Ultrasonic Instruments and Devices II Reference for Modem Instrumentation, Techniques, and Technology
PHYSICAL ACOUSTIC...
43 downloads
658 Views
26MB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
Ultrasonic Instruments and Devices II Reference for Modem Instrumentation, Techniques, and Technology
PHYSICAL ACOUSTICS Volume XXIV
Ultrasonic Instruments and Devices II Reference for Modem Instrumentation, Techniques, and Technology
PHYSICAL ACOUSTICS Volume XXIV
CONTRIBUTORS TO VOLUME XXlV ARTHUR BALLATO BRUCE B. CHICK AARON J. GELLMAN ROBERT S. GILMORE NEIL J. GOLDFINE ROBERT S. HARRIS FRED S. HINKERNELL WILLIAM LORD BRUCE MAXFIELD CLYDE G. OAKLEY EMMANUEL P. PAPADAKIS STEPHEN R. RINGLEE ALAN SELFRIDGE RICHARD A. STERN SATISH UDPA JOHN R. VIG
Ultrasonic Instruments and Devices II Reference for Modem Instrumentation, Techniques, and Technology Edited by
R. N. THURSTON
.
BELL COMMUNICATIONS RESEARCH, INC. RED BANK, NEW JERSEY
ALLAN D. PIERCE PENNSYLVANIA STATE UNIVERSITY UNIVERSITY PARK, PENNSYLVANIA
Volume Editor
EMMANUEL P. PAPADAKIS QUALITY SYSTEMS CONCEPTS, INC, NEW HOLLAND, PENNSYLVANIA
PHYSICAL ACOUSTICS" PRINCIPLES AND METHODS VolumeXXIV
ACADEMIC PRESS San Diego London Boston New York Sydney Tokyo Toronto
This b o o k is p r i n t e d on acid-free p a p e r ( ~ COPYRIGHT 9
1999 BY ACADEMIC PRESS.
ALL RIGHTS RESERVED. NO PART OF THIS PUBLICATION MAY BE REPRODUCED OR TRANSMITTED IN ANY FORM OR BY ANY MEANS, ELECTRONIC OR MECHANICAL, INCLUDING PHOTOCOPY, RECORDING, OR ANY INFORMATION STORAGE AND RETRIEVAL SYSTEM, WITHOUT PERMISSION IN WRITING FROM THE PUBLISHER. ALL BRAND NAMES AND PRODUCT NAMES MENTIONED IN THIS BOOK ARE TRADEMARKS OR REGISTERED TRADEMARKS OF THEIR RESPECTIVE COMPANIES. THE APPEARANCE OF THE CODE AT THE BOTTOM OF THE FIRST PAGE OF A CHAPTER IN THIS BOOK INDICATED THE PUBLISHER'S CONSENT THAT COPIES OF THE CHAPTER MAY BE MADE FOR PERSONAL OR INTERNAL USE, OR FOR THE PERSONAL OR INTERNAL USE OF SPECIFIC CLIENTS. THIS CONSENT IS GIVEN ON THE CONDITION, HOWEVER, THAT THE COPIER PAY THE STATED PER COPY FEE THROUGH THE COPYRIGHT CLEARANCE CENTER, INC. (22 ROSEWOOD DRIVE, DANVERS, MASSACHUSETTS 01923), FOR COPYING BEYOND THAT PERMITTED BY SECTIONS 107 OR 108 OF THE U.S. COPYRIGHT LAW. THIS CONSENT DOES NOT EXTEND TO OTHER KINDS OF COPYING, SUCH AS COPYING FOR GENERAL DISTRIBUTION, FOR ADVERTISING OR PROMOTIONAL PURPOSES, FOR CREATING NEW COLLECTIVE WORKS, OR FOR RESALE. COPY FEES FOR PRE-1997 CHAPTERS ARE AS SHOWS ON THE CHAPTER TITLE PAGES; IF NO FEE CODE APPEARS ON THE CHAPTER TITLE PAGE, THE COPY FEE IS THE SAME AS FOR CURRENT CHAPTERS.
ACADEMIC PRESS
A Division of Harcourt Brace and Company 525 B Street, Suite 1900, San Diego, C A 9 2 1 0 1 - 4 4 9 5 , U S A http://www.apnet.com
ACADEMIC PRESS 2 4 - 2 8 Oval Road, L o n d o n N W 1 7 D X http://www.hbuk/co.uk/ap/ International S t a n d a r d Series N u m b e r : 0 8 9 3 - 3 8 8 X International S t a n d a r d B o o k N u m b e r : 0 - 1 2 - 4 7 7 9 4 5 - X P R I N T E D IN T H E U N I T E D S T A T E S O F A M E R I C A 9 8 9 9 0 0 01 0 2
DS
9 8 7 6 5 4 3 2 1
Contents ix xi
CONTRIBUTORS PREFACE
1 The Process of Technology Transfer and Commercialization
ESSAYI
ACHIEVINGSUCCESSFUL TECHNOLOGY TRANSFER, AARONJ. GELLMAN ESSAYI1 DIFFICULTIES IN TECHNOLOGY TRANSFER, EMMANUEL P. PAPADAKIS ESSAY111 COMMERCIALIZATION: FROMBASICRESEARCH TO SALES TO PROFITS, NEIL J. GOLDFINE ESSAYIV PERSPECTIVES ON TECHNOLOGY TRANSFER AND NDT MARKETS,STEPHEN R. RINGLEE ESSAYv TEAMING-A SOLUTION TO THE PROBLEM OF INTEGRATING
1 7
15
20
SOFT SKILLS AND INDUSTRIAL INTERACTION INTO
ENGINEERING CURRICULA, WILLIAMLORD,SATISHUDPA, AND ROBERTs. HARRIS ESSAYVI INNOVATIVE TECHNOLOGY TRANSFER INITIATIVES, ARTHURBALLATO AND RICHARD STERN
24
33
2 Fabrication and Characterization of Transducers
EMMANUEL P. PAPADAKIS, CLYDEG. OAKLEY, ALANSELFRIDGE, AND BRUCEMAXFIELD
I. 11. 111.
INTRODUCTION MONOLITHIC PIEZOELECTRIC PLATETRANSDUCERS COMPOSITE TRANSDUCERS V
44 45
76
Contents
vi IV. V. VI.
PVDF FILMTRANSDUCERS ELECTROMAGNETIC ACOUSTICTRANSDUCERS (EMATs) SUMMARY
107 116 129
3 Surface Acoustic Wave Technology: Macrosuccess through Microseisms
FREDS. HICKERNELL
I.
INTRODUCTION MEASURES OF SUCCESS SURFACE ELASTICWAVES IV. PRELUDE TO THE SAW ERA (THEEARLYRUMBLINGS) THE INTERDIGITAL TRANSDUCER, MATERIALS, AND FABRICATION V. VI. INTERDIGITAL TRANSDUCER CONTROLLED SAW DEVICES VII. ELECTRODE CONFIGURED MATCHED FILTERDEVICES VIII. SIGNAL PROCESSING THROUGH THE PASSIVE CONTROL OF SAW PROPAGATION IX. ACOUSTOELECTRIC SIGNALPROCESSING X. ACOUSTO-OPTICS XI. SAW SENSORS XII. FUTURESUCCESS ACKNOWLEDGMENTS REFERENCES APPENDIX A. SAW PUBLICATIONS APPENDIX B. SAW CONFERENCES APPENDIX C. SAW APPLICATIONS APPENDIX D. WORLDWIDE SAW ACTIVITIES APPENDIX E. THESAW ENGINEER’S ROLEAS AN ARTISAN 11. 111.
136 138 141 145 148 156 170 174 183 186 186 187 189 190 194 197 203 204 206
4 Frequency Control Devices
JOHN R. VIG AND ARTHURBALLATO I. INTRODUCTION 11. APPLICATIONS 111. FREQUENCY CONTROL DEVICEFUNDAMENTALS IV. RELATEDDEVICES V. FOR FURTHERREADING REFERENCES
209 210 222 267 269 269
vii
Contents
5 Industrial Ultrasonic Imaging/Microscopy
ROBERT
s. GILMORE
275 SUMMARY 277 11. INTRODUCTIONAND HISTORICAL REVIEW 288 111. LISTOF SYMBOLS AND ABBREVIATIONS IV DESCRIPTION AND THEORY OF ACOUSTIC IMAGING/MICROSCOPY 289 295 ROLEOF IMAGEDMATERIAL: PERMITTED RESOLUTION V 323 VI. APPLICATIONS 343 VII. CONCLUSIONS AND FUTUREWORK 344 ACKNOWLEDGMENTS 344 REFERENCES
I.
6 Research Instruments and Systems
BRUCEB. CHICK I.
HISTORICAL BACKGROUND ATTENUATION MEASUREMENTS 111. VELOCITYMEASUREMENTS I\! ATTENUATION AND VELOCITY MEASUREMENTS V NONLINEAR MEASUREMENTS VI. THINFILMMEASUREMENTS VII. ACOUSTICEMISSIONMEASUREMENTS REFERENCES
347 348 348 351 355 357 358 36 1
SUBJECT INDEX
363
11.
This Page Intentionally Left Blank
Contributors
Numbers in parentheses indicate the pages on which the authors’ contributions begin.
ARTHURBALLATO (33, 209) U.S. Army CECOM Fort Monmouth, NJ 07703-5201 BRUCEB. CHICK(347) RITEC, Inc. Warwick, RI 02886 AARONJ. GELLMAN (1) Northwestern University Evanston, IL 60208 (275) ROBERTS. GILMORE General Electric, Co. Schenectady, NY 12309 (15) NEILJ. GOLDFINE JENTEK Sensors, Inc. Watertown, MA 02 172 ROBERTS. HARRIS(24) Iowa State University Ames, IA 5001 1 FREDS. HICKERNELL (135) Motorola, Inc. Scottsdale, AZ 85257 WILLIAMLORD(24) Iowa State University Ames. IA 5001 1 ix
Contributors
BRUCE MAXFIELD (43) Industrial Sensors & Actuators San Leandro, CA 94577 CLYDE G. OAKLEY (43), Tetrad Corp. Englewood, CO 80112
EMMANUEL P. PAPADAKIS(7, 43) Quality Systems Concepts, Inc. New Holland, PA 17557 STEPHEN R. RINGLEE (20) E-Markets, Inc. Ames, IA 50010 ALAN SELFRIDGE (43) Ultrasonic Devices, Inc. Los Gatos, CA 95033 RICHARD A. STERN (33)
U.S. Army CECOM Fort Monmouth, NJ 07703-5201 SATISH UDPA (24) Iowa State University Ames, IA 50011 JOHN R. VIG (209) U.S. Army CECOM Fort Monmouth, NJ 07703-5201
Preface The purpose o f this b o o k is to show examples o f the successful commercialization o f devices and instruments arising from research in ultrasonics carried out over previous years. M u c h o f the research has been reported (in the research stage and in the m o d e o f research reports) in earlier v o l u m e s o f this series, Physical Acoustics: Principles and Methods. Basically, there is progression from idea through research, development, technology transfer, and commercialization to application or use by a set o f customers. The "Water Slide" diagram in Figure 1 illustrates this progression (Papadakis, \
i ...
New
Laboratory (Conception)
~ Developer
(Technology Transfer)
l
Industrial
(Commercialization)
Figure 1 "Water Slide" diagram showing risk as a function of time as more value is added step by step to an idea to bring it to the status of a product. Because an idea may be rejected at any step of the process or may fail in the field with customers, the number of products is much smaller than the number of ideas initially. (MaterialsEvaluation. Used by permission.) xi
xii
Preface
1992). The purpose here is to demonstrate that research in physical acoustics has led to successful commercialization of devices and systems useful to the public in the broad sense. The basic idea is that research, development, technology transfer, commercialization, and sales are part of a "food chain," so to speak, in which each step is necessary and all steps are interdependent. Research can lead to development if the results of the research seem potentially useful. A conscious decision must be made by someone with resources to enter the development stage. Some developments with real utility go on to technology transfer, which means that the developed item is turned over to a user organization and actually employed. However, technology transfer has been defined as successful only when the transfer is a financial transaction between a buyer who is ready, willing, and able to buy and a seller who is ready, willing, and able to sell ~ just as in buying a house (See Professor Gellman's essay). When either the buyer or the seller is a "captive audience," technology transfer has not yet happened. For instance, a prototype handed over as the "deliverable" on a sole-source contract does not constitute technology transfer. Neither does the installation in a factory of a system developed by the R&D department of the same company~unless the factory's management has the power to refuse instead of just an obligation to accept. The idea of the captive audience and the consequence of incomplete technology transfer does not deprecate in any way the quality of the science and engineering that went into the development, nor does it mean that the item is not actually used. In this book, the meaning of "commercialization" goes two steps further. Technology transfer is not commercialization even if money changes hands for the delivery of the first copy of an item. Even the sale of the second copy of the item is not defined as commercialization for the purpose of this book. This treatise does not deal in the delivery of only one or two copies of an item. Instead, commercialization is defined as the sale of three or more copies in arm's-length transactions. Subsequent numbers may be subject to modification, improvements, or customization, but the principle is there. The idea of the captive audience is superseded when a vertically integrated organization uses many copies of an internally developed item. Then it is implicitly assumed that the organization would have carried out a "buymake" decision vis-fi-vis competitive items before it actually chooses its own. In a case of thorough-going vertical integration, a policy decision to buy only internally might have been made. Commercialization could still happen. The most well-known examples of a commercially successful firm with total vertical integration was the Bell Telephone System before the divestiture
Preface
xiii
ruling. That firm as a regulated monopoly commercialized items (telephones) by renting without even selling them. It is still arguable that such regulated monopolies are the best type of economic organization. Every Sophomore economics book offers a proof that the opposite ~ namely, perfect competition~leads to zero economic profit, instability, bankruptcies, and a return to monopoly as only the strongest survives (see, for instance, Samuelson & Nordhaus, 1989). However, economics per se is not the subject of this book. The concern here is with developed items that have been commercialized by selling three or more copies. The seller may or may not be the inventor. Sometimes an inventor also has business acumen and starts a company to reap the benefits of the invention. Often the invention is transferred to another organization for sale. In the present era of downsizing and outsourcing, progressively fewer items are invented, developed, and utilized internally. Commercialization involving sales on the open market with competition, not sole-sourcing, is very relevant. In Chapter 1, several authors address the processes of technology transfer and commercialization from the point of view of "how-to" and successful examples. This chapter introduces the concepts and points out difficulties. Following the lead chapter, there are chapters on various classes of ultrasonic devices and systems that have come to fruition. Included are medical ultrasonic diagnostics, nondestructive testing (NDT), process control, surface acoustic wave (SAW) devices, frequency control devices, research instruments, transducers, and ultrasonic microscopes. The exact title of any chapter may vary from this list a bit. The chapters are liberally illustrated with pictures of actual commercial objects that are or have been in use. The list is not all-inclusive; this is a book and not an encyclopedia. One may object, for instance, that bulk wave delay lines are not given a chapter. One reason is that they were very well covered by John May in Volume 1A of this Physical Acoustics Series. A second reason is that they were subject to the captive audience phenomenon. Almost all were produced ad hoc for some government project in radar or sonar, although a few were utilized as volatile memories in early computers such as Univac. The third reason is that largescale commercialization did not happen since the development of silicon technology for memories occurred just as the bulk wave delay line was poised to enter that commercial market. Two chapters that were under consideration are not included. Their subjects are mentioned here for completeness. One is acoustic emission, which depends on the phenomenon of the generation of sonic and ultrasonic
Preface
xiv
waves as a crack propagates. Other sources such as pressurized gas leaks also emit acoustic emission. The reader is probably familiar with audible acoustic emission from an ice cube fight out of the freezer when it cracks as it is put into a beverage. Acoustic emission is often classed under nondestructive testing because materials and structures can be tested under stress (not destructive in extent) to determine by "listening" whether cracks propagate. Some monorails and amusement tides, among other things, are tested by acoustic emission when loaded with sandbags. Cracks that are acoustic emission sources can be located by triangulation with appropriate instrumentation. The second potential subject not covered is the uses of high-intensity ultrasound. This technology is used from medicine (to break up gallstones), to automobiles (to weld plastics into multicolored tail lights). No single chapter is all-inclusive in its coverage of all inventions, all scientists, or all manufacturers in its domain. Inclusion of manufactured items as examples in a chapter should be taken as paradigms, not as recommendations for the items or as slights for other items not shown. The book is not a catalog of available merchandise. We hope that this book will show the present success of much past research and will assist in the process of bringing research output into the marketplace, to the benefit of customers. EMMANUEL P. PAPADAKIS
August 1998 References Papadakis, E. P. (1992). Research and real world relationships. Materials Evaluation 50(3), 352. Samuelson, P. A., and Nordhaus, W. D. (1989). "Economics", 13th Edition. McGraw-Hill, New York.
The Process of Technology Transfer and Commercialization Essay I Essay II Essay III
Achieving Successful Technology Transfer, Aaron J. Gellman Difficulties in Technology Transfer, Emmanuel R Papadakis Commercialization: From Basic Research to Sales to Profits, Neil J. Goldfine Essay IV Perspectives on Technology Transfer and NDT Markets, Stephen R. Ringlee Essay V Teaming--A Solution to the Problem of Integrating Soft Skills and Industrial Interaction into Engineering Curricula, W. Lord, S. Udpa, and Robert S. Harris Essay VI Innovative Technology Transfer Initiatives, Arthur Ballato and Richard Stern
Essay I Achieving Successful Technology Transfer PROE AARON J. GELLMAN Northwestern University, Evanston, IL 60208
Introduction
'Technology transfer' has become a popular phrase and a subject of great interest in myriad quarters. Not surprisingly, it has taken on various meanings. To consider it in any detail in the space available, the concept must be bound in several ways. First, technology transfer can be internal--that is, within the same enterprise (e.g., between the corporate R&D organization and an operating profit center). It can also be external (e.g., from one firm to another through, say, a licensing or joint venture arrangement). Second, it is assumed that external transfers follow strictly arm's-length negotiations. Third, presumably all transfers are undertaken with the exception that the technology will be
PHYSICAL ACOUSTICS, VOL. XXIV
Copyright 9 1999 Academic Press Essay V Copyright 9 1996 by Academic Press All rights of reproduction in any form reserved. ISBN 0-12-477945-X $30.00
2
Prof. Aaron J. Gellman
utilized in the market, again as the result of arm's-length bargaining. Also, successful transfer requires diffusion of the technology (or of what it can produce) and not just market introduction. It therefore follows that the interest of this paper is actually "technology transfer and utilization" (TTU). Technology Transfer and the Process of Innovation 'Technology transfer' is but the outcome of a process called innovation. This process begins with an invention, an idea, or a concept and concludes with the introduction of a product or service in the marketplace on the basis of an arm's-length transaction. Joseph A. Schumpeter, the first to appreciate the importance of innovation for economies and societies, ultimately concluded that innovation can best be defined as "something newly tried." "Something" can be a product or a p r o c e s s ~ in modem vernacular, hardware or software, product or service. "Newly" refers to the market in which the "something" is to be advanced. Innovation is market-specific. The same product introduced in a different market represents the culmination of a separate process of innovation. "Tried" conveys that the innovative product or service is not a test article or prototype but rather the practical manifestation of a product or service, even if the underlying technology has many prior, contemporaneous and subsequent deployments. Much innovation takes place through the transfer of technology. As noted, such transfer can occur within an organization or involve different organizations (e.g., sellers and buyers). In a large enterprise, a technology or technique can be transferred between, say, its central R&D facility and a profit center unit of the firm. Or a firm (or govemment agency) can transfer technology by licensing it to another enterprise, public or private, in the same geographical area or in another. For the most part, this chapter will deal only with external transfers (although there are many attributes shared by the two types of transfer). Supply-Push and Demand-Pull For many elements of the process of innovation from invention to market introduction, the driving force is either supply-push or demand-pull. Supplypush can be characterized as "I have, don't you need?" while demand-pull is reflected in "I need, don't you have?" Without doubt, the demand-pull force is the stronger of the two for moving technology into the market place through innovation. It is important to recognize that when considering technology
1
The Process of Technology Transfer and Commercialization
3
transfer, the transferor represents supply-push and the transferee represents demand-pull. Thus it is far better when a potential transferee approaches an enterprise asking for help. (This is the case within firms as well as between enterprises.) Often technology transfer is easier to achieve once the innovation has proved itself in some market. Finding a licensee or a joint-venture partner in another country or market becomes less difficult under such conditions. Put another way, achieving technology transfer is generally more challenging when supply-push is the primary force than when demand-pull is at work. But once success results from a supply-push effort, demand-pull becomes easier to galvanize for subsequent innovations. Depending on the nature of the market for a technology, it can be easier or more difficult to link supply-push and demand-pull in this way. For example, if there is a highly sophisticated, highly aggregated market to be served through transferring technology, supply-push has a higher probability of working than in markets at the other extreme of sophistication and aggregation. (This is one of the reasons that technology transfer has often been very successful when the transfer is to large government agencies or to very large firms producing a range of technology-intensive products and services.) Notwithstanding the preceding, technology transfer is most readily accomplished through the exploitation of a demand-pull force. One of the more effective ways to generate such a force is to look for public or private enterprises that publish performance specifications for their inputs rather than design specifications. Performance specifications open the door to technology transfer and innovation in a way not otherwise possible. Indeed, if an enterprise lives or dies on the basis of the success it enjoys in transferring technology, that enterprise may well be wise to induce the target market to switch to performance specifications.
Promoting Technology Transfer Achieving efficient, profitable technology transfer requires recognition of many of the fundamental "facts of life" regarding innovation and, therefore, of technology transfer. For example, one of the most effective ways to promote innovation and technology transfer and to achieve market diffusion beyond market introduction is to find a champion for the innovation or technology. Innovation and technology transfer are people-processes; no matter how technology-dependent, no matter how technologically sophisticated, at base
4
Prof. Aaron J. Gellman
these processes must involve people who will be put in positions either to promote or to thwart them. Among the more powerful "people forces" available for advancing technology through its transfer to different settings is the champion for the technology or innovation. In fact, in most instances of technology transfer there is a need for a champion from the originating enterprise as well as one in every transferee enterprise. Without these very special people, innovation is very difficult, if not impossible, to achieve based upon technology transfer. But it is not the champions alone that matter. All along the path of technological innovation and transfer (and even diffusion) there are individuals who at one time or another can promote or obstruct, depending on how they are managed. It should always be recognized that it is easier to "prove" a cost associated with any given proposed transfer than it is to calculate the benefits that can be expected. This alone is sufficient to underscore the leverage individuals possess where innovation and technology transfer are concerned.
~
M e c h a n i s m s a n d Catalysts
The mechanisms and catalysts supporting the extemal transfer of technology fall into two categories: (1) overt and explicit mechanisms and (2) indirect or covert mechanisms. Technology can be transferred to another entity through the outright sale of such technology or through a once-for-all payment that transfers title or the fight to use such technology in all markets or in defined markets only. Then there is the licensing of technology through which the provider of the technology receives payment in one or a combination of forms, some of which are usually based on the market success of the transferee. Again, joint ventures can be a means for such transfer where the technology itself forms all or part of the equity of the transferor; similarly, a wholly-owned foreign subsidiary can be established explicitly to receive and exploit a technology. Somewhat more complicated is transfer through the relationship that a prime bears to its suppliers of inputs. For example, the producer of a highly complex and technologically sophisticated product will often have developed designs and manufacturing techniques for components which are to be supplied by firms other than itself. Under such circumstances technology is transferred down the chain of supply rather than horizontally. One of the more effective indirect or covert mechanisms for technology transfer has been patent documentation. Such documentation has proved to be highly catalytic for technology transfer in many cases. It is especially effective
1
The Process of Technology Transfer and Commercialization
5
where the transfer is between different countries, given the great expense the unwilling (and probably unwitting) transferor must bear in order to pursuethe matter in court. And, of course, there is industrial espionage, which everyone knows is quite ubiquitous but few are willing to discuss. Intemational setting of standards for products and processes often results in unintended transfers of technology. While the social benefits of such transfers may be substantial, individual generators of the technological possibilities that are plundered certainly suffer economic harm. Again, in such circumstances, it is usually very difficult for firms to pursue the matter given the cost and other constraints playing on the scene. Reverse engineering is a time-honored if morally reprehensible mechanism for technology transfer. Over many years, even decades, some countries' economic performance has been substantially based on successful and unauthorized reverse engineering of products from other countries. More of an indirect than a covert mechanism of technology transfer is the exchange of industrial personnel between firms in different countries. The exchange of academic faculty between universities can produce a similar result as can students pursuing studies abroad. A number of well-documented cases where exceptional graduate students from developing countries (and even those from developed ones) have taken home with them not only a diploma but also some commercially valuable ideas based on scientific outcomes and technological possibilities they picked up while abroad. Diplomats serving in the roles of commercial or scientific attach6s have often been a source of intellectual capital leading to technology transfer back to their home countries. Rarely have such people been engaged in industrial espionage, but it has happened. Still, in most instances they operate legitimately but are nonetheless frequently invaluable conduits for the international transfer of technology. Surprisingly, perhaps one of the most effective technology transfer mechanisms is the open literature, including scientific and engineering publications and trade journals. The power of technological intelligence derived strictly through such means to influence the course of technological innovation in a country or industry has been demonstrated time and time again.
Some Concluding Observations Technology transfer, as technological innovation itself, faces many resistances. It is better to recognize and understand such resistances than pretend they do not exist. First among them is the general resistance to change that is
6
Prof Aaron J. Gellman
found universally in both organizations and individuals. Innovation is always an uphill battle and so is technology transfer. And natural resistance to change is the primary reason why this must always be the case. Both the successful champion for a technology and the skilled technology transfer agent learn how to nullify or overcome such resistance. It is more difficult to transfer technology where there is a need for system integration as contrasted with a technology or innovation that can be introduced on a stand-alone basis. Using the railroad industry as an example, it has been extremely difficult to introduce a new braking system for railway freightcars in North America because of the necessity to interchange freight equipment freely throughout the continent over many different railroads. Consequently, a system that was adopted more than half a century ago remains the basic standard today; the possibility that one can transition to a new and better form of braking without disrupting the freeflow of cars throughout the system is only now emerging. Had advanced braking systems been applicable to the fleet in a drawn-out manner, such technology transfer would have taken place many decades ago. Another considerable barrier arises when a firm has to write off capital investment remaining on the books to exploit a transferred technology. Firms do not like to take capital losses; some even forbid it as a matter of policy. Thus they may not embrace new technology in many cases as quickly they should for their own benefit. Insufficient data and information about the technology involved presents another barrier to transfer. Both technical and economic data and information must be adequate to support the case for the technology and to overcome the general resistance to change previously discussed. An exceptionally powerful resistance comes from the strong propensity to avoid risk that is characteristic of many firms and some industries. Intelligent risk-beating is too scarce in many firms and governments; this slows technological advance, which of course reduces opportunities for technology transfer. Perhaps only education can overcome this particular resistance and therefore incorporating material that stresses the value and importance of technological advance in appropriate academic curricula may be a matter for priority consideration. Finally, market structure extremes present problems for technology transfer and innovation. The "pure" competitor does not have the resources to pursue technological advance even when the technology is available for transfer at no cost (which is usually not the case). At this end of the competitive spectrum, there simply are no excess funds. Such firms are operating at a subsistence
1
The Process of Technology Transfer and Commercialization
7
level and cannot invest in innovation. Toward the monopoly end of the spectrum, firms generate increasing profits and could, if they so chose, deploy some of those profits to acquire or develop technology in the quest to make even more money. When the extreme is reached, however, the monopolist is so successful that it usually sees no reason to receive or develop technology with which to innovate. Although these remarks about the two extremes may be somewhat overdrawn, they do make the point that the most likely candidates to be transferees are firms in the middle of the spectrum between pure competition and monopoly. In summary, it therefore can be observed that if a technology is to be transferred, whether internally or extemally, the following must be true. 9 The motives must be present in both the transferor and transferee. 9 The technology must be available; that is, the people and resources necessary to accomplish the transfer must be present to ensure successful transfer and exploitation. 9 The technology must be credible; that is, the data and information supporting the case for the technology to be transferred must be comprehensive and believable. 9 The technology must be relevant to one or more of the markets the transferee seeks to serve. 9 The price the transferee has to bear in receiving and exploiting the technology must be acceptable given the potential for profit generated by the endeavor.
Essay II
Difficulties in Technology Transfer: A Perspective EMMANUEL R PAPADAKIS, PH.D. Quality Systems Concepts, Inc., 379 Diem Woods Drive, New Holland, Pennsylvania
This essay presents the author's opinions on a few of the difficulties experienced in technology transfer. In doing so, it addresses some factors that may be causes of these difficulties. As technology transfer is a prerequisite step to commercialization, it is valuable to see the process of technology
8
Emmanuel R Papadakis
transfer from the perspective of different people who have had experiences with the process. Other essays treat the subjects of technology transfer and commercialization from the didactic point of view to show ways and means to accomplish the goal of commercialization. Taking all the essays in this chapter together, the reader may be able to form an opinion on viable ways to bring products to the users through the marketplace. The various types of organizations engaged in research and development experience varying degrees of difficulty in effecting technology transfer. For instance, the small company that decides to build a salable object can bring it to market relatively rapidly, provided it has the capital required. The large, vertically integrated company can do likewise. University professors seem to have the most difficulty, although they sometimes find an easier path if they happen upon an advantageous research topic. So what is the secret of finding the optimum research subject? Observing a deeply felt need on the part of a genuine potential customer may be the key. A great deal of very interesting research has neither need nor customer in this sense. (Indeed, much pure research is not intended for market.) Even when the need is there, all books on marketing point out that of the multiplicity of ideas investigated at some early stage, only a few reach the stage of commercialization. Perhaps this book can show its readers ways to maximize potentials for successful commercialization, beginning with the observation of a market need. As an example, consider Prof. Nicholas A. Milas (1896-1971) of MIT, the organic chemist who synthesized vitamin A and vitamin D (Johnson, 1996). Milas is a "Little Immigrant" success story. Prof. Milas (shortened from Miladakis) was an immigrant boy from Greece. He obtained a 4th-grade education in Greece, received some brief tutoring in math and German in Iowa, and began his college career at Coe College in Cedar Rapids, Iowa, before the United States entered World War I (Papadakis, 1977). He worked his way through, graduating Magna Cum Laude in 1922, and went on to Chicago for this Ph.D. in chemistry, after which he was given a National Research Fellowship at Princeton. Unfortunately, I do not have documentation on the method he used for technology transfer. However, his research undeniably bore the characteristic of having observed a real need with many potential customers: His synthesized molecules have found their way into almost every container of milk in the world as well as into vitamin pills. Universities experience varying degrees of difficulty in getting their research into and through the process of technology transfer. Part of the difficulty must be ascribed to the sources of the ideas being brought to the
1
The Process of Technology Transfer and Commercialization
9
attention of university researchers. Since this book concerns the commercialized products that originated as ideas somewhere and went through the research part of their life cycles in a university (or other) laboratory, any factor impeding the process or wasting resources calls for scrutiny. "Where ideas originate" for university professors and graduate students is likely to be from Requests for Proposals issued by the funding agencies. To some degree, discussions in meetings of the industrial consortia in Research Centers may also contribute. Professional society meetings also play a role in disseminating ideas, but not nearly the primary role they once did. Universities have been the sources of research in the scientific community ever since the Enlightenment. They have been joined more recently (in the 1850-1975 time frame) by a few great industrial and government laboratories. During and after World War II, universities were handed the responsibility of doing much of the scientific and engineering work of the United States government. Note, for instance, the Manhattan Project at the University of Chicago and the Radiation Laboratory at MIT. Now universities are also being asked to take on the responsibility of doing much of the industrial R&D as many industrial firms downsize and eliminate their research capability. This means that universities are under pressure to have more of their output realized as technology transfer leading to commercialization. As companies deliberately eliminate their capabilities in certain technical areas, they must rely on supplier companies, universities, and/or the government to provide the technical know-how for their businesses to survive. The downsized companies reach a point where they are not doing the work but rather are choosing among suppliers and issuing contracts for the work to be done. The government has been doing just this for many years by closing facilities such as arsenals and shipyards while issuing more contracts to industry and universities for technology and its hardware. This leads to some degree of inadequacy in the knowledge base among funding agencies in the most knowledge-intensive field of all, R&D. Many businesses, meanwhile, try to gain economically through technology without complete internal capability by applying for benefits under the "dual-use" doctrine for military facilities, under CRADA (Cooperative Research and Development Agreement) arrangements, and under the concept of "leveraging of resources" at government-sponsored Research Centers in universities in which industrial consortia participate. Thus, university professors get a multiplicity of competing ideas. This might seem to be an advantageous new situation, but is it, the point of view of doing something relevant leading to technology transfer?
10
Emmanuel P. Papadakis
As a source of ideas, government has developed a degree of variability and fluctuation over time in the recent past. Government agencies have tended to find new subjects to emphasize on a rather frequent basis. For example, extending the lifetime of the infrastructure of the nation has come up as a research initiative; in previous years the repair of the infrastructure would have been treated by an infusion of tax money through an appropriation or an expenditure of corporate funds to improve profits. Many government agencies have found it necessary to present new ideas and to change the fundamental emphases of the agencies to satisfy needs that previously would have been handled by other means. For instance, the National Science Foundation (NSF), once only handling pure graduate-level and post-doctoral research, has developed a program to help junior college students transfer into degree programs at universities. Agencies have been caught up to a greater or lesser degree in the Areopagus Syndrome, in which it is necessary to address "some new thing" just to seem relevant. Much of the emphasis in some quarters is to provide "seed money" for a short period of time rather than to decide what is really important to do and to form a commitment to do it (fund it). Thus the nation has great difficulty generating the staying power to carry through a plan worthy of attention. The university professor as a consequence is bombarded with new ideas and funding possibilities that may not last long enough for the completion of relevant research to produce candidate inventions for technology transfer. In this context the university professor must choose something relevant and important to work on, or at least something that will bring in money. Sometimes the requirements for bringing in funding are predominant. It is quite possible for a professor to propose, accept, and carry out research on a concept that his best judgment tells him will not work, his supportive contract monitor has funding to pursue it. Untold millions of dollars are spent at universities and elsewhere on ideas that are infeasible or useless or too dangerous. A classic case of one too dangerous was the ANP (Aircraft Nuclear Propulsion) Project of the AEC (Atomic Energy Commission) in which a sodium-cooled breeder reactor was to supply the heat to power the jet engines of an airplane. (Although the work was done in-house at the AEC for security reasons circa 1955, professors and even students got special "ANP" clearances on top of "Q" clearances to carry on work at AEC facilities.) After a B-25 had crashed into the Empire State Building in 1945, many people believed that having a nuclear reactor flying over populated areas would be too dangerous to contemplate. Yet money was spent on the idea.
1
The Process of Technology Transfer and Commercialization
11
An example of an infeasible idea arose at about the time of Sputnik and the Mohole: It was reported that an RFP had been issued by a defense agency for the orbiting Mole. A tunnel was to be dug secretly around the circumference of the earth in an essentially circular ellipse. When the tunnel had been evacuated, an earth satellite was to have been set into orbit within it for surveillance purposes, i.e., spying. It was further reported that big defense contractors privy to secret RFP rushed to quote on this project to procure defense dollars. While it is very likely that the Orbiting Mole story is somewhere between apocryphal and fantasy, I take it as a fable (like Aesop's Fables) that shows the propensity for seeking and spending money from any spigot. Much more recently, research on thick composites for submarine hulls was sponsored. Various academics worked on this research despite the fact that the builders and repairers of submarines knew that repairs to the intemal mechanisms of submarines require the cutting of large sections out of the hulls. Whereas a metal hull can be welded after such repairs, there is no way (and none is envisioned at the present time) of patching such a gap in a composite hull. Hence the composite hull could not be built in the forseeable future, so the research was useless for its stated intent, namely, the hull of the next-generation manned nuclear submarine. The only possibility for technology transfer of the thick composites would be some serendipitous spin-off. It can be argued that serendipity leads to great new things. NASA argued that the space program accomplished just this in the way of "spin-offs" for the general public. It even works in reverse. For example, the background radiation from the "Big Bang" was discovered by AT&T engineers who were looking for the source of static in microwave telephone transmissions. Having difficulty heating Aunt Bertha on the telephone led to a possible confirmation of a scientific theory of fundamental importance ~ n a m e l y , that there is a point in time (the Big Bang) before which "before" has no meaning and "before which" scientists can measure nothing and hypothesize nothing physical. Thus serendipity is wonderful, but camouflaging the desire for it as nuclear airplanes, moon landings, or plastic submarines calls for scrutiny. The professor must look for the serendipity "in advance" if he or she is concerned with technology transfer opportunities. University professors are put in a particularly difficult position by the mixed signals they receive about what is worth working on and why it is important. First, pure science for the sake of knowledge will always be worth working on. But the output of pure science may never be commercialized and hence may never be reported in a book like this. (The closest some pure
12
Emmanuel P. Papadakis
research may come to popularization is in a book such as A Brief History of Time (Hawking, 1988).) Second, in the university there is the always "publish or perish" dictum. In the minds of some professors, however, worthy ideas that may require years of work prior to publication vie for importance with other ideas that are likely to lead to salable products. It is likely that a professor will yield to the pressures to conform to academic practices and eschew practical endeavors; the dean wants papers published in high-quality journals, not practical ones. Third, once projects have been started, the professors Oust as all people) are likely to become fixated on their personal ideas and not see them as impractical even is such a condition were to be pointed out. Fourth, their funding agent is not interested in sales. FitCh, the professors rarely have the ongoing, vital feedback from a customer with a real need. The funding agent may perceive a real need, but is several layers of administration removed from the real customer. By contrast, consider the millieu at a major corporation--one that I have experience with, the Ford Motor Company. The Vice President of Research was actively cognizant of the funder/customer relationship. The VP held an annual budget planning meeting to ascertain the value of research projects. He wanted to fund some and terminate others. Principal Investigators would present their work very succinctly (say, in 41 minutes) in the format of Problem/Value/Approach/Status/Plans. The Problem was the real company problem being addressed by the Investigator; the Value was the dollar figure that could be saved from costs or added to profit by solving the Problem; the Approach was Technical in a sentence or two; the Status included the percentage of completion of the R&D project; and the one-year (budget) Plans also defined the projected completion date and the probability of success. The Investigator is held to all these plans rigorously, just as the input to this little presentation had to be rigorously assembled and understood by the Investigator. (One year I had to fly home from a family vacation to give such a presentation. It was serious business.) The customer was not only the Vice President but also the Division of the Company that had the problem and had enlisted the support of Research; the investigator has plenteous feedback. Even more direct was the feedback on one development project carried out at Panametrics, Inc., a small R&D company and builder of instruments. In 1970, I proposed building a commercial Pulse-Echo Overlap ultrasonic velocity metrology instrument of great versatility and accuracy. Edmund H. Carnevale, the President of Panametrics, agreed to underwrite the development expense after one copy had been sold. He put up the marketing money
1
The Process of Technology Transfer and Commercialization
13
for this attempt. The NRC (National Research Council) in Canada ordered one sight-unseen; then design and production commenced. About 15 were sold in the first three years. Other feedback is available in industry, also. In industry m a practical place m i d e a s that seem promising to researchers often are squelched by factory managers at the very outset so that little resource is expended upon them. One example is an X-ray system that would have been the ideal technical solution to an on-line quality monitoring problem in an automative plant. But the use of X rays was rejected because the labour union would have refused to have workers in the vicinity even with good shielding of the machine. Another example is a fluorescent additive that could have been added to oil in shock absorbers to make any leakage visible to an automatic UV camera. This idea was rejected because of the cost and delay that would have been needed to qualify the new oil mixture as noncorrosive to materials in the device. A final example is an ultrasonic device using attenuation to measure grain size that was proposed to a brass company. The company's response was that the small expense of having a person polish a small tab of the brass sheet and examine it under a microscope would render ultrasound cost-ineffective. These reasons may seem crude and Philistine relative to the elegance of research, but that is the way researchers must learn to think in order to get research into production and not "spin their wheels" uselessly on never-to-be-wanted products. At AT&T (the Bell system) before the divestiture ruling, technology transfer was frequently carried out by reassigning personnel to a more advanced division as the development progressed. A scientist in research might be moved to a device development department and then to a systems integration department, and so on, until finally finding himself in the long lines department making his device fit into a system to transmit telephone calls coast-to-coast. Although this personnel transfer did not happen often as a percentage of personnel, it happened often numerically in such a large establishment. By contrast, the university professor is not moved and generally does not want to move. He or she is held to pure research by the "publish-or-perish" philosophy of academic life. He or she must "toss the research over the wall" to the next stage of development, as Deming so aptly says. It is claimed (Deming, 1982) that even in industry this "walled enclave" mentality regards progress between research, development, engineering, manufacturing planning, production, and sales. How much more is the professor hemmed in! And feedback does not come back "over the wall," either.
14
Emmanuel R Papadakis
In the context of the mixed signals coming as inputs, it is easy to see the difficulties that academics have in directing their output into the channels of technology transfer. Faced with a professor "wearing the hat" of a salesman trying to market some new invention, the industrialist occasionally makes the judgement that he or she is being presented with a "solution looking for a problem." This means that the professor was doing something interesting and publishable while being under the impression that it would also be useful. Had he been in industry, his market research department could have found out through surveys or by consulting factories whether the idea was useful. But the university system has no direct mechanism for such feedback. This is not criticism of universities--they were set up in the Middle Ages for an entirely different purpose, namely, saving and increasing knowledge while providing intellectual freedom for the professors. Few university founders (except the founders of the Colorado School of Mines in 1874 and the builders of the new MIT campus in 1916) ever envisioned government and industry being so dependent on the university nor, indeed, the university being so dependent on government and industry. So, in this book, the items being reported have run the gauntlet from the idea phase through research, development, technology transfer, and commercialization into use. Many ideas have dropped by the wayside, being overwhelmed by negatives. This book reports on successes without any pejorative opinions about the ideas that have been rejected at any stage. And, as mentioned earlier, completeness of coverage is not claimed. REFERENCES Deming, W. Edwards (1982). "Quality, Productivity, and Competitive Position." Center for Advanced Engineering Studies, MIT, Cambridge, Massachusetts. Hawking, Stephen W. (1988). "A Brief History of Time: From the Big Bang to Black Holes." Bantam Books, New York. Johnson, Jean (1996). Private communication, Alumni Director, Coe College, Cedar Rapids, Iowa. Information in Coe tracking system for alumni. Dr. Nicholas A. Milas was awarded his B.S. there in 1922. Papadakis, Philippos E. (1977). Private communication.
1
The Process of Technology Transfer and Commercialization
15
Essay III Commercialization: From Basic Research to Sales to Profits NElL J. GOLDFINE JENTEK Sensors, Inc., Watertown, Massachusetts
Initially, when Emmanuel Papadakis asked me to write an essay on commercialization for this book, I questioned whether readers of an ultrasoundoriented book would benefit from the perspectives and experiences of an eddy current and dielectric sensing company. Emmanuel, however, convinced me that describing my company's commercialization steps would provide insight for researchers and entrepreneurs hoping to commercialize products originating at universities. I founded JENTEK Sensors, Inc. in January 1992 to develop and market a new sensor and measurement technology. This technology, including the Meandering Winding Magnetometer T M (MWM) and the model-based measurement grid approach, was originally developed at the MIT Laboratory for Electromagnetic and Electronic Systems by Prof. James R. Melcher (now deceased) and me (U.S. Patent Numbers 5,015,951, and 5,453,689). The MWM is an advanced eddy current sensor that can either be scanned across a surface or surface mounted like a strain gage. In its array format, the MWM can build images of cracks, coating thickness variations, or early stage fatigue damage for metal components. The measurement grid approach uses models of the MWM to generate look-up tables for properties of interest such as coating thickness, conductivity as a function of depth from a part surface, or magnetic permeability. The MWM modeling and sensor design research began in the early 1980s. This was proceeded by over 20 years of basic research by Prof. Melcher. After more than ten years of focused effort on the MWM, JENTEK is now successfully selling GridStation T M Measurement Systems with MWM probes and measurement grids for a wide range of applications in the aerospace, energy, and manufacturing sectors. (Note: A second, dielectric (capacitive) sensor has also been commercialized by JENTEK. This dielectrometer technology may be used to measure coating thickness; detect anomalies or disbonds; and measure moisture profiles, cure state, and other microstructure changes in relatively insulating materials such as polymers,
16
Neil J. Goldfine
ceramics, soil, paper, paint, plastic, concrete, electronic materials, and glass fiber composites.) Through this essay, I hope I can provide some insight into the difficult path of building acceptance for a "theoretically complex" product by describing the path I chose for the commercialization of the MWM sensor technology. When I first publicly presented the MWM technology at NIST in 1992, I understood little about the barriers to new products erected by the entrenched vendors in the nondestructive evaluation (NDE) industry. Of course, such barriers must be overcome to successfully market any new product. I was fortunate early on to meet industry experts who helped me navigate the painful process of converting the skeptics and building customer awareness of our new capabilities. For example, one mentor, with the U.S. Navy, provided an unusual understanding of the needs and problems of a large portion of the traditional NDE industry. She immediately told me that if I was going to sell to this community, I must remove the theoretical focus (i.e., "get rid of all those equations") and focus instead on the customer's needs. She also introduced me to several key players in the industry, opening doors that may otherwise have taken years to access. The following is a brief description of the five steps along JENTEK's chosen commercialization path. A plot of the path is shown in Figure 1.
STEP 1: BASIC RESEARCH The commercialization path begins with basic research and development. For fundamentally new technology, this step generally occurs over a period of ten to twenty years, or more. For JENTEK's technology, this research began at the Cash Flow
Second Product Launch . . . . .
First
Innovation
\
Product Launch '
time
,,
Basic Research U•
10-20 years
|
i
Transfer
7"
1 Z2:
~,e
SS"
~ ,, ~ ~
.
~
J
"'" ,Profitabilit ss~SS~
~==
[
I
-7
FIG. 1. Commercializationpath.
Profita bil ityy ,,._ time
Continued Investment in Product Enhancements
1
The Process of Technology Transfer and Commercialization
17
MIT Laboratory for Electromagnetic and Electronic Systems and continued at JENTEK. The purpose o f basic research along the commercialization path is to provide the foundation for an innovation that can be "built out" into a product that provides a substantial technology lead over existing products in the market or provides solutions to a range o f problems that cannot currently be solved. JENTEK products are founded upon the capability to accurately model the interaction of magnetic and electric fields with multiple layered media, so that sensors can provide quantitative measurement of properties such as conductivity, permeability, dielectric constant, and layer thickness. These properties are then related to customer properties of interest such as coating thickness, cure state, crack size and depth, shotpeen quality, fatigue life, and other representations of quality or aging. Such a well-defined technology "asset" is critical to the sustainable success of a leading-edge company. STEP 2: INNOVATION The innovation phase is the most illusive of the commercialization path. While I was completing my Ph.D., I worked for several years for an investment banking firm. There I learned two things about innovation: (1) fundamental innovations that can sustain a profitable company by providing a maintainable competitive advantage are very rare and (2) innovations almost always require both substantial financial investment and a skilled champion to become a successful product. Unfortunately, you will not know if you have a true innovation that can result in a profitable product until you have moved far along the commercialization path. Many have tried (with limited success) to define and teach innovation. I will not attempt to elaborate on this most important step, beyond the following. Innovation begins during basic research and, at some difficult-to-define point, research "data" evolves into a technology that provides new value that did not exist before. Simply put, innovation, within the context of commercialization, results in development of a new technology "asset" that meets a clearly defined customer need. STEP 3: TECHNOLOGY TRANSFER INVESTMENT Many misguided entrepreneurs think that if they have a great idea for a product, investors or customers will come. They won't. You must "build out" your technology with a substantial investment directed at meeting specific identified customer needs with substantial revenue potential in a market that can sustain profit margins while permitting continued product enhancement
18
Nell s Goldfine
and supporting new product development. For almost all technologies, this build out requires both time (3-10 years, depending on the nature of the innovation) and money (from a few hundred thousand dollars to several million dollars, depending on how capital-intensive the technology is and how large the technological lead). This is what I define as the technology transfer investment. In other words, technology transfer does not occur when a company licenses a technology from a university or national laboratory; it occurs during the several years that follow. This step results in the evolution of the technology asset produced by the innovation into a well-positioned and well-defined product or product line that meets specific customer needs and specifications. Cost and pricing issues must be dealt with in this step. For new technologies with fundamentally new capabilities, pricing strategies should consider objectives relating to market penetration, customer's return on investment, sustained profit margins, and initial cash flow requirements. During the technology transfer in vestment step, companies must cover their cash flow requirements with outside sources (e.g., private investment, venture capital, strategic partners, loans ). At point A indicated in Figure 1, these cash flow requirements can begin to be alleviated by customersupported projects. Sources for this might be SBIRs, service revenues, or R&D funding from commercial customers. It is critical not to "sell" your furore profits at this point by giving away excessive royalties or other similar (% of sales) payments. STEP 4: PRODUCT LAUNCH The product launch is the key to profits. Unfortunately, you never know if your product will meet your customers' needs until you are well along the commercialization path (i.e., steps 1-4 are completed). Only when you are selling and delivering product, and your customers are saving money or improving the quality of their products and services, will you know that you have met your customers' needs. Successful commercialization is not about luck, unless you subscribe to the philosophy that "luck is when opportunity meets preparation." Your product must be properly positioned in the market (preparation) to meet specific customer needs (opportunity). This is where you find out if your product is truly innovative. JENTEK has been fortunate; the level of interest in our GridStation product with both of our sensor technologies has been overwhelming. For example, we anticipate that our MWM sensor will become a standard for in-service inspection and "health monitoring" of difficult-to-access locations for
1
The Process o f Technology Transfer and Commercialization
19
machinery, aircraft, and power plants. By mounting the MWM sensor like a strain gage, customers can monitor locations that currently require costly disassembly for inspection. Providing solutions to "hot" applications like health monitoring, combined with our core focus on manufacturing quality control systems, provides us with a balanced approach to building profitable product lines based on our new technologies. STEP 5: THE SEARCH FOR PROFITS
Selling is only a small part of the mission of a business; the true goal is profitability. Achieving profitability requires as much innovation and tenacity as developing new technologies. There are numerous small companies in the NDE market that sell small numbers of systems each year, yet they never manage to grow and build profitability. Companies need both a leader with the vision and capability required to build a profitable enterprise and a truly innovative product that can generate substantial profits over a long period of time. It is also important to remember that the limits of your market are only defined by the potential of your technology and your capabilities to meet customer needs. The traditional NDE market may have been somewhat confining, but, the emerging NDE market is expansive and continues to grow. Depending on which report you read and how you define its limits, the current worldwide NDE market ranges in size from $500 million to $1.5 billion or more. If you include fringe markets such as landmine and unexploded ordnance detection (estimated to have a latent market demand of over $33 billion), medical instruments, and electronic media inspection, then future "NDE" markets could clearly be placed among the largest in the land! To those of you who plan to embark on the entrepreneurial path with a new technology you have developed, I can only hope that you have the support of a spouse, friends, and family as terrific as mine. You will need them every step of the way.
Stephen R. Ringlee
20
Essay IV Perspectives on Technology Transfer and NDT Markets STEPHEN R. RINGLEE E-Markets, Inc., 125 South Third Street, Ames, Iowa
Overview
Major nondestructive test (NDT) research programs at institutions such as Iowa State and Johns Hopkins were established not merely to advance the state of the technologies in various NDT arts, but to transfer these advances to commercial markets. Despite the best efforts of these institutions, the NDT technology transfer record remains mixed at best. In fact, NDT markets may be too small, too specialized, and too conservative to absorb many new technical developments from university and institute laboratories. These institutions would be well advised to manage their technology transfer expectations down to more achievable goals.
NDT Markets
Few reliable market studies exist of the various industrial NDT markets. These markets are extremely diverse and include everything from equipment sales to inspection services, software, and engineering consulting. Available equipment includes disparate technologies such as ultrasound, acoustic emission, eddy current, particle and penetrants, x-ray and other radiography, and magnetic or particle emission. In addition, these technologies are used across a variety of end-use markets, such as aerospace, utility, chemical, electronics, energy, metalworking, and transportation. The firms selling equipment or services tend to be small (under $50 million in sales) and are frequently subsidiaries of larger technology companies. While the market has been growing consistently, this growth varies considerably by submarket and is in only a few cases above ten percent per year. As a result, the NDT market can be understood not as a distinct market, but as a collection of solutions to various material and component testing problems. Each of these problems is unique, calling for an individual technical solution. Market participants must have a great deal of technical
1
The Process of Technology Transfer and Commercialization
21
sophistication and flexibility to adapt various approaches to each problem. Even the largest NDT equipment marketers sell equipment for use in a number of end-use applications and must offer engineering assistance and customized software probes, or other devices to help end user to solve the particular technical problem at hand. This assistance requires that NDT sales have margins high enough to justify the expense of customer service and customized engineering assistance. In many cases, these margins are achieved by the cost avoided by the customer from downtime, accidents or loss of use (as in the utility and aerospace industries). The price of inspection services and equipment can also be justified through lower life-cycle costs and increased equipment longevity (as demonstrated in military aircraft and ships). Many NDT firms have also achieved higher margins through the bundling of inspection services and equipment. Unfortunately, higher margins and prices have limited the size of the market for equipment and services. Customers will usually adopt the technique of least cost and effort ~ preferring lower cost (but less reliable) NDT technologies over higher cost, newer and more reliable solutions-unless they have a compelling reason to do otherwise. Regulation paradoxically tends to lock NDT solutions to a "lowestcommon-denominator." Technical gatekeepers such as aircraft manufacturers and govemment agencies limit the ability of end users to employ other than "approved" NDT equipment. These gatekeepers are resistant to new approaches unless they are of compelling utility or meet immediate needs of safety or reliability. Limitations on Technical Innovation
The market forces that define the NDT markets--including the market variety, the intensity of customer service, the need for high margins, and the effects of regulation~all combine to limit the extent to which new technologies can be successfully introduced to end users. Technical developers, including those firms now selling NDT technologies, must justify the investments made to develop, perfect, manufacture, and market new designs. In a market characterized by small sub markets, limited growth, and frequent user resistance to new technology, an equipment or software developer must very carefully discriminate among NDT investment projects. The path that most have chosen, and that tends to yield the highest return on investment, is one of small, incremental improvements to existing technologies. In these cases, the investments required are limited to the engineering efforts needed to
22
Neil J. Goldfine
adapt equipment to a new problem and perhaps the development and manufacture of a new probe or transducer. The case histories of companies such as Zetec, Krautkramer-Branson, and Panametrics illustrate this path.
The Frustrations of University-to-Market Technology Transfer Universities have created "centers of excellence" at institutions such as Iowa State to advance the state of the NDT art. These centers are supported through a mix of state and federal government, university, and industry funds. By design, the leading centers have sought active industrial sponsors, not merely for the incremental funding (which in most cases is modest) but also for the connection with industrial problems and existing solutions. The centers encourage a dialog among their supporting groups through regular conferences, publications, and informal conversations. They also solicit industry and government problems and propose research and development projects to solve these problems. Unfortunately, these centers have not had the success at technology transfer that was expected when they were first created. The author's experience with the Center for NDE at Iowa State illustrates many of the reasons for this disappointment. Technologies developed at the university usually require substantial reengineering or adaptation before they can be used in commercial applications. In many cases, they are not designed with the product economics in mind. In one instance, an ultrasonic instrument developed to measure the texture of rolled sheet steel was designed to a price point of $90,000 in 1989 economics assuming manufacture in batches of 10. Built to order, one instrument would have a price point of about $200,000 in 1995 economics. Although the instrument was of industrial quality and had technical merit as well as on-line potential, if neither materially improved the measurement desired by the customer nor saved the user any time compared with the existing off-line mechanical tester, which sold for approximately $12,000. Consequently, no users or potential NDT equipment makers showed any interest in buying, licensing, or developing this ultrasonic instrument. In other cases, the university-developed technologies are embryonic and require substantially more refinement before they can be reliably used by lesssophisticated customers. Another development, a pulsed eddy-current flaw detector, had great technical merit but required almost six years of additional work in the laboratory before commercial users showed interest in licensing it. In many cases, the funding for the additional developmental work is difficult for a university lab to either obtain or justify, leading to many half-done
1
The Process of Technology Transfer and Commercialization
23
projects that remain on a lab bench for years. This added investment is likewise hard for a commercial user to justify, and significantly limits the interest they show in licensing-developed ideas. Further, many university-incubated projects are begun for motivations other than commercial success. Research papers, student projects, and personal interests are all very important in the academic culture. NDT developments occur during this research work, but are not often pursued through the difficult, expensive, and boring requirements of application development. This application development is a necessary precondition to commercial success. Although a researcher may have achieved significant results from a research project, the incremental work required to mm the results into a licensable technology or even a commercial product is not work at which the university culture excels, enjoys, or believes it can justify. These cultural and economic imperatives create a gap between the university labs and the commercial NDT marketplace. Few technologies cross this gap. Those that do are introduced into a conservative NDT marketplace with growth rates low compared with markets such as communications, computing, and software. There are successes in NDT technology transfer, such as Krautkramer-Branson's license of eddy-current instrumentation technology developed by NASA for crack detection. However, the history of the last 15 years of university-based centers of NDT research illustrates the difficulties and infrequency of transfer of NDT technologies to industry. Those centers that advertise their success at NDT research need to keep this in mind when approaching sponsors. Center managers as well need to hold technology transfer expectations that are achievable, and should work with their funding sponsors to educate them to the realities of the NDT markets.
24
William Lord, Satish Udpa and Robert S. Harris
Essay V Teaming A Solution to the Problem of Integrating Soft Skills and Industrial Interaction into Engineering Curricula WILLIAM LORD AND SATISH UDPA Electrical and Computer Engineering Department, Materials Characterization Research Group
ROBERT S. HARRIS Center for Advanced Technology Development, Iowa State University, Ames, Iowa This essay describes two unusual R&D projects that have resulted in the development of a new paradigm for engineering education. The first project, funded by the Gas Research Institute, consists of a consortium of Battelle (Columbus), Southwest Research Institute, Iowa State University, and a number of gas pipeline inspection companies whose overall goal is to improve the state of the art in gas transmission pipeline inspection. The second project, funded by a Japanese company, Takano, Co., Ltd., involved the design and development to industrial prototype stage of an acoustic microscope. O'his project with Takano has since been followed by two other projects for development of other NDE instruments.) Both the GRI and the Takano projects involved large teams of Ph.D., M.S., and undergraduate students that had to interact on a daily basis with faculty, visiting engineers, and postdoctoral researchers while balancing the hard deadlines imposed by the industrial partners and the academic concerns of working toward a degree. Issues relating to project reports, presentations, intellectual property, technology transfer, and industrial interaction were dealt with as a team, which has led to careful consideration of such teams as an integral part of the educational experience for all engineering students. Details of this new paradigm are presented together with suggestions for incorporating it into engineering curricula.
Introduction
Since its founding in 1858, Iowa State University (ISU) has developed a tradition of outreach excellence in fulfilling its role as the state's land grant institution. In keeping with this tradition, DOE's Ames Laboratory has been used as the core from which to spin off more specialized centers with targeted research, development, and outreach activities in the physical sciences and engineering (Snow, 1994). Eleven centers now form the university's Institute
1
The Process of Technology Transfer and Commercialization
25
for Physical Research and Technology (IPRT), which has strong, interdisciplinary ties with the university's colleges and departments. Interactions with private industry are encouraged through IPRT's Center for Advanced Technology Development (CATD) and the Engineering College's Center for Advanced Technology Development (CATD) and the Engineering College's Center for Industrial Research and Service (CIRAS). Although this structure is not vastly different from those existing at many other universities, it does provide a rich and varied R&D backdrop from which effective ties to industry can be developed and from which the ABET Engineering Criteria 2000 program outcomes and assessment for the university's Engineering College can be met. There is no doubt that the pressures for change in graduate education (Holden, 1995; Brill and Larson, 1995), with their emphases on industrial relevance, flexibility, soft skills, etc., are equally present at the undergraduate level. This has been clearly recognized in the ABET Engineering Criteria 2000, where the necessary program outcomes and assessment delineates the expected abilities of graduating B.S. students. From work on two recent R&D projects involving both teams of research organizations and teams of researchers, it is the authors' contention that such team-based activities can themselves be the vehicle for a radical restructuring of the graduate/undergraduate experience that will meet the needs of a modem engineering education.
The Projects Two large projects provided the basis for developing the new paradigm. The first project was supported by the Gas Research Institute (GRI) to develop new methodologies for analyzing data from nondestructive evaluation tools used for inspecting gas transmission pipelines. The second project was funded by a Japanese company, Takano Co., Ltd., to design and develop a state-ofthe-art acoustic microscope. The projects provided a rich environment that called for extensive interaction, not only among a large group consisting of graduate students, post-doctoral fellows, undergraduate students, and faculty, but also with a number of external agencies. Although the needs of both projects were addressed using a team approach, differences in the nature of the sponsoring organizations, team structure, and deliverables contributed to subtle variances in the emphasis and character of the two project groups.
26
William Lord, Satish Udpa and Robert S. Harris
GRI PROJECT The GRI project was charged with the task of improving the methods employed for detecting and characterizing flaws in underground gas transmission pipelines (Udpa et al., 1996). Natural gas is a major source of energy that is normally transported from well sites to consumer locations through a vast network of pipelines. The pipelines are inspected periodically using devices called "pigs" that travel through them to gather information concerning their integrity. Most inspection devices currently use electromagnetic methods to test the pipeline. GRI funded a consortium consisting of Battelle, Iowa State University, Southwest Research Institute, and two pipeline inspection services vendors (Vetco Pipeline Services and Pipetronix) to investigate methods for improving inspection technology. The two inspection services vendors were deliberately included in the consortium to ensure that the work done is relevant to the industry needs and to allow for absorption of new technology. The consortium was required to meet several times a year to appraise each other and GRI of the progress made during the reporting period. Inaddition, GRI required monthly progress reports to be submitted. A review of the project early on revealed the need for a systematic and comprehensive approach to solving the problem. First, it was felt that a full and thorough understanding of the physics underlying the inspection process was required to develop appropriate signal processing strategies for characterizing defects. This was accomplished through the use of complex numerical models adapted specifically for stimulating the physical process. The next important task involved the development of some new signal processing schemes for extracting information from the electromagnetic transducers in the pig. The sophisticated nature of the signal processing schemes gave rise to the third task, namely, the development of user-friendly software that the inspection services vendors could use without undergoing major retraining. The ISU team was tailored to meet the needs of the project. At its peak, the group consisted of nine graduate students, two post-doctoral fellows, a visiting faculty member, three undergraduate students, and three faculty members. Each of the graduate students specialized in one of three areas; electromagnetics/numerical analysis, signal processing/pattern recognition, and software engineering. Students were also expected to be reasonably familiar with the two disciplines outside their area of specialization to facilitate free flow of ideas and develop an appreciation for the roles played
1
The Process of Technology Transfer and Commercialization
27
by others in the team. This was accomplished by persuading students to take courses in all the subject areas. The team met regularly every week and all students were expected to present brief summaries of their accomplishments, the problems they faced, and plans for the future. In addition, one of the students was given an opportunity to present his/her contributions in greater detail. The meetings provided the students an opportunity to develop a considerable level of confidence in giving presentations and face critical audiences. The graduate students were also recruited to submit monthly reports. These reports were integrated by the post-doctoral fellows and reviewed by the faculty members before being submitted to GRI every month. Each undergraduate student was mentored by a graduate student and charged with specific responsibilities by the graduate student in consultation with a post-doctoral fellow and faculty members. The tasks assigned to the undergraduate student were usually a subset of the tasks required to be completed by the graduate students. The undergraduate students were rotated among the graduate students to provide them with a diverse set of experiences and capture the excitement that accompanies research. The undergraduate students gave presentations summarizing their work at the end of each semester. It must be mentioned that all the undergraduate students have since elected to pursue graduate studies and stay on at Iowa State University. The interaction between the team and other members of the consortium provided an opportunity for the students to improve a number of soft skills. The consortium meeting locations were rotated through participant sites. Consequently, ISU had the opportunity of hosting the meeting about once a year. Each of the students presented a summary of their contributions to the project at these meetings. The presentations were rehearsed, and every effort was made to handle the event as professionally as possible. These meetings considerably improved the morale of the students and allowed them to take pride in their accomplishments. Sharing their results with the "outside world" gave them a proper perspective of their own contribution toward solving a real problem and prepared them for making presentations at technical conferences. Students also visited several of the consortium member facilities. A visit to Battelle provided first-hand exposure to the extensive test facilities located in West Jefferson, Ohio. The students were also heavily involved in transferring technology to the inspection services vendors. Thus the students had to take several trips to industry sites and spend the time and effort to implement their systems in a real environment. The "leap" from an academic concept to a
28
William Lord, Satish Udpa and Robert S. Harris
working software module is a long one, and the students had to overcome several challenges before seeing results. Apart from absorbing a number of soft skills, perhaps the most important lesson learned by the students was the need to work within the framework of a fixed time schedule. TAKANO PROJECT The Takano project received support to design and develop and acoustic microscope for inspecting small ceramic components and integrated circuits. The project called for the construction of a full prototype embodying both new hardware and software and required Iowa State to transfer all the technology and assist Takano in building a commercial product. Takano also signed an agreement with Battelle for evaluating ultrasonic transducers and requested mutual collaboration in finalizing the transducer specifications. This required periodic exchange of information and coordinating plans for testing and evaluating transducers. The project team was structured in a manner similar to the GRI team. However, there was an additional group devoted to the development of hardware. A total of twelve graduate students, two post-doctoral fellows, one visiting faculty member, five undergraduate students, and three faculty members were involved in the project. The hardware group consisting of four students was responsible for building an advanced computer-controlled pulser-receiver with performance characteristics and features heretofore unrivalled in the industry. The students designed and assembled the circuits, and evaluated them before developing a printed circuit board layout. After the boards were fabricated by an outside agency, the prototypes were assembled and tested. The mechanical design of the system was also developed by the hardware group. The signal processing group was responsible for developing all the signal and image processing algorithms and pattern recognition routines. The software group was charged with the responsibility for developing the operating system software and integrating all the systems. The numerical analysis group was involved in simulating the test geometry both for estimating the parameters necessary for developing image restoration algorithms and for optimizing sensor location and characteristics. As in the case of the GRI group, students were expected to be familiar with disciplines outside their own area to facilitate the free flow of ideas. The team met every week to discuss progress and plans for the future. Each student presented a brief summary, and students were encouraged to participate in discussions. The graduate students were required to submit a
1
The Process of Technology Transfer and Commercialization
29
substantial quarterly report. The reports were integrated by the post-doctoral fellows and reviewed by all participants including faculty members before submission to Takano. The undergraduate students played a key role in the design and development activities, particularly in assembling and testing the hardware. This required them to gain familiarity with advanced circuit design and layout as well as simulation packages. The undergraduate students were required to submit a report at the end of each semester. As in the previous case, all the students eventually elected to pursue graduate studies at ISU. Representatives from Takano paid quarterly visits to ISU to stay abreast of the progress being made by the team. All graduate students were required to make detailed presentations of their contributions. A Japanese translation of the presentation was provided by an interpreter. These presentations were particularly useful in developing an appreciation of the differences in the cultural backgrounds since the presentations had to be tailored to suit the Japanese corporate style. In the process, the students developed a remarkable level of sensitivity toward the Japanese culture and an ability to avoid committing social and cultural snafus. The awareness was heightened even further during the process of transferring technology when several students spent a few weeks in Japan to train Japanese engineers. The project was an educational experience for both ISU participants and Takano. Both entities gained in the process. Takano personnel gained a healthy respect for the creative abilities of U.S. students and the university system that nurtures such talents. The ISU team was impressed by the industriousness of their counterparts and the environment in Japanese industry that brings out the best in their personnel. The direct benefits to both parties were truly manifold. As a consequence of their positive experience with this project, Takano decided to fund additional projects and pursue a long-term relationship with Iowa State University, endow a fellowship at ISU, and establish a manufacturing and marketing facility in Ames, Iowa. Lessons Learned
Following is a list, in no particular order of importance, of the lessons learned through these two projects. 9 Cooperation and teaming are not skills that are easily learned--perhaps more so for faculty than for students. A good example of this is the initial suspicion held by the consortium participants on the GRI project, who had previously had to complete for the same R&D funds.
30
William Lord, Satish Udpa and Robert S. Harris
9 Undergraduate students who experience a good project involvement are more likely to stay on for graduate school. 9 Graduate students who work on industrial projects are more likely to obtain job offers. 9 It takes time and effort to adjust to industrial deadlines and the industrial mind-set. Faculty must be particularly alert to the need to protect the intellectual quality of graduate studies and to ensure that Ph.D. dissertations and M.S. theses are of appropriate academic standard. 9 Academic administrators must learn how to evaluate and reward faculty participation in industrial project activity. 9 Industrial-based projects are an excellent vehicle from which to develop student communication skills in that the need for both oral and written presentations is self-evident for the success of the project and not merely an academic exercise. 9 Students underestimate the difficulty of dealing with the "bureaucracy" in terms of placing orders, getting parts made, etc. 9 The mix of nationalities in most engineering graduate programs automatically ensures the development of social skills as well as sensitivity to cultural issues. 9 Everyone underestimates the complexities of technology transfer and the degree of "hand-holding" that must take place in moving ideas from the lab to the field. 9 Teaming and industrial project activity tends to be low cost in terms of university commitment in that R&D is a normal faculty function. Such interaction also fits in well with a land grant institution's mission and often fulfils the state's need for university/industry cooperation. 9 Patent, intellectual property and ethical issues arise as a natural part of university/industry cooperation, allowing effective integration of special topical seminars and case studies. 9 The synergistic effect of teaming is essential for project success. Three students in a team achieve much more than three students working individually.
The New Paradigm An effective way to integrate soft skills into the engineering curriculum is via teaming (faculty, post-doctoral students, graduate students, and undergraduate students) on industrially relevant R&D projects. Such teams, depending on
1
The Process of Technology Transfer and Commercialization
31
the degree of project complexity, could include all levels of graduate and undergraduate students. The exact logistics would vary from college to college, depending on contacts with industry, faculty involvement in R&D activities, numbers of graduate and undergraduate students, interdepartmental and center cooperation, etc. However, an appropriate hierarchy of involvement would be faculty and industry engineers, Ph.D. students, M.S. students, seniors, and undergraduates in the first three years of their degree programs. Student responsibilities at each level would be as follows: 9 Freshmen, Sophomores, and Juniors would be the equivalent of "apprentices" on the projects and would be required to take "project" hours each semester. Their involvement with a project would grow with experience to include the building and conducting of experiments, data acquisition, analysis of data, and general "gopher duties" associated with all projects. They would be required to write short semester reports documenting their experiences and to give short oral presentations to their "host team." 9 Seniors would be required to carry out a full two-semester "design project" that would include the identification, formulation, and solution of an engineering problem requiring the writing of formal technical reports and the presentation of formal seminars to their peers. Wherever possible, projects would be drawn from industry contacts or have relevance to faculty R&D interests. 9 Graduate students, particularly at the M.S. level, would play an integrating and management role, organizing the team meetings, ensuring adequate progress toward project goals, and evaluating undergraduate work and development of team skills. 9 Faculty and industry engineers would provide the motivation and overall organizational skills needed to ensure a quality educational experience. Each department would need to identify a "project czar" to organize the senior project seminar series and maintain uniform grading procedures. Coordination with class advisers would also be required to handle the complex logistics and to set up appropriate seminar series to cover some of the more general topics relating to professional and ethical responsibility, engineering and society, international issues, etc. Implicit in this new paradigm is the concept that not all students need to spend the rest of their professional lives in industry. A high degree of adaptability and flexibility is required on the part of each college and
32
William Lord, Satish Udpa and Robert S. Harris
department to match the project experiences at all levels to the students' own needs and expectations. A university, after all, is perhaps the only institution that can still provide a nurturing environment for the unique, eccentric, and unusually gifted individual.
Concluding Comments Teaming and work on relevant industrial projects is essential in modem engineering curricula to provide all levels of students with an opportunity to experience first-hand the reduction of scientific principles to practice for the betterment of humankind. Such an educational framework encourages the development of those soft skills needed for survival in tomorrow's dynamic, fast-changing engineering profession. Of paramount importance in the implementation of a project-oriented curriculum is both the commitment of the university and college administration to such an endeavor, and perhaps most importantly, the enthusiasm of each individual faculty member for the need to include this type of experience in the curriculum. The ABET Engineering Criteria 2000 goes a long way toward providing the necessary motivation for change. REFERENCES Snow, Joel A. (1994). National research trends. In "Strategic Planning Position Papers." Iowa State University. Holden, Constance et al. (1995). Careers 95: The future of the Ph.D. Science 270, 121-130. Brill, Arthur S. and Larson, Daniel J. (1995). Are we training our students for real jobs? Academia, Nov-Dee, 36-38. Udpa, L. et aL (1996). "Developments in gas pipeline inspection technology. Materials Evaluation, 54 (4), 407-472. ACKNOWLEDGMENTS
The authors are indebted to the College of Engineering at Iowa State University for providing the environment in which to carry out university/industry projects. Financial support and encouragement from the Gas Research Institute (Harvey Haines, Project Manager) and Takano Co., Ltd., is also gratefully acknowledged. BIOGRAPHY
Dr. William Lord is the first faculty member to hold the Palmer Chair in Electrical Engineering. His interests are in the area of nondestructive testing of materials and the application of numerical modeling techniques to the
1
The Process of Technology Transfer and Commercialization
33
understanding of energy/defect interactions. Professor Lord was editor-inchief of the Institute o f Electrical and Electronics Engineers (IEEE) Transactions on Magnetics from 1991 to 1995. He is a Fellow of the IEEE, IEE (UK), and the British Institute of Nondestructive Testing, and served as a National Direction of the American Society for Nondestructive Testing from 1990 to 1994. Dr. Lord was Associate Dean for Research and Graduate Studies in the Engineering College from 1991 to 1995. In 1995, Dr. Lord was made Anson Marston Distinguished Professor of Engineering. He received the B.S. (1961) and Ph.D. (1964) from the University of Nottingham, England. Dr. Satish Udpa's research interests lie in the broad areas of systems theory and numerical analysis. He is involved with the development of signal processing and pattern recognition techniques for solving inverse problems relating to nondestructive testing. He is also engaged in the application of numerical techniques for modeling a wide variety of physical processes underlying nondestructive evaluation methods. Dr. Udpa received the B.S. (1975) from J.N.T. University, India and the M.S. (1980) and Ph.D. (1983) from Colorado State University. Robert Harris has over 30 years experience in the fields of applied research and development, technology assessment, licensing, and new venture start-up. His recent positions of responsibility include: Interim Director of the Center of Advanced Technology Development at Iowa State University (1996 to present), Manager of Industrial Outreach for the Ames Laboratory, (1995 to present), Director of the U.S. West/SBIR program for the state of Iowa, (which is administered by CATD), Associate Director of the Center for Advanced Technology Development (CATD) and Director, Office of Contract Research at Iowa State University.
Essay VI Innovative Technology Transfer Initiatives ARTHUR BALLATO and RICHARD STERN us Army CommunicationsmElectronics Command, Fort Monmouth, New Jersey The Physical Sciences Directorate (PSD) of the Army Research Laboratory (ARL), Fort Monmouth, N J, has achieved significant success in the technology
34
Arthur Ballato and Richard Stern
transfer~strategic alliance arena by blending technology transfer statutes, regulations, and practices into its corporate culture. The essay describes the process and procedure PSD utilized in achieving that success, along with the lessons learned in introducing new and innovative methodologies, techniques, and approaches in transferring federal laboratory technology to the private sector
Introduction
Physical Sciences Directorate (PSD) is one of eleven directorates of the Army Research Laboratory, with headquarters in Adelphi, MD. This directorate consists of 239 individuals, 177 of whom are scientists and engineers (S&Es), including 59 Ph.D.s. PSD is the Army's focal point for research and development in the physical sciences and generates more than 50 issued patents each year. It provides enabling technologies to solve critical barrier problems in photonics, optoelectronics, microelectromechanics, smart materials, solid-state and nanoscience, electrochemistry, energy science, bioscience, high-frequency electronics, rf acoustics, manufacturing science, and electrophysical modeling, advancing the physical sciences technology base consistent with combat needs of the Army. PSD has developed a well-focused, broad-based technology transfer program that creates both strategic and tactical alliances with business, industry, and academia. The Directorate is committed to an "open laboratory" policy where entrepreneurial S&Es from the private sector may engage in onsite cooperative efforts with PSD's S&Es, utilizing the Directorate's intellectual property, unique and sophisticated facilities, and high level of expertise to develop or improve commercial products and processes of interest to both the private sector and the Army. PSD has implemented more than 60 Cooperative Research and Development Agreements (CRDAs) and 18 Patent License Agreements (PLAs) since the Technology Transfer Act became law in 1986, with additional agreements in various stages of preparation. An income stream, established from these agreements, is being used to expand the PSD technology transfer program effort, cover CRDA operating expenses, and reward contributing S&Es. The goal of PSD is to exploit the latest scientific advances to meet military needs while, at the same time, fostering the creation and improvement of commercial products and services within the civilian economy to foster both the economic and military success of the United States.
1
The Process o f Technology Transfer and Commercialization
35
Changing the Culture Early in the 1970s, PSD management recognized that the preponderance of R&D resources and technology needed by the Army existed outside the organization. Management therefore determined that it was essential for PSD to identify and interact with the best available outside research resources and that this could only be done successfully if the interaction was structured to be to the mutual benefit of all concerned. This approach required individual S&Es to develop an awareness of and familiarity with reseach and researchers throughout the world. Toward that end, S&Es at PSD are required to complete a minimum of 80 hours of annual training in the areas of technology, technology transfer, intellectual property, and marketing. PSD provides a continual string of lecturers from business, industry, and academia who provide workshops, lectures, and seminars to supplement regular college and university curricula. As a result of this approach, informal cooperative efforts were initially developed between PSD and technical contractors, colleges, and universities, and through technical working relationships borne of out personal interactions established at technology conferences, symposiums, seminars, workshops, etc. The aim of these teaming efforts was to address specific technical barrier problems using a larger and a more diverse R&D work force that approached a critical mass. The advent of the Technology Transfer Act of 1986 and the related Executive Orders of 1987 provided, at that point in time, a mechanism for formalizing, expanding, and strengthening informal cooperative efforts that already existed at PSD. The first CRDAs established at the Directorate were carried out with organizations with which PSD had informal or contractual relationships. This approach minimized difficulties and obstacles involved in establishing the initial formal collaborative agreements. Implementation of each of these new CRDAs, however, required continual nurturing by the PSD Technology Transfer Office; otherwise they would have languished.
Defining Technology Transfer To define what is meant by "technology transfer," it is sometimes best to explain what technology transfer is not. Effective technology transfer is not merely the process of making known the availability of technology. It is not "throwing technology over the transom" for others to pick up and possibly utilize without further explanation or discussion. Little is accomplished by
36
Arthur Ballato and Richard Stern
this hands-off approach, and disappointment in the process and its accomplishments will surely follow. Some consider the transfer of federal funding to the private sector for the development of dual-use technology a form of technology transfer. Although useful technology is developed through contractual efforts, this should not be considered transfer of existing federal laboratory technology for the benefit of both the laboratory and the private sector. Likewise, establishing a CRDA for its own sake without specific, focused objectives (no matter what resources are made available for that collaboration) cannot be considered a l~otentially productive technology transfer effort. Without an objective and statement-ofwork beneficial to all parties and to which all parties are committed, disappointment can be expected. A fully functional technology transfer program is one where S&Es are motivated to transfer/transition federal laboratory technology (in-house technology, methodologies, expertise, capabilities, facilities use) to the private sector (business, industry, and academia) by means of a CRDA or PLA in a way that benefits all participating parties. The process should be reciprocal, with both federal and nonfederal partners contributing to the effort. A winwin situation should be a driving objective.
Technology Transfer Advancement There are a number of positive steps that every federal laboratory can take to promote technology transfer. First, it is important that a laboratory intending to develop an effective technology transfer program have leadership that champions technology transfer. A laboratory whose management is unsympathetic to the cause is one where the technology transfer program will neither reach its potential nor satisfy any of those involved. Further, the lab must foster a strategic view of the technology transfer process, seeing technology transfer as a tool used by all elements of the lab to assist the lab in meeting its mission goals. An environment supportive of technology transfer needs to be created. Led by example, S&Es trained in the generation and use of CRDAs and PLAs by a technology transfer manager and a user-friendly legal staff will understand the value of intellectual property and appreciate the value of cooperative efforts. By encouraging S&Es to attend and network at technical meetings, conferences, workshops, and seminars and to generate working relationships with their peers outside of their laboratory, even marketing soon becomes a distributed activity. The use of database searches, publishing and
1
The Process of Technology Transfer and Commercialization
37
presenting technical papers, and S&E exchanges all contribute to the open laboratory concept. Certain ingredients are necessary for operating a successful technology transfer program. Among these are motivation, flexibility, trustworthiness, unique technical expertise, infrastructure equipment and easy access thereto, identification of technology ripe for commercialization, realistic expectations, clearly delineated responsibilities, fast response time, local negotiation authority, simple procedures, reasonable CRDA and PLA terms, and userfriendly attomeys.
Collaborative Research and Development Agreements PSD has now established nearly all of its technology transfer and collaborative efforts under formalized CRDAs and PLAs. These businesslike contracts are useful because they define the technology transfer process--including allocation of intellectual property fights, termination procedures, dispute resolution, and publication rights m a n d they limit federal liability through indemnification. Successful technology transfer efforts usually exhibit the following characteristics: participants have a strong vested interest in the outcome; all participants benefit in fair and equal portions; all parties are involved early on; the technology user/customer is involved in the early stages, with simultaneous development of technology and product; and the customer-oriented product is actively marketed. Conversely, barriers to technology transfer include a perceived lack of faimess, suspicion or lack of trust, NIH syndrome/arrogance/pride, a legalistic mind-set, and short-term focus. Each can hinder successful technology transfer. The PSD Technology Transfer Office processes, on the average, one new CRDA a week. Although the content of these CRDAs is very similar, each CRDA has its own personality. Some CRDAs involve large corporations, such as Martin Marietta or Texas Instruments, whereas others establish partnerships with small start-up businesses. A CRDA may be a supplement to an ongoing contract, allowing the Army to assist in the advancement of the R&D effort, or may be initiated to aid a company that needs help in commercialization under an existing PSD PLA. Some CRDA participants provide new and unique electronic materials to PSD in exchange for test and evaluation of those materials by PSD's sophisticated analysis equipment, facilities, technologies, and techniques.
38
Arthur Ballato and Richard Stern
Some CRDAs represent a "spin-on" to the Army, where a small business wants and needs technical assistance in the development of a new patentable idea that may be useful to the Army as well as the private sector. Several of the CRDAs with academia (Princeton, Rutgers, Stevens, and NJIT) involve placement of graduate students, post-doctoral resident associates, and professors into the PSD laboratory where they have the advantage of working together with PSD S&Es and using state-of-the-art facilities and equipment not readily available in academia. PSD has the advantage of enhancing its work force through the use of these individuals, thereby creating a win-win situation. CRDAs have also been used to formalize working relationships with industrial partners in the creation of collaborative team efforts in carrying out DARPA TRPs (Technology Reinvestment Programs) or other federal agency contract awards. A PSD CRDA consists of approximately 12 pages and can be created in as little time as a few days, with Army formal approval received in fewer than 30 days. In the absence of federal acquisition regulations governing CRDAs, the CRDA contract terms anticipate conflict situations such as potential problems involving data fights and intellectual property. When establishing a CRDA, PSD particularly looks for specific elements, including a mission-related project, a win-win situation, a leveraging opportunity, income, dual-use applicability, and the potential for a follow-on PLA. With these elements, the CRDA is one of the most effective and perfected tools in the PSD technology transfer repertoire.
Patent Licensing PSD has found patent licensing to be a very efficient means of transferring technology and has granted licenses of PSD technology to both small and large companies. A number of valuable lessons have been learned from entering the patent licensing market. The first and probably most important consideration in licensing is the breadth of the patent claims. The quality of patent claims is, in fact, that for which the licensee is paying. Claims that are not well crafted will not prevent others from "patenting around" those claims. Patents are too often rendered useless for licensing because their claims are too narrow. This situation arose from the old mind-set that considered patenting from a purely defensive standpoint: it was only used to prevent others from charging the government royalties for its own research. Sometimes, however, an older patent with typically narrow defensive claims can be rehabilitated by applying for new
1
The Process of Technology Transfer and Commercialization
39
and related patents that expand around (picket fence) the claims and uses of the existing invention/patent. This is part of patent portfolio management. Second, federal laboratories should be sensitive to the commercialization requirement (Code of Federal Regulations 37, Part 404) of any patent license they issue. This regulation requires that the licensee be capable of commercializing and intending to commercialize broadly the government patent in a reasonable time frame. This condition is not difficult for a large corporation to meet. Often, however, when dealing with a small business where this may be a problem, a federal laboratory should lean toward issuing nonexclusive licenses, at least until the licensee can show, either by internal growth or strategic partnerships, the capability of supplying the marketplace with the licensed technology. Further, milestones need to be a part of the patent licence to ensure that the licensee is bringing the technology to the market without delay. Failure to meet milestones can be a clear indication that effective commercialization will not be achieved. In such a case, the licence can either be terminated or converted to a less exclusive mode so that others may have the opportunity to license and advance commercialization. Options for greater exclusivity can be incorporated into nonexclusive licenses to allow a small business license to gain exclusivity at a later date when it is in a better position to prove its ability to commercialize the technology. The PSD Technology Transfer Office usually requires an up-front fee when licensing its technology to business and industry. This fee is charged to partially cover the cost of applying for the patent and the related maintenance fees, and as a measure of good faith of the licensor. Up-front licensing fees charged to small business are usually paid in increments distributed over a period of years. Royalty rates vary with the technology field and can sometimes be partially written off against the initial up-front licensing fee. Knowing that the bottom line is the timely commercialization of the technology, PSD uses flexibility in licensing terms to bring PSD technology to the market as quickly as possible. Each of the patent license issues raised here involves negotiations. Negotiation is a talent that must be developed through education, good legal advice, and experience. This seems to be one of the most difficult aspects of technology transfer with which federal technology transfer offices have to deal. Fortunately, a federal laboratory often has more maneuvering room in negotiating than private sector firms because its priority is technology commercialization.
