premetallizing the ceramic substrate at 900 °C (1650 °F) under the same vacuum conditions, with a AgCuInTi filler metal, the decomposition is avoided. It also enhances wettability at elevated temperatures (Ref 89). • The palladium and platinum filler metals showed significantly similar microstructural features. The 55Pt-43Cu-2Ti filler metal exhibited better oxidation resistance than did 58.2Pd-38.8Ni-3Ti. • The meltspun 90Co-10Ti filler metal, after an early sudden increase in oxidation, passivated and remained unchanged after approximately 2 h. It ultimately showed better oxidation resistance than the 58.2Pd-38.8Ni-3Ti filler metal. Silicate Systems. Brazing using a silicate interlayer is similar to metal brazing but differs in that achieving wetting does not generally pose a problem (Ref 90, 91). Ceramic brazes often offer better environmental compatibility than metals but, in general, still not equal to that of the base material (Ref 90). Ceramic brazes
are usually less tolerant of thermal expansion coefficient mismatch (Ref 90, 92). The effect of joint thickness and thermal expansion mismatch on the mechanical properties of joints made with silicate brazes was studied (Ref 93, 94). This study consisted of finiteelement analysis to estimate the stresses involved (Ref 93) and mechanical testing of alumina-to-alumina butt joints with interlayers of various thicknesses and with varying degrees of thermal expansion mismatch (Ref 94). The stress analysis (Ref 93) for brazes with thermal expansion coefficients lower than that of the adherend revealed that axial stresses are tensile and may result in lower fracture stresses. Several glass sealants with joining temperatures in the range of 450 to 1500 °C (840 to 2730 °F) have been developed that are suitable for joining alumina to alumina or to sapphire. The hightemperature sealants (Ref 95) reported were based on kaolin and on potassium feldspar, while the low-temperature glasses (Ref 95) were all borate compositions. A study (Ref 96) was started to identify and characterize silicate systems as brazes for alumina ceramics. Choosing the proper systems can result in good wetting characteristics, impurity insensitivity, and a refractory nature through a postprocessing crystallization anneal. The two silicate systems examined were based on either talc, Mg3Si4O10(OH)2, or anorthite, CaAl2Si2O8. Two talc-based brazes were examined in an effort to extend prior work by other researchers. These brazes were pure talc and a 50-50 (by weight) mixture of talc and alumina. Limited mechanical testing using the four-point bend method showed that alumina-to-alumina joints using the 50-50 mixture were relatively strong (~20 to 50 MPa, or 3 to 7 ksi). However, the talc-based brazes were characterized by unacceptable levels of porosity. Several processing steps were attempted, including calcination of powders and raising brazing temperature from 1500 to 1600 °C (2730 to 2910 °F). These modifications were tried and had some effect on pore morphology but did not significantly reduce porosity. As a result, the talc-based systems were abandoned. The anorthite-based system has proven much more promising. A glass with the same chemical composition as anorthite, termed An-glass (CaAl2Si3O8), can be readily made by melting together equal molar amounts of whiting, CaCO3, and kaolin,
Chapter 5: Brazing Filler Metals / 209
Al2Si2O5(OH)4, at 1600 °C (2910 °F), ~50 °C (90 °F) above the melting point of anorthite (1553 °C, or 2827 °F), for 0.5 h and cooling to room temperature. Additionally, An-glass can be fully crystallized by annealing at 1100 °C (2010 °F) for 1 h. Alumina brazed with An-glass at 1600 °C (2910 °F) leads to joints with several favorable characteristics. First, a good bond develops between the alumina adherends and the Anglass. In contrast to the talc-based brazes, the porosity problem within the braze layer is minimal. Furthermore, acicular crystals (~100 µm long, 2 µm in diameter) are observed to form within the An-glass matrix. Many of these crystals tend to be aligned roughly perpendicular to the braze layer direction and may be significant contributors to the joint toughness. This system proved successful, and average bend strengths of ~145 MPa (21 ksi) were achieved with vitreous joints, while room-temperature strengths were retained to 500 °C (930 °F). Brazing of the same type of alumina used in this study with a filler metal having a composition of 45Cu26Ag-29Ti (at.%) resulted in joints with an average four-point bend strength of 222 MPa (32 ksi) at 25 °C (75 °F) (Ref 96). Figure 5.3 depicts the sequence of steps in silicate brazing (Ref 96). A rapid infrared processing technique, characterized by low cost, short processing time, and easy operation, for forming good joints between SiC-SiC has been developed (Ref 97). The infrared processing technique has been used to join SiC to itself, using a silicon foil as the interlayer material, at temperatures of 1450 °C (2640 °F) for times of 15 to 60 s. The joint cross sections were examined using optical microscopy and scanning electron microscopy (SEM) techniques. Results indicated that joints can be formed, and that the range in thickness from 3 to 20 µm shows very good wetting and exhibits almost no voids. A shear in compression testing method was performed to evaluate the mechanical performance of the joint. Shear strengths on the order of 115 MPa (17 ksi), comparable to those produced by conventional processes, were achieved by the infrared process. Titanium-Zirconium-Copper-Nickel. Researchers (Ref 98) investigated the microstructure and mechanical properties of commercially pure titanium (CPTi) and Ti-6Al-4V alloy joints brazed with newly developed titanium-base amorphous filler metals. Among the developed filler metals were three kinds: Ti-37.5Zr-15Cu-
10Ni, Ti-35Zr-15Cu-15Ni, and Ti-25Zr-50Cu, whose melting points were approximately 100 °C (180 °F) lower than those of conventional titanium-base filler metals (Table 4.18). The use of these filler metals makes it possible to braze at below α/β transformation and β transus temperatures of CPTi and Ti-6Al-4V alloys, respectively. As a result, joints having sufficient tensile properties, as compared to those of the base metals, can only be made by holding for a short time at the brazing temperature. Therefore, in the case of brazing CPTi and Ti-6Al-4V alloy at below α/β transformation and a β transus temperature of each base metal, the original structures of the base metals are completely preserved, and the brazed regions are distinct. The fatigue properties of Ti-6Al-4V alloy joints brazed at 900 °C (1650 °F) for 10 min and 950 °C (1740 °F) for 5 min approach that of the base metal at maximum stresses below 590 MPa (86 ksi). Brazing at 1000 °C (1800 °F) above β transus temperature, the joints exhibit less favorable fatigue properties. These brazed joints have excellent corrosion behavior, so that no reduction in tensile strength occurs after immersion in a 5% NaCl solution for 1000 h.
Fig. 5.3
Schematic of the sequence of steps involved in silicate brazing. (a) Initially, a powdered glass layer of constant thickness is placed between the adherends. (b) On heating, the glass is melted, densifies, and penetrates the grain boundaries, drawing the surfaces together. (c) Crystallization of the interlayer and grain-boundary phase is accomplished through a postjoining controlled heat treatment. Source: Ref 3
210 / Brazing, Second Edition
Three other filler metals have been developed (Ref 99) for wetting Ti-6Al-4V. The three filler metals were 30Pd-60Cu-10Co, 30Pd-40.1Au39.9Cu, and 65.3Ni-11.1Cr-7.6W-4.3Fe-2.6B1.5Si. None exhibited any cracks, voids, and inclusions in the brazed joint. Brazing took place in vacuum (0.013 Pa, or 2 × 10–6 psi), and holding time was 5 min at 1135 °C (2075 °F) for the first filler metal listed previously, 1050 °C (1920 °F) for the second, and 1448 °C (2638 °F) for the last filler metal. Filler metal 1 and 2 were the most suitable in brazed Ti-6Al-4V tests, mainly because of their lower melting points and minimal interaction and erosion due to brazing. Brazing studies conducted (Ref 100) on Ti13.4Al-21.2Nb have led to an understanding of some of the factors that determine microstructures and mechanical behavior of brazed components. The following conclusions were drawn from the results of this work:
• Commercial laminated foil of composition Ti-15Cu-25Ni produced joints with particularly high nickel segregation. Titanium-Zirconium-Beryllium. Brazing is excellent for fabricating assemblies of refractory metals, in particular those involving thin sections. However, only a few filler metals have been specifically designed for both high-temperature and high-corrosion applications (Ref 101). Those filler metals and pure metals used to braze refractory metals are given in Tables 4.27 and 5.14. Low-melting filler metals, such as silver-copper-zinc, copper-phosphorus, and copper, have been used to join tungsten for electrical contact applications, but these filler metals cannot operate at high temperatures. Nickel-base and precious-metal-based filler metals have also been used to join tungsten. Various filler metals will join molybdenum. The effect of brazing temperature on base-metal recrystallization must be considered. When brazing above the recrystallization temperature, brazing time must be kept short. If high-temperature service is not required, copper and silverbase filler metals may be used (Tables 4.11, 4.12, 5.14). Niobium and tantalum are brazed with a number of refractory or reactive-metal-based filler metals. The metal systems Ti-Zr-Be and Zr-Nb-Be are typical, as are platinum, palladium, platinum-iridium, platinum-rhodium, titanium, and nickel-base filler metals. Zinc-Base Filler Metals. In the production of semiconductor devices and integrated circuits, mounting silicon crystals on to the substrate (the base housing) is one of a number of important process operations. Normally, this operation is carried out using contact-reactive brazing utilizing filler metals of gold or its alloys. Under these conditions, a gold-silicon eutectic with a melting point at 370 °C (700 °F)
• The use of filler metals based on the Ti-Cu-Ni alloy system (Table 5.15) can produce braze joints with room-temperature tensile strength comparable to α-2 titanium aluminide parent material. However, at test temperatures of 649 and 760 °C (1200 and 1400 °F), joints produced with these filler metals have only 70 to 80% of base-metal tensile strength. • After a brazing cycle of 1 h at 982 °C (1800 °F), laminated and meltspun foils of the same filler metal were found to produce joints with nearly identical joint compositions, although the distribution of elements within the joints differed. Meltspun foils, particularly at higher solute compositions, produced joints having a more uniform solute distribution than their laminated counterparts. • Laminated and meltspun Ti-15Cu-15Ni foils produced joints with similar microstructures and chemical compositions, using a brazing cycle of 982 °C (1800 °F) for 1 h.
Table 5.15 Melting characteristics of titanium-copper-nickel filler metals Thickness Filler metal
Ti-15Cu-15Ni Ti-20Cu-20Ni Ti-15Cu-25Ni Source: Ref 101
Solidus
Liquidus
Form
µm
in.
°C
°F
°C
°F
Meltspun Laminated Meltspun Meltspun Laminated
40 39 38 38 46
0.0016 0.0015 0.0015 0.0015 0.0018
902 912 915 901 912
1656 1674 1679 1654 1674
932 1007 936 914 1007
1710 1845 1717 1677 1845
Chapter 5: Brazing Filler Metals / 211
is formed at the contact point. However, there is a need to replace filler metals that contain precious metals for use in these applications. A group of scientists at the Materials Research Institute (Ref 102) have shown an interest in eutectic zinc-aluminum and zinc-germanium zinc-base filler metals that have eutectic melting points at 382 and 398 °C (720 and 748 °F), respectively. Their research showed that tapes obtained from the Zn-Al-Ge filler-metal system may be used in semiconductor devices and integrated circuits as a multilayered tape overlaid coating for brazing silicon crystals, as well as localized coatings (using foil electrical detonation) and also as liners.
Specialized Brazing Filler Metals and Materials Brazing materials for high-temperature applications can be selected from ceramic, refractory-compound, or metallic systems. Ceramic brazes are typically glassy oxides and are not suitable for carbon structures at elevated temperatures because of the reaction of oxygen and carbon to form carbon monoxide gas. Refractory-compound brazes for ultrahigh-temperature use are typically refractory silicides, borides, and/or carbides. Metallic filler metals can be composed of noble metals, active metals, refractory metals, or combinations of these. Because of the unlimited available carbon in carbon-carbon composite materials, graphites, and carbons, a eutectic braze of a metal carbide plus graphite is also possible for use at ultrahigh temperatures. Graphite is not readily wet by most conventional filler metals. Filler metals used to join graphite should contain strong carbide formers, because the bonding mechanism depends on carbide formation. Carbides, Composites, and Ceramics. Noble-metal filler metals can be made from combinations of gold, silver, platinum, palladium, cobalt, nickel, and copper. They generally do not wet graphite, but if the graphite is pretreated with an active or refractory metal to form carbides, then these filler metals may wet the graphite (carbide) surface and form a satisfactory joint. Table 5.16 lists several commercially available filler metals that may be suitable for brazing a surface pretreated with an active or refractory metal to form carbides.
Active filler metals can be formulated by combining varying amounts of the carbideforming metals with the noble filler metals in Table 5.16. Potential moderate- and high-temperature filler metals containing active and/or refractory metals are listed in Table 5.17. These filler metals have potential use on carbon-carbon composite materials. Researchers (Ref 103) performed brazing experiments at 750 °C (1380 °F) for 2 h between Ag-Cu-In-Ti filler metal and SiCw/alumina. They found that the first clearly nonbraze layer consisted of an oxide layer of metallic composition 33Ti-31Al-22Cu-14Si. Areas adjacent to the SiC whiskers were of a different composition. A thin, continuous layer on the alumina portion of the composite appeared to be γ TiO. The SiC whiskers were preferentially consumed and underwent reductions in diameter of ~40%. Observed “knobby” whisker morphologies may be related to SiC stacking faults. The η-type phases detected near the silver-
Table 5.16 Commercially available noblemetal filler metals that may wet pretreated graphite
Filler Metal
Re Ru Rh Pt Pt-40Ir Pt-40Rh Pt-20Pd-5Au Pt-60Cu Pd Pd-70Ni Pd-36Ni-10Cr Pd-40Ni Pd-35Co Pd-82Cu Au Au-35Pd Au-25Pd Au-13Pd Au-34Pd-36Ni Au-25Pd-25Ni Au-65Cu Au-62.5Cu Au-18Ni Ag-33Pd-3Mn Ag-20Pd Ag-27Pt Ag-7.5Cu Ni-4.5Si-3.1B Ni-4Si-2B-1Fe Ni-23Mn-7Si-4Cu Cu
Solidus temperature
Liquidus temperature
°C
°F
°C
°F
AWS designation
3180 2500 1968 1769 1950 1935 1645 1200 1552 1290 1232 1238 1230 1080 1063 1427 1380 1260 1135 1102 990 990 950 1149 1070 995 779 982 982 982 1082
5756 4532 3574 3216 3542 3515 2993 2192 2826 2354 2250 2260 2246 1976 1945 2601 2516 2300 2075 2016 1814 1814 1742 2100 1958 1823 1435 1800 1800 1800 1980
3180 2500 1968 1769 1990 1950 1695 1250 1552 1320 1260 1238 1235 1090 1063 1440 1410 1305 1169 1121 1010 990 950 1232 1160 1160 891 1038 1066 1010 1082
5756 4532 3574 3216 3614 3542 3083 2282 2826 2408 2300 2260 2255 1994 1945 2624 2570 2381 2136 2050 1850 1814 1742 2250 2120 2120 1635 1900 1950 1850 1980
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... BAu-3 BAu-1 BAu-4 ... ... ... BAg-19 BNi-3 BNi-4 BNi-8 BCu-1
AWS, American Welding Society
212 / Brazing, Second Edition
copper eutectic portion of the joint appeared to consist of titanium-copper-aluminum-siliconoxygen and Ti3Cu3O. Etching appears to be a useful supplementary technique in the characterization of ceramic-metal joints (Fig. 5.4). Several Japanese scientists (Ref 104) developed a paste containing molybdenum and TiN powders that were printed on AlN substrates. The TiN-Mo does not adhere to the grainboundary phase in the AlN substrate nor to the surface oxide layer but to the AlN grain itself. This method, therefore, seems to be applicable to any kind of AlN substrate that may have different grain-boundary oxide phases and thermal conductivities. This TiN-Mo metallized AlN substrate was tried as a replacement for a BeO heat sink, which has been used for radio frequency power transistors. There was no trouble in assembling the AlN heat sinks into transistors. Thermal resistance and electrical properties for transis-
tors with AlN heat sinks were almost equal to those with BeO heat sinks. The TiN-Mo metallized AlN substrates were found to be suitable replacement BeO substrates as the heat sinks for semiconductor devices. An active filler metal, such as Ti- Ag- Cu, has been used to adhere AlN ceramics to metal (Ref 105), using the previously mentioned AlN-Mo metallizing method. Intermetallics. Researchers (Ref 106) successfully used planar-magnetron-sputtered (PMS) silver interlayers in conjunction with uniaxial solid-state bonding techniques to join a high-coefficient-of-thermal-expansion (CTE) material, alloy 718, to single-crystal silicon, Si(100), as well as to alumina (Al2O3). The CTE difference between alloy 718 and Si(100) of approximately 10 × 10–6/°C (18 × 10–6/°F), combined with the brittleness and reactivity of the silicon, was considered to be a worst-case test for the use of compliant interlayers when joining
Table 5.17 Potential filler metals containing active and/or refractory metals Liquidus temperature Filler-metal composition
Pd-Mo Pt-55Mo Pt-30W Co-Cr-Si-Ni Mo-Ru Ta-Ti-Zr Ta-V-Nb Ta-V-Ti Ta-25Cr Ta-66Cr Ti-Zr-Nb Ti-V-Nb Ti-V-Mo Ti-V-Cr W-Cr Mo-Cr Co-Cr-Ni-Si-W Ni-Cr-Si-Fe-B Ni-Cr-Si-Fe-B Ni-Cr-Si Ni-Cr-P Ti-15Cu-15Ni Ni-Cr-Si-Fe-B Ni-W-Cr-Si-Fe-B Co-Cr-Ni-Si-W-B Au-Ni-Mo Ti-48Zr-4Be Ti-28V-4Be Zr-19Nb-6Be Au-19/60Cr Cu-25/68Cr Cu-25/68Ti
Solidus/eutectic temperature
°C
°F
°C
°F
1455–1760 2205 2300 1900 1900 2095 1815–1925(a) 1760–1845(a) 1980 1700 1600–1700(b) 1650(b) 1650(b) 1540–1650(b) 1870–3370 1815–2620 1150 1040 1075 1135 890 960 1160 1095 1150 ... (c) (d) (e) (f) (f) (f)
2650–3200 4000 4170 3450 3450 3800 3300–3500(a) 3200–3350(a) 3596 3092 2912–3092(b) 3000(b) 3000(b) 2800–3000(b) 3400–6100 3300–4750 2100 1900 1970 2075 1630 1760 2120 2000 2100 ... (c) (d) (e) (f) (f) (f)
1370–1760 2080 2175 ... ... ... ... ... 1980(b) 1700(b) ... ... ... ... ... ... 1120 975 975 1080 890 910 970 970 1105 925 min (1540) (1606) (1745) 1162 1075 880 min
2500–3200 3780 3950 ... ... ... ... ... 3596(b) 3092(b) ... ... ... ... ... ... 2050 1790 1790 1975 1630 1670 1780 1780 2025 1700 min (2800) (2923) (3170) 2124 1970 1620 min
(a) Liquidus temperature will vary with composition, and the remelt temperature was higher when tested using molybdenum (TZM, Mo-0.5Ti-0.1Zr) coupons. (b) The remelt temperature will be raised as carbides are formed. (c) Beryllium is a strong melting-point depressant used to minimize the braze temperature; if Be is eliminated, the solidus becomes 1543 °C (2809 °F). (d) As in (c), if the Be is removed, the solidus becomes 1606 °C (2923 °F). (e) As in (c), if the Be is removed, the solidus becomes 1742 °C (3168 °F). (f) Liquidus temperature will vary with composition.
Chapter 5: Brazing Filler Metals / 213
materials with differing CTEs. As-cast titanium aluminide (TiAl) was also bonded to itself using PMS silver interlayers to show the usefulness of the technique when joining materials that are inherently difficult to join by standard methods. All materials evaluated were successfully joined at temperatures of 500 °C (930 °F) or less using a silver interlayer to facilitate bonding. However, the alloy 718/Si(100) and alloy 718/Al2O3 bonds could not survive thermal cycling from room temperature to temperatures much above the bonding temperature. Ultrasonic pulse-echo techniques revealed that both these bonding systems were severely degraded after thermal cycling to 800 °C (1470 °F). The bond strength of the TiAl joints exceeded the fracture strength of the base material (240.3 MPa, or 34.9 ksi). Attempts were made with these bonds to dissolve the silver interlayer after bonding by appropriate heat treatments. Results showed very little bulk diffusion of silver into the TiAl, with the majority of silver diffusion occurring along the TiAl grain boundaries. Solid-state bonding using compliant silver interlayers appears to have potential as a viable joining technique when low joining temperatures are desired and may be significantly improved on by incorporating hot isostatic pressing methods into the procedure. Metals and Ceramics. Researchers (Ref 107) increased the debonding strength of metalceramic joints by 18 to 28% with the addition of 8.4 vol% short-metal-coated carbon fibers to an active filler metal. The filler metals used were active filler metals in paste and sheet forms. The filler-metal paste (type Cusin-1 ABA, Wesgo Metals) contained 63 wt% Ag, 34.25 wt% Cu, 1.75 wt% Ti, and 1.0 wt% Sn (the melting temperature range of which is 780 to 815 °C, or 1435 to 1500 °F).
Ag-rich phase
Removed by etchant
Cu-rich phase
η-type phases 1)Ti 3 Cu 3 O 2)Ti-Cu-Al-Si compound
TiO Al 2 O 3
Fig. 5.4
SiC
Al 2 O 3
Schematic representation of the reaction zone between SiCw/Al2O3 and Incusil (Wesgo Metals) active braze alloy filler metal. Source: Ref 103
The filler metal in sheet form (0.1 mm, or 0.004 in., thick) was CB2 and contained 96 wt% Ag and 4 wt% Ti (the eutectic temperature of which is 970 °C, or 1780 °F). The carbon fibers were pitch-based (Thornel P-100, BP Amoco Chemicals), which helped to strengthen the filler metal and to slightly decrease the thermal stress at the brazing interface. The carbon fibers were either uniformly distributed in the brazing layer or concentrated near the ceramic side of the metal-ceramic brazing interface. The latter resulted in a lower thermal expansion in the part of the filler metal near the ceramic and gave superior joints, such that the debonding occurred in the part of the filler metal without carbon fibers. The titanium in the active filler metal was segregated at the interfaces between the filler metal and the ceramic, between the filler metal and the metal (steel), and between the carbon fibers and the matrix of the filler metal. The amount of titanium at the interface between the filler metal and the ceramic was smaller when carbon fibers were present in the filler metal. The bare carbon fibers gave joints comparable in quality to the metalcoated carbon fibers. The carbon fibers also served to lower the cost of the filler metal. Si3N4 ceramics with Al2O3 and Y2O3 as additives were joined with an 80Ni-20Cr (wt%) filler-metal sheet as an insert layer. Joining was performed by hot pressing between 1000 and 1350 °C (1830 and 2460 °F) in argon and under uniaxial pressures in the range of 50 to 100 MPa (7 to 14.5 ksi). The average strength, evaluated by four-point bending, was large enough (>300 MPa, or 43.5 ksi) for some industrial applications (Ref 108). However, the scatter of the joint strengths was large. It was probably due to the formation of pores at the joining interface or near the surface. The influence of nitrogen gas partial pressure in the joining atmosphere and uniaxial pressure on the formation of pores at the joint interfaces was confirmed microscopically. Chromium coating the Si3N4 ceramic before joining was effective in reducing the scatter of joint strengths. The oxidation resistance of the joint was excellent up to 800 °C (1470 °F) in air (Ref 108). Researchers (Ref 109) have been investigating the eutectic bonding of nickel to 3 mol% yttria-stabilized zirconia (YSZ), using the eutectic melt of nickel and nickel oxide that exists at 1440 °C (2625 °F). This eutectic point is 15 °C (25 °F) below the melting point of pure nickel. In fact, it is the liquid eutectic that is used as an
214 / Brazing, Second Edition
intermediate layer to bond the solid nickel to the ceramics without the nickel member losing its original shape. The major advantage of the eutectic bonding technique is the excellent wettability of the nickel/nickel oxide eutectic liquid on the ceramic substrate (Ref 110, 111). Liquid eutectic bonding can be used to join zirconia and nickel foil. This liquid phase relaxes the stress caused by differential shrinkage between the metal and the ceramics during bonding and also wets the interface well. On solidification of the eutectic melt, a strong bond is established. The strength is higher than that without the eutectic layer. The eutectic bonding employed proved to be superior to the pure nickel-YSZ bonding at 1450 °C (2640 °F) and to NiO-YSZ bonding in air at 1450 °C (2640 °F) (Ref 109). Rapid Solidification (RS) Technology. Several recently published reports have renewed investigations on the wetting and spreading of nickel-phosphorus filler metals and the wettability of nickel filler metals with boron using RS Metglas brazing foil (Honeywell
International Inc.) (MBF) 35, 1005, 80, and 85 (Table 5.18) (Ref 112, 113). Many nickel-base filler metals themselves contain significant proportions of phosphide and boride phases and thus tend to be brittle. Until recently, this has meant that the compositions of nickel-base filler metals that could be fabricated as foils and wire were extremely limited. Instead, the compositions that are brittle were applied in the form of pastes, that is, alloy powder in a fluid binder. Organic binders burn during the brazing cycle, which tends to introduce voids and carbonaceous residues that weaken joints. This restriction has been overcome by the development of RS casting technology and liquid metal quenching. Because of the low melting point of the filler metals in relation to that of the constituent elements, particularly when the filler metal contains boron and/or phosphorus, and because during RS the cooling rate typically exceeds 105 °C/s (2 × 105 °F/s), these nickel-base filler metals generally solidify with an amorphous structure. Selected filler metals that are available commercially as amor-
Table 5.18 Examples of commercially available rapidly solidified filler metals Melting range Composition, wt%
°C
°F
Structure
Typical applications
779 775–790 577 278 363 139 640–700 610–645 714 770–925 1120–1150 880 940–990 970–1075 1020–1065 712–745
1432 1425–1455 1071 535 685 282 1185–1290 1130–1195 1317 1420–1695 2050–2100 1615 1725–1815 1780–1965 1870–1950 1314–1375
Microcrystalline Microcrystalline Microcrystalline Amorphous Microcrystalline Microcrystalline Amorphous Microcrystalline Amorphous Microcrystalline Amorphous Amorphous Amorphous Amorphous Amorphous Amorphous
Pb-62Sn Pb-5In-2.5Ag Pd-38Ni-8Si
183 300 830–875
361 570 1525–1605
Microcrystalline Microcrystalline Amorphous
Sn-3.5Ag Sn-25Ag-10Sb Ti-15Cu-15Ni Ti-20Zr-20Ni Zr-17Ni Zr-16Ti-28V
221 240–290 902–932 848–856 961 1193–1250
430 465–555 1656–1710 1558–1573 1762 2179–2280
Microcrystalline Microcrystalline Amorphous Amorphous Amorphous Amorphous
Most engineering materials Engineering ceramics Aluminum alloys Sealing electronics packages Microelectronic die attach Hermetic solder seals Copper alloys and mild steel Copper alloys and mild steel Copper alloys and mild steel Copper alloys and mild steel Cobalt-base alloys and superalloys Steels, stainless steels, superalloys Steels, stainless steels, superalloys Steels, stainless steels, superalloys Steels, stainless steels, superalloys Steels, stainless steels, superalloys, cemented carbides Electronics systems fabrication Electronics systems fabrication Stainless steels, superalloys, cemented carbides Electronics systems fabrication Sealing electronics packages and die attach Superalloys and engineering ceramics Superalloys and engineering ceramics Titanium-base alloys Titanium-base alloys
Ag-28Cu Ag-28Cu-5Ti Al-13Si Au-20Sn Au-3Si Bi-43Sn Cu-10Mn-30Sn Cu-10Ni-4Sn-8P Cu-8P Cu-20Sn Co-19Cr-19Ni-8Si-1B Ni-10P Ni-32Pd-8Cr-3B-1Fe Ni-14Cr-5Si-5Fe-3B Ni-15Cr-3B Ni-41Pd-9Si
Source: Ref 1
Chapter 5: Brazing Filler Metals / 215
phous strip are included in Table 5.18. This family of filler metals covers melting temperatures that range from approximately 700 °C (1290 °F) to more than 1100 °C (2010 °F). Although the nickel-base filler metals have particularly benefited from RS technology, other filler metals that are normally brittle when cast conventionally have also been upgraded by this preparation technique. These include filler metals based on copper, palladium, and cobalt, with the copper filler metals being developed as cheaper alternatives to the nickel-bearing filler metals and the palladium- and cobalt-base filler metals for more demanding application environments (Ref 3, 114, 115). The main fillermetal compositions that have been commercially developed in the form of RS foils, wire, and powder are listed in Table 5.18. These are produced with either an amorphous or a microcrystalline microstructure. The use of this casting technology for producing foil and wire preforms not only allows filler metals with new compositions to be manufactured but also confers a number of associated benefits. First, the filler metal is comparatively ductile in the amorphous state, because there are no discrete phases or grain boundaries in the microstructure that might be sources of embrittlement. On heating an amorphous filler metal to approximately half its melting point in degrees Kelvin, the microstructure will revert to a crystalline form. After melting and solidification, the braze fillet will have a conventional microstructure. However, there is strong evidence in the published literature that the microstructure of joints made with filler metals prepared by RS are finer and more uniform than joints made with conventionally prepared filler metals. The reasons for this are not fully understood but are likely to be associated with a degree of atomic ordering in the liquid phase. Evidence for this explanation comes from the gray/white allotropic transformation of tin. Second, the absence of a dendritic microstructure in both amorphous and microcrystalline metals means that a RS alloy is metallurgically homogeneous, as cast, akin to a solid solution. The filler metal therefore melts in a highly uniform manner, which helps to minimize local fluctuations in the erosion of the parent metals due to an absence of segregated phases in the filler metal. Third, filler metals prepared by RS tend to be
cleaner both on the surface and within the bulk, because there is only a single processing step and that is usually carried out in a protective atmosphere. Metallic contamination from rolling and drawing machinery is eliminated, as is carbonaceous contamination from lubricating fluids emanating from the machinery. Finally, the homogeneous filler-metal preforms have a much narrower melting range compared with the equivalent composition filler metal prepared as a trifoil or clad wire, comprising a core of titanium with a cladding of the other constituents. Thus, the Ti-15Cu-15Ni filler metal has a melting range of 902 to 932 °C (1656 to 1710 °F) as a homogeneous filler metal and 912 to 1007 °C (1674 to 1845 °F) when prepared as a trifoil in the form of a titanium layer sandwiched between foils of Cu-50Ni and heated at 10 °C/min (18 °F/min) (Ref 1). When using nickel-base filler metals containing metalloids, there is a tendency for embrittling nickel-metalloid compounds to precipitate in the joint if the width of the braze is more than approximately 50 µm (2 mils). Where wide joints are made, this problem may be avoided by fitting into the joint gap a porous shim of a metal that will rapidly soak up the filler metal and dilute the metalloid. This technique is particularly useful in repair work (Ref 1–3, 116, 117). There are many other families of filler metals. For example, families of Cu-Mn-Sn and Cu-NiSn-P filler metals have been developed as inexpensive alternatives to silver-containing filler metals. These have melting ranges and physical properties in joints that are similar to those of the Ag-Cu-Zn-base filler metals, and the phosphorus-containing filler metals are self-fluxing when used to join copper and steel, because the phosphorus is present at a concentration of approximately 8%. Other filler-metal families, including alloys of magnesium, cobalt, titanium, zirconium, and tungsten, will not be dealt with here, because they are not widely used and their behavior as filler metals is largely similar to that of the filler-metal compositions in Table 5.18; in addition, the metallurgical principles on which they were developed are not significantly different. Nevertheless, it should be borne in mind that these filler metals are offered to the manufacturer because they provide specific property advantages for certain applications (Ref 118–125).
216 / Brazing, Second Edition
Researchers (Ref 118) joined Si3N4 to a 1.25Cr-0.5Mo steel using a (Cu, 5–25Ni, 16–28Ti, traces of B) filler metal (HTB2) in the form of RS foils. The maximum joint strength (three-point bend) at room temperature was 261 MPa (38 ksi). The value was maintained until 723 K (268 MPa, or 39 ksi). As the test temperature was raised, the joint strengths decreased. By means of a SEM with a wave-dispersive spectrometer, the interfacial metallurgical behavior between the filler metal and Si3N4 or the interlayers and its effects on the joint strength was studied. It was found that when the nickel platelet was employed as the buffer layer next to the Si3N4, it was difficult to improve the joint strength, but if the steel platelet was employed as the interlayer instead of nickel, the joint strength could be greatly augmented. Researchers (Ref 119) reported on a group of new high-chromium-containing amorphous filler metals in foil form that has been developed for applications in highly corrosive and/or hightemperature environments. These new filler metals (Table 5.19) contain 10 to 16 wt% Cr; 1.2 to 1.6 wt% B, that is, the minimal amount needed for amorphability; 7.0 to 7.5 wt% Si; 0 to 5 wt% Mo; and nickel as the balance. Their solidus temperature is within 975 to 1030 °C (1785 to 1885 °F), and their liquidus temperature is within the 1090 to 1140 °C (1995 to 2085 °F) range. The filler metals exhibit low erosion of base metal and show no significant detrimental effects on base-metal strength, because boron concentration is kept to a minimum. It was found that optimal combination of strength, ductility, and fatigue resistance of joints is obtained when brazing is combined with hightemperature annealing in one extended cycle. The joints after such treatments are practically free of the brittle central eutectic line, and the base metals adjoining the braze area have only limited and localized segregation of chromium borides. The joint itself is uniform and has a strong and ductile microstructure, with only one nickel-chromium-base solid-solution phase. The fracture of brazements occurs predominantly in the base metal, with the joint ultimate strength higher than the yield strength of the virgin 316L base metal. The brazements have a high corrosion resistance in seawater and water solutions of ammonia and phosphoric acid. This new series of filler metals is currently produced on a regular basis by RS technology as a ductile filler-metal foil having up to 200 mm (8 in.) width and 50 to 60 µm thickness. The foil has
already been used successfully in brazing of hundreds of large, corrosion-resistant 316L heat exchangers. Researchers (Ref 120) found dramatic changes in joint width, microstructure, and Charpy impact energy of copper-to-copper joints torch brazed under load with copper-phosphorus-base Metglas MBF-2005 and BCuP-5 filler-metal foil. It was shown that brazement thickness decreases under low/medium loading, up to a few dozen pounds per square inch, reaching approximately 20 µm. This shrinkage was due to the outflow of filler metal, which began to liquefy early in the brazing cycle. At this point, increasing the load could not promote further liquid expulsion. The decrease in volume of retained liquid and, particularly, the removal of low-melting phosphides from the liquid filler metal resulted in the crystallization of a new ductile joint microstructure. The central eutectic zone, which is present in many conventional brazements and is responsible for joint brittleness, disappeared. The amount of ductile copper-base phase also increased drastically. The major result of brazing under load was substantial—up to a tenfold increase in Charpy impact energy, with particularly strong effects being achieved on MBF2005 joints. A model was proposed that explained reasons for the observed enhancement in joint strength, which is valid for similar filler-metal systems. According to this model, some of the liquid phase, enriched in low-melting constituents, was ejected out of the brazing joint. Rapid saturation of the remaining filler metal by ductile base-metal constituents then occurred, followed by crystallization into a ductile, strong structure. From a practical point of view, loading of joints is beneficial for a wide variety of base and filler metals. Substantial increase in joint strength was achieved using a relatively small load. Brazing of electrical contacts, cemented carbide tips of cutting tools, and so on were the first applications where this technique could be easily implemented. Researchers (Ref 121) were able to produce RS ribbons of silver-base BAg-series filler metals in brittle and ductile states over a wide range of compositions. The brittle fracture mode observed in these normally very ductile materials was due to dispersion hardening resulting from the appearance of RS metastable phases and crystalline morphologies in thin (30 to 50 µm) strip. Both states may have an important
Cr
... 13.0 AWS BNi-2/AMS 4777 7.0 AWS BNi-3/AMS 4778 . . . AWS BNi-5a 19.0 ... 5.3 AWS BNi-6 ... ... 15.2 ... 15.0 ... 10–16
AWS and AMS classifications
4.2 3.0 ... ... ... ... ... ... ...
Fe
4.5 4.5 4.5 7.3 7.3 ... ... 7.25 7.2–7.4
Si
0.03 0.06 0.06 0.08 0.08 0.10 0.06 0.06 0.06
C(a)
2.8 3.2 3.2 1.5 1.4 ... 4.0 1.4 1.4
B
... ... ... ... ... 11.0 ... ... ...
P
Nominal composition, wt%
1.0(a) ... ... ... ... ... ... ... ...
Co
bal bal bal bal bal bal bal bal bal
Ni
965 969 984 1052 950 883 1048 1030 975–1030
°C
°F
°F
2017 1875 1929 2091 1904 1690 1996 2059 1994–2066
Liquidus
1103 1024 1054 1144 1040 921 1091 1126 1090–1130
°C
Melting temperature
1769 1776 1803 1926 1742 1621 1918 1886 1787–1886
Solidus
Metglas is a trademark of Honeywell International Inc., AWS, American Welding Society; AMS, Aerospace Material Specification. (a) Maximum concentration. Source: Ref 119
15 20 30 50 55 60 80 51 5x
Metglas brazing foil alloy
Table 5.19 Nickel-base amorphous filler metals
1135 1005 1085 1170 1070 950 1120 1195 ...
°C
g/cm3
Density
2075 7.51 1840 7.46 1985 7.94 2140 7.49 1960 7.72 1740 7.91 2050 7.80 2183 7.51 . . . 7.49–7.5
°F
Brazing temperature (approx.)
Chapter 5: Brazing Filler Metals / 217
218 / Brazing, Second Edition
practical application as powder and foil filler metals. Recently, a new amorphous filler metal appeared on the market for joining applications where silver-containing materials, particularly Cu-15Ag-5P filler metal having the BCuP-5 designation, had been used for a long time. This Cu5.8Ni-9.2Sn-6.4P quaternary filler metal, designated as Metglas MBF-2005, has good melting and flow characteristics, contains no silver, and has joint resistivity similar to or better than that of BCuP-5. Among potential applications where the new filler metal may be beneficially used, the brazing of copper bus bars for distribution and power transformers is of particular importance. Bus bars have a service life of close to 30 years. Under normal conditions, service temperatures are below 150 °C (300 °F). However, it is assumed that, in some cases, approximately half a year of service time at temperatures of approximately 280 °C (535 °F) can be accumulated during a transformer lifetime. It is desirable that the resistance and strength of the joints would not change substantially after such temperature exposure. So far, there have been no complaints about the aging behavior of BCuP-5 filler metal. Because this filler metal was in service for a long time, there is no standard accelerated testing procedure to follow in order to evaluate a projected change in properties of joints manufactured with a new filler metal (Ref 122). The use of the new filler metal resulted in a more stable brazement than that made with BCuP-5. The changes in resistivity and the Charpy impact energy of the former are less rapid. The possible reason for such a result is the presence of nickel in the new filler metal. Nickel has a few times larger enthalpy of phosphide formation than either copper or silver, thus decreasing the rate of phosphorus dissolution into a pure copper-base material (Ref 123). In conclusion, it is worthwhile to point out (Ref 122) that the new amorphous alloy used as a filler metal results in joints more stable than that of BCuP-5. Because there have been no complaints about the aging of joints manufactured from BCuP-5 over the decades, there should be no such problems as the appearance of joint brittleness and the decrease of conductivity when using the MBF-2005 filler metals. Researchers (Ref 124) reported that titaniumpalladium, ASTM grade 7 and Ti-6Al-4V alloys were brazed in a vacuum furnace, each to itself, by using a new RS amorphous 25Ti-25Zr-50Cu filler-metal foil. Joint tensile strength, fatigue
resistance, and microstructure were determined, the latter by x-ray diffraction analysis, scanning electron microscopy/energy-dispersive spectroscopy, (SEM/EDS), and scanning transmission electron microscopy (STEM) methods. Joint tensile strength was close to that of each base metal. Fatigue properties of titanium-palladium, grade 7 joints did not differ from those of this base metal. The microstructure and mechanical properties of the brazed joints were dependent on the brazing cycle conditions: a fine lamellar eutectic joint microstructure, consisting of α-titanium and γ-[Ti(Zr)]2Cu-tetragonal MoSi2-type phase, was observed after brazing of Ti-6Al-4V alloy at 900 °C (1650 °F) for 10 min, followed by fast cooling. Such brazing operations resulted in high-strength joints. Brazing at temperatures higher than 900 °C (1650 °F) and/or with a relatively low cooling rate resulted in a coarse dendritic microstructure consisting of γ-[Ti(Zr)]2Cu and hexagonal λ-laves Cu2TiZr phases. Finally, it was shown that fast cooling suppresses formation of λ-laves brittle phase, thus resulting in high mechanical properties of the brazed joint. In a subsequent program, researchers (Ref 125) employed the 25Ti-25Zr-50Cu amorphous filler metal, using argon as a shielding gas, to induction braze the Ti-6Al-4V alloy. The brazing cycles were rather short: radio frequency inductor power was supplied for only 40 to 60 s. The tensile strength and ductility of the brazed joint are at the level of that of Ti-6Al4V base metal. According to scanning electron microscopy/energy-dispersive x-ray analysis and scanning transmission electron microscopy phase analysis, two main joint microstructures are observed: either a very fine lamellar/cellular eutectoid consisting of a mixture of α-Ti+γ(Ti,Zr)2Cu tetragonal MoSi2-type phase, or martensitic α-titanium with dispersed γ-phase. The partial presence of a fine, not coarse, Widmanstätten microstructure in the joints did not seem to detrimentally affect joint mechanical properties. The final copper concentration in the joint area was considered as one of the critical parameters for joint microstructure formation. The optimal eutectoid microstructure characteristic of a ductile joint was obtained under brazing conditions that resulted in average copper joint concentrations within the 10 to 12 wt% range. A variety of metallurgical paths along which base and filler metal may interact were proposed, explaining the mechanisms of formation of vari-
Chapter 5: Brazing Filler Metals / 219
ous braze microstructures and the mechanical properties related to them. From the practical point of view, it was proven that induction brazing of compatible samples located in a simple closed chamber may be carried out as an effective and inexpensive process and have a substantial advantage over vacuum furnace brazing, because the latter results in joints having a high strength but, unfortunately, poor ductility. Various metallurgical paths of Ti-6A1-4V base-metal interaction with RS 25Ti-25Zr-50Cu filler metal have been proposed, taking into consideration the construction of the ternary Ti-ZrCu paternal phase diagram and the schematic time-temperature-transformation diagrams of binary titanium-copper alloys. These paths explain the appearance of observed microstructures and related mechanical properties.
Filler-Metal Selection In choosing a filler metal for a specific brazing application, the following information should be taken into consideration. • The base metals being joined • The method of heating to be used. Filler metals with narrow melting ranges of less than 25 °C (45 °F) between solidus and liquidus can be used with any heating method, and the filler metal may be preplaced in the joint area in the form of rings, washers, formed wires, shims, powder, or paste. Alternatively, such filler metals may be manually or automatically face fed into the joint after the base metal is heated. Filler metals that tend to liquate are used with heating methods that bring the joint to brazing temperature quickly or allow introduction of the filler metal after the base metal reaches the brazing temperature. • The brazing temperature required. Low brazing temperatures are usually preferred to economize on heat energy, to minimize heat effects on the base metal (annealing, grain growth, warpage, etc.), to minimize filler-metal/basemetal interactions, and to increase the life of fixtures and other tools. High brazing temperatures are preferred in order to take advantage of a higher-melting, but more economical, filler metal; to combine annealing, stress relief, or heat treatment of the base metal with brazing; to permit subsequent processing at elevated temperatures; to promote filler-metal/basemetal interactions that increase the joint-
remelt temperature; or to promote removal of certain refractory oxides by vacuum or an atmosphere. • Service requirements of the brazed assembly. Compositions should be selected to suit operating requirements, such as service temperature (high or cryogenic), thermal cycling, life expectancy, stress loading, corrosive conditions, radiation stability, and vacuum operation. Differential thermal analysis (DTA) can give valuable information about a material as it is heated and cooled over a range of temperatures. Testing can reveal an endothermic reaction, which is an indicator of melting, or an exothermic reaction, which is an indicator of solidification. Both of these reactions are important, especially with filler metals, in predicting the onset of liquidus and solidus properties of a material. The data generated by these tests could be used as a quality-assurance tool by the customer, and there are those who feel manufacturers should certify the solidus and liquidus points of their materials through DTA testing. A problem with this, though, is the controversy as to whether DTA can be used to predict the exact melting and solidifying temperatures of filler metals. There is lack of agreement among scientists as to what procedures should be followed in melting and solidifying test samples, how thick the test samples should be, and the interpretation of the thermal curves generated by the analysis. There is also disagreement as to whether the test samples should be in solid or powder form. Because of the oxidation that may occur when heating powder, as well as the nonuniformity of heat transfer, the results when testing with the powder form may be inconsistent. There have been some who advocate melting the powder into a solid sample and then testing it, but the objection is that this premelted sample is not what the customer is actually using. It is agreed that standards should be established that will ensure reproducibility of results the world over. Only after standards are established will DTA become a universally accepted means of quality assurance (Ref 126).
Filler-Metal Forms Filler metals are available as rod, ribbon, powder, paste, creams, wire, sheet, and pre-
220 / Brazing, Second Edition
forms (stamped shapes, washers, rings, or shaped wires) shaped to fit a particular part. Depending on the joint design, heating method, and level of automation, the filler metal can be preplaced before the heating cycle starts or face fed after the work is heated. High-production brazing, such as furnace, flame, or induction brazing, which typically involves a high level of automation, usually requires preplacement of the filler metal. Rod and wire forms are usually used for manual face feeding. Many special shapes and forms designed for specific applications are used as preplaced or preformed filler metals. Such preplacing of the filler metal ensures that there is a uniform amount of filler metal in the correct position on each assembly. Preplaced filler metal may be useful with manual brazing but is usually a fundamental part of any mechanized brazing procedure. Where joint areas are large, filler metal may be located between the faying surfaces. Brazing rings are sometimes inserted into grooves machined into the work. Totally enclosed rings in such grooves may be necessary for long sleeve joints or for salt bath dip brazing, where it is desirable to avoid melting of the ring before the work is heated to the brazing temperature. Although the use of preforms or automatic filler-metal feed is virtually mandatory for mechanized brazing, there is still a wide choice in the form in which to apply the filler metal. The most suitable form for any particular application must be decided on the basis of the following factors: • Joint design (size of assembly, depth of braze joint, cross section of components, joint complexity) • Heating method • Desired degree of automation • Desired appearance of completed assembly • Number of assemblies to be manufactured • Range of assembly types to be manufactured Competitive forces have pressured metalworking manufacturers to reevaluate their manufacturing processes. As a result, demands on brazing operations have increased to improve efficiency, produce higher-quality products, and reduce costs. There are a number of factors that can affect the performance of brazing operations. Of particular importance is the proper selection of the correct form of filler metal.
There are three general categories of filler metals in production applications. They are: • Nonfabricated wire and strip: available on spools that contain a specified amount of filler metal at a standard size • Fabricated wire and strip forms: Commonly known as preforms, this filler metal can be fabricated as rings, washers, discs, shims, or other engineered shapes. • Paste filler metals: a combination of atomized filler-metal powder, a neutral binder, and, depending on the heating method to be used, a flux. The binder is used to keep the components of the paste in suspension and to facilitate dispensing. Each of these filler-metal forms has its specific advantages and disadvantages. Preforms. Of all the filler-metal forms, preforms offer the most precise control over the amount of filler metal placed in a joint; such control can reduce excess filler-metal consumption. For example, a wire preform 1.6 mm (0.06 in.) in diameter in the form of a ring with a 25.4 mm (1 in.) inside diameter contains more material than the joint needs, as evidenced by a large fillet in the joint area. Reducing the wire diameter to 1.4 mm (0.05 in.) and retaining the 25.4 mm (1 in.) inner ring diameter reduces the amount of material used by 13%. Where workers are applying excessive amounts of filler metal, the cost of preforms will likely be less than hand application of bulk filler metal. Higher production rate, improved joint quality, and reduced cost of postbraze cleanup contribute to further cost reductions. Preforms can be buried in the joint to improve inspectability. A preform placed in the bottom of a joint will melt and flow through the joint area, forming a fillet at the top of the joint. The operator can see that the filler metal has flowed through the joint area completely (Fig. 5.5). If the filler metal is applied to the top of the joint, the operator can determine, after heating, that the filler metal has melted and flowed down into the joint but not whether joint penetration is complete. This approach increases inspectability of the joint. Preforms generally produce joints of good appearance. This feature can reduce or eliminate finish machining, which is otherwise necessary to improve the appearance of brazed joints. A problem in the assembly of a line of pressure regulators used by the semiconductor fabrication industry was solved with preforms.
Chapter 5: Brazing Filler Metals / 221
The production of the regulators required the joining of an ultrathin stainless steel diaphragm to a relatively thick actuator of the same metal. However, conventional means of metal joining—welding or powder brazing—were unsatisfactory. Welding required high temperatures of over 1371 °C (2500 °F), which deformed the thin diaphragm. Powder brazing compromised the purity and continuity of the seal due to contaminating residue from the organic binder of the powder. To solve the problem, the use of nickel-base preforms was the solution. In the joining process, 3 mil thick filler-metal foil preforms were placed between the parts to be joined and fixtured by tack welding. The complete assembly was then heated in a vacuum furnace to 1066 °C (1951 °F). The method allowed for the achievement of a leak-free, noncorrosive joint at a substantially lower temperature than that required by welding, and without contaminating residues. The approach also facilitated batch production of hundreds of assemblies at a time. After the assemblies were moved from the furnace, there was no cleanup. Because the filler-metal foils were chemically homogeneous, they produced predictable performance over the life of the project. Most filler metals for high-temperature brazing are eutectic compositions formed by transition elements such as nickel, iron, or chromium in combination with metalloids such as silicon, boron, and phosphorus. When they are composed of the conventional crystalline structure, all these materials are inherently brittle and cannot be produced in continuous forms such as foil and wire. Therefore, they have been available only as powders or its derivatives (Ref 127).
Fig. 5.5
Schematic of filler-metal preform placement
On the other hand, the very presence of metalloids at or near the eutectic concentration promotes RS conversion of such filler metals into a ductile amorphous foil. One of the first practical applications of RS technology was in the production of ductile amorphous filler-metal foil from filler metals having compositions that previously could be used only in powder form or as powder-filled pastes. The most important advantages of RS amorphous microcrystalline filler metals are their flexibility and ductility. Because a ductile amorphous filler-metal foil may be applied as a preplaced preform, large brazement gaps are not necessary (as they are for pastes) to achieve complete filling of the braze cross section. In this case, amorphous filler-metal foil has a particular edge over powder and polymer-bonded strip forms because of its superior flow characteristics. One reason for the superiority of the foil is the fact that gas-atomized powder has a very large total surface area, with resulting large amounts of surface oxides. These oxides prevent, to a certain degree, fusion of individual powder particles into a uniform liquid pool. The RS filler metal flows more freely than any powder form. A smaller clearance also promotes improved retention of bulk metal properties because of curtailed erosion by the smaller volume of filler metal. For these reasons, a preplaced, self-fluxing, thin filler-metal foil preform is superior to powder-containing paste, which requires larger clearances for filling joint cross sections. The paste also results in deleterious effects on properties because of a coarser joint grain size, more fully developed intermetallic compounds, and the presence of substantial amounts of contaminants. High cooling rates (approaching 106 K/s) are used in the manufacture of RS products. This high cooling rate enables the stabilization of certain alloys into an amorphous solid state having a spatial distribution of atoms similar to that of liquids. Most RS filler metals have this amorphous structure, with a random, spatially uniform arrangement of the constituent atoms. Because RS amorphous materials are compositionally much more uniform even after crystallization, they melt over a narrow temperature range under transient heating. This is a consequence of the shorter distances over which atoms of different elements have to diffuse in order to form a uniform liquid phase. The resulting instant melting of RS materials is only one of their important features. This is
222 / Brazing, Second Edition
particularly important when brazing fine-gage honeycomb cores, for example, which have to be protected from erosion by molten filler metals during joining. A shorter brazing time is also beneficial in cases wherein base-metal parts may lose their inherent strength due to annealing during the brazing operation. The joining of cold-deformed stainless steels and of precipitation-hardened superalloys are good examples in which a short brazing time can be critically important. The absence of the residual organic solvent bases evident in powder pastes/tapes correspondingly eliminates soot formation and furnace fouling. The low level of gaseous impurities in amorphous filler-metal foil, due to the specific characteristics of its production technology, is an attractive feature for vacuum furnace brazing. As a consequence of these unique properties, amorphous filler-metal foil has become the preferred advanced filler metal for applications related to the aerospace industry, precise machinery and tools, and modern medical equipment. When brazing parts with amorphous fillermetal foil (MBF), two potential sources of incomplete brazing should be considered. First, gaps may exist between MBF foil preforms placed side-by-side during the product assembly. Second, MBF foil may have small holes. Both the gaps and the holes threaten to leave voids within the brazements. To combat the formation of voids, the gaps and holes must be filled during the brazing process. This may be an important barrier when considering applications with brazed cross sections having both width and length dimensions larger than 200 mm (8 in.). This dimension is a current limit for the width of ribbon produced by today’s planar casting technology. If the gaps and holes are moderate in size, up to 3 mm (0.12 in.) maximum dimension, the liquid filler metal will fill them during brazing. Therefore, multiple preforms may be suitable for joining parts, thus permitting assemblies of virtually unlimited dimensions. As stated previously, one of the most advantageous features of MBF is its superior flow when compared to powder filler-metal products. Wire Forms. In terms of the volume of filler metals used in industry today, nonfabricated filler metals are the most widely used. However, while the cost of this material is probably the lowest of the three basic forms, many of its other associated costs are among the highest.
Many companies use nonfabricated filler metals for hand brazing operations. However, when labor costs are added, hand brazing can be quite expensive compared to automated brazing methods. In addition, research and experience have shown that hand brazing operations will use anywhere from 10 to 40% more filler metal per joint than operations that use preformed filler metal. The repeatability of making brazed joints with hand-fed filler metal is generally lower— sometimes significantly lower—than the repeatability of brazed joints made with preforms or paste. Using filler metals in wire form with automatic wire-feeding equipment may sometimes overcome the problems mentioned. Generally, however, problems do exist, and the hidden costs have a negative impact on both productivity and profitability. When using wire-feeding equipment to braze assemblies, cold and unfluxed wire are applied to components that have been heated. When the filler metal is then introduced, it has a tendency to chill the joint area. The most common way to overcome this potential problem is to overheat the assembly, so that it has enough heat present to bring the filler metal in wire form up to the temperature it needs to melt and flow properly. Because the wire is unfluxed and the assembly is often overheated, many people use liberal amounts of flux to ensure that enough will be left when the brazing actually takes place. This often leads to extensive postbraze cleaning requirements, which is why this type of brazing is most commonly used on simple assemblies that do not require strict cosmetic standards. Researchers (Ref 128) tried a new approach to improving the properties of brazed joints by applying a new, parallel-wire reinforcement, along with an appropriate working procedure, in the brazing of metallic materials. They offered solutions to several problems related to brazed joints. All reactions between constituents of the brazed joint were noteworthy. Joining of the reinforcement to the base metal without a filler metal at the contact face between the reinforcement and the base metal was required. In this case, the reinforcement had a decisive influence on tensile strength, toughness, and resistance to crack propagation in the reinforced brazed joint, because the reinforcement was situated in the brazed-joint plane. Coalescence of the reinforcement with the base metal was obtained in two ways, that is, in some cases by diffusion brazing if the combina-
Chapter 5: Brazing Filler Metals / 223
tion of the filler metal, the base metal, and the reinforcement permitted, and in other cases by diffusion welding of the base metal and the reinforcement. In the latter case, the filler metal was inert as far as the reinforcement and the base metal were concerned. In both cases, application of compressive force was required to permit coalescence of the reinforcement wire with the base metal over a large area. The composite brazed joint, with the new stiff and continuous reinforcement composed of parallel wires, resulted in the desired arrangement of microstructure phases in the brazed joint. The filler metal in the joint was trapped between the individual wires. In spite of the presence of filler metal, the reinforcement strongly affected the mechanical properties of the joint. The influence of the reinforcement was controlled by the width of coalescence of the wires and the base metal. The latter determined the active volume fraction of the reinforcement. Filler Pastes and Dispensers. Preforming filler metals or paste are mandatory in applications where furnaces are used as the heating method. For all their attributes, there are situations where preformed filler metals are at a disadvantage. In low-volume runs involving a wide variety of assemblies, for example, the necessary inventory of custom-manufactured preforms can be difficult to manage. Similarly, the automated dispensing of complex preforms in applications that require total automation can be relatively expensive compared to other forms, such as paste filler metals. Manufacturers of dispensing systems for preform and paste filler metals may provide the final answer. Filler metals in paste form also offer users a unique set of advantages. All the elements required to produce a brazed joint—filler metal and flux—are delivered to the joint at one time as a paste deposit. This feature alone can make paste an attractive filler-metal form for those metalworking manufacturers who wish to automate the placement of their filler metal. The precise amount of filler metal needed in the joint area can be applied with relatively inexpensive equipment (Fig. 5.6). Another application of paste, where automated dispensing has proven superior to foot control (Ref 129), has been achieved through the programmable logic controller.
A hand-operated switch added to a hand-held filler-metal-paste dispenser (Fig. 5.7) yielded improved, more consistent filler-metal beads. The dispenser had been used to apply the paste between adjacent tubes in a heat exchanger. Previously, a technician used a foot switch to control the flow of paste from the dispenser as the technician’s hand moved the dispenser nozzle along a recess between adjacent tubes. A precise distribution of paste, suited to the taper of the tubes, was needed for a good braze joint. However, it was difficult to coordinate the foot control with hand motions to coordinate the volume of paste dispensed, along with the position and velocity of the dispenser to achieve the required distribution of paste. In contrast, the hand switch enabled the technician to control the flow of paste and the position and velocity of the dispenser all with one hand; this enhances coordination, providing a more consistent and controlled distribution of paste. The hand switch was resistance welded to a small band clamp and wired with small-diameter shielded conductor cable. It was mounted directly on a dispenser by use of the band clamp and plugged into the dispenser control unit in place of the foot switch. The control unit supplied pulses of pressurized air to the dispenser in
Fig. 5.6
Application of filler-metal paste using hand-held applicator gun/dispenser
224 / Brazing, Second Edition
response to the technician’s commands relayed via the hand switch. The switch was reliable and safe and could be removed and reattached easily, so that cartridges could be interchanged easily. Building rocket engines for the space program is a labor-intensive effort, involving limited-lot manufacture of specialized units. When Rocketdyne, a California-based division of Rockwell International, was looking for a way to automate the process and reduce costs, it found the solution in a new sensor for robotic applications. Through a cooperative research and development agreement, Rocketdyne (Ref 130) incorporated multiaxis seam tracking (MAST)—a fast, reliable, and inexpensive sensor technology—into a robotic operation that replaced manual dispensing of braze paste and powder on rocket thrust chamber tube seams. This sensor technology allowed for a reduction in labor and materials costs. A typical thrust chamber is cone shaped, 152 cm (60 in.) in diameter at its widest point, and 244 cm (96 in.) long (Fig. 5.8). It contains hun-
Switch cable
dreds of tubes running the length of the inside surface. Fuel circulates through these tubes to cool the thrust chamber jacket and preheat the fuel before combustion. To bond the tubes, an operator applied nickel powder and palladiumsilver filler-metal paste along each tube seam. After the tubes have been prepared with the powder and filler-metal paste, the entire structure is heated in a large, high-temperature furnace to form a solid assembly. The manual application of the powder and filler-metal paste is expensive, time-consuming, and subject to inconsistency. After evaluating several different robotic approaches for replacing the manual operation, Rocketdyne selected Sandia’s MAST capacitive sensor approach. The MAST sensor contains small electrodes that generate multiple electric fields. These fields interact with the surface of the thrust chamber tubes to track the seams between the tubes. The MAST system detects changes in the electric fields as capacitance variations and converts them into voltages to control the position of the robot (Ref 130) (Fig. 5.9, 5.10). Filler metals in paste form provide versatility. It is possible to use one or two filler metals for a variety of applications, thus reducing fillermetal inventory. Paste filler metals are normally custom-formulated to meet exact requirements. A product can be ideally suited to specific requirements by varying the amount of filler metal, powder size, flux percentages, and binder percentages.
Switch
Dispenser cartridge
Dispenser nozzle
Fig. 5.7
With just one hand, a technician controls the flow of filler-metal paste from the dispenser. The switch (under the operator’s forefinger) controls the flow of compressed air to the dispenser cartridge. Source: Ref 129
Fig. 5.8
The cone-shaped fuel-tube assembly for thrust chambers manufactured by Rocketdyne. Source: Ref 130
Chapter 5: Brazing Filler Metals / 225
Most pastes have a long shelf life. Store them at room temperature, keeping them away from temperature extremes, and rotate the inventory. Suppliers formulate paste filler metals with different percentages of powdered metal and flux to produce different results. They also can tailor paste viscosity to requirements. When using filler-metal pastes for small production runs and different joint configurations, try to select one or two formulations to limit inventory. Several additional characteristics of paste filler metals should also be considered. In many applications, paste filler metals are placed on the outside of the joint area. As a result, they are exposed to direct heat throughout the brazing cycle. Careful placement of the filler metal and attention to the heating method and time are necessary to ensure that the filler metal in paste form does not melt and flow away from the joint area prior to the components reaching proper temperature. Many paste filler-metal systems exist for heating methods that use torches, induction
Fig. 5.9
Fabricated from an inexpensive, multilayer printed circuit board, the multiaxis seam tracking (MAST) sensor measures 15 cm (6 in.) long and 2.5 cm (1 in.) wide. The lower tip of the board holds four capacitor electrodes that send tracking information to signal conditioning electronics. Source: Ref 130
Fig. 5.10
The multiaxis seam tracking (MAST) sensor, interfaced to a Fanuc S-700 robot arm, tracks tube seams on a thrust chamber section that precisely replicates the geometries of thrust chamber components. The MAST system has been successfully integrated with production robotic equipment at Rocketdyne for dispensing powder and filler-metal paste along tube seams. Source: Ref 130
226 / Brazing, Second Edition
heat, resistance heating, and atmosphere furnaces. There are fewer products that perform reliably in vacuum furnaces. Special care must be taken to evaluate paste filler metals intended for use with vacuum furnaces, so as not to damage the equipment. Transfer Tapes. There are a number of factors to be considered when selecting the form of the filler metal for any particular application. In evaluating the total cost of the different fillermetal forms, careful consideration of all forms is necessary to make the best choice. A new approach to controlled application of filler metal is the use of transfer tapes and preforms. Filler-metal transfer tapes are produced in either rolls or preforms. Incorporating a prefabricated layer of filler metal, transfer tapes have extreme uniformity in thickness and density as well as ease of application. The tapes consist of four layers (Fig. 5.11). A thin, plastic foil carrier layer serves to carry the filler-metal layer until actual application takes place. It is backed by a filler-metal layer that can be composed of any powdered filler metal mixed with selected organic binders. The thickness of this layer can be varied from 0.01 to 0.06 mm (0.0004 to 0.002 in.), with a thickness tolerance of ±5%. The third layer, a pressure-sensitive adhesive film, transfers the filler metal to the metal surface, and, during the brazing and firing cycle, it decomposes without leaving a residue. A paper fourth layer protects this film and is peeled off before application. In manual applications, the tape is first cut to the size and shape required. The protective paper is then removed, and the tape is firmly pressed, either by hand or by use of a rubber roller, against the area of cleaned metal surface to be coated. During this step, the filler metal is transferred to the base metal. The process is completed by simply stripping the carrier layer from the filler-metal layer. Transfer tape has been used to braze turbo-
machinery seals for gas turbines and has eliminated problems of filler-metal distribution and excessive infiltration during brazing. In another application, the use of transfer tape significantly improved a process for assembling sheet-metal turbine blades. Conventional methods of joining a two-piece blade and an internal cooling duct using nickel-base filler metal had resulted in poor control over the location and amount of filler metal applied. The use of transfer tape solved this problem and also permitted simplification of the fixtures required for the brazing operation. In some cases, gold-nickel transfer tape has replaced filler-metal foil. Although prepared and applied similarly to foil, the tape provides more accurate placement and ease of application. In brazing of honeycomb sections, particularly large, fine-cell honeycomb, the use of transfer tape has solved the problem of migration of filler metal and, in some instances, has reduced the time required to prepare large sections for brazing by as much as 85%. Researchers (Ref 131) investigated and developed a casting process for amorphous tapes of NiP11 filler metal based on the planar flow casting method. They found that the NiP11 filler metal in amorphous tape form was an appropriate material to easily design brazed joints and achieve mechanical properties. They were in agreement with the general rules of joint construction leading to the maximal reduction of gap distance but not so much as to give rise to the danger of not filling the space between jointed parts (Ref 132, 133). The mechanical testing results explicitly showed the usefulness of the application of a narrow gap, in this case, 0.03 mm (0.001 in.). It showed that one can obtain brazed joints having tensile strength as well as shear strength of approximately 40% higher when a tape of 0.03 mm (0.001 in.) thickness had been applied
D C A
Fig. 5.11
The four layers used in transfer tape
B
A – Carrier layer B – Filler-metal layer C – Adhesive film D – Protective paper
Chapter 5: Brazing Filler Metals / 227
instead of a tape of 0.06 mm (0.002 in.) thickness. The successful correlation between the results of mechanical testing, structure morphology, and phase composition of brazed joints was also observed. Plating. When an assembly is made of copper, an excellent method of brazing with the copper-silver eutectic composition in any orientation is to silver plate the joint area to a thickness of 0.0025 to 0.010 mm (0.0001 to 0.0004 in.) to produce a tight push fit. Although silver melts at 960 °C (1760 °F), heating above the eutectic temperature of 777 °C (1431 °F) is sufficient to cause the whole joint interface to become molten. The excess plating around the joint provides sufficient additional filler metal to eliminate voids and produce small fillets at the customary joint interface. Larger amounts of filler metal may be obtained by the use of copper-silver filler-metal wire rings in conjunction with the silver plating, the latter ensuring excellent and reliable penetration completely through the joint. Various other plating combinations are possible, for example, gold plating on nickel, nickel plating on zirconium, and nickel plating on tantalum, all of which produce a lower-meltingpoint eutectic. Electroless nickel (a nickel-phosphorus alloy) and a whole range of pure metals, such as copper, silver, gold, and platinum, may be deposited on joint components of highermelting-point stainless steels and used as filler metals at their normal melting points without necessarily forming eutectics. It is worth noting that where electroplating techniques are used, the expertise usually exists to dissolve material from the joint area in a highly controllable manner by either chemical or electrochemical methods. Such techniques have been used to overcome the tight-tolerance problem sometimes encountered with copper components. Nickel plating may also be used in brazing practice to ensure good wetting of stainless steels by low-melting-point filler metals in protective-atmosphere furnaces or to prevent oxidation where good control of the dewpoint of the furnace atmosphere is not possible. For this purpose, plating at least 0.025 mm (0.001 in.) thick is desirable to delay diffusion of chromium in the base metal to the surface. Additionally, it should be noted that nickel plating is usually required on the brazed joint areas of stainless, corrosion-, or heat-resistant
steels or superalloys having a percentage by weight equal to or greater than that shown for the following alloying elements: • Titanium, 0.70% • Aluminum, 0.40% • Titanium plus aluminum, 0.70% A 0.01 to 0.02 mm (0.0004 to 0.0008 in.) thickness is the nickel plate normally applied. Electroplated and electroless nickel coatings offer the following advantages for brazing applications (Ref 134): • The amount of filler metal is controlled during the deposition process. • The coating is metallurgically bonded to the surfaces of the parts immediately at the brazing interface, and the components are sweated together. This requires none of the capillary action that is often relied on for successful brazing. • Parts can be pressed together, if necessary; this provides the advantage that arises from a very thin interface layer that diffuses totally into the bulk material during the brazing process. This can result in extremely high joint strengths. Coatings for brazing applications have been classified into three main types (Ref 135): • Active coatings melt and then wet surfaces prior to joining components together on solidification. • Passive coatings are pressed out of the joint on melting and expose and clean the surface, allowing solid-state bonding reactions to occur. • Barrier coatings neither melt nor are pressed out of the joint clearance. These coatings form a physical barrier that protects the substrate from combining with molten filler metal. Typical active/passive coatings include copper, electroless nickel, tin, and zinc. Copper coatings are used extensively, because they are easily squeezed out of the joint during resistance joining and constitute a passive coating similar to tin and zinc coatings. Electrodeposited nickel has too high a melting point (1453 °C, or 2647 °F) for use in resistance joining applications (Ref 135). Copper and silver are used as active coatings in furnace brazing.
228 / Brazing, Second Edition
When these coatings melt, they take on the role of filler metal between closely fitting parts. Electroless nickel deposits are quite versatile active coatings for resistance joining applications. They typically contain between 8 to 10% P and melt in the range of 882 to 1041 °C (1620 to 1906 °F). Although capable of being deposited very uniformly, they are more expensive than other electroplated coatings and can be quite brittle. When used as filler metal, the thickness of these coatings should be approximately 0.5 µm (Ref 135). Copper coatings serve as a barrier during brazing of steel parts and prevent wetting of the steel by low-cost copper-phosphorus-(silver) filler metals. Because the filler metal does not contact the steel substrate, the formation of brittle iron and phosphide films is avoided (Ref 135). Electrodeposited nickel and electroless nickel coatings serve as barrier coatings in soldering and brazing applications involving copper wire and zinc-aluminum alloys. Electrodeposited nickel is also used on free-machining brass prior to furnace brazing to prohibit dezincification (Ref 135); on free-machining steel to avoid contamination of the filler metal with lead and/or manganese sulfide, which are the freemachining constituents (Ref 135); and on Inconel (Special Metals Corp.) prior to brazing to titanium (Ref 136). It is important to note that only electrodeposited nickel has a high enough melting point for the coating to remain intact when furnace brazing with copper- or nickel-base filler metals (Ref 135). For example, (a) the strength of copper-steel parts can be increased somewhat by using a barrier nickel coating on the steel, which prevents penetration of iron into the filler metal at a brazing temperature of 950 °C (1740 °F) (Ref 137); and (b) with Nitronic 60, which contains nitrogen as a strengthening agent, nickel plating can be used to prevent nitrogen outgassing from creating problems during brazing. For this application, electroless nickel is not acceptable, because the brazing is done at approximately 1100 °C (2010 °F), a temperature much higher than the 890 °C (1635 °F) melting point of electroless nickel containing 10 wt% P. Reportedly, even the higher-melting electroless nickel coatings, for example, those low in phosphorus content, have not been satisfactory for this type of application (Ref 138). Molybdenum and tungsten, which have excellent high-temperature properties, are often
brazed to iron for various applications. However, the direct brazing of iron to molybdenum or tungsten tends to cause exfoliation of the brazed joint in service, due to formation of brittle intermetallic compounds such as Fe7Mo6 and Fe7W6. Nickel deposits 1 to 4 µm thick on the low-carbon steel base metal restrain the formation of these brittle intermetallics, thereby noticeably improving the mechanical properties and shear strength of the brazed joints. For example, the shear strengths of 0.11% C steelBCu-Mo and -W joints were 301 and 307 MPa (44 and 45 ksi), regardless of heating time. With nickel plating, the shear strengths of the joints increased to 393 and 349 MPa (57 and 51 ksi), respectively (Ref 139). A variety of alloys that, after joining, require high strength and corrosion and oxidation resistance up to 649 °C (1200 °F) are nickel plated prior to brazing, in accordance with Aerospace Material Specification (AMS) 2675F (Ref 140). Table 5.20 from this specification lists the plating thickness requirements for various alloys. Because this specification calls for only joint areas to be coated, brush plating is well suited for these types of applications, particularly for parts that are too large for most conventional plating tanks (Ref 141). A gold-nickel electroplated alloy is being used to braze push rods to stainless steel diaphragms that serve as pressure sensors (Ref 142). To form the gold-nickel filler metal, nickel (2 µm thick) is plated from a sulfamate solution, followed by gold plating (5 µm thick) from a citrate solution. The heat of the brazing operation is relied on for forming the filler metal.
Table 5.20 Nickel-plating thickness requirements for furnace atmosphere brazing Nickel-plating thickness, µm Alloy
Vacuum
Nonoxidizing gas
Precipitation-hardenable iron alloys Precipitation-hardenable nickel alloys Ti + Al content under 4% nominal Ti + Al content 4% nominal or greater Nonprecipitation-hardenable nickel alloys with Ti + Al content under 1% nominal All other alloys
2.5 to 15
10 to 15
2.5 to 15
10 to 15
2.5 to 15
20 to 30
15 max
2.5 to 15
15 max
15 max
Source: Ref 132
Chapter 5: Brazing Filler Metals / 229
A new brazing process, based on the vaporphase reaction between the base metal and a metal vapor, was reported (Ref 143). Copperplated carbon steel was reacted with zinc vapor in a protective atmosphere in order to braze the steel with brass filler metal. With the vaporphase brazing approach, the brazing temperature was lower than that for copper brazing, and the brazement was uniformly coated with a goldcolored copper-zinc alloy. The fillets formed at brazed joints by this process were faultless and by no means inferior to those obtained by ordinary brazing methods (Ref 144). Clad and Coat. Precision cladding operations have been developed to metallurgically bond filler metals to base metals (Ref 145). The resulting new materials lower production costs by eliminating shim and paste or brazing flux for operations that do not use vacuum furnaces. The clad self-filler metal is optimized for maximum joint strength in the 0.05 to 0.10 mm (0.002 to 0.004 in.) range. By comparison, typical brazing operations require a shim in the 0.15 mm (0.006 in.) range. This new commercial product is to be used for typical aerospace and automotive applications. Researchers (Ref 146) developed a process that supplies wear-stressed workpieces with a hard material coating by a brazing process. Known as braze coat, this process makes it possible to produce wear-resistant coatings with a high percentage of hard material. The brazecoat process has already been approved as a furnace process and is used industrially in continuous furnaces. To coat wear-stressed workpieces by the braze-coat process, hard material and fillermetal cloths are needed. In order to produce these cloths, the hard material and the filler metal are mixed as powders with an organic binder. They are then fabricated by pressing and rolling. The thickness of the cloths ranges between 0.5 and 3 mm (0.02 and 0.1 in.). The hard-material cloths are disposable WC and Cr3C2. The filler metal acts as the nickelbase filler metal BNi-2. The cloths are fixed on the substrates with an organic adhesive. During heatup, the organic binder evaporates at 500 °C (930 °F). The evaporation creates a definite porosity in the hard-material cloth. This effects a capillary attraction onto the upper-melting filler metal. The filler metal melts at a brazing temperature of 1100 °C (2010 °F), infiltrates the porous hard-material cloth, and joins it to the substrate. Recently, a hard-material coating was
produced that contains hard-material coatings in a filler-metal matrix (Fig. 5.12). Foils and Sheets. Three groups of fillermetal compositions have emerged as brittle filler metals and are prepared as ductile foils by RS technology. The filler metals fall into three groups: the eutectic-type alloys, the peritectic alloys for vacuum brazing, and the copper-silver eutectics (Ref 147). High-technology applications of thin-sheet products frequently require a higher level of dimensional control, finer features, improved quality of cut, and greater cleanliness of the finished product. A supplier of filler metals used in jet aircraft engine component repair has manufactured a superalloy and filler-metal mixture that is used in the repair of stator vanes of high-pressure turbines in jet aircraft engines. The final product is cut out of sintered plates of the mixture that range in thickness from 0.38 to 1.78 mm (0.015 to 0.070 in.) (Ref 148). Traditionally, these plates were cut using either waterjet or punch and die. The company found these conventional technologies, and even flowing-gas laser systems, to be inadequate and expensive for meeting thin-sheetmetal manufacturing requirements. Waterjet was expensive, left stains on the finished product, and required cleaning before shipment to the customer. The cleaning process added further cost to the product and increased manufacturing time, thereby reducing productivity. Punches also had been used in the past, but they were of limited value and could only be used for removing simple shapes (squares, rectangles, etc.). In addition, they were unusable for removing fine features or objects with small separations, and the punch dies were expensive to purchase for more complicated shapes.
Cloths
Brazed layer system
filler metal hard material
Substrate (a)
Fig. 5.12
Substrate (b)
Two sequences of the braze-coat process. (a) A substrate with hard material and filler-metal cloths before brazing. (b) A substrate with braze-coat layer system after brazing. Source: Ref 146
230 / Brazing, Second Edition
Realizing that conventional methods were delivering unsatisfactory results, the company initiated a search into modern sealed CO2 lasers that employ slab-discharge technology. The slab-discharge technology employed in sealed lasers has been successfully used in machining thin-sheet filler metals with lasers (Ref 148). As the name implies, sealed CO2 users permanently seal the required gas mixture within the laser cavity, eliminating the need for gas replacement. Unlike flowing-gas lasers, sealed CO2 laser heads require no service for up to 25,000 h of continuous operation. Consequently, sealed CO2 lasers eliminate maintenance downtime, thereby increasing productivity and reducing costs. Overall, the lower facility requirement results in an average hourly operating cost that is less than half that of a slow-flow laser and less than one-fifth that of a fast-flow laser. Researchers (Ref 149) announced their development of amorphous strip filler metals for high-temperature brazing. These researchers claim that their type-AMA (Amorphous Metal Alloys) filler metals represent one of the most promising materials of this group for use in practice. The technology developed for producing these filler metals makes it possible to manufacture them in the form of thin, ductile flexible strips 20 to 100 µm thick from low-deformability alloys. The filler metals produced from these strips at MIFI (Moscow Engineering Physics Institute) are usually used in the form of powders and pastes with an organic filler. In contrast to them, the ductile flexible strip of AMA filler metal is used in strictly metered amounts. Its chemical and microstructural homogeneity results in narrow meltingpoint ranges, higher strength, and corrosion resistance of brazed joints as well as improving diffusion activity and flow into the gap. Their AMA filler metals have been used for brazing a wide range of materials in any combination: copper and copper alloys, nickel and its alloys, stainless steels, titanium and its alloys, zirconium, beryllium, refractory metals, hard alloys, oxide ceramics, graphite, and so on. Flexible thermal insulation blankets made of ceramic fibers can be protected against weather and handling by attaching thin-metal facesheets. In applications in which the blankets are exposed to gas flows, the facesheets also afford protection against flow-induced stresses and
help reduce aerodynamic drag by providing smoother flow surfaces. Typically, a metal foil is attached to a ceramic blanket that has a thickness of 5 mils (~0.13 mm, or 0.005 in.) or less and is made of titanium, aluminum, chromium, niobium, or alloys of these elements. The blanket can be made of fibers of silica, aluminoborosilicate, SiC, and/or other ceramic materials. Optionally, in preparation for attachment of the metal foil, the ceramic fabric on the attachment surface of the blanket can be precoated with a thin layer of nickel to improve its bonding properties. Small dots of a metal or ceramic filler metal are placed on the attachment surface of the blanket (Fig. 5.13). Preferably, the dots are between approximately 3 to 6 mm (0.12 and 0.24 in.) square and positioned either randomly or in a regular pattern at intervals of approximately 2.5 cm (1.0 in.). The metal or ceramic filler metal can be any of several commercial formulations that both wet the ceramic fabric and form metallic bonds with the metal foil when heated to the brazing temperature. Suitable ceramic filler metals include ceramic-precursor adhesives based on silica, alumina, and/or zirconia. Suitable metal filler metals include copper-silver, copper-gold, and copper-silver-gold alloys that contain titanium and/or vanadium as wetting agents (Ref 150). The metal foil is placed over the dots, then the resulting sandwich is heated to a temperature of approximately 980 °C (1800 °F) in a reducing atmosphere or in a vacuum to effect brazing. Finally, the sandwich is cooled to room temperature, leaving the metal foil strongly bonded to the blanket at the dots. Another convenient method of supplying filler metal is to use filler-metal sheets that consist of a core of aluminum base-metal alloy and a coating of the lower-melting filler metal. The coatings are aluminum-silicon filler metals and may be applied to one or both sides of the basemetal sheet. The coating is normally roll bonded to the core during the mill fabrication. Thus, the filler-metal sheet is a product that can be formed by drawing, bending, or other normal metalworking processes without removing this coating. The formed part can be assembled and brazed without placing additional filler metal in the joint. Filler-metal sheet is frequently used as the member of an assembly with the mating piece made from unclad brazeable alloy. The filler metal on the filler-metal sheet flows by
Chapter 5: Brazing Filler Metals / 231
capillary action and gravity to fill the joint at contact sites. In addition, filler-metal-clad thinwall tubing has been made by continuous seam welding filler-metal sheets.
Case Histories and Problem-Solving Examples Example 1: Filler-Metal Remelt. There are many questions on the availability of information regarding filler-metal remelt properties. With nickel brazing, once the boron has diffused, the remelt temperature is approximately 55 °C (100 °F) higher than the original braze temperature. The remelt temperature of a filler metal depends on the following brazing variables: quantity of filler metal (joint clearance), brazing temperature, time at the brazing temperature, and the degree of mutual solubility between the base and filler metals.
