William Andrew is an imprint of Elsevier The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, UK Radarweg 29, PO Box 211, 1000 AE Amsterdam, The Netherlands Copyright Ó 2010 Elsevier Inc. All rights reserved No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means electronic, mechanical, photocopying, recording or otherwise without the prior written permission of the publisher Permissions may be sought directly from Elsevier’s Science & Technology Rights Department in Oxford, UK: phone (+44) (0) 1865 843830; fax (+44) (0) 1865 853333; email:
[email protected]. Alternatively you can submit your request online by visiting the Elsevier web site at http://elsevier.com/locate/permissions, and selecting Obtaining permission to use Elsevier material Notice No responsibility is assumed by the publisher for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions or ideas contained in the material herein. Because of rapid advances in the medical sciences, in particular, independent verification of diagnoses and drug dosages should be made British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library Library of Congress Cataloging-in-Publication Data A catalog record for this book is availabe from the Library of Congress ISBN 13: 978 1 4377 7830 4
For information on all William Andrew publications visit our web site at books.elsevier.com Printed and bound in the USA 10 11 12 13 14 10 9 8 7 6 5 4 3 2 1
Preface
The purpose of this new book series, Developments in Surface Contamination and Cleaning, is to provide a continuous state-of-the-art critical look at the current knowledge of the behavior of both film-type and particulate surface contaminants. The first volume was published in 2008 and covered 22 critical topics dealing with the fundamental nature of contaminants, their characterization, and various techniques for their removal. The present book is the second volume in the series. The individual contributions in the present book provide state-of-the-art reviews by subject matter experts of different critical topics in surface contamination and cleaning. The very complicated phenomenon of particle deposition onto surfaces in a confined environment is frequently encountered in numerous industrial and environmental settings. The transport behaviors of particle deposition are governed by the nature of airflow turbulence near surfaces, the surface characteristics and, ultimately, the particle size. Direct measurements of particle deposition onto surfaces of interest are a challenging task to perform, and contributions from various potential deposition mechanisms under realistic scenarios make modeling work difficult. De-Ling Liu in her contribution presents a review of the literature on the important physical factors that influence the behavior of particle deposition in enclosures, and the available experimental techniques as well as modeling approaches used for characterizing the rate of particle deposition. Further progress to elucidate the processes of particle deposition will require continued efforts in conducting carefully controlled experimental investigations, in which the particle size, near-surface airflow conditions and nature of the surface are well characterized. Jacques van der Donck describes a systems analysis approach to contamination control that influences a wide variety of industrial processes. For complex systems, contamination control offers a way for finding workable solutions to improve process or product functionality impacted by contaminants. Central to this approach is the system, i.e. a product, a process or a production tool. Contamination control has its focus on the system as a whole. A solution must be compliant with every aspect of the system. In the system analysis, all contamination-sensitive objects and process steps are described and the system requirements are listed. This is followed by identification of the relevant types and levels of contamination. Failure modes and effects analysis (FMEA) is used to rank the failure modes and effects, which helps in developing a solution. The three main options for the solution are: prevention, monitoring, and cleaning. The choice among the three options depends on the .
ix
x
Preface
maturity of the process and how much the actually observed particle contamination levels differ from the requirements. The elements of the FMEA occurrence, severity, and detection can show the impact of different solutions on the process performance. The earlier the contamination risks are taken into consideration during the development of the product, the better they can be counteracted. Particle contamination that deposits on the surface of the substrate during the fabrication process of integrated circuits (ICs) creates defects and thus impacts the yield of the devices. The challenges for the semiconductor industry are in the removal of these particles during manufacturing without damaging the device structure, as well as prevention of particles from depositing on critical surfaces. Current yield models, used to predict the economic aspects of defects, do not include knowledge of the deposition or the removal mechanisms of particles. In their contribution, Martin Knotter and Faisal Wali provide an overview of current yield models, followed by a description of an advanced model that explains the removal of particles in three process steps: Detachment of the particle from the surface by undercutting or applying shear forces. Diffusion of the particle through the boundary layer into the bulk of the process liquids. Transport of the particle away from the wafer environment by filtration or draining. Experimental results illustrate the impact of model parameters on the removal of particle contamination. It is often difficult to justify the large capital investment necessary to install a continuous contamination or electrostatic charge monitoring system. Often, this is the result of faulty historical data, where sampling practices preclude obtaining a true characterization of the workplace. There is a fear that hasty installation of a system will result in placement of sensors in locations where they are not needed. Finally, there are often questions about the types of sensors that should be used, the required resolution and other technical concerns that render decision making difficult. The chapter by Roger Welker describes a method to overcome these difficulties. Several examples illustrate the method as applied to particle sampling. The first step in this method is installation of sampling hardware on workstations that conforms to the requirements for sampling during critical and busy operational periods. Data are collected to determine if the traditional sampling method has achieved an accurate measure of the conditions at the workstation. Thereafter, sampling may continue using the manual optical particle counter, electrostatic charge monitor or other workstation monitor, with a modified sampling protocol to collect comparative data. Sampling may also continue using the previous protocol to provide control data. Data collected with the new protocol are then compared with the
Preface
xi
historical database using the old protocol and the historical database. Generally, this uncovers a number of sample points where the old protocol grossly underestimates the particle concentrations or static charge levels present. Workstations that can be brought under control and maintained using a reasonable manual sampling frequency do not need continuous monitoring. Workstations that repeatedly show unacceptable conditions under manual sampling are candidates for continuous monitoring. Other examples also illustrate the evaluation of the need for continuous horizontal-flow monitoring in a vertical unidirectional-flow cleanroom, electrostatic charge monitoring, and continuous monitoring in cleaning machines. Strippable coatings are used to remove surface contaminants from highquality parts and to protect the cleaned parts from surface damage. This is a low-cost, effective method for protecting and cleaning high-quality surfaces. One successful application has been removal of dust and debris from precision optical surfaces, such as coated lenses and mirrors. Other applications include decontamination of surfaces by removal of hazardous and radioactive contaminants. Rajiv Kohli provides an overview of the types of strippable coatings and their properties, and discusses some of the applications of these coatings for removal of surface contaminants. The contribution by Sami Awad and Ramamurty Nagarajan focuses on ultrasonic cleaning. Ultrasound is an effective technique for cleaning a variety of surfaces, from the delicate removal of contaminant particles on semiconductor wafers to the removal of scale and oxides from steel strip. To remove submicrometer particles, semiconductor and microelectromechanical systems, optical and hard-disk drive components, micro-inertial sensors, laser gyroscopes, compact discs, and optical coating applications all employ liquid ultrasonic cleaning baths for cleaning steps. Ultrasonic cleaning involves the use of high-frequency sound waves in the range between 40 and 200 kHz to remove a variety of contaminants from parts immersed in aqueous media. Ultrasonic cleaning can be very effective, but a number of performance features of ultrasonic cleaning need consideration. The tendency of the ultrasonic energy to damage parts is a consideration in selection of frequency and power density in the cleaning tank, and can influence design of the equipment and design of the process. The contributions in this book provide a valuable source of information on the current status and recent developments in the respective topics on surface contamination and cleaning. The book will be of value to government, academic, and industry personnel involved in research and development, manufacturing, process and quality control, and procurement specifications in microelectronics, aerospace, optics, xerography, joining (adhesive bonding), and other industries. We would like to express our heartfelt thanks to all the authors in this book for their contributions, enthusiasm, and cooperation. Our sincere appreciation
xii
Preface
goes to Matthew Deans, our publisher, who has strongly supported publication of this book and the future volumes in this series. Melanie Benson and the editorial staff at Elsevier have been instrumental in seeing the book to publication. Rajiv Kohli would also like to thank Jody Mantell for her tireless efforts in locating obscure and difficult-to-access reference materials. Rajiv Kohli Houston, Texas
Kash Mittal Hopewell Junction, New York
About the Editors
Dr Rajiv Kohli is a leading expert with The Aerospace Corporation in contaminant particle behavior, surface cleaning, and contamination control. At the NASA Johnson Space Center in Houston, Texas, he provides technical support for contamination control related to ground-based and manned spaceflight hardware for the Space Shuttle, the International Space Station, and the new Constellation Program that is designed to meet the United States Vision for Space Exploration. His technical interests are in particle behavior, precision cleaning, solution and surface chemistry, advanced materials, and chemical thermodynamics. Dr Kohli was involved in developing solvent-based cleaning applications for use in the nuclear industry and he also developed an innovative microabrasive system for a wide variety of precision cleaning and microprocessing applications in the commercial industry. He is the principal editor of the new book series Developments in Surface Contamination and Cleaning; the first volume in the series was published in 2008 and the present book is the second volume in the series. Previously, Dr Kohli co-authored the book Commercial Utilization of Space: An International Comparison of Framework Conditions, and he has published more than 200 technical papers, articles, and reports on precision cleaning, advanced materials, chemical thermodynamics, environmental degradation of materials, and technical and economic assessment of emerging technologies. Dr Kohli was recently recognized for his contributions to NASA’s Space Shuttle Return to Flight effort with the Public Service Medal, one of the agency’s highest awards.
.
xiii
xiv
About the Editors
Dr Kashmiri Lal ‘‘Kash’’ Mittal was associated with IBM from 1972 to 1994. Currently, he is teaching and consulting in the areas of surface contamination and cleaning, and in adhesion science and technology. He is the Editor-in-Chief of the Journal of Adhesion Science and Technology and is the editor of 98 published books, many of them dealing with surface contamination and cleaning. Dr Mittal was recognized for his contributions and accomplishments by the worldwide adhesion community, which organized in his honor on his 50th birthday the 1st International Congress on Adhesion Science and Technology in Amsterdam in 1995. The Kash Mittal Award was inaugurated in his honor for his extensive efforts and significant contributions in the field of colloid and interface chemistry. Among his numerous awards, Dr Mittal was awarded the title of doctor honoris causa by the Maria Curie-Sklodowska University in Lublin, Poland, in 2003.
Contributors
Number in parentheses indicates the page on which the author’s contribution begins
Sami B. Awad (225), Crest Ultrasonics Corporation, P.O. Box 7266, Trenton, NJ 08628, USA Jacques C.J. van der Donck (57), TNO Science and Industry, P.O. Box 155, 2600 AD Delft, The Netherlands
D. Martin Knotter (81), NXP Semiconductors, Gerstweg 2, 6534 AE Nijmegen, The Netherlands
Rajiv Kohli (177), The Aerospace Corporation, 2525 Bay Area Boulevard, Suite 600, Houston, TX 77058 1556, USA
De-Ling Liu (1), Space Materials Laboratory, The Aerospace Corporation, P.O. Box 92957 M2/248, Los Angeles, CA 90009 2957, USA
Ramamurthy Nagarajan (225), Department of Chemical Engineering, IIT Madras, Chennai 600036, India
Faisal Wali (81), NXP Semiconductors, Gerstweg 2, 6534 AE Nijmegen, The Netherlands Roger W. Welker (121), R.W. Welker Associates, 19060 Brasilia Drive, Northridge, CA 91326, USA
.
xv
Chapter 1
Particle Deposition onto Enclosure Surfaces De-Ling Liu The Aerospace Corporation, Los Angeles, CA, USA
1. Introduction 2. Background 2.1. Relevance to Contamination Control in Space Applications 2.2. Nature of Airflow and its Near Surface Characteristics 2.3. Evaluating Particle Deposition: Deposition Flux 3. Mechanisms of Particle Transport 3.1. Brownian Diffusion 3.2. Turbulent Diffusion 3.3. Drag Force 3.4. Gravitational Settling 3.5. Thermophoresis 3.6. Electrostatic Force 3.7. Turbophoresis 4. Parameters for Particle Deposition Characterization 4.1. Deposition Velocity, vd 4.2. Particle Deposition Rate, b 4.3. Comparison of vd and b
5. Methods for Measurement of Particle Deposition 5.1. Measuring Particle Deposition Rate, b 5.2. Measuring Particle Deposition Velocity, vd 6. Review of Experimental Studies 6.1. Particle Characteristics Affecting Particle Deposition 6.2. Airflow Characteristics Affecting Particle Deposition 6.3. Surface Characteristics Affecting Particle Deposition 7. Modeling Particle Deposition and the Experimental Validations 7.1. Homogeneous Turbulence Model 7.2. Three Layer Model 7.3. Mass Transfer Model 8. Summary Acknowledgements References Appendix
Developments in Surface Contamination and Cleaning Copyright Ó 2010 Elsevier Inc. All rights of reproduction in any form reserved.
1
2
Developments in Surface Contamination and Cleaning
1. INTRODUCTION The study of particle deposition from aerosol flow onto surfaces in enclosed spaces has attracted considerable attention in the past few decades owing to its significance in a great variety of technological applications, such as microcontamination control in the semiconductor industry [1,2], aerosol production in chemical reactors [3], indoor air pollution [4 6] and soiling of artworks [7], as well as nuclear reactor safety analysis [8,9]. Both desirable and undesirable outcomes may result from the processes of particle deposition. For instance, the thin-film coating techniques used to generate unique surface properties are an application of the desirable outcome [10,11]. In contrast, there are many examples of undesirable particle deposition, which causes surface contamination or even material damage. One of the well-known examples is the reduced yield of integrated circuits in the wafer manufacturing process. Particulate contamination on critical surfaces of a space telescope leading to optical obscuration due to light scattering and hence degrading performance is another unwanted consequence [12]. To develop effective strategies to control either desirable or undesirable deposition of particles, a sound understanding of the underlying particle transport mechanisms and the associated physical factors influencing the processes is required to gain insights into particle deposition behaviors. A significant body of scientific research has been devoted to studying the dynamic behavior of particles under well-defined configurations and airflow conditions examples include aerosol-laden air flowing through fibrous filters [13,14], particle deposition from turbulent channel flow [15,16], and particle deposition from laminar, vertical flow onto a circular plate [2,17 19]. In contrast, not many published studies consider particle deposition onto surfaces in a three-dimensional confined space, probably due in part to the complexity of airflow patterns associated with enclosures of various geometrical shapes. The goal of the chapter is to provide a state-of-the-art literature review concerning particle deposition onto surfaces within an enclosed volume. The knowledge gleaned from the available scientific publications presented here can be applied to an enclosure scale as small as a spacecraft payload cavity, or as large as a cleanroom in a manufacturing/processing facility. The surfaces of interest for particles to deposit on could be walls, the floor, the ceiling, or any contamination-sensitive surfaces within an enclosed volume relevant to various engineering applications. The chapter begins with the characterization of airflow adjacent to a surface and how it relates to the core region of an enclosure. Particle deposition flux, a parameter used to evaluate the rate of deposition onto surfaces, is briefly introduced as the concept will be mentioned throughout the chapter. Next, the fundamental physical processes governing the transport of particles are discussed, and the definitions of the parameters commonly reported in the literature for characterizing deposition are introduced, along with the available
Particle Deposition onto Enclosure Surfaces
3
experimental techniques for their measurement. The dependence of particle deposition rate on particle size, airflow and surface characteristics is examined from the existing published experimental investigations relevant to enclosures. Lastly, modeling developments for predicting the rate of particle deposition and associated experimental validations are presented.
2. BACKGROUND Knowledge of particle transport and deposition within an enclosed volume has been applied in numerous practical scenarios. In the beginning of this section, special attention is given to the relevance of particle deposition to contamination control in the context of space applications. Since aerosol particles are transported to the vicinity of surfaces by air currents, understanding the nature of airflow and its near-surface characteristics is the first step to gain insights into the processes of particle deposition onto surfaces. In addition, a quantitative approach using the concept of flux and the underlying physical mechanisms that govern particle deposition are briefly described.
2.1. Relevance to Contamination Control in Space Applications In space applications, the concern of particle deposition on contaminationsensitive surfaces arises from the undesirable effects of surface obscuration. For one thing, the presence of particles on optical reflective surfaces, e.g. mirrors and focal planes, interferes with the proper transmission of light to the next optical components, hence reducing the signal strength. Particles residing on an absorptive surface, such as baffles inside a telescope assembly, cause light scattering, which in turn overwhelms the signal and ultimately may even damage other optical components. As regards particles on thermal control surfaces, their presence causes alterations of solar absorbance and/or emissivity, which leads to an altered equilibrium temperature and may further deteriorate the thermal control function. In light of the adverse effects due to particulate contamination, the aim of contamination control is to prevent performance degradation by minimizing the deposition of particles on critical and sensitive surfaces of spacecraft components. Meeting this objective is particularly important for high-profile remote sensing spacecraft, such as the Hubble Space Telescope. As a consequence, tremendous efforts and resources are dedicated to contamination control during various phases of spacecraft processing: design, manufacturing, assembly, testing, integration, storage, shipping, launch-site ground activities, and inflight operations. The development of effective mitigation strategies to minimize particle contamination requires knowledge of particle transport and deposition as well as the associated physical factors affecting the processes. Spacecraft surface
4
Developments in Surface Contamination and Cleaning
contamination as a result of particle fallout may take place at any level of engineering practice for instance, component fabrication in a cleanroom, system assembly and testing within a payload cavity, and spacecraft integration inside a payload fairing. Note that the terms in italics refer to the corresponding enclosed volumes under the example scenario considered; thus, one can see that the study of particle deposition onto surfaces in enclosures is of strong relevance to a variety of circumstances in space applications. The insights from studying the physical processes that influence the rate of particle deposition within an enclosure are expected to provide a strong scientific foundation to benefit various aspects of aerospace contamination control needs.
2.2. Nature of Airflow and its Near-Surface Characteristics Inside an enclosed volume, air is mixed through the following two mechanisms: natural convection and turbulent mixing (forced convection). Natural convection is driven by temperature gradients, i.e. temperature difference (DT) across the space inside the enclosure, induced by heat transfer at surfaces. Turbulent mixing or force convection, on the other hand, is associated with external energy input into a system such as fan mixing or fluid flushing. In an enclosure without the mechanism of forced convection, natural convection becomes the only cause for fluid mixing. Small temperature variations within the enclosure may induce convective currents, which in turn cause aerosol concentration to become mixed throughout the volume. It is estimated that the convective flow velocity can reach approximately 1 cm s 1 in a 1-m-high chamber when the floor or wall surfaces are warmer than the air by 0.01 C [20]. Study also has shown that DT of 0.1 C can induce homogeneous mixing for particles up to 15 mm in diameter [21]. On the other hand, buoyancy can also impede mixing. For instance, a warm ceiling and cool floor will promote stratification of air masses within an enclosure. Some empirical evidence about near-surface airflows in rooms can be found in Nazaroff et al. [22]. One widely accepted approach to model airflow in enclosures is to assume that the air in the core region is homogeneously and isotropically turbulent, behaving like an ideal nonviscous fluid. Adjacent to the interior surface, the air is assumed to behave as a viscous fluid within a thin layer, which is also known as the boundary layer. Inside the momentum boundary layer, the fluid velocity drops sharply to zero at the surface from its mainstream peak value. The thickness of the momentum boundary layer depends on the momentum diffusivity of the fluid. For instance, the typical thickness is about 1 2 cm for room air motion [6], given air kinematic viscosity of 0.15 cm2 s 1 at 293 K and 1 atm. The boundary layer concept can also be used to capture the near-surface characteristics in terms of thermal and contaminant concentration profiles. For air, the thickness of the thermal boundary layer is comparable to that of the momentum boundary layer because the thermal diffusivity (~0.2 cm2 s 1 for air) is of the same order of magnitude as the momentum diffusivity. For gaseous
Particle Deposition onto Enclosure Surfaces
5
contaminants with molecular weights close to N2 and O2, the thickness of the concentration boundary layer would be similar to that of the momentum and thermal boundary layer because the molecular diffusivities are comparable to momentum and thermal diffusivities. Airborne particles, however, have small diffusivities (or diffusion coefficients) compared to molecular contaminants, which leads to much thinner1 particle concentration boundary layers than the momentum and thermal boundary layers. Okuyama et al. [23] estimated from their experiments in a 2.6-L non-stirred cylindrical tank that the particle concentration boundary layer thickness was approximately 0.3 cm when the particle diffusivity is 10 3 cm2 s 1. Considerable information about particle concentration boundary layer thickness can be found in Refs [24] and [25]. In addition to aerosol diffusivities, the nature and the intensity of the nearsurface airflow also play a role in determining the thickness of the concentration boundary layer. Scale analysis [26] that is used to approximate particle boundary layer thickness adjacent to surfaces indicates that, given the same diffusivity, the boundary layers are thinner when the fluid outside them is fast moving and thicker when the fluid moves slowly. In general, the air within an enclosure of arbitrary shape can be visualized to consist of two parts: a core zone where the air is well mixed2 and a thin quiescent boundary layer adjacent to the inside surface of the enclosure where little air motion exists in the direction perpendicular to the surface (see Figure 1.1). In the core zone the particle concentration is often assumed to be spatially uniform due to well-mixed air flows, and large-scale turbulent (or eddy) diffusion is responsible to bring aerosols to the vicinity of the surface for subsequent deposition.3 Upon arrival at the boundary layer, aerosols may migrate through the thin layer to the surface by mechanisms such as Brownian/ turbulent diffusion, gravitational settling, thermophoresis, inertial drift, and electrostatic attraction. These transport processes, as will be discussed in Section 3, commonly control the rate of particle deposition onto surfaces in enclosures. In summary, the boundary layer concept used to characterize aerosol transport processes from mainstream fluid onto surfaces has been applied with great successes in a large variety of airflow scenarios, such as external flows around cylinder or spheres, laminar and turbulent pipe flows, and enclosures with quiescent or turbulent flows. Theoretical developments based on this
1
As will be shown later in Section 6.3.2, the implication of a thinner concentration boundary layer is that small scale surface roughness can play a significant role in affecting particle transport across the boundary layer, but have negligible effects on momentum and heat transfer. 2 There is emerging literature, mainly arising from the study of transmission of airborne patho gens, that challenges the well mixed assumption, particularly for large supermicron particles (e.g. 10 mm, as they are more readily removed from the gas phase due to gravity) [27]. 3 Also known as the ‘‘homogeneous turbulence model’’, its theoretical representations will be addressed in Section 7.1.
6
Developments in Surface Contamination and Cleaning
FIGURE 1.1 Schematic of the core zone and the boundary layer in enclosures of different geometrical shapes.
boundary layer concept to evaluate the extent of particle deposition within an enclosed volume along with the associated experimental validations will be discussed in more detail in Section 7.1.
2.3. Evaluating Particle Deposition: Deposition Flux The rate of particle transport onto a surface is commonly quantified in terms of a deposition flux, with the dimensions of mass (or number) per unit time per unit area. In the event that transport through the boundary layer is dominated by the combined effects of turbulent and Brownian diffusion, the flux of particles onto a smooth, isothermal, and electrically neutral vertical surface can be characterized by the following one-dimensional steady-state continuity equation: dC Flux ¼ 3p þ D dy
(1)
where 3p and D are the turbulent (or eddy) and Brownian diffusivities for aerosols respectively, C is the aerosol concentration in air, and y is the distance from the surface. Equation (1) can be viewed as a modified form of Fick’s first law of diffusion, which describes the linear relationship between the flux of aerosols and the concentration gradient dC/dy, with the term (3p þ D) being the proportionality constant. Physically, this suggests that the net transport of particles, due to diffusion only, always takes place from regions of high to low particle concentration. Equation (1) has long been used to evaluate particle deposition flux. Since the airflow outside the boundary layer is assumed to be homogeneously mixed, a particle concentration gradient only exists within the boundary layer. Other assumptions for this theoretical representation include: (1) particles are completely retained once they collide onto the surface, i.e. the surface acts as a perfect sink for particle deposition; (2) no mechanisms of coagulation,
7
Particle Deposition onto Enclosure Surfaces
condensation, and evaporation are involved during the transport process (i.e. there are no sources or sinks for particles within the boundary layer); thus the particle flux is constant throughout the particle concentration boundary layer. In the presence of particle transport mechanisms other than diffusion, the particle flux can be evaluated by adding terms to account for external force fields acting on the particles, and that gives: Flux ¼ ð3 þ DÞ
dC ve C dy
(2)
where ve is the particle steady-state drift velocity under the action of external body forces counteracted by fluid drag. The external force fields of most interest in particle deposition processes include gravitational, electrical, and thermal (generated by a temperature gradient in the fluid), as will be presented in the next section.
3. MECHANISMS OF PARTICLE TRANSPORT Airborne particles are transported onto surfaces through a variety of physical processes, so-called deposition mechanisms. The fundamental physics behind particle transport or movement from one point to another is universal, regardless of the configurations of the system for instance, dust accumulation in ventilation ducts, gas cleaning by particle collection on fibrous filters, and scavenging of particulate matter from the atmosphere. Below, the underlying physical processes governing particle motion in air are addressed.
3.1. Brownian Diffusion Brownian diffusion is the characteristic random wiggling motion of small airborne particles in still air, resulting from constant bombardment by surrounding gas molecules. Such irregular motions of pollen grains in water were first observed by the botanist Robert Brown in 1827, and later similar phenomena were found for small smoke particles in air. In the early twentieth century, the relationships characterizing Brownian diffusion based on kinetic theory of gases were first derived by Einstein, and later verified through experiments [20]. The Brownian diffusivity D (or diffusion coefficient), which relates the gas properties and the particles through fluid drag, can be evaluated by the Stokes Einstein expression [28,29]: D ¼
kTCc 3pmdp
(3)
where k is Boltzmann’s constant (1.38 10 23 J K 1), T is the absolute temperature in K, m is dynamic viscosity of air, dp is particle diameter, and Cc is
8
Developments in Surface Contamination and Cleaning
a slip correction factor for small particles (see Section 3.3 for details). The value of D depends on the particle size and fluid properties. For instance, the diffusivity of a 0.01 mm aerosol particle is 20,000 times higher than that of a 10 mm aerosol particle. The larger the value of D, the more rapid the mass transfer process to drive particles moving from regions of high to low concentration. Brownian diffusion is the dominant particle deposition mechanism for small particles (< 0.1 mm) over short distances.