40
Arthur Ballato and Richard Stern
Technology Marketing There are many means by which PSD can and does market its technology. PSD showcases its technology at technical expos, advertises in trade journals, utilizes technology brokers, distributes booklets describing PSD patents available for licensing, provides inputs to major databases that are broadly available to the public, publishes in the Federal Register, examines citations to its patents that appear in other newly issued patents to expose potential areas of infringement, publishes and presents technical papers at conferences and symposiums, sponsors workshops and seminars, and has its technologists network with their peers in the private sector. By far, PSD's major marketing successes have been achieved through contacts made by its technologists networking with their peers in business, industry, and academia. Most PSD S&Es now operate in an entrepreneurial-like mode in dealing with potential technology users in the private sector. The results indicate that PSD S&Es are highly motivated to bring in new technology partners/users; act as consultants to the private sector; and deliver unique prototype designs, components, and devices to business and industry. These technologists, understanding the benefits of technology marketing, have changed the way PSD does business. Examples of PSD CRDAs pertinent to the area of ceramics include: 9 "Laser Ablation of Ferroelectrir and High-Temperature Superconducting Thin Films," Rutgers University, New Brunswick, NJ. 9 "Development of Smart Ceramic Materials," Rutgers University, New Brunswick, NJ. 9 "Development of Hermetic Coatings for Optical Waveguides," Rutgers University, New Brunswick, NJ. 9 "New Piezoelectric Materials with Application to Frequency Control," RF Monolithics, Dallas, TX. 9 "Development of a Permanent Magnet System for a Microwave Tube," Martin Marietta, Rancho Bernardo, CA. Because PSD has spent several years developing an effective, efficient, and productive technology transfer program, the major effort of the PSD Technology Transfer Office today is on managing the program rather than developing it. Areas such as marketing, licensing, negotiation, program tracking, exploring innovative approaches to technology commercialization, and managing the intellectual portfolio have now become the main tasks of the Technology Transfer Office at PSD.
1
The Process of Technology Transfer and Commercialization
41
The SBIR Program and Technology Transfer The PSD Technology Transfer Office has developed a new and innovative method of promoting the commercialization of its technology: the Small Business Innovation Research (SBIR) program, which addresses the development, licensing, and commercialization of PSD patents. Proposals that are aimed at productizing specific PSD patents that have already been prototyped and demonstrated by Army technologists/inventors are solicited. These prototype designs still require commercially oriented development and optimization, with consideration given to manufacturing and transitioning into actual utilization for both military and private sector applications. SBIR contractors who successfully productize the subject patents are licensed by the Army to produce and market the technology. The SBIR program appears to yield significant advantages. First, the small business is given funding to commercialize a relatively low-risk invention. This productization venture could in turn become the mainstay and contribute to the healthy growth and success of that particular company. Second, those inventions chosen to be commercialized under the SBIR program are handpicked, with strong consideration given to the need for that particular technology in both the military and private sector. Third, this methodology provides for the development of "niche" technologies where large corporations have either no capability or no desire to initiate development efforts and where small businesses have no funding base to pursue the technology development on their own. Lastly, using this approach, Army technology is fully developed, demonstrated, and effectively utilized for commercial applications, with patent licensing fees and royalties being paid to the inventors and the Army laboratory where the invention was conceived. This new commercialization program effort is now in progress under two SBIR contracts and is advancing well. Even before completion of the SBIR contracts, the small business contractors are already producing results and expect to start delivering on orders from the private sector in the near future.
Follow-Up One of the most important facets of awell-oiled technology transfer program is follow-up. Follow-up on personal contacts made at expos, conferences, seminars, etc. leads to new partnering arrangements under CRDAs and PLAs. Tracking milestones, royalty payments, commercialization plans, etc. with respect to existing patent licenses ensures that laboratory technology stays on the road to commercialization. Follow-up on progress and payments (if
42
Arthur Ballato and Richard Stern
required) under existing CRDAs, and follow-up on investigative measures taken to identify possible infringements on patents are equally important. Continuous follow-up in all areas of the technology transfer cannot be overemphasized. Conclusion
Even when a federal laboratory technology transfer program has been successfully developed and is in full operation, one cannot assume that technology and money will flow automatically. Marketing efforts must expand apace and technologists must be continually educated in the value of their technologies and their intellectual property. S&Es must be encouraged to become marketeers, while attorneys must be encouraged to become userfriendly. All participants--legal, technology-transfer, and marketing--must learn to be better communicators to nurture more effectively and harvest more fully the fruits of the strategic alliance that has become our new culture. REFERENCES American Technology Preeminence Act of 1991 (PL 102-245) Army Regulation 70-57, "RD&A Military-Civilian Technology Transfer," 7/91 Bayh-Dole Act of 1980 (PL 96-517) Cooperative Research Act of 1984 (PL 98-462) Defense Authorization Act for FY 1991 (PL 101-510) Exec. Orders 12591 & 12618 (1987): Facilitating Access to Science & Tech Federal Technology Transfer Act of 1986 (PL 99-382) Intermodal Surface Transportation Efficiency Act of 1991 (PL 102-240) Japanese Technical Literature Act of 1986 (PL 99-502) Malcom Baldrige National Quality Improvement Act of 1987 (PL 100-107) National Competitiveness Technology Transfer Act of 1989 (PL 101-189) National Department of Defense Authorization Act for 1993 (PL 102-25) National Department of Defense Authorization Act for FY 1993 (PL 102-484) National Department of Defense Authorization Act for 1994 (PL 103-160) NIST Authorization Act for 1989 (PL 100-519) Omnibus Trade and Competitiveness Act of 1988 (PL 100-418) Small Business Innovation Development Act of 1982 (PL 97-219) Small Business Technology Transfer Act of 1992 (PL 102-564) Stern, R. and Wittig, T. (in press). Technology transfer: Lessons learned--Preparation for the future. In "Proceedings of the Technology Transfer Society Annual Conference." Washington, D.C., July 1995 Stevenson-Wydler Technology Innovation Act of 1980 (PL 96-480) Trademark Clarification Act of 1984 (PL 98-620) Water Resources Development Act of 1988 (PL 100-676)
Fabrication and Characterization of Transducers E M M A N U E L P. PAPADAKIS Quality Systems Concepts, Inc., New Holland, Pennsylvania
CLYDE G. OAKLEY Tetrad Corp., Englewood, Colorado
A L A N R. SELFRIDGE Ultrasonic Devices, Inc., Los Gatos, California
BRUCE MAXFIELD Industrial Sensors, San Leandro, California
I.
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
44 44
B. Types . . . . . . . . M o n o l i t h i c Piezoelectric A. F u n d a m e n t a l s . . . . B. C o n s t r u c t i o n . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Transducers . . . . . . . . ................ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
44 45 45 46
C. B a c k g r o u n d . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D. T h e o r y . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E. E x p e r i m e n t a l M e t h o d s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E S u m m a r y on M o n o l i t h i c Piezoelectrics . . . . . . . . . . . . . . . . . . . . . . . . III. C o m p o s i t e Transducers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
47 48 62 73 76
II.
. . . . Plate . . . . . . . .
IV
A. Introduction to Piezoelectric C o m p o s i t e Transducers . . . . . . . . . . . . . . . . B. S o m e A d d i t i o n a l B a c k g r o u n d on Transducers . . . . . . . . . . . . . . . . . . . . C. C o m p o s i t e F u n d a m e n t a l s . . . . . . . . . . . . . . . . . . . . . . . . . . D. C o n s t r u c t i o n o f C o m p o s i t e s . . . . . . . . . . . . . . . . . . . . . . . . E. C o m m e r c i a l i z a t i o n . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . F. S o m e C o m m e r c i a l i z e d Piezoelectric C o m p o s i t e Products . . . . . . . . . . . . . . P V D F F i l m Transducers . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . .
76 77 83 94 95 98 107
V.
A. P V D F . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. H y d r o p h o n e s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. B r o a d b a n d Sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D. P V D F Air Transducers . . . . . . . . . . . . . . . . . . . .............. E l e c t r o m a g n e t i c A c o u s t i c Transducers ( E M A T s ) . . . . . . . . . . . . . . . . . . . . .
107 107 112 116 118
43 PHYSICAL ACOUSTICS, VOL. XXIV
. . . . . . . . . . . . . . . . . .
Copyright 9 1999 Academic Press All rights of reproduction in any form reserved. ISBN 0-12-477945-X $30.00
l~mmanuel R Papadakis et al.
44
A. Fundamentals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. General Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Cases Being Considered . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VI. Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I.
A.
118 119
122 129 129
Introduction
GENERAL
Ultrasonic transducers have two functions: transmission and reception. Depending on the system and its mission, there may be separate transducers for each function or there may be a single transducer for both functions. A transducer array may be used in either function. By analogy, the transmitter is akin to an audio speaker and the receiver to the human ear. Or, the transmitter is like a radio galaxy and an array receiver is like a phase array of radio telescopes. In transmission, a voltage (or a current) is applied to the output. In reception, a stress wave is sensed by the receiving transducer and an electrical signal is generated for analysis by the system. The circuitry ahead of the transmitting transducer and following the receiving transducer is not the subject of this chapter but will be mentioned as needed in the analysis of transducer behavior. B.
TYPES
Transducers that have reached commercialization can be listed in four categories. These categories are differentiated by materials, structures, and interaction with matter. The categories are as follows: 1. Transducers in which the transmitting element and/or receiving element is a plane parallel plate of a piezoelectric material. These will be termed "monolithic piezoelectric plate transducers." They may have other structural elements incorporated into a functioning device such as plating, electrical connections, backing materials, front layers, cases, etc. 2. Transducers in which the radiating element and/or receiving element is a diced piezoelectric plate with filler between the elements. These are termed "composite transducers" to account for the two disparate elements, the piezoelectric diced into rods and the compliant adhesive filler. 3. Transducers in which the active element is a film of polyvinylidene difluoride (PVDF).
2
Fabrication and Characterization o f Transducers
45
4. Electromagnetic Acoustic Transducers (EMATs). These are current operated, inductive transducers. A coil induces currents in an adjacent metal surface in the presence of a static or quasi-static magnetic field. EMATs can operate on magnetic metals such as steel as well as on nonmagnetic metals. Once generated by an EMAT, an elastic wave behaves just like an elastic wave launched by any transmitting element of identical amplitude, phase, and source diffraction. EMAT generation of elastic waves is different in magnetic and nonmagnetic metals even though the transducers, in some instances, appear to be identical. EMATs almost invariably have a higher insertion loss (lower power efficiency) than piezoelectric transducers generating the same elastic wave. This means that EMATs should only be used when their primary advantages - - couplant-free operation, and the abilityto generate elastic modes that are otherwise difficult--are required by the user. Such applications include couplant-free generation of plate, surface and Lamb waves for high-speed defect detection and for high-temperature (HT) ultrasonic measurements. As an example, if the proper construction materials, bonding techniques, and cooling methods are used, EMATs can easily operate when adjacent to surfaces as high as 1000~ The major intrinsic limitation of EMATs is that the elastic wavelength being generated must be small compared to the electromagnetic skin depth of the radio frequency (rf) currents that are generating the elastic wave. For most metals, a practical upper frequency is in the region of 5 to 20 MHz. These four types of transducers will be explained and analyzed in the remaining sections of this chapter.
II.
A.
Monolithic Piezoelectric Plate Transducers
FUNDAMENTALS
The piezoelectric plates are cut from piezoelectric crystals or are formed from ferroelectric ceramics that are poled (electrically polarized) in the proper directions. The useful cuts and directions are specified for two types of waves, longitudinal and shear (transverse). Longitudinal plates vibrate with particle motion in the thickness direction and generate longitudinal waves propagating normal to their major faces. Shear plates, on the other hand, vibrate with particle motion in one direction in the plane of the major faces and generate
Emmanuel P. Papadakis et al.
46
shear waves also propagating normal to their major faces. To produce ultrasonic beams from such plates, the lateral dimensions must be many wavelengths. For more details conceming piezoelectricity and piezoelectric plates, see Berlincourt et al. (1964), Cady (1946), IEEE (1987), Jaffe and Berlincourt (1965), Jaffe et al. (1971), Mason (1950), Mattiat (1971), and Meeker (1996). Piezoelectricity was first used in sonar in France during World War I. Piezoelectric elements are reciprocal. An applied voltage generates a deflection, and an impinging stress generates a voltage. This physical condition leads to the use of piezoelectric elements, typically plates, as transducers from electrical signals to stress signals (waves) and from stress waves to electrical signals. In other words, the piezoelectric elements can be used as transmitters and receivers for stress waves. Lindsay (1960) has termed this subject of useful stress waves mechanical radiation. In NDT, the term transducers refers to piezoelectric plates with backing and frontal elements to modify their vibration characteristics. These assemblies are potted inside cases to protect them and provide means for gripping them by hand or for mounting them in systems. These potted transducers are sometimes referred to as "search units," although this nomenclature is disappearing from use. Transducers of this type will be treated in this section. Piezoelectric plates many wavelengths in diameter generate beams of ultrasound when they are caused to vibrate by an electric field applied between their electrodes. The beams are not confined to cylinders but spread because of the finite size of the plate source (Roderick and Truell, 1952; Seki, Granato, and Tmell, 1956; Papadakis, 1959, 1963, 1964, 1966, 1971 a, 1972, 1975; Papadakis and Fowler, 1971; Benson and Kiyohara, 1974; ASNT, 1959, 1991). Sometimes the spreading is useful and sometimes it is deleterious. The spreading can be corrected for, sometimes rigorously and sometimes approximately. B.
CONSTRUCTION
The construction of NDT transducers of the most frequently found type is shown in Fig. 1. (However, composite transducers are also finding their way into NDT.) The construction of the transducer includes electrical connections, a case, protective elements (wearplate), and damping elements (backing) as well as the piezoelectric element. For inexpensive mass production, somecomponents are not strictly optimized. The pulser design is generally not optimum, either, from the point of view of being a predictable and indepen-
2
Fabrication and Characterization o f Transducers
l
.
47 I
\T
E
G
j '
B
j
J
J
,,...
B
s
_
-'-QJ
1~.p~
XTAL
"
.................-<-w; ........... (a)
FIG. 1. Transducer construction some designs. XTAL: piezoelectric backings, S: insulating shields, C: connector. FIG. 1(a) is the side, and
WP j
K 1.oo1" AV., OO00" MIN., .0002" MAX.)
(b)
with the components, including the ground strap present in element, P: plating, WP: wear plate, G: ground strap, B: case, T: top cover, HV: high voltage lead, E: electrical FIG. 1(b) is the front.
dent systems component, because its behavior is modified by the electrical characteristics of the transducer on its output.
C.
BACKGROUND
A literature search has demonstrated that the theory of transducer performance is fully known (Redwood, 1963; Sittig, 1972; Papadakis and Meeker, 1969), both in respect to the generation and reception of ultrasonic pulses by loaded piezoelectric elements and in respect to the propagation of ultrasonic pulses in most media. A large body of computational work has yielded results for practical cases. A great variety of measurement methods are available to characterize ultrasonic transducers and their component parts. Efforts have been expended successfully in the communications industry to bring performance (in ultrasonic delay lines) into conformance with theory (Redwood, 1963; Sittig, 1972; Papadakis and Meeker, 1969; Papadakis and Fowler, 1971). A similar effort was carried out in NDT to characterized monolithic piezoelectric plate transducers in the most prevalent NDT configuration (Papadakis, 1983). In that work, recapitulated here, a computer model (Sittig, 1972) developed for the characterization of multilayer ultrasonic structures was adapted and
Emmanuel R Papadakis et al.
48
utilized to analyze NDT transducers. The original computer model could analyze both ultrasonic delay lines in the pulsed mode and resonators in the continuous wave (cw) mode. An NDT transducer in the pulse-echo mode of operation using a fiat reflector is almost identical to the delay line with two transducers, the proper layers and electrical terminations simply needed to be introduced. The electrical and mechanical parameters of the terminations and layers became the variables that could be studied by means of the program. Other parameters such as electrical pulse length and ultrasonic beam spreading were studied also. One limitation of the computer model was an initial assumption of a rectangular input voltage pulse. This was characterized by its spectrum, which was truncated at twice the nominal transducer frequency to prevent aliasing. For damped NDT transducers, this truncation provided no problem. However, the assumption of the rectangular voltage pulse does not correspond to the input from most NDT flaw detection instruments. The rectangular pulse is ideal for driving a transducer, however, and should be adopted. Further, various methods of evaluating the performance of transducers were surveyed. Several experiments utilizing some of these methods were described. The following subsection presents the results of extensive calculations using the computer model as well as the experimental results on transducers. In addition, several evaluation methods in current use are mentioned. Note that some of the methodology for monolithic piezoelectric plate transducers has relevance for other types as well.
D.
THEORY
1.
Transmission Line Approach: Time and Frequency Domains
The theory to be used was formulated by Sittig (1972) for ultrasonic delay lines and resonators. Sittig's theoretical approach can be used directly on transducers to find their loop response in the time and frequency domains. In this theory, the piezoelectric plate is treated as a three-port device with one electrical port and two mechanical ports (its two major faces). (See Fig. 2.) Then, the other components are treated as transmission line elements to find their effect on the waves being propagated. In the case of an NDT or medical transducer, the principal elements are the backing on one side, the piezoelectric plate itself, the wear plate on the other side, and then the propagation medium. One such structure is used to represent transmission, and a second such structure, reversed, is used to represent reception. The piezoelectric plates are represented by Mason's equivalent circuit (Mason, 1948). Electrical
2
b'abrication and Characterization of Transducers
Co V
_!u iT
z~
l
Zo/Y=
Co
49
C
F,C
I
!
1 XDCR J
-4~--U2 0
OF2
XDC.,R IS INSERTED INTO "" BACK
i
o--
U2
V ~ ?~1 U1 XDCR
I:::1 E:::1 E::
I
v
'
~ PLATE~ C
I
PIECE, /
9. TRANSMISSION 9 LINE.
FIG. 2. The incorporation of the equivalent circuit of the piezoelectric plate into a transmission line representing the structure of the transducer and the transmission medium. Sittig's theory takes this approach.
signals are inserted through circuit elements (terminations) and observed across other terminations, as in Fig. 3. For loop response of an NDT or medical transducer, the terminations would be those of the pulser/receiver. The propagation medium could be a liquid or a solid. The theory can also account for thin layers of couplants, bonds, and plating. Sittig wrote a computer program embodying his theoretical work. The author has used this program with minor modifications to calculate the loop response of many transducers, both real and hypothetical. The program uses a unit amplitude electrical input pulse that is rectangular and equal in duration to a half-period of the resonant frequencyJ~ of the unbonded (before bonding) monolithic piezoelectric plate. In the frequency domain calculation, the spectrum is truncated at twice the resonant frequency to eliminate a problem in numerical analysis. The truncation is a fair approximation because at 2j~ the passband has lowered the spectral amplitude much below its value at j~. A typical set of calculations in the frequency domain is plotted in Fig. 4. One finds conductance, susceptance, phase, and transmitted amplitude as functions of frequency. Then the program does an inverse Fourier transform to find the time domain picture of the transmitted pulse as finally observed after the output terminations. Typical calculated pulses of various bandwidths are shown in Figs. 5(a) to 5(e). Typical results show that the center frequency of the pulse observed upon two transductions is lower than the resonance
Emmanuel P. Papadakis et a l .
50
Ri
Li
""i:"Lo R -.o..-r'fT'f~
E
os
DELAY
MEDIUM
Ro
D S
(a) T DELAY MEDIUM
E
|
Co R
Ro _
--(2
(b)
|
FIG. 3. Possible terminations representative of possible test circuits for transducers. For loop response of one transducer, terminations compatible with the pulser/receiver must be used.
.014
~
l
:~.
i
~
l
o,o-, ~,,. ' : / : ~ _ "7
.008 -
I
'
X
I
i
1
\
~_..~
36
~'.oo6 ' ~-"
2 0 0 0 --
I,
\
-!~~176/',,~,,
_l _g,o / .oo,,,~ /
0
0
2
1 4
I 6 f, MHz
FIG.
4.
/
I
1 8
1,, I0
_J
~\\
14,ooo I
,k.____]
12
14
0
-'--
Frequency domain output of the Sittig computer program; a typical case.
2
Fabrication and Characterization o f Transducers
51
I-. e~
z
~. 0.01 z ar 0 Q p..I
a_ 0.00
<
o o
~._~.__L______I-----~s
50
0
1O0
150 200 250 TIME, NANOSECONDS (a)
300
350
400
0.04
~- 0.03 r
z
0.02
p-
~_ 0.01 ,'~ 0.00 0
"--0.01
U.J
~-o.o2 .,I
a--O.03
9
<[
a.-O.04
9
0
o_o.o5 -0.06
0
50
1O0
150 200 250 TIME, NANOSECONDS
300
350
(b)
FIG. 5. A time domain pulse as computed by the Sittig program. (a) k = 0.35, wear plate = 0 thick, terminations resistive and mismatched by 20:1, Zs=Zr=ZM=18.6, f o = 1 2 M H z ; computation 70F5 of Table 1. (b) k=0.35, wearplate=0 thick, terminations restive and mismatched by 1.25:1, ZB=Zr=ZM= 18.6, f o = 12MHz; computation 70F6 of Table 1. (c) k=0.137, wearplate=0 thick, terminations resistive and mismatched by 1.00:1, ZB=0, Zr = 10.2, ZM = 8.29, fo = 30 MHz; computation 69B 1 of Table 1. (d) k = 0.60, wearplate = 2/8 thick, terminations resistive and mismatched by 1.00:1, Zs= 23, Zr= 27, Zwa=40, ZM = 17.3, fo = 10 MHz; computation 60A2 of Table 1. (e) k = 0.60, wearplate = 2/8 thick, terminations resistive Type 1 and 4 and mismatched by 28.1:1, Z s = 2 3 , Z r = 2 7 , Zwa=40, ZM= 17.3,fo= 10MHz; computation 60A1 of Table 1.
Emmanuel P. Papadakis
52
0.0:3
"
'
'
'
I
'
'
'
'
I
'
'
'
;
I.L
z_ 0.02 z
,~
0.01
O u_ D
0.00
~
~ 1 '
'
_J Q.
~
-
~. - 0 . 0 1 O O .J i
-0.02
,
0
,
,
I
5O
i
,
,
,
,
I
1O0
,
,
, .. ,
150
TIME, NANOSECONDS
(c)
,-- o.,----'
1 ,----r
, r--'--r--'--
t
0.3 0.2 ~
0.1
~
0.0
~'0.!
~
-0.2
o-0.3 .J
-11.4
--
0
100
200 300 TIME, NANOSECONDS
400
500
400
500
(d)
Ok.
0.05 ! 0.04
_z
o.o3l
~
o.o2
I,-
~
F
ee 0.01 ' 0 u. 0.00 o-0.01
9
~-0.02
"
a_
~-0.03
9
~. - 0.04
9
O
~-0.05
9
-0.06
9
0
1O0
200 300 TIME, NANOSECONDS (e) FIG.
5.
(continued)
et
al.
2
Fabrication and Characterization of Transducers
53
frequency of the piezoelectric plate and that the apparent wavelength in the pulse is longer than that expected from the nominal plate frequency. These changes are due to the impedance-matched backing that lowers the plate frequency. Specifically, Fig. 4 illustrates the conductance G and the capacitance C of a highly damped transducer nominally cut for 12 MHz. G and C represent the transducer acting as a transmitter attached to termination 1 in Fig. 3. The insertion loss (loop response) is for the transducer performing a double transduction, i.e., transmitting and receiving. Salient features are the highly damped behavior of the resonant element and the pulling of the spectral response of the double transduction, which in turn is far below the nominal 12 MHz frequency. This graph is an extreme example of the general behavior of damped transducers. Calculations have been made for several realizable transducers. The results are shown in Table 1. In each case, the electrical pulse was rectangular and equal in length to a half-period of the nominal resonant frequency of the piezoelectric plate before bonding. The active element diameter was held at 1.26 cm. Attenuation and diffraction were not included. From other calculations, it is known that these inputs change the results minimally for reasonable values of the parameters. Plating thickness, couplant thickness, and bonding layer thickness were all taken as negligible. The symbols used in Table 1 and all subsequent tables are listed here: LMN: Y-Quartz: PZT: k:
Lead metaniobate ceramic Y-cut crystalline quartz Lead zirconate titanate ceramic Electromechanical coupling constant Half-wavelength frequency of piezoelectric element as cut
Z:
Specific acoustic impedances
ZB Specific acoustic impedances for backing ZT: Specific acoustic impedances for transducer material ZWp :
ZM: twp:
(piezoelectric element) Specific acoustic impedances for wear plate Specific acoustic impedances for propagation medium Thickness of wear plate
TABLE 1
SITTINGPROGRAMCALCULATIONS FOR CERTAIN REALIZABLETRANSDUCERCOMBINATIONS Input Parameters
Results Impedances kg/s - m2 x
Line 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
Computation Piezoelectric Number Material 70F6 70F5 70F4 70F2 70A2 69B1 70D1-1 69A 1 69A2 69A3 70G 1 70G2 70G3 70G4 70G5 70G6 70G7
LMN LMN MN LMN LMN Y-Quartz LMN PZT-5 PZT-5 PZT-5 LMN LMN LMN LMN LMN LMN LMN
Termination Type
h k
(MHz)
ZB
0.35 0.35 0.35 0.35 0.35 0.137 0.35 0.60 0.60 0.60 0.35 0.35 0.35 0.35 0.35 0.35 0.35
12 12 12 12 20 30 12 10 10 10 12 12 12 12 12 12 12
18.6 18.6 14.5 -
-
17.0 23.0 23.0 23.0 18.6 18.6 18.6 18.6 18.6 18.6 18.6
Z7- ZWP ZM 18.6 18.6 18.6 18.6 18.6 10.2 19.6 27.0 27.0 27.0 18.6 18.6 18.6 18.6 18.6 18.6 18.6
-
40.0 40.0 40.0 40.0 -
-
18.6 18.6 14.5 18.6 14.5 8.3 1.5 17.3 17.3 17.3 18.6 18.6 18.6 18.6 18.6 18.6 18.6
tw P (mm) Input Output
R
1 1 1 I 1 1 6 1 1 1 1 1 1 1 1 1 1
1.25 20.00 1.25 1.25 1.25 1.00 1.70 28.10 1.00 0.036 5.00 2.00 1.00 0.80 0.50 0.20 0.04
-
-
0.127 0.127 0.127 0.127 -
-
4 4 4 4 4 4 6 4 4 4 4 4 4 4 4 4 4
fM
A,,,
0.057 0.012 0.063 0.150 0.155 0.025 0.012 0.037 0.308 0.039 0.031 0.052 0.055 0.051 0.040 0.019 0.004
dB~,p 25.0 38.4 24.0 16.5 16.2 32.0 38.6 28.6 10.3 31.9 30.4 25.7 25.3 25.8 28.0 34.4 48.0
L
(MHz) (MHz) 8.6 2.6 9.7 10.0 18.2 27.5 7.6 7.3 6.6 6.1 5.0 7.4 9.2 9.7 10.4 11.2 11.4
9.1 4.7 9.8 10.2 18.0 28.5 7.7 7.3 6.7 6.2 6.2 8.2 9.6 10.2 10.8 11.3 11.5
%
&/A7
BW
0.76 95 0.39 167 0.82 79 0.85 68 0.90 49 0.95 41 0.64 86 0.73 25 0.67 39 0.62 26 0.52 129 0.68 102 0.80 90 0.85 84 0.90 83 0.94 80 0.96 79
h 3
I
0
$
k
2
$
5 @
2
2
Fabrication and Characterization of Transducers
55
Input and output:
Code numbers for termination types shown in Fig. 2 and used in the Sittig program. In all the calculations in this paper, L i - " L o - - C i - - C o "-" O . R: Ratio of Ri and Ro of Fig. 2 to the electrical impedance of the clamped capacitance of the piezoelectric element at fo AMAX: Amplitude of the largest half-cycle in the calculated response for a unit voltage pulse z/2 long where z is usually Zo, the period of the nominal frequency
fo. dBLooP:
f~: f~: fc/J0: %BW:
DF:
Loop response (insertion loss) of the transducer, i.e., 201Ogl0(1/AMAx), calculated from AMAX, for two transductions Frequency of maximum response in the calculated spectrum Center frequency midway between the 3 dB points in the calculated spectrum Ratio frequency pulling toward lower frequencies due to backing, coupling, etc. Percentage bandwidth between the 3 dB points in the calculated spectrum, fc is the denominator. Damping factor defined as half the number of halfcycles that exceed in magnitude the first half-cycle of the received signal in the time domain
In what follows, the piezoelectric elements as-poled have been assumed to be one half-wavelength thick. For any piezoelectric element, the actual thickness needed to achieve 2/2 at a frequency f is a function of k, the electromechanical coupling coefficient achieved by poling (Onoe et al., 1963). Because frequency-thickness constants are quoted for the poled condition, their use when cutting piezoelectric plates to thickness will ensure the proper thickness. The following observations can be made from the results in Table 1: 1. Comparisons of Lines 8, 9, and 10 and also Lines 11 through 17 show that the loop response is best (lowest-lossdBLooP, highest-amplitude AMAX) when the ratio R is unity. This means that the input series resistance and the output shunt resistance are equal to the impedance of the clamped capacitance of the piezoelectric element.
Emmanuel P. Papadakis et al.
56
2. Comparison of Lines 5, 6, and 9 shows that the loop response is better for higher electromechanical coupling coefficients. The relationship seems monotonic. 3. Comparison of Lines 1 and 4 shows that a perfectly matched backing will lower the loop response by 8 to 9 dB relative to no backing. The simplifying assumption here was that the transducer was coupled firmly with no wear plate to a matching propagation medium. 4. Comparison of Lines 1 and 3 shows that lowering the backing impedance somewhat (about 22 percent) improves the loop response by only 1 dB and lowers the percentage bandwidth from 95 to 79, still broadband. To answer remaining questions, sets of transducer responses were calculated to determine the effects of varying one or more parameters at a time as follows: (1) acoustic impedance of the backing, (2) electromechanical coupling factor, and/or (3) length of the rectangular input pulse, relative to
Zo12. In the following tables, the nonvaried parameters are listed in footnotes. Results for varying the backing impedance alone are shown in Table 2 for three values of k. The following observations can be made: 1. Loss is maximum when ZB is 1.5 to 2.0 times Zr. 2. The bandwidth is maximum with Zs equal to or slightly less than Zr. 3. As ZB increases, the center frequency falls at an accelerated rate, falling below J~/2 when ZB is in the vicinity of 1.5Zr. This is the known effect of a plate's becoming a quarter-wave source when bonded to a higher impedance load. The wear plate also lowers f~. 4. Bandwidth performance degrades badly when Zs is much less than Zr. For k = 0 . 1 and 0.3, the bandwidth is less than 30 percent by ZB/Zr=0.5; for k = 0 . 6 , the ratio is 0.25. The bandwidth also shows up in the damping factor, which is greater than 2 when the bandwidth is less than 30 percent. A damping factor of 2 indicates four subsequent half-cycles as large as or larger than the first half-cycle in the received signal. In communications work, the bandwidth must be over 55 percent to ensure that the first and fifth half-cycles are low enough for close packing of digital information in delay lines. Results for varying the electromechanical coupling coefficient alone are shown in Table 3, from which the following observations can be made:
TABLE 2 BACKING IMPEDANCE VARIED, WITH OTHER ITEMS CONSTANT a
ZB O. 1
dBLooP
AMAX
fM
40.0
0.0100
Case I: k=O.1 4.35 4.35
1.0
fc
f./f, 0.870
%BW
1.8
--
2.0 4.0 8.0 12.0 16.0 20.0 24.0 28.0 30.0 32.0 48.0 64.0 96.0
45.4 49.0 53.4 56.1 58.2 59.1 59.8 60.4 60.7 61.0 63.1 64.8 62.1
0.0054 0.0035 0.0021 0.0016 0.0012 0.0011 0.0010 0.0010 0.0009 0.0009 0.0007 0.0006 0.0008
O. 1 1.0 2.0 4.0 8.0 12.0 16.0 20.0 24.0 28.0 30.0 32.0 48.0 64.0 96.0
25.5 27.5 -31.8 35.5 37.8 39.4 40.2 40.9 41.6 41.9 42.3 44.5 46.3 42.3
0.0531 0.0422 0.0257 0.0168 0.0129 0.0107 0.0098 0.0090 0.0083 0.0080 0.0077 0.0060 0.0048 0.0077
0.1 1.0 2.0 4.0 8.0 12.0 16.0 20.0 24.0 28.0 32.0 48.0 64.0 96.0
20.7 21.5 22.3 23.8 25.8 27.0 28.1 29.2 30.2 31.1 32.1 33.6 31.0 28.6
0.0923 0.0841 0.0767 0.0646 0.0513 0.0447 0.0394 0.0347 0.0309 0.0279 0.0248 0.0209 0.0282 0.0372
DF
8 8
4.40 4.35 4.40 4.38 4.40 4.38 4.40 4.40 4.40 4.40 4.40 4.40 4.40 4.40 4.40 4.40 4.35 4.40 4.30 4.38 2.35 2.58 2.30 2.32 2.20 2.25 Case II: k = 0 . 3 4.30 4.25 4.30 4.28 4.25 4.30 4.30 4.30 4.30 4.30 4.25 4.20 3.90 2.30 2.20 2.15 Case III: k 3.95 3.95 3.95 4.00 4.00 4.00 4.05 4.05 3.90 2.65 2.15 1.85 1.80 1.80
4.30 4.30 4.30 4.32 4.30 4.32 4.28 4.25 4.22 2.42 2.20 2.15 0.6 3.98 3.95 4.00 4.00 4.00 4.02 4.05 4.00 3.80 3.72 2.92 1.92 1.85 1.82
0.870 0.875 0.875 0.880 0.880 0.880 0.880 0.880 0.880 0.875 0.515 0.465 0.450
4.6 8.0 12.6 18.2 25.0 31.8 40.9 56.8 68.2 83.0 35.2 23.9 15.9
8 8 6 3 1/2 2 1 1/2 1 1 1/2 1/2 1/2 1 1 1/2
0.850 0.855
4.7 5.8
0.860 0.860 0.860 0.865 0.860 0.865 0.855 0.850 0.845 0.485 0.440 0.430
9.3 16.3 23.3 28.9 37.2 49.7 69.0 82.4 98.2 59.8 45.5 32.6
8 8 8 8 4 1/2 2 1/2 2 1 1 1/2 1/2 1/2 1/2 1/2 1 1/2
0.795 0.790 0.800 0.800 0.800 0.805 0.810 0.800 0.760 0.745 0.585 0.385 0.370 0.365
16.4 17.7 20.0 25.0 30.0 38.5 49.4 65.0 94.7 122.1 107.7 59.7 48.5 41.1
5 4 3 1/2 2 1/2 1 1/2 1 1/2 1/2 1/2 1/2 1/2 0 1/2 1
aConstant items: ZT= 32, Z G L U E = 2.93, Z w p - - 38, ZM= 1.5 (water), Glue = 0.0001 in., WP = 2/18 thick = 0.005 in., fo = 5 MHz, terminations 1 and 4, impedance ratio 1:1.
Emmanuel P. Papadakis et al.
58 TABLE 3
COUPLING CONSTANT VARIED, WITH OTHER ITEMS CONSTANTa k
dlLooP
AMAX
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
55.1 43.2 36.6 32.4 29.4 26.5 24.4 23.5 22.9
0.0018 0.0069 0.0148 0.0240 0.0339 0.0473 0.0603 0.0668 0.0716
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
59.1 47.1 40.2 35.4 31.8 29.2 27.6 26.6 23.3
0.0011 0.0044 0.0098 0.0170 0.0257 0.0347 0.0417 0.0468 0.0684
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
60.7 48.8 41.9 37.3 33.8 31.6 30.1 26.8 23.6
0.0009 0.0036 0.0080 0.0136 0.0204 0.0263 0.0313 0.0457 0.0661
fM
fc
Case I: Z B = 10 4.35 4.38 4.35 4.35 4.30 4.30 4.20 4.22 4.15 4.15 4.05 4.02 3.90 3.90 3.70 3.68 3.60 3.00 Case II: ZB = 20 4.40 4.40 4.35 4.38 4.35 4.30 4.25 4.25 4.15 4.15 4.05 4.00 3.80 3.60 1.65 3.18 0.90 1.08 Case III: Z8 = 30 4.35 4.40 4.30 4.35 4.20 4.30 3.95 4.10 3.15 3.95 2.30 3.40 1.80 2.20 1.35 1.52 0.90 1.01
f~/fo
%BW
DF
0.875 0.870 0.860 0.845 0.830 0.805 0.780 0.735 0.600
14.9 16.1 18.6 22.5 26.5 33.5 46.2 66.7 143.3
41/2 4 3 1/2 21/2 2 1 1/2 1 1/2 0
0.880 0.875 0.860 0.850 0.830 0.800 0.720 0.635 0.215
31.8 33.1 37.2 42.4 50.6 65.0 100.0 146.5 107.0
1 1/2 1 1/2 1 1 1/2 1/2 1/2 0 0
0.880 0.870 0.860 0.820 0.790 0.680 0.440 0.305 0.202
68.2 73.6 79.1 97.6 113.9 132.4 100.0 88.5 86.6
1/2 1/2 1/2 1/2 1/2 1/2 0 0 0
aConstant items: Zr = 32, ZGLUE - - 2.93, Zwp = 38, Zm = 1.5 (water), Glue = 0.0001 in., WP = 2/18 thick = 0.005 in., fo = 5 MHz, terminations 1 and 4, impedance ratio 1: 1.
1. The loop response is monotonic in the coupling coefficient, with dBLooP decreasing as k increases. 2. The center frequency fc drops dramatically as k increases beyond 0.6. 3. The percentage bandwidth is high throughout, but not monotonic for Z8 of 20 or higher when Zr= 32. Results of putting in an electrical pulse different from Zo/2 are shown in Table 4. To adapt the Sittig program to give these results, Sittig's t-parameter
2
Fabrication and Characterization of Transducers
59
TABLE 4 PULSES DIFFERENT FROM ONE HALF-PERIOD OF THE PIEZOELECTRIC a Time to Lobe (Units: to)
Pulse Length (Units: %)
First
Second
Time Between (Units: Zo)
dBLooP
AMAX
2 1.5 1.0 0.75 0.50
1.000 0.825 0.650 0.545 0.415
2.000 1.575 1.220 1.070 0.915
1.000 0.750 0.570 0.525 0.500
30.2 30.3 31.6 33.6 37.9
0.0309 0.0305 00263 0.0209 0.0128
aConstant items: ZB = 30, ZT= 32, ZCLUE= 2.93, Zwp = 38, ZM= 1.5 (water), Glue = 0.0001 in., WP = 2/18 thick = 0.005 in., k = 0.6, fo -- 5 MHz, terminations 1 and 4, impedance ratio 1:1.
was made different from 1.0 and the time axis of the response was scaled to make it equivalent to the use of a constant thickness ( t - 1.0) and a different pulse. As explained earlier, Sittig's program uses a t-value to relate the thickness of the piezoelectric plate to the center frequency j~ and pulse length of the electrical input pulse. A value of t--1 implies that the piezoelectric plate has a thickness of 2/2 relative to the center frequency of the input pulse. Two columns in Table 4 are particularly significant: the one labeled "time between" the first and second half-cycles and the one labeled "dBLooP ." The time between lobes is asymptomatic to %/2 as the pulse gets shorter, but the loop response loss increases rapidly for shorter input pulses. For the highly damped transducer with high coupling coefficient represented here, the time between lobes follows the input pulse length for long pulses at least up to 2Zo where the time between lobes is 1.0Zo, i.e., double the expected time of a fleefree high-Q plate. The Sittig program incorporates a provision for approximating the diffraction loss but not the diffraction phase shift in the field of a transducer. There is also a provision for accounting for loss terms that are both linear and quadratic in frequency. These loss terms make small perturbations in the resulting pulse shapes in the time domain, although they introduce a loss into the loop response of a transducer. It was shown previously (Papadakis and Meeker, 1969) that the violent loss peaks and phase fluctuations in multimode guided-wave propagation can have large effects on pulse shape (Carome and Witting, 1961; Del Grosso and McGill, 1970; Papadakis, 1969a; Papadakis, 1969b). To eliminate such effects, flee-field propagation should be used for transducer evaluation.
Emmanuel R Papadakis et al.
60
2.
Pressure in the Field: Space Domains
The above theories do not calculate the pressure as a function of position in the field of the transducer but are essentially plane wave theories with some perturbations to account for diffraction to a greater or lesser degree. The theory for the pressure as a function of position has been calculated for a piston transducer activated by a continuous wave (cw) source (Seki et al., 1956). The Rayleigh integral (Strutt, 1945) was used. Pressure profiles in the field have been reproduced in handbooks (McMaster, 1959). The concentric maxima and minima for cw radiation were predicted much earlier by Schoch (1941). Schoch showed that the radiation of a piston radiator could be decomposed into a plane wave plus a ring wave emanating' from the perimeter of the radiating element. The ring wave interferes with the plane wave to generate the concentric spatial pattern, which is characteristic of the lobes of the transducer radiation pattern as they form up on the near field. The Schoch theory is valid within the right circular cylinder defined by the perimeter of the radiator as directrix. According to Schoch, the strongest maximum in pressure will be the outermost ring of constructive interference at the perimeter of the radiator and in its near field. This is because this ring provides stationary points for the contributions to the radiation field by elements of arc of the source of the ring wave. This situation is illustrated in Fig. 6. The concentric pressure rings were observed by Dehn (1960) in an ingenious photochemical experiment in which the acoustic radiation in a tank of developer was used to nucleate the development of photographic film
(a)
(b)
FIG. 6. Stationarypoints in the Schoch theory. Points E~ and E2 as well as all other points around the perimeter of the transducer T radiate a "ring wave" to all points P in the field area E In view (b), point E1 is a stationarypoint with respect to P, yielding a maximum pressure around the perimeter of E
2
Fabrication and Characterization o f Transducers
61
placed in the tank in the field of the transducer. The outermost ring appeared at the perimeter of the transducer and was the most pronounced. Continuous wave ultrasonic radiation was used. The application to pulses has only come more recently. The author (Papadakis and Fowler, 1971) calculated the field of a broadband transducer on the basis of a superposition of cw results weighted by a hypothetical bandpass. Beaver (1974) calculated the field of a pulsed transducer on the basis of several hypothetical waveforms it might put out. Beaver's calculation integrated over the transducer area to find the resulting waveform and its pressure at the field points. One case of this calculation is shown in Fig. 7. m
1.00
0.50 0.0
V
-0.50 - 1.00
0.O
1.0
2.0
|
....
3.0
4.0
(a) INPUT PULSE 2.0 1.5 1.0
-
,,c_
~
S = ~
z = 12.0),
_
m 0'5 1 " 0.0 / , . I..,,,
=0.48
i a i
I
"
~
-
-
'
_J
-1.s 1.0
z)~
~
0.5~ I o.o !.. 0
, 1
i 2
i i I .I ; ~" 3 4 5 6 7 8 RADIAL DISTANCE
9
10
(b) FIELD PRESSURE PROFILES FIG. 7. A pulse used by Beaver and the resultant field pressure profile at two different distances away from the transducer.
Emmanuel P. Papadakis et al.
62
The input waveform is a modulated sine wave. The resultant pressure profile consists of a central plateau surrounded by maximum and minimum tings, terminating at a strongest maximum near the perimeter of the transducer. This is the behavior to be expected in the near field (Seki et al., 1956; Strutt, 1945; McMaster, 1959; Schoch 1941). The author's superposition calculation indicated similar behavior, but Beaver's tings are probably a better approximation to the actual performance of transducers. E.
EXPERIMENTALMETHODS
1.
Overview
For a complete test of transducer performance, one must know: 9 The time domain response, i.e., the pulse shape and length, which defines the resolving power of the transducer for finding adjacent reflectors. 9 The frequency domain response, giving the bandwidth and the center frequency as well as the amplitude versus frequency. These are related to the damping and apparent wavelength in the pulse time domain picture. Phase information would be useful also, as would the electrical network properties of the transducer as functions of frequency. 9 The spatial distribution of pressure amplitude in the wave field of the transducer. One would want to visualize all three planes, X-Z, Y-Z, and X-Y (parallel to the face of the transducer in what follows). 9 The amplitude and phase of motion of the face of the transducer wearplate. These quantities would determine the pressure in the field by the Rayleigh integral (Seki et al., 1956; Strutt, 1945). Experimental methods for studying these four domains will be listed and explored in the following subsections. 2.
Time Domain Response
The time domain response should be tested with instrumentation capable of reproducing either the theories to be tested or the transducers in their real-use environment. Otherwise, the theories should be worked out to accommodate the available equipment. In the case of transducers and the Sittig program, one would want a rectangular pulsed voltage source or current source with appropriate terminations to generate one of the test systems shown in Fig. 3. The author knows of only one transducer test carried out in this manner for NDT transducers (Papadakis and Fowler, 1971), although the delay line
2
Fabrication and Characterization of Transducers 1
ta
T
1
T
63
~
i
i....3
a. 3[
0
ILl
/
/
/
1,1
-2
0
50
I00 TIME,
150
200
2 50
nsec
FIG. 8. Time domain response of a highly damped transducer with theoretical prediction confirmed by experimental findings.
industry performs such tests routinely. Theory and experiment are compared in Fig. 8. There, the input was a half-period long, and the transducer was highly damped. Quantitative agreement was obtained. In the more usual case, the transducer is tested with a pulser that applies a spike of voltage by discharging a capacitor onto the transducer; the charge trickles off through a parallel resistor that is variable and acts as part of the input impedance to the receiver amplifier in a pulser receiver. This ad hoc system is full of variability and nonoptimum conditions as far as its use for the exact confirmation of theory by experiments. However, such a system can be used for relative measurements and provides the basis for broadband spectra measurements as well as time domain measurements if a spectrum analyzer is also employed.
3.
Frequency Domain Response
Although not as close to ideal as the rectangular voltage pulse, the spike/decay voltage permits a broad spectrum to be applied to the transducer for frequency domain evaluation purposes. With a system including a
Emmanuel R Papadakis et al.
64
~::.~--
.......................
9
t
9.
.
9,...,: .-',"
.. ,~!.'
ii~:. 84184 , :~ ,.;.
~
.
~~: {7~: '~: U . . . .
.
.
.
....~,
. .
.
.
"~41:.
'
iii:'il. :/
.,
...~ . ,~.,)
.,.
.:.
. 9
. {.
| FIG. 9. Time and frequency domain pictures of echoes in viscous liquids. Viscosity increasing top to bottom, causing loss of high-frequency components. Top picture, with very low viscosity, represents the transducer itself.
2
Fabrication and Characterization of Transducers
65
spectrum analyzer, one can find the bandwidth and center frequency of a transducer at a glance. One can the relate the bandwidth to the damping factor defined by the NDT community as half the number of half-cycles in the echo after the first half-cycle larger in amplitude than the first half-cycle. Figure 9 contains time and frequency domain pictures of a pulse propagated through three liquids of different viscosity. It can be seen that the attenuation increasing with frequency tended to lower the spectrum in the more viscous specimens. Typically, one finds that the center frequency of a transducer is lower than that resonance frequency of the piezoelectric plate from which it was made, that the frequency spectrum contains substantial amplitude down toward zero frequency, and that the spectrum dies out substantially by twice the resonance frequency of the piezoelectric plate. These results agree with the Sittig theory. 4.
Space Domain Response
a. General For a complete characterization of the transducer beam, one would like to be able to plot cross sections in space on planes parallel to the transducer face and also to view the beam from the side. If the propagation direction is Z, then X-Y, X-Z, and Y-Z planar plots of the beam are desired. b. Methods for X - Y Plots. Following is an outline of the several methods that are available for obtaining the desired X-Y plots. 1. C-Scan with Ball Target. Use of a commercial C-Scan apparatus with a ball target has been reported by Mansour (1979). In this method, the transducer is scanned over a ball beating in water. A signal is recorded when the echo is greater in amplitude than a preset limit, for instance, 6 dB down from the maximum amplitude attainable over that particular transducer's face. A two-level trace pattem is formed because the recording is either "on" or "off." This pattem actually represents the product of the field strength at the pole of the ball beating and the sensitivity of a small area of the transducer immediately above the ball (see Fig. 10). Because of specular reflection at the surface of the ball, the only information reflected back to the transducer is from the pole of the ball beating. Other wave segments are lost at large cone angles. At the pole of the ball, the reflected wave becomes essentially a spherical wavefront as its retums to the transducer. Upon reaching the transducer, only the portion of this wavefront immediately above the ball is received. The portion beyond a
Emmanuel P. Papadakis et al.
66
certain radius is lost by destructive interference as the spherical wavefronts intersect the transducer surface at phase shifts of 180 ~ per half-wavelength. Beyond the first half- wavelength, all of the remaining wave integrates to zero. In the example shown in Fig. 10, only a -~-in. (0.318-cm) disk-shaped area is effective upon reception, although the field being sampled is radiated by the entire 1/2-in. (1.27-cm) diameter transducer. The implication of this analysis is that one can see a superposition of the wave field of the transducer and the point-by-point sensitivity of the transducer simultaneously. This assertion is borne out in experimental C-Scan observations (see Fig. 11). One can see Beaver's tings, a feature of the radiation of the whole transducer. One can also see a stripe running across the diameter of the transducer face. This is the copper ground strap, ~-in. (0.159-cm) wide and 0.0001-in. (0.000254-cm) '=
,,,
- ~
I''
\r--,. i
_
-f II /, J ,! [i./
DATA: f ~, D h
= = = =
10 MHz 0.006 in. (0.015 cm) 0.50 in. (1.27 cm) 0.60 in. (1.52 cm)
S = D,/a'
.I
v-I xl---v'---I
11~ ~~
D
(b)
(c) RESULT: X : 0.12 in. (0.30 cm) O-X : 0.38 in. (0.97 cm)
-- 0 . 0 0 5 8
FIG. 10. Waves propagating in the C-Scan geometry with the ball reflector. Only the field at the pole of the ball is sampled, and only a small area above the ball performs the sampling because of destructive interference of the outer wave segments.
2
Fabrication and Characterization o f Transducers
67
thick, between the piezoelectric plate and the wear plate. The remaining area is covered with epoxy, nominally 0.0001-in. (0.000254-cm) thick, but possibly wedged. The ground strap is supposed to be bonded to both the piezoelectric plate and the wear plate. (See the construction diagram in Fig. 1.) 2. C-Scan with Microprobe. Posakony (1981) has reported the construction and use of a piezoelectric microprobe for making C-Scan measurements of the field of transducers. The transducer is stationary while the microprobe is scanned through its field. Because this is a throughtransmission measurement, no echoes are involved; only the radiation field is sampled. When piezoelectric microbes become widely available this method will become preferable to the ball-reflection C-Scan method for sampling the transducer radiation field singly. 3. Holography with Flexible Pellicle. Workers at RCA (Mezrich et al., 1974) have reported a holographic system employing laser readout of the motion of a flexible pellicle in the field of the transducer in a liquid. It is not clear at the present time whether this system will be an improvement over the two C-Scan methods mentioned above.
c. Methods f o r X - Z and Y-Z Plots. Side views of the transducer beam can reveal anomalies in is performance resulting in inhomogeneities in the beam. All four optical methods to be outlined suffer the drawback of taking an average through the beam, not a slice. In the present notation, Z is the propagation direction. 1. Schlieren Method. In the Schlieren method (Fitch, 1964; Whaley et al., 1967; Greer and Cross, 1970), a beam of parallel light is sent through a tank of water normal to the ultrasonic beam in the water. The pressure in the ultrasonic beam changes the water density, hence changing the optical index. The light passing through the ultrasonic beam is refracted out of the parallel light beam and misses the beam stop placed at the focus of the condensing lens after the tank. The refracted light is imaged by a camera. With the ultrasonic beam and the light both run on a continuous wave basis, beam pictures such as Fig. 12 can be obtained. One can observe details of the side-lobe structure and the axial zeroes of pressure in the field of a close-to-ideal transducer. On the other hand, poor radiation
Emmanuel P. Papadakis et al.
68
TRANSDUCER C- SCANS
0.60 INCHES IN WATER 0.5 INCH OIAM. ACTIVE ELEMENTS ,,
MFGR.
XDCR.
FREQ., MHZ.
FOCUS, INCHES ,
,
,,,
BEAM PROFILE 3dB
e a
15 15
NONE
15
NONE
6dB
9 ......~
15
3
15
NONE
15
NONE
10
NONE
15
NONE
15
NONE
C
15
NONE
0
* 0.25 INCH DIAM.
r17 ~' ,
Q
9 Q G @
ACTIVE ELEMENT
FIG. 11. C-Scan results on commercial transducers. One can see Beaver's rings predicted by theory. One can also observe the ground strap across the transducer face.
2
Fabrication and Characterization of Transducers
69
FI6. 12. Schlieren picture of the beam from an ultrasonic transducer with both light and ultrasound run clockwise. Details of the side lobes and the axial zeroes of pressure can be detected. (Photo courtesy of J. T. McElroy, Southwest Research Institute.)
patterns can be detected and recorded, as can focused beams and the reflection of beams from obstacles. Pulsed Schlieren systems are also possible (Newman, 1973). With both the ultrasound and the light pulsed and synchronized, the individual cycles of a broadband pulse can be observed. 2. Photoelastic Method. Photoelastic materials that are made optically active by the application of stress can be used in place of the liquid tank. With this substitution, shear waves as well as longitudinal waves
70
Emmanuel R Papadakis et al.
can be made visible. A pulsed optical system has been demonstrated (Wyatt, 1975) for imaging ultrasonic probe beams in solids. The light pulse must be synchronized with the ultrasonic pulse. An example of a broadband pulse image from an angle beam transducer is shown in Fig. 13. 3. Bragg Refraction. As in the Schlieren method, the ultrasonic wave modulates the index of refraction of the transparent propagation medium (Berry, 1966). Maxima and minima in index follow the crests and troughs of the ultrasonic wave. The light sent in from the side is
FIG. 13. Photoelastic picture of a broadband pulse from an angle beam transducer on a transparent solid. Individual wavefronts can be seen with pulsed light/pulsed ultrasound systems. (Photograph courtesy of R. C. Wyatt, Central Electricity Generating Board, U.K.)
2
Fabrication and Characterization of Transducers
71
refracted by the periodically varying index by Bragg refraction as with X-rays, where n2 = 2d sin 0. The first order refraction, n = 1, is observed. The formalism can be set up in terms of conservation of momentum in which the light and sound propagation vectors add. In this way, shear as well as longitudinal waves can be studied by Bragg refraction. For transducer evaluation (Cohen and Gordon, 1965; Dixon, 1967; Cohen, 1967; Dixon, 1970), the parallel light is incident upon the side of the ultrasonic beam at the Bragg angle away from the normal. The receiver is a photomultiplier tube with a lens system whose axis is also at the Bragg angle, so the light is bent through an angle of 20. The optical system can be translated across the ultrasonic beam, tracing out a cross-sectional intensity plot. The side lobes can be studied by rotating the ultrasonic coordinate system through the Bragg angle with respect to the light axis. The angular width of the diffracted light peak is related to the physical width of the ultrasonic beam (Cohen and Gordon, 1965).
4. Liquid Crystal Scanner. Recently, a Dutch firm (Neratoom) has developed a cholesteric liquid crystal scanner (the Neravite| which permits the visualization and recording of the sound field of a transducer radiating into water. The Neravite can make X-Y plots directly at different values of Z and can make X-Z and Y-Z plots by scanning along Z. The output is color coded according to intensity, yielding semiquantitative results.
d. Methods for Evaluating the Transducer Surface Motion. Next is an outline of several methods for evaluating the surface motion of transducers.
1. Scanning Electron Microscope. The scanning electron microscope (SEM) has been used to observe the motion of the surface of piezoelectric resonators used for electromechanical filters (Gerdes and Wagner, 1970, 1971; Hafner, 1974). The method depends on the electric field generated by the stress in the piezoelectric layer as it is forced into oscillation. The electric field modulates the secondary electron emission from the piezoelectric layer in the SEM. The secondary electron emission intensity can be displayed as brightness modulation on the Z-axis (cathode) of the SEM or as vertical displacement on the Y-axis of the SEM, as in Fig. 14. In this experiment on resonators, the specimen was a quartz plate with a ring electrode on each side. A fifth overtone in flexure is shown.
72
Emmanuel R Papadakis et al.
FIG. 14. SEM picture with Y-axis modulation of a resonator. An unplated piezoelectric coating on the wear plate of the transducer might give an analogous result. (Photo courtesy of R. J. Gerdes, Scanatlanta Research.)
It is suggested that this method could be adapted to study the motion of the surfaces of ultrasonic NDT and medical transducers. To accomplish this, a thin piezoelectric layer could be bonded to the wearplate of the transducer or evaporated onto it. This layer would be strained by the motion of the wearplate when the transducer is energized, and the corresponding strain pattern would appear on the SEM. For a piston radiator, the strain pattern would be a ring at the perimeter of the active area of the transducer itself. The ground straps would represent perturbations in the otherwise blank central portion of the pattern. Although NDT and medical transducers have not yet been studied experimentally by this method, it is expected that valuable information could be gained.
2
Fabrication and Characterization of Transducers
73
2. Other Methods. Several other methods could be used to study the motion of the transducer face. a. C-Scan. The C-Scan method with the ball reflector in a liquid (Mansour, 1979) gives an indication of the surface motion of the transducer by showing the receiving sensitivity over a limited area. b. RCA Pellicle Holography. This method could be used to give a picture of the motion of the transducer face by focusing the transducer face on the pellicle by means of an acoustic lens in the liquid bath. c. Laser-Pulse Shocking. A method has been reported (von Gutfeld, 1977) for generating ultrasonic waves in a material covered with a constraining layer by pulsing this layer with laser radiation. This method is a possible candidate for evaluating the sensitivity of transducer surfaces point by point. The wearplate would be the constraining layer, and the active piezoelectric element would be excited readily over small areas comparable with the laser beam diameter. The ultrasonic wave would then dissipate in the backing. The output would be a voltage at the regular cable connection of the transducer. A C-Scan system could be devised to plot the transducer face sensitivity. Care would have to be taken to keep from overheating the transducer.
E
SUMMARYON MONOLITHIC PIEZOELECTRICS
A computer program due to Sittig (1972) and first used to design transducers for ultrasonic delay lines has been used to analyze ultrasonic monolithic piezoelectric transducers. The NDT construction was the principal configuration studied. The piezoelectric principles are the same, and the added layers are analogous, so the delay line analysis can be carried over directly into NDT analysis. The various transducer layers such as couplant, wearplate, adhesive layers, plating layers, piezoelectric plate, and backing are modeled by the computer program. The delay medium becomes the propagation medium in NDT parlance. The NDT engineer is just as interested in information as is the delay line memory or signal processing engineer--information about a flaw, information about velocity, and so on. If the design of transducers and pulser/receivers were approached from the point of view of information, a considerable degree of optimization could be achieved with a savings in power and possibly other parameters. The computer program also gives a theoretical basis of comparison for analyzing transducers experimentally.
74
Emmanuel R Papadakis et al.
Several experimental methods have been summarized for evaluating transducers in four domains--time, frequency, space, and surface motion. The evaluation may be purely empirical to determine whether transducers meet specifications. On the other hand, the evaluation may be for the purpose of comparison with the fifth " domain" - - theory. Theory should be confirmed by the proper set of experiments. Indeed, theory when confirmed by experiment should lead to the writing of realistic specifications. When standards for transducer evaluation are written, it is important to specify an adequate set of tests. The set should have the property of completeness. This does not mean an exhaustive performance of all possible tests but rather a succinct performance of a few tests that yield complete information. For instance, a set composed of the following information would be complete: (1) the terminations and the shape of the input pulse from a definite source specified, (2) the time domain response of the transducer used for two transductions (send and receive) in a definite medium with a specified reflector, and (3) experimental determination of the amplitude and phase of motion of the transducer face (as loaded by the transducer medium) at several frequencies in its band. Given these three items, the other quantities could be calculated. However, these three may not be the most convenient for practical tests. Thus one might want time domain, frequency domain, and three orthogonal planar pictures in the space domina. The latter could be a C-Scan with an acoustic microprobe and two Schlieren pictures, for instance. It is clear that the present tests produced routinely by transducer manufacturers are not a complete set. The terminations and pulsers are not well specified. The resulting time domain and frequency domain pictures, while helpful, are not exact from the point of view of comparing theory with experiment. The use of one or two traverses across the diameter of the transducer to profile its intensity are inadequate when compared with the simplest C-Scan results. The desires of some workers to write very simple test specifications for limited purposes may be commendable, but such desires are short-sighted from the point of view integrating all five "domains" considered in this chapter. To set up a transducer evaluation facility using equipment familiar to most ultrasonics specialists and to the electronics engineers who might be working with them, one would want: 1. Electrical input and output means of definite, known characteristics. These would include a rectangular pulse generator of known type (voltage source or current source) with adjustable amplitude and pulse
2
Fabrication and Characterization o f Transducers
75
length, adjustable known terminations, and an amplifier of known input impedance. 2. Means for making time domain measurements for echo shape and loop response for two transductions in a specified transmission medium, using the electrical apparatus discussed in (1). 3. Means for finding the frequency domain response of the transducer operated as in (1) and (2). This may be done with a gate and a spectrum, analyzer, or with a digitizer and a Fourier computer. 4. A C-Scan commercial system with a ball beating as the target for the transducer and with an ultrasonic system as in (1) and (2) incorporating a gray scale or pseudo-color to plot signal amplitude. Some commercial flaw detection instruments are a practical compromise for this function, but they are not optimized from the point of view of (1) and (2). 5. A Schlieren system for looking at side views of the beams. 6. A network analyzer capable of making single-ended measurements of impedance parameters versus frequency on devices having one electrical port.
FIG. 15. Picture of several commercially available monolithic piezoelectric transducers. (Panametrics, Inc. Used by permission.)
Emmanuel R Papadakis et al.
76
FIG. 16. Some NDT transducers of the monolithic piezoelectric type. (Panametrics, Inc. Used by permission.) An adjunct of (4) would be an ultrasonic microprobe. To be added at a later date would be a system to probe the motion of the surface of the transducer. This might be holographic, SEM, or pulsed laser shock. Illustrations of the appearance of commercial monolithic piezoelectric transducers for various purposes are shown in Figs. 15 and 16.
III. A.
Composite Transducers
INTRODUCTIONTO PIEZOELECTRIC COMPOSITE TRANSDUCERS
Composites form a large and growing class of materials that combine the properties of two or more pure materials to achieve characteristics that are in some sense superior to the characteristic of any of the constitutive materials alone. Piezoelectric composites, a small subset of the broad composite category, are constructed to have piezoelectric and other properties that enhance their ability to be used for transmitters, sensors, or actuators. The piezoelectric composites that have come to have the widest acceptance for
2
Fabrication and Characterization o f Transducers
77
thickness mode ultrasound transducers are 1-3 piezoelectric-polymer composites. These consist of a periodic array of small, finely spaced piezoelectric posts extending through the thickness of the resonator and surrounded on the sides by a polymer. The result is a piezoelectric resonator that maintains the high piezoelectric activity of the ceramic posts in thickness mode but has reduced density due to the presence of the polymer. The soft polymer also tends to isolate one post from another, thus reducing undesirable lateral resonances. For many applications, these features make 1-3 composites superior to any other type of piezoelectric material including single crystals, ceramics, and piezoelectric polymers. Other composite types that have been commercialized will also be discussed below, including 2-2 composites, which consist of alternating strips of ceramic and polymer bonded sided by side, and 0-3 composites, which consist of a piezoelectric powder embedded in a polymer. This section attempts to cover the three areas in which composite transducers have been commercialized. Those areas are naval sonar applications, water-coupled transducers for NDE, and transducers for medical imaging. We have made an attempt to present information on all these applications and their respective histories of commercialization, however, because one of the authors participated only in the development of transducers for medical imaging, that area will be emphasized. This section begins with some transducer information that is important for discussion of piezoelectric composites, including an analysis of the most critical performance parameters for a piezoelectric material. Then composite connectivity is defined and the basic composite properties are illustrated using modeled results. Techniques for making composites are then addressed. The section concludes with information about the commercialization of composites. It includes a table of manufacturers of composites transducers and pictures of some representative products. The techniques for measuring composite materials and composite arrays are the same as those used for ceramics and ceramic arrays. These techniques are covered earlier in this chapter.
B.
SOME ADDITIONAL BACKGROUND ON TRANSDUCERS
Section II of this chapter discusses the structure of transducers comprising a single electrical element, or in some cases, one element for transmitting and one for receiving. Most medical applications, as well as some NDE and naval
78
Emmanuel R Papadakis et al.
applications, use arrays of elements. Since many of the composites of commercial importance are in arrays, array types will be briefly reviewed.
1.
Array Transducers for Medical and Some NDE Applications
Figure 17 shows the kinds of arrays to be discussed in this section. The annular array consists of a central disk and annular rings that are electrically independent. It is typically used with a motor that steers the array in various directions and thus creates a sector-shaped, two-dimensional image, the edges of which are denoted in the figure by dashed lines. The array may be used in conjunction with switches to vary the aperture size and electrical delays which vary the effective focal point. A single focal point must be chosen on transmit, but the focal point may be varied dynamically on receiver to place the best focal point at the depth from which information is being received. The linear sequenced array consists of a line of between 64 and 512 elements. No physical movement is necessary to create a two-dimensional image. Switches form a connection between the imaging electronics and a contiguous set of elements to be used for the creation of a single line in an image. Aperture size and the effective focal point can be varied as in the annular array except that control can be exercised only in the scan plane. A complete image is created by selecting a different set of contiguous elements for each image line. The curved sequenced array works in the same way as a linear array but creates a sector-shaped image that widens with increasing distance from the transducer. The number of elements also varies between 64 and 512. Additional delay time is necessary to focus this kind of array since the beam naturally diverges from the convex aperture. The linear phased (or phased) array is a linear array that differs from the linear sequenced array in that it is shorter in length, has elements at a finer pitch, and is made to work with a delay system capable of steering the beam in various directions to create an image. The phased array differs from the sequenced arrays in that the entire array may be used in the creation of a single display line. Aperture size and focus may be varied as for the linear and curved sequenced arrays. The fifth transducer shown in the figure comprises a two-dimensional array of elements. The figure shows the array as a linear sequenced array, but the principle of operation can be applied to the curved sequenced and linear phased arrays as well. The width of the array is broken into separate elements to allow the aperture size and focal point to be varied in the plane
2
Fabrication and Characterization o f Transducers
79
IIIHIIIIIIIIIIIIlIIIIIIIIIIIIll Annular Array I (Mechanically Steered) I
Linear Sequenced Array ]
[lll[IJlllllUJII1[lJIJIIIItl 11[ rlflrll /
\~'"\
FJHN|llltllulllil ..................
"....,........,.........
[ CurvedSequencedArray ]l CurvedPhased(Steered)Array ]
Illlllll[lllll[ll]llllllll[I
llllllllll[llllllIlllllttl]11111 ' . . . .
' ...........
i
I 1.5DLinear SequencedArray I FIG. 17. Types of arrays used for medical imaging. The top view shows the shape of the elements as viewed from the back. The bottom view shows the arrays from the side view with the beam shape for a single scan line and single focus depth traced out as a solid line. The dotted lines at the edges denote the edges of the scan area.
perpendicular to the image plane (elevation plane). Since the array is not being sequenced or steered in the elevation plane, the width of the elements can be considerably larger in this plane than in the scan plane. Twodimensional arrays with much wider pitch in the elevation plane are often referred to in the literature as 1.5D arrays. For completeness, it should be mentioned that there are also two-dimensional arrays in which the pitch is the same in both directions. These are usually made with a fine enough pitch that the beam can be steered in three dimensions. These 2D arrays are not considered to be good candidates for composites, for reasons discussed below.
Emmanuel R Papadakis et al.
80
The typical noncomposite construction of a linear sequenced array and a linear phased array are shown in Fig. 18. The center-to-center spacing of the electrical elements in the phased array must be kept less than one halfwavelength at center frequency to prevent the formation of grating side lobes at center frequency when the array is steered at an angle. This constraint usually results in an electrical element in which the piezoelectric material is about twice as tall as it is wide. The linear sequenced array has not traditionally been steered (some sequenced arrays are now steered in the more complex imaging systems, but the arrays then must be built more like the linear phased arrays) and the center-to-center spacing has traditionally been much wider. Some early linear sequenced arrays were built by leaving the piezoelectric material continuous and separating the electrode only, but these arrays demonstrated strong interference between elements. Later arrays were built with physical separations isolating the electrical elements only, but the temporal response of these arrays was quite poor because the lateral Active Electrode
.... / /
~
~
~
///
S
~"
~*"
~"
~ - ~ 1 ~ ~ ] ~ ~ ~
Piezoelectric
~" ~
Inner Matching Layer
~
-~-~///jr __~J
~
Matching Layer
Lens for Focusing
Out-of-Plane
Sub-element Electrical Width ElementWidth Linear Sequenced Array . ,
Active Electrode ~
Piezoelectric Inner Material Matching I.ayer
E
Outer Matching Layer Ground lectrode
! ! Electrical Element Width I
FIG. 18.
Lens for Out-ofPlane Focus
Linear Phased (Steered) Array
Typical construction for linear s e q u e n c e d and linear phased arrays.
2
Fabrication and Characterization o f Transducers
81
vibration of the elements coupled and interfered with the desired resonance. Almost all ceramic linear sequenced arrays with a wide center-to-center spacing of the electrical elements are now subdiced into smaller subelements that are electrically connected in parallel. This is significant because the subdiced element is very similar in form to a 2-2 composite. This subdicing is considered by some to be the first commercial application of composites, although it was first applied before the theory and benefit of composites had been thoroughly researched (McKeighen, 1983). It should also be noted that both of the arrays have two layers on the front surface. In the NDE transducers discussed in the first section of this chapter these are referred to as wearplates since their primary function in many NDE applications is the protection of the piezoelectric material. In medical transducers and in water-coupled NDE applications, it is necessary to place layers between the ceramic and the medium that are approximately one quarter-wavelength in thickness at the center frequency and that step gradually from the relatively high specific acoustic impedance of the ceramic to the relatively low specific acoustic impedance of water or the human body. The design of these layers is critical to achieving broad bandwidth and high sensitivity simultaneously. Additional details regarding matching layer design can be found in the references (Kossoff, 1966; Goll and Auld, 1975; DeSilets et al., 1978; Kino and DeSilets, 1979; Kino, 1987). 2.
Critical Parameters in Piezoelectric Materials
To understand the advantages of using piezoelectric composites in transducer design, it is also necessary to understand the characteristics of piezoelectric materials that are most critical for achieving a high level of performance. Here we will discuss the importance of the coupling constant, the specific acoustic impedance, and the dielectric constant or permittivity. The coupling constant of a piezoelectric material is equal to the square root of the fraction of energy converted from the electrical domain to the mechanical domain (or vice versa) in a single electromechanical cycle. For a resonator that has large dimensions and is poled and resonating in the thickness direction, the appropriate coupling is kr, the thickness coupling constant. For a resonator that has small lateral dimensions and is poled and resonating along its length, the appropriate coupling is designated ~3(IEEE, 1978). For a resonator that has one lateral dimension that is small and another lateral dimension that is large compared with the dimension in the poling direction, the coupling for vibration in the poling direction is designated k~
82
Emmanuel R Papadakis et al.
The coupling that applies in a particular transducer or composite is determined by the shape of the piezoelectric resonator and by the poling direction. For PZT-5H 1 type materials, which are often used in medical arrays, kr, ~3 and kw have values of approximately 0.5, 0.75, and 0.7 respectively. These differences are enough to have a significant impact on transducer performance. We will see below that, because of these differences, it is possible to create composites with higher thickness coupling than the piezoceramic of which they are made. Coupling is a critical parameter in that it limits the maximum gain and/or bandwidth that can be achieved in the transducer. If a transducer is heavily loaded (has loss mechanisms, such as a damping resistor or high impedance backing, that absorb much of the energy in the cycle), the low coupling will result in low sensitivity even though the bandwidth may be quite high. If a transducer is lightly loaded, then low coupling will result in low bandwidth. Another important parameter is the specific acoustic impedance of the piezoelectric material. The reflection of energy that is incident normal to an interface is determined entirely by the ratio of the specific acoustic impedances of the materials on each side of the interface. The better the match, the larger the energy transmission. The impedance of a disk of PZT-5H acting in thickness mode is approximately 36 MRayls, whereas the impedance of water is approximately 1.49 MRayls and that of the human body about 1.54 MRayls. Consequently, if a wave is generated in the piezoelectric ceramic and reaches an interface with water, most of the energy is reflected back into the ceramic. If the ceramic is heavily loaded, the reflected energy is absorbed leading to low sensitivity. If the ceramic is lightly loaded, the reflected energy resonates leading to limited bandwidth. The bandwidth and sensitivity can be simultaneously increased by use of matching layers as discussed above. If two matching layers are used with a low impedance backing, bandwidths of approximately 70% can be achieved with insertion losses of less than 6 dB. If a piezoelectric material is available with a lower specific acoustic impedance, broader bandwidths and/or higher sensitivities can be achieved. One of the advantages of composite materials is that they are always lower in specific acoustic impedance than is piezoelectric ceramic alone. The last critical parameter is the permittivity of the piezoelectric material. The clamped permittivity (permittivity of the material when no displacement 1. PZT-5H is a product of Morgan Matroc, Inc., Electro Ceramic Division, 232 Forbes Road, Bedford, Ohio 44146.
2
Fabrication and Characterization of Transducers
83
of the material is allowed) along with the area and thickness determine the clamped capacitance of the transducer element. The capacitive reactance is inversely proportional to the electrical impedance. The electrical impedance is important because the transducer must interact with transmitters and receivers, often through cables. Elements with high electrical impedance require high transmit voltages to transmit a given amount of power and lose a substantial amount of signal on receive when used with cables. Elements with low electrical impedance produce low voltage levels on receive. It is important to match the electrical impedance to the electrical circuitry. The permittivity of piezoelectric composites is always lower than that of piezoelectric ceramic alone. While this may be an advantage of elements with a large radiating area, it is often a limitation for elements with small radiating area, such as the array elements shown in Fig. 18. C.
COMPOSITEFUNDAMENTALS
The first few papers regarding piezoelectric composites appeared in the early 1970s and arose from researchers attempting to make flexible piezoelectric materials (Pauer, 1973). In the late 1970s a comprehensive program of research on piezoelectric composites was funded by the Office of Naval Research and carried out by the Materials Research Laboratory at Penn State. A fundamental paper discussing the importance of connectivity in piezoelectric composites was written by Newnham et al. (1978). Klicker et al. (1981) researched 1-3 composites with round ceramic rods embedded in polymer for hydrophone applications and reported encouraging results for hydrostatic applications. Gururaja et al. (1981) reported on the use of 1-3 composites for thickness mode transducers. At about the same time, Savakas et al. (1981) reported the discovery of the dice-and-fill method of manufacturing. Over the next several years, North American Philips Briarcliff Laboratories, The Pennsylvania State University Materials Research Laboratory, and Stanford Ginzston Laboratory performed extensive research in the use of 13 composites for medical ultrasound transducers. The first paper resulting from the collaboration appeared in 1983 (Auld et al.) and the first comprehensive report was Gururaja's doctoral thesis, which was published in 1984. In late 1984, all the participating laboratories in the collaboration mentioned above and Hitachi Central Laboratories, which had also been researching piezoelectric composites for several years, presented their work at the IEEE Ultrasonic Symposium (Smith et al., 1984; Shaulov et al., 1984; Takeuchi et
Emmanuel R Papadakis et al.
84
al., 1984; Auld and Wang, 1984). Banno (1983) had already reported his work on 0-3 piezoelectric rubber, which had been developed by the NTK Ceramics Division of NGK Spark Plugs during the previous year. Gururaja et al. summarized the early work on 1-3 composites in two papers published in 1985 (1985a, 1985b). An excellent review article of all work prior to 1989, including an extensive bibliography, was presented by at the 1989 Ultrasonics Symposium by Wallace Smith who had directed much of the development work discussed above (Smith, 1989). A summary of composites theory, experimental data, and composite work prior to 1991 can be found in Oakley (1991a). In the thirteen years since the 1-3 composite papers for medical ultrasound transducers were first presented, hundreds of additional papers have been published and presented on composites and composite transducers. A representative list of papers is given in references. 1.
Composite Connectivity
The designation of a composite material as a 1-3, 0-3, or 2-2 composite indicates the "connectivity" of the materials that comprise the composite. Connectivity may be understood by considering a cubic sample of a composite material made of two materials (A and B), which has dimensions that are large relative to the scale of the composite microstructure. If it is possible to find an orientation of the cubic sample relative to the composite symmetry such that a path may be found that enters the sample through one of the sides perpendicular to the X-axis and exits the opposite side without leaving material A, then material A is said to be connected in the X-direction. If material A is connected in only one coordinate direction, then it is said to have a connectivity of' 1'. If an orientation can be found such that the material A is connected in two directions simultaneously, then its connectivity is '2'. Each material in a composite will have a connectivity ranging from 0 to 3. A composite with two components may be classified by the connectivity of each of the components. The connectivity for 0-3, 1-3, and 2-2 composites is shown in Fig. 19. The direction of polarization for the piezoelectric composites used in commercial transducers is always along the Z-axis in the figure, but it should be recalled that the connectivity designation is, in general, independent of polarization. The most common form of the 0-3 composite consists of a piezoelectric ceramic powder embedded in a continuous polymer matrix. (Note that the connectivity of the piezoelectrically active material is usually given as the first number, but this is not universal.) The 2-2 composite consists of piezoelectrically active plates alternately stacked with polymer
2
Fabrication and Characterization of Transducers )~
Z
Z
~ ,i~g Oo* 0e oo00 D ~ ~
~
85 Z
Y
X
I O-3Composite J
I 2-2Composite ]
1-3Composite
FIG. 19. Typical geometries for composites with 0-3, 2-2, and 1-3 connectivity. The poling direction should be assumed to be along the Z-axis.
plates. The 1-3 composites usually consists of piezoelectrically active isolated pillars surrounded on the pillar sides by a continuous polymer. Also of some importance, but not shown, is the 3-3 composite connectivity in which both materials are continuously connected in three dimensions, similar to the structure of an open cell foam. The importance of connectivity is that it governs which of the material's properties (electrical, mechanical, and piezoelectric) will dominate the properties in a particular direction. This is determined by whether material properties tend to sum in series or in parallel and is thoroughly discussed in Newnham's original paper (Newnham et al., 1978).
2.
Composite Modeling
There are several methods of modeling composite performance. When the scale of the composite is small with respect to a wavelength, the material can be characterized by calculating effective properties. There are several methods for performing these calculations (Hashimoto and Yamaguchi, 1986; Smith and Auld, 1991; Shui et al., 1995). These effective properties can be used in a 1D transducer model, as discussed in the section II.D. 1. Many significant features of composites related to the lateral mode properties cannot be characterized by using effective properties. Piezoceramics used as a plate in thickness mode also have resonances and harmonics or overtones associated with the lateral plate dimensions. In a well-designed piezoelectric composite, these lateral modes can be more readily damped, reduced in frequency, and made with lower coupling constants, thus reducing interference due to these modes. However, regular piezoelectric composite structures
Emmanuel R Papadakis et al.
86
(1-3 and 2-2 composites) possess high-frequency resonances that are not present in piezoelectric ceramic resonators. A substantial amount of modeling has been done to characterize these lateral modes (Auld et al., 1983; Auld et al., 1984; Alippi et al., 1988a; Alippi et al., 1988b; Oakley et al., 1990; Oakley, 199 lb; Geng and Zhang, 1997). Finite element modeling can be used to predict effective properties, lateral mode vibrations, and full transducer performance (Lerch, 1990; Hossack and Hayward, 1991; Wojcik et al., 1993; Hayward et al., 1995). Recent work has demonstrated good agreement between FEA predictions and composite displacements as measured using laser interferometry (Reynolds et al., 1996). Recent theoretical work has also combined the prediction of effective properties with the prediction of lateral modes (Shui et al., 1995; Geng and Zhang, 1997).
3.
Composite Properties
A summary of some important effective composite properties calculated using a modified version of the Hashimoto and Yamaguchi (1986) method are given in Figs. 20 through 22. Figure 20 shows the specific acoustic impedance of the three composite types. Specific acoustic impedance can be calculated as the square root of the product of density and an appropriate elastic stiffness constant. The density is the dominate term and varies linearly as a function of volume fraction from the density of the polymer to the density of the piezoelectric ceramic for all the composite types. The variations in effective stiffness for the various composite types explains the differences among the curves. For 2-2 and 1-3 composites the material acts (in the Z direction) more or less like stiff springs (the ceramic) in parallel with compliant springs (the polymer) that are displaced by the same amount. The effective spring constant can be thought of as the material stiffness times the cross-sectional area perpendicular to the displacement. When the material is uniformly compressed, it is the stiff spring that dominates the compressibility. The composite stiffness, in the Z direction, builds rapidly with volume fraction, since the relatively stiff ceramic makes the composite difficult to compress even when the volume fractions are low. The stiffness of the 0-3 composite, on the other hand, builds up quite slowly with volume fraction since this material acts more like springs in series. When this material is compressed, it is the compliant spring (the polymer) that dominates the compressibility. The stiffness of the polymer tends to dominate until the density of participles is
2
Fabrication and Characterization o f Transducers
87
40
A
35
w
/
>,~ m
r 30 :E ......
o
r
,- 25
r
"o
Q
Q.
0-3 Comp .4-1-3 Comp -,=.9 2-2 Comp
E 20 ._(2 ~ <
15 J
,I"
.~- 10 D.
0
0.1
0.2
J
0.3
...,., v "
0.4
0.5
0.6
0.7
0.8
0.9
1
Ceramic Volume Fraction
F~G. 20. Specific acoustic impedance as a function of ceramic volume fractions for 0-3, 2-2, and 1-3 composites.
1400 1200
///
"g 1000 .=_ ...
E o
~.
800
o D.
E
m o o ._ m Q
/-
600 400
lg
200
J 0
/
/
/
,/
/ x..,,, , /
/ / 0.1
0.2
0.3
0.4
I ~ 0-3 Comp 1-3 Comp 2-2 Comp
=
0.5
~
0.6
__...4p-- ~ v''"
0.7
0.8
0.9
1
Ceramic Volume Fraction
FIG. 21. Dielectric constant (relative clamped permittivity) as a function of ceramic volume fraction for 0-3, 2-2, and 1-3 composites.
Emmanuel R Papadakis et al.
88 0.80 f
0.70
~
._.-=
:
=
,._ 0.60 0
U
u.
I
0.50
.== =.= D.
= 0
0.40
o
el r
9 0.30
._~
-,--0-3 C o m p
0.20
,,,,,
0.10 0.00
-
0
9
-
0.1
~,~._..., ~
0.2
0.3
/
-g- 2-2 Comp
1-3 C o m p
t
-----'*'-~ " ' " * " ~
0.4 Ceramic
0.5 Volume
0.6
0.7
0.8
0.9
1
Fraction
FIG. 22. Thickness coupling factor as a function of ceramic volume fraction for 0-3, 2-2, and 1-3 composites.
quite high. This explains why the 0-3 impedance curve falls far below the curve of the other two ceramics. Figure 21 shows the dielectric properties for the three types of composites under discussion. The ceramic and polymer in 1-3 and 2-2 composites act (in the Z direction) like capacitors in parallel, meaning that the piezoelectric ceramic, which has a dielectric constant many hundreds of times higher than that of the polymer, will dominate the behavior over a large range of volume fractions. It behaves like a large capacitor in parallel with a small polymer capacitor. Consequently, the effective dielectric constant is nearly proportional to the volume of ceramic in the composite. The 0-3 composite, on the other hand, is not connected in the 3 direction so that the ceramic acts like a capacitor in series with the polymer. This means that the low dielectric constant of the polymer dominates over much of the volume fraction and the dielectric constant of the ceramic only becomes significant when the ceramic particles are dense enough to be in physical contact with each other. Figure 22 shows the thickness coupling factor for 0-3, 1-3, and 2-2 composites as a function of ceramic volume fraction. Several major features are evident. Note that over a large range of volume fractions 1-3 and 2-2 composites achieve a higher thickness coupling than does the piezoelectric ceramic alone (where volume fraction equals 1). The 1-3 composite
2
Fabrication and Characterization o f Transducers
89
approaches the 0.75 coupling factor achievable in a long post. The 2-2 composite approaches the 0.7 coupling factor achievable in a strip. For both the 1-3 and 2-2 composites, the coupling drops at low volume fraction. The coupling characteristics can be understood by considering where the energy concentration is for the mechanical and electric fields. To achieve high coupling in the composite, it is necessary that both the mechanical energy and the electrical energy be concentrated in the piezoelectrically active material (the ceramic) and that the material be capable of effectively converting one form to the other. The drop in coupling at low volume fractions for the 1-3 and 2-2 composites can be understood from the perspective of mechanical energy conservation by again thinking of the material as springs in parallel. At low volume fractions, the effective spring constant for the ceramic is low because of the low cross-sectional area that the ceramic occupies. Even though the ceramic is stiffer than the polymer, the polymer holds more energy because there is so much of it. Energy in the polymer (represented by the compliant strings) cannot be converted from electrical to mechanical form thus limiting the coupling. The coupling at high volume fractions drops off in 1-3 and 2-2 composites because the close crowding of the ceramic regions begins to limit the lateral movement of the ceramic portions, making it more like a solid block of ceramic. A similar argument exists for the energy concentration in electric field. Although the regions of high permittivity (the ceramic) contain more electric energy per area, at low volume fractions the area is so small that more electrical energy is contained in the polymer regions. The coupling for the 0-3 composite, also shown in Fig. 22, is much lower than that of the 2-2 and 1-3 composites because it is not continuous through the resonator. Its coupling stays low through most of the volume fraction range and only begins to become significant when the volume fraction approaches 1.0. Even at 90% volume fraction the coupling is still only about two-thirds that of a ceramic plate. This again can be understood by considering the energy contained in springs and capacitors in series. Neither the electrical nor the mechanical energy resident in the composite resides primarily in the ceramic where it can be converted from one form to another. This is one of the major reasons that 0-3 composites are seldom used in medical and NDE transducers. The literature demonstrates that the choice of polymers and ceramics with different properties can result in substantial quantitative differences among these curves (Oakley et al., 1990; Geng and Zhang, 1997). The general shapes, however, remain much the same.