In fact, a nickel-boron-filler-metal brazed joint that is only held for 1 to 3 min at heat at a low brazing temperature will remelt at the original brazing temperature of the filler metal. One method for determining the remelt temperate of a filler metal/base metal is to butt braze rods with an induction coil around the specimen. An optical pyrometer can monitor the temperature and determine the remelt temperature. Brazing filler metals having silicon (Ni-Si+) or phosphorus (Ni-P+) contain elements with larger atomic size. Therefore, they diffuse at a much slower rate. In general, diffusion is directly proportional to the time and temperature and inversely proportional to the quantity of filler metal for a given base-metal/filler-metal combination. Thus, the higher the brazing temperature, the faster diffusion takes place. Conversely, the lower the temperature, the slower diffusion occurs. With a given filler metal, each base metal will have a different rate of diffusibility, and differ-
Metal sheet
A
Dots of brazing material
A
Ceramic blanket Bonds formed by brazing
SECTION A-A (ENLARGED)
Fig. 5.13
Metal sheet and ceramic blanket joined by brazing at the dots. Source: Ref 149
232 / Brazing, Second Edition
ent filler metals have different rates of diffusibility on the same base metal. With all these variables, it is not possible to state a specific temperature and time that will completely diffuse a chosen combination. To solve this problem, it would be necessary to run one or more tests using a variable joint-clearance specimen to determine the specific length of time at the temperature chosen to braze a combination of materials. It is suggested to consult American National Standards Institute/ American Welding Society (ANSI/AWS) A5.8, “Specifications for Filler Metals for Brazing and Braze Welding,” and to obtain the opinions of designers and metallurgists or other experts about the specific applications and use. The filler-metal quantity must be adequate to fill the entire joint and possibly some excess. Braze at the prescribed temperature and for a specific length of time, which could be 1 or 2 h at heat to begin with to complete the brazing cycle. Reference 151 discusses the point where a secondary phase shows up in the center of the braze joint. The clearance at this point is considered the maximum brazing clearance point to obtain full diffusion. In many cases, the filler metal can be completely homogenized with the base metal, leaving no indication of the filler-metal presence. Example 2: Brazing of 17-7 PH versus 174 PH. A manufacturer was brazing 17-7 PH
Fig. 5.14
stainless steel in a vacuum of 0.13 Pa (2 × 10–5 psi) with BNi-2 filler metal at 1063 °C (1945 °F). The filler metal did not flow adequately and balled up. On occasion, the surface of the 17-7 PH had a dark gray color, while at the same temperature, 17-4 PH stainless steel brazed with the same filler metal and these parts came out quite well, with the filler metal flowing into the joint. Why are there the major differences in flow/wetting between the previously mentioned two base metals? The major difference between these two base metals is composition in that the 17-7 PH contains 0.75 to 1.50% Al, along with 0.009% C maximum, 16.00 to 18.00% Cr, 1.00% Mn maximum, 6.50 to 7.75% Ni, and 1.00% Si maximum, as well as 0.04% maximum each for phosphorus and sulfur. Aluminum is the culprit in the 17-7 PH, which readily oxides in the range of 538 to 926 °C (1000 to 1700 °F). At this temperature range, the atmosphere, whether it be vacuum or pure dry hydrogen, is oxidizing to elements such as chromium, aluminum, titanium, and those to the right of chromium in the metal/metal-oxide chart (Fig. 5.14) (Ref 152, 153). To eliminate this problem, there are several methods that brazers use to improve the wetting and flow of filler metal on 17-7 PH. The most common method is to nickel electrolytically plate the 17-7 PH in order to hide the aluminum so that the atmosphere does not see it. A good
Metal/metal oxide equilibria in pure hydrogen atmospheres. Source: Ref 153
Chapter 5: Brazing Filler Metals / 233
plating thickness to obtain these results is 0.01 mm (0.0004 in.) of nickel (thinner coatings may be used if the atmosphere quality is very good). A second method to braze this alloy is with a very good, clean furnace and a higher vacuum to produce a lower partial pressure of oxygen in the furnace. This will prevent oxidation on the aluminum on the part but will allow it to vaporize off the 17-7 PH. A newer method of handling the parts is to blast them with a special Ni-Fe-Cr-Si-B alloy that has been effectively used to eliminate nickel plating of that particular alloy and should work on 17-7 PH. Blasting is very effective on thicker materials. However, if the parts are very thin, blasting can cause distortion. The 17-4 PH alloy brazed satisfactorily because it does not contain any aluminum. Example 3: Closed versus Open Gaps. Joining 250 µm thick plate and 75 µm thick fins made of 436 AISI stainless steel with filler metal of 37 µm thick MBF-20 foil was prepared with two kinds of base-metal/filler-metal settings: • The first group of specimens was prepared from plates and foils. The MBF-20 foil was tack welded to AISI 436 stainless steel coupons. The foil contained holes having dimensions of up to 2 to 3 mm (0.08 to 0.12 in.) wide and up to 5 mm (0.2 in.) long. The seldom-observed holes of these dimensions are typical in MBF filler metals. Also, pieces of MBF-20 were tack welded to the stainless steel coupons in a side-by-side configuration, with the pieces separated by 0.5 and 1.0 mm (0.02 and 0.04 in.) gaps. Both the holes and the gaps not covered by foil and fins are called open gaps. • The second group of specimens consisted of plates, foils, and fins. The MBF-20 foil was tack welded between stainless steel 436 coupons and 25.4 mm (1 in.) square sections of fins. Again, the pieces of MBF-20 foil had holes and were positioned with gaps very similar to that of the first setting. The joints had at least one small dimension within a few dozen micron range, which is a typical dimension for brazing gaps, and the most favorable dimension for the capillary wetting by liquid filler metals. These gaps can be called closed gaps. All specimens tested were brazed in a furnace with an atmospheric pressure less than approximately 2.6 × 10–3 Pa (4 × 10–7 psi). Specimens
were brazed at 1060 °C (1940 °F) for 15 min, with two intermediate temperature holds made at 300 and 900 °C (570 and 1650 °F) during the heating part of this brazing cycle. All the brazed samples had very clean, bright, lustrous surfaces without any traces of oxidation, as expected under good vacuum conditions. In summary, optical microscopy revealed that liquid MBF-20 filled in all closed gaps in the plate-foil-fin assemblies, resulting in complete, full fillets and sound joints. A very different picture emerged from specimens prepared with open holes and gaps. Here, the ability of molten MBF-20 to heal the defects was substantially moderated. For example, the brazing process converted a 1 mm (0.04 in.) wide virgin hole into a smaller but similarly shaped unbrazed spot. Therefore, when vacuum brazing 436 AISI stainless steel plate-fin structural parts with MBF-20 ribbon having 37 µm thickness, sound joints with complete fillets are formed in places where gaps and holes appear in these ribbons. The dimensions of gaps and holes may be as large as at least 3 mm (0.12 in.). The same picture is observed with other base and filler metals. Therefore, manufacturers who braze in vacuum or inert gas environments may braze multiple set side-by-side preforms made of MBF ribbons and separated by gaps, yet still fabricate joints that have no leaks and high overall integrity and strength. Example 4: Nickel Plating and Its Advantages. The question often arises relative to the proper thickness of nickel plating for nickeland cobalt-base metals containing titanium and aluminum. Most parts come out bright and clean, but some come out with different colors. Why are there differences? In brazing, one must consider variables and be able to know about and understand them. Base Metals Containing Aluminum and Titanium. The amount of aluminum and/or titanium in the base metal will require the plating thickness to be varied. However, as the aluminum and/or titanium increases, the thickness of nickel plating, in general, has to be increased. Electrolytic versus Electroless Nickel Plating. The common electroless nickel plating is a nickel-phosphorus alloy that starts to melt at approximately 871 °C (1600 °F) and, when molten, ceases to be a barrier coating because the liquid metal dissolves the aluminum and titanium and brings it to the surface where it oxidizes. The electrolytic nickel plating is gen-
234 / Brazing, Second Edition
erally used for a barrier coating to hide the aluminum and titanium from the atmosphere at all temperatures. See the previous section “coatings” in this chapter. Plating-Thickness Uniformity. The footprints in a brazing furnace readily show the variations in the plating thickness on the aluminum- and titanium-containing base metals. The variation in color indicates the variation in thickness. The thicker the coating, the brighter and cleaner it is. The thinner the coating, the darker the color. Of course, this is keeping all other variables constant. Atmosphere Type. In general, the atmosphere quality in vacuum is better than the gas atmospheres. In general, less plating thickness can be used in the vacuum than in the gas atmosphere processing. Thus, when brazing in pure dry hydrogen, thicker plating is generally required. (See Chapter 6, “Fluxes and Atmospheres,” in this book.) Atmosphere Quality. The atmosphere quality can vary considerably in any one of the given atmosphere types, but, in general, aluminum will vaporize in a very good-quality vacuum or pure dry hydrogen atmosphere as well as argon and others. This requires a very low partial pressure of oxygen, so that the aluminum is not oxidized during the brazing cycle. Titanium, on the other hand, appears to oxidize in a gas atmosphere, such as pure dry hydrogen. However, in vacuum with a very good, high-quality atmosphere, one of the oxides of titanium will vaporize, leaving the surface bright and clean. Brazing Temperature. The rate of diffusion through the nickel plating is dependent on the temperature. Time at Brazing Temperature. The longer a part is held at the brazing temperature, the thicker should be the nickel plating. Recommended Plating Thickness. In general, if the aluminum plus titanium is less than 4%, 0.0025 to 0.015 mm (0.0001 to 0.0006 in.) of plating will be satisfactory for vacuum processing. In pure dry hydrogen, however, 0.01 to 0.015 mm (0.0004 to 0.0006 in.) of nickel thickness is recommended. In base metals containing greater than 4% Al and 4% Ti, 0.08 to 0.015 mm (0.003 to 0.0006 in.) is suggested, and in hydrogen, 0.020 to 0.0027 mm (0.0008 to 0.00011 in.) of nickel plate would be recommended. Example 5: To Plate or Not to Plate? In brazing alloy 738, which contains 3.4% Ti and 0.4% Al, the question of whether to nickel plate or not arises.
It has been reported that some of the newer furnaces can braze high-aluminum-and -titanium-containing base metal under shop floor conditions. In brazing alloy 738, it is recommended to use nickel plating as a barrier coating to effectively braze the high-titanium- and -aluminumcontaining base metals. Electrolytic nickel plating should be used as the barrier coating. The time at the brazing temperature is important because the aluminum and titanium diffuse through the nickel plating when the temperature is above 982 °C (1800 °F). The color showing the bright, clean nickel of the plated surface after completing the brazing cycle is a footprint that ensures a good braze. Example 6: How to Properly Position Filler Metal. A firm was brazing 304L stainless steel with BNi-2 filler metal in a vacuum furnace at 0.13 Pa (2 × 10–5 psi). The joints at the base of the component reflect a well-made braze that flows into the joint. There is a problem, however, with the overhead-type joints, where the filler metal is applied to the bottom side of the joint as positioned in the furnace, and there is nothing underneath it to hold it in place. The filler metal is powder in paste form; therefore, the problem is determining why the filler metal falls off and how it can be prevented from happening in the future. When using paste in brazing, the filler-metal paste should not stay on an overhead joint while going up to the brazing temperature. The binder or cement could be gone by the time the part reaches the brazing temperature of 538 °C (1000 °F), and, therefore, the filler metal would fall off an upside-down joint. To ensure that the filler-metal paste is going to stick on the bottom side of a joint, it must be properly applied. The surface area between the filler metal and the paste must be large enough to support the filler metal. This is indicated in Fig. 5.15 on the left side. As shown on the right side, the filler metal was applied as a filament coming out of the syringe and laid at the joint when the surface was right-side up; thus, there
Fig. 5.15
Cross section of applied filler metal
Chapter 5: Brazing Filler Metals / 235
is very little contact between the filler metal and base metal for the large amount of filler metal. The water-based binders are better since they support the filler-metal powder in the syringe or cartridge and eliminate the settling out of the powder. No filler metal, when applied properly, should fall off of a part. Example 7: Free-Flowing Braze Filler Metals. For normal brazing operations, it is desirable to have the filler metal as fluid as possible so that it goes into the recommended joint clearance and flows all the way through the joint by capillary action. For normal brazing operations, it is preferred to have the brazing temperature at least 10 to 38 °C (18 to 68 °F) above the liquidus temperature of the filler metal to produce adequate flow through the joint. Because it is so close to the melting point of, for example, aluminum base metal, the brazing temperature must be held very close to the melting point of the filler metal. The melting range for most of the filler metals of aluminum is only –1.1 to 4.4 °C (–2 to 8 °F). For example, the liquidus of BAlSi-7 is only –9 °C (–16 °F) below the maximum brazing temperature range, which is quite narrow. Fortunately, the filler metal will allow brazing at 588 °C (1090 °F), while the liquidus of BAlSi-7 is 595 °C (1100 °F). Brazing below the liquidus is possible with a number of filler metals including the copper, silver, and nickel-base families because their melting ranges are comparatively wide and the majority of the filler metal melts at the lower end of the liquidus/solidus temperature range. Why would one want to braze below the liquidus? By brazing below the liquidus temperature, the filler metal is a liquid plus solid (the mushy range), which makes a less fluid material that will fill up wider root openings. Table 1A of ANSI/AWS A5.8-92 shows the liquidus and solidus temperatures and brazing temperature range, as well as a number of filler metals in which the lower temperature of the brazing range is below the liquidus of the filler metal. In the aluminum brazing series, five of the filler metals indicate that the lower temperature of the brazing range is below their liquidus temperature. In the BCuP series, all of the lower temperatures of the brazing ranges listed are below the liquidus of the filler metals. In fact, for BCuP-5 filler metal, the brazing temperature range is 79.3 °C (142.7 °F) below its solidus. For several filler metals in the BAg classification, the lower temperature of the brazing
range is as much as 40 °C (72 °F) below the liquidus of the filler metal. The same is true for several of the nickel-base filler metals, where, again, the lower temperature of the brazing temperature range is below the liquidus of the filler metal. Thus, it is possible to change the fluidity of the filler metal by brazing at the higher-temperature range to get more fluidity or by brazing toward the lower range to get less. There are occasions when it is even desirable to go below the recommended brazing temperature range for very specific brazing applications where the minimum flow is required. Example 8: Avoiding Chloride Stress-Corrosion Cracking. A small, brazed, plate-andframe type 316 stainless steel heat exchanger in hot-water service failed every 6 to 12 months. The exchanger used steam at 862 kPa (125 psi) and 121 °C (250 °F) to heat city water from ambient to 75°C (165 °F) for use in chemical processing. Occasionally, the steam valve was not closed when the exchanger was idle, although this practice was discontinued. Two exchangers were installed in parallel to reduce the flow rate and limit erosion-corrosion (as suggested by the manufacturer), but this did not stop the failures. A defective exchanger was evaluated by visual and metallographic examination and confirmed that the failures were due to chloride stress-corrosion cracking (Cl– SCC) of the type 316 stainless steel frame, with the cracking most severe around the steam inlet and outlet nozzles. Given the susceptibility of type 316 stainless steel to Cl– SCC, and the concentration of chlorides caused by the many crevices typical of this exchanger design, it is not surprising that failure by Cl– SCC took place. The conclusion was that the penetration was caused by liquid metal corrosion (LMC) because the filler metal was copper-base. Although LMC did not cause the exchangers to fail, it did cause grain-boundary penetration, which increased the local stress and the potential for fatigue crack initiation. Because of an increase in the volume of hot water required for processing, a shell-and-tube exchanger, which has inhibited brass tubes and tube sheets, and a carbon steel shell installed in place of the type 316 stainless steel exchanger, Cl– SCC should no longer be of concern when brazed with copper filler metals or silver filler metals containing copper. Molten copper would previously penetrate stainless steel along the
236 / Brazing, Second Edition
grain boundaries. Even if the penetration is relatively shallow (as in this case), the affected boundaries would provide sites for fatigue and thermal fatigue crack initiation. When stainless steel brazing is required, nickel-base filler metals should be specified.
REFERENCES
1. G. Humpston and D.M. Jacobson, Principles of Soldering and Brazing, ASM International, 1993 2. M.M. Schwartz, Brazing, ASM International, 1987 3. M.M. Schwartz, Brazing for the Engineering Technologist, Chapman & Hall/ Kluwer Publications, 1995 4. T. Funamoto, S. Kokura, S. Kato, et al., A Study on Formation of Alloyed Layer of Low Melting Temperature on Bonding Surface of Inconel 738 LC with Sputtering, Q. J. Jpn. Weld. Soc., Vol 3 (No. 4), 1985, p 207–212 5. M. Kato, T. Funamoto, H. Wachi, et al., A Study of the Alloyed Layer of Low Melting Temperature Bonding Surface of IN 738 LC with Boron Pack Cementation, Q. J. Jpn. Weld. Soc., Vol 5 (No. 3), 1987, p 84–87 6. T. Funamoto, M. Kato, M. Wachi, et al., Diffusion Weldability of IN 738 LC with Alloyed Layer on Bonding Surface by Boron Pack Cementation, Q. J. Jpn. Weld. Soc., Vol 5 (No. 3), 1987, p 87–93 7. S. Liu, D.L. Olson, G.P. Martin, et al., Modeling of Brazing Processes That Use Coatings and Interlayers, Weld. J., Vol 70 (No. 8), Aug 1991, p 207–215 8. R.E. Tressler, T.L. Moore, and R.L. Crane, Reactivity and Interface Characteristics of Titanium-Alumina Composites, J. Mater. Sci., Vol 8, 1973, p 151–161 9. M.G. Nicholas, T.M. Valentine, and M.I. Waite, The Wetting of Alumina by Copper Alloyed with Titanium and Other Elements, J. Mater. Sci., Vol 15, 1980, p 2197–2206 10. M.G. Nicholas, Active Metal Brazing, Trans. J. Br. Ceram. Soc., Vol 85, 1986, p 144–146 11. M.G. Nicholas and D.A. Mortimer, Ceramic/Metal Joining for Structural Applications, Mater. Sci. Technol., Vol 1, 1985, p 657–665
12. M.M. Schwartz, Ceramic Joining, ASM International, 1990 13. M.G. Nicholas, Joining of Ceramics, Chapman and Hall, 1990 14. R. Wysopal and E.R. Bangs, “The Importance of Braze Alloy Application in High Temperature Brazing,” presented at the Second International American Welding Society/Welding Research Council Brazing Conference and Colloquium, 52nd Annual American Welding Society Meeting, 26–30 April 1971 (San Francisco, CA) 15. M.M. Schwartz, Metals Joining Manual, McGraw-Hill, 1979 16. D.F. Paulonis, D.S. Duvall, and W.A. Owczarski, U.S. Patent 3,678,570 17. D.S. Duvall, W.A. Owczarski, and D.F. Paulonis, T.L.P.: A New Method for Joining Heat-Resisting Alloy, Weld. J., Vol 53 (No. 4), April 1974, p 203–214 18. B.F. Kuvin, Pick the Right Brazing Alloy, Weld. Des. Fabr., April 1989, p 41–44 19. E.R. Perry, Joining by Brazing: High Quality Filler Alloys are Essential, Weld. Met. Fabr., Vol 59 (No. 1), Jan/Feb 1991, p 11–12 20. G. Humpston, D.M. Jacobson, and S.P.S. Sangha, Low Melting Point Aluminium Alloy Braze, U.K. Patent 9,218,404.3, 1992 21. D.M. Jacobson, G. Humpston, and S.P.S. Sangha, A Low-Melting-Point Aluminum Braze, Weld. J., Aug 1996, p 243–250 22. G. Humpston, S.P.S. Sangha, and D.M. Jacobson, New Filler Metals and Process for Fluxless Brazing of Aluminum Engineering Alloys, Mater. Sci. Technol., Vol 11 (No. 11), Nov 1995, p 1161–1167 23. W. Shultze and N. Schoer, Fluxless Brazing of Aluminum Using Protective Gas, Weld. J., Vol 52 (No. 9), 1973, p 644–651 24. H. Schoer, Aluminum Brazing Alloys for Fluxless Brazing of Aluminum, VAW, Light Metals, P.G. Campbell, Ed., Minerals, Metals and Materials Society, 1988, p 699–702 25. A.A. Suslov, Using Metallic Coatings for Brazing Aluminium Alloys, Weld. Int., Vol 10 (No. 7), 1996, p 563–564 26. A.A. Suslov, Contact Interaction of Brazing Melts with Parent Material when Vacuum Brazing Aluminium Alloys, Weld. Int., Vol 9 (No. 8), 1995, p 651–653 27. T. Takemoto, T. Ujie, H. Chaki, and A.
Chapter 5: Brazing Filler Metals / 237
28. 29.
30. 31. 32.
33.
34.
35.
36.
37.
38.
39.
40.
Matsunawa, Influence of Oxygen Content on Brazeability of a Powder Aluminum Braze Filler Metal, Weld. J., Nov 1996, p 372–378 B. Koetzing, NTIS Alert, Vol 97 (No. 2), 15 Jan 1997, p 23 T. Kayamoto and T. Onzawa, Development of Low Melting Temperature Fillers for Aluminum Alloys, Weld. Int., Vol 10 (No. 5), 1996, p 363–369 T. Kayamoto, K. Kin, C. Saito, and T. Onzawa, Q. J. Jpn. Weld. Soc., Vol 12 (No. 4), 1994, p 495–501 T. Kayamoto, K. Kin, C. Saito, and T. Onzawa, Q. J. Jpn. Weld. Soc., Vol 13 (No. 4), p 450–456 Hoogovens USA Aluminum Corp., HighStrength Aluminum Braze Alloys Resist Corrosion, AM&P, Vol 152 (No. 2), Aug 1997, p 14 K. Suganuma, Y. Miyamoto, and M. Koizumi, Joining of Ceramics and Metals, Annual Review of Materials Science, Vol 18, R.A. Huggins, Ed., Annual Reviews, 1988, p 47–73 B. Dalgleish, K. Trumble, and A.G. Evans, “Microstructural and Mechanical Characterization of Alumina-Aluminum Bonds,” Paper 12-B-88, presented at the 90th Annual Meeting of the American Ceramic Society, 2 May 1988 (Cinncinnati, OH) M. Naka, M. Kubo, and I. Okamoto, Joining of Silicon Nitride with Al-Cu Alloys, J. Mater. Sci., Vol 22 (No. 12), 1987, p 4417–4421 M. Ueki, M. Naka, and I. Okamoto, Joining and Wetting of CaO-Stabilized ZrO2 with Al-Cu Alloys, J. Mater. Sci., Vol 23 (No. 8), 1988, p 2983–2988 M. Ueki, M. Naka, and I. Okamoto, Wettability of Some Metals Against Zirconia Ceramics, J. Mater. Sci. Lett., Vol 5 (No. 12), 1986, p 1261 Y.R. Mahajan and S. Kirchhoff, Rapidly Solidified Structures of an Al-Zr-Gd Alloy, Scr. Metall., Vol 21 (No. 8), 1987, p 1125–1130 R.C. Rathner and D.J. Green, Joining of Yttria-Tetragonal Zirconia Polycrystal with an Aluminum-Zirconium Alloy, J. Am. Ceram. Soc., Vol 73 (No. 4), 1990, p 1103–1105 W.D. Rupert, Braze Life after Cadmium, Weld. J., Nov 1995, p 59–61
41. D.A. Wigley, P.G. Sandefur, Jr., and P.L. Lawing, “Preliminary Results on the Development of Vacuum Brazed Joints for Cryogenic Wind Tunnel Aerofoil Models,” Int. Cryogenic Materials Conf., 10–14 Aug 1981 (San Diego, CA) 42. A.Y. Krasnopevtsev and A.A. Fedorov, Properties and Special Features of Joints in Vapour-Gas Brazing with Multicomponent Brazing Alloys of the Cu-Zn System, Weld. Int., Vol 10 (No. 6), 1996, p 498–500 43. V.E. Khryapin, A Handbook of Brazing, Publ. Mashinostroenie, Moscow, 1981 44. I.E. Petrunin et al., Physical Metallurgy of Brazing, Publ. Metallurgiya, Moscow, 1976 45. A. Lee, International Copper Association, AM&P, June 2001, p 17 46. JETRO, Brazing Agent with 60% Greater Bonding Strength for Ceramic Bonding, 96-06-001-03, June 1996, Japan External Trade Organization, p 16–17 47. T. Takamori and M. Akanuma, Possible Braze Composition for Pyrolytic Graphite, Ceram. Bull., Vol 48, 1969, p 734– 736 48. D.A. Mortimer and M. Nicholas, The Wetting of Carbon and Carbides by Copper Alloy, J. Mater. Sci., Vol 8, 1973, p 640–648 49. D.H. Kim, S.H. Huang, and S.S. Chun, The Wetting and Bonding of Si3N4 by Copper-Titanium Alloys with Other Elements, Ceram. Int., Vol 16, 1990, p 333–347 50. W.D. Rupert, Copper-Phosphorus Alloys Offer Advantages in Brazing Copper, Weld. J., May 1996, p 43–45 51. S.C. Dev, P. Basak, I. Singh, R.K. Dubey, and O.N. Mohanty, A Copper-Base Brazing Alloy for Electronics Industries, J. Mater. Sci., Vol 27 (No. 24), 1992, p 6646–6652 52. N.J. DeChristofaro and A. Datta, in Proceedings of the TMS-AIME on Rapidly Solidified Crystallized Alloys, S.K. Das, B.H. Kear, and C.M. Adam, Ed., May 1985, Allied Corporation, 1985, p 263 53. H. Okamoto and T.B. Massalski, Ed., Phase Diagrams of Binary Gold Alloys, ASM International, 1987, p 76–81 54. A. Datta, A. Rabinkin, and D. Bose, Rapidly Solidified Copper-Phosphorus
238 / Brazing, Second Edition
55. 56. 57.
58. 59.
60.
61.
62.
63. 64.
65.
66.
67. 68.
Base Brazing Foils, Weld. J., Vol 63 (No. 10), Oct 1984, p 14–21 Precious Metals Boost Brazing Uses, Weld. Eng., Jan 1970, p 38–41 M.H. Sloboda, Industrial Gold Brazing Alloys, Gold Bull., Vol 4 (No. 1), Jan 1971, p 2–8 N.P. Fairbanks, “High-Temperature Braze for Superalloys,” Technical Report AFML-TR-76-155, AFSC, Wright-Patterson Air Force Base, OH, 1976 A.M. Tasker, Gold-Containing Brazing Filler Metals, Gold Bull., Vol 16 (No. 4), Oct 1983, p 111–113 J.J. Stephens and F.A. Greulich, Elevated Temperature Creep and Fracture Properties of the 62Cu-35Au-3Ni Braze Alloy, Metall. Mater. Trans. A, Vol 26 (No. 6), June 1995, p 1471–1482 F.M. Hosking, J.J. Stephens, and J.A. Rejent, Intermediate Temperature Joining of Dissimilar Metals, Weld. J., April 1999, p 127–136 F.M. Hosking, C.H. Cadden, N.Y.C. Yang, S.J. Glass, et al., Microstructural and Mechanical Characterization of Actively Brazed Alumina Tensile Specimens, Weld. J., Aug 2000, p 222–230 S. Kang and H.J. Kim, Design of HighTemperature Brazing Alloys for CeramicMetal Joints, Weld. J., Sept 1995, p 289–295 H. Mizuhara and E. Huebel, U.S. Patent 5,273,832, 28 Dec 1993 E.J. Ryan, J.R. Doyle, and D.H. King, Brazing Tomorrow’s Aircraft Engines, Weld. Des. Fabr., Vol 52 (No. 3), March 1979, p 108–114 W.E. Morgan and P.L. Bridges, “Development of a Brazing Alloy for the Mechanically Alloyed High Temperature Sheet Material Incoloy Alloy MA 956,” Final Report TR-3381 (AD-A118252), AFWAL Contract F49620-79-C-0008, Wiggin Alloys Ltd., Hereford, England, Sept 1981 E. Lugscheider and T. Cosack, HighTemperature Brazing of Stainless Steel with Low-Phosphorus Nickel-Based Filler Metal, Weld. J., Oct 1998, p 215– 226 R.L. Peaslee and P.F. Walter, How Oxygen Affects Nickel Brazing Filler Metals, Weld. J., Dec 1994, p 61–66 C.R. Heiple, S.S. Christiansen, D.L.
69.
70.
71.
72.
73. 74.
75.
76.
77.
78.
79.
Keller, et al., Wettability Differences of Braze 508 on 304 Stainless Steel, Weld. J., Vol 69 (No. 9), Sept 1990, p 41–43 R. Hardesty, “High Temperature Be Panel Development,” NASA Contract 18177, NASI-18613, National Aeronautics and Space Administration, May 1989 S.V. Lashko and V.I. Sokopov, A Brazing Alloy for Brazing Corrosion-Resisting Steels Operating at Cryogenic Temperatures, Weld. Prod., Jan 1983, p 37–41 E. Lugscheider and W. Tillman, Development of New Active Filler Metals in a AgCu-Hf System, Weld. J., Vol 69 (No. 11), Nov 1990, p 416–421 M.G. Nicholas, D.A. Mortimer, L.M. Jones, et al., Some Observations on the Wetting and Bonding of Nitride Ceramics, J. Mater. Sci., Vol 25, 1990, p 2679–2689 A.P. Xian and Z.-Y. Si, Joining of Si3N4 Using Ag57Cu38Ti5 Brazing Filler Metal, J. Mater. Sci., Vol 25, 1990, p 4483–4487 L. Ljungberg, R. Warren, and C.-H. Li, A Novel Method for Identifying Phases Formed in Brazed and Soldered Joints, J. Mater. Sci. Lett., Vol 9, 1990, p 1316–1318 C. Peytour, P. Berthet, F. Barbier, et al., Interface Microstructure and Mechanical Behaviour of Brazed TA6V/Zirconia Joints, J. Mater. Sci. Lett., Vol 9, 1990, p 1129–1131 F. Barbier, C. Peytour, and A. Revcolevschi, Microstructural Study of the Brazed Joint between Alumina and Ti-6Al-4V Alloy, J. Am. Ceram. Soc., Vol 73 (No. 6), 1990, p 1582–1586 R. Johnson, International Institute of Welding subcommittee IA report brazing group, 1990 Autumn Meeting SCIA-B185/90, International Institute of Welding M.L. Santella, J.A. Horton, and J.J. Pak, Microstructure of Alumina Brazed with a Silver-Copper-Titanium Alloy, J. Am. Ceram. Soc., Vol 73 (No. 6), 1990, p 1785–1787 J. Kivilahti, M. Paulasto, and M. Turunen, The Influence of Titanium Content of the Filler Alloy and Brazing Parameters on the Microstructures and Properties of Titanium/Al2O3 Joints, International Conf. on Evolution of Advanced Materials, Associazione Italiana di Metallurgia, Italy, 1989, p 347–352
Chapter 5: Brazing Filler Metals / 239
80. L. Martens, W. Tillmann, and E. Lugscheider, Strength and Microstructure of Brazed Cemented Carbide and SiliconNitride Joints, J. Mater. Process. Technol., Vol 58 (No. 1), 1 March 1996, p 13– 23 81. H.W. Grunling, High-Temperature-Resistant Ceramic-Metal Joints, Schweissen Schneiden, Feb 1971 82. S.K. Chatterjee and Z. Mingxi, Tin-Containing Brazing Alloys, Weld. J., Vol 69 (No. 10), Oct 1990, p 37–42 83. “Silver Brazing Alloys—Cadmium Containing, Health and Safety,” The Sheffield Smelting Co. Ltd., Dec 1980 84. P.M. Roberts, Recent Developments in Cadmium-Free Silver Brazing Alloys, Weld. J., Vol 57 (No. 10), Oct 1978, p 23–30 85. P.M. Roberts, Weld. Des. Fabr., Jan/Feb 1979, p 35–46 86. V.R. Miller and W.L. Falke, “Report of Investigations 8783,” U.S. Department of the Interior, 1983 87. J.B. Long and C.J. Evans, Tin Uses, Vol 118, 1978, p 12–16 88. W. Tillmann, E. Lugscheider, E. Schlimbach, K. Manter, and J.E. Indocochea, Heat-Resistant Active Brazing of Silicon Nitride, Part 2: Metallurgical Characterization of the Braze Joint, Weld. J., March 1998, p 103–109 89. K. Suganuma, Joining Non-Oxide Ceramics, Ceramics and Glasses, Vol 4, Engineered Materials Handbook, ASM International, 1991, p 523–531 90. R.W. Rice, Joining of Ceramics, Advances in Joining Technology, J.J. Burke, A.E. Gorum, and A. Tarpinian, Ed., Brook Hill, 1976 91. J.D. Cawley, Joining of Ceramic Matrix Composites, Am. Ceram. Soc. Bull., Vol 68 (No. 9), 1989, p 1619–1623 92. M.P. Borom, The Mechanical and Chemical Aspects of Glass Sealing, Glass Ind., Vol 59 (No. 3), 1978, p 12–16 93. H.P. Kirchner, J.C. Conway, Jr., and A.E. Segall, Effect of Joint Thicknesses and Residual Stresses on the Properties of Ceramic Adhesive Joints, Part I: Finite Element Analysis of Stresses in Joints, J. Am. Ceram. Soc., Vol 70 (No. 2), 1987, p 104–109 94. W.A. Zdaniewski, J.C. Conway, Jr., and H.P. Kirchner, Effect of Joint Thickness
95. 96. 97. 98.
99.
100.
101.
102.
103.
104.
105.
106.
and Residual Stresses on the Properties of Ceramic Adhesive Joints, Part II: Experimental Results, J. Am. Ceram. Soc., Vol 70 (No. 2), 1987, p 110–118 W.E. Hauth III, Ceramic-to-Ceramic Sealing of Large Shapes, Am. Ceram. Soc. Bull., Vol 58 (No. 6), 1979, p 584–586 J.D. Cauley and J.T. Knapp, “Silicate Brazing of a High-Purity Alumina,” MR9105, EWI, May 1991 R.A. Blue and R.Y. Lin, Ind. Heat., April 1996, p 43 T. Onzawa, A. Suzumura, and M.W. Ko, Brazing of Titanium using Low-MeltingPoint Ti-Based Filler Metals, Weld. J., Vol 69 (No. 12), Dec 1990, p 462–467 Y. Guezawa, Advances in Joining Newer Structural Materials, International Institute of Welding Meeting and Proceedings of International Conf., (Montreal, Canada), 1990, p 223–228 C.H. Cadden, N.Y.C. Yang, and T.H. Headley, Microstructural Evolution and Mechanical Properties of Braze Joints in Ti-13.4Al-21.2Nb, Weld. J., Aug 1997, p 316–325 A.A. Rabinkin, H. Lieberman, S. Pounds, T. Taylor, F. Reidinger, and S. Lui, Amorphous Ti-Zr-Base Metglas Brazing Filler Metals, Scr. Metall., Vol 25, 1995, p 399–404 V.S. Chernov, A.S. Exteev, V.A. Tatarinov, F.E. Pashchenko, et al., Properties and Use of Ts Ag Zinc-Based Microcrystalline Brazing Alloy, Met. Sci. Heat Treat., Vol 34 (No. 9–10), 1992, p 584–586 E.-S. Park, J.J. Lannutti, and J.D. Cauley, Microstructural and Reactions of SiCwReinforced Alumina with Ag-Cu-In-Ti, J. Am. Ceram. Soc., Vol 78 (No. 1), 1995, p 15–20 H. Asai, F. Ueno, N. Iwase, et al., Titanium Nitride-Molybdenum Metallizing Method for Aluminum Nitride, IEEE Trans. Compon., Hybrids, Manuf. Technol., Vol 13 (No. 2), June 1990, p 457–461 K. Iyogi et al., Pin Brazing Technology for Aluminum Nitride, Proc. Autumn Nat. Convention, Instit. Electron. Inform. Commun. Eng., Vol C-2 (No. C-51), 1988, p C-2-52 K.D. Challenger and T.W. Cordea, “Low Temperature Solid-State Bonding of
240 / Brazing, Second Edition
107.
108.
109. 110.
111. 112.
113.
114.
115. 116. 117.
118.
119.
Materials Using Compliant Ag Interlayers,” EWI-MR-9404, EWI, June 1994 M. Zhu and D.D.L. Chung, Active Brazing Alloy Containing Carbon Fibers for Metal-Ceramic Joining, J. Am. Ceram. Soc., Vol 77 (No. 10), 1994, p 2712–2720 M. Nakamura and S.D. Peteves, SolidState Bonding of Silicon Nitride Ceramics with Nickel-Chromium Alloy Interlayers, J. Am. Ceram. Soc., Vol 73 (No. 5), 1990, p 1221–1227 Y.C. Wu and J.G. Duh, Eutectic Bonding of Nickel to Yttria-Stabilized Zirconia, J. Mater. Sci. Lett., Vol 9, 1990, p 583–586 M. Wittmer, Materials Research Society Symposium Proceedings, E.A. Giess, K.N. Tu, and D.R. Uhlmann, Ed., Vol 40, Materials Research Society, 1984, p 393 M. Wittmer, C.R. Boer, and P. Gudmundson, J. Am. Ceram. Soc., Vol 65, 1982, p 149 J.C. Ambrose, M.G. Nicholas, N. Young, et al., Wetting and Spreading of Ni-P Brazes: Effects of Workpiece and Braze Composition, Mater. Sci. Technol., Vol 6 (No. 10), 1990, p 1021–1031 W.F. Gale and E.R. Wallach, Wettability of Nickel Alloys by Boron-Containing Brazes, Weld. J., Vol 70 (No. 3), March 1991, p 76–79 D. Bose, A. Datta, and N. DeCristofaro, Comparison of Gold-Nickel with Nickel Base Metallic Glass Brazing Foils, Weld. J., Vol 60 (No. 10), Oct 1981, p 29–34 Brazing with Foil Cuts Costs, Speeds Assembly, Strengthens Jet Engine Parts, Assem. Eng., March 1980 A. Rabinkin, New Applications for Rapidly Solidified Brazing Foils, Weld. J., Vol 68 (No. 10), Oct 1989, p 39–46 J. Colbus and K.F. Zimmerman, Properties of Gold-Nickel Alloy Brazed Joints in High Temperature Materials, Gold Bull., Vol 7 (No. 2), April 1974, p 42–49 C.G. Wan, H.P. Xiong, and Z.F. Zhou, Joining of Si3N4 to 1.25Cr-0.5Mo Steel Using Rapidly Solidified CuNiTiB Foils as Brazing Filler Metals, Weld. J., Dec 1997, p 522–525 A. Rabinkin E. Wenski, and A. Ribaudo, Brazing Stainless Steel Using a New MBF-Series of Ni-Cr-B-Si Amorphous Brazing Foils, Weld. J., Feb 1998, p
120. 121.
122.
123. 124.
125.
126. 127. 128. 129. 130. 131.
132. 133.
134.
66–75; International Institute of Welding Document IA-1014-9, 1999 A. Rabinkin and S. Pounds, Effects of Load on Brazing with Metglas MBF-2005 Filler Metal, Weld. J., May 1988, p 37–45 A. Rabinkin, F. Reidinger, J. Marti, and L. Bendersky, Processing, Structure and Performance of RS “Classical” SilverBase Brazing Alloys, Mater. Sci. Eng., Vol A 133, 1991, p 256–260 A. Rabinkin, Stability to Aging of Copper-to-Copper Joints Brazed with Metglas MBF-2005 and BCuP-5 Filler Metals, Weld. J., Oct 1988, p 29–30 A. Rabinkin, U.S. Patents 4,746,379 and 4,928,872 O. Botstein and A. Rabinkin, Brazing of Titanium-Based Alloys with Amorphous 25Ti-25Zr-50Cu Filler Metal, Mater. Sci. Eng., Vol H-3657, Feb 1994 O. Botstein, A. Schwarzman, and A. Rabinkin, Induction Brazing of Ti-6A14V Alloy with Amorphous 25Ti-24Zr50Cu Brazing Filler Metal, Mater. Sci. Eng., Vol H-4269 7 April 1995 D. Kay, Differential Thermal Analysis for Brazing Filler Metals, Heat Treat., Vol 24 (No. 10), Oct 1992, p 18–20 A. Rabinkin, AM&P, June 2001, p 65–67 B. Zorc and L. Kosec, New Approach to Improving the Properties of Brazed Joints, Weld. J., Jan 2000, p 24–31 Hand-Controlled Brazing-Paste Dispenser, NASA Tech. Briefs, Sept 1994, p 124–125 P. Molley, D. Schmitt, and J. Novak, Rocket-Building Robots, Sandia Technol., Feb 1995, p 8–9 D. Szewieczek and J. Tyrlik, Designing the Brazed Joint Properties with Application of Amorphous Tapes as a Filler Metal, J. Mater. Process. Technol., Vol 53, 1995, p 405–412 Handy and Harman, The Brazing Book, 1977 N. DeCristofaro and D. Bose, Brazing and Soldering with Rapidly Solidified Filler Metals, ASM 1986 International Conference on Rapidly Solidified Materials, (San Diego), American Society for Metals, p 415 R.E. Thrillwood, Vacuum Brazing, Eng., Vol 219 (No. 11), 1979, p 1468
Chapter 5: Brazing Filler Metals / 241
135. J.W. Yardy, Microstruct. Sci., Vol 19, 1998, p 607 136. D.D. Berger, Weld. J., Vol 74 (No. 11), 1995, p 35 137. J.W. Dini, Weld. J., Vol 61 (No. 11), 1982, p 13 138. R.L. Peaslee, Weld. J., Vol 70 (No. 1), 1991, p 118 139. T. Yoshida and H. Ohmura, Weld. J., Vol 61 (No. 11), 1982, p 363 140. “Nickel Alloy Brazing,” AMS 2675F, Society of Automotive Engineers, International, Aug 1994 141. J. Hurrell, “Brush Plating: The Process and Application for Improved Brazeability of Inconel,” Aerospace Symposium, American Electroplaters and Surface Finishers Society, 1992 142. I. Rajagopal, K.S. Rajam, and S.R. Rajagopalan, Met. Finish., Vol 90 (No. 8), 1992, p 59 143. T. Osawa, Weld. J., Vol 60 (No. 11), 1981, p 215 144. J.W. Dini, The Role of Electroplated Coatings in Metal Joining, Weld. J., May 1996, p 47–49 145. M. Dunn, Des. News, 1 Nov 1993, p 16 146. E. Lugscheider, K. Gundlfinger, and H. Schmoor, Braze Coat Process Combines
147. 148. 149.
150. 151.
152. 153.
with Induction Heating for Deposition of Wear-Resistant Materials, Weld. J., May 1993, p 55–59 M. Hunt, Amorphous Metal Alloys, ME, Nov 1990, p 35–38 Coherent Photonics Group, Laser Div., Machining Thin-Sheet Brazing Alloys with Lasers, MAN, Jan 2002, p 26–27 B.A. Kalin, V.T. Fedotov, O.N. Sevryukov, and A.E. Grigor’ev, Amorphous Strip Brazing Alloys for High-Temperature Brazing, Experience with Developing Production Technology and Application, Weld. Int., Vol 10 (No. 7), 1996, p 578–581 D.J. Rasky, P.M. Sawko, P. Kolodziej, and D.A. Kourtides, NASA Tech. Briefs, July 1999, p 49–50 E. Lugschieder and K.-D. Partz, High Temperature Brazing of Stainless Steel with Nickel Base Filler Metals BNi-2, BNi-5, and BNi-7, Weld. J., Vol 62 (No. 6), June 1983, p 160–164 R. Peaslee, American Welding Society Detroit Brazing and Soldering Division Newsletter; Weld. J., Jan 2000, p 75, 77 D.R. Milner and R.L. Apps, Introduction to Welding and Brazing, Pergamon Press, 1969
Brazing Second Edition Mel M. Schwartz, p243-287 DOI: 10.1361/brse2003p243
Copyright © 2003 ASM International® All rights reserved. www.asminternational.org
CHAPTER 6
Fluxes and Atmospheres FLUXES, GAS ATMOSPHERES, AND VACUUM promote the formation of brazed joints. They may be used to surround the work, exclude reactants, and provide active or inert protective atmospheres, thus preventing undesirable reactions during brazing. Under some conditions, fluxes and atmospheres may also reduce oxides that are present. Caution must be observed in the use of atmospheres, because some metals are embrittled by various gases. Notable among these metals are titanium, zirconium, niobium, and tantalum, which become permanently embrittled when brazed in any atmosphere containing hydrogen, oxygen, or nitrogen. Also, hydrogen embrittlement of copper that has not been thoroughly deoxidized must be avoided. The use of any flux or atmosphere does not eliminate the need for thorough cleaning of parts prior to brazing. Metals, when exposed to air, tend to react with various constituents of the atmosphere to which they are exposed. The rate of these chemical reactions is generally accelerated as the temperature is raised. The most common reaction is oxidation, but nitrides and carbides are also sometimes formed. The rate of oxide formation varies with each metal composition and the nature of the oxide. Tenacity, structure, thickness, and resistance to removal and/or further oxidation must be considered. Oxide formation on some metals in air is, for practical purposes, instantaneous, even at or below room temperature. These reactions result in conditions such as oxides or other compounds that hinder the production of consistently sound brazed joints.
Atmospheres One way to control the formation of oxides during brazing and also reduce oxides present after precleaning is to surround the braze area with an appropriate controlled atmosphere. Like fluxes, controlled atmospheres are not intended to perform primary cleaning for removal of oxides, coatings, grease, oil, dirt, or other foreign materials. All parts for brazing must be subjected to appropriate pre-braze cleaning operations, as dictated by the particular metals. When flux is used, a controlled atmosphere may be desirable to extend the useful life of a flux and to minimize postbraze cleaning. In controlled atmosphere applications, postbraze cleaning is generally not necessary. Controlled atmospheres are used extensively for high-temperature brazing. While performing the same basic function as fluxes (i.e., the prevention of oxidation during the brazing cycle), they have several advantages over fluxes: • The joint members are maintained in a clean, oxide-free condition when brazing is done in a controlled atmosphere. After brazing, the brazement can often be used in the asbrazed condition or finish machined without cleaning. • Controlled atmosphere brazing is particularly useful in joining complex assemblies such as heat exchangers, thrust chambers, and honeycomb sandwich structures. Complete removal of fluxes from such assemblies after brazing is difficult or impossible.
244 / Brazing, Second Edition
• Problems associated with flux entrapment in the brazed joint can be avoided if controlled atmospheres are used. Although many controlled atmospheres are available, those used primarily for brazing fall into three broad categories: reducing, inert, and vacuum. The reactions resulting from the use of these different atmospheres are diverse. Certain conditions, however, apply to all three. The general techniques of atmosphere brazing can involve: • A gaseous atmosphere alone • A gaseous atmosphere together with a solid or liquid flux preplaced at the interfaces • High vacuum • Combinations of vacuum and gas atmospheres
Joining Atmospheres Many types of assemblies demand furnace brazing under a protective atmosphere, including assemblies intended for service in a vacuum environment, which must be free from volatile contaminants, and parent-metal components that are disfigured by oxide scale. The categories of joining atmospheres that are available and their interrelationships are shown in Fig. 6.1. Generally, fluxes are needed only when carrying out the joining operation in air or other oxidizing environments.
Fig. 6.1
Categorization of joining atmospheres. Source: Ref 1
Two distinct types of gaseous atmosphere are used for joining: • Chemically inert gas atmospheres (e.g., argon, nitrogen, helium, vacuum). These function by excluding oxygen and other gaseous elements that might react with the components to form surface films and inhibit flowing of and wetting by the filler metal. • Chemically active atmospheres, both gases and fluxes, that are designed to react with surface films present on the components and/or the filler metal during the joining cycle and remove them in the process. These atmospheres may either decompose surface films (as does hydrogen when acting on certain oxide or sulfide layers, for example) or react with the films to produce compounds that can be displaced by the molten filler metal. An example of the latter is magnesium vapor that is introduced during the furnace brazing of aluminum. The vapor reacts with the alumina surface coating to form a complex aluminummagnesium oxide spinel that is readily broken up by molten filler metals. Controlled gas atmospheres require a confining vessel, which invariably means a furnace of some type. Furnace joining also offers other advantages: • The process may be easily automated for either batch or continuous production, be-
Chapter 6: Fluxes and Atmospheres / 245
cause the heating conditions can be accurately controlled and reproduced without the need for much operator skill. • Furnace joining allows uniform heating of components of almost any geometry and is suitable for parts that are likely to distort if heated locally. • The atmospheric protection afforded leads to economies with regard to the use of flux and finishing operations, such as cleaning and the removal of flux residues. Against this must be considered the following potential disadvantages: • Capital costs of the equipment, including the associated gas-atmosphere-handling or vacuum system, may be significant in relation to processing costs. • The entire assembly is heated during the process cycle, which can result in a loss of mechanical properties, even to components removed from the joint area. • The range of permissible parent materials and filler metals tends to be restricted to elements and chemicals of low volatility to avoid contamination of the furnace. For a similar reason, most fluxes are undesirable. Certain metals are embrittled on heating in the presence of standard gas atmospheres (oxygen, nitrogen, hydrogen, and carbon-containing gases) and must therefore be joined in a vacuum furnace. These are principally the refractory metals beryllium, molybdenum, niobium (columbium), tantalum, titanium, vanadium, and zirconium. On the other hand, tungsten, which is also a refractory metal, can be brazed in air under cover of mild fluxes. Thus, the requirements of each metal or alloy must be individually assessed (Ref 1). Oxidizing Atmospheres. Air is the most common oxidizing atmosphere. The principal advantages of joining in air are that no special gas-handling measures are required and there are no difficulties associated with access to the workpiece during the joining process. However, because most component surfaces and those of the filler metal are likely to form oxide scale when heated in air, fluxes normally must be applied to the joint region. An active flux is capable of chemically and/or physically removing an oxide film. The flux may be applied either as a separate agent or may be an integral constituent of the joining filler metal. Fluxes are discussed later in this chapter.
An oxidizing atmosphere is occasionally desirable during brazing. Not only do some fluxes require the presence of oxygen in order to work, but it is a prerequisite for successful joining that oxygen be present in some instances. An example is provided by the copper/copper oxide eutectic brazing process in which copper is brazed to ceramic materials, such as alumina, by a eutectic that is formed in situ between copper and copper oxide just below the melting point of copper (Ref 2). Inert Atmospheres. From a practical point of view, an atmosphere is either oxidizing or reducing. This is because it is not possible to remove and then totally exclude oxygen from the workpiece, except perhaps under rigorous laboratory conditions. Thus, when defining an atmosphere as inert, it must be taken as meaning that the residual level of oxygen present is not sufficient to adversely affect the joining process under consideration. Because the inertness of an atmosphere is judged relative to the specific application, it is necessary to define a quantitative measure of the oxygen present. This parameter is the oxygen partial pressure. Partial pressure provides a measure of the concentration of one gas in an atmosphere containing several gases. The partial pressure of a gas in a mixture of gases is defined as the pressure it would exert if it alone occupied the available volume. Thus, dry air at atmospheric pressure (0.1 MPa, or 0.015 ksi) contains approximately 20% O2 by volume, so that the oxygen partial pressure in air is 0.02 MPa (0.003 ksi). Typical inert atmospheres among the common gases include nitrogen, argon, and hydrogen. Hydrogen is included here because it is not capable of reducing the oxides present on the majority of metals at normal brazing temperatures. The oxygen partial pressure in standard commercial-grade bottled gases is on the order of 1 mPa. Higher-quality grades are available, but their cost is usually too prohibitive to permit their use in most industrial applications. System Types. An atmospheric brazing system (ABS) offers an innovative approach to brazing copper, steel, and nickel components using silver, copper, or nickel filler metals. The system uses precise induction heating to braze conductive materials in a controlled atmosphere. Individual parts are processed with speed, accuracy, and economy without the problems associated with traditional flame brazing or batch vacuum furnaces.
246 / Brazing, Second Edition
Flame brazing in a normal atmosphere causes oxidation, scaling, and carbon buildup on the parts. To clean flame-brazed parts, joint-weakening flux must be applied, or expensive acid cleaning baths may be required. Batch vacuum furnaces have successfully addressed these problems; however, they have significant limitations of their own because of their large size, low efficiency, and lack of quality control. Bell jar atmospheric brazing systems are able to produce clean parts with precision and repeatability, although their low throughput makes them impractical for high-production applications. As a precision atmospheric brazing system designed for high-speed, continuous operation, ABS (Ref 3) successfully overcomes all these problems. The system has a temperature range of 788 to 1204 °C (1450 to 2199 °F) to accommodate a wide variety of parts and processes, including silver and nickel brazing. For brazing silver, copper, and brass alloys, a controlled oxygen-free atmosphere is used, consisting of standard atmospheric gases. The system is suitable for processes that require vacuum (5 × 10–3 Pa, or 7.3 × 10–7 psi, or less) and/or high temperature (to 1204 °C, or 2199 °F), such as brazing steel or nickel alloys with nickel. Because the parts are brazed in an inert atmosphere, oxidation and scaling are eliminated, and no flux or acid cleaning bath is needed. For copper brazing of steel components, ABS uses a gas quenching system during brazing. This maintains the hardness of the part to specification and eliminates the need for secondary processing to restore hardness. The ABS is compact in comparison to large vacuum furnaces and is designed for 24 h operation as part of a manufacturing cell with a high level of quality control, real-time monitoring, and statistical process control. The system is designed for continuous operation as part of the production flow without the need for off-line heating in a batch vacuum furnace. After the initial preparation of the atmosphere within the chamber, a continuous feed system handles the parts. Energy use is considerably lower, particularly in comparison to batch vacuum furnaces that must be kept running for long periods with heat applied to the entire part. With ABS, because heat develops inside the part within a fraction of a second, cycle times can be reduced dramatically. Safety has been a central ABS design consideration. Operator safety is maximized by iso-
lating the heating system, using safety interlocks on the vacuum chamber and manual override access, an emergency stop button on the operator interface, and additional warning systems built into the software and hardware. For nearly two decades, liquid-nitrogen-base atmosphere systems have enabled commercial heat treaters and manufacturing companies to braze parts with good and consistent quality, increase plant safety and throughput, meet justin-time delivery requirements, and enter new markets. Because of these benefits, many heat treaters and manufacturing companies have switched from endothermically and exothermically generated atmospheres to liquid-nitrogenbase atmospheres, such as blends of nitrogenhydrogen and nitrogen-methanol (Ref 4–6). Several low-cost, nitrogen-base atmosphere systems to produce brazing atmospheres from on-site, noncryogenically generated nitrogen have been developed. These systems, which are marketed under the trade name Purifire (BR AtmosphereSystems), provide companies with the capability to produce atmospheres equal in quality and performance to liquid-nitrogen-base systems for brazing carbon steel components. The brazing atmospheres are produced by pretreating mixtures of on-site, noncryogenically generated nitrogen and natural gas using a proprietary purification system. The noncryogenic nitrogen is generated by using a pressure swing adsorption (PSA) or membrane system. In a typical PSA nitrogen-generation system, clean, dry, compressed air is passed through one of the carbon molecular sieve beds. The molecular sieve bed preferentially adsorbs oxygen, allowing nitrogen to pass through the bed to a storage vessel. After a preset time and before the bed is saturated with oxygen, the flow of compressed air is switched over to the second molecular sieve. Meanwhile, the first bed is regenerated and readied for the next cycle by depressurizing and venting it to the atmosphere. The cycle is repeated automatically, thereby producing a continuous stream of nitrogen. The nitrogen gas generated by the previously mentioned noncryogenic techniques contains 0.1 to 5.0% residual oxygen, the presence of which is generally not desirable during brazing. It is, therefore, mixed with a predetermined amount of natural gas and treated using a proprietary purification system to produce the nitrogen-base brazing atmosphere. Two different nitrogen purity levels, 99.0 and 99.5% pure nitrogen, have been used along with 1.5% natural gas
Chapter 6: Fluxes and Atmospheres / 247
to produce nitrogen-base brazing atmospheres. These purity levels were selected because they were well within the safe limits specified by the National Fire Protection Association 86C safety standards for furnace atmospheres. The purified noncryogenic nitrogen-base brazing atmospheres provided less than 3 ppm oxygen and a dewpoint of –37 °C (–35 °F), both in the heating and cooling zones of the furnace. These oxygen and moisture levels are close to the values typically associated with liquid-nitrogen-base brazing atmosphere systems. In addition to providing low oxygen and moisture levels, the purification system produced sufficient amounts of reducing gases such as hydrogen and carbon monoxide, which are generally required for maintaining reducing potential in the furnace and providing bright surface finish to the brazed components. To evaluate the system, high-manganese carbon steel components (Society of Automotive Engineers, or SAE, 1541) containing 0.4% C were brazed at 1107 °C (2025 °F) in a 60.96 cm (24 in.) wide continuous mesh belt furnace using the nitrogen-base brazing atmospheres. These components were brazed with a copper-base filler-metal paste specially formulated to prevent soot formation in low-dewpoint nitrogen-base atmospheres. The total brazing time, including time spent by components in the heating and cooling zones of the furnace, was approximately 45 min. The brazed components were analyzed for braze flow, fillet quality of the brazed joint, and fillet formation as well as for strength of the
Copper fillet
brazed joint and the extent of surface decarburization. The SAE 1541 high-manganese steel components brazed had a bright surface finish without any signs of oxidation. These atmospheres facilitated good wetting and spreading of the copper filler metal on the steel base metal. They also provided good flow in the space of the lap joint and fillet formation (Fig. 6.2). The use of noncryogenically generated nitrogen-base atmospheres was instrumental in brazing high-manganese carbon steel components without soot formation. Additionally, these atmospheres helped to prevent surface decarburization of these steel components. In addition to the work done with brazing SAE 1541 steel components, low-carbon steel components were brazed with atmospheres produced from noncryogenically generated nitrogen. These components were subjected to torsion and peel tests. In all cases, the steel portion of the components tore prior to failure of the brazed joint, demonstrating excellent strength of the braze joint. More importantly, no void formation was observed in the brazed joints in the peel tests. These results clearly show that nitrogen-base atmospheres produced by pretreating mixtures of noncryogenically generated nitrogen and natural gas, using a purification system, can generate a furnace atmosphere for brazing steel components without surface decarburization and with good quality of braze flow and braze joint strength.
Copper fillet
Carbon steel Carbon steel
Carbon steel
Fig. 6.2
(b)
I
(a)
Carbon steel
200 m
Micrographs of Society of Automotive Engineers (SAE) 1541 steel components brazed in atmospheres produced from (a) 99.0% and (b) 99.5% pure noncryogenically generated nitrogen showing good braze flow and fillet formation in lap joints
248 / Brazing, Second Edition
Researchers (Ref 7) studied the brazeability of aluminum in vacuum-nitrogen partial-pressure atmosphere brazing. In their studies in vacuum brazing, they selected an Al-10Si-1.5Mg filler metal. The filler metal melted at 559 °C (1038 °F) solidus temperature; liquidus temperature was 591 °C (1096 °F), and magnesium in the filler metal actively evaporated. The magnesium gas was the effective getter of contaminants such as H2O and O2, which form an oxide film on the surface of aluminum alloys, lowering brazeability. Volatile elements also evaporate and material properties change in high-vacuum brazing. For example, heat exchangers made with aluminum alloys use aluminum-zinc alloy for the cathodic corrosion protection of other aluminum alloys. The vapor pressure of zinc in the aluminumzinc alloy is high because zinc is a volatile element, but aluminum-zinc alloy does not melt at the brazing temperature, which is approximately 600 °C (1110 °F), and zinc does not evaporate actively, compared with magnesium. However, evaporation of volatile elements and change in material properties can be minimized in vacuum-nitrogen partial-pressure atmosphere brazing, and aluminum-zinc alloy may be used as a sacrificial alloy in products made with aluminum alloys. Therefore, in their study, the brazeability in vacuum-nitrogen partial-pressure atmosphere was investigated using T-joints with horizontal aluminum-manganese or aluminum-zinc alloy sheet and vertical A4004-clad A3003 fillermetal sheet. Specimens were brazed over a wide range of brazing pressures and N2 carrier gas flow rates. The brazing temperature and brazing time were 600 °C (1110 °F) and 5 min, respectively. Gas contaminants in brazing atmospheres were measured using a quadruple mass spectrometer. It was found that a higher carrier gas flow rate gave better brazeability. Also, many magnesium-rich swells grew at the surface of the fillermetal sheets under the low N2 gas flow rate conditions, and the surface of the filler-metal sheet was very rough. On the other hand, the partial pressure of H2O increased according to the increase in brazing pressure, but it did not change according to the increase in the carrier gas flow rate at the same brazing pressure. Therefore, brazeability was not decided by the partial pressure of H2O in the brazing atmosphere, and it was found that
brazeability was improved by the higher evaporation of magnesium from the filler-metal sheets according to the increase in the carrier gas flow rate. In another effort to determine the effect of nitrogen on furnace-brazed joint quality, an examination (Ref 8) was conducted concerning the problem of producing poorly brazed and discolored parts when vacuum furnace brazing Multimet (Haynes International, Inc.) alloy to 304L stainless steel, using BNi-2 filler metal in a vacuum of 1.3 Pa (1.9 × 10–4 psi). Some base metals containing nitrogen (for example, Multimet and 304LN) do not braze well in a vacuum furnace even if the furnace is clean and the vacuum atmosphere is 1.3 Pa. Nitrogen, if present in the base metal, can have a significant effect on the ability to braze a joint. This is particularly true with filler metals containing boron. Multimet and 304LN are examples of alloys containing nitrogen as a strengthening agent. The nitrogen level in these alloys may range from 0.02 to 0.30% and have the following effects: • The nitrogen in the base metal will combine with the boron in the filler metal and thus deplete the melt depressant from the filler metal. This condition will raise the liquidus temperature and inhibit filler-metal flow. The degree to which this happens depends on the amount of nitrogen in the base metal, mesh size of the filler-metal powder, furnace heating rate, and the brazing temperature. • Nitrogen also can produce an iridescent bluish-gray film on the base-metal surface. This film will prevent the filler metal from wetting and flowing across the base metal or into the joint. Formation of these films not only is dependent on the nitrogen content of the base metal but also on the pumping rate in the vacuum furnace, gas atmosphere flow rates, and the amount of parts in the load. The adverse effects of nitrogen in the base metal can be minimized by the following procedures: run the part to be brazed through a furnace cleaning cycle, adjust the heating rate during brazing, and nickel plate the base metal. The furnace cleaning cycle is similar to a braze cycle except that the part is held at high temperature for ample time to allow sufficient outgassing. The cleaning cycle should work satisfactorily on base metals with low nitrogen
Chapter 6: Fluxes and Atmospheres / 249
contents; however, this may not work if the nitrogen content is too high (0.08 to 0.18%) or if the joint has very tight clearances. There is an axiom that states, “The tolerable amount of contaminants in the furnace atmosphere is directly proportional to the heating rate.” Thus, the higher the nitrogen content of a base metal, the higher the rate of heating required to produce a properly brazed joint. Another way to minimize the problem with nitrogen in the base metal is by nickel plating both surfaces of the joint by the electrolytic process. It is important to note that some joints may not be suitable for electrolytic plating (that is, blind holes), due to the poor throwing power of nickel. In such cases, internal electrodes may be used. Electroless nickel is not suitable because it is a nickel-phosphorus composition that melts at significantly lower temperatures (approximately 870 °C, or 1600 °F). Approximately 0.01 to 0.03 mm (0.0004 to 0.0012 in.) of electroplated nickel should suffice. Plating thickness is dependent on the heating rate, brazing temperature, percentage of nitrogen in the base metal, and type of filler metal. Difficulties due to nitrogen also arise when using case-hardened (nitrided) steels as base metals. Alloys that tend to absorb nitrogen during annealing in dissociated ammonia experience similar problems. In these situations, boron-containing filler metals, such as Nicrobraz LM (Wall Colmonoy Corporation) (BNi2), should be avoided due to their affinity for nitrogen. Filler metals without boron, such as Nicrobraz 30 (BNi-5) and Nicrobraz 50 (BNi7), may be used instead. Table 6.1 is a guide to the selection of filler metals for specific atmospheres. Tests for Braze Atmospheres (Ref 9). The best solution to evaluate the quality of braze atmospheres is to use a stainless steel T-specimen. By placing the T-specimen in the furnace during brazing of production assemblies, an excellent measuring stick for quality control is developed. Making T-Specimens. Reference 9 gives the complete details on specimen size and suggested type of stainless steel. In interpreting the results of the braze test, a furnace operator must observe and determine if the furnace-atmosphere quality has diminished. Table 6.2 indicates how a braze furnace operator might interpret five different situations that can occur inside the brazing furnace when using nickel-base filler metals. An inspector checking
surface condition can use the chart to assess furnace-atmosphere quality. Reducing Atmospheres. A reducing atmosphere is one that is capable of chemically removing surface contamination from metals. Gases that provide reducing conditions are generally proprietary mixtures that liberate halogen radicals. Specific gas-handling systems are usually needed for these in order to satisfy health and safety legislation. For a few metals, hydrogen is satisfactory as a reducing atmosphere. No less important for meeting its functional requirement than the oxygen partial pressure of the gas is its water content. Hydrogen is a relatively difficult gas to dry, and the water vapor present can present a serious problem. There is also the risk of explosion when dealing with hydrogen at high temperatures, and hydrogen can embrittle some materials. Vacuum Atmospheres. Vacuum is frequently used as a protective environment for filler-metal joining processes. Vacuum offers several advantages compared with a gas atmosphere, particularly the ability to measure and control the oxygen partial pressure more readily. In a substantially leak-free system, the oxygen partial pressure is one-fifth of the vacuum pressure, which is relatively easy to determine, as compared with direct measurement of oxygen partial pressure. Although a roughing vacuum of 10 mPa will provide an atmosphere with the same oxygen partial pressure as a standard inert gas, it is possible to improve on this value, by several orders of magnitude, by using a highvacuum pumping system. Alternatively, a low oxygen partial pressure may be achieved by obtaining a roughing vacuum, backfilling with an inert gas, and then roughing out again. The effect of the second pumping cycle is to reduce the oxygen partial pressure to less than typically one-thousandth of that in the inert gas, that is, approximately 10 µPa. This estimate assumes that the furnace chamber is completely leaktight and does not outgas from interior surfaces, nor does any oxygen or water vapor backstream through the pump (Ref 10). The disadvantages of using a vacuum system for carrying out a joining process are, principally, restricted access to the workpiece and the inadvisability of using either fluxes or filler metals with volatile constituents, such as cadmium, because the vapors can corrode the vacuum chamber, degrade its seals, and contaminate the pumping oils. This problem is not limited to the well-known volatile elements. Many metals that
250 / Brazing, Second Edition
Table 6.1 Brazing filler-metal selection chart For brazing jet, aerospace engine, automotive, and nuclear reactor components; airframe structures; honeycomb; heat exchangers; dairy, chemical, medical, and food processing equipment Designation
AWS A5.8
Suitability for specific brazing applications(a) For hightemperature, high-stress moving engine components
For heavy, nonmoving structures (variable gaps)
A A B B C A A A C A
A A A B B A A A A A
C C B B B B C C C B
A
B
... ...
C C B
4783
A
AMS
For nuclear For honeycomb reactor and other thin core materials assemblies
For large, machinable or softer fillets
For use in contact with NaK
For use with tight or deep joints
(b) (b) (b) (b) (b) (b) (b) (b) (b) (b)
B B C C A C B B A C
A(c) A A(c) A(c) B(c) A B A B A
C C B B C B C C C C
A
A
C
A(c)
B
C C C
A A A
B A A
C C C
C A(c) A(c)
A A A
B
A
(b)
C
A
B
Nickel-base with boron BNi-1 BNi-1a BNi-2 BNi-3 BNi-4 BNi-9 BNi-10 BNi-11 Nicrobraz 160(d) Nicrobraz 200(d)
4775 4776 4777 4778 4779 ... ... ...
Nickel-base with silicon BNi-5
4782
Nickel-base with phosphorus BNi-6 BNi-7 Nicrobraz 51(d) Cobalt-base BCo-1
Comparative properties (1 = highest)
Joint strength(e)
Solution and diffusion with base metal
Fluidity
Resistance
°C
°F
1 1 1 2 3 1 1 1 3 1
1 1 2 2 3 2 3 3 3 3
3 4 2 2 3 2 4 3 4 3
1 2 2 2 3 1 1 1 4 2
1205 1205 1090 1090 980 1205 1205 1205 925 1090
2200 2200 2000 2000 1800 2200 2200 2200 1700 2000
1
4
2
2
1205
2200
... ...
4 2 2
4 5 4
1 1 2
5 5 ...
775 860 855
1400 1575 1570
4783
1
4
2
1
1205
2200
Designation AWS A5.8
AMS
Oxidation resistance (f ) of joints up to indicated temperature
Nickel-base with boron BNi-1 BNi-1a BNi-2 BNi-3 BNi-4 BNi-9 BNi-10 BNi-11 Nicrobraz 160(d) Nicrobraz 200(d)
4775 4776 4777 4778 4779 ... ... ...
Nickel-base with silicon BNi-5
4782
Nickel-base with phosphorus BNi-6 BNi-7 Nicrobraz 51(d) Cobalt-base BCo-1
(continued)
Chapter 6: Fluxes and Atmospheres / 251
Table 6.1 (continued) Designation AWS A5.8
Suggested brazing temperature
Brazing range(g)
AMS
Solidus(h)
Liquidus(h) Recommended atmosphere(i)
°C
°F
°C
°F
°C
°F
°C
°F
1065–1205 1065–1205 1010–1175 1010–1175 1065–1205 1065–1205 1150–1205 1150–1205 1150–1205 1065–1175
1950–2200 1950–2200 1850–2150 1850–2150 1950–2200 1950–2200 2100–2200 2100–2200 2100–2200 1950–2150
1175 1175 1040 1040 1175 1175 1175 1175 1190 1120
2150 2150 1900 1900 2150 2150 2150 2150 2175 2050
970 970 970 980 990 1055 970 970 970 975
1780 1780 1780 1800 1810 1930 1780 1780 1780 1790
1040 1075 1000 1040 1055 1055 1105 1095 1160 1040
1900 1970 1830 1900 1935 1930 2020 2000 2120 1900
I, V I, V I, V I, V I, V I, V I, V I, V I, V I, V
1150–1205
2100–2200
1190
2175
1080
1975
1135
2075
I, V, D
925–1095 980–1095 980–1095
1700–2000 1800–2000 1800–2000
980 1065 1065
1800 1950 1950
875 890 880
1610 1630 1620
875 890 950
1610 1630 1740
I, V, D, E I, V, D I, V, D
1150–1230
2100–2250
1175
2150
1105
2025
1150
2100
I, V
Nickel-base with boron BNi-1 4775 BNi-1a 4776 BNi-2 4777 BNi-3 4778 BNi-4 4779 BNi-9 ... BNi-10 ... BNi-11 ... Nicrobraz 160(d) Nicrobraz 200(d) Nickel-base with silicon BNi-5
4782
Nickel-base with phosphorus BNi-6 ... BNi-7 ... Nicrobraz 51(d) Cobalt-base BCo-1
4783
AWS, American Welding Society; AMS, Aerospace Material Specification. (a) A, best; B, satisfactory; C, least satisfactory. (b) Contains boron; has high neutron absorption. May be used in nuclear plant equipment but not in core. (c) Tested and approved by U.S. Department of Energy laboratories and by private industry manufacturers of nuclear reactors. Tests were conducted on brazed joints of types 304 and 301 stainless steel and Inconel base metals. (d) Trademark of Wall Colmonoy Corporation. (e) Joint strength depends on brazing cycle, joint design, joint clearance, base metal, and so on. (f) Tests conducted on Inconel base-metal joints. Exposed 500 h in still air at temperature indicated. No deterioration of fillet. BNi-10 tests conducted on Hastelloy X. (g) The exact brazing temperature for any specific joint depends on the joint and base-metal properties desired. It will also depend on the different base-metal, filler-metal, and joint design combinations. Consequently, it may sometimes be necessary to determine the ideal brazing temperature by experiment. (h) Data taken from cooling curves prepared in Wall Colmonoy Corporation laboratories. (i) Recommended atmospheres for brazing filler metals; stainless steels and high-chromium base metals require I, V, or D. I, pure dry hydrogen or inert gases; V, vacuum; D, disassociated ammonia, nitrogen atmosphere (–50 °C, or –60 °F, dewpoint or drier); E, exothermic, rich, unpurified 6:1 air-to-gas ratio, or purified and dried. Source: Wall Colmonoy Corporation
Table 6.2 Furnace-atmosphere conclusions drawn from T-specimen tests Production assembly Appearance
Filler-metal flow
Bright Dark Dark Dark
Good Poor or none Good Poor or none
Bright
Poor or none
T-specimen Appearance
Conclusions
Filler-metal flow
Atmosphere
Filler metal
Other
Bright Bright Dark Dark
Good Good Good Poor or none
Good Good Good during heating Bad during heating
Good Good Good ?
Bright
Poor or none
Good
Poor?
... Base-metal problem Quench leak or removed hot Furnace leak? Contaminated furnace? Atmosphere leachout of B, Si, or P?
252 / Brazing, Second Edition
have a negligible vapor pressure at normal ambient temperatures will volatilize during high-temperature brazing processes (>1000 °C, or 1830 °F), particularly when these entail using reduced-pressure atmospheres. Manganesecontaining filler metals and base materials fall into this category, because the vapor pressure of this element is 1 Pa at 1000 °C (1830 °F). Another source of oxidizing contamination in a vacuum system is oil vapor mixed with air and water vapor backstreaming from a rotary pump. This can occur whenever the pressure inside the vacuum chamber drops below 1 Pa but can be largely eliminated by employing a foreline trap or by isolating the pump from the chamber once the required pressure reduction has been obtained. Vacuum System Types and Applications. Researchers (Ref 11) designed several vacuum furnaces that were used to braze electric power interrupter assemblies of the vacuum arc suppression type. These interrupters are used extensively in electric power industries by utilities and heavy electric power users in electrical switch gear. What makes these furnaces unique by industry standards is that they attain deep vacuum without prolonged vacuum pumping time. The 12.70 cm (5 in.) diameter by 15.24 cm (6 in.) deep vacuum chamber and hot zone reach 1 × 10–5 Pa (1.5 × 10–9 psi) in approximately 1 h after work loading and maintain a pressure in the 0.00013 Pa (1.89 × 10–8 psi) range during the brazing cycle to 871 °C (1600 °F). Instead of the standard high-vacuum diffusion pump, the cryogenic Cryopump (CTI Division, Helix Technology Corp.), used for each facility, and other system modifications allow deep-vacuum processing. The high-vacuum operation is accomplished by means of the Cryopump without residual oil backstreaming—a special requirement of this process. Each furnace is microprocessor controlled with programmable logic controller and a vacuum controller for automatic regeneration of the Cryopump, all interlocked with a programmable controller for temperature and process cycle control. The furnaces also include a separate external recirculating argon gas blower and gas-to-water heat-exchanger arrangement, with high-vacuum seal-off gate valves to allow increased cooling of the product below 538 °C (1000 °F) to ambient temperature, to reduce process time.
A typical cycle for vacuum furnace brazing of electric power interrupter assemblies of stainless steel material is shown in Fig. 6.3. Another distinct type of brazing facility is the 500 kPa (5 bar) rapid gas-quenching vacuum furnace (Abar Ipsen H4848 5 bar TurboTreater). Complete with a work zone of 914 mm (36 in.) high by 762 mm (30 in.) wide by 1220 mm (48 in.) in length and a load capacity of 1150 kg (2535 lb), it is capable of quenching with argon or nitrogen from pressures of 1 to 5 bar (100 to 500 kPa). Rarely is there a requirement for the 5 bar pressure capability, and normally, most work is run in the 3 bar range—this equates to a true cost advantage in the amount of cooling gas used on a per-run basis. Additional savings occur as a result of the unique internal gasquenching system found on the furnace (it uses 40% less backfill gas as compared to conventional furnace designs with external quenching systems) (Ref 12, 13). A study was conducted (Ref 14) on brazing titanium-vapor-coated ZrO2. The ZrO2 surfaces were coated with titanium by radio frequency sputtering and electron beam evaporation. The titanium-coated ZrO2 was easily vacuum brazed to itself and to copper-plated nodular cast iron with Ag-30Cu-10Sn wt% filler metal. Braze joints made with titanium-sputter-coated ZrO2 contained high levels of porosity. In contrast, joints made with ZrO2 that was titanium coated with the electron beam evaporation process were free of porosity.
980 ˚C Vac. cool 850 ˚C 400 ˚C 30 ˚C
Argon quench
20 ˚C 50 ˚C 1
2
3
4
5
6
7
8
9
10
Time
Fig. 6.3
Diagram of a typical vacuum furnace brazing cycle. 1. Hold part at room temperature, 20 °C (68 °F), for 3 min. 2. Increase temperature 2 °C/min (4 °F/min) to 30 °C (86 °F). 3. Hold and soak part for 1 min at 30 °C (86 °F). 4. Increase temperature 15 °C/min (27 °F/min) to 850 °C (1560 °F). 5. Hold at 850 °C (1560 °F) for 5 min. 6. Increase temperature 12 °C/min (22 °F/min) to 980 °C (1800 °F). 7. Hold at 980 °C (1800 °F) for 10 min. 8. Decrease temperature to 400 °C (750 °F) and hold for 2 min. 9. Continue to cool to 50 °C (120 °F) in 100 min. 10. Cool part to room temperature (30 min) and remove. Source: Ref 11
Chapter 6: Fluxes and Atmospheres / 253
The titanium coatings reacted with the ZrO2 during the brazing operation, as evidenced by a darkening of the ceramic beneath the coated surfaces. Thermodynamic analysis predicted that the most likely reaction between the ZrO2 and titanium was oxygen diffusion from the ZrO2 into the titanium coating. Flexure bars for four-point bend testing were made from both ZrO2-ZrO2 and ZrO2-Fe joints and tested at 25, 200, 400 and 575 °C (77, 390, 750, and 1065 °F). The highest strengths were recorded for room-temperature tests: 571 MPa (83 ksi) for ZrO2-ZrO2, and 399 MPa (58 ksi) for ZrO2-Fe. In both cases, failures in the highstrength joint specimens were associated with fracture of the ZrO2 beneath the titanium-vaporcoated surfaces. The strength of the joints decreased at elevated temperatures. For the ZrO2-ZrO2 joints, the decrease in strength was accompanied by an increase in the amount of metallic debonding at the ZrO2/vapor coating interfaces. At 400 °C (750 °F), no fracture of the ZrO2 was observed, and failure occurred predominantly by debonding. At 575 °C (1065 °F), joint failure in both types of joints occurred through the filler-metal layer. This study confirmed that vapor coating with titanium is an effective way of promoting the brazing of ZrO2. Analysis of braze joint test specimens indicated that adhesion of the titanium vapor coating to the brazed ZrO2 surfaces was generally very good. The data further suggested that obtaining good coating adherence is a prerequisite to obtaining high-strength braze joints, and that improvements in adhesion would improve joint strength at all test temperatures. Aluminum brazing in vacuum furnaces is not a new concept. People have been brazing aluminum since the 1940s in atmosphere or salt and in vacuum since the 1950s. Aluminum brazing in high vacuum generally has been done with a getter such as magnesium. This technique has been applied to the high-production brazing of complex heat exchangers, such as in the aluminum evaporator assembly for automobile air conditioners. Another approach is to use a flux and a partial-pressure vacuum atmosphere, particularly for low-production or job shop applications. This technique and several applications are discussed in the section “Fluxes” later in this chapter. The most difficult part about aluminum brazing is the removal or reduction of the aluminum
oxide layer. The parts must be extremely clean before applying any filler metal. Usually, an acid pickling of some kind is used to remove the aluminum oxide layer. The cleaning process can be quite involved, with many different pickling and rinsing operations. After the parts are cleaned, it is very important that they be assembled and brazed as soon as possible to minimize oxidation. Aluminum will form an oxide layer at room temperature, and this layer is very difficult to reduce (break down). A typical guideline is a maximum of 4 h from the time the parts are cleaned to when the parts are placed in the furnace. Researchers have conducted work (Ref 15) on fluxless brazing with aluminum. The first method discussed previously involves the fluxless aluminum brazing in high vacuum using a getter. A getter is a material, usually magnesium, that is placed in the furnace to react with the remaining oxygen. Therefore, the getter is used to purify the atmosphere. The magnesium may even react with the oxygen in the aluminum oxide layer to reduce it. This reaction with the oxide layer is a hypothesis and is noted in the literature, but it offers another reason for including magnesium as an alloy in some of the clad fillermetal sheets and wire. The magnesium will flash off at 454 °C (849 °F) and, in doing so, will trap or use up the remaining oxygen in the chamber. Following is a typical brazing run: 1. Clean the parts. 2. Assemble the parts with gloves (the fillermetal sheet is usually foil, wire, or a clad material). 3. Place a small crucible of magnesium in the furnace along with the parts to be brazed. 4. Install work thermocouples in the part, usually type K. 5. Pump furnace to 1 × 10–2 Pa (1.4 × 10–6 psi) or lower. 6. Heat furnace to 260 °C (500 °F). 7. Outgassing will occur, and the vacuum will deteriorate. Hold at 260 °C (500 °F) until vacuum returns to 1 × 10–2 Pa or lower and the work is above 177 °C (351 °F). 8. Increase temperature to 538 °C (1000 °F). Hold until the part reaches 454 °C (849 °F). The vacuum should be 1 × 10–2 Pa or lower. 9. Increase temperature to 604 °C (1119 °F); when the part reaches 593 °C (1099 °F), hold 1 min. Turn off hot-zone power vacuum, cool to 454 °C (849 °F) or below, then
254 / Brazing, Second Edition
backfill furnace with nitrogen and quench out. This is an example of a cycle that has worked for brazing blower wheel assemblies, among others. The cycle most likely will have to be modified for certain other applications, depending on the parts configuration and the furnace condition. The cleanest atmosphere possible is necessary for aluminum brazing. The suggestions that follow for improving the furnace atmosphere may not be necessary for production brazing, but they are used as an assurance on one-timejob-type runs. Multiple pumpdowns by backfilling with an inert gas will help (Ref 16). An alternate method is to pump into high vacuum (1 × 10–2 Pa, or 1.5 × 10–6 psi), run partial pressure (1300 Pa, or 1.9 psi) for 10 min, and then back into high vacuum. This partial pressure purge may be repeated, and then pump at least 1 h in high vacuum before applying heat. The longer the pump, the cleaner the atmosphere will be. Also, the furnace should be heated slowly (not over 6.67 °C/min, or 12 °F/ min) to reduce any large outgassing and to keep the pressure low (better vacuum). Vacuum furnaces used for aluminum brazing are predominantly of the cold-wall design. Water cooling keeps the chamber wall temperature within safe operating limits. Water in the furnace jackets is allowed to run warm, so that the chamber wall temperature is above the atmospheric air dewpoint, thus preventing condensation. Batch vacuum brazing furnaces are run with brief unloading and reloading cycle times. In this way, the hot-zone internals remain hot, generally above 260 to 316 °C (500 to 601 °F). This limits the adsorption of atmospheric air and moisture, making subsequent vacuum cycles rapid and contamination-free, thereby providing excellent conditions for brazing. Although it may be argued that aluminum brazing in a vacuum can be accomplished at 0.2 µm or low four-scale pressures, the mass of contaminants flowing cannot be tolerated when that high. Good vacuum practice for aluminum brazing dictates that the true leak rate, excluding outgassing, for a 1.42 m3 (50 ft3) chamber should be in the range of 5 µm/h. The gas load from this leak rate would be 5 µm/h × 1.43 m3 = 250 µm ft3/h. The sensitivity range used on vacuum furnaces for aluminum brazing would be 5 × 10–10
std cm3/s. That is, no individual leak would exceed this value. A researcher (Ref 17) also investigated fluxless brazing of aluminum alloys in vacuum. He developed a method whereby the oxide film was removed using an activating metal (magnesium). This metal may be included in the composition of brazed aluminum alloys or was introduced into the alloy or placed in the working space of the vacuum furnace in which brazing was carried out. His theory on the mechanism of failure and removal of the oxide film when fluxless brazing aluminum alloys in vacuum in the presence of the activating metal (magnesium) may be described as follows (Fig. 6.4). In heating during brazing, magnesium rapidly evaporates, and its vapors efficiently bond oxygen in the working space of the furnace. The resultant magnesium oxides are gradually removed by a vacuum pump. As the oxygen content decreases and the partial pressure of the magnesium vapors increases, suitable conditions are created for reducing aluminum from its oxides by magnesium vapors. A porous layer of the magnesium oxide forms on the aluminum surface and enables the access of magnesium vapors to the aluminum surface. Consequently, magnesium vapors are adsorbed
Al2O3 Filler metal Al
Al (a)
Mg Al2O3 Mg (b) Mg
Filler metal
Al2O3
Mg Al2O3
(c)
Fig. 6.4
Mechanism of vacuum brazing aluminum alloys in magnesium vapors. (a) 20–300 °C (70–570 °F). (b) 300–560 °C (570–1040 °F). (c) 600 °C (1110 °F)
Chapter 6: Fluxes and Atmospheres / 255
on the aluminum surface with the formation of a low-melting liquid phase, which, spreading below the oxide film, ignites and fractures it. During melting, the filler metal, which spreads on the surface of the parent material, forms a brazed joint, and the oxide film is removed. Fluxless brazing of aluminum alloys was carried out with aluminum-silicon filler metals deposited on the surface of the parent metal by cladding as well as by the contact-reactive method by depositing copper and silver. In contact melting with aluminum at 548 and 550 °C (1018 and 1022 °F), respectively, these metals form a liquid phase of eutectic composition, which is also used as the filler metal. Copper and silver are deposited on the surface of aluminum alloys, clad with aluminum-silicon filler metals, by electroplating or thermal vacuum spraying to reduce their melting point. The chemical composition and mechanical properties of the filler metals, with an allowance made for alloying them with magnesium (activating metal), are presented in Table 6.3. The researcher also believes (Ref 17) that a special feature of fluxless brazing sections of aluminum alloys in vacuum is the presence of an atmosphere of magnesium vapors. Their concentration and uniformity of distribution depend on the residual pressure, the temperature gradient in the component, and the size of the working volume of the furnace in which the sections are brazed. The optimal conditions in his work were a vacuum of 10–2 to 10–3 Pa (10–6 to 10–7 psi), uniform heating with a temperature gradient of ±10 °C (±18 °F), and the formation of a minimum vacuum volume around the brazed section. The most efficient procedure was to use standard vacuum furnaces for brazing, ensuring the opti-
mal conditions with regards to temperature and vacuum, or to develop special vacuum heating systems for brazing sections of aluminum alloys.