3.2. Turbulent Diffusion For particle transport in greater physical scales (i.e. larger distance), turbulent diffusion is far more effective than the thermal-based Brownian diffusion. Mass transfer of particles from one point to another by turbulent (or eddy/convective) diffusion is analogous to momentum transfer by fluctuating velocity components in turbulent flows. Unlike momentum transfer inside the thin stagnant layer of air adjacent to a surface where viscous forces are dominant and turbulence fluctuations are negligible, mass transfer of particles is attributed to the combination of turbulent and Brownian diffusion. A general expression for the particle flux to the surface due to both turbulent and Brownian diffusion was described in equation (1), with the derivations shown in Ref. [28]. Throughout the core of enclosures, particle turbulent diffusivities can vary over large ranges in accordance with the intensity of air motion. Inside the boundary layer, the particle turbulent diffusivity, 3p, is expected to gradually diminish to zero at the surface owing to the physical constraints imposed by the solid boundary. In fact, it is extremely difficult to determine the value of 3p explicitly because it is strongly associated with not only the degree of turbulence with respect to the airflow structure in the system, but also the size of particles (e.g. heavy particles cannot faithfully follow the fluid eddies). Hinze [30] has shown that over a sufficiently long time scale, the particle turbulent diffusivity equals the fluid turbulent viscosity by solving the equation of motion for a particle in a homogeneous turbulent flow field. A similar argument made by Fuchs [20] is that the motion of larger particles is more persistent due to their larger mass, which in turn contributes to the nearly identical average distance traveled by large and small particles over a long time limit. Additionally, numerical simulation of turbulent diffusion of particles also indicated that particle and fluid turbulent diffusion are comparable for particles approximately smaller than 170 mm [31]. In the air mixing experiments [32,33], the air turbulent diffusivity in an enclosed space (31 m3) can be inferred to be in the range of 8 300 cm2 s 1, depending on the air turbulence intensities. In brief, the turbulent diffusivity for particles can be reasonably approximated as the same order of magnitude as that of gas fluid, which is far greater than the particle Brownian diffusivities. In other words, particles tend to migrate closer to the surface by turbulent diffusion before Brownian diffusion becomes important.
Particle Deposition onto Enclosure Surfaces
9
3.3. Drag Force Fluid drag on a particle is the resistance force exerted by the surrounding fluid when there is relative motion between the particle and the fluid. In other words, the drag force is always present as long as the particle is not traveling in a vacuum. The net effect of the drag force is to impede the movement of particles relative to the fluid. The drag force FD on a spherical particle can be characterized by Stokes’s law: FD ¼ 3pmVdp
(4)
where V is the particle speed relative to the local fluid speed. Stokes’s law is derived from the solution of the Navier Stokes equations, assuming4 that inertial forces are negligible compared to viscous forces and that the fluid speed at the particle’s surface is the same as the particle’s. Owing to the low velocities and small particle sizes involved, most aerosol motions occur at low particle Reynolds numbers (Rep 1), where Rep ¼ dpV/n, expressing the ratio of inertial to viscous forces on particles. Table 1.1 shows the calculated Rep for spherical particles of different sizes falling at their steady-state settling velocity in air due to gravity. By looking at Rep, clearly Stokes’s law holds true for particles smaller than 20 mm. For aerosols that are relatively large and move through a fluid rapidly, the inertial forces become dominant as compared to viscous forces, thus the drag force exerted to the particle is calculated as: p 1 2 (5) rV Cd FD ¼ dp2 4 2 where Cd is the drag coefficient, given by Ref. [34]: 8 24 > > Re < 0:1ðStokes’s lawÞ > Re > > > > > > > < 24 1 þ 3 Re þ 9 Re2 lnð2ReÞ 0:1 < Re < 2 Re 16 160 Cd ¼ > > > 24 > > 1 þ 0:15Re0:687 2 < Re < 500 > > > Re > > : 0:44 500 < Re < 2 105 (6) When the size of an aerosol particle approaches the mean free path of gas molecules (0.066 mm for air at 1 atm and 20 C), the assumption of zero relative 4
Other assumptions include incompressible fluid, no walls or other particles nearby, constant motion, zero fluid velocity at the particle’s surface, and rigid spherical particles. These assump tions work well in most cases for aerosols. See Hinds [29] for more detailed discussions.
10
Developments in Surface Contamination and Cleaning
TABLE 1.1 Particle Reynolds Numbers (Rep) Calculated for Particles of Different Diameters Falling at their Terminal Settling Velocities in Air at 20 C, Pressure ¼ 1 atm, and Gravitational Acceleration ¼ 980 cm s2 Particle diameter (mm)
Rep
0.1
5.8 10
9
1
2.3 10
6
10
2.0 10
3
20
1.6 10
2
50
0.25
100
2
300
20
fluid velocity right at the particle surface fails, which could lead to significant errors. In this case, the actual drag force is smaller than predicted by Stokes’s law. A correction parameter, the Cunningham correction factor Cc, is introduced to correct for this ‘‘slip’’ phenomenon for small particles, and the corrected Stokes’s law becomes: FD ¼
3pmVdp Cc
dp l 2:514 þ 0:800 exp 0:55 Cc ¼ 1 þ l dp
(7) (8)
where l is the mean free path of gas molecules. The empirical formula of equation (8) allows the extension of Stokes’s law to aerosol size below 0.01 mm. It is important to include the slip correction factor for particles whose diameter is smaller than about 10 times the mean free path of gas molecules.
3.4. Gravitational Settling Gravity imposes an overall downward drift on particles, which contributes an enhanced deposition flux on surfaces with an upward component to their orientation. The gravity force exerted on a particle is: Fg ¼
p 3 d p rp r a g 6
(9)
where g is the acceleration due to gravity (980 cm s 2 at the Earth’s surface), and rp and ra are particle and air densities respectively. The
11
Particle Deposition onto Enclosure Surfaces
buoyancy effect usually can be neglected because ra is commonly much smaller than rp. The terminal settling velocity of a particle, vts, is established when the gravity force is balanced by the fluid drag force, and it is expressed as [28,29]: vts ¼
Cc rp dp2 g 18m
(10)
As suggested in equation (10), the particle terminal settling velocity increases rapidly with particle size as it is proportional to the square of particle diameter for supermicron particles (where Cc ¼ 1). Assuming unit particle density, for instance, the settling velocities for 1 and 10 mm particles are 3.50 10 3 and 3.05 10 1 cm s 1 respectively.
3.5. Thermophoresis In the presence of a temperature gradient, aerosol particles are driven from the high to low temperature regions; this transport process is known as thermophoresis. The phenomenon was first reported in the nineteenth century [35] and its quantitative descriptions were published by Watson in 1936 [36] and Zernik in 1957 [37]. Thermophoretic force arises from asymmetrical interactions of an aerosol with the surrounding gas molecules in a temperature gradient, where gas molecules on the warm side bombard the particle with higher average momentum than those on the cooler side. Acting in the direction of decreasing temperature, thermophoretic force, Fth, can be expressed as [38]: Fth ¼
3pm2 dp H dT ra T dy
(11)
where dT/dy is the temperature gradient and H is the thermophoretic force coefficient, given by: ka =kp þ 4:36l=dp 2:34 (12) Hy 1 þ 2ka =kp þ 8:72l=dp 1 þ 6:84l=dp where ka and kp are the thermal conductivities of air and the particle material respectively. When the thermophoretic force on a particle is balanced with fluid drag, the thermophoretic velocity is obtained as: vth ¼
Cc nH dT T dy
(13)
Contrary to gravitational settling and Brownian diffusion, which are strong functions of particle size, the thermophoretic velocity is nearly independent of
12
Developments in Surface Contamination and Cleaning
particle size for particles smaller than 1 mm. This occurs because both the thermophoretic force and fluid viscous drag have approximately the same dependence on particle diameter. For larger particles, the thermophoretic velocity decreases with increasing particle size in a manner that depends on the relative thermal conductivities of the particle and of air.
3.6. Electrostatic Force A charged particle migrates in an electrical field due to the Coulomb force FC, which is given by: FC ¼ qE ¼ ne eE
(14)
where q is the charge on the particle, E is the electric field strength, ne is the number of electrons of deviation from the electrically neutral state (including sign), and e is the charge of a single electron (1.6 10 19 C). In the absence of an electrical field, charged aerosols will migrate towards or away from a conducting surface owing to the image force, dielectric force, and dipole dipole force, although these forces are much weaker than the Coulomb force. The overall electrostatic force on a charged particle can be predicted by [39]: FE ¼ qE
qEdp3 3p30 dp6 E2 q2 þ 16p30 y2 16y3 128y4
(15)
where 30 is the permittivity of a vacuum (8.85 10 12 C2 N 1 m 2), and y is the normal distance from a surface. The terms on the right-hand side of equation (15) account for the Coulomb force, image force, dielectric force, and dipole dipole force respectively. When an electric field is present, the Coulomb force dominates and both the dielectric and dipole dipole forces are negligible. In the absence of an electric field, however, the only electrostatic force responsible for particle motion is the image force, as suggested in equation (15). The image force, which only occurs near a conducting surface, is always directed toward a surface and may dominate over Brownian diffusion and turbulent dispersion when extremely close to a surface. On electrically insulating materials, charges may accumulate on the surfaces, and this gives rise to electric fields, which in turn affect deposition of charged particles. It should be noted that aerosol particles acquire or lose their charges through random collisions with airborne ions, which are formed by ubiquitous ionizing radiation. In the absence of an electric field, the processes of aerosol acquiring and losing charges will eventually lead to an equilibrium charge state called Boltzmann equilibrium. The maximum number of charges carried by a particle depends on the size of particles. Table 1.2 shows the distribution of charges carried by aerosols at Boltzmann equilibrium. Take 10 mm particles, for
13
Particle Deposition onto Enclosure Surfaces
TABLE 1.2 Distribution of Charge on Aerosols at Boltzmann Equilibrium Particle diameter (mm)
Percentage of particles carrying the indicated number of charges 1
0
þ1
0.01
0.3
99.3
0.3
0.02
5.2
89.6
5.2
0.6
19.3
60.2
19.3
0.6
0.3
4.4
24.1
42.6
24.1
4.4
0.3
< 3
3
0.05 0.1
2
þ2
þ3
>þ3
0.2
0.3
2.3
9.6
22.6
30.1
22.6
9.6
2.3
0.3
0.5
4.6
6.8
12.1
17.0
19.0
17.0
12.1
6.8
4.6
1.0
11.8
8.1
10.7
12.7
13.5
12.7
10.7
8.1
11.8
2.0
20.1
7.4
8.5
9.3
9.5
9.3
8.5
7.4
20.1
5.0
29.8
5.4
5.8
6.0
6.0
6.0
5.8
5.4
29.8
10.0
35.4
4.0
4.2
4.2
4.3
4.2
4.2
4.0
35.4
Reproduced with permission from Hinds [29], John Wiley, 1999.
example: at equilibrium, only 4.3% of the particles are electrically neutral, and nearly 70% of the aerosol particles carry more than three charges on a single particle (either positive or negative). For highly charged aerosols in an electric field, the drift velocity resulting from electrostatic forces on particles can be thousands of times greater than that from the gravity force.
3.7. Turbophoresis Turbopheresis is a phenomenon in which the net migration of particles occurs from regions of high to low turbulence intensity, i.e. toward surfaces where the turbulent velocity fluctuations decrease to zero. Physically, particles in regions of high turbulence intensity acquire sufficient fluctuating velocity components, which enable them to drift to regions where the turbulence intensity is too low to send them back with sufficient momentum for the return journey. Based on the analogy between Brownian motion and turbulent diffusion, as well as energy transfer from the fluid to the particle, the mathematical expression for turbophoretic velocity Vt was first proposed by Caporaloni et al. [40] as: d v0y 2 Cc rp dp2 (16) Vt ¼ 18m dy
14
Developments in Surface Contamination and Cleaning
where v0y is the fluctuating particle velocity normal to the surface. Reeks [41] deduced the term for the turbophoretic effect and arrived at the same expression in a much more rigorous approach (a special closure of the Liouville particle equation of motion). Distinctly different from turbulent diffusion, turbophoresis is attributed to the interaction between particle inertia and the inhomogeneous turbulent flow field, and the mass transfer takes place against the gradient in turbulence intensity, even in the absence of a concentration gradient. The effect of turbophoresis is only significant for particles with sufficiently high inertia and is typically only relevant in the vicinity of a surface where the gradient in turbulence intensity is high. For near-surface inertial particles under highly turbulent flow scenarios, particle transport models accounting for turbophoresis have shown good agreements with experimental measurements [42 44].
4. PARAMETERS FOR PARTICLE DEPOSITION CHARACTERIZATION To characterize the rate of particle deposition in an enclosure, deposition velocity (vd) and particle deposition rate (b) are the two parameters commonly reported in the literature. Here their definitions, derivations, physical implications as well as the distinctions between b and vd are discussed.
4.1. Deposition Velocity, vd Particle deposition, a mass transfer process of airborne particles on to surfaces, occurs by a first-order, irreversible mechanism. Being ‘‘first order’’ suggests that particle deposition flux is linearly proportional to the airborne concentration, and being ‘‘irreversible’’ means that particles simply adhere once they collide with surfaces. The aerosol deposition rate can be parameterized in terms of a mass transfer coefficient, known as the deposition velocity, vd. It is considered as the proportionality constant between the net aerosol flux J and the free-stream airborne concentration CN: vd ¼
J mass=area time length ½¼ ½¼ CN mass=volume time
(17)
The aerosol concentration CN is determined at a position sufficiently far away from the surface, i.e. core zone in an enclosure, so that the concentration should not vary greatly with positions. With dimensions of length per time, deposition velocity appears to represent an effective velocity, which incorporates all of the complexities of the particle deposition process. In other words, deposition velocity, as an aggregated term, comprises every aspect of particle deposition processes, including (1) aerodynamic transport of particles by turbulent diffusion from the well-mixed core
15
Particle Deposition onto Enclosure Surfaces
region to the thin boundary layer of air adjacent to the surface; (2) mass transfer of particles across the boundary layer, and subsequent uptake by the surface through a variety of particle deposition mechanisms. The magnitude of deposition velocity depends on factors that govern particle transport: particle size, the near-surface air turbulence, surface characteristics including orientation and texture (smooth vs. rough), air surface temperature difference, and the presence of an electric field near the surface. A literature review of the existing experimental findings concerning the influence of these factors on particle deposition will be presented in Section 6.
4.2. Particle Deposition Rate, b Another common parameter used to quantitatively characterize the particle deposition rate b in an enclosure is defined as [20]: dC dp ; t ¼ bC (18) dt where C(dp,t) is the aerosol number or mass concentration as a function of time t in the core region of the enclosure, and b is the particle deposition rate with the unit of time 1. Particle deposition rate is also commonly referred to as the aerosol removal rate or wall loss rate, because the deposition process removes particles from the gas phase onto the available wall surfaces. Equation (18) is written in the form of mass balance on particles suspended in air within an enclosure, with the following conditions: (1) no or negligible sources for particle generation (such as particle influx in the incoming air flow and nucleation); (2) no particle size change due to coagulation,5 condensation, and evaporation; and (3) deposition onto surfaces is the only particle loss mechanism. Note that the particle deposition rate is expected to be a function of particle size, as is the case for particle deposition velocity. Also note that equation (18) only accounts for particle loss attributable to deposition. In an enclosure, there are typically other processes occurring in parallel, such as air exchange, that influence the particle concentration and should be taken into account [6]. Experimentally, the aerosol concentration decays with time as expressed by equation (18) after initial mixing of aerosols in an airtight enclosure. Mathematically, the rearrangement of equation (18) gives: Ct ¼ C0 e
bt
(19)
where C0 is the initial aerosol concentration in the enclosure and Ct is the aerosol concentration at time t. When the air is well mixed, plotting ln C as 5
Particle coagulation, which contributes to aerosol size distribution change, occurs when the aerosol concentration is sufficiently high. When that occurs, coagulation may be an important mechanism to incorporate in equation (18), as seen in Okuyama et al. [23].
16
Developments in Surface Contamination and Cleaning
FIGURE 1.2 An illustrative example of aerosol concentration decay measurements of various particle diameters as a function of time in a 165 L cylindrical chamber. The particle deposition rates, obtained from the values of the slope, were 0.058, 0.14, 0.088, 0.28, 0.44, and 0.71 h 1 for particle diameters of 0.305, 0.073, 0.91, 1.74, 2.02, and 2.99 mm, respectively. Reprinted with permission from Chen et al. [45].
a function of time gives rise to a linear plot with the negative of the slope yielding the particle deposition rate b. Figure 1.2 represents an example of particle concentration decay as a function of time and the associated deposition rate. Similar to particle deposition velocity vd, b incorporates all the deposition processes that remove aerosols from being suspended in the enclosure.
4.3. Comparison of vd and b The major distinction between vd and b is that vd depends on surface orientations with respect to gravity (a ‘‘local’’ parameter), while b represents an average term over all available surfaces for deposition within an entire enclosure (a ‘‘global’’ parameter). For instance, in the context of a rectangular enclosure, the magnitude of vd for supermicron particles varies strongly with surface orientation, i.e. vd,floor > vd,wall > vd,ceiling, owing to the dominant contribution of gravitational settling over diffusion and other mechanisms.
Particle Deposition onto Enclosure Surfaces
17
The value of b can be related to the values of vd for all enclosure surfaces. By material balance, the deposition rate, b, is the surface-area-averaged vd multiplied by the surface-to-volume ratio (S/V) of the enclosure: Z dC C S ¼ vd ðsÞds ¼ C vd ¼ bC (20) dt V V where vd(s) is the particle deposition velocity onto the surface s. The integration is to be carried out over all the enclosure interior surfaces with the total surface area S, and vd is the area-weighted mean deposition velocity. It can be clearly seen that b, unlike vd, eliminates the explicit spatial dependence of particle deposition.
5. METHODS FOR MEASUREMENT OF PARTICLE DEPOSITION Owing to its importance for numerous engineering systems, the processes of particle deposition on surfaces have been widely studied experimentally. Here the general approach for measuring b and vd with respect to an enclosed system is presented with the aim of helping to gain insights into the experimental findings that will be presented in Sections 6 and 7 respectively.
5.1. Measuring Particle Deposition Rate, b Experimentally, the particle deposition rate b as a function of particle diameter is inferred from aerosol concentration decay over time, following equation (18) and as shown in Figure 1.2. To do so, monodisperse aerosols are usually generated in elevated concentrations so that sufficient time is available to take multiple data points on aerosol concentrations during the course of deposition experiments. When the test aerosols are polydisperse, the aerosol-measuring instruments will need to have particle size-resolved capability in order to determine the concentration decay rate as a function of particle diameter. The operation principles for some common particle sizing instruments, e.g. optical particle counter, time-of-flight aerosol spectrometer, electrical aerosol mobility analyzer, as well as aerosol generation methods, can be found in Refs [29,46,47]. Note that the degree of air turbulence plays an important role in influencing the rate of particle deposition. Therefore, the stirring of air within the experimental chamber, by either impeller mixing or fluid flushing, should be controlled consistently throughout the experiment. The scheme of determining particle deposition rate by monitoring airborne particle concentration decay over time does not involve direct measurements of deposited particles on surfaces, thus the measured deposition rate is ‘‘inferred’’. This type of experiment is relatively simple to perform and data analysis is
18
Developments in Surface Contamination and Cleaning
straightforward. However, the drawback of this indirect approach is the inability to differentiate the contribution of particle deposition on surfaces with respect to orientations and locations.
5.2. Measuring Particle Deposition Velocity, vd As suggested in equation (17), particle deposition velocity is determined by normalizing the deposition flux with the average aerosol concentration over the duration of particle deposition. Distinct from measuring the deposition rate constant b, evaluating vd requires a direct measurement of particle deposition, which involves the simultaneous determination of particle deposition flux and airborne particle concentration. In other words, the deposited particle mass will need to be recovered from the surface (or otherwise detected and quantified), and the aerosol concentration outside the boundary layer during the course of deposition will also need to be determined. For a typical experiment of this type, monodisperse aerosols need to be generated and injected into the experimental chamber to provide the size-specific deposition flux. To obtain the deposition velocity as a function of particle diameter, the same experimental procedures are repeated for other particle sizes of interest. The advantage of this direct measurement, albeit time-consuming and labor-intensive, is that spatially resolved particle deposition can be determined. A key challenge of this approach, however, is the need to accurately determine trace amounts of particles deposited on surfaces. The earliest method was performed by microscopically counting the particles on a small area of a test surface [48]. Despite the recent rapid advancement of optical and digital imaging techniques, it remains a tedious task to perform the counting, considering that only a very small surface area is available per microscopic frame, which in turn limits the possibility of sampling larger surface areas. To quantitatively determine deposited particle mass on surfaces, aerosols can be labeled to facilitate subsequent surface analysis. Two types of tracer aerosol detection techniques have been employed for such purposes: (1) fluorescence spectroscopy technique and (2) neutron activation analysis (NAA) or proton-induced X-ray emission (PIXE). With respect to the former method, monodisperse aerosols are generated with fluorescent materials. After deposition, the fluorescent particles on the test surface are extracted with a solution of known volume. The resulting fluorescence intensity of the solution is proportional to the collected particle mass from the surface [49 51]. In order to accumulate sufficient fluorescent particle mass for analysis, the experimental deposition time, particularly for submicron particles, could easily exceed 100 hours [49]. Contrary to the chemistry-based fluorescence technique, NAA or PIXE relies on the physics of atoms in the aerosol materials. In NAA, the energy of the induced radioactivity through neutron bombardment is characteristic of a particular element, while PIXE involves the measurement of the characteristic
Particle Deposition onto Enclosure Surfaces
19
X-ray emission via high-energy proton excitations of the elements in the sample. Methods using NAA and PIXE are semi-invasive, i.e. only the particlebearing surface needs to be removed for the analysis from the test chamber, and removal of deposited particles from the particle-bearing surface is not required. To minimize interfering effects from various elements, rare earth elements such as dysprosium (164Dy) and indium (115In) can be incorporated into the aerosols for the purpose of subsequent NAA analysis [52 54]. Owing to the extremely high sensitivity, for instance, a 164Dy mass of the order of 10 10 g can be easily detected, hence the experimental deposition time can be significantly reduced (e.g. 15 20 min [53]). The drawback of this technique, however, is that specialized facilities, such as a nuclear reactor for neutron activation, are required.
6. REVIEW OF EXPERIMENTAL STUDIES A significant body of experiments has been devoted to exploring particle deposition in enclosures since the late 1940s due to the relevance to material deterioration and the implications to human health risk, etc. Table 1.A1 in the Appendix highlights a chronological summary of the laboratory investigations pertaining to particle deposition in an enclosed environment. The processes of particle deposition onto surfaces are complex, multifaceted phenomena, as they vary strongly depending on the characteristics of particles, airflow patterns, as well as surface properties. In brief, the processes of particle deposition consist of two mechanisms in series: (1) particle transport from the core region in the enclosure to the boundary layer adjacent to the surface, and (2) subsequent deposition onto the surface. Consequently, both properties of airflow and the surface play a crucial role in governing the rate of particle deposition in steps (1) and (2) respectively. The degree of airflow turbulence controls how rapidly particles migrate to the proximity of a surface, while the surface characteristics, such as orientations and roughness, determine how readily particles interact with the surface. As will be seen soon, particle size is the most important parameter governing the motions of particles in a gas phase associated with the transport properties. Here the important physical factors that influence particle deposition onto enclosure surfaces are summarized from the existing literature.
6.1. Particle Characteristics Affecting Particle Deposition Particle size is key in the determination of particle deposition. The dimension of particles in the gas phase is commonly characterized in terms of aerodynamic diameter, which refers to an equivalent diameter of a unit density spherical particle with the same terminal settling velocity as the particle being measured [28,29]. Therefore the concept of aerodynamic diameter has taken particle shape and density into account, and it has successfully captured the
20
Developments in Surface Contamination and Cleaning
FIGURE 1.3 Comparison of experimentally obtained particle deposition rates as a function of particle diameter. The values of particle deposition rate b exhibit distinct minima for 0.1 1 mm particles, as indicated in the V shaped curves.
aerodynamic behavior of airborne particles in many systems of interest.6 The notion of aerodynamic diameter has been commonly used in the aerosol literature, including this chapter. Figure 1.3 illustrates the characteristic V-shaped curves of experimentally determined particle deposition rates as a function of particle diameter. The large variability of the data, which can be up to two orders of magnitude, reflects the fact that other factors, such as airflow turbulence and surface characteristics, also play a significant role in influencing the extent of particle deposition in enclosures. As mentioned earlier, when particles are not electrically charged, the rate of particle deposition is governed by gravitational settling, Brownian motion of particles, as well as turbulent motion of fluid in 6 The behavior of nonspherical particles in shear flows might not be properly characterized using the concept of aerodynamic diameter owing to their complex rotational and translational motions not accounted for in spherical particles. Thus the prediction and measurement of deposition using aerodynamic diameters for nonspherical particles under this scenario would be an average outcome contributing from their stochastic behaviors in the turbulent flow.
Particle Deposition onto Enclosure Surfaces
21
which particles are entrained. For particles smaller than 0.1 mm, Brownian and turbulent diffusion control particle deposition. On the other hand, gravitational settling becomes dominant in removing particles from the gas phase with the increase of particle size, e.g. >1 mm. As a result, a minimum of particle deposition rate occurs for particles of 0.1 1 mm, which are generally referred to ‘‘accumulation mode’’ particles, when neither of the mechanisms works effectively to cause particle deposition (Figure 1.3). With respect to particle composition, experiments have shown that the material of particles makes no difference in the measured particle deposition rates [23]. Also, as seen in Table 1.A1, various aerosol materials have been employed for particle deposition experiments, and the choice of aerosol compositions does not alter the outcome.
6.2. Airflow Characteristics Affecting Particle Deposition Carried by large-scale eddy currents, airborne particles are brought from one point to another by turbulent diffusion. Eddy currents can be induced by natural convection (e.g. temperature gradient) and forced convection (e.g. mechanical stirring and fluid flushing).