Emmanuel R Papadakis et al.
90
4.
The Effect of Composite Parameters on Transducer Performance
The variations of the parameters discussed above result in substantial variations in transducer performance. A one-dimensional model (Krimholtz et al., 1970; Kino, 1987; Oakley 1997) has been used to predict the bandwidth and sensitivity (the peak-to-peak amplitude of the received broadband signal divided by the pulser voltage) for a single array element with an area of 30 square wavelengths. The array elements were thickness mode transducers, with two matching layers, incorporating various volume fractions of the three composite types under discussion. The results are shown in Fig. 23 in the form of a gain-bandwidth product. Three of the curves were generated by matching the source impedance and the load impedance to the impedance of the transducer. The fourth curve was generated by keeping the source and load impedance at 50ohms, which would be typical of many systems using 50-ohm cables. The highest gain-bandwidth product is achieved by the 1-3 composite. Note that at its peak it is over three times the product for pure ceramic (100% volume fraction). This occurs because of the increased coupling constant and
6
0.35
- //
0.3 q o
e=
0.25 o
,-. 4 ,9.0
3
o 1= "o
.o 3
=
,-
r m
--~ 1-3 Comp --~2-2Comp
O.15 ~
~ 0-3 Comp
~
I
'1o
"~ 0.2 ~
2
1-3 Comp Ze = 50 ohms
II1
0.1 ~
.g
m
0
'
0
'
..jd"" ,,..,.>
0.1
0.2
0.3
f
0.05
9
0.4
0.5
0.6
0.7
0.8
0.9
,0 1
Ceramic Volume Fraction
FIG. 23. The modeled gain-bandwidth product as a function of ceramic volume faction assuming a double matching layer that is scaled in specific acoustic impedance. The element area is held constant at 30 square wavelengths. The electrical source and load impedances are matched to the transducer except as marked.
2
Fabrication and Characterization o f Transducers
91
the low specific acoustic impedance. The 2-2 composite is nearly as good as the 1-3 composite. Note that the performance of transducers made from widerange 1-3 and 2-2 composite volume fractions is better than that of transducers made from ceramic plates. The gain-bandwidth product of the 0-3 material, even at its highest level, is an order of magnitude lower than that of the 1-3 composite. Over most of the volume fraction range, the product for the 0-3 is several orders of magnitude lower than that of the 1-3. This does not imply that 0-3 composites do not have value. For naval applications, for example, the low cost for materials covering large surface areas and the conformability to various shapes is very important and the performance is acceptable. The curve in the figure labeled "1-3 Comp Z e = 50ohms" shows the importance of the dielectric constant and of matching the electrical impedance of the transducer to that of the electrical circuits. It shows the gain-bandwidth product for a 1-3 composite element with an electrical impedance much higher than the 50-ohm source and load impedances throughout the volume fraction range. This is typically true in sequenced and phased array elements used for medical imaging. Note that the gain-bandwidth product for the 50ohm source and load is much lower because of the electrical mismatch. It is also interesting that the best volume fraction for this application is approximately 80%. This demonstrates that modeling is very important in determining the optimum composite design. A different assumption on the size of the elements or electrical impedances would change the location of the gainbandwidth peak. The importance of the dielectric constant and electrical impedance of the element has played an important role in the types of transducers on which composite materials have had the most impact; this will be further discussed below. 5.
Lateral Modes in Composites
Some understanding of the higher-frequency lateral vibration modes of composites is essential for the design of 1-3 and 2-2 composites for thickness mode transducers. However, a detailed discussion is beyond the scope of this section. The benefit of using a 1-3 composite in transducers with lateral dimensions that are only a little wider than the thickness can be seen in Figs. 24 and 25. In Fig. 24, the electrical impedance of a square ceramic plate with lateral dimensions 10 times thicker the thickness is shown. Note that many lateral modes and overtones, indicated b y the dips and peaks in magnitude and the peaks in phase, are present at low frequency. These reduce in intensity
92
Emmanuel R Papadakis et al. 100 80
3.5
E o
.1=
3
~i~
!ii
g
6o
'
'
40
~
20
-~
:
0
~
i!
-20
a.
"~" 2.5 ~
gm
2
x |
"
=u 1.5
& ...-E
I :i
1
!i: .,
:: i;
iii ii
J
-40
[..... Ma~176 I Phase Angle
0
C
m
0
i!
~i
i
-60 _E
i
,. ,,-,,.__.....~ -8o
',,_p ,,,_., ,_.,; ,J ' , A _~,., rX.': "
-100 0
0.5
1
1.5
2
2.5
3
Frequency*Thickness (MHz*mm)
FIG. 24. Low-frequency lateral modes in a ceramic plate shown as the electrical impedance (magnitude and phase angle) of a ceramic square with lateral dimensions 10 times the thickness dimension.
5
80
4.5 E J=
60
r",9!
,,,-,,,
4
40
0
',A "0 =I
:
3.5
9
:
r
0
~
-20
~
O
2
.------Magnitude . . . . . P h a s e Angle
I1.
-8 ~. 1.5
0 =..,
20
3
o~ 2.5
E n
.o 0}
e-
8 r
A
-40 ,.
-60
1
.i
0.5
j. i
~
C 'O
&
E
-80
',,,.~
-100 0
0.5
1
1.5
2
2.5
3
Frequency*Thickness (MHz*mm)
FIG. 25. Low-frequency lateral modes in a ceramic plate shown as the electrical impedance (magnitude and phase angle) of a 1-3 composite square with lateral dimension 10 times the thickness dimension. Note that the low-frequency lateral modes are much less noticeable than they are in FIG. 24.
2
Fabrication and Characterization of Transducers
93
as the frequency increases but are still significant enough to interfere with the thickness mode resonance seen between 2 and 2.5 MHz. In Fig. 25, the electrical impedance of a square plate of 1-3 composite is shown. Here the lateral modes die out at much lower frequencies and do not interfere with the thickness mode. In a transducer, this results in a smoother spectrum and quicker ring-down. The reasons for this are that in composites the lateral coupling factors are lower, the mechanical damping for lateral modes is greater, and the periodicity of the structure results in bands of frequencies in which lateral propagation is highly attenuated due to constructive and destructive interference of scattered sound. At higher frequencies the constructive and destructive interference caused by reflections from the periodic ceramic-polymer interfaces produces resonances, some of which are coupled to the thickness vibration and can be seen electrically. These can be seen in the electrical impedance plot of a 2-2 composite shown in Fig. 26. One of the high-frequency resonances occurs when the polymer and ceramic vibrate 180 ~ out of phase in the thickness mode. The resonance frequency associated with this mode may be roughly approximated by assuming that the largest lateral polymer dimension between posts or strips is approximately one half-wavelength, where the wavelength is calculated using the shear wave velocity of the polymer. It is critical to design the composite with a fine enough periodicity that these resonances are above the usable frequency passband of the transducer being designed. 4 .-- 3.5
E
--9
3
F:esonance due to oeriodicil:y Ceramic' k
80
60
2
~
1 8c - 1.5
20
l'hicknes~Mode ~ Third Harmonic
o
N
Magnitude o ~ -20 ix. ~ ...... P h a s e u
-40 E
80
Reson,3nce due to perio(licity ~ \ (Polyrler)
/!\
~ I=
a.
A
r
~" 2.5 =
100
Thk;kness M ~de Ftlndamen al
1
:
:1 ',;
;
D.
;
-60 E ,=,,
0.5
'
t
!\
-80 -100
0
2
4
6
8
10
12
14
Frequency (MHz)
FIG. 26. Resonance of a 2-2 composite showing the high-frequency modes resulting form the periodicity of the composite.
Emmanuel R Papadakis et al.
94 1.6
\ i
1.4 o
,m
\
n,, 1.2 Q
c
\
1
\
\i
t ~
\/ IL
1-3 Ceramic Limit 1-3 Polymer Limit ..... 2-2 Ceramic Limit 2-2 Polymer Limit
0 ,m
0
.E 0.6 0 r Q.
.." ...'""" ~ ,.2.21J ~
0.4
J
0.2
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
Ceramic Volume Fraction
FIG. 27. Center-to-center spacing requirements for 1-3 and 2-2 composites as a function of volume fraction for a composite made from PZT-5H and a moderately stiff epoxy. The lighter lines at the left represent the limitations set primarily by the polymer shear properties, and the heavier lines at the right represent the limitations set by the lateral longitudinal velocity of the ceramic.
Another resonance occurs when the lateral displacement of the ceramic and polymer are 180 ~ out of phase. The resonance frequency associated with this mode may be roughly approximated by assuming that the lateral ceramic dimension is one half-wavelength long where the wavelength is calculated using the lateral longitudinal wave velocity of the ceramic (Oakley 1991b; Geng and Zhang, 1997). A rule of thumb often followed in the medical and NDE transducer industries is to place the lowest high-frequency lateral resonance at about twice the nominal center frequency of operation. Figure 27 shows how the center-to-center spacing must be varied as a function of ceramic volume fraction to avoid interference from unwanted lateral resonances arising either from the ceramic or the polymer. D.
CONSTRUCTION OF COMPOSITES
Composite materials may be made in a variety of ways. 0-3 composites are almost always made by mixing piezoelectric powder into a thermoset polymer (Banno, 1989). When the polymer cures, the powder particles are held in place. Early 2-2 and 1-3 composites were made by preparing the ceramic
2
Fabrication and Characterization o f Transducers
95
strips or posts in advance, holding them in the proper locations and potting around them with polymer (Klicker et al., 1981). 2-2 composites were also made by laminating prepared ceramic and polymer plates together. An altemative approach to making 1-3 composites is similar. 2-2 composites are cut into plates and them laminated alternately with polymer plates (Zola, 1985; Zola, 1986). From the time of its development, the 'dice-and-fill' method (Savakas et al., 1981) has been the primary method for making 2-2 and 1-3 composites both for research and for commercial development. This process consists of dicing a ceramic plate in one (for 2-2) or two (for 1-3) directions such that the base of the ceramic remains continuous. Polymer is poured into the kerfs and cured. Then the excess polymer is lapped off the top and the excess ceramic is lapped off the bottom, leaving the desired structures. More recently, a method has been reported for injection modeling of 1-3 and 2-2 composites (Bowen et al., 1993; Gentilman et al., 1995; Pazol et al., 1996). This is done by heavily loading a thermoplastic binder with ceramic particles, injection molding the heated mixture into a negative of the desired composite, burning out the binder, and sintering the resulting structure into a ceramic preform that resembles the ceramic from the dice-and-fill method after the dicing. The remaining steps are the same as those used for dice-andfill process. Injection molding has the advantage that large areas of material can be made at relatively low cost, which is critical for many naval applications. Injection molding may also reduce cost for medical and NDE composites as well. This material has now been commercialized and is being offered for sale. Other advanced methods are still being investigated (Lubitz et al., 1993; Janas and Safari, 1995), but we are not aware of any commercial company offering composites made by methods other than dice-and-fill and injection molding. E.
COMMERCIALIZATION
After the research and development work was reported in 1984, a substantial amount of effort went into the commercial development and design of composites. Since one of the stated objectives of this volume is to discuss the time and effort required to bring ideas into commercialization, we will present as anecdotal evidence the history of commercialization of 1-3 composites at Echo Ultrasound (now ATL-Echo) where one of the authors was then employed.
96
Emmanuel P. Papadakis et al.
Intemal development of 1-3 composites transducers began at Echo Ultrasound almost immediately after the 1984 Ultrasonics Symposium. Development work began by contacting the traditional ceramic suppliers and requesting that they start to supply composite materials made to the design rules set forth in the early papers. When no supplier could be located, Echo contacted researchers at North American Philips Briarcliff Laboratories to propose a joint effort, but the proposal was not accepted. Having failed to find an external supplier, Echo began a modest program to evaluate 1-3 composites. Within a few months, Echo was making prototype quantities of 1-3 composite material that worked well. By late 1986, singleelement transducers were being manufactured that were higher in sensitivity and wider in bandwidth than those being made with ceramic materials alone. However, progress was hindered by internal skepticism about the potential of composite material. This skepticism was fed by concerns similar to those that face any new technology. One concern centered on the cost of processing composite material compared with the level of improvement being observed and on the cost of converting to a more difficult technology. Another concern centered on the fact that no other company had commercialized the material. The existing market did not require that the new technology be used to be competitive, and there was some fear that the failure of competitors to introduce the technology might be the results of some unknown obstacle to commercialization. Sufficient support was obtained to show the composite transducer performance to some of Echo's customers. The first commercial customer began ordering prototype quantities in 1986. A substantial amount of processing work was required to manufacture these early units and substantial reliability problems were solved through an aggressive program of dissection and analysis. Encouraged by the results, the technical group at Echo was anxious to move the technology into annular arrays, but the difficulties that had been encountered in supplying the first prototypes had damped the enthusiasm of some of the other influential decision makers. The motivation to pursue composite annular arrays aggressively did not come from inside Echo, but from the report that a competitor had introduced composite annular arrays that worked substantially better than ceramic annular array. Although it was soon discovered that the competitive arrays were made from 2-2 composites instead of 1-3 composites, management was now almost universally in favor of pursuing 1-3 composite annular arrays with the hope of retaining a large share of the annular array market.
2
Fabrication and Characterization o f Transducers
97
The existence of a sister company in need of annular arrays with superior performance, and the performance achieved in Echo's early prototypes provided the motivation for the completion of an aggressive program of manufacturing 1-3 composite annular arrays. Over the next few years, Echo put into place design programs such that all designers could design using customized composite materials and manufacturing processes capable of supplying composite arrays in quantity. Although Echo had done a substantial amount of prototyping earlier, the first commercial shipment of composite transducers, as defined in this volume, was made in early 1988, culminating in 31/2 years of internal development and design. Competition was quick to develop, with the introduction of composite single elements and annular arrays by Precision Acoustic Devices in the United States and by Vermon in France. Sound Technology, founded in 1987 by former Echo Labs employees and now a. subsidiary of Acuson, also continued to develop and commercialize composite transducers. Commercialization of 0-3 composites predated the commercialization of 1-3 composites by several years. Under the technical direction of Dr. Banno, NDK introduced Piezorubber some time prior to 1981. This was purchased for and still continues to be used in some naval applications (Ting, 1986). Single-element and annular arrays were a natural first place for 1-3 composites to be applied (Smith, 1989). The increased coupling and lower acoustic impedance resulted in the achievement of boarder bandwidth and, in some cases, additional sensitivity. In annular arrays, the effects were even more dramatic since the outer array elements of equal area arrays have widths that are just a few times wider than their thickness. When built with traditional PZT ceramics, the lateral vibrations results in poor performance of these elements. It is possible to use lead titanate that has been modified to have very low lateral coupling but this is usually done at the expense of bandwidth and sensitivity. Linear and curved sequenced arrays with a 2-2 composite-like structure have been created at least since the early 1980s (McKeighen, 1983). Single electrical elements had widths ranging from 1 to 3 times their thickness which resulted in the ceramic lateral mode interference discussed above. The problem was resolved by subdicing the electrical elements into smaller subelements. This produces a structure with alternating plates and polymers and produces some of the benefits of a 2-2 composite. For example, the coupling is increased to that of k~3. Whether or not the acoustic impedance is decreased depends on the properties of the polymer filler. Often a filler is used
98
Emmanuel R Papadakis et al.
that isolates one subelement from the next so that the electrical element does not vibrate as a unit but each subelement vibrates alone in phase with its neighbor. Because of such subtle differences, there is little agreement on whether such a structure should be referred to as a 2-2 composite array. For the purposes of this section, subdicing is considered to a 2-2 composite in Table 5, which lists companies offering composites or composite transducers for commercial sale. However, many companies who subdice their arrays do not consider them to be composites and are not listed. Composites with a 1-3 structure have also been commercialized in linear and curved sequenced arrays. However, for these arrays, there are some problems as well as some benefits when composites are used. The increased coupling and lower specific acoustic impedance result in broader bandwidths being achieved. However, the increase in coupling of a 1-3 composite over a subdiced array is modest, and the element areas in these arrays are much smaller than the areas for single-element and annular array transducers. The lower dielectric constant of the composite materials results in high electrical impedances, which in turn results in a loss of sensitivity when loaded with a resistance near the 50ohms required for effective use of a coax cable. Nonetheless, for processing reasons 1-3 composite linear and curved sequenced arrays have now become quite common. Linear phased array have elements that are already quite narrow. Consequently, subdicing the element is unnecessary. There might be some benefit in using a composite structure in the elevation dimension, but the electrical impedances of phased array elements are already high and use of a composite in elevation would increase the electrical impedance even further. Two-dimensional arrays that are phased in both directions must have elements that consist of small posts. They are built like a 1-3 composite that has a separate electrical connection to each ceramic post. Although it is not impossible to divide these posts up into a finer composite, the benefits are highly questionable since the electrical impedance of these elements is very high.
E
SOME COMMERCIALIZED PIEZOELECTRIC COMPOSITE PRODUCTS
Since the first major composite work was introduced, composites have become a critical component in many transducer applications. Table 5 lists the companies (known to the authors) that manufacture and offer composites and/or composites transducers for sale. Some companies that are important
TABLE 5 COMPANIES OFFERING PIEZOELECTRIC COMPOSITES OR COMPOSITE TRANSDUCERS FOR SALE Company
Piezoelectric Composite Products Offered
Acoustic Imaging Technologies Corporation 10027 S. 51st. St. Phoenix AZ 85044 (602) 496 6681; (602) 598 9031 FAX ATL/Echo 1 Echo Drive, Reedsville, PA 17084-9772 (717) 667-5000; (717) 667-5001 FAX Blatek, Inc. PAD 2820 E. College Ave. Suite F, State College, PA 16801 (814) 231-2085; (814) 231-2087 FAX Diasonics Vingmed Ultrasound 2860 De la Cruz Blvd., Santa Clara, CA 95050 (408) 496 4700; (408) 496 3565 FAX General Electric Medical Systems P.O. Box 414, EA-54, Milwaukee, WI 53201 (414) 647-4000; (414) 647-4090 FAX Hitachi Medical Corporation of America 660 White Plains Rd. Tarrytown, NY 10591 (914) 524-9711; (914) 524-9716 FAX Imasonic 15 Rue Alain Savary, F-25000 BESANCON, FRANCE (33) 3 81 80 51 71; (33) 3 81 80 17 21 FAX Krautkramer Branson Inc. 50 Industrial Park Rd. Lewistown, PA 17044 (717) 242-0327; (717) 242-2606 FAX Materials Systems Incorporated 521 Great Road, Littleton, MA 01460-1208 (508) 486-0404; (508) 486-0706 FAX NGK Spark Plugs, Co. LTD., Ise Factory 871-6 Hosogo, Enza-cho, Ise-shi, Mie-prefecture 516-1196 JAPAN (81) 596-39-1630; (81) 596-39-1632 FAX Parallel Design 2430 West 12th St., Suite 6, Tempe AZ 852816931 (602) 966-6768; (602) 966-6543 FAX Sound Technology (Acuson) P.O. Box 8071, State College, PA 16803 (814) 234-4377; (814) 234-5033 FAX Tetrad Corporation 357 Inverness Dr. S. #A, Englewood, CO 80112 (303) 754-2301; (303) 754-2329 FAX Thomson Microsonics 399, Route des Crates B.P. 232, 06904 SOPHIA ANTIPOLIS CEDEX (33) 04 92 96 40 00; (33) 04 92 96 31 90 FAX Vermon 180, rue du General Renault B.P. 3813 37038 TOURS Cedex France (33) 2 47 37 42 78; (33) 2 47 38 15 45 FAX
Annular Arrays, Linear and Curved Sequenced Arrays--Medical OEM and Direct sales with system Annular Array, Doppler, Linear and Curved Sequenced Arrays--Medical OEM and Direct sales with ATL systems Single-element, Annular array, Doppler, Linear Sequenced Arrays--Medical and NDE OEM only Linear and Curved Sequenced Arrays, 1.5D Arrays - - Medical Direct sales with systems only Linear and Curved Sequenced Arrays-- Medical Direct sales with systems only Linear and Curved Sequenced Arrays--Medical Direct sales with systems only Single-element, Annular, Linear and Curved Sequenced, 2-D Arrays, High Intensity, High Temp--Medical and NDE; OEM only Single Element, Annular, Linear and Curved Sequenced Arrays--Medical and NDE OEM and direct sales with NDE systems 1-3 and 2-2 composite materials, Undersea panels-- Medical, NDE, Undersea System OEM and Transducer manufacturers PiezorubberTM - - Undersea OEM only
Linear and Curved Sequenced A r r a y s - Medical OEM only Single-Element, Annular, Linear and Curved Sequenced Arrays-- Medical OEM and direct sales with Acuson systems Linear and Curved Sequenced A r r a y s - Medical OEM and direct sales with systems Linear and Curved Sequenced Arrays, 1.5-D and 2-D - - Medical OEM only
Single-Elements, Annular, Linear and Curved Sequenced Linear Phased Arrays--Medical OEM only
1O0
Emmanuel P. Papadakis et al.
users of composite materials but do not manufacture the materials are not listed. Contact information and the types of products offered are also listed. It is impossible to show all the commercial products made using piezoelectric composites in the space provided. The products shown in the following figures were chosen based on the companies that provided the most rapid input, the graphics on unique products, or the highest quality photographs. The inclusion of product photographs is not necessarily an endorsement of those products or companies. Figure 28 shows Piezorubber, which is manufactured and sold by NGK Spark Plug Co. Ltd. Figure 29 is a picture from an Echo Ultrasound brochure, originally published in 1988, showing piezoelectric ceramic, 1-3 composite, and a single-element transducer made from piezoelectric composite. Figure 30 was provided by Blatek and shows a variety of medical imaging transducers made from composites, including single elements for imaging and Doppler, annular arrays, and one linear array. Figure 31, provided by Krautkramer Branson, shows composite transducers in their final packaged form. These include single-element transducers for imaging and Doppler, linear and curved sequenced arrays, and a curved array packaged for intracavity imaging. Figure 32, from Acoustic Imaging and Vermon, shows
FIG. 28. Photograph of Piezorubber from NGK Spark Plug's Company, NTK Technical Ceramics. Note the flexibility of the 0-3 material formulatedwith a flexible polymer.
2
Fabrication and Characterization o f Transducers
101
FIG. 29. A 1988 advertisement from Echo Ultrasound showing an unprocessed ceramic disk at the top, gold plated 1-3 composite disks, and an early single-element transducer.
FIG. 30. Composite transducers from Blakek. Composites for a linear sequenced and an annular are in front. Behind are linear sequenced array modules with a variety of single-element and annular array transducers.
102
Emmanuel R Papadakis et al.
FIG. 3 I. Composite medical transducers offered by Krautkramer Branson. These include linear and curved sequenced array modules for external imaging, a long probe for endovaginal imaging (upper right), a small single-element probe for measuring portions of the eye (black, pencil-like probe), and a variety of single-element imaging and Doppler transducers.
the wide variety of medical transducers that benefit from composite technology. The next two figures (Fig. 33 and Fig. 34), also provided by Krautkramer Branson, show a variety of composite NDE transducers. Figure 33 shows angled beam contact transducer and Fig. 34 shows end-radiating transducers, which are made both for contact scanning and for scanning parts immersed in water. Figure 35 shows a 1.5D array, offered by Thompson, that has a compositelike configuration built as part of the array. Figure 36 shows a number of injections modeled composites made by MSI, and Fig. 37 shows some naval transducers made from those composites. It is clear that the research that demonstrated the value of composites in transducer performance has had a dramatic effect on the ultrasound transducer industry. It is likely that piezoelectric composites will continue to play a larger role in transducers design. For example, research is currently being carried out in stacked composites (Mills and Smith, 1996), in which ceramics and polymers are aligned, bonded, and connected in parallel to increase the element capacitance in arrays. Figure 38 shows a stacked composite made
2
Fabrication and Characterization o f Transducers
103
:~,~ .~
~~
.... . . . . .
~,
,~ ~ , ~
FIG. 32. Medical imaging transducers. On the top fight is a set of external imaging linear sequenced arrays and on the lower left a variety of curved sequenced arrays for endocavity imaging. These are from Acoustic Imaging. On the fight are a linear phased array probe for imaging from the oesophagus (top), a curved sequenced array on a probe for laparoscopic imaging (center), and a set of sequenced arrays for imaging during open surgery (bottom). These transducers are from Vermon.
by Tetrad. Also of interest are composites made from single-crystal materials (Lopath et al., 1996) and composites as a part of "smart systems" (Fiore et al., 1997). These areas of research are likely to increase the number of applications for which composites are the piezoelectric material of choice. Increased usage will also occur due to work in reducing the cost of composite material, which will make it attractive for applications involving large-area coverage. In less than 20 years, piezoelectric composites have been converted from a laboratory curiosity supported by naval money into a major material
104
Emmanuel P. Papadakis
et al.
FIG. 33. Composite single-element NDE transducers from Krautkramer Branson. These are mounted to refracting lenses to steer the beam at an angle to the surface.
FIG. 34. Composite single-element NDE transducers from Krautkramer Branson. Some are for direct contact with a material to be tested and some are for testing a material through water.
2
Fabrication and Characterization of Transducers
105
FIG. 35. Photograph of a Thompson 1.5D array showing the focal characteristics of the array. The array uses a proprietary composite-like structure.
FIG. 36. An injection-molded preform from Materials Systems Incorporated for making a 1-3 composite.
106
Emmanuel R Papadakis et al.
FIG. 37. A SonoPanel TM for underseas applications from Materials Systems Incorporated, made using injection-molded 1-3 composite material.
FIG. 38. A close-up view of a stacked 2-2 composite from Tetrad. The dark line in the center is an extra electrode. The layers are connected electrically in parallel to increase the capacitance of an element.
2
Fabrication and Characterization o f Transducers
107
component in several ultrasound transducer industries. This has been the result of a tremendous amount of work occurring in the university research labs, the research labs of major companies, and in many other large and small companies around the world. It appears likely that the evolution is not completed and that the on-going research will continue to provide products using piezoelectric composites that are more cost-effective and have better performance than devices currently made from other materials.
IV.
A.
PVDF
1.
Unique Properties
P V D F Film Transducers
PVDF (polyvinylidenedifloride and its copolymers) is a plastic, polymer material with some very interesting piezoelectric properties. It has found widespread use in the ultrasound industry, due primarily to its flexibility and to the fact that its acoustic impedance is so much lower than piezoelectric ceramics. It is relatively lossy compared to ceramics, has relatively low electromechanical coupling and a lossy dielectric constant that decreases with frequency, but in spite of all this it remains firmly rooted in the industry as a material of choice for a number of applications. We will look at a few of these applications in this section.
B.
HYDROPHONES
1.
Introduction and Uses
Hydrophones are a good example of an ultrasound product that uses PVDE Hydrophones have found widespread applications in both industry and academia as they are a fundamental tool to measure acoustic pressure at a point in an ultrasonic field. There is not much PVDF in this p r o d u c t n j u s t a disk, typically 9-/~m thick and 400-/~m diameter, which comprises the active element. This active element is bonded down to a stainless steel backing, and ground connection is made to it via an evaporated Cr/Au electrode. One reason that PVDF is useful in this application is the fact that it can be cut and shaped into tiny disks that are useful at the tips of hydrophones, and the relatively good impedance match of PVDF to water aids in obtaining a flat frequency response. Also, it is important that the active element in a
Emmanuel R Papadakis et al.
108
hydrophone operates in a subresonant mode. With a 9-#m thickness, the active element will have a thickness mode around 133 MHz, making the hydrophone useful for measurements up to and above 50 MHz. It would be difficult to fabricate or bond a ceramic active element with this high of a resonant frequency. A CAD drawing of a typical needle-type hydrophone is shown in Figure 39. One use for a device like this would be the measurement of acoustic pressure and intensity produced by a commercial medical ultrasound unit. The FDA in the United States requires manufacturers of ultrasonic equipment to measure such quantities and supply information about the ultrasonic output both to the FDA and to the end user of the equipment. The modem state of the art in these measurements involves the use of hydrophones and a fairly elaborate measurement protocol. Another emerging use of hydrophones is to measure the performance of ultrasonic cleaning baths. Ultrasonic cleaning has become much more widespread in recent years due to the Montreal Protocol, which bans the use of ozone-depleting substances. This use of hydrophones has proven to be a bit challenging, owing to the fact that the pressure in a cleaning tank is time variant and in fact somewhat chaotic, and the cavitation in this media can damage hydrophones. Hardened hydrophones developed specifically for this application utilize a ceramic active element (not PVDF) and have the edges of the active element protected with metal. To still limit the size of the active element, these ceramics can be spot poled, i.e., made active only in very small, highly localized areas. Many other applications exist for hydrophones, including photoacoustics, sonoluminescense, NDT, hyperthermia, therapeutic ultrasound, and lithotripsy. Hydrophones prove useful any time the acoustic pressure in a liquid needs to be measured.
RG174 Cable I_
\
1.5000
Female SMC Connector
~[ 0.5000 [---
Active Element
FIG. 39.
l
Cable length 8"
j-I
(CAD) drawing of a needle-type hydrophone.
,
2 2.
Fabrication and Characterization o f Transducers
109
Discussion and Theory
A hydrophone works by converting the acoustic pressure at a "point" into a voltage that can be measured with a tool like an oscilloscope or spectrum analyzer. In reality, the active element integrates the pressure over its entire active area to produce this voltage, hence the size of the hydrophone element, in terms of wavelengths, is of paramount importance. Ideally, one wants a hydrophone with an active element diameter on the order of a quarterwavelength in diameter at the highest frequency of interest. This would assure a broad angular response and no frequency filtering of the response of they hydrophone due to diffraction. There are, however, trade-offs to be made between the hydrophone element area and sensitivity and cost; smaller apertures are both more costly to build and less sensitive. In the medical industry, where frequencies are typically in the 3 MHz range, a typical hydrophone diameter chosen might be 400/~m corresponding to slightly less than a wavelength at the center frequency. Compromises are even worse, relatively speaking, when one needs to measure a catheter at 30 MHz. The smallest commercially available hydrophones, to this author's knowledge, are around 100 or 150/~m, which would correspond to 2 or 3 wavelengths in water at the frequency of interest. Clearly, in these types of measurements the hydrophone alignment to the acoustic beam under test is of critical importance. Trade-offs between measurement standards and current state of the art in hydrophones is an ongoing concern. New devices, which will be able to push hydrophone active element sizes down below 20 #m, are under development by the author but are still far from commercial availability. Interesting work in this area has been published recently by Lum at HP Labs (Lum et al., 1996). Another key design goal with a hydrophone is to ensure that the output voltage produced is proportional to pressure (i.e., the device is linear) and that the constant relating pressure in to voltage out is constant with frequency. A fundamental property of a piezoelectric material is that the electric field in the material is proportional to stress through it. This fact is utilized in making a hydrophone, with care being taken to ensure that the piezoelectric element need not develop a great deal of current and that it is operating well below resonance. Typically, a ceramic active element mounted on a steel backing might have no resonance at any frequency, and a 12-micron-thick copolymer film mounted on a steel backing might have a quarter-wave resonance around 100 MHz and a reasonably flat frequency response up to frequencies as high as 40MHz. Since a piezoelectric element operating below resonance is
Emmanuel R Papadakis et al.
110
basically a capacitor, an appropriate receiver input should also be a capacitor to ensure that the voltage sensitivity of the device is flat with frequency. Typically, the voltage out of a hydrophone is sensed with a special circuit--a preamplifier--so that the hydrophone is only loaded by a short cable and the input capacitance of the preamp, which is carefully engineered to be minimal. In some more recent hydrophone designs, preamplifiers made with the tiny surface mount components are built right into the hydrophone body, so as to minimize the capacitance seen by the active element. Cables, front end impedance of a preamp, housings, and other sources of parallel capacitance work to reduce hydrophone sensitivity in a very predictable manner (known as capacitive voltage dividing). As aperture size is reduced, all of the above effects become more important. A KLM model can be used to predict the output sensitivity of a hydrophone as a function of frequency. This model is similar to what was developed by Sittig, although it uses a "T" matrix approach to analysis to simplify the proper treatment of lossy components (Selfridge and Gehlbach, 1985). Here the model is used to investigate the effect of electrical load resistance on hydrophone output. The graph in Fig. 40 shows the voltage into the preamplifier as a function of frequency, assuming that an acoustic wave with an amplitude of 1 mW/cm 2 is exciting the hydrophone. As can be seen from this graph, the voltage is constant with frequency, up to about 40 MHz,
O.I
ZS= 106/7 LLI s <:1: I-._J O
I0 woo kHz FIG. 40. loads.
:
_
_
A
.
_ _ : _
.
:
:
:
:
; ; ; .
FREQUENCY
:
:
:
:
:
. : : = !
MHz
Predicted frequency response of a needle-type hydrophone with various electrical
2
Fabr&ation and Characterization o f Transducers
111
provided that the real part of the preamplifier input impedance is on the order of 100 k ~ or above. The piezoelectric layer in this hydrophone was resonant around 40 MHz. PVDF hydrophones can have thinner active elements and remain flat up to frequencies around 100 MHz. Current information about this model can be found on the Internet at http://www.ultrasonic.com/ products/software. In real life, the situation is a bit more complicated than is modeled by the above theory. First, the pressure produced on a hydrophone by a propagating plane wave is not even a constant function of frequency. As stated previously, a typical hydrophone might be 400 microns in diameter. At 1 MHz, this is less than a third of a wavelength, but at 30 MHz, it is over ten wavelengths. Consequently, at 30 MHz the 400-micron aperture looks like a mirror, and a plane wave incident from a water bath is reflected off a rigid boundary and sees a factor of 2, or 6 dB gain. At 1 MHz, the plane wave, although scattered a small amount by the hydrophone, passes basically undisturbed, and the hydrophone senses the "free-field" pressure of the plane wave field. None of these effects are modeled by the graph above, nor is any account made for the angular response of the hydrophone at 30 MHz. A typical experimentally measured sensitivity vs frequency response for PVDF-type hydrophone is shown in Fig. 41. Although the voltage sensitivity is basically flat, some variations due to nonideal construction, diffraction, and other effects are clearly present.
-248 -
,,-..., 13_
tO9
-258
0
i 5
I 10
ll5
I 20
Frequency (MHz)
FIG. 41. Measured frequency response of a needle-type hydrophone.
Emmanuel R Papadakis et al.
112
3.
Conclusion
Hydrophones represent an evolving art, where the design engineer is forced to weigh many factors such as sensitivity, cost, and diffraction effects to come up with a workable comprise. Fortunately, many of the developments being made in other fields, such as surface mount technologies, are allowing the development of hydrophones with better features and fewer compromises than what could be made previously.
C.
BROADBANDSOURCES
1.
Introduction and Uses
Often times it is necessary to transmit over a wide range of frequencies, as opposed to just receiving as with a hydrophone. Once again the unique properties of PVDF or its copolymers can be utilized, and in this case the relatively good impedance match of PVDF to water is useful. Broadband sources (BBSs) are used for calibrating hydrophones, and have a variety of other applications. Topics of contemporary interest are the characterization of microbubbles, useful for enhancing the use of ultrasound in medicine, as well as determination of participle size distributions and concentrations in liquid suspensions. The latter is the subject of a patent granted to Alba (1992). 2.
Discussion and Theory
A typical BBS will utilize a layer of PVDF bonded down to a matched backing. The backing will typically consist of a filled epoxy, as it is not difficult to make such a material with impedance in the 4 to 5 MRayl range required to match PVDE The Internet web site http://ultrasonic.com/tables lists many such materials, a particularly nice one in this application is a dielectric absorber material from Emerson Cummings known as MF110. A cross-sectional schematic of a large-aperture BBS is shown in Fig. 42. A device like that shown here will have an insertion loss, when driven and received by 50 ~, of around 44 dB. It will have a - 6 dB working range from below 2 MHz to above 10 MHz. The low-frequency end can be extended down to much lower frequencies by driving the transmitter with a higher impedance and receiving with a high-impedance preamplifier-type circuit. The 50-~ receivers tend to roll off the low end. A high-impedance transmitter can be as simple as a 1-k~ resistor in series with a big ENI amplifier, or possibly implemented as a custom-designed circuit. The main point is that at
2 Fabrication and Characterization of Transducers -
113
1.9980
S
110 lam PVDF copolymer film --3"
MF110 Backing
1.5000
UHF Female Connector
FIG. 42.
Cross section of large-aperture, low-frequency BBS.
low frequency, current must be crammed into a reluctant, high-impedance capacitor. On the receive end, the concept to keep in mind is simply the fact that one capacitor dumping into another is flat with frequency, but a capacitor driving a 50-f~ resistor will have a zero at zero, or a low-frequency roll off at a predictable frequency. Measurement of insertion loss as above can be made much simpler with the use of a time-gated network analyzer. Such a system can be composed of an HP8165A programmable function generator, used to generate tone-bursts, and a Tektronix TDS724A oscilloscope, used to digitize tone-bursts received at time was arbitrarily delayed relative to the transmitted tone-burst. Custom software (such as program VOF described at http://www.ultrasonic.com/ products/software) can then be used to iteratively change frequency and record amplitude of the received tone-burst. For BBSs that are going to be used at high frequencies, the metal electrodes needed to generate the electric field in the copolymer must be kept thin. At 100 MHz even a micron-thick later of gold can cause very undesirable frequency effects when positioned between an active piezoelectric later and a matched backing. Given thin enough layers of gold--i.e., on the order to 2000 angstromsmfrequency responses like that shown in Fig. 43 are predicted with theory.
114
Emmanuel R Papadakis et al. 60
,,._.,, t.-
.o_
50
t-
....................
<
40
8.2
FIG. 43.
6 dB . . . . . . . . . . . . . . . . . .
50
T 66.5 F r e q u e n c y (MHz)
1 O0
106.1
I 150
Theoretical frequency response of high-frequency BBS.
This theory assumes that the copolymer is 12-#m thick, has an area of 0.712 cm 2, has a stiffened acoustic impedance of 4.45 MRayls, a Qm of 20, a g r o f 3.2, loss tangent of 0.2, stiffened acoustic velocity v~ of 2.47 mm/ms, and a k2 = 0.078. It further assumes that this active element is loaded with 1300 A of gold on each face, then water on the front and impedance 4.2 MRayls on the back. The assumed electrical drive is 50 f], and the assumed receiver input impedance is 10pE Given these assumptions, one predicts more than a decade of bandwidth, two way, given - 6 - d B cutoffs. It should be noted that devices like these are completely nonresonant. This means that the active element is mounted on a matched backing and that the backward-traveling wave, generated by the active element, goes into the backing, hopefully to never be seen again. Without the benefit of this backward-traveling energy, nonresonant devices will have lowered sensitivity, e.g., a - 4 0 - d B minimum insertion loss in the case modeled above. (Note: Insertion loss is only defined when the 10 pF is replaced with a 50-f~ load.) In the case of PVDF element on a matched backing working into water, 65% of the generated ultrasound will go into the backing and 35% will go forward. In the case of a PT (lead titanate) element on a matched backing working into water, 93% of the generated ultrasound will go into the backing and 7% will go forward. Herein lies another inherent advantage to using PVDF, owing directly to its lowered acoustic impedance relative to ceramics. Measuring broad frequency responses is another story in itself. Typical ultrasonic pulsers often fail to have broad enough frequency responses to do justice to devices such as these. Alba (1992) came up with an interesting means for measuring broadband transfer functions using a network analyzer,
2
Fabrication and Characterization of Transducers
115
provided such measurements can be made on device pairs. (The technique does not lend itself to the measurement of single transducers, a topic of ongoing research.) In his method, one aligns two broadband sources with each other, with a ~5-mm water gap between them. One then hooks one device to the swept local oscillator of the network analyzer and the other to the receiver section. Then these are swept in frequency, a very complicated (peaky) spectra results due to the standing waves that exist between the transducers in the water path between them. Alba collected this spectra, including its complex phase, and performed an inverse Fourier transform on it. The resulting function looks similar to what one would see on an oscilloscope, complete with multiple reverbs due to the standing waves and with remarkably broad frequency response. He then gated out the "first arrival" in time, performed the Fourier transform on this part of the signal, and obtained the combined spectra of the transducers and the water path for a single transit. The water path, by the way, can have a significant attenuation of its own at these higher frequencies (ASTM, 1996). At 100 MHz, for example, a typical attenuation would be 31.26 dB/cm with an f2 dependence. Typical measurement results for a pair of broadband sources are shown in Fig. 44. This result, while similar to Fig. 43, is not intended as a direct comparison between theory and experiment. The theory in Fig. 43 does not take into account the frequency dependence of the attenuation in water, the fact that the piezoelectric properties of PVDF vary significantly with frequency, and the fact that the receiver in Fig. 44 is loaded with 50 f~.
60
rn
o
"~
.................
6 dB
50
r
-47 dB
40
FIG. 44.
9.4
36
,,I I 50 100 115 Frequency (MHz)
Actual two-way transfer function measured on a pair of high-frequency BBSs.
116 3.
Emmanuel R Papadakis et al. Conclusion
PVDF lends itself to use in broadband applications in water due to its relatively low acoustic impedance and internal damping. Devices made in such applications have been challenging to measure because of their very broad bandwidths. In addition, the frequency dependence of the piezoelectric properties of PVDF over broad bandwidths make it challenging to model.
D.
PVDF AIR TRANSDUCERS
1.
Introduction and Uses
The strongest piezoelectric response one will find in PVDF is in response to stretching it. The e31 coupling is typically much higher than e33. This property of the material can be utilized when attempting to make efficient transducers for use in air. Air has a plane wave acoustic impedance of approximately 0.0004 MRayls as compared to 1.5 MRayls for water, approximately 4 MRayls for PVDF, and approximately 35 MRayls for PZT. Clearly, the production of ultrasound in air is difficult to do over a wide bandwidth when pushing on it directly with thickness mode devices. One clever way around the problem is the concept of the "singing drum." In one of the simplest implementations of this concept, a sheet of PVDF is laid over a metal plate with an array of holes in it. Pressure or vacuum is then applied to one side of the plate to create an array of PVDF diaphragms, each with the diameter of the hole under it, which are under uniform tension and can be driven in parallel. A schematic cross section of such a device is shown in Fig. 45. Note that this concept is in the process of being patented (Selfridge and Khuri-Yakub, 1997).
Metal Plate Vacuum FIG. 45. Schematiccross section of 'singing drum' air transducer.
2
Fabrication and Characterization o f Transducers (t)
E
>
0.5 _
_
/
117
93.562 kHz 169.60 mVpp
0
>
0.1
Q.
= o
-
",..,.
t0 ..c e~
._r
0.01 ...... 70
|
]
!
1 I II i 1001 86.517 102.12 Frequency (kHz)
FIG. 46. A i r pressure in front of an array o f 'singing drums' microphone.
2.
I 120
as measured w i t h a
Discussion
Devices such as those described above have been built and tested. Unfortunately, the theoretical analysis has not been carried out to a satisfactory level. Just the prediction of the lateral stress in the membrane (as a function of back pressure) has proven to be a complicated affair that interrelates film compliance and thickness to radius of curvature and ultimately lateral stress. More work is needed in this area, possibly with the use of finite element analysis. The measurement of the singing drum has been considerably simpler than its theoretical analysis, given a laser vibrometer. To measure the bandwidth and absolute displacement of a singing drum transducer, one only needs to focus a laser vibrometer on the vibrating surface. One can then sweep the frequency and record the amplitude and phase of the resultant vibration. Alternatively, one can use a microphone (i.e., a B&K 4138) at a point in the far-field from such a device to measure the remarkable, smooth passband that can be obtained. A typical device showing an approximately 14%, - 6 - d B bandwidth around 93 kHz in air is measured in Fig. 46. This device utilizes 25-micron film on a plate with 1-mm holes with a back pressure of 1 atm.
3.
Conclusion
PVDF has its strongest electromechanical coupling to stretch. This fact can be useful when designing transducers that need to drive low acoustic impedance media such as air.
118
Emmanuel R Papadakis et al.
V.
Electromagnetic Acoustic Transducers (EMATs)
EMATs are current-operated, inductive transducers. A coil induces currents in an adjacent metal surface in the presence of a static or quasi-static magnetic field. This form of electromechanical transduction works on any metal. EMATs operate on both magnetic metals such as steel and nonmagnetic metals such as aluminium and stainless steel. Once generated by an EMAT, an elastic wave behaves just like an elastic wave launched by any transmitting element of identical amplitude, phase, and source diffraction. EMAT generation of elastic waves is, however, different in magnetic and nonmagnetic metals even though the transducers, in some instances, appear to be identical. The important differences between operation on magnetic and nonmagnetic metals is given at the end of each case discussed below. For the theoretical background for this discussion of EMATs see, for example, Maxfield and Fortunko (1983), Maxfield et al. (1987), Alers and Burns (1987), Alers et al. (1990), and Thompson (1990). All the discussions in this section relate to single-element transmitters or receivers. Some work has been done with arrays but these are, for the most part, still devices in development. EMATs almost invariably have a higher insertion loss (lower power efficiency) than piezoelectric transducers generating the same elastic wave. This means that EMATs should only be used when their primary advantages - - couplant-free operation or the ability to generate elastic modes that are otherwise difficult--are required by the user. Such applications include couplant-free generation of plate, surface, and Lamb waves for high-speed defect detection and for high-temperature (HT) ultrasonic measurements. As an example, if the proper construction materials, bonding techniques, and cooling methods are used, EMATs can easily operate when adjacent to surfaces as high as 1000~ The major intrinsic limitation of EMATs is that the elastic wavelength being generated must be large compared to the electromagnetic skin depth of the radio frequency (rf) currents that are generating the elastic wave. For most metals, a practical upper frequency is in the region of 5 to 20 MHz.
A.
FUNDAMENTALS
EMATs are made by combining wire or printed circuit coils and permanent magnets or electromagnets to generate the desired wave mode. In some ways, the different magnet and wire combinations are analogous to the crystal cuts or polarizations of piezoelectrics. Five different EMAT types have been built
2
Fabrication and Characterization of Transducers
119
for commercial or laboratory use; they are listed here according to the wave types that are generated: 1. Bulk, normal beam, shear horizontal (SH) (radial or linear polarization) waves 2. Bulk, angle beam, SH waves, surface-skimming SH waves, and SH plate waves 3. Rayleigh waves, angle beam shear vertical (SV) waves (peaked around 37 ~) and Lamb and other plate waves 4. Bulk longitudinal waves 5. SH waves in magnetic metals B.
GENERAL CONSIDERATIONS
In all cases, a magnetic field interacting with the current induced in the metal by an adjacent coil generates a surface stress via the Lorentz magnetic force. In some cases, this stress is relatively constant over the surface while in others, a periodic current or magnetic field is used to produce a periodic surface stress. In magnetic metals, additional stresses are generated by magnetostriction. Because they use a biasing magnetic field, EMATs behave differently on magnetic and nonmagnetic materials. Also, elastic waves are generated in magnetic materials by two separate and quite different mechanisms: (1) magnetostriction and (2) the Lorentz force on induced or eddy currents. The behavior of EMATs on magnetic materials can be quite varied depending on the strength and orientation of the biasing magnetic field. It is beyond the scope of this section to deal with the detailed considerations of EMAT design in different magnetization regimes. Careful attentionto the design details for some situations can result in a 10- to 25-dB increase in the signal-to-noise ratio (SNR). On the other hand, some types of EMATs operate on magnetic metals in much the same manner as on nonmagnetic metals. The discussion and descriptions given below are technically exact for EMATs on nonmagnetic metals. At the end of each case description, we comment on how operation may differ on magnetic metals. The exception to this is Case 5, which applies only to magnetic metals. As is the case in designing any ultrasound transmitting transducer where a directed, forward, or angle beam is required, source diffraction or beam spread plays dominant factor in the design of EMATs. Since elastic wave generation takes place by the interaction of induced surface currents with a static or quasi-static magnetic field, the physical size of both the rf coil and the source
Emmanuel R Papadakis et al.
120
of the magnetic biasing field must be taken into account. This is sometimes complicated by the fact that the induced current distribution from a coil is often spread over an area much larger than the coil dimensions (induced currents loops must close). Three fairly simple cases serve to illustrate how to calculate (or at least estimate) the dimensions of the elastic wave source, which is the dominant geometrical factor in determining beam spread: 1. A spiral coil biased by a permanent magnet; here the source size is about the coil radius (Fig. 47(a)). 2. An elongated spiral coil combined with a rectangular magnet pole cap to generate linearly polarized shear waves in a ferromagnetic metal such as illustrated in Fig. 47(b); here the source dimensions are nearly those of the magnet pole cap. 3. A shielded and shaped coil as illustrated in Fig. 47(c); here the sources size is approxiatmely that of the linearly polarized current region beneath the coil. The desired or optimum physical size of the receiving transducer is governed by quite different parameters than those for the transmitting element. For normally incident waves, the size can be determined by the required spatial resolution. Within this restriction, it is customary to use a coil having as many turns as possible consistent with a self-resonant frequency that is at least somewhat above the highest required operating frequency. The output voltage from the receiver coil is given by
VR- J E. dr,
(1)
where E is the surface electric field that is generated when an elastic wave either is reflected from the surface adjacent to coil or passes under the coil. The line integral is over the length of wire in the coil. For a rectangular coil such as the one mentioned in example (2) above, one has V R = ~ o B N U W 2,
(2)
where N is the number of turns in the coil of width W2 (see Fig. 47(b)), B is the value of the bias magnetic field, co is the angular frequency, and U is the elastic wave displacement. For nonnormal incidence, the voltage is always lower; VR= 0 when there is a phase difference of 2~N in the elastic wavefront fields over the face of the coil.
MAGNET
RF
COIL
BNC
(a)
I
~\
S ec f io n
A- A "I]'>l" " "////"'"
..I...
PERMANENT MAGNET \
P O L E CAP
#.:~&~A~ ~
/.'
/
~~,
"1"I;///'I" I''/I//~ S /T/ E/ E/ L/ / S/ ,"U/ R/ FA C E //>';~. ;''V'//V/>S'; . ," / I / / . / / / / / / / / / / ///., / / / / / / / / / / ' / ,,-.,-/'..
///
......
,,. . . . . . . . . .
.. , . / . . /
,/.....,,,-1
~ ;~...
8<:#
t A
t
il
\,Fil Z!,/ COIL
POLE
W I
CAP
(b) Seciion
A-A
Secfion
B --B
-~\~&\\\~
~ Conducting shield
L__ . . ............ ...... '
"/,/.J>?x,<,: ..::~.,~;:: ,S.TE.E,L
///
1/II.'..'.i/
/
.SU.R,rA c E "./..,> :>:'; -"::::F.:;: .'.":;/.,< ::
...........
""
"//),.......//i///li
(c)
FIG. 47. (a) Sketch of a permanent magnet EMAT for the generation of radially polarized shear waves in nonmagnetic or magnetic metals. (b) Two elongated spiral coils connected as shown in series, biased by a magnet with a pole cap of width W1 and length W2 and placed near a magnetic metal. The shield must be grounded. (c) Two elongated spiral coils connected in series as shown in (b). The central region is closer in elevation to the metal surface than the remainder of the coil. This generates a somewhat rectangular region of approximately linear current density surrounded by a much larger region of "returning" current. The shield should be grounded and at least 50% larger than the total coil size.
Emmanuel R Papadakis et al.
122
When used with magnetic materials, pole caps on electromagnets or permanent magnets can be used to define spatially the region over which elastic wave generation takes place. This field focusing is discussed below for the generation of shear waves in steel. Impedance matching and tuning are extremely important concepts when working with EMATs (ARRL, 1997). To achieve the lowest insertion loss, it is necessary to get the largest possible current to flow from the transmitter element driver electronics into the transmitter EMAT. This normally requires impedance matching for maximum power transfer. Since the EMAT load is reactive (inductive), this is a frequency-dependent matching criterion. Impedance matching coupled with tuning of the EMAT inductance determines the EMAT operating frequency and bandwidth. Since there is always some power lost in any impedance matching network, the lowest insertion loss usually comes when the EMAT rf coil is tuned so that the coil and tuning network are impedance matched to the transmitter electronics output characteristics. Normally this means the EMAT transmitter electronics (EMAT driver) must have an output resistance in the range of 1 to 50ohms. Drivers with low output resistance (high output current) should be placed as close as possible to the EMAT rf coil. This has led to the concept of active EMATs, ones that incorporate significant electronics (both transmit and receive). Receiver impedance matching is usually simpler in concept and easier to implement. The resistance of the EMAT, as seen by the input of the EMAT low-noise preamplifier, should produce a noise voltage at least as large as the preamplifier noise referred to the input. This means that the system noise is dominated by the signal source, (the EMAT), and hence the system noise performance cannot be improved without redesigning the sensor (the EMAT). Typical good commercial low-noise amplifiers have a input noise-equivalent resistance of 40 to 100ohms. Thus, the transformed EMAT ac resistance (always bigger than the dc resistance) should be around 100 ohms to achieve the best signal-to-noise ratio (SNR) using that particular receiver. Sometimes this requires as much as a 1:100 impedance step-up from the EMAT coil (tenfold increase in the EMAT coil voltage). Transmission line impedance transformers are usually best for these purposes (Sevick, 1987).
C.
CASES BEING CONSIDERED
1.
Case 1
Bulk, normal beam, SH waves are generated by placing a current-carrying coil in a magnetic field perpendicular to the surface in which elastic waves are
2
Fabrication and Characterization o f Transducers
123
~Bo ~
J
~ ~F
L
FIG. 48. Pictorial representation of the magnetic field and inducing coil configuration needed to generate shear waves through the Lorentz force mechanism.
being excited. One configuration for accomplishing this is shown in Fig. 48. The Lorentz force is parallel to the surface and the resulting shear wave propagates away from the surface along the surface normal. A linearly polarized induced current (such as that obtained over a portion of the coil face in example (3) above) can be achieved locally by using the proper coil geometry. This will generate a linearly polarized shear wave if the magnetic field is localized to the region of linearly polarized currents. (The divergence of the current is zero, so all current loops must eventually close. Consequently, it is not possible to have a coil that generates a linearly polarized current everywhere.) A spiral coil, which really consists of concentric loops of wire, will generate a nearly circular current pattern in the metal surface (these current loops close in themselves). In this case, the Lorentz force is radially outward, thereby generating a radially polarized shear wave. Behavior on magnetic and nonmagnetic metals is similar except that field focusing can only be used on magnetic metals.
2.
Case 2
The periodic permanent magnet (PPM) and elongated spiral or racetrack coil EMAT shown in Fig. 49 can generate three different types of waves. The current, flowing in opposite directions in the fight and left halves of each of the elongated spiral coils, experiences a periodic magnetic field of opposite polarity. The magnet poles are arranged such that the Lorentz force is essentially constant along any line perpendicular to the wave propagation
Emmanuel R Papadakis et al.
124
direction (which is parallel to the induced current). This produces a spatially periodic surface shear force at the frequency of the alternating current in the coil. When this EMAT is driven at a frequency given by f = Vs/2D (where Vs is the bulk shear wave velocity and 2D is the magnet spatial period), a surface-skimming SH wave is generated. This is not a classic surface wave, but rather, a surface-skimming bulk SH wave. At a higher frequency, an SH wave is generated that propagates at an angle 0 to the surface normal, where 0 - s i n -1 [Vs/(2Df)]. The amplitude of this wave is almost independent of the angle. Note that the angle can be tuned electronically by varying the frequency. This magnetic structure works, but sometimes not very well on magnetic metals. The geometry discussed in Case 5 is recommended for SH waves in magnetic metals having significant magnetostriction. The EMAT in Fig. 49 also generates SH plate waves. Note that SH plate waves are dispersive. In other words, the wave velocity depends on the plate thickness.
"poxySpacer S mm)
Magnet
Shunt
Gless/Epoxy Sp~ (1.3 ram)
Magnet Shunt
~' Spacer . 9. . . . . . . 1)
,32 AWG.
FIG. 49. pickup.
Cutaway view of a PPM EMAT that has good rejection to common-mode electrical
2
3.
Fabrication and Characterization of Transducers
125
Case 3
Perhaps the most widely used and broadly useful EMATs are ones using a spatially periodic current distribution and a spatially constant bias magnetic field. Since the periodic induced currents are frequently produced using a serpentine or meander-line coil such as the one shown in Fig. 50, these are often referred to as meander-line (ML) EMATs. When placed on a thick material (many wavelengths thick) and driven at a frequency f = VR/(2D) (where VR is the Rayleigh wave velocity and 2D is the spatial periodicity of the induced currents), the ML EMAT generates a Rayleigh wave that propagates perpendicular to the current lines and has a beam divergence characteristic of an end-fire antenna system that has N elements placed at half-wavelength intervals, where N is the number of current line elements (Maxfield et al., 1987). At a frequency given approximately by f = 1.6 Vs/(2D), a shear vertical (SV) wave is generated at 37 ~ to the vertical (the equation that gives the exact angle as a function of excitation frequency is not very useful because there is only significant amplitude available in a 10~ band centered around 33~ When placed on material less than a wavelength or so in thickness, and tuned to the correct frequency, an ML EMAT generates Lamb or other plate waves that have both SV and L particle displacements. (See the dispersion curves in Fig. 51, which shows that the plate wave velocity depends on both the plate thickness and the excitation frequency.)
t
..,
9. . . . .
'
,
FIG. 50. A pictorial representation of a meander-line EMAT using a bias field parallel to the surface. Perpendicular bias fields can also be used. The bias field can be supplied by either PMs or EMs.
Emmanuel R Papadakis et al.
126
>,
o
5
o" 1.1_
4
0 00
E
3
a
2
8
1
2
3
4
5
6
?
j~ Dimensionless Wave Number FIG. 51. Dimensionless frequency as a function of dimensionless wave number of symmetric Lamb modes. The slope is the wave group velocity. For reference purposes, the three standard velocities are shown.
When the different material B-field is taken into account, these EMATs work about the same on magnetic and nonmagnetic metals.
4.
Case 4
Bulk longitudinal (L) wave EMATs are the highest insertion loss EMATs, primarily due to the requirement for a large magnetic bias field parallel to the surface. For nonmagnetic metals, this typically requires a gap in the magnetic circuit that is several wavelengths, say, 1.0 cm at an operating frequency of 1 MHz. Also, for L wave generation, it is seldom possible for the EMAT rf coil to be inside the magnet gap. This had led to EMAT pole designs that tend to "push" the magnetic flux out one side of the magnet where the rf coil is placed. The flux line plot shown in Fig. 52 corresponds to the possible magnet pole configuration shown in Fig. 53.
2
Fabrication and Characterization of Transducers
127
FIG. 52. A flux line plot for a magnet that is useful for generating L waves in nonmagnetic metals. The dark region represents the magnet iron in the magnetic circuit. The rf coil is placed inside the physical confines of a magnet (as shown in FIG. 53) but the nonmagnetic metal being investigated is not restricted by the magnet geometry.
RF Coil
Cooling Grooves
FIG. 53. A HT longitudinal wave EMAT rf coil placed to use the maximized fringing field from a modified C-shaped electromagnet. The coil sits very close to the test surface temperature. The cooled surface plate keeps the magnet pole caps at an acceptable temperature and protects electronics sometimes mounted beneath the coil holder.
Emmanuel R Papadakis et al.
128
Additional complicating factors enter for magnetic materials when either generating or receiving L waves, particularly at or near normal incidence to the surface. A discussion of L wave EMATs in magnetic metals is beyond the scope of this article; the reader is urged to proceed with careful thought to EMAT design when trying to transmit or receive L waves in a magnetic metal, especially ferromagnetic metals at fields below magnetic saturation where both magnetostriction and the Lorentz force contribute to the generation of elastic waves.
5.
Case 5
SH surface and plates waves can be generated very efficiently in magnetic metals using a meander-line (ML) coil and a magnetic field parallel to the surface with the coil and magnet geometry shown in Fig. 54. This form of EMAT uses magnetostriction (in this case, the rf-induced currents produce magnetic fields that interact with the magnetic domains in the metal surface), so the applied or bias magnetic field requirement is modest, from 30 to 300 mT (Davidson and Alers, 1997). Since this field must exist in the surface where rf currents are located, it can be particularly helpful to use a timedependent bias field to take advantage of the electromagnetic skin effect. The if-pulsed current that actually generates the elastic wave is triggered just prior S H W a v e Path
\
~ CARRIAGE X
EMAT
B U'I-IWELD
"x
~,
9
\
'\
.
~X~ EMATs
Pulsed Magnets
FIG. 54. A magnet and coil structure that is useful for generating SH waves by magnetostriction. Small, pulsed electromagnets are used to generate a biasing magnetic field parallel to the surface and at an angle to the propagation direction so as to maximize the transduction efficiency.
2
Fabrication and Characterization of Transducers
129
to the peak in the pulsed bias field. Generally, the bias field has a duration of 10 to 100 #sec, so the same bias field pulse may be used for generation and detection that occurs within the time frame of the bias pulse. Obviously, this configuration only works on magnetic metals that have significant magnetostriction. Source diffraction is governed primarily by the coil width and the number of wavelengths in the transmitting coil.
VI.
Summary
As RKO Pathe news was "the eyes and the ears of the world," so transducers are the eyes and the ears of most ultrasonic systems. From simple beginnings in piezoelectric crystals, transducer technology has branched out into the use of electromagnetic coils, polymer films, and finely partitioned piezoelectrics to take advantage of particular properties useful in certain situations. Research has led to many improvements and many new devices. Coils and magnets can work on metals in a noncontact mode. PVDF films match well into liquids and can radiate into air effectively because of their high coupling coefficient in stretch, which can be translated by geometrical construction into a drumhead sort of radiator. The finely partitioned (sliced, diced, molded) piezoelectrics have a higher coupling coefficient for longitudinal waves and minimize unwanted radial motion. Arrays can be made directly from the diced parts with proper electrical connections. This chapter has given details of theory, manufacture, and analysis of transducers. Examples have been given, but for complete listings of manufacturers and parts, the reader should consult NDT advertising and buyers guides.
References Alba, E (June 16, 1992). "Method and Apparatus for Determining Particle Size Distribution and Concentration in a Suspension Using Ultrasonics." U.S. Patent No. 5,121,629. Alers, G. A., and Burns, L. R. (1987). EMAT designs for specific applications. Mater. Eval. 45, 1184-1189. Alers, G. A., Maxfield, B. W., Monchalin, J. R, Salzburger, J. J., and Thompson, R. B. (1991). Other ultrasonic techniques. In "ASNT Nondestructive Testing Handbook," Columbus, OH: ASNT, Paul McIntyre ed., 2nd Edition, Vol. 7, Section 10. Alippi, A., Craciun, E, and Molanari, E. (1998a). Piezoelectric plate resonances due to first Lamb symmetric mode. J. Appl. Phys. 64(4), 2238-2240. Alippi, A., Craciun, E, and Molanari, E. (1998b). Stopband edges in the dispersion curves of Lamb waves propagating in piezoelectric periodical structures. Appl. Phys. Lett. 53, 1806-1808. ARRL (American Radio Relay League) (1987). "ARRL Handbook for Radio Amateurs." ARRL, 225 Main Street, Newington, Connecticut 06111.
130
Emmanuel P. Papadakis et al.
ASNT (1959). "Nondestructive Testing Handbook," 1st Edition, Vol. 2 (R. C. McMasters, ed.). Roland Press, New York, pp. 12-19. ASNT (1991). "Nondestructive Testing Handbook", 2nd Edition, Vol. 7, (P. McIntire, ed.). American Society for NOndestructive Testing, Columbus, Ohio. ASTM (1996). Standard Guide E1065-96, "Standard Guide for Evaluating Characteristics of Ultrasonic Search Units (Transducers)." Auld, B. A., and Wang, Y. (1984). Acoustic wave vibrations in periodic composite plates. Proc. 1984 IEEE Ultras. Syrup. pp 528-532. Auld, B. A., Kunkel, H. A., Shui, Y. A., and Wang, Y. (1983). Dynamic behavior of periodic piezoelectric composites. Proc. 1983 IEEE Ultras. Symp. pp 554-558. Auld, B. A., Shui, Y. A., and Wang, Y. (1984). Elastic wave propogation in three-dimensional periodic composite materials. J. de Physique 45, 159- 163. Banno, H. (1983). Recent developments of piezoelectric ceramic products and composites of synthetic rubber and piezoelectric ceramic particles. Ferroelectrics 50, 3 - 1 2 . Banno, H. (1989). Peizoelectric composites. 1989 Yearbook Encyclopedia of Physical Science and Technology, 219 - 226. Beaver, W. L. (1974) "Sonic nearfields of a pulsed piston radiator" J Acoust. Soc. Am. 56, 1043 - 1048. Benson, G. C., and Kiyohara, O. (1974). Tabulation of some integral functions describing diffraction effects in the ultrasonic field of a circular source. J. Acoust. Soc. Am. 55, 184- 185. Berlincourt, D. A., Curran, D. R., and Jaffe, H. (1964). Piezoelectric and piezomagnetic materials and their function in transducers. In "Physical Acoustics: Principles and Methods," Vol. 1A (W. P. Mason, ed.). Academic Press, New York. Berry, M. V. (1966). "The Diffraction of Light by Ultrasound." Academic Press, London. Bowen, L. J., Gentilman, R. L., Pham, H. T., Fiore, D. E, and French, K. (1993). Injection molded finescale piezoelectric composite transducers. Proc. 1993 IEEE Ultras. Syrup. 4 9 9 - 503. Cady, W. G. (1946). "Piezoelectricity." McGraw-Hill, New York. Carome, E. F., and Witting, J. M. (1961). Theory of ultrasonic attenuation in cylindrical and rectangular waveguides. J. Acoust. Soc. Am. 33, 187- 197. Cohen, M. G. (1967). Optical study of ultrasonic diffraction and focusing in anisotropic media. J Appl. Phys. 38, 3821-3828. Cohen, M. G., and Gordon, E. I. (1965). Acoustic beam probing using optical techniques. Bell Systems Tech. J. 44, 693 - 721. Davidson, P. K., and Alers, G. A. (1997). Proceedings of the ASNT Spring Conference, Houston, TX, March 1997. Also U.S. Patent No. 5,537,876. Dehn, J. T. (1960). Interference patterns in the near field of a circular piston. J. Acoust. Soc. Am. 32, 1692 - 1696. Del Grosso, V. A., and McGill, R. E. (1970). Sound speed dispersion in liquid cylinders. J. Acoust. Soc. Am. 48, 1294- 1296. DeSilets, C. S., Fraser, J. D., and Kino, G. S. (1978). The design of efficient broadband piezoelectric transducers. IEEE Trans. SU-25, 115 - 125. Dixon, R. W. (1967). Acoustic diffraction of light in anisotropic media. IEEE J QE-3, 85-93. Dixon, R. W. (1970). Acoustooptic inter-actions and devices. IEEE Trans. ED-17, 229-235. Fiore, D., Gentilman, R., Pham, H., Serwatka, W., McGuire, P., Near, C., and Bowen, L. (1997). 1-3 piezocomposites SmartPanels TM. SPIE Proc. Smart Structures and Materials 1997, Industrial and Commercial Applications of Smart Structure Tech. 391 - 396. Fitch, C. E. (March 1964). An optical schlieren system for ultrasonic imaging. Mater. Eval. 22(3), 124-127. Geng, X., and Zhang, Q. M. (1997). Evaluation of piezocomposites for ultrasonic transducer applications--influence of the unit cell dimensions and the properties of constituents on the performance of 2-2 piezocomposites. IEEE Trans. UFFC 44(5), 857-872.
2
Fabrication and Characterization of Transducers
131
Gentilman, R. L., Bowen, L. J., Corsaro, R. D., and Houston, B. H. (1995). Piezoelectric composite panels for underwater acoustic control. Proc. Design Eng. Tech. Conf. DE-84(2), 489-497. Gerdes, R. J., and Wagner, C. E. (April 1970). Scanning electron microscopy of oscillating quartz crystals. Proc. 3rd Ann. SEM Symp. IITRI, Chicago, IL. Gerdes, R. J., and Wagner, C. E. (April 1971). Study of frequency control devices in the scanning electron microscope. Proc. 25th Ann. Freq. Contr. Symp. U.S. Army Electronics Command, Ft. Monmouth, New Jersey. Goll, J., and Auld, B. A. (January 1975). Multilayer impedance matching schemes for broadbanding of water loaded piezoelectric transducers and high Q resonators. IEEE Trans. SU-22, 5 3 - 55. Greer, A. S., and Cross, B. T. (1970). Schlieren techniques for NDT. Nondestr. Testing 3, 169-172. Gururaja, T. R. (May 1984). Piezoelectric composite materials for ultrasonic transducer applications. Ph.D. Thesis, The Pennsylvania State University. Gururaja, T. R., Schulze, W. A., Cross, L. E., Newnham, R. E., Auld, B. A., and Wang, Y. J. (1985a). Piezoelectric composite materials for ultrasonic transducer applications. Part I: resonant modes of vibration of PZT rod-polymer composites. IEEE Trans. SU-32, 481-498. Gururaja, T. R., Schulze, W. A., Cross, L. E., and Newnham, R. E. (1985b). Piezoelectric composite materials for ultrasonic transducer applications. Part II: evaluation of ultrasonic medical applications. IEEE Trans. SU-32, 499- 513. Gururaja, T. R., Schulze, W A., Shrout, T. R., Safari, A., Webster, L., and Cross, L. E. (1981). High frequency applications of PZT/polymer composite materials. Ferroelectrics 29, 1245 - 1248. Hafner, E. (1974). Crystal resonators. IEEE Trans. SU-21, 220-237. Hashimoto, K. Y., and Yamaguchi, M. (1986). Elastic, piezoelectric and dielectric properties of composite materials. 1986 Proc. IEEE Ultras. Symp. 697-702. Hayward, G., Bennett, J., and Hamilton, R. (1995). A theoretical study on the influence of some constituent material properties on the behavior of 1-3 connectivity composite transducers. J. Acoust. Soc. Am. 98, 2187-2196. Hossack, J. A., and Hayward, G. (1991). Finite element analysis of 1-3 composite transducers. IEEE Trans. UFFC 38(6), 618 - 629. IEEE (1978). ANSI/IEEE Standard # 176-1978, "IEEE Standard on Piezoelectricity." IEEE (1987). Standard # 176-1987, "IEEE Standard on Piezoelectricity." Jaffe, H., and Berlincourt, D. A. (1965). Piezoelectric transducer materials. Proc. IEEE 53, 1372-1386. Jaffe, B., Cook, W. R., and Jaffe, H. (1971). "Piezoelectric Ceramics." Academic Press, New York and London. Janas, V. E, and Safari, A. (1995). Overview of fine-scale piezoelectric ceramic polymer composite processing. J. Amer Cer. Soc. 78(11), 2945-2955. Kino, G. S. (1987). "Acoustic Waves: Devices, Imaging, and Analog Signal Processing." PrenticeHall, Englewood Cliffs, New Jersey. Kino, G. S., and DeSilets, C. S. (1979). Design of slotted transducer arrays with matched backings. Ultras. Imag. 1, 189- 209. Klicker, K. A., Biggers, J. V., and Newnham, R. E. (January 1981). Composites of PZT and epoxy for hydrostatic transducer applications. J Am. Cer. Soc. 64(1). Kossoff, G. (March 1966). The effects of backing and matching on the performance of piezoelectric ceramic transducers. IEEE Trans. SU-13(2), 20-30. Krimholtz, R., Leedom, D., and Matthaei, G. (1970). New equivalent circuits for elementary piezoelectric transducers. Elect. Lett. 6, 398- 399. Lerch, R. (May, 1990). Simulation of piezoelectric devices by two- and three-dimensional finite elements. IEEE Trans. UFFC 37(3), 2 3 3 - 2 4 7 . Lindsay, R. B. (1960). "Mechanical Radiation." McGraw-Hill, New York.
132
Emmanuel P. Papadakis et al.
Lopath, E D., Park, S. E., Shung, K. K., and Shrout, T. R. (1996). Ultrasonic transducers using piezoelectric single crystal perovskites. Proc. l Oth IEEE Int. Symp. Appl. Ferroelectrics, 543 - 546. Lubitz, K., Wolff, A., and Preu, G. (1993). Microstructuring technology. Proc. 1993 IEEE Ultras. Symp. 513-524. Lum, P., Greenstein, M., Grossman, C., and Szabo, T. L. (1996). High-frequency membrane hydrophone. IEEE Trans. UFFC 43(4), 536-544. Mansour, T. M. (June 1979). Evaluation of ultrasonic transducers by cross-sectional mapping of the near field using a point reflector. Mater. Eval. 37(7), 50-54. Mason, W. P. (1948). "Electromechanical Transducers and Wave Filters," 2nd Edition. D. van Nostrand-Reinhold, Princeton, New Jersey, p. 205. Mason, W. P. (1950). "Piezoelectric Crystals and Their Application to Ultrasonics." Van Nostrand, New York. Mattiat, O. E. (1971). "Ultrasonic Transducer Materials." Plenum Press, New York and London. Maxfield, B. W., and Fortunko, C. M. (1983). The design and use of electromagnetic acoustic-wave transducers (EMATs). Mater. EvaL 41, 1399. Maxfield, B. W., Kuramoto, A., and Hulbert, J. K. (1987). Evaluating EMAT designs for selective applications. Mater. Eval. 45, 1166 - 1183. McKeighen, R. E. (March 1983). Basic transducer physics and design. Seminars in Ultrasound 4(1), 50-59. McMaster, R. C. (1959). "Nondestructive Testing Handbook, II," The Ronald Press Company, New York, Sec. 44, pp. 12 - 19. M.D. Editor (May 12, 1977). News Trends. Mach. Des. 8. Meeker, T. R. (1996). Publication and proposed revision of ANSI/IEEE standard 176-1987 ANSI/IEEE standard on piezoelectricity. IEEE Trans. UFFC 43(5), 717-772. Mezrich, R. S., Etzold, K. E, and Vilkomerson, D. H. R. (1974). System for visualizing and measuring ultrasonic wavefronts. RCA Rev. 35(4), 4 8 3 - 519. Mills, D. M., and Smith, S. W. (1996). Combining multi-layers and composites to increase SNR for medical ultrasound transducers. Proc. 1996 IEEE Ultras. Syrup. 1509 - 1512. Newman, D. R. (1973). Ultrasonic Schlieren system using a pulsed gas laser. IEEE Trans. SU-20, 282-285. Newnham, R. E., Skinner. D. P., and Cross, L. E. (1978). Connectivity and piezoelectric-pyroelectric composites. Mat. Res. Bull. 13, 525- 536. Oakley, C. G. (May 1991a). Analysis and development of piezoelectric composites for medical ultrasound transducer applications. Ph.D. Thesis, The Pennsylvania State University. Oakley, C. G. (May 1991b). Geometric effects of the stopband structure of 2-2 piezoelectric composite plates. Proc. 1991 IEEE Ultras. Symp. 657- 660. Oakley, C. G. (May 1997). Calculation of ultrasonic transducer signal-to-noise ratios using the KLM model. IEEE Trans. UFFC 44(5), 1018- 1026. Oakley, C. G., Huebner, W., and Liang, K. (1990). Design considerations for 1-3 composites used in transducers for medical ultrasonic imaging. Proc. 7th Int. Symp. Appl. Ferroelectrics. Onoe, M., Tiersten, H. E, and Meitzler, A. H. (1963). Shift in the location of resonant frequencies caused by large electromechanical coupling in thickness-mode resonators. J. Acoust. Soc. Am. 35, 36 -42. Papadakis, E. P. (1959). Correction for diffraction losses in the ultrasonic field of a piston source. J Acoust. Soc. Am. 31, 150-152. Papadakis, E. P. (1963). Diffraction of ultrasound in elastically anisotropic NaCI and some other materials. J. Acoust. Soc. Am. 35, 490-494. Papadakis, E. P. (1964). Diffraction of ultrasound radiating into an elastically anisotropic medium. J. Acoust. Soc. Am. 36, 414-422.
2
Fabrication and Characterization
of Transducers
133
Papadakis, E. P. (1966). Ultrasonic diffraction loss and phase change in anisotropic materials. J. Acoust. Soc. Am. 40, 863-876. Papadakis, E. P. (1969a). Effect of multimode guided wave propagation on ultrasonic phase velocity measurements: problems and remedy. J. Acoust. Soc. Am. 45, 1547-1555. Papadakis, E. P. (1969b). Variability of ultrasonic shear wave velocity in vitreous silica for delay lines. IEEE Trans. SU-16, 210 - 218. Papadakis, E. P. (1971 a). Effect of input amplitude profile upon diffraction loss and phase change in a pulse-echo system. J. Acoust. Soc. Am. 49, 166-168. Papadakis,. E. P. (1971b). Nonuniform pressure device for bonding thin stabs to substrates. J. Adh. 3, 181-194. Papadakis, E. P. (1972). Ultrasonic diffraction loss and phase change for broad-band pulses. J. Acoust. Soc. Am. 52, 847-849. Papadakis, E. P. (1975). Ultrasonic diffraction from single apertures with application to pulse measurements and crystal physics. In "Physical Acoustics: Principles and Methods," Vol. XI. (W. P. Mason and R. N. Thurston, eds.). Academic Press, New York, pp. 151- 211. Papadakis,. E. P., and Fowler, K. A. (1971). Broad-band transducers: radiation field and selected applications. J. Acoust. Soc. Am. 50(Pt. 1), 729-745. Papadakis, E. P. (1983). Use of computer model and experimental methods to design, analyze, and evaluate ultrasonic NDE transducers. Mater. EvaL 41(2), 1378-1388. Papadakis, E. P., and Meeker, T. R. (Sept. 24-26, 1969). Digital pulse propagation for pulse-echo measurements. Paper I • 5, 1969 IEEE Ultras. Symp. Pauer, L. A. (1973). Flexible piezoelectric material. IEEE Int. Conv. Rec., 1 - 5. Pazol, B. G., Bowen, L. J., Gentilman, R. L., Pham, H. T., Serwatka, W. J., Oakley, C. G., and Dietz, D. R. (1996). Ultrafine scale piezoelectric composite materials for high frequency imaging arrays. Proc. 1996 IEEE Ultras. Symp. 1263-1268. Posakony, G. J. (1981). Private communication. Redwood, M. (1963). A study of waveforms in the generation and detection of short ultrasonic pulses. Appl. Mater. Res. 2, 7 6 - 84. Reynolds, P., Hyslop, J., and Hayward, G. (1996). The influence of constructional parameters on the practical performance of 1-3 composite transducers. Proc. 1996 IEEE Ultras. Symp. 967-970. Roderick, R. L., and Truell, R. (1952). The measurement of ultrasonic attenuation in solids by the pulse technique and some results in steel. J. Appl. Phys. 23, 267-279. Savakas, H. P., Klicker, K. A., and Newnham, R. E. (1981). PZT-epoxy piezoelectric transducers: a simplified fabrication procedure. Mat. Res. Bull 16, 677-680. Schoch, A. (1941). (No title given), Akust. Z. 6, 318-337. Seki, H., Granato, A., and Truell, R. (1956). Diffraction effects in the ultrasonic field of a piston source and their importance in the accurate measurement of attenuation. J. Acoust. Soc. Am. 28, 230- 238. Selfridge, A. R., and Gehlbach, S. (1985). KLM transducer model implementation using transfer matrices. Proc. 1985 Ultras. Symp. San Francisco. Selfridge, A. R., and Khuri-Yakub, P. (1997). Co-inventors, partial rights assigned to American Technology Corporation, San Diego, CA. Sevick, J. (1987). Transmission line transformers. ARRL (1987). Shaulov, A. A., Smith, W. A., and Signer, B. M. (1984). Performance of ultrasonic transducers made from composite piezoelectric materials. Proc. 1984 IEEE Ultras. Symp. 545- 548. Shui, Y., Geng, X., and Zhang, Q. M. (1995). The theoretical modeling of resonant modes of composite ultrasonic transducers. IEEE Trans. UFFC 42(4), 766-773. Sittig, E. K. (1972). Design and technology of piezoelectric transducers for frequencies above 100 MHz. In "Physical Acoustics: Principles and Methods, IX" (W. P. Mason and R. N. Thurston, eds.). Academic Press, New York.
134
Emmanuel R Papadakis et al.
Smith, W A., (1989). The role of piezocomposites in ultrasonic transducers. Proc. 1989 IEEE Ultras. Syrup. 755 - 766. Smith, W. A., and Auld, B. A. (1991). Modeling 1 - 3 composite piezoelectrics, thickness-mode oscillations. IEEE Trans. UFFC 38( 1), 40 - 47. Smith, W. A., Shaulov, A. A., and Singer, B. M. (1984). Properties of composite piezoelectric materials for ultrasonic transducers. Proc. 1984 IEEE Ultras. Symp. 539- 544. Strutt, J. W., Lord Rayleigh (1945). "The Theory of Sound," 2nd Edition. Dover Publications, New York, p. 105. Takeuchi, H., Nakaya, C., and Katakura, K. (1984). Medical ultrasonic probe using PZT/polymer composite. Proc. 1984 IEEE Ultras. Symp. 507 - 510. Thompson, R. B. (1990). Physical principles of measurements with EMAT transducers. In "Physical Acoustics: Principles and Methods," Vol. XIX (R. N. Thurston and A. Pierce, eds.). Academic Press, New York, pp. 157-200. Ting, R. Y. (1986). Evaluation of new piezoelectric composite materials for hydrophone applications. Ferroelectrics 67, 143. von Gutfeld, R. J. (Oct. 26-28, 1977). Thermoelastically generated MHz elastic waves from constrained surfaces. Paper G-1 presented at the 1977 Ultrasonics Symposium, Phoenix, Arizona. Whaley, H. L., Cook, K. V., McClung, R. W., and Snyder, L. S. (May 1967). Optical methods for studying ultrasonic propagation in transparent media. Proc. 5th Int. Symp. on Nondestr. Testing. Wojcik, G. L., Vaughan, D. K., Abboud, N., and Mould, Jr., L. (1993). Electromechanical modeling using explicit time-domain finite elements. Proc. 1993 IEEE Ultras. Syrup. 1107- 1116. WIS (1997). "LambSolver, Version 2.1." WIS, Inc., 117 Quincy Street, N.E., Albuquerque, New Mexico. Wyatt, R. C. (Sept. 1975). Imaging ultrasonic probe beams in solids. Brit. J. Nondestr. Testing 17, 133140. Zola, J. (April, 1985). "Method for Fabricating Composite Transducers: Laminated Piezoelectric Material and Passive Material." U.S. Patent No. 4,514,247. Zola, J. (February 1986). "Transducer Comprising Composite Electrical Materials." U.S. Patent No. 4,562,981.
3 Surface Acoustic Wave Technology Macrosuccess through Microseisms FRED
S. H I C K E R N E L L
Motorola Inc., Space and Systems Technology Group, Scottsdale, Arizona I. II. III. IV. V.
VI.
VII.
VIII.
IX. X. XI. XII.
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Measures o f Success . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Surface Elastic Waves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Prelude to the SAW Era (The Early R u m b l i n g s ) . . . . . . . . . . . . . . . . . . . The Interdigital Transducer, Materials, and Fabrication . . . . . . . . . . . . . . . . A. The Interdigital Transducer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. SAW Materials C. Processing and Fabrication . . . . . . . . : .................... Interdigital Transducer Controlled SAW Devices . . . . . . . . . . . . . . . . . . . A. The Two-Port Delay Line . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. The Multiple-Port Delay Line . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. The Bandpass Filter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . " .... D. SAW Resonators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E. SAW Oscillators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Electrode Configured Matched Filter Devices . . . . . . . . . . . . . . . . . . . . . A. Correlators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Pulse Expander-Compressors . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. The Programmable Transversal Filter . . . . . . . . . . . . . . . . . . . . . . . . Signal Processing T h r o u g h the Passive Control o f SAW Propagation . . . . . . . . A. Waveguides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. The Multistrip Coupler . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Reflection Gratings . . . . . . . . . . ; ....................... D. U n i f o r m Dielectric F i l m Layers . . . . . . . . . . . . . . . . . . . . . . . . . . . Acoustoelectric Signal Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acousto-optics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . SAW Sensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Future Success . . . . . . . . . . . . . . . . . . . . . . . . . . . . .......... Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A p p e n d i x A. SAW Publications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A p p e n d i x B. SAW Conferences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A p p e n d i x C. SAW Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . A p p e n d i x D. Worldwide SAW Activities . . . . . . . . . . . . . . . . . . . . . . . . A p p e n d i x E. The SAW Engineer's Role as an Artisan . . . . . . . . . . . . . . . .
136 138 141 145 148 148 149 153 156 156 159 160 167 169 170 171 172 173 174 175 177 179 182 183 186 186 187 189 190 194 197 203 204 206
135 PHYSICAL ACOUSTICS, VOL. XXIV
Copyright 9 1999 Academic Press All rights of reproduction in any form reserved. ISBN 0-12-477945-X $30.00
136
Fred S. Hickernell
This chapter gives an overview of three decades of technology development in surface acoustic wave (SAW) ultrasonics. SAW technology, as applied to modern electronic systems, was born with the concept of a thin-metal interdigital electrode transducer (IDT) on a polished piezoelectric plate and spent its youth exploring the limits of time and frequency domain signal processing functions. Since then it has matured as a manufacturing technology in consumer electronics, found economic success in frequency selectivity for telecommunications, and continues to grow and support a variety o f wireless and sensor applications. Its secret to success has been the slow wave velocity accorded elastic waves, its accessibility to surface displacements and electric fields, its passive device nature, its high-frequency capability, the availability of a large dynamic range, and the simplicity of its manufacture. A technology with an explosive beginning, it has evolved into a respected and much needed component for time and frequency control in electronic systems and has a promising applications-filled future.
I.