Atmosphere Application One type of controlled atmosphere is the product of combustion of a torch flame; a neutral or reducing flame is normally used. Separately supplied controlled atmospheres may also be used with induction or resistance brazing, but controlled atmospheres are most commonly used in furnace or retort brazing operations. In fact, furnace brazing requires the use of a suitable atmosphere to protect assemblies against oxidation and, in the case of steels, against decarburization during brazing and during cooling, which is accomplished in chambers adjacent to the brazing, especially where titanium, zirconium, and refractory metals are concerned. The principle followed in the use of controlled gas atmospheres involves the preparation of a special protective gas and its introduction into the furnace or brazing retort at pressures above atmospheric. As the gas is continuously supplied to the furnace and circulated through it, the furnace becomes purged of air. The protective gas atmosphere is maintained at a slight pressure, which prevents air from seeping into the brazing retort or furnace. In some operations, work is placed in a cold retort or furnace prior to purging, and the retort or furnace is not opened until the brazing cycle is completed. Where parts must be fed continuously or periodically into a furnace that is at brazing temperature, gas curtains or intermediate chambers are provided to avoid contamination of the furnace atmosphere.
Table 6.3 Chemical composition and mechanical properties of filler metals used for fluxless brazing of aluminum Content, %
Melting temperature (Tm)
Filler metal
Ag
Cu
Si
Mg
°C
Al-Mg-Ag Al-Cu-Mg
26–28 ... ... ... ... ... ... ... ...
... 28–32 ... ... ... 25–28 8–20 8–10 3–5
... ... 6–8 8–10 10–12 6–8 8–10 10–12 10–12
4–6 4–6 1–2 1–2 1–2 1–2 1–2 1–2 1–2
540–545 505–510 605–610 590–600 575–580 525–530 550–560 555–565 570–575
Al-Si-Mg
Al-Cu-Si-Mg
Note: Balance is aluminum.
°F
1000–1015 940–950 1120–1130 1095–1110 1070–1075 975–985 1020–1040 1030–1050 1060–1065
Bending strength (σB) MPa
110–130 70–90 80–100 100–120 120–140 70–90 100–120 100–120 110–130
ksi
16–19 10–13 12–15 15–17 17–20 10–13 15–17 15–17 16–19
256 / Brazing, Second Edition
The ability to control the composition and therefore the effectiveness of a furnace atmosphere depends not only on the condition and proper operation of the atmosphere-producing equipment but also on the proper setup and operation of the furnace being used. When certain types of controlled atmospheres, such as those containing hydrogen, are employed, extreme care must be taken to prevent the formation of explosive mixtures of gas. Mixtures of hydrogen with air ranging from 4 to 75% H are explosive. As a safety precaution when potentially explosive gas atmospheres are used, the furnace or retort should be thoroughly purged with the gas to ensure the removal of all air before heat is applied. Waste gases from the furnace can be either continuously burned or directed into the open air outside the building. Some atmospheres, such as those containing carbon monoxide, are toxic. Proper burning off or disposal of the waste gases from these atmospheres is especially important for safety. In brazing of toxic metals such as beryllium, waste gases should be carefully filtered or piped to an outside area. Many brazing atmospheres are generated by passing metered mixtures of hydrocarbon fuel gas and air into a retort for reaction. Most of these atmospheres are rich exothermic mixtures in which the heat liberated from the reaction is sufficient to continue it. A rich exothermic atmosphere is the least expensive of the generated atmospheres, is adequately reducing for many applications, has relatively low sooting potential, and requires a minimum of generator maintenance. Approximately 70 to 80% of all brazing atmospheres are exothermic, and they are generally used to braze mild steel or lowcarbon steel. Recently, a firm attempted unsuccessfully to furnace braze carbon steel components in an exothermic atmosphere using BAg-1 filler metal. They found that the following filler metals were not acceptable, because they contained cadmium and/or zinc, and no atmosphere—even vacuum—would allow for a good braze joint: BAg-1a, BAg-2, BAg-2a, BAg-3, and BAg-27. The cadmium in a filler metal will normally vaporize in an atmosphere and will drop out as a fine dust in the furnace, to be stirred up later, or it will be carried out into the room if there is an inadequate exhaust system. In a vacuum furnace, however, the cadmium and zinc will be evaporated from the filler metal and deposited
in the heat shielding and on the colder electrical insulators. This can cause considerable problems. Furthermore, this will occur even with the use of a partial pressure in the vacuum furnace. Thus, cadmium should not be used in a vacuum/inert atmosphere furnace. Filler metals of silver-copper that contain zinc (but no cadmium) will also vaporize in an atmosphere, whether it has a dewpoint of 26.7 °C (80.1 °F) or down to –62 °C (–80 °F). Sometimes, flux is used to reduce vaporization, and, while this works, it is not completely satisfactory. Furthermore, flux also introduces a cleaning operation. It is recommended that, if needed, silver filler metals containing elements other than cadmium and zinc be used in atmosphere furnaces. In a vacuum furnace, it is necessary to use a partial pressure, particularly at higher temperatures, to prevent vaporization of silver and copper. Exothermic or endothermic gases are chiefly made by the controlled combustion of natural or synthetic gases with air to form a mixture essentially composed of nitrogen, hydrogen, methane or ethane, carbon dioxide, carbon monoxide, and water vapor. As the ratio of fuel gas to air is increased, a mixture becomes endothermic; it requires the addition of heat and a catalyst for combustion to occur. Endothermic gas mixtures are used in brazing medium and high-carbon steels and sometimes mild steels.
Atmosphere Composition The compositions of controlled atmospheres recommended for brazing cover a wide range (Table 6.4). These data are not intended as a comprehensive tabulation of atmosphere-metal combinations but rather as a general outline of some of the more widely used combinations. Dewpoint Control. The combustion of gas mixtures results in a controlled atmosphere containing entrained moisture, which is largely undesirable in brazing. The moisture can sometimes be removed by condensation. The use of certain filler metals, however, requires cooling in conjunction with absorption-type driers to reduce the dewpoint to satisfactory levels. Accurate dewpoint control is especially important when dry hydrogen atmospheres are required because of the sensitivity to moisture of the metals usually brazed in this type of atmosphere. Dissociated ammonia atmospheres do not always require such accurate control.
Combusted fuel gas, dried (carburizing) Dissociated ammonia
Cryogenic or purified N2 Cryogenic or purified N2 Deoxygenated and dried hydrogen Heated volatile materials (inorganic vapors—zinc, cadmium, lithium, volatile fluorides) Purified inert gas (e.g., helium, argon)
Vacuum above 266.6 Pa (2 torr) Vacuum from 66.65 to 266.6 Pa (0.5 to 2 torr) Vacuum from 0.13 to 66.65 Pa (0.001 to 0.5 torr) Vacuum of 0.13 Pa (10–3 torr) and lower
4
6A 6B 7
10 10A ... ...
...
... ...
...
...
... ...
...
...
1–30 2–20 100
–68 °C (–90 °F) –29 °C (–20 °F) –59 °C (–75 °F)
...
75
38–40
15–16
14–15
0.5–1
H2
–54 °C (–65 °F)
–40 °C (–40 °F)
–40 °C(–40 °F)
20 °C (68 °F)
20 °C (68 °F)
Maximum dewpoint of incoming gas
...
...
... ...
...
...
70–99 70–97 ...
25
41–45
73–75
70–71
87
N2
...
...
... ...
...
...
... 1–10 ...
...
17–19
10–11
9–10
0.5–1
CO
Composition of atmosphere, %
...
...
... ...
...
...
... ... ...
...
...
...
5–6
11–12
CO2
BNi, BAu, BAlSi, titanium alloys
BCu, BAg
BCuP, BAg BCu, BAg
Same as 5
BAg
BAg(b), BCuP, RBCuZn(b), BCu, BNi Same as 5 Same as 5 Same as 5
Same as 2
BCu, BAg(b), RBCuZn, BCuP Same as 2
BAg, BCuP, RBCuZn
Filler metals
Heat- and corrosion-resisting steels, aluminum, titanium, zirconium, refractory metals
Carbon and low-alloy steels, copper
Copper Low-carbon steels, copper
Same as 5 plus titanium, zirconium, hafnium
Copper(c), brass, low-carbon steel, nickel, Monel, medium-carbon steel(d) Same as 2 plus medium- and high-carbon steels, Monel, nickel alloys Same as 2 plus medium- and high-carbon steels Same as 1, 2, 3, 4 plus alloys containing chromium(e) Same as 3 Same as 4 Same as 5 plus cobalt, chromium, tungsten alloys and carbides(e) Brasses
Copper, brass(b)
Base metals
...
...
Special purpose. Parts must be very clean, and atmosphere must be pure. ... ...
Special purpose. May be used in conjunction with 1 thru 5 to avoid use of flux.
... ... ...
...
Carburizes
...
Decarburizes
...
Remarks
(a) Types 6, 7, and 9 include reduced pressures down to 266.6 Pa (2 torr). (b) Flux required in addition to atmosphere when alloys containing volatile components are used. (c) Copper should be fully deoxidized or oxygen-free. (d) Heating time should be minimized to avoid objectionable decarburization. (e) Flux must be used in addition to the atmosphere if appreciable quantities of aluminum, titanium, silicon, or beryllium are present.
10C
10B
9
8
5
3
2
Combusted fuel gas (low hydrogen) Combusted fuel gas (decarburizing) Combusted fuel gas, dried
Source(a)
1
Brazingatmosphere number
Table 6.4 Atmospheres for brazing
258 / Brazing, Second Edition
The ability of pure hydrogen to reduce metal oxides is determined by the temperature, the oxygen content (measured as dewpoint), and the pressure of the gas. Because furnaces typically operate at atmospheric pressure, only temperature and dewpoint play a part (Ref 2). The diagram presented in Fig. 6.5 is a plot of the dewpoint at which the oxide and the metal are in equilibrium at various temperatures. The 20 curves shown in this diagram define the equilibrium conditions for 20 pure metal/metal oxide systems. The positions of 13 additional elements whose curves fall outside the chart are also indicated. The oxides chosen for the calculations of this diagram represent the most difficult-to-reduce oxide of each metal. The metal/metal oxide equilibrium curves slope upward and to the right for each metal. The region above and to the left of each curve represents conditions that are oxidizing for that metal. All points below and to the right of each curve cover the conditions required for reducing the oxides. The diagram therefore illustrates that the higher the processing temperature, the
higher the dewpoint (or oxygen content) that can be used for any particular metal. In other words, a given purity of hydrogen becomes progressively more reducing at progressively higher temperatures, or, to put it another way, the higher the brazing temperature, the lower the H2:H2O ratio can be for any given metal (Ref 2). Use of the diagram in Fig. 6.5 for practical purposes requires, first, that the correct curve be selected. For processing of any alloy, the element having the most stable oxide (farthest to the right) is the governing curve. If copper is to be brazed to stainless steel, for example, then a ratio suitable for reducing the chromium oxide must be selected. The chromium oxide curve applies, because chromium oxides are more stable than those of iron or nickel. Generally, it has been found that when the most difficult-toreduce constituent of an alloy is present in more than approximately 1 at.%, a continuous film of its oxide is formed, and its curve therefore is applicable. Alloys having a concentration progressively lower than 1 at.% of the most stable
Temperature, ˚F
+80
500
100
1000
1500
2000
2500
3000
3500 Pa
+40 +30
Au, Pt, Ag, Pd, Ir, Cu, Pb, La Co, Ni, is more Sn, Os, difficult to Bi reduce than those plotted
Mo O W O 2 2
+20 +60 +10
Torr
Metals easier to reduce than those plotted:
+40 0
103
101
–10
–30 Mo
–60 –80 –100
re
–40 Mo
re
–50
red
ox id uc
izin g
ing
O5 Ta 2
O
Mn
Cb
2 SiO
O3 B2
–60 –70
O3 Al 2
–80 O Ti
–120
Ba O
–90 –140
–100 100
500
1000 Temperature, ˚C
Fig. 6.5
Metal/metal oxide equilibria in hydrogen atmospheres. Source: Ref 2
10o
VO
O
O
–40
102
O3 Cr 2
101
10–1
10o
10–2
10–1
10–3
10–2
10–4
M gO
–20
–20
Na 2O
0
Zn
Dewpoint of hydrogen
+20
1500
O2 Zr
UO 2
O Ca O Be O2
Th
2000
Partial pressure of water vapor (May be read as vacuum furnace pressure)
+100
˚C
Fe O
˚F
Chapter 6: Fluxes and Atmospheres / 259
oxide-former appear to lie progressively closer to the curve of the next most stable oxide-former. From Fig. 6.5 it can be seen that chromium oxide can be reduced at 815 °C (1540 °F) if the dewpoint of hydrogen is lower than –56 °C (–69 °F); at 1095 °C (2005 °F), oxide reduction will occur in a hydrogen atmosphere with a dewpoint lower than –29 °C (–20 °F). When a pure metal is brazed, the curve representing that metal is used to determine the temperature and hydrogen dewpoint at which oxide reduction will occur. In the case of an alloy, the curve for the element that forms the most stable oxide is used to determine the conditions for oxide reduction. The need to control the dewpoint of the hydrogen atmosphere is evident from these considerations, and many devices and systems have been developed to accomplish this objective. To ensure accuracy, dewpoint measurements should be made at the furnace outlet rather than at the inlet. In practice, it is necessary to use hydrogen that has a somewhat lower dewpoint than that indicated by the curve for any given metal, partly because the surface becomes oxidized during heating until the temperature is reached that corresponds to the equilibrium temperature for that dewpoint. The reduction of these oxides formed during heating requires that sufficient time be allowed at conditions sufficiently below, or to the right of, the equilibrium curve. It is also necessary in practice to provide a continuous flow of hydrogen into the work zone during processing to sweep out the outgassed contaminants and thus maintain the necessary atmosphere purity at the metal surface. A practical example of dewpoint is one involving an automotive fluid coupling, which is a bowl-shaped part with a number of vanes brazed to the bowl and a stamping on the top part of the vane. The copper filler metal is only right at the brazed joint and in the fillet but does not flash out over the rest of the low-alloy steel. The brazing takes place in a continuous furnace using a copper oxide paste, BCu-2. In the past, engineers considered that the flowing out of the copper from the joint, which is sometimes called flashing or blushing, indicated a better braze. In reality, both can have equal strength, and, on occasion, if the filler metal is retained at the braze joint, there may be a somewhat larger fillet on a horizontal. However, not
much difference can be obtained on a vertical joint, which is quite small due to gravity. The primary controlling factor that determines the wetting of the filler metal across the alloy steel surface is the dewpoint of the atmosphere, everything else being equal. Brazing specialists found that with a 7-to-1-ratio exothermic atmosphere, a standard dewpoint of approximately 16 to 21 °C (61 to 70 °F) would cause copper to flow from the joint across the surface for some distance. To reduce the flow and keep it in a large clearance gap, it was necessary to increase the atmosphere dewpoint in the furnace to 32 °C (90 °F) and above, so this would hold the copper in the braze joint and still produce good, high-strength copper brazed joints in the press fit joint. In most furnaceapplication experiences, the higher dewpoint was obtained by reducing the flow of atmosphere into the furnace, thus allowing the incoming parts to bring in the oxygen and moisture on their surfaces and internally, when the part was so designed. With less atmosphere flow, as the parts outgassed on heating, less of the partial pressure of oxygen was removed from the furnace with the flow of the atmosphere, thus increasing the dewpoint. With everything else being held constant, the flow can also be altered by reducing the percentage of hydrogen, which means changing the ratio from 7 parts air to 1 part gas to 8 or 9 parts air to 1 part gas. The same copper flow effect would also be noted in a nitrogen/hydrogen atmosphere, in which there is normally 2% H2 in the nitrogen carrier gas. In this atmosphere, the change in dewpoint in the furnace will also change the copper flashing characteristics of the copper on the carbon steel. Many years ago, it was noted that in a hydrogen atmosphere, as the dew-point increased to the 27 to 32 °C (81 to 90 °F) dewpoint, the copper stayed close to the joint and did not flash out onto the carbon steel. Conversely, as the dewpoint was lowered, more flashing occurred across the carbon steel. When dewpoints reached the –51 to 57 °C (–60 to 135 °F) range, it was found that the copper in a 0.025 mm (0.001 in.) thick joint in carbon steel could flow out of the joint and across the surface of the carbon steel. Thus, while very low dewpoints were needed to braze chromium-containing steels, to prevent oxidation of chromium, it was impractical to run carbon steel parts brazed with copper
260 / Brazing, Second Edition
in the same load, because difficulties were encountered in keeping the copper in the joint area. When running carbon steel in hydrogen, the flow of drier hydrogen was reduced, thus allowing the buildup of the partial pressure of oxygen in the retort system so the copper would stay in the joint. The copper brazing of carbon steel has been used for many years to fabricate large production quantities of parts and has been a very good tool for the brazing industry. However, a better understanding of the atmosphere, particularly on the shop floor, would be helpful in obtaining the desired results. Some people use a 21 °C (70 °F) dewpoint atmosphere with very good success, and others feel that it is necessary to have a –1.5 °C (29.3 °F) dewpoint to get proper results. In certain cases, these lower dewpoints are required, particularly if there is a higher chromium content in the low-alloy steel. One last point on dewpoint control concerns a 300-series stainless steel being brazed with a brazing atmosphere of nitrogen with 20% hydrogen. When dewpoint readings were taken, a different dewpoint was found when the flow of atmosphere into the furnace was increased. Therefore, the problem was, “Why did the dewpoint change with different flow rates of atmosphere through the piping before it even reached the furnace?” High-purity gases with low dewpoints—as would be coming from a liquid nitrogen tank and a hydrogen trailer, which originally came from a liquid system—work much differently than one would normally expect gases to behave. The change in dewpoint with a change in flow rate through the piping system is a classic “footprint,” which is stating that there is a leak in the system. In directly stating the problem, any leakout of a system of high-purity gases is also a leak into the system. Because the potential across the leak is essentially with the same dry gas on the inside and the same humid, oxygen-containing air on the outside of the leak, the diffusion potential is the same all the time. As the flow is changed in the piping system, more gas is passing through the pipe, with the same amount of oxygen and moisture coming through the leak. Thus, the amount of oxygen and moisture per cubic foot of gas flowing through the pipe is diluted. This is recognized at the dewpoint sensor as a lower dewpoint. Thus, when a given dewpoint is at a specific flow rate and the flow rate is increased, the dewpoint would normally get better.
Noting this change in dewpoint with flow rate allows one to have a good indicator of leaks in the system, by just measuring the dewpoint at two widely different flow rates. If there is no difference in dewpoint, then the system does not have any leakage. However, if there is a difference in the dewpoint, this is an indicator that there are leaks in the system, and maintenance of the piping system is required. The dewpoint instrument can be a useful tool in detecting leaks in the dry gas supply system. Practice has shown that certain metal oxides can be reduced, to an extent, in a high-purity inert gas environment due to the low partial pressure of oxygen in the atmosphere. A continuously pumped vacuum roughly equivalent to the partial pressure of water vapor, as presented on the right-hand ordinate of Fig. 6.5, also gives similar results in practice. Because the equilibrium diagram is presented only for H2/H2O atmospheres, all the oxygen (O2) in the hydrogen atmospheres must be converted to H2O before the dew-point is determined. Such curves as those shown in Fig. 6.5 and others that have been published (Ref 18, 19) aid in comprehending the actions of hydrogen and water vapor in respectively reducing and oxidizing metal oxides and metals, but they do not portray the complete story involved in the use of controlled atmospheres. They indicate neither the rate at which reduction will occur nor the physical form of the oxide. Oxides of aluminum, titanium, beryllium, and magnesium cannot be reduced by hydrogen at ordinary brazing temperatures. If these elements are present in small amounts, satisfactory brazing can be done in gas atmospheres. When these elements are present in quantities exceeding 1 or 2%, the metal surface should be plated with a pure metal that is easily cleaned by hydrogen, or a flux should be used in addition to the hydrogen. Accurate control of dewpoint cannot be overemphasized in gas atmosphere brazing if sound and completely bonded joints are to be produced in metals that form oxides of high stability (Ref 20).
Atmosphere Components (Gases) The components of brazing atmospheres have individual characteristics that affect their suitability for brazing various metals and alloys. Carbon monoxide (CO) is an active agent for the reduction of some metal oxides (e.g.,
Chapter 6: Fluxes and Atmospheres / 261
those of iron, nickel, cobalt, and copper) at elevated temperatures. Carbon monoxide can serve as a source of carbon, which may be desirable in brazing some carbon steels but is undesirable in other applications. When decomposed, it may release oxygen, which is undesirable in many controlled atmospheres for brazing. Carbon monoxide can be generated from oil on the parts at brazing temperatures. Carbon monoxide is toxic, and adequate ventilation must be provided unless waste gas is trapped and burned. Carbon dioxide (CO2) is neutral to most metals and is an inert constituent of some brazing atmospheres, except when it is decomposed to carbon monoxide or carbon and oxygen, all of which are reactive with metals. It will oxidize iron, however, and some alloying elements, such as chromium, manganese, and vanadium. At high temperatures, CO is more stable; at low temperatures, CO2 forms preferentially. Its presence may be undesirable as a source of oxygen, carbon, and carbon monoxide when decomposed. In CO-CO2 atmospheres, the carbon dioxide content of a furnace atmosphere can be undesirably increased by air leakage. In applications such as the brazing of carbon steels, it must be removed from the atmosphere to avoid oxidation and decarburization of the metal surfaces. Methane (CH4) may come from the atmosphere gas as generated from organic materials left on the part by inadequate cleaning. It can serve as a source of carbon and hydrogen. Methane is sometimes added to certain atmospheres to balance decarburizing gases present. Oxygen (O2). Free oxygen in the brazing atmosphere is always undesirable. In addition to the sources already mentioned, oxygen may come from gases adsorbed on surfaces in the heating chamber. Nitrogen (N2) is used in a controlled atmosphere to displace air from the furnace and to act as a carrier gas for the other atmosphere constituents. The typical high purity of nitrogen allows low levels of reducing gases to be used. A nitrogen atmosphere is applicable whenever exothermic gas or dissociated ammonia is used as the reducing agent. Nitrogen is inert to most metals, but high levels of nitrogen should be used cautiously when working with metals that are susceptible to nitriding, such as chromium and molybdenum (Ref 21). Proper use of nitride-inhibiting atmosphere constituents can minimize nitrogen pickup where it is a concern. Nitrogen is noncombustible and nonexplosive
and therefore is desirable from a safety standpoint (Ref 22). There are several other advantages of nitrogen-base atmosphere. Cryogenic nitrogen has a very low dewpoint and is a very dry gas, so, when hydrogen (from exothermic reaction or from dissociated ammonia) is added, the resulting H2:H2O ratio is relatively high, which makes for a high reducing capacity, or good fluxing. In fact, a nitrogenbase atmosphere usually permits the required amount of hydrogen to be reduced to below the explosive level of the mixture. Another advantage of nitrogen-base atmospheres stems from the elimination of chemical fluxes when they are used primarily to reduce oxides. The use of fluxes requires larger joint clearances to allow flux to escape and be displaced by the filler metal. This may produce a weaker joint. A particular advantage of a nitrogen-base atmosphere is that it can be tailored to provide just the right level of reduction, depending on the material being processed or the stage within the brazing cycle. For example, it may be desirable to have a slightly oxidizing atmosphere in the preheating section of a furnace to help burn off organic compounds used in paste filler metals. As a result of the previously mentioned aspects, provision for adjustments in furnaceatmosphere composition can be made by introducing different compositions at different points in the cycle or, in the case of continuous furnaces, in different zones. Finally, it should be noted that a nitrogenbase atmosphere containing methanol has been developed that results in a very dry atmosphere and subsequently, in better wetting by filler metals (Ref 23). A firm was brazing fuel manifold assemblies and encountered inconsistent results. Some parts brazed well, while in others, the filler metal balled up, would not wet the surface, and would fall off. The materials being joined were American Iron and Steel Institute (AISI) 347 and N155 stainless steels in a vacuum of approximately 1 µm at 1010 °C (1850 °F). The filler metal, 82Au-18Ni, successfully wetted the 347 stainless steel but did not wet the N155. This result tended to indicate that there was no problem in satisfactorily wetting the 347 stainless steel; however, the footprint left by the N155 gave a clue relative to where to attack the problem. The main difficulty in using N155 is the nitrogen content (0.10 to 0.12%) in the steel.
262 / Brazing, Second Edition
This is a lot of nitrogen; depending on the pumping system and atmosphere quantity, a low nitrogen content may or may not braze adequately. Usually a high nitrogen content of 0.20% could be expected to cause problems. Because this part was brazed in vacuum furnace, some outgassing of the nitrogen will occur; however, enough nitrogen remains at the surface to prevent wetting and flow of the filler metal. One difficulty in brazing this type of material is the consistently varying results, depending on the number of pieces in the furnace, quality of the atmosphere, heating rate, and nitrogen content in the base metal. When production quantities were put into the furnace, they would not braze at all. This inconsistency occurs when the production quantity dumps out a lot of nitrogen into the furnace and causes a nitrogen buildup on the surface of the N155, whereas, in the case of a single part in the furnace, assuming that the atmosphere and other variables are good, the nitrogen layer will be sufficiently removed from the surface, and the part can achieve adequate brazing. If the vacuum furnace is set up with a nitrogen backfill and there happens to be a small leak, there may be a good enough vacuum for normal brazing. However, with extra nitrogen in the atmosphere, this could cause some variation in the existing problem. If the nitrogen-containing base metal has a fairly high content of nitrogen on the surface, it will appear to have an iridescent bluish-gray color. Under these circumstances, one should not expect adequate wetting and flow on the surface of this base metal. In some base metals with low nitrogen content, a pretreatment, consisting of a bakeout of the detail parts in a separate vacuum furnace load, at or above the brazing temperature can outgas sufficient amounts of nitrogen to allow adequate brazing. Unfortunately, on some base metals and furnace equipment, the precleaning bakeout cycle has not proven adequate to remove enough nitrogen to allow suitable brazing. Experiments to ensure that the proper thickness of nickel plating is used with various nitrogen contents in the base metal should be conducted. If there is nitrogen in the atmosphere from leakage or backfilling for partial-pressure brazing, the thickness of electrolytic nickel may have to be greater. A good starting point would be with a 0.01 mm (0.0004 in.) thickness of the electrolytic nickel plating. Finally, it is also good practice to run a proto-
type set of parts to determine if adequate vacuum coverage and filler-metal selection are proper and suitable. Inorganic Vapors. In equipment designed for their use, vapors such as those of zinc, cadmium, lithium, and fluorine compounds can serve to reduce metal oxides and scavenge the atmosphere of oxygen. They are useful for replacement of constituents of alloys evolved during brazing. Such vapors are toxic, and proper safety precautions should be used (Ref 24). Hydrogen (H2). Reducing atmospheres not only prevent the formation of surface oxides on the base metal at the brazing temperature but also reduce residual surface oxides and the oxides that form during the low-temperature stages of the heating process. Although some reducing atmospheres can be used to braze metals whose oxides are easily reduced, hydrogen and atmospheres containing large amounts of hydrogen (e.g., cracked ammonia) are most suitable for high-temperature brazing. Hydrogen is one of the most active agents for reducing the oxides of many metals during brazing. If an oxidized metal is heated to a sufficiently high temperature in a dry hydrogen atmosphere, the oxide will be reduced, and water vapor will form. Oxide reduction continues until the amount of water vapor increases to the point where the ratio of H2O to H2 reaches equilibrium for the metal oxide at that particular temperature; further oxide reduction will not occur unless the moist hydrogen is replaced by dry hydrogen. Although all metal oxides can be reduced, some are more difficult to reduce than others; for example, hydrogen with a much lower dewpoint at a given temperature is required before oxide reduction can occur. Figure 6.5 is the most useful diagram in determining the condition in pure dry hydrogen (temperature and hydrogen dewpoint) under which a particular metal oxide will be reduced. Heat-resistant base metals that contain appreciable amounts of aluminum and/or titanium are difficult or impossible to braze in a hydrogen atmosphere, because the oxides of these metals cannot be reduced at the temperatures used for brazing. From Fig. 6.5 it can be seen that a hydrogen dewpoint of –90 °C (–130 °F) or lower is required to reduce titanium oxide at 1370 °C (2500 °F); even lower dewpoints are required for reduction of aluminum oxide. Although hydrogen dewpoints below –73 °C (–99 °F) can be obtained under laboratory conditions, they are difficult to obtain and maintain
Chapter 6: Fluxes and Atmospheres / 263
in production. However, several approaches to the problem of brazing metals alloyed with titanium and/or aluminum can be considered: • The joint can be brazed in a vacuum. • The surfaces of the joint members can be plated with nickel, copper, or another metal that does not form a refractory oxide and is easily wetted by the filler metal. • The joint members can be oxidized in a wet hydrogen atmosphere, and then the oxides of titanium or aluminum can be leached from the joint surfaces in a nitric acid/hydrofluoric acid solution. Surfaces relatively free from titanium and aluminum oxides can be obtained in this manner. Long brazing cycles should be avoided, because aluminum and titanium may diffuse to the surface and reoxidize. • A high-temperature flux can be used to prevent the formation of oxides and promote wetting. Of the techniques described previously, plating is the most straightforward method for production brazing of heat-resistant alloys containing aluminum and/or titanium. Several grades of hydrogen are available for brazing, and the grade that is appropriate to the application should be selected. Dissociated ammonia has high hydrogen content and is, therefore, a very reducing atmosphere. It is used mostly in brazing of stainless steels or mild nickel alloys. Boron-containing filler metals are not suitable for use in N2 containing furnace atmospheres. Boron combines with N2 to produce a black boron nitride. Depending on the percent nitrogen as well as the time and temperature, there may be only a small reaction or a reaction sufficient to put a heavy, black layer on the top of the filler metal and to prevent all filler metal from flowing. Taking the work to 760 °C (1400 °F) with N2 as an atmosphere certainly would not be recommended for boron-containing filler metals. Purging at room temperature may also be objectionable when the flow rate of the atmosphere is not sufficient to remove all of the N2 by 538 °C (1000 °F) or before (preferably 204 °C, or 400 °F). The B-N2 reaction, like so many others, is a time-rate reaction. Experience has shown the following reactions: • Dissociated NH3 (25%N2-75%H2) is suitable for a fast induction brazing operation yet is not suitable for furnace brazing with a heat-
ing rate of 15 min from room temperature (RT) to brazing temperature. • 0.4% N2 in pure dry H2 is suitable for furnace brazing with a heating rate of 15 min from RT to brazing temperature but is not suitable for furnace brazing with a heating rate of 60 min from RT to brazing temperature. • 0.2% N2 in pure dry H2 is suitable for furnace brazing with a heating rate of 60 min from RT to brazing temperature and is not suitable for furnace brazing with a heating rate of 7 h from RT to brazing temperature. • 0.01% N2 in H2 was needed for furnace brazing with a heating rate of 7 h from RT to brazing temperature yet is not suitable with a heating rate of 7 days from RT to brazing temperature. The previous observed examples are based on filler metals containing from 2 to 3.5% B, and it is not known how this reaction decreases with decreasing B content. When nickel brazing in N2-base atmospheres, filler metals containing silicon or phosphorus will be suitable as long as the surface is not nitrided or a light blue-gray N2 film is not present. This film will occur from N2-containing base metals or base metals heat treated in a N2base atmosphere where interstitial N2 is picked up in the base metal. Boron-containing nickel filler metals should be used in pure dry H2, Ar-H2 or vacuum atmospheres. Water vapor is objectionable because its presence can promote oxidation or cause decarburization. Water vapor may be added intentionally by controlled humidification or unintentionally due to air leakage, air carried into the furnace with the work, reduction of metal oxides, leakage from water jackets, contaminated gas lines, diffusion of oxygen through inadequate flame curtains, and other less obvious sources. However, a carefully controlled amount of water vapor will aid in cleaning of carbonaceous material from brazed surfaces and in removal of binder left behind by filler metals. In addition, water vapor can be used to inhibit fillermetal flow where filler-metal containment is desirable. This is particularly beneficial in brazing of wide-gap joints. The amount of water vapor required in the latter instance depends on the amount of hydrogen that is present in the atmosphere. The reducing ability of a hydrogen-base atmosphere depends primarily on the H2:H2O
264 / Brazing, Second Edition
ratio, which must be higher than 10 to 1 if the atmosphere is to be effective. The amount of water in an atmosphere is specified by the dewpoint, the temperature at which moisture in the gas will condense. The relationship between dewpoint temperature and moisture content of gases is shown in Table 6.5. Argon (Ar) and Helium (He). Inert gas atmospheres can be used to braze most metals. However, they are most useful in brazing base metals whose properties are adversely affected by exposure to hydrogen. For example, alloys of titanium, zirconium, niobium, and tantalum are extremely sensitive to the presence of minute quantities of hydrogen and become embrittled. Such metals can be brazed satisfactorily in a controlled atmosphere of argon or helium; they can be brazed equally well in a vacuum. The parts to be brazed must be cleaned and handled carefully, because the primary purpose of the inert gas atmosphere is to prevent the formation of oxides during brazing. Inert gases such as helium and argon form no compounds with metals. In equipment designed for their use, they inhibit evaporation of volatile components during brazing and, when used at ambient or partial pressure, reduce the evaporation rate of volatile elements relative to vacuum. Although both argon and helium can be used for controlled atmosphere brazing, argon is most frequently used. Commercial argon is available at a guaranteed as-delivered purity of 99.996%, or no more than 40 ppm total impurities. Assuming that water vapor is the only impurity, this impurity level corresponds to a dewpoint of approximately –50 °C (–60 °F). The actual dewpoint may be considerably lower, because impurities other than water vapor may be present in the gas. As a result, argon can often be used as a protective atmosphere without further purification.
Table 6.5 Relationship between dewpoint temperature and moisture content of furnace atmospheres Dewpoint temperature °C
–18 –34 –51 –62 –73 Source: Ref 2
Moisture content
°F
vol%
ppm
0 –30 –60 –80 –100
0.150 0.0329 0.0055 0.0014 0.0002
1500 329 55 14 2
Vacuum. Vacuum brazing has made great strides in the past decade, and there is every reason to believe that the usefulness of this process will continue to grow in the future. The growth of vacuum brazing is due in part to improvements in equipment design and performance. Vacuum brazing is particularly well suited for joining: • Heat-resistant nickel- and iron-base alloys that contain aluminum and/or titanium • Reactive metals • Refractory metals • Ceramics Vacuum conditions are especially well suited for brazing very large, continuous areas where solid or liquid fluxes cannot be removed adequately from the interfaces during brazing, and where gaseous atmospheres are not completely efficient because of their inability to purge occluded gases evolved at close-fitting brazing interfaces. Vacuum is also suitable for brazing many similar and dissimilar base metals, including titanium, zirconium, niobium, molybdenum, and tantalum. The characteristics of these metals are such that even very small quantities of atmospheric gases may result in embrittlement and sometimes disintegration at brazing temperatures. These metals and their alloys may also be brazed in inert gas atmospheres if the gases are of sufficiently high purity to avoid contamination and the resultant loss in properties of the metals. Compared with other types of brazing, vacuum brazing has advantages and disadvantages, described as follows. Vacuum removes essentially all gases from the brazing area, thereby eliminating the necessity of purifying a supplied atmosphere. During brazing, the pressure within the furnace is maintained at a level such that oxidation of the workpieces does not occur. Commercial vacuum brazing facilities operate in the range from 0.013 to 0.00013 Pa; the gas impurity level corresponding to a pressure of 0.013 Pa (1.89 × 10–6 psi) is approximately 0.3 ppm. The actual pressures used depend on the materials being brazed, the filler metals being used, the area of the brazing interfaces, and the degree to which gases are expelled from the base metals during the brazing cycle. Certain oxides of base metals will dissociate in vacuum at brazing temperatures. Vacuum is
Chapter 6: Fluxes and Atmospheres / 265
used widely to braze stainless steel (Ref 25, 26), superalloys, aluminum alloys, and refractory materials by special techniques. The mechanism of oxide removal in a vacuum is not clearly understood. Oxide films can be removed by evaporation, dissociation, diffusion, or a combination of diffusion and chemical reaction. The low pressure existing around the base and filler metals at elevated temperature removes volatile impurities and gases from the metals. Frequently, the properties of the base metals themselves are improved. This characteristic is nevertheless a disadvantage where elements of the filler metal or base metals volatilize at brazing temperatures because of the low surrounding pressure. This tendency can be corrected by adherence to proper vacuum brazing techniques. Many vacuum furnaces have the ability to operate under a partial pressure of inert gas. There are then two general types of vacuum brazing: brazing in a high vacuum and brazing in a partial vacuum. High vacuum is particularly well suited for brazing of base metals containing hard-to-dissociate oxides (Ref 20). Partial vacuums are used where the base metal or filler metal, or both, volatilize at brazing temperatures under high-vacuum conditions. The lowest pressure at which the metals will remain in the solid or liquid phase at the brazing temperature is determined by calculation or experimentation. The brazing chamber is evacuated to high-vacuum conditions. The heating cycle proceeds under high vacuum until just below the temperature where vaporization would begin. High-purity argon, helium, or, in some instances, hydrogen is gradually introduced in sufficient amounts to overcome the vapor pressure of the volatile metals at brazing temperature. This technique appreciably widens the range of materials for which vacuum brazing is effective. Vacuum purging prior to high-purity dry hydrogen brazing is frequently employed where extra precautions must be taken to ensure optimal freedom from foreign or contaminating gases. Similarly, dry hydrogen or inert gas purging prior to evacuation is sometimes helpful in obtaining improved brazing results in a highvacuum atmosphere. Zirconium, titanium, and other elements with high affinities for oxygen and other gases are sometimes strategically placed close to, but not in contact with, the part being brazed in a highvacuum atmosphere. These so-called getters
rapidly absorb very small quantities of oxygen, nitrogen, and other occluded gases that may be evolved from the metals being brazed and thus improve the quality of the brazing atmosphere. Another method of reducing contamination in vacuum is by another gettering technique. Lithium, magnesium, sodium, potassium, calcium, titanium, and barium can all be vaporized in the chamber to reduce the volume of oxides and nitrides present in the vacuum atmosphere. These materials may condense on the chamber walls; therefore, care must be taken. Their disadvantage is that most of them will either react with the workload or form a coating on the wall when exposed to atmospheric moisture. Successful brazing in vacuum depends on the presence of a promoter—either a metal or a reactive gas. The key action of metal promoters in vacuum brazing is to chemically reduce the oxide films to permit wetting by the filler metal. In addition, they must also scavenge remaining oxygen and moisture in the vacuum, but these are not the key mechanisms. Many metals can fulfill the function of a braze promoter, but magnesium is the best. Magnesium contained in the filler metal does double duty. In the case of aluminum, as it vaporizes, magnesium tends to disrupt the aluminum oxide at the aluminum boundary layer, and later, as a vapor in the vacuum chamber, it reacts with oxygen and oxides (mainly water) to reduce or eliminate the formation of additional aluminum oxide surface films that could inhibit good wetting and capillary flow. Table 6.6 shows the metal activators and their actions in vacuum brazing.
Atmospheres, Base Metals, and Filler Metals The protective atmospheres most commonly used in furnace brazing with silver alloy filler metals are rich exothermic gas, endothermic gas, dissociated ammonia, dry hydrogen, and commercial nitrogen-base atmosphere blends, principally with hydrogen. Even when a flux is used, an atmosphere is usually employed to minimize or prevent oxidation and discoloration of the base metals and to ensure that the flux performs its functions. Exothermic and endothermic atmospheres are less expensive than dissociated ammonia or dry hydrogen; they are used in furnace brazing of steel-to-steel or to oxygen-free copper or copper alloys, using a flux and BAg-1a or BAg5 filler metal.
266 / Brazing, Second Edition
Provided that dissociation is complete (100%), dissociated ammonia can be used in nearly all furnace-brazing applications involving the use of silver alloy filler metals.
Fluxes The primary purpose of brazing fluxes is to promote wetting of the base metal by the filler metal. The efficiency of flux activity, which is commonly referred to as wetting, can be expressed as a function of brazeability (Ref 27). Flux must be capable of dissolving any oxide remaining on the base metal after it has been cleaned and any oxide films on the liquid filler metal. It is important to realize that most fluxes are not designed or intended for the primary removal of grease, oil, or dirt and cannot take the place of proper precleaning operations. However, in some instances, fluxes may serve to suppress the volatization of high-vapor-pressure constituents in a filler metal. Some filler metals, when in the molten stage, are self-fluxing on certain alloys. To effectively protect the surfaces to be brazed, the flux must be applied as an even coating and must completely cover and protect the surfaces until the brazing temperature is reached. It must remain active throughout the brazing cycle. Because the molten filler metal should displace the flux from the joint at the brazing temperature, the viscosity and surface tension of the flux and the interfacial energy between the flux and the surfaces of parts are important. Therefore, recommended fluxes should be used in their proper temperature ranges and on the materials for which they are designed. Additionally, the role of flux joining is one of controlled corrosion. The corrosive attack centers on the dissolution and dispersion of oxide
tarnish, but the surface layer of metal atoms may also be removed. This attack is rapid because of the elevated operating temperatures and because the oxide capacity of the flux is considerable. The interaction of the flux melt with the oxide layer is central to obtaining a good, clean, fast joint (Ref 28). Certain filler metals contain alloy additions of deoxidizers, such as phosphorus, lithium, and other elements that have strong affinities for oxygen. For example, phosphorus-copper filler metals act as fluxes on copper and silver. In some instances, these additions make such filler metals self-fluxing without the application of prepared fluxes or controlled atmospheres. These filler metals are self-fluxing only in the molten state and will themselves oxidize during the heating cycle. In other cases, they are used in conjunction with protective atmospheres or fluxes to increase wetting tendencies. When large sections are to be brazed or where prolonged heating times are contemplated, the use of additional flux is advisable.
Flux Constituents Many chemical compounds (Ref 2) are used in the preparation of fluxes, and many proprietary fluxes on the market are formulated to offer specific properties. When fluxes are heated, reactions take place between the various chemical ingredients, forming new compounds that are quite different chemically and physically from the unreacted constituents. For instance, if a fluoborate is an ingredient in a flux, fluorides may be formed as the ingredients react. During brazing, the chemistry is especially transient. Reaction rates of the flux with oxygen, base metals, filler metals, and any foreign materials present increase with temperature. Composition of the flux must be carefully tailored to suit all the factors of the brazing cycle, including dwell time. Attack of the flux on the metals must be limited,
Table 6.6 Possible promoters (activating metals) in vacuum brazing Activating metal
Rare earths, beryllium, scandium, yttrium Magnesium, calcium, strontium, lithium Barium, sodium, zinc Antimony, bismuth Source: Ref 2
Removes oxygen
Removes water vapor
Vaporizers
Reacts with aluminum oxide
Promotes vacuum brazing
Yes Yes Yes Yes
Yes Yes Yes No
No Yes Yes Yes
Yes Yes No No
Yes Yes No No
Chapter 6: Fluxes and Atmospheres / 267
because the flux must react promptly with metal oxides or other tarnish to enable the joint to be satisfactorily formed. Active halides, such as chlorides and fluorides, are, for instance, necessary in fluxes for alloys containing aluminum or other highly electropositive metals. The most common ingredients of chemical fluxes are: Borates (sodium, potassium, lithium, etc.) Fused borax Elemental boron Fluoborates (potassium, sodium, etc.) Fluorides (sodium, potassium, lithium, etc.) Chlorides (sodium, potassium, lithium) Acids (boric, calcined boric) Alkalis (potassium hydroxide, sodium hydroxide) • Wetting agents • Water (either as water of hydration or as an addition for paste fluxes) • • • • • • • •
Most brazing fluxes are proprietary mixtures of several of the previously mentioned ingredients. The ingredients are mixed and reacted in ways that give satisfactory results for specific purposes. Their functions are described as follows. Borates are useful in formulating the fluxes that melt at higher temperatures. They have good oxide-dissolving power and provide protection against oxidation for long periods. Most borates melt and are effective at temperatures of approximately 760 °C (1400 °F) or higher. They have a relatively high viscosity in their molten condition and therefore must be mixed with other salts to increase fluidity. Fused borax is another high-temperature melting material that is active at high temperatures. It is little used in lower-temperature brazing processes. Elemental boron powder is added to increase overall fluxing action. Silver brazing fluxes that contain elemental boron offer improved protection on carbides and on materials that form refractory oxides, such as chromium, nickel, and cobalt. Fluoborates react similarly to other borates in many respects. Although they do not provide protection from oxidation to the same extent as other borates, they flow better in the molten state and have greater oxide-dissolving properties. Fluoborates are used with other borates or with alkaline compounds, such as carbonates.
Fluosilicaborates. Another class of compound is fluosilicaborates, which have somewhat higher melting points than fluoborates and provide good coverage and surface adherence. Their high melting points limit their use. Fluorides react readily with most metallic oxides at elevated temperatures and therefore are used extensively in fluxes as cleaning agents. They are particularly useful when refractory oxides, such as those of chromium and aluminum, are encountered. Fluorides are often added to increase the fluidity of molten borates, thereby facilitating their displacement and improving the capillary flow of the molten filler metal. Fluorides can generate dangerous fumes, however, and so their use warrants strict attention to good safety practices. Fluorides, up to 40% in flux content, give silver brazing fluxes their characteristically low melting points (560 °C, or 1040 °F) and high activity for dissolving metal oxides. Chlorides function in a manner similar to fluorides but have a lower effective temperature range. Chlorides must be used with caution because, at lower temperatures, they are used to depress the melting points of fluoride-base fluxes. As seen in Table 6.7, aluminum and magnesium brazing fluxes contain alkaline chlorides or fluorides. Lithium salts give these fluxes low melting points (540 to 615 °C, or 1000 to 1140 °F) and high chemical activity, enabling the fluxes to dissolve stubborn aluminum oxide. Boric acid is a principal constituent used in brazing fluxes because it facilitates the removal of the glasslike flux residue left after brazing. Its melting point is below that of borates but higher than that of fluorides. Silver brazing fluxes contain boric acid and potassium borates, combined with complex potassium fluoborate and fluoride compounds. High-temperature fluxes, based on boric acid and alkaline borates, sometimes contain small additions of elemental boron or silicon dioxide to increase activity and protection, good up to 1204 °C (2200 °F). Fluoride content of these fluxes is usually low, at most 2 to 3%. These braze ferrous and high-temperature alloys and carbides. Alkalis, such as potassium and sodium hydroxides, are used sparingly, if at all, to elevate the useful working temperature of the flux. Their drawback is that they are deliquescent; even small amounts in other flux agents can
Aluminum brazing
Aluminum brazing Aluminum bronze
FB1B
FB1C
FB3A
FB2A
FB3K
Silver brazing
High-temperature brazing High-temperature brazing High-temperature brazing Magnesium brazing
FB3I
FB3J
High-temperature brazing
FB3D
FB4A
Aluminum brazing
Flux category
FB1A
AWS specification
Paste
Flammable liquid Powder
Powder(e)
Slurry(d)
Paste(d)
Paste
Powder
Powder
Powder
Form
Copper, ferrous and nickel alloys, carbides
Copper, ferrous and nickel alloys, carbides Copper, ferrous and nickel alloys, carbides Copper, ferrous and nickel alloys, carbides Magnesium alloys
Brazeable base metals containing aluminum (aluminum brass, aluminum bronze, Monel K500); may also have application when minor amounts of titanium or other metals are present that form refractory oxides Copper, ferrous and nickel alloys, carbides
Aluminum alloys
Aluminum alloys
Aluminum alloys
Base materials
Table 6.7 Characteristics of brazing fluxes
Manual
Dip brazing
Manual
Manual
Application method
BAg, BCuP(c)
Heat source
Torch, induction
Salt bath
Torch
Torch, furnace
Torch
Torch, furnace, induction
Torch, furnace, induction
Salt bath
Furnace
Torch, furnace
(continued)
Manual, automatic
BAg, BCu, BNi, Automatic BAu, RBCuZn BAg, BCu, BNi, Manual BAu, RBCuZn BAg, RBCuZn Manual, automatic BMg Dip brazing
BAg, BCu, BNi, Manual, BAu, RBCuZn automatic
BAg,BCuP(c)
BAlSi
BAlSi
BAlSi
Filler metals
Typical ingredients (b)
4,5,6
6,7
...
...
...
4,5,6
Borates, fluorides, boron
Chlorides, fluorides
...
...
...
Borates, fluorides
4,5,6,7 Chlorides, fluorides 4,5,6 Chlorides, fluorides, borates
4,5,6,7 Chlorides, fluorides
4,5,6,7 Chlorides, fluorides
Joining methods (a)
565–870
480–620
760–1205
760–1205
760–1205
760–1205
595–870
540–615
560–615
580–615
°C
Notes
... 1050–1600 Water, 35% max; paste may be thinned with water if desired; usually applied by brushing or dipping the work into the flux.
900–1150
1400–2200 . . .
1400–2200 . . .
1400–2200 . . .
1400–2200 . . .
1075–1140 For torch or furnace brazing; water or alcohol may be added as the flux is used. 1040–1140 For furnace brazing; water or alcohol may be added as the flux is used. 1040–1140 For chemical bath dip brazing 1105–1600 Water, 35% max; paste may be thinned with water if desired; usually applied by brushing
°F
Active temperature range
General-purpose paste flux for most ferrous and nonferrous alloys
...
...
...
Paste flux used for controlled atmosphere furnace brazing ...
Powder flux for dip brazing Paste flux
Powder flux for furnace brazing
Powder flux for torch or furnace brazing
Flux description
Silver brazing
Silver brazing
Silver
Silver brazing
B3E
B3F
B3G
B3H
Slurry brazing Slurry(e)
Powder
Waterbased liquid
Paste(e)
Form
Copper, ferrous and nickel alloys Copper, ferrous and nickel alloys
Copper, ferrous and nickel alloys
Copper, ferrous and nickel alloys, carbides
Copper, ferrous and nickel alloys, carbides
Base materials
BAg
BAg, ferrous
BAg, BCuP(c)
BAg, BCuP(c)
BAg, BCuP(e)
Filler metals
Automatic, BCuP(c) Automatic Torch
Manual
Manual, automatic
Manual, automatic
Application method
...
Torch
Torch, furnace
Torch, furnace
Torch, induction
Heat source
...
...
4,5,6
...
...
Joining methods (a)
565–925
...
Borates, fluorides
...
...
Typical ingredients (b)
1050–1700
565–870
650–870
565–870
565–925
°C
Notes
1050–1700 Water, 35% max; paste may be thinned with water if desired; usually applied by brushing or dipping the work into the flux. 1050–1600 All brazeable ferrous and non-ferrous metals except those containing aluminum or magnesium 1200–1600 All brazeable ferrous and nonferrous metals except those containing aluminum or magnesium 1050–1600
°F
Active temperature range
Paste flux
Paste flux
General-purpose paste flux for most ferrous and nonferrous alloys
Flux description
AWS, American Welding Society. (a) 4, Apply dry powder to joint; 5, Dip heated filler-metal rod in powder or paste; 6, Mix flux to paste consistency with water, alcohol, or other carrier; 7, Dip or immerse in molten flux bath. (b) Fluxes in form of aqueous paste may also contain wetting agents. (c) Used with copper and copper alloy base metals only. (d) May contain elemental boron or silicon dioxide. (e) Boron-modified. Note: Pastes have high viscosities and are typically applied by brushing. Slurries have low viscosities and can be sprayed or automatically dispensed. Source: Ref 2
Silver brazing
Flux category
FB3C
AWS specification
Table 6.7 (continued)
270 / Brazing, Second Edition
cause problems in humid weather and can severely limit the storage life of the flux. Alkalis elevate the useful working temperature of the flux. Wetting agents are used in paste and liquid fluxes to facilitate the flow and spreading of the flux onto the workpiece prior to brazing. Water is present in brazing fluxes either as water of hydration in the chemicals used in formulating the flux or as a separate addition for making a paste or liquid. Water used in forming a paste must be evaluated for suitability, and hard waters should be avoided.
Groups of Fluxes There is no single flux that is best for all brazing applications. Fluxes are classified (Table 6.7) according to their performance on certain groups of base metals in rather specific temperature ranges. The five categories of brazing fluxes are aluminum, aluminum-bronze, silver, magnesium, and high-temperature flux. Within each type and class, there are numerous commercial and proprietary fluxes available, and selection of an appropriate flux must be done by careful analysis of the properties or features required for a particular application. Reference to Table 6.7 is not a substitute for thorough evaluation in selecting an optimal flux for a specific high-production joint. For successful use, a flux must be chemically compatible with all the base metals and filler metals involved in the brazement. It must be active across the entire brazing temperature range and throughout the time at brazing temperature. If the brazing cycle is long, a less active but more protective flux should be selected. Conversely, if the cycle is short, a more active flux, which will promote quick filler-metal flow at the minimum temperature, may be used. Where more than one flux is suitable for the application, other considerations, such as safety and cost, should be evaluated (Ref 2).
Flux Selection Criteria Base-material type determines flux selection more than any other factor. To braze aluminum alloys, coat parts with aluminum brazing fluxes. Similarly, aluminum-bronze and magnesium fluxes braze only with their respective base metals. To braze ferrous alloys and nickel alloys, two flux types can be used: silver brazing or high-temperature fluxes. Which of the two is
better depends on base- and filler-metal type, brazing conditions, and cost. Fabricators call on silver brazing fluxes, more expensive than hightemperature fluxes, to minimize heat input and distortion to the work. These also braze copper alloys. To braze carbides—for example, tungsten carbide infiltrated with cobalt to impart high strength with toughness—coat with boronmodified fluxes and fill the joint with silver filler metals containing nickel. High-temperature fluxes and filler metal also braze carbides, when the carbide-steel combination can tolerate the high brazing temperatures (1093 °C, or 2000 °F). Within a particular flux type, there are several criteria for choosing a specific flux for maximum efficiency: • For dip brazing, water (including water of hydration) must be removed, usually by preheating prior to immersion in the salt bath. • For resistance brazing, the flux must permit the passage of current. This usually requires a wet, dilute flux. • The effective temperature range of the flux must include the brazing temperature for the specific filler metal being used. • Controlled atmospheres may modify flux requirements. • Ease of flux residue removal should be considered. • Corrosive action on the base metal or filler metal should be minimized. Flux/Temperature Range Specification. To be effective, flux must be molten and active before the filler metal melts, and it must remain active until the filler metal flows through the joint and solidifies on cooling. Therefore, filler-metal solidus determines minimum working temperature of the flux, and filler-metal liquidus dictates maximum brazing temperature that the flux must withstand. Generally, select a flux that is active approximately 30 °C (54 °F) below the solidus of the filler metal and that remains active at least 90 °C (162 °F) above the filler-metal liquidus. If overheating is likely to occur during brazing, as when torch brazing, select a flux active at 120 to 175 °C (216 to 315 °F) above the fillermetal liquidus. This gives the flux a wide temperature range to remove surface oxides before the filler metal melts and will keep it effective at brazing temperatures.
Chapter 6: Fluxes and Atmospheres / 271
Brazing time affects flux performance. Molten flux forms a semiprotective blanket that prevents oxidation only for a finite period— oxygen will eventually diffuse through the flux to the base materials. Flux must continue to remove newly formed oxide until the end of the heating cycle. Because flux can dissolve only a limited amount of oxide, the longer the heating cycle, the greater the likelihood that the flux will become saturated with oxide; this condition is called flux exhaustion. Rated temperature range of a flux, which depends on brazing temperature, flux type and volume, and base-material type, assumes a brazing cycle of 15 to 20 s. With a longer heating cycle, flux exhaustion may occur even when brazing below the maximum operating temperature, because, over time, the flux becomes saturated with metal oxide. To avoid flux exhaustion over prolonged heating cycles, switch to a flux with a higher working-temperature range. When the heating cycle is short, a fabricator can braze with a flux above its maximum rated working temperature. Using a low-temperature flux above the maximum working temperature eases flux removal, because these fluxes are more soluble in water than are high-temperature fluxes.
Flux Application Ideally, flux is applied to both joint surfaces; for some applications, coating only one surface suffices—the flux will transfer to the mating surface on assembly. Application methods depend on joint design, production volume, and joint-heating technique. Operators brush to apply paste flux to the joint and to surrounding surfaces, or they may dip parts into a container of flux. Flux for dipping is of a thinner consistency than that used for brushing. In some cases, parts are dipped in boiling flux solutions in which the solids are completely dissolved. Automatic application of flux can be carried out by spraying, pumping, blotting, or dipping. Fluxes for brazing are generally available in the form of powder, paste, slurry, or liquid. The form selected depends on the individual work requirements, the brazing process, and the brazing procedure used. Fluxes are most commonly applied in paste form because of the ease with which pastes can be applied to small parts and their adherence in any position. The particle size of paste or dry flux should be uniform and small,
for the most effective application. It is frequently helpful to heat the paste slightly before application. A low-viscosity slurry or diluted paste flux is used when the flux is to be sprayed on a joint. Certain fluxes (types FB3A and B) will completely dissolve in water to produce a liquid solution called liquid flux. Automatic torch brazing has been made possible by the development of face-feeding machines. One of these is the paste feeder, which applies a mixture of flux and filler-metal powder. Methods of applying fluxes, and techniques employed in the use of fluxes, are as follows. Dipping. This is the most popular production method for applying flux. Preformed rings should be in position prior to dipping to ensure a thorough and uniform coating of flux on the ring and to avoid having the operator touch the flux with his fingers. Continual contact with flux may cause a skin disease known as dermatitis. Flux may be thinned for dipping by dilution, by heating, or both. Flux heated to 60 to 70 °C (140 to 160 °F) will adhere to the metal much better than will cold flux. Heating also reduces spattering when the water is boiling out during the brazing operation. Flux pots are available with thermostatic controls that maintain desired temperatures. Spraying. Thin layers of flux may be applied with a standard paint sprayer. The container should be an integral part of the spray gun to simplify cleaning and to prevent flux from caking in the lines. Air pressure keeps the ports clear. Between applications, immerse the container in hot water. Following the operation, the container should be filled with hot water and the ports blown free of all flux. The spraying station should be well ventilated, and all other precautions should be taken to keep the operator from inhaling the spray mist. Brushing of flux on parts has several advantages. For example, the scrubbing action of the bristles helps wet the metal with flux—a very important feature when working with dense materials such as tungsten carbide and highchromium stainless steels. Brushing does not lend itself particularly well to automatic applications. However, it has been applied successfully by using indexing turntables, with the parts rotating. The brushes dip into the flux, rise, and advance to the parts. Rotation of the parts under the brushes ensures complete coverage. Pressure Oil Can Application. On parts with chamfers or concave surfaces, flux can be applied quite easily with a simple pressure-type
272 / Brazing, Second Edition
oil can. The flux must be thinly diluted to ensure good coverage around the entire well. This method is especially good for small joints. Thin coatings may be applied with a sponge set in a bath of flux. The method is particularly helpful where flux is desired on only one surface or on projections from the surface. Improving Filler-Metal Flow. To obtain an effective brazed joint, the filler metal must displace all of the flux in the joint. Generally, this is not difficult, because the filler metal melts at one point in the assembly and, by capillary action between closely fitted members, flows through the joint, flushing the flux ahead of it. On sharp shoulders and close fittings emanating in both directions from the shoulder, the filler metal may have some difficulty in flowing. In this case, the sharp corner mating to the shoulder should be broken to assist flux displacement. Unless this is done, the flux will boil up through the filler metal, causing pinholes, or the filler metal may separate, wetting both mating surfaces but leaving a layer of flux between. Fluxing of Large, Flat Surfaces. On large, flat surfaces, a very thin coating of flux should be applied. Often, it may be necessary to wipe the flux from the surface, leaving only the pores of the metal filled with flux. On flat surfaces, the oxygen is usually excluded, and not much flux is required. Also, because shim stock is generally used, the same flushing action does not occur as with a preformed ring, and a heavy coating of flux may result in too many entrapped flux islands. A heavier shim or washer of smaller area will flush out the flux with fewer voids. A heavy coating of flux can be applied around the perimeter to prevent the entrance of oxygen to the joint. Flux Removal. Flux subjected to atmospheric oxygen is difficult to remove. To aid removal, flux should be used with less dilution. Various methods for reducing scaling should be tried to find the most effective. Flux that has not reacted generally comes off easily in hot water. Quenching of parts after the filler metal has set, but while they are still warm, aids flux removal considerably. Parts that cannot be quenched must be permitted to cool slowly. If flux removal is difficult, a warm bath consisting of 10% sulfuric acid should be used. Dilution of Fluxes. Generally, paste fluxes are supplied so thick that they must be diluted with water to a suitable consistency. When oxidation is light, as on copper and silver, the flux may be thinned considerably. When the flux is
spent and the filler metal does not flow properly, the paste should be heavier. When a heavier flux does not help, brazing conditions must be changed, or another flux must be used. Copper and copper alloys require a considerably diluted flux to remove their light oxides. A heavier flux and general, rather than locally applied, heat are required for carbon steels. A fluoride flux should be used for alloy steels containing chromium, vanadium, or manganese; stainless steels require an active, almost undiluted flux. Particle Size. The particle size of dry flux or paste flux is important, because better fluxing action will result when all constituent particles of a flux are small and thoroughly mixed. Stirring, ball milling, or grinding of a flux mixture is helpful if the flux has become lumpy. Preheating of the paste or liquid flux may facilitate application. Powdered Fluxes. Powdered flux can be applied to the joint in four different ways: • Dry • Mixed with water and alcohol to form a paste • By dipping the heated filler-metal rod into the flux as needed (in torch brazing) • Sprinkled on the joint Mixtures of powdered filler metal and flux are sometimes used where it is desirable for both flux and filler metal to be preplaced. Liquid fluxes (type FB3D), in which fluxing ingredients are completely in solution, may be sprayed on the joint or entrained in the fuel gas. Gas/Flux Mixture. Liquid fluxes are sometimes used in torch brazing. The fuel gas is passed through the liquid flux container, thus entraining the flux in the fuel gas, and the flame and flux are applied where needed. Usually, a small amount of additional preplaced flux is used for the joint surroundings. Slurries. Most slurries are water based; some organic-based fluxes—petroleum- or polyethylene-glycol based, for example—suit precision dispensing due to lower evaporation rates and better viscosity control. Hot rodding, used to braze weld, plunges a hot brazing rod into powder flux. Heat from the rod causes a small amount of flux to adhere to the rod surface. This method is best suited to brazing of shallow joints, up to 6.4 mm (0.25 in.) in steel, because it results in poor capillary penetration in deep joint areas.
Chapter 6: Fluxes and Atmospheres / 273
General Considerations in Flux Application. The paste and liquid flux should adhere to clean metal surfaces. If the metal surfaces are not clean, the flux will ball up and leave bare spots. Thick paste fluxes can be applied by brushing. The proper consistency depends on the types of oxides present as well as the heating cycle. For example, ferrous oxides formed during fast heating of the base metal are soft and easy to remove, and only limited fluxing action is required. However, when joining copper or stainless steel or when the heating cycle is long, a concentrated flux is required. Flux reacts with oxygen, and once it becomes saturated, it loses all its effectiveness. The viscosity of the flux may be reduced without dilution by heating it to 50 to 60 °C (120 to 140 °F), preferably in a ceramic-lined flux or glue pot with a thermostat control. Warm flux has low surface tension and adheres to the metal more readily.
Fluxes and Specific Processes (Ref 29, 30) Flux is required for induction brazing. The flux used should decompose oxides without corroding the base metal or the filler metal, should be extremely active because of the short brazing times employed, and should be easy to remove after brazing. Type FB3A flux is used for an estimated 95% of the induction brazing applications that involve steel. Paste and liquid fluxes are most often applied to the joint by brushing. A flux is used in almost all resistance brazing. It serves the same purposes in resistance brazing as in other brazing processes: • Providing a coating to prevent or minimize oxidation of the work metal during heating • Dissolving oxides that are present or that may form during heating • Assisting the molten filler metal in wetting the work metal to promote capillary flow The flux in resistance brazing, however, has the additional function of serving as an electrical conductor to permit passage of the brazing current through the joint; most dry fluxes are nonconductors and must be mixed with water in order to conduct current. The flux is usually applied as a dilute, waterbased paste shortly before the parts and filler metals are assembled for brazing. If the filler
metal is in powder form, flux can be combined with it in fine-particle paste. The same fluxes are used for resistance brazing as for other brazing processes on the same work metal. Type FB3C fluxes are general-purpose fluxes suitable for most metals that are commonly resistance brazed (although type FB4A flux is needed for copper alloys that contain tin, aluminum, or silicon); type FB1A, B, and C fluxes are used on aluminum alloy work metals. The two general situations in which a flux is not used in resistance brazing are brazing in a vacuum or protective reducing gas or inert atmosphere, and brazing of copper with a BCuP filler metal. A flux is not ordinarily needed in resistance brazing of copper when a BCuP filler metal is used, because these filler metals are self-fluxing on copper by virtue of their phosphorus content. Noncorrosive Flux Braze Process (Nocolok). A noncorrosive aluminum brazing process, Nocolok (Alcan Aluminum Ltd.), has become widely accepted throughout the world as the preferred process to braze aluminum heat exchangers. The Nocolok flux is primarily potassium-aluminum fluoride (no chlorides), is noncorrosive, and is chemically inert on the surface of the aluminum, both before and after brazing. The postbraze residue is compatible with freon, oils, and engine coolants and has no adverse effect on the heat-transfer capability of the components (Ref 31, 32). Although brazing with Nocolok can be successfully achieved using flame and induction heating, the process is at its most effective using a continuous furnace. In common with all fluxes used for brazing aluminum, the action of Nocolok flux is to remove the tough, persistent oxide film from the metal surface and promote filler-metal wetting and flow. The potassium flux operates at brazing temperatures by melting and dissolving the oxide film. The flux melts at 562 °C (1044 °F), just below the eutectic temperature of the aluminum-silicon filler metal (577 °C, or 1071 °F). In contrast, chloride-salt-based fluxes work by melting, followed by penetration and separation of the oxide film from the metal substrate. The benefits of using Nocolok flux include, in essence, that Nocolok flux and its flux residue are noncorrosive, unreactive with aluminum, and almost insoluble in water. The flux residue is known to improve the corrosion resistance of brazed components, both in the as-brazed and
274 / Brazing, Second Edition
painted conditions. In commercial terms, the noncorrosive nature of the flux together with its tolerance to brazing assembly fit-up and flexible process control ensure that Nocolok flux brazing is one of the lowest-cost methods for the joining of aluminum heat exchangers. The best braze quality and lowest flux loadings (3 to 5 g/m2) result from continuous furnace brazing with a dry gas atmosphere having a dewpoint below –40 °C (–40 °F) and an oxygen content of less than 1000 ppm. Most systems currently in use include a flux applicator, a dryoff oven, and a brazing furnace, consisting of an entrance vestibule, brazing chamber, water-jacketed cooling chamber, exit vestibule, and air-cooling chamber. Peripheral equipment includes atmosphere-sampling and -analyzing equipment and a scrubber to remove particulate and hydrogen fluoride vapor generated by the flux in the braze chamber. The fluxer uses a high-purity water system together with powdered Nocolok flux to generate a flux slurry for application on the individual cores in a heat exchanger. Depending on the product being processed, either one or two concentration levels are applied. Typically, the cores are introduced horizontally on a continuous conveyor into a core flux-application stage, followed by a blowoff, and, if required, the cores then proceed to a second stage where a higher concentration of flux is applied to specific locations requiring more aggressive aluminum oxide removal. Customer requirements relative to fluxers have evolved from the need to process only onesized part into the need to handle a variety of different parts in terms of physical dimensions. This has necessitated that the fluxer be able to apply the flux according to the part actually being processed, without manual intervention to change nozzle positions for either flux application or for blowoff. A newly developed system identifies the part being processed as it is being introduced into the system and automatically repositions the flux application and blowoff nozzles to accommodate the changing part profile. Further developments include a refined delivery system that supplies the flux from the mixing reservoirs to the spray nozzles with uniform pressure. Commonly used in the industry today are pneumatic pumps. The newest mechanical delivery system evens out the pump delivery pressure, reducing flux-settling tendencies, and provides a uniform flow. One of the most critical aspects of the brazing process is the flux
application, because excess flux can cause potential problems with corrosion in the braze furnace, poor part finish, and flux buildup within the braze chamber itself. Recent changes to blowoff technology use a combination of volume and velocity to optimize the removal of excess flux. A newly designed dryoff oven employs an air-delivery system that has reduced the drying cycle time by over 40% while at the same time improving part temperature uniformity. The brazing system has also seen a number of improvements. When first introduced, the brazing furnace consisted of two separate chambers connected together. The first consisted of a forced convection preheat chamber followed thereafter by a muffle chamber without convection, both of which employ a nitrogen atmosphere. The current design is 100% radiant heat transfer, and this radiation furnace also has proven to provide the most uniform heating, even better than forced convection. The braze furnace design strategically places the heating elements not only for uniform distribution of energy throughout the entire length but also uses the concept of employing the greater amount of energy at the entrance and a lesser amount as the parts continue through the furnace. This compares favorably to alternative designs that use a modular concept where all zones have the same kilowattage. The latter concept has the unfortunate problem that insufficient energy is available to the product when first entering, thereby slowing the heating cycle. Secondly, such modular designs have too much energy available in the latter portion of the furnace, which consequently is not used. As a result, modular designs generally have more connected kilowattage and a slower heated rate, which translates into higher operating costs (the higher the connected kilowattage, the higher the peak demand) and longer heating cycles. Improvements were made on the nitrogen delivery system to the braze furnace, using the thermal properties of the cold incoming nitrogen to absorb energy from the exiting conveyor and product, raising the nitrogen temperature to that of the brazing furnace chamber temperature prior to entering into the braze chamber. This design also enhances the removal of HF and water vapor, which are the two contaminants most undesirable, both in terms of their effect on the product and on the alloy components of the furnace. As to the environment, the potassium-aluminum fluoride aluminate flux in the molten
Chapter 6: Fluxes and Atmospheres / 275
condition generates a hydrogen fluoride vapor. This vapor, together with any chemically released water (there is some chemically bound water in the Nocolok flux), is directed toward the front end of the furnace to be taken off through the scrubber system. This scrubber not only absorbs the hydrogen fluoride vapor, but also removes a significant portion of KAlF4, which condenses during cooling. In addition, the scrubber removes other contaminants, such as remnants of curtain materials, which may be used to pressure seal the furnace. This concept has substantially lengthened the service life of the activated alumina in the scrubber, resulting in material cost savings as well as reduced maintenance labor expense and greater uptime.
Application Quantity One must apply enough flux to coat the joint faces and adjacent surfaces with a thin layer. Excess flux will not compromise joint quality and may even assist flux removal, because residues will be less loaded with metal oxide and more soluble in water. Also, applying flux to surfaces adjacent to the joint helps to prevent oxidation of the workpiece and may act as a flux reservoir, draining flux into the joint. Using too little flux, however, can lead to premature flux exhaustion and inadequate coverage, producing unsound or unsightly brazed joints. It is better to err on the side of too much, rather than too little, flux. The choice of heat source has little effect on flux selection. Exceptions include salt bath heating, which requires dip brazing fluxes; specialized high-temperature torches using a flammable-liquid flux; and furnace brazing, which often calls for a powder flux to minimize the amount of vapor. Boron-modified fluxes are often preferred for induction heating. Torches usually apply one of the high-temperature fluxes, FB3K. This flux is a flammable liquid containing trimethyl borate. A dispenser installed in the fuel gas line feeds flux vapor into the flame.
Base Metal/Filler Metal/Flux Combinations The filler metal/flux combination can be either a brazing paste or a flux-coated rod. Pastes, mixtures of filler-metal powder and flux, and sometimes, an organic binder to ease dispensing work well for automated processes; aluminum, silver, and high-temperature brazing pastes are most popular. Flux-coated rods per-
form brazing and braze welding. The most common flux-coated filler-metal rods are silverbrazing and low-fuming bronze, used primarily to braze weld. The variety of base metals and alloys that are joined by brazing has prompted the development of many different fluxes in addition to those listed in Table 6.7. Two fluxes have been used successfully for furnace or induction brazing of beryllium with good results: a 60%LiF40%LiCl flux and a tin chloride flux. It is important from a safety standpoint that beryllium and beryllium compounds are toxic as flux residues. Only approved installations should consider brazing of beryllium, regardless of the methods used. The FB2-type flux has been used to clean the surface of magnesium, which permits capillary flow. Because of the corrosive nature of this flux, complete removal is of utmost importance if good corrosion resistance is to be obtained in brazed joints. In the refractory metal family, tantalum may be brazed in air using fluxes normally used for brazing with aluminum filler metals or fluxes that are suited to the particular filler metal being used. However, tantalum and niobium require protective coatings, such as nickel or copper electroplate, to induce wetting during brazing. Conventional low-temperature fluxes have been used in brazing tungsten for electrical contact applications when silver- and copper-base filler metals are used. In brazing molybdenum with an oxyacetylene torch, fair protection may be obtained by using a combination of fluxes—a commercial borate- or silver-base brazing flux plus a hightemperature flux containing calcium fluoride. The temperature range over which these fluxes are active is the range from 565 to 1425 °C (1050 to 2600 °F). The molybdenum is first coated with the commercial silver brazing flux, and then the high-temperature flux is applied. The silver flux is active at the lower end of the active temperature range; the high-temperature flux then takes over and is active up to 1425 °C (2600 °F). When BAg filler metals are used for brazing nickel-base alloys, FB3A- and FB3B-type fluxes are suitable for most alloys not containing aluminum, whereas FB4 flux may be used with aluminum-containing nickel alloys. Because most cast irons are brazed at relatively low temperatures, the filler metals used are almost exclusively silver-base filler metals. Of these, BAg-1 is most often used for brazing
276 / Brazing, Second Edition
of cast iron, principally because it has the lowest brazing temperature range. A fluoride-type FB3A flux is usually used with the BAg-1 filler metal, whereas FB3B flux is used with other filler metals in the BAg series. The selection of flux for brazing low-alloy and carbon steels depends on the filler metal (Ref 2). Fluxes FB3A, FB3B, and FB4A are suitable for BAg filler metals; type FB4A normally is used with RBCuZn filler metals. Fluxes and atmospheres may also be used together. The flux can be in either paste or powder form or can be combined with the filler metal. For example, in a face-fed operation, the hand-held filler metal can be coated with the appropriate flux. Typical fluxes employed for prefluxing of low-alloy and carbon steels that are brazed in a neutral chloride salt bath are FB3A and FB3B. Generally, the application of flux to an assembly is not necessary when a cyanide bath is used. In brazing of copper-aluminum alloys (aluminum bronzes), the formation of refractory aluminum compounds creates difficulty in wetting, and, as a result, strong fluxes are required. Aluminum bronzes can be brazed with silverbearing filler metals and type FB4A flux. The RBCuZn filler metals may be used for brazing of coppers and of copper-nickel, coppersilicon, and copper-tin alloys. However, they are not useful for brazing aluminum bronzes, because the required brazing temperatures destroy the effectiveness of the fluxes required for these base metals (Table 6.7). With the copper-zinc filler metals, care should be taken not to overheat the metal, because volatization of the zinc causes voids in the joint. In torch brazing, an oxidizing flame will reduce zinc fuming, and FB3D brazing flux should be used. The BCuP filler metals are useful for brazing high-leaded cast brass pipe fittings, if precautions are taken to flux properly and avoid overheating. Brasses containing aluminum or silicon require treatment similar to aluminum or silicon bronzes. Lead added to brass to improve machinability may alloy with the filler metal and cause brittleness. Major brazing difficulties occur when the lead content is over 2 or 3%. To maintain good flow and wetting during brazing, leaded brasses require complete flux coverage to prevent the formation of lead oxide or dross. Additionally, the FB3A and FB3B fluxes are suitable for use with BCuP and BAg filler metals in brazing all the copper base metals except aluminum bronzes.