6.2.1. Forced Convection Early experiments observed that increasingly turbulent air motion in an enclosed space was related to enhanced particle deposition onto surfaces, and the empirical parameters describing such phenomena were obtained as a function of the air stirring intensity [55]. Corner and Pendlebury [56] later explained these empirical observations based on the theoretical grounds, and concluded that air turbulence was one important factor that influenced the rate of particle deposition. Subsequent experiments help to shed light on the dependence of particle deposition rates on the air turbulence level in the enclosures. Figure 1.4 represents four laboratory results of particle deposition rates with respect to different air mixing scenarios. It can be seen that enhanced particle deposition is attributed to increased air turbulence, which was done by either fluid flushing [57] or fan stirring [23,58,59]. The enhancement of particle deposition onto enclosure surfaces results from more effective aerodynamic mass transfer at higher airflow turbulence, which brings aerosol particles more rapidly toward a surface. In addition, as the turbulent fluctuation of air motion becomes significant, particle inertia may contribute to particle deposition, and inertial transport of particles through the boundary layer could be potentially enhanced by nearsurface turbulent bursts [44]. Inertially induced turbulent deposition could be particularly important for large particles.
22
Developments in Surface Contamination and Cleaning
FIGURE 1.4 Dependence of experimentally measured particle deposition rates on the degree of airflow turbulence, as characterized by the fan stirring rate (in rpm, revolutions per minute) and the air exchange rate (incoming airflow rate divided by the enclosure volume, in units of h 1). The interior surface area to volume ratio, S/V, is indicated for each test chamber. Within the same experimental system, enhanced particle deposition is consistently observed at higher air turbulence levels.
6.2.2. Natural Convection Natural convection arises from temperature differences among air parcels, or heat transfer at surfaces (i.e. surface-to-air temperature difference). In the absence of forced convection, natural convection becomes the only means of air mixing inside enclosed spaces. Cheng [59] conducted particle deposition experiments under isothermal and still-air conditions with no apparent air movements detected7 in the chamber, but still found the deposition data well described by the homogeneous turbulence model assuming the core region was well mixed (see Section 7.1 for further discussion). Figure 1.5 shows a comparison of particle deposition rates as a function of particle size under natural convection conditions from different experiments [23,45,59 61]. First, notice that particle size remains the predominant parameter governing the deposition rates, as indicated by the distinct V-shaped 7
The detection limit in these experiments was about 1 cm s 1.
Particle Deposition onto Enclosure Surfaces
23
FIGURE 1.5 Comparison of experimentally measured particle deposition rates as a function of particle size under natural convection conditions. The interior surface area to volume ratio, S/V, is indicated for each experimental chamber. DT in Chen et al. [45] refers to the temperature difference between the top and bottom walls (e.g. Tb Tt 10 C) of the test chamber.
curves. Secondly, the scatter of the data in these experiments may be attributed to: (1) the S/V ratio, as higher surface area normalized by volume translates to more surfaces available for particles to deposit on, and (2) temperature gradient within the enclosure, as it dictates the extent of convective mixing. Chen et al. [45] and Cheng [59] specifically documented the temperature measurements inside their test chambers, ensuring that the experimental systems were either isothermal or otherwise reported. The temperature data from the other deposition measurements are either insufficient or unavailable to draw any useful information.
6.3. Surface Characteristics Affecting Particle Deposition After airborne particles enter a boundary layer, they may be transported to the surface by means of a variety of deposition mechanisms, as described in Section 3. It remains a good approximation and has been demonstrated experimentally that adhesion of micrometer-sized (or smaller) particles is complete and irreversible once they come into contact with a surface [62]. The van der Waals force is the predominant adhesion force between an aerosol particle and any surface (including another particle), although electrostatic and capillary forces may be important as well [29].
24
Developments in Surface Contamination and Cleaning
Surface characteristics play a role influencing particle deposition as they often affect particle surface interactions within the boundary layer. The dependence of particle deposition on surface characteristics has been explored experimentally by manipulating various factors such as surface orientations and roughness. Below, an overview is provided concerning experimental deposition studies with respect to various surface characteristics.
6.3.1. Surface Orientation: Horizontal vs. Vertical The effect of gravity contributes to the major difference in terms of particle deposition flux onto surfaces of various orientations, e.g. a horizontal upward-facing surface (floor) as opposed to a vertical (wall) or a horizontal downward-facing surface (ceiling). Briefly, the mechanism of gravitational settling is primarily responsible for particle deposition on the floor, while turbulent and Brownian diffusions are dominant deposition processes to walls and ceiling. Particle flux with respect to surfaces of various orientations can only be studied by means of spatially resolved deposition velocity measurements, as described in Section 5.2. Byrne et al. [52] employed the NAA technique to recover the particle mass deposited on various surfaces under forced convection in their cubic test chamber (2 2 2 m3). They reported that proportionately more of the overall particle deposition flux was found on the floor (and thus less on walls) as particle size increases, as summarized in Table 1.3. Thatcher et al. [49] used fluorescent tracer particles to measure particle deposition velocity with respect to different surfaces under natural convection flow conditions in their cubic enclosure (1.2 1.2 1.2 m3). In addition to examining the effect of different surface orientations, the interior surface temperatures of the chamber were independently controlled at fixed surface-toair temperatures to investigate the thermophoretic effects on particle deposition (see Section 6.3.3 for more discussion). In brief, the experimental chamber was
TABLE 1.3 Relative Contributions of Particle Deposition Flux Experimentally Determined for the Horizontal and Vertical Surfaces as a Function of Particle Diameter in Byrne et al. [52] Flux on horizontal surface (floor)
Flux on vertical surface (one wall)
0.7
57%
9%
4.1 10
5
2.5
68%
8%
6.2 10
5
4.5
72%
7%
1.1 10
4
5.4
80%
5%
2.0 10
4
Particle size (mm)
*Accounting for particle deposition to all chamber surfaces.
Average deposition velocity* (m s1)
Particle Deposition onto Enclosure Surfaces
25
heated on the floor and one vertical wall, and cooled on the ceiling and the opposite wall. As long as the surface-to-air temperature difference is kept identical, the thermophoretic effect is considered to be very similar, allowing comparisons to be made concerning particle deposition on surfaces of different orientations. Figure 1.6 presents a comparison of particle deposition velocity experimentally obtained from vertical and horizontal surfaces at the same
FIGURE 1.6 Comparison of experimentally measured deposition velocity for surfaces of various orientations in a cubic chamber under natural convection. The surface temperatures in (a) and (b) were kept at 1.5 K higher and 1.5 K lower than that of air respectively. The values of deposition velocity plotted here were obtained from Table 2 in Thatcher et al. [49]. The asterisk in (b) denotes that the deposited particle mass on the ceiling was under the detection limit.
26
Developments in Surface Contamination and Cleaning
temperature (þ1.5 K in (a) and 1.5 K in (b) relative to the air temperature). One of the distinctive features in Figure 1.6a is that particle deposition velocity onto the horizontal upward-facing surface (i.e. floor) increases as particle size increases, in agreement with theoretical predictions and experimental results in Byrne et al. [52]. Also note in Figure 1.6a that the measured deposition velocities onto vertical surfaces (i.e. wall) were increasingly lower than those on the floor by one to three orders of magnitude with the increase of particle size. Due to the diminishing contribution from gravitational settling, the deposition velocity to the horizontal downward-facing surface (i.e. ceiling) also decreases for increasing particle size, as shown in Figure 1.6b.
6.3.2. Surface Texture: Smooth vs. Rough Surface texture complicates the process of particle deposition. The microscale roughness elements can alter the near-surface turbulent structures and reduce the boundary layer thickness, which in turn influences particle deposition. Roughness can also influence diffusive deposition even when the fluidmechanical properties are not disturbed. The roughness elements can extend into and across the particle concentration boundary layer (which can be much thinner than the viscous sublayer), and this can expose the elements to higher particle concentration, thus enhancing the deposition rate. In addition, particle deposition may be further enhanced by more sites within the roughness elements available for impaction provided that particle inertia is sufficient. Experimental data regarding the effect of surface roughness on particle deposition within an enclosure are sparse. Nonetheless, a large collection of literature exists for particle deposition onto smooth and rough pipe walls, allowing some clues to be inferred. Sippola [15] provides a comprehensive review of this issue. Based on some assumptions and the measurements from a stirred 2.6-L cylindrical chamber, Shimada et al. [63] proposed a semi-empirical model to explain the dependence of surface roughness height on particle Brownian and turbulent diffusive deposition, and to estimate the rate of particle deposition. Figure 1.7 illustrates the influence of surface roughness level on the experimentally measured deposition rates for 0.01 0.2 mm particles under the same turbulent mixing conditions (500 rpm). They concluded that as the surface roughness becomes significant, particle deposition tends to be influenced by the turbulent diffusion very close to the surface, leading to the enhancement of particle deposition. Later Shimada et al. [65] further refined the model, and reported that the experimentally determined deposition rates agreed well with those reproduced in model calculations when the two-dimensional configuration of surface roughness was taken into account in calculating particle concentration above a rough surface. Since their modeling approach focused on diffusive deposition of 0.01 0.2 mm particles, the mechanisms of gravitational settling and inertial impaction were not included.
Particle Deposition onto Enclosure Surfaces
27
FIGURE 1.7 The dependence of experimentally measured particle deposition rates on the surface roughness. The length scale of 11.2, 119.2, and 204.8 mm in Shimada et al. [63] refers to the average height of roughness of the sandpapers placed on the interior surface of the enclosed tank. The S/V ratios of the experimental chamber in Shimada et al. [63] and Abadie et al. [64] are 47 and 10 m 1 respectively.
Thatcher and Nazaroff [50] studied the influence of different surface roughness on particle deposition under natural convection in a 1.8 m3 cubic enclosure. They found that deposition of small particles (~0.2 mm) was relatively insensitive to surface textures, but more pronounced effects were observed with increasing particle size. In their experiments, the measured deposition onto the roughest vertical8 surface for 1.3 mm particles was five times greater than deposition to a vertical smooth surface. They also noted that the surface roughness had more substantial effects for vertical surfaces than horizontal, and for warm surfaces than cool. Various wall treatments on all interior surfaces were placed in a cubic chamber (0.6 0.6 0.6 m3) to measure deposition rates in Abadie et al. [64] for particles of 0.7, 1.0, and 5.0 mm under fan mixing conditions. As shown in Figure 1.7, the measured particle deposition rates were found to increase with the following order of increasing surface roughness level: linoleum, smooth wall paper, rough wall paper, and carpet. Lai et al. [66] conducted deposition experiments in an 8 m3 test chamber under different fan speeds using monodisperse particles from 0.7 to 5.4 mm. 8
Vertical surfaces are studied to exclude the direct influence of gravity on deposition.
28
Developments in Surface Contamination and Cleaning
They observed that, at the highest fan speed, the average ratio of particle deposition flux onto a vertical rough surface over smooth surface increased from 1.05 for 0.7 mm particles to 1.1, 1.6, and 2.4 for particle sizes 2.5, 4.3, and 5.4 mm respectively. Since gravity does not play a role in particle deposition onto vertical walls, surface roughness and perhaps particle inertia for larger particles are believed to contribute to the enhanced particle deposition in the experiments. In a room-sized setting (2.2 2.7 2.4 m3), Thatcher et al. [67] reported on the influence of particle deposition rates upon different furnishing levels (bare, carpeted, and fully furnished with chairs, curtains, etc.) for particle sizes from 0.5 to 10 mm. The use of carpeting and furniture can be considered to increase the average roughness of surfaces (although the airflow pattern will be different for the furnished scenario due to additional obstructions inside the room; also increased surface area-to-volume ratio). Across all particle sizes measured, the measured deposition rates were observed to have a consistently descending trend with respect to the furnishing or roughness level: furnished > carpeted > bare. Lai and Nazaroff [51] placed glass plates and sandpapers of different grades on the vertical walls of a 1.8 m3 cubic chamber to measure the deposition velocities of supermicron particles under forced convection conditions. They reported that the experimentally determined deposition velocity appeared to reach a fairly steady value for particles larger than 7 mm. In addition, the deposition velocity was observed to increase with increasing roughness grade, albeit the increments with increasing surface roughness were not as significant as the influence of particle size on deposition.
6.3.3. Surface Temperature: Warm vs. Cold As discussed in Section 3.5, thermophoretic forces on airborne particles can be induced by temperature differences between a surface and air. In the presence of a temperature gradient, particles always move toward the direction of lower temperature. As a consequence, particles tend to preferentially deposit onto a cold surface over a warm one due to thermophoresis. Numerous experimental studies have examined the effectiveness of exploiting thermophoresis to minimize particle deposition on silicon wafers for microcontamination control [17 19]. However, experimental data for particle deposition on cold and warm surfaces with respect to an enclosed volume are sparse. The experiments performed by Thatcher et al. [49] provided valuable insights into particle deposition under the influence of surface-to-air temperature differences within an enclosure. Figure 1.8 represents the experimentally measured deposition velocity to the vertical surfaces as a function of the surface-to-air temperature difference (1.5 and 10 K respectively) for five different particle sizes. The deposition velocity predicted theoretically by Nazaroff and Cass [24,68] was also plotted for comparison.
29
Particle Deposition onto Enclosure Surfaces
cool wall
warm wall
10-4
10-7
* 10-10 10-4
0.1 μm
Deposition velocity, m/s
10-7
10-10 10-4
0.5 μm Experimentally measured deposition velocity
10-7
10-10 10-4
Theoretical predictions for isolated vertical flat plate 0.7 μm
10-7
10-10 10-4
1.3 μm
10-7
10-10
2.5 μm 10
1.5
1.5
10
Surface-to-air temperature difference, K FIGURE 1.8 Experimentally measured deposition velocities to the vertical walls of a 1.2 1.2 1.2 m3 enclosure as functions of particle diameter and surface to air temperature difference. Depo sition velocities predicted theoretically by Nazaroff and Cass [24,68] for a vertical isolated flat surface are also plotted for comparison. Each diamond symbol represents the average value obtained by 10 local surface extractions over the central portion of the cool or the warm wall from a single experiment. The error bars span the maximum and minimum values of measurements at each location, while an arrow is used to indicate that the lower limit of the 95% confidence interval for the mean was less than the detection limit. The asterisk denotes that the deposited particle mass was under the detection limit. Reprinted with permission from Thatcher et al. [49].
As shown in Figure 1.8, the thermophoresis effect on deposition velocity appears to be diminishing as particle size increases. For instance, the experimentally obtained deposition velocity to the cool vertical surface (10 K) for 0.1 mm particles is more than three orders of magnitude higher than to the warm
30
Developments in Surface Contamination and Cleaning
vertical surface (þ10 K). On the other hand, the measured deposition velocities to the cool and warm walls for larger particles (1.3 and 2.5 mm) are approximately the same. Furthermore, for particles larger than 0.5 mm, the measured deposition velocity on the warm wall at þ10 K is higher than at þ1.5 K. This observation is counterintuitive since the warmer surface is anticipated to enhance thermophoretic repulsion, thus reducing particle deposition. In summary, the deposition velocity determined experimentally in Thatcher et al. [49] shows relatively good agreements with the predictions for all five particle sizes studied when the wall surface temperature is cooler than the air. In contrast, the deposition velocity to the warm wall obtained in the experiments appears to be much higher than predicted, particularly for increasing particle size. The discrepancy between experimental findings and theoretical predictions may be attributed to the fact that the existing model does not account for all the factors influencing particle deposition under the settings of the experiments. The surface-to-air temperature difference influences not only particle thermophoretic velocity, but also the near-surface airflow pattern, which further introduces more complexities toward the understanding of the deposition processes. The complicated airflow pattern and flow instability in their chamber experiments are believed to play a role in contributing to the variations of the deposition velocity measurements. Other factors such as particle inertia and the effect of corners on airflow not being addressed in the model may be of importance to influence the deposition process.
6.3.4. In the Presence of an Electric Field Electrostatic forces may play an important role in influencing particle deposition. When the interior surface of an enclosed volume has the tendency to acquire electrostatic charge (e.g. made of poor conducting materials), a local near-surface electric field may develop, which in turn enhances the deposition of charged particles.9 Early experiments have shown that the presence of electric charges on surfaces can introduce additional variability to particle deposition measurements [69]. McMurry and Rader [70] demonstrated in their 60 m3 Teflon chamber experiments under natural convection that the enhanced deposition for 0.07 1 mm particles was attributed to electrostatic effects, as seen in Figure 1.9. In another set of Teflon film bag (250-L) experiments, the deposition rates of singly charged particles could be significantly increased as compared to particles at Boltzmann distribution [70], as shown by the symbols V and ; in Figure 1.10. For instance, a nearly 30 times deposition enhancement was observed for the 0.1 mm singly charged particles in comparison to particles with Boltzmann charge distribution. In addition, their experimental data indicated that the 9 As mentioned in Section 3.6, airborne particles become charged by collision with air ions and carry certain charges depending on their particle size according to Boltzmann distribution.
Particle Deposition onto Enclosure Surfaces
31
FIGURE 1.9 Comparison of measured particle deposition rates in a 60 m3 Teflon smog chamber and ke denote the estimated average with modeling predictions by Crump and Seinfeld [71]. E electric field strength on the chamber interior surface and the parameter of airflow turbulence intensity respectively. Note that the experimental data exceeded predicted values for 0.07 1 mm particles when electrostatic effects were not taken into account. Reprinted with permission from McMurry and Rader [70].
deposition rates were nearly identical for oppositely charged particles of the same size. The laboratory results agreed satisfactorily with their model, which incorporated electrostatic drift as an additional transport mechanism. As explained in the model, deposition for particles larger than 1 mm and smaller than 0.05 mm were still dominated by gravitational settling and Brownian/ turbulent diffusion respectively, while electrostatic effects were important factors influencing deposition of 0.05 1 mm particles. Shimada et al. [72] performed deposition measurements for 0.02 0.2 mm particles in a stirred metal tank where the turbulence intensity and the electric field could be controlled. They found no difference in deposition rates for charged and uncharged particles when the surfaces were grounded. As the electric field strength increased, however, the enhancements of particle deposition rates were observed, as indicated in Figure 1.10. By directly collecting particle deposition mass onto vertical surfaces in a 1.8 m3 cubic chamber, Lai [73] reported that the measured particle deposition velocities onto acetate sheets for 3.5 9 mm particles were more than an order of magnitude higher than those onto glass and copper plates, owing to the electrostatic effects. The laboratory data indicated that deposition velocities for
32
Developments in Surface Contamination and Cleaning
FIGURE 1.10 Experimentally determined particle deposition rates as a function of particle diameter under various influence of electrostatic attractions. The data points of McMurry and Rader [70] were obtained in a 250 L Teflon film bag (natural convection; S/V 10.6 m 1), in 1 which the average electric field was estimated as 45 V cm via data fitting to their model. The electrostatic attractions in Shimada et al. [72] were established by applying known voltage (F) at the bottom of the enclosed tank (forced convection; S/V 47 m 1).
both glass and copper surfaces were comparable, and the use of antielectrostatic spray on the surfaces was found to lead to reduced particle depositions by minimizing the coulombic effect.
7. MODELING PARTICLE DEPOSITION AND THE EXPERIMENTAL VALIDATIONS Two major types of models, both theoretical-based and semi-empirical, have been proposed to quantify, predict, and explain particle deposition in enclosures. The first type, as will be discussed in Sections 7.1 and 7.2, involves first modeling the air turbulent structure adjacent to the enclosure surfaces, and particle transport is subsequently formulated accounting for gravity, diffusion, and other deposition mechanisms. One key challenge in this modeling approach lies in the determination of the near-surface particle eddy diffusivity, which is postulated to be related to the air turbulence intensity within the enclosure as well as the distance to the wall. The second type of modeling
Particle Deposition onto Enclosure Surfaces
33
approach, as will be described in Section 7.3, applies the well-known analogy between mass and heat transfer, and deposition rate of particles from the gas phase to the surface due to diffusion is estimated via the mass transfer correlation. In this section, the modeling developments based on these two approaches for predicting particle deposition rates in enclosures are summarized, together with the available experimental investigations to compare against the modeling analyses as well as the limitations associated with these models.
7.1. Homogeneous Turbulence Model Modeling efforts for studying particle deposition from turbulent flow onto enclosure surfaces were initiated in the early 1950s. Corner and Pendlebury [56] developed the first theoretical model accounting for particle deposition onto the surfaces of horizontal and vertical orientations in a rectangular enclosure where the air was homogeneously turbulent. Their model was derived based on the following key assumptions: (1) the air outside the boundary layer is homogeneously turbulent and aerosol concentration is uniform (as illustrated in Figure 1.1); (2) within the boundary layer of thickness d, the fluid motion is turbulent with random fluctuations, but the mean fluid motion is parallel to the surface; (3) the velocity gradient is linear within the boundary layer; (4) the mechanisms of gravitational settling and Brownian/turbulent diffusion are responsible for particle transport through the boundary layer; (5) turbulent diffusion dominates Brownian motion at the edge of the boundary layer; and (6) particle transport is quasi-steady state in the boundary layer. Under the above assumptions, the aerosol concentration in the boundary layer adjacent to the surface is governed by: dC v vC 3p þ D vts k$ ¼ 0 (21) vy dy vy where vts is the particle terminal settling velocity and k is the unit normal vector in the vertical direction. The first and second terms of equation (21) account for the processes of turbulent/Brownian diffusion and gravitational settling respectively. The boundary conditions for the above equation are: C ¼ 0 at y ¼ 0 C ¼ CN at y d
(22)
where CN is the bulk aerosol concentration outside the boundary layer. The particle concentration profile in the boundary layer can thus be solved from
34
Developments in Surface Contamination and Cleaning
equations (21) and (22). In the Corner and Pendlebury model, the particle eddy diffusivity is approximated using Prandtl’s mixing length theory as: 3p ¼ ke y2
(23)
where ke is the turbulence intensity parameter used to characterize the degree of turbulent mixing inside the enclosure. There is no a priori way to estimate the values of ke with available data; hence it is usually determined empirically by fitting experimental data into the theory. Corner and Pendlebury [56] suggested that ke be evaluated as: ke ¼ k
dU dy
(24)
where k is the von Ka´rma´n constant (usually taken as 0.4) and U is the component of the mean flow velocity parallel to the surface. The velocity gradient, dU/dy, is approximated by means of fluid drag force balance for a flat plate in Corner and Pendlebury’s model. Following the seminal work of Corner and Pendlebury [56], several additional model extensions have been proposed. Crump and Seinfeld [71] derived expressions for estimating particle deposition rate in an enclosure of arbitrary shape, and the analytical solutions for a spherical vessel were specifically formulated. They showed that, in an enclosure having only horizontal and vertical surfaces, the transport processes of gravitational settling and diffusion can be treated independently by vectorially summing the gravitational settling velocity to the deposition velocity associated with the diffusion process. For the inclined surfaces of a sphere, however, their derivations indicated that these two mechanisms are always intimately coupled to each other and cannot be separated. They proposed that the value of ke can be evaluated from the fluid energy dissipation rate, instead of velocity gradient in the boundary layer as suggested in Corner and Pendlebury [56]. Moreover, Crump and Seinfeld [71] derived the particle deposition rate with a more general form of 3p ¼ keyn, where n can be any number, which could be obtained by empirically fitting experimental data. They also demonstrated that the use of the exponent n ¼ 3, as suggested by Friedlander [28], could produce analytical expressions to predict deposition that are analogous to those with n ¼ 2. In summary, Crump and Seinfeld [71] calculated the particle deposition velocity onto the surface of an enclosure to be: vd ðqÞ ¼
v cos q q ts exp pvts cos q n sin pn n ke 3np
1
(25) 1
where q is the angle between normal vector of wall and gravity direction (rad).
Particle Deposition onto Enclosure Surfaces
35
To validate their own theoretical derivations, Crump et al. [74] performed particle deposition experiments in a spherical vessel. By fitting the data and using n ¼ 2, the turbulence intensity parameter ke in their study was evaluated in terms of the volumetric airflow rate into the chamber, which was related to the turbulence energy dissipation rate, as suggested in Okuyama et al. [75].10 The experimental results in Crump et al. [74] showed good agreement with the analytical expressions with respect to the dependence of particle deposition rates on particle size and turbulence intensity. Based on the existing model developed by Crump and Seinfeld [71], McMurry and Rader [70] incorporated electrostatic effects as an additional particle deposition mechanism to evaluate the particle deposition rate in an enclosure. Experiments were also performed to measure the deposition rate of neutral and singly charged aerosols in a 250-L Teflon bag and a 60 m3 Teflon smog chamber respectively. A good fit between the experimental results and the model predictions was obtained by setting n ¼ 2 and adjusting the values for ke and the mean electric field. Okuyama et al. [23] studied the deposition loss of monodisperse aerosols with particle diameters of 0.006 2 mm in a stirred cylindrical vessel. The turbulence intensity parameter ke in their experiments was calculated by the average energy dissipation rate per unit mass of air, and the parameters k and n were obtained by fitting the experimental data. The experimental deposition rates compared well with the model calculations by Crump and Seinfeld [71] when the eddy diffusivity was assumed to be proportional to the 2.7th power of the distance from the surface (i.e. n ¼ 2.7). In addition, more deviations from the model were observed for increasing particle size or flow turbulence intensity, and this is likely attributed to effects of enhanced particle inertia, which were not considered in the model. Cheng [59] measured the particle deposition rates of monodisperse particles ranging from 0.005 to 2 mm in a spherical chamber. Both the chamber temperature and air velocity profiles under various turbulence conditions were measured. The turbulence intensity in the experiments evaluated using both methods, including the velocity gradient as shown in equation (24) and energy dissipation rate [23,75], gave reasonable estimates of ke as well as the particle deposition rates. The data were well explained by Crump and Seinfeld’s model [71], except that the best-fitted estimate of n was approximately 2.8, instead of n ¼ 2 as supported in Crump et al. [74], McMurry and Rader [70], and Chen et al. [45]. By vectorially adding the deposition flux, Nazaroff and Cass [68] incorporated the thermophoresis effect into the model of Corner and Pendlebury [56] to estimate deposition velocity onto a vertical isolated flat plate in the The expression ke fð3=nÞ1=2 suggested in Okuyama et al. [75] originates essentially from Prandtl’s mixing length formula, where (3/n)1/2 is proportional to the r.m.s. velocity gradient in the boundary layer. The symbol 3 refers to the turbulence energy dissipation rate and n is the kinematic viscosity of the fluid. 10
36
Developments in Surface Contamination and Cleaning
presence of surface-to-air temperature difference. Their analysis showed that, as particle size increases, an increasingly pronounced difference in deposition velocity is predicted for the vertical warm relative to the cool walls. The experimental findings in Thatcher et al. [49] indicated that the model appeared to greatly underestimate the deposition velocity for a warm surface, especially for supermicron particles, as was explored in Section 6.3.3. In summary, the existing experimental data agree reasonably well with the theory of Corner and Pendlebury [56], including its extensions. However, there are still some issues that must be carefully examined before applying the model to various scenarios. For example, can ke and n be determined in any particular enclosure, without resorting to fitting the results of sophisticated particle deposition experiments? These two parameters are key and vary with experimental conditions; they ultimately influence the magnitude of the particle eddy diffusivity 3p, which directly affects the calculation of deposition rate and flux to the surface. Table 1.4 summarizes the values of n and ke determined from various particle deposition experiments. The scenario of n ¼ 2 is regarded as the ‘‘classical’’ form of Corner and Pendlebury [56] and Crump and Seinfeld [71]. This is also supported by some experimental data [45,70,74]. In other experimental findings, however, the best fit of the data to the model occurs at n ¼ 2.6 2.8 [23,59,60,76], close to n ¼ 3 as suggested by Friedlander [28] as well as Pandian and Friedlander [77] based on a theoretical perspective of the analogy between mass and heat transfer. In the definition of particle eddy diffusivity (in cm2 s 1, 3p ¼ keyn) when n ¼ 2, the dimension of ke(s 1) has the dimensions of a rate constant. Noninteger values of n lead to ke with a dimension of L2 nT 1, which not only lacks a solid physical foundation but also causes conceptual and practical problems when ke is to be evaluated based on information of velocity gradient or turbulent energy dissipation rate (both methods assuming n ¼ 2). Based on the rules of dimensional analysis, Benesˇ and Holub [78] suggested a modified formulation to avoid the dimensional inconsistency problem: n 2 y (26) 3p ¼ ke d d where d is the boundary layer thickness. Using this new expression has shown to yield good agreements with data from one experimental study [59]. However, it remains unresolved with respect to the evaluation method for n. Nevertheless, Crump and Seinfeld [71] noted that the choice of n value is of little importance from the standpoint of their theoretical derivations. Van Dingenen et al. [60] suggested that, after re-examining the data of Crump et al. [74] and McMurry and Rader [70], the exact value of n is trivial as long as an appropriate ke is determined in an independent way for instance, the evaluation of near-surface velocity gradient or energy dissipation rate, as suggested in Okuyama et al. [23,75].