Introduction
It is undeniable that surface acoustic wave (SAW) devices have had overwhelming success in an electronic world where integrated digital semiconductor circuits dominate. SAW devices exist in nearly every aspect of your life. Throughout your home, they are tucked away in your television set, your cable box, your cordless phone, and your audio system. As you exit your home, SAW as resonant frequency control elements provide the means to unlock your car and swing your garage door open. On your drive to work, a SAW-controlled vehicle ID tag allows your automobile to move swiffiy by the toll gates. Your cellular satellite phone with its array of radio frequency (RF), interstage, and intermediate frequency (IF) SAW filters place you in contact with people throughout the world. A SAW ID tag lets you into the office. At work, SAW devices control your computer, wireless local area network, and measurement instruments. Meanwhile, overhead sophisticated SAW devices in avionics and satellites that process radar and electronic intelligence signals protect your freedom. SAWs stand ready at your country's boundaries as frequency control elements of a missile intercept system. Secure communications in the presence of noise are assured through SAW matched filters. Orbiting satellites carry SAW devices as filters for communications and signal processing elements to monitor the heavens above and the earth below. They exist as critical filter elements in deep space satellite transponder equipment throughout our solar system and beyond, helping to display our neighboring planetary systems and to map the features of our solar system. In short, SAW filters can be found from within a few inches of your heart tucked away in your shirt-pocket phone to millions of miles away in outer space.
3
Surface Acoustic Wave Technology Macrosuccess
137
The sections that follow document, from the author's experience and perspective, the evolutionary success of SAW devices. First we consider the question, how do you measure the success of SAW technology? Then we look at the properties and diversity of surface acoustic waves and the prelude that led to the SAW era. This is followed by the explosion of SAW application ideas engendered by the concept of the interdigital transducer (IDT) electrode on a polished piezoelectric plate with the control of the phase and amplitude of signal waveforms leading to electrode-controlled signal processing devices. SAW propagation control in the space between IDT electrodes led to expanded time-bandwidth device concepts and practical device implementations for signal processing. The surface waveguide was envisioned as a major step toward microminiaturization of circuits, which were analogs of mircowave circuits. The interaction of the traveling stress field and accompanying electric field at the surface of a piezoelectric plate with an adjacent semiconducting layer led to the development of active acoustoelectric devices including amplifiers, convolvers, and correlators. Acousto-optic and acoustomagnetic devices were also developed. Initially, there were strong military and government incentives to explore the limits of time and frequency domain signal processing. Consumer electronics companies reaped the fruit of the volume production of inexpensive SAW filter elements to replace tunable coils and capacitors. SAW resonators and oscillators were developed as high-frequency replacements for traditional bulk acoustic wave (BAW) counterparts. Today, low-loss, lowcost SAW filters have met the challenge of the wireless communications revolution, which continues to grow and expand. The concepts and development of SAW sensors are growing. The spread spectrum and matched filter concepts developed earlier are being revisited as a means of complementing their digital signal counterparts. All in all, SAW technology is continuing to grow and expand. It is an exciting field that the author has been blessed to be a part of for over 30 years. It is a field difficult to encompass with any brief set of words and a bibliography. The published work, the conference presentations, the theses, the patents, and the product catalogs are vast. The author cannot possibly acknowledge all the contributors, their contributions, and their respective organizations, and for this he is very apologetic. The photographs chosen to illustrate device concepts in this Chapter are of SAW devices developed at Motorola. Over the past 30 years, the SAW applications at Motorola evolved from military and space applications to commercial and consumer telecommunication products. The photographs have been chosen to acquaint any
138
Fred S. Hickerneil
reader who is new to the field with the basic device concepts that have impacted the electronics world. Although the written word cannot thoroughly convey the heart and soul of SAW technology and the personalities involved, perhaps some sense of the adventure leading up to the present day activities can be imparted. The SAW community consisting of scientists, engineers, technicians, and support personnel have made the success possible. Some made significant early contributions and then moved on to other fields. Many of the original workers are still making significant contributions, having tied their lifetime careers to the SAW field. New workers are bringing fresh ideas and insights into the field. It is with great admiration that this article is dedicated to the international SAW community and, especially, to the numerous workers I have met and interacted with over these past 32 years. II.
Measures of Success
In its early years SAW technology provided intrinsic and extrinsic measures of success. The researcher found satisfaction in discoveries that have added to the scientific understanding of the basic properties of matter. The university professor found a wide variety of useful problems in SAW technology that were fundable and publishable. The graduate student found a research project guaranteeing a thesis, a graduate degree, and job opportunities. The government, faced with real-time processing of a wide range of signal waveforms, funded and developed devices with long delay times and large time-bandwidth products. Systems engineers incorporated SAW devices with their electronics to realize unique signal processing functions. Industry- and government-supported laboratories and institutes developed SAW components and modules to enhance their radar and communication system products. Consumer and commercial product businesses produced SAW devices as replacements for existing electronic counterparts and, by so doing, reduced labor and component costs. The early expectations for SAW devices were met and thus contributed to their success. Today SAW devices have moved into a world of competing technologies. A systems designer or producer chooses SAW devices using the following measures: (1) Will it perform to specification? (2) Will it be of good quality and reliable? (3) Is it the only or best alternative to other devices? (4) Does it satisfy form factor and weight requirements? (5) Will it cost less than other alternatives? In a very dynamic electronics market the user, who may have alternative device sources, will also measure success on how quickly the
3
Surface Acoustic Wave Technology Macrosuccess
139
manufacturer responds with a working prototype. On the other side, the SAW device manufacturer looks for market opportunities where volume production will yield a high return. He or she looks for economic success and promotes the technology into as many applications as possible. Economic success follows if the manufacturer totally satisfies the customer with existing products and then anticipates and promotes the customer's future needs. Following are some quantitative measures to use as a check on the present health of SAW technology and the outlook for its continued success. 1. A Vibrant Technology as Measured by Publications and Conferences. Publications and conferences reflect the fact that new ideas and device implementations are being generated and shared. There have been well over 20,000 SAW-related publications indexed since 1965. The total number of SAW-related publications abstracted from scientific and engineering journals worldwide since 1975 has averaged between 700 and 800 per year. Appendix A lists selected books, monographs, and review articles relating to SAWs, together with graphs on SAW journal and theses publications. International symposiums with SAW-related sessions have occurred in China, Western and Eastern Europe, Japan, Russia, the United States, and Canada within the past five years. The annual IEEE International Ultrasonics Symposium, sponsored by the Ultrasonics, Ferroelectrics, and Frequency Control Society (UFFC-S), has been the leading conference for SAW presentations. Since 1970, an average of 80 SAW-related papers have been presented and published in conference proceedings each year. Universities continue to account for one-third to one-half of the papers. Since 1990, most of the SAW papers presented at Ultrasonic Symposiums have come from outside the United States. In response to this, the symposium is now moving to sites outside the United States. Appendix B describes conference and symposium activities and their related statistics. 2. The Maintenance o f a High Level o f Patent Activity. Over 5000 SAW patents have been issued worldwide. Patents issued today range from 200 to 300 per year, indicating that new ideas continue to flow. SAW patents are now being filed primarily by SAW manufacturing businesses to protect the unique design geometries, fabrication procedures, and crystal cuts for their products. Some manufacturers prefer to protect their intellectual property through closely guarded trade secrets. These trade secrets are easily equal in number to the patented ideas.
140
Fred S. Hickernell
3. Acceptance as an Important Technology for Electronic Systems Development. In their early development, SAW devices were found to be an important component replacement for bulkier and more costly electronic components such as delay lines and filters. Their time delay and frequency capabilities in a small form factor made them ideally suited to perform complex signal processing functions that digital techniques could not realize. Today, they are well entrenched in consumer and commercial products. The more costly and complex military-related signal processing SAWs are now a small part of SAW technology development. New opportunities for sensors and improved telecommunications products will lead the way in the future as consumer and commercial products continue their growth. Presently, there are more than 500 million SAWsproduced annually, with the majority going into video and telecommunication products. This represents a billion dollar business. It is predicted that in telecomunications alone, the required number of filters in the year 2001 will be over 600 million. More than 400 million SAW filter and resonator components for commercial and consumer products will also be needed. Present and future application areas are noted in Appendix C. 4. The Number of Businesses Being Sustained in the Manufacture of SAW Products. There are more than 50 SAW companies supplying a worldwide market and many smaller companies engaged in the production of specialty SAW products. Japan claims the most SAW businesses, many of which are embedded in larger industries with a large market share. With 500 million SAW devices now being produced each year and a predicted average growth of around 20 percent per year, the number of employees engaged in the SAW business area will continue to increase. Currently the number of employees engaged in SAW device design, processing, manufacturing, and support functions is well over 3000. Some companies now facing capacity problems have the difficult choice of whether to expand facilities (a costly capital expenditure) or to find manufacturers with semiconductor process equipment willing to produce wafers in a foundry mode of operation. Appendix D presents a list of present suppliers of SAW devices. 5. The Recognition Given Scientists and Engineers for Work in SAW Technology. National and international awards and recognitions have been given to several SAW scientists and engineers. A monarch, an emperor, and presidents have recognized SAW technology contributors. A number of these distinguished individuals have been elected to the National Academies
3
Nurface Acoustic Wave Technology Macrosuccess
141
of Science and Engineering of their respective countries and several have received awards from professional engineering and science organizations. Over 40 contributors from several different countries have been elected as Fellows of the Institute of Electrical and Electronic Engineers (IEEE). The success of SAW goes far deeper than the facts and figures of yesterday and today. SAWs have succeeded because talented people with vision have worked hard to make their dreams come true. They were caught up in an exciting technology with expanding horizons. The concept of a simple transducer structure and an elastic wave released from the inner confines of a solid brought out the creative spirit in countless scientists and engineers. A saying attributed to G. K. Chesterton is, "The whole difference between construction and creation is this: that a thing constructed can only be loved after it is constructed; but a thing created is loved before it exists." This is the creative spirit that has permeated SAW technology development. There is the satisfaction that your tiny SAW component, interrupting the electron flow in a maze of conductors and integrated circuits, performs a time or frequency function that brings better communications to the world, improves the standard of living, secures a more healthy environment, gives a greater understanding of the world and a broader view of the universe, and even saves lives. Then there are the side benefits of such a technology, which is recognized and applied worldwide and which brings people together in cooperative and collaborative efforts. New friendships are formed, barriers are broken down, and the world is brought closer together in peace and understanding. This is the heart and spirit of SAW technology. Creativity, innovation, cooperation, and just plain good business sense have played important roles in making SAW technology the success it is today.
III.
Surface Elastic Waves
Surface elastic waves have been rumbling across our planet for millennia, signaling the movement of the earth's crust as it rearranged to form our present-day land and sea masses. The major destructive component in these earthquake waves are elastic waves traveling across the earth's surface. Volcanic action generates surface waves that can be analyzed by scientists to predict potential eruptions. Heroes and villains of the old Western movies dropped to their knees and put their ears to the ground or to the railroad track to detect the surface waves made by the hoofbeats of approaching horses or an oncoming train. Heavy vehicles and modern-day construction equipment that disturb the underlying surface continually generate surface elastic waves.
142
Fred S. Hickernell
Surface elastic waves characteristically travel along the surface of a solid with wavelengths that can extend over 10 orders of magnitude from a few kilometers (nature's seismic waves) to submicrometers (high-frequency surface acoustic waves on a piezoelectric crystal). The waves characteristically exhibit decay in amplitude with depth and therefore are confined within a few wavelengths of the surface. At a free surface, they exhibit a nondispersive slow-wave-velocity characteristic dependent on the density and elastic constants of the solid. An average of 3.3 million detectable macroseisms (earthquakes) of magnitude 1.0 or greater occur annually while 3.3 billion microseisms (SAWs) now occur daily. Nondestructive evaluation (NDE) of plate structures using surface waves was an early useful application of SAW properties in ultrasonics. Today SAWs are used in acoustic microscopy to study the features of surfaces and obtain quantitative physical properties. The naturally occurring elastic waves induced by thermal processes at the surface of a solid are optically probed to obtain the basic elastic properties of materials. In 1885, it was John Strutt, the third Baron Lord Rayleigh, whose mathematics demonstrated the existence of wave propagation confined to the surface of an elastic solid [ 1]. Rayleigh's mathematics defined a nondispersive acoustic wave propagating along the stress-free boundary of a semiinfinite elastic half-space with the energy confined at the surface. He insightfully suggested that these waves, now called Rayleigh waves, were associated with earthquakes. A cross section of the displacements for a Rayleigh wave in units of wavelength in the sagittal plane near the free surface of an isotropic solid is shown in Fig. 1. The polarization of surface waves can be conveniently described in terms of the sagittal plane defined by the normal to the surface and the propagation direction of the surface wave. Tracing the movement of the wave, the displacements at the surface are retrograde elliptical containing a strong shear component normal to the direction of propagation and compression-extension along the surface in the direction of propagation. Within a quarter-wavelength depth, the longitudinal component changes phase and the elliptical motion reverses direction. An exponential decay of the displacements with depth occurs and a large fraction of the energy is confined within a wavelength of the surface. It was Rayleigh wave propagation along the principle axes of piezoelectric crystals that was used for early SAW device development. From this modest beginning, a rich variety of surface elastic waves were predicted and have reached their pinnacle of existence in recent times at
3
Surface Acoustic Wave Technology Macrosuccess
143
~Z 0.25
ug
tu >,<
0.5
,,I,, I,,n 0.75 uJ C1
1.0
!
L 0
0.25
v
v
0.5
v
0.75
1.0
1.25
DISTANCE (WAVELENGTH) Fig. 1. A cross section of the particle displacements in a Rayleigh wave traveling along the surface of an isotropic solid.
microwave acoustic frequencies with piezoelectric crystalline solids. In 1905, A. E. H. Love identified a surface wave with displacements transverse to the sagittal plane that occurred when a slower-velocity layer overlaid a fastervelocity substrate. Six years later (1911), Love published a comprehensive monograph on geodynamics, which included the earlier work [2]. Such waves, which are frequency dispersive, have been used as waveguide structures and dispersive delay lines for surface acoustic wave devices. In 1917, H. Lamb considered the propagation of waves in elastic plates with finite thicknesses [3]. Such plates can support a number of modes, symmetric and antisymmetric, depending on the ratio of plate thickness to acoustic wavelength. Such structures and waves have been used in recent years for SAW sensor applications. Robert Stoneley contributed to the mathematical understanding of elastic wave propagation at the boundary between two adjoined semi-infinite solids in 1924 [4]. In addition, he was the first to introduce the concept of crystalline anisotropy into the computation of elastic surface wave characteristics [5]. This set off a flurry of mathematical investigations of the existence of surface waves on anisotropic surfaces. Kraut [6] has detailed the mathematical work between 1955 and 1968 in which several authors dealt with the propagation of surface elastic waves in anisotropic crystalline media. Although certain authors claimed that there were forbidden directions in which surface waves did not exist, Lim and Farnell [7] carried out a detailed computer search for these forbidden
144
Fred S. Hickernell
directions. They determined that there is always a surface wave solution satisfying the free surface boundary condition. The bottom line--surface acoustic waves can exist on any solid, with the propagation properties being determined by the crystallographic orientation. Surface acoustic waves on a piezoelectric solid have unique features not found on nonpiezoelectric solids. On a piezoelectric solid, a traveling electric field accompanies the surface wave, which facilitates the means for excitation and detection and interactions with electromagnetic fields and electrons. The wave velocity depends on the piezoelectric, dielectric, and elastic properties of the solid and their temperature and pressure dependencies. The presence of piezoelectricity can cause an oscillatory decay of displacement amplitude with depth. A propagation feature unique to piezoelectric solids, having no nonpiezoelectric counterpart, is the existence of waves that can propagate with displacements entirely transverse to the sagittal plane in certain crystal cuts. The prediction of these waves, called Bleustein-Gulyaev (BG) and Bleustein-Gulyaev-Shimizu (BGS) waves, occurred in the late 1960s. Their independent discovery was reported in published articles in the United States [8], in Russia [9], and in Japan [10]. The U.S. and Russian papers were theoretical; the Japanese paper included experimental verification of the existence of the waves. The transverse displacements of the BGS waves extend many wavelengths below the surface and have the advantage of no mode conversion upon reflection from a substrate edge. The waves have been experimentally investigated in hexagonal crystals and in lead zirconate titanate (PZT) ceramic materials. One aspect of the variety of propagation modes that has become important to SAW devices are the piezoelectrically active pseudo-SAW (PSAW) or leaky surface waves, which propagate with a velocity higher than that of the slowest transverse bulk wave and radiate elastic energy into the interior of the solid. On certain cuts of piezoelectric crystals, these waves can have high coupling factors and high velocities, and their leaky wave attenuation can be minimized. In recent years, such cuts on lithium tantalate and lithium niobate have been extensively used for bandpass filter development and manufacture. There also exist high-velocity pseudo-SAWs (HVPSAWs) with velocities near those for longitudinal bulk mode propagation. Various characterization schemes have been used to classify the SAW propagation modes. Farnell and Adler have classified the various wave types according to crystal symmetry for both piezoelectric and nonpiezoelectric crystal classes [11]. The rigorous definition relating to the condition of nonzero terms in the elastic and piezoelectric matrices is given in their
3
Surface Acoustic Wave Technology Macrosuccess
145
paper together with type classifications for principle propagation directions in the various crystalline solids. Their classification scheme is helpful for understanding the various names and acronyms encountered in the literature for particular surface acoustic waves. In summary, there is a wide variety of surface waves that can be generated on piezoelectric crystals and that have been used extensively for device applications. A Rayleigh wave propagates when the sagittal plane is a symmetry plane and there are only two orthogonal displacement components. If the sagittal plane is not a symmetry plane, the particle motion of the surface wave may include three displacement components, which characterize a generalized SAW (GSAW) with pseudo-SAW wave propagation. The presence of a film or grating can alter the propagation loss characteristics and thereby reduce the effect of leaky wave losses into the bulk of the crystal. While Rayleigh wave propagation remains a strong mode for device development, pseudo-SAW and leaky wave modes have captured the attention of the SAW designers of telecommunication filters in recent years. The shallow bulk acoustic wave, (SBAW), surface-skimming bulk wave (SSBW), and surface transverse wave (STW) represent surface elastic waves with dominant shear horizontal displacements. The presence of a film layer (e.g., metal or dielectric) or topographical grating can confine these waves near the surface. Bleustein-Gulyaev-Shimizu waves are shear horizontal waves with deep penetration below the surface of the displacement. Placing a thin, slowvelocity film layer on a fast substrate can result in a generalized surface wave or a special Love mode with transverse displacements. Finally, these are the plate modes or Lamb waves confined within a plate with free boundaries on either side that exist in membrane structures. Thus the SAW designer has a variety of surface elastic waves with which to fashion devices for use in electronic systems development.
IV.
Prelude to the SAW Era (The Early Rumblings)
The 20-year period before 1960 wimessed a growing interest in surface and plate waves for nondestructive evaluation (NDE) but little interest in electronic applications. The first indications of possible electronic applications was in the early 1960s. In Russia, there was activity in the investigations of transduction techniques for the generation and detection of surface waves above 1 MHz and the investigation of the velocity and attenuation properties of surface waves on solids [12]. Victorov [13] records experiments with early transducers using bulk wave transducer geometries. Some typical transduction
146
Fred S. Hickernell
schemes that use bulk mode transducers for surface wave generation are shown in Fig. 2. The most efficient were those that relied on the creation of a periodic deformation of the surface of the solid. The wedge technique, using a shear mode bulk transducer, was the most efficient and flexible method for the excitation of surface waves and was used extensively in NDE work. Work was underway in the early 1960s at Bell Laboratories on the concept of guided surface wave components, the analogs of microwave strip-line components. In Great Britain, Mortley [ 14] had developed an interdigital transducer concept for generation of bulk waves with a dispersive delay capability. It was this transducer concept, when applied to surface waves, that would launch the SAW era. The introduction of the concepts of modem day SAW devices were first proposed in the United States at the Ultrasonics Symposiums of 1964 and 1965. The concept of a dispersive delay line with metallic grating lines for detection was proposed by John Rowen [15] in the 1964 Ultrasonics Symposium. John May Jr. [16] showed a picture of Rowen's surface wave delay line and its realizations with measured data by E. K. Sittig in 1965. Details of the device are discussed in Rowen's patent filed in December of 1963 [ 17]. It was reported that the basic idea of a tapped delay line similar in structure with surface electrodes was reduced to practice in Russia at the Academy of Aerospace Instrumentation in Leningrad in the early 1960s, but this was not published or patented for security reasons. SAW work at the
Jt .
.
.
.
~"","'""-"'"
.
.
.
.
.
:'"l','r: :':":: :"]
v
,
,
Fig. 2. Typical transduction schemes that used bulk acoustic wave transducers for the excitation of surface acoustic waves.
3
Surface Acoustic Wave Technology Macrosuccess
147
institute was not reported in the open literature until 1965 [18]. At the December 1965 Ultrasonics Symposium, R. M. White described surface wave amplification and attenuation using transducers consisting of vacuum-deposited interleaved conducting combs. While the results of an acoustoelectronic interaction were interesting, the interdigital transducer concept used to generate the surface waves, which had appeared in the literature two months earlier, was what caught the interest of the listeners. As often happens, the birth of a technology can be associated with a simple idea and a subsequent experimental event. In 1965, when White and Voltmer [ 19] reported results of the launching of a surface wave on quartz with a thin film aluminum interdigital transducer, those working on microwave acoustics realized the possibilities presented by this simple, efficient way of controlling surface waves on a piezoelectric crystal. The time domain was captured in millimeter space through surface waves traveling at acoustic velocities, and this new technology was nurtured by those who saw great potential for its signal processing and filter capabilities. The budding high-frequency surface wave technology had an identity crisis (usually one names the baby at birth). It was being wrested from the hands of mathematicians, seismologists, and low-frequency NDE proponents for its nurture and growth, into the hands of electrical engineers for application to electronic systems developments. Up until around 1970, the terms ultrasonic Rayleigh waves, surface elastic waves, elastic surface waves, acoustic surface waves, and surface acoustic waves were in vogue. (It never suffered from the painful history of microwave acoustics, which went through such names as hypersonics, supersonics, and pretersonics.) Surface elastic waves (SEW) and surface acoustic waves (SAW) were the most popular terms, and it was SAW that became the accepted acronym (with only a passing reference to its being the more masculine sounding). The care, feeding, and growth of the new technology was undertaken by universities, government laboratories, industry, and institutes of science all over the world. In universities, mathematics and physics departments saw opportunities for an exhaustive extension of the theoretical existence of surface waves on crystals and experimental work on propagation properties. University engineering departments, along with industry, concentrated on design models for transducer structures and means of controlling surface wave propagation. Government laboratories funded and also explored signal processing applications for military purposes. Industry looked to SAW for components to complement their electronic communication and radar systems. Out of the theoretical and experimental work came devices that
148
Fred S. Hickerneil
extended and improved the signal processing functions of electronic systems. This all occurred within the first ten years with an explosive growth in publications and presentations beginning in 1970. After this time, business units within and outside industry entered the picture with volume applications for the technology to further its growth. Consumer electronics and wireless telecommunications spurred the robust growth of the technology as it stands today.
V. A.
The Interdigital Transducer, Materials, and Fabrication
THE INTERDIGITALTRANSDUCER
To gain some appreciation for the simplicity and versatility of the role of the interdigital transducer introduced by White and Voltmer [19] for surface acoustic wave technology, consider the representation of a SAW device shown in Fig. 3. The generating structure consists of a series of interdigital metallic lines fabricated on the surface of a polished piezoelectric plate. Alternate lines are connected together to form a one-terminal-pair driving point. When this array is driven at a frequency such that the physical distance between alternate lines corresponds to the wavelength of the surface wave, a strong bidirectional
ACOUSTIC SUR FACE WAVE
ELECTRICAL OUTPUT
ACOUSTIC ABSORBER
ELECTRICAL INPUT
INTERDIGITAL ELECTRODE
PIEZOELECTRIC PLATE
Fig. 3. SAW device configuration with interdigital transducer electrodes on a piezoelectric substrate.
3
Surface Acoustic Wave Technology Macrosuccess
149
surface wave emanates from the transducer. This surface wave, as indicated previously, is characterized by elastic energy confinement in depth within the first few wavelengths of the surface, retrograde elliptical motion of particles at the surface, and low-loss, nondispersive slow wave velocity. The frequency response of the transducer is a design variable; it can be controlled by the number of finger pairs, the width and spacing of adjacent fingers, and the overlap region of the fingers. Thus, there is a two-dimensional geometric design freedom for the electrodes that corresponds to specific time and frequency responses in the electronic domain. Basically, there is a one-toone correspondence between the IDT spatial geometry and the time domain. The frequency domain is but a Fourier transform away. Modeling of the interdigital transducer structure and device-associated propagation geometries has been a very intense and very important part of the successful development of SAW devices. Transducer modeling has developed from simple and highly effective equivalent circuit models [20, 21 ], through impulse response [22, 23] and coupling of mode models [24], to computer-intense Green's Function modeling [25]. A simple and effective closed-form mathematical model was also developed for bandpass filter design [26]. Ongoing efforts continue on the modeling of second-order effects such as triple transit echo, reflections between electrodes, bulk wave scattering, electromagnetic feedthrough, and diffraction of waves. Today the effects of wire bonds, parasitics, component matching, and package structures are being incorporated into the models for SAW devices. Most companies have developed their own specific proprietary software for the computer-aided design of the transducer structures for production SAW components. An assessment of various transducer modeling techniques was reported by Ruppel et al. [27]. B.
SAW MATERIALS
One of the most important features of device modeling is to have accurate data on the materials being used for the devices. SAW device technology requires the characterization of piezoelectric and associated materials. Various crystalline and ceramic piezoelectrics have been investigated for SAW device development. The piezoelectric ceramics possess high coupling factors but can only be used at very low frequencies (usually less than 100 MHz) because their granular nature causes scattering of the waves and resulting high propagation losses. Crystalline materials are much better suited for development because of their high material Q. Figure 4 shows a photograph of a
150
Fred S. Hickernell
Fig. 4. Czochralski-grownboule of lithium niobate.
boule of lithium niobate, which along with cultured quartz, is one of the most commonly used SAW materials. The original IDT experiments by White and Voltmer in 1965 [19], were done on a Y-cut, Z-propagating quartz plate. SAW experiments that shortly followed used the higher-coupling-factor lithium niobate (LiNbO3) and lithium tantalate (LiTaO3) substrates, also using Y-cut, Z-propagating plates. The SAW world was the beneficiary of the development and characterization of LiNbO3 and LiTaO3 by Bell Laboratories in the 1960s. These three crystalline materials have been extensively used for SAW device development over the past 33 years with an extension away from standard cuts to rotated cuts. The move to rotated cuts takes advantage of velocity, coupling factor, propagation loss, temperature, and pressure enhancements. Quartz is favored where narrowband, high-temperature stability operation is required in resonators and filters. Lithium niobate is used for low-loss broadband applications to delay lines and filters. Lithium tantalate has properties intermediate between the two in terms of coupling efficiency and temperature coefficient
3
Surface Acoustic Wave Technology Macrosuccess
151
of frequency. Bismuth germanium oxide (Bi12GeO20), lithium tetraborate (LizB407), berlinite (A1PO4), and langasite (La3GasSiO14) are other highresistivity piezoelectric materials that have been used as SAW substrate materials. Piezoelectric semiconductors such as gallium arsenide (GaAs) and cadmium sulfide (CdS) have also been used where the combination of piezoelectrically generated fields and electron mobility can be used for acoustoelectronic interactions. The earliest cuts used SAW propagation along crystal symmetric axes. Shortly thereafter rotated cuts, which enhanced particular properties, were used. The Y-rotated X-propagating ST cut of quartz identified by Schulz et al., at Raytheon [28] supplanted the Y-Z cut, taking advantage of its first-order zero temperature coefficient of frequency. The Y-cut, Z-propagating Rayleigh wave on lithium niobate was extensively used for broadband delay lines and filters and particularly for large time-bandwidth reflective array compressors. The 128 ~ Y-rotated, X-propagating cut of lithium niobate identified in 1976 [29] had a higher velocity and coupling efficiency than the Y-Z cut. The 112 ~ Y-rotated, X-propagating cut of LiTaO3 [30] produced a lower temperature coefficient than the standard Y-Z lithium tantalate cut. These rotated cuts represented a more generalized SAW but with improved characteristics. Other piezoelectric materials both in substrate and thin film form have also been investigated. Although pseudo-SAW (PSAW) mode propagation had been characterized theoretically and observed experimentally in the late 1960s and early 1970s, it was not until the late 1980s and early 1990s that such modes were extensively incorporated into devices. It was necessary to identify cuts with very low leaky wave attenuation in the presence or absence of a thin metal layer or periodic surface discontinuity to meet low-loss device development. The demands of increasing frequency and wider bandwidths for wireless communication systems led to the use of rotated cuts of niobate and tantalate, whose dominant particle motions were in the plane of the substrate. A parallel effort by the SAW sensor community investigated similar modes for liquid sensing. Table 1, which lists the properties of selected piezoelectric crystals, shows the velocity, coupling factor, attenuation, and temperature coefficient of frequency along with two different propagation loss f a c t o r s ~ o n e representing viscous loss in dB/cm for regular cuts and the other the leaky wave loss in dB per wavelength. Also shown is a film substrate combination that extended the capabilities of lithium tantalate in terms of temperature coefficient. The main cuts used for wideband bandpass filter development today are the 64 ~ and 41 ~ Y-cut, X-propagating lithium niobate and the 36 ~ Y-cut, X-propagat-
Fred S. Hickernell
152 TABLE 1
S A W PROPERTIES OF PIEZOELECTRIC SUBSTRATES Crystal
Quartz YX ST ST-PSAW ST-HVPSAW Lithium Tantalate YZ 112 YX 36 YX PSAW 36 YX HVPSAW Lithium Niobate YZ 128 YX 64 YX PSAW 41 YX PSAW Gallium Arsenide (001 )(110) Cadmium Sulfide (001 )(100) Zinc Oxide (001)(100) Lithium Tetraborate XZ SiO2/YZ LiTaO3 Langasite Bismuth Germ. Oxide (001)(110)
Velocity (m/s)
2Av/v
3159 3158 5078 5745
0.18 0.12 0.033 0.011
8.2 9.8
3230 3288 4227 6978
0.72 0.6 5.6 2.1
3.5 3.3
3488 3992 4692 4752 2868 1725 2690 3542 3435 2600 1681
4.5 5.3 10.8 17.2 0.072 0.47 1.0 1.0 1.7 0.3 1.4
3.1 2.7
(%)
Loss* (dB/cm)
Leaky (dB/L)
TCF (ppm/~
24 0 7.8e-2 1.2e-3 -35 -18 2. le-4 0.12
-94 -75 5.2e-2 2.4e-4
14.0 20.0 9.5 17 12.0
-52 -37 0 0 0 -120
*Viscous propagation loss values in air at 1.0 GHz
ing lithium tantalate. The regular ST cut of quartz is still the substrate of choice for temperature-stable filters and resonators. Lithium tetraborate devices are in production, and langasite with its related compounds are waiting in the wings as possible substrates. As mentioned earlier, a very important aspect of device development is the availability of accurate material constants for the substrate and associated materials. There has been a continuing refinement of material constants over the years. An extensive set of constants and associated SAW properties were originally developed by the Air Force Cambridge Research Laboratories in the early 1970s [31, 32]. These have been refined and added to, as for example, the most recent work on lithium niobate and lithium tantalate [33]. It is essential to maintain an updated list of materials and their constants, which can be done with the elastic, piezoelectric, and dielectric properties in matrix form. It then is possible through the use of software available from McGill
3
Surface Acoustic Wave Technology Macrosuccess
153
University [34] to generate the basic engineering design parameters for arbitrary cuts and propagation directions of substrates. The McGill software, which uses a matrix boundary value approach, can be used for film-layered as well as free surface substrates. It gives the displacement components and potential for a piezoelectric substrate with depth. This is particularly important in understanding the nature of the surface wave propagation for a particular substrate. The software can be adapted to include viscosity as part of the elastic matrix, although in practice very few complete sets of viscosity constants exist for SAW crystals. C.
PROCESSINGAND FABRICATION
The most basic surface wave device consists of a polished piezoelectric plate with a metal IDT structure on its surface. This is a simple single-layer metal process, and the IDT electrode can be replicated on the crystal by standard semiconductor photolithographic techniques for devices in the VHF, UHE and lower microwave frequency regions with line resolution to 500 nm. For this reason universities, industries, government laboratories, and small businesses could very quickly fabricate and evaluate SAW devices. The electrode metal most commonly used is aluminum, which because of its low density and moderate acoustic impedance introduces the least perturbation on the surface. The thin metal film (60-400nm) is deposited on the substrate by standard vacuum deposition techniques such as evaporation or sputtering prior to etching the desired patterns. The electrode patterns, which relate to a particular functional device, are contained on a mask to be used in contact or projection to sensitize the photoresist used to pattern the metal. The removal of unwanted metal from substrate and the retention of the electrode pattern protected by the resist is accomplished by a wet chemical or reactive ion etch. The lift-off process is also in common use. In this case, the photoresist is applied to the bare wafer and developed, followed by metal deposition and lift-off of the unwanted metal. A thin layer ( ~ 20nm) of titanium is commonly used prior to aluminum deposition to enhance surface adhesion. At this point, the device processing is complete for devices that use only single-level mask patterning. In some cases, additional process steps are applied to the substrate (e.g., ion milling of grooves in the case of reflective array devices and/or the deposition of dielectric films for precise tuning, surface protection, and/or temperature compensation). Visual inspection or electrical probing of wafers may be used prior to dicing and device assembly. If processes are well controlled and yield
154
Fred S. Hickernell
rates are high, (i.e., high volume production of simple filters), wafer probing or visual inspection may not be cost-effective and a final performance test of the packaged device is used to screen out the bad devices. The separation of individual SAW devices is normally accomplished by sawing with a thin, high-speed, small-diameter diamond blade saw. The devices usually cannot be separated by a scribe-and-break technique. For most cuts of piezoelectric crystals, there are no natural cleavage planes along the required directions like those commonly found in silicon and gallium arsenide. The conventional assembly of a SAW device in a package involves three basic steps: (1) securing the crystal die to the package, (2) wire bonding the interconnections, (3) dampening the ends of the crystal die, if necessary, to prevent edge reflections. Resilient bonding compounds have been used to provide a sturdy nonvolatile bond that does not stress the crystal. Alloy metal bonding and epoxy bonding are also used. Interconnection with 1-mil aluminum wire ultrasonically bonded to the aluminum pad metal on the crystal and the package leads gives a reliable, high-conductivity bond with pull strengths averaging greater than 4.0 grams. Gold ball bonds have also been used in some cases. Presently, work is being done on flip-chip bonding to reduce package size and fabrication steps. For dampening material at the ends of the crystal die, a resilient compound is used to prevent edge reflections of the surface wave. It may not be necessary to use dampening material if the end edges are slanted at a shallow angle or the wave is contained within the substrate by some reflective mechanisms. Considerable accelerated life test data has been taken to verify the reliability of these assembly processes. The packaging of surface acoustic wave devices takes many shapes and forms. The most direct approach to packaging small devices such as bandpass filters has been the use of standard IC metal and ceramic packages with pin leads. This has generally given way to small surface mount packages in volume production, which are amenable to tape and reel mounting for subsequent assembly on a printed circuit board. Figure 5 displays a set of these SAW filter packages whose measurements extend from 20 by 6 by 2.2 mm down to 3.8 by 3.8 by 1.6 mm. For longer delay lines and matched filters, custom metal housings are made from aluminum or stainless steel with rf connectors attached to be assembled in a system. The package may be designed to accommodate matching components such as inductors and capacitors. The SAW device can be mounted with electronic components on a common alumina substrate to form a hybrid circuit module such as an
3
Surface Acoustic Wave Technology Macrosuccess
155
.........
s
9
.,.: :..
Fig. 5. Typical surface mount packages for IF, interstage, and rf SAW filters for telecommunication systems.
oscillator. The packaging and interconnections must also ensure that no induced stress is placed on frequency-sensitive SAW devices. Special packaging studies have also been made to ensure that g-forces and vibration do not affect device operation. Another important factor of packaging is that of providing a moisture seal. This is usually accomplished by the hermetic seal of standard metal or dielectric packages. Because package costs can become a major part of volume-produced devices, there is a continuous effort to use plastic mold techniques. The concem in such packaging is to keep the surface clear of any internal or external condensation, which could dissipate the surface wave energy. Where volume production is in the tens of thousands per week, batch processing of wafers is carried out and the assembly lines are highly automated for die placement, wire bond, package seal, and final testing. In the case of engineering prototypes, the time required from design to packaged device can be done in one to two weeks. This takes advantage of highly developed computer-aided design and layout tools, rapid contact or projection mask fabrication, and the simplicity of the aforementioned device processing and fabrication.
156
Fred S. Hickernell
The cost of simple SAW devices, such as bandpass filters in volume production, range from 50 cents to 2 dollars. Piezoelectric wafer costs are very competitive in large quantities, with a 3-inch quartz wafer less than 20 dollars, lithium niobate in the 20 to 30 dollar range, and lithium tantalate under 50 dollars. At 900 MHz for rf filters, 800 to 900 filters can be processed on a single wafer, leading to substrate cost per die of a few pennies. For nonstandard large customized wafers, the cost is much higher, depending on the size and shape. Often, the cost of packaging dominates the material device costs for SAW filters and delay lines. As the requirement for higher operating frequencies in the gigahertz region increases with critical line dimensions on the order of tenths of microns, the demands on the quality of the piezoelectric wafers increases. Variations from stoichiometry will affect the elastic properties. Mechanical defects induced at the polished surface of the wafer will affect the propagation and coupling properties of the surface waves. Wafer thickness variations and wafer bow will affect the resolution capabilities of the photolithographic processes. The dimensions of the patterns will be approaching the dimensions of the SAW dicing blade, causing more wasted material area. In the highly competitive world of SAW devices produced in the hundreds of thousands, costs are looked at in terms of tenths of cents. Thus, the old adage of Benjamin Franklin, "A penny saved is a penny eamed," holds just as true today for SAW device costs. Vl.
Interdigitai Transducer Controlled SAW Devices
The concept of an interdigital metal electrode on the surface of a low-loss piezoelectric plate was the key to the rapid implementation of SAW devices in electronic systems. The IDT served the dual role of excitation/detection and time/frequency signal processing. Surface wave devices were developed for delay lines, bandpass filters; matched filters, and resonators using the twodimensional geometric aspects of periodic metal stripe configurations. The IDT offered a simple means of fabricating both customized devices and large quantities of two-port and multiple-port microacoustic devices reliably and inexpensively. The following paragraphs present some simple examples of the early applications of surface wave devices with interdigital transducer electrodes as the major features of these signal processing functions. A.
THE Two-PORT DELAY LINE
Figure 6 depicts the basic parts of a simple two-port SAW delay line. It consists of a polished piezoelectric plate, two interdigital metal electrodes
3
157
Surface Acoustic Wave Technology Macrosuccess
INTERDIGITAL ELECTRODE INPUT SIGNAL
BONDED WIRE CONTACT ~~ ~ ~
/ l~ ~
OUTPUT SIGNAL \ \ \
.
,EZOELE , ,C L TE
ACOUS TERMINATION
~
~ , ~ ~
~NP SIGNAL
Fig. 6. A basic two-port SAW delay line. with bonded wire contacts for the input and output signal, and acoustic terminations for absorbing the bidirectionally launched surface wave at each end of the crystal. The separation of the electrode patterns determines the time delay, the number of finger pairs determines the bandwidth, and the spacing between finger pairs determines the frequency. This basic geometry was used to develop the early two-port and multiple-port tapped delay lines. Figure 7 is a photograph of a two-port surface wave delay line used in airborne coherent transponder equipment. The line served a twofold purpose in the transponder system: (1) to provide a fixed coherent signal reference delay and (2) to act as a bandpass filter. It had a total time delay of 2.35 gs, a center frequency of 60MHz, and a 3-dB bandwidth of approximately 12 MHz. The average loss across the operating band was 13 dB. The line provided greater than 50 dB rejection at frequencies below 35 MHz and above 85 MHz. To simultaneously obtain the required signal bandwidth and out-ofband rejection characteristics, a graded pitch electrode pattern was used. The pattem was synthesized from a computer program based on an equivalent electrical circuit model for a surface wave interdigital transducer. The agreement between the predicted frequency loss characteristic and the measured characteristic was excellent. The crystal was mounted in a standard dual in-line integrated circuit package with input and output timing coils and was hermetically sealed for environmental protection. It replaced a larger and more costly bulk wave device. The operating frequencies for simple two-port delay lines have extended from 10 MHz to 2 GHz. Typical time delays have been from 0.3 to 30.0 Its and bandwidths from 1 to 300 MHz. Piezoelectric crystal lengths 10 cm or less are
158
Fred S. Hickernell
9 9149 ~9 9 9149 9 9149 9149,! , 9149 I~9149
9
........
....
9
'~
::~:!~i~i
~.... i.:~ ~.!~'%
,
....
9 1 4 99149
.........~=.i~i
~ , ~.=.~.
. '
~:~ 5
.
'5 9
.,
~
4; ~
: i~. . ~ ~
.
.
.. 9
Fig. 7. Two-port SAW delay line in dual in-line package with tuning coils, used for a coherent transponder application.
required for these delay times. To achieve much longer delays (e.g., >50 ~ts), different techniques have been investigated [35]. Delays up to a millisecond have been achieved using a helical line with the surface waves propagating from the front to the back side of the crystal with rounded edges as illustrated in Fig. 8. Such a line requires good finishing techniques and a crystal with a
Fig. 8.
Helical SAW delay line on a piezoelectric crystal with rounded edges.
3
Surface Acoustic Wave Technology Macrosuccess
159
reasonably isotropic velocity surface to ensure low-loss and good directionality. Guiding structures can be added to confine the surface waves over the long path lengths. Reflecting structures on the crystal surface either by metal or grooved geometries have produced delays in the 25 to 100 microsecond region. These techniques are noted in Section VIII.
B.
THE MULTIPLE-PORT DELAY LINE
By adding IDT patterns to the substrate along the propagation path, a multiple-port or tapped delay line can be developed. With a tapped delay line structure, electronic switching circuitry can be used to address the taps and select a particular delay time. The switchable multiple-port SAW delay line developed in the early 1970s as shown in Fig. 9 was used in a radar application for timing and coherent signal processing. This is an example of the hybrid integration of a tapped SAW delay line with simple microelectronic switching circuitry. The tapped delay line has ten aluminum interdigital electrode transducer patterns on a polished lithium niobate crystal. The IDT electrodes are spaced to obtain a range of delay times between 400ns and
Fig. 9. Multiple-port SAW delay line with hybrid chip and film electronic switching circuitry.
160
Fred S. Hickernell
6.0 ~S. The ten output ports are connected to switching diodes mounted along the side of the crystal. The diode biasing network contains a thick film, screened resistor network and chip rfblocking capacitors. The particular delay time required is selected by turning on the associated diode. The entire assembly is mounted on a self-supporting alumina substrate. The delay line operates over a 100-MHz bandwidth in the upper VHF region. Loss at any tap across the 100 MHz bandpass is less than 30 dB. This basic hybrid-tapped delay line has been produced in large quantities over the past 20 years [36]. C.
THE BANDPASS FILTER
The periodic nature of the interdigital transducer provides a natural bandpass function. The bandpass filter has become the most mass produced SAW device for application in communication systems at VHF, UHF, and microwave frequencies. Easily fabricate& the surface wave filter is a simple compact planar structure, amenable to large-scale chip manufacture, and easily integrated into hybrid microelectronic circuitry. It is an extremely versatile component, permitting a wide range of desired passband and reject bandshape responses to be developed. One of the earliest and commonly used design techniques for shaping bandpass filter responses was reported in 1969 [37] and involved the use of the weighted overlap IDT structure for developing a specific frequency passband and reducing adjacent side-lobe levels. The principle is that of the Kallman or transversal filter [38], with the interdigital electrode spacings the delays, and the electrode overlaps the amplitude weighting functions. This early work on SAW filters was in the low VHF region. A concentrated effort by workers in the United States, Europe, and Asia was directed toward a simple, low-cost SAW filter as a replacement for the coils and capacitors in television sets. SAW filters were developed for audio and video consumer product applications. Filters were developed for radar and communication systems and used for frequency synthesis and frequency selection and analysis. The work on surface wave bandpass filters has remained strong and broadened in scope over the past several years. The important factors of loss, bandshape, tipple, temperature dependence, side-lobe suppression, and frequency capability have been under extensive investigation. New transducer configurations evolved to achieve loss levels of under 3 dB, suitable for the front end ofwireless communication products. To illustrate an amplitude-weighted (often termed apodized) SAW filter structure, a photograph of a three-transducer electrode pattern is shown in Fig. 10(a). The IDT metallic electrodes are photoetched on the surface of a
3
Surface Acoustic Wave Technology Macrosuccess
161
polished piezoelectric plate. The interdigital pattern shown uses split finger electrodes to reduce in-phase reflections and has two broadband outer transducers and a center transducer with three symmetrical side-lobes. An input signal drives the apodized center electrode, which launches a bidirectional acoustic surface wave detected by the adjacent unweighted IDT electrode structures. The type of amplitude weighting and the number of finger pairs of the central transducer are determining factors in the frequency response of the filter. The time impulse response of this filter shown in Fig. 10(b) is an excellent replication of the apodized pattern, illustrating the oneto-one correspondence of the spatial electrode domain and the time domain. The frequency response, which is the Fourier transform of the time impulse response, is shown in Fig. 10(c). The agreement between theory and experiment is excellent. In this example, the filter is centered at 70 MHz, with a fiat bandpass characteristic over a 12-MHz region. Ban@ass ripple is
(b)
(a)
(c)
Fig. 10. A 70-MHz SAW bandpass filter. (a) electrode pattern, (b) time impulse response, (c) frequency response.
162
Fred S. Hickernell
less than -+-0.75 dB. The out-of-band rejection is greater than 45 dB with a transition width to the 45 dB point of approximately 4 MHz. SAW bandpass filters of this type have been implemented for IF filtering in a range of products and systems from the living room to outer space. One of the earliest implementations of bandpass filters in volume production was in television IF filters [39]. This was one of the earliest commercial successes for SAW bandpass filters. Figure 11 shows a packaged SAW bandpass filter with an apodized pattern on YZ lithium niobate used in a television IF to replace the coils and capacitor circuits indicated in the adjacent enclosure. This dramatically illustrates the size reduction that surface wave technology brought to electronic circuitry. Implicit in the introduction of SAW filters into television sets was the reduction in cost of the end-of-the-line manual tuning operation. By 1979, it was estimated that well over 10 million SAW IF filters were being produced by two manufacturers in the United States, four in Western Europe, and seven in Japan [40]. There were also at least three major manufacturers in Eastern Europe and Russia. More detailed information on the fabrication and performance of early TV SAW filters is contained in two
Fig. 11.
Television IF filter that replaces coils and capacitor network shown to the left.
3 Surface Acoustic Wave Technology Macrosuccess
163
articles in a special issue of the Proceedings of the IEEE [41, 42]. The major producers of SAW filters for television today are Siemens and Murata, which manufacture over 200 million per year. Siemens uses lithium niobate as the substrate [43], whereas Murata has a unique approach using sputtered zinc oxide films on a glass substrate [44]. SAW bandpass filters have been used in the IFs of transponders for over 30 deep space and earth satellite missions [45]. The performance characteristics for filters in the 100 to 400 MHz region were achieved using amplitude weighting of the interdigital electrodes. The filters underwent very rigid qualification specifications and have performed reliably in space environments. A photograph of the type of the filters used for the Voyager missions to the planets of our solar system is shown in Fig. 12. The bandpass filters were on lithium niobate and quartz and had simple apodized central transducers. The maximum in-band insertion loss was less than 12 dB with bandwidths of 3 to 5 MHz and rejection levels of 60 dB. Other developments of SAW filters for satellite communications have been reported by workers in France [46] and Canada [47].
Fig. 12.
SAW IF filter used in the deep space transponders for the Voyager mission.
164
Fred S. Hickernell
A logical extension of SAW filter applications was to electronic warfare for the frequency sorting of signals by the use of contiguous SAW filter banks [48]. SAW bandpass filters staggered in frequency with their response overlap near the 3 dB points can be multiplexed to achieve this frequency sorting. The filters require triple transit levels greater than 60 dB and strong out-of-band rejection. Several filters with different frequencies can be processed on a single substrate. The filter bank can also be used for frequency synthesis [49]. In this case, the filter bank is coupled to diode switches for the frequency selectivity and' permits frequency hopping in an efficient and compact configuration. Many other transducer configurations which have been developed to lower the insertion loss, narrow or broaden the bandpass, improve the shape factor, and lower the out-of-band rejection. One alternative to apodization weighting is withdrawal weighting, where fingers are removed in relationship to the desired time impulse response [50]. In both withdrawal weighting and apodization, it is necessary to account for the phase changes that can occur due to velocity differences for the waves traveling under unmetallized and metallized regions to have a clean filter response. To lower the loss, the IDT patterns can be cascaded in-line to reduce bidirectional losses. This technique is used by the interdigitated interdigital transducer IIDT [51 ]. There are configurations associated with unidirectional transducers, the most commonly used being the single-phase unidirectional transducer SPUDT [52]. Resonator, waveguide-coupled, and impedance element IDT structures can all be configured for low loss because they effectively contain the surface wave energy within their respective IDT patterns. Some of these resonator structures have counterparts in the bulk acoustic wave frequency selectivity world. Comparisons of SAW filter technology with competing technologies have been made [53]. However, the small wavelengths compared to electromagnetic still favor the SAW filter for size and weight. A SAW bandpass filter chip 1.4 mm on a side, which operates near 2.0GHz, is shown in Fig. 13. Digital techniques are advancing, but their requirement for power and their lower frequency capabilities have not impacted the passive nature and highfrequency capabilities of SAW filters. From the electrical point of view, the ability of SAW filters to synthesize phase and amplitude independently permits bandpass and band-reject functions. In addition, the SAW filter is extremely compact and easily integrated with other microelectronic circuit elements. The planar geometry, simple processing steps, and low-cost multiple-chip manufacture make the surface acoustic wave approach very enticing from the fabricational point of view.
3
Surface Acoustic Wave Technology Macrosuccess
Fig. 13.
165
A 2.0-GHz SAW bandpass filter on a 1.4-mm square die.
SAW filters in the tens of millions are being produced each year at an estimated rate of more than 10 per second from manufacturers throughout the world. The shapes of the IDT patterns have changed considerably from the simple two- and three-element apodized structures to the example in Fig. 14, which shows the mask pattern for a 9-element ladder filter. SAW filters are
ItlllIIIIIII~ItlIiWIHtWII~'~
~Htl"HZilltHlllH~ll
Fig. 14.
IH H
-
-
~
SAW bandpass ladder filter pattern with electrically connected resonators.
Fred S. Hickernell
166 ~ : ~i~,~i~:~i ~ ~,~%,~
9 . . 9,. ~,;~. 9 ~, ~~".,~
Fig. 15.
~
".: ~ : " . ~ )
'
.~,
SAW packaged IF and rf filters used in handheld cellular telephones.
low cost and packaged in form factors that are easy for board assembly. Examples of IF and rf filters manufactured in large quantities and surface mount packaged for Use in cellular phones are shown in Fig. 15. Table 2 provides information on the range of typical parameters for SAW bandpass filters. Obviously, any choice of one particular performance parameter will not be independen t of the other parameters since trade-offs take place between insertion loss, bandwidth, bandshape, etc., which are functions of the piezoelectric substrate, the design, and manufacturing tolerances. Catalogs from SAW suppliers contain a myriad of available bandpass filter TABLE 2 SAW FILTER PERFORMANCE CHARACTERISTICS Performance Parameter Center frequency (MHz) Relative bandwidth (%) quartz lithium tantalate lithium niobate Insertion loss (dB) Out-of-band rejection (dB) Shape factor (3 to 40 dB) Passband amplitude tipple (dB peak) Passband phase ripple (degrees peak) VSWR
Typical
Extended
20-2500
10-5000
0.5-5 5-10 10-30 4-30 40-50 > 1.5 >0.5 >3.0 2-2.5:1
0.05-10 1-30 5-70 1-3 60-80 1.1 0.1 1.0 1.5:1
3
Surface Acoustic Wave Technology Macrosuccess
16'/
products. Custom-designed devices with reasonable specifications can be quickly developed.
D.
SAW RESONATORS
Since SAWs do not perfectly reflect at an edge boundary (there is a decomposition into bulk and surface waves), they cannot be fabricated in the same form as the common bulk wave plate resonator. Thus it is necessary to create a collective reflector made up of quarter-wavelength periodic discontinuities to provide the reflection mechanism. Such reflectors, if fabricated carefully, can achieve the very high Q factors that approach the basic Q of the material being used. The key advantage of the SAW resonator is its capability for operation at fundamental frequencies ten to one hundred times that of corresponding fundamental mode bulk acoustic wave (BAW) resonators. Quartz is commonly used for SAW resonators because of its high material Q (low SAW propagation loss) and its temperature stability. The concept of gratings as reflectors was set forth by Ash in 1970 [54]. Simple one-port and two-port resonator configurations are shown in Fig. 16. In the one-port geometry, there is a central transducer bounded by quarterwavelength reflective gratings. For the two-port geometry, two centralized broadband transducers are bounded by the gratings. There are several different ways in which the transducers and gratings can be configured to form resonators. In the basic one-port configuration, the SAW resonator serves as a resonant impedance element. In the two-port configuration, the centralized transducers serve as input and output and the response is equivalent to that of a
Illii Illlllllllllll--Illl
Illl.. Illlllllllll
ONE P O R T
tttt., tlltlttittt
r
-
Illlllf ttltttttIllll-.llll I
TWO PORT Fig. 16.
IDTs and grating reflectors for basic 1- and 2-port resonator configurations.
168
Fred S. Hickerneli
narrowband low-loss filter. There is an effective cavity length that occurs at a point well within the grating. Typically, the reflection coefficient of the grating elements is 0.01 or less, and the number of elements in the grating commonly exceeds 100. Fabrication techniques will vary in order to achieve the best possible Q values. The easiest fabrication method is to use deposited metal for the gratings and transducers. Higher Q values have been achieved using ion milled gratings and the combination of ion milling with metal filling of the grooves. The Q factors that can be achieved are over 10,000 in the lower UHF region and in excess of 1000 for frequencies in the GHz region. In their earliest implementation, SAW resonators were used as replacements for BAW resonators at higher frequencies. The early packaging and contact methods were similar to those of BAW resonators. The photograph in Fig. 17 shows such a SAW resonator with a center frequency of 150 MHz. Resonators can be surface mounted in standard metal or ceramic IC packages. In 1981, SAW workers at Hewlett Packard discussed the applications of SAWs and described the first SAW resonator used in HP instruments, which replaced a lower-frequency bulk crystal resonator [55]. Other applications to instrumentation followed [56]. Bell and Li reviewed the early status of SAW resonators in 1976 [57]. Coldren and Rosenberg discussed design and performance
Fig. 17. A 150-MHz SAW resonator with mounting and packaging for replacement of BAW resonators.
3
Surface Acoustic Wave Technology Macrosuccess
169
trade-offs for various configurations of resonators [58]. A review of SAW resonator filter technology that has proved to be successful in large-scale production was given by Wright [59] in 1992. SAW resonators, as stand-alone components, are used for several applications but primarily as frequency- and/or time-controlling elements. For example, SAW resonators are used to generate data and clock signals in fiber-optic data links, in computers, keyless entry systems, tags, and telecommunication applications. The SAW resonator concept has been used extensively in recent years for the creation of very low-loss impedance element filters by cascading resonators in ladder and lattice configurations [60]. E.
SAW OSCILLATORS
The motivation for the use of SAW delay lines and resonators as the frequency-determining elements in oscillators, like resonators, is their highfrequency capability compared to bulk acoustic wave (BAW) oscillators. SAW oscillators can be operated at frequencies from 100 MHz to over 2 GHz and eliminate the need for multipliers and filters, which increase size, weight, power, and cost. The basic oscillator circuit consists of a SAW delay line or resonator in an amplifier loop. The SAW device is the frequency-determining element and quartz is the usual substrate for temperature stability. When the amplifier gain exceeds the loss of the delay line, oscillation occurs. A phase shifter is often incorporated in the loop with the delay line controlled circuit, adding the capability for voltage control of oscillation over a small range of frequencies. The use of a SAW resonator as the frequency-controlling element gives a very stable fixed frequency and very high Q factors of well over 1000 to 10,000, depending on frequency. Early work with SAW oscillators was reported by Lewis [61 ]. Work at Raytheon over several years has been on the All Quartz Packaged (AQP) SAW hybrid circuit oscillators in the 150 MHz to 1 GHz region [62]. These have pushed the performance limits in what can be achieved in terms of performance, frequency trimming, packaging, aging, noise, vibration, and g-forces in SAW oscillators. For airborne radar applications, where low Doppler shifts and small cross section targets need to be detected, the use of the high-frequency SAW resonator permits a higher fundamental frequency with lower phase noise. Very high quality voltage-controlled SAW oscillators are sold as commercial products with frequencies generally in the UHF region. They can be obtained as hybrid integrated circuits with the SAW device, amplifier, control circuitry, and interconnects within the same package. Typical applications are
170
Fred S. Hickernell
frequency synthesis and translation, and data retiming and synchronization as part of a phase locked loop.
VII.
Electrode-Configured Matched Filter Devices
Radar, communication, and navigation systems receive and process information in the form of amplitude, frequency, and phase modulation on a carrier signal. This information can be distorted in amplitude, phase, and time by the transmission media and circuits through which it passes. Noise can be inadvertently or purposefully introduced into the signal channel. To extract the desired information, it is essential to maximize the signal-to-noise ratio for optimum detection. The maximum signal-to-noise ratio is obtained when the receiver has a filter characteristic whose impulse response is the time inverse of the received signal. This is called a matched filter. The technique of matched filtering constitutes the optimum linear processing of signals. Matched filtering can be easily implemented by manipulating the IDT electrode configurations for SAW devices. Figure 18 illustrates some possible variations in electrode fe'atures by which a SAW matched filter could be implemented. The upper part of the figure shows a central IDT with electrode variations bounded by two broadband standard IDTs that could be placed on the surface of a piezoelectric crystal to form a matched filter pair. In the central pattern, there is a weighting of the finger overlap regions, a phase change, and a variability in the finger spacings.
INTERDIGITAL TRANSDUCER
IMPULSE RESPONSE
Fig. 18.
Example of an arbitrary IDT pattern and impulse response for a SAW matched filter.
3
Surface Acoustic Wave Technology Macrosuccess
171
The two outer patterns have uniform finger spacing and overlap. The impulse waveform response in time of the center pattern as detected by the uniform pattern to the left is shown at the bottom of the figure. The time response replicates in amplitude and frequency (phase) the interdigital pattern. The waveform has amplitude modulation, a point of phase reversal, and frequency modulation characteristics. If this time waveform is generated, transmitted, and subsequently received by an identical surface wave device using the interdigital pattern to the right as the receiving transducer, there will be a correlation peak when the time waveform is physically aligned with the center transducer. This is the basis for SAW matched filter pair devices such as fixed correlators and chirp filters using interdigital electrodes to govern the time/ frequency response.
A.
CORRELATORS
For illustrative purposes, Fig. 19 shows the implementation of a 7-bit phasecoded Barker sequence. At the top of the figure, there is the central interdigital transducer pattern with two-cycle phase coding to give the 7-bit Barker sequence (1, - 1, 1, 1, - 1, - 1, - 1). The time domain impulse response waveform of this pattern is shown at the middle of the figure. The autocorrelation function that results as this waveform travels through the center interdigital pattern is shown at the bottom of the figure.
INTERDIGITAL ELECTRODE 1
-1
1
1
-1
-1
-I
SURFACE WAVE
Yl
"
6'
-7
Fig. 19.
j =
-6
'
-5 -4 -3 -2-1 0 1 2 3 AUTOCOR R E I.~TION
-
4
5
6
7
A 7-bit biphase-coded Barker sequence SAW IDT pattern and correlation function.
172
Fred S. Hickernell
Figure 20(a) shows the surface wave substrate and package housing for a 100-tap biphase-coded filter developed for a pulse-compression memory-test system. The tapped surface wave structure is on temperature-stable quartz and has a 10.0-MHz bandwidth and 10-microsecond expanded phase-coded pulse. The operating frequency of the device is 165 MHz. Scope photographs of matched filter pair performance are shown in Figures (b) and (c). The upper photograph (b) shows an expanded view of the leading and trailing edge of the biphase-coded signal produced by pulse excitation of the center interdigital structure. The lower photograph (c) shows the compressed pulse under a correlated condition. The amplitude of the lower trace has been raised to indicate the side-lobe level, which is approximately 20 dB below the peak of the compressed pulse.
B.
PULSE EXPANDER-COMPRESSORS
A pulse expander-compressor (chirp filter system) can be developed using a simple series of linearly graded interdigital lines. Figure 21(a) shows the
(b)
(a) (c)
Fig. 20. A 165-MHz 100 tap biphase-codedSAW correlatorwith 10 gs time delay. (a) SAW device and package, (b) time domain impulse response, (c) time domain correlation response.
3
Surface Acoustic Wave Technology Macrosuccess
173
(b)
(a) (c) Fig. 21.
SAW chirp filter at 60 MHz with a 20 MHz bandwidth and 5 Its time delay. (a)
SAW device and package, (b) time domain impulse response, (c) time domain compression response.
surface wave substrate and the package that houses the substrate and matching network for a 100:1 chirp filter. The filter operates at 60 MHz with a 20-MHz bandwidth and 5-microsecond expanded pulse. The chirp matched filter is developed on lithium niobate and has approximately 1300 individual tap electrodes. Scope photographs with expanded and compressed pulse responses in time are shown in Figures (b)and (c). The lower trace of the compressed pulse photograph has been expanded 28 dB in amplitude to show the time side-lobe response. The device was developed for use in an advanced solid state radar system.
C.
THE PROGRAMMABLE TRANSVERSAL FILTER
The concept of the multiport delay line with appropriate phasing of the taps and a tap-selectable switching circuit can be transformed into a programmable transversal filter. In 1940, Kallman introduced the concept of a universal filter, called a transversal filter, composed of a tapped delay line whose discrete
174
Fred S. Hickernell
.--o
,~oUTPUT IN
.....
<
<
Fig. 22. Two-channelSAW delay line with external switchingfor a binary phase programmable transversal filter. taps, spaced in time, could be amplitude- and phase-weighted and then summed [38]. Arbitrary time impulse responses could be synthesized, leading to the desired signal responses in the frequency and time domains. One of the requirements for synthesis was that a tap would weakly sample the propagating wave with negligible attenuation and distortion. This concept was a natural for implementation in the surface acoustic wave propagation mode with the interdigital transducer concept on a piezoelectric substrate. The direct extension of SAW tapped delay lines to produce programmable matched filters can be accomplished by providing external switching configurations to address a particular sequence of taps. Figure 22 shows a simple schematic of a two-channel phase-coded line with external switching and summation. The selection of taps can be done with addressable external digital circuits. Wideband 128-tap and 256-tap programmable SAW matched filters developed by Hazeltine illustrate the capabilities that can be achieved by this hybrid technology, which incorporated SAW and full-custom LSI technology [63]. VIII
Signal Processing Through the Passive Control of SAW Propagation
Between points of excitation and points of detection, the propagation of the SAW can be controlled through modification of the intervening substrate surface region. The surface can be configured using films and topographical structures to enhance the signal processing capabilities of transducer-config-
3
Surface Acoustic Wave Technology Macrosuccess
175
ured devices and to perform useful signal processing functions independent of the transducer structures. Metal films on the piezoelectric substrate can be configured for use as a waveguide, to redirect the surface waves on alternate paths, and to selectively reflect the surface waves. Periodic topographical grooves on a substrate can collectively reflect and also redirect surface wave flow. Dielectric films have electrical and physical properties that can enhance the performance of a SAW device.
A.
WAVEGUIDES
One of the original visions for SAW signal processing was the control and confinement of surface waves through waveguides that relied on structures where the propagation velocity in the guiding structure was reduced to a value less than the free surface velocity and any bulk mode velocities. The two main categories of guides investigated were topographic guides, where the velocity reduction is achieved by shaping the substrate surface, and thin film guides, where the velocity reduction is effected by the elastic or electric properties of a deposited film. Examples of these types of waveguides are shown pictorially in Fig. 23. An early idea was to develop "microsonic" circuits, a highly compact SAW circuit technology which were analogous to microwave microstrip circuits. These microsonic circuits were envisioned as having guidance,
TOPOGRAPHIC 9
.
:.
T
RECTANGULARRIDGE
WEDGE
LAYERED
v-" STRIP i ~
......
L
-
:~ -
Fig. 23.
~
~
v.j
A
Y
E
.Ss R
soBsT.AT LOC,TY v
SLOT
SAW topographic and film-layered waveguides.
VELOCITY
176
Fred S. Hickernell
splitting, coupling, short circuit, and open circuit functions using surface waves on a substrate within distances much less than their microwave counterparts. Although considerable speculation and theoretical work was given to this concept, in experimental practice only a few of the device functions were demonstrated. The major barrier was radiation losses in trying to confine the elastic wave energy in curved and terminated (open) structures. Lagasse and coworkers [64] have reviewed the status of SAW waveguides and their application potential. An example of a SAW film waveguide structure on silicon used to study the properties of guided wave propagation and the potential for long delays is shown Fig. 24. Thermally grown oxides of five different thicknesses (in the range of 3 to 11 microns) on silicon were patterned in a spiral configuration, and wedge transducers used to excite and detect both Rayleigh and Love waves. The structure was used to study the velocity characteristics and minimum turning radius of the Rayleigh and Love modes. Rayleigh waves with minimum dispersion could be propagated in the slots, and Rayleigh and Love waves were propagated in the oxide strip waveguide mode. The characterization was with ZnO diffusion layer wedge transducers with frequencies in the 30 to 90 MHz region. Schmidt and Coldren [65] reviewed the general aspects and capabilities of the strip, the slot, and, in particular, the Av/v waveguidance. The use of a thin
Fig. 24. Silicondioxide waveguide on a silicon substrate with wedge SAW transducers.
3
Surface Acoustic Wave Technology Macrosuccess
177
metal film to form a Av/v waveguide by shorting the piezoelectric property and providing a slow wave channel for SAW propagation was investigated extensively and employed for device development. The waveguide configuration requires only a thin metal film, such as aluminum, less than 100 nanometers thick and a few wavelengths in width. It works best on high coupling factor material such as lithium niobate. One application envisioned for a Av/v waveguide structure was for long delay lines for signal storage in which the delay line confined the wave over long distances [35]. A second and even more important application for the Av/v waveguide was to enhance the energy density and nonlinear interaction of two counterpropagating acoustic waves to form an elastic convolver. There are two basic types of convolvers: the elastic and the acoustoelectronic. The acoustoelectronic convolver uses a semiconductor medium as the nonlinear medium. Work at Lincoln Laboratories by Yao [66] yielded high-performance convolvers. Morgan [67] has reviewed convolver technology with several key references included. The elastic convolver has typical performance characteristics of 100-MHz bandwidths, 10-kts integration, and times and efficiencies in the - 7 0 dBm region. This gives TB products of 1000 (processing gain of 30 dB) and an output signal-to-noise ratio near 50 dB. B.
THE MULTISTRIP COUPLER
The SAW multistrip coupler (MSC), introduced by Marshall and Paige in 1971 [68], performs the function of a SAW directional coupler on the free surface of a piezoelectric substrate through the use of an array of parallel metal strips. The theory and design of the MSC has been reviewed by Marshall and coworkers [69]. A pictorial representation of some the basic operations that can be performed by a multistrip coupler are shown in Fig. 25. In Fig. 25(a), the set of parallel metallic strips overlapping the path of the SAW generated by the transducer on the first piezoelectric substrate generates a SAW on the second piezoelectric substrate although the substrates are interconnected only by metallic strips. The transfer of the SAW can be complete or partial, depending on the number of metal stripes and the coupling factor of the piezoelectric substrates. In Fig. 25 (b), the SAW from an IDT with a large aperture is successively compressed by a series of multistrip couplers to a narrow beam. It is evident that the surface wave displacement need not be lateral but can be in any direction and relies primarily on electrical interconnection. In Fig. 25(c), the SAW is displaced normal to its original direction of propagation. The folding of the MSC lines
Fred S. Hickernell
178
(a)
(c)
-II
,-I
M_._
.,11
(b) Fig. 25. change.
MSC components with (a) track change, (b) pulse compression, and (c) directional
around an IDT can give it unidirectionality as well as reflector and tap functions, as illustrated in Fig. 26. Other unique components can be developed [70]. The MSC has proven to be a useful tool for enhancing signal processing functions. The MSC can be easily incorporated in a SAW device and processed as a single metal layer together with the IDT electrodes.
Fig. 26.
Multistrip couplers for unidirectional transduction, reflection, and tapping.
3 C.
Surface Acoustic Wave Technology Macrosuccess
179
REFLECTION GRATINGS
Reflection gratings--like those used for SAW resonators but in the region between transducers--gave rise to some very unique devices. Williamson [71] has reviewed the properties and applications of reflective array devices, which have included resonators, oscillators, tapped delay lines, bandpass filters, filter banks, and dispersive delay lines. The reflection gratings that have found the most use for signal processing are those that are angled near 45 ~ to the propagation path and reflect the surface waves 90 ~ to the original propagation path. The reflection mechanisms can be due to topographic features such as ion-milled grooves or films with mass loading and piezoelectric shorting features such as metal stripes or dots. Sometimes it is possible to modify the surface region through doping to produce a sufficient elastic property change to induce reflections. Some examples of reflective array devices are shown in Fig. 27. By far the most impressive device for large time-bandwidth signal processing has been the reflective array compressor or RAC device first developed at MIT Lincoln Laboratories by Williamson and Smith [72]. It is illustrative of the basic concepts of reflective array devices. A diagram of the RAC device is shown in Fig. 28. A surface wave launched by one of the transducers encounters a series of ion-milled gratings with increasing or decreasing
1111~1I11111
Illl[Li![!
1iiiill
RESONATOR
"
O
FILTER
RING FILTER
TRACK CHANGER
Fig. 27. Examples of reflective array devices using metal stripes or grooves for the reflection of surface waves.
180
Fred S. Hickernell ETCHED GRATING
INPUT TRANSOUCER
9 OUTPUT TRARSDUCER
Fig. 28.
Reflective array compressor (RAC) filter.
periodicity. The surface wave is strongly reflected at a 90 ~ angle when the periodicity of the grooves match the wavelength of the SAW. The wave then experiences a second reflection in a symmetrically placed mirror image grating, which directs it back to the second receiving transducer. The path length is characteristic of a particular frequency and increases or decreases with frequency dependent on whether the grating spacing is increasing or decreasing. Techniques for phase and amplitude weighting can be incorporated in the design of the RAC device. Very large time-bandwidth ( ~ 104) pulse expanders and compressors developed by workers at Hughes Aircraft have been demonstrated with this technology [73]. The photograph of a reflective array compressor with a 6-in long Y-cut, Z-propagating LiNbO3 housed in an aluminum package is shown in Fig. 29. The RAC device has a center frequency of 160 MHz, a bandwidth of 40 MHz, and a total time of 50 microseconds. This 2000 time-bandwidth RAC used two 2.5 finger pair transducers with over 4000 ion-milled grating elements with depths extending from 100 nm at the high-frequency end to 400 nm at the low-frequency band edge. The reflective dot array was one such device proposed as an alternative to the etched gratings [74]. The technique was applied to three distinct device classes: (1) bandpass filters with low ripple and high out-of-band rejection, (2) pulse expansion and compression filters, and (3) a 255-bit PN code correlator. The large time-bandwidth (TB "~ 104) reflective array dispersive delay lines that have been demonstrated are bounded by bandwidths of 500 MHz with
3
Surface Acoustic Wave Technology Macrosuccess
181
Fig. 29. Reflective array compressive filter with ion-milled grooves on LiNbOs having a time-bandwidth product of 2000.
20 Its delay and 100 MHz with 100 Its delay. Smaller TB products fall within these ranges. Available from SAW manufacturers today are dispersive delay devices with time-bandwidth products extending from 10 to over 10,000, center frequencies in the 30 to 1000 MHz range, and relative bandwidths from 1 to 60%. Processing gains in excess of 30 dB can be easily realized. Figure 30 shows a sampling of the TB product capabilities and limits of linear FM SAW filters using interdigital electrodes and reflective array techniques. 1000,
"
- '%
-
.1
N
',lillll
-.
SlZE[~,
,c'l'-I"t-H'l'i"=--
_%
" ~ I!1+,
100'% i,
i
""
\
IIIL lllr"
%.
",.~, ;
~
illl II .~
:::i1!
'
III
~.J
I111
"~: ~',ill !!!ll-i~
~
~m
]
9 = Ill
.... %
~n -~ :::1.
i
:
-, %
-
"
9: i i i
II
ll~
Illl
,i
-Jill
-~-;]lt.[ , "~ <',l i l t
11~I l[)JI IIJLI
"- ',~.. -
]LJ] Illl
Iill
N%__ [111
o
S T D ID Q U A R T Z
9 RAC LiNbO 3 ,~ I M C O N
_
10 \
-%,
-,]
~, x
~'I',11 \ I I 11
: ; :- , , ~ "~ =
9 1 1
10 BANDWIDTH
:'.t',ll .: .: .: .I l l
\
--!11 . 10o (MHz)
O: %
~:~',,,
",
\
~
] I
lOO0
Fig. 30. Time-bandwidth capabilities of linear FM SAW filters using IDT electrode and reflective array techniques.
182
Fred S. Hickerneli
The use of the analog SAW matched filter for pulse expansion and compression is being replaced for some systems by high-speed digital fast Fourier transform (FFT) processors for moderate TB product devices [75]. The power required to implement such digital solutions has been decreasing and TB products from the tens range to several hundred can be achieved with very good side-lobe performance. It is anticipated that continually improving digital FFT architectures will increase TB products while maintaining reasonable power requirements. A second competing approach for obtaining very wide band compressive devices is the use of thin film, high-transition-temperature superconductors [76]. The technology supports multigigahertz bandwidths and time-bandwidth products in excess of 100. D.
UNIFORM DIELECTRIC FILM LAYERS
The presence of a uniform dielectric film layer on the free surface of a SAW substrate alters the propagation characteristics of the surface wave, which can provide an enhancement of substrate properties. It was determined early in the investigation of SAWs that there were velocity, loss, coupling factor, and temperature coefficient improvements to be gained through the use of thin films on piezoelectric substrates [77]. Although this kind of film technology has been used sparingly in SAW device product development, it still represents a potential source for enhancing such products. For example, a glass film that has a strong positive temperature coefficient of frequency (TCF =80), when combined with a high coupling factor piezoelectric substrate with a negative TCF, can improve and even reduce to zero the first-order temperature coefficient without substantially reducing the overall coupling efficiency. Parker and coworkers at Raytheon demonstrated this with glass compensation of lithium tantalate [78]. The TCF was reduced to zero and the coupling efficiency was maintained. The resulting temperature dependence was better than that of ST quartz. Thin glass films have also been used as a practical matter on SAW components as a protective coating, to prevent shorting by metal particles, to reduce metal migration due to electric and acoustic fields, and for frequency trimming. The presence of a dielectric film layer also presents the possibility for Love wave propagation and for higher-order SAW modes. Love wave propagation and the accompanying velocity dispersion of a glass film on silicon was used by Lardat, Maerfeld, and Tournois to develop a large time-bandwidth dispersive delay line [79]. The higher-order SAW modes increase the
3
Surface Acoustic Wave Technology Macrosuccess
183
frequency capability of the SAW propagation for a given transducer electrode spacing. The effects of velocity dispersion, reduced coupling efficiency, and increased propagation loss may have to be accounted for in the use of these higher-order modes. It is interesting to note that the enhancement of coupling factor of a thin film on a piezoelectric substrate is not only affected by the film's elastic properties but also strongly by the dielectric property. Field enhancement at the boundary of the film/substrate can give a significant increase in coupling factor from the value of the permittivity of the film. For example, a ZnO film placed on a YZ lithium niobate substrate almost doubles the coupling, irrespective of the presence of a piezoelectric property in the ZnO film.
IX.
Acoustoelectric Signal Processing
A strong acoustoelectronic effect was first observed with bulk acoustic waves traveling in II-VI and III-V semiconductors in the early 1960s. The piezoelectrically generated electric fields from the elastic wave motion resulted in current flow and a resulting attenuation of the acoustic wave. The application of a dc field in the direction of the wave propagation, which gave a drift velocity to the mobile electrons greater than the phase velocity of the acoustic, resulted in energy being delivered to the acoustic wave and its amplification. This acoustoelectronic effect was studied in the surface acoustic wave mode using piezoelectric semiconductor substrates and combinations of semiconductor materials on or in proximity to piezoelectric substrates. Some combinations of substrate and film are shown in Fig. 31. The arrangement of a high mobility semiconductor in proximity to a strong piezoelectric and a semiconductor with a piezoelectric film were the combinations most used. The original amplifier work was quickly expanded to explore other aspects of current and field interactions such as the strong nonlinear acoustoelectronic interaction between an acoustic wave and semiconductor. An important class of parametric type devices evolved--convolvers and correlators that take the product of two signals and form the convolution or correlation integral of the signals. Also, it was possible to store signals in the adjacent semiconductor and take the correlation at a later time in a memory correlator. Imaging devices were also developed using the photoexcitation of carriers and the subsequent readout by the traveling surface wave. Acoustoelectronic effects were also carried out externally with individual delay line taps connected to semiconductor elements such as diodes and amplifiers. Some very innovative devices were developed and demonstrated, with a major part of the work
184
Fred S. Hickernell SEMICONDUCTING FI LM
/
J
!--'-1
r"--!
r'-"l
PIEZOELECTRIC
PIEZOELECTRIC FILM SEMICONDUCTOR
PIEZOELECTRIC
\
/
FILM
SEMICONDUCTING
FILM
DIELECTRIC
Fig. 31. Combinations of substrates and film layers for the development of SAW acoustoelectronic devices.
being done at Stanford University and the Lincoln Laboratories of Massachusetts Institute of Technology. Kino [80] summarized the first ten years of the work and the breadth of applications. Reible [81] has highlighted the capabilities of gap-coupled convolvers, which have processing gains of 33 dB and dynamic range capabilities of 50 to 60 dB. The configuration effectively used was a semiconductor wafer in proximity to a piezoelectric crystal, as shown in Fig. 32. Under these conditions the properties of semiconductor and piezoelectric could be controlled separately, and the semiconductor could be device-patterned for specific signal processing functions.
Piezoelec rystal m
m
m
m
~
Se uctor mm
m
m l
BasePlate Fig. 32
Separated media acoustoelectronic SAW device.
mm
m
3
A'urJilceAcoustic Wave Technology Macrosuccess
185
Another aspect of acoustoelectronic interactions identified early was the piezoresistive effect that can occur when a traveling SAW stresses the channel of a MOSFET on silicon and alters the current flow of the electrons in the channel region [82]. The illustration of such a device showing ZnO film transduction and MOSFET detection on silicon is shown in Fig, 33. The effect was used to demonstrate a series of programmable filters whose taps were programmed through control of the gate field bias [83]. This permitted tap selectivity and amplitude weighting for biphase and quadraphase correlators. ZnO film transducers were used to generate the surface waves on the MOSFET patterned silicon. Such features as stored codes, random input codes, and contiguous code detection were demonstrated. A major drawback to the implementation of these programmable devices was the power required to operate the devices in an efficient detection mode. The efficiency of the MOSFET detection can be enhanced by the presence of a piezoelectric film such as ZnO on the gate electrode. The acoustic charge transport (ACT) device was an acoustoelectronic device that relied on the capabilities of the piezoelectric field to transport charge carriers of a semiconductor. Thetechnology was given strong support in the use of gallium arsenide, which is both semiconducting and piezoelectric, and it was possible to demonstrate several programmable features
~THIN
FILM -
..... Fig. 33. detectors.
OXX IDEI D E ~
MOSFET
Basic monolithic correlator on silicon with ZnO film transducers and MOSFET
186
Fred S. Hickernell
[84]. The ACT device came very close to being commercialized and produced in volume. Several acoustoelectronic devices were tested in systems and performed well. However, the only acoustoelectronic device productized and developed commercially in significant quantities was a zinc oxide on silicon convolver developed by Clarion for spread spectrum telecommunication systems [85]. X.
Acousto-optics
The periodic expansions and compressions of the traveling surface acoustic wave changes the index of refraction, which will scatter a guided optical wave at the surface of a solid. This effect has been employed to demonstrate a number of signal processing devices including Bragg spectrum analyzers and optical computing. Unique functional devices have been demonstrated but have not been developed into products. Lewis [86] has compared optical and acoustical signal processing techniques including acousto-optics. XI.
SAW Sensors
Sensors provide an interface between electronic equipment and the physical world. They convert physical or chemical quantities into electrical signals. Surface wave sensors sense changes in the characteristics of the path over which the surface acoustic wave travels. These changes may be a velocity change and/or propagation loss due to a sensing response. For detecting thermal properties such as temperature, the change in elastic constants result in a velocity change or phase shift. Similarly for mechanical changes such as force, pressure, or acceleration, the elastic properties of the SAW device will produce a propagation velocity change. For detecting chemicals, magnetic, or radiant energy, a film that is sensitive to the particular quantity being sensed is applied to the surface in the propagation path and the resultant velocity and loss change is measured. The major markets where sensors are presently in demand are in the automotive, industrial, biomedical, and environmental areas. SAW sensors have primarily been used in of chemical, biomedical, and environmental sensing. A recent book on sensors authored by a group of scientist engaged in acoustic wave sensor work contains information on SAW sensors and their application [87]. As the substantial increase in the number of SAW sensor publications in the past five years shows, SAW sensors are a very important area for future development.
3
Surface Acoustic Wave Technology Macrosuccess XII.
187
Future Success
The future success of SAW technology will rely on its capabilities as a unique signal processing device that works hand in hand with digital circuitry. The strengths of SAW technology are its capabilities of passive signal processing in the time and frequency domain and the access to, and sensitivity of, its surface. In its early history, SAW technology found applications as replacements for electronic components, expanded the time-bandwidth dimensions of signal processing, and made a platform available for interactive signal processing. It now tides on the crest of a telecommunication bonanza that needs its frequency sorting capabilities. As telecommunications move to higher frequencies with features requiting greater bandwidth, part of the future success of SAW technology will be to meet this challenge. This will put greater demands on material selection, processing yields, and the ingenuity of the SAW design engineer. In 1985, six areas of technology improvement were identified as essential to the significantly expanded use of SAW technology [88]. Each one of those areas have been aggressively pursued and for the most part have accounted for the increased growth in SAW market areas. The SAW technology challenges were (1) improved low-loss filters, (2) improved packaging, (3) microwave frequency devices, (4) programmable filters, (5) more accurate filter design techniques, and (6) improved materials and processing. For improved low-loss filters, SAW resonator, IIDT, SPUDT, and similar techniques that retain the SAW energy within the electrode structures have brought loss levels under 3 dB. High-power capability is still evolving, with 1-3 W power levels demonstrated over significant time periods. Low-cost packaging with hermetic capabilities have been achieved with single metal layer ceramics (ceramic chip carriers) sealed with solder or epoxy capable of direct surface board mounting. Lower-cost plastic molded packaging and chip packaging are being pursued to further reduce the cost of volume-produced SAW devices. SAW devices have been demonstrated for frequencies as high as 10 GHz. Volume production with good reliability and yield is now possible for devices operating in the 1-3 GHz region. SAW programmable filters and associated spread spectrum techniques have not been implemented in any large volume communication systems. They remain in customized systems, although digital technology has replaced many of their functions. There still is a strong need for programmability of frequency and bandwidths in SAW filters as the requirement for cell phones that work anywhere in the world become more in demand. The design of filters, including second-order effects,
188
Fred S. Hickernell
parasitics, and packaging, has evolved into the three-dimensional modeling realm with very sophisticated mathematical and computational techniques to account for and predict performance. Finally, though few new materials have been investigated in the past ten years, better cuts and surface processing of standard SAW materials have improved performance capabilities. This, coupled with better automated wafer processing and manufacturing techniques, has given high yield at achieved low cost. What are the challenges and possibilities of the next ten years? Certainly the mood is for higher frequencies, wider bandwidths, and higher power handling, while maintaining low-loss for filters in the telecommunication area. Higher-frequency resonators and oscillators will become available, as will new lower-loss packaging. Another look will be given at convolvers and correlator geometries for spread spectrum communication systems. SAW tags will be pursued to become more price competitive with competing technologies. The simple coding and encoding features of SAW tapped delay lines need to be exploited as much as possible for identification tagging, keyless entry, food processing, and security-related functions. SAW sensor activity will substantially increase and businesses will evolve to service the automotive, chemical, and environmental needs. Independent SAW businesses will continue to spin off from universities and industry, will often find distinct market niches that will add to the growth of the SAW area. The signal processing capabilities of SAW devices may be still underutilized because of the lack of understanding of its capabilities and entrenched digital technologies. Architectures have been suggested, which working together with digital integrated circuits, enhance performance and reduce power. There will continue to be the requirement for low-volume, custommade SAW devices, which when coupled with digital technology give a distinct advantage in the signal processing functions of sorting and identification. These signal processing architectures will take advantage of the wide bandwidth passive SAW devices capability, minimizing the power requirements for the overall system. It is incumbent on the SAW community to recognize these opportunities, educate the system designers, and promote these integrated technologies where applicable. Because of the "chip" nature of SAW devices, they are a natural to be integrated into functional modules in hybrid form or as multilayer microwave integrated circuits. The promise of a semiconductor IC chip with integrated acoustic filter and resonator functions using piezoelectric films remains to be fulfilled. The sensor area is a rapidly growing business, and SAWs can bring some unique features to this area. SAW devices with superior sensing properties
3
Surface Acoustic Wave Technology Macrosuccess
189
have already been demonstrated in the environmental, chemical, and biological areas. The challenge is to create compact, low-cost smart sensors as a part of a complex of microsensors interfacing with semiconductor logic circuitry. SAW engineers need to bring the unique features of SAWs to complement other microsensor technology. Another aspect of the future success of SAWs relates to the cooperative efforts that are possible between SAW businesses and university and government laboratories and even among businesses that may have some unique technical competencies to be shared. These cooperative efforts are now taking place across country boundaries and continents. It is important that support continue to those countries, such as Eastern Europe and the former Soviet Union, where political and economic conditions have drastically reduced support of scientific activities and technological developments. There remains a resource of very highly qualified scientists and engineers whose contributions can benefit both the supporter and the supported. It is a win-win situation. An excellent example of what has been done with SAW technology is contained in a recent publication that documents the 30+ years of SAW contributions at the University of Trondheim in Norway and the resulting businesses and application areas [89]. In the introduction of 'From Science to Enterprise,' Kjell Ingebritsen states, "The success of applied research is measured through the applications . . . . It gives specific satisfaction to see applications of surface acoustic waves being the technological fundament for industrial ventures in Norway." The application areas referred to extend from sea bottom exploration, through land-based vehicle and personnel identification, to earth resource satellite surveillance. SAW technology does have a furore. It is important to look beyond the economic success enjoyed by the high-volume production of the filters and resonators of today's world and envision a world where the microseisms of SAW technology rumbling across piezoelectric plates enable new device applications for furore advanced electronic systems.