Refractory oxides form easily on aluminum bronzes, and the more active FB4-type fluxes are needed to cope with them. The effectiveness of type FB3A flux may be reduced rapidly at the temperatures needed for brazing with RBCuZn filler metals and is completely destroyed in brazing with BCu. Type FB5 flux may be used with these filler metals, except in brazing of aluminum bronze or beryllium-copper. More active fluxes are needed for these base metals, and mixtures of FB4 and FB5 fluxes may be found satisfactory for the few applications of this kind. In brazing of copper, the copper-phosphorus and copper-silver-phosphorus filler metals are self-fluxing. Flux is beneficial, however, for heavy assemblies where prolonged heating would otherwise cause excessive oxidation. The special coppers that contain small additions of silver, lead, tellurium, selenium, or sulfur (generally no more than 1%) are brazed readily with the self-fluxing BCuP filler metals. Wetting action is improved when a flux is used and when there is a small amount of shearing motion between the components while the filler metal is molten. Finally, FB3-type flux is suitable for most applications in which copper-nickel alloys are brazed. Aluminum forms a natural refractory oxide that is remarkably stable and tenacious. It is mechanically durable, with a hardness that is inferior only to that of diamond, and its high melting point (2050 °C, or 3720 °F) reflects its high degree of physical stability. Alumina is also chemically stable to the extent that it cannot be directly reduced to the metal by aqueous reagents. On exposure to air, a layer of alumina will form almost instantaneously on the surface of aluminum and will grow to an equilibrium thickness of between 2 and 5 nm at ambient temperature. On heating to 500 to 600 °C (930 to 1110 °F), the thickness of this surface coating will increase to approximately 1 µm. Therefore, special fluxes have been formulated for use with aluminum alloys. These have to be particularly effective in protecting the metal from oxidation. The aluminum fluxes are divided into two categories: those that are suitable for use with solders at temperatures below 450 °C (840 °F), and those that can be used at higher temperatures with brazes. A commonality between the aluminum soldering and brazing fluxes is that they all contain halide compounds. These are highly corrosive, especially in the presence of
Chapter 6: Fluxes and Atmospheres / 277
moisture, including humid atmospheres. Therefore, all flux residues must be removed as completely as possible. The cleaning processes are very laborious and costly, and there is always a danger that some residues will survive the cleaning procedures, resulting in corrosion in the vicinity of the joint. Aluminum Brazing Fluxes. Two principal types of fluxes are used for brazing aluminum (Ref 33): • Chloride formulations: The active ingredients are chlorides of the alkali earth metals. These fluxes operate by infiltrating cracks in the alumina; on reaching the metal, they proceed to undermine the oxide layer and mechanically displace it. The flux residues left on the workpiece surfaces are highly corrosive and must be completely removed. • Fluoride formulations: Many of the wellknown fluxes of this type contain a mixed sodium-aluminum fluoride that, when molten, can dissolve alumina, but the residues are the source of severe corrosion if left on the component surfaces. However, by using potassium rather than sodium in the formulation, the flux can be made neutral without compromising its ability to dissolve alumina. The proprietary Nocolok flux comprises a eutectic between K3AlF6 and KAlF4, which melts at 562 °C (1044 °F) (Ref 34, 35). The Nocolok flux is not hygroscopic, and its residues do not corrode aluminum. Therefore, it can be applied to joint surfaces and left there. It does not spall off during thermal or other forms of stressing (Ref 34). Researchers (Ref 36) developed a brazing method that permits aluminum heat-exchanger tubes coated with 4 to 8 µm of zinc to be brazed to unclad aluminum alloy finstock. Deposition of a layer of zinc on aluminum is well established, but only for soldering and only at thicknesses of 30 µm or more. The new technology enables sound joints to be made over a wide range of heating conditions, with preferred temperature cycles similar to industry standards for brazing. In the technology, zinc is coated onto aluminum heat-exchanger tubing prior to brazing to provide filler metal for the joints and to diffuse into the surface during heating. The resulting outer zinc-rich layer is anodic to the metal below it and provides progressive corrosion
protection to the substrate. Although zinc diffuses very slowly into aluminum at temperatures below 400 °C (750 °F), diffusion rates increase rapidly as the temperature approaches 600 °C (1110 °F), the normal peak temperature for brazing aluminum. Developers modeled the thermal cycles of the procedure on industrial brazing standards and chose the conventional fluoride-base flux. Therefore, customary practices and equipment can be used. Another research and development program (Ref 37) developed a novel approach that eliminates the need for cladding aluminum components to be joined with a filler metal. In other respects, however, the process is identical to conventional nitrogen furnace brazing. In this technique, one aluminum surface in each joint is coated with a thin layer of silicon powder (average particle diameter 1 to 100 µm) and the potassium fluoroaluminate flux powder mixture (average particle diameter 1 µm) in a ratio by weight of silicon to flux powder of 1:1 to 1:3 (Fig. 4.5a). The powder mix is most easily deposited from water-based slurry by dipping, but other techniques, such as electrostatic spraying or deposition using a volatile binder, can be used. The amount of silicon powder used to cover the surface can range from a few to several tens of grams per square meter, depending on the joining application. Brazing is carried out by heating at a rate of approximately 6 °C/min (10 °F/min) in nitrogen at near-atmospheric pressure for a few minutes. During the temperature rampup, the flux melts (at 562 °C, or 1044 °F) and dissolves the surface film of aluminum oxide (Fig. 4.5b). The dissolution of this surface film allows the silicon particles to come into intimate contact with the bare aluminum. The large elemental concentration gradients at the aluminum-silicon interface cause interdiffusion between the aluminum and silicon (Fig. 4.5c). At temperatures above the aluminum-silicon eutectic-reaction temperature of 577 °C (1071 °F), the silicon particles diffuse rapidly into the aluminum surface, generating in situ a layer of aluminum-silicon alloy with a composition near the eutectic (Fig. 4.5d). The filler metal penetrates the joint by capillary action and forms a thin strip (fillet), thereby producing a metallurgical bond on cooling. Because of this capillary flow, brazing requires minimal contact force at the joint interface. Any unused filler metal remains on the aluminum surface to form a layer of aluminum-silicon
278 / Brazing, Second Edition
alloy with a composition near the eutectic (Fig. 4.5e). This mechanism is not limited to the combination of aluminum and silicon. It occurs whenever the interface between two metals is heated above their eutectic-reaction temperature. In principle, therefore, aluminum joints can be brazed using any intermediary metal that is characterized by having a relatively low eutectic-reaction temperature with aluminum, such as copper, germanium, or zinc. Similarly, other metals, such as copper, brass, steel, and so on, can be brazed using a powder of an appropriate eutectic-forming intermediary metal. The use of a flux is essential, because it eliminates oxide interdiffusion barriers. The use of silicon for this aluminum joining process has been found to be the most cost-effective. To successfully apply this novel brazing technique, the metal surfaces must be uniformly coated (1 to 8 × 10–3 kg/m2) with the silicon flux. However, in order to deposit a uniform coating from water-based slurry, the aluminum surface must first be chemically cleaned to remove any organic contaminants that would prevent uniform wetting of the surface by water. This is achieved by first immersing the surface to be coated in a weak solution of sodium hydroxide (NaOH) for a few seconds, then in nitric acid, followed by rinsing in water, and then drying for a few minutes in an air furnace. More information on cleaning is provided in Chapter 7, “Fixturing, Tooling, Stopoffs, Parting Agents, Surface Preparation, Surface Cleaning, and Repair.” This brazing technique can be used with a wide range of aluminum alloys, provided that their magnesium content is <0.1 wt%. A higher content impairs the ability of the fluoroaluminate flux to dissolve aluminum oxide surface films, because the formation of magnesium fluoride (MgF2) at elevated temperatures prevents aluminum-silicon interdiffusion and joining. However, magnesium-containing alloys can be brazed by cladding their surface with any magnesium-free aluminum alloy, thereby generating filler metal from the cladding material. One advantage of this new brazing process is that other materials can be introduced into the surface to enhance specific properties. When zinc powder is added to the silicon-flux mix, zinc diffuses into the coated surface without adversely affecting the generation of filler metal. For example, the deposition of a silicon-
flux-zinc powder mix from water-based slurry to produce a zinc surface coverage of 4.3 × 10–3 kg/m2 results in an appreciable amount of dissolved zinc over a distance of approximately 100 µm into the aluminum surface following brazing. This zinc concentration was found to enhance the resistance of the aluminum component to corrosion. Aluminum-copper joints can be brazed using this technique by depositing the silicon-flux mix on the aluminum surface. During heating, liquid metal is first produced by dissolution of silicon into aluminum to form an aluminum-silicon eutectic liquid above 577 °C (1071 °F), as explained earlier. Wetting of the copper surface by this liquid metal immediately causes the copper to dissolve, forming a molten ternary alloy. This alloy then constitutes the filler metal, which has a composition from Cu0.3Al0.62Si0.08 to Cu0.02Al0.97Si0.01, depending on the temperature. Copper-copper and copper-brass joints can also be brazed using the silicon-flux powder mix. In this case, the filler metal, which consists of a copper-silicon liquid of near-eutectic composition, is generated in a way similar to that illustrated in Fig. 4.5, by dissolution of silicon onto the copper surface at a temperature above the copper-silicon eutectic-reaction temperature of 803 °C (1477 °F). The use of a silicon-flux powder mix to braze aluminum components, thereby eliminating the need for cladding metal, has been successfully tested, and the process is highly cost-effective, because it eliminates the use of relatively expensive clad brazing sheet.
Postbraze Cleaning and Flux Removal There are seven major reasons for removing residual flux after brazing: • The joint cannot be inspected for soundness until the cover of flux residue has been removed. • The joint may be bound together by the flux in the semblance of a brazed joint, only to break apart later in service. • In fluid or pressure service, the flux may block pinholes that might withstand a pressure test but would leak soon after being placed in service. • If left on the joint, the flux attracts available water, resulting in oxidation and corrosion.
Chapter 6: Fluxes and Atmospheres / 279
• Painting, coating, or plating cannot be done satisfactorily on areas covered with flux residue. • If a joint is overheated, excess heat impairs flux removal; spent flux residues, saturated with metal oxides, are most difficult to remove. • To avoid flux exhaustion, apply excess flux to ease removal of residues from the base material. If parts have been well cleaned before brazing and not overheated during brazing, flux residue can usually be removed by a hot water rinse followed by thorough drying. To avoid corrosion, flux removal should be delayed no more than 48 h. A quick method of removing glasslike residues is to quench the joint in cold water after brazing and thus crack off the deposit by thermal shock. However, in some applications, such treatment may cause distortion of the brazed assembly. Scrubbing, applying a steam jet, and most of the standard abrasive techniques, such as wire brushing and abrasive blasting, are also used to dislodge stubborn flux residues, provided the operation does not impair the function of the assembly. When the flux cannot be removed from steel assemblies by rinsing in cold or hot water, a cold 5% solution of sulfuric acid will prove more effective. This works for nonferrous metals and mild steels. More aggressive solutions are needed for stainless steels and high-temperature alloys. The solution may be warmed to accelerate the action, provided that care is taken to prevent excessive attack on the assembly. A small addition of sodium dichromate to the solution makes the action even faster, but the time of immersion must be carefully controlled to avoid the greater risk of etching the steel. Phosphate solutions similar to those used for cleaning steel are effective flux removers and have the added advantage of giving carbon steel assemblies a temporary protective coating. However, the coating will hamper subsequent brazing operations. Boric acid, as applied in gas fluxing, can be removed by washing in clean water heated to at least 65 °C (150 °F). (Boric acid is only slightly soluble in cold water.) Mixed borax and boric acid fluxes are more difficult to remove than other types. Fortu-
nately, moisture absorption and corrosion are minimal with borax fluxes. In fact, rather than risk damage to delicate assemblies, such as electronic components, when mixed borax and boric acid fluxes are used, the flux is sometimes allowed to remain after brazing. These fluxes can be removed by quenching, shot blasting, sand blasting, chipping, filing, scraping, and wire brushing. The rate of solution in water is slow, and even if dilute sulfuric acid is used, the necessary period of immersion may be inconveniently long for production work. Fluoride fluxes are soluble in water and are much easier to remove than borax fluxes. Holding under running cold water while brushing with a wire or bristle brush will usually suffice. Alternatively, the assemblies may be boiled in water for a few minutes and then rinsed in cold water. Brazed parts should be quenched while still hot, if at all possible, but the filler metal must be completely solidified before any water quenching begins. If the flux has not been too heavily oxidized, quenching and agitating the part in hot water should be sufficient to remove all flux residues. Dilute sulfuric acid solutions and phosphate solutions can also be used for quicker results. The residue of fluoride fluxes is hygroscopic, and if the assembly is not quenched after brazing, it is often advantageous to postpone postbrazing cleaning for 24 h. Under normal atmospheric conditions, the residue will absorb moisture during this period and will become more readily soluble in any of the solvents previously mentioned. It is helpful to think of flux as performing like an ink blotter. The blotter can absorb only so much ink, after which it becomes saturated and unable to absorb any more ink. Similarly, flux will absorb oxides, which are continuously generated during the brazing cycle (oxides are not generally formed in brazing methods that exclude oxygen-bearing atmospheres) but only to its point of saturation. Once that point has been reached, the flux becomes useless, and further heating will only oxidize the joint. Fully saturated fluxes are generally hard to remove. In such cases, acidifying the water will accelerate removal of the flux residue. Flux residue should be removed completely to avoid corrosion by the residual active chemicals. The residue obtained from a flux, particularly when considerable oxide removal has occurred, is a form of glass. The less flux
280 / Brazing, Second Edition
required to clean the metal, the less will be the formation of glass, and the easier the task of removing the flux residue. The thermal shock cracks off the residue. By the same token, the use of an adequate amount of the proper flux will usually make residue removal easier, because spent flux has high glass content. Flux removal from properly cleaned, brazed parts usually can be accomplished by rinsing in hot water accompanied by light brushing. Preferably, this rinse should be done immediately after the brazing operation, and thorough drying following rinsing is recommended. Tenaciously adhering flux particles may be removed either by dipping the parts in one of several proprietary chemical dips or by mechanical means such as fiber brushing, wire brushing, shot blasting, or chipping. The method used should be compatible with the required properties of the base metals. If, for example, stainless steel parts are cleaned with a wire brush, a stainless steel brush should be used. Soft metals, such as aluminum and copper, should be cleaned by methods that will not damage or roughen their surfaces. Cleaning with nitric acid is not recommended where filler metals containing silver or copper have been used. Parts should be thoroughly washed and dried after chemical dipping. As brazing technology becomes more sophisticated, methods of detecting flux residues not removed by traditional chemical methods are being developed. Ion-scattering spectrometry (ISS) and secondary ion mass spectrometry (SIMS) have successfully been adapted for detecting residual flux in brazed joints in service. Further, ISS and SIMS can differentiate
between chloride and fluoride fluxes. These new methods offer an opportunity to literally fingerprint flux residuals. ISS and SIMS analysis also can detect residuals and contaminants solely on the surface. Then, by depth profiling, the braze deposit can be analyzed independently of the surface and/or the base metal, and finally, the braze/base metal joint can be analyzed. Fluxes used in brazing of aluminum alloys can cause corrosion if allowed to remain on the parts. It is therefore essential to clean joints after brazing. A thorough water rinse followed by a chemical treatment is the most effective means of complete flux removal. As much flux as possible should be removed by immersing the parts in an overflowing bath of boiling water just after the filler metal has solidified (Ref 38). If such a quench produces distortion, the parts should be allowed to cool in air before immersion, to decrease the thermal shock. When both sides of a brazed joint are accessible, scrubbing with a fiber brush in boiling water will remove most of the flux. For parts too large for water baths, the joints should be scrubbed with hot water and rinsed with cold water. A pressure spray washer may be the best first step. A steam jet is also effective in opening passages plugged by flux. Any of several acid solutions (Table 6.8) will remove any flux after washing. The choice depends largely on the thickness of the brazed parts, the accessibility of fluxed areas, and the adequacy of flux removal in the initial water treatment. Agitation and turbulence improve the efficacy of any flux-removal treatment. Ultrasonic cleaning is effective for cleaning inaccessible
Table 6.8 Solutions for removing brazing flux from aluminum parts Operating temperature Type of solution
Composition
°C
°F
Procedure(a)
Immerse for 10 to 20 min; rinse in hot or cold water Immerse for 10 to 15 min; rinse in cold water; rinse in hot water; dry Immerse for 5 to 10 min; rinse in cold water; immerse in nitric acid solution (first entry in table); rinse in hot or cold water Immerse for 10 to 15 min; rinse in hot or cold water Immerse for 5 to 30 min; rinse in hot water
Nitric acid
58 to 62% HNO3, H2O
20
68
Nitric-hydro-fluoride acid
58 to 62% HNO3, 48% HF, H2O
20
68
Hydrofluoride acid
48% HF, H2O
20
68
Phosphoric acid/chromium trioxide Nitric acid/sodium dichromate
85% H3PO4, CrO3, H2O
82
180
58 to 62% HNO3, Na2Cr2O7.2H2O, H2O
60
140
(a) Before using any of these solutions, it is recommended that the assembly first be immersed in boiling water to remove the major portion of the flux. Source: Ref 2
Chapter 6: Fluxes and Atmospheres / 281
areas, decreases the immersion time, and reduces the possibility of attack on the aluminum. Checking for complete flux removal should be a routine inspection procedure. To detect the presence of flux, a few drops of distilled water are put on the surface to be tested and left there for a few seconds. The water is then picked off with an eyedropper and placed in an acidified solution of 5% silver nitrate. If the solution stays clear, the metal is clean. If a white precipitate clouds the solution, flux is still present on the surface. Flux-removal procedures must then be repeated until the brazed assembly tests clean. Complete removal of the flux is essential, because it is corrosive to aluminum in the presence of moisture (Ref 39).
Case Histories and Problem-Solving Examples Example 1: Leaking in a Copper and Brass Valve Assembly. A brazing firm encountered a problem of a high leak rate on a copper and brass valve assembly. The main body and fitting of the assembly were brass, and the tube was copper. The filler metal was BAg-1, and the parts were brazed in a continuous furnace at 704 to 732 °C (1300 to 1350 °F). The atmosphere used was dissociated NH3, with a liquid flux to assist brazing. The parts were previously torch brazed successfully with a minimum number of rejections; however, production demands shifted the brazing method to furnace. Why then was there a problem? Zinc- and cadmium-containing filler metals should not be used in a controlled atmosphere furnace; thus, BAg-1 filler metal is not suitable for this furnace atmosphere. The first problem with the filler metal is that it contains approximately 24% Cd, which is considered to be a poisonous material. It is readily dissociated from the filler metal and either deposited in the furnace as a powder or carried out in the atmosphere; both cases lead to a hazardous condition. This is not the proper use of this type of filler metal. Secondly, the 16% Zn also dissociates, thus leading up to 40% loss in the volume of filler metal applied to the joint. A 1.59 mm (0.063 in.) diameter wire was used in torch brazing and was reduced to a 1.0 mm (0.04 in.) diameter wire for furnace brazing; thus, the filler metal has been decreased by 59%. Up to an
additional 40% of filler metal can be lost by evaporation of the zinc and cadmium from the BAg-1 filler metal, leading to an overall reduction of 75.4%, which is an insufficient quantity of filler metal to fill the large joint properly. An additional problem could be caused by continual vaporization of the zinc and cadmium, adding to some porosity in the joint area. The liquid flux used in this instance ran down on the fitting part in one area and appeared to keep the base metal bright and clean in that area, indicating the flux was doing its job. It also left a residue on the filler-metal fillet, indicated by black spots on the surface of the fillet, meaning additional cleaning was required. Flux used in an atmosphere furnace will vaporize partially and build up a flux residue in the cooler areas of the furnace. Because the fluxes are usually hygroscopic materials, they will pick up moisture when the furnace is shut down, causing considerable difficulty when a dry atmosphere is required. Therefore, in summary, the filler metal should be changed to a silver-copper or silvercopper-tin composition, which would not have any elements that would vaporize in a dissociated NH3 atmosphere. A better-quality part and joint would be obtained if the fitting part and casting (main body part) were brass plated, particularly in the joint area. Plating the inside diameter of a hole, even though it is approximately 19 mm (0.75 in.) in diameter, is difficult, because the throwing power is not too good, and internal electrodes may have to be used. With more filler metal in the joint, no bubbling from vaporization, and covering up the brass with copper-tin plating, including in the joint area, the quality of the brazement should improve. However, more filler metal is the most important variable. Example 2: Brazing of Tungsten Carbide. Copper readily wets some carbides but tends to ball up and not wet the surface on others. Why? If the surface is ground off before brazing tungsten carbide (WC), then copper wets and flows profusely. Another method that has been used is to etch WC in a mixture of 10 to 20% nitric acid and 1 to 2% hydrofluoric acid while another system that was developed some years ago was a salt bath treatment in which the surfaces were oxidized and said to remove the WC. Plating is a viable means of ensuring adequate wetting and flow. Nickel is the best choice to use for the plating material.
282 / Brazing, Second Edition
Example 3: Residue on Steel Parts After Brazing. A 1018 steel was being brazed in a continuous furnace with BCu-2 copper paste filler metal on WC in an exothermic atmosphere and approximately +30 °C (+86 °F) dewpoint atmosphere. Most of the time, parts have come out bright and clean, but they sometimes have a black residue where the filler metal is applied. Why does this occur, and how can it be prevented? The black residue comes from the organic binders used to make the paste viscous so it can easily be dispensed from a cartridge. To control the atmosphere in the furnace, a dewpoint instrument should be used to measure the amount of moisture, or water vapor, in the atmosphere. The oxygen molecule of the water vapor is the controlling element, because the process is very sensitive to the amount of oxygen in the atmosphere. Because there is hydrogen in an exothermic atmosphere, the oxygen drawn into the furnace with the parts will, at temperature, combine with the hydrogen to produce more water or moisture. The easiest way to determine the oxygen content of the atmosphere so it can be controlled means measuring the moisture or water as dewpoint of the atmosphere in the furnace. To properly control the atmosphere in the brazing furnace, it is essential for a brazing company to have a recording dewpoint instrument that takes the atmosphere from the location where the parts are expected to be at or near their maximum temperature. Example 4: Atmosphere Control for Brazing Stainless Steel and Copper. When brazing stainless steel with copper, the chromium in the stainless steel becomes the first element to oxidize. Therefore, the atmosphere must be controlled so there is a low enough partial pressure of oxygen to ensure dissociation of the chromium oxide on the surface at higher temperatures. Preferably, this is a dewpoint of –46 °C (–51 °F) or below. When brazing with the BNi-2 filler metal, it is necessary to have a lower partial pressure of oxygen to get suitable results. Example 5: Furnace Preparation for Brazing René 77. How should one handle a furnace for brazing René 77? Should it be cleaned in pure dry hydrogen or by using vacuum? The concern is that some parts being brazed are castings, which cause some difficulty if not furnace cleaned; other parts are overhauled with cracks in them and are to be furnace cleaned and then brazed.
Furnace cleaning can be divided into: • Type A: Metals that do not contain aluminum, titanium, or other similar elements with oxides that cannot be dissociated in hydrogen or vacuum at the normal brazing temperature of 1008 to 1120 °C (1846 to 2050 °F) • Type B: Base metals that contain aluminum or other similar readily oxidizable elements where the surface or the crack has been oxidized during engine operation • Type C: Base metals of type A, which are clean sheet metal or machined parts that have been degreased, can be brazed without furnace cleaning. If brazing is not satisfactory, the furnace atmosphere and surface contaminants should be investigated. Stainless and nickel-alloy brazing, generally in an atmosphere of 0.13 to 10–3 Pa (1.89 × 10–5 to 1.45 × 10–7 psi, can have a good or bad atmosphere, regardless of the vacuum pressure. The best way to monitor the atmosphere is with a T-specimen and the nickel-base filler metal (see T-specimen tests, Ref 9). With each load of parts, a T-specimen, with a sufficient amount of the nickel-base filler metal applied to one side of one end, should be included. The amount of filler metal remaining at the point of application will reveal the atmosphere. If there is no filler metal remaining at the point of application, the atmosphere is good for that specific filler metal. If quantities remain, the increase in remaining filler metal indicates a degrading of the atmosphere. Under certain conditions, type A base metals can benefit from furnace cleaning. Typical of these base metals are castings containing nitrogen, which react with boron in some filler metals. Similar parts run in dissociated ammonia will pick up enough nitrogen to prevent the filler metals from adequately melting and flowing in boron-containing filler metals. Type B produces a titanium or aluminum oxide on the surface when furnace cleaned in hydrogen or vacuum. These base metals should not be furnace cleaned in the standard hydrogen or vacuum systems, for example, René 77, which contains 3.5% Ti and 4.4% Al. If furnace cleaning of these base metals is required, it can be accomplished with the fluoride-ion cleaning process (see Chapter 7), which removes the oxides from the surface and inside the cracks. Fluoride-ion cleaning will also remove titanium and aluminum from the base-metal surface to a
Chapter 6: Fluxes and Atmospheres / 283
depth of 0.03 to 0.05 mm (0.001 to 0.002 in.). This removal from the surface allows the part to be brazed without nickel plating and along with nonaluminum or nontitanium parts. Most braze operators prefer pure dry hydrogen for cleaning parts, because it has been seen that hydrogen diffuses through solid base metal, which removes various unwanted elements, such as sulfur, from combustion chambers that have been in aircraft flying at very high levels. While vacuum may be suitable, at the present time, there is no test data indicating vacuum furnace cleaning is equivalent to pure dry hydrogen. Example 6: Brazing with Nickel-Base Filler Metals. Within the nickel-base filler metals, there is a varying degree of sensitivity to atmosphere, depending on the chemistry. The BNi-5 filler metal, with a high chromium and silicon content, is at the top of the list of filler metals that are sensitive to variations in atmosphere. The BNi-2 filler metal is toward the middle of the filler-metal sensitivity list. Therefore, an atmosphere that will run BNi-2 is satisfactorily used and successfully brazed may not be suitable for BNi-5. Likewise, Inconel 625, which contains aluminum and titanium, is also more sensitive to the partial pressure of oxygen in the atmosphere than is a 304 stainless steel. These oxygen-sensitive materials require a much better atmosphere quality in the furnace than filler metals and base metals that are less sensitive to atmosphere quality. If a vacuum gage indicates a good vacuum does not mean there is a good brazing atmosphere in the furnace. An atmosphere indicating a vacuum of 0.13 to 0.00013 Pa may not be suitable for obtaining good braze quality. One must consider other variables. When a part comes out of a furnace with a gray discoloration and other parts are bright and clean, this indicates substantial outgassing coming from the heat shielding and other fixturing in the furnace that is increasing the partial pressure of oxygen in the system. If a part had a green color using the BNi-5 filler metal, there was sufficient oxygen present in the system to allow green chromium oxide to form on the filler metal. The atmosphere in the best of furnaces is still oxidizing between 538 and 927 °C (1000 and 1701 °F), and this also holds true for vacuum atmospheres in the 0.13 to 0.00013 Pa range. The footprints indicating the brighter surface in some parts in the furnace are the outgassing, which is attaching to the line-of-sight surfaces
and come from the furnace heat shielding, grates, and fixtures. There are two potential solutions. (a) Wrap the entire part in a stainless steel foil. The outgassing will stick to the foil. (b) Clean up the furnace, thus eliminating the outgassing and making sure that all minute leaks are sealed. Example 7: Brazing Aluminum Bronze to Naval Brass. Some manufacturers would like to braze aluminum bronze to naval brass in pure dry hydrogen with a BAg-8 silver-copper eutectic filler metal. How? Aluminum bronze base metal has an aluminum content of 7 to 10%. Because aluminum oxidizes readily, even in the driest of hydrogen atmospheres, this is undoubtedly the major problem. Likewise, if the atmosphere is quite dry, some aluminum could vaporize from the surface and aluminize the filler metal (either in wire or powder form). This also would cause the filler metal to not melt properly and to leave more residue. One of the simplest methods for solving this problem is to use a grade of flux specifically formulated to handle aluminum bronze base metals. It is recommended to use (Table 6.7) American Welding Society (AWS) brazing flux type 4. If it is desired not to have any flux in the joint area, then it is possible to plate the aluminum bronze with either nickel or copper. However, the higher the temperature and the longer the time to come up to heat, the thicker the plating required. It is suggested that 0.01 mm (0.0004 in.) thickness would be a good starting place. In brazing naval brass in a hydrogen atmosphere, the vapor pressure of zinc is increased to the point that dezincification can occur. This is true whether the atmosphere is exothermic, pure nitrogen, pure hydrogen, nitrogen/hydrogen, or any of the gas atmospheres. Because the BAg-8 silver-copper eutectic is a potential filler metal, the brazing temperature would have to be approximately 788 to 816 °C (1450 to 1501 °F). To reduce the amount of dezincification, it would be better to use the BAg-18 filler metal, which has a lower melting point and thus will lower the amount of zinc vaporization that may occur. The brazing range on this filler metal starts at 718 °C (1324 °F), and 760 °C (1400 °F) would be a suitable brazing temperature. Dezincification will be noted as a change from the brass color to a whiter surface, or, depending on the furnace, brown fumes coming out of the furnace. Protective atmospheres are used in furnace
284 / Brazing, Second Edition
brazing of copper and copper alloys, although numerous exceptions exist. Exothermic-based, endothermic-based, dissociated ammonia, and other suitable prepared atmospheres are widely used to protect copper and copper alloys that are not adversely affected by hydrogen at elevated temperatures. Depending on the base-metal/filler-metal combination, the use of a brazing flux may be avoided by brazing in either a dry hydrogen or a prepared nitrogen-base atmosphere. These are among the more expensive furnace atmospheres; however, at low dewpoints, they have the advantage of being highly reducing. Prepared exothermic-based atmospheres are considerably less expensive and are effective in preventing oxidation at elevated temperatures, but the reducing potential of these atmospheres is limited, because hydrogen content does not exceed approximately 13%, and consequently, they cannot be used as substitutes for chemical fluxes. Vacuum atmospheres are suitable for brazing of copper and copper alloys when the alloys contain few or no elements having high vapor pressures at the brazing temperature (lead, zinc, etc.). The filler metal should also be restricted to vacuum-grade materials containing low contents of high-vapor-pressure elements, such as zinc and cadmium (Table 6.9). Example 8: Brazing of Refractory Metals. All of the refractory metals have been successfully brazed in inert, reducing, or vacuum atmospheres. The environment in which the refractory metals are brazed is determined by the reactivities of these metals with oxygen, hydrogen, and nitrogen as well as the effects of these elements on the mechanical properties of the refractory metals. All of the refractory metals react with oxygen at moderately elevated temperatures, but they force different types of oxides. Niobium and tantalum form hard, adherent oxides at temperatures above 205 and 400 °C (400 and 750 °F), respectively. On the other hand, molybdenum and tungsten form volatile oxides at temperatures above 400 and 510 °C (750 and 950 °F), respectively. In either case, the surfaces of the refractory metals must be protected during brazing to ensure wetting by the filler metal. Also, these metals must be coated with an oxidationresistant material if they are exposed in air at elevated temperatures. For such service conditions, the filler metal must be compatible with both the base metal and the coating.
Niobium and tantalum are embrittled by the presence of hydrogen at relatively low temperatures. In contrast, molybdenum and tungsten can be brazed in a hydrogen atmosphere. Niobium, molybdenum, and tantalum are embrittled by nitrogen at high temperatures; however, reactions with nitrogen begin at relatively low temperatures. For example, nitrogen is dissolved in molybdenum at temperatures as low as 595 °C (1100 °F), but severe embrittlement does not occur until the temperature exceeds approximately 1095 °C (2000 °F). Tantalum behaves in much the same manner. Tungsten can be brazed in an inert gas atmosphere (helium or argon), a reducing atmosphere (hydrogen), or vacuum. Two precautions should be observed for vacuum brazing. The vapor pressures of the compositional elements of the filler metals should be compatible with the soundness of the base metals, and the deposited filler metal should be evaluated. Purified dry hydrogen and inert gas (helium and argon) atmospheres have been found suitable for brazing molybdenum. For brazing pure molybdenum, the purity of the hydrogen atmosphere is not critical. A dewpoint of 26.7 °C (80.1 °F) can be tolerated in hydrogen when reducing molybdenum oxide. A low dewpoint (–46 °C, or –51 °F, or lower) at 1205 °C (2200 °F) is required for brazing of titanium-bearing molybdenum alloys. Vacuum furnaces have been used for brazing of molybdenum; however, precautions should be taken in selecting a proper filler metal that will not volatilize during the brazing cycle. Pressures of less than 0.13 Pa are desirable for brazing of molybdenum alloys containing titanium. Inert and vacuum atmospheres are satisfactory for brazing of titanium, zirconium, and beryllium. A vacuum atmosphere of 0.13 Pa or better is required for brazing of titanium, whereas the dewpoint of an argon or helium inert atmosphere should be –57 °C (–71 °F) or less to prevent discoloration of the titanium and produce satisfactory brazed joints at temperatures from 760 to 925 °C (1400 to 1700 °F). Fabricators of titanium brazements generally favor liquid argon from cryogenic storage vessels because of its inherent purity. Although information is limited on brazing of zirconium as compared with brazing of titanium, these two metals are very similar chemically. In general, the brazing techniques used are applicable to both systems.
30 24 20 65 62 ... 28 28 50 27 32 20 28 21 ... ... ... 20 50 62.5 65 6 15.5 0.05 99.95 25 ... ...
Cu
... ... ... ... 3 18 ... 0.5 ... ... ... ... ... ... ... ... ... ... ... ... ... ... 3 ... ... ... ... 0.06
Ni
10 ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ...
Sn
... ... 75 35 35 82 ... ... ... ... ... ... ... ... ... ... ... 80 50 37.5 35 94 81.5 ... ... 50 92 ...
Au
... ... ... ... ... ... ... ... ... 5 10 15 20 25 5 10 20 ... ... ... ... ... ... ... ... 24 8 65
Pd
0.001 0.001 ... ... ... ... 0.001 0.001 0.001 0.001 0.001 ... ... ... ... ... ... 0.001 ... ... ... ... ... 0.001 0.001 0.001 0.001 0.001
Zn
0.001 0.001 ... ... ... ... 0.001 0.001 0.001 0.001 0.001 ... ... ... ... ... ... 0.001 ... ... ... ... ... 0.001 0.001 0.001 0.001 0.001
Cd
Chemical composition, %
0.002 0.002 ... ... ... ... 0.002 0.002 0.002 0.002 0.002 ... ... ... ... ... ... 0.002 ... ... ... ... ... 0.002 0.002 0.002 0.002 0.002
Pb
0.002 0.002 ... ... ... ... 0.002 0.002 0.002 0.002 0.002 ... ... ... ... ... ... 0.002 ... ... ... ... ... 0.002 0.002 0.002 0.002 0.002
P
0.005 0.005 ... ... ... ... 0.005 0.005 0.005 0.005 0.005 ... ... ... ... ... ... 0.005 ... ... ... ... ... 0.005 0.005 0.005 0.005 0.005
C
... 14.5 ln ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... 0.06 Co ... 35 Co
Other
600 625 885 1000 975 950 780 754 780 805 825 852 880 900 970 1000 1070 890 950 990 1000 965 900 961 1083 1102 1200 1230
°C
1115 1155 1625 1832 1787 1742 1435 1390 1436 1481 1517 1566 1616 1652 1778 1832 1958 1634 1742 1814 1832 1769 1652 1761 1981 2015 2190 2245
°F
Solidus (melting point)
720 705 895 1020 1030 950 780 795 855 810 852 900 900 950 1010 1065 1175 890 975 1015 1020 990 910 961 1083 1121 1240 1235
°C
1325 1300 1640 1865 1885 1740 1435 1465 1571 1490 1566 1652 1652 1742 1850 1949 2147 1634 1787 1859 1868 1814 1670 1761 1981 2050 2265 2255
°F
Liquidus (flow point)
BVAg-18(b) BVAg-29(b) ... ... ... BVAu-4 BVAg-8(b) BVAg-8b(b) BVAg-6b(b) BVAg-30(b) BVAg-31 ... ... BVAg-32 ... ... ... BVAu-2 ... ... ... ... ... BVAg-0 BVCu-lx BVAu-7 BVAu-8 BVPd-1
Class(a)
(a) All vacuum-grade filler metals for which American Welding Society (AWS) classifications are given are considered grade 1 filler metals (i.e., BVAg-18, grade 1). A grade 2 variety is also available for each vacuum-grade filler metal with an AWS classification. Each grade 2 filler metal has the exact same composition as the corresponding grade 1 variety, except that its zinc and cadmium contents are 0.002%. Grade 2 varieties also have 0.002% P. (b) Contain 0.02% P. Source: Ref 1, 2
60 61.5 5 ... ... ... 72 71.5 50 68 58 65 52 54 95 90 80 ... ... ... ... ... ... 99.95 0.05 ... ... ...
Ag
Table 6.9 Vacuum-grade filler metals
Chapter 6: Fluxes and Atmospheres / 285
286 / Brazing, Second Edition
Beryllium is most commonly brazed in purified argon, although helium and vacuums of less than 0.13 Pa are also suitable. Finally, where premetallized ceramics are to be brazed, the operation can be carried out in high-purity inert gas, hydrogen, or vacuum atmospheres. If the ceramic is to be brazed directly—that is, without being premetallized— a vacuum atmosphere is preferred. The current increased interest in ceramics (Ref 40–54) and cemented carbides (Ref 55) has created new and expanded work in dewpoint curves, gas mixtures, and reducing atmospheres to allow wetting by filler metals to take place. REFERENCES
1. G. Humpston and P.M. Jacobson, Principles of Soldering and Brazing, ASM International, 1993 2. M.M. Schwartz, Brazing, ASM International, 1987 3. Precision Atmospheric Brazing for HighSpeed, In-Line Operations, MAN, Dec 2000, p 66, 69 4. B.B. Bonner, D. Garg, and P.T. Kilhefner, Brazing in a Nitrogen-Based Atmosphere Offers Advantages, Weld. J., Oct 1993, p 31–33 5. B.B. Bonner, D. Garg, and P.T. Kilhefner, On-Site Nitrogen Generation for Effective Furnace Brazing Atmospheres, Ind. Heat., April 1993, p 58–59 6. D. Eichelberger, B.B. Bonner, and D. Garg, Novel Approach for Furnace Brazing Low-to-Medium-Carbon Steel Components, Heat Treat. Met., Vol 1, 1994, p 15–17 7. T. Hattori, S. Sakai, A. Sakamoto, and C. Fujiwara, Brazeability of Aluminum in Vacuum-Nitrogen Partial-Pressure Atmosphere Brazing, Weld. J., Oct 1994, p 233–240 8. R. Peaslee, Effect of Nitrogen on Furnace Brazed Joint Quality, Ind. Heat., April 1997, p 45–47 9. D. Kay, Weld. Des. Fabr., April 1993, p 35–36 10. W.R. Jones, Vacuum—Another Atmosphere?, Heat Treat., Oct 1986, p 39–41 11. W.R. Jones and R.J. Fradette, Brazing of Electric Power Interrupters in Vacuum Furnace Specially Designed for the Process, Ind. Heat., Aug 1994, p 32–33
12. Vacu-Braze, a Small Company with a Big Future, Ind. Heat., Jan 1994, p 14 13. A.B. Craven, Energy Dominates in Job Costing, Ind. Heat., Feb 1996, p 33–35 14. M.L. Santella and J.J. Pak, Brazing Titanium-Vapor-Coated Zirconia, Weld. J., April 1993, p 165–172 15. W. Hoke III and C. Amenheuser, Techniques for Aluminum Brazing in Vacuum Furnaces, Weld. J., Oct 1993, p 65–67 16. L.L. Ashburn, Fluxless Vacuum Furnace Brazing of Aluminum Particularly Advantageous for More Critical Applications: Part II, Ind. Heat., April 1994, p 43–46 17. A.A. Suslov, Fluxless Welding Sections of Aluminium Alloys in Vacuum, Weld. Int., Vol 9 (No. 5), 1995, p 406–407 18. A.T. Sibley, R.K. Ahuja, and D.M. Buck, Choosing a Nitrogen-Based Atmosphere, Met. Powder Rep., Vol 36 (No. 1), 1981 19. Brazing Handbook, 4th ed., AWS C3 Committee on Brazing and Soldering, American Welding Society, 1991 20. A. Sakamoto, Wetting in Vacuum-Inert Gas Partial Pressure Atmosphere Brazing, Weld. J., Vol 64 (No. 10), 1983, p 272–281 21. Nitrogen Furnace Atmosphere Particularly Favorable in Commercial Brazing of Stainless Steel Tubing, Ind. Heat., Aug 1982, p 46–47 22. T.S. Bannos, The Effect of Atmosphere Composition on Braze Flow, Air Products and Chemicals, Inc., 1983 23. W. Whitman, T.A. Nelson, and J. Solomon, Furnace Brazing with Nitrogen Atmospheres, Weld. J., Vol 61 (No. 10), 1980, p 21–25 24. “Safety in Welding and Cutting,” American National Standard Z49.1, American Welding Society 25. W.T. Hooven, Vacuum Furnace Brazing of Missile Guidance Component, Ind. Heat., Aug 1982, p 43–44 26. H.E. Pattee, “High-Temperature Brazing,” WRC Bulletin 187, Welding Research Council Sept 1973 27. G.M.A. Blanc, J. Colbus, and C.G. Keel, Notes on the Assessment of Filler Metals and Fluxes, Weld. J., Vol 42 (No. 5), 1961, p 210s–222s 28. “Research Studies of Brazing Fluxes,” Bulletin F-37, United Wire and Supply Corp., 1952
Chapter 6: Fluxes and Atmospheres / 287
29. Y. Baskin, Penton Publishing, 1992, Weld. Des. Fab., 4p 30. Y. Baskin, Penton Publishing, 1991 Weld. Des. Fab., 2p 31. Continuous Furnace Brazing of Aluminum Alloy Radiator Cores at State-ofthe-Art Facility, Ind. Heat., Dec 1993, p 50–51 32. G.H. Willett, Non-Corrosive Flux Brazing, Ind. Heat., Aug 1996, p 33–35 33. Aluminum Brazing Handbook, 4th ed., Aluminum Association, 1990, p 27–35 34. W.E. Cooke, T.E. Wright, and J.A. Hirschfield, “Furnace Brazing of Aluminum with a Non-Corrosive Flux,” SAE Technical Paper Series 780300, Society of Automotive Engineers, 1978 35. S.A. Urban, Manufacturers Capitalize on Brazing Flux Breakthrough, Weld. J., Sept 2000, p 37–39 36. Adv. Mater. Process., May 1993, p 14, 16 37. R.S. Timsit, Cost-Effective Brazing Technique for Aluminum Parts Eliminates the Use of Alloy Cladding, Mater. Technol., Vol 9 (No. 11–12), 1994, p 240–242 38. Classifying Brazing Fluxes, 554, AWS Brazing Handbook, American Welding Society, and AS 3-92 Specification for Fluxes for Brazing and Braze Welding 39. W.D. Kay, Postbraze Cleaning of Silver Brazed Joints, Weld. J., Vol 57 (No. 10), 1976, p 872–873 40. W.H. Chang, A Dew-Point/Temperature Diagram for the Metal/Metal Oxide Equilibria in Hydrogen Atmosphere, Weld. J., Vol 35 (No. 12), 1956, p 622–624 41. M.C. Rey, D.P. Kramer, et al., DewPoint/Temperature Curves for Selected Metal/Metal Oxide Systems in Hydrogen Atmosphere, Weld. J., Vol 65 (No. 5), 1984, 162–166 42. R.O. Williams, “Thermodynamics of Copper-Nickel Alloys Containing Aluminum, Silicon, Titanium and Chromium Relative to Their Use in Ceramic Brazing,” ORNL-6072, Contract DE-AC05840R21400, Metals and Ceramics Division, Oak Ridge National Laboratory, Nov 1984 43. M.G. Nicholas, and D.A. Mortimer, Ceramic/Metal Joining for Structural Applications, Mater. Sci. Technol., Vol 1 (No. 9), 1985, p 657–665 44. H.E. Pattee, “Joining Ceramics to Metals
45.
46.
47.
48. 49. 50.
51. 52.
53.
54.
55.
and Other Materials,” WRC Bulletin 178, Welding Research Council, 1972 H. Mizuhara and E. Huebel, Joining Ceramic to Metal with Ductile Active Filler Metal, Weld. J., Vol 65 (No. 10), 1984, p 43–51 M. Vilpas, “Joining of Ceramics for High-Temperature Applications,” NASA TT-20030, N87-29678, NTIS HC A03/ MF A01, National Aeronautics and Space Administration, Oct 1987; translation of “Korkeissa Lampotiloissa Kaytettavien Keraamisten Materiaalien Liittaminen,” Report UTT-TIED-481, Technical Research Center of Finland, Espoo, Aug 1985 J.P. Hammond, S.A. David, and M.L. Santella, Brazing Ceramic Oxides to Metals at Low Temperatures, Weld. J., Vol 67 (No. 10), 1988, p 227–232 H. Mizuhara, Vacuum Brazing Ceramics to Metals, Adv. Met. Process., Feb 1987 R.W. Rice, Joining of Ceramics, Advances in Joining Technology, Brookhill Publishing Co., 1976, p 69–111 R. Morrell, “Ceramics in Modern Engineering,” 0305-4624/84/050252, Institute of Physics, London, U.K., 1984, p 252–261 C.R. Weymueller, Braze Ceramics to Themselves and to Metals, Weld. Des. Fabr., Aug 1987, p 45–48 P.F. Becher and S.A. Halen, Joining of Si3N4 and SiC Ceramics via Solid-State Brazing, Proceedings of the DARPA/ NAVSEA Ceramics Gas Turbine Demo. Engine Prog. Rev. (Columbus, OH), MCIC, 1978, p 649–653 R.E. Loehman, Transient Liquid Phase Bonding of Silicon Nitride Ceramics, Surfaces and Interfaces in Ceramics and Ceramic-Metal Systems, Materials Science Research, Vol 14, Plenum Press, 1981, p 701–711 R.L. Tallman, R.M. Neilson, Jr., J.C. Mittl, et al., “Joining Silicon NitrideBased Ceramics: A Technical Assessment,” ECG-SCM-6572, DE84-011356, DOE Contract DE-AC07-761D01570, EG&G Idaho, Inc., March 1984 K.A. Thorsen, H. Fordsmand, and P.L. Praestgaard, An Explanation of Wettability Problems when Brazing Cemented Carbides, Weld. J., Vol 65 (No. 10), 1984, p 308–315
Brazing Second Edition Mel M. Schwartz, p289-311 DOI: 10.1361/brse2003p289
Copyright © 2003 ASM International® All rights reserved. www.asminternational.org
CHAPTER 7
Fixturing, Tooling, Stopoffs, Parting Agents, Surface Preparation, Surface Cleaning, and Repair THIS CHAPTER covers some of the most significant aspects of the various brazing processes. They include fixturing and tooling for the various brazing processes; stopoffs and parting agents, which are primarily liquid oxide coatings; the preparation and cleaning of material surfaces and braze filler materials; and finally, the need, at times, for the repair of an assembly/component/part.
Fixturing and Tooling Fixture design is one of the most important factors in ensuring high-quality brazed joints. Fixture design is important in determining the efficiency and productivity of the brazing system. A fixture is built to cradle, hold, or secure the assembly being joined. In assembling some components for brazing, the assembly may require additional positioning or support that cannot be provided by self-jigging alone, and so, the use of auxiliary fixtures is unavoidable. These fixtures can take the form of simple baskets or wire stands, machined graphite blocks, clamps, wrapped wire, differential thermal expansion components, or cast supports. Depending on the brazing process, the fixture material will vary. Low-carbon steel is commonly used for fixtures for short runs in furnace or torch brazing and sometimes in induction brazing. Steel has the advantage of low cost and the disadvantage of low strength at brazing temperatures. For long production runs, stainless
steels and wrought and cast heat-resisting alloys are used. Stainless steels and Inconel (Special Metals Corp.) are used for springs and clips in dip brazing fixtures. It is mandatory that these materials be used, due to the corrosive nature of the flux bath and its attack on low-carbon steel. Figure 7.1 shows a method of deadweight loading for vacuum furnace brazing that improves the quality of the joint. In the application for which it was developed—fabrication of stainless steel cold plates—the method decreased the rate of rejection from 57% to zero. Prior to this improvement, the plates to be brazed were pressed together under a one-piece glide plate. The glide plate became distorted from the furnace heating and cooling (temperature gradients) and therefore did not apply its weight uniformly over the surfaces of the plates to be joined. In the improved method, the plates were weighted with heavy stainless steel blocks. The blocks act independently and thus were immune to distortion. Besides being uniformly distributed, the force they applied to the plates was larger and was repeatable from one brazing operation to the next. Larger blocks were used on the edges of the plates, where more thermal mass was needed to reduce the differences between the temperatures at the interior and the edge as the temperature of the furnace was varied—a refinement that was not possible with the one-piece glide plate. The cold plate lay-up consisted of a stack of stainless steel sheets separated by filler-metal tape 0.1 mm (0.004 in.) thick, all on a stainless steel baseplate (Fig. 7.1). A stopoff coat of zir-
290 / Brazing, Second Edition
Blocks
Stainless-steel sheet 0.016 in. (0.41 mm thick)
Cold-plate doubler and its extension
Core
Brazing alloy
Stainless-steel sheet 0.012 in. (0.30 mm thick)
Cold-plate doubler and its extension
Baseplate Cross section expanded vertically to show layup Thermocouples Thermocouples
Baseplate
Radiation shield (one of four)
Cold-plate stack
Gap to vent doubler slot Vent holes Typical layup with weighting blocks
Fig. 7.1
Stainless steel blocks in two sizes weigh a stack of layers to be brazed together to make a cold plate. Radiation shields, two of which have been removed to show the stack, prevent overheating at the edges. Thermocouples at selected locations monitor brazing temperatures. Source: Ref 1
conium oxide powder was sprayed onto those surfaces that were in contact with each other but were not intended to be brazed together (e.g., between the lay-up and the baseplate), to ensure separation after brazing. The smaller blocks were placed on most of the top surface of the stack. The larger blocks were placed at the periphery of the top surface. The blocks were separated from each other by gaps 0.4 mm (0.016 in.) wide. Shields made of stainless steel sheet were placed at the edges of the stack to prevent overheating of the edges by direct radiation. The shields were tack welded in position. The assembly was brazed at a temperature between 1040 and 1085 °C (1900 and 1985 °F). Fixturing has always been considered to be just another standard part of each furnace load, regardless of the contents of the load. Since braze fabricators are always considering the economics of their production operation, they must always consider the importance of fixturing design (see Fig. 7.2 and 7.3).
Fixture Design Challenge A rule of thumb for the total weight of all fixtures should not exceed 50% of the total weight of the assemblies being brazed per furnace run (Ref 2).
Fig. 7.2
This vacuum furnace contains typical fixture baskets used for brazing or heat treating. The top (fourth) layer uses only lightweight wire mesh baskets to hold the parts. Although the lower three baskets are ideal for stacking, this kind of basket adds extra weight to the furnace run and consumes more energy. Photo courtesy of Vacuum Furnace Systems Corp., Souderton, PA. Source: Ref 2
Chapter 7: Fixturing, Tooling, Stopoffs, Parting Agents, Surface Preparation / 291
Considerations in Fixturing Although a number of different fixtures could be designed for any given assembly, there are specific considerations that determine the best type to use. Fixture design should adhere to the following principles: • Fixtures should allow easy insertion of assembly components and easy removal of the brazed assembly. • Fixtures should support assembly components to permit expansion and contraction during the heating and cooling cycles. External fixtures should expand more quickly and internal fixtures more slowly than the assembly. In applications where tight clamping is required, the reverse is true. • The assembly should be supported at points away from the heat zone to prevent the fixture from becoming a heat sink. • Fixtures should permit heat to be directed around the entire joint area, so that the heating pattern designed for the system is free to cause flow of the filler metal throughout the joint. • Gravity should be used to assist capillary action wherever possible. If the shape and weight of the part permits, the simplest way to hold them together is by gravity. Otherwise, additional weight can be added. If a number of assemblies are to be brazed and their configuration is too complex for self-support or clamping, a good idea is to use a brazing support fixture. In planning such a fixture, design it for
the lowest mass; the contact area between the fixture and the assembly should be at a minimum (Fig. 7.4). A fixture that contacts the area broadly will conduct heat away from the joint area. If the fixture is to be used in a torch application, clearance for the open flame to reach the joint area without restriction should be allowed. Materials should be chosen that are resilient to high temperature and thermal cycling, such as stainless steels, nickel-base superalloys, or ceramics. If a fixture is needed close to the joint, where there is risk of brazing the assembly to the fixture, a nonwetting material, such as titanium, should be used. Additional considerations in fixture design are: • Alignment and dimensional stability of assembly components should be maintained until the filler metal solidifies. • Fixtures should be sufficiently flexible to accommodate other similar assemblies, where possible. • Component pieces of the assembly should be self-locating, so that the fixture only supports and cradles the components to the degree necessary to achieve good results (Fig. 7.4). • Fixtures should be designed for minimum surface contact with the assembly; point or line contact is preferable to overall surface contact. Even with a minimum number of contact points between the fixture and the components to be brazed, it is sometimes difficult to prevent the fixture from being wetted by the filler metal and sticking to or being brazed to the assembly.
Fig. 7.3
These heavy outer baskets add extra weight to this furnace load. Lightweight separators between the layers of small parts would be sufficient. Photo courtesy of Vacuum Furnace Systems Corp., Souderton, PA. Source: Ref 2
Fig. 7.4
Self-locating aluminum alloy components fixtured prior to dip brazing. Source: Ref 3
292 / Brazing, Second Edition
• If the contact surface between the fixture and the components must be extensive, selection of a fixture material that resists wetting becomes critical. • Sections of fixtures should be as thin as possible and consistent with required rigidity and durability. • For efficient furnace brazing, the mass of the fixture should be held to a minimum; as the weight of the fixture is increased, the fuel efficiency of the furnace is reduced. • In induction brazing, fixtures should be located well away from the work coil so that they will not act as heat sinks or interfere with the magnetic field. • Fixtures should allow room for the induction coil to heat the joint area uniformly (Ref 4).
Designing Effective Fixtures In the design of fixtures, designers must recognize their use (Ref 2) whether (a) to hold parts (such as baskets or trays), (b) to hold individual parts in alignment during brazing, and (c) to act as dead weights during brazing. Baskets and Trays. Usually baskets are made from stainless steel rods welded together. These baskets must be durable and be capable of abuse through many different furnace cycles. Fixtures for Parts Alignment. Commonly used alignment fixturing, in order from most desirable to least desirable, is:
direct the flow of the filler metal into the joints to best advantage and to minimize distortion or movement of the parts. These points are generally easy to determine by cut-and-try methods. When a proper procedure is found, the filler metal can be made to flow into all joints, leaving neat fillets and clean surrounding surfaces, and the job can usually be done without distortion (Ref 6). Some of the various methods of holding assemblies together within the furnace are as follows. Laying Parts Together. Perhaps the simplest method for joining two parts is simply to lay one on top of the other, with filler metal placed either between the members or wrapped around one of the members near the joint. In this method, either the weight of the upper member must be sufficient to ensure good metal-tometal contact or a weight can be added to ensure such contact. This technique sometimes lacks the advantage of having a definite means of indexing or keeping the parts from moving in relationship to one another. Pressing Parts Together. The most common method for assembling parts for vacuum furnace brazing is simply to press them together. In general, regardless of the degree of tightness, some scheme is usually employed to prevent slippage of the parts when they become
• Self-fixturing • Self-fixturing with some mechanical assistance • Heavy external fixturing, using heavy-duty baskets Figure 7.5 shows before and after views of assemblies to illustrate the technique of how fixturing might be improved (Ref 5). By using materials property data and in selecting appropriate metals and ceramics for fixturing, one can be very unique and creative and take advantage of the expansion/contraction characteristics of materials (Ref 2, 5).
Specialized Fixturing for Vacuum Brazing In designing an assembly for vacuum furnace brazing, one must keep in mind how the assembly will be held together within the furnace and how it will be set up in the furnace, so as to
Fig. 7.5
Improvement in fixturing. (Before) Note heavy fixture plates on each side of brazed assembly, and the extra weights added on top. (After) Replace the heavy blocks and the two large fixture plates (top and bottom) with molybdenum bar stock.
Chapter 7: Fixturing, Tooling, Stopoffs, Parting Agents, Surface Preparation / 293
heated in the furnace, particularly if the joint has a vertical axis. Figure 7.6 shows a shoulder formed on one member to accomplish this stability. In pressing parts together, the usual tolerances used in machining the parts naturally result in variations in the amount of press fit, which cannot be avoided. An effort should be made, however, to have a snug fit at all times, if possible. Sometimes a heavy press fit causes distortion of the parts by stretching them beyond their elastic limit when hot; this results in weakening of joints and assembly. Spot Welding and Tack Welding. Spot welding is frequently employed for maintaining definite relationships between parts assembled for vacuum furnace brazing. It is a fast, inexpensive, and generally efficient operation. Auxiliary Fixtures. All of the foregoing methods of holding assemblies together are used for vacuum furnace brazing primarily; however, the methods are applicable to other brazing methods. The choice of any method depends on the characteristics of each individual product. In some instances, however, it is impractical to use any of the suggested methods, and it is then necessary to resort to auxiliary fixtures to properly locate the members with respect to one another during furnace brazing. These fixtures sometimes take the form of graphite blocks, ceramic assemblies, heat-resisting superalloy and/or refractory alloy supports, or clamps. Auxiliary fixtures have several disadvantages. They constitute additional mass that must be heated, are subject to warpage that might make them unsuitable for repeated operations, and present an extra item of maintenance ex-
Fig. 7.6
Vacuum-brazed ordnance projectiles. Source: Ref 6
pense. However, two examples where auxiliary fixtures have been used to advantage are as follows: • Blocks: This type of tooling, consisting of weighted blocks, has been successfully used to produce brazed refractory and diffusionwelded reactive-metal components. The tools are usually made of the same material as the parts being brazed, thus minimizing differential thermal expansion problems (Fig. 7.1). • Pellets: This new approach to tooling has resulted in a quantum step in removing the distortion problems associated with elevatedtemperature brazing and diffusion welding (at 870 to 2480 °C, or 1600 to 4500 °F) and produces a fluid-type pressure over the assembly to be joined. Thus, control is attained over the mass of the fixturing, the weight of the fixturing, and continuous use of the fixturing without thermal warpage. Additionally, this flexweight-tooling concept can be used in joining titanium alloys, steels, superalloys, and refractory metals. The pellets used have been tungsten, graphite, and alumina. The versatility of this technique allows flat, curved, and cylindrical panels to be brazed. It should be further noted that an alternative to pellets, in some cases, is the use of mesh materials in the form of netting or screen. Materials such as 0.10 mm (0.004 in.) molybdenum wrapped around a titanium honeycomb sandwich panel and a stainless steel tooling mandrel can also successfully produce 360° panels. The economic advantage of pellets over mesh is significant and usually governs the choice of fixturing method. Therefore, in selecting fixtures for assembling parts for vacuum brazing, the following factors should be considered: • The mass of the fixture should be kept to the minimum value that will adequately accomplish the intended purpose. The fixture should be designed to provide minimum interference with even heating of the parts by removing heat by conduction from the brazing area. It is also important that the fixture not hamper the flow of the filler metal. • Vacuum is one of the determining factors in the selection of the material to be used in the
294 / Brazing, Second Edition
fixture, and these materials must withstand the temperatures involved without being appreciably weakened, distorted, or vaporized. • Consideration should be given to the expansion and contraction of the fixture in relation to the part being brazed, to ensure a combination that will maintain proper joint clearance and alignment at the brazing temperature. Therefore, the coefficients of expansion of the fixture material and the parts should be considered. Example: Fixtures for Brazing of Copper Tubing. There was a need to join, without gaps, 17.1 m (56 ft) of 9.5 mm (0.37 in.) diameter coiled copper tubing to the inside diameter of a 61 cm (24 in.) diameter copper cylindrical plate, 9.5 mm (0.37 in.) thick, that functioned as a pipe/jacket around the tubing. The cylinder also had an irregular welded scarf seam with numerous gaps that needed to be filled (Ref 7.8). In order for the copper tube and pipe assembly to function properly, a continuous bond between the 9.5 mm (0.37 in.) copper tubing and the inside wall of the cylindrical plate was essential for an effective heat exchange. This specification was consistent with the vacuum brazing requirement for a minimal gap of 0.01
Fig. 7.7
to 0.13 mm (0.0004 to 0.005 in.) between components. This is essential so that the filler metal will diffuse into the two parent metals for a successful braze. Although the tubing was precoiled to a larger diameter than the inside diameter of the cylindrical wall, once placed into position, there was not enough spring in the soft copper tubing to keep it tight against the wall. In fact, the tubing would become softer as the temperature was elevated during the brazing process. Also, there were perforations through the cylinder wall positioned approximately every 51 mm (2 in.) around and down the plate. The coiled tubing needed to avoid blockage of these holes with filler metal. Because of these factors, fixturing became critical to achieve the proper gap for brazing the part assembly. To achieve the proper braze gap, small-diameter copper all-threads were tacked and positioned at four quadrants and nine steps down the inside wall (Fig. 7.7). Washers and nuts were used to pull in the tubing to the wall. After brazing, the all-thread fixturing could easily be removed. This still left a span of 45.75 cm (18 in.) of tubing between the all-thread fixtures that varied in distance from the wall from 0.03 to 1.76 mm (0.0012 to 0.07 in.). In order to
Copper tubing fixtured on the cylindrical wall for application of nickel filler metal. Source: Ref 7
Chapter 7: Fixturing, Tooling, Stopoffs, Parting Agents, Surface Preparation / 295
bridge the gap between the wall of the vessel and the tubing between the all-thread fixtures, 2.2 mm (0.09 in.) copper wire segments, 51 to 76 mm (2 to 3 in.) in length, were prepared and strategically placed in areas where the gap was larger than 0.13 mm (0.005 in.). On each end of the tubing coil were two fitted stainless steel adapters, connected using a more customary braze joint. To join the two copper components, American Welding Society/American National Standards Institute (AWS/ANSI) BNi-6 filler metal was chosen, because it would do an excellent job of wetting both the copper as well as the stainless fittings. The paste was applied to the top and bottom of the tubing but was dispensed heavier over the wire segment on top of the tubing coil. White stopoff was painted on at the bottom of the jacket, both inside and outside. The use of stopoff is a precaution used to protect against filler metal that might run down the wall with gravity while at the 980 °C (1800 °F) brazing temperature. Should runoff occur, the combination of the protective stopoff and the fact that the jacket is resting on silica cloth would help minimize, if not eliminate, the possibility of the filler metal accumulating and sticking to the bottom edge. The result showed that all of the braze contact, tube to wall, was continuous around the 17.1 m (56 ft) of copper tubing. Also achieved were a relatively moderate fillet on the top side of the tubing and a small, fine fillet underneath. The stainless steel adapters had both braze continuity as well as being leaktight.
Graphite Fixturing Some problems can occur with graphite fixtures. While graphite fixtures are excellent in retaining flatness and are readily machinable, the thermal expansion of graphite is much lower than most other materials being brazed, such as American Iron and Steel Institute (AISI) type 321 stainless steel. If a graphite fixture were to be used where the 321 stainless steel part was to be shouldered in a groove in the graphite fixture, the groove would have to be machined at least 7.6 mm (0.30 in.) larger than the flange (at room temperature) to allow for the larger growth of the AISI 321 part flange at the brazing temperature. There is another important caution regarding graphite fixtures. Graphite fixtures should not be used in gas atmospheres containing hydro-
gen when brazing or processing stainless steels. Carbon monoxide and methane can be formed, and these gases will carburize the low-carbon stainless steels, thus making the stainless steel susceptible to corrosion. In atmospheres of argon, nitrogen, or vacuum, this carbon transfer does not occur. In all cases, however, the graphite fixture must not come in direct contact with the metal part. There are certain other requirements for brazing fixtures that are peculiar to the vacuum brazing process. For brazing fixtures that will be used in vacuum, materials should be selected that will not expel gas or otherwise contaminate the inner-furnace environment. Whereas graphite and various steels and superalloy heat-resisting materials are satisfactory at brazing temperatures up to 1095 °C (2000 °F), the tooling materials for use at 1650 °C (3000 °F) are limited. Refractory metals and their alloys, ceramics, graphite, newly developed carbides and borides, and refractory-coated graphite are the only materials available for fixturing at such temperatures. Of these materials, ceramics such as alumina, zirconia, and beryllia have exhibited reactions with refractory metals, contamination, and problems of thermal expansion and contraction. Graphite is another unsatisfactory material. It has excellent dimensional stability, but it embrittles the refractory metals that are usually used for the heating elements of the furnace. The newly developed carbides and borides (ZrB, ZrC, TiC) are still in limited use. This leaves only the refractory metals and their alloys.
Stopoff Materials and Parting Agents Frequently, it is necessary to prevent the filler metal from wetting portions of assemblies, fixtures, and metallic supports. The materials customarily used for this purpose are refractory oxides such as levigated alumina, magnesium oxide, magnesium hydroxide, and titanium dioxide, which are used as extremely fine powders suspended in alcohol, lacquer, acryloid cement, water, or acetone. Two types of stopoff mixtures are available. One is fast drying and behaves much like a commercial lacquer. The second is a nonwicking type, composed of oxides in a gelled vehicle that does not settle out on standing; this type dries slowly. Slurries usually are brushed on with a paintbrush or roller. The use of an artist’s brush is ideal for precision
296 / Brazing, Second Edition
applications in fine areas, although this operation is time-consuming and requires considerable skill. Two developments in stopoff use have recently been reported (Ref 1, 3). Ceramic stopoff materials, such as zirconium oxide, aluminum oxide, and boron nitride, are used to protect areas from unintentional brazing. However, if applied directly to a metal surface, these materials tend to spall during the brazing process. The problem is easily corrected by using a nickel alloy precoat (e.g., Nichrome, Harrison Alloys Inc.; or Inconel X, Special Metals Corp., or the equivalent) before the ceramic is applied. The precoat improves adhesion of the ceramic and makes the ceramic more tolerant to deformation. The precoat is applied in a plating process as a film and is approximately 0.0001 to 0.001 mm (0.000004 to 0.00004 in.) thick. The stopoff coating does not interfere with the brazing process and, when used in a vacuum, does not contaminate the pumping system. This process has been successfully used in brazing of tubes inside the Space Shuttle main engine nozzle. In the second development, thin sheets of alumina-enriched paper prevent the workpiece from becoming attached to the tooling in brazing operations. Used in fluxless vacuum brazing of stainless steel parts, the paper acts as a barrier that prevents bonding of the filler metal to the tooling (Fig. 7.8). Because of the high chemical stability of alumina, the paper does not react with the parts or the tooling, even at the high temperatures and pressures required for brazing. The alumina barrier is especially useful in brazing of parts with perforated or otherwise irregular surfaces, because it prevents the filler metal from extruding through the perforations and contacting the platens of the brazing press. Unlike other common barrier materials (various powders and solids), the alumina paper does not disintegrate in the press. Because the paper does
Fig. 7.8
not outgas, there is no contamination of the filler metal, thus ensuring uniform, reliable joints without voids. In practice, flow may not stop when the joint is filled, and filler metal may flow onto areas where it is not wanted. For example: • In brazing of a threaded stud into a part, the filler metal is likely to follow the threads and render them out of tolerance. • Support points on fixtures used in furnace brazing may become wetted by the filler metal, producing an unwanted braze and perhaps resulting in loss of the fixture and assembly, because it may be impossible to separate them without damage. • Some parts, such as turbine and compressor brazements, are designed to close tolerances, and excess filler metal may be dimensionally objectionable. • Tubular assemblies, particularly small capillary tubes (≤1.57 mm, or 0.062 in., inside diameter), can easily become partly or completely blocked with filler metal. • Excess filler metal may be unacceptable because of appearance. • In production brazing, where it may be necessary to use more filler metal than called for to allow for variations in fit-up between parts, some joints will have excess filler metal, which will flow away from the joint area. Although stopoff coatings can prevent wetting and flow of filler metals on portions of assemblies that contact the coated surfaces, they will not necessarily form barriers against creep of filler metal on the assemblies, and filler metals often creep beneath the coatings. This problem becomes more acute in vacuum brazing. Because many stopoff materials are oxides, the vacuum removes them, leaving only a slight residue that is ineffective as a stopoff material.
Sheets of alumina paper placed between the parts to be brazed and the heating platens of the press. Source: Ref 1, 3
Chapter 7: Fixturing, Tooling, Stopoffs, Parting Agents, Surface Preparation / 297
When zirconium oxide is mixed with a suitable nitrocellulose lacquer, it produces an effective parting agent for brazing in vacuum above 870 °C (1600 °F). Therefore, zirconium oxide is a suitable material to prevent flow of braze filler metal where not wanted in a vacuum environment. In certain types of work, it is necessary to confine the flow of filler metal to definite areas. This may sometimes be accomplished by controlling the amount of filler metal used and its placement in the assembly. Stopoff Materials. For brazing of carbon and low-alloy steels in the more commonly used atmospheres such as exothermic-based atmospheres, milk of magnesia painted on the appropriate areas is an effective stopoff. Also, painting of fixtures with a water solution of chromic acid and then heating them to the brazing temperature renders them resistant to wetting by the filler metal, because a thin layer of chromium oxide forms. For brazing in a hydrogen atmosphere or in a vacuum, commercial materials are used that are composed of graphite or oxides of aluminum, titanium, and magnesium prepared in the form of a water slurry or organic binder mixture. Recently, a boron nitride additive was added to various stopoff formulations for brazing applications above 1200 °C (2200 °F). Application. For large areas, the use of a brush or roller is satisfactory. This method is used to protect touchpoints of metal fixtures. Often, it is advisable to repaint the fixture prior to each use, because portions of some stopoff materials crack off during each heating cycle. Use of a medical syringe makes it possible to obtain extremely fine detail in stopoff application. Needles with an inside diameter of 0.25 mm (0.01 in.) are often used. With a small needle, a drop of stopoff can be applied at a precise point. With fast-drying stopoff, there is always a danger that some of it will inadvertently run into the joint area. If this happens, the assembly must be taken apart and all unwanted stopoff removed. A nonwicking stopoff can be applied by conventional equipment designed for application of liquid plastics, paste-type filler metals, and other organic compounds. It remains stable over long periods of time and does not clog the valves and tubing in the system. Removal. Wire brushing, air blowing, or water flushing can remove brazing stopoff mate-
rials of the parting compound type. The surface reaction type can best be removed by a hot nitric acid/hydrofluoric acid pickle, except when the brazed assemblies contain copper or silver, which will be attacked. Solutions of sodium hydroxide or ammonium bifluoride can be used in all applications, including copper and silver. Other stopoff materials can be removed by dipping in a 5 to 10% solution of either nitric or hydrofluoric acid.
Surface Cleaning and Preparation Cleaning of all surfaces that are involved in the formation of the desired brazed joint is necessary to achieve successful and repeatable braze joining. All obstructions to wetting, flow, and diffusivity of the molten filler metal must be removed from both surfaces to be brazed prior to assembly. The presence of contaminants on one or both surfaces may result in formation of voids, restriction or misdirection of filler-metal flow, and inclusion of contaminants within the solidified brazed area, all of which reduce the mechanical properties of the resulting brazed joint. Grease, oil, dirt, residual fluorescent penetrant fluids, pigmented markings, residual casting or coring materials, and oxides prevent the uniform flow and bonding of the filler metal, and they impair fluxing action, resulting in voids and inclusions. With the refractory oxides or critical atmosphere brazing applications, precleaning must be more thorough, and the cleaned components must be preserved and protected from contamination. In many applications, precleaning and cleaning are considered as the same step, whereas in some most-critical applications of brazing, precleaning is performed, and then fitup takes place; the parts are stored in a clean room, and subsequently, another cleaning takes place in the future, before use. Therefore, precleaning must generally use a clean room for the final handling and assembly of the cleaned parts. Although many fluxes have some cleaning effect, this is not the primary reason for their use, and complete reliance should not be placed on them for this function. Fluxes are used primarily to prevent the formation of oxides during brazing, to reduce the surface tension of the filler metal, and to form a protective covering of slag over the solidifying brazed joint. Their effectiveness in removing existing oxides is only slight and incidental. The type of contaminant that must be re-
298 / Brazing, Second Edition
moved in the cleaning process should be identified. The five categories of lubricants are: • • • • •
Mineral oils Emulsions Semisynthetics Synthetics Natural oils
Mineral oils are petroleum distillates that may be compounded with various additives. Emulsions consist of mineral oils, additives, and surfactants. They are diluted with water for use and will typically contain 5 to 10% mineral oil in this form. This is a large emulsion and will appear as a white, opaque (milky) solution. Semisynthetics start as a concentrated oil-inwater emulsion. They are further diluted with water prior to use to form a stable microemulsion that appears as a gray, translucent solution, although some contain dyes. Synthetics contain no oil and are typically polyglycol or polyisobutylene based. They contain additives and emulsifiers and appear as transparent solutions. Natural oils can be lard- or fat-based lubricants. In addition to creating foaming problems in spray washers, natural oils are generally not recommended, because they are not very effective lubricants (Ref 9). Precleaning. It is important to realize that most metal surfaces actually consist of a thin layer of metal oxide crystals formed by reaction of the metal with oxygen in the air. It is only after penetration of the metal oxide layer that atoms of the metal itself are encountered. In addition to being covered with oxide crystals, metal surfaces are also characterized by the existence of absorbed moisture, oxygen, and possibly other gases from the atmosphere. They also frequently are coated with various amounts of oil, grease, wax, perspiration, die lubricants, and mill scale. Mill scale is a combination of salts and oils that form on the surface during the process of making the metal. Many of these substances are loosely held and therefore make a poor substrate for bonding. As a result, a large portion of the effort expended in surface preparation prior to bond formation is devoted to removing these materials. The length of time that cleaning remains effective depends on the metals involved, the atmospheric conditions, the amount of handling the parts may receive, the manner of storage,
and similar factors. It is recommended that brazing be done as soon as possible after the parts have been cleaned. Maintaining part flow (first in, first out) is essential, particularly for atmosphere or vacuum brazing processes. Protection from contamination during storage is equally important. Maximum storage time and conditions must be determined for each product. Solvent wiping, degreasing, hot alkaline cleaning, and physical abrasion methods, or combinations thereof, are all intended to remove the loosely held contaminating materials and to expose a chemically inert and physically strong surface suitable for use as a base for adhesion. The key to success of brazing operations is the removal from the surface of extraneous substances by formation of relatively pure and nonreactive structures (suitable for wetting by braze filler metals). Cleaning is commonly divided into two categories: chemical and mechanical. Both chemical and mechanical cleaning methods are used to clean metal components for brazing, but chemical methods are the more widely used. Chemical cleaning methods vary from simple manual immersion to complex multistage operations. Chemical methods include alkaline cleaning, solvent cleaning, vapor degreasing, and acid pickling. The mechanical methods most commonly used are dry and wet abrasive blast cleaning. If warranted, machining or grinding may be used to obtain the necessary joint cleanliness and to ensure satisfactory wetting by the filler metal (Ref 10). Degreasing is generally done first. The following degreasing methods are commonly used, and their action may be enhanced by mechanical agitation or by applying ultrasonic vibrations to the bath: • Solvent cleaning: soak or spray operation with petroleum or chlorinated solvents • Vapor degreasing: chlorinated and trichlorotrifluoroethane solvents that clean by soaking and condensation of the hot vapor on the work • Alkaline cleaning: commercial mixtures of silicates, phosphates, carbonates, wetting agents, and, in some cases, hydroxides • Emulsion cleaning: mixtures of water, hydrocarbons, fatty acids, and wetting agents • Electrolytic cleaning: anodic, cathodic, and periodic reversal (Table 7.1)
Chapter 7: Fixturing, Tooling, Stopoffs, Parting Agents, Surface Preparation / 299
Chemical Cleaning Methods. Alkaline cleaning methods, including soak, spray, and barrel cleaning, are widely used for removing oily, semisolid, or solid soils from metal components before brazing. They employ commercial mixtures of silicates, phosphates, carbonates, detergents, soaps, wetting agents, and, in some cases, hydroxides. They are generally satisfactory for removing most cutting and grinding fluids, grinding and polishing abrasives, and some pigmented drawing compounds. Solvent cleaning is capable of removing oil, grease, loose metal chips, and other contaminants from metal components. Parts are immersed and soaked in a petroleum solvent or chlorinated hydrocarbon. Spray methods can also be employed. Vapor degreasing, a cleaning process usually performed first, uses stabilized trichloroethylene or stabilized perchloroethylene. To supplement the cleaning action of the vapor, some
degreasing units are equipped with facilities for immersing the work in the hot solvent or for spraying it with clean solvent. Scale and oxide removal can be accomplished mechanically or chemically. Prior degreasing allows intimate contact of the pickling solution with the parts, and vibration aids in descaling with any of the following solutions: • Acid cleaning: phosphate-type acid cleaners • Acid pickling: sulfuric, nitric, and hydrochloric acid • Salt bath pickling: electrolytic and nonelectrolytic The selection of chemical cleaning agent will depend on the nature of the contaminant, the base metal, the surface condition, and the joint design. For example, base metals containing copper and silver should not be pickled with nitric acid. In all cases, the chemical residue
Table 7.1 Prebraze cleaning procedures for selected base metals Mechanical cleaning Base metal
Degreasing agent
Wire brush
Grind
Blast
Sand
Chemical pickling mediums and methods
Phosphate-type, sulfuric, nitric, or hydrochloric acids or salt bath Same as low-carbon steel Acid: Dip in 10% nitric and 0.25% hydrofluoric acids for up to 5 min. Hot or cold water rinse. Dry Caustic: Dip in 5% sodium hydroxide at 60 °C (140 °F) for 60 s. Cold water rinse. Dip in 50% nitric acid for 10 s. Water rinse. Dry Copper: Dip in cold 5 to 15% sulfuric acid
Low-carbon steel
Any(a)
✓
✓
✓
✓
Stainless steel Aluminum alloy
Any(a) Solvents
... ...
✓ ...
✓ ...
✓ ...
Copper alloy
Solvents, alkaline cleaners
✓
...
...
✓
Nickel
Any(a)
...
✓
...
✓
Aluminum-bronze: Dip in cold solution of 3% sulfuric and 2% hydrofluoric acid. Follow with dip in 5% sulfuric at 27 to 49 °C (80 to 120 °F) Copper-silicon: Dip in 5% sulfuric at 27 to 49 °C (80 to 120 °F). Follow with dip in 5% sulfuric plus 2% hydrofluoric solution Brass: Dip in cold 5% sulfuric Nickel and nickel-aluminum: Dip in nitric acid, sulfuric acid, sodium chloride solution at 21 to 38 °C (70 to 100 °F). Follow with 1 to 2% sodium hydroxide dip Nickel-copper and nickel-copper-aluminum: Dip in 21 to 38 °C (70 to 100 °F) solution of nitric acid and sodium chloride. Follow with dip in 21 to 38 °C (70 to 100 °F) solution of nitric acid. Neutralize in 1 to 2% sodium hydroxide solution Nickel-chromium-iron: Dip in 49 to 60 °C (120 to 140 °F) solution of nitric and hydrochloric acid. Neutralize in 1 to 2% sodium hydroxide solution
(a) Vapor degrease with trichloroethane or perchloroethylene vapors; wash with petroleum or chlorinated hydrocarbon solvents, alkaline cleaners, detergents, or emulsifiers
300 / Brazing, Second Edition
must be removed by thorough rinsing to prevent formation of other equally undesirable films on the joint surfaces or subsequent chemical attack of the base metal. Mechanical Cleaning. Mechanical methods are less widely used than chemical methods in cleaning for brazing. However, they are usually preferred for removing heavy scale and may be indispensable in removing the more tenacious lubricants, such as pigmented drawing compounds. Mechanical abrading or roughening may be required on very smooth surfaces to promote filler-metal wetting and flow. Chemical cleaning also roughens the mating surfaces to enhance capillary flow and wetting by the filler metal. Mechanical cleaning methods, such as grinding, filing, machining, blasting, and wire brushing, also are used to remove objectionable surface conditions and roughen faying surfaces in preparation for brazing. If a power-driven wire wheel is used, care should be exercised to prevent burnishing. Burnishing can result in surface oxide embedment, which interferes with the proper wetting of the base metal by the filler metal. When rolling, fine grinding, or lapping has produced a base-metal surface that is too smooth, the filler metal may not effectively wet the faying surfaces. In this case, the parts can be roughened slightly by rubbing with 30- to 40grit emery cloth for improved wetting. Provision should be made to adjust and control joint clearances after machining or blasting operations. Cutting oils used in machining must be removed prior to brazing. When faying surfaces of parts to be brazed are prepared by blasting techniques, there are several factors that should be understood and considered. The purpose of blasting parts to be brazed is to remove oxide films and to roughen the mating surfaces, so that capillary attraction of the filler metal will be increased. The blasting media must be clean and must not leave a deposit on the surfaces to be joined that will restrict filler-metal flow or impair brazing. The materials should be fragmented rather than spherical, so that the blasted parts are lightly roughened rather than peened. The operation should be done in such a way that delicate parts are not distorted or otherwise harmed. Recommended and nonrecommended blasting materials and methods, along with their advantages and disadvantages for use in preparing surfaces for brazing, are as follows:
Recommended blasting materials
Chilled cast iron and hardened-steel fragmented shot are recommended, because the roughening they produce and the residual iron they leave on the faying surfaces both promote the flow of filler metal. However, the residual iron is also a disadvantage, because it may rust on standing. Also recommended are stainless steel grits and powders, modified nickel-base filler-metal grits, and glass beads. The advantages of these materials are that they promote fillermetal flow on stainless steel surfaces and leave nonrusting residues. Their disadvantage is that they do not clean or roughen the surfaces as readily as do fragmented iron and steel shot. Nonrecommended Nonmetallic materials, such as alumina, blasting materials zirconia, silica, silicon carbide, and other and methods similar materials, are not recommended for prebraze preparation. Although excellent for cleaning and roughening, they can be undesirable, because they can become embedded in the surfaces and may retard filler-metal flow. Wet blasting methods, such as vapor blasting, generally are not recommended, because they contaminate the surfaces with minerals from water, rust inhibitors, and refractory oxides.
Thermal treatments are used to clean and modify surfaces by heating in furnaces with specific atmospheres that reduce oxides and remove objectionable contaminants. Examples of this practice are bright annealing carbon steel parts in a controlled atmosphere and precleaning stainless steel in a dry hydrogen or vacuum atmosphere. Vacuum furnace operations are also conducted to remove objectionable contaminants from small capillary spaces and cracks that are to be repaired by filling with filler metal. Precoating and Finishing. For some applications, parts to be brazed are precoated by electrodeposition, hot dip coating, flame spraying, and cladding methods. Precoatings and finishes frequently are used to ensure wetting and flow on base metals that contain constituents such as aluminum, titanium, or other additions that are difficult to wet. Precoatings also protect clean surfaces and prevent the formation of oxides on base metals in storage and during the heating process. When brazing dissimilar metals, precoatings can reduce the tendency of a filler metal to wet and flow preferentially on one of the base metals. Occasionally, precoatings are used on refractory metals and several superalloys to prevent the rapid diffusion of filler-metal constituents into the base metal and the subsequent formation of brittle intermetallic compounds.