Experimental Flow mixing method chamber
Particle size (mm)
n
ke (s1)
Method to evaluate ke
2
0.028 and 0.068 at two flow rates
Data fitting and later correlated ke to flow rates
Convective mixing by feeding air continuously
Spherical glass vessel, 0.024 118 L 0.794
Crump et al., 1983 [74]
2
6.4 10 3 for 250 L bag and 0.12 for 60 m3 smog chamber
Data fitting
Natural convection
250 L Teflon film bag 0.02 1.8 and 60 m3 Teflon film smog chamber
McMurry and Rader, 1985 [70]
0.004 and 0.02 at 0 and 10 C
Data fitting
Convective mixing by temperature gradient
Pyrex glass cylinder, 165 L
Chen et al., 1992 [45]
2
DT
0.04 3.0
Investigators
2.6
3.5 10
2
Data fitting
Natural convection
Spherical borosilicate 0.02 0.2 glass chamber, 230 L
Van Dingenen et al., 1989 [60]
2.6
7.1 10 107
3
Data fitting
Natural convection
Spherical glass chamber, 0.23 m3
0.001 0.3
Holub et al., 1988 [76]
Turbulent mixing (fan)
Cylindrical chamber, 3.9 m3
2.7
Ranging from 0.05 to 264, depending on the fan speeds
Estimated from fluid energy dissipation; assuming k 0.3
Turbulent mixing (fan) Cylindrical tank, and natural convection 2.6 L
0.006 2
Okuyama et al., 1986 [23]
2.8
0.0152 1.0 depending on the fan speeds
(1) Velocity gradient near the chamber surface; (2) energy dissipation by mixing; and (3) data fitting
Turbulent mixing (fan) Spherical aluminum and natural convection chamber, 161 L
0.005 2
Cheng, 1997 [59]
Particle Deposition onto Enclosure Surfaces
TABLE 1.4 Summary of the Values of n and ke Determined from Various Particle Deposition Experiments
37
38
Developments in Surface Contamination and Cleaning
7.2. Three-Layer Model Incorporating the information on the structure of near-surface turbulent diffusivity, Lai and Nazaroff [25] proposed a model in which only one parameter, friction velocity u*, is required to evaluate particle deposition under homogeneously turbulent conditions. The friction velocity is defined by: u ¼
1=2 sw 1=2 dU
¼ n ra dy y¼0
(27)
where sw is the shear stress at the surface, ra is the air density, and n is the kinematic viscosity of air. The parameter u* is intended to capture the turbulent characteristics in the vicinity of the surface. As suggested in equation (27), the friction velocity is related to the velocity gradient at the surface, dU/dy, which can be evaluated by a freestream air speed UN and a characteristic length of enclosure surfaces L [79]:
2 U L dU
0:074 ra UN N ¼ 2 dy y¼0 ra n n
1=5
(28)
Another approach of estimating friction velocity is to measure the velocity profile in the logarithmic flow region near the surface, i.e. the Clauser-plot method [80]. Within the turbulent boundary layer, the time-averaged velocity U as a function of distance from the surface y is expressed by: U 2:5u yUN þA (29) ¼ ln UN n UN where A is a constant. Thus the friction velocity can be inferred from the slope of the line by plotting measurements of U/UN versus the logarithm of (yUN/n). Commonly known as the law of the wall, a turbulent boundary layer consists of three distinct zones according to the velocity distribution as a function of the distance perpendicular to the wall [79]. The approach used by Lai and Nazaroff [25] to analyze particle deposition from turbulent flow is to examine the turbulent flow structure zone by zone, and to formulate particle transport equations for each zone, assuming (1) gravitational settling, Brownian and turbulent diffusion are responsible for the particle deposition processes; (2) constant particle flux in the concentration boundary layer; (3) particle eddy diffusivity is well represented by fluid turbulent viscosity; and (4) negligible surface roughness effects. Across the boundary layer, the expression for the deposition velocity was then integrated for each zone. The deposition velocity was evaluated for vertical, upward, and downward horizontal surfaces, and the firstorder deposition rates were derived for rectangular and spherical cavities respectively. This approach is somewhat complicated but remains practical to use.
39
Particle Deposition onto Enclosure Surfaces
Table 1.5 summarizes the equations required for calculating particle deposition along with explanations of parameters used in the model. Their model predictions compared well with published experimental data for deposition of 0.001 2 mm particles in a spherical enclosure [59], as seen in Figure 1.11. Lai and Nazaroff [25] indicated that their model calculations yield the best agreement with the model of Crump and Seinfeld [71] with n ¼ 2.95, very close to n ¼ 3, which is used to characterize convective heat transfer [77]. It is expected that the analogy between heat and mass transfer should hold when particle inertia is insignificant. Lai and Nazaroff [51] performed laboratory experiments to measure deposition of 0.9 9 mm particles from turbulent flow onto vertical surfaces of a cubic aluminum chamber. The experimentally measured particle deposition velocities for 9 mm particles were higher by a factor of 30 150 as compared to their own model predictions for friction velocities of 2.9 9.8 m s 1. The chamber walls were grounded, thus electrostatic force was considered negligible. They
TABLE 1.5 Summary of Equations Used for Particle Deposition Analysis in the Three-Layer Model Parameters
Equations
Integral I* I a b Deposition velocity, vertical surface
vd;v
Deposition velocity, upward horizontal surface
vd;u
Deposition velocity, downward horizontal surface
vd;d
3:64Sc2=3 ða
bÞ þ 39
p 1 ð10:92Sc1=3 þ 4:3Þ3 8:6 10:92Sc1=3 þ 3 tan1 p ln 1 2 Sc þ 0:0609 3 10:92Sc1=3 p þ 3 1=3 þ 1 ð10:92Sc þr Þ 10:92Sc1=3 1 2r p þ 3 tan ln 2 3 10:92Sc1=3 Sc1 þ 7:669 104 ðr þ Þ3 u I
vts 1
expð
vts I u Þ
vts expð
vts I u Þ
1
Continued
40
Developments in Surface Contamination and Cleaning
TABLE 1.5 Summary of Equations Used for Particle Deposition Analysis in the Three-Layer Modeldcont’d Parameters
Equations
Deposition rate, b rectangular enclosure
vd;v Av þ vd;u Au þ vd;d Ad V
Deposition rate, b spherical y enclosure
3 u 1 ð Þ½2D1 ðxÞ þ x where x 2R I 2
vts I u
Reprinted with permission from Lai and Nazaroff [25]. Nomenclature: Sc n/D particle Schmidt number n kinematic viscosity of air D Brownian diffusivity of particles rþ dpu*/2n dp particle diameter u* friction velocity vts terminal settling velocity of particles Av area of vertical surfaces Au area of upward-facing surfaces Ad area of downward-facing surfaces V enclosure volume R radius of spherical enclosure D1(x) Debye function defined by: R 1 x t dt D1 ðxÞ x 0 et 1
* The integral is evaluated analytically under the approximation that particle Brownian diffusivity, D, is negligible compared with eddy diffusivity for y+ 4.3, where y+ is the normalized distance from the surface. This approximation is accurate to 1% or better for particle diameters larger than 0.01 m m. For smaller particles, the integration must be carried out numerically. See the following for results:
Particle diameter, dp (mm)
Integral, I
0.001 0.0015 0.002 0.003 0.004 0.005 0.006 0.007 0.008 0.009 0.01
29.1 49.1 71.0 120.3 174.9 234.2 297.4 364.0 432.7 504.5 579.3
y
In the limit of small particles (negligible influence of gravitational settling), the expression simplifies to b ¼ 3u*(RI) 1. In the limit of large particles (negligible influence of Brownian diffusion), the expression simplifies to b ¼ 3vts(4R) 1.
41
Particle Deposition onto Enclosure Surfaces
Particle deposition rate β,s-1
10-2 Lai and Nazaroff [25] Cheng [59] 0 RPM u*= 0.9 cm/s 300 RPM u*= 0.2 cm/s u*= 3.1 cm/s 1000 RPM u*= 5.1 cm/s 1800 RPM
10-3
10-4
10-5
0.001
0.01
0.1
1.0
Particle diameter, μm FIGURE 1.11 Predictions of particle deposition rate as a function of particle diameter in the three layer model of Lai and Nazaroff [25] compared with experimental data [59]. Reprinted with permission from Lai and Nazaroff [25].
note that the discrepancy between experimental observations and model calculations is attributed to the inadequacies of the model, in which the key transport and deposition processes for supermicron particles may not be addressed appropriately. For example, particle inertia and shear-induced lift force11 are postulated to be potentially important players to enhance transport of larger particles through the relatively thin boundary layers and lead to subsequent surface deposition. Lai [81] later incorporated the mechanism of particle inertia into the three-layer model, by adopting the commonly accepted electrical resistance analogy in atmospheric dry deposition modeling [34], to simulate supermicron particle deposition from turbulent flow to a vertical surface. Treating the frictionvelocity as the fitting parameter and incorporating analyses accounting for surface roughness effects [82], the agreements between experimental data and modeling calculations were good for smooth surfaces, but less satisfactory for rough surfaces. The influence of electrostatic drift was considered as an additional particle deposition mechanism in the inertia-incorporated three-layer model by Chen and Lai [83]. The experimental findings involved with electrostatic effects in Lai [73] again demonstrated that the particle deposition is a complicated 11 A particle in a shear flow field may experience a lift force perpendicular to the main flow direction. This life force arises owing to particle inertia and is important for large particles. Lai and Nazaroff [51] postulated that shear induced lift force may be important in their experiments, in which a significant velocity gradient adjacent to the surface is expected owing to the parallel airflow pattern along the vertical walls of the chamber.
42
Developments in Surface Contamination and Cleaning
process, particularly when mechanisms other than diffusion and gravitational settling are involved. More detailed experiments or numerical simulation on near-surface airflow structure will be helpful to shed light on the dynamics of particle transport.
7.3. Mass Transfer Model As mentioned in Section 4.1, particle deposition velocity is equivalent to a mass transfer coefficient with the units of length per time (L T 1). Based on the analogies for mass and heat transfer [84], the correlations for transfer coefficients in mass and heat transport should be of the same form for particle mass transfer from turbulent flow to the enclosure surfaces by diffusion, and for convective heat transfer in jacketed vessels, assuming that the average roughness heights are immersed within the viscous sublayer (i.e. for smooth walls). Following this analogy, Pandian and Friedlander [77] proposed a semi-empirical mass transfer expression to estimate the rate of particle deposition due to turbulent and Brownian diffusion, in the form of the Sherwood Reynolds Schmidt correlation: Sh ¼ aðReÞ2=3 ðScÞ1=3
(30)
where Sh is the Sherwood number ¼ kmDc/D, km is the mass transfer coefficient or particle deposition velocity (m s 1), Dc is the chamber diameter (m), D is particle diffusivity (m2 s 1), Re is the Reynolds number for stirring ¼ NDs/n, N is the impeller speed (rpm, revolutions per minute) in a stirred chamber, Ds is the stirrer diameter (m), n is the kinematic viscosity of fluid in the chamber (m2 s 1), Sc is the particle Schmidt number ¼ n/D, and a is a constant. By fitting the experimental data in Okuyama et al. [23] to equation (30), Pandian and Friedlander [77] obtained a ¼ 0.63 as the best fit for particles smaller than 0.1 mm in diameter and Re no greater than 3000. Other laboratory results were also used to compare with equation (30) and good agreements were found, as shown in Figure 1.12. Using different sets of data, the values of fitted a have been found to be slightly different. By equating the expression of particle deposition rate in Crump and Seinfeld [71] to equation (30), Cheng [59] showed that the mass transfer equation proposed by Pandian and Friedlander [77] has an equivalent form as the Crump and Seinfeld model with n ¼ 3. The particle deposition rate estimated from the mass transfer correlation thus can be estimated as: b ¼ a
D S 2=3 1=3 Re Sc Dc V
(31)
where S and V are the interior surface area and the volume of the chamber respectively.
Particle Deposition onto Enclosure Surfaces
43
FIGURE 1.12 Mass transfer correlation of particle diffusional deposition in a smooth walled stirred chamber, with Sherwood number (Sh) plotted as a function of Re2/3Sc1/3. The dashed line of a 0.63 was obtained from fitting the data in Okuyama et al. [23], and the straight line of a 0.54 was obtained from fitting the data in Okuyama et al. [23] and Cheng [59]. Reprinted with permission from Cheng [59].
Note that this mass transfer correlation indicated in equation (30) is only applicable for evaluating particle deposition rate in the diffusional regime (e.g. dp < 0.1 mm) in small cylindrical or spherical vessels with smooth surfaces, assuming that Re can be defined as stated in equation (30). For larger particles and higher fluid turbulence, deviations from equation (30) were observed because diffusion was no longer the only mechanism for deposition and other processes, such as inertial drift, may contribute to particle transport to the surfaces.
8. SUMMARY The phenomenon of particle deposition onto surfaces in a confined environment is frequently encountered in numerous industrial and environmental settings, either as a desirable or an undesirable outcome. This chapter presented a literature review on the important physical factors that influence particle deposition in enclosures, as well as the available experimental techniques and modeling approaches used for characterizing the rate of particle deposition. Experimental findings compared with the model calculations were presented, and caveats with respect to the models were discussed. As was already known, and further substantiated by the scientific evidence presented here, the process of particle deposition is a complicated phenomenon. Transport behavior of particles to surfaces is governed by the nature of airflow
44
Developments in Surface Contamination and Cleaning
near surfaces, including the details of turbulence structure, the surface characteristics, and, importantly, particle size. The experimental studies reviewed in this chapter have revealed that the deposition rate varies broadly across conditions, and the measurement results are affected by various factors acting simultaneously. Direct measurements of particle deposition onto surfaces of interest are a challenging task to perform, and the contribution from various potential deposition mechanisms under realistic circumstances makes modeling work difficult. Nevertheless, further progress to elucidate the processes of particle deposition will require continued efforts to conduct more carefully controlled experimental investigations, in which the particle size, near-surface airflow conditions and the nature of surfaces are well characterized.
ACKNOWLEDGEMENTS The author thanks Professor William W. Nazaroff at UC Berkeley for his valuable comments.
REFERENCES [1] D.W. Cooper, Particulate contamination and microelectronics manufacturing: an introduc tion, Aerosol Sci. Technol. 5 (1986) 287. [2] B.Y.H. Liu, K.H. Ahn, Particle deposition on semiconductor wafers, Aerosol Sci. Technol. 6 (1987) 215. [3] S.E. Pratsinis, T.T. Kodas, M.P. Dudukovic, S.K. Friedlander, Aerosol reactor design: effect of reactor type and process parameters on product aerosol characteristics, Ind. Eng. Chem. Process Des. Dev. 25 (1986) 634. [4] J.D. Spengler, K. Sexton, Indoor air pollution: a public health perspective, Science 221 (1983) 9. [5] A.V. Nero, Controlling indoor air pollution, Scientific American 258 (1988) 42. [6] W.W. Nazaroff, A.J. Gadgil, C.J. Weschler, Critique of the use of deposition velocity in modeling indoor air quality, in: N.L. Nagda (Ed.), Modeling of Indoor Air Quality and Exposure, ASTM STP 1205, ASTM, Philadelphia, PA, 1993, pp. 81 104. [7] W.W. Nazaroff, G.R. Cass, Protecting museum collections from soiling due to the deposition of airborne particles, Atmos. Environ. 25A (1991) 841. [8] A. Wright, Primary system fission product transport and release, Report NUREG/CR 6193, ORNL/TM 12681, Oak Ridge National Laboratory, Oak Ridge, TN, 1994. [9] F. Gelbard, MAEROS User Manual, US Nuclear Regulatory Commission, Washington, DC, 1982. [10] F.E. Kruis, H. Fissan, A. Peled, Synthesis of nanoparticles in the gas phase for electronic, optical, and magnetic applications a review, J. Aerosol Sci. 29 (1998) 511. [11] T.T. Kodas, Generation of complex metal oxides by aerosol processes: superconducting ceramic particles and films, Adv. Mater. 1 (1989) 180. [12] A.C. Tribble, Fundamentals of Contamination Control, SPIE Press, Bellingham, WA, 2000. [13] B.Y.H. Liu, K.W. Lee, Experimental study of aerosol filtration by fibrous filters, Aerosol Sci. Technol. 1 (1981) 35. [14] L. Spielman, S.L. Goren, Model for predicting pressure drop and filtration efficiency in fibrous media, Environ. Sci. Technol. 2 (1968) 279.
Particle Deposition onto Enclosure Surfaces
45
[15] M.R. Sippola, Particle deposition in ventilation ducts, Ph.D. dissertation, University of California at Berkeley, Berkeley, CA, 2002. [16] P.G. Papavergos, A.B. Hedley, Particle deposition behavior from turbulent flows, Chem. Eng. Res. Des 62 (1984) 275. [17] F. Schmidt, K. Gartz, H. Fissan, Turbulent particle deposition on a horizontal circular plate, J. Aerosol Sci. 28 (1997) 973. [18] Y. Ye, D.Y.H. Pui, B.Y.H. Liu, S. Opiolka, S. Blumhorst, H. Fissan, Thermophoretic effect of particle deposition on a free standing semiconductor wafer in a clean room, J. Aerosol Sci. 22 (1991) 63. [19] Y. Otani, H. Emi, C. Kanaoka, K. Kato, Determination of deposition velocity onto a wafer for particles in the size range between 0.3 and 0.8 mm, J. Aerosol Sci. 20 (1989) 787. [20] N.A. Fuchs, The Mechanics of Aerosols, Pergamon Press, Oxford, UK, 1964. [21] J.F. Van De Vate, Investigations into the dynamics of aerosols in enclosures as used for air pollution studies, Report ECN 86, Netherlands Energy Research Foundation, 1980. [22] W.W. Nazaroff, M.P. Logocki, T. Ma, G.R. Cass, Particle deposition in museums: comparison of modeling and measurement results, Aerosol Sci. Technol. 13 (1990) 332. [23] K. Okuyama, Y. Kousaka, S. Yamamoto, T. Hosokawa, Particle loss of aerosols with particle diameters between 6 and 2000 nm in stirred tank, J. Colloid Interface Sci. 110 (1986) 214. [24] W.W. Nazaroff, G.R. Cass, Particle deposition from a natural convection flow onto a vertical isothermal flat plate, J. Aerosol Sci. 18 (1987) 445. [25] A.C.K. Lai, W.W. Nazaroff, Modeling indoor particle deposition from turbulent flow onto smooth surfaces, J. Aerosol Sci. 31 (2000) 463. [26] A. Bejan, Convective Heat Transfer, John Wiley, New York, 1984. [27] A.C.K. Lai, K. Wang, F.Z. Chen, Experimental and numerical study on particle distribution in a two zone chamber, Atmos. Environ. 42 (2008) 1717. [28] S.K. Friedlander, Smoke, Dust, and Haze: Fundamentals of Aerosol Dynamics, second ed., Oxford University Press, Oxford, UK, 2000. [29] W.C. Hinds, Aerosol Technology: Properties, Behavior, and Measurement of Airborne Particles, second ed., John Wiley, New York, 1999. [30] J.O. Hinze, Turbulence, McGraw Hill, New York, 1975. [31] W. Uijttewaal, Particle motion in turbulent pipe flow, Report MEAH 128, TU Delft Lab for Aero and Hydrodynamics, Delft University of Technology, Delft, The Netherlands, 1995. [32] A.V. Baughman, A.J. Gadgil, W.W. Nazaroff, Mixing of a point source pollutant by natural convection flow within a room, Indoor Air 4 (1994) 114. [33] A.C. Drescher, C. Lobascio, A.J. Gadgil, W.W. Nazaroff, Mixing of a point source indoor pollutant by forced convection, Indoor Air 5 (1995) 204. [34] J.H. Seinfeld, S.N. Pandis, Atmospheric Chemistry and Physics, second ed., John Wiley, New York, 1998. [35] J. Aitken, On the formation of small clear spaces in dusty air, Trans. Roy. Soc. Edinburgh 32 (1884) 239. [36] H.H. Watson, The dust free space surrounding hot bodies, Trans. Faraday Soc. 32 (1936) 1073. [37] W. Zernik, The dust free space surrounding hot bodies, Br. J. Appl. Phys. 8 (1957) 117. [38] L. Talbot, R.K. Cheng, R.W. Schefer, D.R. Willis, Thermophoresis of particles in a heated boundary layer, J. Fluid Mech. 101 (1980) 737. [39] A. Li, G. Ahmadi, Aerosol particle deposition with electrostatic attraction in a turbulent channel flow, J. Colloid Interface Sci. 158 (1993) 476. [40] M. Caporaloni, F. Tampieri, F. Trombetti, O. Vittori, Transfer of particles in nonisotropic air turbulence, J. Atmos. Sci. 32 (1975) 565. [41] M.W. Reeks, The transport of discrete particles in inhomogeneous turbulence, J. Aerosol Sci. 14 (1983) 729.
46
Developments in Surface Contamination and Cleaning
[42] S.T. Johansen, The deposition of particles on vertical walls, Int. J. Multiphase Flow 17 (1991) 355. [43] A. Guha, A unified Eulerian theory of turbulent deposition to smooth and rough surfaces, J. Aerosol Sci. 28 (1997) 1517. [44] J. Young, A. Leeming, A theory of particle deposition in turbulent pipe flow, J. Fluid Mech. 340 (1997) 129. [45] B.T. Chen, H.C. Yeh, Y.S. Cheng, Evaluation of an environmental reaction chamber, Aerosol Sci. Technol. 17 (1992) 9. [46] P.A. Baron, K. Willeke, Aerosol Measurements: Principles, Techniques, and Applications, second ed., John Wiley, New York, 2001. [47] P.H. McMurry, A review of atmospheric aerosol measurements, Atmos. Environ. 34 (2000) 1959. [48] S.K. Friedlander, H.F. Johnstone, Deposition of suspended particles from turbulent gas streams, Ind. Eng. Chem. 49 (1957) 1151. [49] T.L. Thatcher, W.A. Fairchild, W.W. Nazaroff, Particle deposition from natural convection enclosure flow onto smooth surfaces, Aerosol Sci. Technol. 25 (1996) 359. [50] T.L. Thatcher, W.W. Nazaroff, Effects of small scale obstructions and surface textures on particle deposition from natural convection flow, Aerosol Sci. Technol. 27 (1997) 709. [51] A.C.K. Lai, W.W. Nazaroff, Supermicron particle deposition from turbulent chamber flow onto smooth and rough vertical surfaces, Atmos. Environ. 39 (2005) 4893. [52] M.A. Byrne, A.J.H. Goddard, C. Lange, J. Roed, Stable tracer aerosol deposition measure ments in a test chamber, J. Aerosol Sci. 26 (1995) 645. [53] A.C.K. Lai, M.A. Byrne, A.J.H. Goddard, Measured deposition of aerosol particles on a two dimensional ribbed surface in a turbulent duct flow, J. Aerosol Sci. 30 (1999) 1201. [54] A.C.K. Lai, M.A. Byrne, A.J.H. Goddard, Aerosol deposition in turbulent channel flow on a regular array of three dimensional roughness elements, J. Aerosol Sci. 32 (2001) 121. [55] G.O. Langstroth, T. Gillespie, Coagulation and surface losses in disperse systems in still and turbulent air, Can. J. Chem. 25 (1947) 455. [56] J. Corner, E.D. Pendlebury, The coagulation and deposition of a stirred aerosol, Proc. Phys. Soc. Lond., Sect. B B64 (1951) 645. [57] Y. Nomura, P.K. Hopke, B. Fitzgerald, B. Mesbah, Deposition of particles in a chamber as a function of ventilation rate, Aerosol Sci. Technol. 27 (1997) 62. [58] M. Shimada, K. Okuyama, Y. Kousaka, Influence of particle inertia on aerosol deposition in a stirred turbulent flow field, J. Aerosol Sci. 20 (1989) 419. [59] Y.S. Cheng, Wall deposition of radon progeny and particles in a spherical chamber, Aerosol Sci. Technol. 27 (1997) 131. [60] R. Van Dingenen, F. Raes, H. Vanmarcke, Molecule and aerosol particle wall losses in smog chambers made of glass, J. Aerosol Sci. 20 (1989) 113. [61] J.F. Van De Vate, The thickness of the stagnant air layer in aerosol containments and the aerodynamic diameter of aggregates of small spheres, J. Colloid Interface Sci. 41 (1972) 194. [62] A.D. Zimon, Adhesion of Dust and Powder, second ed., Consultants Bureau, New York, 1982. [63] M. Shimada, K. Okuyama, Y. Kousaka, K. Ohshima, Turbulent and Brownian diffusive deposition of aerosol particles onto a rough wall, J. Chem. Eng. Japan 20 (1987) 57. [64] M. Abadie, K. Limam, F. Allard, Indoor particle pollution: effect of wall textures on particle deposition, Building Environ. 36 (2001) 821. [65] M. Shimada, K. Okuyama, Y. Kousaka, J.H. Seinfeld, A model calculation of particle deposition onto rough wall by Brownian and turbulent diffusion, J. Colloid Interface Sci. 125 (1988) 198. [66] A.C.K. Lai, M.A. Byrne, A.J.H. Goddard, Experimental studies of the effect of rough surfaces and air speed on aerosol deposition in a test chamber, Aerosol Sci. Technol. 36 (2002) 973.