Acknowledgments The author expresses his thanks for the conversations and discussions with the engineers of the SAW Business Unit at Motorola and to a group, to numerous to name, of engineers, technicians, and process personnel over the past 32 years, who have supported the SAW work at Motorola. Dr. Vladimir Anisimkin, Professors Igor Yakovkin and Sergei Kulakov, and Academician
Fred S. Hickernell
190
Yu. V. Gulyaev kindly supplied information on the development of SAW technology in Russia. Several university professors in Japan were helpful in supplying information on SAW developments in that country. Several of the SAW businesses worldwide kindly sent their catalogue and company profiles. A recent presentation by David Morgan on the history of SAW devices will appear in the proceedings of the 1998 IEEE International Frequency Control Symposium [90]. The author has been privileged to be associated with such a vibrant technology over so many years, and has been blessed to personally know so many of the contributors to SAW technology. References 1. Lord Rayleigh. (1885). On waves propagating along the plane surface of an elastic solid. Proc. London Math. Soc. 17, 4-11. 2. Love, A. E. H. (1911). "Some Problems of Geodynamics." Cambridge University Press, London, Dover 1967. 3. Lamb, H. (1917). On waves in an elastic plate. Proc. Roy. Soc. (London) Ser. A 93, 114. 4. Stoneley, R. (1924). Elastic waves at the surface of separation of two solids. Proc. Roy. Soc. A106, 416. 5. Stoneley, R. (1955). The propagation of surface elastic waves in a cubic crystal. Proc. Roy. Soc. A232, 447. 6. Kraut, E. A. (1971) Surface elastic waves ... a review. In "Acoustic Surface Wave and Acoustooptic Devices" (Thomas Kallard, ed.). Optosonic Press, New York. 7. Lim, T. C., and Farnell, G. W. (1968). Search for forbidden directions of elastic surface wave propagation in anisotropic crystals. J. Appl. Phys. 39, 4319. 8. Bleustein, J. O. (1968). A new surface wave in piezoelectric crystals. Appl. Phys. Lett. 13, 411-413. 9. Gulyaev, Y. V. (1969). Electroacoustic surface waves in solids. Sov. Phys. JETP Lett. 9, 37-38. 10. Ohta, Y., Nakamura, K., and Shimizu, H. (1969). Piezoelectric surface shear wave. Tech. Rep. Instr. Electr. and Commun. Eng. of Jpn US69-3, 1, (in Japanese). 11. Farnell, G. W., and Adler, E. L., (1972). Elastic wave propagation in thin layers. In "Physical Acoustics," Vol. IX (W. P Mason and R. N. Thurston, eds.) Academic Press, New York pp. 35-127. 12. Vinogradov, K. N., and Ul'yanov, G..K. (1959). Measurements of the SAW propagation velocity and attenuation in solids. Sov. Phys. "Acoustics 5, 290-293. 13. Victorov, I. A. (1967). "Rayleigh and Lamb Waves--Physical Theory and Applications." Plenum Press, New York. 14. Mortley, W. S., (May 1963). British patent 988,102. 15. Rowen, J. H. (1964). High frequency dispersive ultrasonic delay lines. 1964 Ultrason. Symp., paper J-6, Santa Monica, CA. 16. May Jr., J. E. (1965). "Ultrasonic traveling-wave devices for communication," IEEE Spectrum, October, 73-85. 17. Rowen, J. H. (November 1966). "Tapped Ultrasonic Delay Line and Uses Therefore." U.S. patent 3,289,114. 18. Smirnov, Yu. G., (1965). Experimental study of the SAW propagation in piezoelectric plates of quartz. Trans. of the Leningrad Institute of Aerospace Instrumentation N-45, 10-16, (in Russian). 19. White, R. M., and Voltmer, E W. (1965). Direct piezoelectric coupling to surface elastic waves. Appl. Phys. Lett. 7, 314-316.
3
S u r f a c e A c o u s t i c Wave T e c h n o l o g y M a c r o s u c c e s s
191
20. Smith, W. R., Gerard, H. M., Collins, J. H., Reeder, T. M., and Shaw, H. J., (1969a). Analysis of interdigital surface wave transducers by use of an equivalent circuit model. IEEE Trans. Microwave Th. and Techn., 17, 856-864. 21. Smith, W. R., Gerard, H. M., Collins, J. H., Reeder, T. M., and Shaw H. J., (1969b). Design of surface wave delay lines with interdigital transducers. IEEE Trans. Microwave Th. and Techn. 17, 865-872. 22. Tancrell, R. H., and Holland, M. G. (1971). Acoustic surface wave filters. Proc. IEEE. 59, 393-409. 23. Hartmann, C. S., Bell Jr., D. T., and Rosenfeld, R. C. (1973). Impulse model design of acoustic surface-wave filters. IEEE Trans. Microwave Th. and Techn. 21, 162-175. 24. Wright, P. V. (1989). A new generalized modeling of SAW transducers and gratings. Proc. 43rd Symp. Freq. Ctrl. 596-605. 25. Milsom, R. E, Reilly, N. H. C., and Redwood, M. (1977). Analysis of generation and detection of surface and bulk acoustic waves by interdigital transducers. IEEE Trans. Son. & Ultrason. 24, 147-166. 26. Vasile, C. E (1974). A numerical Fourier transform technique and its application to acousticsurface-wave bandpass filter synthesis and design. IEEE Trans. Son. & Ultrason. 21, 7-11. 27. Ruppel, C. C. W., Ruile, W., Scholl, G., Wagner, K. Ch., and Manner, O. (1994). Review of models for low-loss filter design and applications. Proc. 1994 IEEE Ultrason. Symp., 313-324. 28. Schulz, M. B., Matsinger, B. J., and Holland, M. G. (1970). Temperature dependence of surface acoustic wave velocity on alpha-quartz. J. Appl. Phys., 14, 2755-2765. 29. Shibayama, K., Yamanouchi, K., Sato, H., and Meguro, T. (1976). Optimum cut for rotated Y-cut LiNbO3 crystal used as the substrate of acoustic-surface-wave filters. Proc. IEEE., 64, 595-597. 30. Schulz, M. B., and Holland, M. G. (1972). Temperature dependence of surface acoustic wave velocity in lithium tantalate. IEEE Trans. Son. & Ultrason., 19, 381-384. 31. Slobodnik Jr., A. J., Conway, E. D., and Delmonico, R. T. (1973). "Microwave Acoustics Handbook: Volume 1A. Surface Wave Velocities." Air Force Cambridge Research Laboratories, Bedford, Massachusetts. 32. Slobodnik Jr., A. J., Conway, E. D., and Delmonico, R. T. (1974). "Microwave Acoustics Handbook: Volume 2. Surface Wave Velocities--Numerical Data." Air Force Cambridge Research Laboratories, Bedford, Massachusetts. 33. Kovacs, G., Amhom, M., Engan, H. E., Vistini, G., and Ruppel, C. W. (1991). Improved material constants for LiNbO3 and LiTaO3. Proc. 1991 IEEE Ultrason. Symp., 435-438. 34. Adler, E. L., Slabosczewicz, J. K., Farnell, G. W., and Jen, C. K. (1990). PC software for SAW propagation in anisotropic multilayers. IEEE Trans. Ultrason., Ferroelec. and Freq. Ctrl. 37, 215-223. 35. Coldren, L. A., and Shaw, H. J. (1976). Surface-wave long delay lines. Proc. IEEE. 64, 598-609. 36. Cho, E, Adamo, M., Hickemell, E, and Yarrington, L. (1976). Manufacturing technology for a SAW hybrid tapped delay line. Proc. 1976 IEEE Ultrason. Symp., 528-531. 37. Hartemann, E, and Dieulesaint, E. (1969). Acoustic-surface-wave filters. Elect. Lett. 5, 657-658. 38. Kallman, H. E. (1940). Transversal filters. Proc. IEEE 28, 302-310. 39. DeVries, A. J., Sreenivasan, T., Subramanian, S., and Wojcik, T. J. (1974). Detailed description of a commercial surface-wave TV IF filter. Proc. 1974 IEEE Ultrason. Symp., 147-152. 40. Hunsinger, W. J. (1979). SAW filter applications in consumer electronics. Proc. 1979 IEEE Ultrason. Symp., 541-544. 41. DeVries, A. J., and Adler, R. (1976). Case history of a surface-wave TV IF filter for color television receivers. Proc. IEEE 64, 671-676. 42. Parker, D. W., Pratt, R. G., and Stevens, R. (1976). A television IF acoustic surface wave filter on bismuth silicon oxide. Proc. IEEE 64, 677-681.
192
F r e d S. H i c k e r n e l l
43. Tobolka, G., Faber, W., Albrecht, G., and Pilz, D. (1984). High volume TV IF filter design, fabrication, and applications. Proc. 1984 IEEE Ultrason. Symp., 1-12. 44. Fujishima, S., Ishiyama, H., Inoue, A., and Ieki, H. (1976). Surface acoustic wave VIF filters for TV using ZnO sputtered film. Proc. 30th Symp. Freq. Ctrl., 119-122. 45. Hickernell, E S. (1988). High-reliability SAW bandpass filters for space applications. IEEE Trans. Ultrason., Ferroelec. and Freq. Ctrl. 35, 652-656. 46. Henaff, J. and Brossard, P. C. (1981). Implementation of satellite communication systems using surface acoustic waves. IEEE Trans. Microwave Th. and Tech. 29, 439-550. 47. Peach, R. C. (1995). SAW based systems for communication satellites. Proc. 1995 IEEE Ultrason. Symp., 159-166. 48. Allen, D. E., Arneson, S. H., and Hickernell, E S. (1980). Surface acoustic waves: a solution to frequency sorting receivers. Proc. 24th Int. Tech. Symp. SPIE 239, 209-219. 49. Laker, K. R., Budreau, A. J., and Carr, P. H. (1967). A circuit approach to SAW filterbanks for frequency synthesis. Proc. IEEE 64, 692-695. 50. Hartmann, C. S. (1973). Weighting interdigital surface wave transducers by selective withdrawal of electrodes. Proc. 1973 IEEE Ultrason. Symp., 423-426. 51. Lewis, M. E (1982). SAW filters employing interdigitated interdigital transducers. Proc. 1982 IEEE Ultrason. Syrup. 12-17. 52. Hartmann, C. S., Wright, R V, Kansy, R. J., and Garber, E. M. (1982). An analysis of SAW interdigital transducers with internal reflections and the application to the design of single-phase unidirectional transducers. Proc. 1982 IEEE Ultrason. Symp. 40-45. 53. Coon, A. (1991). SAW filters and competitive technologies. Proc. 1991 IEEE Ultrason. Symp., 155-160. 54. Ash, E. A. (1970). Surface wave grating reflectors and resonators. Digest of the IEEE Symp. on Microwave Th. and Techn., 385-386. 55. Mierzwinski, M. E., and Terrien, M. E. (1981). "Hewlett-Packard Journal," Vol. 32. (R. P. dolan, ed.) 15-16. 56. Bray, R. C. (1986) Applications of SAW devices in high performance instrumentation. Proc. 1978 IEEE Ultrason. Symp., 299-307. 57. Bell Jr., D. T., and Li, R. C. M. (1976). Surface-acoustic-wave resonators. Proc. IEEE, 64, 711-721. 58. Coldren, L. A., and Rosenberg, R. L. (1978). SAW resonator filter overview: design and performance trade-offs. Proc. 1978 IEEE Ultrason. Symp., 422-432. 59. Wright, P. V. (1992). A review of SAW resonator filter technology. Proc. 1992 IEEE Ultrason. Symp., 29-38. 60. Hickernell, T. S. (1994). Development of a SAW ladder filter for a portable phone system. Proc. 16th Piezoel. Dev. Conj.', 106-111. 61. Lewis, M. E (1974). The surface acoustic wave oscillator--a natural and timely development of the quartz crystal oscillator. Proc. 28th Symp. Freq. Ctrl., 304-314. 62. Montress, G. K., Parker, T. E., and Andres, D. (1994). Review of SAW oscillator performance. Proc. 1992 IEEE Ultrason. Syrup., 43--54. 63. Lattanza, J., Herring, E G., Krencik, P. M., and Clerihew, A. E (1983). 240 MHz wideband programmable SAW matched filter. Proc. 1983 IEEE Ultrason. Symp., 143-150. 64. Lagasse, R E., Mason, I. M., and Ash, E. A. (1973). Acoustic surface waveguides - - analysis and assessment. IEEE Trans. Microwave Th. and Techn. 21,225-236. 65. Schmidt, R. V., and Coldren, L. A. (1975). Thin film acoustic surface waveguides in anisotropic media. IEEE Trans. Son. & Ultrason. 22, 115-122. 66. Yao, I. (1981). High performance elastic convolver with extended time-bandwidth product. Proc. 1981 IEEE Ultrason. Symp., 181-185. 67. Morgan D. R (1985) "Surface-Wave Devices for Signal Processing." Elsevier Science, New York, Paperback Edition (1991 ), 294-311.
3
Surface Acoustic
Wave Technology Macrosuccess
193
68. Marshall, E G., and Paige, E. G. S. (1971). Novel acoustic-surface-wave directional coupler with diverse applications. Elect. Lett. 7, 460-462. 69. Marshall, E G., Newton C. O., and Paige, E. G. S. (1973). Theory and design of the surface acoustic wave multistrip coupler. IEEE Trans. Microwave Th. and Techn. 21,206-215. 70. Marshall, E G., Newton, C. O. and Paige, E. G. S. (1973). Surface acoustic wave multistrip components and their applications. IEEE Trans. Microwave Th. and Techn. 21,216-225. 71. Williamson, R. C. (1976). Properties and applications of reflective array devices. Proc. IEEE 64, 702-710. 72. Williamson, R. C., and Smith, H. I. (1973). The use of surface elastic wave reflection gratings in large time-bandwidth pulse-compression filters. IEEE Trans. Microwave Th. and Techn. 21, 195-205. 73. Gerard, H. M., Otto, O. W., and Weglein, R. D. (1974). Development of a broadband reflective array 10,000:1 pulse compression filter. Proc. 1974 IEEE Ultrason. Symp., 197-201. 74. Solie, L. E (1977). Reflective dot array devices. Proc. 1977 IEEE Ultrason. Syrup., 579-584. 75. Tortoli, P., Guidi, E, and Atzeni, C. (1994). Digital vs. SAW matched filter implementation for radar pulse compression. Proc. 1994 IEEE Ultrason. Syrup., 199-202. 76. Lyons, W. G., Arsenault, D. R., Anderson, A. C., Gerhard Sollner, T. C. L., Murphy, P. G., Seaver, M. M., Boisvert, R. R., Slattery, R. L., and Ralston, R. W. (1996). High-Tc superconductive wideband compressive receivers. The Lincoln Laboratory J., 33. 77. Hickernell, E S. (1973). The role of layered structures in surface acoustic wave technology. Proc. Int. Specialist Seminar on Component and Systems AppL of Surface Acoustic Wave Dev., 11-21.
78. Parker, T. E., and Wichansky, H. (1979). Temperature-compensated surface-acoustic-wave devices with SiO2 film overlays. J. Appl. Phys. 50, 1360-1369. 79. Lardat, C., Maerfeld, C., and Toumois, P. (1971). Theory and performance of acoustical dispersive surface wave delay lines. Proc. IEEE. 59, 355-368. 80. Kino, G. S. (1976). Acoustoelectric interactions in acoustic-surface-wave devices. Proc. IEEE. 64, 724-748. 81. Reible, S. A. (1981). Acoustoelectric convolver technology for spread-spectrum communications. IEEE Trans. Son. & Ultrason. 28, 185-1195. 82. Claibome, L. T., Staples, E. J., and Harris, J. L. (1971). MOSFET surface wave detectors for programmable matched filters. Appl. Phys. Lett. 19, 58-60. 83. Hickernell, E S., Adamo, M. D., De Long, R. V., and Hinsdale, J. G. (1980). SAW programmable matched filter signal processor. Proc. 1980 IEEE Ultrason. Symp., 104-108. 84. Hoskins, M. J., and Hunsinger, B. J. (1982). Recent developments in acoustic charge transport devices. Proc. 1986 IEEE Ultrason. Symp., 439-450. 85. Miagawa, S., Okamoto, T., Niitsuma, T., Mitsutsuka, S., Tsubouchi, K., and Mikoshiba, N. (1984). Sezawa wave correlator using monolithic ZnO/SiO2/Si structure. Proc. 1984 IEEE Ultrason. Symp., 298-302. 86. Lewis, M. E (1985). A comparison of optical and acoustical signal processing techniques. Proc. IEEE Ultrason. Symp., 114-123. 87. Ballantine, D. S., White, R. M., Martin, S. J., Ricco, A. J., Zellers, E. T., Frye, G. C., and Wohltjen, H. (1997). "Acoustic Wave Sensors, Theory, Design, and Physico-Chemical Applications." Academic Press, New York. 88. Hartmann, C. S. (1985). Future high volume applications of SAW devices. Proc. IEEE Ultrason. Symp., 64-73. 89. Ingebritsen, K. A. (Ed.) (1992). "From Science to Enterprise." University of Trondheim. 90. Morgan, D. P. (1998). History of SAW devices. Proc. IEEE Intl. Frequency Control Symposium, Pasadena, May 1998.
194
Fred S. Hickerneli Appendix
A.
SAW Publications
There have been over 20,000 SAW publications indexed since 1965. Figure A-1 graphically shows the number of publications by year from 1960 through 1996. Shortly before 1970, there began a dramatic rise in publications that leveled near 800 per year for about 15 years and then displayed another rise. The rise after 1990 is due in part to the increase in sensor-related papers. SAW theses from universities, colleges, and institutes in the United States, Canada, and in Russia are shown in Fig. A-2. The numbers for Russia are taken from a presentation by Professor Mozhaev at the 1994 Ultrasonics Symposium in Cannes that was published in those conference proceedings. It is interesting to note that there was some SAW activity in the former Soviet Union as early as 1949, which continued into the 1960s when the first theses were published in the United States. The theses work in Russia exceeded that of the United States and Canada in the 1970s and 1980s before dropping off in the 1990s. The United States and Canadian contributions came from 58 universities, with 81 contributions from Stanford, Purdue, University of Califomia at Berkeley, Rensselaer Polytechnic Institute, and the University of Illinois. The total number of universities and institutes in the former Soviet Union that contributed theses based on publications is around 50. 1400
~iiiii~~ i!
1200lOOO
,,~
iiiiiiii
!
~ ii
I
,
ili
ii
i!i
i
::
i
!
........................ i ............... ~............... !-.-.-t.---',--.-t..--t...-t...-f...i-....~.....u...t-.--.i : i ~ ~ ~ ~ ~ : ~: ~ ~ j I9 li. i
:
= ~ ~ !
J
J
J J
i
i
i
i
;
i
' [
, t
II
il
I
, +
,
.--.- t I " i l l l,
........ f ............ -..'..................... , ...... i ' ! ~ I ; J[ | i t! J
i
,:
[
,,,
:
I:
:!
I
:i
i
i
.
i /'
|
l" ...... lm= ,
l=ilil '
............................i---.~----.[d--..!...............i....!.+.f-+++..+.--f--.-{ ...............f.-i.--..f+++++.f...i.-...i.-..4=;IH "
,,,,
ti
t
tl
:i
! 'I Iiiil i ~ ' !........ ! , ~+. r . ~ t l .
i i l l
uJn"
800-
~ '!'+~T , t r t~: ...............................
Z
ill ~J i ! ! Jliilj~i],] 6 0 0 ............................................... j ..........j.........I" ........~i
[
i
=~ 400-
...................................... [ f i f
200-
.......................................... i............. i..4....~
i i
,
' iiiiii!il 1961
! ,
~i
i
1966
!" J
"
~
' i
1971
'
j .
I
"I
t ~1
'
;
' !,,it,
j - ~ i i' ~i , l i i '1 t '~ J : i i i~i iiJiJ
1976
1981
1986
!"
i,
i
i l
) }I
1991
,
1I
!
~ [t ii Ii.i " i! !
!
!'i! 9
i
i
I
. I If ,
~Ji
l]
I
II
:
,ii[
]t
j
'
' I
. . . .
. l t .......... . i f .......................} t ~ ..........
i
......... f+"f-"'
i
[],
alUll ~ "1
:
'
j
[
~t i
i i'l
! i ! i
,)~ ! I !I!
,
1996
YEAR
Fig. A-1.
Surface acoustic wave literature publications from 1960 to 1996.
3
Surface Acoustic Wave Technology Macrosuccess 40
iiiii!iiiiiiiiii!ff!!iiiiiiiiii!i!iii
30-
,
.
.
2 5 ........................ri-!TTI.",:.I-i-!Ii ! l l t i i l f l i i i i i i i l
I
~'
.
.
.
.
.
II
iii!iiiiiiiiiii!iiiiiiJii~iillil
[i
1
5
iiiiiiiii
tiilfiiii
i ii
i i I ili
liiiiiii!i!!!ii~',!iiiiii~iii
i
i i
i '~ -ifl
iiiiiiii!iiltiiiii
i(i
t li
i
Iii
i ! i
i
i
i! !i ii ii ii iilill
ill
iii
!ill
'
i I t i !
..............-'.--.v.+--~,--...t-:..---+-~~.+--i-ii--.i-i !i
5
..i
I
_
i
iillliiiti
J"ii
1956
196~!.
!
l i l l ' i i i , ' l
_llu $i,,d~ m i.i iI i I I I I I I I i i i i i i
1948
i
i.
ili i
J
I
"
'il
i
'i
i i
,
i i i i i i i i I
1972
i i
i
i'
i i
~o...................7]-K!-i-f--H--f-Ki-! ......P.I.-I...K-I--i !' ii " i i
iii
i:jii[ l-i
ii
it
20....................ii-iii-i-ii}-i)ti-iiKiilfl-it7
iiiiiliti
.
........iTTirT--".;--rl-~-7!-i-r.-! i i i!i ill
ii!iiiiill
=3
.
195
i
"
I
~
i
'il
I !~
t-i
Ii
...... tt-i-~ ...................... ~ i
"
'
ii.l!ii
-t-i-i~ ....................
i'iill
'
i i
i
'
t
I
l
i
i i i i
iI
i
i
I
l
~
t
i i i i i i i 1 i
1988
!
'!ili~i
J J i
1980
t
i
.-' .................. ' ...................... i i i i
1996
YEAR
Fig. A-2. Russia.
SAW-related theses and dissertations published in the United States, Canada, and
References The following is a list of SAW-related books and monographs that have been written since 1967. They cover a wide area of SAW technology from collections of papers, to textbooks, to review articles, to applications, etc. This list is followed by a series of journal review articles. Each is a glimpse in time of the state of the art as SAW technology has advanced. The importance of these longer publications is the valuable references in areas of particular interest to the reader.
BOOKS AND MONOGRAPHS FOR APPENDIX A 1. Viktorov, I. A. (1967). "Rayleigh and Lamb Waves: Physical Theory and Applications." Plenum Press, New York. 2. Kallard, T. (Ed.) (1971). "Acoustic Surface Waves and Acousto-optic Devices." Optosonic Press, New York. 3. Auld, B. A. (1973). "Acoustic Fields and Waves in Solids." Vol. 2. Wiley, New York. 4. Morozov, A. I., Proklov, V. V., Stankovskii, B. A., and Gingis, A. D. (1973). "Piezosemiconductor Transducers and Their Applications." Energy, Moscow, (in Russian). 5. Morgan, D. P. (Ed.) (1976). "Key Papers on Surface Acoustic Wave Passive Interdigital Devices." Peter Peregrinus, Stevenage.
196
F r e d S. H i c k e r n e l l
6. Karinskii, S. S. (1975). "SAW Devices for the Signal Processing." Soviet Radio, Moscow, (in Russian). 7. Matthews, H. (Ed.) (1977). "Surface Wave Filters: Design, Construction, and Use." John Wiley & Sons, New York. 8. Oliner, A. A. (Ed.) (1978). "Acoustic Surface Waves." Springer-Verlag, Berlin. 9. Yakovkin, I. B., and Petrov, D. V. (1979). "Light Diffraction by Acoustic Surface Waves." Nauka, Novosibirsk, (in Russian). 10. Dieulesaint E., and Royer D. (1980). "Elastic Waves in Solids." John Wiley & Sons, New York. 11. Rechtskii, V. I. (1980). "Acoustoelectronic Components: Elements and Devices Based upon SAW." Soviet Radio, Moscow, (in Russian). 12. Viktorov, I. A. (1981). "Sound Surface Waves in Solids." Nauka, Moscow, (in Russian). 13. Morozov, A. I., Proklov, V. V., and Stankovskii, B.A. (1982). "Piezoelectric Transducers for the Radioelectronic Devices." Radio and Communication, Moscow, (in Russian). 14. Shaskolskaia, M. P. (Ed.) (1982). " Acoustic Crystals." Science, Moscow, (in Russian). 15. Balakirev, M. K., and Gilinskii, I. A. (1982). "Waves in Piezoelectric Crystals." Nouka, Novosibirsk, (in Russian). 16. Ristic, V. M. (1983). "Principles of Acoustic Devices." John Wiley & sons, New York. 17. Orlov, V. S., and Bondarenko, V. S. (1984). "SAW Filters." Radio and Communication, Moscow, (in Russian). 18. Rechitskii, V. I. (1984). "SAW Radiocomponents." Radio and Communication, Moscow, (in Russian). 19. Morgan, D. P. (1985). "Surface-Wave Devices for Signal Processing." Elsevier, Amsterdam. 20. Dmitriev V. V., Akpambetov, V. B., Bronnikova, E .G., et al. (XXXX). "Integral Piezoelectric Devices for the Signal Filtering and Processing." Radio and Communication, Moscow. (in Russian). 21. Zelenka, J. (1986). "Piezoelectric Resonators and Their Applications." Elsevier, Amsterdam. 22. Datta, S. (1986). "Surface Acoustic Wave Device." Prentice-Hall, Englewood Cliffs, New Jersey. 23. Kino, G. S. (1987). "Acoustic Waves: Devices, Imaging, and Analog Signal Processing." Prentice-Hall, Englewood Cliffs, New Jersey. 24. Malischewsky, P. (1987). "Surface Waves and Discontinuities." Elsevier, Amsterdam. 25. Kochemasov V. N., Dolbnya V. I., and Sobol, V. (1987). "Acoustoelectronic Fourier-Transformers." Radio and Communication, Moscow, (in Russian). 26. Campbell, C. (1989). "Surface Acoustic Wave Devices and Their Signal Processing Applications." Academic Press, San Diego. 27. Feldmann, M., and Henaff, J. (1989). "Surface Acoustic Waves for Signal Processing." Artech House, Boston. 28. Parker, D. E, and Maugin, G. A. (Eds.) (1989). "Recent Developments in Surface Acoustic Waves." Springer-Verlag, Berlin. 29. Biryukov, S. V., Gulyaev, Yu. V., Krylov, V. V., and Plessky, V. P. (1991). "Surface Acoustic Waves in Inhomogeneous Media." Nauka, Moscow, (in Russian; English translation will be published by Springer-Verlag). 30. Kress, W. and de Wette, E D. (Eds.) (1991). "Surface Phonons." Springer-Verlag, Berlin. 31. Briggs, A. (1992). "Acoustic Microscopy." Oxford University Press, New York. 32. Garcia-Moliner, E, and Velasco, V. R. (1992). "Theory of Single and Multiple Interfaces." World Scientific Publishing Company, Singapore. 33. Biryukov, S. V., Gulyaer, V. V., Krylov, V. V., and Pleasky, V. P. (1995). "Surface Acoustic Waves in Inhomogeneous Media." Springer-Verlag, Berlin. 34. Campbell, C. (1998). "Surface Acoustic Wave Devices for Mobile and Wireless Communications." Academic Press, San Diego.
3
Surface Acoustic Wave Technology Macrosuccess
197
The following comprehensive reviews from joumals, either authored or edited, promoted SAW interest in the wider technical community. Along with these technical contributions there were a number of articles that were technical, tutorial, and publicity minded in nature and that appeared in trade magazines starting in 1969 and extending through the 1970s and early 1980s. These articles, which also generated wide interest in SAW device applications, appeared as a series or individual articles in Electronics, Microwave Journal, Microwave Systems News, Physics Today, Spectrum, and IEEE Magazine, among others. In May 1973, Ultrasonics began a series of articles, primarily by British authors, that extended through 1975. JOURNAL R E V I E W ARTICLES 1. Special Issue on Microwave Acoustics. IEEE Trans. Microwave Th. and Techn. 17, November 1969. 2. White, R M. (1970). Surface elastic waves. Proc. IEEE 58, 1239-1276. 3. Joint Special Issue on Microwave Acoustic Signal Processing. IEEE Trans. Microwave Th. and Techn., 21, April 1973 and IEEE Trans. Son. Ultrason. 20, March 1973. 4. Special Issue on Surface Acoustic Wave Devices and Applications. Proc. IEEE 64, May 1976. 5. Special Issue on Computer Aided Design of SAW Devices. Wave Electr. 2, July 1976. 6. Joint Special Issue on Surface Acoustic Wave Devices and Applications. IEEE Trans. Microwave Th. & Tech. 29, and IEEE Trans. Son. & Ultrason. 28, May 1981. 7. Special Issue on Acoustic Microscopy. IEEE Trans. Son. & Ultrason. 32, May 1985. 8. Special Issue on SAW Convolvers and Correlators. IEEE Trans. Son. & Ultrason. 32, September 1985. 9. Special Issue on Acoustic Sensors. IEEE Trans. Ultrason., Ferroelec., and Freq. Cntr. 34, March 1987. 10. Special Issue on SAW Applications. IEEE Trans. Ultrason., Ferroelec., and Freq. Cntr. 35, November 1988. 11. Campbell, C. K. (1989). Application of surface acoustic and shallow bulk acoustic wave devices. Proc. IEEE 77, 1453-1484. 12. Special Issue on Thin Films for Acoustoelectronics. IEEE Trans. Ultrason., Ferroelec., and Freq. Ctrl. 42, May 1995.
Appendix B.
SAW Conferences
The advent of the interdigital transducer and associated device developments stimulated the need for workers in SAW technology to share their work in a more direct and personal way. In the period from 1965 to 1970, SAW activity was described in the established scientific and engineering conferences dealing with ultrasonics and microwave electronics. Most of these contributions were not recorded in conference proceedings but later appeared in internal and government reports and in the open literature. In the United States, most of the SAW work was reported at the annual IEEE Ultrasonics
198
Fred S. Hickernell
Symposium. In 1970, SAW contributions dominated that symposium, with 59 papers on surface acoustic waves out of the total of 105 papers presented. Figure B-1 shows the contributions to symposium publications by industry, universities, government-supported institutes and laboratories, and independ e n t S A W businesses. Industries and universities have contributed fairly equally over the years, followed by government-supported organizations, with slowly increasing participation by private SAW businesses. In many cases, such as in Japan, the SAW business units are embedded within an industrial complex. As indicated by the graph, the most symposium published work, which is an indicator of the activity level in SAW technology, was strong in the 1970s and early 1980s. Since then it has maintained a consistent level of activity. Although government spending accounted for the majority of the early support of SAW activities, more support now comes from business and industrial entities. The IEEE Ultrasonics Symposium has continued to have a strong component of SAW work, with the majority of papers now being presented by workers outside the United States and Canada. The graph of Fig. B-2 shows the number of invited and contributed papers on SAW technology given by the two groups. Since 1990, the international contributions have exceeded
i
1 O0
~ J i : I i
i
i
! ! i i :, i
i i
i i J J i '. ! i ! i i
~ i i i ~ i
~ F ! ! ! i
I ! i
I i
i ! i i i i i i' : Ji i J : ..... !ii ! J i i I i i ........................... i._i......=...L...~.....~....~.....~....L...1~ ~ ~ , ~ , ~ ~ , , ......................................................., ~ .....r...r....~.....,.......r....r....r...T....r..., i ~ ~ ,~ ~ ~ ~ , ~ :
~ i i !
i i ! i
~ i ! t
i ! ! i i i i i i i i i :: i i i i j !! !ii I'' I i i i i i ~
.!..!, ! i
!.....!.....i.....i.....
iiili!
rr LU
i i i 1
,
!
j i
~ j
J i i J J J J Ii t! iii t i
BUS .
INST ee UNIV
. . . . . .
i i i i i l ! i J l J j 60
............................!......i......',.....!.....~....., .....
iiiii
Z 40
i
i
i
i
i
i
!
!
!
!
IND
Jii ............................~----..i---.-f...+...i.....
.:. .. .. :.. 9
,
:
:
!
!
I
2 0 ...........................~----"'---:---'.....
121
.
.
.
.
.
.
.
.
1964
1970
1 9 7 6 1 9 8 2 1 9 8 8 1994 YEAR Fig. B-1. The number of SAW ultrasonic symposium proceedings papers with the work supported by industry, universities, government laboratories and institutes, and businesses.
3
199
Surface Acoustic Wave Technology Macrosuccess
I
~176 I iiii!iii!!i!iiiiiii!ili!!i!iiiii!
80 ...............................................................................................................................................................
!i!!!ilii'ii!''~!'l!iiiii o o
.....................
.....
INTL
ii
i...-i....i..,.i..,!.,,.i-..i.i--.i ........!.i .... i..i4q ........i..-.i.--..i. . .i. -.-i.---i.-.-i.-H-!4
.......................
w
i iii ii ii i iiiii ii !i ii 50 ............................ i,--i-.,--!.-+-i----.i! i-i J i~ ..i i.-i ,i .~---i--.F.-..-~--.i--.!---i----i---i-----i,.--i ~-,;- ..................
~
4o ......................i.-..i.,...i..,-i,-.-i.-...i .... i .....iii...i!.iii.14.-i
~-
ca
US-CAN
~o........................ i--i-i--i--~i'"iI i'i 1964
1969
1974
......144.. i.... i. i.i .................
ii' '~ ii ii ..............
1979 1984 YEAR
1989
1194
Fig. B-2. The number of SAW ultrasonic symposium proceedings papers from the United States, Canada, and countries outside North America.
those from the United States and Canada. The majority of international contributions have come from Western Europe and Japan (Fig. B-3). Contributions from Eastem Europe and Russia have been greater recently due to the availability of travel support. Contributions from China are growing. Figure B-4 displays the technical emphasis at the Ultrasonics Symposium by six category groups. Material and propagation properties of substrates and films has been a continuous and steady activity along with delay lines and bandpass filters, which represent about 50 percent of the papers presented. Although resonator and oscillators have been a much smaller but continuing activity, with major contributions between 1974 and 1990, they represent a major business area today. Passive signal processing devices such as correlators, convolvers, and chirp filters, and active acoustoelectronic and associated signal processing devices had strong support for development in the 1970s and early 1980s. Since 1985, SAW sensor activity has grown and is continuing at a high level. In the early 1970s, there were three major specialized conferences in the Western world dealing with SAW technology. The conferences were intended to characterize the present state of the technology, explore the limits of the technology, and to wrestle with such questions as where the technology
200
Fred S. HickerneU 90-
)i
ii
i
i i i ) ) ) ! ( i ! ) i i : : : . . . . . .
~ i ~ : ,
)
)iiiii
80-
i i i I i i ~.; ! l ! ) ! ! !
Z
"
"ii ) )
i )
i i ! . ,"~;
'
i
i
~
i i l)
I
i i i i l ( ~ ! i i
i
~
"
i
ASIA
i
~p~1-i~+~.~.-`~.-+~-b~4 .....
~
~
i)
i
~
!
I
i
i
i
i
"
l
" I
E-EUR
........................ l-.-.-fP ~.---!.---+..-.---I-.---+.--.~.-.ff -+.......... +-l.--i--,.. ~..-.-!,.--.i,-.--.i.-.--!-.---i--.-.l-----~-----i--.--+-,.-~-.. ----~.--~--..-~--.-
i !i i( !! ii i
(111i
60-
:E
i!
.......................+ ~ i ~ . ~ + ~ ; ; . ~ + ~ . F ~ F ~ F ~ . ~ ! ~
70-
IT LU m
)
!
ii
~i i ~ i i
i1
i!ii
11 i ) 1 1 1 ( ~ i
ii i ~i ii i ~ ) i ~) !
i
i i i i ) 9 : ~ ~. i i I I t ! ! i (i i i i ~.~.L.~.~4~.~-~*-.:...~:~.L~.~.~ i ! i i i ~ ! i ! t I t i i i i
i
! i i
! I
!
i~!!
( JAPAN
! i ! ! !!! i i i i!i i:i !t !
, ! ) ) I ( ) i I i ) i ( ) (
)
I
, ) ) ! ) ) ) ) )
) I i
! ! ! ! i i i i " i i i ! i i ) i I I I i j i : ' ' .......................... l.....[ ......... : : .......... , ,L...L....[ 9 : ........ , -I.....L..J......[ : , 9 l L....L.I..L....L...L....L..L..J,. . . . . . , . ......... , .L...:.....L...L..I....J.....) l t , / '
5040-
........................
! ii
ii
i
i
: i
~ i ~
) i
! "
i i
1
i~ ~ !i! ,i
I i {
t
I
t
!
I I t
i ~ : : i i I
~ !
!
i
i
~
i
! I i
i
.... 9
W-EUR
: t i
9
t
1-,, I i i ),-il I ) i..... 1"--'i'1111 ) ) )'......... (' i
30-
........................9- r ; . - ; . , - ,
.....
. . . . . . . . . . . . . . . . . . . . . . . . .
.
~'''iilii
20-
.
.
.
i
.
.
.
.
.
.
!!!!!! ilii11
10-
.......................i.i-.i~.!"-.|i.! ~i'!! 1
i
ii
1964
i! i
i
I
I
1969
i
i
1
I
i
1974
I
I
i
i
1979
i
i
i
i
i
1
1984
I
I
i
i
I
1989
I
I
I
i
1994
i
I
YEAR
Fig. B-3. The number of SAW ultrasonic symposium proceedings papers by geographic location in Western Europe, Japan, Eastern Europe and Russia, and Asia.
120]
. . . . . . . . . . . . i
100J
,oi,,
/
.
!
~ :
t
~ i
i :
i ), : ii !I i !I !I I
i
)
i
! iI ~! I ;i i
!
i
i
i
,
...........
i
i
!J,
{
I
i i ! ~! !
~ I
i!iiiiiiiiil,ii~itliiii:~ii~
;,i ' ~ i i i ) ) ; ! : ! !i
~',!illi',
'.:!',
i
~il~
ii
i
!
ii
i ~ i i ~ ! ; l ~ ) l ! i : l i : i i i l l ~
I
~i
i
i
~
)i t
iiiil ..... ii~ilili
~
t.,t
1965
1970
i
i!
i
~:~ i
~
iil'
''
1975
t
I
r
i;ii:
,
I '
1980
I
;
~
l~
i
i iI~
,,~
t ) , , ~
' i ,
1990
SENSOR
'
.i
',
RES-OSC DL-BPF
[
i'
ACOUSTEL SIG-PROC
'
1''
1985
)
l
iII~i i
!{
i I
I
....... J....... " . . . . .
i;i!
I:'
i
t
~ :
",
iii 1'i!
,,!
:
iiiii.._i
if!
lii
I i
ii
..................... L_.i_LL~i,~"'!i,!j
ii 2 0 ..................... l~-i-~'~
o
i
.................. ~"'mllmll'~"....... ~'11~ "'~''~''J''''L''~'L'''--'L''i--~''~P''--''~'''~'----''''Li--'~"
~!lli!
60
:,
)
MAT-PROP
i
1995
YEAR
Fig. B-4. The number of SAW ultrasonic symposium proceedings papers by technical categories of materials and propagation, delay lines and bandpass filters, resonators and oscillators, passive signal processing devices, active accoustoelectronic devices, and sensors.
3
Surface Acoustic Wave Technology Macrosuccess
201
should be applied. The first major conference in the United States was in March 1970 when IBM hosted a two day International Symposium on Acoustic Surface Waves at the Thomas J. Watson Research Center, Yorktown Heights, New York. Thirty-seven papers were presented in two-and-a-half days with representation from Canada, England, France, Germany, the Netherlands, Norway, and Switzerland. A total of 158 people attended the conference, with 14 attendees from outside the United States. From within the United States there were 24 industrial companies represented, 14 universities, and 10 government agencies and government-supported institutes. This was an indication of the interest in the technology, particularly by industry. There were no conference proceedings, but a two-page summary of the conference highlights was published in Ultrasonics [ 1 ]. In September of 1973, the Intemational Specialist Seminar on Component Performance and the Systems Applications of Surface Acoustic Waves was held in Aviemore, Scotland. There were 39 presentations, with representation from the United States and Western Europe. There were 18 papers from industry, 8 from universities, and 13 from government agencies and government-supported institutes. The majority of the papers recognized support from their respective governments. The proceedings of all papers presented were published [2] by The Institution of Electrical Engineers (IEE). In April 1974, The Polytechnic Institute of New York sponsored a Microwave Research Institute Symposium on Optical and Acoustical Micro-Electronics in New York City. Twenty-seven of the fifty papers presented related to SAW technology. In addition to Western authorship of papers, there were four papers authored by distinguished scientists from the USSR. The papers were published in a bound book proceedings by Polytechnic Press of the Polytechnic Institute of New York [3]. During the 1970s and 1980s, SAW papers werepresented at sessions in a number of professional society sponsored conferences throughout the world. In addition, there were university- and government-sponsored conferences. Special tutorials with tuition fees were sponsored by private training businesses throughout the United States. In-house industry seminars to acquaint systems engineers did much to encourage the use of SAW devices. In July 1985rathe hundredth anniversary of Lord Rayleigh's classic paper on surface w a v e s - - a conference was held at the Royal Institution in London, England, where Lord Rayleigh had worked and lectured. The Rank Prize Funds sponsored the three-day event, which drew approximately 100 participants from England, Western Europe, Japan, and the United States. This heralding of the start of the second Rayleigh wave century was accomplished
202
Fred S. Hickernell
by the presentation of 22 papers, which then were published by SpringerVerlag in their series on Wave Phenomena [4]. There was also a formal banquet held in Goldsmith Hall. In Western Europe, SAW papers had been presented at Microwave and Communication Conferences over the past 30 years and this practice continues today. In the former USSR, the major conference was the AllUnion Conference on Acoustoelectronics and Physical Acoustics of Solids. It was held every two years in different cities of the USSR starting in 1965. The last conference of this series took place in Leningrad in 1991. In 1994, the conference was opened to international participation and held on a cruise ship on the waterways between Moscow and St. Petersburg. It was held in conjunction with the third International Conference on Surface Waves in Solids and Layered Structures (ISWASS III), and proceedings were published [5]. The first ISWASS was held in Academic City (Academgorodok) outside Novosibirsk in July of 1986 and was the first to invite a group of Western scientists. There were approximately 300 in attendance at ISWASS I with a program of 25 invited papers, 40 contributed oral papers, and 125 poster papers. The papers were presented in Russian and English with simultaneous translation. The proceedings of this symposia were published in three volumes [6]. The second ISWASS was held in Bulgaria in September 1989, in conjunction with the biennial Bulgarian Acoustoelectronics Conferences. This was just prior to the dissolution of the Eastern Bloc countries. There was a total of 212 participants from 15 countries, with the most participants from Eastern Europe and the former Soviet Union. There were 210 papers (57 oral and 153 poster), with 23 invited. The oral presentations were published by World Scientific [7]. The Bulgarian Acoustoelectronics Conferences began in 1983 and were held every two years, with the last one in 1993. The conferences were organized by workers at the Institute of Solid State Physics of the Bulgarian Academy of Sciences and the Faculty of Physics, Sophia University, both in Sophia, Bulgaria. In 1987, the conferences were opened to international participation. Conferences and conference sessions have been held in Japan on acoustoelectronics and SAW developments for several years. At each Ultrasonics Electronics Symposium over the past 17 years, several SAW technology papers were presented. The papers were then published in the Japanese Journal of Applied Physics in the spring of the following year. In December 1992, The International Symposium on Surface Acoustic Wave Devices for Mobile Communication was held in Sendai, Japan. This specialized confer-
3
Surface Acoustic Wave Technology Macrosuccess
203
ence reflected the impact that SAW devices were making on mobile communications and viewed the future of the technology. There were 26 invited papers presented and a proceedings was published [8]. The symposium drew 120 participants, including 35 from outside Japan. The 1998 IEEE Ultrasonics Symposium will be held in Sendai, Japan, in October 1998. Throughout the history of SAW, including the present, special SAW sessions have been held in conjunction with acoustic, microwave, frequency control, and communication conferences throughout the world. The annual IEEE International Frequency Control Symposium, the annual IEEE International Microwave Symposium, and the annual Symposium on Ultrasonic Electronics in Japan have regularly had sessions on surface acoustic waves.
References 1. Dawes, M. I. (Ed.) (July 1970). Ultrasonics 9, 192-193. 2. Proc. International Specialist Seminar on Component Performance and the Systems Applications of Surface Acoustic Waves. 366 pages, Conference Publication 109. The Institution of Electrical Engineers, Herts, England, 1973. 3. Proc. Symposium on Optical and Acoustical Microelectronics, (Fox, Jerome, Eds), 646 pages.
Microwave Research Institute Symposia Series, XXIII, 1974. 4. Ash, E. A., and Paige, E. G. S. (Eds.) (1985). "Rayleigh-Wave Theory and Applications." Springer-Verlag, Heidelberg, Germany. 5. Proc. International Symposium on Surface Waves in Solids and Layered Structures and National Conference on Acoustoelectronics. 50 pages. St. Petersburg Academy of Aerospace Instrumenta-
tion (Russia) (1994). 6. Proc. I International Symposium on Surface Waves in Solids and Layered Structures "86. 960
pages. USSR Academy of Sciences, Siberian Branch, Novosibirsk, Russia, 1986. 7. Proc. II International Symposium on Surface Waves in Solids and Layered Structures and IV International Scientific Technical Conference Acoustoelectronics '89. 468 pages, World Scientific,
Singapore, 1990. 8. Proc. International Symposium on Surface Acoustic Wave Devices for Mobile Communication. 241
pages, Shibayama, K., and Yamanouchi, K. (Eds.). The 150th Committee in Japan Society for the Promotion of Science, Tokyo Press Co. Ltd., Tokyo, Japan, 1992.
Appendix C.
SAW Applications
The applications of SAW delay lines, bandpass filters, resonators, and oscillators and related modules are widespread use in the automotive, commercial, computer, consumer, environmental, industrial, residential, and telecommunication markets. SAWs are also still being used for military and space applications but to a much lesser degree. (It is estimated that less than 10 percent of the total SAW devices are being used for military and space purposes). SAW devices have been highly competitive with other approaches,
204
Fred S. Hickernell
particularly in the wireless market areas where stable, high-frequency sources and filters are in great demand. As microprocessor clock rates continue to grow, so will the need for high-frequency control modules. Growth will continue in the broad area of wireless remote data communications. In telecommunications, cellular-based phones, with base stations nearby or in satellites above the earth, will continue to proliferate through our communication-conscious, faster-paced world. Examples of the areas of application and future application opportunities for SAW filters, resonators, delay lines, matched filters, oscillators, transmitter/receiver modules, and signal processing subsystems include the following: 9 Telecommunications. Cellular, personal communication systems and networks, cordless, mobile, pagers, wireless local loops, base stations, global satellites. 9 Data Communications. Digital radio, wireless local area networks, handheld data terminals, global positioning systems, remote and airborne weather instrumentation. 9 Video Applications. Television, cable television, video cassette recorders, high-definition TV, direct broadcast satellite TV, interactive TV, cable modems, video games. 9 Automotive and Home. Remote keyless entry, theft deterrent systems, alarms, meter readings, home appliances, audio equipment. 9 Military and Space. Electronic warfare, signal intelligence, battlefield communications, avionics, fuzing, orbital and deep space satellite systems, surveillance systems, missiles. 9 Other Markets. Sensors, test equipment, identification tags, highfrequency clocks, computers, commercial avionics.
Appendix D.
Worldwide SAW Activities
Within the past ten years the SAW developments in particular countries and areas of the world have been the subject of publications, primarily from presentations at Ultrasonic Symposiums. References to these papers are given at the end of this appendix [ 1-7]. Major SAW manufacturers by country who actively produce products for the open market are shown in the list below. Many of the businesses produce filters in volume for the consumer and telecommunications market. Others produce custom SAW devices in smaller quantities. Some of these manufacturers have SAW businesses resulting
3
Surface Acoustic Wave Technology Macrosuccess
205
originally from the need to produce components for their own products. Other independent businesses were spin-offs from major industries or universities. In some cases the SAW business was an extension of a related business, such as quartz bulk wave resonator filters. There are also several smaller businesses, either separate or within larger companies, that produce SAW devices but have not been listed. Some government research institutes and industries have retained the capabilities to produce high-quality custom SAW devices. New smaller businesses with niche markets, such as in the sensor area, do not appear on the listing. However, although the list is incomplete, it is reasonably representative of present worldwide SAW activities.
SAW Component and Module Suppliers China. Bejing Chang Feng, Chongqing Xingtei, Jiangxi Jinghua, Nanjing Electronic Devices Inst., Zhe Jian De Qian Electronic Components. Europe. C-MAC, GEC-Plessey, Micronas, Racal, Siemens, SiemensMatsushita, Tele-Filter, Thompson Microsonics. Japan. Clarion, Epson, Fuji, Fujitsu, Hitachi, Horiku, JVC, Japan Radio, Japan Energy, Kinseki, Kyocera, Matsushita, Meidensha, Mitsubishi, MuRata, NDK, NEC, Nikko Kyodo, OKI, Sanyo, Seiko, Soshin, Taiyo Yuden, Toko, Toyocom, Tokyo Denpa, Toshiba. Korea. Daewoo, Samsung, LG Electro-Components Ltd. North America. Andersen Laboratories, ComDev, Motorola, Nortel, Phonon, Raytheon, RF Monolithics, Sawtek, Texas Instruments, TRW, Vectron, Zenith. Russia. Avangard Elionica, Fonon, Institute of Radioengineering and Electronics, Institute of Semiconductor Physics, Moscow Radiocommunication Research Institute, ONIIP, Tomsk Academy of Control Systems and Radioelectronics.
References 1. Ebata, Y., and Satoh, H. (1988). Current applications and future trends for SAW in Asia. Proc. 1988 IEEE Ultrason. Symp., 195-202. 2. Gautier, H., and Maerfeld, C. (1988). Current applications and trends of SAW components in France. Proc. 1988 IEEE Ultrason. Syrup., 67-75. 3. Sharif, M. A. (1991). Development and capabilities of SAW technology in Brazil. Proc. 1991 [EEE Ultrason. Symp., 189-192.
206
Fred S. Hickernell
4. Yakovkin, I. B. (1991). Investigation of fundamentals, current status, and trends of SAW in Siberia. Proc. 1991 IEEE Ultrason. Syrup., 1375-1386. 5. Buff, W. (1991). SAW technology in Eastern Europe. Proc. 1991 IEEE Ultrason. Syrup., 107-110. 6. Heighway, J., and Morgan, D. P. (1991). 1991 and beyond: the current status and future trends for SAW in Europe. Proc. 1991 IEEE Ultrason. Syrup., 217-223. 7. Spassov, L. (1992). Acoustoelectronics in Bulgaria. Proc. 1992 IEEE Ultrason. Syrup., 179-188.
Appendix E.
The SAW Technologist as an Artisan
The myriad of patterns that have been imprinted on the smooth surface of a piezoelectric plate to create a SAW device brings to mind the work of an artist--a painter, and a sculptor, and a musician. SAW technology is about the romance and creative instincts of engineers and scientists presented with a polished piezoelectric p l a t e - - a solid canvas on which to produce shimmering geometric patterns and shadowy impressions. The artisan patterns in the spatial-time domain and his images spring to life with the application of oscillating electrical energy, which sends microacoustic waves skittering across the surface. As painter, the artisan's palette contains metal and dielectric films that are applied to the surface atom by atom. As sculptor, the artisan chisels impressions to precise depths in the surface, deftly removing material atom by atom. Patterns emerge as material is applied and removed to form the finished work of art. With the patterning complete, the artisan launches, propagates, and detects the microacoustic waves, fashioning their phase and amplitude, focusing and expanding them, directing their flow and storing them. As musician, the artisan creates the microseismic rhythms that will play in the frequency and time domain of the electronic world. At first, the fledgling artisans explored the frequency and time regimes from microseconds to gigahertz with simple in-line rectangular geometrical patterns. Then came the means of changing wave direction as waves went back and forth and across the plate and even around to the back side to extend wave lifetime. Then, with an understanding of collective reflections, the surface waves were made to dance in place. The waves were guided and focused with films and geometric shapes. Signals were easily transformed from the frequency to the time domain by the patterning on the piezoelectric plate. By producing finer and finer patterns, the microseismic pulsations reached into the billions of times per second. As the requirements for new and improved microseismic rhythms grew, new piezoelectric materials were sought and new ways of cutting the solid canvas from the gemlike piezoelectric crystals were identified. From this new freedom of choice, the artisan
3
Surface Acoustic Wave Technology Macrosuccess
207
gained better control over velocity, attenuation, coupling factor, temperature, and stress stability of the microacoustic waves. Not content to restrict the waves to a piezoelectric plate alone, the artisans applied piezoelectric thin films to other types of solid canvases to bring wave motion to the canvas surfaces. Interactions of the microacoustic waves with each other on piezoelectric plates and in the presence of semiconductors provided unique signal storage and matched filter operations. By adding vapor- and fluid-sensitive films on the piezoelectric canvas, the traveling surface waves felt, sniffed, and tasted specific substances and identified their presence. When the SAW artisan completes a masterpiece, a suitable protective frame is found and the artwork is mounted in a gallery amidst electronic components and interconnections. The interconnection with its neighboring electronic components breathes life into the artwork, and its rhythmic expansions and contractions assume control over the time and frequency functions within the gallery. The artisan takes pride in the resulting work of art, but then turns and eagerly seeks out other polished piezoelectric plate canvases on which to create new and more beautiful patterns.
This Page Intentionally Left Blank
Frequency Control Devices JOHN R . VIG and ARTHUR BALLATO U S. Army Communications -Electronics Command. Fort Monmouth. NJ 07703 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209 Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .210 A . Communication Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212 B. Navigation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .216 C . Surveillance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218 D. Identification-Friend-or-Foe (IFF) Systems . . . . . . . . . . . . . . . . . . . . . . 219 E . Electronic Warfare . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220 F. Missile Guidance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .221 G . Battery Consumption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .221 H . Survivability under Radiation and High Acceleration . . . . . . . . . . . . . . . . 222 I . Logistics Costs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222 111. Frequency Control Device Fundamentals . . . . . . . . . . . . . . . . . . . . . . . . . 222 A . Crystal Oscillators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223 B. Oscillator Categories . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236 C . Oscillator Circuit Types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 238 D. Oscillator Instabilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239 E . Oscillator Comparison and Selection . . . . . . . . . . . . . . . . . . . . . . . . .261 F. Failure Mechanisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 266 G . Specifications, Standards, Terms and Definitions . . . . . . . . . . . . . . . . . . 266 IV Related Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .267 A . Crystal Filters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .267 B . Sensors and Transducers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .268 V ForFurtherReading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 269 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 269 I. I1.
I
.
Introduction
Frequency control devices provide the precise time and frequency on which modem electronics depends. A vibrating quartz crystal. i.e., a quartz resonator. is the "heart" of nearly all frequency control devices . Quartz clocks provide accurate time and quartz oscillators are the sources of precise frequency. The hdamental roles these devices play in the modem world can be seen by considering what would happen if all the quartz crystals in the world suddenly stopped vibrating . All modem communication systems (telephones.
PHYSlCAL ACOUSTICS. VOL. XXlV
ISBN 0-12-477945-X $30.00
210
John R. Vig and Arthur Ballato
radios, TV stations, air traffic control systems, etc.) would stop functioning, all but the oldest transportation systems (automobiles, trucks, airplanes) would cease operating, and all computers would stop. The consequences would be catastrophic. Time is important not only for the daily schedules of human beings, but also, for example, for determining the sequence of events that take place inside computers, and for time-tagging the information that flows through communication systems. Frequency sources are essential for determining the frequencies of radio and TV transmissions, radar systems, communication and navigation systems, etc. Frequency control technology took a great leap forward in the 1920s when quartz was first utilized to realize crystal resonators for the stabilization of oscillators, thereby launching the field of modem frequency control. With the introduction of quartz control, timekeeping moved from the sun and stars to small, man-made sources that exceeded astronomy-based references in stability. Since then, the applications of devices based on quartz have expanded dramatically. The quartz resonator has continued to evolve to become a device capable of precision one million times greater than the original. It also serves as the "flywheel" in atomic frequency standards. Atomic standards make frequency the most accurate entity known. Of the man-grown single crystals, quartz is second to silicon in quantity grown. About 2500 to 3000 tons of quartz crystals are grown per year (about three to four times as much silicon is grown). The major applications are oscillators for clocks and frequency sources. Other important applications are sensors, and filters used for frequency selection. About 2 • 109 bulk acoustic wave (BAW) quartz resonators, and several hundred million quartz surface acoustive wave (SAW) devices are manufactured annually.
II.
Applications
The major applications of quartz crystals are shown in Table 1. The applications in the fourth and fifth columns comprise most of the annual production, but the applications in the first and second columns are the most demanding. The military applications have been, and continue to be, the "drivers" of the technology. Civilian applications usually follow at a later date. For example, the military developed spread-spectrum techniques for jamming resistance and communication security. Civilian applications of spreadspectrum technology followed, such as the cellular telephone systems in
Military and Aerospace Communications Navigation IFF Radar Sensors Guidance systems Fuzes Electronic warfare Sonobuoys
Research and Metrology
Industrial
Consumer
Automotive
Atomic clocks Instruments Astronomy and geodesy Space tracking Celestial navigation
Communications Telecommunications Mobileicellularlportable radio, telephone and pager Aviation Marine Navigation Instrumentation Computers Digital systems CRT displays Disk drives Modems Taggingiidentification Utilities
Watches and clocks Cellular and cordless phones, pagers Radio and hi-fi equipment Color TV Cable TV systems Home computers VCR and video camera CB and amateur radio Toys and games Pacemakers
Engine control, stereo, clock Trip computer
PZ
o'
2
John R. Vig and Arthur Ballato
212
which spread-spectrum techniques are used for maximizing the number of simultaneous users in the assigned frequency band. In the United States, the genesis of the quartz crystal industry can be traced to the decision in 1939 to make large-scale use of crystal control in military communication systems [ 1]. In early military systems, controlling the carrier frequency of radio communications systems was the primary application. The typical (normalized frequency) accuracy in World War II systems was 200ppm [2]. In systems of the 1960s and 1970s, the typical accuracy requirements ranged from 40 to 0.5 ppm [3]. In systems that are currently in production or development, the accuracy requirements range from 5 ppm in some tactical radios to parts in 1012 in some navigation, electronic warfare, and strategic communication systems. In addition to possessing high accuracy, the frequency sources of modem systems must also exhibit low noise characteristics and must remain stable in extreme environments. In the following section, the most demanding applications are reviewed. Many of these are military applications; some of these have parallels in the civilian world. In early military systems, controlling the carrier frequency for improved spectrum utilization was the principal driver of frequency control technology. In modem systems, the major drivers are spread-spectrum systems that require ever-higher clock accuracies, surveillance systems that require lownoise oscillators in the presence of platform vibrations, and tactical (handheld) systems that require ever-higher accuracies with the lowest possible battery consumption and in the smallest possible volume.
A.
COMMUNICATIONSYSTEMS
In communication systems, the accuracy and stability of oscillators and clocks affect important system performance parameters, such as the spectrum utilization, resistance to unintentional and intentional (i.e., jamming) interference, signal acquisition speed, autonomy period, and bit error rates.
1.
Spectrum Utilization
In the field of communications electronics, the subject of frequency control is intimately related to the subject of frequency spectrum utilization [4]. Historically, in both commercial and military systems, to allow for more users in a given frequency band, it was necessary to reduce the channel spacings, which required the tightening of the frequency tolerances allowed in
4
Frequency Control Devices
213
both the transmitters and receivers. As the number of users grew, and as technology allowed the allocation of higher frequency bands, the frequency tolerances became tighter and tighter. The same channel spacing at a higher frequency, of course, requires tighter frequency tolerance. The frequency accuracy requirements of tactical radios prior to the advent of spreadspectrum techniques were typically 10 to 50ppm. Radios that employ spread-spectrum techniques typically require 5- to 0.001-ppm oscillators. In digital communication systems, not only must the oscillators possess high accuracy, they must also have low noise characteristics, for reasons that are discussed below. The noise of oscillators can also limit the capacity of communication systems. Since the noise from a transmitter in one channel extends to neighboring channels, as the number of transmitters grows, the noise accumulates to the point where receivers can no longer function properly. For example, in one L-band satellite communication system, the vibrationinduced phase noise [5, 6] is a serious limitation on the number of users per transponder when the users are on vibrating platforms, such as aircraft, trucks, etc. The noise from a typical commercial oscillator (2 x 10 -9 per g vibration sensitivity) limits the number of users to less than 100 per transponder, whereas a state-of-the-art oscillator (2 x 10-1~ per g vibration sensitivity) can allow as many as 1200 users per transponder. Since the rental of a transponder costs more than $1 million per year, the economic impact of oscillator noise can be significant [7]. 2.
Resistance to Jamming
Spread-spectrum techniques are used in military systems primarily for rejecting intentional and unintentional jamming, and for communication security [8, 9]. In spread-spectrum systems, the transmitters and receivers contain clocks that must be synchronized. For example, frequency hopping is a spread-spectrum technique used in several evolving military communication systems. In such systems, the transmitters and receivers must hop to the same frequency at the same time. The faster the hopping rate, the higher the jamming resistance, and the more accurate the clocks must be. For example, for a system with a hopping rate of 1000 hops per second, the dwell time at each frequencyis 1 millisecond. For such a system to operate properly, the clocks must remain synchronized to about 100 microseconds. When several radio nets operate in an area, self-jamming can be a problem if the nets operate independently of one another, i.e., if the nets are not
214
John R. Vig and Arthur Ballato
orthogonal. Radios of neighboring nets can then occasionally hop to the same frequency at the same time, thus producing self-jamming. When the nets are orthogonal, i.e., when the neighboring nets are synchronized and use codes that ensure that radios do not hop to the same frequency at the same time, the radios must not only be synchronized within a net but also to those of neighboring nets. This requires an even higher clock accuracy. With the availability of fast spectrum analyzers and synthesizers, it is possible to jam frequency hopping systems [9, 10]. If the jammer is fast enough, it can detect the frequency of transmission and tune the jammer to that frequency well before the radio hops to the next frequency. However, with a good enough clock, it is possible to defeat such follower jamming. As is illustrated in Fig. 1, even a "perfect" follower jammer can be defeated if a good enough clock is available to allow a very fast hop rate. (A perfect jammer is defined as one that can identify the frequency of a received signal, tune a synthesizer to that frequency, and transmit the jamming signal in zero time.) Because radio waves travel at the speed of light, the radio-to-jammerto-radio (R1 to J to R2) and radio-to-radio (R1 to R2) propagation delays are 3.3 kts per km. Therefore, if the hop-rate is fast enough for the propagation delay difference to be greater than 1/hop-rate, i.e., if the radios can hop to the
I tl
Examvle _
Let R1 to R2 = 1 km, R1 to J = 5 km, and J to R2 = 5 km. Then, since propagation delay = 3.3 ~ts/km, t ! = t 2 = 16.5 las, t R = 3.3 laS, and t m < 30 ~ts.
t2
Radio
[
R1
I"
,, . tR
~.
[ Radio
"-
I
R2
To defeat a "perfect" follower jammer, need a hop-rate given by" t m < (t 1 + t 2)- tR, w h e r e t m -- msg. d u r a t i o n / h o p = 1/hop-rate FIG. 1.
Allowed clock error -- 0.2 t m = 6 l,ts. For a 4 hour resynch interval, clock a c c u r a c y r e q u i r e m e n t is:
4 x 10 -1~
Clock required for a jamming-proof very fast frequency hopping radio.
4 Frequency Control Devices
215
next frequency before the jamming signal reaches the receiver, then the radios are jamming-proof (for follower jammers). In the example of Fig. 1, the propagation delays tl, t2, and tR imply that the message duration tm must be less than 30 gs. Since the clock accuracies required by frequency hopping systems are usually 10 to 20% of tin, the allowed clock error is about 6 gs. In a military environment, such accuracies can be maintained for periods of hours and longer only with atomic clocks. The requirement for C 3 systems to be interoperable places yet another stringent requirement on accuracy. For example, when an Army unit calls for air support from an Air Force unit that may be many hundreds of kilometers away, the clocks in the respective units' radios must be synchronized for the units to be able to communicate. Maintaining synchronization for extended periods among independent clocks that are widely separated requires very high quality clocks.
3.
Signal Acquisition Speed
The speed with which a communication link can be established depends on the speed with which a transmitter's signal can be acquired, which is strongly dependent on the frequency difference between transmitter and receiver [ 11 ]. In spread-spectrum systems, it is also dependent on the time difference between the transmitter's clock and the receiver's clock. The larger these differences, the longer it takes to search and acquire. While searching, the system is more vulnerable to interception and jamming than at other times. For acquiring weak signals, the noise of the receiver's reference oscillator can also affect the acquisition. For example, in a tactical radio system, the time it takes for a radio to enter a net depends on the radio's frequency and time errors. Similarly, in a satellite communication system, the time it takes for a terminal to acquire the satellite depends on the terminal's frequency and time errors. (Range errors also affect the relative time between transmitter and receiver.) In a navigation system, such as the Global Positioning System (GPS), the time to first fix is strongly dependent on the receiver's frequency error [12]. Minimizing acquisition time is especially important in submarine and special operations forces electronic systems, because avoiding detection is of paramount importance in such systems. In secure communication systems, precise synchronization and low noise is necessary to be able to recover encrypted data.
John R. Vig and Arthur Ballato
216
4.
Autonomy Period
Autonomy period, also called "radio silence interval," is important in modem warfare. For example, to remain undetected, submarine and special operations forces must, at times, refrain from communicating over the air for extended periods. When clocks are not resynchronized and resyntonized (i.e., refrequency calibrated), time and frequency errors increase with increasing mission duration. The better the long-term stability of the systems' oscillators, the longer can be the allowable autonomy period, and the shorter will be the subsequent acquisition time.
5.
Digital Communications
Digital communication systems, whether commercial or military, must be synchronized and have the same data rates. Synchronization plays a critical role in such systems because it ensures that information transfer is performed with minimal buffer overflow or underflow events, i.e., with an acceptable level of "slips." Slips cause problems, e.g., missing lines in FAX transmission, clicks in voice transmission, loss of encryption key in secure voice transmission, and data retransmission [ 13, 14]. The phase noise of oscillators can lead to erroneous detection of phase transitions, i.e., to bit errors, when phase-shift-keyed (PSK) digital modulation is used. In digital communications, for example, where 8-ary PSK is used, the maximum phase tolerance is +22.5 ~ of which -1-7.5~ is the typical allowable carrier noise contribution [ 14]. Due to the statistical nature of phase deviations, if the RMS phase deviation is 1.5 ~ for example, the probability of exceeding the -t-7.5~ phase deviation is 6 • 10 -7, which can result in a bit error rate that is significant in some applications. Shock and vibration can produce large phase deviations even in "lownoise" oscillators [5, 6]. Moreover, when the frequency of an oscillator is multiplied by N, the phase deviations are also multiplied by N. For example, a phase deviation of 10 -3 radian at 10 MHz becomes 1 radian at 10 GHz. Such large phase excursions can be catastrophic to the performance of systems, e.g., those that rely on phase locked loops (PLL) or phase sift keying. Lownoise, acceleration-insensitive oscillators are essential in such applications. B.
NAVIGATION
Precise time is essential to precise navigation. Historically, navigation has been a principal motivator in man's search for better clocks. Even in ancient
4
Frequency Control Devices
217
times, one could measure latitude by observing the stars' positions, but determining longitude is a problem of timing. Since the earth makes one revolution in 24 hours, one can determine longitude from the time difference, At, between local time (which was determined from the sun's position) and the time at the Greenwich meridian (which was determined by a clock): longitude in degrees = (360~ hours)• At in hours. Today's military (and civilian) navigation systems require ever-greater accuracies. Modem navigation systems utilize ultraprecise clocks and radio transmissions of precisely timed navigation signals. The Global Positioning System (GPS), developed by the U.S. Department of Defense, is the most accurate worldwide navigation system available. As the price of GPS receivers has declined, the number of civilian applications of GPS has increased. Today, the number of GPS receivers sold for civilian applications far exceeds the military sales. In GPS, navigation is accomplished by one-way time measurements [15-18]. Since electromagnetic waves travel 0.3 meters (i.e., a foot) per nanosecond,, if, for example, a vessel's timing was in error by a microsecond, a navigational error of 300 meters would result. In GPS, atomic clocks in the satellites and quartz oscillators in the receivers provide nanosecond-level accuracies. The resulting (worldwide) navigational accuracies are about 10 meters and differential accuracies can be as good as a few millimeters. Each GPS spacecraft contains four high-performance atomic clocks. (Only one is tumed on; the others are for backup.) The military GPS receivers contain oven-controlled crystal oscillators (OCXO); the less expensive commercial receivers contain less expensive temperature-compensated crystal oscillators (TCXO). The spacecraft clocks provide a time accuracy of better than 100 nanoseconds [17, 18]. The oscillator is a key component in the receiver [ 12]. In military receivers, especially, the oscillator's performance has a direct influence on system performance. The noise of the oscillator affects the navigation accuracy and the performance (i.e., jamming margin) in a highjamming environment; the medium-term (10 to 1000 second) stability affects the reacquisition capability, system integrity monitoring, and performance in a high-jamming environment; the long-term stability affects the time to subsequent fix and the capability to operate with less than four satellites; the warmup time of the oscillator affects the time to first fix; and the power requirement and the size of the oscillator affect the receiver's battery life, mission duration, and weight.
218
C.
John R. Vig and Arthur Ballato
SURVEILLANCE
In surveillance, Doppler radars especially require low-noise oscillators [19, 20]. The velocity of the target and the radar frequency are primary determinants of the phase noise requirements. Slow-moving targets produce small Doppler shifts, therefore, low phase noise close to the carrier is required. To detect fast-moving targets, low noise far from the carrier is required. For example, when using an X-band radar to detect a 4 km/hour target (e.g., a slowly moving vehicle), the noise 70 Hz from the carrier is the important parameter, whereas to detect supersonic aircraft, the noise beyond 10 kHz is important. When a radar is on a stationary platform, the phase noise requirements can usually be met with commercially available oscillators. A good quartz crystal (bulk acoustic wave, BAW) oscillator can provide sufficiently low noise close to the cartier, and a good surface acoustic wave (SAW) oscillator can provide sufficiently low noise far from the carrier [21]. Very far from the carrier, dielectric resonator oscillators (DRO) can provide lower noise than either BAW or SAW oscillators. A combination of oscillators can be used to achieve good performance in multiple spectral regions [22], e.g., a DRO phase locked to a frequency-multiplied BAW oscillator can provide low noise both close to the carrier and far from the carrier. The problem with achieving sufficiently low phase noise occurs when the radar platform vibrates, as is the case when the platform is an aircraft or a missile [5, 6]. The vibration applies time-dependent stresses to the resonator in the oscillator, which results in modulation of the output frequency. The aircraft's random vibration thereby degrades the phase noise, and discrete frequency vibrations (e.g., due to helicopter blade rotation) produce spectral lines that can result in false target indications. The degradation in noise spectrum occurs in all types of oscillators (BAW, SAW, DRO, atomic frequency standards, etc.) Figure 2 shows an example of a typical aircraft random vibration envelope (in the upper right-hand comer) and the resulting phase noise degradation [5, 6]. Such a large degradation can have catastrophic effects on radar performance. In a coherent radar, the platform-vibrationinduced phase noise can reduce the probability of detection to zero. Most air defense systems cannot cope with stealth aircraft. To detect stealthy targets, the radar systems must compensate for the smaller reflections by significantly increasing the transmitted power, which is often not feasible, or by significantly improving the radar receiver's sensitivity. Higher sensitivity results in receiving more clutter and false targets. Very low noise reference oscillators are required for detecting targets under such conditions.
4
Frequency Control Devices
219
-o.
.87
+
-90 ' I
300
~-110
frequency (Hz)
-140
II
1 K 2K
J"~
-150
!
-IGO I
I
l
5 380 IK Offset Frequency (Hz)
I
2K
FIG. 2. A typical aircraft random vibration envelope (in the upper right-hand comer), and the resulting phase noise degradation. Conventional (i.e., "monostatic") radar, in which the illuminator and receiver are on the same platform, is vulnerable to a variety of countermeasures. Bistatic radar [23], in which the illuminator and receiver are widely separated, can greatly reduce the vulnerability to countermeasures such as jamming and antiradiation weapons. The transmitter can remain far from the battle area, in a "sanctuary." The receiver can remain "quiet." The timing and phase coherence problems can be orders of magnitude more severe in bistatic than in monostatic radar, especially when the platforms are moving. The two reference oscillators must remain synchronized and syntonized during a mission so that the receiver knows when the transmitter emits each pulse, and so that the phase variations will be small enough to allow a satisfactory image to be formed. Low-noise crystal oscillators are required for short-term stability; atomic frequency standards are often required for longterm stability. (A combination of a low-noise crystal oscillator and an atomic standard can provide both low noise and good long-term stability.)
D.
IDENTIFICATION-FRIEND-OR-FOE(IFF) SYSTEMS
In modem warfare, when the sky is filled with friendly and enemy aircraft, and a variety of advanced weapons are ready to fire from both ground and airborne platforms, reliable identification of friend and foe is critically
John R. Vig and Arthur Ballato
220
important. Friendly-fire casualties due to lack of adequate IFF systems have been a major problem in recent wars [24]. Precise timing can play a major role in solving this problem. For example, cooperative IFF systems use an interrogation/response method that employs cryptographically encoded spread-spectrum signals. The interrogation signal received by a friend is supposed to result in the "correct" code being automatically sent back via a transponder on the friendly platform. The "correct" code must be changed frequently to prevent a foe from recording and transmitting that code ("repeat jamming"), thereby appearing to be a friend. The code is changed at the end of what is called the code validity interval, (CV). The better the clock accuracy, the shorter can be the CVI, the more resistant the system can be to repeat jamming, and the longer can be the autonomy period for users who cannot resynchronize their clocks during a mission. The CVI chosen is usually dictated by the accuracies achievable with low-power oscillators. E.
ELECTRONICWARFARE
The ability to locate radio emitters is important in modem warfare. One method of locating emitters is to measure the time difference of amval of the same signal at widely separated locations. Emitter location by means of this method depends on the availability of highly accurate clocks, and on highly accurate methods of synchronizing clocks that are widely separated. Since electromagnetic waves travel at the speed of light (30 cm per nanosecond), the clocks of emitter-locating systems must be kept synchronized to within nanoseconds to locate emitters with high accuracy. (Multipath and the geometrical arrangement of emitter locators usually result in a dilution of precision.) Without resynchronization, even the best available militarized atomic clocks can maintain such accuracies for periods of only a few hours. With the availability of GPS and using the "GPS common view" method of time transfer, widely separated clocks can be synchronized to better than 10 ns [17]. An even more accurate method of synchronization is "two-way time transfer via communication satellites," which, by means of very small aperture terminals (VSATs) and pseudonoise modems, can attain subnanosecond time transfer accuracies [25]. An important application for frequency sources is the ELINT (ELectronic INTelligence) receiver. These receivers are used to search a broad range of frequencies for signals that may be emitted by a potential adversary. The frequency source must be as noise-flee as possible so as not to obscure weak
4
Frequency Control Devices
221
incoming signals. The frequency source must also be extremely stable and accurate to allow accurate measurement of the incoming signal's characteristics. E
MISSILEGUIDANCE
When a missile is guided by ground radar, the radar is vulnerable to antiradiation missiles and other countermeasures. Placing the radar onboard the missile can greatly reduce the vulnerability, but at the expense of placing much greater demands on missile components, especially the reference oscillator. As previously discussed, the missile's high vibration levels degrade the oscillator phase noise by a wide margin. Vibration-insensitive low-noise oscillators are required for on-board radar systems. Such systems could benefit greatly from improvements in vibration-resistant low-noise oscillators [5]. G.
BATTERYCONSUMPTION
Tactical military electronic systems are usually powered by batteries. In many of these systems, precise timing plays an essential role. When the system is not being used, everything except the clock can usually be turned off. As a result, the power requirement of the clock is a major determinant of battery consumption. For example, to power the time and frequency unit in one tactical satellite terminal during the required 10-day standby period, a battery pack weighing 18 kg was required. By replacing the power-hungry, ovencontrolled crystal oscillator of the original design with a Microcomputer Compensated Crystal Oscillator (MCXO) [26], which is a much lower power oscillator of similar accuracy, 12 kg of battery weight could be saved. The cost savings resulting from reducing the power requirements of oscillators can be large. For example, a calculation estimates that for one model of tactical radio, the cost savings resulting from reducing the power requirement for the radio's 20-year life, (assuming peace-time usage of 2 hours per day) is $48,000 per milliwatt per 10,000 radios [27]. Since more than 200,000 of these radios may eventually be produced, the eventual cost savings when this radio is fully fielded may be more than $1 million per mW of power saving! During a conflict, the radios would, of course, be used 24 hours per day, for an even greater saving; however, the real benefit of lower power radios is a lighter, more mobile force, which can operate for longer periods without needing battery replenishment.
222
John R. Fig and Arthur Ballato
Another reason for minimizing the power dissipation of oscillators is that the dissipation often produces undesirably large infrared signatures, which makes the system easier to detect by an adversary. H.
SURVIVABILITYUNDER RADIATION AND HIGH ACCELERATION
Survivability under ionizing radiation and high shock and vibration conditions is primarily a military (and space) requirement. Gun-hardened oscillators are required, for example, for smart munitions, air-dropped and artilleryemplaced sensors, fuzes, and space defense systems. The highly shockresistant oscillators have been developed that can withstand the shock of being launched from a howitzer [28, 29]. Radiation-hardening of oscillators used to be a major issue in many military systems because a high-intensity pulse of nuclear radiation stops clocks and causes large temporary, and smaller permanent, frequency offsets in frequency standards [30]. With the threat of nuclear war receding, radiation-hardening is less of an issue today for military systems (although it remains an important issue for space systems). I.
LOGISTICSCOSTS
The long-term stability and the lifetime of oscillators often have a significant impact on logistics costs. As the oscillator's frequency ages (and all but cesium beam frequency standards do age), or as, e.g., a cesium beam frequency standard nears end of life, at some point, the oscillators must be recalibrated or replaced. A need for frequent recalibration or replacement has a significant adverse impact on the life cycle cost of equipment. Lower aging oscillators do cost more initially; however, the increased cost is often recovered rapidly through a decrease in logistics costs. An important goal of research aimed at reducing oscillator aging is to provide systems with calibration-flee life. III.
Frequency Control Device Fundamentals
The fundamentals of quartz oscillators are reviewed in this section, with emphasis on quartz frequency standards (as opposed to inexpensive clock oscillators). The subjects discussed include: crystal resonators and oscillators, oscillator types, and the characteristics and limitations of temperaturecompensated crystal oscillators (TCXO) and oven-controlled crystal oscilla-
4
223
Frequency Control Devices
tors (OCXO). The oscillator instabilities discussed include: aging, noise, frequency vs temperature, warm-up, acceleration effects, magnetic-field effects, radiation effects, and atmospheric pressure effects. Interactions among the various effects are also considered. Guidelines are provided for oscillator comparison and selection. Discussions of specifications are also included, as are references and suggestions for further reading.
A.
CRYSTAL OSCILLATORS
1. Oscillator Basics
Figure 3 is a greatly simplified circuit diagram that shows the basic elements of a crystal oscillator [31-33]. The amplifier of a crystal oscillator consists of at least one active device and the necessary biasing networks; it may also include other elements for band limiting, impedance matching, and gain control. The feedback network consists of the crystal resonator and may contain other elements, such as a variable capacitor for tuning. The frequency of oscillation is determined by the requirement that the closed-loop phase shift equal 2nn, where n is an integer, usually 0 or 1. When the oscillator is initially energized, the only signal in the circuit is noise. That component of noise, the frequency of which satisfies the phase condition for oscillation, is propagated around the loop with increasing amplitude. The rate of increase depends on the excess loop gain and on the bandwidth of the crystal network. The amplitude continues to increase until the amplifier gain
Tuning Voltage
IDI-
....
L__I
Crystal
Resonator
Output Frequency Amplifier FIG. 3.
Crystal oscillator m simplified circuit diagram.
John R. Pig and Arthur BaUato
224
is reduced, either by the nonlinearities of the active elements (in which case it is self-limiting) or by an external level-control method. At steady state, the closed-loop gain is 1. If a phase perturbation A~b occurs, the frequency of oscillator must shift by a Af to maintain the 2nzc phase condition. It can be shown that for a series-resonance oscillator
af f
2QL'
where QL is the loaded Q of the crystal in the network [31 ]. ("Crystal" and "resonator" are often used interchangeably with "crystal unit," although "crystal unit" is the official name. See references 3 to 6 for further information about crystal units.) Crystal oscillator design information can be found in references 31, 32, 35, and 36. The abbreviation for crystal oscillator is XO.
2.
Crystal Unit Equivalent Circuit
A quartz crystal unit is a quartz wafer to which electrodes have been applied, and which is hermetically sealed in a holder structure. (The wafer is often referred to as the "blank," or the "crystal plate.") Although the design and fabrication of crystal units comprise a complex subject, the oscillator designer can treat the crystal unit as a circuit component and just deal with the crystal unit's equivalent circuit. The mechanically vibrating system and the circuit shown in Fig. 4 are "equivalent," because each can be described by the same differential equation Z~ ,)
Spring
C L
Mass r- -I
R
Dashpot . n
FIG. 4. Equivalentcircuit of a mechanicallyvibrating system.
4 Frequency Control Devices
225
[37]. The mass, spring, and damping element (i.e., the dashpot) correspond to the inductor, capacitor, and resistor. The driving force corresponds to the voltage, the displacement of the mass to the charge on the capacitor, and the velocity to the current. A crystal resonator is a mechanically vibrating system that is linked, via the piezoelectric effect, to the electrical world. Figure 5 shows a (simplified) equivalent circuit (of one mode of vibration) of a resonator, together with the circuit symbol for a crystal unit. A load capacitor CL is shown in series with the crystal. Co, called the "shunt" capacitance, is the capacitance due to the electrodes on the crystal plate plus the stray capacitances due to the crystal enclosure. The R1, L1, C1 portion of the circuit is the "motional arm," which arises from the mechanical vibrations of the crystal. The Co to C1 ratio is a measure of the interconversion between electrical and mechanical energy stored in the crystal, i.e., of the piezoelectric coupling factor, k. Co~C1 increases with the square of the overtone number; the relationship of C o / C 1 to k and N is 2Co/C1 = [rcN/2k] 2, where N is the overtone number. When a dc voltage is applied to the electrodes of a resonator, the capacitance ratio Co/C~ is also a measure of the ratio of electrical energy stored in the capacitor formed by the electrodes to the energy stored elastically in the crystal due to the lattice strains produced by the piezoelectric effect. Figure 6 shows the reactance versus frequency characteristic of the crystal unit. The Co~C1 is also inversely proportional to the
CL
Symbol for crystal unit
Co I(_
CL C1 ~f fs _ _
/,v
FIG. 5.
L1
C1 2(Co + CL)
R1 ~ 1. Voltage control (VCXO) 1
1 2. Temperature compensation (TCXO)
Equivalent circuit of crystal unit with load capacitor.
John R. Vig and Arthur Ballato
226
f
Antiresonance
]
Frequency
Area of Usual "Parallel
Resonance"~ 9
"~ cD
O
series / /
Resonance i
27tfCo FIG. 6.
R e a c t a n c e versus frequency o f a crystal unit.
antiresonance-resonance frequency separation (i.e., the pole-zero spacing), which is an especially important parameter in filter applications. The slope of the reactance vs frequency curve near ~ is inversely proportional to C1, i.e., A X / ( A f / f ) ~ l / T c f C l near ~., where X is the reactance. C1 is, therefore, a measure of the crystal's "stiffness," i.e., its tunabilitymsee the next equation. When the load capacitor is connected in series with the crystal, the frequency of operation of the oscillator is increased by a Af' , where Af' is given by
Af'
C1
f
2(C0 + CL)
When an inductor is connected in series with the crystal, the frequency of operation is decreased. The ability to change the frequency of operation by adding or changing a reactance allows for compensation of the frequency vs temperature variations of crystal units in TCXOs and for tuning the output frequency of voltage-controlled crystal oscillators (VCXO); in both, the frequency can be changed by changing the voltage on a varactor. For the simple RLC circuit of Fig. 4, the width of the resonance curve is inversely proportional to the quality factor Q, but in a crystal oscillator, the situation is complicated by the presence of Co and by the fact that the operating Q is lower than the resonator Q. For a quartz resonator, Q=(27cf~CiR1) -1. References 3, 5, and 6 contain further details on the equivalent circuit.
4
Frequency Control Devices
227
Some of the numerous advantages of a quartz crystal resonator over a tank circuit built from discrete Rs, Cs, and Ls are that the crystal is far stiffer and has a far higher Q than what could be built from normal discrete components. For example, a 5-MHz fundamental mode AT-cut crystal may have C 1 - 0 . 0 1 pF, L 1 - 0 . 1 H, R 1 - 5 f~, and Q - 10 6. A 0.01-pF capacitor is not available, since the leads attached to such a capacitor would alone probably contribute more than 0.01 pE Similarly, a 0.1-H inductor would be physically large, would need to include a large number of turns, and would need to be superconducting to have a resistance <5 f~. 3.