Chapter 7: Fixturing, Tooling, Stopoffs, Parting Agents, Surface Preparation / 301
The selection of the precoating depends on the base metal, filler metal, and brazing technique. Electroplatings of copper on steel and low-stress nickel on stainless steels are coatings that often are used. Brush plating is one technique used to apply electroplates locally at the joint areas; conventional plating with masking is also used. Proper procedures must be followed to activate the surface prior to applying the electrodeposit. Electroless deposition (autocatalytic) of nickel should not be employed to precoat materials that are to be brazed at temperatures above 870 °C (1600 °F). The phosphorus content of the nickel plating provides a eutectic at 880 °C (1615 °F), and melting of the coating may interfere with wetting and flow. In some brazing operations, the basic metals are clad with filler metal. Examples of this practice include aluminum alloys clad with BA1-Si filler metal, copper clad with BCuP filler metal, and copper clad with BAg-3 for sandwich brazing of carbide cutting-tool tips. Recently, a firm was induction brazing nickelplated 17-4 PH stainless steel to a tungsten carbide L-shaped part with BAg-24 silver filler metal. In testing the brazed joint, the L-shaped end of the tungsten carbide was pushed off (Fig. 7.9). Unfortunately, the braze joint failed at the surface of the 17-4 PH base metal, leaving a dark surface with no apparent filler metal on that side of the joint. The rest of the part appeared to
Fig. 7.9
Schematic of induction-brazed dissimilar-metal joint
be all right, so what could be the problem with this area? Why the filler metal did not wet the 17-4 PH base metal, and why there was no evidence of silver filler metal on the 17-4 PH surface where failure occurred was baffling. Because the part was nickel plated, it is possible this was the cause of the problem. One method for testing nickel plating is to put the parts into a vacuum or pure dry hydrogen furnace and take them up to 980 °C (1800 °F) for 5 min, cool them down, and remove them from the furnace. On inspecting the parts, blisters on the plated surface indicate that the 17-4 PH had not been adequately cleaned prior to the plating. Blister spots indicate poor bond strength where the filler metal is brazed to the nickel. The braze strength has been compromised, because there is a poor bond between the nickel and the steel. Plating, being molecular particles laid one on top of the other, makes a structure of lamellar layers. However, when it is run in the furnace at 980 °C (1800 °F), the lamellar structure changes to a crystalline structure, and the bond strength is increased. Currently, parts are plated and then directly brazed. If the plated parts were set aside for a period of time before brazing, the bond strength of the plating may improve. It would be necessary to determine how long it would take to see a beneficial effect, whether it was a day, a week, or longer, due to subsequent postbraze aging of the base metal at room temperature. Another interesting point to consider is how much hydrogen is in the 17-4 PH from the plating operation, and does this have a bearing on the bond strength of the nickel plating to the 17-4 PH? Specialized Processes. Research recently examined and evaluated the use of autodissolution (dissolving surface impurities automatically) and ion bombardment as a procedure for surface cleaning steels and high-purity irons (Ref 11). The work demonstrated that ion bombardment was a more effective cleaning method than simple heating to 500 °C (930 °F) and was even more effective when a hot substrate was treated. It is therefore probable that surfaces ion bombarded at 700 °C (1290 °F) and immediately joined were much cleaner than the present analyses suggest, but in all the cases studied, steels appeared to be easier to clean than highpurity irons. The technique used in the research followed the form of ion bombardment using argon at 0.013 Pa (10–4 torr) and approximately
302 / Brazing, Second Edition
4 kV potential, followed by electron bombardment, and was very effective for high-vacuum diffusion welding.
Surface Preparation for Specific Base Metals For some applications, the parts to be brazed are precoated by electrodeposition, hot dip, or flame-spraying methods. Another technique uses electron beam evaporation and the deposition of films on the substrate surfaces, which greatly enhances liquid metal wetting, spreading, and brazing. Aluminum has been suitably used to wet and join beryllium. Precoatings are used to retain clean surfaces and prevent the formation of oxides on base metals that contain refractory alloy additions. For brazing of dissimilar metals, precoatings reduce the tendency of the filler metal to wet and flow preferentially on one of the base metals. Occasionally, precoatings are used on refractory metals to prevent the rapid diffusion of filler-metal constituents into the base metal and the subsequent formation of brittle intermetallic compounds. The use of nickel plating facilitates brazing of some alloys, but it may not always be feasible because of the manufacturing sequence, the size of the items involved, or the lack of suitable plating equipment. An alternative surface-preparation method is plasma spraying of a 0.03 to
0.04 mm (0.0012 to 0.0015 in.) layer of fine nickel powder or filler metal directly onto the surface. In some recently completed work, it was found that this technique could be used for applying filler metals on A-286, Inconel 713 and 718, MAR-M-246, René 41, and others. These base-metal alloys contain aluminum and titanium and are among the most difficult alloys to braze. Unlike wet methods (paste filler-metal techniques), which allow little control and do not work with some difficult-to-wet alloys, the plasma-sprayed deposit can withstand handling under manufacturing conditions. Moreover, the sprayed filler metal does not contaminate furnace atmospheres. Other techniques used to clean and prepare surfaces for brazing difficult-to-wet alloys include envelopment of parts in a chromium fluoride atmosphere or in a fluorocarbon gas atmosphere. The latter technique has been used extensively in preparing all types of brazeable γ nickel-base superalloys and has been successfully used in repair of crack damage in gas turbine engine components (Ref 12, 13). Finally, selection of the precoating method depends on the base metal, the filler metal, and the brazing technique (Fig. 7.10, 7.11, 7.12). An example of the use of a combination of surface-preparation methods is in braze repair of aircraft gas turbine nozzles (Ref 13). An activated diffusion healing (ADH) process (Fig. 7.10) for maintaining aircraft gas turbine engine
Diffusion zone Low-melting matrix
Airfoil cross section Superalloy substrate Diffusion zone
ADH-alloy powder
Fig. 7.10
Superalloy powder
Original surface of superalloy powder particle
In the activated diffusion healing (ADH) repair process, a superalloy powder is cast into the crack using an assist from a lower-melting-point braze filler metal. The starting slurry also contains a standard brazing binder. Shown here: crack prior to casting and, inset, after vacuum brazing. Note diffusion of braze filler metal into both powder and substrate. Source: Ref 13
Chapter 7: Fixturing, Tooling, Stopoffs, Parting Agents, Surface Preparation / 303
parts was developed for General Electric aircraft engines. The first stage in the ADH process is a four-step cleaning operation that prepares parts for brazing:
4. A vacuum-cleaning cycle extracts any residual fluoride ions, resulting in an extremely clean part ready for the ADH process.
1. An alkaline solution removes oxides and soils. 2. An acidic chemical solution strips parts of their protective aluminide coating. 3. Fluoride-ion cleaning removes any remaining complex oxides, including those embedded in the smallest cracks.
The filler metal typically contains chromium, aluminum, tantalum, and cobalt and a 2.4% addition of boron that serves as a melting-point depressant. Vacuum brazing typically takes approximately 30 min at approximately 1200 °C (2200 °F) (Ref 16). (See the section “Repair Techniques with Cleaning Agents” in this chapter.) The following section covers cleaning procedures and compositions of cleaning solutions for some of the numerous brazeable metals and alloys. Many proprietary cleaning solutions are equally satisfactory.
Exhaust gas Inlet gas port
Scrubber
Weld fabricated stainless steel retort
NaOH solution O-ring
H2 CH4
Water cooling jacket Insulation Furnace Load Spreader can
H2, CH4, HF
Fluorocarbon
Aluminum and Aluminum Alloys In the case of aluminum, precleaning requirements vary, depending on surface condition, material thickness, the alloys to be brazed, and finally, the liquid- or gastightness required in the finished joint. Solvent-type cleaning operations for removal of surface lubricants have been quite satisfactory. Solvent cleaning is mandatory for fluxless brazing, particularly where a large surface area is exposed during brazing. Etchanttype cleaning may be done with a caustic or acid cleaner. Two cleaning procedures that are applicable to dip brazing and furnace flux brazing of aluminum alloys are as follows:
Fig. 7.11
Fluoride-ion cleaning process. Fluoride ion is produced by thermal decomposition of tetrafluoroethylene. Source: Ref 14
Caustic cleaning
• Degrease (solvent or vapor) • Dip in 5 wt% sodium hydroxide for up to 60 s at 60 °C (140 ° F) • Rinse in cold water • Dip in 50 vol% cold nitric acid for 10 s • Rinse in hot or cold water • Dry
Acid cleaning
• Degrease (solvent or vapor) • Dip in 10 vol% cold nitric acid plus 0.25 vol% hydrofluoric acid for up to 5 min • Rinse in hot or cold water • Dry
Heat
H2
MF + M + HF + H2 + H2O
H2 + MF + H2O
H2 Cr + 2NH4F CrF2 + N2 + 4H2 Heat
Fig. 7.12 Source: Ref 15
CrF2 pack fluoride-ion cleaning diagram. Fluoride ion is supplied by secondary chemical reaction.
In a new technique, aluminum is cleaned of its oxide film and is sealed immediately with a polymeric material, making it suitable for vacuum brazing. The time between cleaning and brazing is no longer a critical factor. First, the surface of the aluminum is degreased with any common degreaser, such as naphtha. After degreasing, the aluminum oxide is removed by chemical cleaning with an alkali wash of so-
304 / Brazing, Second Edition
dium hydroxide and sodium bicarbonate. A water rinse at 60 to 70 °C (140 to 160 °F) and an acid wash follow. After the acid treatment, the aluminum is rinsed in distilled water. It is then immersed in an all-organic solvent miscible with water, such as acetone, to remove all water from the surfaces. Immediately after this step, the clean surfaces are coated with a sealer. One of the best sealers comprises polystyrene in toluene and acetone. Sealed aluminum surfaces can be stored for several days without appreciable surface oxidation.
Beryllium Beryllium, like titanium, reacts with oxygen at conventional brazing temperatures. Beryllium also reacts with atmospheric nitrogen. Because the presence of an oxidized or nitrided surface impairs the wetting and flow properties of filler metals, high-temperature brazing usually is done in an argon atmosphere or in a vacuum after thorough acid cleaning of the basemetal surfaces. Fluxes have been used to prevent oxidation during low-temperature brazing of beryllium in air. Beryllium surfaces are also sometimes plated with silver to improve filler-metal wetting and flow.
Cast Iron Special preparation methods are available that are designed to improve wetting on the ascast brazing surfaces of cast iron. One such method is a proprietary electrochemical treatment that will remove graphite, silica, and other oxides. The process uses a molten salt bath operating at 460 to 480 °C (860 to 900 °F). The bath composition consists of 75% sodium hydroxide, 5% sodium chloride, 5% sodium fluoride, 14% sodium carbonate, and 1% potassium carbonate (Ref 17). Optimal cleaning is obtained by a 19 min direct-current electrolysis treatment (reduction for 4 min, oxidation for 10 min, and reduction for 5 min). Preparation is completed by rinsing in hot water to remove the salt, and then drying. Oncxe treated in this bath, machined surfaces rust rapidly. The rust film must be removed before brazing. Another choice is flame cleaning, which results in strong brazements when done properly. Flame cleaning with an oxyacetylene torch, however, calls for expertise in playing the oxidizing flame over the entire surface being brazed. Burning with a reducing flame, to
reduce surface iron oxide to elemental iron, is done next. Grit blasting of the faying surfaces may also be used. Finally, chemical treatment in fused sodium and potassium nitrate salts is not recommended, because it is a complex electrolytic treatment and decreases brazement strength.
Copper and Copper Alloys Standard solvent or alkaline degreasing procedures are suitable for cleaning of copper and copper-base metals of organic films, and mechanical methods (wire brushing, abrading, sanding, etc.) may be used to remove oxides. Complete chemical removal of oxides requires proper selection of the pickling solution. Typical procedures used for chemical cleaning are as follows: Copper Aluminum bronzes
Copper-silicon alloys
Brass and nickelsilver alloys Beryllium-copper
Chromium-copper
Copper-nickel
Immerse in cold 5 to 15 vol% sulfuric acid Successively immerse in a cold mixture of 2% hydrofluoric acid and 3% sulfuric acid and then in a solution of 5 vol% sulfuric acid at 27 to 49 °C (80 to 120 °F) and repeat until clean. Electroplating with copper at least 0.013 mm (0.0005 in.) thick on surfaces to be brazed will aid wetting. Immerse in hot 5 vol% sulfuric acid, then in a cold mixture of 2 vol% hydrofluoric acid and 5 vol% sulfuric acid Immerse in cold 5 vol% sulfuric acid Immerse in 20 vol% sulfuric acid at 70 to 80 °C (160 to 175 °F), water rinse, then quick dip (less than 30 s) in cold 30 vol% nitric acid solution followed by immediate and thorough rinsing Immerse in hot 5 vol% sulfuric acid, then in a cold mixture of 15 to 37 g/L sodium bichromate with 3 to 5 vol% sulfuric acid. Subsequent copper plating may facilitate wetting. Standard solvent or alkaline degreasing procedures are used to remove sulfur or lead from the surface, because these elements might cause cracking during the brazing cycle. Oxides are removed by abrading or by pickling in hot 5 vol% sulfuric acid, followed by immediate and thorough rinsing.
Magnesium and Magnesium Alloys These can be satisfactorily cleaned by the mechanical method of abrading with aluminum oxide cloth or steel wool. A chemical cleaning method that has been successful consists of dipping for 5 to 10 min in hot alkaline cleaner, followed by dipping for 2 min in ferric nitrate bright pickle solution.
Chapter 7: Fixturing, Tooling, Stopoffs, Parting Agents, Surface Preparation / 305
Nickel and Nickel Alloys Precleaning of nickel alloys just prior to brazing is particularly important, because they are subject to attack by low-melting-point elements, particularly lead and sulfur, at elevated temperatures. Because grease, oil, paint, shop dirt, and other foreign materials usually contain these harmful elements, they must be entirely removed before brazing. The oxide films formed on nickel alloys are tenacious, and wire brushing may not remove them; however, they may be removed with emery cloth or by grinding. Uniform oxide removal by pickling cannot be expected unless the high-nickel alloy being processed is first thoroughly cleaned of all foreign material. Additional methods of chemical cleaning to remove oxides and other adherent metallic contaminants include immersion in phosphate acid cleaners. Care must be taken in selecting time of exposure for both acid cleaning and pickling of heatresistant nickel-base alloys and superalloys. Overexposure during chemical cleaning can lead to excessive metal loss, grain-boundary attack, and selective phase-structure attack. As a last step in chemical cleaning, ultrasonic cleaning in alcohol or clean, hot water is recommended.
Refractory Metals Surface oxide removal is mandatory before brazing of refractory metals. The cleaning operation should be performed immediately before brazing to prevent contamination. Degreasing should be used to remove oil, fingerprints, and grease. Both mechanical and chemical cleaning methods are satisfactory. Molybdenum and Its Alloys. For removing heavy oxide films from molybdenum and molybdenum alloys, molten salt baths, such as 70% sodium hydroxide and 30% sodium nitrite at 260 to 370 °C (500 to 700 °F) or commercial martempering salt (mixture of sodium and potassium nitrates) at 370 °C (700 °F), have achieved good results. The former bath should be controlled closely, because it attacks molybdenum. Gross attack has not been noted with the latter bath. Electrolytic etchants can be used to remove surface oxides from simple parts; however, grain-boundary attack of molybdenum by such etchants can be severe. Chemical etching is the most popular cleaning method, and three successful techniques are given in Ref 1.
Tantalum and its alloys can be cleaned by both mechanical and chemical methods. Hot chromic acid (glass-cleaning solution) is quite satisfactory. Hot caustic cleaning solutions will attack the metal and should not be used. Prior to chromic acid cleaning, tantalum can be blast cleaned. This procedure, however, should be followed by an immersion in a hydrochloric acid solution to dissolve the iron particles. The glass-cleaning solution then becomes more effective. Abrasion and other usual mechanical cleaning methods have been found to be acceptable. Tantalum has a tenacious oxide film that reforms immediately on exposure to air or vapor after any cleaning treatment. Another method of preparing tantalum is described subsequently in the section “Niobium and Its Alloys” in this chapter. Niobium and Its Alloys. One method of preparing niobium (as well as tantalum) prior to brazing is to electroplate either copper or nickel onto an acid-cleaned surface. The deposits are bonded to the niobium (or tantalum) by diffusion, and, in the case of copper, melting actually occurs. Brazing is subsequently accomplished by using the plated surface as a base. Niobium can be cleaned by both mechanical and chemical methods (one chemical method is given in Ref 1). Tungsten and Its Alloys. Thorough cleaning prior to brazing is essential for tungsten, and both mechanical and chemical cleaning can accomplish this purpose. Cleaning methods that are acceptable for tungsten are given in Ref 1. The most effective cleaning procedure will depend on the tenacity of the oxide film. In cases where wrought tungsten sheet has been mill cleaned, degreasing is sometimes the only cleaning operation necessary prior to brazing. However, the optimal conditions for preparation of tungsten should be determined for each particular application. In some cases, electroplating of tungsten with nickel and other elements has been used satisfactorily to stop diffusion of elements that form brittle intermetallic compounds with the base metal. A hydrogen atmosphere furnace cleaning operation at 1065 °C (1950 °F) for 15 min has been effective in reducing light oxide films.
Low-Carbon and Stainless Steels For best results with low-carbon steels, faying surfaces should be cleaned mechanically or chemically to ensure essentially nearly complete absence of oxides and organic matter.
306 / Brazing, Second Edition
Stainless steels require more stringent precleaning than do carbon steels. The tenacious oxide films that impart corrosion resistance to stainless steels are more difficult to remove by fluxes or reducing atmospheres than the oxide films that form on carbon steels. Precleaning of stainless steels for brazing should include a degreasing operation to remove any grease or oil films. The joint surfaces should also be cleaned mechanically or with an acid pickling solution. Wire brushing, however, should be avoided, especially with a carbon steel wire brush. The best practice is to braze parts immediately after cleaning.
Carbides The most commonly used carbides are tungsten carbide particles in 10 to 15 vol% Co binder (hence the term cemented carbides). Much of the difficulty in brazing carbides is the result of improper cleaning. The carbide surfaces should be grit blasted or ground on a silicon carbide or diamond wheel to remove any surface carbon enrichment, because such surfaces are not readily wetted by the filler metal. The usual precaution of degreasing the surface prior to brazing also should be taken. Occasionally, some of the more difficult-to-wet carbides, such as titanium carbide, are coated with copper oxide or nickel oxide and then fired in a reducing atmosphere to fuse the copper or nickel onto the surface. This surface is readily wetted by the common filler metals.
Ceramics The inherent porosity associated with many ceramic bodies necessitates the use of very strict cleaning procedures before brazing. They are usually fired in air at a temperature from 800 to 1000 °C (1470 to 1830 °F) to permit outgassing. Suitable alkaline cleaning solutions may be used, followed by immersion in dilute nitric acid and subsequent rinsing in a neutralizing solution.
Powder Metals The use of powder metallurgical (P/M) products is increasing throughout industry. Existing and many new industries are taking advantage of P/M processing for new parts and for replacing existing parts made by other methods and materials. The same or better parts can often be produced at the same or lower cost, especially
when manufacturing complex parts in large quantities (Ref 18). Unfortunately, there are certain restrictions in form and size of the P/M parts due to the rigid P/M tooling and limitations to the pressing force available; for example, it is impossible to press undercuts, hollow grooves, and holes transverse to the pressing direction. However, there would be a much wider market if reliable and competitive joining processes, for example, brazing, were available to join P/M parts to bigger components and/or components with a high complexity. Brazing is recognized as a reliable and competitive process for the joining of many materials and components; however, when brazing P/M materials, it is often seen that much of the filler metal infiltrates the porous compacts instead of filling the joint. Because the strength and quality are drastically affected by the presence of voids, it is important to control and limit the infiltration of the compacts. The infiltration of a porous P/M component by a filler metal is closely related to a series of parameters, such as the size and amount of porosity and whether it is open or closed internally, as well as to the faying surface, the viscosity and amount of the melted filler metal, together with its ability to wet and to solidify isothermally by reactions with the P/M material. Researchers (Ref 19) conducted a study on P/M parts, because they felt that surface condition is a very important factor in attaining sound brazements. Powder metallurgy parts may be difficult to braze, because molten filler metal may infiltrate porous compacts, resulting in voids in the joint area. The influence of base-material composition and brazing parameters on infiltration and joint filling was studied on compacts of a pure iron powder and an alloyed powder (Fe-1.75Ni1.5Cu-0.5Mo-0.5C). The filler metal was AWS/ ANSI BNi-8. The effect of brazing temperature, filler-metal amount, and compact density was established on as-sintered compacts as well as on compacts that were surface modified by coining, lapping, and grit blasting. Coined and as-sintered compacts, with open pores, resulted in increased infiltration, but by using compacts with higher densities and a brazing temperature just above the liquidus temperature of the filler metal, the infiltration was reduced due to isothermal solidification. In this way, good-toacceptable joints were obtained.
Chapter 7: Fixturing, Tooling, Stopoffs, Parting Agents, Surface Preparation / 307
The researchers concluded that when one is going to braze compacts produced by a P/M route, the processing of the compacts is of the highest importance. The resulting surface quality has a great influence on the degree of infiltration by the filler metal and, therefore, the amount of filler metal left in the joint. Lapping and grit blasting closed surface pores, so no infiltration occurred, and brazing of the compacts was similar to brazing fully dense parts. When brazing as-sintered and coined compacts, close attention should be given to parameters such as brazing temperature, amount of filler metal, and compact density. By varying these parameters, it is possible to create joints between these porous parts. It is easier to control infiltration for higher-density compacts relative to lower-density compacts, thereby giving a more consistent joint quality. Coined compacts require approximately 50 µm of filler metal for the higher-density compacts (7.1 g/cm3, or 0.26 lb/in.3) and 100 µm of filler metal for the lower-density compacts (6.7 g/cm3, or 0.24 lb/in.3) (Ref 19, 20). As-sintered compacts required at least 100 µm of filler metal for the higher-density compacts (7.1 g/cm3) and 200 µm of filler metal for the lower-density compacts (6.7 g/cm3). It might be necessary to use even higher amounts of filler metal, because infiltration is very hard to control for these compacts. The composition of the investigated compacts did not have an effect when brazing with filler metal BNi-8. However, strong effects could be expected in other material and fillermetal combinations, especially where the effect of isothermal solidification is high (Ref 20).
from the rest of the shop and provided with means for controlling the atmosphere and cleanness within the area. In almost all instances, clean rooms are air conditioned. The objective is to reduce or eliminate airborne contamination. Workers in the clean room may be required to wear special lint-free clothing. Fresh clothing usually is provided daily. Special shoes also may be necessary. To avoid carrying outside contamination into the clean room, it may be advisable to provide dressing rooms between the shop area and the clean room, where street clothes are exchanged for clean room clothes. Workers in clean rooms may be required to wear hair coverings, but this is not necessary in every case. One of the most important functions of the clothing used in clean rooms is protection of the parts from contamination by handling with bare hands. Cleaned parts can become so contaminated by bare hands that the flow of filler metal is impaired or prevented even in a satisfactory atmosphere. Therefore, workers in a clean room should be required to wear gloves. White, lintfree cotton gloves have proven most satisfactory. These gloves are changed several times a day, depending on the amount of soil they accumulate. In some operations, the cotton gloves are supplemented by nylon gloves worn over them. Clean rooms have been widely used as assembly areas for brazed honeycomb assemblies and brazed stainless steel heat exchangers used in connection with jet engine fuel systems. They also are used as assembly areas for electronic components, such as vacuum tubes and solid-state devices.
Clean Rooms
Vacuum Brazing Cleaning
Within the past decade, more and more brazing has been performed in atmospheres without flux, and therefore, the cleanness of the parts is of great importance. Although chemical fluxes are capable of removing residual oxides that are generated during the brazing cycle, atmospheres are not. Precleaning, therefore, must be more thorough, and the components must be preserved and protected in the clean condition. One of the more common techniques is to use a clean room for final handling and assembling of the cleaned parts. Clean rooms are areas physically separated
A critical characteristic and benefit of the vacuum brazing process is its cleaning and brightening capability. These benefits are immediately evident after the parts are removed from the furnace. Stainless and carbon steel parts are brighter and cleaner, because oxides have been removed from the surfaces of brazed assemblies. The cosmetic improvement has particular importance for the instrument and medical device industries. During the vacuum brazing process, the part assembly simultaneously goes through a bakeout that removes oils and other contaminants
308 / Brazing, Second Edition
from the assembly—in some cases, better than chemical cleaning. This cleaning reaches machined part crevices and the internal dimensions of tubing and short capillary tubes. Productivity improves, because no flux or other contaminants need to be removed after the vacuum brazing process (Ref 21, 22).
Postbrazing Treatments Parts that are brazed in a suitable atmosphere should be bright and clean and should require no further processing. However, if flux or stopoff material is present, it should be thoroughly removed. If the brazement requires heat treatment, it must be done below the solidus or remelt temperature of the filler metal (Table 7.2).
Repair Techniques with Cleaning Agents Fluoride-ion cleaning, using hydrogen fluoride (HF) gas as the active agent, is rapidly becoming established as a cost-effective method of preparing nickel- and cobalt-base superalloys for braze repair in the aerospace industry. These alloys include René 80, IN-100, IN-738, MARM-509, B-1900+Hf, and MAR-M-200+Hf. The method, as described (Ref 23), can be practiced using relatively simple heat treat equipment, such as a hydrogen atmosphere retort in an air atmosphere furnace. Alternative fluoride-ion cleaning techniques have been developed (Ref 14) and patented (Ref
15). This process depends on the thermal decomposition of tetrafluoroethylene as the source of the HF gas and the direct use of CrF2 that produces less gas evolution than the chromium/ ammonium fluoride reaction, thus reducing the exhaust overload problem (Fig. 7.11, 7.12). Chemically, the HF gas cleaning process is identical to the previously described processes but has one significant difference: instead of obtaining HF gas through secondary reactions, a small quantity of HF gas is introduced directly into the reactor system through a precision electronic gas mass flow meter. Figure 7.13 is a representation of the system configuration. It therefore appears that the fluoride-ion cleaning technique, using HF gas, offers a simpler, more precise alternative to other available, more complex techniques for jet engine superalloy components.
Case Histories and Problem-Solving Examples Example 1: Surface Preparation of 304L Stainless Steel Plates. A firm was trying to braze two large 304L stainless steel plates
Gas inlet HF + H2 Main exhaust
Furnace
Table 7.2 Postbraze cleaning procedures for selected base metals Steels
Stainless steel
Aluminum
Copper
Flush with hot water, or quench still-hot brazements. Dip in 50% hydrochloric acid. Water rinse Dip in 20% hydrochloric plus 20% sulfuric solution at 77 to 82 °C (170 to 180 °F). Follow with dip in 10% nitric. Water rinse Immerse still-hot brazement in boiling water, and scrub with fiber brush. Finish with one of the following procedures: • Dip in nitric acid for 10 to 20 min. Follow with hot, then cold water rinse • Dip in nitric and sulfuric acid solution. Immerse in hot water for 10 to 15 min. Dry • Dip in phosphoric and chromium trioxide solution at 82 °C (180 °F) for 10 to 15 min. Water rinse Flush with hot water. Dip in solution of 10 to 25% hot sulfuric and 10% potassium dichromate
Superalloy turbine components
HF + H2
Fig. 7.13
CrF2 pack fluoride-ion cleaning diagram. The fluoride ion is supplied by secondary chemical reaction. Source: Ref 23
Chapter 7: Fixturing, Tooling, Stopoffs, Parting Agents, Surface Preparation / 309
together. The circular plates were ground to 30 to 80 µin. RMS and a 0.05 mm (0.002 in.) thick copper shim was placed between the plates and was used as the filler metal. The plates and foil were assembled in the furnace and loaded to approximately 5.358 g/cm3 with 15.24 cm (6 in.) of steel blocks. The part was run under vacuum of 0.13 to 1.3 Pa (2 × 10–5 to 2 × 10–4 psi) at up to 815 °C (1500 °F). At this temperature, nitrogen was added as a partial pressure to 1000 µm (1 torr, or 130 Pa), and the part was heated to 1120 °C (2050 °F) for copper brazing. It was held at the brazing temperature for 15 min, to ensure uniform heating, and the furnace cooled to 815 °C (1500 °F) and blower cooled from there. The part appeared to braze well, and the filler metal flowed. However, when the plates were ultrasonically inspected, it was found that there were unbrazed void areas. What caused the many voids, and why are there many voids? Large, flat plates are one of the more difficult brazements to make. In brazing large plates, one must consider the roughness and waviness as well as the adequacy of sufficient filler metal. Care must also be taken on both heating and cooling to prevent overstressing the part and causing distortion. Example 2: Removing Synthetic Oils to Prevent Rusting after Brazing of Stainless Steels. One current problem for the brazing industry is the effects of the new synthetic oils. These synthetic oils can be removed right after the machining or stamping operations; however, when the oil remains on the surface of the part for 8 h or longer, the oil cannot be removed by vapor degreasing. At the present time, the layer of oil is removed by a saturated solution of sodium hydroxide heated to a temperature of above 82 °C (180 °F). The parts have to stay in this solution a suitable length of time to allow the oils to be completely digested off the surface of the parts. Example 3: Troubleshooting Brazement Quality of Copper-Brazed 1018 Steel. A manufacturer of copper-brazed 1018 steel parts had a varied quality of parts. Was the surface of the parts contaminated, does the furnace atmosphere vary, or could something else be affecting braze quality? The manufacturer did not have a dewpoint indicator; therefore, how can one determine the cause of the problem? While there are many variables in the brazing operation, it is assumed that the parts were clean and free from scale, dirt, oils, heavy oxides, and other foreign materials. Another assumption is
that the furnace atmosphere was adequate to protect the part from oxidation and thus allow the copper filler metal to adequately wet and flow on the joint surfaces of the part. If the copper flows and spreads well on the clean surface but does not flow well or balls up on the uncleaned area, the footprints would indicate the surface was contaminated. Example 4: Use of Graphite Fixturing in a Hydrogen Atmosphere. Graphite fixtures and support plates in vacuum furnaces have been successfully used. However, when brazing in hydrogen, can the graphite or carbon fixturing and support plates be used in the hydrogen atmosphere? There are a number of variables that enter into the solution, with only a limited amount of information. For example, if the parts being brazed were pure copper or a nickel-copper alloy that does not pick up carbon, graphite would be suitable. However, in brazing stainless steel with 0.03% maximum carbon, there definitely would be a carbon pickup that would affect the corrosion resistance of the low-carbon and stainless steel. Normally, problems do not occur in vacuum furnaces, because stopoff or other material is placed between the parts and the graphite. Parts that can pick up carbon should never be placed directly on the carbon or graphite fixtures, particularly at higher temperatures, such as the copper brazing temperatures. If wet hydrogen is used rather than dry hydrogen, it should be pointed out that the moisture in the atmosphere would react with the carbon to form carbon monoxide and hydrogen. Carbon monoxide is a very good carrier of carbon to carburize the parts in the load. If the hydrogen has a high dewpoint, there will be more carbon monoxide formed and more reaction, thus the fixtures will deteriorate at a faster rate. Use caution when putting carbon steels in with low-carbon stainless steel alloys, particularly those having 0.03% or less carbon. To ensure that the carbon steel fixtures are adequately decarburized, a simple test is to place a 0.08 mm (0.003 in.) thick low-carbon stainless steel shim between two pieces of the carbon steel and run a simulated brazing cycle. Then, the stainless steel strip area between the carbon steel parts can be chemically analyzed for the amount of carbon in the stainless steel, along with a piece that has not been through the furnace cycle. This test will indicate whether there is any pickup of carbon in the stainless steel and
310 / Brazing, Second Edition
if the steel parts have been adequately decarburized. Atmospheres such as vacuum, pure dry argon, and pure dry helium do not show a pickup of carbon by migration. However, it is still important to use a stopoff (or some other material) between the carbon or graphite and the parts that are placed on it. Graphite and carbon fixtures and supports are very good, because they hold their dimensions much better than the normal stainless steels and will retain their flatness. However, using graphite or carbon is not recommended in atmospheres containing hydrogen, except in special cases where there is little or no pickup of carbon at the temperatures with the base metal being processed. Example 5: Troubleshooting Problems Associated with Tack Weld Fixturing. Tack welding as a means of fixturing is very convenient, but it can present difficulties in some instances. There are a number of tungsten electrodes used for gas tungsten arc welding (GTAW), which is the most convenient way to make tack welds. They include pure tungsten, tungstencesium oxide, tungsten-lanthanum oxide, tungsten-thorium oxide, and tungsten-zirconium oxide. The thorium oxide electrodes have been used the longest. However, there are problems where the thorium oxide is blown into the joint around the weld area, which causes the braze to stop before reaching the tack weld, creating porosity close to the edge of the tack weld. If the pure tungsten electrode is ground to a point, the arc is sufficiently stable, and the tack welds are readily made. Cracking of tack welds is caused by stresses in the part during the heating cycle during brazing. If it is desirable to tack weld parts, metal-tometal, it is essential that the faying surfaces to be brazed be blasted with a clean, special, nickel blasting material that will prestress the surface. This will allow the molten filler metal to be pulled through the joint. Otherwise, a definite clearance is required for the BNi-2 filler metal. Example 6: Flow of Filler Metal Around Tack Welds. When tack welding parts to hold them during the brazing operation, some technicians experience a problem with the filler metal not wanting to flow around the tack welds. The filler metal may flow through the joint up close to the tack welds, but it leaves a void on either side of them. With tack welds made using GTAW with argon shielding gas on 304L base
metal and BNi-2 filler metal, diffusion brazing was accomplished in a vacuum furnace at 10–2 Pa at 1065 °C (1950 °F) and held for 1 h. When the pure tungsten electrode is used the filler metal flows around the tack weld. Therefore, it is recommended as in Example 5 above that only pure tungsten electrodes be used for tack welding.
REFERENCES
1. M.M. Schwartz, Brazing for the Engineering Technologist, Chapman & Hall, 1995 2. W.D. Kay, Braze-Fixturing: A Key to Higher Productivity or a “Thief ”?, Ind. Heat., Jan 1994, p 32–34 3. M.M. Schwartz, Brazing, ASM International, 1987 4. A. Belohlow, Understanding Brazing Fundamentals, Am. Weld., Sept/Oct 2000, p 11–13 5. Brazing Handbook, 4th ed., American Welding Society, 1991, p18 6. M.M. Schwartz, Brazing in a Vacuum, WRC Bulletin 244, Welding Research Council, Dec 1978 7. C.F. Burns, Jr., Vacuum Brazing: A Tube and Pipe Application, Ind. Heat., Sept 1998, p 107–108 8. T.M. Anandan and R.R. Pfouts, Vacuum Furnace Brazing of Component for Baking Appliance, Ind. Heat., Nov 1994, p 63–65 9. C.J. Smithells, Ed., Metals Reference Book, 5th ed., Butterworth and Co., 1976, p 1267 10. D. Peterson, Eliminate Ozone-Depleters from Your Brazing and Welding Jobs, Weld. J., Sept 1994, p 39–42 11. I.A. Bucklow, Joining a Ni-Based Creep Resistant (ODS) Alloy by Brazing, The Welding Inst., International Institute of Welding Meeting and Proceedings on Joining, (Montreal, Canada), 1990, p 293–298 12. J.W. Chasteen, “Development and Evaluation of Wide Clearance Braze Joints in Gamma Prime Alloys,” Cont F33(615)79-C-5033, AFWAL-TR-82-4016, University of Dayton Research Institute, Dayton, OH, 1 May 1979 to 1 May 1981 13. W.A. Demo and S.J. Ferrigno, Brazing Method Helps Repair Aircraft Gas-Tur-
Chapter 7: Fixturing, Tooling, Stopoffs, Parting Agents, Surface Preparation / 311
14. 15. 16.
17. 18.
bine Nozzles, Adv. Met. Process., Vol 141 (No. 3), March 1992, p 43–45 J.W. Chasteen, “The U.D.R.I. Fluorocarbon-Cleaning Process,” 1980 J.W. Chasteen, U.S. Patent 4,405,379, 1983 H. Zhuang and Y. Li, Cleaning and Repair Brazing of Cracks in Turbine Vanes, DVS 125, Brazing, High Temperature Brazing and Diffusion Welding Symposium, 19–20 Sept 1989 (Essen, GDR), Deutscher Verband fur Schweisstechnik e.V., p 103–106 S.M. Riad and A. El-Naggar, Brazing of Gray Cast Iron, Weld. J., Vol 62 (No. 10), Oct 1981, p 22–24 O. Andersson, Joining of P/M Parts by Brazing, DVS Berichte 148, Brazing, High Temperature Brazing and Diffusion Welding, 1992, p 43–46; Deutscher Verband fur Schweisstechnik e.V., Dusseldorf, Germany
19. J. Christensen, K. Gotthjelp, and P. Kjeldsteen, The Effect of Surface Treatments on the Brazing of Iron-Based Powder Metal Compacts, Weld. J., June 1997, p 239–244 20. E. Lugscheider, W. Tillmann, et al., Joining of Porous PM Materials, DVS Berichte 148, Brazing, High Temperature Brazing and Diffusion Welding, 1992, p 163–167; Deutscher Verband fur Schweisstechnik e.V., Dusseldorf, Germany 21. C.F. Burns, Jr. and R. Lacock, Vacuum Brazing: A Three in One Process That Provides Efficiency, Ind. Heat., Oct 1996, p 43–45 22. C.F. Burns, Jr. and R. Lacock, Three-inOne Joining Process Streamlines Production, Mach. Des., 19 June 1997, p 87–89 23. A.L. Clavel and J.A. Kasperan, VaporPhase, Fluoride-Ion Processing of Jet Engine Superalloy Components, Plat. Surf. Finish., Nov 1991, p 52–57
Brazing Second Edition Mel M. Schwartz, p313-338 DOI: 10.1361/brse2003p313
Copyright © 2003 ASM International® All rights reserved. www.asminternational.org
CHAPTER 8
Joint Design IF A JOINT DESIGN WORKS RELIABLY in the application for which it is intended, it is a good joint design, whether it is a very simple or a very sophisticated design. However, the design of a brazed joint requires some special considerations, dictated by the nature of the joining process: • Composition and strength of the filler metal: Generally, the bulk strength of the filler metal is lower than that of the base metals, and so, a correctly designed joint is required to obtain adequate mechanical strength. • Capillary attraction: Because brazing depends on the principle of capillary attraction for distribution of the molten filler metal, joint clearance is a critical factor affecting the brazing process. • Flux and air displacement: Not only must the filler metal be drawn into the joint, but flux and air must be displaced from it. This requirement influences joint design and clearances. • Type of stress: In general, it is preferred that any load on a brazed joint be transmitted as shear stress rather than tensile stress. Different styles of joints subject the filler-metal film to different stresses and therefore alter its behavior under stress. These stresses usually are tensile stresses in butt joints, shear stresses in lap joints, and tensile and/or shear stresses in scarf joints. • Composition and strength of the base metals: In a joint made according to recommendations between high-strength members, the filler-metal film in the joint may actually be stronger than the base metal itself. The rigidity and freedom from yielding of the members of the joint confine the filler-metal film between them, causing the film to have prop-
erties different from those of the bulk filler metal. Therefore, it is possible for joints in high-strength base metals to be stronger than joints of the same type in lower-strength base metals.
Types of Joints One design factor that can be altered for best brazing results is the type and bonding area of the joint. This is an important factor, as well as clearance between members of the joint. Both of these affect not only the strength of the completed joint but also the ease of brazing. Several factors influence the selection of the type of joint to be used: fabrication techniques prior to brazing, the number of items to be brazed, the method of applying the filler metal, and the ultimate service requirements of the joint. Design of a brazed joint begins by building mechanical strength into the assembly, using overlaps, interlocks, and flanges that support areas of high stress. The final brazed joint creates a single fabricated part with respect to shear, compression, and dynamic and static loading. There are basically only two types of brazed joints: butt and lap. All other joints are really only modifications of these two basic types. Common types of brazed joints are shown in Fig. 8.1. Butt and Lap Joints (Ref 2). The butt joint has the advantage of a single thickness at the joint. Preparation is relatively simple, and the joint has sufficient strength for many applications. However, the strength of any joint depends, in part, on the bonding area available, and in a butt joint, this area is determined by the
314 / Brazing, Second Edition
thinnest member of the joint. The thinnest member, therefore, dictates the maximum strength of the joint. Advantages of the butt joint are ease of preparation and single thickness at the joint, which reduces stress concentrations. Another drawback of the butt joint is that almost the entire load is transmitted as tensile stress, which is not very desirable. For these reasons, a butt joint should be chosen only when the thickness of the joint is a critical consideration and strength requirements are secondary. Butt joints are used where the thickness of lap joints would be objectionable and where the strength of the brazed joint will satisfactorily meet the service requirements. The strength of a properly brazed butt joint may be sufficiently high so that failure will occur in the base metal, or it may be below the strength of the base metal, so that failure will occur in the braze. Joint strength will depend on the strength of the filler metal and on the filler-metal/base-metal interactions that take place during the brazing cycle. High efficiency will not be obtained with the butt joint when the filler metal in the joint is much weaker than the base metal. To obtain the best efficiency, the brazed joint should be free of defects (no flux inclusions, voids, unbrazed areas, pores, or porosity). Another means of obtaining high butt-joint strength is to use minimum joint clearances compatible with the base and filler metals involved as well as with the brazing process to be used. Although the minimum joint clearance produces optimal joint strength, producing such clearances is sometimes considered economically impractical. However, with the current
sophisticated metalworking techniques, maintaining proper clearances is not ordinarily a major problem. It is important to point out that if high-quality, high-reliability brazements are to be manufactured, it is imperative that clearances be controlled. The bonding area of a lap joint can be made larger than that of a butt joint. In fact, the area of overlap may be varied so that the joint is as strong as the weaker member, even when a lower-strength filler metal is used or when small defects are present in the final braze. The lap joint has a double thickness at the joint, but the load is transmitted primarily as shear stress, which is desirable. As a rule of thumb, an overlap of at least three times the thickness of the thinner member usually yields maximum joint efficiency (Fig. 8.2). Longer overlaps waste preparation time and filler metal and do not increase joint strength. For an exact determination of the length of a lap for maximum strength (Fig. 8.3), use one of these formulas:
1 Flat: L = F T × S
t(D – t) Tubular: L = F T × S
where L is the length of the lap, F is the factor of safety, T is the tensile strength of the thinner member, t is the wall thickness of the thinner member, S is the shear strength of the filler metal, and D is the diameter of the lap.
3t1 t2
t1< t2 Thickness, t1
Lap joint 3t Thickness, t Scarf joint
3t
Thickness, t
Butt joint modified
Fig. 8.2
Fig. 8.1
Types of brazed joints. Source: Ref 1
Determining bond area for different joint configurations. Bond area should be three times the thickness of the thinnest member.
Chapter 8: Joint Design / 315
Strength is only one reason that the majority of brazed joints are of lap design. Lap joints can be readily designed to be self-jigging or, as in the case of tubing, self-aligning. Also, preplaced filler metal can be held in position better with such joints. Lap joints do, however, have disadvantages. They can interfere with fit or function in some applications, they result in increased metal thickness at the joint, and they create stress concentration at the edges of the lap where there is an abrupt change in cross section. Butt-Lap and Scarf Joints. Variations of the two basic joint designs include the butt-lap joint and the scarf joint. The butt-lap joint is an attempt to combine the advantage of a single thickness with maximum bonding area and strength. It requires more preparation than straight lap or butt joints and may not be applicable to thin members. The butt-lap joint is generally easier and less costly to prepare than the scarf joint, which requires more fixturing. However, both find use with flat and tubular parts. The scarf joint represents another attempt to increase the cross-sectional area of the joint without increasing its thickness. Angling of the butting surfaces (Fig. 8.1) increases the effective bonding area. Such joints are relatively difficult to prepare properly and are even more difficult to align, particularly with thin members. Because the scarf joint is at an angle to the axis of tensile loading, its load-carrying capacity is similar to that of the lap joint and greater than that of the butt joint. Finally, the designer wishing to distribute the stresses in a joint, and thereby maximize the mechanical strength of the joint by reducing stress concentration, can determine where the greater stress falls and then counter such stress concentration (Fig. 8.4) by:
Fig. 8.3
Flat and tubular lap joints. Source: Ref 1
• Thickening the thinner members at the point of stress concentration • Reshaping the thicker member to spread the transfer of loads • Thickening both parts at the joint to enlarge the joint area and reduce stress • Changing the joint type • Moving the joint location • Adding reinforcement In assemblies of small or thin parts, a fillet should be used to distribute stresses and strengthen the joint. To produce a fillet, a little more than the minimum amount of filler metal is applied or a more sluggish filler metal is used, or both.
Joint Clearance The single most important design consideration in achieving good brazements is joint clearance—the distance between the faying surfaces to be joined. Joint clearance affects the mechanical performance of a brazed joint in several ways. It affects: • The purely mechanical effect of restraint of plastic flow of the filler metal offered by the greater strength of the base metal • The possibility of voids • The capillary force, which accounts for fillermetal distribution • Amount of intermetallic phases present in the joint The actual proper clearance for a brazed joint depends on the type of flux, the surface finish of the mating parts, the base-metal/filler-metal interaction, the base metal, the filler metal, the preplacement of the filler metal, and the type of brazing process to be used. In general, the smallest acceptable clearance will result in the strongest joint. Why should small clearances be used? The smaller the clearance, the easier it is for capillarity to distribute the filler metal throughout the joint area and the less will be the likelihood that voids or shrinkage cavities will form as the filler metal solidifies. Small clearances and correspondingly thin filler-metal films make sound joints. According to some experts, the ideal clearance for production work (Fig. 8.5) is 0.05 to 0.13 mm (0.002 to 0.005 in.). Joint clearances up to 0.13 to 0.20 mm (0.005 to 0.008 in.) are
316 / Brazing, Second Edition
good when silver-base filler metals are used. However, some metals actually require interference fits, whereas others require clearances as great as 0.25 mm (0.010 in.) (Ref 3, 4). As a rule, mineral fluxes, strong interactions between filler metal and base metal, long joints, widemelting-range metals, and easily oxidized base metals call for increased joint clearances. A factor that complicates determination of the proper design of a brazed joint is that the clearance must be proper at the brazing temperature. Although this is easily attained when both joint members are made of the same base metal, brazing of dissimilar metals with different coefficients of expansion can lead to complications. Table 8.1 may be used as a guide for determining the optimal clearance range at brazing temperatures when designing brazed joints for maximum strength.
Fig. 8.4
For example, if a brass bushing is to be brazed into a steel sleeve, the fact that brass expands much more than steel when heated to the same temperature must be considered. A clearance of, say, 0.08 mm (0.003 in.) at room temperature may disappear completely at a brazing temperature of 725 °C (1340 °F). With similar metals of approximately equal mass, the room-temperature clearance is a satisfactory guide. The joint clearances indicated in Table 8.1 are radial clearances for tube-type lap joints. The clearances should be used as diametral clearances for some applications if there is no provision in the design to ensure alignment and concentricity of the parts. Excessive joint clearance will result in voids in the joint, particularly when a gas flux is used. Joint strength is related to test specimen design and testing method. Thus, tests must be
Design changes that reduce stress concentrations. Source: Ref 1
Chapter 8: Joint Design / 317
conducted in accordance with the proposed production joint design and brazing procedures to obtain specific design strength data. Variations in brazing procedures and joint design will alter the effect of joint clearance on strength properties. For example, a free-flowing filler metal used for brazing in a high-quality atmosphere will adequately flow through a joint having a very small clearance. However, when the atmosphere deteriorates, it is often necessary to use a larger clearance to obtain adequate flow. When designing a brazed joint, the brazing process to be used and the manner in which the
filler metal will be placed in the joint should be established. In most manually brazed joints, the filler metal is simply fed from the face side of the joint. For furnace brazing and high-production brazing, the filler metal is pre-placed at or in the joint. Automatic dispensing equipment may perform this operation. Figure 8.6 gives some assembly recommendations. Figures 8.7 and 8.8 illustrate methods of preplacing filler metal in wire and sheet forms. When the base metal is grooved to accept preplaced filler metal, the groove should be cut in the heavier section. When computing the strength of the intended joint, the groove area should be subtracted from the joint area, because the filler metal will flow out of the groove and into the joint interfaces, as shown in Fig. 8.9. Powdered filler metal can be applied in any of the locations indicated in Fig. 8.7. It can be applied dry to the joint area and then wet down with binder, or it can be premixed with the binder and applied to the joint. The density of powder is usually only 50 to 70% of a solid metal, so the groove volume must be larger for powder. Where preplaced shims are used, the sections being brazed should be free to move together when the shims melt. Some type of loading may be necessary to move them together and force excess filler metal and flux out of the joint. Many assemblies are simple and require only a push fit. It is essential to have well-fitted joints with square corners on the female and male parts, so that capillarity is continuous throughout the joint. Furnace brazing and other high-produc-
Joint thickness, mm 150
1034
130
897
110
758
90
621
70
483
50
345
Tensile strength, MPa
Tensile strength, psi
0.08 0.15 0.23 0.30 0.38 0.45 0.53 0.61 0.69
207
30 0.003 0.006 0.009 0.012 0.015 0.018 0.021 0.024 0.027
Joint thickness, in.
Fig. 8.5
Tensile strength versus joint clearance. Source: Ref 1
Table 8.1 Brazing joint clearances Joint clearance(a) Filler-metal group
BAlSi BCuP BAg BAu BCu BCuZn BMg BNi
mm
0.05 to 0.20 0.20 to 0.25 0.03 to 0.13 0.05 to 0.13 0.00 to 0.05(b) 0.05 to 0.13 0.00 to 0.05(b) 0.00 to 0.05(b) 0.05 to 0.13 0.10 to 0.25 0.05 to 0.13 0.00 to 0.05
in.
0.002 to 0.008 0.008 to 0.010 0.001 to 0.005 0.002 to 0.005 0.000 to 0.002(b) 0.002 to 0.005 0.000 to 0.002(b) 0.000 to 0.002(b) 0.002 to 0.005 0.004 to 0.010 0.002 to 0.005 0.000 to 0.002
Comments
For laps less than 6.4 mm (0.25 in.) long For laps more than 6.4 mm (0.25 in.) long ... Flux brazing (mineral fluxes) Atmosphere brazing (gas-phase fluxes) Flux brazing (mineral fluxes) Atmosphere brazing (gas-phase fluxes) Atmosphere brazing (gas-phase fluxes) Flux brazing (mineral fluxes) Flux brazing (mineral fluxes) General applications (flux or atmosphere) Free-flowing types, atmosphere brazing
(a) Values given are radial clearances when rings, plugs, or tubular members are involved. For some applications, it may be necessary to use recommended values as diametral clearances to prevent excessive clearance when the entire joint gap is on one side. Excessive clearances will produce voids, particularly when brazing is performed in a high-quality atmosphere (gas-phase fluxing). (b) For maximum strength, a press fit of 0.001 mm/mm (0.001 in./in.) of diameter should be used.
318 / Brazing, Second Edition
tion brazing methods, however, frequently involve preplacement of the filler metal and may also incorporate some sort of automatic dispensing equipment. Provisions must be made to allow the components to accept the preform or paste and to secure it properly in position.
Fig. 8.6
Examples of preplacement of filler-metal wire and shims are shown in Fig. 8.7 and 8.8. In some applications, additional filler metal is added by extending the filler-metal shim beyond the joint. In dip brazing aluminum, joint clearances should be from 0.10 to 0.25 mm (0.004 to 0.010
Recommendations for assembly of brazed joints. Source: Ref 5
Chapter 8: Joint Design / 319
in.), and the preplaced filler metal is subject to very uniform preheating and subsequent immersion in the salt bath. Preheating is necessary to drive off any moisture on the assembly as well as to minimize thermal shock and flux entrapment. Very close attention must be given to joint clearances for induction brazing, because in this process, the heat may be induced in only one component of the joint, causing it to heat more quickly and thereby expand more rapidly than the other. This uneven expansion could produce a very undesirable, uneven heating condition and cause a change in joint clearance during the brazing cycle, which should be compensated for in the initial joint clearance. Finally, as a rule of thumb for torch, furnace, or mechanized torch or induction brazing with laps of 6.4 mm (0.25 in.) or less, clearances of 0.10 to 0.25 mm (0.004 to 0.010 in.) can be used. Clearances up to 0.6 mm (0.02 in.) are used for longer laps, because the filler metal changes composition by dissolving the base metal and becomes sluggish as it flows through long lap joints. The correct clearance for any given joint is best determined by trial. It should be kept in mind that, in most cases, a properly designed and produced high-strength brazement will ultimately fail in the base metal; the braze is by no means the weakest link. In general, loading of a brazed joint demands the same design considerations given to any joint or change in cross section. In that regard, any joint design that moves the high stress concentration
Fig. 8.6 (continued)
from the joint to the base metal is a good design. Whenever possible, joints should be designed to be stressed in shear or compression, rather than in tension.
Design for Assembly For a sound brazed joint, parts must be positioned securely during brazing and cooling cycles. If possible, parts should be designed so that gravity keeps them in position during brazing. If their shapes and weights permit, parts should be positioned so that they will stay put or move into position when the filler metal melts. The parts to be brazed should be assembled immediately after fluxing, before the flux has time to dry and flake off. Assemblies designed to be self-locating and self-supporting are the most economical. When fixtures are needed to maintain alignment or dimensions, the mass of a fixture should be minimized. It should have pinpoint or knifeedge contact with the parts, away from the joint area. Sharp contacts minimize heat loss through conduction to the fixture. The fixture material must have adequate strength at brazing temperature to support the brazement. It must not readily alloy at elevated temperatures with the work at the points of contact. In torch brazing, extra clearance will be needed to access the joint with the torch flame as well as the filler metal. In induction brazing, fixtures are generally made of ceramic materials to avoid putting extraneous
320 / Brazing, Second Edition
metal in the field of the induction coil. Ceramic fixtures may be designed to serve as a heat shield or a heat absorber. When gravity or gravity-aided fixturing cannot do the job, other methods of fixing parts in position must be considered. Such methods include press fitting, expanding, crimping,
Fig. 8.7
Methods of preplacing filler metal wire
Fig. 8.8
Preplacement of filler metal shims
Fig. 8.9
Brazed joints with grooves for preplacement of filler metal. Note that after the brazing cycle, the grooves are void of filler metal.
peening, swaging, staking or pinning, and tack welding (Fig. 8.10). Self-jigging is the method of assembly in which the component parts incorporate design features that will ensure that the parts, when assembled, will remain in proper relationship throughout the brazing cycle, without the aid of auxiliary fixtures. This is the preferred method of assembly, because it eliminates the initial and replacement costs of auxiliary fixtures and the cost of heating them during brazing, and it usually is a more effective method of holding the components. Self-jigging can be accomplished by the various methods mentioned previously (Ref 3, 6). Gravity Locating. Perhaps the simplest method of assembling two components is to rest one on top of the other, with the filler metal either wrapped around one component near the joint or placed between the components. The principal disadvantage of gravity locating is the lack of a dependable means of orienting the components or keeping them from moving in relation to one another. Nevertheless, some production components are assembled in this manner, especially those in which the upper component is relatively heavy. Interference or Press Fitting. This assembly method requires expansion or contraction of mating component surfaces and provides a very tight fit—sometimes called a tight press fit. Most interference fits require considerable force to achieve assembly, a force generally provided by an arbor press or similar tool. Thus, an interference fit is a press fit. Lighter interference fits may provide zero clearance or a very slight gap between the mating surfaces of the components. These fits also require some external force, such as that provided by an arbor press, to achieve assembly. Fits with zero clearance are referred to as sizeto-size fits. Some method is used to prevent slippage when the components are heated in the furnace, particularly if the joint has a vertical axis. A shoulder on one of the components can be used to ensure stability. Knurling. In high-production manufacturing, there is considerable variation in joint clearance among the assemblies being brazed. Typical brazed assemblies in which a round male member is fitted into a female member are subject to either of two conditions: (a) the male part is off-center, or (b) the male member is out-ofround. Knurling of the male member is sometimes used to correct these conditions and
Chapter 8: Joint Design / 321
obtain uniformity among brazed joints. For a detailed discussion of knurling and examples of its use in solving a brazing problem involving a stainless cryogenic valve body, see Ref 7 and Fig. 8.11 and 8.12. Staking. Figure 8.10 shows how staking will effectively lock two components in position. Burrs are turned in on the shaft by driving a punch into it. This method, of which there are a number of modifications, is commonly used to retain the orientation of such assemblies as cams, levers, and gears on shafts or on common hubs. It is sometimes a substitute for tack welding, knurling, or interference fitting. Expanding. This method is commonly used in assemblies of tubes to tube sheets. A tubular component is pressed into a header sheet and expanded in the hole to lock the assembly, as shown in Fig. 8.10. Spinning. When the diameter of a hole in an assembly may not be altered during assembly, as when a hub is fastened to a lever, the assembly can be locked together by spinning in a riv-
RSW or GTAW
Fillermetal ring
Gravity located
Riveted
Fig. 8.10
(a)
Fig. 8.11
eting machine. The spinning method of assembly is used for parts for various types of business machines, many of which were formerly assembled by cross drilling and pinning of hubs. Swaging. An inexpensive and effective method of assembling a stud in a hole in a hollow body is to swage it in place on the detail parts or in the punched hole, because the swaging operation forces the components into intimate contact. Crimping. Figure 8.10 shows the assembly of a disk, a shell, and a filler-metal ring in which the disk and ring are held in place by crimping the end of the shell. In general, it is preferable to set an assembly of this type on end in the furnace, so that the filler metal will flow downward through the joints. Tack Welding. The tack welding method of assembly usually requires careful investigation to determine the most strategic point(s) for placing the weld(s). For economy, the number of tack welds per assembly should be held to a minimum.
Interference filled
Spun
Tack welded
Swaged
Knurled
Crimped
Staked
Expanded
Peened
Typical self-fixturing methods for brazed assemblies. RSW, resistance spot welding; GTAW, gas tungsten arc welding
(b) Knurling. (a) Schematic of a typical knurled surface showing vertical capillary paths for filler metal. (b) Closeup of knurl pattern showing diameter before and after knurling. Exaggerated size for clarity. Source: Ref 7
322 / Brazing, Second Edition
Peening. Assembly of two hollow shells by the peening method is shown in Fig. 8.10. Components are pressed together, and the outer shell is peened. For application of filler metal to an assembly of this type, spraying on the joint interfaces before assembly is preferred. Riveting and Folding or Interlocking. Riveting could be considered to be a modification of the spinning and swaging methods in which a rivet is used as part of the assembly. It is widely used to assemble the vanes to the outer disks of fan wheels before furnace copper brazing. Several methods and designs of folding or interlocking can be used to secure joints. These methods are widely used in the manufacture of brazed tubing or tubular assemblies.
Effects of Brazing Variables on Clearance Mineral and Gas-Phase Fluxes. Use of mineral-type or gas/atmosphere-type fluxes, or a combination of both, will have an important bearing on joint clearance. When the clearance is too small, the mineral flux will be held in the joint, and displacement by the liquid filler metal may be difficult or impossible. Thus, joint defects may be produced. When the clearance is too large, the liquid filler metal will travel around pockets of flux, thus giving rise to excessive flux inclusions. Gas-phase fluxes are atmospheres that also affect optimal clearances for specific brazements. Gas fluxes permit lower clearances for optimal strength, and so, the load-carrying capacity of the joint can be higher. In atmosphere furnace brazing of joints in the vertical
Knurled area
Fillermetal loading groove
Fig. 8.12
Filler-metal loading groove with knurled design (exaggerated gaps for clarity). Source: Ref 7
position, free-flowing filler metals will flow out of joints having clearances in excess of 0.08 mm (0.003 in.). Surface Finish. In general, the filler metal is drawn into the joint by capillary attraction. Thus, if the surface finish of the base metal is too smooth, the filler metal may not distribute itself throughout the entire joint and may leave voids. To ensure adequate filler-metal flow throughout the joint, particularly when the clearance is zero or is a press fit, the faying surfaces of the joint should be roughened, preferably with a clean metallic grit compatible with the base metal. Surfaces that are too rough also result in lower joint strength, because only the points may be brazed, or because the average clearance is too large. A surface roughness of 0.8 to 2 µm root mean square is generally acceptable but is not to be considered optimal for all basemetal/filler-metal combinations. Tests must be conducted to ensure optimal conditions for a specific brazement. Base-Metal/Filler-Metal Interaction. The mutual solubility of many base metals and filler metals causes interaction to take place through solution of the base metal by the liquid filler metal and through diffusion in the liquid and solid states. Such interaction affects the permissible clearance for a specific brazement. If the interaction is low, the clearance may, in general, be small; however, when the interaction is high, which usually occurs in the liquid phase, then the clearance will have to be larger. When the joint is long and interaction is large, the clearance should be increased. For example, when aluminum base metal is brazed with a lowermelting aluminum-base filler metal at a temperature close to the base-metal melting point, interaction can be expected. Base metals often contain one or more elements whose oxides are not easily dissociated in a specific atmosphere or by a specific mineral flux. Because the metal/metal oxide dissociation of a given base metal by a specific mineral flux or atmosphere is dependent on many factors, it is important to match the proper clearance with the brazing process and flux or atmosphere. For example, for small aluminum additions, the clearance may have to be increased. With larger aluminum additions, brazing may not take place unless the atmosphere is made more active (lower dewpoint or lower vacuum pressure) or the surface is protected with a barrier coating, such as an electrolytic nickel coat-
Chapter 8: Joint Design / 323
ing. The clearances would then require reappraisal to obtain optimal joint properties. Coefficients of thermal expansion will, of course, have an effect on joint clearance at brazing temperature when the base metals are dissimilar. Filler Metals. Free-flowing filler metals generally require smaller clearances than sluggish filler metals. Filler metals that have single melting points, such as copper, silver, eutectic filler metals, and self-fluxing filler metals, will usually be free-flowing, particularly when there is very little interaction with the base metal. Variations in the quality of fluxes and atmospheres (low-quality or oxidized fluxes, lowpurity atmospheres) can enhance the free-flowing characteristic of the filler metal or can result in no flow at all. Thus, clearances may appear to be improper for a given set of brazing conditions when, in fact, the flux or atmosphere requires improved control. Liquid Metal Infiltration. A technique of liquid infiltration reduced the intermetallicforming elements and improved the mechanical properties of brazements with large joint clearances (Ref 8). High-temperature brazing of nickel-base alloy Inconel 625 (Special Metals Corp.) by liquid infiltration was investigated. Nickel-phosphorus, nickel-boron, and Ni-41.2Pd-8.8Si (Metglass MBF-1006, Honeywell International Inc.) were used as melting-point depressants. They formed wetting liquids at the bonding temperature and subsequently infiltrated the nickel-chromium powder interlayers. The infiltrated liquid caused shrinkage of the powder interlayer (Ref 1, 9–12). In this study, a process called liquid-infiltrated powder interlayer bonding (LIPB) was used to produce brazements with large joint clearances using nickel-base filler metals. The filler metals consisted of coatings and thin foils on the faying surfaces. A powder interlayer was placed between the two filler metals. At brazing temperature, the filler metals melted and infiltrated the powder interlayer. Dissolution and rearrangement of powder particles occured subsequently, followed by homogenization of the joint. By controlling the amount of liquid formed, brittle phases and porosities in the powder interlayer were reduced to a minimum, while a clearance of up to several millimeters could be joined. Isothermal solidification time was much shorter, as compared to the diffusion brazing process, because the diffusion path was
greatly decreased in the powder interlayer. Thus, the LIPB process relaxed the limitation on joint clearance set by the diffusion brazing process. The process can be divided into four consecutive stages, which are illustrated in Fig. 8.13. The following results were obtained: • Brazements with large joint clearances could be made by the LIPB process, using nickelphosphorus and nickel-boron eutectics as the infiltrants. Intermetallic compounds were greatly decreased because of the much smaller amount of phosphorus or boron contained in these joints, compared to those joined with commercial filler metals. • Nickel-phosphorus eutectic was found to cause shrinkage of the nickel-base powder interlayer and to contribute to elimination of residual pores in the joint, whereas nickelboron eutectic showed less benefit in this regard. • Infiltration using MBF-1006 as the infiltrant was slow, due to high solubility of palladium and silicon in the nickel-base powder interlayer. A proper mixture of the infiltrant and the interlayer is important to produce a fully infiltrated joint. • Tensile tests showed 60% joint efficiency up to 700 °C (1290 °F) when Inconel 625 was the base metal and nickel-phosphorus eutectic was the infiltrant. Plastic deformation was observed, although the ductility of the joint was still much lower than that of the base metal. Joints infiltrated by nickel-boron eutectic revealed poor tensile strength, due to isolated pores in the interlayers. Wide-Gap Brazing. Unlike conventional brazing, where only the filler metal is used, in wide-gap brazing, the filler metal is often a mixture of filler metal and a high-temperature-melting powder, the latter being more commonly referred to as the gap filler. During brazing, the gap-filler particles remain largely unmelted, thus providing the necessary capillary forces to retain the molten filler metal, which would otherwise be too fluid to bridge the gap faying surfaces. Such a mixture would thus behave like a slurry, with sufficient bridging power to fill large defects and wide cracks and to rebuild large surface areas of aerofoils. In most of the studies reported on wide-gap brazing, the gaps to be brazed were prepacked with loose braze-mix powders, with an addi-
324 / Brazing, Second Edition
tional supply of filler metal or braze-mix powder deposited outside the gap (Ref 13–15). In other studies, partially sintered inserts of filler-metal mixes were used (Ref 16, 17). Although the preplacement technique was used in a few instances in which the filler metal was only deposited in the reservoir area outside the gap to be brazed, these studies were more concerned with the microstructure and resultant mechanical properties of joints brazed with filler metals only (Ref 18–20). To date, limited information is available regarding the effect of material and process parameters on the physical soundness of widegap brazed joints produced by such a technique. Researchers (Ref 21) conducted a study whereby wide gaps were brazed using the preplacement technique. The latter technique of braze-mix application is more suitable for repairing cracks sustained on thin-walled aero-
Fig. 8.13
engine components. It may not be easy to braze repair such cracks using the prepacking technique, because it is often difficult to access the underside of the cracks during prepacking. Commercial filler-metal and gap-filler powders, Nicrobraz LC and Nicrogap 116 (Wall Colmonoy Corp.), respectively, were used in the study; the compositions of the previously mentioned filler metals are given in Table 8.2. After brazing, the joints were sectioned depthwise, prepared metallographically, and examined by optical microscopy to detect the presence of macrovoids, the latter being defined as voids of characteristic length greater than the mean diameter of gap-filler particles in joints made with filler-metal mixes, or larger than onetenth of the width of the gap in joints made with filler metals only. The results showed that three major types of macrovoid could be identified:
Schematic showing the stages of the liquid-infiltrated powder interlayer bonding process. The widening stage is absent, because the dissolution mainly occurs between the powders and the melting-point depressant (MPD) foils. The stages may overlap, depending on the rates of infiltration and diffusion. Source: Ref 8
Chapter 8: Joint Design / 325
• Type 1: irregularly shaped macrovoids throughout the longitudinal section of the joint. Type 1 macrovoids dominated at low brazing temperatures over the range of gap widths and gap-filler contents that were studied. • Type 2: irregularly shaped macrovoids at the tail end of the joint. Type 2 macrovoids occurred frequently in joints produced at intermediate and high brazing temperatures with filler-metal mixes with gap-filler contents of less than 20%. For a fixed gap-filler content, the frequency of occurrence of such macrovoids increased appreciably with increasing brazing temperature. • Type 3: spherical macrovoids in the fillermetal deposit and adjacent joint area. Type 3 macrovoids were common features in joints produced at intermediate and high brazing temperatures with filler-metal mixes with gap-filler contents greater than 40%. With increasing gap-filler content, type 3 macrovoids, which appeared initially in the fillermetal deposit, spread into the adjacent joint area. The formation of type 3 macrovoids can be suppressed by brazing at a higher temperature. Sound wide-gap brazed joints free from the various types of macrovoids described previously were reproduced using the preplacement technique using filler-metal mixes with gap-filler contents of 30 to 40% over a range of brazing temperatures from 1150 to 1200 °C (2100 to 2190 °F). The present results were compiled in the form of braze quality-control maps, delineating regions of sound joints from those containing specific types of macrovoid as functions of brazing temperature, filler-metal mix ingredients, gap width, and gap depth. A subsequent study (Ref 20) examined the effects of material and process parameters on the formation of various types of constituent void,
namely, interfacial, interstitial, and shrinkage voids, in nickel-base wide-gap filler metals produced by prepacking with dry filler-metal-mix powders of Nicrobraz LC and Nicrogap 116. The results showed that the size and number of interfacial and interstitial voids increased with gap depth and width and with gap-filler content in the filler-metal mix but decreased with increasing brazing temperature. At intermediate and high gap-filler contents, shrinkage voids were not readily observable, because they were absorbed into other constituent voids. At low gap-filler contents, shrinkage voids were numerous and small at high brazing temperatures but could become isolated and large at low brazing temperatures. The findings were compiled into braze quality-control charts delineating the regions of gap-filler content and brazing temperature over which a gap of a certain depth could be soundly brazed without the formation of constituent macrovoids. The study showed that gap depth was an important parameter that must be considered in wide-gap brazing with prepacking of nickel-base filler-metal mixes. Large interfacial and interstitial voids were common features for gaps 6 mm (0.24 in.) or larger in depth, unless brazing was conducted at sufficiently high temperatures with filler-metal mixes with gap-filler contents of less than 30%. Wide joint gaps >500 µm can sometimes be used to minimize the effects of expansion mismatch between two components. The filler metal must have high viscosity in order to fill such wide joints. This is achieved either by using a filler metal with a wide melting range and performing the joining process at below the liquidus temperature, so that the filler metal is not fully molten, or by mixing in metal powder. Spacers are required to control the joint gap. Wide joints can also be achieved by inserting porous shims. One particular merit of wide joints to ceramic components is that they obviate the need to closely machine the mating surfaces of the components, which tends to be
Table 8.2 Nominal compositions of materials used in a wide-gap brazing study Composition, wt% Material
IN-625 Nicrobraz LC Nicrogap 116 Source: Ref 21
Ni
Cr
Fe
Si
B
S
Mn
Al
Ti
Mo
Nb
C
61.0 73.9 80.0
21.5 14.0 20.0
2.5 4.5 ...
0.25 4.5 ...
... 3.0 ...
0.008 ... ...
0.25 ... ...
0.2 ... ...
0.2 ... ...
9.0 ... ...
3.65 ... ...
0.05 ... ...
326 / Brazing, Second Edition
costly and can weaken the material by creating subsurface cracks. However, because the joint is wide, the mechanical properties of the joint are essentially those of the bulk filler metal. A variety of mechanical schemes are available to assist in overcoming the problem of thermal expansion mismatch. Several approaches that have proved successful are described subsequently. Interlayers. One route toward reducing the mismatch stress concentration that develops in soldered and, in particular, brazed joints involves a redesign of the joint to accommodate one or more interlayers. Two basic configurations are described in the literature (Ref 9). In the first approach, a compliant interlayer is inserted that will yield when the joint is placed under stress, thereby reducing the forces acting on the components. Optimal stress reduction is normally achieved when the thickness-to-length ratio of the interlayer is in a certain range, which is determined by the combination of materials used and the dimensions of the joint. For such a joint with a length of 15 mm (0.6 in.) between Al2O3 and steel employing a Ag-Cu-Ti filler metal and a nickel interlayer, the interlayer should be between approximately 1.5 and 3 mm (0.06 and 0.12 in.) thick. If the interlayer is much thinner than the prescribed minimum thickness, it is unable to absorb a significant proportion of the applied stress, whereas if it falls outside the upper limit, the interlayer will not yield to any extent. Interlayers are used to relieve mismatch stresses more frequently in brazed joints than in soldered assemblies, for the following reasons. First, the modulus of compliant metals that are most effective in accommodating stress is too close to that of many solders to provide much relief, because the solder will tend to yield in preference to the interlayer. Second, solders tend to form hard, interfacial phases with most engineering metals and alloys, which will confer a high modulus to the interlayer and actually exacerbate the situation. By contrast, common filler metals are more likely to form solid solutions, which are more ductile. High-purity copper is generally used for interlayers in brazed joints, because it combines a low elastic modulus with good wetting characteristics and is also inexpensive to fabricate to the desired geometry. An alternative approach is to redistribute the stresses across a much wider zone, so that they are within tolerable levels everywhere in the assembly. One method of achieving this gradu-
ated redistribution of stress is to insert into the joint one or more thick shims or plates that have thermal expansion coefficients that are intermediate between those of the abutting components. The plates must be sufficiently thick (generally not less than 5 mm, or 0.2 in.), so that they are not significantly distorted by the imposed stresses. An assembly containing a single plate with an intermediate thermal expansivity is shown in Fig. 8.14. This approach is particularly suitable where there is a need to join metals to ceramics and other ceramic-like nonmetals. If the intermediate plate is selected to have a thermal expansion coefficient that is close to that of the nonmetal, then it is possible to transfer the major proportion of the stress to the more robust metallic part of the assembly. Where the two components have greatly different thermal expansivities, it may be necessary to use a graduated series of plates to reduce the mismatch stresses in each joint to an
Si3N4 (α = 3 x 10– 6/°C) Ag-Cu-Ti Molybdenum (α = 6 x 10 – 6/°C) Ag-Cu-Ti Ductile iron (α = 12 x 10 – 6/°C)
Fig. 8.14
Graded seal assembly. Source: Ref 9
Ceramic
Metal
αC α4 α3 α2 α1 αM α = thermal expansivity
αC <α4 <α3 <α2 <α1 <αM
Fig. 8.15
A graded series of plates (1 to 4) designed to reduce the mismatch stress between ceramics and metals to an acceptable level. α, thermal expansivity. Source: Ref 9
Chapter 8: Joint Design / 327
acceptable level. A schematic illustration of an assembly of this type is shown in Fig. 8.15. Self-Brazing Metals. Self-brazing material, base metal with filler metal that is roll bonded to it on one or both sides, produces brazements without application of filler-metal shims, pastes, or fluxes. During brazing, the clad layer, a source of filler metal, fills joints by capillary action. Producers of self-brazing materials offer custom and standard clad materials—ferrous, copper, or aluminum alloys clad on one or both sides with copper, bronze, or zinc. Clad metal forms and stamps as would base metal alone, using the same tooling. The presence of soft, copper-base filler metal on stainless and carbon steel substrates can reduce tool wear. With no need to apply shims, powders, fluxes, or pastes or to clean brazements of residue, self-brazing materials bring significant savings, especially in high-volume production and in manufacture of complex brazements that contain several braze joints. Producers of clad base metals control roll bonding of the composite material to create uniform clad-layer thicknesses. A cladding thickness ratio indicates the thickness of the clad layer and base metal as percentages of the total composite thickness. A ratio identified as 8:84:8 represents a composite material with the base metal making up 84% of the thickness and a clad layer with a thickness of 8% on both sides of the base metal. Typical cladding ratios are 5 to 30%. Once the filler metal is clad to the base metal, subsequent rolling or forming will not alter this ratio. Manufacturers offer self-brazing clad sheet and foil 0.10 to 2.5 mm (0.004 to 0.10 in.) thick in widths of 3.2 to 610 mm (0.13 to 24 in.). Selfbrazing stock with filler-metal layers down to 0.13 mm (0.005 in.) thick are possible without porosity and discontinuities in the clad alloy or separation between the filler-metal layer and the base-metal layer. Braze cladding 0.05 to 0.10 mm (0.002 to 0.004 in.) thick produces comparable braze-joint thicknesses that deliver optimal joint strength. Shims in this thickness range are expensive to fabricate and too fragile to handle practically. Typical filler-metal shims have thicknesses of 0.13 to 0.20 mm (0.005 to 0.008 in.), which is greater than recommended joint thickness for maximum braze-joint strength. Self-brazing materials work well with contoured joints and with brazements of fine detail and great complexity. When designing with
clad materials, designers need not worry about joint configurations for introducing filler metal to the joint. Removing this constraint expands design flexibility. Joints that take filler-metal shim inserts cannot switch to clad-metal substrates without some redesign to realize the narrower joint thicknesses that clad materials provide. Because self-brazing materials result in smaller volumes of filler metal at the joint interface, brazements experience less shrinkage. Fixturing and tooling must accommodate the narrower joint thicknesses and the subsequent decrease in joint expansion and contraction during brazing.
Strength Loading. Joint strength is a function of joint design, brazing temperature, amount of filler metal applied, location and method of fillermetal application, heating rate, and brazing technique. The designer must evaluate expected service stress on the brazement and create joints that preferentially transmit loads as shear rather than tensile stress. For high-performance joints, consider increasing the alloying surface areas, not the fillermetal-layer thickness. When possible, include interlocking joints or flanged and countersunk parts. Properly designed joints result in fillermetal films that are stronger than the base metal, due to the mechanical constraints imparted to the filler metal. Therefore, joints in high-strength base metals can be stronger than similarly designed joints in lower-strength base metals. Exceptions may occur, however, with hardenable base metals. Brazement tensile strength increases with increasing base-metal strength to a greater degree than with shear strength. Stress Concentrations. The designer can redirect stress from the joint into the base metal by thickening the thinner member at the point of stress concentration, reshaping the thicker member at the concentration, changing the joint type, moving the joint away from stress concentrations, or adding reinforcement. Relocating the joint away from stress concentrations may enable use of thinner members for reduced assembly weight. Bonding Area. Joint strength is a function of overlap and braze-joint thickness, the latter determined by joint clearance. Greater overlap produces greater shear strength in lap and buttlap joints. Longitudinal shear stress on braze-
328 / Brazing, Second Edition
ments is limited when overlaps are kept small. Testing should be done to evaluate the appropriate joint overlap for the intended service of the brazement. When calculating the strength of a joint that incorporates a gap or void for preplacement of the filler metal, the groove area must be subtracted from the overall joint area, because bonding of the base components does not exist across the groove. Distortion. Part temperatures should be uniform during heating and cooling, so that thermal stresses within the components do not exceed base-metal yield point. Proper design and braze-temperature control are key to controlling distortion. For precise dimensions, allow extra stock for machining after brazing, or preshape pieces in a compensating direction to offset expected distortion. Creep Stress. At brazing temperature, materials may have low elevated-temperature creep strength. This property, typically lower than the material yield strength, distorts heavy parts in the downward direction. Another indication of creep during brazing is indentations where the part was supported or fixtured. To avoid creep, design and orient parts such that the heaviest or strongest section supports the assembly weight. Thin members should not experience compressive forces. When evaluating the joint design and fixturing, analyze components as simple stress-analysis problems using common beam-stress formulas. Use material-strength values at brazing temperature. Joint Length and Configuration. Joint length affects clearance, particularly when there is interaction between the filler metal and the base metal. As the filler metal is drawn into a long joint, it may pick up enough base metal to freeze before it reaches the other end. The more interaction that exists with a specific basemetal/filler-metal combination under given brazing conditions, the larger the clearance must be as the joint becomes longer. This is only one of a number of reasons why it is important to make the joint length as short as possible, consistent with optimal joint strength. Dissimilar Base Metals. In designing joints where dissimilar base metals are involved, the joint clearance at the brazing temperature must be calculated from thermal expansion data. The brazing temperature must be also taken into consideration (Ref 22, 23). When there is high differential thermal expansion between two details, the filler metal must be strong enough to
resist fracture, and the base metal must yield during cooling. Some residual stress will remain in the final brazement as a result of joining at the brazing temperature and subsequent cooling to room temperature. Thermal cycling of such a brazement during service will also stress the joint area, which may or may not shorten service life. Whenever possible, the brazement should be designed so that residual stresses do not add to the stress imposed during service. In a few specific cases, base materials (such as carbides) do not possess sufficient ductility, and it is necessary to use a soft, ductile spacer, such as nickel or copper, which will yield during the cooling cycle to prevent high stress between two high-strength base metals, for example, tungsten carbide brazed to a nickel- or cobalt-base stainless steel that is heat and corrosion resistant. Problems can also be expected in brazing of low-ductility materials, such as carbides and ceramics, to heat treatable base metals requiring high-to-moderate quenching rate or to base metals that undergo volume increases as a result of transformation (Ref 24). Dissimilar-Metal Joints. Different thermal expansion characteristics mandate care in designing joint clearance, heating technique, and fixturing. Unequal rates of expansion at brazing temperature alter joint clearances and can result in thermal stresses that lead to distortion. The result is deformation, braze failure, residual stresses, or lack of bonding. For joints between dissimilar materials, the designer should use thermal expansion curves rather than values for coefficients of expansion. This allows the designer to prepare the joint for the entire brazing-temperature cycle during heating and cooling. To design for adequate clearance, begin with a fit-up at a specified ambient temperature, then adjust for expected differences in thermal expansion. High differential thermal expansion between joint components requires a filler metal with strength enough to resist fracture and base metals that yield during cooling. The final brazement will contain residual stress. Joint design should ensure that this residual stress does not add to expected service stress, thereby reducing the life of the assembly. Techniques for alleviating thermal stresses of dissimilar joints include sandwiching into the joint a soft-metal layer that will yield during brazing and limit shear forces on the base metals. A more difficult approach is to design fixturing that applies force to the parts to compensate
Chapter 8: Joint Design / 329
for expansion differentials. Fixture force resizes or positions components to desired dimensions. Applicable only to ductile materials, this technique results in plastic deformation of one or both parts. Brazing brittle components, such as intermetallics, ceramics, or carbides, to metal alloys can result in strain cracking of the brittle member due to stresses that build up during cooling from brazing temperature. This is due to their low expansion and ductility characteristics. Avoid strain cracking by using multiple inserts, bonding only one contacting surface, or leaving gaps between adjacent surfaces to allow contraction of the alloy. Tube-to-socket joints require the material having the larger coefficient of expansion to be the outside tube. In brazing a brass tube into a steel socket, the greater expansion of the brass may close the clearance and squeeze molten filler metal out of the joint, thus preventing a bond. A steel tube in a brass socket may result in excessive joint clearance because of the expansion differential, again resulting in lack of bonding. For tongue-in-groove joints, the tongue is typically the high-expansion material. Joint clearances required for tube-to-tube or tube-to-socket joints may be too stringent to obtain by machining. Manufacturing allowances on IDs and ODs may result in excessively tight or loose fit-ups for brazing. Techniques for ensuring tube centering and alignment include knurling, crimping, swaging, expanding, and staking. As an example, inner tubes may have the bonding area knurled with long, vertical,
Fig. 8.16
side-by-side indentations. The indentations allow capillary paths for the filler metal and provide easily machined press-fit clearances. These techniques provide self-fixturing and alleviate machining tolerances; however, they still must allow capillary action and filler-metal flow.