Particle Deposition onto Enclosure Surfaces
47
[67] T.L. Thatcher, A.C.K. Lai, R. Moreno Jackson, R.G. Sextro, W.W. Nazaroff, Effects of room furnishings and air speed on particle deposition rates indoors, Atmos. Environ. 36 (2002) 1811. [68] W.W. Nazaroff, G.R. Cass, Mass transport aspects of pollutant removal at indoor surfaces, Environ. Int. 15 (1989) 567. [69] A. Lieberman, J. Rosinski, Behavior of an aerosol cloud in a plastic chamber, J. Colloid Interface Sci. 17 (1962) 814. [70] P.H. McMurry, D.J. Rader, Aerosol wall losses in electrically charged chambers, Aerosol Sci. Technol. 4 (1985) 249. [71] J.G. Crump, J.H. Seinfeld, Turbulent deposition and gravitational sedimentation of an aerosol in a vessel of arbitrary shape, J. Aerosol Sci. 12 (1981) 405. [72] M. Shimada, K. Okuyama, Y. Kousaka, Y. Okuyama, J.H. Seinfeld, Enhancement of Brownian and turbulent diffusive deposition of charged particles in the presence of electric field, J. Colloid Interface Sci. 128 (1989) 157. [73] A.C.K. Lai, Investigation of electrostatic forces on particle deposition in a test chamber, Indoor Built Environ. 15 (2006) 179. [74] J.G. Crump, R.C. Flagan, J.H. Seinfeld, Particle wall loss rates in vessels, Aerosol Sci. Technol. 2 (1983) 303. [75] K. Okuyama, Y. Kousaka, Y. Kida, T. Yoshida, Turbulent coagulation of aerosols in a stirred tank, J. Chem. Eng. Japan 10 (1977) 142. [76] R.F. Holub, F. Raes, R. Van Dingenen, H. Vanmarcke, Deposition of aerosols and unattached radon daughters in different chambers; theory and experiment, Rad. Protect. Dosimetry 24 (1988) 217. [77] M.D. Pandian, S.K. Friedlander, Particle deposition to smooth and rough walls of stirred chambers: mechanisms and engineering correlations, Physicochem. Hydrodynam. 10 (1988) 639. [78] M. Benesˇ, R.F. Holub, Aerosol wall deposition in enclosures investigated by means of a stagnant layer, Environ. Int. 22 (Suppl. 1) (1996) S883. [79] H. Schlichting, Boundary Layer Theory, seventh ed., McGraw Hill, New York, 1979. [80] H.H. Brunn, Hot Wire Anemometry: Principles and Signal Analysis, Oxford University Press, Oxford, UK, 1995. [81] A.C.K. Lai, Modeling indoor coarse particle deposition onto smooth and rough vertical surfaces, Atmos. Environ. 39 (2005) 3823. [82] N.B. Wood, A simple method for the calculation of turbulent deposition to smooth and rough surfaces, J. Aerosol Sci. 12 (1981) 275. [83] F. Chen, A.C.K. Lai, An Eulerian model for particle deposition under electrostatic and turbulent conditions, J. Aerosol Sci. 35 (2004) 47. [84] R.B. Bird, W.E. Stewart, E.N. Lightfoot, Transport Phenomena, John Wiley, New York, 1960. [85] A.W. Harrison, Quiescent boundary layer thickness in aerosol enclosures under convective stirring conditions, J. Colloid Interface Sci. 69 (1979) 563. [86] J. Bigu, Radon daughter and thoron daughter deposition velocity and unattached fraction under laboratory controlled conditions and in underground uranium mines, J. Aerosol Sci. 16 (1985) 157. [87] G.A. Schmel, Particle deposition from turbulent air flow, J. Geophys. Res. 75 (1970) 1766. [88] G.P. Fotou, S.E. Pratsinis, A correlation for particle wall losses by diffusion in dilution chambers, Aerosol Sci. Technol. 18 (1993) 213. [89] L. Morawska, M. Jamriska, Deposition of radon progeny on indoor surfaces, J. Aerosol Sci. 27 (1996) 305.
48
APPENDIX TABLE 1.A1 Summary of Experimental Investigations Pertaining to Particle Deposition Measurements in Enclosed Environments Investigators Langstroth and Gillespie, 1947 [55]
Particle diameter 0.6 4 mm
Particle composition Ammonium chloride
Test chamber 3
1.12 m cube
Aerosol analysis method
Flow conditions
Electrostatic and thermal precipitation
Natural convection and turbulent mixing
Key findings
Lieberman and Rosinski, 1962 [69]
Polydisperse 1 8 mm
Zinc cadmium sulfide
310 L Lucite sphere
Optical particle counter
Natural convective mixing
The presence of electric charges on poor conducting surface (e.g. plastic) can introduce additional variability in particle deposition measurements
Van De Vate, 1972 [61]
0.09 1.3 mm
Polystyrene latex particles (PSL)
1 m3 (height 2 m, S/V 7 m 1)
Electrostatic precipitation and microscopy
Natural convective mixing
The particle boundary layer thickness (0.85 mm) was estimated from the particle concentration decay measurements, assuming aerosols were deposited by gravity and diffusion only
Harrison, 1979 [85]
Monodisperse 0.234 2.02 mm
Polystyrene latex particles
58 58 58 cm3 plywood chamber
Light scattering particle counter
Natural convective mixing
Particle deposition rates were found to depend on the surface roughness (rough latex paint vs. smooth aluminum foil surface) Particle concentration boundary layer thickness was found to be a function of particle size and surface roughness
Developments in Surface Contamination and Cleaning
The logarithm of the aerosol mass concentration was found to decrease linearly with time due to deposition to chamber surfaces The rate of particle deposition showed dependence on turbulent flow conditions
Monodisperse 0.024 0.794 mm
Bigu, 1985 [86]
McMurry and Rader, 1985 [70]
Monodisperse 0.02 1.8 mm
NaCl and polystyrene latex particles
Roughly spherical glass vessel with a volume of 118 L
Electrical aerosol analyzer for particle <0.2 mm and optical particle counter for larger particles
Turbulent mixing (continuous flow)
Radon and thoron progeny attached to airborne particles
26 m3 test facility
Plate out experiments followed by a activity measurements
Turbulent and natural convective mixing
The measured deposition velocities for Rn and Th progenies with the fan on (turbulent mixing) were significantly higher than those measured with the fan off (convective mixing)
NaCl
250 L Teflon film bag and 60 m3 Teflon film smog chamber
Condensation nucleus counter (CNC)
Natural convective mixing
Gravitational settling is the predominant particle loss mechanism for particles >1 mm Brownian/turbulent diffusion is the dominant mechanism for removing particles <0.05 mm Electrical forces were shown to contribute particle deposition significantly for smaller charged particles (0.05 1 mm), depending on the field strength near the surface Positive and negative particles were lost to the surface at the same rate when electrical force was the dominant transport mechanism
The dependence of deposition rates on particle size and turbulent intensity was shown to agree well with the theoretical derivations by Crump and Seinfeld [71] A single value of turbulence parameter ke* provided a moderately good fit to the measurement results for 0.02 0.8 mm particles
Particle Deposition onto Enclosure Surfaces
Crump et al., 1983 [74]
49
Continued
50
TABLE 1.A1 Summary of Experimental Investigations Pertaining to Particle Deposition Measurements in Enclosed Environmentsdcont’d Investigators
Particle diameter
Okuyama et al., Monodisperse 1986 [23] 0.006 2 mm
Aerosol analysis Test chamber method
Flow conditions
NaCl, diethylhexyl sebacate (DEHS), and polystyrene latex particles
Cylindrical tank with a volume of 2.6 L; acrylic resin interior
Turbulent mixing and no mixing
Light scattering particle counter for particles >0.5 mm; particle size magnifier and light scattering particle counter for particles <0.5 mm
Key findings
Shimada et al., 1987 [63]
Monodisperse 0.01 0.2 mm
NaCl, diethylhexyl sebacate (DEHS)
2.6 L CNC cylindrical tank with sand roughened interior walls
Turbulent mixing
No difference in deposition rate was found due to particle materials Enhanced deposition rates were observed with increasing fluid turbulent intensity Particle deposition rates exhibited a strong function of particle size with minimum deposition rate at 0.2 (no mixing) to 0.6 mm (vigorous mixing) Good agreements between model and experiments were obtained for 0.006 0.2 mm particles, and for ke > ~10 s 1 Diffusion boundary layer thickness increased proportionally with 1/3rd power of Brownian diffusion coefficient Enhanced particle deposition onto rough surfaces was found compared to that onto smooth surfaces The enhancement of particle deposition, caused by turbulent diffusion very close to the surface, was more pronounced with increasing turbulent intensity and particle size
Developments in Surface Contamination and Cleaning
Particle composition
Monodisperse 0.01 0.2 mm
NaCl, diethylhexyl sebacate (DEHS)
2.6 L CNC cylindrical tank with interior walls lined with various roughness elements
Turbulent mixing
The effect of deposition rate enhancements due to surface roughness showed good agreements with the model predictions Increased particle depositions were observed with increasing particle size, stirring speed, and surface roughness
Holub et al., 1988 [76]
0.001 0.3 mm
Radon progeny attached to airborne particles
Spherical Diffusion battery, glass CNC chamber (BOM), 0.23 m3; cylindrical chamber (SUG NPL), 3.9 m3
Natural
convection (BOM); turbulent mixing by fan (SUG NPL)
The higher turbulence intensity in the SUG NPL vessel played a role in the greater deposition rate measured for the unattached Rn than that in the BOM chamber by a factor of ~30 The experimentally determined deposition rates in both chambers agreed with modeling predictions (Crump and Seinfeld [71]) with n 2.61
Shimada et al., 1989 [58]
Monodisperse 0.1 2 mm
Polystyrene latexparticles
Cylindrical chamber, 3.9 m3
Turbulent
mixing by fan
The effect of inertia on enhanced deposition rates was found for particles >0.2 mm The particle deposition rates were shown to depend on turbulent intensities and particle sizes, with the minimum deposition rates for 0.3 0.5 mm particles The enhancement of particle deposition rates due to inertia was satisfactorily explained by the model of Schmel [87]
Optical particle counter and CNC
51
Continued
Particle Deposition onto Enclosure Surfaces
Shimada et al., 1988 [65]
52
TABLE 1.A1 Summary of Experimental Investigations Pertaining to Particle Deposition Measurements in Enclosed Environmentsdcont’d Investigators Shimada et al., 1989 [72]
Particle composition Test chamber
Aerosol analysis method
Flow conditions
Charged and uncharged monodisperse 0.02 0.2 mm
NaCl
CNC
Turbulent mixing by impeller
Two cylindrical metal chambers with volumes 2.65 and 0.57 L
Key findings
Van Dingenen Monodisperse et al., 1989 [60] 0.02, 0.024, 0.03, 0.05, 0.1, 0.2 mm
NaCl
Spherical borosilicate glass with a volume of 230 L
Condensation nucleus counter (CNC)
Natural convective mixing
For charged and uncharged particles, no differences in measured deposition rates were found when the chamber walls were grounded Deposition rates of charged particles were significantly enhanced as the elec tric field strength increased, or as the particle size decreased The deposition rates could be reasonably predicted by taking into account Brownian/turbulent diffusion, and coulombic forces acting on the particles The observed particle deposition coefficients agreed with Crump and Seinfeld [71] taking n 2.6 Re evaluation of experimental data from literature indicated that an appropriate value of ke was the most important factor in determining particle deposition rates, while the exact value of n is of minor importance
Developments in Surface Contamination and Cleaning
Particle diameter
Monodisperse 0.04 3 mm (0.04, 0.073, 0.215, 0.305, 0.481, 0.62, 0.76, 0.91, 1.09, 1.74, 2.02, 2.99)
Pyrex glass cylindrical chamber, 165 L
Monodisperse 0.01 0.2 mm
NaCl
Byrne et al., 1995 [52]
0.7, 2.5, 4.5, 5.4 mm
Labeled with 2 2 2 m3; dysprosium aluminum (164Dy) and surface indium (113In)
0.006 0.35 mm Radon progeny attached to airborne particles
Natural convective mixing
Fotou and Pratsinis, 1993 [88]
Morawska and Jamriska, 1996 [89]
CNC and aerodynamic particle sizer (APS)
50 L glass spherical and 50 L polyethylene cylindrical vessels
Cubical chamber with a volume of 3 m3 (S 12.5 m2)
Minimum deposition loss was observed for 0.2 0.3 mm particles The flow turbulence intensity increased in the presence of temperature gradient, resulting in higher deposition rates for particles smaller than 1 mm The best fit of data and Crump and Seinfeld’s model [71] occurred when n 2
CNC
Turbulent mixing (continuous flow)
The experimental data agreed with the mass transfer correlation of Pandian and Friedlander [77] as well as the model of Crump and Seinfeld [71]
Neutron activation analysis (NAA) for airborne particles collected on filters and deposited particles on surfaces
Turbulent mixing by fan
The measured particle deposition flux on floor, wall, and ceiling surfaces compared well with particle deposition velocities estimated by the concentration decay method Enhanced particle deposition flux was observed with surfaces of increasing roughness
Diffusion battery for Rn progeny, and differential mobility particle sizer for all airborne particles
Natural convective mixing
Particle Deposition onto Enclosure Surfaces
Chen et al., 1992 [45]
Particle deposition rates exhibited a minimum for particles between 0.1 and 0.25 mm, increased significantly for smaller particles, and slightly for larger particles
53
Continued
54
TABLE 1.A1 Summary of Experimental Investigations Pertaining to Particle Deposition Measurements in Enclosed Environmentsdcont’d Particle diameter
Particle Test composition chamber
Aerosol analysis method
Flow conditions Key findings
Thatcher et al., 1996 [49]
0.1, 0.5, 0.7, 1.3, 2.5 mm
Ammonium fluorescein
Optical particle counter and filter samples for airborne particles; surface extraction followed by fluorometric analysis for deposited particles
Natural convective mixing
1.22 1.22 1.22 m3; aluminum chamber
Thatcher and Nazaroff, 1997 [50]
0.1, 0.5, 1.3 mm
Ammonium fluorescein
1.22 1.22 1.22 m3; aluminum chamber
Optical particle counter and filter samples for airborne particles; surface extraction followed by fluorometric analysis for deposited particles
Natural convective mixing
For horizontal surfaces, gravity played a dominant role in particle deposition for particles >0.1 mm at 1.5 K (surface to air temperature difference); at 10 K, thermophoresis was important for submicron particles, but not significant for supermicron particles For vertical surfaces, gravity, inertia, and thermophoresis affect deposition velocity and the relative importance varies with particle size and surface to air temperature difference Non uniform particle deposition was observed along the vertical surfaces Deposition of small particles (0.1 and 0.2 mm) was relatively insensitive to surface textures; increased deposition onto rough surfaces was found with particle size The surface roughness effect on particle deposition was greater for vertical surfaces than horizontal, and for warm surfaces than cool Particle deposition rates were influenced by surface roughness, surface orientation, surface to air temperature difference, and particle size
Developments in Surface Contamination and Cleaning
Investigators
Monodisperse 0.005 0.14 mm for silver and 0.1 2 mm for polystyrene latex particles
Silver and polystyrene latex particles
161 L spherical aluminum chamber
CNC for submicron particles and aerodynamic particle sizer (APS) for supermicron particles
Turbulent mixing (impeller) and no mixing
Diffusional deposition and gravitational settling were dominant deposition mechanisms for particles <0.1 mm and >0.5 mm respectively Increased turbulence was observed to enhance diffusional deposition, but not for particles >1 mm when gravitational settling was the dominant mechanism The experimental data suggested that n 2.6 2.8 yielded the best fit
Nomura et al., 1997 [57]
Monodisperse 0.02 2 mm
Soot, NaCl, carnauba wax, oil, and fluorescein particles
0.75 0.75 1.8 m3 (S/V 6.44 m 1)
CNC for <1 mm and OPC (optical particle counter) for >1 mm
Turbulent mixing by continuous flow input
Particle deposition rates were observed to depend on particle size and air exchange rate (i.e. air turbulence intensity)
Abadie et al., 2001 [64]
Monodisperse 0.7, 1, and 5 mm
Polystyrene latex and dry powder
0.6 0.6 0.6 m3
Optical particle counter
Turbulent mixing by a fan
Significantly higher deposition rates onto rough surfaces (e.g. carpet) were measured than those onto smooth surfaces (e.g. linoleum)
Lai et al., 2002 [66]
Monodisperse 0.7, 2.5, 4.5, 5.4 mm
Labeled with dysprosium (164Dy) and indium (113In)
22 2 m3; aluminum surface
Neutron activation Turbulent analysis (NAA) for mixing airborne particles collected on filters and deposited particles on surfaces
The ratio of particle deposition on rough surfaces relative to smooth surfaces increased with particle size and the level of air turbulence
Particle Deposition onto Enclosure Surfaces
Cheng, 1997 [59]
Continued
55
Investigators
Particle diameter
Particle Test Aerosol analysis composition chamber method
Flow conditions
Thatcher et al., 2002 [67]
Polydisperse 0.5 10 mm
Olive oil
2.2 2.7 Aerodynamic 2.4 m3 particle sizer (APS)
Natural
convection and turbulent mixing by a fan
Ammonium fluorescein
1.22 Optical particle Turbulent 1.22 counter and filter mixing 1.22 m3; samples for airborne aluminum particles; surface chamber extraction followed by fluorometric analysis for deposited particles
Key findings
Lai, 2006 [73]
Monodisperse 3.5 9 mm
Ammonium fluorescein
1.22 Optical particle Turbulent 1.22 counter and filter mixing 1.22 m3; samples for airborne aluminum particles; surface chamber extraction followed by fluorometric analysis for deposited particles
*ke: turbulence intensity parameter; see discussion in Section 7.1.
Increased air mixing speed, surface roughness as well as surface area resulted in enhanced particle deposition rate Submicron particles exhibited more marked increase of deposition rates as a result of the furnishing level than supermicron particles Deposition velocities on smooth and rough vertical surfaces were determined for particles 0.9 9.1 mm The measured deposition velocity increased as particle size increases The measured deposition velocity increased with increasing roughness scale Particle size had more pronounced effects over surface roughness on deposition Due to electrostatic effects, the measured deposition velocities on to acetate sheets were over an order of magnitude higher than those on to glass surfaces Particle deposition could be significantly reduced by applying anti electrostatic spray by 93% for acetate sheets and 68% for glass surfaces
Developments in Surface Contamination and Cleaning
Lai and Nazaroff, Monodisperse 2005 [51] 0.9, 1.6, 2.2, 3.5, 5.0, 7.0, 7.8, 9.1 mm
56
TABLE 1.A1 Summary of Experimental Investigations Pertaining to Particle Deposition Measurements in Enclosed Environmentsdcont’d
Chapter 2
Contamination Control: A Systems Approach Jacques C.J. van der Donck TNO Science and Industry, Delft, The Netherlands
1. Introduction 2. A Systems Approach 2.1. Step 1: Define the System and the Effect of Contamination 2.2. Step 2: Setting Priorities for the Contaminating Steps 2.3. Step 3: Reduce RPN Numbers by Working on Severity, Occurrence and Detection
3. Effect of Contamination Control Measures on RPN Scores 3.1. Particle Contamination of Lithographic Reticles: Working on Severity or Occurrence? 3.2. Carbon Contamination of EUV Optics: Working on Occurrence 4. Pitfalls 5. Conclusions Acknowledgements References
1. INTRODUCTION In many industrial systems the functionality of products, equipment, and processes can be deteriorated by contamination. Examples can be found in pharmaceuticals [1,2], food [3], textiles [4], space [5,6], semiconductors [7 11], health [12 15], and other fields. Figure 2.1 shows images from the solar panels of the Mars Exploration Rover before (left image) and after 4 years of operation (right image). The particle contamination has increased to such a level that only one-third of all light is able to penetrate through the dust layer and is converted to electrical energy. As a consequence, the power supply is not strong enough to transmit all data to earth. Also, in the semiconductor industry, contamination is an important source of production errors and yield loss [7 10]. Figure 2.2 shows two examples of Developments in Surface Contamination and Cleaning Copyright Ó 2010 Elsevier Inc. All rights of reproduction in any form reserved.
57
58
Developments in Surface Contamination and Cleaning
FIGURE 2.1 Solar panels from NASA’s Exploration Rover on Mars: prior to launch (left image) and after 4 years of operation (right image) [16].
particles on semiconductor surfaces. When these particles are present during one of the production steps of the integrated circuit (IC), short circuits or missing pattern elements will be obtained during the next process steps and as a result non-functioning dies will be produced. Yield loss is an important threat to the economic viability of the semiconductor production process. Two examples of contamination by particles are given above. Other types of contaminants, such as airborne molecular contaminants (AMCs) [20], biological contaminants and films are as important as particles. The systems approach described here can be applied on all types of contaminants. In this chapter, examples of both particle contamination and molecular contamination are given. Considerable effort is spent in counteracting the negative effects of contamination using cleaning and other means. This is generally referred to as contamination control. Contamination control is the collective effort to control contamination to such a level that it guarantees, or even improves, process or product functionality. In the holistic approach to contamination
FIGURE 2.2 Particles of different dimensions (left: sub micrometer size; right: micrometer size) present on IC surfaces after back end of line (BEOL) cleaning [17 19].
Contamination Control: A Systems Approach
59
control, the whole process chain of the production process and/or product is taken into consideration. Therefore, usually more than one contaminating step will be identified. In order to select the step in which the highest performance gain can be achieved, failure mode effect analysis (FMEA) can be used to rank the contaminating steps. This chapter describes the basics of contamination control and how FMEA fits in the process chain. Finally, a series of examples are provided on how contamination control measures can contribute to improving the performance of products and production processes. Contamination control is usually applied to existing processes or products. We will show that the same principles can also be used in an early stage of the design of new equipment and products.
2. A SYSTEMS APPROACH In the introduction some situations were mentioned where contamination leads to serious problems. In many cases, cleaning is chosen as an automatic option for solving the contamination problem. A good example of the general approach for contamination issues is provided by Moreno et al. [21]. In the Mars Exploration Rover Mission, some UV sensors are present that have become contaminated by Mars dust (see Figure 2.3). The solution presented is a mechanical brush that cleans the sensor surfaces. In the conclusion, Moreno and co-authors mention that ‘‘Since the removal efficiency is over 93%, the error in the UV sensor is lower than 7%’’ [21]. In the approach of Moreno et al. [21] two factors are missing: What is the maximum error level allowed for the UV sensor? If the specification on the error level is unknown, it cannot be estimated whether the proposed solution is sufficient for good functioning of the sensors. Do options other than cleaning exist? Cleaning is only one way to approach the problem. Other options might be to change the geometry in such a way that no buildup occurs, or application of a non-sticking coating.
FIGURE 2.3 Contamination of UV sensors by Mars dust, which reduces the transparency of the window: sensors with dust (left), detail of the sensors (middle), and the cleaning solution with a magnetic brush (right) [21]. (see colour plate section at end for coloured version)
60
Developments in Surface Contamination and Cleaning
These two questions are illustrative of the contamination control approach. Central in this approach are the product, process or production equipment and their needs. The product, process or production equipment together are regarded as a system. The whole system is considered when a contamination problem is under investigation. All solutions will be checked against the requirements of the product, process or production equipment. The contamination control approach involves considerable effort. For simple systems an obvious solution may be directly present and an instantaneously workable result may be achieved. But then the contamination control approach is too much effort for the result achieved. Contamination control is, in particular, effective for improvement and design of complex systems with a tight process window and a high risk of interference between different process steps. It enables more flexibility for encountering the effect of contamination. For effective contamination control, a systematic strategy can be followed that is based on three subsequent and often repeating steps. By using FMEA and improving the item with the highest priority on a cyclical basis, a continuous increase in performance can be achieved. The three main steps are described in the following paragraphs.
2.1. Step 1: Define the System and the Effect of Contamination When a process or piece of equipment increases in complexity, the number of key factors that must be taken into account also increases. Therefore, to get a good overview of these factors a system analysis must be carried out in which those essential components are identified that can be affected by contamination and requirements can be defined for contamination levels and process conditions. The system analysis consists of a number of steps: a.
Describe the system. Describe the equipment and the processes that take place and identify products and process units that are sensitive to contamination. Process conditions and sequences are addressed at this stage and functional properties of all relevant substrates and process units are described. b. Define requirements. Every product or production unit has key functionalities that are necessary for good performance. These properties may vary within certain limits, called the requirements. If for any production step an out-of-requirement situation is present, loss of yield, throughput or quality will occur. Therefore, knowledge of these requirements is essential in order to estimate the impact of contamination and find improvements for the system. Examples of such requirements are: transmission efficiency of windows between 72% and 75%, temperature in the range of 64 67 C, process time, etc.
Contamination Control: A Systems Approach
61
c.