Stability versus Tunability
In most crystal oscillator types, a variable-load capacitor is used to adjust the frequency of oscillation to the desired value. Such oscillators operate at the parallel resonance region of Fig. 5, where the reactance vs frequency slope (i.e., the "stiffness") is inversely proportional to C1. For maximum frequency stability with respect to reactance (or phase) perturbations in the oscillator circuit, the reactance slope (or phase slope) must be maximum. This requires that the C1 be minimum. The smaller the C1, however, the more difficult it is to tune the oscillator (i.e., the smaller is Af' for a given change in CD. The highest stability oscillators use crystal units that have a small C1 (and a high Q). Since C1 decreases rapidly with overtone number, high-stability oscillators generally use third- or fifth-overtone crystal units. Overtones higher than fifth are rarely used, because R1 also increases rapidly with overtone number, and some tunability is usually desirable to allow setting the oscillator to the desired frequency. Wide-tuning-range VCXOs use fundamental mode crystal units of large C1. Voltage control is used for the following purposes: to frequency or phase lock two oscillators; for frequency modulation; for compensation, as in a TCXO (see below); and for calibration (i.e., for adjusting the frequency to compensate for aging). Whereas a high-stability, ovenized 10-MHz VCXO may have a frequency adjustment range of 4-5 x 10 -7 and an aging rate of 2 x 10 -8 per year, a wide-tuning-range 10-MHz VCXO may have a tuning range of -t-50 parts per million (ppm) and an aging rate of 2 ppm per year. In general, making an oscillator tunable over a wide frequency range degrades its stability because making an oscillator susceptible to intentional tuning also makes it susceptible to factors that result in unintentional tuning. For example, if an oven-controlled crystal oscillator (OCXO) is designed to have a stability of 1 • 10 -12 for a particular averaging time and a tunability of
John R. Vig and Arthur BaUato
228
1 x 10 -7, then the crystal's load reactance must be stable to 1 x 10 -5 for that averaging time. Achieving such load-reactance stability is difficult because the load-reactance is affected by stray capacitances and inductances, by the stability of the varactor's capacitance vs voltage characteristic, and by the stability of the voltage on the varactor. Moreover, the 1 x 10 -5 load-reactance stability must be maintained not only under benign conditions, but also under changing environmental conditions (temperature, vibration, radiation, etc.). Therefore, the wider the tuning range of an oscillator, the more difficult it is to maintain a high stability.
4.
Quartz and the. Quartz Crystal Unit
A quartz crystal unit's high Q and high stiffness (small C~) make it the primary frequency- and frequency-stability-determining element in a crystal oscillator. The Q values of crystal units are much higher than those attainable with other circuit elements. In general purpose crystal units, Qs are generally in the range of 104 to 106. A high-stability 5-MHz crystal unit's Q is typically in the range of two to three million. The intrinsic Q, limited by internal losses in the crystal, has been determined experimentally to be inversely proportional to frequency (i.e., the Q f product is a constant for a given resonator type). For AT- and SC-cut resonators, the maximum Qf = 16 million when f i s in MHz. Quartz (which is a single-crystal form of SiO2) has been the material of choice for stable resonators since shortly after piezoelectric crystals were first used in oscillators--in 1918. Although many other materials have been explored, none has been found to be better than quartz. Quartz is the only material known that possesses the following combination of properties: 1. It is piezoelectric ("pressure electric"; piezein means "to press" in Greek). 2. Zero temperature coefficient resonators can be made when the plates are cut along the proper directions with respect to the crystallographic axes of quartz. 3. Of the zero temperature coefficient cuts, one, the SC-cut (see below), is "stress compensated." 4. It has low intrinsic losses (i.e., quartz resonators can have high Qs). 5. It is easy to process because it is hard but not brittle, and, under normal conditions, it has low solubility in everything except the fluoride etchants.
4 Frequency Control Devices
229
6. It is abundant in nature. 7. It is easy to grow in large quantities, at low cost, and with relatively high purity and perfection. The direct piezoelectric effect was discovered by the Curie brothers in 1880. They showed that when a weight was placed on a quartz crystal, charges appeared on the crystal surface; the magnitude of the charge was proportional to the weight. In 1881, the converse piezoelectric effect was illustrated; when a voltage was applied to the crystal, the crystal deformed due to the lattice strains caused by the effect. The strains reversed when the voltage was reversed. Of the 32 crystal classes, 20 exhibit the piezoelectric effect (but only a few of these are useful). Piezoelectric crystals lack a center of symmetry. When a force deforms the lattice, the centers of gravity of the positive and negative charges in the crystal can be separated so as to produce surface charges. The piezoelectric effect can provide a coupling between an electrical circuit and the mechanical properties of a crystal. Under the proper conditions, a "good" piezoelectric resonator can stabilize the frequency of an oscillator circuit. Quartz crystals are highly anisotropic, that is, the properties vary greatly with crystallographic direction. For example, when a quartz sphere is etched in hydrofluoric acid, the etching rate is more than 100 times faster along the fastest etching rate direction, the Z direction, than along the slowest direction, the slow-X direction. The constants of quartz, such as the thermal expansion coefficient and the temperature coefficients of the elastic constants, also vary with direction. That crystal units can have zero temperature coefficients of frequency is a consequence of the temperature coefficients of the elastic constants ranging from negative to positive values. The locus of zero-temperature-coefficient cuts in quartz is shown in Fig. 7. The X, Y, and Z directions have been chosen to make the description of properties as simple as possible. The Z-axis in Fig. 5 is an axis of threefold symmetry in quartz; in other words, the physical properties repeat every 120 ~ as the crystal is rotated about the Z-axis. The cuts usually have two-letter names, where the "T" in the name indicates a temperature-compensated cut; for instance, the AT-cut was the first temperature-compensated cut discovered. The FC-, IT-, BT-, and RT-cuts are other cuts along the zero-temperaturecoefficient locus. These cuts were studied in the past (before the discovery of the SC-cut) for some special properties, but are rarely used today. The highest-stability crystal oscillators employ SC-cut or AT-cut crystal units.
230
John R. Vig and Arthur Ballato 90or~... ,
60"
I
?AT
30 ~ k _
0
,
I
FC r
IT r
f
"
oLC _
-30"
_ RT
-60" -90o I -- 0 o
I
I
I0 ~
!
I 20 ~
,
30 ~
0
Y
The AT, FC, IT, SC. BT, and RT-culs are on the loci ol zero temperature coefficient culs. the LC is a "linear coefficient" cut that is used in a t h e a t e r . Y-cut:
= +90 ppm/OC
(thickness-shear mode)
X-cut: z -20 ppm/OC (extensional mode)
FIG. 7.
Zero-temperature-coefficient cuts of quartz.
Because the properties of a quartz crystal unit depend strongly on the angles of cut of the crystal plate, in the manufacture of crystal units the plates are cut from a quartz bar along precisely controlled directions with respect to the crystallographic axes. The orientations of the plates are checked by means of x-ray diffraction. In some applications, the orientations must be controlled with accuracies of a few seconds of angle. After shaping to required dimensions, metal electrodes are applied to the wafer. Circular plates with circular electrodes are the most commonly used geometries, although the blanks and electrodes may also be of other geometries. The electroded wafer is mounted in a holder structure [38]. Figure 8 shows the two common types of holder structures used for resonators with frequencies greater than 1 MHz. (The 32-kHz tuning fork resonators used in quartz watches are packaged typically in small tubular enclosures.) Because quartz is piezoelectric, a voltage applied to the electrodes causes the quartz plate to deform slightly. The amount of deformation due to an alternating voltage depends on how close the frequency of the applied voltage is to a natural mechanical resonance of the crystal. To describe the behavior of a resonator, the differential equations for Newton's laws of motion for a continuum, and for Maxwell's equations, must be solved with the proper
4
231
Frequency Control Devices Two-point Mount Package
Three- and Four-point Mount Package
Quartz ~-~ . . . . . . . . . . . ."~ Blank ~ ~ - - x ~1-- Electrodes
QBT:nrt:
~ . . . . . -~. . . . . .~-. . . . .~
.on
Cover---~. N [ ~ / ~ I ~ . Bonding
A~ea
Base ~
t ~ ~ J H
Pins
-3__ Topview of cover (a) (b) FIG. 8. Typical constructions of AT-cut and SC-cut crystal units: (a) two-point mount package; (b) three- and four-point mount package.
electrical and mechanical boundary conditions at the plate surfaces [39]. Because quartz is anisotropic and piezoelectric, with ten independent linear constants and numerous higher-order constants, the equations are complex and have never been solved in closed form for physically realizable threedimensional resonators. Nearly all theoretical works have used approximations. The nonlinear elasticconstants, although small, are the source of some of the important instabilities of crystal oscillators, such as the acceleration sensitivity, the thermal-transient effect, and the amplitude-frequency effect, each of which is discussed in this chapter. In an ideal resonator, the amplitude of vibration is maximum at the center of the electrodes; it falls off exponentially outside the electrodes, as shown in the lower fight portion of Fig. 9. In a properly designed resonator, a negligible amount of energy is lost to the mounting and bonding structure, i.e., the edges must be inactive for the resonator to be able to possess a high Q. The displacement of a point on the resonator surface is proportional to the drive current. At the typical drive currents used in (e.g., 10-MHz) thickness-shear resonators, the peak displacement is on the order of a few atomic spacings. (The peak acceleration of a point on the electrodes is on the order of 1 million g.)
John R. Vig and Arthur Ballato
232 _ Metallic Electrodes
\
~
~
/ ~
/"
Resonator Plate Snbstrate (the "blank")
....
In[
[ Conventional resonator geometry and amplitude distribution, [u[
FIG. 9. Resonator vibration amplitude distribution for a circular plate with circular electrodes. As the drive level (the current through a. crystal) increases, the crystal's amplitude of vibration also increases, and the effects due to the nonlinearities of quartz become more pronounced. Some of the many properties that depend on the drive level are resonance frequency, motional resistance R1, phase noise, frequency-vs-temperature anomalies (called activity dips), and frequency jumps, which are discussed in other sections of this chapter. The drive-level dependence of the resonance frequency, called the amplitudefrequency effect, is illustrated in Fig. 10 [40]. The frequency change with drive level is proportional to the square of the drive current; the coefficient depends on resonator design [41 ]. Because of the drive-level dependence of frequency, the highest-stability oscillators usually contain some form of automatic level control to minimize frequency changes due to oscillator circuitry changes. At high drive levels, the nonlinear effects also result in an increase in the resistance [35]. Crystals can also exhibit anomalously high starting resistance when the crystal surfaces possess such imperfections as scratches and particulate contamination. Under such conditions, the resistance at low drive levels can be high enough for an oscillator to be unable to start when power is applied. The drive-level dependence of resistance is illustrated in Fig. 11. In addition to the nonlinear effects, a high drive level can also cause a frequency change due to a temperature increase caused by the energy dissipation in the active area of the resonator. Bulk acoustic wave (BAW) quartz resonators are available in the frequency range of about 1 kHz to 500 MHz. Surface acoustic wave (SAW) quartz
4
Frequency Control Devices
,.~
60
Oz.,
40
233
3
oc~ r ,.~
r~
;>,, u C o ::::1 c~
diopter
Mifz SC
diopter Mtiz SC 20
~
0
I diopter 0 HHz SC
-20
j
, 4
~oo
A
;oo
,,
a
~oo
,;o
*
~oo
~o0
700
Crystal Current (p amps) FIG. 10.
Anomalous starting resistance
r
Drive-level dependence of frequency.
Drive level effects
Normal operating range
"" ,,,.,.~
~J ~
iI
I
10-3
/I / I
%%
~'-,
10 -2
.......
I
I
10 -I
I
........
I ........
I0
I• (mA) FIG. 11. Drive-level dependence of crystal unit resistance.
/
/
I
I00
John R. Vig and Arthur Bailato
234
r
......
-1'
q-7--
---]z
xl
Flexure Mode
Extensional Mode
Face Shear Mode
Thickness Shear Mode
Fundamental Mode Thickness Shear
Third Overtone Thickness Shear
FIG. 12. Modesof motion of a quartz resonator.
resonators are available in the range of about 150 MHz to 1.5 GHz. To cover the wide range of frequencies, different cuts mvibrating in a variety of modes m a r e used. The bulk wave modes of motion are shown in Fig. 12. The AT-cut and SC-cut crystals vibrate in a thickness-shear mode. Although the desired thickness-shear mode usually exhibits the lowest resistance, the mode spectrum of even properly designed crystal units exhibits unwanted modes above the main mode. The unwanted modes, also called spurious modes or spurs, are especially troublesome in filter crystals, in which energytrapping rules are employed to maximize the suppression of unwanted modes [34]. These rules specify certain electrode geometry to plate geometry relationships. In oscillator crystals, the unwanted modes may be suppressed sufficiently by providing a large enough plate diameter to electrode diameter ratio, or by contouring (i.e., generating a spherical curvature on one or both sides of the plate). Above 1 MHz, the AT-cut is commonly used. For high-precision applications, the SC-cut has important advantages over the AT-cut. The AT-cut and SC-cut crystals can be manufactured for fundamental mode operation up to a frequency of about 300 MHz. (Higher than 1 GHz units have been produced on an experimental basis.) Above 100 MHz, overtone units that operate at a
4
235
Frequency Control Devices
selected harmonic mode of vibration are generally used, although higher than 100 MHz fundamental mode units can be manufactured by means of chemical polishing (etching) techniques [42]. Below 1 MHz, tuning forks, X-Yand NT bars (flextural mode), +5 ~ X-cuts (extensional mode), or CT-cut and DT-cut units (face shear mode) can be used. Tuning forks have become the dominant type of low-frequency units due to their small size and low cost. A large number of tuning-fork crystals ('~ 10 9) are produced annually for the worldwide watch market and other applications. These tuning forks must be small, low cost, rugged, stable (as a function of temperature, time, and shock), and must allow a long battery life [43]. The requirements are met with tuning forks operating at 32,768 Hz (which is 215 Hz). This frequency is a compromise among size, power requirement (i.e., battery life), stability, and manufacturing cost. In an analog watch, for example, the 32,768 Hz is divided by two 15 times, resulting in a 1 pulse per second output. These pulses drive a stepping motor that advances the second hand by 6 ~ (i.e., 1/60th of a circle) once every second.
u X
(a)
~t
z l
X
(b)
.. ,. ,****IP u
(c)
FIG. 13. (a) The natural faces and crystallographic axes of quartz; (b) orientation of the tuning fork with respect to the crystallographic axes; (c) flexural vibration mode of the tuning
fork.
John R. Vig and Arthur Ballato
236
Figure 13(a) shows the natural faces and crystallographic axes of quartz, and Fig. 13(b) shows the orientation of the tuning fork with respect to these axes. After processing the tuning fork into a resonator, including the deposition of appropriate electrodes and hermetic sealing into an enclosure, and upon excitation with an appropriate oscillator circuit, the tuning fork vibrates in the flexural vibration mode shown in Fig. 13(c).
B.
OSCILLATOR CATEGORIES
A crystal unit's resonance frequency varies with temperature. Typical frequency vs temperature ( f vs T) characteristics for crystals used in frequency standards are shown in Fig. 14. The three categories of crystal oscillators, based on the method of dealing with the crystal unit's f vs T characteristic, are XO, TCXO, and OCXO (see Fig. 15). A simple XO does not contain means for reducing the crystal'sfvs T variation. A typical X O ' s f vs T stability may be 4-25 ppm for a temperature range o f - 5 5 to +85~ In a TCXO, the temperature-dependent variations of a capacitor external to the crystal compensate for the crystal's f vs T characteristic [44]. The capacitance variations produce frequency changes that are equal and opposite to the frequency changes resulting from temperature changes; in other words,
1---~, BT-cut
AT-cut
15
10
E
s
/
/8'~~"~.'~"~,
:/ /-~/"/
Y-bar quartz
/
Y
,-.
-5
-10
15
/.//!zl/i
~ : i I I I I I I I IN
/ ~ i //~-,,_,!
19 = 350 20 + A8, g)= 0
/~ ";~-; ~. for 30-25-20-15-10
-5
0
5
10
15
20
~
-j
AT-cut. . . . . ,-~,L-_u.-I
fundamental modo pl ,ano-olar,K) AT-cut
/ -45-40-35
5th overtone
~_~
-- ~ \ " "-q... --_..~". ,':
25
Temperature
30
35
40
(oc)
~ 45
50
55
(SO E5
e'{ 70
"/5
8(]
85
FIG. 14. Frequency versus temperature versus angle-of-cut characteristics of AT-cut crystals, with inset showing AT- and BT-but plates in Y-bar quartz.
4
Frequency Control Devices
237
!
Voitage
~~~n%
I/
-45'r br
Output
0 Crystal Oscillator (XO) m .
,;
,
~~-~+
_
Sensor
I
Networkor I
,
1l
co,,,p=o~
_
-45'c,,
Ioxio_61 ::
_
.
.
[~
~
\._.i~~,.. /
ftl
~. . T. ,+. . I X. lO-~
: : ~ ,
9
~ Tem,perature Com pensa,ted (TCXO)
21
I
I 9
I
......
rxl -
-
i
Temperature
I0 X 10-6
"
/'~._\
~
]
+=00'o
-IxlO -8
Oven-~-"--- ' - - Zn
/F---I-J(
xo
I:
IfCO~176L....... i;Temperature I' II F ;I Se.~o, lJ I L
J
U._ __.
Sl
O Oven Controlled (OCXO) FIG. 15. Crystal oscillator categories based on the method of dealing with the crystal unit's frequency versus temperature characteristic.
the capacitance variations compensate for the crystal's f vs T variations. Analog TCXOs can provide about a twentyfold improvement over the crystal's f vs T variation. A good TCXO may have an f vs T stability of -61 ppm for a temperature range o f - 5 5 to +85~ In an OCXO, the crystal unit and other temperature-sensitive components of the oscillator circuit are maintained at a constant temperature in an oven [44]. The crystal is manufactured to have a n f v s T characteristic that has zero slope at the oven temperature. To permit the maintenance of a stable oven temperature throughout the OCXO's temperature range (without an internal cooling means), the oven temperature is selected to be above the maximum operating temperature of the OCXO. OCXOs can provide more than a thousandfold improvement over the crystal's f vs T variation. A good OCXO may have a n f v s T stability of better than -65 • 10 -9 for a temperature range o f - 5 5 to +85~ OCXOs require more power, are larger, and cost more than TCXOs.
238
John R. Vig and Arthur Ballato
A special case of a compensated oscillator is the microcomputer-compensated crystal oscillator (MCXO) [45]. The MCXO overcomes the two major factors that limit the stabilities achievable with TCXOs: thermometry and the stability of the crystal unit. Instead of a thermometer that is external to the crystal unit, such as a thermistor, the MCXO uses a much more accurate "self-temperature sensing" method. Two modes of the crystal are excited simultaneously in a dual-mode oscillator. The two modes are combined such that the resulting beat frequency is a monotonic (and nearly linear) function of temperature. The crystal thereby senses its own temperature. To reduce the f vs T variations, the MCXO uses digital compensation techniques: pulse deletion in one implementation, and direct digital synthesis of a compensating frequency in another. The frequency of the crystal is not "pulled," which allows the use of high-stability (small C1) SC-cut crystal units. A typical MCXO may have an f vs T stability of +2 x 10 -8 for a temperature range of - 5 5 to +85~
C.
OSCILLATORCIRCUIT TYPES
Of the numerous oscillator circuit types, three of the more commonly discussed ones, the Pierce, the Colpitts, and the Clapp, consist of the same circuit except that the rf ground points are at different locations, as shown in Fig. 16. The Butler and modified Butler are also similar to each other; in each, the emitter current is the crystal current. The gate oscillator is a Pierce-type
~
IS
4 Colpitts
Clapp
T T 2//F//Z///I.//'///////F//////f/
2/i/ll///lll/lllllill/l//i/lil 1
Modified Butler
Gate
Pierce
m //////////////////~///////////
Butler
FIG. 16. Oscillatorcircuit types.
4 Frequency Control Devices
239
that uses a logic gate plus a resistor in place of the transistor in the Pierce oscillator. (Some gate oscillators use more than one gate.) Information on designing crystal oscillators can be found in references 1, 2, 5, and 7. The choice of oscillator circuit type depends on such factors as the desired frequency stability, input voltage and power, output power and waveform, tunability, design complexity, cost, and the crystal unit's characteristics. In the Pierce family, the ground point location has a profound effect on the performance. The Pierce configuration is generally superior to the others, e.g., with respect to the effects of stray reactances and biasing resistors, which appear mostly across the capacitors in the circuit rather than the crystal unit. It is one of the most widely used circuits for high-stability oscillators. In the Colpitts configuration, a larger part of the strays appears across the crystal, and the biasing resistors are also across the crystal, which can degrade performance. The Clapp is seldom used because, since the collector is tied directly to the crystal, it is difficult to apply a dc voltage to the collector without introducing losses or spurious oscillations. The Pierce family usually operates at parallel resonance (see Fig. 6), although it can be designed to operate at series resonance by connecting an inductor in series with the crystal. The Butler family usually operates at (or near) series resonance. The Pierce can be designed to operate with the crystal current above or below the emitter current. Gate oscillators are common in digital systems when high stability is not a major consideration. (See the references for more details on oscillator circuits.) Most users require a sine wave, a TTL-compatible, a CMOS-compatible, or an ECL-compatible output. The latter three can be simply generated from a sine wave.
D.
OSCILLATORINSTABILITIES
1. Accuracy, Stability, and Precision Oscillators exhibit a variety of instabilities. These include aging, noise, and frequency changes with temperature, acceleration, ionizing radiation, power supply voltage, etc. The terms accuracy, stability, and precision are often used in describing an oscillator's quality with respect to its instabilities. Figure 17 compares the meanings of these terms for a marksman and for a frequency source. (For the marksman, each bullet hole's distance to the center of the target is the "measurement.")
John R. Fig and Arthur Bailato
240
Precise but not accurate f
Not accurate and not precise
Time
Time
Not stable and not accurate
Accurate and precise f
f
t"
Stable but not accurate
Accurate but not precise
Time
Accurate but not stable
Time
Stable and accurate
FIG. 17. Accuracy, stability, and precision examples for a marksman (top) and for a frequency source (bottom).
Accuracy is the extent to which a given measurement, or the average of a set of measurements for one sample, agrees with the definition of the quantity being measured. It is the degree of "correctness" of a quantity. Frequency standards have varying degrees of accuracy. The International System (SI) of units for time and frequency (second and Hz, respectively) are obtained in laboratories using very accurate atomic frequency standards called primary standards. A primary standard operates at a frequency calculable in terms of the SI definition of the second: "the duration of 9,192,2631,770 periods of the radiation corresponding to the transition between the two hyperfine levels of the ground state of the cesium atom 133" [46]. Reproducibility is the ability of a single frequency standard to produce the same frequency, without adjustment, each time it is put into operation. From the user's point of view, once a frequency standard is calibrated, reproducibility confers the same advantages as accuracy. Stability describes the amount something changes as a function of parameters such as time, temperature, shock, and the like. Precision is the extent to which a given set of measurements of one sample agrees with the mean of the set. (A related meaning of the term is used as a descriptor of the quality
4 Frequency Control Devices
241
of an instrument, as in a "precision instrument." In that context, the meaning is usually defined as accurate and precise, although a precision instrument can also be inaccurate and precise, in which case the instrument needs to be calibrated.)
2. Aging "Aging" and "drift" have occasionally been used interchangeably in the literature. However, recognizing the "need for common terminology for the unambiguous specification and description of frequency and time standard systems," the International Radio Consultative Committee (CCIR) adopted a glossary of terms and definitions in 1990 [47]. According to this glossary, aging is "the systematic change in frequency with time due to internal changes in the oscillator," and drift is "the systematic change in frequency with time of an oscillator." Drift due to aging plus changes in the environment and other factors external to the oscillator. Aging, not drift, is what one denotes in a specification document and what one measures during oscillator evaluation. Drift is what one observes in an application. For example, the drift of an oscillator in a spacecraft might be due to (the algebraic sum of) aging and frequency changes due to radiation, temperature changes in the spacecraft, and power supply changes. Aging can be positive or negative [48]. Occasionally, a reversal in aging direction is observed. At a constant temperature, aging usually has an approximately logarithmic dependence on time. Typical (computer-simulated) aging behaviors are illustrated in Fig. 18, where A(t) is a logarithmic function and B(t) is the same function but with different coefficients. The curve showing the reversal is the sum of the other two curves. A reversal indicates the presence of at least two aging mechanisms. The aging rate of an oscillator is highest when it is first turned on. When the temperature of a crystal unit is changed (e.g., when an OCXO is turned off and turned on at a later time), a new aging cycle starts. (See the section concerning hysteresis and retrace below for additional discussion of the effects of temperature cycling.) The primary causes of crystal oscillator aging are stress relief in the mounting structure of the crystal unit, mass transfer to or from the resonator's surfaces due to adsorption or desorption of contamination, changes in the oscillator circuitry, and, possibly, changes in the quartz material. Because the frequency of a thickness-shear crystal unit, such as an AT-cut or an SC-cut, is inversely proportional to the thickness of the crystal plate, and because a
242
John R. Vig and Arthur Ballato A(t) = 5 Ln(0.5t+l)
Time
<1 A(t) + B(t)
B(t) = -35 Ln(0.006t+l)
FIG. 18. Computer-simulatedtypical aging behaviors, where A(t) and B(t) are logarithm functions with different coefficients.
typical 5-MHz plate is on the order of 1 million atomic layers thick, the adsorption or desorption of contamination equivalent to the mass of one atomic layer of quartz changes the frequency by about 1 ppm. Therefore, to achieve low aging, crystal units must be fabricated and hermetically sealed in an ultraclean, ultra-high-vacuum environment. As of 1998, the aging rates of typical commercially available crystal oscillators range from 5 to 10 ppm per year for an inexpensive XO, to 0.5 to 2 ppm per year for a TCXO, and to 0.05 to 0.1 ppm per year for an OCXO. The highest precision OCXOs can age a few parts in 1012 per day, i.e., less than 0.01 ppm per year.
3.
Noise in Frequency Standards
a. The Effects o f Noise. Sometimes the suitability of oscillators for an application is limited by deterministic phenomena. In other instances, stochastic (random) processes establish the performance limitations. Except for vibration, the short-term instabilities almost always result from noise. Long-term performance of quartz and rubidium standards is limited primarily by the temperature sensitivity and the aging, but the long-term performance of cesium and some hydrogen standards is limited primarily by random processes.
4 Frequency Control Devices
243
Noise can have numerous adverse effects on system performance. Some of these effects are: (1) it limits the ability to determine the current state and the predictability of precision oscillators (e.g., the noise of an oscillator produces time prediction errors of ~'CCry(V) for prediction intervals of v); (2) it limits synchronization and syntonization accuracies; (3) it can limit a receiver's useful dynamic range, channel spacing, and selectivity; (4) it can cause bit errors in digital communications systems; (5) it can cause loss of lock, and limit acquisition and reacquisition capability in phase locked loop systems; and (6) it can limit radar performance, especially Doppler radar. To characterize the random components of oscillator instability, appropriate statistical measures are necessary. Noise characterization has been reviewed [34, 49, 50] and is also the subject of an IEEE standard [51 ]. The two-sample deviation, denoted by Cry(Z), is the measure of short-term instabilities in the time domain. The phase noise, denoted by ~ ( f ) , is the measure of instabilities in the frequency domain. It is related to the phase instability, denoted by S~(f), by ~(f)=_ 89 ).
b. Noise in Crystal Oscillators.
Although the causes of noise in crystal oscillators are not fully understood, several causes of short-term instabilities have been identified. Temperature fluctuations can cause short-term instabilities via thermal-transient effects (see Section III.D.4.b concerning dynamic f vs T effects), and via activity dips at the oven set point in OCXOs. Other causes include Johnson noise in the crystal unit, random vibration (see Section III.D.6 concerning acceleration effects in crystal oscillators), noise in the oscillator circuitry (both the active and passive components can be significant noise sources), and fluctuations at various interfaces on the resonator (e.g., in the number of molecules adsorbed on the resonator's surface). In a properly designed oscillator, the resonator is the primary noise source close to the carrier and the oscillator circuitry is the primary source far from the carrier. The noise close to the carrier (i.e., within the bandwidth of the resonator) has a strong inverse relationship with resonator Q, such that ~ ( f ) c ~ 1/Q 4. In the time domain, Cry(Z)~ (2 x 10-7)/0 at the noise floor. In the frequency domain, the noise floor is limited by Johnson noise, the noise power of which is kT = - 174 dBm/Hz at 290 K. A higher signal (i.e., a higher resonator drive current) will improve the noise floor but not the close-in noise. In fact, for reasons that are not understood fully, above a certain point, higher
244
John R. Fig and Arthur Ballato
drive levels usually degrade the close-in noise. For example, the maximum "safe" drive level is about 100 ga for a 5-MHz fifth overtone AT-cut resonator with Q ~ 2.5 million. The safe drive current can be substantially higher for high-frequency SC-cut resonators. For example, 5 ~ -- - 180 dBc/Hz has been achieved with 100-MHz fifth overtone SC-cut resonators at drive currents 10 mA. However, such a noise capability is useful only in a vibration-free laboratory environment. If there is even a slight amount of vibration at the offset frequencies of interest, the vibration-induced noise will dominate the quiescent noise of the oscillator (see Section III.D.6). When low noise is required in the microwave (or higher) frequency range, SAW oscillators and dielectric resonator oscillators (DROs) are sometimes used. When compared with multiplied-up (bulk acoustic wave) quartz oscillators, these oscillators can provide lower noise far from the carrier at the expense of poorer noise close to the carrier, poorer aging, and poorer temperature stability. SAW oscillators and DROs can provide lower noise far from the carrier because these devices can be operated at higher drive levels, thereby providing higher signal-to-noise ratios, and because the devices operate at higher frequencies, thereby reducing the "20 log N losses" due to frequency multiplication by N. Noise floors at 5 ~ = - 180 dBc/Hz have been achieved with state-of-the-art SAW oscillators [50]. Of course, as is the case for high-frequency BAW oscillators, such noise floors are realizable only in environments that are free of vibrations at the offset frequencies of interest. Figures 19(a) and 19(b) show comparisons of state-of-the-art 5-MHz and 100MHz BAW oscillators and a 500-MHz SAW oscillator, multiplied to 10 GHz. Figure 19(a) shows the comparison in a quiet environment, and Figure 19(b) shows it in a vibrating environment. The short-term stability of TCXOs is temperature (T) dependent and is generally significantly worse than that of OCXOs, for the following reasons: 9 The slope of the TCXO crystal's frequency ( f ) vs T varies with T. For example, the f vs T slope may be near zero at ~20~ but it will be "~ 1 ppm/~ at the T extremes. T fluctuations will cause small f fluctuations at laboratory ambient Ts, so the stability can be good there, but millidegree fluctuations will cause "~ 10-9ffluctuations at the Textremes. The TCXOs f vs T slopes also vary with T; the zeros and maxima can be at any T, and the maximum slopes can be on the order of 1 ppm/~ 9 AT-cut crystal's thermal transient sensitivity makes the effects of T fluctuations depend not only on the T but also on the rate of change
4
Frequency Control Devices
:
~~~
-20
-40
245
BAW = bulk-acoustic wave ] ~ ~ 1 7 6 SAW = surface acoustic wave oscillator
.-=
\
b,.
x
l
.
,
.
.
.
.
5 MHz x 2000
~l_~
1 0
,'~______~
BAWis I
10 1
,
~
,
10 0
I 200
I
101
10 2
BAW
\
lower noise
.i
10 4
BAW
100 MHz X 100 SAW
lower noise
5500, ..
1.03
....
500 MHz x 20
,
I
10 5
10 6
Offset frequency in Hz (a)
~
-20 ~-
-40 -
a:~q
-60 -
~,
-80
m
.,-,
~ . ~ ~ ~ \ . \
~
\
/ - - - - V i b r a t i o n induced phase noise dominates the phase / noise of both (whichever has lower acceleration ./sensitivity will have lower phase noise; currently, ~ 9 BAW can orovide lower sensitivity than SAW ) ~ Illustration assumes 1 x 10"9/I] acceleration ~ ~ sensitivity for both BAW and SAW, and 0.01 . "\ ~ e2/Hz random vibration power spectral ~ . \\ \ ~ density at all vibration frequencies.
\ \
5 MHz x _20~..__.~_.__ B AW
-100
\"
-120
--. \ -
100 MHz x 100
-140 -
500 MHz x 20
-160 =_
l
I
I
1
10-I
10~
101
102
103
~ I
104
I
105
,
I
106
Offset frequency in Hz (b)
FIG. 19. (a) Low-noise SAW and BAW multiplied to 10 GHz (in a nonvibrating environment); (b) low-noise SAW and BAW multiplied to 10 GHz (in a vibrating environment).
John R. Vig and Arthur Ballato
246
of T (whereas the SC-cut crystals typically used in precision OCXOs are insensitive to thermal transients). Under changing T conditions, the T gradient between the T sensor (thermistor) and the crystal will aggravate the problems. 9 TCXOs typically use fundamental mode AT-cut crystals, which have lower Q and larger Cl than the crystals typically used in OCXOs. The lower Q makes the crystals inherently noisier, and the larger C1 makes the oscillators more susceptible to circuitry noise. 9 AT-cut crystals' f v s T often exhibit activity dips. At the Ts where the dips occur, the f v s T slope can be very high, so the noise due to T fluctuations will also be very high, e.g., 100 x degradation of Cry(Z) and 30-dB degradation of phase noise are possible. Activity dips can occur at any T.
c. Frequency Jumps. When the frequencies of oscillators are observed for long periods, occasional frequency jumps can be observed. In precision oscillators, the magnitudes of the jumps are typically in the range of 10 - l l to 10 -9. The jumps can be larger in general purpose units. The jumps occur many times a day in some oscillators, and much less than once a day in others. The frequency excursions can be positive or negative. The causes (and cures) are not well understood. The causes are believed to include nearby spurious resonances, stress relief, changes in surface and electrode irregularities, and noisy active and passive circuit components. The effect can depend on resonator drive level; in some units, frequency jumps can be produced at certain drive levels (but not below or above). Aging affects the incidence. Well-aged units show a lower incidence of jumps than new units. Environmental effects can also produce jumps. Magnetic field, pressure, temperature, and power transients can produce sudden frequency excursions, as can shock and vibration. It is not unusual, for example, to experience shock and vibration levels of >0.01 g in buildings as trucks pass by, heavy equipment is moved, boxes are dropped, etc. [0.02 g x 1 0 - 9 / g - 2 • 10-11]. 4.
Frequency versus Temperature Stability
a. Static Frequency versus Temperature Stability. As an illustration of the effects that temperature can have on frequency stability, Fig. 20 shows the effects of temperature on the accuracy of a typical quartz wristwatch. Near the wrist temperature, the watch can be very accurate because of the frequency of the crystal (i.e., the clock rate) changes very little with temperature. However,
4
Frequency Control Devices
247
Temperature coefficient of frequency = -0.035 pprn]*C2 r~
0
~D
20 t
t
-55 ~ Military "Cold"
-10 ~ Winter
FIG. 20.
t
t
+28 ~ +49 ~ Wrist Desert Temp.
t
+85 ~ Military "Hot"
Wristwatch accuracy as it is affected by temperature.
when the watch is Cooled to -55~ or heated to +100~ it loses about 20 seconds per day, because the typical temperature coefficient of frequency of the tuning-fork crystals used in quartz watches is -0.035 ppm/~ 2. The static f v s T characteristics of crystal units are determined primarily by the angles of cut of the crystal plates with respect to the crystallographic axes of quartz [33-35]. "Static" means that the rate of change of temperature is slow enough for the effects of temperature gradients (explained later) to be negligible. As Fig. 14 illustrates for the AT-cut, a small change in the angle of cut can significantly change the f vs T characteristics. The points of zero temperature coefficient, the "turnover points," can be varied over a wide range by varying the angles of cut. The f vs T characteristics of SC-cut crystals are similar to the curves shown in Fig. 14, with the inflection temperature (T,.) shifted to about 95~ (The exact value of T,. depends on the resonator's design.) Other factors that can affect thefvs T characteristics of crystal units include the overtone [52]; the geometry of the crystal plate; the size, shape, thickness, density, and stresses of the electrodes; the drive level; impurities and strains in the quartz material; stresses in the mounting structure; interfering modes; ionizing radiation; the rate of change of temperature (i.e., thermal gradients) [53]; and thermal history. The last two factors are important for understanding the behaviors of OCXOs and TCXOs, and are, therefore, discussed separately.
John R. Vig and Arthur Ballato
248
The effect of harmonics, i.e. "overtones," o n f v s T is illustrated for AT-cut crystals in Fig. 21 [52]. This effect is important for understanding the operation of the MCXO. The MCXO contains an SC-cut resonator and a dual-mode oscillator that excites both the fundamental mode and the third overtone of the resonator. The difference between the fundamental mode f v s T and the third overtone f vs T is due almost exclusively to the difference between the first-order temperature coefficients. Therefore, when the third overtone frequency is subtracted from three times the fundamental mode frequency, the resulting "beat frequency" is a monotonic and nearly linear function of temperature. This beat frequency enables the resonator to sense its own temperature. Interfering modes can cause activity dips (see Fig. 22), which can cause oscillator failure [54]. Near the activity dip temperature, anomalies appear in both the f vs T and resistance (R) vs T characteristics. When the resistance increases at the activity dip, and the oscillator's gain margin is insufficient, the oscillator stops. Activity dips can be strongly influenced by the crystal's drive level and load reactance. The activity dip temperature is a function of CL because the interfering mode usually has a large temperature coefficient and a Cl that is different from that of the desired mode. Activity dips are troublesome in TCXOs, and also in OCXOs when the dip occurs at the oven temperature. The incidence of activity dips in SC-cut crystals is far lower than in AT-cut crystals. An important factor that affects the f vs T characteristics of crystal oscillators is the load capacitor. When a capacitor is connected in series I 50 40
|
I
I
I
1
I
I
I/ilt'
~/t
q
;+,,o-~
30
20 Af
10
T
-
l,y
o -10 -20 -30
~
g
l
~
-40 -50
/at/ i~. -100
-80
e
-
o
o
f
-
c
u
( 0 ) is about the overtone m o d e s .
t
r
-
the reference 0-angle-of-cut is about 3 0 . I -60
m i n u t e s 1 I -40
FIG. 21.
-20
I 0
higher.) i= = 20
i
i
40
60
_
-
J 80
Effects of harmonics on f vs T.
AT
4 Frequency Control Devices
249
>, o c 4) :l o" 4) u.
o
4) o r
RLZ
,
'
H,L, ....
.~_
"7
.1~
-40
..... A , _ _
4)
I -20
I 0
I,
I 20
Temperature
I
I 40
I
I 60
I
I 80
!
I 100
(~
FIG. 22. Activity dips in the frequency versus temperature and resistance versus temperature characteristics, with and without CL.
with the crystal, the f v s T characteristic of the combination is rotated slightly from that of the crystal alone. The temperature coefficient of the load capacitor can greatly magnify the rotation [55]. The f vs T of crystals can be described by a polynomial function. A cubic function is usually sufficient to describe thefvs TofAT-cut and SC-cut crystals to an accuracy of q- 1 ppm. In the MCXO, in order to fit thefvs T data to 4-1 • 10 -8, a polynomial of at least seventh order is usually necessary [56,57]. b. Dynamic Frequency versus Temperature Effects. Changing the temperature surrounding a crystal unit produces thermal gradients when, for example, heat flows to or from the active area of the resonator plate through the mounting clips. The static f v s T characteristic is modified by the thermal-transient effect resulting from the thermal-gradient-induced stresses [53]. When an OCXO is turned on, there can be a significant thermal-transient effect. Figure 23 shows what happens to the frequency output of two OCXOs, each containing an oven that reaches the equilibrium temperature in six minutes. One oven contains an AT-cut, the other, an SC-cut crystal. Thermal gradients in the AT-cut produce a large frequency undershoot that anneals out several minutes after the oven reaches equilibrium. The SC-cut crystal, being "stress-compensated" and thereby insensitive to such thermal-transient-
250
John R. l/ig and Arthur Ballato 10-3
~ 1 0-4 =
AT
0
._ 10-5 0
~ 10.8 " ~
SO
e-
I'-
~1
o"- 10.7
tie
-2 x 10"7 s/K2
for a typical AT-cut resonator.
C
.2 ~
Deviation from static f vs.= T = ~ ~-'t ' where, for example, ~
E
10 .8
|
'
0 @
g -lo "8 lU
~Oven
:
~
"
Warmup Tim
"
8
9
~__-J~~
~' -10 -7 0
(3
" -10
FIG. 23. Warm-upcharacteristics of AT-cut and SC-cut crystal oscillators (OCXOs).
induced stresses, reaches the equilibrium frequency as soon as the oven stabilizes. In addition to extending the warm-up time of OCXOs, when crystals other than SC-cuts are used, the thermal-transient effect makes it much more difficult to adjust the temperature of OCXO ovens to the desired turnover points, and the OCXO frequencies are much more sensitive to oven-temperature fluctuations [5 8]. The testing and compensation accuracies of TCXOs are also adversely affected by the thermal-transient effect. As the temperature is changed, the thermal-transient effect distorts the static f vs T characteristic, which leads to apparent hysteresis [57]. The faster the temperature is changed, the larger is the contribution of the thermal-transient effect to the f vs T performance. c. Thermal Hysteresis and Retrace. The f vs T characteristics of crystal oscillators do not repeat exactly upon temperature cycling [58]. The lack of repeatability in TCXOs, "thermal hysteresis," is illustrated in Fig. 24. The lack of repeatability in OCXOs, "retrace," is illustrated in Fig. 25. Hysteresis is defined [59] as the difference between the up-cycle and the down-cyclefvs T characteristics and is quantified by the value of the difference at the temperature where the difference is maximum. Hysteresis is determined
4
Prequency Control Devices
231
1.0
0.5
~oo
i per tur
~
O.
-I.(3 FIG. 24. Temperature-compensated crystal oscillator (TCXO) thermal hysteresis, showing that the firstfvs T characteristic upon increasing temperature differs from the characteristic upon decreasing temperature.
15
|
'
-
'
I .
.
OVEN -.e- 1.4 D A Y S - - ~ OFF I
10
.
-.
.
.
.
~
~
~ oven oN .
0
.
.
.
.
.
.~
..
|
(a) t~
<1 [ ~
15
OSCILLATOR I O F F -.e-14 D A Y S " ~ I0 . . . . . , -... . . . ,,,
i
0j,
l
.
.
~
"'-/
..
I
I
~ O S C I L| L A T O R ON9
(b) FIG. 25. Oven-controlled crystal oscillator (OCXO) retrace example, showing that upon restarting the oscillator after a 14-day off-period, the frequency was about 7 x 10 .9 lower than what it was just before turn-off, and that the aging rate had increased significantly upon restart. About a month elapsed before the pre-turn-off aging rate was reached again. (The figure shows A f / f in parts in l 0 - 9 VS time in days.)
252
John R. Vig and Arthur Ballato
during at least one complete quasi-static temperature cycle between specified temperature limits. Retrace is defined as the nonrepeatability of the f vs T characteristic at a fixed temperature (which is usually the oven temperature of an OCXO) upon on-off cycling an oscillator under specified conditions. Hysteresis is the major factor limiting the stability achievable with TCXOs. It is especially so in the MCXO because, in principle, the digital compensation method used in the MCXO would be capable of compensating for thefvs T variations to arbitrary accuracy if the f vs T characteristics could be described by single-valued functions. Retrace limits the accuracies achievable with OCXOs in applications where the OCXO is on-offcycled. Typical values of hysteresis in TCXOs range from 1 to 0.1 ppm when the temperature-cycling ranges are 0 to 60~ and - 5 5 to +85~ Hysteresis of less than 1 • 10 -8 has been observed in SC-cut (MCXO) resonators [56]. The typical MCXO resonator hysteresis in early models of the MCXO was a few parts in 108 [56, 57]. Typical OCXO retrace specifications, after a 24-hour off period at about 25~ range from 2 • 10 -8 to 1 • 10 -9. Low-temperature storage during the off period, and extending the off period, usually make the retrace worse [58]. The causes of hysteresis and retrace are not well understood; the experimental evidence to date is inconclusive [58]. The mechanisms that can cause these effects include strain changes, changes in the quartz, oscillator circuitry changes, contamination redistribution in the crystal enclosure, and apparent hysteresis or retrace due to thermal gradients.
5.
Warm-Up
When power is applied to a frequency standard, it takes a finite amount of time before the equilibrium frequency stability is reached. Figure 23, discussed above, illustrates the warm-up of two OCXOs. The warm-up time of an oscillator is a function of the thermal properties of the resonator, the oscillator circuit and oven construction, the input power, and the oscillator's temperature prior to turn-on. Typical warm-up time specifications of OCXOs (e.g., from a 0~ start) range from 3 to 10 minutes. Even TCXOs, MCXOs, and simple XOs take a few seconds to "warm up," although these are not ovenized. The reasons for the finite warm-up (i.e., stabilization) periods are that it takes a finite amount of time for the signal to build up in any high-Q circuit and that the few tens of milliwatts of power dissipated in these oscillators can change the thermal conditions within the oscillators.
4
6.
Frequency Control Devices
253
Acceleration Effects
Acceleration changes a crystal oscillator's frequency [5, 6]. The acceleration can be a steady-state acceleration, vibration, shock, attitude change (2-g tipover), or acoustic noise. The amount of frequ_ency change depends on the magnitude and direction of the acceleration A, and on the acceleration sensitivity of the oscillator F. The acceleration sensitivity is a vector quantity. The frequency change can be expressed as ~• =f F . A ~
f
Typic~ values of IFI are in the range of 10-9/g to 10-1~ For example, when F - 2 x 10-9/g and is normal to the earth's surface, and the oscillator is turned upside down (a change of 2 g), the frequency changes by 4 x 10 -9. When this oscillator is vibrated in the up-and-down direction, the timedependent acceleration modulates the oscillator's output frequency at the vibration frequency, with an amplitude of 2 x 10-9/g. When an oscillator is rotated 180 ~ about a horizontal axis, the scalar product of the gravitational field and the unit vector normal to the initial "top" of the oscillator changes from - 1 to + 1 g, i.e., by 2 g. Figure 26 shows actual data of the fractional frequency shifts of an oscillator when the oscillator was rotated about three mutually perpendicular axes in the earth's gravitational field. For each curve, the axis of rotation was horizontal. The Axis 3
-10.00oMru z~kf
x 10 "9 f 9
:
4
tipover lest - ppb OF(max) - F(min))/2 = 1.889 -09 (ccw) (F(max) - F(min)}/2 = 1.863 -09 (r delia THETA = 106.0 deg.
(F(max) - F(min))/2 = 6.841 -10 (of w) (F(max) - F(min))/2 = 6.896 -10 (cw) della T H E T A = 150.0 deg.
4[ 2t
Axis 2 4 I
g FIG. 26.
-41 Axis 3
(F(max) - F:(min))/2 = 1.882 -09 (ccw) (F(max) - F(min))/2 = 1.859 -09 (cw~
-
2-g tipover test ( A f v s attitude about three axes).
~
254
John R. Vig and Arthur Ballato
sinusoidal shape of each curve is a consequence of the scalar product being proportional to the cosine of the angle between the acceleration-sensitivity vector and the acceleration due to gravity [6]. In the frequency domain, the modulation results in vibration-induced sidebands that appear at plus and minus integer multiples of the vibration frequency from the carrier frequency. Figure 27 shows the output of a spectrum analyzer for a 10-MHz, 1.4 x 10-9/g oscillator that was vibrated at 100 Hz and 10 g. For sinusoidal vibration, the "sidebands" are spectral lines. When the frequency is multiplied, as it is in many applications, the sideband levels increase by 20 dB for each 10 x multiplication. The increased sideband power is extracted from the carrier. Under certain conditions of multiplication, the carrier disappears, i.e., all the energy is then in the sidebands. The acceleration sensitivity can be calculated from the vibration-induced sidebands. The preferred method is to measure the sensitivity at a number of vibration frequencies to reveal resonances. Figure 28 shows an example of the consequence of a resonance in an OCXO. In this case, the resonance was at 424 Hz, and it amplified the acceleration sensitivity seventeenfold. The effect of random vibration is to raise the phase noise level of the oscillator. The degradation of phase noise can be substantial when the oscillator is on a vibrating platform, such as on an aircraft. Figure 2
T h e "sidebands" are spectral lines at + f, from the carrier frequency (where f, = vibration frequency). The lines are broadened because of the linite bandwidth of the spectrum .
J
NOTE:
[[~ -!
L(f)
-
lOg amplitude @ 100 Hz IF[ = 1.4 x 10 .9 per g
I
~ |
.
=
FIG. 27.
.
'~
.
.
.
.
Vibration-induced "sidebands" (i.e., spectral lines).
f
~
4
255
Frequency Control Devices
10-8
"~ 109
"
~;~
9 " a v n a m ~c
'~a~ge
9 anaWz e~ "5Pe~;t~um
10-10
~o
2;o
~oo-
~oo
~oo-
'
'
:
1000
Vibration Frequency (Hz) FIG. 28.
Resonance in the acceleration sensitivity vs vibration frequency characteristic.
showed a typical aircraft random-vibration specification (power spectral density [PSD] vs vibration frequency) and the resulting vibration-induced phase noise degradation. Acoustic noise is another source of acceleration that can affect the frequency stability of oscillators. The peak phase excursion, (])peak, due to sinusoidal vibration is _-> _.+
Af ~)peak -- Z
F . A )~ radians. --
fv
Upon frequency mu_Jlti~ication, (Dpeak increases by the multiplication factor. For example, if F . A = 1 x 10 -9 , f 0 = 1 0 M H z , and f ~ = 1 0 H z , then ~ p e a k - - 1 x 10 -3 rad. If this oscillator's frequency is multiplied to 10GHz (e.g., in a radar system), then at 10GHz, ~bpeak= 1 rad. Such large phase excursions can be catastrophic to many systems. Figure 29 shows how the probability of detection for a coherent radar system varies with the phase noise of the reference oscillator [60]. The phase noise requirement for a 90% probability of detection of a 4 km/hr target is - 1 3 0 dBc per Hz at 70 Hz from the carrier, for a 10-MHz oscillator. Such as phase noise is well within the capability of 10-MHz oscillators, provided that the oscillators are in a quiet environment. However, when the oscillators are on a vibrating platform, such as an airborne radar system, the phase noise of even the best available oscillators (as of 1997) is degraded by an amount that reduces the probability of detection to zero.
John R. Vig and Arthur Ballato
256
10 0
e
-.
.
.
.
~-80
~
I~,N
- -
ft., 9 60
To "see" 4 km/h targets, low phase noise 70 Hz from the carrier is required. Shown is the probability of detection of 4 km/h targets vs. the phase noise 70 Hz from the carrier of a 10 Ml-lz reference oscillator. (After multiplication to 10 GHz, the phase noise will be at least 60 dB higher ) The phase noise due to platform vibration, e.g., on an aircraft, reduces of detection of slow-moving targets to zero.
-
Low Noise
-140
'
I
-135
High Noise '1-
I
-
I
,
-130 -125 -120 -115 Phase Noise (dBc/Hz) (at 70 Hz fromcarrier,for 4 km/h targets)
-110
FIG. 29. Coherentradar probability of detection as a function of reference oscillator phase noise. During shock, a crystal oscillator's frequency changes suddenly due to the sudden acceleration, as is illustrated in Fig. 30. The frequency change follows the expression above for acceleration-induced frequency change except if during the shock some elastic limits in the crystal's support structure or electrodes are exceeded (as is almost always the case during typical shock tests), the shock will produce a permanent frequency change. Permanent frequency offsets due to shock can also be caused by change in the oscillator circuitry (e.g., due to movement of a wire or circuit board), and the removal of (particulate) contamination from the resonator surfaces. Resonances in the mounting structure will amplify the shock-induced stress. If the shock level is sufficiently high, the crystal will break; however, in applications where high shock levels are a possibility, crystal units with chemically polished crystal plates can be used. Such crystals can survive shocks in excess of 30,000 g and have been fired successfully from howitzers [28, 29].
7.
Magnetic-Field Effects
Quartz is diamagnetic; however, magnetic fields can affect magnetic materials in the crystal unit's mounting structure, electrodes, and enclosure. Time-
257
4 Frequency Control Devices ,',f XlO s I
t
3-
2-
I
0-
-t
-2
to FIG. 30.
The effect o f a shock at t = tl on oscillator frequency.
varying electric fields will induce eddy currents in the metallic parts. Magnetic fields can also affect components such as inductors in the oscillator circuitry. When a crystal oscillator is designed to minimize the effects of magnetic fields, the sensitivity can be much less than 10 -1~ per oersted. Magnetic-field sensitivities on the order of 10 -12 per oersted have been measured in crystal units designed specifically for low magnetic-field sensitivity [61 ]. 8.
Radiation Effects
Ionizing radiation changes a crystal oscillator's frequency primarily because of changes the radiation produces in the crystal unit [62, 63]. Under certain conditions, the radiation will also produce an increase in the crystal unit's equivalent series resistance. The resistance increase can be large enough to stop the oscillation when the oscillator is not radiation hardened. Figure 31 shows a crystal oscillator's idealized frequency response to a pulse of ionizing radiation. The response consists of two parts. Initially, there is a transient frequency change that is due primarily to the thermal-transient effect caused by the sudden deposition of energy into the crystal unit. This
John R. Vig and Arthur Ballato
258
f, = original, preirradiation
frequency
f . = steady-state frequency (0.2 to 24 hours after
exposure)
Af,, = sleady-state frequency
offset
w
ft = frequency at time t
1
to
t
Time
[ 10 "11 for natural quartz (an R increase can stop the oscillation) Af,.,/rad* = { 10 "12 for cultured quartz t 10"1~ for swept cultured quartz ~ for a 1 megarad dose (coefficients are dose dependent)
FIG. 31. Crystaloscillator's response to a pulse of ionizing radiation:fo = original preirradiation frequency, Afss=steady-state frequency offset (0.2 to 24hours aider exposure), ft = instantaneous frequency at time t. effect is a manifestation of the d y n a m i c f v s Teffect previously discussed. The transient effect is absent in SC-cut resonators made of high-purity quartz. In the second part of the response, after steady state is reached, there is a permanent frequency offset that is a function of the radiation dose and the nature of the crystal unit. The frequency change versus dose is nonlinear, the change per rad being much larger at low doses than at large doses. At doses above 1 kilorad (SiO2), the rate of frequency change with dose is quartzimpurity-defect dependent. For example, at a 1-megarad dose, the frequency change can be as large as 10 ppm when the crystal unit is made from natural quartz; it is typically 1 to a few ppm when the crystal is made from cultured quartz, and it can be as small as 0.02 ppm when the crystal is made from swept cultured quartz. The impurity defect of major concern in quartz is the substitutional A13+ defect with its associated interstitial charge compensator, which can be an H +, Li +, or Na + ion, or a hole. This defect substitutes for a Si4+ in the quartz lattice. Radiation can result in a change in the position of weakly bound compensators, which changes the elastic constants of quartz and thereby leads to a frequency change. The movement of ions also results in a decrease in the crystal's Q, i.e., in an increase in the crystal's equivalent series resistance, especially upon exposure to a pulse of ionizing radiation. If the oscillator's gain margin is insufficient, the increased resistance can stop the oscillation for periods lasting many seconds. A high-level pulse of ionizing radiation will
4
Frequency Control Devices
259
produce photocurrents in the circuit, which result in a momentary cessation of oscillation, independent of the type of quartz used in the resonator. In oscillators using properly designed oscillator circuitry and resonators made of swept quartz, the oscillator recovers within 15 gs after exposure [64, 65]. Sweeping is a high-temperature, electric-field-driven, solid-state purification process in which the weakly bound alkali compensators are diffused out of the lattice and replaced by more tightly bound H + ions and holes [66, 67]. In the typical sweeping process, conductive electrodes are applied to the Z surfaces of a quartz bar, the bar is heated to about 500~ and a voltage is applied so as to produce an electric field of about 1 kilovolt per centimeter along the Z direction. After the current through the bar decays (due to the diffusion of impurities) to some constant value, the bar is cooled slowly, the voltage is removed, and then the electrodes are removed. Crystal units made from swept quartz exhibit neither the radiation-induced Q degradation nor the large radiation-induced frequency shifts. Swept quartz (or low-aluminumcontent quartz) should be used in oscillators that are expected to be exposed to ionizing radiation. At low doses (e.g., at a few rads), the frequency change per rad can be as high as 10 -9 per rad [68]. The low-dose effect is not well understood. It is not impurity dependent, and it saturates at about 300rads. At very high doses (i.e., at >>1 Mrad), the impurity-dependent frequency shifts also saturate because, since the number of defects in the crystal are finite, the effects of the radiation interacting with the defects are also finite. When a fast neutron hurtles into a crystal lattice and collides with an atom, the low dose effect is scattered like a billiard ball. A single such neutron can produce numerous vacancies, interstitials, and broken interatomic bonds. The effect of this "displacement damage" on oscillator frequency is dependent primarily on the neutron fluence. The frequency of oscillator increases nearly linearly with neutron fluence at rates of: 8 x 10 -21 neutrons per square centimeter (n/cm 2) at a fluence range of 10 l~ to 1012 n/cm 2, 5 x 10 -21/n/cm 2 at 1012 to 1013 n/cm 2, and 0.7 x 10-21/n/cm 2 at 1017 to 1018n/cm 2.
9.
Other Effects on Stability
Ambient pressure change (as during an altitude change) can change a crystal oscillator's frequency if the pressure change produces a deformation of the crystal unit's or the oscillator's enclosure (thus changing stray capacitances and stresses). The pressure change can also affect the frequency indirectly through a change in heat-transfer conditions inside the oscillator. Humidity
260
John R. Vig and Arthur Ballato
changes can also affect the heat-transfer conditions. In addition, moisture in the atmosphere will condense on surfaces when the temperature falls below the dew point, and can permeate materials such as epoxies and polyimides, and thereby affect the properties (e.g., conductivities and dielectric constants) of the oscillator circuitry. The frequency of a properly designed crystal oscillator changes less than 5 • 10 - 9 when the environment changes from one atmosphere of air to a vacuum. The medium- and long-term stability of some oscillators can be improved by controlling the pressure and humidity around the oscillators [69, 70]. Electric fields can change the frequency of a crystal unit. An ideal AT-cut is not affected by a dc voltage on the crystal electrodes, but "doubly rotated cuts," such as the SC-cut, are affected. For example, the frequency of a 5-MHz fundamental mode SC-cut crystal changes 7 x 10 -9 per volt. Direct-current voltages on the electrodes can also cause sweeping, which can affect the frequencies of all cuts. Power-supply and load-impedance changes affect the oscillator circuitry and, indirectly, the crystal's drive level and load reactance. A change in load impedance changes the amplitude or phase of the signal reflected into the oscillator loop, which changes the phase (and frequency) of the oscillation [70]. The effects can be minimized through voltage regulation and the use of buffer amplifiers. The frequency of a "good" crystal oscillator changes less than 5 x 10- ~ofor a 10% change in load impedance. The typical sensitivity of a highquality crystal oscillator to power-supply voltage changes in 5 • 10-11N. Gas permeation under conditions where there is an abnormally high concentration of hydrogen or helium in the atmosphere can lead to anomalous aging rates. For example, hydrogen can permeate into "hermetically" sealed crystal units in metal enclosures, and helium can permeate through the walls of glass-enclosed crystal units.
10. Interactions among the Influences on Stability
The various influences on frequency stability can interact in ways that lead to erroneous test results if the interfering influence is not recognized during testing. For example, building vibrations can interfere with the measurement of short-term stability. Vibration levels of 10 -3 to 10-2g are commonly present in buildings. Therefore, if an oscillator's acceleration sensitivity is 1 • 10-9/g, then the building vibrations alone can contribute short-term instabilities at the 10 -12 to 10 -11 level.
4 Frequency Control Devices
261
The 2-g tipover test is often used to measure the acceleration sensitivity of crystal oscillators. Thermal effects can interfere with this test because, when an oscillator is turned upside down, the thermal gradients inside the oven can vary due to changes in convection currents [6]. Other examples of interfering influences include temperature and drive-level changes interfering with aging tests; induced voltages due to magnetic fields interfering with vibrationsensitivity tests; and the thermal-transient effect, humidity changes, and the effect of the load-reactance temperature coefficient interfering with the measurement of crystal units' static f vs T characteristics. An important effect in TCXOs is the interaction between the frequency adjustment during calibration and the f vs T stability [71]. This phenomenon is called the trim effect. In TCXOs, a temperature-dependent signal from a thermistor is used to generate a correction voltage that is applied to a varactor in the crystal network. The resulting reactance variations compensate for the crystal's f vs T variations. During calibration, the crystal's load reactance is varied to compensate for the TCXO's aging. Since the frequency vs reactance relationship is nonlinear, the capacitance change during calibration moves the operating point on the frequency vs reactance curve to a point where the slope of the curve is different, which changes the compensation (i.e., compensating for aging degrades the f v s T stability). Figure 32(a) shows how, for the same compensating CL vs T, the compensating f vs T changes when the operating point is moved to a different CL. Figure 32(b) shows test results for a 0.5-ppm TCXO that had a 4-6 ppm frequency-adjustment range (to allow for aging compensation for the life of the device). When delivered, this TCXO met its 0.5 ppm f vs T specification; however, when the frequency was adjusted 4-6 ppm during testing, the f vs T performance degraded significantly. E.
Oscillator Comparison and Selection
The discussion that follows applies to wide-temperature-range frequency standards (i.e., to those designed to operate over a temperature range that spans at least 90~ Laboratory devices that operate over a much narrower temperature range can have much better stabilities than those in the comparison that follows. Commercially available frequency sources cover an accuracy range of several orders of magnitude--from the simple XO to the cesium-beam frequency standard. As the accuracy increases, so does the power requirement, size, and cost. Figure 33, for example, shows the relationship between accuracy and power requirement. Accuracy versus cost would be a similar
John R. Vig and Arthur BaUato
262
Af
fs
/ t
\
F ----
Af
C1
fs
2(C0 + CL)
Compensating f vs. T
~-~
?
Compensating CL vs. T (a)
rag 9 adjustment
2
&
~'~~..
-
-
o
~ + 6
ppm aging adjus,ment (b)
FIG. 32. (a) Change in compensating frequency vs temperature due to CL change; (b) temperature-compensated crystal oscillator (TCXO) trim effect.
4
263
Frequency Control Devices
10-12
[ ccl
1 gs/day
/
/
1 ms/year
/
10-]0
/ /
1 ms/day
/
10 .8
1 s/year
<
J J f
10-6
J
1 s/day 7
10-4 ,e"
0.001
. . . . . l'
0.01
I
0.1
......
I
1
. . . .
10
100
Power (W) FIG. 33. Relationship between accuracy and power requirements (XO: simple crystal oscillator, TCXO: temperature-compensatedcrystal oscillator, OCXO: oven-controlled crystal oscillator, RB" rubidium frequency standard, CS: cesium beam frequency standard).
relationship, ranging from about $1 for a simple XO to about $50,000 for a cesium standard (1997 prices). Table 2 shows a comparison of salient characteristics of frequency standards. Characteristics are provide in Table 2 for atomic oscillators: rubidium and cesium frequency standards and the rubidium-crystal oscillator (RbXO). In atomic frequency standards, the output signal frequency is determined by the energy difference between two atomic states rather than by some property of a bulk material (as it is in quartz oscillators). An introductory review of atomic frequency standards can be found in reference 74, and reference 34 is a review of the literature up to 1983. (Reference 74 reviews both atomic and quartz frequency standards; parts of this chapter are based on the quartz portion of that document.) The RbXO is a device intended for applications where power availability is limited, but where atomic frequency standard accuracy is needed [72, 73]. It consists of a rubidium frequency standard, a low-power, high-stability crystal oscillator, and control circuitry that adjusts the crystal oscillator's frequency to that of the rubidium standard. The rubidium standard is turned on periodically (e.g., once a week) for the few minutes it takes for it to warm up and correct the frequency of the crystal oscillator. With the RbXO, one an approach the long-term stability of the rubidium standard with the low (average) power requirement of the crystal oscillator.
Comparison of Frequency Standards' Salient Characteristics Quartz Oscillators
Accuracy* (per year) AgingIYear Temp. Stab. (range, "C) Stability, q,,(r) (r = 1 s) Size (cm3) Warm-up Time (min) Power (W) (at lowest temp.) Price (-$)
Atomic Oscillators
TCXO
MCXO
OCXO
Rubidium
RbXO
Cesium
2x 5 lo-' 5 lo-' (-55 to +85) 1 10 0.1 (to 1 x 1 0 P ) 0.04 10-100
5x 10P 2 x lo-8 3 x lo-* (-55 to +85) 3 x 10-1° 30 0.1 (to 2 x 0.04 tlOOO
1 x lo-8 5 lo-9 1 (-55 to +85) 1 x 10-l2 20-200 4 (to 1 x 10-8) 0.6 200-2000
5 x 10-1° 2 x lo-'0 3 x 10-lo (-55 to +68) 3 x 10-12 300-800 3 (to 5 x 10-lo) 20 2000-8000
7 x 10-1° 2 x 10-lo 5 x 10-1° (-55 to +85) 5 x 10-l2 1000 3 (to 5 x 10-lo) 0.65 < 10,000
2 x lo-" 0 2 x lo-" (-28 to +65) 5 x lo-" 6000 20 (to 2 x lo-") 30 40.000
'Including environmental effects (note that the temperature ranges for rubid~umand cesium are narrower than for quartz).
6 3
3
2a a
rl
b
& T
4
Frequency Control Devices
265
The major questions to be answered in choosing an oscillator include: 1. What frequency accuracy or reproducibility is needed for the system to operate properly? 2. How long must this accuracy be maintained, i.e., will the oscillator be calibrated or replaced periodically, or must the oscillator maintain the required accuracy for the life of the system? 3. Is ample power available, or must the oscillator operate from batteries? 4. What warm-up time, if any, is permissible? 5. What are the environmental extremes in which the oscillator must operate? 6. What is the short-term stability (phase noise) requirement? 7. What is the size constraint? In relation to the second question, what cost is to be minimized: the initial acquisition cost or the life-cycle cost? Often, the cost of recalibration is far higher than the added cost of an oscillator that can provide calibration-free life. A better oscillator may also allow simplification of the system's design. The frequency of the oscillator is another important consideration, because the choice can have a significant impact on both the cost and the performance. Everything else being equal, an oscillator of standard frequency, such as 5 or 10 MHz, for which manufacturers have well-established designs, will cost less than one of an unusual frequency, such as 8.34289 MHz. Moreover, for thickness-shear crystals, such as the AT-cut and SC-cut, the lower the frequency, the lower the aging [48]. Since at frequencies much below 5 MHz, thickness-shear crystals become too large for economical manufacturing, and since all the highest stability oscillators use thickness-shear crystals, the highest-stability commercially available oscillator's frequency is 5 MHz. Such oscillators will also have the lowest phase noise capability close to the carrier. There are also some excellent 10-MHz oscillators on the market; however, oscillators of much higher frequency than 10 MHz have significantly higher aging rates and phase noise levels close to the carrier than do 5 MHz oscillators. For lowest phase noise far from the carrier, where the signal-to-noise ratio determines the noise level, higher-frequency crystals (e.g., 100 MHz) can provide lower noise because such crystals can tolerate higher drive levels, thereby allowing higher signal levels.
J o h n R. Pig a n d A r t h u r Ballato ~
266
F.
FAILUREMECHANISMS
Crystal oscillators have no inherent failure mechanisms. Some have operated for decades without failure. Oscillators do fail ("go out of spec") occasionally for reasons such as: 9 Poor workmanship and quality control--e.g., wires that come loose at poor-quality solder joints, leaks into the enclosure, and random failure of components 9 Frequency ages to outside the calibration range due to high aging plus insufficient tuning range 9 TCXO frequency vs temperature characteristic degrades due to aging and the "trim effect," and OCXO frequency vs temperature characteristic degrades due to shift of oven set point. 9 Oscillation stops or frequency shifts out of range or becomes noisy at certain temperatures, due to activity dips 9 Oscillation stops or frequency shifts out of range when exposed to ionizing radiation, due to use of unswept quartz or poor choice of circuit components 9 Oscillator noise exceeds specifications, due to vibration induced noise 9 Crystal breaks under shock, due to insufficient surface finish.
G.
SPECIFICATIONS,STANDARDS, TERMS AND DEFINITIONS
There are numerous specifications and standards that relate to frequency control devices. The major organizations responsible for these documents are the Institute of Electrical and Electronics Engineers (IEEE), the International Electrotechnical Commission (IEC), the International Radio Consultative Committee (CCIR), and the U.S. Department of Defense, which maintains the Military Specification (MIL-SPEC) system. A listing of "Specifications and Standards Relating to Frequency Control" can be found in the final pages of the Proceedings of the IEEE Frequency Control Symposium and its predecessor, the Proceedings of the Annual Symposium on Frequency Control. In the 1996 Proceedings, for example 79 such documents are listed [75]. Many of the documents include terms and definitions, some of which are inconsistent. Unfortunately, no single authoritative document exists for terms and definitions relating to frequency standards. The terms and definitions in the CCIR glossary [47], in IEEE Std. 1139 [51 ], and in MIL-O55310s Section 6 [59] are the most recent; they address different aspects of
4 Frequency Control Devices
267
the field, and together form a fairly good set of terms and definitions for users of frequency control devices. The most comprehensive document dealing with the specification of frequency standards is MIL-O-55310 [59]. The evolution of this document over a period of many years has included periodic coordinations between the government agencies that purchase crystal oscillators and the suppliers of those oscillators. The document addresses the specifications of all the oscillator parameters discussed above, plus many others. This specification was written for crystal oscillators. Because the output frequencies of atomic frequency standards originate from crystal oscillators, and because no comparable document exists that addresses atomic standards specifically, MIL-O-55310 can also serve as a useful guide to specifying atomic standards. MIL-STD-188-115, Interoperability and Performance Standards for Communications Timing and Synchronization Subsystems, specifies that the standard frequencies for nodal clocks shall be 1 MHz, 5 MHz, or 5 x 2NMHz, where N is an integer. This standard also specifies a 1-pulse-per-second timing signal of amplitude 10 V, pulse width of 20 gs, rise time less than 20 ns, fall time less than 1 gs; and a 24-bit binary coded decimal (BCD) time code that provides Coordinated Universal Time (UTC) time of day in hours, minutes, and seconds, with provisions for an additional 12 bits for day of the year, and an additional four bits for describing the figure of merit (FOM) of the time signal. The FOMs range from BCD character 1 for better than 1 ns accuracy to BCD character 9 for "greater than 10ms of fault" [76]. IV.
A.
Related Devices
CRYSTALFILTERS
Quartz crystal units are used as selective components in crystal filters [77]. With the constraint imposed by the equivalent circuit of Fig. 3, filter design techniques can provide bandpass or bandstop filters with a range of characteristics. Crystal filters exhibit low insertion loss, high selectivity, and excellent temperature stability. Filter crystals are designed to have only one strong resonance in the region of operation, with all other responses (unwanted modes) attenuated as much as possible. The application of energy-trapping theory [39,77] can provide such a response. If electrode size and thickness are selected in accordance with that theory, the energy of the main response is trapped between the electrodes, whereas the unwanted modes are untrapped and propagate toward the edge of the crystal resonator, where their energy is dissipated. It is
John R. Vig and Arthur Ballato
268
possible to manufacture AT-cut filter crystals with greater than 40-dB attenuation of the unwanted modes relative to the main response. Filters made of piezoelectric ceramics, usually of PZT compositions, cover the range of about 50 Hz to above 10 MHz. As is the case for quartz filters, these devices range from simple resonators to multielectrode resonators to coupled sets of devices. PZTs, having a high coupling coefficient and moderate Q, allow the use of these devices in medium- to high-bandwidth applications. Ceramic filters have poorer frequency vs temperature and longterm stability than do quartz crystal filters. SAW bandpass filters are used in applications such as television receiver intermediate frequency circuits. Such filters can provide stop-band rejection of better than 60 dB and in-band response flat to 0.1 dB. SAW devices can also perform signal processing functions, such as pulse compression and sidelobe suppression that can be performed by dispersive filters and reflective array compressor (RAC) filters, and a chirp transform can provide the Fourier transform of an input signal. B.
SENSORSAND TRANSDUCERS
Whereas in frequency control and timing applications of quartz crystal devices the components are designed to be as insensitive to the environment as possible, resonators made of quartz crystals (and of other piezoelectric materials) can also be designed to be highly sensitive to environmental parameters such as temperature, mass changes, pressure, force, and acceleration. The sensitivity of resonators to mass loading has been exploited in chemical and biological sensors. The sensor applications of piezoelectric resonators were the subjects of special sessions at the 1997 IEEE International Frequency Control Symposium [75] and is the subject of a special issue of the IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control [78]. Quartz crystal transducers can exhibit unsurpassed resolution and dynamic range. For example, one commercial "quartz pressure gage" exhibits a 1-ppm resolution, i.e., 60 Pa at 76 MPa (0.01 lb/in 2 at 11,000 lb/in2), and a 0.025% fullscale accuracy. Quartz thermometers can provide microdegrees of resolution and millidegrees of absolute accuracy over wide temperature ranges. Quartz sorption detectors can detect a change in mass of 10 -12 g. Quartz accelerometer/force sensors are capable of resolving 10 -7 to 10 -8 of full scale. Piezoelectric ceramics are also used for sensors. In some applications, ceramics (e.g., various compositions of lead zirconate titanate, PZT) are preferred over quartz and other single crystal piezoelectrics due to their higher
4 Frequency Control Devices
269
piezoelectric coupling, which can result in a larger signal when a force is applied. For example, ceramics are used in some accelerometers wherein the acceleration of a mass applies a compressional or shear force to the piezoelectric ceramic. The signal conditioning circuitry, which must amplify what is usually a small signal from a high-impedance source, is often part of the device.
V.
For Further Reading
Reference 34 contains a thorough bibliography on the subject of frequency standards to 1983. The principal forum for reporting progress in the field has been the Proceedings of the IEEE Frequency Control Symposium (which was called the Proceedings of the Annual Symposium on Frequency Control prior to 1992) [75]. Other publications that deal with frequency standards include IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control, IEEE Transactions on Instrumentation and Measurement, Proceedings of the Annual Precise Time and Time Interval (PTTI) Applications and Planning Meeting [79], and Proceedings of the European Frequency and Time Forum [80]. Review articles can be found in special issues and publications [34, 81-
85]. Frequency control information can also be found on the World Wide Web at www.ieee.org/uffc/fc.
References 1. Bottom, V. E. (1981). A history of the quartz crystal industry in the USA. Proc. 35th Ann. Symp. Freq. Control., NTIS Accession No. AD-Al10870, pp. 3-12. Copies of the proceedings are available from NTIS 5285 Port Royal Road, Springfield, VA 22161. 2. Havel, J. M. (1956). Crystal requirements for future military equipment. Proc. lOth Ann. Symp. Freq. Contr., NTIS Accession No. AD-A298322, pp. 440-454. 3. Hafner, E. (1972). Frequency control aspects in Army communications and surveillance. Proc. 26th Ann. Symp. Freq. Contr., NTIS Accession No. AD-A771043, pp. 15-19. 4. Kinsman, R. G., Gailus, P. H., and Dworsky, L. N. (1992). Communication system frequency control. In "The Froehlich/Kent Encyclopedia of Telecommunications," Vol. 3 (E E. Froehlich and A. Kent, eds.). Marcel Dekker, New York, pp. 93-121. 5. Filler, R. L., and Vig, J. R. (1990). Low-noise oscillators for airbome radar applications. AGARD Conf. Proc. No. 482, Advances in Components for Active and Passive Airborne Sensors. This paper has been reprinted as U.S. Army Laboratory Command Technical Report SLCET-TR-91-26, same title, October 1991, NTIS Accession No. AD-A242264. 6. Vig, J. R. et al. (1992). Acceleration, vibration and shock effects. Proc. 1992 IEEE Freq. Contr. Symp. 763-781. This paper has been reprinted as U.S. Army Laboratory Command Technical Report SLCET-TR-91-3 (Rev. 1), "The Effects of Acceleration on Precision Frequency Sources (Proposed for IEEE Standards Project P1193)," July 1992, NTIS Accession No. AD-A255465.
270
J o h n R. V i g a n d A r t h u r B a U a t o
7. Bloch, M. Frequency Electronics, Inc. (February 1993). Private communication. 8. Sass, P. E (July 1993). Why is the Army interested in spread spectrum? IEEE Comm. Mag., 23-25. 9. Dixon, R. C. (1976). "Spread Spectrum Systems." John Wiley & Sons, New York. 10. Robertson, A. D., and Painter, E C. (September 1985). Tactical jamming. Def. Sci. and Eng., 20-28. 11. Hellwig, H., Stein, S. R., Walls, E L., and Kahan, A. (1978). Relationships between the performance of time/frequency standards and navigation/communication systems. Proc. l Oth Ann. PTTI Applications and Planning Meeting, 37-53. 12. Hessick, D. L., and Euler, W. C. (1984). GPS user receivers and oscillators. Proc. 38th Ann. Symp. Freq. Contr., NTIS Accession No. AD-A217381, pp. 341-362. 13. Abate, J. E. et al. (April 1989). AT&T's new approach to the synchronization of telecommunications networks. IEEE Comm. Mag., 35-45. 14. Kartaschoff, P. (1977). Frequency control and timing requirements for communications systems. Proc. 31st Ann. Symp. Freq. Contr., NTIS Accession No. AD-A088221, pp. 478-483. 15. Hum, J. (1989). "GPS: A Guide to the Next Utility." Trimble Navigation ltd., 645 North Mary Ave., Sunnyvale, CA 94088. 16. Van Dierendonck, A. J., and Birnbaum, M. (1976). Time requirements in the NAVSTAR global positioning system (GPS). Proc. 30th Ann. Symp. Freq. Contr. NTIS Accession No. AD-046089, pp. 375-383. 17. Lewandowski, W., and Thomas, C. (July 1991). GPS time transfer. Proc. IEEE, 79, 991-1000. 18. Fox, C., and Stein, S. R. (1991). GPS time determination and dissemination. Tutorials from the 23rd Ann. PTTI Applications and Planning Meeting, NTIS Accession No. AD-A254745. 19. Skolnik, M. I. (1990). "Introduction to Radar Systems," 2nd Edition. McGraw-Hill, New York. 20. Leeson, D. B., and Johnson, G. E (1996). Short-term stability for a Doppler radar; requirements, measurements, and techniques. Proc. IEEE, 54(2), 244-248. 21. Parker, T. E., and Montress, G. K. (1992). Spectral purity of acoustic resonator oscillators. Proc. 1992 IEEE Freq. Contr. Symp., 340-348. 22. Walls, E L., Felton, C. M., and Martin, T. D. (1990). High spectral purity X-band source. Proc. 44th Ann. Symp. Freq. Contr., 542-548. 23. Willis, N. J. (1990). Bistatic radar. In "Radar Handbook" (M. I. Skolnik, ed.), Chapter 25. McGraw-Hill, New York. 24. Lippermeier, G., and Vernon, R. (1988). IFFN: solving the identification fiddle. Def. Electr., 83-88. 25. Kirchner, D. (1991). Two-way time transfer via communications satellites. Proc. IEEE, 79, 983-990. 26. Schodowski, S. S. et al. (1989). Microcomputer compensated crystal oscillator for low power clocks. Proc. 21st Ann. PTTI Applications and Planning Meeting, NTIS Accession No. ADA224769, pp. 445-464. Details of the MCXO are also described in a series of six papers in the Proc. 43rd Ann. Symp. Freq. Contr. (1989), NTIS Accession No. AD-A235629, pp. 2-36. 27. Reiss, E. H. and Leong, E (unpublished, 1991). U.S. Army Electronics Technology and Devices Lab., Ft. Monmouth, NJ. 28. Vig, J. R., LeBus, J. W., and Filler, R. L. (1977). Chemically polished quartz. Proc. 31st Ann. Symp. Freq. Contr., NTIS Accession No. AD-A088221, pp. 131-143. 29. Filler, R. L. et al. (1980). Ceramic fiatpack enclosed AT and SC-cut resonators. Proc. 1980 IEEE Ultras. Symp., 819-824. 30. Suter, J. J. et al. (1992). The effects of ionizing and particle radiation on precision frequency sources. Proc. 1992 IEEE Freq. Contr. Symp., IEEE Cat. No. 92CH3083-3. 31. Smith, W. L. (1985). Precision oscillators. In "Precision Frequency Control," Vol. 2 (E. A. Gerber and A. Ballato, eds.). Academic Press, New York, pp. 45-98.
4
Frequency Control Devices
271
32. Frerking, M. E. (1978). "Crystal Oscillator Design and Temperature Compensation" Van Nostrand Reinhold, New York. 33. Bottom, V. E. (1982). "Introduction to Quartz Crystal Unit Design." Van Nostrand Reinhold, New York. 34. Gerber, E. A., and Ballato, A. (Eds.) (1985). "Precision Frequency Control." Academic Press, New York. 35. Parzen, B. (1983). "Design of Crystal and Other Harmonic Oscillators." John Wiley & Sons, New York. Chapter 3 of this book, Piezoelectric Resonators, by A. Ballato, is an oscillator-application oriented treatment of the subject. 36. Benjaminson, A. "Computer-Aided Design of Crystal Oscillators." U.S. Army Laboratory Command R&D Technical Report DELET-TR-84-0386-F, August 1985. AD-B096820. "Advances Crystal Oscillator Design." U.S. Army Laboratory Command R&D Technical Report SLCET-TR-85-0445-F, January 1988, AD-B121288. "Advanced Crystal Oscillator Design." U.S. Army Laboratory Command R&D Technical Report SLCET-TR-88-0804-1, February 1989, AD-B 134514. "Advanced Crystal Oscillator Design." U.S. Army Laboratory Command R&D Technical Report SLCET-TR-88-0804-E December, 1991, ADB163808. 37. Hafner, E. (1985). Resonator and device measurements. In "Precision Frequency Control," Vol. 2 (E. A. Gerber and A. Ballato, eds.). Academic Press, New York, pp. 1--44. 38. Kusters, J. A. (1985). Resonator and device technology. In "Precision Frequency Control," Vol. 1 (E. A. Gerber and A. Ballato, eds.). Academic Press, New York, pp. 161-183. 39. Meeker, T. R. (1985). Theory and properties of piezoelectric resonators and waves. In "Precision Frequency Control," Vol. 1 (E. A. Gerber and A. Ballato, eds.). Academic Press, New York, pp. 47-119. 40. Kusters, J. A. (1981). The SC-cut crystal--an overview. Proc. 1981 Ultras. Syrup., 402-409. 41. Filler, R. L. (1985). The amplitude-frequency effect in SC-cut resonators. Proc. 39th Ann. Symp. Freq. Contr., NTIS Accession No. AD-A217404, pp. 311-316. 42. Vig, J. R., LeBus, J. W., and Filler, R. L. (1977). Chemically polished quartz. Proc. 31st Ann. Symp. Freq. Contr., NTIS Accession No. AD-A088221, pp. 131-143. 43. Momosaki, E. A brief review of progress in quartz tuning fork resonators. Proc. 1997 IEEE Int. Freq. Contr. Symp. IEEE Cat. No. 97CH36016, pp. 552-565. 44. Frerking, M. E. (1985). Temperature control and compensation. In "Precision Frequency Control," Vol. 2 (E. A. Gerber and A. Ballato, eds.). Academic Press, New York, pp. 99-111. 45. Schodowski, S. S. et al. (1989). Microcomputer compensated crystal oscillator for low power clocks. Proc. 21st Ann. PTTI Applications and Planning Meeting, 445--464. Available from the U.S. Naval Observatory, Time Services Department, 34th and Massachusetts Avenue, NW, Washington, DC 20392. Details of the MCXO are also described in a series of five papers in the Proc. 43rd Ann. Symp. Freq. Contr. (1989), NTIS Accession No. AD-A235629. 46. XIIIth General Conference of Weights and Measures, Geneva, Switzerland, October 1967. 47. Intemational Radio Consultative Committee (CCIR) (1990). Recommendation No. 686, Glossary. In "CCIR 17th Plenary Assembly," Vol. 7. Standard Frequencies and Time Signals (Study Group 7), CCIR, Geneva, Switzerland. Copies available from International Telecommunications Union, General Secretariat--Sales Section, Place des Nations, CH1211 Geneva, Switzerland. 48. Vig, J. R., and Meeker, T. R. (1991). The aging of bulk acoustic wave resonators, filters, and oscillators. Proc. 45th Ann. Symp. Freq. Contr., IEEE Cat. No. 91CH2965-2, pp. 77-101. 49. Rutman, J., and Walls, E L. (1991). Characterization of frequency stability in precision frequency sources. Proc. IEEE, 79(6), 952-960. 50. Parker, T. E. (1987). Characteristics and sources of phase noise in stable oscillators. Proc. 41st Ann. Syrup. Freq. Contr., NTIS Accession No. AD-A216858, pp. 99-110.