Joint Design and Ceramics The mechanical behavior of ceramics is often a key factor in their technological application. For structural applications, strength, thermal shock resistance, fracture toughness, reliability, and lifetime are invariably critical issues. Even in nonstructural applications, mechanical behavior is still of importance in producing an optimal design and in degradation mechanisms. Since the early 1980s, there have been substantial advances in understanding the mechanical behavior of ceramics, especially the way this behavior is influenced by processing and microstructure. At the same time, the scientific framework for describing mechanical properties has become more sophisticated, and new testing techniques have been developed for the measurement of critical parameters. In order to understand fracture, flaw, size, shape, and so on, various developments have occurred. Detailed information on fracture mechanisms, joint design, and associated joining processes can be found in Ref 25 through 32 as well as in Fig. 8.16. Electroforming is a process by which join-
Ceramic-to-metal joint configurations. (a) Butt and lap seal joint designs. (b) Joint designs for transition to thick, all-metal members. (c) Backup of ductile metal seal with blank ceramic
330 / Brazing, Second Edition
ing is obtained by electroplating a metallic layer on the surface of a joint, or a layer of metal is deposited on a form. Ceramic joints joined in this manner require that the ceramic first be metallized and plated with copper and nickel or gold. After masking off conductive areas on which plating is not desired, the joint is plated with copper in an electroplating bath. After plating, the form is removed to leave a shell of metal whose inside configuration matches that of the form. The technique has been used to produce ceramic-to-metal seals in a traveling wave tube (Ref 33). Typical joint designs are shown in Fig. 8.17. Research (Ref 33) resulted in the fabrication of 41 electroformed seals joined by electroplating. The materials used included molybdenum, copper, and Kovar; the ceramic was a high-Al2O3 body. Filler Metals and Coatings. Most filler metals do not wet ceramics easily, unless their surfaces are treated in a manner to promote wetting. Such difficulties are to be anticipated when one recalls that oxide ceramics (Al2O3, BeO, ZrO2, and so on) comprise the largest group of these structural materials.
(a)
(b)
(c)
(d)
Electroform copper Ceramic Metal component
Fig. 8.17
Joint designs for electroformed seals. (a) Vee design of metal-to-ceramic electroform seal. (b) Step design of metal-to-ceramic electroform seal. (c) Step design using plain ceramic cylinders and metal sleeves. (d) Vee design using ceramic disc and metal cylinder. Source: Ref 33
Although the metallizing of ceramic surfaces is costly and time-consuming, the brazing of metals to such surfaces is a relatively straightforward process, because the metallized layer ensures wettability of the ceramic by the filler metal. However, certain metals and hydrides possess the ability to wet ceramic surfaces that have not been metallized, and active-metal and active-hydride processes based on this characteristic have been developed for producing ceramic-to-metal joints and seals. The joining of ceramics to metals with the active-metal or active-hydride processes dates back to the middle 1940s, when titanium hydride was used for this purpose (Ref 34). Fine titanium hydride powders (300 mesh) suspended in a suitable binder were painted on the area to be joined; they were then dried, and the ceramic and metal parts were assembled, with a silver-base filler metal in contact with the hydride area. The assembly was heated to 900 to 1000 °C (1650 to 1830 °F) in a vacuum or a H2 atmosphere. Pure titanium remained on the ceramic surface after the titanium hydride dissociated. When the silver filler metal melted, it alloyed with the titanium to form a silver-titanium alloy that bonded strongly with the metal and the areas of the ceramic that were coated with titanium hydride. Researchers (Ref 35) reported that in studying the bonding of ceramics with active metals and their hydrides and extending the work outlined in Ref 34, the hydrides of zirconium, tantalum, and niobium were just as effective as titanium hydride in ceramic-to-metal joints. The effectiveness of various filler metals in making bonds with Al2O3, synthetic sapphire, BeO, and ThO was evaluated. In addition, it was found that titanium and zirconium, produced in reducing titanium and zirconium hydride, could also be used in powder form for ceramic-to-metal joints, thus marking the beginning of the active-metal joining process. In developing experimental filler metals, it was noted that excellent bonds to ceramics, diamonds, sapphires, and other materials were made with a filler metal containing 85Ag-15Zr; aluminum-zirconium, aluminumsilver-zirconium, and silver-titanium were also evaluated as filler metals. The effects of various brazing environments, including vacuum and gases (H2, argon), were also investigated. Since these early investigations (Ref 35–37), joining ceramics to metals by the active-metal or active-hydride process has advanced significantly. The strengths of joints made by this
Chapter 8: Joint Design / 331
process are as great as those obtained with joints made by the moly-manganese process. Some difficulty has been experienced in making seals by the active-metal or active-hydride process in dry H2. The dewpoint of H2 must be extremely low to prevent oxidation of titanium. Producing ceramic-to-metal seals in a vacuum is advantageous in that the parts are outgassed during brazing. The concept of fabricating ceramic-to-metal joints and seals by the active-metal or activehydride process was first applied in the electronics industry. In recent years, however, these joining processes have found other uses to meet the need of high-temperature vacuumtight seals in the nuclear and aerospace industries. The characteristics of the moly-manganese, active-metal, and active-hydride processes are as follows:
to permit metals to be joined to ceramics without the ceramic materials being metallized. Some of these silver-base filler metals (Cusil and Incusil, Wesgo Metals) are ductile and adaptable to brazing metals to such materials as Si3N4, partially stabilized zirconia, transformation-toughened Al2O3, and SiC as well as many other refractory materials (Ref 40, 42). Geometrical Considerations. Because most metals and their alloys have a coefficient of thermal expansion greater than most ceramics, rigid joints between ceramics and metals are stressed on cooling. The magnitude of residual stresses depends on:
• The moly-manganese process: a multistep sealing process in which the ceramic surface is metallized and plated with one or two metals before brazing can take place. The operations are conducted at a high temperature in a controlled atmosphere of H2. Hydrogen firing may discolor some ceramics and produce conductive surfaces. Despite the number of steps required to produce a seal, the molymanganese process can be automated quite readily, and minor deviations in the process variables can be tolerated (Ref 38–42). • The active-hydride process: essentially a single-step process in which hydride reduction and brazing proceed simultaneously. Joining in a vacuum or in a controlled atmosphere of H2 or an inert gas is accomplished at relatively low temperatures, permitting a fast brazing cycle. This process is more difficult to automate than the moly-manganese process, particularly if the joints are produced in a vacuum. Careful control must be exercised in coating the ceramic with the hydride. Even though the process is considered a onestep process, the hydride process has been supplanted by the active-metal process. • The active-metal process: may be a one-step operation like the active-hydride process. Joining proceeds at high temperatures in a vacuum or in a controlled atmosphere; vacuum joining is not readily automated.
Consideration of thermal expansion mismatch is often not necessary when using soft solders or soft-metal components. As the temperature capability of the joint is raised by using more refractory metals and higher brazing temperatures, however, not only is there a greater mismatch on cooling to ambient temperature, but there is also a lesser ability to relax stresses. It is therefore desirable to try to match the thermal expansions of ceramic and metal over the temperature range from ambient to the brazing temperature and to ensure that the ceramic components are thick walled compared with the metal attached (Fig. 8.18a). It is desirable to place the ceramic in slight compression by allowing the metal to clamp down on it by relative thermal contraction from the brazing temperature, as in a disc seal (Fig. 8.18e). If the metal component is made too stout, the clamping stress produces excessive axial tension in the ceramic and weakens the product. In the tube seal (Fig. 8.18e), the stresses are predominantly in shear at the interface between ceramic and braze. Here, a successful joint requires a strong metallized layer and a strong braze not subject to embrittlement. A balancing ring of ceramic may be required (Ref 40). With internal seals involving, for example, tubes or rods (Fig. 8.18c, d), the metal ideally should be closely matched to the ceramic in thermal expansion coefficient. Thin-walled tubes of soft metal can be sealed without difficulty, because the metal yields on cooling (Ref 40). With solid rods, the stresses developed are
New active filler metals are constantly undergoing changes and modifications in composition to meet the ever-demanding requirements
• • • •
Joint geometry Relative thickness of ceramic and metal Ability of metal and braze to relax stresses Temperature at which joint solidifies
332 / Brazing, Second Edition
Tube or rod Thin matched expansion metal cap
Ceramic
(a) Ceramic balancing ring (not always essential)
Thin soft copper Ceramic (b)
Thin walled metal tube
Soft metal pin, cu, pt, max. diam. 0.5 mm (0.019 in)
Ceramic
Internally metallied hole in ceramic (d)
(c)
Low expansion metal pin, W or Mo
Ceramic
High expansion tube placing ceramic disc in compression
(e)
Fig. 8.18
Various types of seals (braze shown dark in exaggerated thicknesses). (a) Leadthrough joints for heavy section conductors. (b) Large-bore tube and disc joints. (c) Tube through a disc. (d) Soft pin in a disc. (e) Hard pin in disc seal. Source: Ref 40
Chapter 8: Joint Design / 333
rather higher. Soft pins of copper or platinum can be brazed into metallized holes in a ceramic disc, such as Al2O3, provided they do not exceed approximately 0.5 mm (0.02 in.) in diameter; otherwise, the metallized layer splits. For larger pins, a low-expansion metal such as tungsten or molybdenum can be used. These contract less than most ceramics on cooling from the brazing temperature and place the interface in compression. They also, however, put the ceramic into hoop tension, which has to be balanced by an outer compression seal on the ceramic disc. This presents no great difficulty, because the outer seal can be used to fix the component to a structure. Tubular bore seals consisting of a thin, ceramic section hermetically sealed to metal members at each end for space power alternators delivering 300 W of electrical power for the system have used designs discussed previously and in Fig. 8.18, as well as active-metal filler metals (Ref 24, 43, 44). Ceramic-Ceramic and Ceramic-Metal Applications. The applicability of the aforementioned brazing methods to marine diesel engine composites is shown in Fig. 8.19 and 8.20. Figure 8.19 shows a basic outline of a ceramic-bonded exhaust valve in which a sintered ceramic component is solid-state diffusion bonded to the valve face. From the result of an exhaust valve damage-simulator test, the exhaust valve in which Si3N4 is used as the face is expected to have a burnout resistance ten to
Fig. 8.19
Ceramic-bonded exhaust valve. Source: Ref 45
fifty times higher than the conventional value (stellite/ stainless steel) (Ref 46, 47). Figure 8.20 shows the ceramic-bonded fuel nozzle being developed. When SiC is solid-state diffusion bonded to the atomizer seat of the nozzle, its wear resistance is several tens of times higher than that of the conventional fuel nozzle. In view of the fact that the working temperature of the seat zone is relatively low, a method of bonding with aluminum as the insert metal is under investigation (Ref 47). The application of the ceramic-metal bonding technology to marine diesel engines described previously may also be applied to the wheel shafts of superchargers and gas turbines and to high-temperature machine components in gas turbine combustion chambers and elsewhere. Titanium-containing filler metals were used to join both 94 and 99+% Al2O3 compositions. Resulting tensile strengths of 77 to 110 MPa (11 to 16 ksi) compared favorably with conventional molybdenum-manganese metallizing (Ref 48). The two different titanium-containing filler metals that were used were essentially copper-silver eutectic compositions that contain small (1 to 3) percentages by weight of titanium. The basic difference between the two filler metals was that one contained 10 wt% In, whereas the second filler metal contained no indium. In order to verify the filler metal as an active braze alloy (ABA), a component was selected (Ref 49) to see how well the component (94% Al2O3) performed when brazed with the
334 / Brazing, Second Edition
indium-containing filler metal. The component chosen was a 94wt% Al2O3 ceramic header with two Mo-Al2O3 cermet electrical feedthroughs. Attached to the connector end of this header were two copper contacts, which were subsequently brazed to the Mo-Al2O3 cermet surface. Shear test results (Ref 50) showed that the filler-metal preforms with 10% In yielded the highest shear strength. As a result, it was found that ABAs provide a simplified method of joining Al2O3 ceramics. Second, comparable tensile strengths can be obtained from ABAs as from conventional molybdenum-manganese metallizing techniques. The ABA sealing results in the migration of titanium from the bulk braze to the ceramic (Al2O3) surface. Filler-metal systems have been developed for joining partially stabilized ZrO2 to nodular cast iron (NCI). The process was developed for advanced-design diesel engines and was termed the active substrate process (Ref 51). These joints call for low brazing temperatures, to avoid the loss of properties of the ZrO2 and the NCI, both of which were heat treated. The process consists of vapor depositing a layer of titanium onto the ceramic, then brazing at 735 °C (1355 °F) with a 60Ag-30Cu-10Sn filler metal. Subsequent work (Ref 51) has shown that by electroplating NCI with copper, enhanced wetting occurs. Adding a transition piece minimizes strain on the ceramic (Fig. 8.21). Tolerances to thermal cycling and shock (argon quench from braze) resulted in no failures after 24 thermal cycles of 375 to 600 °C
(710 to 1110 °F) at average heating and cooling rates of 56 °C/min (100 °F/min) (Ref 51). Shear strengths of 136 MPa (20 ksi) were attained for the NCI-ZrO2 joints. In regard to the transition piece to minimize thermal strain on the ceramic in the joints between the NCI and the ZrO2, a titanium metal transition piece was introduced. The results showed that the thermal expansion coefficient differential at the ceramic surface was held down, and differential strain during braze cooldown was minimized. Testing of the ZrO2NCI joints showed that the various braze interfaces had excellent shear strength and resistance to thermal shock and cycling. In addition, the process can be applied for joining a number of other ceramics and metallic materials for structural applications, including Al2O3 and dispersion-toughened Al2O3 for the ceramic component, and cast iron, titanium, Nb1Zr alloy, and TiC-Ni/Mo cermet as the nonceramic component.
Fig. 8.21
Fig. 8.20
Ceramic-bonded fuel nozzle. Source: Ref 45
Schematic drawing of principal approach of active substrate process. NCI, nodular cast iron; PSZ, partially stabilized zirconia; FM, filler metal; TP, transition piece. Source: Ref 51
Chapter 8: Joint Design / 335
Another application illustrating the use of a transition piece is shown in Fig. 8.14. If the thermal expansion difference between metal and ceramic is large, brazing with a graded seal can be used. For instance, this process has been used to join Si3N4, molybdenum, and stainless steel rings together simultaneously (Fig. 8.14). A two-step (step braze) joining technique can also be used, with a stainless steel/molybdenum braze first, followed by a lower-temperature braze to a ceramic ring with a lower-temperature filler metal. This technique minimizes the joint stresses on cooling to room temperature. Finally, a tapered joint for joining a Si3N4 turbocharger blade to a 410 stainless steel shaft has been successfully used (Ref 52). The turbocharger brazed joint should withstand up to 400 °C (750 °F), and the tapered joint allows the alignment of both ceramic and metal shaft axes. Recent work in the development of a hightemperature ceramic-to-metal seal was disclosed, whereby 52 procedures were developed for fabricating vacuumtight metal-to-ceramic ring seals between Inconel 625 and MgO3Y2O3 (wt%) tubes metallized with CaO29Al2O3-35SiO2 (wt%) glass containing 50 vol% molybdenum filler. A filler metal of Au25Pd-25Ni (wt%) was found to be the most reliable braze for joining Inconel to metallized MgO-3Y2O3 bodies. Another filler metal, Au34Pd-36Ni (wt%), has also been used successfully. The temperature program for brazing was significant: • Heat to 928.6 °C (1703.5 °F) in 4.5 h • Increase temperature by 27 °C (48 °F) over 5 min interval • Cool rapidly to 782 °C (1440 °F) and hold 1h • Cool to 582 °C (1080 °F) at 67 °C/min (120 °F/min) • Furnace cool from 582 °C (1080 °F) to room temperature The reliabilities of joint design and processing procedures similar to those mentioned previously, the material systems evaluated and proven to be successful, and the leak test procedures for 3000 h without failures in vacuum or air resulted in a prototype electrical feedthrough. With the advent and commercial availability of Ag-Cu-Ti filler metals with the Ag-Cu-In-Ti family of filler metals, the one-step vacuum
brazing of steels, superalloys, and ceramics became practical. Step brazing, possible with the previously mentioned filler metals, is a process in which the subsequent braze is made with a lower-melting filler metal, such as AgCu-In-Ti, after an initial braze of Ag-Cu-Ti. In selecting the filler-metal composition, several criteria must be met. The filler metal must have suitable ductility, allow sufficient wetting on both ceramic and metal surfaces, and allow controlled flow to a preplaced position. Minimum filler-metal flow (blushing) on both the ceramic and base surface is required. The degree of blushing depends on the filler metal and metal-substrate compositions, brazing temperatures, and atmosphere. The thermal expansion of the base metal and the ceramic should follow similar behavior from room temperature to the solidus of the filler metal. Consequently, the typical brazing cycle is as follows: • Heat to approximately 50 °C (90 °F) below the solidus temperature • When melting begins, hold at this temperature for a given time until all parts, including braze fixtures, reach a uniform temperature • Increase temperature above liquidus (25 to 50 °C, or 45 to 90 °F) to obtain complete melting • Hold at this temperature up to 10 min, then cool Sandwich Seal Joining. Because the major problem in joining ceramics to metals is primarily their differences in thermal expansion, steps must be taken to alleviate the variation and differences (Ref 53, 54). In most cases, the metal parts to be joined are small (less than 12.7 mm, or 0.5 in., in diameter) and the stress problems are minimized if a ductile filler metal is selected. If the parts being joined are larger (up to 127 mm, or 5 in., in diameter or greater), the
Alumina ring Cusil ABA Cupro nickel Cusil ABA
Alumina cylinder
Assembly
Fig. 8.22
Brazed assembly
Ceramic-to-ceramic sandwich seal. ABA, active braze alloy. Source: Ref 52
336 / Brazing, Second Edition
problem becomes more complicated. The combination of low brazing temperature, ductile filler metal, and proper engineering design is required for successful joining (Fig. 8.22, 8.23). The sandwich seal in Fig. 8.22 increases joint reliability and uses a backup ceramic ring with two purposes: it eliminates the bending movement in the metal and distributes the shear stress equally between the two ceramic faces. In Fig. 8.23, the metal part is attached edgewise to permit concentric distortion, with stress distributed across the total ceramic face by forming a full filler-metal fillet.
Case Histories and Problem-Solving Examples Example 1: Clearances and Applying Filler Metal. A firm was brazing cobalt-base valve seats to carbon steel valve bodies. They applied BNi-2 paste filler metal 360° around the top of the braze joint. However, there was a problem in judging the quantity of filler metal required for this operation. At times, there was excess filler metal, and, at other times, there was not enough. The shop personnel, therefore, were questioning whether there was a way to determine how much filler metal should be applied, or is there a better way of applying it? Some of the inherent variables that can affect the amount of filler metal necessary during the processing of the valve bodies include variations in the furnace atmosphere quality, heating rate, furnace loading, and filler metal application and quality inspection. For clearance variations, it was determined that the BNi-2 filler metal joint clearance was 0.03 to 0.10 mm (0.001 to 0.004 in.). Therefore,
430 Stainless steel end cover Cusil ABA
Alumina cylinder
Assembly
Fig. 8.23
Brazed assembly
Ceramic-to-ceramic sandwich seal. ABA, active braze alloy. Source: Ref 52
the operator would have to apply 300% more filler metal for the 0.10 mm (0.004 in.) clearance than when applying filler metal to the 0.03 mm (0.001 in.) clearance. Therefore, it can be expected that there will be a slight excess of filler metal with a clearance of 0.10 mm (0.004 in.), and a much larger excess of filler metal when the parts only have a clearance of 0.03 mm (0.001 in.). Example 2: Choosing the Right Filler Metal. An auto engine manufacturer uses BNi7 filler metal in lieu of copper for an auto engine compartment brazement. The question of concern is, does the low ductility of BNi-7 make it unsuitable when compared to the ductility of copper? Although it is true that none of the nickelbase filler metals brazed for a short time or at low temperature have the ductility of copper, the nickel-base filler metals processed to take advantage of the diffusion brazing process will produce a joint that is adequately strong. BNi-7 has been satisfactorily used to manufacture many thousands including those used in tank heat exchangers and in the automotive, nuclear, vacuum, and aircraft industries. Example 3: Partial Brazing Checklist to Secure Good Braze Quality. The solenoid body being brazed had a plate and top part made from 1018 steel, with a 304L tube brazed between them. The manufacturer examined four major areas to improve his braze quality: cleanliness, atmosphere, knurling, and the filler metal form and placement in order to overcome his problems.
REFERENCES
1. M. Schwartz, Brazing, ASM International, 1987 2. R.M. Trimmer and A.T. Kuhn, The Strength of Silver-Brazed Stainless Steel Joints—A Review, Brazing Soldering, No. 2, spring 1982, p 6–13 3. Welding, Brazing, and Soldering, Vol 6, Metals Handbook, 9th ed., American Society for Metals, 1983, p 941–944 4. H. Zhuang, E. Lugscheider, and J. Chen, “Wide Gap Brazing of Stainless Steel with Nickel-Based Brazing Alloys,” Document SCIA-B-133, International Institute of Welding Meeting, Sept 1985 (Strasbourg, France)
Chapter 8: Joint Design / 337
5. Met. Constr., Data Sheet Series 3, March 1986, p 166–168 6. H.H.H. Watson, Fluid-Tight Joints for Exacting Applications, Weld. Met. Fabr., Sept 1976, p 491–495 7. Nicrobraz News, Vol 2 (No. 1), winter 1993, p 1–5 8. W.P. Zhuang and T.W. Eager, High Temperature Brazing by Liquid Infiltration, Weld. J., Dec 1997, p 526–531 9. G. Humpston and D.M. Jacobson, Principles of Soldering and Brazing, ASM International, 1993, p 47–50 10. S. Liu, D.L. Olson, G.P. Martin, and G.R. Edwards, Modeling of Brazing Process That Uses Coatings and Interlayers, Weld. J., Vol 70 (No. 8), 1991, p 207–215 11. W.D. MacDonald, “Kinetics of Transient Liquid Phase Bonding,” Ph.D. dissertation, Massachusetts Institute of Technology, 1993 12. W.D. Zhuang, “Applications of Metal Powders for Large Gap Joining,” Ph.D. dissertation, Massachusetts Institute of Technology, 1997 13. E. Lugscheider, V. Dietrich, and J. Mittendorff, Weld. J., Vol 40 (No. 2), 1988, p 47–51 14. P.R. Mobley and G.S. Hoppin, Weld. J., Vol 40 (No. 6), 1961, p 610–617 15. H. Zhuang, J. Chen, and E. Lugscheider, Weld. World, Vol 24 (No. 9/10), 1986, p 200–208 16. E. Lugscheider, T. Schitney, and E. Halmoy, Weld. J., Vol 68 (No. 1), 1989, p 9–13 17. E. Lugscheider and T. Schitney, Brazing Soldering, Vol 14, spring 1988, p 27–29 18. A. Sakamoto, C. Fujivara, T. Hattori, and S. Sakai, Weld. J., Vol 68 (No. 3), 1989, p 63–71 19. E. Lugscheider and T. Cosack, Weld. J., Vol 67 (No. 10), 1988, p 215–221 20. S.K. Tung and L.C. Lim, Wide-Gap Brazing with Prepacks of Nickel-Base Braze Mixes, Mater. Sci. Technol., Vol 11 (No. 9), Sept 1995, p 949–954 21. L.C. Lim, W.Y. Lee, and M.O. Lai, Nickel-Base Wide-Gap Brazing with Preplacement Technique, Part 1: Effect of Material and Process Parameters on Formation of Macrovoids, Mater. Sci. Technol., Vol 11 (No. 9), Sept 1995, p 995– 960
22. How to Specify Proper Joint Clearance for Brazing Dissimilar Metals, Ind. Heat., Nov 1961, p 2162–2164 23. L.A. Biagi, G.W. Koehler, and J.A. Patterson, The Fabrication and Brazing of 15 A, 120 keV Continuous Duty Accelerator Grid Assemblies, Weld. J., Vol 61 (No. 10), Oct 1980, p 33–35 24. H.E. Pattee, Joining Ceramics to Metals and Other Materials, WRC Bulletin 178, Welding Research Council Nov 1972 25. M.M. Schwartz, Ceramic Joining, ASM International, 1990 26. R.E. Loelunan and A.P. Tomsia, Joining of Ceramics, Am. Ceram. Soc. Bull., Vol 67 (No. 2), 1988, p 375–380 27. C.H. Bates, M.R. Foley, G.A. Rossi, et al., Joining of Non-Oxide Ceramics for High-Temperature Applications, Am. Ceram. Soc. Bull., Vol 69 (No. 3), 1990, p 350–356 28. S.M. Johnson, “The Formation of High Strength Silicon Nitride Joints by Brazing,” Report D88-1208, SRI International, Menlo Park, CA, Sept 1987 29. M.G. Nicholas and D.A. Mortimer, Ceramic Metal Joining for Structural Applications, Mater. Sci. Technol., Vol 1, 1985, p 657–665 30. R.E. Loehman, Interfacial Reactions in Ceramic-Metal Systems, Am. Ceram. Soc. Bull., Vol 68 (No. 4), 1989, p 891– 896 31. M.G. Nicholas and R.M. Crispen, Brazing Ceramics with Alloys Containing Titanium, Ceram. Eng. Sci. Proc., Vol 10 (No. 11–12), 1989, p 1602–1612 32. N. Iwamoto, Y. Makino, and H. Miyata, Joining Silicon Carbide Using NickelActive Metal (or Hydride) Powder Mixtures, Ceram. Eng. Sci. Proc., Vol 10 (No. 11–12), 1989, p 1761–1767 33. M. Hare, R.F. Keller, and H.A. Meneses, “Electroformed Ceramic-to-Metal Seal for Vacuum Tubes,” TR453-3, Contract DA-36-039-SC-73138, Stanford University, Menlo Park, CA, 17 Nov 1958 34. R. Bondley, Metal-Ceramic Brazed Seals, Electron., Vol 20 (No. 7), 1947, p 97– 99 35. C.S. Pearsall and P.K. Zingeser, “Metal to Nonmetallic Brazing,” Technical Report 104, MIT Research Laboratory of Elect., Cambridge, MA, 5 April 1949
338 / Brazing, Second Edition
36. F.C. Kelly, Metallizing the Bonding of Nonmetallic Bodies, U.S. Patent 2,570, 248, 9 Oct 1951 37. C.W. Fox and G.M. Slaughter, Brazing of Ceramics, Weld. J., Vol 43 (No. 7), July 1964, p 591–597 38. I.G. Kutzer, Joining Ceramics and Glass to Metals, Mater. Des. Eng., Jan 1965, p 106–110 39. Brazing Manual, 3rd ed., American Welding Society, 1976, p 262–263 40. R. Morrell, Joining to Other Components, Part 1: An Introduction for the Engineer and Designer, Handbook of Properties of Technical and Engineering Ceramics, Her Majesty’s Stationery Office, London, 1985, p 267–278 41. G.R. Van Houten, Ceramic-to-Metal Bonds, Mater. Des. Eng., Dec 1958, p 112–114 42. C.R. Weymueller, Braze Ceramics to Themselves and to Metals, Weld. Des. Fabr., Aug 1987, p 45–48 43. H. Mizuhara and E. Huebel, Joining Ceramic to Metal with Ductile Active Filler Metal, Weld. J., Vol 65 (No. 10), Oct 1986, p 43–51 44. G.G. Hoop, “Generator Development, SPUR Program, Part 11, Generator Stator Bore Seal,” TDR-63-677-Part 2, Contract AF33(657)-10922 and AF33(615)-1551, Westinghouse Electric Corp., Feb 1967 45. T. Yamada et al., Diffusion Bonding SiC or Si3N4 to Nimonic 80A, High Temp. Technol., Vol 15 (No. 4), Nov 1987, p 193–200
46. T. Yamada et al., “Development of Ceramic Exhaust Valves,” D-82, 16th CIMAC, 1985 47. T. Yamada et al., collected abstracts, 33rd Meeting and Symposium of the Marine Engineering Soc., Japan, 1986 48. R.T. Cassidy et al., “Bonding and Fracture of Titanium-Containing Braze Alloys to Alumina,” Monsanto Research Corp., MLM-3431 (OP) and MLM-3394 and DE87002197 and DE87009195, U.S. Dept. of Energy Contract De-ACOE76DP00053, Oct 1987 and Oct 1986 49. H. Mizuhara, Vacuum Brazing Ceramics to Metals, Adv. Met. Process., Vol 131 (No. 2), Feb 1987, p 53–55 50. H. Mizuhara and K. Mally, Ceramic-toMetal Joining with Active Brazing Filler Metal, Weld. J., Vol 64 (No. 10), Oct 1985, p 27–32 51. J.P. Hammond, S.A. David, and M.L. Santella, Brazing Ceramic Oxides to Metals at Low Temperatures, Weld. J., Vol 67 (No. 10), Oct 1988, p 227–232 52. The Promise of Ceramics, Adv. Met. Process. Met. Prog., Vol 131 (No. 1), Jan 1987, p 44–50 53. B.J. Dalgleish, M.C. Lu, and A.G. Evans, The Strength of Ceramics Bonded with Metals, Acta Metall., Vol 36 (No. 8), Jan 1988, p 2029–2035 54. M. Naka et al., Influence of Brazing on the Shear Strength of Alumina-Kovar Joints Made with Amorphous Cu50Ti50 Filler Metal, Trans. JWRI, Vol 12 (No. 2), 1983, p 181–183
Brazing Second Edition Mel M. Schwartz, p339-346 DOI: 10.1361/brse2003p339
Copyright © 2003 ASM International® All rights reserved. www.asminternational.org
CHAPTER 9
Evaluation and Quality Control of Brazed Joints INSPECTION OF BRAZED ASSEMBLIES involves many factors. Of fundamental importance are the principles of the entire brazing process and of the brazing operation itself, including the basic physical and metallurgical properties of brazed joints and base metals, testing methods, interpretation of drawings, and specifications. Inspection of the completed assembly or subassembly is the last step in the brazing process and is essential for ensuring satisfactory and uniform quality of the brazed unit. This operation also provides a means of evaluating the adequacy with which the prior steps in the process have been carried out with regard to ultimate integrity of the brazed joint (Ref 1). This chapter outlines the requirements and methods associated with the inspection of brazements and emphasizes the incorporation of these requirements into the overall quality system. This is especially important, because each brazing application has a unique set of requirements. The process capabilities, technical specifications, practical limitations, and the applicability of the selected process to a particular manufacturing environment must be thoroughly understood. Specifically, acceptance limits, design limitations, and both nondestructive and destructive inspection techniques are reviewed (Ref 2).
Design and Quality System The design of the brazement is extremely relevant to inspection and should be such that the
completed joints can be inspected by techniques that provide the degree of part reliability suitable for the service requirements. If the part is intended for a critical application, it must be designed not only so that brazements of the required quality can be made, but also so that the brazed assemblies can be readily inspected to ensure that those quality requirements have been met. The inspection methods chosen to evaluate the brazing procedure and the serviceability of the product are largely dependent on the service requirements of the brazed assembly. In many cases, inspection methods are specified by the ultimate user or by regulatory codes or standards of quality for brazed joints; the same approach used in setting up standards for any other phase of manufacturing should be employed. These standards should be based, if possible, on known requirements that have been established by prior service tests or history. Inspection of brazed joints may be conducted on either test specimens or finished brazed assemblies. Although process controls and examination of test specimens may be an acceptable method for control of noncritical components, they are not a substitute for examination by nondestructive testing of the actual components intended for critical applications. A critical component is defined as one whose primary failure would cause significant danger to persons or property or would result in significant operational penalties. Persons responsible for design, manufacture, and inspection of critical brazed components should obtain the American Welding Society (AWS) Recommended Practices Document (Ref 3).
340 / Brazing, Second Edition
Tests may be nondestructive or destructive. Each method of testing has a particular capability to disclose one or more brazing defects. Brazing defects are of three general types: • Those associated with drawing or dimensional requirements • Those associated with structural discontinuities in the brazed joint • Those associated with the filler metal or the brazed joint Nondestructive inspection methods are applicable to brazements made by all of the brazing processes and are essential to the quality control of brazements for critical applications. The size, shape, complexity, and degree of critical application dictate the particular inspection method or methods that are most suitable. If no accurate and dependable method of inspecting a critical brazed joint can be found, either the part should be redesigned to permit inspection or another, more inspectable joining technique should be used. Neither periodic destructive inspection nor process qualification requirements are completely acceptable substitutes for nondestructive inspection of the actual hardware entering critical service. When the acceptance limit for any type of brazed defect is defined, the following must be considered: shape; orientation; location in the brazement, including surface versus subsurface; and the relationship to other imperfections. Judgments for disposition of discrepant components should be made by persons competent in the fields of brazing metallurgy and quality assurance who fully understand the function of the component. Such dispositions must be documented. The documentation that is essential to the quality of the assembly must be prepared and implemented. It should include the assignment of “responsibility, authority, and the interrelation of all personnel who manage, perform, and verify work affecting the quality of the assembly” (Ref 4). The procedure for reviewing the brazing process and modifying the appropriate documentation, as required, must be agreed on. Coordinating such activities with all involved parties helps avoid common pitfalls in the quality system and reduces the chance of a nonconformity in the brazement. Catastrophic failure is most often caused by an essential operation not being done at all, rather than by some essential operation not being done precisely.
The factors that affect the strength, durability, and suitability of a brazed assembly for service can be divided into three categories: design factors, inspectable factors, and process control factors. Design factors include the proper selection of base materials, filler metals, and brazing process; designing for ease of manufacture; and designing so that necessary inspections of the brazed joints can be readily and dependably performed. Inspectable factors include dimensional configuration, cosmetic requirements, and, in most cases, a percentage of the actually brazed joint area. Factors that are usually not inspectable on the completed assembly (except as can be inferred from the percentage of joint area brazed) include brazed joint surface preparation, clearances between faying surfaces of the joint, and the actual brazement temperature profile that occurred during the brazing operation. The existence of this last factor is the reason that the manufacturing process, as well as the brazed product, must be specified and controlled.
Quality Standards for Brazing and Brazing Processes The AWS C3.3 document (Ref 5) was written to provide a basic guide to what must be done to ensure the suitability of a brazed component for a critical application. According to Ref 6, a critical component is one that would either cause a significant danger to persons or property or would result in a significant operational penalty if it were to fail. Although such applications vary widely, there are certain common considerations in materials, design, manufacture, and inspection that must not be overlooked. This document lists and explains these common considerations and the best techniques for dealing with them. These recommended practices describe those procedures that should be followed in the design, manufacture, and inspection of brazed joints for critical components in order to ensure their reliability in service. The recommended procedures represent the best current practice, in the opinion of the AWS, and are necessary for the control of brazed joint quality. At the present state of the art, all practices may not be applicable to all products or all brazing processes. However, when some of these practices are omitted on critical components, it should be the result of a rational decision rather than the result of a lack of knowledge of the best practice (Ref 7).
Chapter 9: Evaluation and Quality Control of Brazed Joints / 341
Other American National Standards Institute (ANSI)/AWS specifications for specific brazing and inspection processes are listed in Table 9.1. Those that are relevant to any particular brazed assembly should be selected and specified on the engineering drawing and related documentation for the assembly. In situations where the specific requirements of these documents are modified or waived, it must be with the written consent of those responsible for product suitability during its service life, as explained in the specification.
Types of Common Defects Nondestructive inspection is used to identify the following types of common defects. The limits of acceptability must be specifically defined. Lack of Fill (Voids, Porosity). Lack of fill can be the result of improper cleaning, excessive clearances, insufficient filler metal, entrapped gas, insufficient temperature, poor or oxidizing atmospheres, stopoff misplacement, insufficient flux, and movement of the mating parts caused by improper fixturing while the filler metal is in the liquid or partly liquid state. This imperfection reduces the strength of the joint by reducing the load-carrying area, and it may provide a path for leakage. Flux entrapment can be found in any brazing operation where a flux is added to prevent and remove oxidation during the heating cycle. The entrapped flux prevents flow of the filler
Table 9.1 Selected American Welding Society (AWS) brazing documents AWS designation
BRM B2.2 ANSI/AWS C3.2 ANSI/AWS C3.3
ANSI/AWS C3.4 ANSI/AWS C3.5 ANSI/AWS C3.6 ANSI/AWS C3.8 AWS A5.8
Title
Brazing Manual Standard for Brazing Procedure and Performance Qualification Standard Methods for Evaluating the Strength of Brazed Joints in Shear Recommended Practices for Design, Manufacture, and Examination of Critical Brazed Components Specification for Torch Brazing Specification for Induction Brazing Specification for Furnace Brazing Recommended Practices for Ultrasonic Inspection of Brazed Joints Specification for Filler Metals for Brazing and Braze Welding
Note: ANSI, American National Standards Institute. Source: The American Welding Society
metal into that particular area, thus reducing the joint strength. The entrapped flux, if corrosive, may severely reduce service life. Noncontinuous fillets are usually found by visual inspection. This type of defect is evidenced by a large void in the fillet. Such voids may or may not be acceptable, depending on the specified requirements of the brazed joint. Base-metal erosion is caused by alloying of filler metal with the base metal during brazing. This results in melting of some of the basemetal constituents, causing undercuts, or the disappearance of the mating surfaces. It may reduce the strength of the joint by changing the composition of the materials and by reducing the cross-sectional area of the base metal. Unsatisfactory Surface Appearance. Excessive filler-metal spread, roughness due to liquation, and excessive filler metal may be detrimental for several reasons. In addition to aesthetic considerations, these defects may act as stress concentrations or corrosion sites or may interfere with inspection of the brazement. Cracks reduce both strength and service life expectancy. They may act as stress raisers, causing premature fatigue failure as well as lowering the mechanical strength of the brazement.
Brazing Process Planning and Control Design and Development Planning. As already noted, the requirements for a serviceable part must be understood so that they are within the capability of the processes selected. The inspection procedures must be considered in design, prototype manufacture, and full-scale production. Performance and service requirements must be both acceptable and achievable. Typical considerations are whether or not the part can be classified as a critical component and whether or not existing technology can be adapted, or new technology developed, in order to inspect the part. The selection of the inspection technique can rely on the performance history of the part or similar parts. The manufacturing facility must be aware of the requirements and must be capable of meeting them at a reasonable cost on an acceptable production schedule. Too often, unrealistically high standards that cannot be met by the processes selected and the available equipment are set during the engineering development phase. They then must be lowered, based on production expediency rather than sound engineering judgment, when manu-
342 / Brazing, Second Edition
facturing cost and schedule commitments must be met. Technical-Business Interface. For a successful development program, team members must understand the activities and constraints of one another. A written assignment of tasks is necessary. Often, interdepartmental or interdisciplinary projects fail at the interface between organizations. Because brazing requires consistent quality from the machining of the detail parts to the application of filler metal to the completion of postprocessing, each of which is usually performed by a separate department or organization, a written document or another form of communication is necessary. Acceptance Limitations and Common Imperfections in Brazed Joints (Ref 8). The basic acceptance limitations that are specifically applicable to brazing must be considered. The relevant factors, when defining the acceptance limit for any type of braze discontinuity, are shape, orientation, location in the brazement (surface or subsurface), and relationship to other discontinuities. Acceptance limits should always be stated as the minimum requirement for acceptability. Judgments as to the disposition of discrepant components should be made by someone who is not only competent in the fields of brazing metallurgy and quality assurance but fully understands the function of the component. Such dispositions should be documented. General Inspection Criteria. The general inspection methods are applicable to brazements made by every brazing process, and they should be used to control the quality of brazements for critical applications. The size, complexity, and degree of the critical application dictate the particular inspection method(s) that is most suitable. If an accurate and dependable method of inspecting a critical brazed joint cannot be found, then the part should either be redesigned to permit inspection or another more readily inspected joining technique should be used. Neither periodic destructive inspection nor process qualification requirements are completely acceptable substitutes for the nondestructive inspection of actual hardware entering critical service. Brazements are inspected prior to assembly by nondestructive methods for quality and conformance to specifications. Brazing procedures should be qualified to meet specification requirements using both nondestructive and destructive inspection methods.
Nondestructive Inspection Nondestructive inspection testing techniques include visual, leak, radiographic, ultrasonic, penetrant, thermal transfer, and other methods, each of which is described subsequently. Detailed information about nondestructive inspection methods is provided in Ref 9. Visual inspection is the most widely used of the nondestructive methods for inspecting brazed joints. It is performed both with and without magnification and can be used to evaluate the external evidence of voids and porosity, noncontinuous fillets, base-metal erosion, surface cracks, fillet size and shape, and general braze appearance. If a brazement can be inspected on only one side, then there is no guarantee that a satisfactory joint exists, even though there is a good fillet. Visual inspection cannot reveal internal imperfections in the brazed joint, such as entrapped flux, porosity, lack of fill, and internal cracks. If there is any chance for misinterpretation, then the inspector should be provided with samples, photos, or sketches to ensure that he knows precisely what conditions he is supposed to recognize and reject. Leak testing recognizes and can constitute either pressure or vacuum testing. This inspection method is recommended when gas or liquid tightness is required of the brazed joints. Assemblies that are subjected to low pressure but not governed by the requirements of the American Society of Mechanical Engineers (ASME) Boiler and Pressure Vessel Code can be tested with air. “Low pressure” is defined in the code and is specific to the particular application. The user must carefully evaluate the situation to determine what leak-test standards are appropriate. Assemblies that are subjected to high pressure and are governed by the code requirements should be tested hydrostatically after initial testing with low-pressure air. Both test pressure and duration should be specified to provide a realistic measure of joint quality. It should be emphasized that the leak-tightness of a joint, when it is originally tested, does not necessarily ensure continued integrity in service after subjection to various and repeated loadings. Pressure testing with air can be accomplished in several ways: • Close all openings, pressurize the assembly, submerge in water, and note any signs of leakage indicated by rising air bubbles.
Chapter 9: Evaluation and Quality Control of Brazed Joints / 343
• Close all openings, pressurize the assembly, brush a soap solution or commercially available indicator over the joint area, and note the location where bubbles are formed. • Close all openings, pressurize the assembly, lock the air under pressure within the assembly by closing the air inlet source, and note any change in trapped gas pressure over a period of time (corrections for temperature and barometric pressure may be necessary). Pressure testing with helium is frequently preferred because of the relative ease of finding very minute leaks. A mass spectrometer is used to detect helium leakage. When testing for very small leaks, it is usually important to thoroughly remove all liquid or vapors from the assembly by a suitable drying operation. Purging with dry gas while heating the assembly above the boiling point of the liquid is one such method. Vacuum testing is generally used when checking assemblies that are part of refrigeration equipment, as well as small assemblies in situations where it is imperative that the most minute leak be detected. The mass spectrometer is used in vacuum testing, with helium as the sensing medium. The assembly being tested can either be placed in a container having a helium atmosphere, or the helium atmosphere or the helium can be flushed over its surface. The mass spectrometer is connected to the vacuum side of the assembly, preferably between the auxiliary diffusion pump and the mechanical pump. If mass spectrometer helium leak testing is to be used, then care should be taken to prevent the use of liquid on the joint prior to the test, because of the potential for masking leaks. Again, it should be emphasized that the leak-tightness of a joint, when it is originally tested, does not necessarily ensure continued integrity in service. Proof testing subjects the completed joint to loads that exceed those loads that will be applied during its service life. The proof-testing loads can be applied by hydrostatic methods, tensile loading, spin testing, or other methods. Occasionally, it is not possible to ensure a serviceable part by any other nondestructive method of inspection, in which case proof testing becomes the most satisfactory method. Much caution should be used when specifying proof testing as the primary method of inspecting brazed joints for critical applications. Because proof testing does not inspect braze quality but, instead, applies a one-time load that may not closely sim-
ulate all the conditions encountered in service, it may not accurately predict service life, especially if cyclic loadings are encountered in service. Acoustic emission (Ref 10), when used alone or in conjunction with proof testing, represents another nondestructive inspection technique for brazed joints. The acoustic emission source is the elastic energy release within the material or component as it undergoes deformation from an applied load, for example, cracking of a stressed brazed joint. Because it is a dynamic process and nondirectional, it can pick up faults from any location in the component when the sensor is properly placed. The complete acoustic emission system includes these major components: sensor, preamplifiers, main amplifiers, measurement circuitry, computer(s) for data acquisition and analysis, and accessory items as required by the application. The acoustic emission system must be carefully selected with respect to each application. Because the acoustic emission phenomenon is an irreversible process in materials, it must be accounted for in the overall inspection procedure. Radiographic inspection is used extensively to examine brazed joints. However, this method is usually not selected for brazed joints requiring radiographic sensitivities of less than 2% and where thickness and x-ray adsorption ratios do not permit differentiation of the joint. Possible differences in adsorption characteristics of base metals and filler metals should always be considered when interpreting the radiographic results. A joint that may be inspectable when brazed with silver filler metal may not be inspectable when brazed with copper or nickel filler metal, because of the differences in the x-ray characteristics of the filler metals. Special techniques are required to reliably inspect joints of varying thicknesses. Views that show two sides of a cylindrical joint, such as a pipe fitting, should be interpreted with particular care. Radiography is used to find internal flaws, cracks, and braze voids that are not discernible by those techniques that are limited to surface observation. It is also useful in determining the depth of surface cracks and voids, if they are within the limits of the x-ray process. Radiographic inspection shows the presence of filler metal in a joint but does not necessarily
344 / Brazing, Second Edition
indicate the presence of a metallurgical bond between the base metal and filler metal. (These must be ensured by process controls.) It should be noted that a radiograph that shows no voids may indicate either 100 or 0% joint coverage, or it can merely indicate a poor x-ray technique. Careful steps should be taken to ensure that film readers are aware of this and are particularly suspicious of joints that appear to show 100% coverage. Ultrasonic inspection of brazed joints generally depends on the reflection of sound waves by surfaces. A transducer emits a pulse and receives echoes in return. In the standard brazed lap joint, these echoes come from the front and rear surfaces of the part at areas of completely bonded joints. At defective areas, the unbrazed faying surfaces of the joint cause a third echo, located between the first two. These echoes either can be displayed on an oscilloscope, or the defect signal can be used to trigger the pen of a recording device to produce a facsimile of the joint that shows bonded and unbonded areas. This inspection method is sensitive to setup variables, part configuration, and materials. Therefore, it should be considered a comparative method of inspection that requires a reference standard that is identical to the part being inspected, including defects of known configuration. This standard is used to set up and calibrate the equipment at specified intervals. The applicability of this inspection method primarily depends on the design of the joint and the configuration of the adjacent areas of the brazed assembly. Ultrasonic techniques have been developed and used for a wide range of brazing applications, and they often provide one of the best methods for evaluating joint quality. Dye and fluorescent penetrant inspections are used for the detection of imperfections that are open to the surface of the brazed joints in both magnetic and nonmagnetic materials. These techniques should not be used to inspect brazed fillets. Cracks and surface porosity can be detected in brazed joints that have been machined. The interpretation of brazed fillets is difficult because of small irregularities. Incomplete flow and partial fillets can also be observed. These inspection methods should not be used if subsequent repairs are to be made, because the penetrants are often either difficult or impossible to remove completely.
Thermal-transfer inspection can be used on certain specific cases. For example, brazed aircraft propeller blades can be photographed a few minutes after leaving the furnace, while still hot. The covering of the skins appears bright red in areas where it has been brazed to the reinforcing rib but is a much darker red, or black, in areas that are unbrazed. A method of inspecting brazed honeycomb panels uses powdered or liquid materials with low melting points to indicate the differences in heat-transfer characteristics when the honeycomb assembly is placed under infrared heat lamps. Temperature variations cause the liquid to be repelled from warm areas and to act as a heat sink, causing the fluid to flow to each area of a good braze. Holography is lensless, three-dimensional photography. It has been used to detect voids, flaws, unbrazed areas, and laminations. Generally, detection of flaws in skin-to-core brazes in honeycomb structures depends on the ratio of skin thickness to flaw width. Thus, deep flaws can be found, if they are broad enough (Ref 11). Other Techniques. More sophisticated techniques use thermally sensitive phosphors, liquid crystals, and other temperature-sensitive materials. Infrared-sensitive electronic imaging devices with a television readout are now commercially available to monitor temperature differences that are characteristic of variations in braze quality. Careful verification of the specific inspection technique should be done prior to using the technique to accept critical components.
Design Testing, Evaluation, and Feedback Once the brazing method and the inspection technique have been selected, their feasibility and production suitability must be proven. The process capability can be honestly evaluated using actual results. Specific requirements for the selected inspection technique can be documented. The testing process must provide unequivocal evidence that the overall manufacturing process can achieve the part requirements, such as joint strength, corrosion resistance, and thermal cycle fatigue. If these tests produce inconclusive or anomalous results, both on prototype and production assemblies, then the entire procedure must be reevaluated.
Chapter 9: Evaluation and Quality Control of Brazed Joints / 345
When evaluating the success or failure of the manufacturing and quality procedure, the final intended use of the assembly must be considered, whether it is one of a kind or, ultimately, a high-volume product. Results in controlled laboratory conditions may be very different from results on the production floor. When a process is turned over to production, adequate training is essential to ensure that process requirements are understood. All process requirements must be explained, including joint clearance, application of filler metal, application of flux, application of stopoff, brazing cycle, postbrazing operations, accessibility, health and safety, and qualifications and the reasoning behind them. Additionally, a rework procedure and limitations must be integrated into the overall system, including testing to ensure reworked assemblies meet the original assembly requirements. The system for rework should account for pre- and postprocessing steps, and it should clearly identify any characteristics that may differ from the original brazed assembly. Care must be taken to assure that metallurgical and mechanical integrity are maintained and that performance requirements are not compromised. If the standard cannot be verified, then a conscious and well-informed decision must be made to modify the existing brazing process for the rework and disposition of nonconforming parts. The written specification must outline the rework procedure and provide a mechanism for its implementation and documentation.
Destructive Inspection and Testing Methods Destructive methods of inspection are employed to ascertain whether a brazement design meets the requirements of the intended service application. Destructive methods may be used for partial sampling and for checking of the nondestructive methods of inspection by selecting production assemblies at intervals and testing them to destruction. The methods and frequency are usually established by qualitycontrol procedures. Destructive testing methods are used for random or lot testing of brazed joints. In lot testing, a specified percentage of all production is tested to destruction. The results of these tests are assumed to apply to the entire production lot, and the various lots or batches are accepted or
rejected accordingly. When used as a check on a nondestructive method of inspection, a production joint may be selected at regular intervals and tested to destruction so that rigid control of brazing procedures is maintained. Rules for selection, testing, and lot acceptance should be established and strictly followed. Destructive testing methods may be conducted on either test specimens or finished brazements. The three major types of destructive testing methods are (a) metallographic, (b) mechanical, and (c) chemical (Ref 8, 12). Metallographic inspection (Ref 13) requires the removal of sections from the brazed joints and their preparation for macroscopic or microscopic examination. This method is useful for detection of flaws, such as porosity, poor flow of filler metal, excessive base-metal erosion, diffusion of filler metal, and improper joint fit-up, as well as evaluation of the microstructure of the brazed joint. When defects are found, it may be an indication that the brazing procedure is out of control or that improper techniques are being used. Mechanical inspection (Ref 14) encompasses a number of techniques, including peel, tensile, shear, fatigue, impact, and torsion tests. Peel tests are frequently used for evaluating lap joints. One member is held rigid, as in a vise, while the other is peeled away from the joint. This test can be used as a means of production quality control. It can determine the general quality of the bond as well as the presence of voids and flux inclusions in the joint. The permissible number, size, and distribution of these discontinuities depend on the service conditions of the joint and may be limited by applicable codes. This method is not suitable for diffusionbrazed joints using nickel-base filler metals, because such a joint does not peel. Tensile and shear tests are used to determine the strength of a joint in tension or in shear. They are generally used for development rather than production quality control. However, selected production samples can be evaluated by these tests. (Refer to AWS C3.2, “Standard Methods for Evaluating the Strength of Brazed Joints in Shear.”) Fatigue testing under cyclic loading is used to a limited extent. In most cases, it tests the base metal as well as the brazed joint. As a general rule, fatigue tests require long times to complete. Therefore, they are very seldom used for quality control.
346 / Brazing, Second Edition
Impact tests, similar to fatigue tests, are generally limited to laboratory work in determining the basic properties of brazed joints. As a general rule, the standard notch-type specimens are not suitable for brazed joints. Special types of joints may be required to obtain accurate results. Standard butt joints with modified notches also can be used. Torsion tests are occasionally used on brazed joints in production quality control. Examples of products that are tested are studs, screws, or tubular members brazed to thick sections. The base member is clamped rigidly, and the stud, screw, or tubular member is rotated to failure in either the base metal or the brazed joint. Torsion testing is used on a sampling basis to verify the brazing procedure. Chemical inspection is ordinarily used to determine chemical composition or corrosion resistance. Tests for chemical composition are often required to ascertain whether the base metal and filler metal meet specifications or codes. REFERENCES
1. M.M. Schwartz, Brazing, ASM International, 1987, p 438 2. M.B. Vollaro and J.A. Miller, Evaluation and Quality Control of Brazed Joints, Welding, Brazing, and Soldering, Vol 6, ASM Handbook, ASM International, 1993, p 1117–1123 3. “Recommended Practices for the Design, Manufacture, and Examination of Critical Brazed Joints,” AWS C3.3, American Welding Society 4. “Quality Systems—Models for Quality Assurance in Design/Development, Production, Installation, and Servicing,” ANSI/ASQC Standard Q91-1987, Subsection 4.1.2 (Organization), 4.1.2.1 (Responsibility and Authority), ANSI, 1987
5. “Recommended Practices for Design, Manufacture, and Inspection of Critical Brazed Components,” ANSI/AWS C3.3, Forward, AWS Committee on Brazing and Soldering, 1980 6. “Recommended Practices for Design, Manufacture, and Inspection of Critical Brazed Components,” ANSI/AWS C3.3, Section 1 (Introduction), Subsection 1.2 (Definition of a Critical Component), AWS Committee on Brazing and Soldering, 1980 7. “Recommended Practices for Design, Manufacture, and Inspection of Critical Brazed Components,” ANSI/AWS C3.3, Section 1 (Introduction), Subsection 1.1 (Scope), AWS Committee on Brazing and Soldering, 1980 8. “Recommended Practices for Design, Manufacture, and Inspection of Critical Brazed Components,” ANSI/AWS C3.3, Section 5 (Inspection), Subsection 5.1 to 5.38, AWS Committee on Brazing and Soldering, 1980 9. Nondestructive Evaluation and Quality Control, Vol 17, ASM Handbook, ASM International, 1989 10. R.K. Miller, and P. McIntire, Ed., Acoustic Emission Testing, Vol 5, Nondestructive Testing Handbook, American Society for Nondestructive Testing, 1987, p 12–19 11. J.A. Mock, Holographic NDT Helps to Keep Materials Defect-Free, Mater. Eng., Aug 1970, p 50–52 12. J.A. Mock, Guide to Destructive Testing, Mater. Eng., Aug 1977, p 18–29 13. Metallography and Microstructures, Vol 9, ASM Handbook, ASM International, 1985 14. Mechanical Testing and Evaluation, Vol 8, ASM Handbook, ASM International, 2000
Brazing Second Edition Mel M. Schwartz, p347-382 DOI: 10.1361/brse2003p347
Copyright © 2003 ASM International® All rights reserved. www.asminternational.org
CHAPTER 10
Applications and Future Outlook JOINING ADVANCED MATERIALS poses many challenges, some of which are quite special. However, the greatest challenge is to learn from history. For example, consider materials diversity. Modern materials engineers and designers cannot be parochial in their approach to materials selection. Quite the contrary, they must be totally open-minded. Just because metals were used exclusively in the past does not mean that they must be used in the future. Nor does it mean, however, that they should be quickly abandoned in favor of a newer material. This is an easy trap to fall into. It happened in the aerospace industry, and it led to some unnecessary difficulties in joining (Ref 1). Examples of these difficulties include problems joining aluminum-lithium alloys, several superalloys, and intermetallic alloys. The ability to produce reliable brazed metalmetal, ceramic-ceramic, ceramic-metal, and matrix-composite joints is a key enabling technology for many production, prototype, and advanced developmental items and assemblies. Design and performance demands are now being made that previously could not be met readily by existing processes and commercially developed metals, alloys, and ceramics (Ref 2). Considerable attention is now being paid to the potential usefulness of these new materials (Ref 3), due to the improvement in atmospheres (fluxless brazing) (Ref 4), cleaning methods (Ref 5), modeling procedures, filler-metal foils (Ref 6) and their improved properties, nontoxic filler metals (Ref 7), and nondestructive evaluation techniques to verify structural integrity, soundness, and suitability. For example, ceramics are generally more refractory and less dense than metals and can be stronger. However, if these properties are to be exploited, it is generally essential for an adequate joining technol-
ogy to be available, and that technology is brazing (Ref 2, 8–15).
Automation The future of braze processing lies in automation. Automation of the brazing process is the key to maximizing production and quality and minimizing costs. Typical machines that are currently available are capable of automatic brazing by torch, furnace, induction, infrared, and resistance processes. Rotary and shuttle machines automatically apply paste and filler metal, heat the parts, torch braze the assembly, cool the brazement, and remove some or most of the flux. Many engineers and manufacturing specialists are reluctant to try to automate brazing operations for fear of upsetting a well-established process and causing quality problems because of a processing and/or method change. Their attitude, in short, is, “If it’s working and producing parts, don’t change it; don’t upset it.” However, brazing processes, when properly understood, are just as adaptable to increased productivity through automation as any other process; on the other hand, of course, they are also subject to similar limitations. Regardless of how attractive the idea of automating a brazing operation may be, the first step usually is to verify and compare the economics. If automation is economically justified, the next step is to investigate the mechanics of how to automate the operation. Major factors requiring serious analysis for automating brazing applications are noted as follows. Base Metals. The joint base metals (steel, copper, aluminum, etc.) each require review
348 / Brazing, Second Edition
from their strength characteristics, brazing compatibility with each other, compatibility with various filler metals, creep, specific heat, and material heat coefficients. They all interact within the brazing process as well as later during an end-service use of the assembly. Size and mass of the assembly base metals have direct influence on the machine concept, type of heat system used, output, parts handling approach, total machine cost, and number of operators required. Joint requirements, such as strength, fatigue, appearance, shock resistance, service temperatures, corrosion resistance, and other special requirements, deserve examination. Incomplete specifications require resolution. The assembly joint should have proper design, fit-up, and bonding surface area to provide the joint strength requirements. A well-designed self-locating joint provides numerous benefits to the overall brazing operation. This is especially true if the joint can be positioned during the heat cycle to take advantage of both gravity and capillary attraction. Furthermore, self-locating assemblies simplify fixture design and reduce machine cost. Also, such assemblies assist the operator to accurately and repeatedly load each fixture correctly for reduced rejects and increased machine output. Filler metals are available in many standard and nonstandard compositions. A thorough examination should be made to ensure use of the correct filler metal for an application. This review should include joint-design strength characteristics with base metals as well as fillermetal service requirements, corrosion resistance, and ability to wet and flow into the joint formed by the base metals within the specified temperature. Paste filler metals are readily adaptable for automatic machine operation. They can be automatically dispensed in various deposit sizes and configurations. Furthermore, because they contain both filler metal and flux, separate fluxing operations (sometimes requiring additional operators) are eliminated. Wire feed systems are also available for machine operation. Metallurgical considerations should include a review of the various intermetallic reactions during and after the brazing process, such as liquation, phosphorus and sulfur embrittlement, carbide precipitation, stress cracking, and so on. Flux activity plays a critical role within the brazing process. The flux must remove light
oxides from the base metals and provide a path for the filler metals to wet and form the joint. Also, the flux must prevent reformation of oxides on the base metals during the brazing heat cycle. Some fluxes provide activity and oxide protection in very high temperature ranges involving long time cycles. Others are used for mild fluxing applications, such as dilute atmosphere furnace operations. Flux residue removal is an important consideration for many brazing operations. Needless to say, when clean parts enter the brazing process, less residue will remain in the joint area for postcleaning. Furthermore, precleaned parts assist the overall operations and provide consistently acceptable brazed assemblies, especially when the application requires sealed joints. The flux must be compatible with base metals, filler metals, the machine heat system, as well as Environmental Protection Agency (EPA) and Occupational Safety and Health Administration (OSHA) requirements. Many times, brazed assemblies require removal of flux residue due to continued flux activity, or the revival of activity due to the hygroscopic characteristics of the residue. It is advantageous to remove flux residues while parts are still warm from the brazing process and have not hardened. Therefore, including the postcleaning process within the automated brazing machine flow often merits serious consideration. Many companies have introduced ultrasonic cleaning systems requiring less-toxic aqueous cleaning chemicals. Such a system uses ultrasonic cavitation to remove contaminants, including surface oxides, ingrained drawing compounds, entrapped metal chips, and so on within very short time cycles. The process cleans assembly surface areas internally as well as externally. This cleaning approach is especially effective when used on complex assemblies containing multiple brazed joints. See Fig. 10.1 for a conceptual brazing machine and postcleaning operation. Steps in Automating a Brazing Process. The following steps are necessary for successful automation: • Make an economic analysis to justify automation (Ref 16) • Break down the total operation into its several parts (Ref 17, 18) and determine which can be eliminated, which can be combined, and which can be carried on in conjunction with other operations.
Chapter 10: Applications and Future Outlook / 349
• Determine the required type of automated equipment (rotary indexing machine, constant rotary machine, in-line conveyor, inline indexing conveyor, shuttle machine, or racetrack-type conveyor). The types of parts to be brazed and the required production rate will normally determine what type of mechanized equipment will be most feasible. • Apply specialized skills and knowledge to the various design and engineering problems involved in realizing the objectives of automation: mechanics (in the selection of a mechanical movement and also in the design of the part-holding fixture), selection and application of flux and filler metals, type and position of heat application, cooling, additional mechanics of automatic loading and unloading, and electrical control systems. Example 1: Automated Induction Brazing of Hydraulic Pump Shoes. Automation of an induction brazing operation has more than doubled production of shoes for hydraulic pumps used to move control surfaces in commercial airliners. The operation of manually brazing a bronze ring onto the steel shoes in the pumps used to be so difficult and time-consuming that it was the bottleneck of the entire pump-manu-
facturing process (Ref 19). Engineers from Aeroquip-Vickers Limited, manufacturers of the pumps, worked with Lepel Corporation to develop a four-station index machine that has reduced cycle time for the operation from 36 to 40 s to 15 to 18 s. In addition, the new machine has reduced the number of operators required for brazing from two to one. Most of the pumps are from a family of steel shoes that are made from several different materials, including 4140, 4160, and 6150 steel. Because of the critical nature of the application, the material and manufacturing processes used to make the shoes are tracked throughout the manufacturing cycle; for example, a piece of steel is cut off from each batch of material and sent to a metallurgical laboratory to verify its properties. The process begins when the operator loads the part into a nest at the first station. Then, the machine indexes the nest into the second station, where the part is automatically centered inside the induction coil. Within 8 s, the parts are heated to 710 °C (1310 °F), and a ceramic guide moves up against the ring and begins spinning while applying pressure to remove the voids. The next station provides a place for the part to cool under an air blast and set. The fourth
PC-controlled Jib crane
Postclean
Load basket
Rinse tank
Ultrasonic tank
4 position cleaning line
Heating Water cool 5 4
6
+
3 2
7
Water cool
8 1
Filler-metal applicator
Load/unload
Brazing machine
Fig. 10.1
Automatic brazing machine with ultrasonic cleaning. PC, personal computer
Air dry
Preclean
350 / Brazing, Second Edition
and final station is where the operator unloads the part. All moving parts of the machine are enclosed with polycarbonate resin shielding, and the machine contains a light curtain, emergency stop button, and ready light to ensure operator safety. Sensors detect any interruption in air and water used for cooling and automatically stop operation to prevent damage to the machine or workpieces. Interchangeable coils and nests and adjustable cycle times give the equipment the ability to accommodate a wide range of shoe sizes. Power output on the machine can be adjusted from 6 to 30 kW, providing appropriate heat for each size shoe while minimizing power consumption. With a cycle time of only 15 to 18 s, the Lepel machine more than doubled the production rate, from 400 to 1000 units per day in the two-shift operation. The quality of parts produced by the machine has been excellent. Yields typically run approximately 98%, and the quality is actually slightly higher than the previous manual operation. Example 2: Automated Flame Brazing. Many companies today have flame brazing applications in which automation may be beneficial, but there are many considerations to analyze before deciding whether some form of automation might be practical and have a reasonable payback. An inexpensive machine that can rotate the part being brazed while multiple flames heat the part rapidly may be a good choice. If rotating is not suitable for the part, such as two tubes joined to an oblong block, then the machine should have the added capability to oscillate the flames. A machine with optical pyrometers can control the temperature of the braze joint. When using a timed heat cycle, a part might be overheated or underheated. This is because there are many variables affecting the temperature of the part being brazed. Natural gas, often used because of low cost and easy availability, has up to 16% moisture and also fluctuates from 900 to 1100 Btu, which changes the amount of heat. The amount of flux, the viscosity of the flux, the wall thickness variation in the parts, fans, the humidity, the fixture temperature (it increases during production), and the temperature of the production parts (it might have been on a cold floor or near a heater) all change the time required to braze efficiently. An optical pyrometer helps to compensate for all of these conditions by removing the flame when the brazing has
been completed. Also, look for ease in changing flame manifolds and torch tip location for different production parts. A machine with adjustable timers makes it easy to change brazing times, optical pyrometer settings, speed of rotation, and oscillation without programming (Ref 20). Example 3: Conveyor Brazing System. An automatic conveyor system was installed in a company for handling random parts for any style air-conditioning cooling cores without stopping the line to change tools. The overhead conveyor incorporates servo-driven burner manifolds that move in and out of the line to braze coils as they pass, while the control system automatically adjusts the heat output of the burner and the heat-zone line speed to accommodate the different coil styles. In addition, the burner manifolds return to a home position and idle at a low output level to conserve fuel gas and extend burner life. The new system replaces a manual brazing operation that was relatively expensive and time-consuming, especially when running a mix of various sizes on one production line. A critical part of the control system uses programmable logic controllers (PLCs) and networks to alert the operator to system anomalies and component failures. The network also connects to a standard phone modem, so engineers can remotely modify the PLC software and add new automation cells. This substantially cuts travel time and costs for engineers and increases uptime. Example 4: Automated Brazing Paste Dispenser. A sensor-based control system has been developed (Ref 21) for locating key features on rocket thrust chamber assemblies for automated braze paste dispensing. The system uses a noncontact multiaxis seam-tracking (MAST) sensor to locate these features. The MAST sensor measures capacitance variations between the sensor and thrust chamber surface to produce four varying voltages for control purposes. The sensor information is used to locate the thrust chamber surface and to guide the robotic paste-dispensing equipment along the seams in real-time. Example 5: Use of Robots in a Brazing Application. A proposed robotic system would position filler-metal-foil preforms accurately and tack weld them in place in or on large workpieces in preparation for brazing. The system would automate the time-consuming, skill-dependent, labor-intensive filler-metal-foil-application procedure. The robotically attached pre-
Chapter 10: Applications and Future Outlook / 351
forms would satisfy specifications better and more consistently than manually installed preforms do. The robotic foil-application system was conceived for use in applying filler-metal foil to the nozzle jacket of the main engine of the Space Shuttle (Fig. 10.2). The preforms were applied in five main bands. Preforms 0.1 mm (0.004 in.) thick were placed in the forward end of the jacket; the rest of the jacket was covered with preforms 0.038 mm (0.0015 in.) thick. The last piece inserted in each band was custom-fitted to allow for dimensional tolerances. Initially, filler-metal-foil preforms were cut manually from foil sheets by use of paper-cutting shears. A technician fit the preforms individually in or on the part to be brazed and trimmed or recut them as necessary. Often, the technician had to shift them to eliminate gaps and had to slit or shift them to prevent wrinkles. The proposed robotic system would use preforms cut automatically by electrical discharge machining. A robot, guided by a machine-vision subsystem and equipped with a vacuum-pickup end effector, would pick up each foil preform, stretch it out smoothly, place it in the proper position on the workpiece, and tack weld it at several points simultaneously. The machinevision subsystem would ensure that gaps be-
Brazing-foil preforms
Robotic end effector for placing and tack-welding preforms
Nozzle jacket Sping-loaded tack-welding Vacuum gripper (one of four) electrodes to with lateral position control conform to countour of nozzle
Bottom view
tween preforms do not exceed specifications. Specially developed software would control the system (Ref 21).
Fluxless Brazing A fluxless bonding procedure (Ref 22) has been developed for the 8090 aluminum-lithium alloy, based on an aluminum foil interlayer containing little or no lithium. This foil distorts during the bonding process; the oxide film thus mechanically disrupts, and lithium diffuses from the parent alloy into the foil. The procedure requires little sophistication, because the parts are simply mechanically abraded and degreased before assembly in the bonding equipment. Bonding parameters of 530 °C (985 °F), 8 to 10 MPa (1160 to 1450 psi) pressure, and 20 to 60 min bonding time have been successfully applied, giving joint shear strengths of approximately 80 MPa (12 ksi). Further work has been done in examining different surface preparations, interlayer materials, and testing techniques. Due to their advantages, fluxless brazing methods for aluminum have increasingly come into common use since the late 1970s. Some aspects concerning the casting and rolling of the filler metal, for example, the purity of the fillermetal alloys and the cleanliness and cladding tolerances of the filler-metal sheets, have been examined as well as the essential brazing characteristics of the different filler metals, for example, cladding ratio, kind and purity of the brazing atmosphere, fillet formation, joint clearance, and magnesium deposition in the furnace. Brazing filler metals in particular have been scrutinized for impurity levels, strict process compliance, and material control to achieve reproducible, sound brazed joints. Research (Ref 23) evaluated the fluxless brazing of various filler metals that range from approximately 7 to 12% Si, but usually, the silicon content is approximately 10%. Currently, three types of filler metals are used worldwide, with the following additional alloying constituents: Filler metal
Side view Detail of end effector
Fig. 10.2
Robotic system for placing brazing-foil preforms in the nozzle jacket of the Space Shuttle main engine
High-magnesium type, 10AlSi-1.5Mg Low-magnesium type, 10AlSi-Mg Magnesium-free type, XL-10AlSi
Alloying constituent
4004 (1.5% Mg) 4104 (1.5% Mg and 0.1% Bi) Modified 4045 (0.15% Mg) Low Bi content
352 / Brazing, Second Edition
With aluminum, it is possible to clad the filler metal directly to the base metal, permitting complex parts to be formed and assembled without the need for replaced filler-metal preforms. The use of filler-metal wires and shims is not common. The base metals most frequently used for cladding are 3003, 3103, 3005, 3105, 6951, and 6063. Aluminum has very favorable properties for heat-exchanger applications, and feasible designs for aluminum heat exchangers and methods for their economical mass production have existed since at least the early 1960s. Despite these encouraging factors, no major automotive application of the aluminum radiator in the United States occurred during the first 30 years of development and application efforts. It was not until the 1980s that mass production of aluminum radiators began in the United States. The ability to join aluminum by fluxless brazing was one of the most important factors determining the suitability of an aluminum substitution for copper to replace the standard radiator. The vacuum-braze process for aluminum requires no fluxing, and the braze is accomplished in one step. Assembly density of parts to be joined is not limited with this process, and the oxide barrier to brazing is overcome by the use of a solute addition. Efforts during the 1990s and beyond focused on producing and improving the reliability of aluminum radiators through improvements in aluminum materials and the brazing process. Use of alternate fluxes that pose no corrosion problems now compete with fluxless processes for reducing costs and improving reliability. Fluxless inert gas brazing competes with the vacuum process for improving process rates and yields. Intelligent processing methods for controlling brazing have been developed and have established new standards for manufacturing reliability.
Novel and Emerging Brazing Processes Microwave Brazing. Microwave joining of sintered SiC has shown that small samples could be joined using a silicon interlayer (applied as pressed powder); scanning electron microscopy (SEM) examination showed a smooth, homogeneous interlayer 50 µm (2 mils) wide. Results showed that the interlayer could be reduced to 10 to 20 µm (0.4 to 0.8 mils) using
an oil-based slurry made from silicon powder, and to less than 5 µm (0.2 mils) by plasma spraying silicon on one of the SiC surfaces. Direct joints were made in reaction-bonded SiC using the residual silicon. Excellent joints with good mechanical properties were obtained in both small specimens and in small-scale tube assemblies, such as heat-exchanger and radiant burner tubes. In situ reaction synthesis from powders to produce a SiC-TiC-SiC joint was demonstrated (Ref 24) as well the feasibility of producing SiC from microwave-assisted decomposition of polymer precursors. Finally, new applicator designs, including a compound adjustable iris and a mitered-bend single-mode cavity, were demonstrated to provide improved heating of larger and longer specimens. This work has provided the foundation for scaleup of microwave joining to SiC components for industrial applications. The follow-on work (Ref 25) developed and identified optimal time-temperature profiles for the microwave joining of SiC and also developed (Ref 25–27) new microwave joining methods that can be applied to accomplish in situ formation of SiC interlayers and to join larger samples required for industrial applications. The work focused on the investigation of the effect of specimen preparation on joining of SiC using polymer precursors to form SiC in situ at the interface. Los Alamos National Laboratory also completed the evaluation of joints that were made by FM Technologies Incorporated, using four different joining temperatures as part of an effort to determine optimal joining temperature. Infrared Brazing. Joining of Hastelloy X to Inconel 718 by a rapid infrared processing technique was investigated (Ref 28) using a nickel-base filler metal, AMS 4777 (Ni-7Cr3Fe-3.2B-4.5Si-0.06C, wt%). With this infrared technique, joining typically was completed in seconds in ambient argon. The effects of the infrared joining time on the joint and base-material microstructures, the elemental distribution within the microstructures, and the resulting joint shear strength were investigated. Results showed that excellent joint shear strengths of as high as 503 MPa (73 ksi) were obtained when processed at approximately 1150 °C (2100 ° F) for 120 s. Microstructural examinations of the joint with an optical microscope and a SEM indicated that good wetting existed between the filler metal and both Hastelloy X and Inconel 718. The braze-affected zone width increased with increasing joining time but did not show a
Chapter 10: Applications and Future Outlook / 353
direct correlation with the joint strength. Hastelloy X and Inconel 718 exhibited no noticeable change in microstructures due to the rapid heating cycle of infrared processing. Laser brazing uses the thermal energy developed by laser beams to perform localized brazing of thin-wall precision parts. Because of the higher costs of brazing with lasers, this method should be considered only when conventional brazing methods are inadequate. The following types of metal joints frequently are difficult, if not impossible, to produce by ordinary brazing methods but have been successfully made by laser brazing: • Joints in miniature precision parts that require minimal heat input during joining operations to maintain dimensional tolerances • Joints in thin base metals, 0.10 mm (0.004 in.) and less, that tend to become eroded and sometimes perforated during conventional brazing operations using certain filler metals or fluxes • Joints in assemblies containing heat-sensitive materials or parts that cannot be removed during joining operations • Brazed joints near glass-to-metal seals, adhesively bonded joints, and other thermally sensitive connections • Connections inside evacuated or pressurized vessels or containers (e.g., within sealed glass vacuum tubes) The main advantage that laser brazing offers over more conventional brazing processes is its ability to produce a brazed connection locally without heating the entire part or component to the flow point of the filler metal (Ref 29). Another advantage is the high degree to which the thermal energy of laser beams can be controlled in terms of intensity, spot size, duration, and precision of location or positioning. Also, because a laser beam is capable of being transmitted through solids that are transparent to its wavelength, brazing can be accomplished within hermetically sealed vacuum or highpressure atmosphere enclosures. Most laser brazing is confined to line-contact types of joints, where fillets are made to bridge across the intersection of two surfaces or of a surface and an edge. Infrequently, a bead is made to connect two abutting edges (butt joint). Because the flow of the filler metal is confined to the region heated by the laser beam (gener-
ally, a circular spot), a seam is brazed by producing a series of overlapping spots. Process Parameters. In most applications, the laser beam is aimed directly at the preplaced filler metal in the joint, and the weld is executed (Ref 29). Generally, laser brazing involves positioning the workpiece under the fixed laser beam. Although it is possible to aim a laser beam at the desired spot on a fixed part or workpiece, it is simply more convenient and requires less costly equipment to fix the laser and move the workpiece, which is usually very small and light. Both continuous-wave CO2 and pulsed lasers can be used, but solid-state pulsed lasers, such as neodymium-doped yttrium-aluminumgarnet (YAG) pulsed lasers, appear to be more adaptable to the progressive overlapping-spot mode of brazing. Their shorter wavelength also seems to couple more efficiently to filler metals of high-conductivity and high-reflectivity metals, such as copper, silver, and aluminum. Brazing filler metal is preplaced in the joint, either in powder form or as a shim sandwiched between the members. Powdered filler metal is easier to use and handle, and most powder-type filler metals remain in place during the series of laser pulses. However, the increased beam power needed for filler metals of high reflectivity or thermal conductivity may require the filler metal to be used in the form of foil or shims sandwiched between the joint members so that it will remain in place during brazing. Adequate atmospheric protection is required for all laser brazing, whether or not a flux is used and even if precautions are taken to keep the joint region free of contaminants. Depending on the composition of the materials, the purity of the filler metal, and other variables, joints may be brazed without flux, but a protective atmosphere (argon shielding or vacuum) should always be used. When a flux is used, it may be mixed with the filler-metal powder into a paste, using water or alcohol, and applied to the joint. The mixture must be thoroughly dried before brazing is attempted. Best results have been obtained by firing the laser beam directly over the joint containing the preplaced filler metal, with the beam axis coinciding with the joint axis. The main objective of tilting the laser beam relative to the joint configuration is to achieve the same thermal energy input in both members and thereby obtain a uniformly contoured fillet profile. Process Limitations. Laser brazing generally should be reserved for precision jobs that
354 / Brazing, Second Edition
require the strength of a brazed joint but, for some reason, cannot tolerate the heat, distortion, or other consequences of more conventional techniques. Another fundamental consideration is the availability of a laser of adequate power and the availability of accessory workpiecepositioning equipment. The rapid rate of solidification characteristic of laser-melted deposits can lead to levels of brittleness that would be unsuitable in joints for many applications made with some filler metals (the BNi series, for example). This can be overcome by varying laser parameters, such as increasing beam width, pulse duration, or braze spot overlap, when using filler metals that are hard and that undergo a further increase in hardness when solidified rapidly. The same considerations also apply in brazing of base materials that are hardenable under conditions of rapid quenching. Process Applications. A miniature pressuresensor assembly requiring a strong, pressuretight (to several hundred mega pascals) joint between sections of 70Cu-30Ni capillary tubing was brazed with a fluoride-paste flux and prealloyed powdered BAg-1 filler metal (Ref 30). Strong, reliable connections can be formed between wires and copper circuit boards using the laser beam brazing process (Ref 31). Laser brazing could be easily adapted to form the microconnections of integrated circuits, whether they are electrical connections or other types of joints. The process could also be interfaced with numerically controlled positioning equipment, making high-speed, high-production, and highprecision applications possible. In continuing studies, researchers (Ref 32) used a neodymium-YAG laser in conjunction with energy-flux concentrators to perform precision brazing of small tube assemblies. Polished copper reflectors in the shape of either a wedge or a cone were used to direct and concentrate infrared laser energy onto the tube joint and filler-metal preform to accomplish the braze. These reflectors were designed using optical ray-tracing software employing the principles of nonimaging optics and multiple beam reflections to create a tuned waveguide that could be adapted to a variety of shapes. The ray-tracing algorithm also calculated the incident energy flux on the tubes. This information was used in a finite-difference thermal model to predict processing times. Experimental trials and metallographic evaluation for brazes made in 304 stainless steel tubing using Au-18Ni filler metal were conducted.