Describe the influence of contaminants on the functional properties. When contaminants are present on a product or on a piece of equipment, the functional properties may be affected. In order to estimate the impact of contaminants on the process or product the change in functional properties by contaminants must be described. d. Derive specifications for contamination levels. By combining the requirements and the way contaminants influence the properties, the requirement for the maximum allowable amount of contaminants can be derived. This can be used as a target for the contamination problem. e. Identify possible sources. From the process layout, possible contamination sources can be identified. With this knowledge, possible contamination issues can be identified and solved in the design cycle before any hardware is built. Steps a e constitute a theoretical study. This can be carried out in an early stage of the design of new process equipment. The development costs can be reduced by taking the key contamination risks into consideration before any hardware is built. Actual contamination levels can then be determined and sources can be pinpointed from experiments.
2.2. Step 2: Setting Priorities for the Contaminating Steps In the first step, events are identified that will cause an out-of-specification situation for the process or product. Since contamination control has a larger scope than only an isolated event, more than one contaminating step can be found. For the sake of efficiency the largest risks should be considered first. This requires a system for setting priorities. A good method, which is widely used in the semiconductor industry for yield optimization, is FMEA [22]. In this method, the process or equipment is divided into subprocesses and for each step or module a failure mode is defined. Finally, a risk priority number (RPN) is calculated for each failure mode. The RPN is calculated from the scores for three different parameters described below: Severity, Occurrence, and Detection. When the failure modes are ranked from high RPN to low RPN, the failure mode with the highest risk will appear at the top of the list. Severity is related to the impact of a failure mode on the functionality of the product or process. The score on Severity increases when the influence of a contaminant on a critical functional property increases. The scores depend on the application and failure mode. In the literature [22] the definitions of the scores for Severity differ slightly. They are all related either to customer dissatisfaction or to damage to equipment or people. Scoring is very different for experimental equipment that is still under development than for improvement of mature processes. For experimental equipment, Severity is often ranked between ‘‘no impact’’ and ‘‘extensive damage to equipment’’.
62
Developments in Surface Contamination and Cleaning
Occurrence is related to frequency of the failure mode. The score for Occurrence increases when the failure mode takes place more frequently. Scoring tables vary slightly from source to source. For the semiconductor industry, the Sematech ratings [22] can be used. Score 1 is defined as ‘‘An unlikely probability of occurrence during operating time interval. Single failure mode (FM) probability < 0.001.’’ Score 10 is defined as ‘‘A high probability of occurrence during the operation interval FM > 0.2.’’ Detection is related to the detectability of the failure mode. If the occurrence of a failure mode can be detected instantaneously, the score for Detection will be low. If no detection method is available for the failure mode, then the score will increase to 10. The RPN is then calculated from the following expression: RPN ¼ Severity Occurrence Detection The accuracy of the RPN is strongly dependent on the accuracy of the input data. For equipment in operation, the experimental data on Severity, Occurrence, and Detection can be gathered. If the equipment is still in the design phase scores can only be estimated. For Severity and Detection, the first estimates can be quite accurate, since they can be derived from the process design and requirements. If Step 1 has been carried out, this information is available. Finding a score for Occurrence is more difficult. If no experimental data are available, only estimates can be made that are mainly based on experience. FMEA is a technology-driven activity. The only way costs are rated is by the score for Severity as ‘‘customer dissatisfaction’’ [22]. An economical evaluation of the costs caused by the failure is not given. An alternative method for prioritizing, which includes the costs of a failure mode, is a Total Value of Ownership analysis [23]. Here, the processes are analyzed in terms of added value of each process step and the costs related to it. With this method, costs related to the failure mode and the economical value for a solution can be made apparent.
2.3. Step 3: Reduce RPN Numbers by Working on Severity, Occurrence and Detection From the FMEA the failure mode with the highest RPN number is selected for improvement. It may be obvious that the highest gain can be achieved by improving the factor with the highest score. In principle, all three standard options for contamination control can be used for this: prevention, cleaning, and monitoring.
2.3.1. Prevention Prevention is aimed at blocking the chain of the contaminating process. Figure 2.4 shows the elementary steps of how a product may get contaminated.
63
Contamination Control: A Systems Approach
source
transport
deposition product
FIGURE 2.4
Schematic view of the contamination process.
Contaminants are released from a source. The contaminant is transported from the source to the product. Finally, the contaminant deposits and adheres to the product. Blocking one of the three steps will result in a successful prevention strategy. 2.3.1.1. Source Control The general assumption for source control is that when contamination levels are low, the frequency of process interference by contamination diminishes. This will result in a reduction in the score for Occurrence in the FMEA. From Step 1 it is known which type of contaminant is relevant to the failure mode at hand. Common sources are impurities in gases or liquids. Other types of sources may be greases from bearings and outgassing products from polymeric materials. Particles are often introduced in a piece of equipment either by gases and liquids or they may also be generated within the equipment itself by frictional wear [24 26]. Some preventive actions are very easy. For instance, the contamination level of gases and liquids can be effectively reduced by application of filters. If the outgassing product of a component is interfering with the process, the material may be changed to a material that shows a lower outgassing rate. Particles that are generated by frictional wear can also be eliminated. Changing mechanical settings to reduce friction or application of materials that are less sensitive to mechanical wear may reduce the particle release considerably [27,28]. 2.3.1.2. Prevention of Transport Sometimes, the sensitive substrate itself is clean but it is positioned next to a source of contamination. The contaminant can then be transported from the source to the substrate. Blocking or diverting the transport mechanism will also prevent the contamination of sensitive surfaces. In the literature, transport of particles is described in great detail [29,30]. Transport phenomena are very much dependent on the geometry of the system
64
Developments in Surface Contamination and Cleaning
and the flow. Simple design choices can reduce the transport probability of contaminants considerably. For instance, it is often possible to design equipment in such a way that the sensitive areas are not facing contamination sources or gas/liquid flows [28]. Another option is to place a barrier, shielding the sensitive surface from the source. This shows that design plays a dominant role in interrupting transport processes. The positions of inlets and vents and the position of barriers are all examples of effective design tools that can be used to prevent the transport of contaminants to the critical surface. 2.3.1.3. Adhesion Prevention In most applications, the presence of a single molecule or particle will not result in degradation of performance. Interference with the functionality can only be expected when the concentration of contaminants surpasses the critical level. The driver for adhesion is the adhesion force of the contaminant to the surface. When the adhesion energy of a contaminant is low, it may be released from the surface and transported away. A method to reduce the adhesion force between contaminants and surfaces is to use materials or coatings that have low interaction energy with other materials. The low surface energy of these materials results in a low adhesion force. Well-known examples are fluorocarbons like Teflon. For molecular contamination, the temperature of the substrate is a driver for release. For AMCs the desorption rate depends on the work of adhesion and the temperature. Therefore, AMCs are more easily released from hot than from cold surfaces. In wet systems, particle surface interactions mainly depend on surface charge and steric barriers. For the electric charge the rule of thumb is: similar charges repel and opposite charges attract. The surface charge depends on the material and pH of the solution. By measuring zeta potential as a function of pH, pH ranges with similar charge signs for particle and surface can be selected. Steric barriers are often caused by the adsorption of soluble polymers and brush coatings. The polymers adsorb on the surface of the particle and the substrate surface, and the hydration of the polymer results in a steric barrier between the two. This way the particle cannot approach the substrate surface close enough to be effectively attracted by the van der Waals forces and the particle stays in dispersion.
2.3.2. Cleaning Cleaning is one of the ways to reduce the impact of contaminants on a production chain. In this manner, the score on Severity in the FMEA can be reduced. It can also provide a solution when prevention does not result in sufficient improvement in the Occurrence. The goal of cleaning is to return the substrate or equipment to such a state that functional properties are well within specifications. For equipment, the economics of the process set limits for time available
Contamination Control: A Systems Approach
65
for cleaning and the interval between two cleaning actions (the mean time between cleaning). Therefore, the level of cleanliness after a cleaning action should be such that it is good enough for a sufficiently long period of operation. In our approach, the functional properties are the starting point. From the functional properties and other system requirements all boundary conditions for cleaning can be derived. The most important parameter is the type of contaminant that should be removed. A variety of techniques are available for different types of contaminants. AMCs can be removed by heating or oxidation techniques like UV/ ozone [31 33] or plasma cleaning [7,34,35]. Particulate contaminants can be removed by a variety of techniques like scrubbing [36,37], ultrasonic cleaning [36,38,39], megasonic cleaning [36,37,39,40], water jetting [41], air jet [42,43], laser cleaning [44 48], cryogenic cleaning techniques [49 54], and many others [55 58]. The choice of a cleaning technique depends not only on the cleaning performance, but damage is one of the other aspects that is at least as important as cleaning performance. Damage risk often increases when the removal efficiency of a cleaning method increases. A good example is the removal of a tacky organic residue with an organic solvent. When the dissolution property of the solvent increases, the risk of damage to tubing and plastic components increases as well. Other important factors that should be taken into account are costs, cleaning time, and interference with process conditions present at that location in the equipment. For instance, application of a wet cleaning technique in a vacuum environment is not likely to be very successful. Also, temperature, cleaning time, cost of ownership, sensitivity to vibrations, sensitivity to water or other chemicals, footprint, technical maturity, availability of experimental facilities, operator skill level and, in the case of development, time to market are factors that should be taken into account. From the system analysis, selection criteria can be chosen that are most relevant for the issue at hand.
2.3.3. Monitoring Monitoring means to measure process conditions or product quality during production. It is important for the following reason: if a production error occurs in an early stage of the process without being noticed, the faulty product goes through the rest of the process. In the end, the product must be scrapped and all costs that are added to the process after the error are wasted. Especially when the added value of the different production steps is very high, it is beneficial to prevent faulty products from going through the production line any further. Monitoring helps to reduce the propagation of errors and gives the system the opportunity to take action when errors occur. In many cases, when monitoring is introduced, not only does the score on Detection improve, but sometimes also an improved score on Severity results. This is the case when an efficient feedback loop is introduced. By being able to
66
Developments in Surface Contamination and Cleaning
take counteractions within a short time frame, the costs that are the result of the contamination can be reduced. The impact is reduced and thus the score on Severity is lower. Monitoring can be applied on products or on process parameters. Products are monitored when the process window is not very well known and variation in product quality occurs. In such cases it is beneficial to test each product. An example is probing of dies in the semiconductor industry [59]. In probing, every die on a wafer is tested. An assembly of fine electrical contacts is placed on the pads of the die and a set of diagnostic tests is carried out. If even a single test does not meet the requirements, the die is rejected. In the back-end process this die will not be used for further assembly. Monitoring of process parameters can only be done when the influence of the parameters on the product quality is well known. Examples are pH and temperature measurements in cleaning tools. In many cleaning processes, it is known what pH and temperature values result in sufficiently clean products without damage. If any excursion of the process window is observed, the pH or temperature may be adjusted in a feedback loop. Monitoring techniques can be used to improve the score on Detection in the FMEA. When an efficient feedback loop is present the scores on both Detection and Severity improve, but it is time-consuming and expensive. It only pays off when the added value of the following process steps is very high and the feedback loop is fast enough. Therefore, in most cases improving the scores for Occurrence or Severity should have a higher priority than improving the score for Detection.
3. EFFECT OF CONTAMINATION CONTROL MEASURES ON RPN SCORES In this section, a number of examples are given where contamination control measures can effectively improve the ranking of failure modes in the FMEA and thus the process performance. The examples are both on how running processes can be improved and on how failure modes in process equipment can be rated in the design phase. The most extreme example of processes that are influenced by contamination is semiconductor production. In this arena, a quarter of all process steps are related to cleaning [60] and considerable effort is spent on yield improvement.
3.1. Particle Contamination of Lithographic Reticles: Working on Severity or Occurrence? The first example shows how different contamination control measures change the scores in the FMEA for current reticles. The second example shows how the FMEA gives input in the design of new equipment and the technological developments that were done based on the FMEA input.
Contamination Control: A Systems Approach
67
3.1.1. Particles on Lithographic Reticles: Working on Severity In the lithography step of the semiconductor production process, the structure of the IC is printed on a silicon wafer. The reticle or photomask is a fused silica plate covered with a pattern of a chromium absorber layer. It is positioned under a lens and the pattern is projected in a silicon wafer with a photoresist layer. If a particle as small as 100 nm is deposited on the pattern area of the reticle, it will be printed in every exposure that is carried out with the reticle. These printing errors may result in non-functioning ICs [61]. If this error is not detected, it will result in a yield loss at the end of the production. In terms of FMEA this problem can be rated as presented in Table 2.1. The impact of a particle on a reticle is very high. Furthermore, the score for Occurrence is also high. A method to reduce the score for Occurrence is by prevention (Table 2.2). In the semiconductor world, a simple but effective prevention strategy is found for reticles: use of a pellicle [63]. Particles deposit on the pellicle, 3 mm away from the patterned area. Since the pellicle is not in focus on the wafer, small particles are not printed. Only particles larger than 3 mm result in printing errors. The pellicle effectively blocks the transport of particles to the patterned area of a reticle. In many wafer fabs, wafers are measured after the lithography step. In this metrology step, parameters related to print quality are determined, but a die-todie comparison can also be carried out. A particle on the reticle can be detected when, in every printed field, a constant difference between dies is observed.
TABLE 2.1 FMEA Score for a Particle on a Reticle Process step
Lithography step with a bare reticle
Failure mode
Particles of 100 nm or larger on pattern
Occurrence
In an ISO Class 3 cleanroom, the number of particles is of the order of 1500 particles/m3 [62]. The deposition probability of a particle is high. Based on this assumption the score for Occurrence is 8
Severity
If a particle is present on a reticle, it can be printed and it will result in a defective die. For a hypothetical six die reticle, 17% of all dies must be scrapped at the end of the process. If wafers are exposed with this reticle, the economic damage is $53,000 per hour (17% yield loss, $2600/wafer, 120 wafers/hour). Score for Severity: 8
Detection
In this example no further detection is carried out. Score for Detection: 10
RPN
8 8 10
Comment
This is a high score. The yield loss is threatening the economic viability of this process.
640
68
Developments in Surface Contamination and Cleaning
TABLE 2.2 FMEA for Particle on a Pellicle Process step
Lithography step using a reticle and pellicle
Failure mode
Particles of 3 mm or larger on the pellicle
Occurrence
The amount of particles larger than 3 mm is about three orders of magnitude smaller than for 100 nm particles. Therefore, the frequency of this occurrence decreases. Score for Occurrence: 2
Severity
This does not differ from Table 2.1. Score for Severity: 8
Detection
This does not differ from Table 2.1. Score for Detection: 10
RPN
2 8 10
Comment
By changing the geometry, the failure mode has changed. Contamination by particles smaller than 3 mm does not interfere with the process and there are no more failure modes.
160
TABLE 2.3 FMEA for Particle on a Pellicle with Metrology Process step
Lithography step with a reticle with pellicle and metrology afterwards
Failure mode
Particles of 3 mm or larger on pellicle
Occurrence
This does not differ from Table 2.2. Score for Occurrence: 2
Severity
The Severity has changed. On average, 1 hour production time is lost. Furthermore, costs of rework must be counted. On average, $7500 damage per event is realistic. Score for Severity: 6
Detection
If the reticle contains only one die per printed field or two different dies, this method does not work and printing defects are not detected. This situation is not common. There is a time delay between the occurrence and detection. Score for Detection: 3
RPN
263
Comment
By introducing a metrology step with a feedback loop both Severity and Detection were improved simultaneously.
36
Contamination Control: A Systems Approach
69
On average, a printing defect can be detected within half an hour after exposure. As a consequence, the reticle will be unloaded and cleaned. The resist will then be stripped from the whole batch of wafers and a new exposure with a clean reticle will be carried out. The litho-tool will be unproductive until exposure on a different product starts. The time required for this, the idle time, depends on the flexibility of the logistics and production planning system of the wafer fab. The Severity for the failure mode has diminished (Table 2.3). However, it is still rather high. Some lithographic tools can be equipped with an Internal Reticle Inspection System (IRIS) [64]. This unit is capable of detecting a 3 mm particle on the pellicle. When such a particle is detected, the reticle is unloaded and cleaned. This prevents rework costs involved with faulty exposure. Furthermore, the loss of productivity is reduced to time required for changing to the production of another product. If the production planning and the logistics in the wafer fab are well organized, the idle time for the lithographic tool can be kept as short as possible. The risk and damage from particle contamination on a reticle are decreased from a frequent damage costing $53,000/hour to an occasional event that costs less than $2500 per event (Table 2.4).
3.1.2. Reticle Handling for Extreme Ultraviolet Lithography: Working on Occurrence In the International Technology Roadmap for Semiconductors (ITRS) [65], protection of extreme ultraviolet (EUV) lithographic masks is mentioned as one TABLE 2.4 FMEA for Particle on a Pellicle with IRIS Process step
Lithography step with a reticle and pellicle with IRIS
Failure mode
Particles of 3 mm or larger on pellicle
Occurrence
This does not differ from Table 2.3. Score for Occurrence: 2
Severity
When a particle is detected, little or no productivity loss is the result. The damage depends on the time required to switch to another product. Costs are lower than $2500. Score for Severity: 2
Detection
A new detection technique is introduced. This technique is very reliable. Score for Detection: 2
RPN
222
Comment
By incorporating Detection in the production chain before exposure, the impact on the score for Severity is further reduced.
8
70
Developments in Surface Contamination and Cleaning
of the difficult challenges for lithography for nodes smaller than 32 nm. This is because of the fact that no suitable pellicle material is available [66]. In 2006 two full-field EUV lithographic tools became operational. The situation at the time of the design was a bare reticle with metrology after the lithographic step. The situation was even slightly worse since EUV is a vacuum technique. The pump-down and vent actions during loading and unloading of a reticle into the vacuum system were considered as a higher risk on particle contamination than handling in atmospheric cleanroom conditions. The scores for the FMEA were estimated as: Occurrence ¼ 9 (worse than the score in Table 2.1) Severity ¼ 6 (similar to the score in Table 2.3) Detection ¼ 3 (similar to the score in Table 2.3) RPN ¼ 162. The most logical steps for improvement were a pellicle and in-line particle detection on reticles. These improvements were not available at the stage of prototype design. Therefore, it was decided to put considerable effort into reducing the score for the Occurrence, keeping the number of deposited particles as low as possible. Loading and unloading the reticle into the vacuum was considered to be one of the most sensitive steps with respect to particle contamination. The reticle handling (RH) unit was used to carry out these actions. Therefore, during the design of the RH unit special attention was paid to avoid the risk of particle contamination [67]. Initially, the high-risk areas were identified. From this, a design strategy was chosen based on the following steps [28]: Minimize particle generation minimize number of handling contacts use materials that show minimal particle generation Prevent transport of particles by design by process conditions. In this approach, two of the three elements of prevention are present: source control and transport. The third element, prevention of buildup, cannot be used, since the design of the mask and the conditions under which it must operate cannot be changed. Particle generation is minimized by controlling two factors: the number of handling contacts and the use of materials that do not generate particles upon contact. In the RH unit, the reticle is picked up by a robot and transported from the loadlock towards the reticle stage. Each time a reticle is picked up particles may be generated by frictional wear. The number of particles generated increases with the number of times the reticle is picked up by the robot. Therefore, a method was used to reduce the number of pickups of the actual
Contamination Control: A Systems Approach
71
reticle. The reticle was placed on an RH frame and the gripper of the robot picked up the RH frame instead of the reticle (see Figure 2.5). Materials that do not generate particles were selected using the TNO method for particle generation, the ticker tool [67]. In ticker-tool tests, point contacts are mimicked between carrier pins of different materials, geometries, and reticle surfaces. Damage and particle formation were investigated with a microscope (see Figure 2.6). In this figure, microscopy images of two materials before and after 100 contacts are shown. Material A shows damage on the point of impact and material B shows smearing of the pin material on the surface. With these experiments, materials were selected that were compatible with the reticle surface. The risk of transport of particles towards the pattern side of the reticle was estimated to be highest during pump-down and venting. In order to keep load unload times as short as possible, pump-down and venting took place in a small volume container, the storage box (see Figure 2.7). The storage box consists of an aluminum top outer shell (gray) with a bottom opening plate (light blue). The RH frame is shown in green and the reticle in orange. The yellow and light orange parts together form the top inner cover. The dark blue areas are pump and vent lines. Three methods were used to prevent transport of particles towards the pattern side of the reticle. The first one used point-of-use filters in the box. These filters were present at three locations: two in the gas inlet system and one in the pump line. The second method was to use a unidirectional gas flow. The design of the storage box was chosen such that the flow direction was always away from the pattern side. The storage box contains an outer and an inner cover and one bottom plate. The reticle is positioned on the RH frame. During pump-down, the exit line in the bottom plate is used. The gas flow near the pattern side is parallel between reticle and frame. Here the RH frame shields the
FIGURE 2.5 [67].
Robot gripper which picks up the frame with reticle instead of the reticle itself
72
Developments in Surface Contamination and Cleaning
FIGURE 2.6 Microscopy images of two materials A and B before (left) and after 100 contacts: middle: light microscopy (scale bar is 20 mm); right: scanning electron microscopy (scale bar for material A is 50 mm; scale bar for material B is 100 mm) [67]. (see colour plate section at end for coloured version)
pattern side from direct flow. The volume between inner and outer covers is pumped down through a filter in the top of the inner cover. The gas flow is directed towards the backside of the reticle, which is far less sensitive to particle contamination. During venting, the inlet channel at the side of the box is used. The gas is introduced through the filters at the side and in the top of the inner covers. Again, the flow is directed towards the less sensitive backside of the reticle. A flow parallel to the surface is present near the pattern. The third protection mechanism is to minimize the gap between the reticle and the RH frame. This keeps the gas volume, and thus the number of particles flowing into the area, as small as possible. Furthermore, the probability of the particle to collide with the reticle or the RH frame increases when the gap is smaller. If particles are present, they deposit closer to the edge of the reticle. These positions are not printed and particles do not result in printing errors.
FIGURE 2.7 Storage box for lithographic (EUV) reticles: hardware (left) [68] and drawing [28] (see text for explanation of the colors). (see colour plate section at end for coloured version)
Contamination Control: A Systems Approach
73
These methods for protection of the reticle against particle contamination were first tested on the laboratory scale. A functional model was then built for determining the optimal design. With the design fixed, the optimal conditions for operation were determined experimentally. By measuring added particles on the pattern side of the reticle, optimal pump-down and venting rates were determined. The systematic approach for design and materials for the RH unit resulted in a successful prevention strategy against particle contamination. A series of multiple load and unload tests in the RH unit showed a slow but steady decline in the number of deposited particles. Finally, it was demonstrated that the unit was capable of particle-free loading and unloading of reticles routinely [28].
3.2. Carbon Contamination of EUV Optics: Working on Occurrence The optical system of EUV lithography is very sensitive to hydrocarbons. These compounds adsorb on mirror surfaces and are decomposed by the EUV radiation into a carbon layer. The carbon layer absorbs EUV radiation, which results in a reduction of the EUV radiation at wafer level. The throughput of the EUV tools depends strongly on the amount of radiation that can be applied on the wafer. Since source power is not up to the desired level, each percentage point of EUV radiation that is absorbed is directly translated into loss of throughput. Furthermore, as the mirrors are positioned deep in the system, replacing the mirrors involves high cost. Therefore, the mirrors should not show more reflection loss than a few percent over their lifetime. The amount of hydrocarbons in a standard vacuum system is high enough to surpass the maximum amount of allowable carbon within a short period. Table 2.5 summarizes the situation in terms of FMEA. Mertens et al. [69] worked on root cause analysis for the contamination source and showed how to reduce its impact on the FMEA. One of the most important sources of hydrocarbons is the photoresist layer on the wafer. During exposure, outgassing products are released from the resist. These hydrocarbons can diffuse towards the mirrors and adsorb on their surface. Upon exposure to EUV radiation, the hydrocarbons are decomposed and remain on the mirror surface as a carbon layer. Since the hydrocarbons are decomposed, no equilibrium is reached between adsorption and desorption and the layer will continue to grow if no action is taken. The proposed solution was to separate the optical system from the wafer compartment (see Figure 2.8). Each compartment can have its own gas conditions with cleanliness levels adjusted to the required functionality. Between the optics and the wafer compartment, a gaslock system was designed. This gaslock was able to suppress the hydrocarbon level by five orders of magnitude. The suppression was calculated by a numerical method and determined experimentally. In terms of FMEA, the scores for Severity and
74
Developments in Surface Contamination and Cleaning
TABLE 2.5 FMEA for Carbon Growth on an EUV Mirror Process step
EUV mirrors in the optical system of an EUV lithographic tool
Failure mode
Reflection loss caused by carbon growth on the mirrors
Occurrence
In a one chamber vacuum system, the maximum allowable level of contamination is surpassed within a short time. Score for Occurrence: 10
Severity
When the maximum allowable level of contamination is surpassed, the EUV lithographic shows non functionality. Score for Severity: 10
Detection
The intensity of the EUV radiation is measured on a regular basis. Score for Detection: 1
RPN
10 10 1
100
Detection remained the same and the score for Occurrence was reduced to a much lower level. Mertens et al. [69] showed an excellent example in which the transport of a contaminant (hydrocarbons) between a source (wafer) and a sensitive area (optical components) was eliminated. By balancing the gas flow in the system the transport of hydrocarbons was effectively suppressed. They mention two important approaches for the design of the gaslock: numerical calculations and analytical models. A number of numerical flow calculation models are available [70 72]. In the literature, it is shown that these models enable one to predict which flows will be present and how contaminants may be transported [73,74]. With this type of information, the elimination of the transport mechanism can be achieved.
FIGURE 2.8 Separation of optical system and wafer compartment by a gaslock system that eliminates transport of contaminants between a wafer and a sensitive optical system [69].