272
J o h n R. V i g a n d A r t h u r B a l l a t o
51. IEEE Standard Definitions of Physical Quantities for Fundamental Frequency and Time Metrology - - Random Instabilities, IEEE Std. 1139. IEEE, 445 Hoes Lane, Piscataway, NJ 08854, U.S.A. 52. Ballato, A., and Lukaszek, T. (1975). Higher order temperature coefficients of frequency of massloaded piezoelectric crystal plates. Proc. 29th Ann. Syrup. Freq. Contr., NTIS Accession No. ADA017466, pp. 10-25. 53. Ballato, A., and Vig, J. R. (1978). Static and dynamic frequency-temperature behavior of singly and doubly rotated, oven-controlled quartz resonators. Proc. 32nd Ann. Syrup. Freq. Contr., NTIS Accession No. AD-A955718, pp. 180-188. 54. Ballato, A., and Tilton, R. (1978). Electronic activity dip measurement. IEEE Trans. Instr. and Meas., IM-27, 59-65. 55. Ballato, A. (1978). Frequency-temperature-load capacitance behavior of resonators for TCXO applications. IEEE Trans. Son. Ultras., SU-25, 185-191. 56. Filler, R. L. (1991). Frequency-temperature considerations for digital temperature compensation. Proc. 45th Ann. Syrup. Freq. Contr., IEEE Cat. No. 87CH2965-2, pp. 398-404. 57. Filler, R. L. (1988). Measurement and analysis of thermal hysteresis in resonators and TCXOs. Proc. 42nd Ann. Symp. Freq. Contr, NTIS Accession No. AD-A217275, pp. 380-388. 58. Kusters, J. A., and Vig, J. R. (1990). Thermal hysteresis in quartz resonators--a review. Proc. 44th Ann. Syrup. Freq. Contr., IEEE Catalog No. 90CH2818-03, pp. 165-175. 59. U.S. Department of Defense, Military Specification, Oscillators, Crystal, General Specification for MIL-PRF-55310. The latest revision is available from Military Specifications and Standards, 700 Robbins Ave., Bldg. 4D, Philadelphia, PA 19111-5094. Full text of this document also available at http ://www.dscc.dla.mil/Programs/MilSpec/ListDocs.asp ?BasicDoc=MIL-PRF-5 5 310 60. Taylor, J. (August 8, 1980). Effects of crystal reference oscillator phase noise in a vibratory environment," Technical Memorandum 2799-1011. Motorola, SOTAS Engineering Development Section, Radar Operations. The Stand-Off Target Acquisition System (SOTAS) was a helicopterborne coherent radar system that was being developed for the U.S. Army; the program was canceled before completion. To meet the 90% probability of detection goal for 4 km/h targets while operating from a helicopter, the system required a reference oscillator acceleration sensitivity of 4 • 10 -12 perg, which was (and in 1997, still is) about hundredfold beyond the state of the art. 61. Brendel, R. (1996). Influence of magnetic field on quartz crystal resonators. IEEE Trans. Ultrason., Ferroelec., and Freq. Ctrl, 43, 818-831. 62. King, J. C., and Koehler, D. R. (1985). Radiation effects on resonators. In "Precision Frequency Control," Vol. 2 (E. A. Gerber and A. Ballato, eds.). Academic Press, New York, pp. 147-159. 63. Suter, J. J. et al. (1992). The effects of ionizing and particle radiation on precision frequency sources. Proc. 1992 IEEE Freq. Contr. Symp., IEEE Cat. No. 92CH3083-3. 64. King, J. C., and Sander, H. H. (1973). Rapid annealing of frequency change in high frequency crystal resonators following pulsed X-irradiation at room temperature. Proc. 27th Ann. Symp. Freq. Contr., NTIS Accession No. AD-771042, pp. 117-119. 65. Paradysz, R. E., and Smith, W. L. (1973). Crystal controlled oscillators for radiation environments. Proc. 27th Ann. Symp. Freq. Contr., NTIS Accession No. AD-771042, pp. 120-123. 66. Martin, J. J. (1988). Electrodiffusion (sweeping) of ions in quarts. IEEE Trans. Ultrason., Ferroelec., and Freq. Ctrl, 35(3), May IEEE CAtalog 88CH2588-2, pp. 228-296. 67. Gualtieri, J. G. (1989). Sweeping quartz crystals. Proc. 1989 IEEE Ultras. Symp., 381-391. 68. Flanagan, T. M., Leadon, R. E., and Shannon, D. L. (1986). Evaluation of mechanisms for lowdose frequency shifts in crystal oscillators. Proc. 40th Ann. Symp. Freq. Contr., NTIS Accession no. AD-A235435, pp. 127-133. 69. Walls, E L. (1988). The influence of pressure and humidity on the medium and long-term frequency stability of quartz oscillators. Proc. 42nd Ann. Syrup. Freq. Contr., NTIS Accession No. AD-A217725, pp. 279-283.
4
Frequency Control Devices
273
70. Walls, E L., and Vig, J. R. (1995). Fundamental limits on the frequency stabilities of crystal oscillators. IEEE Trans. Ultras., Ferroelec., and Freq. Ctrl, 42, 576-589. 71. Filler et al. (1989). Specification and measurement of the frequency versus temperature characteristics of crystal oscillators. Proc. 43rd Ann. Symp. Freq. Ctrl, NTIS Accession No. AD-A235629, pp. 253-256. 72. Vig, J. R., and Rosati, V. J. (1984). The rubidium-crystal oscillator hybrid development program. Proc. 16th Ann. PTTI Applications and Planning Meeting, 157-165. 73. Riley, W. J., and Vaccaro, J. R. (1987). A rubidium-crystal oscillator (RbXO). IEEE Trans. Ultrason., Ferroelec., and Freq. Ctrl, UFFC-34, 612-618. 74. Stein, S. R., and Vig, J. R. (1992). Communications frequency standards. In "The Froehlich/Kent Encyclopedia of Telecommunications," Vol. 3 (E E. Froehlich and A. Kent, eds.). Marcel Dekker, New York, pp. 445-500. A reprint of this chapter is available under the title "Frequency Standards for Communications," as U.S. Army Laboratory Command Technical Report SLCET-TR-91-2 (Rev. 1), October 1991, NTIS Accession No. AD-A243211. 75. The Proceedings of the IEEE International Frequency Control Symposium and its predecessor, the Proceedings of the Annual Symposium on Frequency Control, have been published since the tenth symposium in 1956. The latest volumes are available from the IEEE, 445 Hoes Lane, Piscataway, NJ 08854. The earlier volumes are available from the National Technical Information Service, 5285 Port Royal Road, Sills Building, Springfield, VA 22161. Ordering information for all the Proceedings can be found on the World Wide Web at www.ieee.org/uffc/fc (e.g., the Proceedings of the 1996 IEEE International Frequency Control Symposium is available from the IEEE, Cat. No. 96CH35935). 76. U.S. Department of Defense, Military Standard, MIL-STD-188-115, "Interoperability and Performance Standards for Communications Timing and Synchronization Subsystems." The latest revision is available from Military Specifications and Standards, 700 Robbins Avenue, Building 4D, Philadelphia, PA ; 19111-5094. 77. Smythe, R. C., and Wagers, R. S. (1985). Piezoelectric and electromechanical filters. In "Precision Frequency Control," Vol. 2 (E. A. Gerber and A. Ballato, eds.). Academic press, New York, pp. 185-269. 78. IEEE Trans. Ultras., Ferroelec., and Freq. Ctrl., to be published in late 1998. 79. The Proceedings of the Annual Precise Time and Time Interval (PTTI) Applications and Planning Meeting are available from the U.S. Naval Observatory, Time Services Department, 34th and Massachusetts Ave., N.W, Washington, DC 20392-5100. The latest volumes are also available from the National Technical Information Service, 5285 Port Royal Road, Sills Building, Springfield, VA 22161. 80. The Proceedings of the European Frequency and Time Forum are available from the Swiss Foundation for Research in Metrology (FSRM), Rue de l'Orangerie 8, CH-2000 Neuchatel, Switzerland. 81. IEEE Trans. Ultras., Ferroelec., and Freq. Ctrl, UFFC-34, November 1987. 82. IEEE Trans. Ultras., Ferroelec., and Freq. Ctrl, UFFC-35, May 1988. 83. Kroupa, V. E (Ed.) (1983). "Frequency Stability: Fundamentals and Measurement." IEEE Press, New York. 84. Sullivan, D. B. et al. (Eds.). "Characterization of clocks and oscillators." National Institute of Standards and Technology Technical Note 1337. National Institute of Standards and Technology, Boulder, CO 80303-3328. 85. Proc. IEEE, Special Issue on Time and Frequency, 79(7), July 1991.
This Page Intentionally Left Blank
Industrial Ultrasonic Imaging/ Microscopy ROBERT S. GILMORE General Electric Company. Research and Development Center; Schenectady. New York
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .275 Introduction and Historical Review . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277 A . Scanning Acoustic Microscopy (SAM). . . . . . . . . . . . . . . . . . . . . . . . 282 B. Photoacoustic Microscopy (PAM) . . . . . . . . . . . . . . . . . . . . . . . . . . . 282 C . Scanning Electron Acoustic Microscopy (SEAM) . . . . . . . . . . . . . . . . . 282 D. Scanning Laser Acoustic Microscopy (SLAM) . . . . . . . . . . . . . . . . . . . 283 E . 1.5 Ultrasonic Testing vs Ultrasonic Imaging . . . . . . . . . . . . . . . . . . . . 285 111. List of Symbols and Abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . 288 IV. Description and Theory of Acoustic Imaging/Microscopy . . . . . . . . . . . . . . . 289 A . Steps Required to Acquire and Display an Acoustic Image . . . . . . . . . . . . 289 B . Focusing Acoustic Beams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 289 V. Role of Imaged Material: Permitted Resolution . . . . . . . . . . . . . . . . . . . . . 295 A . Type of Data To Be Acquired . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 299 B. Imaging Surfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 300 C . Subsurface Imaging of Volumes . . . . . . . . . . . . . . . . . . . . . . . . . . . 308 D. Measuring Resolution in Scanned Images . . . . . . . . . . . . . . . . . . . . . . 313 E . Image Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 321 VI . Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 323 VII . Conclusions and Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 343 Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 344 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 344 I. I1.
I. Summary Ultrasonic imaging and scanned acoustic microscopy are terms used to describe similar imaging processes at different magnifications and frequencies. (A typical ultrasonic imaging/microscopy system is shown schematically in Fig. 1). Both processes form images by acquiring spatially correlated measurements of the interaction of high-frequency sound waves with materials. With the exception of the interference measurement, called V(z). and the gigahertz frequencies used by the higher-frequency scanning acoustic microscopes. it is difficult to establish operational differences between them . This is especially true since almost all commercial ultrasonic imaging systems use LIJ
PHYSlCAL ACOUSTICS. VOL . XXlV
Adapted and reprinted From Joumal of Physics. D~visionD: Applied Physics Vol . 29. 1996. pages 1389-1417 . ISBN 0-12-477945-X $30.00
Robert S. Gilmore
276
Iz =Y FOCUSED FOCUSEU~RASONICTRANSDUCER ULTRASONIC TRANSMITTTER/ RECEIVER ELECTRONICS
SCAN CONTROLLER
I DATA ACQUISITION/D ISPLAY COMPUTER
MULTI-AXIS SCANNER
,,]
I
MAGNETIC/ OPTICAL STORAGE
FIG. 1. Schematicof a typical ultrasonic imaging/microscopysystem. transducers producing focused beams and can display magnified highresolution images. Ultrasonic C-Scan imaging was developed largely by the ultrasonic nondestructive testing industry. The development was gradual and evolutionary. Over a 50-year period, better and better broadband transducers, electronics, and scanners were developed for operation at progressively higher frequencies, now ranging from 1.0 to 100 MHz. Conversely, scanning acoustic microscopes made a relatively sudden appearance 20 years ago on the campus of Stanford University. The first scanning acoustic microscopes operated at gigahertz frequencies and used microwave electronics that produced acoustic tone-bursts with many wavelengths per pulse. Three factors control resolution in an acoustic image: 9 Diameter of the acoustic beam or its point spread function (PSF) 9 Size and spacing of the pixels making up the image 9 Signal-to-noise ratio (contrast) of the feature being resolved The beam diameter, or PSF, is controlled by the frequency of the ultrasonic pulse and the focal convergence of the beam (or focal length to diameter ratio, Z/d). In the coupling fluid, the Z/d ratio is determined by the transducer diameter and lens, but in the material, the Z/d is established by the material's ultrasonic velocities. Pixels are the squares of color or gray scale that make up computer displays of scanned images. Following Nyquist's criterion, the
5 Industrial Ultrasonic Imaging~Microscopy
277
resolution of those images is twice the size and spacing of the pixels. It follows, therefore, that to support the resolution of an ultrasonic beam, the pixels must be no larger than half that beam diameter. Finally, the contrast of the feature being studied must be (at least) a clear shade of gray above the background produced by the image noise. The noise can be due to the material or the electronics. Written to support industrial ultrasonic inspection of materials, this discussion will emphasize the similarities between imaging and microscopy rather than the differences. The roles of the focusing lens, the pulse frequency and the material being imaged, with respect to the final resolution of an acoustic image, will be considered in detail. It will be shown that additional improvements in resolution can be achieved with image processing. Finally, applications studies in metals, ceramics, composites, attachment methods, coatings, and electronic assemblies will be used to demonstrate specific roles for imaging/microscopy in nondestructive testing. II.
Introduction and Historical Review
Because both ultrasonic C-Scan imaging and scanning acoustic microscopy (SAM) describe processes used to acquire images of sound waves interacting with materials, considerable effort has gone into defining differences between them. The early acoustic microscopes operated at gigahertz frequencies, and ultrasonic imaging was done at megahertz frequencies, but recently work with even that distinction has become blurred, and there are many more similarities than differences between them. The terms scanned ultrasonic microscopy and scanned acoustic microscopy can be used interchangeably, as can ultrasonic C-Scan imaging and scanned acoustic imaging. All are processes that form images of sound wave interactions with materials and involve scanning an acoustic source by electronic or mechanical means. If ultrasonic~acoustic microscopy refers to processes that create "magnified" images of objects, then it follows that ultrasonic or acoustic imaging refers to processes that create images of objects without magnification. Because magnified images involve acquiring the acoustic data from many small areas of the object or region, either the interrogating acoustic beam must be focused, or the energy used to create the acoustic signal must be focused, in order to insonify each of these small areas independently. However, it is not unusual for ultrasonic beams at essentially any frequency to be focused to improve the resolution and contrast of the acoustic image acquired with that beam. Acoustic microscopy has been used to describe
278
Robert S. Gilmore
images acquired at 1 to 3 Mhz and ultrasonic imaging for images at 75 to 100 Mhz. To further the confusion, product names have been introduced by various companies with acronyms that are very similar to C-Scan and SAM. Historically, industrial applications for ultrasonic imaging have evolved toward broadband rather than narrowband (tone-burst) signals. When broadband systems are used, the short time duration of the pulses obscures the near entry surface material for only a few round-trip wavelengths; therefore, thinner material sections can be evaluated. Most of the ultrasonic imaging/ microscopy applications described here were done with broadband pulses and systems. These data can be acquired during uninterrupted mechanical scanning by peak-detecting a single maximum amplitude, detecting both the amplitude and the phase of the largest pulse or by digitizing the entire reflected waveforms from a region in the sample. All three types of image data can be acquired with many commercially available ultrasonic imaging systems, and the images can be displayed as C-Scans, B-Scans, or by solid model imaging formats. In industrial quality control, almost all ultrasonic imaging systems use broadband electronics, detect time-resolved signals, and acquire images with mechanical scanning. One exception would be the use of the phase interference method V(z)). This method requires narrowband systems, and when mechanically scanned, produces phase interference images. At present, and after years of development, ultrasonic C-scan immersion inspection continues to provide significant assistance in materials evaluation. It provides hardcopy and/or computer records of test results. Producing grayscale or color images, it simplifies detection by displaying both high- and lowamplitude signals and discriminates sound material from regions containing flaws on the basis of signal amplitude. In addition, C-Scan-type imaging spatially correlates multiple signals to indicate a single flaw when successive scan lines and pulses provide multiple signals. Table 1 shows that the use of high-frequency acoustic waves for visualizing the interiors of opaque solid materials was first recognized in Russia by Sokolov and in Germany by Muhlhauser in 1929. ~'3 Muhlhauser filed for a German patent for the concept of ultrasonically inspecting and visualizing material volumes in 1931. Until 1940, only continuous wave transmission techniques were used in all ultrasonic research and development. When Bergman published Ultrachall in 1937,1 he included approximately 600 papers on various continuous wave ultrasonic investigations. Several of these papers mentioned imaging concepts. An additional 90 papers were added when the English translation of Bergmann's book was published in 1938.
5
Industrial Ultrasonic Imaging~Microscopy
279
TABLE 1 ACOUSTIC IMAGING AND MATERIALS EVALUATION: A TIMELINE Year
Event
1929
Muhlhauser and Sokolov independently propose ultrasonic waves for materials evaluation. Scanned imaging is suggested outright in manuscripts by both. 1,2,3 Muhlhauser obtains German patent for ultrasonic testing of materials using continuous wave transmission. 1,2,3 Sokolov Tube patented and published; Sokolov observes 3-GHz acoustic images equal optical resolution. 1,2,3,4 Bergmann writes Ultraschall (in German) with 600 references (English translation in 1938). 1 Firestone invents pulse-echo ultrasonic testing (patent awarded 1942). SNT (later ASNT) chartered to provide a professional forum for nondestructive testing. 2'3 Sperry acquires Firestone patent (ultrasonic reflectoscope). Erdman, Krautkramer, Pringle, and Smack develop ultrasonic C-Scan equipment. Hastings, using an Erdman system, makes gray-scale C-Scan images: film, paper. 3 First A S N T Handbook, R. C. McMaster, ed. C-Scans, focused probes, scanned images, CRT gray-scale, etc. 3 Dunn proposes scanned ultrasonic absorption microscope. 5 Jacobs et al. add electron multiplier to Sokolov Tube. 6 Korpel et al. at Zenith Corp. invent scanning laser acoustic microscope. 7 First International Symposium on Acoustical Holography ("Holography" later changed to "Imaging"), A. Metherell, ed. 8 Batalle founded Holotron Inc. (later Holosonics Inc.) to market acoustic holography systems. K. Fowler (Panametrics Inc.) introduces and markets a quartz buffer-rod lensfocused 50-MHz transducer. 9 Lemons and Quate invent and introduce 1-GHz SAM. 1~ Stanford group includes G. Kino, P. Khuri-Yakub, and B. Auld. Sonoscan Inc. founded by L. Kessler to market SAM. ~ E. Ash builds SAM group at University College, London, UK that includes C. Tsai and H. Wickramasinghe. Tsai builds second SAM group in U.S. at Carnegie-Mellon University.12 E. I. Leitz Ltd. and Olympus Ltd. introduce scanning acoustic microscopes to international market. Imaging is now seeing rapid growth. More than 30 firms manufacture industrial acoustic imaging/microscopy systems for an international market.
1931 1936 1937 1940 1945 to 1958
1959
1963 1966 1967 1969
1971 1973 1974 1977 1980 1990s
Sokolov was the first to produce a device to make acoustic images and also to realize that 3-GHz acoustic images would rival optical images in resolution and detail. The Sokolov Tube, first described in 1936, was an image converter that produced a television display by scanning the back surface of a (quartz) piezoelectric receiver element with an electron beam. The acoustic amplitude pattern or shadowgraph on the receiver was produced by passing an acoustic beam through the object being imaged. The early images produced by Sokolov Tubes were at best uninspiring. This Was because of poor resolution
280
Robert S. Gilmore
and contrast, and very low signal strength. Pohlmann (1939 in 3) also recognized the "possibility of an acoustic image in analogy to an optical image." Pohlmann Cells depended on the intensity of the ultrasonic waves developing patterns in liquid suspensions of aluminum powder. The resulting images showed little improvement in comparison to the Sokolov Tube, and they required a longer development time to form the particles into patterns. The work of both authors was described by Bergmann 1 and others. 2'3 Firestone's 1940 invention and development of devices to produce pulsed acoustic waves at ultrasonic frequencies, 3 when combined with the development of mechanical scanning systems, initiated rapid development in ultrasonic pulse-echo testing and C-Scan immersion testing. The scanned ultrasonic systems developed by Erdman, Krautkramer, Pringle, and Smack during the 1950s were summarized and described in ASNT's first Nondestructive Testing H a n d b o o k , Volume 11.3 This work also referenced gray-scale images that were originally published in 1955 by Buchanan and Hastings. 13 Most of these scanning systems recorded ultrasonic amplitude data on electrostatic paper. The gray-scale images were recorded on photographic film through Z-axis modulation of the electron beam of an X-Yoscilloscope. Polaroid camera backs became the most popular method for recording the brightness of the oscilloscope displays. However, because the Polaroid images were limited to a 4 • 5 inch format and had to be scanned and then viewed, they were more useful for research and development studies than for nondestructive evaluation. In 1963 Jacobs et al. 6 added an electron multiplier to the Sokolov Tube for a factor of 10 increase in signal strength. Even then Sokolov Tube images showed poor resolution and contrast in comparison to the ultrasonic C-Scan images that were being developed during the same period, but they had the advantage of providing a real-time image. In his presentations, Jacobs usually showed moving pictures because object motion gave the observer better apparent contrast than still images. Korpel et al. 7 developed the scanning laser acoustic microscope during this period. C-Scan testing continued to develop. Because electrosensitive paper was manufactured in widths up to 75 cm (30 in.) large parts could be scanned and imaged with mechanically linked plotters. This made paper the most popular C-Scan recording medium for industrial testing. The pulse-echo or pulsetransmission images were usually of flaw echoes that exceeded an amplitude threshold. For many systems, these images were binary (either black or white), typically at 1 • magnification, and consisted of shaded areas with flaws and white areas without flaws.
5
Industrial Ultrasonic Imaging~Microscopy
281
Thresholds used to detect the flaw signals are usually set to some fraction of the amplitude from a reference reflector. The reference was made the same size as the flaw to be detected. Reference reflectors are still the most commonly used sensitivity references for ultrasonic nondestructive evaluation (NDE). They are typically slots, flat-bottomed holes (FBH), or side-drilled holes and are machined into the same material used to manufacture the part under evaluation. This assures that the attenuation in the calibration sample is the same as in the part, and that the reference flaw is placed in an equivalent microstructure. Acoustic noise from the material is excluded by the threshold. In the late 1960s, gray-scale displays using electrosensitive paper or thermal paper began to replace the binary displays of the preceding decade. But it was the introduction of small computers in the 1970s that provided both acoustic imaging and acoustic microscopy with vastly improved capabilities for industrial NDE, materials science, and medicine. In 1973, work at Stanford University under the direction of C. E Quate 1~ introduced a new dimension to scanned acoustic imaging. The Stanford scanning acoustic microscopes operated at gigahertz frequencies, produced images with optical microscope resolution, and gave a new excitement and focus to the research effort in acoustic imaging and microscopy worldwide. Four decades after Sokolov's prediction, images to rival optical microscopy became a fact. Unlike the lower-frequency C-Scan system, these gigahertzfrequency acoustic microscopes used tone-burst systems having only a few percent bandwidth. During this period broadband commercial transducers and electronics appeared with center frequencies up to 100 MHz. Possibly the most popular transducer design was the broadband 50 to 100 MHz ultrasonic transducer with a single-surface quartz lens, introduced by K. Fowler in 1971.19 Tone-burst excitation was also used in SAM studies at frequencies down to 100 MHz. However, with the exception of almost all ultrasonic industrial NDE, Tsai, 12 Gilmore et al. 14 and the pulse compression acoustic microscopy by Nikoonanhad et al. ~5 most SAM studies until the middle 1980s utilized acoustic pulses containing multiple wavelengths and narrowband systems. Calculations for material properties require amplitude and phase measurements at several heights of the acoustic transducer above the sample. Algorithms for the use of these V(z) data have become highly sophisticated, as reported in the work of Weglein and Wilson, 16'17 Lemons and Quate, 18 Long et al. 19 Bertoni,20 Kushibiki and Chubachi, 21 Wichramasinghe,a2 Somekh et al. 23 Ash, 24 and others. 25'26'27'28 Between 1973 and 1980, other methods for generating and scanning highly focused acoustic waves were developed and used to make acoustic
282
R o b e r t S. Gilmore
imaging/microscopy systems. Most acoustic microscopy may be summarized under four methods that are used to generate, focus, and scan the acoustic energy: scanning acoustic microscopy, photoacoustic microscopy, scanning electron acoustic microscopy, and scanning laser acoustic microscopy. To make a complete introductory survey, it is necessary to list, describe, and reference these four methods.
A.
SCANNINGACOUSTIC MICROSCOPY (SAM)
In scanning acoustic microscopy, images are formed of the surface or interior regions of materials by mechanically scanning piezoelectric transducers that produce focused acoustic beams. These image data are acquired scanned line by scanned line. The resolution of the image is controlled by the diameter of the focused acoustic beam and by the size and spacing of the pixels that display the assembled data. A single focused transducer may be used to transmit and received reflected signals, or two transducers may be used to transmit and receive signals passing completely through the object 24'25'26'27'28 (Fig. 2). Figure 2c shows an example of an industrial scanning acoustic microscope for pulse-echo operation. 29
B.
PHOTOACOUSTICMICROSCOPY (PAM)
Short pulses of light, typically from a laser, are focused and scanned over the surface of the object. The light pulses are formed by chopping or pulsing the light source. Acoustic pulses result from the rapid localized heating of the object surface by each of the light pulses. The acoustic pulses are in turn monitored by a piezoelectric receiver or by laser-sensing the time-delayed surface displacements due to the reflected acoustic pulse. Although the transmitted acoustic pulses may be used to monitor the interior of an object, the best resolution is obtained at or very near the surface of the material 24'25'29 (Fig. 3).
C.
SCANNINGELECTRON ACOUSTIC MICROSCOPY (SEAM)
Electron acoustic microscopy is similar to photoacoustic microscopy in that a pulsed or chopped electron beam is used to produce the acoustic waves. The apparatus is typically made by incorporating a piezoelectric receiver on the stage of a scanning electron microscope (SEM). Scanning electron acoustic
5
Industrial Ultrasonic Imaging~Microscopy
283
SAPPHIRE TRANSMITTER MYLAR LENS SUPPORT SAMPLE
\/
~ /
SENDING TRANSDUCER
/
SDUCER
II - \ II .y SAPPHIRE
[ ..........
I
t MECHANICAL SCAN
SCAN CONTROL AND CRT DRIVE CIRCUITRY
CIRCULATOR PULSED OSCILLATOR
SENDING AND RECEIVING TRANSDUCER
SAMPLE ---..---b. ]
RECEIVER] SYNC
/
YX ]
MECHANICAL SCANNER
I SCAN CONTROL
AND CRT DRIVE CIRCUITRY
l-
FIG. 2. Schematic diagrams of scanning acoustic microscopes. (a) Pulse-transmission mode. (b) Pulse-echo mode. (C. Quate and L. Lam, StanfordUniversity) microscopy is also most effective on very-near-surface conditions, and the samples must be small enough to fit on a SEM stage 31'32 (Fig. 4).
D.
SCANNINGLASER ACOUSTIC MICROSCOPY (SLAM)
The scanning laser acoustic microscope was the first acoustic imaging system that used the term acoustic microscope in the description of the apparatus. This method introduces a broad and uniform acoustic beam through one surface of the object being imaged and then scans the displacement patterns of the opposite surface with a focused laser. While the resulting displays are often clear and informative, they are transmission shadowgraphs, like those formed by the Sokolov Tube and Pohlmann Cells, rather than images. As with
@
0. (D
FIG. 2. (c) This production Ultrasonic Immersion Test System was designed for frequencies to 100 MHz. The stainless steel tank with scanning bridge is anchored onto the anti-vibration base support table and is driven by lead screws. Capable of performing complex scans like those needed for tapered or cone shaped parts, the system has six motorized axes including X, Y, Z, gimbal, swivel and a turntable. This system has a 15 inch turntable that can support a 200 pound load. Courtesy of Sonix Inc., 800 Morrissette Drive, Springfield, VA 22152.
3 b Q
3
c
2
5
Industrial Ultrasonic Imaging~Microscopy
285
FOCUSED
s~,,e ~COUST.C / /,EOE.W, J
/
~
SAPPHIRE
------t FOCUSED / ~ LIGHT SPOT //~[
I\ v YTRRECE/V~NGRS
SOUND ~" II t~, ............. WAVES ~,~x
9
MECHANICAL SCANNER
cat
LDISPLAYJ
SCAN CONTROL AND CRT DRIVE CIRCUITRY
SYNC
] SYNC
I
LASER PROBE (RECEIVER) BEAM
~L
SIGNALOUT
!
~.iSQUARELAW "-I DETECTOR
1
AND LASER
I SCANNER
@
FIG. 3. Schematic diagrams of photoacoustic microscope systems. (a) C. Williams, Stanford University. (b) Y. Martins and E. Ash, University College, London.
all shadowgraphs, the resolution of a feature is dependent on its depth from the scanned surface 7'25'26 (Fig. 5). Figure 5b a scanning laser accoustic microscope by Sonoscan Inc., the company that developed it.
E.
1.5 ULTRASONIC TESTING VS ULTRASONIC IMAGING
Manual observations of ultrasonic waveforms are still the most widely used industrial ultrasonic test method. This has been primarily due to the low cost of the instrumentation in comparison to the cost of an imaging system.
286
Robert S. Gilmore MODIFIED SCANNING ELECTRON MICROSCOPE
r
ELECTRON
I
STAGE
I PIEZOELECTRIC TRANSDUCER RECEIVER
SCAN CONTROL AND CRT DRIVE CIRCUITRY
I
"1
SOUND WAVES
RECEIVER AMPLIFIER
I-
FIG. 4. Schematic diagram of a scanning electron acoustic microscope. (S. Cargill32) Scanned imaging systems have steadily expanded in industrial use, especially where the highest probability of detection for flaws is required. This is because the spatial correlation provided by C-Scan images gives a higher probability of detection than is possible with manual observations, even with the use of automatic flaw alarms and mechanical scanners. Because scanned acoustic microscopy utilizes the same technology as ultrasonic C-Scan imaging, acquiring magnified images of detected flaws is becoming a natural extension of ultrasonic immersion testing. Until the development of gray-scale imaging, little ultrasonic surface wave scanning was done because of the sensitivity of sharply inclined beams to finish roughness, finish direction, and flaw orientation. This surface finish noise made ultrasonic surface wave testing one of the least dependable methods of interrogating surfaces. The surface waves generated by scanning acoustic microscope techniques are also sensitive to surface roughness, but because it images the material structure, surface finish, and any detected flaws, it is less sensitive to the direction of the finish or the orientation of a
5
Industrial Ultrasonic Imaging~Microscopy
287
SCAN CONTROL AND CRT DRIVE CIRCUITRY PHOTODETECTOR I
~-~
I
l LASER SCANNER
,UO,TE
,~,~ ,,yFOVE~-S,,~
OsOSCOt
FLAW SAMPLE
ACOUSTIC FREQUENCY GENERATOR
9 ~i
m
mm l
i~,~i
9
PIEZOELECTRIC INCIDENTSOUND TRANSDUCER
~.~....~
l'IL 2m
ii . . . . .~
~:
:.:~...~
o
:
.....~:
,~,.
~
...~.
~.
,. ;:-.:
~ i ~ : , ~ : ~ ~ . 7 , ..::: : : , ~ ' ~ : : ~ J ........ ~: ~
: ii
i.,i ~
~ ,
9....... 9
.:.
.,..
.
.
9
.
.
.
:.~.
FIG. 5. Schematic diagram o f a scanning laser acoustic microscope. (L. Kessler, Sonoscan Inc.) (b) Scanning Laser Acoustic Microscope, courtesy o f Sonoscan Inc.
Robert S. Gilmore
288
flaw. Therefore, scanning acoustic microscopy methods have greatly contributed to surface inspection at all ultrasonic frequencies. In addition to the detection of flaws, acoustic images also contain substantial information for materials characterization. Grain size can be measured with standard observational techniques applied to acoustic gray-scale images of surfaces. Weglein, 17 Liang et al., 19 Briggs,28 Kushibiki and Chubachi, 21 and others 18'25'27 have developed analytic methods for tone-burst microscopy, and similar calculations have been and are being developed for broadband systems. These developments will be surveyed and summarized for NDE and materials characterization. III. CL Cs CR Cl
C2 C3 DR d
Ex ER Ez Z
Z2 Z3 f 2 22 NDE
~)L | OR SAM PAM SEAM SNR X
List of Symbols and Abbreviations
Velocity of longitudinal wave (mm/las) Velocity of shear or transverse wave (mm/las) Velocity of surface wave (Rayleigh Wave) (mm/~ts) Longitudinal velocity in acoustic lens (mm/las) Longitudinal velocity in coupling (focusing) medium (mm/las) Velocity in test material--this could be Cl, Cs, or CR, depending on application or equation (mm/las) Diameter of entry circle for mode-converted surface wave (mm) Diameter of lens or transducer generating acoustic wave (mm) Lateral acoustic beam diameter (mm) Diameter of surface wave focus (mm) Depth of focus for acoustic beam (mm) Focal length of lens or curved element transducer (mm) Focal length in couplant medium with velocity C2 (mm) Focal length/depth in test material with velocity C 3 (mm) Frequency of acoustic wave (Hz) Wavelength of acoustic wave (mm) Wavelength in velocity C2 (mm) Nondestructive evaluation, a term becoming more commonly used than nondestructive testing First critical angle (longitudinal waves) Second critical angle (shear waves) Rayleigh critical angle (surface waves) Scanning acoustic microscopy--a focused piezoelectric transducer scanned to image a material surface/volume Photoacoustic microscopy--a pulsed light source scanned to image a material surface/volume Scanning electron acoustic microscopy--an electron beam scanned to image a material surface/volume Signal-to-noise ratio Lateral coordinate direction, X = 0, specified at the intersection of the material surface with the axis of symmetry of the acoustic beam Vertical coordinate direction, Z = 0, also at the intersection of the material surface with the beam axis
5
Industrial Ultrasonic Imaging~Microscopy
289
IV. Description and Theory of Acoustic Imaging/Microscopy A.
STEPS REQUIRED TO ACQUIRE AND DISPLAY AN ACOUSTIC IMAGE
To understand the process of acoustic imaging/microscopy, consider the steps required to acquire and display an acoustic image: 1. A sample is selected for imaging and mounted on a scanning stage, and the type of imaging is selected, i.e., surface or subsurface. 2. A transducer (size, frequency, and focus) is selected on the basis of 9 Type of imaging to be done 9 Acoustic velocities in the sample 9 Resolution required for the image 3. Based on the beam diameter (resolution) produced by the transducer, a pixel size and spacing (of no more than half the beam diameter) is selected, and a scan plan is loaded into the scan controller. 4. Data for the image are acquired in the form of signal amplitude, amplitude and phase, time of flight from a reference, or fully sampled wave-forms that can be processed to give all of the preceding information. 5. Color or gray scale is selected to display the image data, and the dynamic range of the acquired/processed data is gain-adjusted to coincide with the dynamic range of the display. To describe the process fully, it is necessary to treat each of these steps in detail, especially the transducers, the materials, and the procedures for surface and subsurface imaging. B.
FOCUSINGACOUSTIC BEAMS
Most commercial focused transducers are constructed with a flat piezoelectric element producing a quasi-plane wave pulse that is focused by a spherical lens. More recently, transducers with spherically curved piezoelectric elements have become available. Both methods produce convergent spherical waves and well-defined focal zones in the coupling fluid. However, when highly focused beams are projected into solid substrates with high acoustic velocities, the beams become badly aberrant. Figure 6 summarizes the focusing of acoustic beams. The point focus produced by a spherically curved piezoelectric element or a very high index lens (Fig. 6(a)) would be subject to at least some spherical aberration when focused by a low-index spherical lens as in Figure 6(b). Regardless of the
Robert S. Gilmore
290 (a)
\ \
HIGH VELOCITY SPHERICAL ~ ~
PARAXIAL RAY
(b)
~ \
LOW VELOCITY S P H E R I C A L ~ PARAXIAL ~ RAY
C2
' DIFFRACTION LIMITED FOCUS
C2
NAL RAY
,,/]~'k~
ABERRATED-.~.[-/~ FOCUS ~/~ "
RA/NAL
VELOC ~-" HIGH SPHERIcA~TY
~-- ASPHERIC DESIGNEDL%NRS
\
\
LENS
c2
(c)
(d)
SUBSTRATE
",,
,J; VELOCITY I SUBSTRATE
,,,' c2
4
VELOCITY SUBSTRATE
ASPHERIC LENS ~
DESIGNED FOR
SUBSTRATE
C2
(e)
PAR ~1~,~'_ ' FOCUS. .WATER" . . . . . . . . ,,',,'WITHj SUBSTRATE (Fpw) REMOVED MARGINAL ASPHERIC FOCUS, WATER WILLNOT FOCUS '
/
FIG. 6. Five general conditions for focusing acoustic beams: (a) Point focus from a spherically curved element. (b) Aberrant focus produced by a low index spherical lens. (c) Aberrant subsurface focus produced by a high-velocity material. (d) Point focus produced by a Fermat lens or element. (e) Distributed focus produced in water by a Fermat lens or element.
5
291
I n d u s t r i a l Ultrasonic I m a g i n g ~ M i c r o s c o p y
focus of the beam in the coupling fluid, it becomes badly aberrant if a highvelocity substrate is inserted between the lens and the focal point (Fig. 6(c)). An aspheric lens or piezoelectric element (Fig. 6(d)) can be designed to produce a well-behaved focus in the solid of a specific velocity and at a specific depth; however, it will not produce a focused beam in the coupling fluid (Fig. 6(e)). Transducers with spherical piezoelectric elements do produce converging spherical waves and diffraction limited beams in the coupling fluid, where spherical lenses produce aberrant beams, hence the term s p h e r i c a l aberration. However, for those spherical lenses with very high indices of refraction to the coupling fluid, the aberration in the fluid is very small. Consider a family of acoustic rays parallel to the axis of symmetry (Fig. 7(a)), where each ray is incident to a spherical surface of radius R c at some angle | If the spherical surface bounds two media characterized by acoustic velocities C 1 and C2, then the refractive relationships between the rays in medium 1 and 2 as C1 sin| 1 -- ~2 sin |
(1)
where Eq. (1) is Snell's law and C 1 / C 2 is the index of refraction for the solidto-fluid interface. Note that high-velocity lens materials such as quartz, silicon, and sapphire produce large values of C 1 / C 2 , the refracted angle | is small, and the refracted ray falls close to the radius of curvature of the lens. If the lateral distance from the axis of symmetry to the ray-lens intersection is X and the distance to the point where the ray crosses the axis is Z, then X = R c sin | Z' = R c ( 1 - cos | Z - Z' =
X
tan(|
_--
R c sin | l
tan(|
1~2)
(2)
-- (~)2)'
and finally, sin| -q- 1 - c o s (~)1)" Z -- R c \tan(O 1 - - 0 2 )
(3)
For small values of O 1, O 2 = [ G / C 2 ] 0 1 , both Z' and ( | 02) are small, cos O 1 = 1.0, and sin(O1 - O2), tan(| - O2), and (O1 - O2) can be taken as equal. Eq. (3) now becomes the "Lens Makers Equation": 9
Zo--R c
1
(1_~12)
.
(4)
Equation (4) is in fact precise for small-angle low-index lenses and for large-angle, high-index lenses.
0
10 20 30 40 50 60 70 80 90 ANGLE OF INCIDENCE 8,
FIG. 7. (a) Ray in medium 1 incident on a spherically curved surface of radius Rc bounding medium 1 and medium 2, where C2 < C1. (b) Aberration curves plotted for the marginal focus Z divided by paraxial focus Z,, produced by rays near the lens center.
5 Industrial Ultrasonic Imaging~Microscopy
293
Comparing the focal lengths determined by Eq. (3) and Eq. (4) show how large angles and low-index lenses can and do produce spherical aberration. Figure 7(b) shows the aberration curves (Z/Zo) for acoustic lenses manufactured from five commonly used lens materials. Note that sapphire and [ 111 ] silicon produces very little aberration and forms a point focus even for lenses with significant curvature. Regardless of how perfectly the beam is focused in the coupling fluid, Figure 6(c) shows that significant aberration will occur upon entry into a highvelocity substrate. In the case of fluid-to-solid or low velocity to high velocity, the high indices of refraction produce greater aberration rather than reduce it. Consider the spherically convergent beam in the fluid (C2) incident on a third higher velocity medium with velocity C3. Snell's law again establishes the relationships:
c3
sin O3 - ~ sin O2.
(5)
When refracted angle O 3 approaches 90 ~ of arc, sin O 3 = 1.0 and sinO 2 -
~33 "
(6)
Because Snell's law is not linear, to produce a well-behaved focus in a highvelocity material, either O3 must be small (actually less than 30 ~ works well) or an aspheric focusing lens must be used (Fig. 6(d)).
1.0
3d
_~ 0 6
6dB -
0.4
02 '" FIG. 8.
1.0
-null
sinc2X c2X
~ ~ 1.2,91 ,
I'11 I i, , ~ ~ 0 +1.0
Schematic showing the point spread function for a circular lens.
Robert S. Gilmore
294
The term point focused is misleading. If all is managed well, the pointfocused transducer produces a diffraction-limited spot size described by its point spread function (PSF). For a circular transducer of diameter d transmitting a wavelength/12, the theoretical beam amplitude profile is 14'19'33'34'35'36 [2J, (X)] 2
P(X)-L (x) J' 0.7
--
0.6
--
(7)
--
15.0
I
0.5
~
12.5
0.4
10.0
z_ ~
17.5
0.3
25
7.5
0.2
5.0 50
0.1
0
.1
5
2.5
10
O0
2.5
.08
2.0
~
.
~ .04
,.o ~
.02
o.5
0
FIG. 9.
~"
5
10
Beam diameters and depths of foci for a range of frequencies and Z/d ratios.
5
Industrial Ultrasonic Imaging~Microscopy
295
where X = 0.5kd(r/Z), k is the wave number (2~z/2), Z is the lens-to-focus distance, and r / Z is the angular distance from the axis of symmetry of the beam. If this function is evaluated at - 1 , - 3 , and - 6 dB with respect to its central maximum, the beam diameters (Fig. 8) are Ex _ K)t2 2 Z
(8)
for K = 0.43, 0.74, and 1.03, respectively. All three numbers are important for acoustic imaging. The - 1 - d B diameter is a good value for the pixel dimension, the - 3 - d B beam is the value producing a single shade of gray in the image, and the -6-dB diameter is the 0.5-amplitude beam used industrywide to specify the transducer resolution. In addition to having a diffraction-limited diameter, the focal zone also has a diffraction-limited depth of fOCUS.14'19'33'34'37 Using the same notation, the -3-dB depth of focus is
-
I 2C- -2. C3
(9)
Figure 9 shows the range in beam diameter and depth of focus for the typical industrial ultrasonic range from 2.25 to 100 MHz.
V.
Role of Imaged Material: Permitted Resolution
Most scanned acoustic imaging systems, including acoustic microscopes, use some form of fluid coupling to introduce their focused acoustic beam into the test material. It is apparent from Eqs. 5 and 6 that a practical understanding is required of the interaction of focused acoustic beams with fluid-solid interfaces if one is to select the correct transducer for acoustic imaging. Consider a planar water-steel interface (Fig. 10). The ambient (22~ longitudinal velocity of deionized water is CL = 1.48 mm/gs. Shear waves in low-viscosity fluids are neglected because their propagation distances are usually less than one wavelength. For this case, the steel is specified to be fine-grained, homogeneous, and isotropic. The longitudinal, shear, and surface wave velocities, respectively, are listed in Table 2. Restricting this discussion to geometrical acoustics and neglecting nonpropagating waves (sometimes called evanescent or head waves), Snell's law (Eq. 5) describes the direction of the incident and refracted rays and the three critical angles at the water-steel interface as shown. The critical angles for the water-steel interface are tabulated in both Figure 10 and Table 2. Clearly the interaction
Robert S. Gilmore
296 r
~qoj.~ -..'~"
~
~L~'~ f
~
LENS
CL= 5.97 mm/ps 0
~
/
/
STEEL C L = 5.90 mm/ps
z PA AXA" OlIU'NAL
I~C_~sk~PARAXIAL/Cs =3.28mm/ps _~_ z
SHEAR
, Cocus N
\\\ \
/
/
FIG. 10. The angular spectrum for a water-steel interface showing the critical angles for longitudinal and shear waves, and the Rayleigh critical angle.
of a spherically focused acoustic beam with a water-steel interface is complicated even though the elastic symmetry of the isotropic material is as simple as can be specified for a solid. The behavior shown in Figure 10 is often referred to as the angular spectrum for steel. The angular spectrum changes with the material. Figure 11 shows three schematics indicating the critical angle behavior for the water-material interface for each of the materials. Five general conclusions can be drawn from a consideration of Snell's law, Table 2, Figure 10, and Figure 11 with respect to the interaction of focused beams with fluid-solid interfaces: 1. Transducers with foci that include longitudinal, shear, or surface wave critical angles will produce aberrated foci. 2. Beyond the first critical angle, | into the test material. 3. Beyond the second critical angle, | the test material.
longitudinal waves do not penetrate shear waves do not penetrate into
4. Beyond the Rayleigh critical angle, O R (surface waves), no acoustic waves penetrate the test medium or interact with the water-material interface, i.e., total reflection occurs for angles of incidence greater than O R.
5
Industrial Ultrasonic Imaging~Microscopy
297
TABLE 2 ACOUSTIC PROPERTIES OF EXAMPLE MATERIALS Materials
Cc
Cs
CR
(mm/ ms)
(mm/ ms)
(mm/ ms)
(degrees of arc)
(degrees of arc)
(degrees of arc)
2.71 3.51 3.51 3.51 1.85 1.87 8.90 2.20 2.33 2.33 3.25 7.89 3.96 3.99 10.50 5.59 1.10 1.23 11.32 19.2 16.5
6.37 17.05 17.63 18.65 13.05 1.10 4.76 5.97 8.90 9.37 11.15 5.90 10.8 11.22 3.70 3.13 2.35 2.67 2.23 5.20 7.02
3.11 11.51 12.81 12.00 8.97 0.59 2.40 3.75 5.84 5.12 6.16 3.28 6.35 6.10 1.69 1.20 1.09 1.12 0.86 2.90 4.10
2.89 9.90 11.10 10.80 7.89 0.55 2.20 3.40 4.91 4.70 5.54 2.96 5.82 5.50 1.57 1.14 1.05 0.95 0.70 2.70 3.90
13.4 5.5 4.8 4.6 6.5 ** 18.1 14.4 9.6 9.1 7.6 14.5 7.9 7.6 23.6 28.2 39.0 33.7 41.6 16.5 12.2
28.4 6.8 6.6 7.1 9.5 ** 38.1 23.3 14.7 ** 13.9 26.8 13.5 14.0 61.1 ** ** ** ** 30.7 21.2
31.0 8.6 7.7 7.9 10.8 ** 42.3 25.8 17.5 ** 15.5 30.0 14.7 15.6 70.5 ** ** ** ** 33.2 22.3
5.9 1.0 0.9 1.6
2.90 1.48 1.45 0.52
Density (gm/cc)
|
|
OR
Solids Aluminum Diamond compact Diamond [ 100]* Diamond [ 111 ]* Beryllium Cesiumt Copper SiO2 (fused) Si [100]* Si [111]* SiN Steel A120 (sintered) Sapphire [001 ]* Silver (Ag) AgC1 Polystyrene Plexiglass Lead (Pb) W WC (4% Co)
Liquids (22~ Gallium (Ga) Water SAE 20 Oil Freon-4
* Longitudinal shear and surface waves are dependent on both direction and polarization in single-crystal and other anisotropic materials. t Cesium (like sodium and lithium) explodes in water. ** Propagation modes in materials with velocities slower than water cannot be produced by mode conversion, and do not produce critical angles.
5. Surface waves cannot be generated by mode conversion when the surface wave velocity is slower than the longitudinal velocity in the fluid coupling medium. (This can be avoided by the use of a lowervelocity fluid.) The coupling fluid typically used in imaging is water, but studies have been done using low-melting-point metals (such as gallium and mercury), mineral
298
Robert S. Gilmore O
O
5.506.8~8 6 ~
I
POLYCRYSTALLINE DIAMOND
--
: :-iJ I
0~ ,~
=68~
0R= 8"6 o
1.4.5 ~
:,'-.,~.";~.";~.~J:~,.'~.!i).;."
~ lii)i~
7
0~
STEEL 0 L = 14.5 ~ 0 s = 26.8 ~ 0 R 29.9~ =
26.8 ~ ~9.9
~
PLASTIC CHIP CARRIER 0L = 28.7 ~
~
: ~,i ! i , . : . i : . i i i ~ ~ 5.3 ~
66.0 ~
0S= 55"3~ OR = 66.0 ~
FIG. 11. Schematicindication of the angular spectra for polycrystalline diamond, steel, and A1203 reinforced plastic (the plastic is used in electronic chip carriers), where water is the incident medium.
or fluorocarbon oils, gases under high pressure, and liquid gases, depending on the frequency of the acoustic beam, the material velocities, chemical reactivity, and temperature, respectively. Regardless of how the beam is focused, the indices of refraction of the coupling fluid to test material controls the imaging process. The highest acoustic velocity measured to date is the extentional velocity in a highly crystalline carbon whisker (19.3 mm/ps). However, for solids not in whisker form, at temperatures near ambient (22~ the longitudinal acoustic velocities are bounded above by the [111] crystallographic direction in diamond (18.6 mm/ps) and below by the [100] direction in cesium (approximately 1.0 mm/ps). Shear or transverse velocities are also bounded by the same materials, the maximum being in the [100] direction in diamond (12.7 mm/~ts) and the minimum being in the [111] direction in cesium
5
Industrial Ultrasonic Imaging~Microscopy
299
(approximately 0.4 mm/ps). Depending on the Poisson's ratio for the material, the surface wave propagates at 0.87 to 0.95 times the velocity of the transverse wave. These data and the acoustic velocities of other materials have been compiled by several authors. 38'39'4~ Eleven fine-grained homogeneous isotropic polycrystalline compacts are tabulated in Table 2. Isotropic compacts have been chosen as example materials because they are the easiest to describe and understand. These materials essentially cover the entire range of acoustic velocity behavior for isotropic solids. Beryllium (Be) is the highest velocity metal and lead (Pb) the lowest velocity material that is encountered in practical situations such as materials joining (soldering). Acoustic velocities in liquids have also been compiled by several investigators. 18'41 The highest velocity is in liquid gallium, at about 30~ and the lowest is in Freon-4, one of the heavier fluorocarbon fluids. Water and SAE 20 motor oil are similar in velocity.
A.
TYPE OF DATATO BE ACQUIRED
Ultrasonic imaging systems usually fall into one of the following four categories, with respect to data acquisition: 1. Single or multiple gates are used to acquire peak detected amplitudes. 2. Single or multiple gates are used to acquire phase and/or amplitude and phase. 3. A single signal gate may be used to acquire amplitude and time of flight. 4. Fully sampled waveforms can be acquired. Systems that record only the maximum amplitude pulse in the signal gate usually use gated peak detectors. Additional circuitry can detect the phase and arrival time of the maximum signal in the gate. Full waveform acquisition requires digitizers with sampling frequencies of at least twice the highest carrier frequency component of the pulses. Once full waveforms are acquired, they can be processed to extract the information in any other acquisition method. The price one pays is in the storage media required to acquire and store the additional information on the signal. Amplitude alone is typically acquired at 8 bits per pixel, and the position in which it is stored on the computer disk corresponds to its spatial position in the image. Amplitude and phase, or amplitude and time of flight, would require at least two 8-bit words per
Robert S. Gilmore
300
ultrasonic pulse. Multiple gate acquisitions in turn require one or more 8-bit words for each gate, each ultrasonic pulse. If full waveforms are acquired, then at least one 8-bit word is acquired for each sample, and often hundreds of samples are acquired for each pixel or image position. For example, a 1024 x 1024 amplitude image can be contained in 1 Mb of computer memory; 2 Mb are required for amplitude and phase or amplitude and time-of-flight images. Finally, 256 Mb are needed if 256 samples are acquired for each waveform/pixel. V(z) images in their most fundamental form are simply scanned amplitude measurements of the phase interference between the direct surface reflection and the surface wave. They may be scanned at constant height to form an image or scanned in various heights (z). Either produces multiple interferences between the surface wave and the direct reflection.
B.
IMAGING SURFACES
Since the discovery of V(z) in the mid 1970s, imaging surface and nearsurface conditions has become synonymous with V(z) imaging. V(z) can only occur when multiwavelength pulses are used, and the direct surface reflection overlaps with the surface wave arrival in time. Figure 12 shows these arrivals, time-resolved with broadband signals. Because the two signals are necessarily of the same frequency, their phase interference is directly controlled by the path length and velocity of the surface wave over the surface of the sample. As will be shown, this is in turn controlled by the height of the transducer above the sample surface and the critical angle for the surface wave mode conversion. Because both time-resolved surface wave imaging and V(z) imaging are controlled by the same wave-material interaction physics, the broadband methods will be developed first and then the same principles applied to V(z). Liang et al. 19 and Briggs 28 have published extensive discussions on tone-burst surface wave imaging or V(z). This discussion will therefore concentrate on time-resolved surface wave imaging, which is the type most used in industrial NDE. Consider the steps required to acquire and display an acoustic image of the surface and near-surface of a sample with an acoustic imaging system such as shown in Figures 1 and 2: 1. The sample is selected, its surface is checked for finish (and is polished if necessary), and the sample is then mounted on a scanning stage.
5
301
Industrial Ultrasonic Imaging~Microscopy I
I~
DL
"1
(6)
(1)
(i)
ZA
er.qR ~
c~ ~
~(3)f"~ /~
~~(4)/
, i v ~v, ~,
CR \
liquid solid
,'
\
V\ ,2~,, \ \ // \
zA
DIRECT REFLECTION(/) SURFACE WAVE
ZB
/
\ \,/
DR
- v -
FIG. 12. Schematic showing 50-MHz surface wave arrival from a superalloy steel at defoci of Z A and Z B or water paths of Z 2 - ZA and Z 2 - Z8. The entry circle ray geometry producing these signals is shown with two different-sized flaws, large and small.
2. A transducer (size, frequency, and F/d) is selected on the basis of the surface wave velocity in the sample and the resolution required for the image. For normal incidence surface wave imaging, the cone of focus must include the surface wave critical angle. For a high-resolution image, a pixel size/spacing is chosen to be 1/6 of the surface wavelength, and the appropriate scan plan is loaded into the scan controller. 3. Depending on system capabilities , data for the image are acquired as signal amplitude, amplitude and phase, time of flight from a reference, or fully sampled waveforms, which can be processed to give all of the preceding information.
302
Robert S. Gilmore
4. A color or gray-scale display is selected, the dynamic range of the acquired/processed data is gain-adjusted to coincide with that of the display, and the image is displayed. Figure 10 shows that when the full angular spectrum of the material is subtended by the cone of focus, then subsurface information from the longitudinal and shear components may be included in a surface wave image. This, however, seldom presents a serious problem. Often the relative amplitude of the pixels displaying a subsurface feature will indicate the nature of the energy forming the image. Because of the pulse transmission nature of surface wave imaging, a region in the image containing amplitudes lower than normally produced by sound material often indicates that a superficial flaw is blocking the surface path for the wave. Conversely, regions containing higher amplitudes often indicate that longitudinal or shear wave echoes from subsurface features are positively interfering with the surface wave signal. Features with low-amplitude edges around a highamplitude center can indicate a near-surface feature initially blocking surface wave paths within the entry circle but then reflecting both longitudinal and shear energy as those much smaller focal zones begin to interact with the feature. The time domain photographs in Figure 12 show two surface wave pulses at two different path lengths (Z A and Zs) in an isotropic material, tungsten carbide. In forming the surface wave arrival, each ray proceeds from the piezoelectric element (1) to the lens (2). Following specifically those rays that are refracted at the Rayleigh critical angle (OR), they proceed in a converging truncated cone of rays to intersect the specimen surface (3) where they intersect it to form the entry circle shown in the lower part of the figure. Those rays at and around OR then mode-convert to surface waves and propagate on a circle diameter to the far side (4), where they reconvert and proceed back to the lens (5) and the piezoelectric element (6). For a buffer rod of length L, a lens of radius R c, focusing at an axial distance Z 2, where it is defocused to axial distances Z A and Z 8, the time (tl) required for the directly reflected pulse, following the ray path on the lens axis of symmetry, to travel from the piezoelectric element and return for Z A is t,
2L
2(Z2 - ZA)
c2
,
(10)
where C1 and C2 are the velocity in the lens and coupling liquid, respectively. The round-trip travel time for the pulse travelling along the ray incident at the
5
303
Industrial Ultrasonic Imaging~Microscopy
Rayleigh critical angle (Fig. 12) is 2L t2 -- ~11 -]-
+
2(1 -4- cos | C1
2(Z2 - ZA) -- 2 R c ( 1 - cos 0L) C 2 cos 0 R
+
2 Z A tan | CR
where | OR, and CR are the lens angle, the Rayleigh critical angle, and the Rayleigh velocity, respectively. From Figure 12 and Snell's law, the diameter (DR) of the entry circle, where the incident pulse in the liquid is mode converted into a leaky Rayleigh wave, and (OR) are D R -- 2 Z A tan O R
and
O R - sin -1 C 2 .
CR
(12)
Writing the differences in round-trip travel time for the direct reflection (Atl) and the surface wave (At2) at the two defocus distances Z A and Z~, combining Eqs. (10), (11), and (12), and solving for CR gives
Equation (13) permits the Rayleigh group velocity to be determined from time measurements at two water path settings where both the velocity in the water and the material are determined. However, for temperature-stabilized water baths, C2 is known and invariant. Therefore, for a constant water velocity, a flat sample, and an invariant waterpath during scanning, the time delay of the surface wave arrival behind the direct reflection tl - t2 can give the surface wave velocity directly. The entry circle schematic (Fig. 12) shows that any discontinuities on the surface that interrupt or change direction of any of the converging bundle of surface wave rays changes the amplitude received by the piezoelectric element. The mechanism by which this occurs can be clarified by considering the two opposing 60 ~ segments of that entry circle, and again by considering only those rays propagating from left to fight. The 60 ~ segment depicts a converging ray bundle point-focused at the center of the entry circle and then diverging to the opposite side of the circle perimeter, what is in fact displayed is a surface wave transmission acoustic microscope that is focused by the curvature of the entry surface circle. Entry circles of three to four surface wavelengths are required to time-resolve the leaky Rayleigh wave from the
304
Robert S. Gilmore
direct reflection. But, unlike subsurface scanning, this does not obscure any portion of the entry surface material. In isotropic materials the surface wave originates at an entry circle and provides a 360 ~ cylindrically convergent, pulse transmission, surface features with equal probability regardless of the direction of their surface strike. Utilizing broadband pulses with center frequencies at 1.0 MHz extends the depth of interrogation of the broadband systems to three surface wavelengths, or 9.0 mm in steels and equivalentvelocity media. This surprising depth is due to the low-frequency components of the broadband pulse. Two parameters describe the resolution of a time-resolved surface wave imaging system. These are the diameter of the entry circle (DR) and the spot size of the surface wave focal zone or crossing zone of focus at the center of the circle. The spot diameter (ER) can be calculated from the amplitude point spread function of the double cylindrical lens system placed on the entry surface by the entry circle. The point spread function produced by a cylindrical lens has been shown by Born and Wolf 35 and Kino 36 to be of the form 2
where Xhas the same definition as Eq. (7) and Fig. 8 for a pulse-echo system. Gilmore et al. 14 used these results to calculate the -3-dB and -1-dB diameters for the point spread function for the crossing zone at the entry circle center: E R =KR2 R,
(14)
where K e equals 0.32 and 0.16, respectively. The surface wave entry circle taken as 180 ~ transmitter/receiver segments may be treated as F/0.5 lenses, which produce very sharply focused spot sizes. The surprisingly small -3-dB diameter (0.32) of the center focus explains the consistently high resolution of mode-converted surface wave images that involve cylindrically convergent surface waves. The general rules for broadband surface wave imaging may be summarized from the inspection of Figure 12. The entry circle defined by D e must be at least three surface wavelengths in diameter to time-resolve the surface wave arrival from the direct surface reflection. Therefore, De is typically nine times the -3-dB diameter of the focus at its center, and is approximately three times the -20-dB diameter (1.0 2e). Cracks, seams, and other linear discontinuities with little width but with lengths that exceed D R are imaged as having a width
5
305
Industrial Ultrasonic Imaging~Microscopy
that approaches that circular diameter. Features with dimensions smaller than DR produce detectable changes in amplitude only when they interact with the crossing zone of focus at the center. They will, however, produce some decrease in amplitude as soon as they become included in the entry circle. In specifying a surface wave scan, half of the -3-dB spot diameter, 0.16, should be used for the line-to-line spacing and the pulse-to-pulse spacing along the line. This will limit the amplitude ripple in the scanned acoustic field and support the -3-dB resolution in the image. For 50-Mhz surface wave images in most materials, this spacing is 0.02 mm or less. The difference in amplitude of the surface wave signals produced by the two water path distances ( Z 2 - ZA) and (Z2 - Z s ) in Figure 12 is inversely proportional to the distance each pulse travels across the surface of the tungsten carbide. A small amount of this amplitude loss is due to material attenuation, but most of the drop in amplitude is due to the continuous radiation of elastic energy, characteristic of leaky Rayleigh waves, back into the water during propagation. The received amplitude decreases proportionately with the length of the entry-exit path. Therefore, when path length D R is specified, this also specifies the energy lost back into the coupling fluid for the range of frequencies that make up the ultrasonic pulse. Equations (11) and (12) show that DR is specified by the ratio of the surface wave velocity to the longitudinal velocity in the coupling medium (usually water) and defocus distance Z. Written in terms of C2, CR, and Z, DR becomes DR - - 2 Z
C2
4C2-C 2
.
(15)
At constant Z, high values for CR result in smaller entry-exit paths (DR); lower CR, on the other hand, requires larger Rayleigh angles to generate the surface wave and result in longer entry-exit paths. When the surface wave velocity varies in the direction of travel in the plane of the surface, such as for images of anisotropic grains, then the travel time along each entry-exit path around the entry circle is different, i.e., the entry figure is no longer a circle. In this case the received amplitude results from the sum of all the arrivals back at the transducer in both amplitude and phase. In the case of an image of a polycrystalline material, if the maximum, minimum, and average velocities for each of the grains on the surface can be uniquely related to the orientation of the surface with respect to the crystal axes of the grain, then the received amplitude (and therefore the gray scale at which each grain is displayed) can be related to the crystallographic orientation. The
306
Robert S. Gilmore
interaction between anisotropic grains and single-frequency signals was first considered by Somekh et al. 23 Figures 13 and 14 show images of fully annealed titanium and copper alloys, specifically Ti83-A15-Sn2-Zr2-Mo4-Cr4 and Cu0.98 Ag0.02. Each titanium grain shows a consistent shade of gray and therefore a consistent
FIG. 13. 20-MHz surface wave image of a 100 mm • 100 mm titanium sample showing a fully annealed grain structure.
5
Industrial Ultrasonic Imaging~Microscopy
307
FIG. 14. 50-MHz surface wave image of a Cu0.98-Ag0.02 alloy sample 30 mm across. Note the intragranular structure.
amplitude. From grain to grain, however, the titanium signals vary over a range of 1 to 64 (6 bits). The Cu-Ag sample shows less grain-to-gain contrast (5 bits) but more intragranular variation. If the surface wave arrivals shown in Figure 12 were produced by a tone-burst generator instead of a halfwavelength impulse, then, depending on the length of the resulting pulse,
Robert S. Gilmore
308
the surface wave arrival would be overlapped by the direct reflection from the surface and the two would interfere. When the surface wave is delayed an even number of half-wavelengths, the interference is positive and the two signals add. When it is delayed an odd number of half wave-lengths, the interference is destructive and the two signals subtract, Raising and lowering the transducer increases and decreases the surface wave path length regardless of the shape of the entry figure or the anisotropy of the material being imaged (Fig. 15). The best contrast is achieved in a V(z) image when the transducer is placed at a height midway between a minimum and maximum where the slope of the received interference amplitude is greatest. Adjusting a signal gate to a position in time where the arrivals overlap and scanning at the described height sill produce the greatest net change in the contrast due to any velocity changes in the sample or flaws blocking the surface wave arrival.
C.
SUBSURFACEIMAGING OF VOLUMES
Following are the steps required to acquire and display an acoustic image of a subsurface region in a sample. 1. Select and mount the sample on the scanning stage. 2. Select a transducer (diameter, frequency, and focal length) on the basis of (a) the range in depth to be imaged, (b) the index of refraction between the sample and the fluid coupling medium, and (c) the resolution required. 3. Based on the beam diameter (resolution) produced by the transducer, select a pixel size and spacing (of no more than half that beam diameter), and load a scan plan into the scan controller. 4. Focus the transducer at the depth of interest (Z3)
where
C2 ( Z 2 - 12,1. Z3---~-t~3
as shown in Figure 16. Acquire data for the image in the form of signal amplitude, amplitude and phase, time of flight from a reference, or fully sampled waveforms, the latter of which can be processed to give all of the preceding information. 5. Select a color or gray-scale display, gain-adjust the dynamic range of the acquired/processed data to coincide with that of the display, and display the image.
(1)
,
(2)
/
iIIIII k /
\ \ \ \ \
(2) SIGNALS BROAD BAND SIGNALS GATE SURFACE WAVE AMPLITUDE
(1)
NARROW BAND SIGNALS
v(z)
(2)
A
/ ~ /~ (1)+(2) SIGNALS
'
INTERFERENCE AMPLITUDE
FIG. 15. Fundamentals of V(z) (surface wave) imaging. 309
310
R o b e r t S. Gilmore DT
Lens C1
DE
Water
c~
I
FOCAL GEOMETRY E7 IN SUBSTRATE (3) "-3
I I
I
and 3
%
FOCAL GEOMETRY IN WATER (2)
'l FIG. 16. Schematic showing the geometry for a diffraction limited beam, Z/d > C3/C2. The figure demonstrates that the beam in the solid refracts to the same diameter as in the fluid, although the depth of the focus is shortened.
5 Industrial Ultrasonic Imaging~Microscopy
311
Volume imaging also requires that an acoustic field be scanned to a uniform amplitude throughout the length, width, and depth of the inspected material. For this to be accomplished in an economical number of scans, the interrogating beam must have a depth of focus (Eq. (9)) that is, if not equal to, at least an appreciable fraction of the material depth. For many industrial parts this could be 1 to 10 cm or considerably greater. Typically the image is scanned to a uniformity of 1 dB, using the 1-dB beam diameter (Eq. (8)) to establish the pixel spacing. However, because it is not unusual for the dynamic range of the data acquisition to be 40+ dB (8 bits is 48 dB), for many industrial inspections the - 3 - d B diameter and depth of focus may be used. For industrial NDE, the flaw size to be detected is the key parameter that establishes the sensitivity for the nondestructive evaluation of a volume of material. This flaw (Fig. 16) must block a sufficient enough fraction of the focused beam that a detectable signal is reflected back to the transducer. The detectable size can be expressed as a fraction of the beam area, or as a fraction of the reflector size used to calibrate the system. Once detected, a flaw may be characterized with respect to its size and shape by imaging with a much higher resolution beam. For small flaws, there may be no economical solution. A beam diameter small enough to detect the flaws may require a considerable length of time to complete a scan of the material volume. In cases of substantial coherent acoustic noise from the focal zone, an alternative would be to specify the detectable flaw in terms of the ratio of its reflected amplitude to that of the noise. The detectable amplitude is usually given as two to five times the peak acoustic noise, depending on the permissible false alarm ratio. This amplitude can then be evaluated with respect to the detectable flaw size by a suitable calibration target. In this case fracture mechanics and probabilistic analysis must be used either to approve or disapprove the test with respect to its noise-limited detection capability. The first concern of an ultrasonic inspector, however, is to make sure that the critical flaw is detected. If a reflecting flaw is smaller than the focused beam, then the reflected amplitude is in proportion to the ratio of the area of the flaw to the area of the beam, Krautkramer and Krautkramer. 37 To a first approximation, the shape of the flaw cross section may be ignored if it is planar, perpendicular to, and totally contained in the beam area. Therefore, this analysis will discuss flat flaws that are located on the central axis of the interrogating beam at some time during the scan. One method of determining the beam diameter required for an inspection is to determine the flaw diameter required to totally block it. A circular reflecting flaw in the shape of a disk of diameter DU--perpendicular to and centered on
312
Robert S. Gilmore
the axis of symmetry of an acoustic beam focused at distance Z, in a material of longitudinal velocity 6'3 producing a wavelength 23, by a lens of diameter d m w i l l totally block the beam if the back-reflected beam half-angle from the disk is equal to the half-angle of convergence of the focused beam. Taking the sine as equal to the angle and setting the half-beam angle 14'37 for a flat circular reflecting flaw as equal to the half-angle of convergence of the focused beam gives 1.2323 _ dC 3 DU -- ~ 2zC 2
or
Of - 2.422 ~ Z.
(16)
Note that Eq. (16) has the same form as Eq. (8) with K -- 2.4. Although the insonificationmby the focused beam across the face of disk D / r e d o e s not have the uniformity that is assumed in the derivations, experimental measurements on flat-bottomed holes show that Eq. (12) does give a reasonable blocking diameter. Similar measurements show that smaller flaws are usually detectable and that for data acquisition systems with at least 32-dB dynamic range, flaws with reflecting areas equal to 1/20th of the beam area can be detected when they reflect echoes that are at least twice the peak acoustic noise in the material. With the critical flaw size assigned and expression obtained for the blocking flaw size (Du) and the diameter and depth of focus for the acoustic beam, a scanning plan can now be established for an inspection volume. Figure 17 shows a representative volume. This is a hot isostatically pressed and sintered sample with a slant layer of alumina spheres, 0.25 mm in diameter. The layer goes from the top surface on the left side of the sample to 12.7 mm depth over its 50-mm diameter, giving it an angle of approximately 15 ~ of arc with respect to the plane of that surface. The longitudinal velocity in the nickel-based super alloy is 6.1 mm/ms. The images shown in Figure 18 were made with three 50-MHz transducers focused at Z / d values of two, three, and four, respectively. These transducers produce progressively larger lateral (-3-dB) beam diameters of 0.75, 0.115, and 0.150 mm, respectively, and hence progressively poorer lateral resolution. They also produce greater (-3-dB) depths of field in proportion to the square of their Z / d ratio, giving diameters of 0.5, 1.0, and 2.0 mm, respectively. Even the Z / D = 4 transducer, however, can focus over only a fraction of the 12.7-mm depth over which the layer ranges. The image shown in Figure 19 was acquired with a 15-MHz, Z / d = 7 transducer producing a depth of field in the sample of 8 mm. This transducer is able to produce detectable signals for two-thirds of the entire depth of the
5
Industrial Ultrasonic Imaging~Microscopy
313
FIG. 17. The nickel-based superalloy sample shown has a slant layer of 250-1am(0.010-in.) alumina (A1203) spheres sintered into it. The layer increases in depth from the surface, upper left, to 12.7 mm (0.5 in.) deep lower right. The sample entry surface is 47 mm x 46 mm (-1.875 in. x 1.750 in.), the sample depth is 25.4 mm (1.0 in.).
slant layer. Note that this increase in detection depth has resulted in a substantial loss in lateral resolution. The beam diameter is now 0.6 mm, or approximately 2.4 times the diameter of the target spheres.
D.
MEASURINGRESOLUTION 1N SCANNED IMAGES
The resolution inherent to and obtainable from scanned acoustic images is basically determined by the beam diameter and by the scanning increment or pixel size. The effect of the pixel size on the image resolution is summarized by Nyquist's theorem, which states that in order to support the spatial resolution of the beam it must be spatially sampled at less than half that dimension. In order to support the spatial resolutions inherent to the beam diameter, the pixel size of the image must be one-half of the ultrasonic beam diameter. In other words, to support the - 3 - d B beam diameter resolution the pixels that make up the image must be less than half that size. As shown in Figure 10, subsurface foci, in high-velocity substrates always contain significant refractive aberration. In addition, the foci are also subject to micro-aberrations due to grain-to-grain anisotropy. The grain-to-grain anisotropy, however, is the mechanism that permits acoustic waves, and
314
Robert ?~'?!.",.%.'" .,..-.,
.................. + ,J .+ ~+ . +++
.
i~l + ? ? .........~~;:i...... ~,+i++
Z / d = 2.0, focused at 1.5 m m depth, E x = 0.075 mm E z = 0.5 mm.
~i.' :
-t"
'
- ++
:y
9
.~,...
+i,i~+ + ++: ++?~it
+
+"+'
+..
9.- ..:.
%
+
+:++
9 .... :
+
++:)!
+ S
i:!;
+. "% ';
+
+.
.,.+.+~
++
+i',+
o
.'./
+
:N
!??:
~i:
+ ,IV
+ " 9"+ ~ E+
+++~I
-'.
-
'+i
L i+i" +: + i:,i+<, ~,~++; + ~ :
' ,'+ ....
~ +'+' ,, : ~
l!r ................... ++~ +I++!++ +i~
- "+ "-
i+ ++ +"
"
. ++
,.,
++' +
~*
i
;<_
+. ++-. ~ - . - " - + _-.. _ _ , . "9%q, , , ".,, ~ .
,
+:
+ : ' ,,j :; ~, ++~+++ / " +
+ .+.-.
++
++:+
S. Gilmore
9
Z / d = 3.0, focused at 8.0 mm depth, Ex = 0.115 mm E+. = 1.0 mm.
.~ +,
....
+
~+|
+
:++ ;~a
i+i+ ....
9
-
+
~
"..,
+~ . .
+
+
.."
:i~+
.
++++
+%...: 9
++.s.-
~i+ .~ +" ~ ." +..... + ;~:+',+,:,+,~
.
9 :'"
+
+ ++
i .......
+.+. . . . . . . . . . .
+
:
:.."+ifi~
j~
i)~
,+,+ +++,+++++.i~,+++*+;~+
FIG. 18. The three image segments of nickel-based superalloy sample in Fig. 17 were made with the three 50-MHz transducers indicated. Eqs. (8) and (9) were used to determine diameters ( E x ) and depths (Ez) of foci.
5
Industrial Ultrasonic I m a g i n g ~ M i c r o s c o p y .....
/ ~
...... ~.., 9
...~
...... ~ .... ~,
.
:~!k;~.....
315
~.:.~~i!!~ I
Z/d
"4
~
~,
Ex =
Ez
focused at mm depth, 0 . 1 5 0 mm 2 . 0 mm.
= 4.0,
12.5
9
=
Q~O 9/
.~':~ '~.~: 9
9
~,~.~
..
:
,~ : , ~
...
, ,iN FIG.
(continued)
18.
,.o "g.'..m.~ . W
9 ~q
~
, ,,_. ~
~ ,r ~:~A".t ~
,w
9 .
~
to
IK"..;.,::
"r
#
-
#"
'" " ~ " . . . . / . ..I. "
~".~'.
. ?..~,
_.
.. "."
9
.... ~y"
.
9
FIG. 19. Superalloy sample imaged with a 15-MHz, F/D = 6 transducer focused at 8.0 mm depth. Eqs. (8) and (9) give Ex = 0.06 mm and Ez = 8.0 mm. Although the ability to separate and image the spheres is lower, they still produce easily detected signals.
316
Robert S. Gilmore
particularly surface waves, to image the microstructure. Surface wave images of micro structure are simply micro-abberation images. The most straightforward method to determine the resolution produced in a high-velocity solid is to scan a resolution target fabricated from the same or a velocity matched material. A number of resolution targets can provide this type of calibration information for an ultrasonic imaging system. Targets such as the U.S. Air Force 1951 target, (Fig. 20) were first used by Gilmore e t al. 14 to measure the resolution in acoustic images. It is also possible to machine flat-bottomed hole targets that provide this information (Fig. 21). Optically transparent resolution targets make it feasible to verify visually that the resolution target has been correctly fabricated. An image of the target with a candidate ultrasonic transducer then permits quantitative image resolution estimates to be made even when the interrogating acoustic beam contains significant refractive aberration. This is important because useful subsurface images can be acquired with the use of acoustic beams that are aberrated to the point that diffraction-limited beam width calculations are meaningless. Military Standard 15-A describes the basic pattern that has become known as the "USAF-1951" or simply the "Air Force" target. Each element consists of two patterns oriented at fight angles to each other, each containing three lines and two spaces. The line and spacing width are equal; the line length is equal to five times the line width. The change in pattern size is based on the sixth root of two, i.e., for every six target elements, the number of line spaces per millimeter doubles. Each six-target-element set is known as a group, and the group number (0, 1, 2, 3, and 4 for the targets used in this chapter) is the power of two, to which the first element in the group is raised to express the Number of Lines per Millimeter in USAF Resolving Power Test Target 1951 : E l e m e n t - 2 - 1 - 1 2 3 1 0.250 0.500 1.00 2.00 4.00 8.0 2 .280 .561 1.12 2.24 4.49 8.98 3 .315 .630 1.26 2.52 5.04 10.1 4 .353 .707 1.41 2.83 5.66 11.3 5 .397 .793 1.59 3.17 6.35 12.7 6 .445 .891 1.78 3.56 7.13 14.3
4 16.0 17.95 20.16 22.62 25.39 28.51
5 6 32.0 64,0 36.0 71.8 40.3 80.6 45.3 90.5 50.8 102. 57.0 114.
7 128.0 144.0 161. 181. 203. 228.
: i
i9
3 ~ :~ti:,~:~~ i ~ . U ."~ , H - ~ ~j ~'; ~' : !~/ ~ ~i i : . .~ 4 t ~ S ~ ! i ~ , . ~ ~ " n~ii~i:i~,;~~ e ~i ,~ _.....
Image format is 1/4 to 228 lines/mm target
FIG. 20. Table of sizes and image format for U.S. Air Force resolving power test target 1951 as provided by Teledyne-Gurley Inc. Troy, N.Y.
5
Industrial Ultrasonic Imaging~Microscopy
317
FIG. 21. Acousticimage of a block with multiple fiat-bottomed holes (FBH). There are four 3 x 3 arrays of holes spaced at the hole diameter, and a single hole at the block center. Each array has nine holes with diameters of 400, 800, 1200, and 1600 microns, respectively. The single hole is 400 microns. Note that the 400-micron holes are detected but not resolved (separated) by the 20 MHz, F/7.0 acoustic beam, which is 500 microns in diameter.
number of lines per millimeter. Therefore, the numbers 0 to 4 for these targets correspond to 1, 2, 4, 8, and 16 lines per millimeter, as shown in Figure 20. Fabricating the targets from transparent materials with similar acoustic longitudinal and shear velocities assures that the refractive aberration in the transparent target will be similar to that of the opaque materials under study. Sapphire can be used to match many structural ceramics, such as polycrystalline SiN, A1203, BeO, MgO, and SiC. Lead borosilicate glass produces targets that match the zirconium alloys and other lower-velocity materials in integrated circuit applications. Fused quartz is a good velocity match to most structural steels, including those used for high-temperature turbine engine disks and blades, and for titanium alloys. It is also a good velocity match to the oxide-loaded glasses used in computer chip carrier applications. The development of standards for establishing the sensitivity and resolution of all ultrasonic NDE inspection methods has proven to be an ongoing
318
Robert S. Gilmore
problem for the testing industry. The traditional sensitivity standard is a flatbottomed hole (FBH), first proposed by the Krautkramers (1959). 37 Drilled into materials that are identical in composition and in microstructure to those being inspected, the FBH provides a reflector of known circular (and therefore mathematically definable) scattering cross section. Since, as a first-order approximation, pulse-echo amplitude is linearly proportional to the area of a compact flat reflecting void perpendicular to the acoustic axis of the interrogating beam, the FBH area establishes the reflecting area that can be detected, and hence the test sensitivity. However, an ultrasonic image of a single hole can only show that it was detected; no resolution information is supplied. Another drawback is that the circular bottom of the hole is always accompanied by a cylindrical shaft connecting it to the drilled surface, and therefore it is surface-connected and only approximates the buried circular void/crack that it is intended to represent. For materials with scattering microstructures (i.e., materials with grains, or more than one material, such as composites), the echoes from the bottoms of small holes add algebraically to the grain boundary reflections, producing significant echo amplitude variations from holes of precisely the same size. Therefore, single-reflector calibration blocks can overcalibrate or undercalibrate, depending on the location of the FBH in the material microstructure. To meet these problems and to try to solve the sensitivity/ resolution dilemma, standards have been developed with multiple hole patterns (Fig. 21). These hole patterns can determine the variance in amplitudeand resolution produced by scattering materials and are easy to make for materials that can be drilled, such as titanium, steel, and zirconium. However, ceramics are difficult to drill, and holes only a few microns in diameter m s u c h as are required to calibrate inspections for small voids in structural ceramics--are especially difficult. Lithography has furnished a new tool for studying and for creating subsurface void arrays with very precise geometries. These arrays may be produced in high-temperature structural ceramics, in many glasses, and at ceramic-metal interfaces. The combination of photolithographic methods with ion beam etching and hot pressing provides the ability first to produce surface features with highly precise geometries and locations and then to transform these features into internal features without losing this detail. The methods summarized in Figure 22 are reported in greater detail elsewhere. 43,44. Figure 23(a) is a magnified image of the target shown in Figure 20. This target contains Groups 0 through 4. The 0.025-mm beam and pixels resolve
5 Industrial Ultrasonic Imaging~Microscopy
319
Sample Preparation 9..:.~:::::~:~:~.~::~::.. ~::..:.,:.::.,:~:..:.~:~:: ....~::..::.:..:::.:...:::,: .:.:::::::.::,.,:,.,:,:,:,.:: :,:~:::,::. :...,::,: ::::, :;:,;: .,..,:: :....:.~.~.::... ~.:;.: ~.::: ..: :: ::: :....:.::::-.:
:.....::..: ::::. ~:..:.. :: ..: ::.....:.:.~:..,...::,,
N~!i~;N~NN~4~2~:!~i~:~:!~i~!`.`.!~:`~!i`~!:!~i!~.i!~!~i!Ni~!!i!:Nii!!i~;i~ii~i~i!ii~!`i;~ii~i~;!~i~i~ii~;)i~i4i;i
-~
selectively expose to visible or uv radiation
Positive photoresist
Substrate
~
,
~
Mask Selectively exposed photoresist Substrate ,onbeametch
~4~;~
Patterned photoresist
~,~,~:~N~+ ~:~;:~:~:~:,~.~
I~
Substrate
i.............l..~....~........ MicrodesignedResi photoresi dualStetchedsurface
.I: i
:: i i
!~i~:::,;:~i~i!i;:ii~ ....:,i .... ii~ ....~ ~ ~:2, :~ ; . I . t: .: . . . . . . t f ....... :
I
/ _. -"
Unetched substrate Microdesigned internal defect structure Etched substrate
FIG. 22. Schematic illustration of the procedure used to produce the subsurface target pattern in single-crystal sapphire blanks 1.5-mm and 3.0-mm thick. (a) Coated substrate. (b) Selective exposure through a USAF 1951 mask. (c) Ion beam etching of substrate and patterned photoresist. (d) Etched photoresist and sapphire. (e) Bonding of etched and unetched sapphire blanks. (Courtesy of Rodel and Glaeser 43)
the largest bar-space patterns in Group 4 (16 lines/mm), but the smallest patterns in the group (28.51 lines/mm) are not resolved. This observation provides one more verification of the Nyquist theorem applied to spatial resolution. Stated simply: in order to resolve dimension d, image data must be acquired at a sampling interval less than or approaching d/2. Applying the Nyquist theorem to the 25-gm acoustic beam, used to acquire the data file for
.......
90 . ~ , ~ ~. . . . . . . . ..+~ ~ . ,
~, ~~.,~ ',2 ~ Y:.o ., .o~ ~.';..;, -, ~; ~.~:'~. ;~.~i,,~,~,,,.".,'..;,>~,~;'~./?..: ~,; ~;~,.'~', ~,,~..~,.~,~,,~.~,,'r ~,..... .,~,~ ~.~ ..... ~...~,~, ,, ..~,~: .
" ~ ~ o:,i. ~'~..'~.~,~~~:'~,~.("~ i.~~:,~ ~ ~ .~'~,q"~ ~.:~~~ ~'.~' , 9
,. ,-~,:;.g' ~ ~
..
'
,
9
~ , ~ ~
! ,
~.
~ ~
~ (a)
%..: ,..
~...
:~ . . . . .
.. ~.
.~
.. " . ~ . , . .
~
,
~.
(b) FIG. 23. (a) Magnified image of Groups 2, 3, and 4 as described in Fig. 20. The 25-l.tm pixels are now visible, as is the loss of resolution as the line pairs approach 32 lines/mm. (b) Image of the same area but undersampled in two directions by a factor of 2. The image now displays Group 3 patterns with the same relative resolution shown for Group 4. 320
5
Industrial Ultrasonic Imaging~Microscopy
321
Figures 23(a) and 23(b), indicates that the micron pixels do not support the 25-lam resolution provided by the beam. To support the beam the pixel size/spacing should be less than 12.5 gm. Figure 23(b) shows the data file in Fig. 23(a) undersampled by two in each direction, increasing the pixel size to 50 gm. Note that it is now not possible to resolve even the largest pattern in Group 4 of the resolution target. However, looking at the next largest group, it can be observed that patterns in Group 3 (upper fight in the Fig. 23(b)) show the same resolution as for Group 4 in Figure 23(a). Since the spatial resolution due to pixel size/spacing is decreased by a factor of two, the displayed decrease in resolution is expected. E.