This method and its enabling models were validated, offering new processing techniques that may be better suited than current methods for some applications. Proposed extensions of this work to other applications included a methodology to adapt brazing using nonimaging optical concentration to complex part geometries. Electron Beam Brazing. Electron beam welding (EBW) equipment and techniques have been used to a limited extent for brazing. The high vacuum used in the work chamber (0.013 to 0.0013 Pa, or 2 × 10–6 to 2 × 10–7 psi) permits adequate flow of filler metal on properly cleaned joints without the need for a reducing atmosphere or flux. Thus, flux entrapment does not occur, and the work does not require cleaning after brazing. The high vacuum and absence of flux provide a brazing environment that avoids the problems associated with prepared atmospheres that are encountered in brazing of some stainless steels as well as the more reactive metals (such as titanium). Electron beam brazing is performed similarly to EBW, except that the beam is defocused to provide a larger beam spot and to reduce the power density or the heating effect on the work. If necessary, the beam-spot diameter can be enlarged substantially, depending on the type of equipment (defocused beam), while providing an adequate amount of heat input for brazing. Work movement can be used if an area substantially larger than the beam spot is to be heated, and the work can be rotated under the beam for uniform heating. Brazing temperatures are reached quickly, and heat can be localized to minimize grain growth, softening of cold-worked metals, and, in austenitic stainless steels, sensitizing of the material by carbide precipitation. Electron beam brazing is a convenient method for brazing small assemblies, such as instrument packages, and combines the versatility and close controllability of electron beam heating with the advantages of vacuum brazing. Packaged devices can be encapsulated with an internal vacuum without damaging the basic package. Applications. Tube-to-header joints in small heat-transfer equipment made of heat-resistant alloys and refractory metals are sometimes electron beam brazed. In one technique, the tube-toheader joint is electron beam welded on the top side of the header, and the heat of the beam causes the filler metal preplaced on the reverse side of the header at the joint to melt and flow. Small-diameter, thin-wall stainless steel tubes
Chapter 10: Applications and Future Outlook / 355
are readily joined by electron beam brazing. The filler metal BCu-1a has been successfully used in this application. Dresser-Rand, EWI, and PTR-Precision Technologies Incorporated developed Ebraze welding for a new line of centrifugal compressors for the petroleum and petrochemical industries (Ref 33, 34). The process, which combines electron beam welding and vacuum furnace brazing, is used to fabricate the compressor impeller. The impeller, the heart of a centrifugal compressor, comprises a disc, shroud, and blades. Usually, a shaft extends through the disc with an interference fit. Welding, machining from a solid forging, or integral casting attaches the blades to the disc. Previously, an electron beam directed through the cover and into the blade joined the cover to the bladed disc, melting the two pieces together. A barrier, created by the unwelded sidewalls of the blade, prevented molten weld metal from flowing away from the weld point. The 6.4 mm (0.25 in.) thick blade had a fused area 4.8 to 5.3 mm (0.19 to 0.21 in.) thick, which became a classic stress raiser. The majority of impeller fatigue failures occur at blade-disc or blade-cover joints or at a blade leading or trailing edge. To achieve 100% fusion, the EBraze process was developed by borrowing an approach from the vacuum-furnace-braze process, where fusion occurs across the complete surface. The filler metal was placed between the cover and blade of Society of Automotive Engineers (SAE) 4330 stainless steel, and an electron beam was used to weld through it. To prove the process practical for fabrication, a 384 mm (15 in.) diameter impeller of 4330 stainless steel was manufactured with a threedimensional blade design. Testing covered corrosion, impact, and use of other impeller materials, such as American Iron and Steel Institute (AISI) 410 stainless steel, 17-4 and 15-5 PH stainless steels, and Cartech 625+. Preliminary results were positive. Analytical work included three-dimensional finite-element analysis (FEA), elasto-plastic FEA, transient-thermal FEA, and fracture mechanics analysis. An electron beam method that permanently bonds a sapphire fiber to a platinum shell to fabricate a high-temperature fiber-optic probe was developed by the National Aeronautics and Space Administration (NASA) Lewis Research Center (Ref 35). Under normal circumstances, a sapphire fiber would be attached to platinum by a ceramic epoxy. However, this assembly was
to be part of a fiber-optic probe that would measure high temperatures (<600 °C, or 1100 °F) in vibrating machinery, such as in jet engine combustion research. The assembly consists of a 0.38 mm (0.015 in.) diameter sapphire fiber attached to a 6.35 mm (0.25 in.) long, 15 mm (0.59 in.) diameter platinum shell. Because of the small size, electron beam brazing was chosen instead of conventional vacuum brazing. Gold reactive filler metal was selected, so that the sapphire would not be affected by the total heat required. First, a copper fixture that acted as a holding and centering device was fabricated. Then, two 0.38 mm (0.015 in.) diameter rings of gold filler metal were placed around the sapphire, and the fiber was inserted into the platinum shell. The fixture was then placed in an electron beam welder with a vacuum of 0.013 Pa (2 × 10–6 psi). A power setting of 1 mA at 55 kV was selected to generate a broad beam sufficient to melt the braze and cause the gold to flow into the joint. The total time required was 5 s. Third-Body Friction Brazing. The Welding Institute has developed a novel variant of the relatively unfamiliar friction brazing technique called third-body friction joining, which produces high-strength joints in a wide variety of materials. The technique involves a dovetailtype re-entrant arrangement and a third-body material with a relatively lower melting point than the materials to be joined. In general, the joint geometry is designed to limit the stress within the third-body material. The basic principle involves the frictional heating and plasticizing of an independent third-body material that fills the gap between the primary components involved. Relative movement, plastic flow, and intimate contact of the third-body between each component during the friction period usually provide sufficient scouring action to avoid the use of a flux or a controlled atmosphere. In addition, although the third-body material is plasticized by friction heating, it is not melted. Consequently, many of the solidification problems associated with conventional brazing techniques are avoided. On cooling after friction heating has been terminated, the plasticized material consolidates, thereby locking both components together. Depending on the compatibility of the materials, the third body of plasticized material may or may not be metallurgically bonded to either or both of the components to be joined. In both cases, the components are locked together by the third-body material to produce a secure
356 / Brazing, Second Edition
joint. Moreover, the joint is usually formed without the deformation of either component. Typically, this joining technique provides a greater cross-sectional area of the third-body material to support the required axial-torsional loads. The resulting mechanical properties can exceed the ultimate tensile strength of the component material. This is in contrast to some brazed joints that essentially only attain the mechanical properties of the filler metal. Figure 10.3 shows a tensile-tested third-body friction joint. A smaller-diameter carbon steel rod remains joined to a larger-diameter carbon steel rod via a third body of an aluminum material. The third-body material stayed intact after withstanding 617 MPa (89 ksi) before failure occurred in the smaller-diameter steel rod. In this case, the third body consists of aluminum (grade 6082) machining swarf (chip mixture) that was plasticized to fill the total space available between the profiled external diameter of the small rod and the profiled internal diameter of the outer bar. The use of opposed helical flat-bottom reentrant grooves has been designed to provide multiaxial and torsional interlock integrity. The Welding Institute has also examined a friction seam-brazing technique with the use of reentrant features, as illustrated in Fig. 10.4. Potential applications for this innovative third-body friction joining technique include the joining of difficult-to-weld materials such as ceramic-ceramic, metal-ceramic, powder-metallurgy-manufactured materials, and thermoset plastic and composites interlocked with thermoplastic intermediate material. Vacuum Arc Brazing. Researchers (Ref 36) examined and developed technological
methods of depositing coatings on titanium alloys by arc brazing in vacuum with an arcdeposited hollow cathode using a powder-composite filler metal that consisted of a mixture of powders of tungsten or chromium carbide and a filler metal of the Ti-Cu-Ni-Zr system, with a melting point of 917 to 957 °C (1683 to 1755 °F) (Ref 37, 38). When depositing high-quality coatings by arc brazing in vacuum with feeding-powder-composite filler metals, it was necessary to satisfy the following requirements: the pressure in the vacuum chamber should be ≤8 × 10–2 Pa (12 × 10–6 psi); the temperature of heating the parent metal should not exceed the temperature of the polymorphous transformations of titanium α, (α + β)-alloys; the temperature of superheating the filler metal should not be higher than the temperature at the start of dissolution of the tungsten carbide in the molten filler metal; and it is permissible to hold the processed surface of the parent metal at 927 to 997 °C (1701 to 1827 °F) for up to 5 s. Plasma Arc Brazing. The suitability of brazing for joining of heat-sensitive materials was recognized a long time ago. However, the real breakthrough for industrial application was achieved with gas metal arc brazing (Ref 39, 40). Bronze alloys were employed for this purpose. Because the melting point of these alloys is relatively low, the necessary energy input is also lower. Because gas metal arc brazing is a stable, nearly spatter-free process, it is most commonly used (Ref 41).
Re-entrant features
Fig. 10.3
Third-body friction joint. The tensile-tested assembly shows that the smaller-diameter carbon steel rod remains joined to a larger-diameter carbon steel rod by means of a third body of aluminum.
Fig. 10.4
Friction seam brazing with re-entrant features
Chapter 10: Applications and Future Outlook / 357
However, in gas metal arc brazing, the energy input and filler-metal feed are coupled, which limits the freedom of design and configuration in feeding and depositing the filler metal. Furthermore, the required current strength is relatively high because of the low energy density of an uncontrolled arc. The resulting wide joints cause severe distortion of the brazed components, especially with thin sheet metal. Plasma arc brazing is an alternative to gas metal arc brazing because of the high energy density. The advantages of this method have been described in several publications and include lower thermal and metallurgical effects on the base metal, high brazing speed, low distortion, and narrow joints (Ref 42–45). The method is highly promising for the joining of higherstrength sheet metal (Ref 42). Specific applications of this technique are described in detail in Ref 46 and 47. Basic Description. During plasma arc brazing, the torch operates with two independently adjustable arcs, the pilot arc and the main arc. The pilot arc, designated as the nontransferring arc, operates between the tungsten electrode and a water-cooled copper nozzle. The pilot, or nontransferring arc, is connected to the negative pole of the power source and is indirectly water cooled. The water-cooled copper nozzle is connected to the positive pole and constricts the plasma jet. The main arc, designated as the transferring arc, is constricted by the directed gas feed and the plasma nozzle and operates between the negatively polarized tungsten electrode and the positively polarized workpiece. Most available industrial torches operate with a pilot current of 15 to 30 A at an arc voltage usually between 15 and 20 V. The operating voltage of the main arc is usually approximately one-half the value of the current strength, which lies in the range of 40 to 80 A. However, current values up to 200 A are also possible (Ref 48). Brazing Filler Metals. Experience in various fields of application shows bronze wires provide economical filler metal. Bronze filler metals usually employed for hard brazing of steels are conventional furnace filler metals. The process temperature, not the filler-metal composition, ultimately controls the wetting behavior of such filler metals. Together with the standard filler metal CuSi3, numerous other versions are currently available on the market and are applicable over a wide range of brazing requirements (Table 10.1). Table 10.1 gives properties of copper-base
filler metals. Brass and nickel-silver filler metals are not included, because they are not suited for arc brazing. Lack of standardization aggravates the difficulty of selecting appropriate copper-base filler metals. Thus, plasma arc brazing provides an alternative and supplementary method for joining of sheet metal, especially in vehicle body construction. Because of the separate heat input and material feed in combination with the higher power density, smooth and spatter-free joints can be produced with excellent properties for applications (Ref 46). Diffusion brazing produces joint properties that are significantly different from those of conventional brazed joints. The main objective of the process is to produce joints having mechanical properties approaching those of the base metal in applications where other joining processes are unacceptable. Some examples are the following: • Cast nickel-base superalloys for high-temperature service, and beryllium alloys • Some dissimilar-metal combinations • Assemblies where a combination joining and heat treating cycle is desirable to minimize distortion • Elevated-temperature applications, such as high-strength titanium alloys in aircraft • Large, complicated assemblies where it is economical to produce many strong joints simultaneously (Ref 51–53) Two approaches to diffusion brazing are used. One uses a filler metal that has a chemical composition approximately the same as the base metal but with a lower melting temperature. Melting temperature is suppressed by adding certain alloying elements to the base-metal composition or to a similar alloy composition. For example, the melting temperature of a nickel-base high-temperature alloy can be lowered by a small addition of silicon or boron. In this case, the filler metal melts and wets the base-metal faying surfaces during the brazing cycle. This approach is sometimes called activated diffusion bonding or transient liquid phase bonding. The second approach uses a filler metal that will alloy with the base metal to form one or more eutectic or peritectic compositions. When the brazing temperature is slightly higher than the eutectic or peritectic temperature, the filler metal and base metal will alloy to produce a
358 / Brazing, Second Edition
low-melting composition. The filler metal itself does not melt, but a low-melting alloy is formed in situ. This method is also known as eutectic brazing. An example is the diffusion brazing of titanium alloys with copper. With either approach, the assembly is held at brazing temperature for a sufficient time for diffusion to produce a uniform alloy composition across the joint. As this takes place, the melting temperature of the filler metal and the strength of the joint increase. The brazing time depends on the degree of homogeneity desired, the thickness of the initial filler-metal layer, and the brazing temperature. The relationship of heating rate to brazing temperature may also be important. A low heating rate will allow more solid-state diffusion to take place, and more filler metal will be required to provide sufficient liquid to fill the joint. Conversely, if large quantities of filler metal and fast heating are used, the molten metal may run out of the joint and erode the base metal. The thick joint so formed will require a longer diffusion time to achieve a suitable composition gradient across the joint. The composition of the filler metal may be important with respect to response to subsequent heat treatment. This is particularly true for
metals that undergo phase transformation during heating and cooling. Alloy composition will determine the transformation temperature and rate of transformation. Therefore, the phase morphology and mechanical properties of the joint can be controlled by the joint design and the brazing cycle. Advantages and Limitations. Diffusion welding and brazing have a number of advantages over the more commonly used welding and brazing processes as well as a number of distinct limitations on their applications. Some of the advantages of the two processes are as follows: • Joints can be produced with properties and microstructures very similar to those of the base metal. This is particularly important for lightweight fabrications. • Components can be joined with minimum distortion and without subsequent machining or forming. • Dissimilar alloys can be joined that are not weldable by fusion processes or by processes requiring axial symmetry, such as friction welding. • A large number of joints in an assembly can be made simultaneously.
Table 10.1 Filler metals for plasma arc brazing Filler metal Specifications
Melting range DIN No.
Yield strength
°C
°F
MPa
ksi
Tensile strength MPa
A5 fracture
ksi
strain, %
Low-alloy, copper-base fillers SG-CuSi3(a), R-CuSi-A(b)
2.1461
910–1025
1670–1875
250
36
380
55
46
SG-CuSn(c), CuSn1Mn, SCF133G(d) L-CuSi3Mn1(c), CF116C L-CuSi2MN(d)
2.1006
1020–1050
1870–1920
230
33
>340
>49
25
... ...
910–1025 1030–1060
1670–1875 1885–1940
2.1022 2.1056 ...
910–1040 825–990 887–1020
1670–1905 1515–1815 1629–1870
230 235 240
33 34 35
2.0921
1030–1040
1885–1905
180
26
380–450
55–65
40
2.0922
1030–1050
1885–1920
290
42
530–590
77–86
30
2.0923
1015–1045
1860–1915
400
58
>600
>87
16
2.1367
945–985
1735–1805
>400
>58
>650
>94
>10
>120 >80(g)
>17(e,f) >12(g)
340–460 290–340
49–67 42–49
40 45
Tin-bronze fillers SG-CuSn6(c) SG-CuSn12(a) SG-CuSn10SiMn(h)
260 320 >350
38 46 >51
20 5 15
Aluminum-bronze fillers SG-CuA18(a), R-CuAlAl(b), CF303G(d) SG-CuAl8Ni2(a), SG 31150C(i) SG-CuAl8Ni6(a), CR CuNiAl(b), CF 310G(d) SG-CuMn13Al7(a), MSG 31-GZ-300-CN(i)
Note: DIN, Deutsche Industrie-Normen A5 fracture strain is percent elongation after fracture for specimen with an original length 5 times its original diameter. (a) DIN 1733. (b) American Welding Society/American Society of Mechanical Engineers. (c) DIN 17833-52. (d) prEN13347. (e) Guide value for minimum, not for approval testing. (f) Unpublished proposed standard, DIN-NAS, 1999. (g) For rods. (h) Not standard. (i) DIN 8555. Source: Ref 49, 50
Chapter 10: Applications and Future Outlook / 359
• Members with limited access can be joined. • Large joint members of base metals that would require extensive preheat for fusion welding can be more readily joined. An example is thick copper. • Defects normally associated with fusion welding are not encountered. The following are some important process limitations: • The thermal cycle is normally longer than that of conventional welding and brazing processes. • Equipment costs are usually high, and this can limit the maximum size of components that can be produced economically. • The processes are not adaptable to a high production rate, although a number of assemblies may be joined simultaneously. • Adequate nondestructive inspection techniques for quality assurance are not available, particularly those that assure design properties in the joint. • Suitable filler metals and procedures have not yet been developed for all structural alloys. • The faying surfaces and the fit of joint members generally require greater care in preparation than for conventional hot pressure welding or brazing processes. Surface smoothness may be an important factor in quality control in the case of diffusion brazing. • The need to simultaneously apply heat and a high compressive force in the restrictive environment of a vacuum or protective atmosphere is a major equipment problem with diffusion welding. Surface Preparation. The faying surfaces of joint members to be diffusion welded or diffusion brazed must be carefully prepared before assembly. Surface preparation involves more than cleanliness. It also includes (a) the generation of an acceptable finish or smoothness, (b) the removal of chemically combined films (oxides), and (c) the cleansing of gaseous, aqueous, or organic surface films. The primary surface finish is ordinarily obtained by machining, abrading, grinding, or polishing. One property of a correctly prepared surface is its flatness and smoothness. A certain minimum degree of flatness and smoothness is required to ensure uniform contact. Conventional metal cutting, grinding, and abrasive polishing methods are usually adequate to produce the needed surface flatness and smoothness. A secondary effect
of machining or abrading is the cold work introduced into the surface. Recrystallization of the cold-worked surfaces increases the diffusion rate across the weld or braze interface. Chemical etching (pickling), commonly used as a form of preweld preparation, has two effects: the first is the favorable removal of nonmetallic surface films, usually oxides; the second is the removal of part of or the entire coldworked layer that forms during machining. The need for oxide removal is apparent, because it prevents metal-to-metal contact. Degreasing is a universal part of any procedure for surface cleaning. Alcohol, acetone, detergents, and many other cleaning agents may be used. Frequently, the recommended degreasing technique is intricate and may include multiple rinse-wash-etch cycles using several solutions. Because some of these cleaning solvents are toxic or flammable, proper safety precautions should always be followed. Heating in vacuum may also be used to obtain clean surfaces. The usefulness of this method depends to a large extent on the type of metal and the nature of its surface films. Organic, aqueous, or gaseous adsorbed layers can be removed by vacuum heat treatment at elevated temperature. Most oxides do not dissociate during a vacuum heat treatment, but it may be possible to dissolve adherent oxides in some metals at elevated temperature. Some metals that may dissociate oxides and dissolve the resulting oxygen at an elevated temperature are zirconium, titanium, tantalum, and niobium. Cleaning in vacuum usually requires subsequent vacuum or inert atmosphere storage and careful handling to avoid the recurrence of surface contamination. Many factors enter into selecting the faying surface treatment. In addition to those already mentioned, the specific welding or brazing conditions may affect the selection. With higher temperature or pressure, it becomes less important to obtain extremely clean surfaces. Increased atomic mobility, surface asperity deformation, and solubility of impurity elements all contribute to the dispersion of surface contaminants. With lower temperature or pressure, better-prepared and -preserved surfaces are more important. Preservation of the clean faying surface is necessary following the surface preparation. One requirement is the effective use of a protective environment during diffusion welding or brazing. A vacuum environment provides continued protection from contamination. A pure
360 / Brazing, Second Edition
hydrogen atmosphere will minimize the amount of oxide formed, and it will reduce existing surface oxides of many metals at elevated temperature. However, it will form hydrides with titanium, zirconium, niobium, and tantalum that may be detrimental. High-purity argon, helium, and sometimes nitrogen can be used to protect clean surfaces at elevated temperature. Many of the precautions and principles applicable to brazing atmospheres can be applied directly to diffusion brazing or welding. Materials Joined. The advantage of diffusion brazing of aluminum and its alloys is that joining can be accomplished at temperatures well below normal brazing temperatures. This has led to the development of a diffusion brazing system for fabricating boron-aluminum structural components in order to minimize the filament degradation that occurs during elevatedtemperature processing. As boron-aluminum technology advanced, it was recognized that secondary fabrication processes, such as brazing, would have to be developed in order to fully use the weight-saving potential of this new material (Ref 27). However, the need to maintain low processing temperatures was a major stumbling block. Application of the conventional aluminum-silicon filler metals was not acceptable. Diffusion brazing offered a solution to this problem. In this case, the approach was to plate the commercial Al-7.5Si filler metal (4343) with copper. A ternary Al-Cu-Si eutectic, with a nominal composition of Al-27.5Cu-5.2Si, will form when this system is heated to 524 °C (975 °F). This technique was used to braze boronaluminum assemblies below 540 °C (1000 °F). The Al-Cu-Si system seemed ideally suited for brazing the popular A356.0 and A357.0 cast alloy compositions because of their similarity to the Al-7.5Si filler metal. The results of this work (Ref 27) showed that the A356.0 casting alloy can be diffusion brazed by plating one of the joint members with copper. A liquid phase is formed at the Al-Cu-Si eutectic temperature of 524 °C (975 °F) through diffusion of copper into the casting alloy during heating. After quenching and aging, the joint strength will equal that of the casting itself. Microstructurally, the brazed joint will be indistinguishable from the casting. Other types of diffusion brazing have been applied to titanium and to superalloys (Ref 54–62). Other Factors: Time, Pressure, Metallurgical, and Equipment/Tooling. The duration of
the diffusion brazing cycle will depend on (a) the brazing temperature, (b) the diffusion rates of the filler metal and the base metal at brazing temperature, and (c) the maximum concentration of filler metal permissible at the joint. The alloy composition at the joint may influence the response to heat treatment or the resulting mechanical properties of the joint. Therefore, the joint must be held at high temperature for some minimum time to reduce the concentration of filler metal to an acceptable value. Conventional brazing requires little or no pressure across the joint. In some cases, fixturing may be necessary to avoid excessive pressure. This is particularly so when the molten filler metal is to flow into the joint by capillary action. When the filler metal is placed in the joint before brazing, excessive pressure may force low-melting constituents to flow out of the joint before brazing temperature is achieved. In that case, the molten filler metal may not be sufficiently fluid to fill interface voids. The metallurgical events that transpire during diffusion brazing are similar to those that occur during diffusion welding. An additional factor is the variation in chemical composition across the joint. Compositional variations can significantly affect the response of a particular alloy to heat treatment. For metals that exhibit an allotropic transformation, the chemical composition affects both the transformation temperature and the rate of transformation. Thus, the response to heat treatment across a diffusionbrazed joint varies with the local chemical composition. For example, copper stabilizes the beta phase in titanium and decreases the beta-toalpha transition temperature. The equipment and tooling used for diffusion brazing are essentially the same as those used for conventional brazing. If furnace brazing is used, the entire cycle can be done in the same equipment or in a dedicated furnace. In some cases, it may be more economical and convenient to braze with one piece of equipment and then follow with a diffusion heat treatment with other equipment. For example, the brazing could be done with resistance welding or induction heating equipment, and the diffusion heat treatment could be performed in a furnace. Step brazing makes use of the different brazing-temperature ranges of related types of filler metals. One section of an assembly is brazed using an appropriate filler metal. Then, following any supplemental operations, another brazing operation is performed using a filler
Chapter 10: Applications and Future Outlook / 361
metal with a lower brazing-temperature range. The filler metals are selected so that the temperature used for the second braze does not impair the braze made at the higher temperature. Step brazing involving multiple sequences also has been used. The brazing process selected may be any of those discussed in this book, but furnace, induction, and resistance brazing have been the most successful. Application Examples. Step brazing is used for making ceramic-to-metal joints in headers (Ref 63) in which ceramic disks serve as baseplates for encapsulated electronic components and must provide hermetic seals capable of withstanding a vacuum of 0.0013 Pa (2 × 10–7 psi). The ceramic pellets are metallized and nickel plated to allow subsequent metal-to-metal braze joining. No flux is required, because the process takes place in an ultrahigh-purity, controlled atmosphere furnace using a very dry hydrogen atmosphere (dewpoint of –62 °C, or –80 °F). A high-melting silver-base filler metal is used in this first brazing step. In the second step, a lowermelting-point silver-copper filler metal is used. Another application of step brazing is the fabrication of a target holder for a multitarget transmission anode x-ray tube involving joining of five different materials (Ref 64). Targets of tungsten, gold, and uranium 238 (D-38) are attached to oxygen-free high-conductivity (OFHC) copper, which is then joined to stainless steel (Fig. 10.5). Two different heating techniques, induction and furnace, were used. The sequence of brazing was as follows: 1. The stainless steel bellows is furnace brazed to the stainless steel flange with OFHC copper filler-metal wire at 1085 °C (1985 °F).
Cusil braze
2. The tungsten target is furnace brazed to the copper rod with 3Ni-35Au-62Cu filler metal at 1030 °C (1885 °F). 3. The previously mentioned bellows and flange are induction brazed to the copper bar with 82Au-12Ni filler metal at 950 °C (1740 °F). 4. The gold and D-38 targets are simultaneously furnace brazed to the copper rod with 72Ag-28Cu filler metal at 780 °C (1435 °F). The liquid hydrogen flowing through the stainless steel tubes that line the Space Shuttle engine jackets removes heat. The jacket liners are intricate brazements made of hundreds of small-diameter tubes of stainless steel brazed to each other and to the jacket backing in a twostep process. Foil-type filler metal (34.5 to 36.0 Au, 13.5 to 14.5 Ni, 9.5 to 10.5 Pd, 14.5 to 17.5 Mn, balance Cu) is attached by resistance spot welding to the jacket of Inconel 718, which is brazed at 1050 °C (1920 °F) to 1080 A-286 stainless steel tubes. This slow brazing cycle takes 24 h, including argon and partial vacuum purges. In the second step, a filler-metal paste (30.0 to 32.0 Au, 9.0 to 10.5 Ni, 9.0 to 10.5 Pd, 14.5 to 17.5 Mn, balance Cu) is used to make the tubeto-tube connections along the entire length of the tubes. Argon is used to purge the chamber below 760 °C (1400 °F), and hydrogen is used from 760 to 980 °C (1400 to 1800 °F). Step brazing is used for fabrication of the front frame of a major jet engine. Looking like a large, spoked wheel, the airfoil-shaped struts are joined to the outer case and to the hub. The forward frame houses the engine bearing and supports the front end of the engine. A total of 240 parts are joined by 75 m (250 ft) of brazed seams to the Inconel 718 frame (Ref 65).
Copper braze
Nicoro braze
Nicoro braze
Tungsten target
Copper rod
D-38 target Gold target
Fig. 10.5
Stainless Steel flange
Stainless Steel bellows
Target holder for a multitarget transmission anode x-ray tube fabricated using step brazing to join five different materials. (Nicoro and Cusil, Wesgo Metals). Source: Ref 64
362 / Brazing, Second Edition
Because it would be difficult to fixture and braze 240 parts in one operation, step brazing is used. Initially, BNi-4, a nickel-base filler metal with a melting point of 1105 °C (2020 °F), is used to braze tubing, brackets, and stiffeners into subassemblies. In the second step, performed at 1065 °C (1950 °F) with BNi-3 filler metal, these subassemblies are joined to flanges, hub castings, and manifolds without remelting the joints previously brazed. The final step, done at 1040 °C (1900 °F) with BNi-2 filler metal, adds manifold covers, extra tubing, and spacers, again without remelting the joints brazed during previous steps. Braze Processing Systems. Research has been conducted (Ref 66) on the aluminum brazing process and its major application to brazing heat exchangers for automobile air conditioners, radiators, and so on. Several different types of furnaces, ranging from single-batch to continuous process types, have been developed. Car manufacturers are using in-line vacuum
Fig. 10.6
brazing furnaces (Fig. 10.6) worldwide, with highly successful results. With increasing requirements for reduced weight of automobiles in recent years, an increasing number of automobiles are being built with aluminum alloy radiators rather than the conventional copper alloy radiators. Aluminum alloy radiators are 30 to 40% lighter than copper alloy radiators. However, the lower resistance to corrosion has not made them very popular until recently, when improvements in the brazing process and improved materials were developed. Aluminum alloy radiators have been fabricated by either the vacuum brazing technique or the noncorrosive flux technique described in Table 10.2. Automotive heat exchangers include those for cooling the engine and evaporator, such as radiators, oil coolers, intercoolers, and other units, and those for air conditioning, such as capacitors, heaters, and other components. Most automobiles of the latest models come equipped
Continuous vacuum brazing furnace dedicated to fabrication of aluminum alloy automobile radiators. Radiators are visible on the platform.
Chapter 10: Applications and Future Outlook / 363
with an air conditioner; the majority of these heat exchangers are made of aluminum. To meet the requirement for reduced weight, aluminum radiators replace copper radiators in automobiles, as shown in Fig. 10.7. It is anticipated that the shift to aluminum radiators will occur more rapidly than in the
Table 10.2 Comparison of brazing processes for aluminum alloy radiators showing the reduction in steps by vacuum furnace brazing
Atmosphere
Noncorrosive flux brazing
Vacuum brazing
Atmospheric pressure, N2 gas
Vacuum, N2 carrier gas
Assembling
Assembling
Degreasing
Degreasing
Process
Corrosion protective treatment
Flux application
Drying
Brazing
Fig. 10.7
Brazing
Vacuum degreasing; brazing
Cooling
Cooling
Surface treatment
Surface treatment
Automobile cooling system including an aluminum alloy radiator
past. Future technological tasks in fabricating aluminum radiators include: • Use of vacuum degreasing technique: Some of the component parts of aluminum radiators are degreased and brazed in the vacuum brazing furnace, and complex processing is conducted, but in the future, it will be necessary to develop technology to apply the technique to all component parts. • Development of high-strength materials: In order to further reduce the wall thickness of component parts, it is also necessary to develop high-strength materials in addition to corrosion-resistance materials. The common method of increasing the strength of aluminum materials is to add magnesium and silicon. The vacuum brazing technique is the most promising method for use of filler-metal materials containing magnesium, and thus, this technique will be used for fabrication of products developed from new compositions in the future. It is claimed (Ref 67) that a novel, moisturefree brazing atmosphere system has been developed to replace the conventional humidified nitrogen-base brazing atmosphere system for brazing carbon steel components with good and consistent brazed joint quality and properties. The system involves adding a small amount of carbon dioxide to a dry nitrogen-hydrogen atmosphere to control braze flow and eliminate soot formation while brazing carbon steel components. It has been installed and commercially demonstrated in a production furnace. The results also show that the addition of a small amount of carbon dioxide is instrumental in overcoming problems normally experienced with conventional brazing atmospheres, such as humidifying a nitrogen-hydrogen atmosphere with the right amount of moisture and finding a suitable location for introducing the humidified nitrogen stream into the brazing furnace. The addition of a mere 0.5% CO2 to the nitrogen-hydrogen atmosphere containing 6.5% H resulted in significantly improved quality and consistency of brazed joints, as shown in Fig. 10.8 and 10.9. The addition eliminated excessive flow of filler metal and formation of soot on brazed joints. Furthermore, it provided excellent penetration of filler metal and fillet formation, as shown in Fig. 10.8 and 10.9. Programmable Logic Controllers. The tool for now and the future in vacuum furnace
364 / Brazing, Second Edition
control systems is the programmable logic controller (PLC) (Ref 68). While PLCs were introduced before the 1980s, their use on vacuum furnace control systems was not widely accepted for several years. Modern PLCs used as control devices on even the most complex vacuum furnace systems are reliable, powerful, and practical. The PLCs have become an important tool to help reduce furnace operator involve-
ment, to produce consistent product quality, and to reduce furnace downtime. Most modern vacuum furnaces equipped with a PLC require only that the operator load and unload the furnace, select the recipe to be run, and push the start button. Today’s PLC systems, including the PLC, operator interface panel, engineering programming, and installation labor, generally cost
Fig. 10.8
Carbon steel component brazed in a moisture-free atmosphere. The addition of 0.5% CO2 to the nitrogen-hydrogen atmosphere containing 6.5% H resulted in significantly improved quality and consistency of the brazed joints.
Fig. 10.9
Carbon steel component brazed in a moisture-free atmosphere. The component exhibits good braze flow, fillet formation, and brazed joint quality.
Chapter 10: Applications and Future Outlook / 365
much less than their predecessor relay/timer/ push-button systems. More importantly, most of the problems associated with the earlier systems, such as dust, mechanical wear, loose wires, and so on, are virtually nonexistent. Additional benefits are possible using PLCcontrolled equipment, especially if the PLC controls all operating parameters of a vacuum furnace, including temperature, vacuum level, cooling pressure, and time (Fig. 10.10). The PLC also can be connected to a modem, so the equipment can be monitored, controlled, and modified from a remote location. This can include service from the furnace manufacturer, which allows a quicker response time, less furnace downtime, and can eliminate the need for an on-site field service call. These benefits translate into cost savings. Metal-Ceramic Joining for Microelectronics Packaging. Brazing processes have been previously confined to high-temperaturefired/cofired alumina substrates using either tungsten- or molybdenum-base refractory metallization. The metallized alumina packages are usually fired (cofired) at approximately 1600 °C (2910 °F). Prior to brazing, the metallized pads and seal rings are normally plated with nickel and heat treated. Conventional brazing of pins, leads, and heat sinks to the metallized alumina substrate is usually carried out in a nitrogenhydrogen atmosphere using a silver-copper eutectic filler metal in the 820 to 900 °C (1510 to 1650 °F) temperature range. Current developments cover the development of silver-, gold-, and copper-base thick-film
Fig. 10.10
Typical furnace brazing control system with a remote computer station. PLC, programmable logic controller; PC, personal computer
paste compositions, filler-metal compositions, pin-lead window-frame and heat-sink surface treatment, and furnace conditions where brazing processes are accomplished in the 550 to 760 °C (1020 to 1400 °F) temperature range in a nitrogen atmosphere. The metal-ceramic joint strengths obtained using low-temperature-fired thick films were found to be comparable to those accomplished using high-temperature-fired tungsten- or molybdenum-base metallization. The capabilities developed by researchers (Ref 69), coupled with ceramic circuitization using existing (a) high electrical conductivity metallization (copper-, gold-, or silver-base) and (b) low dielectric constant and low coefficient of thermal expansion dielectric thick-film pastes and tapes, offer the package designer extended electrical, thermal, performance, and reliability capabilities that were not available previously. Kovar pins were attached to tungsten-metallized, alumina-ceramic packages with silvercopper filler metals, using a wide range of brazing process parameters that were systematically varied (Ref 70). It was found that major changes in brazing process factors, including the thermal cycle, furnace atmosphere, filler-metal composition, and filler-metal quantity, have shown no detrimental effect on cofired ceramic package performance, as judged by the results of metallographic inspection and by mechanical tests (pinpull strength) performed on samples exposed to vibrational fatigue, thermal shock, salt spray corrosion, and temperature-humidity cycling. This robustness was attributed to the combined quality of the alumina-tungsten cofired ceramic materials, fabrication processes, brazing pins, and package design. The observed corrosion behavior was attributed to sample handling, test variability, or variability in the plating process. Despite this high level of packaging system performance, product quality must be monitored as changes are made to any of the key packaging raw materials, processes, or design elements. Modeling Behavior of Brazing Processes and Materials. In the past several years, especially with the assistance of computational equipment and software, a variety of modeling techniques evaluating solidification, dissolution, filler-metal concentrations, and so on have taken place. These studies, as well as others, will produce the finite differences in the many brazing processes that currently exist and also
366 / Brazing, Second Edition
will offer a better understanding of the melting, wetting, and solidification of filler metals and base-metal effects (Ref 71). The future appears to offer an exciting time for the braze specialists and technicians. A study (Ref 72) was able to predict the tensile strength of resistance-brazed joints using a model equation based on the electric current variation. The results of the work provided the following conclusions: • During resistance brazing, the electric current passing through the primary circuit of the transformer changes in relation to the three extremum values, i1, i2, and i3. The electric current rapidly increased to the first maximum value, i1, then gradually decreased to the minimum value, i2, and finally, gradually increased to the second maximum value, i3. • The tensile strength of resistance-brazed joints was predicted using a model equation expressed in the form of a polynomial in three variables of i1, i1-i2 and i2/i3. The predicted tensile strength of Ti-10Zr (wt%) alloy joints resistance brazed with Ti-20Zr-20Ni-20Cu (wt%) or Ni-50Cu (wt%) filler metal agreed well with experimental data. The values of the multiple correlation coefficient adjusted for the degrees of freedom were 0.883 for Ti20Zr-20Ni-20Cu filler metal and 0.884 for Ni-50Cu filler metal. In Situ Reaction Joining of Fiber-Reinforced SiC Composites. Methods for in situ reaction joining of SiC-SiC composites have recently been developed (Ref 73). The SiCfiber-reinforced SiC-matrix composites (SiCSiC) produced by chemical vapor infiltration are being developed for use in structural applications at temperatures approaching 1000 °C (1830 °F) (Ref 74–76). These composites contain approximately 40 vol% SiC fibers (Nicalon, Dow Corning Corporation) and are infiltrated to approximately 85% of the theoretical density with SiC. In order to fully realize the advantages of these materials, practical joining techniques are being developed. Successful joining methods will permit the design and fabrication of components with complex shapes and the integration of component parts into larger structures. These joints must possess acceptable mechanical properties and exhibit thermal and environmental stability comparable with the composite that is being joined.
Joining of SiC has been accomplished by a variety of techniques, including: • Direct diffusion bonding (Ref 77, 78) • Codensification of interlayer and green bodies (Ref 79) • Diffusion welding or brazing with boride, carbide, and silicide interlayers (Ref 77) • Hot pressing of sinterable SiC powder (Ref 80) • Bonding with polymeric precursors (Ref 81) • Brazing with oxide (Ref 82) or oxynitride materials (Ref 83) • Reactive metal bonding (Ref 84) • Active metal brazing (Ref 85) Although varying degrees of success have been achieved, these joining methods must be improved on to withstand the intended service temperatures (1000 °C, or 1830 °F). The initial studies have identified two material systems with potential for joining SiC-SiC composites by reaction methods. Focus has been aimed on joints produced using TiC-Ni and SiC + Si interlayers. The microstructures of joints have been characterized, and the results appear promising. Further work is needed to optimize joint microstructures, understand interfacial reactions, and assess the mechanical properties of the joined components (Ref 73). A glue invented by researchers at the U.S. Department of Energy’s Ames Laboratory, Ames, Iowa, allowed manufacturers to join SiC composite parts for the first time. The rugged ceramic composites are considered possible replacements for steel and the superalloys used in the aerospace industry, because they can withstand higher temperatures, do not melt, and are less susceptible to corrosion. The composites consist of SiC fibers woven together like a mat and then encased in a SiC matrix. Just as steel rebar strengthens concrete, the fibers strengthen the matrix material. Previously, it was impractical to use composites for complex objects such as fans, heat exchangers, or fuel cells, because there was no way to form reliable joints between the parts. Silicon-bearing polymers and an aluminumsilicon alloy powder make up the paste-like glue. When heated to 480 °C (900 °F), polymers in the glue begin to break down into SiC and excess carbon. As the heat increases to 595 °C (1100 °F), the alloy powder starts to melt. Silicon from the alloy reacts with the excess carbon
Chapter 10: Applications and Future Outlook / 367
to form more SiC, while the aluminum reacts with available oxygen to form alumina. The additional ceramic particles, or whiskers, diffuse to strengthen the joint in the same way the fibers toughen the composite. Tested ceramic joints made with the glue demonstrated strength up to 99.5 MPa (14.4 ksi) at temperatures to 1200 °C (2190 °F). For comparison, steel has little strength at temperatures above 705 °C (1300 °F). For application, the glue can be heated with a propane torch and cures in a regular atmosphere without clamping pressure (Ref 86). Joining of Abrasive Tool Materials. In recent years, direct brazing of a monolayer of diamond crystals on a steel substrate with active filler metals has gained tremendous importance in the industry, with a view to developing tools that can outperform the conventional galvanically bonded diamond tools. An existing proprietary process uses specially prepared nickelchromium filler metal to facilitate its application on a steel substrate. The brazing is done either in a vacuum or a dry hydrogen furnace. Studies (Ref 87) have shown that a commercially available nickel-chromium hardfacing alloy, flame sprayed on a steel substrate with an oxyacetylene gun, could be used for direct brazing of diamond particles. During the induction brazing in an argon atmosphere, the chromium present in the filler metal segregated preferentially to the interface with diamond to form a chromium-rich reaction product promoting the wettability of the filler metal. It has been further revealed that under a given set of brazing conditions, the wettability of the nickel-chromium hardfacing alloy toward diamond grits primarily depended on its layer thickness (Ref 87). Joining with Metallic Amorphous Glass Foils. Rapidly solidified (RS) amorphous and microcrystalline filler metals are currently used in a wide variety of brazing applications. The RS materials, typically cast to foil form for direct use in metal joining, offer superior purity and chemical and microstructural homogeneity when compared with conventionally formed filler metals. This homogeneity, in turn, manifests itself in uniform melting, flow in the joint area, and solidification during the brazing process. Accurate control of brazing in this manner permits the production of uniform joint microstructures that are free of voids and macroscopic segregation. The results are dramatic reductions in reject rates and superior joint properties. Over a broad range of base
metal/filler metal combinations, the use of RS filler metals yields joints with superior mechanical properties and improved resistance to thermal fatigue and corrosion. Moreover, the use of RS technology uniquely permits the formation of foils in many filler-metal systems that are brittle and unformable in the crystalline state. The basic difference between crystalline and glassy metals is in their atomic structures. Crystalline metals are composed of regular, threedimensional arrays of atoms that exhibit a longrange order. Metallic glasses do not have long-range structural order. Atoms are packed in a random arrangement similar to that of a glass or a liquid metal. Despite this vast structural difference, crystalline and glassy metals of the same composition will have nearly identical densities. Typically, a metallic glass will be a few percent less dense than its crystalline counterpart. Metallic glasses also lack the microscopic structural features common in crystalline metals. In the absence of crystallinity, grains, grain boundaries, grain orientations, and additional phases do not exist. The glassy state is essentially one phase, possessing complete chemical homogeneity (Ref 88). A whole family of RS filler metals has now been produced, including Cu-Ni-Sn-P (78Cu10Ni-4Sn-8P and 77Cu-6Ni-10Sn-7P), RSNi-2 (82.5Ni-7Cr-3Fe-3B-4.5Si), RSNi-3 (92Ni0.5Fe-3B-4.5Si), RSNi-Pd, and Al-Mg-Si filler metals (Ref 89, 90). The applications for these RS materials range from high-temperature brazing of superalloys in critical assemblies within gas turbine engines to low-temperature soldering of semiconductors and lead frames in microelectronic devices. Applications include the brazing of honeycomb Inconel 625 exhaust nozzles, cones, plugs, fan ducts, and tail pipes for jet engines. In all cases, the foil is applied to the facesheets using resistance tack welding with roller wheel electrodes. For complex contours, precut foil is chosen. The filler-metal foil chosen was the nickel-base American Welding Society (AWS) BNi-2 or Aerospace Material Specification (AMS) 4777B. The metallic glass foil is really metal with an amorphous, glasslike atomic structure. It is produced by very rapid quenching of a stream of liquid metal into a highly ductile ribbon. The supercooling bypasses the nucleation and grain growth stages altogether, producing an extremely strong material.
368 / Brazing, Second Edition
The future of RS technology lies in the joining of ceramics to metals in structural applications such as internal combustion engines and gas turbine components. Two approaches have been adopted in the development of ceramic brazing systems: alteration of the ceramic substrate by metallization, and use of titanium-zir-
conium filler metals, which react with ceramic materials and promote wetting. Diffusion of Boron in Ductile Foils. Another form of filler metal that has recently been developed is used and produced by diffusion of boron into the surface of ductile foil (Fig. 10.11) or wire (Fig. 10.12). The filler
foil Ingot Boron diffusion
Boride
Ductile core
Boride
Fig. 10.11
Basic steps used to produce ductile brazing filler-metal foil with boron diffused into the foil surface
Fig. 10.12
Basic steps used to produce filler-metal wire preforms with boron diffused into the wire
Chapter 10: Applications and Future Outlook / 369
metal is produced in its final configuration prior to the introduction of boron, thereby eliminating the ductility problem that accompanies the presence of boron. Boron is then added by diffusion into the surface to produce a composition suitable for brazing. As with foil, nickel-base wire can be drawn in the boron-free ductile state. Preforms can be produced in any desired configuration. Boron is subsequently diffused into the surface to achieve the proper finished filler-metal composition. Ductile-based nickel foils, including (but not limited to) alloys such as AMS 4775, 4776, 4777, 4778, and 4779, have been developed. With the diffusion of boron directly into the surfaces to be joined, the surface assumes the composition of the base metal plus boron, the melting-point depressant. The surface to be bonded is thus converted to a brazelike filler metal, permitting diffusion brazing. Although a number of applications for these materials are based on the substitution of foil or wire for conventional powder-based filler metals, many areas of interest have developed that are not currently using brazing techniques. The wire and foil preforms described previously are being used in the transient liquid phase (TLP) bonding process, in which specifically tailored foil compositions and bonding cycles are used to produce high-quality bonds in Inco 713C turbine vanes in a turbofan engine used for commercial wide-body transports (Ref 54, 56). Transient liquid phase bonding (Ref 54), also called activated diffusion bonding (ADB) (Ref 91), is a bonding process that combines the manufacturing ease of brazing with the high efficiency of solid-state diffusion bonding. The TLP bonding is applicable to nickel-base superalloys that are difficult to bond by conventional fusion welding because of their fusion cracking troubles. This process is carried out with vacuum furnace brazing using filler metals with a specific composition, usually nickel-base filler metals containing boron. These filler metals temporarily melt and then resolidify at the bonding temperature according to the boron diffusion into the base metals. Moreover, it is said that by the postbond heat treatments, the elements composing the base metal diffuse into the TLP bonds and make a similar bond as the base metal (Ref 92). Researchers (Ref 93) examined the TLP bonding of nickel-base superalloys Mar-M247 and IN-939, using filler metals spe-
cially designed and fabricated into flexible coils by RS processes. The TLP bonds, using those specially designed filler metals for Mar-M247 and IN939, reflect the following results: • Microstructures of Mar-M247 TLP bonds were almost the same as those for the base metal and were attained with the filler metal containing adequate quantities of strengthener 10.8Co-8.8Cr-3.9W-3.0Ta-3.0Al-2.5B. • Sufficient bonding pressure was also indispensable for the Mar-M247 TLP bonds to achieve high stress-rupture properties. • The IN-939 TLP bonds showed better bonding efficiency than the Mar-M247 TLP bonds. Transient Liquid Insert Metal (TLIM) Diffusion Bonding. The TLIM bonding consists of three processes: dissolution of base metal, isothermal solidification, and homogenization. The advanced TLIM bonding process uses an amorphous filler metal and a metal powder sheet. In this new process, both the time necessary to complete the isothermal solidification process and that for the homogenizing process are shortened, compared with those of conventional processes. The mechanism to shorten the TLIM bonding process is the use of a powder sheet with an insert metal. The morphology and size of the powder, kind of powder, and the thickness of powder sheet were also factors to control the mechanical properties of the bonded joints. Researchers (Ref 94) have successfully applied the previously mentioned theories in a TLIM process to the oxide-dispersion-strengthened (ODS) alloys MA-754, Mar-M247, and alloy 713C. The MBF-80 (Ni-15.5Cr-3.7B) amorphous foil insert material was used as the intermediate filler metal. The process was developed especially to join hot-cracking-susceptible nickel-base cast superalloys, dispersion-strengthened alloys, singlecrystal alloys, and ODS alloys and is similar to the previously mentioned TLP and ADB processes (Ref 54, 91). Evaluation of Structural Defects. Since the 1960s, nondestructive evaluation methods to detect braze defects have grown from visual means to x-ray methods to ultrasonic to eddycurrent examination and even to several heating methods where all of these nondestructive techniques are available in production processes.
370 / Brazing, Second Edition
Now and in the future, metals and ceramics will be joined to each other in various fields of industry, and structural defects and integrity of the joints will require more sophisticated means of verification. Due to the difference between the thermal expansion coefficients of ceramics and metals, high thermal stresses can nucleate cracks, which then propagate either in the ceramic or along the ceramic-braze interface. Because applied stresses are added to the residual stresses in the joints, it is of great importance to reliably detect defects in the ceramic-to-metal joints before use, using nondestructive test methods. As a result, new equipment has been developed and applied in the evaluation of the integrity of the brazed ceramic-to-metal joints. The equipment includes scanning acoustic microscopy (SAM), C-mode scanning acoustic microscopy (C-SAM) (Ref 95), x-ray computer tomography (CT) (Ref 96), scanning laser acoustic microscopy (SLAM) (Ref 96), and scanning photoacoustic microscopy (SPAM) (Ref 96). The results of comparative C-SAM and SEM studies (Ref 95) show that structural defects in brazed ceramic-to-metal assemblies can be reliably detected by using a reflection-type C-SAM operating at frequencies up to 100 MHz. For the detection of disbonding of the interface and laminar and surface-opening cracks in the tested nitride ceramic, different focusing techniques must be used. X-ray CT is a bulk characterization technique that can display real-time, two-dimensional x-ray sections of complex parts, such as turbocharger rotors and engine valve components. Current x-ray systems take several minutes for the beam to scan a part, but the use of a conebeam x-ray system to collect all the data at once, coupled with powerful computer processing, will make future systems much faster. Conventional x-ray can only detect 1 to 2% density variations, whereas x-ray CT is 100 times more sensitive and can detect density variations ranging from 0.01 to 0.02%. In addition, x-ray CT can be used to determine the quality of a ceramic part from the beginning to the end of processing. The SLAM scans the surface perturbations continuously with a focused laser beam. The use of a scanning laser beam to detect surface displacements (namely, the sound field) allows images of the transmitted and scattered-mode converted sound fields to be visualized independently. This type of acoustic microscope provides images through the entire thickness of
a part. For a typical advanced ceramic, a 10 MHz beam will penetrate a few millimeters of material with resolution of 250 µm (10 mils), whereas a 500 MHz beam generates a resolution of 5 µm (0.2 mils) (Ref 96). SPAM shows excellent potential for detecting surface and near-surface flaws in opaque ceramics (Ref 96). In conclusion, at the present time, 5 µm (0.2 mils) flaws are not detectable, 50 µm (2 mils) flaws are always detectable, and flaws of intermediate size may be detectable. In addition, surface flaws as small as 1 µm (0.04 mils) are detectable. There is no single technique that will detect all flaws. A number of techniques must be used, and these must be carefully optimized for the material, part, and application. X-ray CT is a much more widely useful technique than neutron radiography, because the latter requires a nuclear reactor to provide enough flux to examine ceramic parts properly. However, ultrasonic techniques may be preferred over x-ray methods, because ultrasonic waves present no hazard to the operator. Robotic Inspection Systems. A new robotic system was developed to inspect arrays of brazed tubes that were fabricated for use in heat exchangers and similar objects. In the application for which the original version of the system was conceived, the arrays would contain coolant tubes to be mounted in the nozzle of the main engine of the Space Shuttle. Other versions might be used to inspect components of terrestrial heat exchangers for power plants, vehicles, and refrigeration equipment, for example. In the Space Shuttle application, the array of coolant tubes must be inspected for dents and scratches on the tubes, filler metal on the crowns of the tubes, wetting of the tubes by the fillermetal fillets, voids and blobs in the fillets, skulling (crust), and discoloration. In addition, the sizes of the fillets must be measured. Heretofore, the arrays of coolant tubes have required manual and visual inspection by multiple technicians. Manual and visual inspection is very tedious and time-consuming. It is also highly subjective, and technicians might occasionally misinterpret what they see and/or leave some areas uninspected; consequently, the results of manual and visual inspections by different technicians can be inconsistent. The proposed robotic inspection system would provide for automated, complete, objective, and efficient inspection of the arrays of
Chapter 10: Applications and Future Outlook / 371
coolant tubes. The system (Fig. 10.13) includes a laser and a video camera that scans the array of tubes under computer control. The camera, the scanning robot, and an image-processing computer have been integrated into one subsystem. The output of the video camera is digitized and fed to the image-data-processing computer, which extracts profiles of the tubes and fillets from the image data. This computer also analyzes the profiles to identify dents and scratches, filler metal on the crowns, wetting voids, blobs, and skulling and to compute the sizes of the fillets. A color version of the system could also look at oxidation and discoloration. A videotape recorder could be used to monitor the inspection process and provide a record of the work performed. Ceramic-Metal-Graphite Joining. The application of ceramics in structural components such as turbine engines has received extensive attention in recent decades due to their excellent high-temperature strength and resistance to corrosion and wear. However, because of their brittle nature, joining of ceramics to metals is frequently required. As a consequence, the lack of joining techniques has, in many cases, limited their use. Normally, conventional fusion welding is not performed, due to the risk of brittle fracture initiation as a result of the high concentration stresses formed on cooling. Hence, solid-state
Fig. 10.13
bonding and various types of brazing are currently applied to maintain the excellent basemetal properties of ceramics. Brazing possesses a major advantage compared with conventional welding, because the base metals do not melt. This allows brazing to be applied in the joining of dissimilar materials that cannot be joined by fusion processes due to metallurgical incompatibility. In general, brazing produces less thermally induced stress and distortion, because the entire component is subjected to heat treatment, thus preventing the localized heating that may cause distortion in welding. In addition, it is possible to maintain closer assembly tolerances without costly secondary operations. Moreover, brazing can be easily adopted for mass production (Ref 97, 98). Therefore, the scientific principles involved in the brazing of ceramics have been explained, discussed, and accepted, and, in general, wetting seems to be the limiting factor to obtain sufficient adherence. This fact has led to the development of various techniques to metallize the ceramic surfaces prior to brazing (indirect brazing) (Ref 30, 96, 99, 100). Recent advances have, however, resulted in new types of filler metals that provide sufficiently low contact angles due to the addition of active elements. These are essentially silver-copper filler metals, silver or copper filler metals with additions of titanium, and tin-base solders (Ref 101).
Robotic system for inspection of brazed coolant tubes. In the proposed system, image-processing algorithms would be used to profile the tubes and braze fillets and to identify any potential flaws.
372 / Brazing, Second Edition
Researchers (Ref 102) evaluated the joining of Si3N4 with thin, sputter-deposited titanium and nickel films to 304 stainless steel using metallic buffers in a series of Si3N4/Ni/Mo/Ni/ 304. Calculations using a finite-element method indicated a marked reduction in thermal stress induced in the joined Si3N4 with increasing thickness of the molybdenum buffer. The conclusion showed that the strong interfacial bond inducing the fracture of the joined Si3N4 was interpreted in terms of a good interfacial reaction, the interdiffusions and the reduction for thermal stress being due to the insertion of the molybdenum buffer. Researchers (Ref 103) verified that direct brazing of a Si3N4-steel joint using Ag-28Cu filler metal with titanium interlayer could probably solve the main difficulties in brazing ceramics to metals: wetting (Ref 104, 105) and residual stresses due to thermal expansion mismatch (Ref 106–108). Future efforts will evaluate direct brazing of ceramic to metal with a reactive interlayer material such as niobium, tantalum, zirconium, or hafnium. Niobium as an interlayer has been used in joining SiC to 304 stainless steel (Ref 109). Niobium as an interlayer material for joining ceramics possesses two superior properties. First, niobium easily reacts with ceramics, alumina (Ref 110), and SiC (Ref 111). Secondly, soft niobium, with a low expansion coefficient, can relax the thermal stress that arises from the difference between ceramics and metal. Researchers (Ref 109) used reaction-sintered SiC with 13 wt% Si, pressureless sintered SiC with a few percent alumina, and niobium with a purity of 99.9 wt%. Using the Ti-Ag-Cu filler metal, graphite can be joined to 95% alumina ceramics. This will provide an important basis for the widening application of graphite (Ref 112). For the existence of carbon, the Ti-Ag-Cu method of brazing is used to make AlAg3 and Ti3Al exist in the physical phase of the sealed region, which is an essential factor in the possible forming of gastight seals between graphite and 95% alumina. For the existence of AlAg3 and Ti3Al, the use of the Ti-Ag-Cu method is a significant advance in the mechanism research of seals between 95% alumina ceramics and oxygen-free copper (Ref 113), Kovar, beryllium, beryllium oxide (Ref 114), tungsten, molybdenum, and stainless steel (Ref 115). Superplastic Forming/Brazing Process. A new method, combining superplastic forming
(SPF) and the brazing process, for the production of complex-shaped Inconel 718 components has been studied (Ref 116–119). The SPF/brazing process has been used as a substitute for the SPF with concurrent diffusion welding process in which welding pressure cannot be applied effectively. The SPF/brazing process can also shorten the total working time and save energy during fabrication. Using this method, an Inconel 718 superalloy sheet was superplastically formed and concurrently brazed with a nickel-base filler metal, MBF-20. The working parameters were as follows: SPF was done at 985 °C (1805 °F) under an argon pressure of 2.45 MPa (355 psi), and the temperature increased to 1040 °C (1900 °F) for 10 min for brazing. The effect of pressure on the brazement during the SPF/brazing process was simulated on a single-lap shear test specimen. The results show that the joint strength increased from 784.2 ± 10.4 to 868.4 ± 12.5 MPa (113.7 ± 1.5 to 125.9 ± 1.8 ksi) as the pressure applied increased from 0 to 2.45 MPa (0 to 355 psi) (Fig. 10.14). Other Applications. The potential use of ceramic-faced steel tappets for automobiles was examined and developed by The Welding Institute. Researchers (Ref 120) found that activemetal filler metals based on silver-copper and containing titanium successfully wet both ceramic (Syalon 101) and metal (0.2C-1Cr-Mo steel) and gave sound joints. Convoluted discs of thin, soft iron made by die pressing were successfully used as interlayers, and they satisfied the initial criteria for the automobile manufacturer. Lightweight, elevated-temperature material assemblies have been evaluated by many aerospace companies for supersonic transport designs (Fig. 10.15). Studies of future reusable space transportation systems (STS) considered both insulated and hot-structure concepts. One such STS study employed the hot-structure, integral tank and fuselage concept shown in Fig. 10.16. This vehicle concept combined the functions of propellant containment, cryogenic insulation, thermal protection, and support of the vehicle thrust and aerodynamic loads. The vehicle, which was designed for 500 missions (500 ascents and 500 entries), used a large wing area to achieve a low wing loading. This design approach resulted in a longer, higher altitude entry trajectory than that flown by the Space Shuttle orbiter, which has a relatively high wing loading. This higher altitude trajectory resulted in a maximum entry temperature over much of the
Chapter 10: Applications and Future Outlook / 373
vehicle of approximately 760 °C (1400 °F), which is considerably less than that experienced on the Space Shuttle and which is within the operating range for superalloy materials such as René 41. The construction of the proposed tank wall for the hot-structure vehicle concept shown in Fig. 10.16 consists of a vacuum-sealed René 41 superalloy honeycomb-core sandwich on the lower surface of the vehicle. René 41 honeycomb sandwich panels were tested to produce combined thermal and mechanical longitudinal stresses that simulate those that would occur in a larger, more complex integral tank and fuselage structure of an Earth-to-orbit vehicle (Ref 121). Top-selling premium golf clubs share two attributes: they perform well and convey status. In the marketing-driven world of golf club manufacturing, nickel has carved a niche as an exotic material that delivers high performance. More nickel clubheads are showing up each year, and there is increasing use of nickel alloys. In the latest use, the clubface insert of a trimetal
Fig. 10.14
woods is vacuum-brazed maraging steel, a highstrength alloy with a high nickel content. The high-density nickel alloy inserts can be made thinner than steel or titanium, rendering them attractive to club designers.
Future Outlook Brazing will not go away (Ref 122–124). Materials will continue to advance. Designers will continue to place new and greater demands on structures and the materials of which they are composed. Lower cost, greater productivity, and better quality will be inseparable goals in future designs. Materials will continue to be more highly engineered. Design optimization will demand greater and more exotic combinations of materials. As a consequence, understanding of joining will have to evolve at least at a comparable pace. The life cycle of a product will not end with use and inevitable disposal but will have to take
Superplastic forming/brazing process. R.T., room temperature
374 / Brazing, Second Edition
into account reuse. Materials, if not products, will have to be rejuvenated, recycled, recovered, reconstituted . . . reborn! What is joined will have to be joined in an environmentally conscious way. Fluxes for brazing will have to be made environmentally acceptable. Pre- and postcleaning for brazing will have to be made environmentally compatible. Brazing of Dissimilar Materials. Brazing is likely to grow the most rapidly of all processes as (and if) ceramics continue to evolve and proliferate. An example of this evolution is the innovative brazing technique called WideGap joining (Material Resources Inc.) for joining of dissimilar materials. This new method (Ref 125) uses powder metallurgy preforms to produce composite braze joints with controlled thickness to compositionally tailor or offset thermal expansion mismatches. These fillermetal powder-based preforms combine nonmelting-temperature filler particles that are infiltrated during the brazing cycle by a lower-
melting-temperature braze matrix. This technique can braze across wide gaps (~0.50 to 2.5 mm, or 0.02 to 0.1 in.), where conventional brazes are normally 0.05 to 0.15 mm (0.002 to 0.006 in.) thick. Proper combination of the filler and filler-metal matrix provides composite metallurgical bonds, thus controlling coefficient of thermal expansion mismatch through the properties of the braze. These joints lower thermal stresses in the joint area and enable dissimilar materials to be brazed at high temperatures. Active elements can be added to the WideGap powder preforms, permitting the joining of ceramic, carbon-base composites and graphite materials to themselves and/or to other metals. High-strength, high-temperature joints result from tailored metallurgical interfaces and the joint composite properties that lower the residual stresses in the joint materials. Preforms are available as flexible mats (polymer-filled powder mixtures) that can be tailored to the composition of joint materials. The mats
Empennage Titanium SPF/DB sandwich Composite sandwich Composite stiffened skin Other Wing strake
Fuselage Outer wing Main wing box Titanium SPF/DB sandwich Mechanically fastened Brazed titanium spar caps
Fig. 10.15
PMC Trusses
Splice plate Sinewave PMC Continuous spar
Materials and processes, including brazing, proposed for supersonic transport vehicles. SPF, superplastic formed; DB, diffusion bonded; PMC, polymer-matrix composite
Chapter 10: Applications and Future Outlook / 375
are placed in the joints and are processed in vacuum or hydrogen brazing furnaces. Table 10.3 offers a comparison of dissimilar joining methods. Applications of WideGap brazing include:
• Chemical/metallurgical refineries: reactors and related seals • Mixers, blades • Hot gas nozzles • Aircraft and aerospace engine components
• Brazing to graphite furnace parts • Pipeline joining of dissimilar or clad materials • Pipeline cladding and repair • Repairs of dies, rolls, and molds • Conductors, busses, and coils • Electrical contacts • Graphite-metal conductors • Carbon-carbon composite shafts and seals • Ceramic-metal joining—thermal protection and management • Wear protection • Chemical attack protection
Control Systems. What will future PLC/ PC/operator-interface-panel systems look like? The next logical progression in vacuum furnace control systems probably will consist of a standalone PC without a PLC. While many users of today’s control systems are wary of committing to total dependence on a PC to control their equipment, several PC-only systems already are in operation. The main concern of PC-only control is the possibility of the PC locking up or crashing. Either of these conditions on a PC-based control system requires shutting off the PC, rebooting,
Vehicle concept
Strut
Vacuumsealed René 41 honeycomb
Titanium Frame
René 41 Slots Tank-and-fuselage cross section
Fig. 10.16
Tank-and-fuselage wall configuration
Integral tank and fuselage hot-structure concept proposed for a space transportation vehicle. Source: Ref 121.
376 / Brazing, Second Edition
and, in some cases, replacing the hard drive/ computer. Control of the equipment and the process is lost during this period, but there are remedies available, which, if not completely foolproof, will at least minimize the impact of these occurrences. A remedy to reduce the possibility of PC lockup is the installation of a more reliable, simplified operating system than that used in most modern PCs. A PC usually locks up because something objectionable exists in the operating system, programming code, software, or memory. The simpler the operating system, the less chance of lockup. Today, the preferred furnace control system still incorporates a PLC as the primary controller because of its proven reliability and simplified programming requirements. However, the PC-based system, with its inherent capability to provide additional functions, may evolve into the future furnace control system of choice (Ref 68).
Case Histories and Problem-Solving Examples Example 1: Increasing Filler-Metal Remelt Temperature. Many people appear to be interested in the remelt temperatures for various filler metals. Additionally, braze technicians wish to know how to guarantee that the filler metal will not remelt on a subsequent brazing operation or in service, causing the brazement to fall apart.
For example, if you look at the filler and base metals that have mutual solubilities and consider iron and silver, there is no solubility of silver in iron, or vice versa. When brazing these two materials together, the phase diagram shows pure silver in being brazed to pure iron would remelt at the same temperature or at temperatures above 961 °C (1762 °F), the melting point of pure silver. To increase the remelt temperature, it is essential to have sufficient mutual solubility of the filler metal, which has a low melting temperature, and the base metal, which has a much higher melting temperature, or to have an element that diffuses out of the filler metal into the base metal and makes the filler metal low melting at the brazing temperature. The result is that both the base and filler metal having a very small percentage of the melting point depressant thus raises the remelt temperature. The following combinations (Ref 126) would be helpful in determining mutual solubility: • Copper plus a silver filler metal • Iron and low-alloy steel brazed with copper • High-nickel base metals with gold-nickel filler metals • Iron- and nickel-base filler metals brazed with nickel filler metals Diffusion brazing can be successfully accomplished if there is a mutual solubility between, for example, silver filler metal and copper base metal. An example of no mutual solubility is in silver brazing of carbon and low-alloy steels.
Table 10.3 Comparison of joining technologies for dissimilar materials Conventional joints Features
WideGap (high temperatures)
Application
Requires vacuum furnace
Processing
Vacuum equipment and batch processing High brazing temperature with thick graded preform minimizes cracking.
Mechanical
Environmental Processing Characteristics
Good, enclosed Vacuum furnaces, pre-placement needed • Strengths over 340 MPa (49 ksi) • Fills wide gaps • Composites lower CTE mismatches • Can include active braze elements
CTE, coefficient of thermal expansion. Source: Ref 125
Active braze (high temperatures)
Requires vacuum furnace and premetallizing Vacuum equipment and batch processing High brazing temperature may lead to cracking of dissimilar and ceramic materials. Good, enclosed Vacuum, pre-placement needed • Strengths over 340 MPa (49 ksi) • Narrow gaps • High CTE mismatches
Premetallizing (Braze)
Slow, multistep Metallizing and vacuum equipment The lower-temperature, flexible nature of the brazing material makes it more forgiving to different coefficients of expansion. Good, enclosed Multistep requires precoating; excellent bonds • Typical strengths 24.03 MPa (3.5 ksi) • Electrically conductive • Thermally conductive • Limited to certain materials
Chapter 10: Applications and Future Outlook / 377
For some excellent examples of diffusion brazing, mutual solubility and filler/base metal remelts, see Ref 126. As a rule of thumb, in most cases, filler metals that have some degree of solubility with the base metal and have their remelt temperature raised will also have their strength increased. Aluminum, copper phosphorus, cobalt, magnesium and some other filler metals fall into the same category. Example 2: Diffusion Brazing for Jet Engine Repair. A company was repairing cracked jet engine parts using BNi-2 filler metal on nickel alloy parts. The brazing that took place was for 10 min at heat. Going back for a second brazing operation, the technicians found that the shorter time of 5 min improved the surface appearance. Why? Does boron in the braze filler metal have an effect? Boron is an element similar to nickel, chromium, and iron. When added to a nickelchromium alloy, it lowers the melting temperature of the alloy, which allows one to use the low-melting nickel-chromium-boron alloy as a filler metal. Boron is a unique element because of its small molecular size. It readily diffuses into the adjoining base metals, changing the chemistry and physical properties of the filler metal in the joint. This process has been defined by some experts as diffusion brazing. Diffusion brazing allows for the alteration of the properties of the braze joint. During diffusion brazing, the remelt temperature is increased and can reach remelt temperatures exceeding 1370 °C (2500 °F). When the filler metal is completely diffusion brazed with a nickel-base metal, the remelt temperature will be the melting point of that base metal. At the same time, the hardness will be equivalent to the base-metal hardness and tensile strength of the butt joint, and the stress rupture will be equivalent to the base metal. After full diffusion in the nickel-base metal, the filler metal will have disappeared, and the grain structure goes right across the joint. In this case, the elements of the base metal have diffused into the braze area, and the filler metal has been homogenized with the base metal. Braze repair of hot-section jet engine parts has flourished and become a multi-million dollar business with the use of nickel-base filler metals and diffusion brazing. Example 3: Braze Repair of Honeycomb Structures. In the application of repairing brazed honeycomb seals in jet engines a company found that when the worn honeycomb was
removed the following appeared: (a) erosion of base metal, (b) small cells were filled with filler metal, and (c) the prevention of flow of filler metal up the intersection where the two foils meet. The brazing of honeycomb is very interesting and started back in 1955 for aircraft wing components (Ref 30). As knowledge of filler metals, base metals, fabrication techniques, and furnaces and processing control expanded and were developed, the basic brazing problems were overcome. To overcome item (a) above, the amount of filler metal applied has to be controlled to avoid erosion. Erosion can be prevented by having only a small fillet between the core and backing. The second problem of filling the core cells is associated with the first problem. Once again, more is not better. With metal-to-metal fitup, very little filler metal is required. This will require more attention to the assembly fitup, and additional inspection may be required before brazing. For item c, we must assume that the parts are clean and the brazing atmosphere is proper; therefore, the filler metal will definitely flow up the nodes of the cells by capillary action. If it is desired to keep the filler metal out of the node, there is a particular stopoff material that will accomplish this (see Ref 127). Other nonboron filler metals that show less tendency to exhibit diffusion, solution, and erosion include BNi-5 and BNi-7 and modifications of these filler metals.
REFERENCES
1. R.W. Messler, Joining of Advanced Materials, SAMPE J., Vol 31 (No. 2), March–April 1995, p 41–45 2. M.G. Nicholas, “Material Aspects of Ceramic-Ceramic and Ceramic-Metal Bonding,” Harwell Laboratory, International Institute of Welding, Sept 1992 3. D.V. Dunford and P.G. Partridge, Strength and Fracture Behaviour of Diffusion-Bonded Joints in Al-Li (8090) Alloy, Part 1, J. Mater. Sci., Vol 25, 1990, p 4957–4964 4. G.E. Crabtree, “Atmospheres and Surface Chemistry Effects on Copper Brazing of Kovar,” Paper B2B, Joining Subgroup of the Interagency Mechanical Operation Groups—U.S. Department of Energy,
378 / Brazing, Second Edition
5.
6.
7.
8.
9.
10.
11.
12.
22nd American Welding Society International Brazing and Soldering Conf., 16–18 April 1991 (Detroit, MI) D. Manente, P. Stoute, and J.P. Immarigeon, “Development and Evaluation of Gaseous Cleaning and Diffusion Braze Repair Techniques for Cobalt and Nickel Base Superalloys,” Paper B4B, Vac-Aero International, Canada, 23rd American Welding Society International Brazing and Soldering Conf., 24–26 March 1992 (Chicago, IL) Y. Miyazawa and T. Asiga, “Brazing of a Nickel Base Alloy and Stainless Steel by Nickel Base Brazing Foils,” Paper B1D, Tokai University, Japan, 23rd American Welding Society International Brazing and Soldering Conf., 24–26 March 1992 (Chicago, IL) P.F. Timmons, “The Development of Non-Toxic Silver-Based Brazing,” Paper B3B, Husky Oil, Canada, 23rd American Welding Society International Brazing and Soldering Conf., 24–26 March 1992 (Chicago, IL) M.L. Santella, “High Strength Silicon Nitride Braze Joints,” Paper B3C, Oak Ridge National Laboratory, Tennessee, 22nd American Welding Society International Brazing and Soldering Conf., 16–18 April 1991 (Detroit, MI) R. Xu and J.E. Indacochea, “Brazing Silicon Nitride to Metals Using Active Filler Metal,” Paper B2A, University of Illinois, 23rd American Welding Society International Brazing and Soldering Conf., 24–26 March 1992 (Chicago, IL) S. Dunkerton, “Brazing of Sialon to Steel for Automotive Applications,” Paper B2E, TWI, Abington, U.K., 23rd American Welding Society International Brazing and Soldering Conf., 24–26 March 1992 (Chicago, IL) H.-K. Lee and J.-Y. Lee, Decomposition and Interfacial Reaction in Brazing of SiC by Copper-Based Active Alloys, J. Mater. Sci. Lett., Vol 11, 1992, p 550–553 M.L. Santella, J.J. Pak, and R.J. Fruehan, “Analysis of Microstructures Formed by Brazing Al2O3 and ZrO2 with Ag-Cu-Ti Alloys,” Paper B1C, Oak Ridge National Laboratory, Tennessee, 20th American Welding Society International Brazing and Soldering Conf., 3–5 April 1989 (Washington D.C.)
13. B. Palm, in Ceramics in Advanced Energy Technologies, H. Krockel, E. Merz, and O. Van der Biest, Ed., Reidel Publishing, 1984, p 178 14. H. Mizuhara and E. Huebel, Joining Ceramic to Metal with Ductile Active Filler Metal, Weld. J., Vol 65 (No. 10), 1986, p 43–51 15. T. Joy, R.M. Crispin, and M.G. Nicholas, High Technology Joining, B.A.B.S., Abington, U.K., 1987, p 23/1 16. C.M. Reyle, “Automation for Soldering and Brazing,” SME-AD-92-219, Society of Manufacturing Engineers, June 1992 17. G. Williams, Increased Productivity Through Brazing Automation, Weld. J., Vol 54 (No. 10), 1973, p 640–643 18. A. Rabinkin, “Fundamental Aspects of the Brazing Process,” Paper B4A Allied Signal, New Jersey, 20th American Welding Society International Brazing and Soldering Conf., 3–5 April 1989 (Washington, D.C.) 19. Automated Induction Brazing Boosts Production of Hydraulic Shoes, Weld. Des. Fabr., Sept 2000, p 53–55 20. C.M. Norlin, What to Consider Before Automating Brazing and Soldering Operations, Weld. J., Oct 1993, p 57–58 21. D.J. Schmitt, J.E. Maslakowski, J.L. Novak, and G.P. Starr, “Sensor-Based Control of Rocket Thrust Chamber Feature Location for Automated Braze Paste Dispensing,” M594-208, Sept 1994, NA SA 22. Connect, TWI, Abington, U.K., Dec 1989, p 6 23. H. Schoer, Aluminum Brazing Alloys for Fluxless Brazing of Aluminum, Light Metals 1989, TMS, p 699–702 24. “Research on Microwave Joining of Silicon Carbide,” final report, LA-SUB-9473, Los Alamos National Laboratory, New Mexico, 30 June 1993; National Technical Information Service Alert, 07/ 01/97-13 25. “Research on Microwave Joining of Silicon Carbide,” final report, LA-SUB95-192-PT.3, Los Alamos National Laboratory, New Mexico, 1995; National Technical Information Service Alert, 07/15/97-14 26. E. Siores and D. DoRego, Microwave Applications in Materials Joining, J. Mater. Process. Technol., Vol 48 (No. 1–4), 15 Jan 1995
Chapter 10: Applications and Future Outlook / 379
27. M.M. Schwartz, Brazing for the Engineering Technologist, Chapman and Hall, 1995 28. C.A. Blue, R.A. Blue, and R.Y. Lin, Joining of Hastelloy-X to Inconel 718 Using an Infrared Process, J. Mater. Process. Technol., Vol 58 (No. 1), 1 March 1996 29. C.E. Witherell and T.J. Ramos, Laser Brazing, Weld. J., Vol 59 (No. 10), Oct 1980, p 267–277 30. M.M. Schwartz, Brazing, ASM International, 1987 31. T.A. Jones and C.E. Albright, Laser Beam Brazing of Small Diameter Copper Wires to Laminated Copper Circuit Boards, Weld. J., Vol 63 (No. 12), Dec 1984, p 34–47 32. V.R. Davé, R.W. Carpenter, D.T. Christensen, and J.O. Milewski, Precision Laser Brazing Utilizing Non-Imaging Optical Concentration, Weld. J., June 2001, p 142–147 33. Electron-Beam Welding and VacuumFurnace Brazing Combine to Fabricate Impellers, Weld. Des. Fabr., Nov 1997, p 8–9 34. Electron Beam Braze-Welds Compressor Impellers, Weld. Des. Fabr., June 1999, p 14 35. M.L. Tuma, Electron Beam Bonds Sapphire Fiber to Platinum Shell, Adv. Mater. Process., Sept 1998, p 14–15 36. V.M. Nerovniyi and K.O. Peremit, Hardening and Restoration of Working Surfaces of Titanium Components by Vacuum Arc Brazing, Weld. Int., Vol 9 (No. 10), 1995, p 815–817 37. I. Onsawa and A. Suzumura, Brazing of Titanium Using Low Melting Point TiBased Filler Metals, Weld. J., Vol 11, 1991, p 462–467 38. S.V. Lashko and E.I. Vrublevskii, Technology of Brazing Components in Engineering: A Handbook of the Designer, Publ. Mashinostroenie, Moscow, 1993 39. M. Ebbinghaus and H.-D. Prinz, “Arc Brazing of Steel Components with Copper-Based Filler Materials,” Report 192, DVS-Bericht, 1998, p 146–149 40. C. Radscheit, Breakthrough in MIG Brazing, Automobilproduktion, June 1998, p 117 41. H. Hackl, MIG Brazing of Zinc-Coated Metal Sheets and Sections, Schweissen Schneiden, Vol 540 (No. 6), 1995, p 351–354
42. B. Bouaifi, Low-Heat Joining of SurfaceCoated Sheet Metal Materials by Plasma Brazing, DVS Jahrbuch Schweisstechnik, 2001, p 115–122 43. R.V. Hughes, G. Dryburgh, and S. Garbett, Plasma Braze Welding in Autobody Production at Jaguar Cars, Weld. Met. Fabr., March 1995, p 110–111 44. C. Radscheit and J. Muller-Rogait, Plasma Brazing—Basic Principles and Application—New Brazing Technologies for Thin Sheet Metal, SLV, Munich, Germany, March 2000 45. R.P. Walduck and J. Lindstrom, Plasma Braze Welding in Autobody Construction, Proceedings from the American Welding Society Annual Conference, April 1997 (Los Angeles, CA), p 215–216 46. B. Bouaifi, A. Ait-Mekideche, and C. Radscheit, Modern Brazing Techniques in the Automotive Industry, Proceedings of the Int. Brazing and Soldering Conference, 2–5 April 2000 (Albuquerque, NM), AWS, p 153–161 47. B. Walduck, Using Plasmabraze in Car Body Fabrication, Weld. Met. Fabr., Aug 1999, p 11–14 48. U. Draugelates, B. Bouaifi, A. Helmich, B. Ouaissa, and J. Bartzsch, Plasma-Arc Brazing, Weld. J., March 2002, p 38–42 49. “Copper and Copper Alloys, Rods, and Wires for Use as Welding and Hard-Brazing Fillers,” DIN 1793352, Deutsche Industrie-Normen, Sept 1998 50. “Hard Brazing Fillers—Copper-Based Brazing Fillers, Composition, Applications, Technical Delivery Conditions,” DIN 8513, Parts 1 and 2, Deutsche Industrie-Normen, Oct 1979 51. J.T. Niemann and R.A. Garrett, Eutectic Bonding of Boron-Aluminum Structural Components, Part I, Weld. J., Vol 53 (No. 4), April 1974, p 175–184 52. J.T. Niemann and R.A. Garrett, Eutectic Bonding of Boron-Aluminum Structural Components, Part II, Weld. J., Vol 53 (No. 8), Aug 1974, p 351–359 53. J.T. Niemann and G.W. Wille, Fluxless Diffusion Brazing of Aluminum Castings, Weld. J., Vol 57 (No. 10), Oct 1978, p 285–291 54. D.S. Duvall, W.A. Owczarski, and D.F. Paulonis, TLP Bonding: A New Method for Joining Heat Resistant Alloys, Weld. J., Vol. 53 (No. 4), Apr 1974, p 302–314
380 / Brazing, Second Edition
55. M.M. Schwartz, Diffusion Brazing Titanium Sandwich Structures, Weld. J., Vol 57 (No. 9), Sept 1978, p 35–38 56. W.A. Owczarski and D.F. Paulonis, Application of Diffusion Welding in the USA, Weld. J., Vol 60 (No. 2), Feb 1981, p 22–23 57. B. Norris and F. Goiny, Joining Processes Used in the Fabrication of Titanium and Inconel Honeycomb Sandwich Structures, Rohr Industries, Inc., First Society for the Advancement of Materials and Process Engineering International Metals Symposium, 18–20 Aug 1987 (Cherry Hill, NJ) 58. L.B. Blackburn, “Effect of LID Processing on the Microstructural and Mechanical Properties of Ti-6Al-4V and Ti6Al-2Sn-4Zr-2Mo Titanium Foil-Gauge Materials,” NASA TP-2677, National Aeronautics and Space Administration Langley Research Center, April 1987 59. P. Adam and L. Steinhauser, Bonding of Superalloys by Diffusion Welding and Diffusion Brazing, AGARDCP-398, Motoren Turbinen Union, 61st Meeting of Structural and Materials Panel of Advisory Group for Aerospace Research and Development, 11–13 Sept 1985 (Oberammergau, Germany), July 1986, p 9-1 to 9-6 60. W. Blair, “Fabrication of Titanium Multiwall Thermal Protection System (TPS) Curved Panel,” NASA-CR-165754, Cont. NASI-15646, Rohr Industries, Inc., Aug 1981 61. P.E. Doherty and D.R. Harraden, New Forms of Filler Metal for Diffusion Brazing, Weld. J., Vol 56 (No. 10), Oct 1977, p 37–39 62. R.V. Sharples and I.A. Bucklow, “Diffusion Bonding of Aluminum Alloys to Titanium,” 7836.01/85/448.3, 307/1986, The Welding Institute, July 1986 63. T.K. Ferguson, AJMing Braze Spatter, Am. Mach., May 1977, p 87–88 64. D.D. Berger, Step Brazing a Multi-Target TRAX, Weld. J., Vol 62 (No. 10), Oct 1983, p 41–43 65. J.R. Kelley, Brazing Makes Frames for Jet Aircraft Engines, Weld. Des. Fabr., July 1984, p 43–45 66. O. Ohkubo, Fabrication of Aluminum Alloy Radiators by Vacuum Brazing in
67. 68.
69.
70.
71.
72.
73.
74.
75.
76.
77. 78. 79.