Contamination Control: A Systems Approach
75
4. PITFALLS Research and development are often a matter of trial and error, even with a systematic approach. Most of the errors are obvious but are difficult to avoid. Some of them are summarized below. 1. Requirements are not (completely) specified. When solutions for a contamination problem are chosen, a trade-off between the different techniques must be made. In the trade-off process, options are ranked by a score for a set of requirements. A workable table can only be obtained when the most important requirements are incorporated. In some cases, a solution violates a requirement that is considered trivial in the first instance, but appears essential in the end. The whole process then starts from the beginning with a re-evaluation of all requirements and possible solutions. 2. No relation between contamination level and functional demand is known. For the question ‘‘how much contamination is acceptable?’’, the answer is very often ‘‘none’’. This is obviously impossible. It shows how little insight is present on how contaminants influence the functionality of the product. This relation is not always necessary for simple systems. However, the need for this information increases when the equipment and systems become more complex. Only then can trade-offs be made between different sources and contaminants. 3. Contamination becomes an issue after all other requirements are set. The focus during design is mostly on process and product functionalities. Also, thermal or chemical properties are taken into consideration. Contamination issues become apparent after all design choices have been made. In that case, a lot of different materials, designs, and equipment have already been chosen. Changing or replacement results in high costs and down time. This limits the choice of solutions for contamination control. 4. The cost/benefit relation is unclear. Contamination control and FMEA are both technology driven. The costs related to yield and productivity loss are often incorporated in an indirect way. This way the benefits of incorporation of a contamination control solution are often not directly clear in terms of process economics. When costs are described as in Section 3.1.1 the benefits of the different options become clearer.
5. CONCLUSIONS Contamination influences a wide variety of industrial processes. For complex systems, contamination control, the collective effort to control contamination to such a level that it guarantees or even improves process or product functionality, offers a way for finding workable solutions. Central in the approach is the system, i.e. a product, a process or a production tool. Contamination control has its focus on the system as a whole.
76
Developments in Surface Contamination and Cleaning
A solution must be compliant with every aspect of the system. Therefore, the basis of every contamination control investigation is system analysis. In this phase, all sensitive objects and process steps are described and the system requirements must be listed, following which the relevant contamination types can be identified and levels can be determined. When these steps are taken, a ranking of all contamination issues can be made by FMEA, after which a solution for the failure mode with the highest score can be developed. Three main directions for solutions can be distinguished: prevention, cleaning, and monitoring. Prevention reduces contamination sources or blocks the transport or deposition mechanism. Here knowledge of material properties, design, and transport phenomena is essential. Cleaning is reducing the amount of contaminants to such a level that it is within specifications. In the selection of a cleaning method, not only removal properties, but also damage and other aspects must be taken into consideration. Monitoring is based on finding the key parameter for product or process quality. When this parameter is checked during the process, process integrity and product quality can be guaranteed. The choice between the three options depends on the maturity of the process and how much the actually observed particle levels differ from the requirements. The elements of the FMEA Occurrence, Severity, and Detection can identify the impact of different solutions on the process performance. Contamination control is a valuable tool for the development of complex systems where contamination is likely to limit the functionality or product quality and yield. The earlier the contamination risks are taken into consideration during the development, the better they can be counteracted.
ACKNOWLEDGEMENTS The current chapter is a summary of TNO knowledge that has been gathered by many colleagues within many projects in the last decade. Special thanks are due to Ton Bastein, Anton Duisterwinkel, and Norbert Koster for their contribution to the development of the Contamination Control approach within TNO and valuable input and discussions on this chapter. Mrs Noreen Harned of ASML is thanked for releasing information on ASML lithography-related topics for publication.
REFERENCES [1] G.A. Willing, T.R. Burk, F.M. Etzler, R.D. Neuman, Adhesion of pharmaceutical particles to gelatin capsules having variable surface physicochemical properties: evaluation using a combination of scanning probe microscopy techniques, Colloids Surfaces A 193 (2001) 117.
Contamination Control: A Systems Approach
77
[2] R.M. O’Donnell, X. Sun, P. de B. Harrington, Pharmaceutical applications of ion mobility spectrometry, TrAC Trends in Analytical Chemistry 27 (2008) 44. [3] S. Le´crigny Nolf, C. Faille, T. Be´ne´zech, Removal kinetics of Bacillus cereus spores from a stainless steel surface exposed to constant shear stress. 1. Experimental system, Biofouling 14 (2000) 287. [4] W.G. Cutler, R.C. Davis (Eds.), Detergency: Theory and Test Methods, Marcel Dekker, New York, 1972. [5] A.C. Tribble, B. Boyadjian, J. Davis, J. Haffner, E. McCullough, Contamination control engineering design guidelines for the aerospace community, NASA CR 4740, National Aeronautics and Space Administration, May 1996. [6] W.D. Kimura, G.H. Kim, Comparison of laser and CO2 snow cleaning of astronomical mirror samples, Proc. SPIE 2199 (2000) 1. [7] F. Wali, D.M. Knotter, A. Mud, F. Kuper, Impact of particles in ultrapure water on random yield loss in IC production, Microelectronic Eng. 86 (2009) 140. [8] M. Okazaki, R. Maas, S. H. Ko, Y. Chen, P. Miller, M. Thothadri, M. Dutta, C. P. Chang, A. Anapolsky, C. Lazik, Y. Uritsky, M. Seamons, D. Padhi, W. Yeh, S. Sinkwitz, C. Ngai, Wafer edge polishing process for defect reduction during immersion lithography, Proc. SPIE 6922 (2008) 3A1. [9] K.A. Reinhardt, Advances in wafer cleaning and particle removal technology, in: K.L. Mittal (Ed.), Particles on Surfaces 9: Detection, Adhesion and Removal, VSP/Brill, Leiden, The Netherlands, 2006, pp. 225 241. [10] M.T. Andreas, Particle removal challenges and solutions in semiconductor fabrication CMP processes, in: K.L. Mittal (Ed.), Particles on Surfaces 9: Detection, Adhesion and Removal, VSP/Brill, Leiden, The Netherlands, 2006, pp. 243 254. [11] K.A. Reinhardt, W. Kern (Eds.), Handbook of Silicon Wafer Cleaning Technology, second ed., William Andrew, Norwich, NY, 2008. [12] Telegraaf, Hospital boxmeer closes operating rooms, 24 October 2008. [13] Volkskrant, Stop on operations in the seven unsafe operating rooms, 18 September 2008. [14] Algemeen Dagblad, Air in operating rooms of hospitals has been found in good condition, 15 October 2008. [15] F. Gyntelberg, P. Suadicani, J.W. Nielsen, P. Skov, O. Valbjørn, P.A. Nielsen, T. Scheinder, O. Jørgnsen, P. Wolkoff, C.K. Wilkins, S. Gravesen, S. Norn, Dust and the sick building syndrome, Indoor Air 4 (1994) 223. [16] NASA, Mars Exploration Rover Mission. Website: http://marsrovers.nasa.gov/spotlight/ 20080420 Spirit.html (accessed 20 April 2008). [17] J.W. Butterbaugh, Using cryogenic aerosol process to clean copper, low k materials without damage, MICRO 20 (February 2002) 23. [18] J.W. Butterbaugh, S. Loper, G.P. Thomes, Enhancing yield through argon/nitrogen cryo kinetic aerosol cleaning after via processing, MICRO 17 (June 1999) 33. [19] T.J. Wagner, K. Kawaguchi, Improved yields for the nano technology era using cryogenic aerosols, Proc. IEEE/SEMI Advanced Semiconductor Manufacturing Conference and Workshop, 2004, pp. 467 471. [20] SEMI Standard F21 1102, Classification of airborne molecular contaminant levels in clean environments, Semiconductor Equipment and Materials International, San Jose, CA, 2002. [21] L. Moreno, R. Caba´s, D. Ferna´ndez, Low mass dust wiper technology for MSL Rover, Proc. 9th ESA Workshop on Advanced Space Technologies for Robotics and Automation, ESTEC, Noordwijk, The Netherlands, 2006. [22] M. Villacourt, Failure Mode and Effects Analysis (FMEA): A Guide for Continuous Improvement for the Semiconductor Equipment Industry, Technology Transfer #92020963B ENG, SEMATECH, Austin, TX, October 2002. [23] K. Hurkens, W. van der Valk, F. Wynstra, Total cost of ownership in the services sector: a case study, J. Supply Chain Management 42 (2006) 127.
78
Developments in Surface Contamination and Cleaning
[24] K. H. Cha, K. H. Chung, D.E. Kim, Effect of slider load on the wear debris contamination tendency of head/slider, IEEE Trans. Magn. 35 (1999) 2355. [25] E. S. Yoon, B. Bhushan, Effect of particulate concentration, materials and size on the friction and wear of a negative pressure picoslider flying on a laser textured disk, Wear 247 (2001) 180. [26] K. Kato, Tribological implications of particles, in: R. Kohli, K.L. Mittal (Eds.), Develop ments in Surface Contamination and Cleaning, William Andrew, Norwich, NY, 2008, pp. 299 327. [27] G. J. Heerens, E. Ham, R. Lansbergen, R. Snel, A. Duisterwinkel, H. van den Berg, H. Werij, R. Moors, B. van der Ven, M. Ganguly, A. Gilmans, S. Del Puero, B. Blum, Particle free EUV reticle handling, Proc. 2nd Int. EUVL Symposium; www.sematech.org/meetings/ archives/litho/euvl/20030930/presentations/5E%20 weij%20EUV%20Symp.pdf, 2003. [28] E. Ham, J.C.J. van der Donck, B. Mertens, R. Snel, M. Oderwald, R. Landsbergen, H. van den Berg, H. Meiling, H. Meijer, T. Nabben, R. Moors, B. Blum, Progress in particle free handling of reticles in the ASML EUV Program, Paper presented at EUV Conference, San Diego, CA; http://www.sematech.org/meetings/archives/litho/7470/Poster/FinalS1/1 RP 01%20Ham TNO%20Poster.pdf, 2005. [29] D.J. Rader, A.S. Geller, Transport and deposition of aerosol particles, in: R. Kohli, K.L. Mittal (Eds.), Developments in Surface Contamination and Cleaning, William Andrew, Norwich, NY, 2008, pp. 189 266. [30] C. H. Lin, C. Zhu, Relevance of particle transport in surface deposition and cleaning, in: R. Kohli, K.L. Mittal (Eds.), Developments in Surface Contamination and Cleaning, William Andrew, Norwich, NY, 2008, pp. 267 297. [31] R.R. Sowell, R.E. Cuthrell, D.M. Mattox, R.D. Bland, Surface cleaning by ultraviolet radiation, J. Vac. Sci. Technol. 11 (1974) 474. [32] J.R. Vig, UV/ozone cleaning of surfaces, J. Vac. Sci. Technol. A3 (1985) 1027. [33] J.R. Vig, UV/ozone cleaning of surfaces, in: K.L. Mittal (Ed.), Treatise in Clean Surface Technology, vol. 1, Plenum Press, New York, 1987, pp. 1 26. [34] D.F. O’Kane, K.L. Mittal, Plasma cleaning of metal surfaces, J. Vac. Sci. Technol. 11 (1974) 567. [35] J. M. Koo, J. B. Lee, Y. Moon, W. C. Moon, S. B. Jung, Atmospheric pressure plasma cleaning of gold flip chip bump for ultrasonic flip chip bonding, J. Phys.: Conf. Ser. 100 (2008) article 012034. [36] A.A. Busnaina, G.W. Gale, I.I. Kashkoush, Ultrasonic and megasonic theory and experi mentation, Precision Cleaning (April 1994) 13 19. [37] C.K. Chang, T.H. Foo, M. Murkherjee Roy, V.N. Bliznetov, H.Y. Li, Enhancing the efficiency of postetch polymer removal using megasonic wet clean for 0.13 mm dual Damascene interconnect process, Thin Solid Films 292 (2004) 462 463. [38] S. Awad, Ultrasonic cavitations and precision cleaning, Precision Cleaning (November 1996) 12 17. [39] B. Kanegsberg, E. Kanegsberg (Eds.), Handbook for Critical Cleaning, CRC Press, Boca Raton, FL, 2001. [40] D.A. Koos, J. Svirchevski, D.J. Vitkavage, D.G. Hansen, K.A. Reinhardt, F. Huang, M. Mitchel, G. Zhang, Development of a non contact post CMP cleaning process for copper, in: K.L. Mittal (Ed.), Particles on Surfaces 9: Detection, Adhesion and Removal, VSP/Brill, Leiden, The Netherlands, 2006, pp. 127 144. [41] D.A. Summers, Waterjetting Technology, first ed., Taylor & Francis, London, 1995. [42] K. Gotoh, Removal of micrometer size particles from solid surfaces by an impinging air jet, in: K.L. Mittal (Ed.), Particles on Surfaces 9: Detection, Adhesion and Removal, VSP/Brill, Leiden, The Netherlands, 2006, pp. 329 355. [43] K. Gotoh, Cleaning using a high speed impinging jet, in: R. Kohli, K.L. Mittal (Eds.), Developments in Surface Contamination and Cleaning, William Andrew, Norwich, NY, 2008, pp. 889 917.
Contamination Control: A Systems Approach
79
[44] Y.F. Lu, W.D. Song, Y. Zhang, Y.W. Zheng, Laser surface cleaning of electronic materials, Proc. SPIE 3618 (1999) 278. [45] M. She, D. Kim, C.P. Grigoropoulos, Liquid assisted pulsed laser cleaning using near infrared and ultraviolet radiation, J. Appl. Phys. 86 (1999) 6519. [46] Y.W. Zheng, B.S. Luk’yanchuk, Y.F. Lu, W.D. Song, A.H. Mai, Dry laser cleaning of particles from solid substrates: experiments and theory, J. Appl. Phys. 90 (2001) 2135. [47] J.M. Lee, K.G. Watkins, W.M. Steen, Surface cleaning of silicon wafer by laser sparking, J. Laser Applics. 13 (2001) 154. [48] A.A. Busnaina, J.G. Park, J.M. Lee, S.Y. You, Laser shock cleaning of inorganic micro and nanoscale particles, Proc. IEEE/SEMI Advanced Manufacturing Conference, 2003, pp. 41 45. [49] J.W. Butterbaugh, S. Loper, G. Thomes, Yield enhancement by cryokinetic cleaning, MICRO (June 1993) 33 43. [50] R. Sherman, D. Hirt, R. Vane, Surface cleaning with the carbon dioxide snow jet, J. Vac. Sci. Technol. A12 (1994) 1876. [51] N. Narayanswami, J.F. Weygand, P. Ruether, K.K. Christenson, J.W. Butterbaugh, S. H. Yoo, B.Y.H. Liu, S. K. Chae, J.J. Sun, Evaluation of particle removal efficiency in wafer cleaning processes, Semiconductor Int. (June 2000) 1 15. [52] J.C.J. van der Donck, R. Schmits, R.E. van Vliet, A.G.T.M. Bastein, Removal of sub 100 nm particles from structured substrates with CO2 snow, in: K.L. Mittal (Ed.), Particles on Surfaces 9: Detection, Adhesion and Removal, VSP/Brill, Leiden, The Netherlands, 2006, pp. 291 302. [53] R. Sherman, Carbon dioxide snow cleaning, in: R. Kohli, K.L. Mittal (Eds.), Devel opments in Surface Contamination and Cleaning, William Andrew, Norwich, NY, 2008, pp. 987 1012. [54] W.T. McDermott, J.W. Butterbaugh, Cleaning using argon/nitrogen cryogenic aerosols, in: R. Kohli, K.L. Mittal (Eds.), Developments in Surface Contamination and Cleaning, William Andrew, Norwich, NY, 2008, pp. 951 986. [55] D.W. Cooper, H.L. Wolfe, J.T.C. Yeh, M.J. Miller, Surface cleaning by electrostatic removal of particles, Aerosol Sci. Technol. 13 (1990) 116. [56] K.L. Mittal (Ed.), Particles on Surfaces 8: Detection, Adhesion and Removal, VSP/Brill, Leiden, The Netherlands, 2003. [57] K.L. Mittal (Ed.), Particles on Surfaces 9: Detection, Adhesion and Removal, VSP/Brill, Leiden, The Netherlands, 2006. [58] J.M. Bennett, D. Ronnow, Test of opticlean strip coating material for removing surface contamination, Appl. Optics 39 (1999) 2737. [59] Semiconductor Manufacturing Tour, Infrastructure Tutorial, Baileys Harbor, WI. Website: www.infras.com/Tutorial/sld018.htm, 1999. [60] K. Qin, Y. Li, Mechanisms of particle removal from silicon wafer surface in wet chemical cleaning process, J. Colloid Interface Sci. 261 (2003) 569. 61] R. Tejeda, R. Engelstad, E. Lovell, K. Blaedel, Particle induced distortion in extreme ultraviolet lithography reticles during exposure chucking, J. Vac. Sci. Technol. B20 (2002) 2840. [62] ISO 14644 1, Cleanrooms and Associated Controlled Environments Part 1: Classification of Air Cleanliness, International Organization for Standardization, Geneva, 1999. [63] M. Fujita, H. Nakagawa, Mask Protective Device, European Patent 0696760, 1999. [64] ASML i Line Stepper Tool PAS 5500 User Guide, Release 8.4.0, ASML, Veldhoven, The Netherlands, June 2000. [65] Semiconductor Industry Association, International Technology Roadmap for Semi conductors, Overview of 2008 Update, 2008, p. 36. [66] S. Hector, P. Mangat, Review of progress in extreme ultraviolet lithography masks, J. Vac. Sci. Technol. B19 (2001) 2612.
80
Developments in Surface Contamination and Cleaning
[67] G. J. Heerens, E. Ham, R. Lansbergen, R. Snel, A. Duisterwinkel, H. van den Berg, H. Werij, R. Moors, B. van der Ven, M. Ganguly, A. Gilmans, S. Del Puero, B. Blum, Particle free EUV reticle handling, Paper presented at 2nd Int. EUVL Symposium, Antwerp, Belgium. Website: www.sematech.org/meetings/archives/litho/euvl/20030930/ presentations/5E%20 weij%20EUV%20Symp.pdf, 2003. [68] N. Harned, P. Kuerz, H. Meiling, B. Mertens, G. van Baars, Review of progress in extreme ultraviolet lithography masks, Paper presented at 3rd Int. EUVL Symposium, Miyazaki, Japan. Website: www.sematech.or/meetings/archives/litho/euvl/20041101euvl/presentations/ day3/To02 Harned Final Handout.pdf, 2004. [69] B.M. Mertens, B. van der Zwan, P.W.H. de Jager, M. Leenders, H.G.C. Werij, J.P.H. Benschop, A.J.J. van Dijsseldonk, Mitigation of surface contamination from resist outgassing in EUV lithography, Microelectronic Eng. 53 (2000) 659. [70] Fluent 5 User’s Manual, Fluent Inc., Lebanon, NH, 1998. [71] G.A. Bird, Molecular Gas Dynamics and the Direct Simulation of Gas Flows, Clarendon Press, Oxford, 1994. [72] R. Dorsman, Numerical simulations of rarefied gas flows in thin film processes, Ph.D. thesis, Delft University of Technology, Delft, The Netherlands, 2007. [73] A.C.K. Lai, F. Chen, Modeling particle deposition and distribution in a chamber with a two equation Reynolds averaged Navier Stokes model, Aerosol Sci. 37 (2006) 1770. [74] L. Tian, G. Ahmadi, Particle deposition in turbulent duct flows comparisons of different model predictions, Aerosol Sci. 38 (2007) 377.
Chapter 3
Particles in Semiconductor Processing D. Martin Knotter * and Faisal Wali y * y
NXP Semiconductors, Nijmegen, The Netherlands University of Twente, Nijmegen, The Netherlands
1. Introduction 1.1. Impact of Particles in IC Manufacturing 1.2. Yield Calculation Models 1.3. Origin of Particles 1.4. Determination of Particle Removal Efficiency (PRE) 1.5. Methods to Remove Particles 2. Theory 2.1. Particle Removal Mechanism 2.2. Model of Particle Removal
2.3. Liquid Particle Counter (LPC) Measurements 2.4. Metal Ion Core Particles 3. Particle Removal Study 3.1. Tank Dynamics, Impact of Particle Counter, and Particle Composition 3.2. Particle Diffusion Out of the Boundary Layer 3.3. Particle Detachment 4. Conclusions Acknowledgements References
1. INTRODUCTION Advances in integrated circuits (ICs) have a high impact on society. These advances result in continuously increasing performance of home personal computers, higher density flash memory chips, faster wireless communication in combination with smaller antennas, and all kinds of combinations of the aforementioned components. The main characteristic of these advances has been the shrinking dimension of the features of which the ICs are made.
1.1. Impact of Particles in IC Manufacturing Every two years the feature size of the new generation of microprocessors is reduced with a O2 factor [1]. Since 2004 the smallest size is in the nanoscale Developments in Surface Contamination and Cleaning Copyright Ó 2010 Elsevier Inc. All rights of reproduction in any form reserved.
81
82
Developments in Surface Contamination and Cleaning
range which is defined by the 100-nm limit. Simultaneously, particles that can cause device damage and that are deposited on the product during its manufacturing are smaller in size. In early road maps, the critical dimension of the particles was assumed to one-tenth of the minimum device feature size. Later, this was relaxed from one-third to currently half of the feature size. Today, it means detection and removal of particles in the range of 10 20 nm. The manufacturing of microprocessors consists of hundreds of process steps, many of which can be sensitive to particle contamination. The process steps that are sensitive to particle contamination can be grouped into several categories (see Figure 3.1): a. b. c. d.
Particles in holes or trenches In-film particles Particles as mask (not related to pattern) Patterning deviations where particles become part of the photo mask.
a
c e-
d
b s
g + ++
d
e-
s
g + ++
d e-
FIGURE 3.1 Cross sectional details of an IC with different impact of random particles on device performance. Color coding: white is the insulator; light gray is the semiconductor; dark gray is the conductor; black is a particle with unknown properties. (a) Particle in a contact hole before the hole is filled up with metal results in an ‘‘open’’ circuit. (b) Particle on the gate area (g) prior to gate definition results in poor transistor performance (s source, d drain). (c) Particle on an area that is implanted with low energy dopants. (d) Patterning problem where the particle is located in the photo mask pattern, resulting in a masked etch and a ‘‘short’’ circuit.
Particles in Semiconductor Processing
83
These categories do overlap, but they all result in different kinds of failure in the final device. The commonality is that all these particles can result in random yield loss. 1. Particles in holes or trenches. The most prevalent failure mode in manufacturing is category (a), where a particle obstructs the conductivity between two metal layers. The reason is the process weakness in the previous steps, i.e. poor definition of holes and trenches in the dielectric layer. This plasma etch step is contaminating and the subsequent cleaning step has a relatively small process window. Plasma etching results in so-called ‘‘sidewall polymers’’ that are actually polymers deliberately deposited on the sidewall to attain a high degree of anisotropy in the etch process (sidewall passivation). Furthermore, the sidewall polymers can contain residue of the reaction product between the plasma gas and the metal layer that is exposed on the bottom of the hole. Exposure of the metal to the plasma gas is poorly controlled, resulting in an unpredictable amount of metal salts in the sidewall polymers. Compared to the cleans earlier in the process, the clean between the plasma etch and the deposition of the next metal layer (to fill the hole) has reduced chemical etch activity as well as physical power, because it is not allowed to etch the exposed metal nor the dielectric material which, in the latest technologies, is made of porous silica-based material. 2. In-film particles. Category (b) process failure has been driving the roadmap for semiconductors with respect to cleaning and cleanliness performance of the manufacturing facilities. The critical area here is defined as the area where the gate is going to be made and it is the number of transistors in the device multiplied by the gate area in each transistor. Surprisingly, this has remained constant over the years, because the gate area has been halved with each new generation while the number of transistors has doubled. This means that the number of allowable particles to achieve a 99% device yield remains constant. However, the challenge here is that the particles now are of a smaller size. Except for the deposition tool itself, there are no typical process-related particles depositing on the wafer before and during these process steps. In the deposition tool the material to be deposited is not only deposited on the wafer but also on the sidewall of the chamber. Several depositions in a row without a layer-removal step can cause these multilayers to crack from the stresses, resulting in airborne particles. These particles will be deposited on the wafer either in the beginning of the deposition, resulting in in-film particles that are non-removable, or at the end of the deposition. If the particles are not yet within the film, a relatively robust cleaning process can be used to remove these particles. It can make use of mechanical forces such as megasonic energy without too much concern for pattern damage. Also, at this stage under-etching is still allowed.
84
Developments in Surface Contamination and Cleaning
3. Particles as mask (not related to pattern). This process failure is similar to the previous example, where particle contamination is present on a wafer without a photoresist pattern, but structures can also be present on the wafer. The critical processes involved can be implantation, oxidation, or a film-removal step. The particles mask the underlying substrate and they will leave a ghost pattern in the final device even after the particle has already been removed. 4. Patterning deviations where particles become part of the photo mask. Before or during the definition of the patterned photoresist layer, particles can deposit and can be positioned such that they are on a location where no photoresist should have been (failure mode d in Figure 3.1). Particles on photomasks that are used to define the patterned photoresist layer can result in a similar failure, but this will be picked up as a systematic yield loss because the image of the particle will be printed on every product at the same position. As in the previous case, the particle acts as a mask, but in this case it will inhibit the etching process on a patterned wafer. The position of the particle defect can occur anywhere and the particle imprint can no longer be removed. In the given example (Figure 3.1), the particle results in an electrical short because the etched metal was not removed completely and the imprint is short-circuiting two metal lines that should have been separated. Other failure scenarios are also possible that can result in an open circuit, if, for example, a dielectric layer is not etched away and the deposited metal layer later cannot fill that space.
1.2. Yield Calculation Models The degree of success in IC manufacturing is measured by yield (Y) that is defined as the ratio of usable devices in respect to potentially usable devices before starting its manufacturing [2]. Usable devices are defined as those that pass several physical and electrical tests during or after completion of the multistep manufacturing process. Knowledge of yield performance of a manufacturing facility or process is used to predict the yield of new products that has a higher degree of integration. The economics of the introduction of such production processes helps in making the decision to build a new fab, upgrade the existing fab, and the identifications of problematic process steps. A large portion of the yield loss is caused by contamination present in the wafer environment and, in this category, particles are the major contributor. To be able to reduce the impact of contamination on manufacturing, defects have to be detected. These defects are initially measured after each process step. The long-term defect probability and yield prediction are related by statistical probability distribution models. There are several yield prediction models in the literature. Most of these models require the detection of defects on the product wafer during the
Particles in Semiconductor Processing
85
multistep process flow and the confirmation of successful or unsuccessful fabrication of the affected product. This assumes that a wafer consists of a number of ICs that are sawed out of the wafer at the end of the process. Each IC consists of a number (N) of transistors. If a single transistor in an IC is not working, the whole IC is malfunctioning and is considered scrap. Finally, defects, especially particles, can be deposited randomly on the wafer, which is a stochastic process.