IMAGEPROCESSING
Clearly, like most other imaging techniques, ultrasonic images are degraded by blurring due to the practical limitations imposed by the beam diameter and the pixel size. The purpose of this section is to demonstrate that digital image processing methods may be used to provide improvements of as much as a factor of two in resolution, 46'47 providing that the spatial resolution at which the original data was acquired supports the final/improved resolution in the processed image. The details for digital image processing are extensive. 48'49 This discussion is limited to Wiener filtering. Figure 24(a) shows a 1024 x 1024 longitudinal wave image of the 25.4 mm • 25.4 mm sapphire target described above. The 50-MHz, F/3.0 transducer used to acquire the image used a 6.37-mm diameter lens, focusing the beam at 19.05 mm in water. This produces half-angle of incidence of 9.5 ~ well beyond the 7.6 ~ critical angle of incidence for longitudinal waves in water incident on a C-axis sapphire substrate (Table 1). The longitudinal focus produced in sapphire is similar to that shown for steel in Figure 10 (between 0 and 14.5 degrees). Clearly aberration, not diffraction, controls the resolution in this image. For the sapphire target used to produce the acoustic images in Figure 24, vacuum hot pressing (1370~ 15 MPa, 2.6 milliPascals) was used to bond a 1.5-mm-thick unetched single crystal [001] sapphire slab to a 3.0-mm-thick etched slab of the same [001] orientation. This produced the USAF-1951 target embedded in a volume of single-crystal sapphire. The target contains bar-space patterns 0, 1, 2, 3, and 4, as described in Figure 20 with an entrysurface-to-target-pattern depth of either 2.0 or 4.0 mm. Referring to Figure 20, these range from 1 to 28.5 lines per millimeter. Note that in addition to
322
Robert S. Gilmore
(a)
(b)
(c) FIG. 24. (a) 50-MHz, F/3.0 image of the sapphire resolution target taken from the 1.5-mm surface-to-pattern depth. The flaws evident in the image were actually very useful. One of the smallest was used to define the point spread beam function used to provide the image enhancement shown in (c). (b) Magnified image of the central portion of (a). (c) Wiener filter enhancement of (b) showing approximately a factor of 2 improvement in resolution. providing the geometric precision required for image calibration, these void arrays are truly buried flaws surrounded by intact solid material. They scatter sound like the subsurface cracks they are. A display of the raw data, as shown in Figure 24, is able to resolve only the first two patterns in Group 2 (a little better than 4.5 lines/mm or 220 lam). The resolution inherent to the high spatial frequency of the image (25 lam pixel size), however, can still be utilized. Figures 24(b) and 24(c) zoom in on the outlined portion of Figure 24(a) and then enhance the resolution using a Wiener filter image processing technique. 45'46 Note that in Figure 24(c) the patterns resolved in Group 3 suggest that the processed image shows twice the
5
Industrial Ultrasonic Imaging~Microscopy
323
resolution (is able to resolve line-space combinations one-half those) of the unprocessed image.
VI.
Applications
In considering the applications for acoustic microscopy, it is important to remember that any industrial ultrasonic inspection system capable of producing magnified C-Scan images can be used as an acoustic microscope. Therefore, although a complete applications summary of ultrasonic imaging/microscopy is beyond the scope of this chapter, Figure 25 indicates what that range in scope could be. Since resolution in an ultrasonic image is dependent on both frequency and focal convergence, it is difficult to separate applications as a function of frequency without some overlap. If most industrial ultrasonic systems can be used as ultrasonic microscopes, it follows that most ultrasonic microscopes can be used to inspect parts, providing the parts are small enough to be scanned. Although few ultrasonic microscopes are constructed large enough to inspect large industrial parts, many are large enough to do inspection development for heavy sections. In these studies the transducer focus and bandwidth are selected for the inspection, as well as the scan index. In addition, the calibration standards for the inspection may be qualified more precisely than would be possible with an industrial inspection system. The broad area applications chosen for this discussion consist of: 9 Evaluation and qualification of calibration blocks at higher resolution and frequency than the inspection frequency 9 Inspection development studies in metals and composites 9 High-resolution/frequency characterization of flaws detected by lowerfrequency inspections 9 Evaluation of attachment integrity of welds and composite structures, as a process development and process control tool as well as a quality assurance tool 9 Inspection of electronic devices and materials Most industrial inspection and medical diagnostic imaging are done at ultrasonic frequencies from 1.0 to 10.0 MHz. At this frequency range, metal and composite sections multiple inches in thickness may b e penetrated, and equivalent depths of penetration are achieved in medical imaging. Almost all in vivo medical imaging is done in this frequency range.
Robert S. Gilmore
324
Large forgings, welds and castings, nuclear pipe welds, pressure vessels Medical: mammography, heart, fetal scans
Industrial NDE: large parts, coarse grained metals, composites of 6.0 mm to 0.5 mm resolution images
Power generation turbine parts
Medical: fragments in eyes, eye dimensions, blood vessel walls, burns
Industrial NDE: flaw detection, quality control, process evaluation and control, materials characterization of 3.0ram to 0.1 mm resolution images
Spot welds, braze inspection, AgCdO electrical contacts
Typical PAM SEAM, and SLAM operating ranges
Biomedical studies of cell structure
High frequency industrial NDE, electronic components, ceramics, attachment integrity of 0.3 mm to 0.00~ mm resolution images
Classic surface wave acoustic microscopy of 0.005-ram to 0.0005-ram resolution
Structural aircraft composites Aircraft rotating parts, railway, locomotive parts, automotive parts
Plastic packaging for chip carriers Solid state bond attachment (diffusion) Thick film adhesion paint, electrolytic coatings Porosity, cracks structural ceramics diamond compacts ceramic Capacitors ceramic IC packages semiconductor heat-sink bonds, laser welds IC die attachment Porosity, cracks fine grain ceramics CVD diamond films thin film adhesion IC line structures substrate cracks VLSl line structure micro-flaws, cracks, fine grain material characterization
z/ FIG. 25.
Applications of ultrasonic imaging/microscopy versus frequency.
Calibration blocks are the key to a successful and well-designed inspection; therefore, they should be carefully evaluated before using them. Figure 26(a) shows a reactor pressure vessel calibration block. The upper surface of the block is clad by 0.3 in. (8.0 mm) of roll-bonded stainless steel cladding, and four 0.062-in. (1.5-mm) side-drilled holes are located at three positions in the clad and just below the cladding-to-base-metal interface. Figures 26(b) and
5
Industrial Ultrasonic Imaging~Microscopy
325
FIG. 26. 2.25- and 5.0-MHz images of a 2.0 in. x 6.0 in. x 3.3 in. (50.8 x 152.4 x 83.8 mm) reactor pressure vessel calibration block. The 5.0-MHz and 2.25-MHz beam diameters are 1.2 mm and 3.1 m m (0.048 in. and 0.125 in.), respectively. The 0.62-in. (1.5-mm) side-drilled holes are easily detected in both images.
326
Robert S. Gilmore
26(c) show 2.25-MHz and 5.0-MHz pulse-echo images of the cladding and clad-to-steel interface. The inspection frequency, 2.25 MHz, clearly shows all four flaws but shows large edge effects at the block ends. The higher resolution provided at 5.0 MHz gives much more detail at the cladding interface and is better for characterizing and qualifying the calibration block. Both images can be acquired by most industrial C-Scan inspection systems. Positioning the focus for successful flaw detection and determining the effect of the material on flaw signals are important for inspection development. Figure 27 shows three transducer scanning depths with respect to an array of sixteen 0.01565-in. (0.4-mm) flat-bottomed holes at 2.0-in. (50.8-mm) depth in forged Ti-6A1-4V. The titanium block was machined from a large forged ring. The transducer is 5.0 MHz, focused to a - 6 - d B beam of 0.096 in. (2.4 mm). For this demonstration the amplitudes reflected by the holes are plotted vertically rather than as a C-Scan image. Note that as the focal zone of the transducer approaches the same depth as the holes, the signal-to-noise ratio of each of the hole echoes becomes much greater, even though the absolute amplitude is the same range (0 to 200 arbitrary units) for all three scan depths. When the focal zone is in the same plane as the flaws, the enhanced signal-to-noise is produced by decreasing the noise with respect to the hole echoes rather than enhancing the hole echoes. The high microstructure noise in the first two plots clearly shows the curvature of the ring forging. Selecting the correct transducer, frequency, and focus for composite inspection requires careful studies of the flaws that would be expected during manufacture. Figure 28 shows four 10-MHz pulse-echo images of four composite samples, each with dimensions of 4.0 in. x 4.0 in. (101.6 mm x 101.6 mm). Three of the samples have intentionally seeded flaws. Each sample consists of two 0.025-in. (0.635-mm) woven plies of Kevlar, five plies of unidirectional carbon (0.005 in./0.127 mm each), and then an additional two plies of Kevlar (same thickness). The total composite thickness is 0.125 in. (3.1 mm). By gating each depth in the sample, the structural segments may be imaged separately. Figure 28(a) shows the first 0.050 in. (1.27 mm) of Kevlar to be without flaws, (b) shows a delamination between the second and third carbon plies, (c) shows a backing paper flaw between the fourth and fifth carbon plies, and (d) shows a Teflon-induced delamination between the bottom two Kevlar plies. Note that even though the 10-MHz image has a depth of focus that is greater than the composite thickness, the images do degrade in resolution because of attenuation of the higher frequencies in the broadband signal. A number of commercial systems
5
Industrial Ultrasonic Imaging/Microscopy
327
2001 Focus at Surface /
150 r
Focus on Surface
lilLi
,i.+
g.iI
'~t,'
I
2.0"
Block nnflrl ,i- _.--..-..fin k'-4 sets of 16 FBH (#1, #'2,#3, #4) Spaced at .5"
I
.5"
200 Focus at 1.3"
Focus
at 1.3"
150
f
loo
:
l,
11+5"
~ ,
I
i ,, ,,!tJ . i~ i,,
% .5"
!
"1
iiil+cusa'20
Focus at 2.0"
~ i
2.0"
rl nnnn~n-TT-nrtn
Focal Positions for 5 MHz F/8.0 Transducers
/+,,u0e
100~
i
g!;ti
!,i!! ~'0
J, D~sXance
FIG. 27. Positioning the focus for successful flaw detection. Three scanning depths with respect to an array of sixteen 0.01565-in. (0.4-mm) flat-bottomed holes at 2.0 in. (50.8 mm) depth in ring-forged Ti-6A1-4V. The 5.0-MHz beam is 0.096 in. (2.4 mm). Ultrasonic amplitude is plotted vertically. Note improved SNR as the focal zone approaches the same depth as the holes. Horizontal distances are in mm.
now marketed permit up to eight images to be acquired simultaneously at increasing depths. Figure 29 shows 5-MHz pulse-transmission images o f the two six-ply carbon-fiber-reinforced epoxy samples. Sample (a) is good; sample (b) shows
Robert S. Gilmore
328
(a)
(b) FIG. 28. Four 10-MHz pulse-echo images of four 100 mm x 100 mm composite samples. Each sample is constructed with two 0.025-in. (0.635-mm) woven plies of Kevlar, five plies of unidirectional carbon (0.005 in./0.127 mm each), and an additional two plies of Kevlar. A good structure is shown in photo (a), Teflon delaminations are shown in photos (b) and (d), and backing paper is shown in photo (c). Beam diameter is 0.75 mm (0.030 in.), 512 x 512 displays.
b
Industrial Ultrasonic Imaging~Microscopy
(c~
(d) FIG. 28.
(continued)
329
Robert S. Gilmore
330 (a)
z:
(b)
FIG. 29. 5-MHz pulse-transmission images are shown of two 100 mm x 100 mm areas in a six-ply carbon-epoxy composite. The good sample shows uniform transmission (a); the dark region in (b) is due to porosity. Beam diameter 1.2 mm (0.48 in.), pixel size 0.2 mm (0.008 in), 512 x 512 display.
5
Industrial Ultrasonic Imaging~MicrOscopy
331
(a) !! ,~:,,..,, . . . . .
,..,..
.
~,%. ,~. ..
:i:: .., ? ~..~ ~: ~. .:i~ ..;
r
) .... ~:.~. ~:~. ~9 . . ~ . ~..:
~.i~, 9 2.j, .. .~:~.
,
~,'~ ... ~'. ~:~,~ .. g~.
~::~:
~,, ~
~..
9
.
(b)
FIG. 30. Silicate flaw in a 1.2-in. (31.2-mm) circular trepan from a Ni-Cr-Mo-V turbine rotor forging. (a) The 15-MHz beam diameter is 0.024 in. (0.6 mm); (b) The 50-MHz beam diameter is 0.010 in. (0.25 mm). The cross is a cursor on the CRT display, 512 x 512 images. (Courtesy of L. B. Burnet, General Electric Company).
332
Robert S. Gilmore
considerable porosity, which attenuates the 5-MHz amplitude to produce a darker image. As the frequency increases above 10 MHz, resolution increases but attenuation increases also. In the range from 10 to 50 MHz, medical applications are pretty much limited to in vitro (excised tissue) studies except for measurements on the eye and studies of the structure of bums. However, there are many industrial applications in this frequency range. Consider the role of ultrasonic microscopy in evaluating flaws trepanned from large steam turbine rotor forgings. Figure 30 shows (a) 15-MHz and (b) 50-MHz images of a flaw that was trepanned from a Ni-Cr-Mo-V medium alloy turbine rotor. The flaw was detected by a visual pulse-echo inspection conducted at 2.25 MHz. Serial sectioning of the largest segment of the flaw shows a silicate inclusion. The composition of the inclusion suggested that it resulted from the liner of either the furnace or the ladle when the steel was poured. Note that the 50-MHz image shows an improvement in both resolution and detection. Figure 31 shows a 50-MHz image of a (electrical resistance heated) solidstate weld in a Ti-6A1-4V alloy. The ultrasonic beam diameter is 0.0024 in. (0.063 mm). The surfaces to be attached were deliberately contaminated by an artificial fingerprint (made with lanolin and NaC1). Note that in the vicinity of this artificial flaw the microstructure shows a much narrower heat-affected zone. This is to be expected since the contamination also decreased the electrical conductivity at the weld interface. Porosity in BaTiO3 capacitor blanks is both common and acceptable, as indicated by the four 12.7 mm (0.5 in.) blanks displayed in the 50-MHz image in Figure 32(a). The upper left blank, however, shows some connec-
%~ :,,
. ,,
~
~
~
FIG. 31. 50-MHz image of an electrical resistance-heated solid-state weld in a titanium alloy. The ultrasonic beam diameter is 0.0024 in. (0.063 mm). The surfaces to be attached were deliberately contaminated by an artificial fingerprint (made with lanolin and NaC1) (1024 x 512). (Weld is 50 mm wide, imaged at 1024 x 256 pixels) (Courtesy of R. Sundell, General Electric Company).
5
Industrial Ultrasonic Imaging~Microscopy
333
(a)
~@~ ~ ~'~
~.............. ~~!i'~':: i~!!~:,~!~"~ ~:,'~i:;~;~~,"
(b) FIG. 32. (a) Four 12.7-mm diameter (0.5 in.) BaTiO3 capacitor blanks imaged at 50 MHz. In (b) and (c), the lower-left blank is zoomed to show high detail, 20-gm beam, and pixel diameter: (b) 1024 x 1024 image, (c) 256 x 256 image.
334
Robert S. Gilmore
(c) FIG. 32.
(continued)
tivity between the pores, and that may lead to cracking. Analysis of the zoomed images shown in (b) and (c) suggest that the pores are all smaller than 100 lam (0.004 in.). A number of commercial image processing programs can pan and zoom across large images. An example of this capability is shown in Figure 33. Here a 2048 x 2048 image is acquired of the reverse face of a 1973 Canadian penny by monitoring the amplitude of a ultrasonic beam focused directly on the surface. The ultrasonic beam used to acquire these 50MHz images is approximately 20 lam (0.0008 in.) in diameter, and the image is acquired with a pixel size of 10 lam. This very high resolution image can then be zoomed for an additional magnification of as much as 8 x, as shown in the displays in Figures 33(b), 33(c), and 33(d). The modem U.S. penny is made with copper-plated zinc. When such a coin is struck, the zinc body sustains many microfractures such as shown in the 50-MHz image in Figure 34(b). This deformation pattern was imaged by first lapping the raised coin surface flat and then focusing the 20-jam beam just below the surface. Electronic devices and assemblies have been one of the main benefactors of the development of high-frequency ultrasonic imaging. Figure 35 shows the
5
Industrial Ultrasonic Imaging~Microscopy
..
335
i~i~
(a)
(b) FIG. 33. Surface of a Canadian penny imaged with a 20-gm diameter 50-MHz beam. The 2048 x 2048 image file permits a series of zoomed images to give highly magnified surface detail, 10-gm pixels.
Robert S. Gilmore
336
(c) .~,~,,~.~,..,~,....~,: ..: ....
i~, ~ B 9
.
,
.~
(d) FIG. 33.
(continued)
," :..~.,%., ,: ,,
5
Industrial Ultrasonic Imaging~Microscopy
337
(a)
..,,,~iiiiiiiiiii!! .. i!ii ~,~,,,......... ~,,,, iiiiii 84184i 84 . .
FIG. 34. Deformation processes are clearly displayed by first lapping a modern copperplated zinc coin flat and then imaging the subsurface microstructure at 50 MHz. Beam diameter and pixel diameter are both 20 gm, image is 1024 x 1024.
complex metal circuitry in and at the surface of a ceramic chip carrier. The pulse-echo image was acquired at 50 MHz with a 20-gm beam diameter. The larger black features at the center are irregularities in the manufacture of the device. Many silicon-controlled power devices require wafers several inches in diameter such as shown in Figure 36(a). Figure 36(b) shows a 50-MHz image of a badly cracked guard ring under an uncracked 0.5-mm [ 100] silicon entry surface. The image also shows the finger structure of the 37.5-mm-diameter thyristor gate array.
338
Robert S. Gilmore
FIG. 35. Complex metal circuitry in and at the surface of a ceramic chip carrier. 50-MHz direct reflection 1024 x 1024 image, 20-1am beam diameter, 10 g pixels. The larger black features at the center are irregularities in the manufacture of the device.
Figure 36(c) shows a surface wave image of a [100] silicon wafer fragment that contained several small surface flaws. Note the complex surface wave structure (Fig. 36(d)) around such a defect, caused by the anisotropy of the silicon wafer. Most of the images displayed so far have been pulse-echo and acquired with mechanically scanned systems. Figure 37 shows a pulse transmission image acquired with a 100'MHz stage in a scanning laser acoustic microscope. The image shows metalized ribbon leads on an A1203 ceramic substrate. The bonded areas (one is circled) are 125 x 125 microns. Both metals and electronic assemblies can be studied in the GHz frequency range. Figure 38 shows three images of a small Inconel sample. Figures 38(a) and 38(c) show an optical and an acoustic image of an unetched surface. The 2.7-GHz surface wave acoustic image has a -6-dB beam diameter of 2 gm. By comparing features in the acoustic image of Figure 38(c) and etched optical image of Figure 38(b), it is apparent that both display the very fine grain structure in the Inconel. Figure 38 was acquired with the mechanically scanned V(z) surface wave imaging technique.
5
Industrial Ultrasonic Imaging~Microscopy
339
(a)
....
,,,,
~,~ ~/t~i~:m (b) FIG. 36. (a) Silicon wafer, 1.5-in. (38-mm) diameter, 0.020-in. (0.5-mm) thick. (b) 50-MHz image of a badly cracked guard ring under the uncracked silicon entry surface in (a), also the 37.5-mm diameter thyristor gate array in that device. (c) 50-MHz surface wave image of a 25mm [100] silicon wafer fragment containing surface flaws. (d) A zoomed image of surface wave structure around the defect caused by [ 100] anisotropy in silicon.
340
Robert S. Gilmore
(c)
(d) FIG. 36.
(continued)
The detail at which an integrated circuit can be imaged at 2.7 GHz is shown by imaging the structure of an FET transistor (Fig. 39). Note that the resolution of this image is indeed very close to that of the best optical microscopes.
5
Industrial Ultrasonic Imaging~Microscopy
9~.%.,:,,~::~.:
.
.
.
. . . . .
....
.
.
"
1:
"
.
...:.
:..
::.:.:.,~:~::.::...;,.::~::;~.~..~,,
.~
.... i.::i::~.;:!::~;ii::?@;~ -:....:::{::..:-:~.~::~:~{@ ......~..~i~}i:~:~{..................... .:..:;.....:. . ...
;i!:.!}i:i..... ~ ~
341
.....
~{:;;'~' ~
(a)
0.5 mm
FIG. 37. 100-MHz pulse transmission image acquired with a scanning laser acoustic microscope, showing metalized ribbon leads on an A1203 ceramic substrate. The bonded areas (circled) are 125 ~tm2. (Courtesy L. Kessler, Sonoscan).
342
Robert S. Gilmore
'
~
~.~.".~,~,,:;~",~~,
,,~. ,,, ~
,.?:,,i:
o, . . . .
,
,
,,
9,r~.;,~.~ff~ ,~
(a)
(b)
(c) 50pro FIG. 38. Three images of a small lnconel sample: (a) and (c) are an optical and an acoustic image of the unetched surface, (b) is an optical image of the surface etched. The 2.7-GHz, V(z) surface wave acoustic image has a - 6-dB beam diameter of 2 ~m. (b) and (c) principally show grain structure. (Courtesy of C. Quate, Stanford University).
5
343
Industrial Ultrasonic Imaging~Microscopy
(a)
I~"--~ 4O ~n
i0 IJm
FIG. 39. 2.7-GHz V(z) image of the structure of an FET transistor. Note that this image resolution is very close to the best optical microscopes. (Courtesy of C. Quate, Standford University).
VII.
Conclusions and Future Work
Several definitive acoustic microscopy and acoustic imaging works have been referenced that have very complete bibliographies. Lemons and Quate, ~8 Bennet, 25 Kessler, 26 Khuri-Yakub, 27 and Briggs 28 are five. Briggs has the most recent and complete bibliography but has written from the aspect of materials characterization rather than nondestructive testing. Lemons and Quate and the papers edited by Bennett summarize the development of the instrumentation and techniques as of 1979 and 1985, respectively. All of these reviews assume that the reader is well versed in the details of acquiring and displaying mechanically scanned images. This chapter has used these details
R o b e r t S. Gilmore
344
as an outline to organize the presentations for surface and subsurface imaging and is written specifically from the aspect of nondestructive inspection. The Acoustic Imaging Symposia, currently at Volume 24, 49 provide an informative annual review of the current state of the art on ultrasonic/acoustic imaging. In addition, no discussion of ultrasonics or acoustics would be well referenced without a careful review of the Physical Acoustics series, 41 many of whose volumes are edited by Warren P. Mason. Future progress in ultrasonic/acoustic imaging and microscopy will almost certainly be driven by the dramatic growth in computers (with respect to both availability and power) for image acquisition, processing, and analysis. Image data based on waveform capture is memory-intensive, but permits the operators developed for ultrasonic spectroscopy to be used in analysis of the data as well as the image processing operators such as Wiener filters and fast Fourier transforms. It seems reasonable to suggest that all of the analytic procedures for V(z), as published by Weglein and Wilson, 16'17 Liang et al., 19 Kushibiki and Chubachi, 21 and Somek et al., 23 have corresponding operators that can be developed to extract the amplitude-frequency information in broadband ultrasonic images. Applicable work in ultrasonic spectral analysis, such as summarized in Fitting and Adler, 42 can also be applied to extract the multifrequency information from broadband ultrasonic/acoustic images. The lately reemerging air-coupled acoustic imaging methods such as originated by Wickramasinghe, 5~ developed by Fortunko 5~'52 and still later developed by Bond and others 53 promises to provide in the forthcoming decade a considerable body of information on plastics and other low-velocity materials. Low-velocity fluids such as the Freons also promise to provide similar opportunities. ACKNOWLEDGMENTS
The author wishes to acknowledge M. Gigliotti, P. Howard, R. Klaassen, K. Mitchell, L. Perocchi, R. Trzaskos, E. Nieters, J. Young, and most of the authors in the Reference section for many helpful discussions over the years. REFERENCES 1. Bergmann, L. (1938). "Ultrasonics (and Their Scientific and Technical Applications)." John Wiley & Sons, New York. 2. Carlin, B. (1949). "Ultrasonics." (2nd Edition, 1960). McGraw-Hill, New York. 3. McMaster, R. C. (1959). "Nondestructive Testing Handbook," Vol. II. Ronald Press, New York. 4. Sokolov, S. (1936), USSR Patent No. 49.
5
Industrial
Ultrasonic Imaging~Microscopy
345
5. Dunn, E, and Fry, W. (1959). Ultrasonic absorption microscope. J. Acoust. Soc. Am. 31, 632-633. 6. Jacobs, J., Berger, H., and Collis, W. (1963). Investigation of the limitations to the maximum attainable sensitivity in acoustical image converters. IEEE Trans. Ultras. 10(2), 83-88. 7. Korpel, A., Adler, R., Desmares, P., and Watson, W. (1966). A television display using acoustic deflection and modulation of coherent light. Appl. Opt. 5, 1667-1675. 8. El-Sum, H., and Larmore, L. (1967). "Proceedings of the First International Symposium on Acoustic Holography" (later "Acoustic Imaging") (A. Metherell, ed.). Plenum Press, New York. 9. Panametrics Inc., 221 Crescent St., Waltham, MA, 01754. (1971). 10. Lemons, R. A., and Quate, C. E (1973). Acoustic microscopy by mechanical scanning. Appl. Phys. Lett. 24, 165-167. 11. Sonoscan Inc., 530 Green Street, Bensen, IL 60106. 12. Tsai, C. S., Wang, S.K., and Lee, C. C. (1977). Visualization of solid material joints using a transmission acoustic microscope. Appl. Phys. Lett. 31,791-793. 13. Buchanan, R. W., and Hastings, C. H. (1955). Ultrasonic flaw plotting equipment: a new concept in industrial inspection. Nondest. Test. 13(5), 17-25. 14. Gilmore, R. S., Tam, K. C., Young, J. D., and Howard, D. R. (1986). Acoustic microscopy from 10 to 100 MHz for industrial applications. Philosophical Trans. Roy. Soc. A320, 215-235. 15. Nikoonanhad, M., Yue, G., and Ash, E. (1985). IEEE Trans. Son. & Ultrason., Special Issue on Acoustic Microscopy SU-32(2), 130-375. 16. Weglein, R. D., and Wilson, R. G. (1978). Characteristic materials signatures by acoustic microscopy. Elect. Lett. 14, 3562-354. 17. Weglein, R. D. (1985). Acoustic micro-metrology. 1EEE Trans. Son. R Ultrason, SU-32(2), 225235. 18. Lemons, R. A., and Quate, C. E (1979). Acoustic microscopy. In "Physical Acoustics," Vol. 14 (W. P. Mason and R. N. Thurston, eds.). Academic Press, New York, pp. 2-90. 19. Liang, K. K., Kino, G. S., and Kuri-Yakub, B. T. (1986). Material characterization by the inversion of V(z). IEEE Trans. Son. & Ultrason. SU-32(2), 213-224. 20. Bertoni, H. L. (1984). Ray-0ptical evaluation of V(z) in the reflection acoustic microscope. IEEE Trans. Son & Ultrason. SU-31(2), 105-116. 21. Kushibiki, J., and Chubachi, N. (1985). Material characterization by the line-focus-beam acoustic microscope. IEEE Trans. Son. & Ultrason., Special Issue on Acoustic Microscopy SU-32(2), 130375. 22. Wichramasinghe, K. (1979). Contrast and imaging performance in the scanning acoustic microscope. J. Appl. Phys. 50(2), 664-672. 23. Somekh, M., Briggs, G. A. D., and Ilett, C. (1984). The effect of elastic anisotropy on contrast in the scanning acoustic microscope. Philosophical Mag. 49A, 179-204. 24. Ash, E. (Ed.) (1980). "Scanned Image Microscopy." Academic Press, New York. 25. Bennett, S. (Ed.) (1985). IEEE Trans. Son. & Ultrason., Special Issue on Acoustic Microscopy SU32(2), 130-375. 26. Kessler, L. (1989). Acoustic microscopy. In "Nondestructive Evaluation and Quality Control Metals Handbook." ASM International, Materials Park, Ohio. 27. Khuri-Yakub, B. (1993). Scanning acoustic microscopy. Ultrasonics 31(5), 361-372. 28. Briggs, A. (1992). "Acoustic Microscopy." Oxford University Press, New York. 29. Martins, Y., and Ash, E. (1982). Photodisplacement microscopy, using a semiconductor laser. Elect. Lett. 18, 763-764. 30. Sonix Inc. 800 Morrissette Drive Springfield, VA 22152. 31. Balk, L., and Domnik, M. (1993). Quantitative scanning electron microscopy of silicon. Scanning Microscopy 7(1), 37-48.
346
R o b e r t S.
Gilmore
32. Cargill, G. S. (1980). Electron acoustic microscopy. In "Scanned Image Microscopy." Academic Press, New York, pp. 319-331. 33. Monk, G. S. (1937). "Principles of Light." McGraw-Hill, New York. 34. O'Neil, H. (1949). Theory of focusing radiators. J. Acoust. Soc. Am. 21,516. 35. Born, M., and Wolf, E. (1980). "Principles of Optics," 6th Edition. Pergamon Press, London. 36. Kino, G. S. (1980). Fundamentals of scanning systems. In "Scanned Image Microscopy" (E. A. Ash, ed.). Academic Press, London. 37. Krautkramer, J., and Krautkramer, H. (1990). "Ultrasonic Testing of Materials." Springer-Verlag, New York, pp. 75-79. 38. Simmons, G., and Wang, H. (1971). "Single Crystal Elastic Constants and Calculated Aggregate Properties: A Handbook." Cambridge University Press, Cambridge, Massachusetts. 39. Chung, D., and Buessem, W. (1968). The elastic anisotropy of crystals. In "Anisotropy in SingleCrystal Refractory Compounds," Vol. II (E Vahldiek and S. Mersol, eds.). Plenum Press, New York. 40. Anderson, O., and Liebermann, R. (1968). Sound velocities in rocks and minerals: experimental methods and extrapolation to very high pressures. In "Physical Acoustics," Vol. IVb (W. Mason, ed.). Academic Press, New York. 41. Mason, W. (Ed.) (1964-1994). "Physical Acoustics," Vols. I-XX. Academic Press, New York. 42. Fitting, D. W., and Adler, L. (1981 ). "Ultrasonic Spectral Analysis for Nondestructive Evaluation." Plenum Press, New York. 43. Rodel, J., and Glaeser, A. M. (1989). Photolithography: a new tool for ceramic science. Proc. Mater. Res. Soc. Symp. 155, 293-306. 44. Rodel, J., and Glaeser, A. M. (1987). Production of controlled-morphology pore arrays: implications and opportunities. J. Am. Cer. Soc. 70, c 172-c 176. 45. Mitchell, K. W., and Gilmore, R. S. (1992). A true Wiener filter implementation for improving the signal to noise in acoustic images. In "Review of progress in Quantitative NDE," Vol. II (D. O. Thompson and D. E. Chimenti, eds.). Plenum Press, New York. 46. Howard, P., and Gilmore, R. (1994). Ultrasonic C-scan images for hard-alpha flaw detection and characterization. In "Review of Progress in Quantitative NDE," Vol. 13 (D. O. Thompson and D. E. Chimenti, eds) Plenum Press, New York. 47. Jain, A. K. (1989). "Fundamentals of Digital Image Processing." Prentice-Hall, Englewood Cliffs, New Jersey. 48. Lim, J. S. (1990). "Two-Dimensional Signal and Image Processing." Prentice-Hall, Englewood Cliffs, New Jersey. 49. "Acoustic Imaging," Vols. 1-24 (Various editors). Plenum Press, New York. 50. Wickramasinghe, K., and Petts, C. (1980). Acoustic microscopy in high pressure gases. Proc. IEEE Ultras. Symp., 668-670. 51. Fortunko, C., Renken, M., and Murray, A. (1991 ). Examination of objects made of wood. Proc. IEEE Ultras. Symp., 1099-1103. 52. Fortunko, C., Dube, W., and McColskey, J. (1993). Gas-coupled acoustic microscopy in the pulseecho mode. Proc. IEEE Ultras. Symp. Page 667. 53. Bond, L., Chang, C. H., Fortunko, C., and McColskey, J. (1992). Absorption of sound in gases between 10 and 25 MHz: argon. Proc. IEEE Ultras. Symp., 1069-1073,
Research Instruments and Systems BRUCE B. CHICK Ritec, Inc., 60 Alhambra Road, Suite 5, Warwick, RI 02886
I. II. III. IV. V. VI. VII.
HistoricalBackground . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . AttenuationMeasurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VelocityMeasurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Attenuation and Velocity Measurements . . . . . . . . . . . . . . . . . . . . . . . . . NonlinearMeasurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Thin Film Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acoustic Emission Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...........
I.
347 348 348 351 355 357 358 361
Historical Background
Instrumentation of various degrees of complexity and sophistication for the ultrasonic examination of materials has been developed over the years since Floyd Firestone was granted a patent for his flaw detection device. His invention was manufactured and distributed by the Sperry Products Co. under the name Reflectoscope. That first commercial approach to using pulsed ultrasound to examine materials for flaws was indeed quite simple when compared to today's elaborate devices. It generally operated over a very narrow range of frequencies and consisted principally of a spike-type pulse and a simple tuned receiver plus time delayed gates. Frequency of operation was primarily determined by the resonant frequency of the transducer employed. The principal application of examining materials for flaws or defects in composition eventually came to be known as nondestructive testing (NDT). This chapter will present representative machines designed primarily for the relatively new field of nondestructive evaluation (NDE) of materials using ultrasound over a very wide range of frequencies. In N D E work, the objective is generally to document the "quality" of the material under test rather than to examine it for gross flaws. Descriptions of some representative N D E instruments developed in the research laboratory and later successfully 347 PHYSICAL ACOUSTICS, VOL. XXIV
Copyright 9 1999 Academic Press All rights of reproduction in any form reserved. ISBN 0-12-477945-X $30.00
Bruce B. Chick
348
brought to the commercial marketplace are discussed in the following sections.
II.
Attenuation Measurements
An early example of an NDE instrument was the "Attenuation Comparator" developed at Brown University in the 1950s (Chick et al., 1960), which was first brought to the marketplace through the Sperry Products Company in Danbury, Connecticut with manufacture being carried out by Madison Industries. It was used primarily to measure attenuation versus frequency. It contained a tuned pulsed oscillator and a superheterodyne receiver. The resultant "pulser/receiver" combination could be manually tuned over a frequency range of 10 to 200 MHz. The tuned pulsed oscillator gave it a welldefined frequency of operation that allowed it to be adjusted to successfully drive single-crystal transducers such as quartz plates on any of the odd harmonics that fell within the range of the instrument. For example, a 10-MHz fundamental frequency device could be operated successfully at 10, 30, 50, 70 MHz, etc. The instrument also included a calibrated, adjustable exponential waveform that could be matched to the decay of a multiple echo field produced from high-quality, low-loss materials. In this way, attenuation vs frequency data could be readily obtained. A later addition to the system was an attenuation recorder (Truell et al., 1959) in which time delayed gates were used to select out two echoes from an echo field to measure changes in attenuation. The instrument provided a readout on a strip chart recorder of the dB ratio of the amplitudes of the two selected echoes. This made possible the convenient acquisition of great amounts of data that were obtained from experiments with basic materials when some outside parameter such as temperature, magnetic field, stress, fatigue, or nuclear radiation was applied to a sample. Instruments of this type are still being marketed by Matec Instruments.
III.
Velocity Measurements
During this same period of time a very unique method of measuring ultrasonic velocity was being developed at Bell Laboratories (May, 1958). In this technique, an oscilloscope is tricked into a display that allows two successive echoes to be overlapped on the oscilloscope screen on successive traces that are initiated at the approximate round-trip time in the sample. By finely tuning a continuous wave (CW) oscillator that determines the triggering time, the
6
Research
Instruments
and
Systems
349
radio frequency display of the two echoes can be precisely overlayed to obtain the round-trip time in the sample (which is the period of the oscillator). This method became known as the pulse-echo-overlap (PEO) method (Papadakis, 1964, 1967). An instrument for making PEO measurements was later marketed by Panametrics (Papadakis, 1976) and is shown in Figure 1. The PEO measurement technique is still thought of as one of the most accurate methods for velocity measurements. The technique is based theoretically on the pulse superposition technique, also developed at Bell Laboratories (McSkimin, 1961). (This latter method was never brought to the commercial marketplace to the knowledge of this author.) A simplified block diagram of a typical PEO system is shown in Figure 2. The object of the system, as indicated above, is to match the period of the CW oscillator (measured with a high-resolution frequency counter) to the roundtrip time in the sample. The output of the CW oscillator is first divided by a fairly large number (typically 100 or 1000, with division by 1000 used as a starting point) to produce an operating repetition rate for the transducer pulse driver and to also initially trigger the time base of the oscilloscope. A normal field of echoes from the sample is then displayed on the cathode ray tube (CRT) of the oscilloscope. The same trigger is also used to initiate two series~.~,';.'.i".~'.;;'~'~"..~ .'.~.'. .'. ~ ~,~,~;~ ~ . z~:~,,~. ,. .~.~:.+.~ . . . . ~-~'~:~% ~,'~ ~.,~";',.',~ : . 9, ~ ~ , ~ - ~ ~ . ~ ~o~ . . .
~.,9~.~., . ..
9~.~ ~ . ,,,,.. .:, , .
. a a - ~ - .,. 9
"~WE" ..
FIG. 1. Pulse-echo-overlap (PEO) measurement instrument marketed by Panametrics. (Courtesy of Panametrics Inc.).
350
Bruce B. Chick ,.
l
l
CW
0SC,
1711
:I I --9 i
l
.
.
.
|
Ir COUNTER
~
TRIG
I
,,
I I I
I
!
t
IPUL~E I
~
PULSES |
I PAUND "----T . . . (3). E . . ~ . . . ; . . H-O;S
ECHOE~ ~
FIG. 2.
I
I
1
(4)
!Z
"J P(5) P(6)
SVNCU
Simplified block diagram of a typical PEO system.
connected, delayed gates where the outputs of the gates are combined and applied to the CRT to produce two intensified segments of the trace. The delays and widths are then adjusted to roughly match the first two echoes of the echo field. Next, the round-trip in the sample is roughly measured from the oscilloscope time base. The reciprocal of this time is used to establish a preliminary frequency setting of the CW oscillator. At this point, the trigger to the oscilloscope is changed to one that is produced at the CW oscillator frequency. Then the time base of the oscilloscope is increased so that one (radio frequency) RF echo will nearly fill the entire screen. With careful manipulation of the trigger controls, it is possible to achieve repetitive triggering of the oscilloscope at this much higher frequency, which is a rate roughly matching the round-trip time in the sample. Small adjustment of the CW oscillator will cause the RF cycles of the two intensified echoes to exactly match at zero crossings of the baseline. Figure 3 shows a typical CRT display of properly matched RF echoes. Because it is possible to match different cycles from the two echoes, a method of matching the same cycles in each echo was developed to obtain proper data
6
Research Instruments and Systems
351
"0 m
s
E
<
i
Time
FIG. 3.
Typical CRT display of properly matched RF echoes.
from the CW oscillator (Papadakis, 1967). To obtain the exact round-trip time, further modification of the measured period of the CW oscillator must be carried out to account for transducer and bond phasing effects (McSkimin, 1961).
IV.
Attenuation and Velocity Measurements
Very sensitive relative measurements of attenuation and phase are possible through the use of phase detection techniques (Blume, 1963). In Blume's approach, the difference in phase between a received echo, A sin(cot + 4)), and a CW reference signal, B sin(cot), is measured using a single phase detector. The phase difference, ~b, is due to electronic delays, phase shifts in the transducer, and, primarily, the propagation time in the sample. Neglecting the electronic delays and transducer phase shifts, ~b = 2nfLy, where L is the sample length and v is the velocity of sound in the sample. By adjusting the frequency, f, it is possible to change the output of the phase detector, which is proportional to sin 4), so that, for a selected echo, 4)= 90 ~ and the phasedetected signal is at a null (or quadrature) condition. By electronically selecting this echo with a time delayed gate, the phase-detected signal is converted to a (direct current) DC voltage with a sample-and-hold circuit (which is sensitive to both positive and negative input signals). The DC voltage is then amplified and fed back to a voltage-sensitive tuning element in
352
Bruce B. Chick
the CW oscillator. Because the sample-and-hold is sensitive to both the positive and negative output variations from the phase detector, the resulting DC voltage can be fed back to the oscillator in a degenerative fashion. In this way the oscillator frequency becomes sensitive to very small changes in the transit time in the sample and very high sensitivities are obtained. The limitations are: (1) changes in transit time in the sample cause changes in operating frequency, which then cause phase changes in the transducer and bond; (2) no amplitude information is obtained from the single phase detector; (3) changes in transit time in any of the system electronics or cables will change the oscillator frequency and the resulting measurements. Very sensitive relative measurements of attenuation and velocity are possible with a more elaborate approach, where the CW source is used as the reference for two phase-sensitive detectors (a phase-sensitive detector is sensitive to both phase and amplitude) where the reference is shifted by 90 ~ to one of the detectors. RITEC, Inc. has taken the basic concept and expanded it into a single instrument identified as the RITEC Advanced Measurement (RAM) system (Fortunko et al., 1992). This system includes a direct digital synthesizer for generation of the CW signal; a high power, gated RF pulse amplifier; a tracking, superheterodyne receiver; a pair of phase-sensitive detectors; and a pair of gated integrators at the output of the phase-sensitive detectors. The integrators are used to convert selected pieces of information into DC voltages for digitization and transfer to a computermwhich is used to control all appropriate functions of the instrument as well as to analyze data. A block diagram of the system is shown in Figure 4, where a super, heterodyne receiver is automatically tuned to the frequency of interest for further amplification and processing. (In a superheterodyne receiver incoming signals are typically first amplified and then converted to an intermediate frequency [IF] through a mixing process with an adjustable, continuous wave [CW] signal, which is usually referred to as the local oscillator [LO].) There are several very unique features in the RAM system. One of the most important is the technique used to accomplish phase-coherent detection in a superheterodyne receiver where all signals are converted to an intermediate frequency (IF). The starting point is a crystal-controlled CW oscillator operating at the intermediate frequency. It is used as the reference for the phase-sensitive detectors and for the additional function of generating the system operating frequency by mixing (multiplier) action with a synthesizer operating at the local oscillator frequency for the superheterodyne receiver. The two RF signals, one from the IF oscillator and the other from the synthesizer, produce sum and difference frequencies at the output of the
6
Research Instruments and Systems
353 DATA ACQUISITION CARD
HIGHPOWERRF GATED AMPLIFIERMODULE
DIGIYAL CONTROL OF ALl_ MOD1 TI,ES
OUTPUT AMPLITL'DE'" ] CONTROL
TIMINGMODULE CON'I'ROL OF RF BURST WIDTH, REPETITION RATE, LNTEGRATE GATE POSITION AND WIDTH
AMPLIFIER GATE CW at frequency F
QUADRA~J~
FREQUENCY SYNTHESIZER [
I
DIRECTDIO1TAL SYNTHESIZER
.....
PHASE SENSITIVE INTEGRATORGATE
[
.... ,
SIGNAL
~ RFV E ,. BROADBAND R
!'
PASS
o.,,oL j
GATED INTEGRATORS
]IF OSCILLATOR
''"~1761 MIXER& W AMPLIFIER
!
,
,~
NOLO--TO DATA ACQUSITIONCARD
f
FIG. 4. Blockdiagram of the RITEC Advanced Measurement (RAM) system. mixer. It is a simple task to select the difference frequency and reject the sum frequency through the use of a low-pass filter. (For example, if a 100 kHz operating frequency is desired and the intermediate frequency is 25 MHz, then the synthesizer is set at 25.1 MHz to produce frequencies of 100 kHz and 50.1 MHz at the output of the mixer. The 100 kHz signal is easily separated from the 50.1 MHz signal by using a low-pass filter.) The selected difference frequency is amplified and shaped into a radio frequency pulse by means of a series of high gain amplifiers that are coherently gated on and off for a predetermined number of cycles. The resulting RF pulse is then used to drive an appropriate transducer coupled to a test piece. An automatic switching diplexer can be used to couple return signals from the sample to the broadband, high gain portion of the superheterodyne receiver, which precedes another mixer. In addition to the signal from the sample, the CW output from the synthesizer, operating as the local oscillator, is also coupled into this mixer. The mixer produces two output frequencies, which are the sum and difference of the two input frequencies. A bandpass filter is used to select the one that is equal to the intermediate frequency. The output of the bandpass filter is amplified by a small amount in a so-called IF amplifier and applied to a signal input pin on each of the two phase-sensitive detectors. (Very low IF amplifier gain is required to prevent breakthrough of
354
Bruce B. Chick
the LO signal from saturating the detectors when operating at very low frequencies where the LO frequency falls within the bandwidth of the IF amplifier filter. Therefore, almost all of the total gain in the superheterodyne receiver in the RAM occurs in the broadband "front end" of the receiver.) The output of the IF oscillator is also applied to the phase-sensitive detectors, but the signal to one of them is phase shifted by 90 ~. In this manner the outputs of the two detectors are A sin(cot + ~b) and A cos(o~t + ~b). These two outputs are frequently referred to as the real and imaginary parts of the signal. A pair of time-gated integrators are next used to convert selected portions of the train of signals from the detectors into DC voltages (I1 and I2), which can be digitized and transferred to the computer. The true amplitude of the time-selected signal is then simply v/(I 2 + I2), and the phase is ~ b tan- l (I2 / I1). There are many advantages to the superheterodyne technique: noise rejection by the relatively narrow IF filter, fast selection of the operating frequency, convenient shaping of the receiver bandpass by adjustment of the IF filter, and quadrature phase-sensitive detection at a fixed frequency. (Phase detection at a fixed frequency permits the phase errors in the system to be adjusted down to approximately 0.2 ~ instead of the typical 3 ~ error that occurs in a broadband receiver.) Another valuable feature of the RAM is that changes in velocity and attenuation can be made using two echoes when available from the sample. When this is done, changes attributable to instrumentation and cables are essentially eliminated through common mode rejection. This then allows much greater sensitivity to be realized under adverse conditions for the instrumentation. In addition to conventional measurement methods, the RAM also functions well in the "interrupted CW method" (Miller and Boloef, 1971) in which a very long pulse is used in an effort to build up mechanical resonance in a thin plate of test material. The synthesizer is programmed to slowly sweep through a wide frequency range to detect the mechanical resonances in a test piece. This method has been found to be extremely useful in thin metal plates when used in conjunction with electromagnetic transducers (EMATs), which do not load on the sample resonances (Petersen et al., 1994) (Hirao et al., 1993). A unique feature of an EMAT is that there is no requirement for surface preparation of the material to be examined nor is there a requirement for an acoustic bond between the transducer and the test piece. The trade-off for these benefits, however, is extremely poor sensitivity. Fortunately, this poor sensitivity is easily overcome in resonance techniques where the frequencies of mechanical resonance typically produce a 40 dB enhancement of signal levels over nonresonant signals.
6
Research Instruments and Systems V.
355
Nonlinear Measurements
An advanced version of the RAM has been brought to the marketplace by RITEC for the study of nonlinear acoustic phenomena (SNAP) in materials. The SNAP unit has the same functions built into it as does the RAM but also has a second high-power gated RF amplifier and three independent but coherent gated synthesizers. (A unique and very valuable feature of currently available direct digital synthesis [DDS] chips is that multiple units can be gated into a fully coherent "on" condition with respect to each other if they are all clocked from the same CW source, which is typically a quartz crystal stabilized oscillator.) The first two DDSs in the SNAP system drive the two gated amplifiers, and the third is used to independently tune the built-in superheterodyne receiver. The instrument also contains a pair of 16-bit analog-to-digital converters to convert data obtained from the gated integrators to digital form so that very high resolution data can be transferred to the computer for analysis. The pulses to be amplified by the gated amplifiers can be independently delayed, as is required for some nonlinear measurement techniques. They still maintain their coherency to each other because the delay circuitry is derived from the same oscillator used to clock the DDSs. Additional features include automatic leveling (ALC) for the gated amplifiers as well as two methods of modulation, including Hanning modulation of the signal amplified by one of the gated amplifiers. Some nonlinear effects for a large-amplitude wave propagating in a material are the generation of second (and higher) harmonics of the fundamental frequency of the wave, and the nonlinear dependence of the amplitude of a received signal on the driving amplitude. In addition, effects occur when two large-amplitude waves, at frequencies fl and j~, propagate in the sample. These effects include the generation of waves at the sum frequency, fl +f2, and at the difference frequency, fl -J2- Because the signals of interest, which are produced by the nonlinearity of the sample, occur at frequencies different than those of the transmitted signals, the receiving system must be capable of selecting specific frequencies of interest and rejecting others that may be larger in amplitude. This is accomplished in the SNAP system with the superheterodyne receiver as described above as well as with appropriate, external, low-pass and high-pass filters. Besides being extremely useful for nonlinear measurements, the SNAP system can also be used for more conventional ultrasonic measurements. A photo of the SNAP system is shown in Figure 5, and a block diagram is shown in Figure 6.
Bruce B. Chick
356
~:::!:':~iiii
9
i / ~ ! ....."": i
.
,,--,
~
..
,-=rr,, ~
FIG. 5. RITEC's study of nonlinear acoustic phenomena (SNAP) system. (Courtesy of RITEC Inc.).
OUTPUT AMPLITUDE CONTROl.
~rr^LfJoc^~o C O N T R O L OF ALL MODULES & DATA O U T P U T TO C O M P U T E R
HIGH POWER RF GATED A M P I J F I E R MODULES
OUTPUT AMPLITUDE CONTROl,
TIMING MODULE
A M P I J F I E R GATES
. . . . . .
CONTROl. OF RF BURST WIDTHS & POSITIONS, REPETITION RATE. AND INTEGRATE GATE POSITION AND WIDTH
START TRIGGER
s..
and
I
L_= =r~
I
.
' . . . . . . . . . .
,
. FREQUENCY
. ~ o ~ . . ~ . ,
""~":_"~"" _
II,~~"~~176
, I
PX~nS~Sl'nV~
|
0o/180o
, ......... I..:!................... S.ELECT,O~ ...... !I
FIG. 6.
I
DATA OUTPUT
. GATED
16 BIT A/D
~ ...... 1! . . . . . . . . CONTR:O'.:..::!.l
Block diagram of the SNAP system.
. .... ...l.:,:I }
i
6
Research Instruments and Systems
VI.
357
Thin Film Measurements
The development of a remarkable new method for examining very thin films at light frequencies began at Brown University in 1948 (Grahn, 1989). Picosecond ultrasonic laser sonar (PULSE) is the first metrology technique capable of measuring multilayer metal films directly on production wafers. PULSE uses ultrafast light pulses to launch ultrasonic waves into such films and then analyzes the resulting echoes to measure the thickness of all layers simultaneously with a precision of better than 1%. Return signals are recovered by means of a second illuminating laser also applied to the material. Because received signals are very weak, extensive averaging is used to recover signals. This can proceed at a very rapid rate, however, because of the megahertz repetition rate of the system. It has been brought to the marketplace by Rudolph Technologies, Inc. in collaboration with the researchers at Brown. A photo of this system appears in Figure 7.
i i 84184184 ~i~ i: 84184184184 ii~
99
ii!iiiii9ii
~:~i(!!! ~!!i!!i~i!i~ii@~i~,i!!!i~iii 9 ii~ii~ii i~ i ~i I i 84 ~iiii~!~!iii~i~i~i~i~i~ii~iiiiiiiii~i~ii~iiiiiii~i~i~ii~i~i~i~i!!~i!~i~i~i~ii~i!!iiiiiiiiiiii~!iii!i~ii!ii~!i~i iiil i
~
~ .........i.i.~.i.i.i.iiii~iii~iiiii!ii~ii~ii!iiii~iii~ii~i~!!~i~i~ii~i~i~i~ii~i~i~i~i~ii~ii~i~i
~
,~,,'
................
iiiiiiiiiiii~iii~iiiiiii~i~i~i ~i~i~i'!~i~i84184
~ ~~
.~.
~i~ ~ii!i!i~i!i ~: ~!i?ii!:!i;i;i!iiii~!!il i!i
i::ii
!i9
}i!i :!!
FIG. 7. Picosecond ultrasonic laser sonar (PULSE) system marketed by Rudolph Technologies, Inc. in collaboration with Brown University researchers. (Courtesy of Rudolph Technologies Inc.).
9358
B r u c e B. C h i c k
The developers of the Pulse technique describe it as conceptually simple: the system generates and detects sound in opaque films with pulses of light. An ultrafast laser produces a brief (0.1 picoseconds) light pulse focused on a region of the film surface approximately 10 microns in diameter. The absorption of this "pump" light pulse causes a 5 to 10~ rise in temperature near the surface of the film. The resulting rapid thermal expansion generates a sound wave that propagates away from the surface at the speed of sound. The heat dissipates into the sample in a few nanoseconds and then plays no further role in the measurement. When the sound wave encounters the interface with a lower layer, some of it reflects back toward the surface as an echo. The remainder continues down into the lower layers of the sample. When the echo from the first interface reaches the surface, it changes the sample's reflectivity. A second light pulse, diverted from the pump beam by a beam splitter, detects the surface change. This detection, or "probe," pulse is delayed relative to the pump pulse by guiding it to the sample over a longer path. The time t for the sound wave to propagate through the film at the speed of sound, v s, and return to the surface is related to the film thickness d by t = 2 d / v s. With just one layer multiple echoes will be produced. The rate of decay will be primarily related to the reflectivity ratio between the materials, but other loss mechanisms can also be present. PULSE, however, uses only the echo times to determine film thicknesses, so its metrological precision is insensitive to echo amplitudes. This behavior allows the technique to be used to monitor individual layers and their interactions within a stack.
VII.
Acoustic Emission Measurements
The last system to be discussed is a new acoustic emission method developed by Gorman (1991) at the Naval Post Graduate School in Monterey. This new technique is known as modal acoustic emission (Modal AE). Knowledge of the modal nature of wave motion in finite or bounded structures has a long history (the discussion of which can be found in many textbooks), but the application to AE is relatively recent. Modal AE, which represents a major advance in the field of acoustic emission, is the study of the ultrasonic wave modes produced by acoustic emission sources in plates, rods, shells, and other thin-walled materials. The basic objectives of Modal AE arc: 1. Determining the nature of wave modes 2. Accurate source location 3. Verifying theoretical models
6
Research Instruments and Systems
359
The fundamentals of Modal AE are fairly well established and can be found in the literature (e.g., Gorman, 1991). Some applications have been documented as well (Prosser, 1994). This new technique has been brought to the marketplace by Digital Wave Corporation. A photo of the system is shown in Fig. 8, and a simplified bock diagram appears in Figure 9. Modal AE was first applied to composite material laboratory specimens, where it was shown that transverse matrix cracks produce a definite wave mode that can be easily measured and counted. This was followed by fatigue testing of 7075 aluminum specimens, where crack growth can be positively determined despite the presence of fastener fretting and test machine noise. Modem Modal AE testing was stimulated by the introduction of the first Modal AE instrument, the Fracture Wave Detector, by Digital Wave Corporation in 1992. Modal AE is now a rapidly growing subject of study at universities and research laboratories around the world. It has been shown that AE waveforms in plates consist of wideband wave modes containing many frequencies, each of which travels at different velocities. Such plate waves had been mentioned before in the literature, but true understanding of the importance of their interpretation has come only recently. Although the spectrum of plate modes is vast, the modes of interest
FIG. 8. The Modal AE system marketed by Digital Wave Corporation. (Courtesy of Digital Wave Corporation)
360
Bruce B. Chick A3000 Gated Pulse
A~aw
Amplifier OO !
B!025 Sens~
PA2040G/A ~ .....
,..
Preamplifier
o
I
RS-232 Serial
.
I FIG. 9.
.
,,
F4000 Rlll~r Tdgger Module
8-12 bit A/D
'\
Simplified block diagram of Digital Wave's Modal AE system.
in Modal AE are the fundamental (i.e., lowest order), extensional (E), and flexural (F) modes. Even more recent is the technological innovation that makes the measurement and analysis of plate waves practical and easy to use in the laboratory. True AE waveforms are captured with high-fidelity, wideband sensors. Therefore, Digital Wave designed a Modal AE instrument to capture and record the actual waveforms to that they can be studied with wave propagation theory (elastodynamics). This Modal AE instrument is characterized by high-fidelity, wideband sensors that capture the actual physical variables of the waveforms. Modal AE analysis methods range from basic visual recognition of the wave modes to advanced computational wave propagation, signal processing, and source location techniques. Modal AE can be applied to laboratory specimens or large production structures. In metal, by detecting the E mode wave, this method has been used successfully to detect and locate cracking in both compact test specimens loaded in constant amplitude fatigue and in large aircraft structural components during spectrum fatigue loading. In-flight applications of the technique to new and aging aircraft are currently being explored. In hat-stiffened composite aircraft structures, delaminations can be detected by the large, low-frequency F mode wave produced when the stiffness debonds as the structure is loaded in compression-compression fatigue loading. Current Modal AE research is focused on fiber bundle breakage and matrix cracking occurring in composite materials based on the frequency content and wave mode shaped detected. By pulsing a particular mode and frequency into a composite material, the material properties can be measured. A Gaussianshaped pulse produced in software is fed into a gated RF amplifier. The
6
Research Instruments and Systems
361
detected wave group or phase velocity is measured and converted to stiffness values. Wide application in the composites industry is foreseen. It is hoped that the brief, somewhat historical description of representative instruments provided in this chapter will aid workers who are interested in new instrumentation approaches to ultrasonic examination of materials. References Blume, R. J. (1963)i Instrument for continuous high resolution measurement of changes in the velocity of ultrasound. Rev. Sci. Instr. 34, 1400. Chick, B. B., Anderson, G. P., and Truell, R. (1960). Ultrasonic attenuation unit and its use in measuring attenuation in alkili halides. J. Acoust. Soc. Am. 32, 186. Fortunko, C. M., Petersen, G. L., Chick, B. B., Renkin, M. C., and Preis, A. L. (1992). Absolute measurements of elastic wave phase and group velocities in lossy materials. Rev. Sci. Instr. 65(1). 3477-3486. Gorman, M. R. (1991). Plate wave acoustic emission, J. Acoust. Soc. Am. 90(1), (Pt. 1), 358-364. Grahn, H. T., Marsi, H. J., and Tauc, J. (1989). Picosecond ultrasonics. IEEE J. Quant. Electr. 25, (12), 2562-2569. Hirao, M., Ogi, H., and Fukuoka, H. (1993). Resonance EMAT system for acoustoelastic stress measurement in sheet metals. Rev. Sci. Instr. 64(11), 3198-3204. May, Jr., J. E. (1958). Precise measurement of time delay, I.R.E. Natl. Conv. Rec. 6, (Pt. 2), B4. McSkimin, H. L. (1961). Pulse superposition method for measuring ultrasonic wave velocities in solids. J Acoust. Soc. Am. 33, 12. Miller, J. G., and Bolef, D. I. (1969). A simplified continuous wave ultrasonic technique and spectrometer. Rev. Sci. Instr. 40, 915. Papadakis, E. P. (1964). Ultrasonic attenuation and velocity in three transformation products in steel. J. Appl. Phys. 35, 1474-1482. Papadakis, E. P. (1967). Ultrasonic phase velocity by the pulse-echo-overlap method incorporating diffraction phase corrections. J. Acoust. Soc. Am. 42, 1045-1051. Papadakis, E. P. (1976). New, Compact instrument for pulse-echo-overlap measurements of ultrasonic wave transit times. Rev. Sci. Instr. 47, 806-813. Petersen, G. L., Chick, B. B., Fortunko, C.M., and Hirao, M. (1994). Resonance techniques and apparatus for Elastic wave velocity determination in thin metal plates. Rev. Sci. Instr. 65(1), 192-198. Prosser, W. H. (1994). Plate mode velocities in graphite/epoxy plates J Acoust. Soc. Am. 96(2), (Pt. 1), 902-903. Truell, R., Chick, B., Picker, A., and Anderson, G. (1959). The use of ultrasonic methods to determine fatigue effects in metals. W.A.D.C. Techn Rep. 59-389.
This Page Intentionally Left Blank
Index
Acuson, piezoelectric composites, 99 Aging, oscillators, 241-242 Aircraft Nuclear Propulsion Project (ANP Project), 10 Air defense systems, 218 Air transducers, PVDE 116-117 Amplifier, crystal oscillator, 223 Amplitude-frequency effect, 232, 233 Angular spectrum, 296 Annular array, 78 Areopagus syndrome, 10 Arrays, 129 1.5D arrays, 79, 102, 105 annual array, 78 curved sequenced array, 78, 79 linear phased array, 78, 79, 80, 98 linear sequenced array, 78, 79, 80 two-dimensional array, 78-79, 98 AT&T, technology transfer in, 13 ATL/Echo, piezoelectric composites, 95, 99 Atomic clocks, Global Positioning System, 217 Atomic oscillators, frequency standards, 263, 264 Atomic standard, 210 Attenuation, measurement, 348, 351-352 Attenuation Comparator, 348 Autonomy period, 216
Academic institutions, technology transfer, 8-12, 14, 22-23 Acceleration, crystal oscillators, 253-256 Accuracy, oscillators, 240 Acoustic beams, focusing, 289-295 Acoustic charge transport (ACT) device, 185~ 186 Acoustic emission, measurements, 358-361 Acoustic Imaging Technologies Corp., piezoelectric composites, 99 Acoustic microscopy, 275,277-278, 284, 285, 343-344 acoustic velocity, 298 applications, 323-343 focusing acoustic beams, 289-295 future work, 344 history, 278-281 image processing, 321-323 imaging surfaces, 300-308, 309 material imaged, 295-299 photoacoustic microscopy (PAM), 282, 285 resolution, 276 measuring, 313, 31 6-321 scanning acoustic microscopy (SAM), 275, 276, 282, 284 scanning electron acoustic microscopy (SEAM), 282-283, 286 scanning laser acoustic microscopy (SLAM), 283, 285, 287 subsurface imaging of volumes, 308, 310313, 314, 315 Acoustoelectric signal processing, 183-186 Acousto-optics, 186 Activity dips, 232
Backing impedance, piezoelectric plate transducers and, 56, 57-58 Battery consumption, military applications, 221-222 363
364 BAW, s e e Bulk acoustic wave Berlinite, 151 Bismuth germanium oxide, 151 Bistatic radar, 219 Blaked, Inc., piezoelectric composites, 99, 101 Bleustein-Gulyaev (BG) waves, 144 Bleustein-Gulyaev-Shimizu (BGS) waves, 144, 145 Bragg refraction, transducer beam evaluation, 70-71 Broadband sources, 112 hydrophone calibration, 112 theory, 112-115 Bulk acoustic wave (BAW) oscillator, 218 Bulk acoustic wave (BAW) resonators, 167, 168, 169, 232 Bulk longitudinal wave EMATs, 126-128 Butler configuration, oscillator, 238, 239
Cadmium sulfide, 151 Carnevale, Edmund H., 12-13 Cesium, frequency standards, 263, 264 Chirp filter, 172-173 Clamped permittivity piezoelectric composites, 87, 88 piezoelectric materials, 82-83 Clapp configuration, oscillator, 238, 239 Colpitts configuration, oscillator, 238, 239 Commercialization piezoelectric composites, 95-98, 99 technology transfer and, 15-20, 40-41 Communication systems frequency control devices, 209-210, 211, 212-216 synchronization, 213, 215, 216 Composites, 76 Composite transducers, 44, 76-78 applications, 98, 100-107 array transducers, 78-81 commercialization, 95-98, 99 composite construction and, 94-95 composite parameters and, 84-91 lateral vibration modes, 91-94 piezoelectric material characteristics, 81-84 Computer-aided design, SAW components, 149
Index Computer software, SAW crystal studies, 152153 Connectivity, piezoelectric composites, 83, 84-85 Continuous-wave (CW) oscillator, pulse-echooverlap (PEO) instrument, 349 Corporate environment, technology transfer and, 12-14 Correlators, 171-172 Coupling constant piezoelectric composites, 88-89 piezoelectric materials, 80-81 Crystal filters, 267-268 Crystal oscillator (XO), 223-236, 237 acceleration, 253-256 magnetic field, 256-257 noise, 243-246 radiation, 257-259 shock, 256, 257 stability vs. tunability, 227-228 Crystal plate, 224 Crystal unit, resonance frequency, 236 Crystal unit equivalent circuit, 224-227 C-scan imaging, 68, 74, 276-278 with ball target, 65-67, 68, 73, 75 history, 280 with microprobe, 67 Curved sequenced array, 78, 79 CW oscillator, s e e Continuous-wave oscillator
DDS, s e e Direct digital synthesis Diasonics Vingmed Ultrasound, piezoelectric composites, 99 "Dice and fill" method, piezoelectric composites, 95 Dielectric constant, piezoelectric composites, 87, 88 Dielectric film layers, SAW substrate, 182183 Dielectric resonator oscillator (DRO), 218, 244 Digital communications phase-shift-keyed (PSK) digital modulation, 216 slips, 216 synchronization, 216 Digital Wave Corp., Modal AE, 359-360 Direct digital synthesis (DDS), 355
Index
Doppler radar, low-noise oscillator, 218, 219 Drift, oscillators, 241 DRO, s e e Dielectric resonator oscillator
Echo Ultrasound, piezoelectric composite transducers, 95-97, 99, 101 Education, collaborative program between university and industry, 24-33 Electrical impedance, piezoelectric materials, 83 Electrode-configured matched filter devices, 170-171 correlators, 171-172 pulse-expander-compressors, 172-173 programmable transversal filter, 173-174 Electromagnetic acoustic transducers (EMATs), 45, 118-129 design, 120-122 disadvantages, 118 mode of operation, 118, 119-120 types, 118-119, 122-129 Electron acoustic microscopy, 282 scanning electronic acoustic microscopy, 282-283, 286 Electronic warfare, 220-221 ELINT (ELectronic INTelligence) receiver, 220 EMATs, s e e Electromagnetic acoustic transducers Energy-trapping rules, 235 Engineering education technology transfer and, 24-33 Gas Research Institute case study, 24, 25, 26-28 Takano Company project, 24, 25, 28-29 Equivalent circuit, crystal unit, 224-227
Finite element modeling, piezoelectric composite performance, 86 Flat-bottomed hole (FBH), 317, 318 Flaw detection, acoustic microscopy, 324-332 Focusing acoustic beams, 289-295 transducers, 289 Ford Motor Company, research budgeting, 12
365 Frequency control devices, 209-210 applications, 210, 212 battery consumption, 221-222 communication systems, 209-210, 211, 212-216 electronic warfare, 220-221 identification-friend-or-foe (IFF) systems, 219-220 missile guidance, 221 navigation, 216-217 surveillance, 218-219 characteristics, 261-264 choosing, 265 failure, 266 filter crystals, 267-268 harsh conditions and, 222 history, 210, 212 logistics costs, 222 oscillators, 223-224 acceleration changes, 253-256 accuracy, 240 aging, 241-242 categories, 236-238 circuit types, 238-239 frequency vs. temperature stability, 246252 instabilities, 223, 240-261 magnetic field effects, 256-257 noise, 213, 242-246 precision, 241 primary standards, 240 radiation effects, 257-259 reproducibility, 241 stability, 240, 241,246-261,259261 warm-up, 252 quartz crystals, 210-212, 267 quartz crystal sensors, 268-269 quartz crystal transducers, 268 quartz crystal unit, 228-236 resonators, 268, 270 specifications and standards, 266-267 theory, 222-262 crystal unit equivalent circuit, 224-227 stability vs. tunability, 227-228 Frequency domain response, piezoelectric plate transducer, 48-59, 62, 63, 65, 74 Frequency jumps, oscillators, 246
366
Gallium arsenide, 151 Gas Research Institute (GRI), collaborative program with Iowa State University engineering department, 24, 25, 26-28 Gate oscillators, 239 General Electric Medical Systems, piezoelectric composites, 99 Generalized SAW (GSAW) waves, 145 Global Positioning System (GPS), 217 Government agencies, technology transfer and, 9-10, 33--42 Grain-to-grain anisotropy, resolution of scanned image, 313 Gray-scale imaging, history, 280, 281 GSAW, s e e Generalized SAW waves Gun-hardened oscillators, 222
High-velocity pseudo-SAWs (HVPSAWs), 144 Hitachi Medical Corporation of America, piezoelectric composites, 99 Hopping rate, 213 Hydrophones, 107-112 Hysteresis, oscillators, 250
Identification-friend-or-foe (IFF) systems, 219-220 IEEE International Ultrasonics Symposium, 139, 146, 198 IF amplifier, 353 IIDT, s e e Interdigitated interdigital transducer Imasonic, piezoelectric composites, 99 Industrial espionage, technology transfer and, 5 Industrial ultrasonic imaging, 275, 275-278, 277-278, 284, 285, 343-344 acoustic velocity, 298 applications, 323-343 focusing acoustic beams, 289-295 future work, 344 history, 278-281 image processing, 321-323 imaging surfaces, 300-308, 309
Index
material imaged, 295-299 photoacoustic microscopy (PAM), 282, 285 resolution, 276 measuring, 313, 316-321 scanning acoustic microscopy (SAM), 275, 276, 282, 284 scanning electron acoustic microscopy (SEAM), 282-283, 286 scanning laser acoustic microscopy (SLAM), 283, 285, 287 subsurface imaging of volumes, 308, 310313, 314, 315 Injection molding, piezoelectric composites, 95 Innovation NDT market, limitations on, 21-22 technology transfer and, 2, 17 Interdigital transducer, 137, 148-149 fabrication, 153-156 materials, 149-153 SAW devices, 156 bandpass filter, 160-167 multiple-port delay line, 159-160 SAW oscillators, 169-170 SAW resonators, 167-169, 232-233 two-port delay line, 156-159 Interdigitated interdigital transducer (IIDT), 164, 170, 187 Iowa State University, technology transfer and engineering education, 24-33
Jamming margin, 218 JENTEK Sensors, Inc., technology transfer case study, 15-20
Krautkramer Branson Inc., piezoelectric composites, 99, 100, 102, 104
Lamb, H., 143, 145 Langasite, 151 Laser-pulse shocking, transducer surface motion evaluation, 73
367
Index
Lateral vibration modes, piezoelectric composites, 85, 86, 91-94 Leaky surface waves, 144, 303-304 "Lens Makers Equation," 291 Linear phased array, 78-80, 98 Linear sequenced array, 78-80 Liquid crystal scanner, transducer beam evaluation, 71 Lithium niobate, 149-152, 163 Lithium tantalate, 150-152 Lithium tetraborate, 161 Lithography, subsurface void arrays, 318, 319 Love, A. E. H., 143, 145 Love wave propagation, 182 Low-noise oscillators, 216, 218, 219, 221, 244
M
Magnetic field, crystal oscillators and, 256257 Marketing, technology transfer, 39-40 Material Systems Inc., piezoelectric composites, 99, 105, 106 MCXO, s e e Microcomputer compensated crystal oscillator Meandering Winding MagnetometerT M (MWW), technology transfer case study, 15-20 Meander-line (ML) coil, 128-129 Mechanical radiation, 46 Medical ultrasound frequency range, 323 transducers, 103 array transducers, 78-81 Microcomputer compensated crystal oscillator (MCXO), 221,238, 252 frequency standards, 263, 264 Milas, Nicholas A., 8 Military applications battery consumption, 221-222 communication systems, 212-216 jamming, 213-215 radio silence, 216 electronic warfare, 220-221 identification-friend-or-foe (IFF) systems, 219-220 missile guidance, 221 navigation, 216-217
surveillance, 218-219 survivability under radiation and high acceleration, 222 Military Standard 15-A, resolving power, 316 Missile guidance, low-noise oscillators, 221 Modal acoustic emission (Modal AE), 358361 Modeling interdigital transducer, 149 piezoelectric composite performance, 85-86 piezoelectric plate transducer performance, 48-59, 73 Modified Butler configuration, oscillator, 238, 239 Monolithic piezoelectric plate transducers, 44, 45-48, 62, 73-76 construction, 46 frequency domain response, 48-59, 52, 63, 65, 74 pressure profile, 60-62 Sittig computer program, 48-59, 73 space domain response, .60-62, 65-73, 74 theory, 48-62 time domain response, 48-59, 62-63, 74 Monostatic radar, 219 MOSFET detection, 185 MSC, s e e Multistrip coupler Multiple-port delay line, 159-160 Multistrip coupler (MSC), 177-179
NASA, space program and technology transfer, 11 National Science Foundation (NSF), technology transfer and, l0 Navigation, frequency control devices, 216217 NDK, piezoelectric composite transducers, 97 NDT, s e e Nondestructive testing Neravite(R), liquid crystal scanner, 71 NGK Spark Plugs Co., piezoelectric composites, 99, 100 Noise, oscillators, 213, 242-246 Nondestructive testing (NDT), 347 acoustic emission measurements, 358-361 attenuation measurements, 348, 351-352 C-scan imaging, 65-68, 73-75, 276, 277, 278, 280 history, 278-281
368 Nondestructive testing (NDT) ( c o n t . ) nonlinear measurements, 355-356 technology transfer, 16, 19, 20-23 thin film measurements, 357-358 transducers array transducers, 78-81 monolithic piezoelectric plate transducers, 46-48, 73, 76 velocity measurements, 348-351 Nonlinear measurements, 355-356
OCXO, s e e Oven-controlled crystal oscillator 1.5D arrays, 79 1.5 ultrasonic testing, 285-286, 288 Orbiting Mole, 11 Oscillators, 209-210 accuracy, 240 applications, 210, 212 communication systems, 209-210, 211, 212-216 electronic warfare, 220-221 identification-friend-or-foe (IFF) systems, 219-220 missile guidance, 221 navigation, 216-217 surveillance, 218-219 categories, 236-238 characteristics, 261-264 choosing, 265 circuit types, 238-239 continuous-wave oscillators, 349 crystal oscillators, s e e Crystal oscillators failure, 266 filter crystals, 267-268 gate oscillators, 239 gun-hardened oscillators, 222 history, 210, 212 instabilities, 223, 240-261 acceleration changes, 253-256 aging, 241-242 frequency vs. temperature stability, 246252 magnetic field effects, 256-257 noise, 213, 242-246 radiation effects, 257-259 warm-up, 252 logistics costs, 222 power requirement, 221-222 precision, 241
Index
primary standards, 240 quartz crystals, 210-212, 267 quartz crystal sensors, 268-269 quartz crystal transducers, 268 quartz crystal unit, 228-236 radiation hardening, 222 reproducibility, 241 resonators, 210, 268 sidebands, 255 specifications and standards, 266-267 stability, 240, 241,246-252, 259-261 theory, 222-261 crystal unit equivalent circuit, 224-227 stability vs. tunability, 227-228 Oven-controlled crystal oscillator (OCXO), 227-228, 238, 241,242, 248, 249, 250, 252, 266 frequency standards, 263, 264 Global Positioning System, 217
PAM, s e e Photoacoustic microscopy Panametrics Inc., pulse-echo-overlap (PEO) instrument, 12-13, 349-351 Parallel Design, piezoelectric composites, 99 Patents, technology transfer and, 4-5, 38-39 Peak phase excursion, 255 PEO methods, s e e Pulse-echo-overlap methods Periodic permanent magnet (PPM), EMAT design with, 123-126 Permittivity, piezoelectric materials, 82-83 Phased array, 78, 79 Phase detection techniques, attenuation and velocity measurements, 351-352 Phase-shift-keyed (PSK) digital modulation, 216 Photoacoustic microscopy (PAM), 282, 285 Photoelastic method, transducer beam evaluation, 69-70 Physical Sciences Directorate (U.S. Army), technology transfer case study, 33-42 Picosecond ultrasonic laser sonar (PULSE), 357 Pierce configuration, oscillator, 238, 239 Piezoelectrically active pseudo-SAW (PSAW) waves, 144, 151 Piezoelectric composites, 76-77, 83-84 0-3 composites, 77, 84-89, 91, 94
Index
1-3 composites, 77, 83, 84-89, 90-93, 9498 2-2 composites, 77, 84-89, 91, 93, 94-95 application, 98, 100-107 commercialization, 95-98, 99 companies selling, 99 connectivity, 83, 84-85 construction, 94-95 lateral vibration modes in, 85, 91-94, 96 literature review, 83-84 modeling performance, 85-86 permittivity, 83 properties, 86-89 transducer performance and, 90-91 Piezoelectric composite transducers, 44, 7678 applications, 98, 100-107 array transducers, 78-81 commercialization, 95-98, 99 composite construction and, 94-95 composite parameters and, 84-91 lateral modes, 91-94 piezoelectric material characteristics, 81-84 Piezoelectric crystals, 229 Piezoelectric effect, 229 Piezoelectric elements, 46 Piezoelectric materials coupling constant, 80-81 permittivity, 82-83 specific acoustic impedance, 82 Piezoelectric plates, ultrasound, 46 Piezoelectric plate transducers, 44, 45-48, 62, 73-76 backing impedance, 56, 57-58 construction, 46 frequency domain response, 48-59, 62, 63, 65, 74 pressure profile, 60-62 Sittig computer program, 48-59, 73 space domain response, 60-62, 65-73, 74 theory, 47-62 time domain response, 48-59, 62-63, 74 Piezorubber, 97, 100 Point focus, 289-290, 294 Point spread function, 304 Precision, oscillators, 241 Precision Acoustic Devices, piezoelectric composite transducers, 97 Pressure vessel calibration block, 324-326 Primary standards, 240 Programmable SAW filters, 174
369 PSAW, s e e Piezoelectrically active pseudoSAW waves PSK, s e e Phase-shift-keyed digital modulation "Publish or perish," technology transfer and, 12 PULSE, s e e Picosecond ultrasonic laser sonar Pulse-echo-overlap (PEO) methods, 12-13, 349-350 Pulse expander-compressors, 172-173 Pulse superposition technique, 349 PVDF (polyvinylidene fluoride), 107, 116 PVDF film transducers, 107, 117, 129 air transducers, 116-117 broadband sources, 112-116 hydrophones, 107-112
Quartz, 228 properties, 228-229, 256 Quartz crystals, 210, 212, 229, 230 applications, 210, 211 properties, 228-231 Quartz crystal transducers, 268 Quartz crystal unit, 228-236 Quartz oscillators, frequency standards, 263, 264
RAC filter, s e e Reflective array compressor filter Radiation, crystal oscillators and, 257-259 Radio communication frequency control devices, 212-213 autonomy period, 216 digital communication, 216 jamming, 213-215 signal acquisition speed, 215 "Radio silence," 216 RAM system (RITEC), 352-354, 355 Rayleigh waves, 142, 145, 303-304 RbXO, s e e Rubidium-crystal oscillator RCA pellicle holography, transducer surface motion evaluation, 73 Reflection gratings, 179-182 Reflective array compressor (RAC) filter, 179180, 181 Reflective dot array, 180
370 Reflectoscope (Sperry), 347 Reproducibility, oscillators, 241 Research acoustic emission measurements, 358361 attenuation measurements, 348, 351-352 nonlinear measurements, 355-356 technology transfer and, 1--42 thin film measurements, 357-358 velocity measurements, 348-351 Resolution, measuring, 313, 316-321 Retrace, oscillators, 250, 252 Reverse engineering, technology transfer and, 5 RITEC, Inc. RITEC Advanced Measurement system (RAM), 352-354, 355 SNAP unit, 355, 356 Rowen, John, 146 Rubidium, frequency standards, 263, 264 Rubidium-crystal oscillator (RbXO), 263, 264 Rudolph Technologies, Inc., PULSE system, 357-358
SAM, s e e Scanning acoustic microscopy Satellite communication systems, signal acquisition speed, 215 SAW, s e e Surface acoustic wave SBAW, s e e Shallow bulk acoustic wave Scanned acoustic imaging, s e e C-scan imaging Scanning acoustic microscopy (SAM), 275, 276, 282, 284 history, 281,284 Scanning electron acoustic microscopy (SEAM), 282-283, 286 Scanning electron microscope (SEM), transducer surface motion evaluation, 71-72 Scanning laser acoustic microscopy (SLAM), 283, 285, 287 Schlieren method, evaluation beam evaluation, 67, 69, 75 Schumpeter, Joseph A., 2 SEAM, s e e Scanning electron acoustic microscopy SEM, s e e Scanning electron microscope Semiconductors, piezoelectric, 151 Sensors
Index
Meandering Winding MagnetometerTM,case study, 15-20 SAW sensors, 143, 186, 188-189 Shallow bulk acoustic wave (SBAW), 145 Shock, crystal oscillators, 256, 257 "Shunt" capacitance, 225 Signal processing frequency control devices, 215 SAW devices, 174-175, 188 acoustoelectronic signal processing, 183' 186 multistrip coupler (MSC), 177-179 reflection gratings, 179-182 uniform dielectric film layers, 182-183 waveguides, 175-177 "Singing drum" air transducer, 116, 117 Single-phase unidirectional transducer (SPUDT), 164, 187 SLAM, s e e Scanning laser acoustic microscopy "Smart systems," 103 Snell's law, 293, 295 Sokolov tube, 280 Sound Technology, composite products, 97, 99 Space domain response, piezoelectric plate transducer, 60-62, 65-73, 74 Specific acoustic impedance piezoelectric composites, 86-88 piezoelectric materials, 82 Sperry Products Co., Reflectoscope, 347 Spread-spectrum systems, 213 SPUDT, s e e Single-phase unidirectional transducer Spurious modes, 235 Spurs, 235 SSBW, s e e Surface-skimming bulk wave Stability, oscillators, 240, 241,246-252, 259261 Standards, technology transfer and, 5 Stealth aircraft, 218 Stoneley, Robert, 143 STW, s e e Surface transverse wave Subsurface imaging, of volumes, 308, 310313, 314, 315 Subsurface void arrays, lithography, 318, 319 Supply-push force, technology transfer, 2-3 Surface acoustic wave (SAW) devices, 136138 acousto-optics, 186 applications, 136, 203-204 art of SAW work, 206-207
Index
3 71
Surface acoustic wave (SAW) devices (cont.) conferences, 139, 146, 197-203 the future, 187-189 history, 142-148 literature review, 194-197 measuring success of, 138-141 patent activity, 139 publications, 139, 194-197 signal processing, 174-175, 188 acoustoelectronic signal processing, 183186 multistrip coupler (MSC), 177-179 reflection gratings, 179-182 uniform dielectric film layers, 182-183 waveguides, 175-177 surface elastic waves and, 141-145 worldwide activities, 140, 204-205 Surface acoustic wave (SAW) filters, 163-167 bandpass filters, 160-167 chirp filter, 172-173 electrode-configured matched filter devices, 170-171 correlators, 171-172 programmable transversal filter, 173-174 pulse-expander-compressors, 172-173 Surface acoustic wave (SAW) ID tag, 136 Surface acoustic wave (SAW) sensors, 143, 186, 188-189 Surface acoustic wave (SAW) transducers interdigital transducer, 137, 148-149 fabrication, 137, 153-156 materials, 149-153 interdigital transducer controlled devices, 156 bandpass filter, 160-167 multiple-port delay line, 159-160 SAW oscillators, 169-170, 218 SAW resonators, 167-169, 232-233, 244 two-port delay line, 156-159 Surface elastic waves, 141-145 Surface-skimming bulk wave (SSBW), 145 Surface transverse wave (STW), 145 Surveillance, low-noise oscillators, 218, 219 Synchronization, communications systems, 213,215,216 T Takano Co., Ltd., collaborative program with Iowa State University engineering department, 24, 25, 28-29
TCXO, see Temperature-compensated crystal oscillator Teaming, collaborative program between university and industry, 24-33 Technology transfer, 1-2, 7-8 academic environment and, 8-12, 14, 22-23 case studies AT&T, 13 Gas Research Institute, 24, 25, 26-28 Iowa State University, 24-33 JENTEK Sensors, Inc., 15-20 Panametrics, Inc., 12-13 Physical Sciences Directorate (U.S. Army), 33-42 Takano Company project, 24, 25, 28-29 collaborative R&D agreements, 37-38 commercialization and, 15-20, 40-41 corporate environment and, 12-14 covert mechanisms of, 4-5 defined, 35-36 education institutions, 37-38 engineering education and, 24-33 government agencies and, 10, 33-42 government and, 9 innovation and, 2, 17 investment in, 10-12, 17-18 marketing, 39-40 market structure, 2-3, 6-7 mechanisms and catalysts, 4-5 nondestructive testing (NDT), 16, 19, 20-23 patents and, 4-5, 38-39 product launch, 18-19 profits in, 19 promoting, 3-4, 36 resistance to, 5-6 source of ideas, 8-10 supply-push and demand-pull, 2-3 system integration, 6 universities, 8-12, 14, 22-23 Temperature-compensated crystal oscillator (TCXO), 227, 236, 237, 242, 244, 246, 248, 250, 252, 261,266 frequency standards, 263, 264 Global Positioning System, 217 Tetrad Corp., piezoelectric composites, 99, 106 Thermal hysteresis, oscillators, 250, 252 Thin films, measurements, 357-358 Thompson Microsonics, piezoelectric composites, 99, 102, 104
372 Time domain response, piezoelectric plate transducer, 48-59, 62-63, 74 Time-gated network analyzer, 113 Transducers, 129 choosing, 301 composite transducers, 44, 76-78 applications, 98, 100-107 array transducers, 78-81 commercialization, 95-98, 99 composite construction and, 94-95 composite parameters and, 84-91 lateral vibration modes, 91-94 piezoelectric material characteristics, 8184 electromagnetic acoustic transducers (EMATs), 45, 118-129 end-radiating transducers, 102, 104 fabrication, 153-156 focusing, 289 functions, 44, 129 interdigital transducers, 137, 148-149 fabrication, 153-156 materials, 149-153 interdigitated interdigital transducer (IIDT), 164, 170, 187 medical ultrasound, 103 array transducers, 78-81 modeling interdigital transducer, 149 piezoelectric composite performance, 8586 piezoelectric plate transducer performance, 48-59, 73 monolithic piezoelectric plate transducers, 44, 45-48, 62, 73-76 construction, 46 frequency domain response, 48-59, 62, 63, 65, 74 pressure profile, 60-62 Sittig computer program, 48-59, 73 space domain response, 60-62, 65-73, 74 theory, 48-62 time domain response, 48-59, 62-63, 74 nondestructive testing, 46-48, 73, 76, 78-81 PVDF film transducers, 107, 117, 129 air transducers, 116-117 broadband sources, 112-116
Index
hydrophones, 107-112 quartz crystal transducers, 268 single-phase unidirectional transducer (SPUDT), 164, 187 theory, 47-62 types, 44--45 Transversal filter, 173-174 Trim effect, 261 Tuning-fork crystals, 235 Two-dimensional arrays, 78-79, 98 Two-port delay line, 156-159
Ultrasonic cleaning, 108 Ultrasonic imaging, s e e also Acoustic imaging C-scan imaging, 65-68, 73-75, 276, 277, 278, 280 data acquisition, 299 Ultrasonic microscopy, s e e Acoustic microscopy Universities, technology transfer, 8-12, 14, 22-23 USAF Resolving Power Test Target, 316 U.S. Army, Physical Sciences Directorate (PSD) case study, 33-42
VCXO, 227 Velocity, measurement, 348-351 Vermon, piezoelectric composite transducers, 97, 99 Volume imaging, 308, 310-313, 314, 315
W Warm-up, oscillators, 252 Waveguides, SAW signal processing, 175-177
XO,
see
Crystal oscillator