Dedicated Furnaces, Ind. Heat., Jan 1993, p 53–58 D. Garg and P.T. Kilhefner, Moisture-Free Atmosphere System for Brazing Ferrous Metals, Ind. Heat., April 1999, p 49– 52 J. Terchila and J. Kowalewski, Vacuum Furnace-Control Systems: Past, Present and Future, Ind. Heat., Oct 2001, p 113– 116 R.L. Keusseyan and J.L. Dilday, “Development of Brazing Interconnection to Low Thermal Expansion Glass-Ceramics for High Performance Multichip Packaging,” Electronics Components and Technology Conference-93, 1–3 June 1993 (Florida) D.A. Weirauch, Jr., Mechanical and Environmental Tests of Brazed Ceramic Multiplayer Packages, Weld. J., May 1994, p 110–118 H. Nakagawa, C.H. Lee, and T.H. North, Modeling of Base Metal Dissolution Behavior During Transient Liquid-Phase Brazing, Metall. Mater. Trans. A, Vol 22 (No. 2), Feb 1991, p 543–555 K. Takeshita, Model Equation for Predicting the Tensile Strength of ResistanceBrazed Joints, Weld. J., Sept 2000, p 261–265 B.H. Rabin, Joining of Fiber-Reinforced SiC Composites by In Situ Reaction Methods, Mater. Sci. Eng. A, Vol 130, 1990, p L1–L5 A.J. Caputo and W.J. Lackey, “Fabrication of Fiber-Reinforced Ceramic Composites by Chemical Vapor Infiltration,” ORNL Technical Memo 9235, Oak Ridge National Laboratory, Tennessee, 1984 D.P. Stinton, A.J. Caputo, and R.A. Lowden, Synthesis of Fiber-Reinforced SiC Composites by Chemical Vapor Infiltration, Am. Ceram. Soc. Bull., Vol 65 (No. 2), 1986, p 347–350 A.J. Caputo et al., Fiber-Reinforced SiC Composites with Improved Mechanical Properties, Am. Ceram. Soc. Bull., Vol 66 (No. 2), 1987, p 368–372 T.J. Moore, Feasibility Study of the Welding of SiC, J. Am. Ceram. Soc., Vol 68 (No. 6), 1985, p C151–C153 M.G. Nicholas and R.M. Crispin, Diffusion Bonding, R. Pearce, Ed., Cranfield Institute of Technology, 1988, p 173 C.H. Bates et al., Joining of Non-Oxide Ceramics for High-Temperature Applica-
Chapter 10: Applications and Future Outlook / 381
80.
81. 82.
83.
84.
85.
86. 87.
88. 89.
90.
91. 92.
93.
tions, Am. Ceram. Soc. Bull., Vol 69 (No. 3), 1990, p 350–356 T. Iseki, K. Arakawa, and H. Suzuki, Joining of Dense Silicon Carbide by Hot Pressing, J. Mater. Sci. Lett., Vol 15, 1980, p 1049–1050 S. Yajima et al., Joining of SiC to SiC Using Polyborosilozane, Am. Ceram. Soc. Bull., Vol 60 (No. 2), 1982, p 253 J.A.P. Gehris, “High Temperature Bonding of Silicon Carbide,” Master’s thesis, New Mexico Institute of Mining and Technology, Socorro, NM, 1989 N. Tamari et al., Joining of Silicon Carbide Ceramics with Si3N4-Y2O3-La2O3MgO Mixture, Jpn. Ceram. Soc. Vol 94 (No. 10), 1986, p 1087–1091 S. Morozumi et al., Bonding Mechanism Between Silicon Carbide and Thin Foils of Reactive Metals. J. Mater. Sci., Vol 20, 1985, p 3976–3982 J.K. Boadi, T. Yano, and T. Iseki, Brazing of Pressureless Sintered SiC Using AgCu-Ti Alloy, J. Mater. Sci., Vol 22, 1987, p 2431–2434 Glue Solves Sticky Ceramic Joint Problem, Mach. Des., 25 Feb 1999, p 47 A.K. Chattopadhyay, L. Chollet, and H.E. Hintermann, Experimental Investigation on Induction Brazing of Diamond with Ni-Cr Hardfacing Alloy Under Argon Atmosphere, J. Mater. Sci., Vol 26, 1991, p 5093–5100 N. DeChristofaro and C. Henschel, Metglas Brazing Foil, Weld. J., Vol 57 (No. 7), July 1978, p 33–38 A. Datta, A. Rabinkin, and D. Bose, Rapidly Solidified Copper-Phosphorus Base Brazing Foils, Weld. J., Vol 63 (No. 10), Oct 1984, p 14–21 A. Datta and N.J. DeChristofaro, Rapidly Solidified Filler Metals, Brazing and Soldering Applications: Rapidly Quenched Metals, Elsevier Science Publishers, 1985, p 1715–1722 G.S. Hoppin III and T.F. Berry, Activated Diffusion Bonding, Weld. J., Vol 49 (No. 11), Nov 1970, p 505–509 “Brazing, High Temperature Brazing and Diffusion Welding,” IBSN3-87155-4308, Second International Conf., 19–20 Sept 1989 (Essen, Germany), International Institute of Welding M. Nakahashi, S. Suenaga, M. Shirokane,
94.
95.
96. 97.
98. 99.
100. 101.
102.
103.
104.
105.
106.
et al., Transient Liquid Phase Bonding for Ni-Base Superalloys, Mar-M247 and IN939, Mater. Trans., JIM, Vol 33 (No. 1), 1992, p 60–65 Y. Nakao, K. Nishimoto, K., Shinozaki, et al., “Transient Liquid Insert Metal-Diffusion Bonding of Nickel-Base Superalloys,” Osaka University and Pusan National University, International Institute of Welding, Sept 1992 P. Kauppinen and J. Kivilahti, Evaluation of Structural Defects in Brazed Ceramicto-Metal Joints with C-Mode Scanning Acoustic Microscopy, NDTE Int., Vol 24 (No. 4), Aug 1991, p 187–190 M.M. Schwartz, Handbook of Structural Ceramics. McGraw-Hill, 1992 N. Yoshimi, H. Nakae, and H. Fujii, New Approach to Estimating Wetting in a Reaction System, Mater. Trans., JIM, Vol 30, Feb 1990, p 41–47 O.M. Akselsen, Review Advances in Brazing of Ceramics, J. Mater. Sci., Vol 27, 1992, p 1989–2000 R. Smith, M. Thorpe, and E. Lugscheider, New Materials Increase Applications for Brazing, Part II, Ind. Heat., Dec 1997, p 43–49 T. Walz, Brazing Ceramics, Ceram. Ind., Dec 1999, p 47–50 A.-P. Xian and Z.-Y. Si, An Improvement of the Oxidation Resistance of Ag-Cu Eutectic-5 at.% Ti Brazing Alloy for Metal/Ceramic Joints, Mater. Lett., Vol 12, 1991, p 84–88 H. Sueyoshi, M. Tabata, and Y. Nakamura, Joining of Silicon Nitride to SUS 304 Stainless Steel Using a Sputtering Method, J. Mater. Sci., Vol 27, 1992, p 1926–1932 A.-P. Xian and Z.-Y. Si, Direct Brazing of Si3N4-Steel Joint Using Ag-28 Cu Brazing Filler Metal with Ti Interlayer, J. Mater. Sci. Lett., Vol 10, 1991, p 1381–1383 R.R. Kapoor and T.W. Eager, Oxidation Behavior of Silver- and Copper-Based Brazing Filler Metals for Silicon Nitride/ Metal Joints, J. Am. Ceram. Soc., Vol 72, March 1989, p 448–454 R.R. Kapoor and T.W. Eager, Tin-Based Reactive Solders for Ceramic/Metal Joints, Metall. Mater. Trans. B, Vol 20 (No. 6), Dec 1989, p 919–924 A.G. Evans, M. Ruhle, and M. Turwitt, On
382 / Brazing, Second Edition
107.
108. 109.
110.
111.
112.
113.
114.
115.
the Mechanics of Failure in Ceramic/ Metal Bonded Systems, J. Phys. Colloque, Vol 46 (No. C4), April 1985, p 613–626 M. Naka, T. Tanaka, and I. Okamoto, Amorphous Alloy and Its Application to Joining, Trans. JWRI, Vol 14 (No. 2), Dec 1985, p 185–195 K. Suganuma, T. Okamoto, M. Koizumi, et al., J. Mater. Sci. Lett., Vol 4, 1985, p 648 M. Naka, and T. Saito, Niobium Interlayer for Joining SiC to Stainless Steel, J. Mater. Sci. Lett., Vol 10 (No. 6), March 1991, p 339–340 S. Morozumi, M. Endo, M. Kikuchi, et al., Bonding Mechanism Between Silicon Carbide and Thin Foils of Reactive Metals, J. Mater. Sci., Vol 20 (No. 11), Nov 1985, p 3976–3982 M. Naka, T. Saito, and I. Okamoto, Proc. of the Fourth International Symposium on the Science and Technology of Sintering, Elsevier Applied Science, 1988, p 1373 C. Qiu and H. Xia, A Research on the Mechanism of Joining Graphite to Ceramics of 95% Al2O3, Advances in Joining Newer Structural Materials, 23–25 July 1990 (Montreal, Canada), Pergamon Press, p 129–135 C. Qiu, Mechanism Research on Seal Between 95% Alumina Ceramics and Activated Metal of Oxygen-Free Copper, Proc. of Fourth Annual Meeting of PRC Elec. Soc., E-Vacu. Assoc., 1982, p 155– 156 C. Qiu, Technical Improvements on the Weld of Beryllium and Beryllium Oxide Ceramics, Proc. of the Experimental Exchange Meeting on Brazing Matl. and Its Applications, Vol 1, 1984, p 8 C. Qiu, Advantages of Titanium and its Alloy in Welding Applications, Proc. of Fifth National Weld Symposium, 1989, AWS, p 8
116. M.S. Yeh and T.H. Chuang, Super Plastic Forming/Brazing Process for Inconel 718 Superalloy Components, Weld. J., May 1997, p 197–200 117. T.H. Chuang, Method for Superplastic Forming with Concurrent Brazing Bonding, BRD Patent DE4200047C2, 1992 118. M.S. Yeh, C.W. Tsau, and T.H. Chuang, Evaluation of Superplastic Formability of SP-Inconel 718 Superalloy J. Mater. Eng. Perform., Vol 5 (No. 1), 1996, p 71– 77 119. A. Rabinkin and S. Pounds, Effects of Load on Brazing with Metglas MBF2005 Filler Metal, Weld. J., Vol 67 (No. 5), 1988, p 33–45 120. I.A. Bucklow, J.H. Potter, and S.B. Dunkerton, “Development of a Brazed Ceramic-Faced Steel Tappet,” TWI Report 450/1992, June 1992 121. J.L. Shideler, R.A. Fields, L.F. Reardon, and L. Gong, “Thermal and Structural Tests of René 41 Honeycomb IntegralTank Concept for Future Space Transportation Systems,” NASA-TP-3145, L16752, National Aeronautics and Space Administration, May 1992 122. B. Irving, Brazing and Soldering: Facing New Challenges, Weld. J., Oct 1998, p 33–37 123. S. Yamada and K. Masubuchi, Advanced Welding Technology Keeps Japan’s High-Speed Trains on Track, Weld. J., Nov 2000, p 48–53 124. “Applications of Diffusion Joining in Industries in Germany,” IA-421-93/OE, Working Group “Diffusion Welding,” Ed., DVS, July 9, 1993 125. Material Resources International, Lansdale, PA, www.materialresources.com 126. R.L. Peaslee, Brazing Q&A, Weld. J., July 1999, p 115 127. R.L. Peaslee, Brazing Q&A, Weld. J., April 2002, p 46, 66
© 2003 ASM International. All Rights Reserved. Brazing (#06955G)
www.asminternational.org
Subject Index For general and specific information on filler metals, see Filler Metal Index.
A Acceptance limits, 340, 341, 342 Acid cleaning, 303, 304 Acoustic emission testing, 343 Activated diffusion bonding (ADB). See transient liquid phase (TLP) bonding Activated diffusion healing (ADH), 302(F), 302–303 Activating metals, 254, 255 Active braze alloy (ABA), 213(F), 334, 335(F), 336(F) Active brazing technology, 132 Active coatings, 227 Active filler metal (ABA), 128–129 Active-hydride process, 330, 331 Active metal brazing (ABA), 141, 196 Active metal processes, 150, 330, 331 Active-metal (reactive) brazing, 177 Active-substrate process, 334(F) Admiralty brass, 80–81 Adsorption-type driers, 256 Aeroquip-Vickers Limited, 349 Aerospace brazed manifolds and tube assemblies, 26(F) Aerospace Material Specification (AMS) for filler metals, 367 Air pressure testing, 342–343 Alkalis, 267 Allotropic transformation, 360 Alloy tool steels, 106 Alloying, 67 Alpha-beta transformation temperature, 209 Alpha2 aluminides, 138 Alumina (Al2O3) filler metals for, 128, 129, 130, 195 filler metals systems for, 129 joints, bending strength of, 131(T) joints TiC strengthened, bending strength of, 132(T) to Kovar, methods for, 142(F) reaction layer technique, 130 and titanium brazing, 133 titanium filler metals test header, 128(F) Alumina-ceramic brazing, 365
Alumina dispersion-hardened copper (ADHC), 82 Alumina enriched paper, 296 Alumina nitride (AlN), 128 Alumina paper, 296(F) Aluminides, 138–139 Aluminum alloy radiator, 363(F) Aluminum and aluminum alloys, 71–77, 303–304 acid cleaning, 303 aluminum-lithium alloys, 73, 74(T) aluminum oxides, polymeric sealing, 303 aluminum silicon castings, 73 base metals used for cladding, 352 brazing under vacuum, 30 caustic cleaning, 303 commercial filler metals for, 72 composition limits for, 75(T) contact-reactive brazing for, 184 diffusion bonding for, 74 dip brazing of, 73, 303, 318 dispersion-strengthened aluminum, 77 fluxless brazing of, 183 furnace brazing of, 244 galvanic corrosion, 183 heat-treatable, 71 magnesium content in, 71 metallic coatings for, 184 non-heat-treatable, 71 oxide films, 184 oxide of, 263 precipitation-hardening of, 183 rapid-solidification powder metallurgy for, 76–77 refractory oxides of, 88 Aluminum and aluminum-silicon filler metal systems, 78 Aluminum bonded with copper-silver alloy interlayer, 75(F) Aluminum brazing, 273, 362–363 fluxless methods for, 351 in high vacuum, 253 methods for, 72–73 techniques for, 277–278
© 2003 ASM International. All Rights Reserved. Brazing (#06955G) 386 / Brazing, Second Edition
Aluminum brazing (continued) in a vacuum partial pressure atmosphere, 248 Aluminum brazing fluxes, 277–278 Aluminum-bronze brazing, 276 Aluminum bronze fluxes, 270 Aluminum bronzes, 81 Aluminum fluxes, 270 Aluminum-matrix composites, 137 Aluminum nitride (AlN), 128 Aluminum nitride (AlN) ceramics, 212 Aluminum oxide, 296 Aluminum-silicon filler metals, 182(T) Aluminum to copper brazing, 70 American National Standards Institute, 341 American Society of Mechanical Engineers (ASME), 342 American Welding Society (AWS) AWS filler metals, 367 brazing documents, 341(T) Recommended Practices for the Design, Manufacture, and Examination of Critical Brazed Joints, 339, 340 Standard Methods for Evaluating the Strength of Brazed Joints in Shear, 345 Amorphous filler metal foil (MBF), 222 Amorphous metal alloys (AMA), 230 Amorphous structure, 214 An-glass (CaAl2Si3O8), 208–209 Argon (Ar) atmosphere for furnace brazing chart, 33(T), 257(T) carbon transfer, 295, 3 10 for certain metals, 95, 99, 112, 117, 284, 286 controlled atmosphere brazing, 32, 36 induction brazing, 82, 367. See also backfilling infrared brazing, 352 partial vacuum, 265 quenching with, 252 uses of, 264 Atmosphere brazing system (ABS), 245–249 Atmosphere type, 234 Atmospheres atmosphere applications, 255–256 atmosphere components (gases), 260–265 atmosphere composition, 256–260 atmosphere control, 60, 282 boron in, 263 for brazing, 257(T) for brazing base and filler metals, 111–112, 248, 282, 284, 286 chromium fluoride atmosphere, 302 controlled atmospheres, 32, 33(T), 36–37, 243–244 dewpoint temperature and moisture content in, 264(T) exothermic atmosphere, 256 fluorocarbon atmosphere, 302 furnace atmospheres, 251(T), 264(T) for furnace brazing, 24–25, 26, 26(F), 32 graphite fixtures, 295, 309–310 hydrogen atmospheres, 232(F), 258(F), 259, 284, 295, 309–310 induction brazing, 36–37 inert atmospheres, 245, 284
www.asminternational.org
joining atmospheres, 244–255 and laser brazing, 353 metal/metal oxide equilibrium curves, 232(F), 258(F) moisture-free, 363, 364(F) nickel-plating thickness requirements for furnace atmosphere brazing, 228(T) nitrogen atmospheres, 261, 263 oxide reduction, 259 oxidizing atmospheres, 245 problem solving tips, 60, 282, 309–310 protection for laser brazing, 353 quality, 234 reducing atmospheres, 245–249 selection of, 32 semi-continuous controlled atmosphere type (CAB), 26(F), 26–27, 27(F) system types, 245–249, 252–255 tests for braze atmospheres, 249 types of, 234, 244(F) vacuum atmospheres, 249–255 vacuum/partial pressure, 248 wetting problems, 302 Atmospheric brazing system (ABS), 245–249 Atmospheric control, 245, 256 Austenitic nonhardenable stainless steels, 106–117 Austenitic stainless steels, 106–107 Automated brazing, 349 Automated induction brazing, 350 Automation, 5, 317, 347–351. See also robotics automatic brazing machine, 349(F) in brazing fluxes and pastes, 274, 348, 350 in brazing system, 24(F) in cleaning processes, 348 economics, 58–59, 179, 347 in flame brazing, 350 in induction brazing, 349–350 machine costs reductions in, 348 multiaxis seam tracking (MAST), 224 numerically controlled positioning equipment, 354 programmable logic controllers (PLC), 364 self-locating assemblies in, 348 steps to, 348–349 in torch brazing, 23 Auxiliary fixtures, 293
B Backfilling, 31–32, 163, 249, 262 Barrier coatings, 180–181, 227, 228 Barrier (tape) curtains, 27, 28(F) Base metals/materials, 347–348 characteristics of, 15 composition and strength of, 313 containing aluminum and titanium, 233 dissolution, 93 effects of brazing variables on clearance, 322–323 erosion, 341 family groups, 70–71 fluxes and filler metals, 66–67, 275–278, 322 hardenability of, 15
© 2003 ASM International. All Rights Reserved. Brazing (#06955G)
heat resistant, 16 heat treatables, 63 heat treatment of, 16 for hypergolic fuel, 39 metallurgical phenomena, 15 prebraze cleaning methods, 299(T) reactions with, 177 relative ease of brazing various, 16(T) residual stresses, 15 temperatures for brazing, 177 Baskets and trays, 292 Batch furnaces, 25, 29, 246 Bending strength of SiC-SiC joints, 125(F) Beryllium dip brazing of, 78 furnace brazing of, 78 induction brazing of, 78, 275 Beryllium and beryllium alloys brazing of, 275, 284, 286 filler metals for, 78, 79 flame cleaning of, 304 low ductility of, 77 wetting problems of, 77 Beta transus temperature, 209 Black boron nitride, 263 Black residue, 282 Blasting, materials and methods, 18, 300 Blocks, specialized fixtures for vacuum brazing, 293 Blowoff technology, 274 Blue-gray nitrogen film, 248, 263 Boiler and Pressure Vessel Code (American Society of Mechanical Engineers), 342 Bond area for different joint configurations, 314(F) Bonding, 8, 179 Bonding area strength, 327–328 Bonding failure and dissimilar-metal joints, 328 Borates, 267 Boric acid, 267 Boron aluminum-boron composites, 137 atmospheres, 263 boron-aluminum brazing, 360 boron effects, 249 ceramic stopoff materials, 296 diffusion brazing, 360, 377 diffusion of, 368–369 discoloration, 263 elemental boron powder, 267 in filler metals, 248–249, 250, 251(T), 257, 377 filler metals for, 157–158 grain boundary diffusion, 197 metal-matrix composite (MMX), 137 and nitrogen, 249, 263 Boron-modified fluxes, 157–158 Boron nitride, 263, 296, 297 Boundary zone, 4 Brasses, 80–81, 160, 189 furnace brazing of, 78 torch brazing of, 80 Braze coat, 229
www.asminternational.org Subject Index / 387
Braze-coat process, 229(F) Braze failure and dissimilar-metal joints, 328 Braze interlayer thickness, 57–58 Braze processing systems, 362–363 Braze repair of superalloys, 308 Braze welding, 53–54 Brazements ceramic-bonded exhaust valve, 333(F) ceramic-bonded fuel nozzle, 334(F) design of, 339 inspection of, 339 marine diesel engine components, 333 quality requirements for, 339 rotors made of copper bars, 84(F) shear strength of, 327 of silicon-carbide, 127(F) tensile strength of, 327 Brazing active, for alumina (Al2O3) to Kovar, 142(F) advantages and limitations of, 3–4 below the liquidus temperature, 235 defined, 1 dissimilar metals, 374–375 filler-metal selection chart, 250–251(T) filler-metal systems joint design for, 20(F) historical development of, 2–3 under load, 216 mechanics of, 4–5 of metal and ceramics, 231(F) metallurgical basis of, 2 methods for aluminum and aluminum alloys, 72–73 vs. other welding processes, 4–5 procedures for copper alloys, 81–83 of SAE 1541 steel, 247 with silicon interlayer, 208 vs. soldering, 1–2 solid solutions in, 2 support fixtures, 291 temperatures, 179, 234 toxic metals, 256 typical cycle of, 254 Brazing control systems, 365, 375–376 Brazing defects. See defects Brazing fluxes. See flux/fluxes Brazing furnace, 274 Brazing joints. See joints Brazing pastes, 43, 185(F) Brazing processes compared with soldering and welding, 2(T) internal joints simultaneously brazed, 4(F) to join large surface areas, 4(F) new method of, 76(F) for nickel and nickel alloys, 88 planning and control, 341–342 space transportation vehicle, 374(F) used to replace complex castings, 3(F) Brazing systems, advantages of, over welding, 37–38 Brazing techniques, 16–17, 89 Brittle intermetallics, 66, 114, 177, 193 Bronzes, 81
© 2003 ASM International. All Rights Reserved. Brazing (#06955G) 388 / Brazing, Second Edition
Brushing, 271 Butt and lap joints, 313–315 Butt-joints, 20, 20(F), 314 Butt-lap and scarf joints, 315
C C-mode scanning acoustic microscopy (C-SAM), 370 Cadmium addition to Ag-Cu-Zn filler metals group, 201 in atmospheric furnace, 256 fumes, 119 standards, 187 substitutes, 187, 189, 191(T) toxicity of, 187 in vacuum furnace, 256 Capillary attraction, 4–5, 11–12, 313 Carbide precipitation, 15, 63–64 Carbides atmosphere for brazing, 257 and ceramics, 117–119 described, 306 filler metals for, 187 fluxes for, 257, 268–269(T), 270 formation in other metals, 64, 95, 159 newly developed types of, 295 Carbides and cermets, 117–119 furnace brazing of, 118 Carbon and carbon brazing coefficient of thermal expansion (CTE), 139 filler metals, 140 shear cracking, 140 Carbon and low-alloy steels cleaning methods for, 305–306 described, 102 dip brazing of, 105 filler metals for, 103(T), 105 furnace brazing of, 102, 104 hardenable, 102 induction brazing of, 145, 188 temperatures for brazing, 104 torch brazing of, 102, 103(T), 104 Carbon dioxide (CO2), 261 Carbon electrodes, 42 Carbon fibers, 213 Carbon monoxide atmosphere, toxics of, 32, 256, 261 Carbon monoxide (CO), 260–261 Carbon pickup, 309 Carbon resistance brazing, 42 Carburizing flame, 23 Carburizing (fluxing) salts, 105 Carrier gas flow rate, 248 Cartridge brass, 160 Case hardened (nitrided) steels, 249 Cast brazing, 146 Cast iron, 100–105 brazing processes for, 100–101 dip brazing of, 100 electro-chemical cleaning of, 304 electrolytic treatment of, 101
www.asminternational.org
furnace brazing of, 101, 102, 159 graphite in, 100 temperatures for brazing, 102 Caustic cleaning, 303 Cavity coalescence, 113 Cavity nucleation, 113 Cemented carbides, 201, 205, 206, 286, 306 Central eutectic zone and joint brittleness, 216 Ceramic stopoff materials, 296 Ceramics, 119–134, 306. See also specific ceramics (e.g., silicon carbide) atmosphere for brazing, 286 brazements of, 333(F), 334(F) brazing, 121–122 ceramic bonding, 191, 193, 330 ceramic brazes, 208 ceramic joints with glue, 367 ceramic materials, by microwave brazing, 56–57 ceramic-to-ceramic brazing, 132, 205, 333–335, 335(F), 336(F) ceramic-to-matrix composite (CMC) materials, 160–161 ceramic-to-metal brazing, 141–144, 150–159, 157(F), 196, 205, 329(F), 331, 333–335 ceramic-to-metal-to-graphite joining, 371–372 coatings and metallizing, 10, 150–151, 153(T), 330 filler metals for, 150–159, 153(T), 160–161, 196, 205, 331, 371 hermetic bonds, 141 interlayers for reducing thermal expansion mismatch, 157(F) joint designs, 329(F) mechanical behavior of, 329 multilayer interlayer, 132 special joining methods, 141–144 surface condition effect on joint strength, 141–142 thermal expansion, 10, 121 wetting, 121, 207, 330, 371 Cerastar RB-SiC (silicon-carbide), 126 Cermets, 117, 118 Charpy impact testing, 159, 216, 218 Chemical bonding, 8 Chemical cleaning, 18, 298, 299–300, 359 Chemical composition and mechanical properties of filler metals used for fluxless brazing, 255(T) Chemical etching (pickling). See chemical cleaning Chemical fluxes, 267 Chemical inspection, 346 Chloride formulations, 277 Chloride stress-corrosion cracking, 235 Chlorides, 267 Chromium carbide, 64 Chromium fluoride atmosphere, 302 Clad materials, 327 Clad self-filler metal, 229 Cladding thickness ratio, 327 Clean rooms, 297, 306–307 Cleaning methods and procedures. See also chemical cleaning acid cleaning, 303, 304 agents, selection of, 299–300
© 2003 ASM International. All Rights Reserved. Brazing (#06955G)
automation of, 348 for base metals, 280(T), 299(T), 303, 304, 305 blasting materials and methods, 18, 300 caustic cleaning, 303 chemical cleaning, 18, 298, 299–300, 359 cleaning methods, 280(T), 299(T), 301–302, 348, 359 degreasing methods, 298 effects of, 359 flame cleaning, 304 fluoride-ion cleaning, 303, 303(F), 306, 308(F) flux removers, 279 fluxes, 18 for fluxless brazing, 303 ion bombardment, 301–302 mechanical cleaning, 18, 298, 299(T), 300 overexposure in chemical cleaning, 305 phosphate acid cleaners, 305 postbraze, 278–281, 308(T) prebraze, 299(T) precleaning, 297–302 precoating and finishing, 300–301 process operations, 49 quality assurance, 305 repair techniques with cleaning agents, 308 solvent cleaning, 303 specialized processes, 300–301 surface cleaning and preparation, 18, 297–302 thermal treatments, 300 toxics in cleaning materials, 348, 359 ultrasonic cleaning, 279 for vacuum brazing, 307–308 by vacuum heat treatment, 359 Coatings and metallizing, 179–181 for alumina (Al2O3) to Kovar, 142(F) for aluminum and aluminum alloys, 184 application methods and techniques, 180, 206 applications of, 180 barrier coatings, 180–181 for brazing applications, 227 ceramics, 10, 150–151, 286, 330 electron beam evaporation, 252 metallizing paint, 142–143 noble metals, 180(F) passive coatings, 227 physiochemical coatings, 180 process of, 142 radio frequency sputtering, 252 reactive metal, 180 transition-metal barrier, 180(F) types of, 180 Cobalt-base alloys brazing techniques for, 89 filler metals for, 89 stress-corrosion cracking (SCC), 89 Coefficient of thermal expansion (CTE), 177 joint design, 19 residual stresses, 15 vs. thermal expansion curve, 328 Cold-wall furnace, 29 Color match, 17, 178. See also discoloration
www.asminternational.org Subject Index / 389
Combining two filler metal families, 189 Commercially available noble filler metals that may wet pretreated graphite, 211(T) Common imperfections in brazed joints, 342 Comparison of brazing properties for filler metals used to replace cadmium-containing filler metals, 191(T) Complex nitrides, 10 Composite materials, 53–54 Composites, 136–138 aluminum-boron composites, 137 aluminum-matrix composites, 137 copper, other metals and composites, 83–85 dispersion hardened composites, 54 dissimilar material combinations, 147–150 fiber-reinforced SiC composites, 366–367 graphite, composites, diamonds, aluminides to metals, 147–150 matrix-metal composites, 53 metal-matrix composites (MMC), 137, 157 Compression of bonding zone, 54 Computer developments, 376 Concentricity problems, 19 Contact angle of a liquid, 7 Contact-reactive brazing, 184, 210 Contact-solid-phase melting, 188 Continuous-type furnace, 27 Continuous-type furnace brazing, 25–26 Continuous vacuum brazing furnace, 362(F) Continuous wave CO2 laser, 353 Control systems, 375–376 programmable logic controllers (PLC), 350, 363–365, 365(F) Controlled atmospheres, 32, 33(T), 36–37, 243–244 Conveyor brazing system, 350 Cooling rate, diagram of, 45(F) Cooling systems, 31–32 Copper and copper alloys, 79–85 brazing procedures for, 81–83 carbon resistance brazing of, 42 to ceramics, brazing of, 245 chemical cleaning, 304 composites of, 83–85 copper-brass (CuproBraze), 85 copper-manganese filler metals, 84 copper-manganese-tin, 190 copper nickels, 81 copper-steel, 83, 84 copper-tin, 193 copper-titanium, 190–193 cracking, 79 dip brazing of, 82–83 embrittlement, 79 filler metals for, 42 flashing of filler metals, 259 fluxes for, 80, 278 furnace brazing of, 81, 265 induction brazing of, 82 precipitation-hardenable copper, 80 resistance brazing of, 82, 83
© 2003 ASM International. All Rights Reserved. Brazing (#06955G) 390 / Brazing, Second Edition
Copper brazing filler metals, 192(T) Copper coatings, 228 Copper plating, 82, 141, 150, 304, 330, 360 Copper to copper brazing, 190 Copper to copper joints, 216 Copper to graphite bonds, 164 Copper to mild steel brazing, 190 Corrosion. See also stress-corrosion cracking (SCC) chloride stress-corrosion cracking, 235 corrosion resistance, 17, 188 crevice corrosion, 70, 114 dissimilar material combinations, 144–145 galvanic corrosion, 115, 183 and halide compounds in fluxes, 276 interface corrosion, 107, 108 intergranular corrosion, 106–107 liquid metal corrosion (LMC), 235 phosphides, 87 stress-corrosion cracking (SCC), 69–70, 89 Corrosion in certain metals cobalt-base alloys, 89 ferritic nonhardenable stainless steels, 107 nickel and nickel alloys, 69–70, 87 titanium and titanium alloys, 115 Costs and cost savings, 179 energy consumption, 27 of furnace brazing, 23 gas expenses, 265 of induction heating, 82 and quality assurance, 341–342 relative cost of furnace brazing, 23 from scrubber application, 275 Crack nucleation, 4 Cracks and cracking, 79, 329, 341, 369 Creep stress, 328 Crevice corrosion, 70, 114 Crimping, 321 Critical applications and components. See also quality assurance classification of, 339–342 Cryogenic nitrogen, 261 Cu3Sn intermetallic compounds, 56 Cubic boron nitride, 157–158 Culinary applications, 113, 178 CuproBraze process, 190 Curie temperature, 112 Cyaniding (fluxing) salts, 105 Cyclic loadings, 343 Cyropump (vacuum furnace), 252
D Decarburization, testing for, 309–310 Defects. See also discoloration; distortion; embrittlement; hydrogen embrittlement; inclusions; porosity; stresscorrosion cracking; voids acceptance limits for, 340, 341, 342 cracks and cracking, 79, 329, 341, 369 detection of, 345 dross, 79, 276
www.asminternational.org
filler metals, excessive flow of, 363 imperfections in brazed joints, 342 macrovoids, 324, 325 types of, 314, 340, 341, 345 Deformation, 328 Degreasing methods, 298 Deoxidizers, 266 Department of Energy (DOE), 366 Design and development control, 341–342 Design and quality system, 339–340 Design factors, 340 Design for assembly, 319–322 Design of joints. See joint design Design testing, evaluation and feedback, 344–345 Designing effective fixtures, 292 Destructive inspection and testing methods, 345–346 Dewpoint atmosphere chart, 258 described, 259–260 in vacuum furnaces, 254, 256, 284 Dewpoint control measurement, 282, 309 for oxidation reduction/control, 107, 227, 262, 264, 331 Dewpoint (oxygen content), 258 Dezincification of naval brass, 160 Diamond/tungsten carbide braze interlayer thickness, 57–58 fillets for, 57 microwave brazing, 57 Diamonds brazing, 12–13, 134–135, 367 carbide formation, 135 Differential thermal analysis, 219 Diffusion, 17, 265 Diffusion annealing, 66 Diffusion bonds and bonding, 74, 138(F), 154–155, 188 Diffusion-brazed joints, 74(F) Diffusion brazing, 54. See also transient liquid phase (TLP) bonding advantages and disadvantages of, 358–359 alloy system requirements, 55 approaches to, 357 for boron-aluminum brazing, 360 compression loading, 56 defined, 377 dissimilar material combinations, 358 equipment and tooling for, 360 equipment for, 360 eutectic compositions, 357 filler metals, 140 materials joined, 360 of NiAl to nickel, 149 other factors: time, pressure, metallurgical, equipment/ tooling, 360 peritectic compositions, 357 schematic of steps, 56(F) surface preparation, 359–360 uses for, 357 Diffusion of boron in ductile foils, 368–369 Diffusion sink brazing, 99
© 2003 ASM International. All Rights Reserved. Brazing (#06955G)
Dilution of fluxes, 272 Dip brazing, 46–50, 270, 318–319 Dip brazing of base metals aluminum and aluminum alloys, 73, 303, 318 beryllium, 78 cast iron, 100 copper and copper alloys, 82–83 low carbon steels, 105 magnesium and magnesium alloys, 86–87 nickel and nickel alloys, 88 Dipping, 271 Direct electric-resistance heating, 44(F) Direct eutectic resistance brazing, 43 Direct resistance brazing, 44, 44(F), 45 Discoloration black boron nitride, 263 black residue, 282 blue-gray nitrogen film, 248, 263 green chromium oxide, 283 of stainless steel parts, 60 Dispersion hardened composites, 54 Dispersion hardening, 216 Dispersion-strengthened alloys, 77, 89 Dissimilar metals/materials, 5, 70 and bonding failure, 328 and braze failure, 328 carbon or graphite, 140 corrosion in, 144–145 and deformation, 328 diametral clearance nomograph, 71(F) diffusion bonding, 147 graphite, composites, diamonds, aluminides to metals, 147–150 interfacial energy, 9 intermediate-temperature joining, 195 joining of, 371 joining technologies comparison, 376(F) joint clearances for, 316 metal-to-metal with coatings, 145–147 precoatings for, 300 pressure brazing, 147 procedures for, 41, 42 and residual stresses, 328 schematic of, 301(F) strain cracking, 329 strength, 328–329 thermal expansion, differences in, 144 thermal stresses of, 328 Dissociated ammonia, 263, 266 Dissociation, 265 Dissolution-solidification coatings, 180, 181 Distortion, 5 from fixturing, 289, 292, 293 gas metal arc brazing, 357 strength, 328 Distribution coefficient for alloy solute, 179 Documentation of destructive inspection and testing methods, 345 of product processing, 340, 342 as to quality assurance, 341, 342
www.asminternational.org Subject Index / 391
Double-wall retort, 28 Dresser-Rand, 355 Dross, 79, 276 Ductile (cast) iron, 101 Ductility, 201 Ductility and fracture toughness, 17 Dupree equation, 11 Dye and fluorescent penetrant inspections, 344
E Ebraze welding, 355 Economics of braze processing automation, 58–59 Economy, 179 Effects of brazing variables on clearance, 322–323, 323 Electric resistances, diagram of, 45(F) Electrically heated batch box-type furnace brazing, 25(F) Electro-chemical cleaning of cast iron, 304 Electrochemical metallizing, 179 Electrodes, 41–42, 43, 49 Electrodynamic circulation, 48, 49(F) Electroformed seals, 330(F) Electroforming, 329–330 Electrolytic coatings, 180 Electrolytic treatment, 101 Electrolytic vs. electroless nickel plating, 233 Electron beam brazing, 354–355 Electron beam welding (EBM), 354, 355 Electron brazing, 354 Electronic circuit packaging, 143–144 Electroplating, 334 Elemental boron powder, 267 Embrittled nickel-metalloid compounds, 215 Embrittlement. See also hydrogen embrittlement copper and copper alloys, 65, 79 filler metals found to cause, 64 hydride formation, 65 interaction of alloying elements, 65 phosphorus, 15, 67 in steel, 65 subsurface, 180 sulfur, 15, 166 by sulfur and low-melting metals, 87 tough pitch coppers, 79 by various gases, 243 Endothermic reaction, 219 Energy consumption, 27 Environmental developments, 374 Environmental Protection Agency (EPA), 348 Equilibrium curves for metal/metal oxides, 232(F), 258, 258(F) Equilibrium phase diagram, 9, 9(F) Equipment. See also robotics; tests/testing for atmospheric control, 245, 256 for automated brazing, 349 for automated induction brazing, 350 for braze welding, 53–54 for brazing control systems, 365, 375–376 for diffusion brazing, 360
© 2003 ASM International. All Rights Reserved. Brazing (#06955G) 392 / Brazing, Second Edition
Equipment (continued) for electron brazing, 354 for exothermic brazing, 51 for flux dispensing, 317–318 for fluxless brazing, 351 for furnace brazing, 23–24 for induction brazing, 32, 34–36, 35(F), 37, 132 for infrared brazing, 50 for laser brazing, 353, 354 for molten chemical (flux) bathdip brazing, 47–48 for plasma arc brazing, 359 for resistance brazing, 41, 42, 83 for salt bath furnace brazing, 47 for torch brazing, 22, 102, 182 for vacuum furnace brazing, 31, 264 Ergodynamics, 223 Eutectic-bonding approach, 75 Eutectic brazing, 358 Eutectic compositions in diffusion brazing, 357 Eutectic transformation, 181 Eutectic-type filler metals, 2 Evaluation of structural defects, 369–371 Evaporation removal of oxide film, 265 EWI, 355 Examples of commercially available rapidly solidified filler metals, 214(T) Exhaust overload problem, 308 Exothermic atmosphere, 256 Exothermic brazing, 51–52 Exothermic reaction, 219 Expanding, 321 Explosive mixtures, 256
F Fatigue testing, 345 Faying surfaces, overlap of, 1 Ferritic nonhardenable stainless steels, 107 Filler-metal cloths, 229 Filler metal flow, 15, 18 Filler-metal forms clad and coat, 229 filler pastes and dispensers, 223–226 foils and sheets, 229–231 plating, 227–229 preforms, 220–222 transfer tapes, 226–227 wire forms, 222–223 Filler metal paste, 223(F), 224(F) Filler metal placement, rules of, 17 Filler-metal rings, 43 Filler metal systems, 78, 129(T) Filler metal types, 181–211 Filler metals. See also specific filler metals (e.g., tin) for alumina (Al2O3), 128, 129, 130, 195 for alumina dispersion-hardened copper (ADHC), 82 for alumina nitride (AIN), 128 for aluminum and aluminum alloys, 72 for beryllium and beryllium alloys, 79 for boron, 157–158
www.asminternational.org
for braze welding, 53, 54 for carbon and low-alloy steels, 103(T) for carbon and low alloy steels, 105 for carbon and low-alloy steels, 105 for cemented carbides, 206 for ceramic-matrix composite (CMC) materials, 160–161 for ceramic-metal brazing, 196 for ceramic-to-ceramic brazing, 205 for ceramic-to-metal brazing, 205, 331 for ceramics brazing, 150–159, 153(T), 371 for cobalt-base alloys, 89 for contact-reactive brazing, 210 for copper and copper alloys, 42 for culinary applications, 113, 178 eutectic-type and crack nucleation, 4 for fluidity, 179 for Hastelloy X to Inconel 718, 352 for magnesium and magnesium alloys, 86–87 melting characteristics of titanium-copper-nickel filler metals, 210(T) for molybdenum and molybdenum alloys, 96 for naval brass, 160 for nickel and nickel alloys, 93(T) for niobium and niobium alloys, 99 for nodular cast iron to ZrO2, 334 for oxide-dispersion-strengthened (ODS) alloys, 91, 369 for plasma arc brazing, 357, 358(T) for refractory metals, 201, 206(T) for semiconductor devices, 210–211 for silicon nitride (Si3N4), 121, 122, 123 for stainless-steel ceramic joints, 153(T) for stainless steels, 108–111, 113–114 for tantalum and tantalum alloys, 99 for TiC strengthened alumina (ATC), 132(F) for titanium and titanium alloys, 114, 115, 116, 163 toxics of, 189 for transient liquid phase (TLP) bonding, 369 for tungsten and tungsten alloys, 100 for tungsten carbide, 153(T) for wide-gap brazing, 375 for zirconium and zirconium alloys, 117, 133–134, 334 Filler metals and coatings, 330–331 Filler metals creep, 296 Fillets, 4, 17, 57, 315, 341 Finite element method, 355, 372 Fire safety standards, 247 Fixturing and tooling baskets, 291(F) for brazing of copper tubing, 294(F), 294–295 clamshell-type, 39, 40(F) considerations in, 291–292 for dip brazing, 291(F) fixture design challenge, 290 improvements in, 292(F) in-place tooling, 39 materials for, 289, 291, 292, 295 nature of, 289
© 2003 ASM International. All Rights Reserved. Brazing (#06955G)
open-coil tooling, 39 for parts alignment, 292 for plate brazing, 290(F) pliers-type, 39–40, 40(F) self-fixturing methods, 321(F) thermal expansion problems, 293–294 Fixturing design, 319–322 Flame brazing, 246 Flame characteristics, 22–23 Flame cleaning, 304 Flow of filler metals, 15 Fluidity, 178–179 Fluoborates, 267 Fluoride formulations, 277 Fluoride-ion cleaning, 303, 303(F), 306, 308(F) Fluorides, 267 Fluorocarbon atmosphere, 302 Fluosilicaborates, 267 Flux dispensing, 317–318 Flux/fluxes and air displacement, 313 application quantity, 275 applications, 271–273 AWS grades of, 283 characteristics of, 268–269(T) constituents, 266–270 for copper and copper alloys, 80 corrosion and halide compounds, 276 described, 266–281 entrapment, 341 exhaustion, 271 filler metals and base metal compatibility, 275–277 flux islands, 272 functions of, 12, 266, 348 in furnace brazing, 24–25 for molten chemical (flux) bathdip brazing, 49 of oxides during brazing, 297 selection criteria, 270–271 situations when not required, 273 and specific processes, 273–275 for surface cleaning, 18 temperature range, 270–271 type FB series, 271–273, 275–276 water soluble, 271 working temperature range of, 271 Flux removers and removal, 278–280 Flux residues, 275, 279, 280, 280(T) Fluxing of large flat surfaces, 272 Fluxless brazing, 183, 254, 351–352 FM Technologies Incorporated, 352 Food industry applications, 113, 178 Formation of intermetallic phases, 178 Free energy of the reaction, 8 Freezing range of filler metals, 15 Friction seam brazing, 356(F) Fuel flame, 23 Fuel gases, torch brazing, 22 Fuel-tube assembly, 224(F) Furnace-atmosphere conclusions drawn from T-specimens, 251(T)
www.asminternational.org Subject Index / 393
Furnace atmospheres, 264(T) Furnace brazing, 23–32 atmospheres, 24–25, 32 batch furnace, 25 continuous-type, 25–26 control systems, 365(F) described, 24 electrically heated batch box-type, 25(F) equipment for, 23–24 fluxes in, 24–25 furnaces types, 24 hump mesh-belt furnace, 25–26 joint clearances for, 319 relative cost of, 23 retort-bell combustion furnace, 25 semi-continuous controlled atmosphere type (CAB), 26 throughput of, 24 in vacuum, 28 Furnace brazing of base metals aluminum and aluminum alloys, 244 beryllium, 78 brasses, 78 carbides and cermets, 118 cast iron, 101, 102, 159 copper and copper alloys, 81, 265 low carbon steels, 102, 104 magnesium and magnesium alloys, 86 nickel and nickel alloys, 88 P/M materials, 139 stainless steels, 106–108, 111–113 steel, 265 titanium and titanium alloys, 116 tungsten, 100 zirconium and zirconium alloys, 117 Furnace joining, 244–245 Furnaces, internally heated, 48(F) Furnaces types, 24 Fused borax, 267 Future outlook, 373–376
G Galvanic corrosion, 115, 183 Galvanic couplers, 70 Gamma aluminides, 138 Gas/flux mixture, 272 Gas metal-arc brazing, 356–357 Geometrical considerations for joint design and ceramics, 331, 333 Getters/gettering of contaminants, 115 defined, 253 effect of magnesium alloys, 77 magnesium gas as, 248 Glass-ceramic-bonded/metal seal, 120(F) Glass-ceramic joining, 119–120 Glass-ceramic/metal seal, 120(F) Gold-base brazing filler metal alloys, 196(T) Gold-copper, 194–195 Gold-nickel-palladium, 194–195
© 2003 ASM International. All Rights Reserved. Brazing (#06955G) 394 / Brazing, Second Edition
Gold plating, 155 Golf clubs, alloy usage, 373 Graded seal assembly, 326(F) Graded series of plates, 326(F) Grain boundary diffusion, 197 Graphite brazing coefficient of thermal expansion (CTE), 139 direct to metal, 147–148 dissimilar material combinations, 147–150 filler metals, 140–141, 157–158, 159 joint shear strength, 148(T) shear cracking, 140 Graphite fibers, radial, 149(T) Graphite fixturing, 295 Gravity locating, 320 Gray (cast) iron, 101 Green chromium oxide, 283
H Hafnium, 204 Hand-fed filler metal, 222 Hand held applicator for filler metal paste, 223(F), 224(F) Hardenability of base-metals, 15 residual stresses, 15 Hardness and machinability of filler metal, 17 Health and safety. See also cadmium; toxics beryllium in flux residues, 275 critical applications and components, 339–342 dermatitis from flux handling, 271 explosive mixtures, 256 fluoride fumes, 267 NFPA standards, 247 OSHA, 187, 348 toxic metals, 256 Heat-affected zone (HAZ), 5, 45(F), 65 Heat-resistant alloys, 89 Heat treatment of base-metals, 16 Heating methods, 21, 21(T) Helium (He) atmosphere for furnace brazing, 33(T) in brazing certain metals, 95, 117, 266, 284 carbon pickup, 310 as inert gas, 244, 265 pressure testing, 113, 264 Hexoloy-SA (silicon-carbide), 126 High pressure leak testing, 342 High speed steels, 106 High-temperature brazing, 323 High-temperature fluxes, 270 Holography, 344 Honeycombs brazing of, 377 controlled atmosphere brazing, 243 exothermic brazing, 52 filler metals, 67, 89, 115, 200 filler metals chart, 198(T), 250(T) infrared brazing, 50
www.asminternational.org
inspection of, 344 nickel-base filler metals, 197 quenching, 115 RS materials, 222, 367 space transportation vehicle, 373, 375(F) transfer tapes, 226 Hot rodding, 272 Hump mesh-belt furnace, 25–26 Hydrides for ceramic bonding, 330 effects of, 65 Hydrogen atmospheres, 262–263 brazing refractory metals in, 284 graphite fixtures in, 295 hydrogen/air ratio, 259 hydrogen reduction process, 262 metal/metal oxide equilibrium curves, 232(F), 258(F) reduction of oxides, 260 Hydrogen embrittlement, 15, 32, 63–65, 82, 264
I Impact tests, 159, 216, 218, 346 Improving filler metal flow, 272 In situ reaction joining of fiber-reinforced SiC composites, 366–367 Inclusions base metal, 108 contaminants, 297 flux, 322, 345 hydrogen, 65 oxides, 30 Inconel X (precoats), 296 Induction brazing advantages of, 32, 34–36 assembly of cast iron and steel components, 35(F) atmosphere brazing system (ABS), 245 atmospheres, 36 automatic temperature control, 41 boron-modified fluxes for, 275 cost of, 82 equipment for, 32, 34–36, 35(F), 37, 132 fixturing, 35–36 flux free brazing, 39 fluxes, 36 hand held, 37 joint clearances, 34, 319 localized heating, 104 sandwich filler strip in, 37 of steel base to cast iron nose, 34(F) tube-in-place, 37–38 type FB3B, 275 types of joints, 34 Induction brazing of base metals beryllium, 78, 275 copper and copper alloys, 82 low carbon steels, 145, 188 of stainless steel, 60–61 stainless steels, 60, 100, 188, 301 titanium and titanium alloys, 116, 219
© 2003 ASM International. All Rights Reserved. Brazing (#06955G)
tungsten, 100 zirconium and zirconium alloys, 116 Inert gas atmosphere, 207, 245, 264, 284. See also argon (Ar); helium (He) Infiltration, 323 Infrared brazing automated brazing, 58, 347 for certain metals, 72, 352–353 equipment for, 50 microstructure changes by, 51 rapid infrared joining (RIJ), 136–137, 138 uses of, 50–51 using argon (Ar), 352 Infrared-sensitive electronic imaging, 27(F), 52(F), 53, 344 Inorganic vapors, 33(T), 257(T), 262 Inspectable factors, 340 Inspection of brazements, 339. See also quality assurance; tests/testing Interface corrosion, 107, 108 Interface Seeding in microwave brazing, 56 Interface voids, 325 Interfacial amorphous phase in silicon-carbide bonding, 126 Interfacial bonding of silicon nitride (Si3N4), 154 Interfacial compound formation, 10 Interfacial energy, 9, 130, 207, 266 Interfacial glassy phases, in silicon-carbide bonding, 126 Interference or press fitting, 320 Intergranular corrosion, 106–107 Intergranular penetration of phosphorus, 180 Interlayer brazes, 112 Interlayers, 124, 205–206, 326–327 Intermediate layers, 326 Intermetallic compounds, 67 Intermetallic phases, 2 Interstitial voids, 325 Ion bombardment, 301–302 Ion plating, 141, 155, 206 Ion scattering spectometry (ISS), 280 Iron-base Fe-Cr-Al ODS alloys, 93 Isothermal solidification, 306, 323, 369
J Joining atmospheres, 244–255 Joining of abrasive tool materials, 367 Joining with metallic amorphous glass foils, 367–368 Joint brittleness, 216 Joint clearances, 261 atmospheric influence of, 317 for BCuP filler metals, 193 for BNi-2 filler metals, 336 dip brazing of aluminum, 318–319 for dissimilar metals, 316 excessive, 329 for furnace brazing, 319 for high butt-joint strength, 314 induction brazing, 34, 319 and joint length, 328
www.asminternational.org Subject Index / 395
large, 323 ranges of, 18 tensile strength of, 317(F) tensile strength vs., 317(T) for torch brazing, 319 values for, 315–316 wide-gap joints, 263 Joint design. See also seals; specific joints (e.g., lap joints) for brazing filler-metal systems, 19(T) coefficient of thermal expansion (CTE), 19 factors influencing, 19 joint-filling capacity of filler metal, 17 joint length and configuration, 328 joint redesigns, 316(F) joint strength, 16, 226, 229 joint thickness, limits of, 1 requirements for, 15, 348 for tube-in-place induction brazing, 38(F) for various types of seals seals, 332(F) Joint design and ceramics, 329–336 Joints, 12 assembly recommendations, 318–319(F) diagrams of, 45, 45(F) irregular fixed by diffusion brazing, 54 open-coil setup and typical joints, 39(F) types of, 20, 313–315, 314(F)
K Knobby whisker morphologies, 149 Knurling, 320–321, 321(F), 322(F)
L Laminated foils, 210 Laminated interlayers of ceramics, 11 Lap joints clearances, 200 described, 20, 313–316, 319 diagram of, 20(F), 315(F) overlap for, 20 shear strength of, 184 testing, 345 Laser brazing, 52, 52(F), 53(F), 353–354 Leaded naval brass, 80–81 Leaded steels, 102 Leak testing, 342–343 Lepel Corporation, 349 Levigated alumina, 295 Liquation, 177, 179 Liquid fluxes, 272 Liquid-infiltrated powder interlayer bonding (LIPB), 323, 324(F) Liquid interface diffusion, 181 Liquid metal corrosion (LMC), 235 Liquid metal embrittlement, 96–97 Liquid metal infiltration, 323 Liquidus. See also specific filler metals; specific metals alloying, 67
© 2003 ASM International. All Rights Reserved. Brazing (#06955G) 396 / Brazing, Second Edition
Liquidus (continued) defined, 1 depressants, 187 filler metals selection, 219 flux selection, 270 melting and fluidity, 178–179 minimum temperature of, 7 Lithium oxide scavenger, 262, 265, 266 as wetting agent, 78, 204 Local overheating, 42, 43 Los Alamos National Laboratory, 352 Low pressure leak testing, 342
M M6X compounds, 131 Machining swarf, 356 Macrovoids, 324, 325 Magnesium and magnesium alloys cleaning methods for, 304 dip brazing of, 86–87 filler metals for, 86–87 furnace brazing, 86 gettering effect of, 77 torch brazing of, 86 Magnesium evolution, 30 Magnesium fluxes, 270 Magnesium gas, 248 Magnesium oxides, 295 Magnetohyrodynamic circulation, 49, 49(F) Malleable (cast) iron, 101 Maraging steels, 112–113 Marine diesel engine components, 333 Martensitic hardenable stainless steels, 107 Matrix-metal composites, 53 Maxwell’s law of electromotive forces, 48 Mechanical cleaning, 18, 298, 299(T), 300. See also cleaning Mechanical inspection, 345–346 Mechanization of torch brazing, 23 Medical applications, 178 Melting and melting point, 177–179, 197, 323 Meltspun foils, 210 Membrane system, 246 Metal-matrix composites, 136, 137 Metal/metal oxide equilibrium curves, 232(F), 258, 258(F) Metal thickness, 3 Metal-to-ceramics brazing of, 195, 365 joints, 365 seals, 335 systems, 10, 151 Metallic bonds, 2, 120, 178 Metallization. See coatings and metallizing Metallographic inspection, 345–346 Metallurgical considerations, 348 Metallurgical phenomena, 15 Metallurgical reactions, 63–70
www.asminternational.org
Methane (CH4), 261 Methanol, 261 Micrographs, 247(F) Microstructure changes by infrared (quartz) brazing, 51 Microstructure effects of brazing, 5 Microstructure of furnace-brazed joint, 29(F) Microwave brazing, 352 Al3O3 ceramic, 57 ceramic materials, 56–57 of diamond/tungsten carbide, 57, 58(F) interface Seeding in, 56 tungsten carbide and polycrystalline diamonds, 57 Milk of magnesia, 297 Mineral and gas-phase fluxes, 322 Mismatch stresses, 326 Modeling behavior of brazing processes and materials, 365–366 Molten chemical (flux) bathdip brazing, 46–50 Molybdenum and molybdenum alloys, 95–97, 275, 284, 305 Molybdenum-manganese process, 144, 331, 334 Mulitmet, 248 Mullite, 127–128 Multiaxis seam tracking (MAST), 224, 225(F), 350 Multicomponent brasses, 189 Multilayer interlayer, 132 Multiple coatings, 180 Mutual solubility, 376
N National Aeronautics and Space Administration (NASA), 355 National Fire Protection Association (NFPA), 247 Naval brass, 80–81, 160 Neodymium-yttrium-aluminum-garnet laser, 353, 354 Nernst-Bruenner theory, 93 Neutral flame, 23 Neutral or reducing flame, 255 Neutral salts, 105 Nicalon, 136, 366 Nichrome (precoats), 296 Nickel and nickel alloys brazing processes for, 88 dip brazing of, 88 dispersion-strengthened, 89 embrittlement, 87 filler metals for, 93(T) furnace brazing of, 88 melting point of, 197 precipitation-strengthened, 88 precleaning, 305 resistance brazing of, 88 stress-corrosion cracking (SCC), 69–70, 87 sulfur embrittlement, 66 torch brazing of, 88 Nickel and TD-NiCr alloys, 93(T) Nickel-base amorphous filler metals, 217(T)
© 2003 ASM International. All Rights Reserved. Brazing (#06955G)
Nickel-base brazing filler metals, 198(T) Nickel plating, 88, 104, 108, 118, 196, 227, 228, 228(T), 233–234, 249, 262, 301–302 Nickel silvers, 81 Niobium and niobium alloys, 97–99, 264, 284, 305 Nitriding, 249, 261 Nitrogen atmosphere, 263 brazing in, 123, 274, 365 described, 261–262 filler metals for, 375 fluxless brazing, 183 and titanium, 161 Nitrogen buildup, 262 Nitrogen in base metal, 248, 249 Nitrogen pickup, 261 Nitronic 60, 228 Noble metals, 180, 180(F), 181, 201, 211 Nocolok process, 73, 273–275 Nodular cast iron to ZrO2, 334 Nominal compositions of materials used in a wide-gap brazing study, 325(F) Noncadmium alloys for carbide brazing, 191(T) Noncontinuous fillets, 341 Noncorrosive flux braze process (Nocolok), 273–275 Nondestructive testing. See also tests/testing acceptance limits, 340, 341, 342 of critical applications, 339, 342 described, 342–344 eddy-current testing, 369 equipment for, 370 flaw detection limits, 370 low pressure leak testing, 342 pressure testing, 113, 264, 342–343 reference standards for, 344 for tube-in-place induction brazing-induction brazing, 41 Nonwetting, 8(F) Novel and emerging brazing processes, 352–373
O Occupational Safety and Health Administration (OSHA), 187, 348 Optical pyrometers, 350 Oxide ceramics, 10 Oxide-dispersion-strengthened (ODS) alloys, 89, 91, 93, 369 Oxide film layers, 12, 253, 265 Oxide reduction, 259 Oxide scavengers, 262, 265, 266 Oxide stability and formation, 66 Oxides of aluminum, 184, 263, 303 equilibrium curves for metal/metal oxides, 232(F), 258, 258(F) formation of, 243 hydrogen reduction of, 260 inclusions, 30 of magnesium, 295 reduction of, 260
www.asminternational.org Subject Index / 397
refractory, 88, 204, 295 of titanium, 263 of zirconium, 296, 297 Oxidizing atmospheres, 245 Oxidizing flame, 23 Oxidizing gas gettering, 30 Oxyacetylene flame, 22(F) Oxygen (O2), 261 Oxygen partial pressure, 245, 249
P Parallel-wire reinforcement, 222 Part flow, maintaining, 298 Partially stabilized tetragonal zirconia (PSTZ), 152 Partially stabilized zirconia (PSZ), 150 Particle size, 272 Passive coatings, 227 Pastes brazement gaps, 221 for brittle filler metals, 214 dispensers for, 223 filler metals in form of, 197, 275 fluxes in form of, 271 inventory of, 225 Peel tests, 345 Peening, 322 Pellets, 293 Peritectic compositions, in diffusion brazing, 357 Phosphate acid cleaners, 305 Phosphides, 87 Phosphor bronze, 81 Phosphorus, 180, 193 Phosphorus embrittlement, 15, 67 Physiochemical coatings, 180 Pickling (chemical etching). See chemical cleaning Planar-magnetron-sputtered (PMS) silver interlayers, 212, 213 Plasma arc brazing (braze welding), 54, 357, 359 Plasma arc welding, 357 Plasma spraying, 352 Plating. See also copper plating; nickel plating gold, 155 ion, 141, 155, 206 silver, 227 Plating thickness, 88, 143, 160, 228, 234, 249, 262 Polyvalent elements, 180 Porosity. See also voids described, 341 in graphite, 139 in PM parts, 306 in powdered metal brazements, 306 quality assurance, 252 testing, 342, 344, 345 Postbraze cleaning and flux removal, 278–281 Postbrazing treatments, 67–70, 308 Powder filler metals, 99, 185, 317 Powder metals (P/M), 76–77, 139, 306–307 Powdered fluxes, 272 Precipitated carbides, 64
© 2003 ASM International. All Rights Reserved. Brazing (#06955G) 398 / Brazing, Second Edition
Precipitation-hardening of aluminum and aluminum alloys, 183 Precipitation-hardening stainless steels, 107–108 Precleaning, 297–302 Precoatings, 296, 300–301 Preform placement, 221(F) Preplacement of filler metals, 320(F) Press fitting, 320 Pressure-brazed joints, 66 Pressure oil can application, 271 Pressure swing adsorption (PSA) system, 246 Pressure testing, 342–343 Pretinning of difficult to braze metals, 46 Problem solving tips application of filler metal for continuous brazing, 59, 60 atmosphere control, 60 atmosphere control for brazing stainless steel and copper, 282 braze repair of honeycomb structures, 377 brazing aluminum bronze to naval brass, 159–160, 283–284 brazing cast iron to dissimilar metals, 159 brazing nickel alloy strip to copper-aluminum bar, 162 brazing of 17–7 PH vs. 17–4 PH, 232–233 brazing of 409 stainless steel, 60 brazing of a ceramic-matrix composite, 161–162 brazing of copper-graphite assemblies, 163–164 brazing of leaded brass, 60 brazing of refractory metals, 284 brazing of stainless steel, 60 brazing of stainless steel to titanium, 160–161 brazing of tungsten carbide, 281 brazing parts exposed to synthetic machine oils, 59 brazing tungsten carbide to carbon steel wheels, 12–13 brazing with nickel-base filler metals, 283 choosing right filler metals, 336 clearances and applying filler metals, 336 closed vs. open caps, 233 copper brazing of stainless steel inserts, 161 diffusion brazing for jet engine repair, 377 filler metal remelt, 231–232 flow of filler metal around tack welds, 310 furnace brazing of leaded heat exchanger to a copper heat exchanger shell, 59 furnace preparation for brazing René 77, 282–283 high-frequency induction brazing of stainless steel, 60 increasing filler metals remelt temperature, 376–377 leaking in a copper and brass valve assembly, 281 magnesium buildup in a a vacuum furnace used to braze aluminum, 60 partial brazing checklist to secure good braze quality, 336 removing synthetic oils to prevent rusting after brazing of stainless steels, 309 residue on steel parts after brazing, 282 surface preparation of 304L stainless steel plates, 308–309 torch brazing of 304L tube to a fitting joint, 59 troubleshooting brazement quality of copper-brazed 1018 steel, 309
www.asminternational.org
troubleshooting problems associated with tack weld fixturing, 310 use of graphite fixturing in a hydrogen atmosphere, 309–310 vacuum brazing of nickel alloy tubes to titanium alloy fittings, 162–163 Process control factors, 340 Process operations, 49–50 Processes (heat sources), 21 Product layer, 181 Programmable logic controllers (PLC), 350, 363–365, 365(F) Promoters (activating metals), 265, 266(T) Proof testing, 343 PTR-Precision Technologies Incorporated, 355 Purifire (commercial brazing system), 246
Q Quality assurance, 220, 222. See also defects; nondestructive testing; porosity; tests/testing batch vacuum furnace problem, 246 braze quality-control maps, 325 burnishing and oxide embedment, 300 carbon pickup, 309 and costs, 341–342 for critical applications, 340 destructive inspection and testing methods, 345–346 dewpoint control, 282, 309 documentation, 340, 342 duty assignments for, 340 factors for, 340 overexposure in chemical cleaning, 305 part flow, 298 product tracking, 349 quality standards for brazing and brazing processes, 340–341 requirements for brazements, 339 residual flux, 281 resistance heating, 45 surface contaminants, 298 tack welds cracking, 310 testing of alumina-ceramic brazing, 365 unbrazed void areas, 309
R Radio frequency induction furnace, 163 Radio frequency inductor power, 218 Radiographic inspection testing, 343–344 Rapid infrared joining (RIJ), 136, 137 Rapid infrared processing technique, 50–51, 208–209, 352–353 Rapid-solidification powder metallurgy, 76–77 Rapidly solidified filler metals, 194, 214–216, 218, 367 Reaction brazing, 10 Reaction layer technique, 130 Reaction products for metal-ceramic systems, 10 Reaction rates, 45, 266 Reaction wetting, 10 Reactive-metal coatings, 180–181
© 2003 ASM International. All Rights Reserved. Brazing (#06955G)
Reactive metals, 181, 196 Recommended Practices for the Design, Manufacture, and Examination of Critical Brazed Joints (American Welding Society), 339, 340 Recrystallization, 5 Recrystallization temperature ranges, 94–95 Recycling, 374 Red brasses, 80 Reducing atmospheres, 245–249 Reducing flame, 23 Reference standards for nondestructive testing, 344 Refractory metals, 94–100, 305 carbides of, 117 defined, 94 embrittlement of, 95 recrystallization temperature ranges for, 94–95 transition-temperature ranges for, 94 wetting agents in filler metals for, 201 Refractory oxides, 88, 204, 295 Relative cost of furnace brazing, 23 Remelt temperature of filler metals, 231 Repair techniques with cleaning agents, 308 Residual flux, 280, 281 Residual gas analyzer, 60 Residual stresses, 154, 328, 331 Resistance brazing advantages and limits of, 41 applicability of, 41 dissimilar metals/materials, 41 equipment for, 41, 42, 83 flux selection, 270 principle of, 135(F) tensile strength, 366 Resistance brazing of base metals aluminum and aluminum alloys, 72 copper and copper alloys, 82, 83 nickel and nickel alloys, 88 tungsten, 100 Resistance-reactive brazing, 84 Retort-bell combustion furnace, 25 Retorts, 28 Reworked assemblies, 345 Riveting and folding or interlocking, 322 Robotics evaluation of structural defects, 370–371 robotic inspection systems, 225(F), 370–371, 371(F) robotic system for brazing, 350–351, 351(F) RS amorphous materials, 221–222 RS conversion of filler metal powders, 221 RS filler metals, 194, 214–216, 218, 367 RS Metglas brazing foils, 214
S Safety hazards. See health and safety Salt bath furnace brazing, 47 Sandwich brazing, 70 Sandwich filler strip, 37 Sandwich seal joining, 335–336 Scanning acoustic microscopy (SAM), 370
www.asminternational.org Subject Index / 399
Scanning electron microscope/energy-dispersive spectroscopy (SEM/EDS), 216, 218 Scanning electron microscopy (SEM), 352 Scanning laser acoustic microscopy (SLAM), 370 Scanning photoacoustic microscopy (SPAM), 370 Scanning transmission electron microscope (STEM), 218 Scarf joints, 315 Seals ceramic to metal seals, 329(F) for electroformed seals, 330(F) joint designs for, 326(F), 332(F), 335(F), 335–336, 336(F) Self-brazing methods, 327 Self-fixturing methods, 321(F) Self jigging, 320 Self-locating assemblies, 348 Semi-continuous controlled atmosphere type (CAB), 26(F), 26–27, 27(F) Semi-continuous furnace, 30, 30(F) Semiconductor devices, 143, 210–211 Sessile drop configuration, 8(F) Sessile drop experiments, 132 Shear cracking, 140 Shear strength of brazement, 327 SIALON, 124 SiC/alumina brazing, 211 SiC brazements, 366 SiC-SiC joints, 125(F) SiC whiskers, 148–149 Silicate brazing, 209(F) Silicon bronzes, 81 Silicon carbide (SiC), 124–127, 126(F), 127(F), 152 Silicon nitride (Si3N4), 131(T), 152, 154–156, 155(F), 213 Silver-copper binary system, 9, 9(F) Silver fluxes, 270 Silver plating, 227 Slab discharge technology, 230 Slurries, 272 Soldering, 2, 2(T) Solid skull, 179 Solid solutions, 2 Solid-state exothermic chemical reaction, 51–52 Solidification shrinkage, 178 Solidus. See also specific filler metals; specific metals alloying, 67 brazing fundamentals, 7, 12 defined, 1 filler metals characteristics, 177 filler metals selection, 219 flux selection, 270 melting and fluidity, 178–179 temperature of, 7 Solute addition, 352 Solutions (chemical) for flux removal from aluminum parts, 280(T) Solvent cleaning for fluxless brazing, 303 Soot formation, 247, 363 Space transportation vehicle, 373, 375(F)
© 2003 ASM International. All Rights Reserved. Brazing (#06955G) 400 / Brazing, Second Edition
Space transportation vehicle (continued) brazing processes for, 374(F) Spinning, 321 Spot welding and tack welding, 293 Spraying, 271 Spreading, 9 Spreading pressure, 7 Squeeze brazing (SQ), 131 Squeeze casting, 146 Stainless steels applications of, 113–114 base-metal inclusions, 108 filler metals for, 108–114 furnace brazing of, 106–108, 111–113 induction brazing of, 60–61, 100, 188, 301 metallurgical considerations in brazing, 108 precipitated carbides in, 64 torch brazing of, 106, 112 types of, 107–108 wetting problems, 106 Staking, 321 Standard Methods for Evaluating the Strength of Brazed Joints in Shear, (American Welding Society), 345 Steels. See also ceramics; specific steels (e.g., stainless steels) atmospheres for brazing, 32 brazing processes for, 100 contact-reactive brazing of, 184 dip brazing of, 46 dissimilar material combinations, 144 furnace brazing of, 265 hydrogen embrittlement, 65 induction heating of, 82–84 resistance brazing of, 41–43 stress corrosion cracking (SCC), 68 Step brazing, 194, 335, 361(F), 361–362 Step-seam heating, 45 Steps in automating a brazing process, 348–349 Stopoff materials and parting agents, 18, 60, 295–297 Strain cracking, 329 Strength, 327–329 Stress concentrations, 178, 315, 316(F), 327 Stress-corrosion cracking (SCC), 68, 69–70, 87, 89, 235–236 Stress cracking, 67–70, 107 Stresses chloride stress-corrosion cracking, 235 creep stress, 328 mismatch stresses, 326 reducing thermal stress, 124 residual stresses, 15, 154, 328, 331 tensile and shear stress tests, 345 thermal stresses, 328, 372 types of, 313 Structure tests, 54 Subsurface embrittlement, 180 Sulfur embrittlement, 15, 66 Superalloys ceramics, brazing to, 154–155 cleaning techniques for, 302, 305, 308
www.asminternational.org
described, 89–94 diffusion brazing of, 357, 360 filler metals for, 199, 214(T) nickel plating for, 227 ODS alloys, 93 precoatings of, 300 single crystal, brazing of, 94(T) transient liquid insert metal (TLIM) diffusion bonding, 369 transient liquid phase (TLP) bonding, 150, 369 Superplastic forming/brazing process, 372–373, 373(F) Surface cleaning and preparation, 17–18, 297–307. See also cleaning methods Surface contaminants, 298 Surface finish, 322 Surface modification coatings, 180 Surface roughness and filler metal flow, 18 Surface tensions, 7 Swaging, 321
T Tack welding, 321 Tack welds, cracking, 310 Tantalum and tantalum alloys brazing of, 275, 284 cleaning methods for, 305 diffusion sink brazing, 99 powder filler metals for, 99 sensitive to hydrogen embrittlement, 264 Technical business interface, 342 Temperature. See also liquidus; solidus barrier (tape) curtains, 27 for brazing, 104 for brazing base materials, 177 for filler metals, 17, 179, 231 flux/fluxes. range of, 270–271 high-temperature brazing, 323 high-temperature fluxes, 270 induction brazing control of, 41 vs. joint strength, 122(F) recrystallization temperature ranges, 94–95 remelt temperatures, 94, 231 and time, effect on wetting, 20–21 torch brazing control of, 86, 270 transition-temperature ranges for refractory metals, 94 Tensile and shear stress tests, 345 Tensile strength, 327, 366 Tests/testing. See also nondestructive testing acoustic emission testing, 343 active filler metal (ABA) tests, 128–129 air pressure testing, 342–343 of alumina-ceramic brazing, 365 atmospheric testing, 60 for braze atmosphere, 249 critical applications, 339, 342 for decarburization, 309–310 design testing, evaluation and feedback, 344–345 destructive inspection and testing methods, 345–346 documentation of, 345 fatigue testing, 345
© 2003 ASM International. All Rights Reserved. Brazing (#06955G)
helium pressure testing, 343 impact testing, 216, 346 leak testing, 342–343 methods of, 340 for nickel plating, 301 other testing techniques, 344 peel tests, 345 pressure testing, 342–343 proof testing, 343 radiographic inspection, 343–344 reference standards, 344 structure tests, 54 tensile and shear stress tests, 345 tests for braze atmospheres, 249 of titanium-containing filler metals for brazing of alumina, 129(T) titanium filler metals test header, 128(F) torsion tests, 346 tube-in-place induction brazing-induction brazing, 41 ultrasonic inspection testing, 344 vacuum testing, 343 visual inspection testing, 342 Thermal coatings, 180 Thermal distortion. See distortion Thermal expansion, 70 ceramics behavior, 10 coefficients of, 370 curve vs. coefficients of, 328 curves for common metals, 68–69(F) differences in, 121 of graphite fixtures, 295 matching of, 121 mismatch of, 331 problems in, 293–294 Thermal stresses, 328, 372 Thermal-transfer inspection, 344 Thermal treatments, 300 Thermal vacuum spraying, 184, 185 Thin coatings, 272 Third-body friction brazing, 355–356, 356(F) Throughput of furnace brazing, 24 Ti3Al, 138 TiC (cermets), 117 TiC strengthened alumina (ATC), 132, 132(F) Time at brazing, 234 Tin in bronzes, 81 as cadmium substitute, 189 in soldering, 1–3 wetting agents, 207 TiN (as cermets), 117 Titanium and titanium alloys brazing atmosphere for, 286 brazing of, 213 brittle intermetallics, 114 contact-reactive brazing, 184 crevice corrosion, 114 diffusion bonding, 138(F), 147 filler metals for, 114, 115, 116, 163 furnace brazing of, 116 galvanic corrosion, 115
www.asminternational.org Subject Index / 401
induction brazing of, 116, 219 oxides of, 88, 263 sensitive to hydrogen embrittlement, 264 titanium active brazing, 133 titanium Al2O3 diffusion bonding, 133 torch brazing of, 111 transient liquid phase diffusion bonding (TLPDB), 148 Titanium dioxide, 295 Titanium foil, 115 Titanium-matrix composite (TMC), 136 Titanium oxide and hydrogen dewpoint, 331 Tool steels, 105–106. See also carbon and low-alloy steels Tooling. See fixturing and tooling Torch brazing described, 21–23 equipment for, 22, 23, 102, 182 flame characteristics, 22–23 fuel gases, 22 hydrogen embrittlement, 62 joint clearances for, 319 localized heating, 24 low equipment cost, 82 mechanization of, 23, 271 stress relieving, 69 torch tips, 22 torches, 22 workpiece clearance for, 319 Torch brazing of base metals brasses, 80 carbon and low alloy steels, 102, 103(T), 104 magnesium and magnesium alloys, 86 nickel and nickel alloys, 88 stainless steels, 106, 112 titanium and titanium alloys, 111 tungsten, 111 zirconium and zirconium alloys, 111 Torsion tests, 346 Tough pitch coppers, 79 Toxicity. See also cadmium; health and safety of carbon monoxide atmosphere, 32, 256, 261 in cleaning materials, 348, 359 of filler metals, 189 in flux residues, 275 in inorganic vapors, 262 Transfer tapes, 226, 226(F) Transient liquid insert metal (TLIM) diffusion bonding, 368–369 Transient liquid phase diffusion bonding (TLPDB), 137, 148 Transient liquid phase (TLP) bonding, 88, 136 defined, 181 in diffusion brazing, 357 filler metals for, 369 for metal-matrix composites (MMC) and ceramic substrates, 157 for metals and ceramics, 156 of NiAl to NiAl, 149–150 Transition-metal barrier, 180(F) Transition pieces, 334, 335 Trimetal sandwiches, 119
© 2003 ASM International. All Rights Reserved. Brazing (#06955G) 402 / Brazing, Second Edition
Tube-in-place induction brazing-induction brazing, 37–41 Tungsten and tungsten alloys, 99–100, 111, 305 Tungsten carbide brazing of, 12–13, 57, 189–190 Tungsten carbide (continued) filler metals for, 153(T) thermal expansion of, 37
U Ultrasonic inspection testing, 344 Uniaxial solid-state bonding techniques, 212
V Vacuum arc brazing, 356 Vacuum atmospheres, 249–255, 284 Vacuum-brazed ordnance projectiles, 293(F) Vacuum brazing advantages and disadvantages of, 264–265 applications of, 28 of ceramic-to-metal seals, 331 described, 264–265 laying parts together for, 292 mechanisms of, for aluminum alloys, 254(F) promoters (activating metals) in, 266(T) specialized fixtures for, 292–295 uses of, 264 vacuum compression system, 83(F) Vacuum brazing cleaning, 307–308 Vacuum furnace brazing, 290(F) backfill gas, 252 batch-type, 28(F) compared to atmospheric brazing, 363(T) controls, 30–31 cooling systems, 31–32 cycle diagram, 251(F) cycle/sequence of events, 29(F) Cyropump, 252 deep vacuum, 252 diagram of, 28(F) equipment for, 31, 264 operating conditions, 254 radiation shields, 30–31, 31(F) rapid gas-quenching type, 252 for refractory metals, 245 Vacuum-grade filler metals, 285(T) Vacuum heat treatment and cleaning methods, 359 Vacuum testing, 343 Vacuum-tube-grade filler metals, 66 Vacuum weld brazed joint, 55(F) van der Walls bonding, 8 Vapor-gas brazing, 189 Vapor-phase brazing, 228 Vapor pressure, 66 Viscosity, 15 Visual inspection testing, 342 Vitreous (glass) bonding, 120, 121 Voids. See also porosity interface voids, 325, 360 interstitial voids, 325 joint clearance, 19, 315, 316, 317(T)
www.asminternational.org
lack of fill, fillet defects, 341 surface contaminants, 297 surface finish, 322 testing for, 342–345
W Water, 270 Water vapor, 263–264 Welding Institute, 355 Welding process compared with soldering and brazing, 2(T) Wetting agents, 201, 204, 270 of ceramics, 207, 330 defined, 12 described, 11–12 and dewetting, 12(F), 13 effects of, 11–12 by electroplating of NCl with copper, 334 factors controlling, 11 interfacial energy, 207 nature in filler metals, 179 nature in fluxes, 266 reaction wetting, 10 sessile drop configurations, 8(F) of the surfaces, 177 Wetting problems of beryllium and beryllium alloys, 77 ceramics brazing, 121, 371 chromium fluoride atmosphere, 302 fluorocarbon atmosphere, 302 stainless steels, 106 Wide-Gap brazing, 263, 323–326, 375 Work of spreading, 8
X X-ray characterization, 343 X-ray computer tomography (CT), 370 XT3 (cermet grade), 117–118
Y Yellow brasses, 80 Yielding design, 326 Young’s equation, 7, 8 Yttria-stabilized zirconia (YSZ), 213–214 Yttrium-tetragonal zirconia polycrystal (Y-TZP), 186
Z Zinc fuming, 80 Zirconium and zirconium alloys brazing of, 284 described, 116–117 filler metals for, 117, 133–134, 334 furnace brazing of, 117 sensitive to hydrogen embrittlement, 264 torch brazing of, 111 Zr-to-Zr joints, 133–134 ZrO2 to nodular cast iron, 334 Zirconium oxide (ceramic stopoff materials), 296, 297
ASM International is the society for materials
engineers and scientists, a worldwide network dedicated to advancing industry, technology, and applications of metals and materials. ASM International, Materials Park, Ohio, USA www.asminternational.org This publication is copyright © ASM International®. All rights reserved. Publication title
Product code
Brazing
#06955G
To order products from ASM International: Online Visit www.asminternational.org/bookstore Telephone 1-800-336-5152 (US) or 1-440-338-5151 (Outside US) Fax 1-440-338-4634 Mail
Customer Service, ASM International 9639 Kinsman Rd, Materials Park, Ohio 44073-0002, USA
Email [email protected]
American Technical Publishers Ltd. 27-29 Knowl Piece, Wilbury Way, Hitchin Hertfordshire SG4 0SX, In Europe United Kingdom Telephone: 01462 437933 (account holders), 01462 431525 (credit card)
www.ameritech.co.uk Neutrino Inc. In Japan Takahashi Bldg., 44-3 Fuda 1-chome, Chofu-Shi, Tokyo 182 Japan Telephone: 81 (0) 424 84 5550 Terms of Use. This publication is being made available in PDF format as a benefit to members and customers of ASM International. You may download and print a copy of this publication for your personal use only. Other use and distribution is prohibited without the express written permission of ASM International. No warranties, express or implied, including, without limitation, warranties of merchantability or fitness for a particular purpose, are given in connection with this publication. Although this information is believed to be accurate by ASM, ASM cannot guarantee that favorable results will be obtained from the use of this publication alone. This publication is intended for use by persons having technical skill, at their sole discretion and risk. Since the conditions of product or material use are outside of ASM's control, ASM assumes no liability or obligation in connection with any use of this information. As with any material, evaluation of the material under end-use conditions prior to specification is essential. Therefore, specific testing under actual conditions is recommended. Nothing contained in this publication shall be construed as a grant of any right of manufacture, sale, use, or reproduction, in connection with any method, process, apparatus, product, composition, or system, whether or not covered by letters patent, copyright, or trademark, and nothing contained in this publication shall be construed as a defense against any alleged infringement of letters patent, copyright, or trademark, or as a defense against liability for such infringement.