1.2.1. Wallmark’s Model In 1960, the first yield prediction model was proposed by Wallmark [3] and is described by: Y ¼ ð1 S=100ÞN
(1)
His model was based on the assumption that the percentage (S) of working transistors that could be produced by a certain process was known from an existing manufacturing process or was determined experimentally. As the number of transistors in an IC was in the range of one to several hundred, determination of such numbers was reliable. The model is based on probability calculations: if one out of ten transistors is not working, then nine out of ten are working. Thus, the probability that a device is made out of two working transistors is 0.92 ¼ 0.81, but with ten transistors the yield would drop to 0.910 ¼ 0.35. An economist could tell if this yield is sufficient to bring such a product to the market, or that the yield has to be improved first before going on to integrate the next generation IC into the device. As the number of transistors in the newer generation of ICs rapidly increased and the integration climbed to a higher level, this model was no longer suitable. The main problems were that the method did not identify the yield-determining process steps and it was not able to predict the impact of the reduction in transistor dimensions.
1.2.2. Poisson Model In 1963, Hofstein and Heiman [4] presented some theoretical characteristics of the ‘‘insulated-gate field-effect transistor’’ (see Figure 3.2). The transistor was described as a control electrode (gate) insulated from a thin conducting channel in the surface of a silicon substrate by an oxide film. They proposed the critical area to be the area under the gate (AG): if a defect occurs in this area, the transistor will fail and a defect outside this area has no or minor impact. The probability that a defect occurs in the critical area is considered to be totally random and thus independent of surface structure, differences in local surface composition, or the presence of other defects. Under these conditions, the Poisson probability function can be used to calculate the yield of a process for which the defect density (D) is known (equation (2)). This model has proven
86
Developments in Surface Contamination and Cleaning
gate L
SiO2 channel Source
drain W Silicon wafer
FIGURE 3.2 A typical layout of a field effect transistor (not to scale). W is the gate width that is considered the critical dimension of a transistor and L is the gate length. The critical area AG WL.
to be very accurate in predicting yield for products with a total die area below 0.25 cm2.1 Y ¼ e
NAG D
(2)
It is the definition of defects resulting in device failure that has been the topic of dispute. D is composed of a measured defect density, where a certain size is considered to be critical, and of the kill ratio of the defect. The kill ratio is the number of particles that causes a defective product divided by the total number of particles in the critical area. Defects are measured and sized with optical techniques. Bare or patterned wafers are scanned with a laser beam and when the beam illuminates a defect the light is scattered. The scatter intensity is a measure for the defect size. However, the scatter intensity not only depends on the defect size but, among other things, it also depends on the defect composition. Defects measured are particles, pits and asperities, and pattern deformation. Only the first category, particles, is of interest in this chapter. Their composition remains unknown. The size at which particles become a killer is rather arbitrary. As mentioned before, the size depends on the most critical dimension (i.e. W) of the device and is defined as a fraction of that size. This fraction is currently 0.5. In fact, it does not matter what fraction is chosen, because the kill ratio will compensate for errors in the chosen size. The kill ratio is determined by actually measuring a certain number of defects of the same category and it determines the number of devices that failed due to these defects. It is a calibration factor. It becomes clear that if the size at which a defect was measured is chosen to be smaller, more defects will be measured and the kill ratio will decrease accordingly. Besides the size 1 Instead of one specifc critical area, the gate area (NAG), most yield models are more general and use A as total area in a certain process step. Therefore, for NAG also read A.
87
Particles in Semiconductor Processing
compensation, the kill ratio is also needed because the tools to measure defect density do not measure the composition of the defect. Some defects are not killer, e.g. organic particles will disappear in a furnace oxidation. Since in manufacturing the compositions of defects in most process steps are unknown, the kill ratio has to be determined for each process separately.
1.2.3. Murphy’s Model The Poisson model is correct for one process with a known fixed defect density. However, the defect density is not constant in time nor is the defect density uniform over the wafer. It is subject to random variation that results in some clustering of defects. The effect of such clustering is that the chance that an IC in such area or time frame is affected by two defects will increase. This IC can only fail once. If these clusters had spread out over the whole wafer or in time these two defects would have hit two ICs and both would have been scraped. Since the relation between defect density and yield is not linear, the long-term yield will be underestimated (see Figure 3.3). Murphy [5] included the long-term variation in defect density in the longterm yield calculations by using a (normalized) distribution function f(D) that describes this variation: ðN e NAG D f D dD (3) Y ¼ 0
1
0.8
Yield (fraction)
0.17 0.6 0.13 0.4
0.2
0
0
10
20
30
40
50
Defect density (AU) FIGURE 3.3 Relation between defect density and yield according to the Poisson model. If the average defect density is 10 and it fluctuates randomly between 5 and 15, then the long term yield will be underestimated by a maximum of 2%.
88
Developments in Surface Contamination and Cleaning
The function can be determined from the manufacturing history or it can be assumed to have a certain shape, such as rectangular, bell curve, triangular, box, or line. Murphy chose the triangular shape and obtained: 2 NAG D (4) Y ¼ 1e NAG D
1.2.4. Negative Binomial Model A more physical meaning can be given to the variation in defect density. Incidents are the reason that defects cluster in place or time. Wafers are hit by splashes, dry-in marks, scratches, or some other localized random events, causing many defects to be created in a localized area. Also, one of the many processes or process tools can run out of control at some moment in time. The key property is that the final defect count is the sum of many processes that can run out of control, adding a little to the total defect count. Since most processes and tools run within control, the defect distribution becomes skewed with a maximum close to zero defects. The statistic that best describes such nonsymmetric variation is the gamma distribution. Therefore, Stapper [6,7] expressed the variation in defect concentration by a gamma distribution function and used that function for f(D) of equation (3). After integration, he derived: NAG D a (5) Y ¼ 1þ a The new parameter a is a clustering parameter that is equal to D2/s2, where s2 is the variance in the defect density. According to Stapper [6], a varies from 0.3 to 5, but is typically between 2 and 3 [8]. More clustering (more incidents) results in a larger variance, smaller values of a and, thus, a larger predicted yield than predicted with Poisson’s model with the same average defect density. As an effect of the gamma distribution function, this model becomes a sort of unifying model. If a becomes larger than 10, this models starts to overlap with the Poisson model (i.e. no clustering). When a is between 4 and 7, the yield prediction is similar to Murphy’s model. With a ¼ 1, this model is equal to Seed’s model [2], which is often used as an alternative to Poisson’s model. In 1991, Stapper [9] emphasized that the negative binomial yield model has found general acceptance in semiconductor manufacturing in Canada, Europe, and the USA. The Poisson yield model is the model of choice for comparing data from single process steps and is used in ITRS roadmap discussions [1].
Particles in Semiconductor Processing
89
1.2.5. Random Defects and Non-Random Deposition Model In all yield models defects are systematic, identified with in-line defect inspection tools, and well classified. Furthermore, the defects are assumed to deposit randomly on any area of the whole wafer. In a recent study [10], defect density in the wafer environment was related to yield. This means that the measured particles in ultra-pure water (UPW), which is used for device manufacturing, relate to the in-line measured defect density on the respective devices. The defect density (D) due to particles present in UPW can be described by our predictive model: D ¼ Np SKR Pd
(6)
Here Np represents the particle concentration (particles cm 3 in UPW), S (L cm 2) is the amount of UPW that contacts the wafer during the fabrication at the critical process steps, KR represents the fraction of killer particles, and Pd is the probability that particles deposit on to critical areas. The value of Np and S can be measured. The unknown value KR varies with the particle composition and size. Pd depends on the process settings of the manufacturing process step generating contamination, particle composition, and wafer surface composition. It is Pd that is the subject of our further investigations as it can have a large impact on yield, if the particle would have a preference to deposit on critical areas. It was shown [11] that particles do deposit on specific areas during the drying sequence of a cleaning procedure.
1.3. Origin of Particles The composition of the particles is diverse and it depends on location and time. Failure analysis of particles using SEM-EDX that resulted in a device failure shows, in many cases, the presence of Si, but whether it is Si, silica, silicon nitride, or organosilicon compounds (e.g. dimethylsiloxane) remains unknown in many cases. Another category of particles detected in failure analysis is particles related to photoresist residue. Organic material is difficult to discern, but fluoride residues can be measured and they are very likely related to the fluoride used in the plasma to etch patterns. These residues are either not removed after the plasma etch step, or they are redeposited out of the cleaning solution used to do the post-etch cleanup. The non-process-related particles originate from the wafer environment, such as cleanroom air and process liquids (chemicals and water), or from the wafer edge. Cleanroom air. Manufacturing of ICs is done in cleanrooms, but the handling of wafers is done in an even cleaner microenvironment. Process tools are completely enclosed with their own filter system and the wafers are stored and transported in closed pods that have standard interfaces to connect the pod to the tool. This enables the loading and unloading of
90
Developments in Surface Contamination and Cleaning
wafers into the tool without exposing the wafer to cleanroom air. Thus, the probability that particles come from the cleanroom air is relatively low. Process liquids. This is an important source of particles. If particles are present in the liquids they have plenty of opportunity to end up on the wafer surface (see Figure 3.4) [11]. During the immersion of the wafers into the liquids, hydrodynamic forces act on particles floating on the liquid whereby the particles are ‘‘stamped’’ on the wafer surface [12]. When the wafers are in the submerged state, particles are deposited due to electrostatic forces and van der Waals forces [13]. The withdrawal of the wafers from the liquid determines the amount of liquid left to dry on the surface [14]. All contamination in the drying liquid will be left on the surface. The final step, drying, does not determine the number of particles left on the wafer, but rather the location on the wafer [11]. Wafers are exposed to liquids during lithography, cleaning, wet etching, galvanic deposition, spin-on layers, and polishing steps. These process steps make up more than 50% of the total number of process steps. Wafer edge. The edge of the wafer is a known source of contamination. At the edge contamination is accumulated and generated. Accumulation can occur because many cleans do not target the cleaning of the wafer edge. Generation is due to mechanical mishandling of the wafer that is done with grippers manipulating the wafer, or during transport where wafers often bump into solid surfaces. Particles are also generated at the edge, because the deposited layer ends at the edge in an uncontrolled way and the layers can peel and flake off. If these wafers are immersed, particles can dislodge and redeposit on the wafer. Usually, the pattern of deposition is recognizable with inspection tools as it occurs as a smear from the edge. Besides the aforementioned resist residues the major source of process-related particles are deposition tools. During deposition, material is deposited on the wafer and on the wall of the reactor. If the layers on the reactor become too thick, these layers can crack and flakes will come loose. In some cases, it is the gas mixture at the start of the deposition process that is of incorrect composition which causes particles to form in the gas. In all these cases, particle deposition
1
3
4
2
FIGURE 3.4 Opportunities for particles to deposit on a wafer during the bath cleaning process: (1) immersion; (2) submersion; (3) withdrawal; (4) drying.
91
Particles in Semiconductor Processing
is unanticipated and the risk can be reduced by increased tool cleaning sequences, or by adding an extra wafer-cleaning step after the deposition step. Chemical mechanical polishing (CMP) is another process step that adds particles to the wafer. However, this is anticipated and there are dedicated postCMP cleaning tools and processes available. Many of these processes do not target the wafer edge properly. If particles deposit on the wafer, it is assumed to be a random process. Yield models are using this boundary condition (see Section 1.2). It has been shown that particles coming out of a liquid do not deposit randomly [11]. If a surface has mixed areas that are hydrophobic and hydrophilic, particles tend to deposit on the hydrophilic areas, depending on the nature of the particles. If the surface has structures, the particles will tend to deposit next to the sidewalls. Convection flows and lateral capillary forces during the drying step drive these processes. A problem for technologies with smaller feature sizes is that the smaller particles that might be critical have a higher tendency to deposit preferentially.
1.4. Determination of Particle Removal Efficiency (PRE) In order to evaluate a cleaning process, a method is required to determine the particle removal efficiency of the process. This has to be done with particlecontaminated wafers. In early manufacturing, these wafers were prepared by putting the wafers through a process or tool that was known to contaminate the wafer. Subsequently, the number of particles was measured (pre-count), the wafers were cleaned, and the number of particles was re-measured (postcount). The PRE was calculated from: PRE% ¼ 100%
ðPre-count Post-countÞ Pre-count
(7)
This was a very pragmatic approach that could indeed give some indication of the performance of two processes under evaluation at the same time. The weakness of this approach is that the particles are of unknown origin and the composition can change on a day-to-day basis. Also, the amount of deposited particles could vary from wafer to wafer. Thus, for scientific purposes, this method is unsuitable. A second drawback of the method is that it does not account for particles added by the cleaning process itself. This means that if the pre-count is relatively small, the determined PRE will be offset by these added particles. An improvement is to use a series of particle-contaminated wafers with variable particle concentrations. If the pre-count and the post-countpre-count differences are plotted on an x y plot, the slope of the linear regression represents the PRE and the intercept represents the amount of particles added by the clean itself. An example of such an experiment is given in Figure 3.5.
92
Developments in Surface Contamination and Cleaning
Pre-count–post-count
1100 900
y = 0.93x 24
700 500 300 100 100
0
250
500
750
1000
Pre-count FIGURE 3.5
Example of a particle removal test: PRE
93% and added particles by the clean is 24.
The methods to contaminate the wafers have been improved by contaminating wafers with particles of known composition. In manufacturing, suspensions are generated by breaking silicon wafers in/above water (to generate silicon particle suspension) or by diluting slurries that are used for the polishing process (silica particles). These suspensions are applied on to wafers by spinning, or the wafers are immersed in these suspensions. For scientific studies, silicon nitride particles tend to be preferred, because these particles are believed to be more difficult to remove than silica or silicon. Problems related to this particle-application method will be discussed in Section 3.3.1. It is not common practice in the industry to monitor particle removal efficiency with particle-contaminated wafers on a regular basis. Instead, practical clean wafers are processed in cleaning tools and the amount of added particles (D[x]) by process and tool is monitored. It is wrongly assumed that the small amount of particles already present on the wafers is not removed. It was demonstrated that from such a monitor program the particle removal efficiency could be determined [15], because the small amount of particles is partially removed. By describing the cleaning process as an equilibrium reaction between particle attachment and detachment the following equation can be derived:
(8) D x ¼ Ad 1 e kr t Here the parameter Ad describes the amount of particles that are added by the process at infinite process time minus the initial particle count. kr is the removal rate constant of the particle detachment process. This formula makes more sense with the following definition of particle removal efficiency: PRE% ¼ 100% 1 e kr t and PRE ¼ 1 e kr t (9)
Particles in Semiconductor Processing
93
The performance of the cleaning process could be calculated by taking Ad and kr to be normally distributed [15].
1.5. Methods to Remove Particles Particle contamination accounts for 90% of the contaminants and is responsible for 80% of the defects [16]. The easiest way to avoid the impact of particles is to prevent the particles from depositing on the wafer. If this is not possible, the particles have to be removed. In the semiconductor industry there are many methods to remove the particles. They are based on chemical principles, physical forces, or a combination of these. Since the structures made on the surfaces are so small and fragile, there is a delicate balance between particle removal and structure damage. The main chemical principle is undercutting of the particle by etching the substrate. Additionally, surfactants can be added to either aid the detachment or prevent the redeposition of particles. Physical methods that aid the chemical processes are ultrasonics or megasonics [17], brush cleaning [18], or centrifugal forces [19]. More or less pure physical methods are bombardment of the contaminated surface with high-speed droplets [20,21], or with ice particles [22,23], or with laser-assisted cleaning [24,25]. Many other methods are under investigation, but have not found serious application in semiconductor manufacturing. Solutions that are used to remove particles are aqueous based. Solutions such as HF etch silicon oxide but not silicon, while such solutions as ammonia hydrogen peroxide mixture etch silicon and silicon oxide at more or less the same rate. The problem with the latter is that it can etch silicon in an uncontrolled way, resulting in increased silicon surface roughness [26]. A disadvantage of the HF solution is that the electrostatic double-layer thickness is minimal and repulsive forces that prevent redeposition of particles are weak. Furthermore, using megasonic cleaning in combination with HF can cause pitting of the substrate [27]. The first systematically developed silicon wafer cleaning process is the RCA clean introduced in 1965 by Kern and Puotinen [28]. It is called the standard clean (SC) and is based on a two-step process with intermediate water rinses. The first step is aqueous ammonium hydroxide hydrogen peroxide water mixture (APM), also called standard clean 1 (SC1), and the second step is hydrochloric acid hydrogen peroxide mixture (HPM), or standard clean 2 (SC2). It is the APM step that targets the removal of particles and the HPM step that adds particles. APM has been the major workhorse for particle removal in the semiconductor industry and has been upgraded throughout the years. Both APM and HPM were originally used at 75 80 C which resulted in excessive hydrogen peroxide decomposition. For this reason, nowadays APM is used at
94
Developments in Surface Contamination and Cleaning
lower temperatures and hydrogen peroxide is left out of the HPM. Also, the concentration has been a target of improvement. APM was originally used in a mixing ratio of 1:1:5 8 (NH4OH (25%) H2O2 (30%) H2O) and this has gone down to as low as 1:1:500 [29]. The addition of ultrasonic energy or megasonic energy allowed higher dilution ratios and lower operating temperatures [30]. Megasonic and ultrasonic cleaning are both acoustic cleaning methods. Megasonic cleaning uses a higher frequency (typically 800 kHz up to 2 MHz) than ultrasonic cleaning (<100 kHz). The principal mechanism of acoustic cleaning is pulsating gas bubbles moving across a particle-contaminated surface and thereby initiating the liftoff of particles [31,32]. Furthermore, these gas bubbles together with the sound waves cause micro-streaming which is a relatively non-harmful mechanism to enhance particle diffusion. However, these same gas bubbles can also collapse completely, creating high-energy hot spots, so-called cavitation [33]. Cavitation causes structure damage. Since in ultrasonic cleaning cavitation bubbles are larger than in megasonic cleaning, the damaged area is larger. Brush cleaning can be applied in two modes: contact mode and non-contact mode. In the contact mode, the brush directly touches the surface and applies a force on the particles present on the surface of the substrate. In the noncontact brush-cleaning mode, the spinning brush generates a fluid velocity that removes the particles by shear forces. Contact brush cleaning is mainly used on flat surfaces, for example, after the chemical mechanical polishing process, since full contact of a brush will damage the device structure. The mechanism of removal by high-speed droplets or high-speed ice particles is slightly different. High-speed ice particles (usually made up of solid CO2 or argon, formed by a sudden rapid expansion of the respective gases) require direct collision with the particle on the surface, whereby impulse is transferred and the particle is knocked away, or explosive evaporation detaches the particle from the surface. In the case of droplets, this mechanism is also possible, but it is also likely that the advancing front of an impinging droplet close to the particle generates enough shear forces to detach the particles. Highspeed droplets are created by aerosol formation with a high pressure of N2 in a dedicated spray nozzle. The most important parameter to obtain high particle removal efficiencies and low substrate damage is the control over the droplet or ice-crystal size distribution [34]. A laser can heat a localized spot very rapidly, which can cause an explosive expansion of the material. It induces either an explosive evaporation of a condensed liquid film between the particle and the substrate, or it generates a shock wave by heating the liquid just above the particle. Commercial tools are available for so-called laser cleaning that use the explosive evaporation mechanism. Some believe that the liquid film is not required and the particle is ejected from the wafer by the rapid expansion of the substrate [35]. Laser shock wave cleaning is in development [25].
Particles in Semiconductor Processing
95
2. THEORY 2.1. Particle Removal Mechanism In order to remove particle contamination from wafer surfaces, the removal forces that act on the particles should be greater than the attractive adhesion forces. Adhesion forces are extensively described in many textbooks [36,37]. However, the most important adhesion forces are van der Waals forces and electrostatic forces [38]. The strength of the van der Waals forces depends on the particle material as well as the particle size. Electrostatic forces depend, in addition to material properties, on the pH of the solution and can be attractive or repulsive. It is a longer range force than the van der Waals force. The major challenge of particle contamination in a semiconductor manufacturing environment is that particles are made up of a collection of different materials (see Section 1.3). Furthermore, some particles are deposited from liquid and others from air; some particles adhere chemically, others adhere purely physically; and some particles dissolve, while others are insoluble in the cleaning liquids. Therefore, the adhesion strength varies and the required force to remove them can be excessive for the majority of the particles. It is the detachment process that gets the most attention of researchers. As described in the previous paragraph, both chemical and physical forces are used to remove particles. There are many ways to apply a force on a particle, but the most important one is by shear forces. Shear forces are forces that act in the plane of the substrate, while the particle moves perpendicular to this plane. Due to the shear forces, the particle can break the strong chemical bonds and can start to roll or slide. The required liftoff forces are introduced if a sliding or rolling particle contacts a surface asperity or structure [39]. A mechanism that is underestimated by the industry is due to surface tension when a particle on a surface passes through a liquid gas interface. This is the case during the immersion of wafers into a liquid (Section 3.2), cleaning with high-velocity droplets, and also megasonic cleaning where gas bubbles move over the wafer surface. A very instructive description is given by Leenaars [40]. In his patent application, a laser creates a gas bubble and this gas bubble is moved over a contaminated surface (see Figure 3.6). The forces generated by the advancing and receding contact angles can result in enough liftoff force to overcome the adhesion force (FA). The maximum lift force (Fl,max) that acts on the particle is given in equation (10), where R is the particle radius, g is the surface tension of the liquid, q is the wetting angle of the particle, and a ¼ the contact angle of the liquid on the substrate.
96
Developments in Surface Contamination and Cleaning
Fl
FH gas
gas
liquid
Fl
FH liquid
FA
FA
a: particle moves into liquid
b: particle moves out of liquid
FIGURE 3.6 Particle removal mechanism with a gas bubble (only the gas liquid solid interface is depicted). When a particle goes through a gas liquid interface it will feel two major forces. FA is adhesion force between particle and substrate and FH is force caused by the surface tension difference at the liquid air interface [41].
Fl;max
q ¼ 2pRg sin cos a 2 2
(10)
From this equation three conclusions can be drawn about the cleaning system: 1. Best cleaning is done with liquids with high surface tension, e.g. water, and without surface tension-reducing agents (surfactants) added. 2. Substrate surface should be wetted by water and thus should be as hydrophilic as possible. Surface with contact angle larger than 90 will draw the particles to the surface. 3. It is much easier to remove hydrophobic particles. Leenaars [40] noted that the speed of the gas bubble should not exceed 10 cm s 1, because the interface of the liquid cannot adapt faster and the cleaning becomes less effective. On the same principle, Kittle [41] used a very large amount of gas bubbles as foam to remove particles.
2.2. Model of Particle Removal The removal of particles is considered to be a three-step process (see Figure 3.7). Physical and chemical parameters, such as temperature, viscosity, pH, and the presence of surfactants, can impact each step differently. The first step (ka) is the detachment of the particle from the surface, which brings the particle into the boundary layer. The second step (kb) is the diffusion of the particle through the boundary layer into the bulk of the process liquids. If this does not occur before the wafer is dried, the particle will redeposit on the wafer surface. Finally, the particle has to be carried away from the wafer environment (kc) by filtration or draining of the contaminated process liquid [42].
97
Particles in Semiconductor Processing
Boundary layer
ka
Bulk liquid
kb
kc
Silicon wafer
Filter
Measure concentration FIGURE 3.7 The particle removal process is described by three consecutive process steps: ka, the detachment; kb, the diffusion; kc, the removal out of the system.
Physical chemical processes control the detachment process. The rate at which particles are detached from the surface (ka) depends on a number of factors: Forces acting on the particle. The higher the force, the faster the particles will start to move and detach. Etch rate. The higher the etch rate of the substrate surface, the faster the undercutting process and the subsequent particle detachment. Redeposition. The lower the redeposition rate, the better the average particle removal efficiency. Making use of repulsive electrostatic forces between the substrate surface and the particles by selecting the proper pH can prevent redeposition. Addition of surfactants can also reduce redeposition. Once the particles are detached they have to diffuse through the boundary layer into the bulk liquid with a linear diffusion rate constant kb. The thickness of the boundary layer determines, to a large extent, the residence time of the particles. Thinner boundary layers will reduce the migration distance, which can be achieved by applying higher liquid flows, for example by using higher flow rates, applying megasonic energy, or using centrifugal forces. The rate of diffusion is also determined by the size of the particles, as well as viscosity and temperature of the liquid. The last process step is the actual removal of particles from the process (kc). For example, in a tank process where wafers are fully immersed in a process liquid, the process liquid is agitated and pumped around in a recirculation loop. It leaves the top of the tank and re-enters the bottom of the tank after passing a pump and a particle filter. If mixing in the tank is thorough, then the particle concentration will decrease exponentially over time. The time constant of this decay is determined by the flow rate and the particle filter efficiency.
98
Developments in Surface Contamination and Cleaning
This full process can be monitored by measuring the particle concentration in the bulk of the liquid with a liquid particle counter (LPC). The LPC uses light scattering for particle detection. To avoid any possible quantification and sizing issues, a monodisperse silica suspension is used for this purpose [43]. It is possible to measure the effect of the above-mentioned physical and chemical parameters on the rate constants ka, kb, and kc.
2.3. Liquid Particle Counter (LPC) Measurements The total removal process of particles can be considered as a chemical reaction process, where the particle concentration on the surface is expressed as [A], in the boundary layer as [B], and in the bulk liquid as [I]. First-order reaction kinetics are assumed and the respective reaction rate constants, k, can be determined from dedicated experiments. kc. If only the bulk of the liquid is filled with particles, some of the particles will diffuse into the boundary layer, but most of them are removed by the filtration action. The removal can be expressed by the differential in equation (11) using the continuously stirred tank reactor (CSTR) model and neglecting the particles that diffuse into the boundary layer: d½I ¼ kc ½I dt
(11)
where kc is the tank turnover frequency, which is determined by the flow rate divided by the tank volume. In this experimental setup, kc for a tank filled with ultrapure water is 0.0098 s 1. In subsequent experiments, kc can vary since the flow rate depends on the viscosity of the cleaning liquid, but it can always be determined independently from the other rate constants. kb. If the boundary layer is filled with a number of particles [B]0, some particles will precipitate on the surface (and become the particle concentration A), but most of them will diffuse out of the boundary layer, where they are subsequently removed. The concentration as a function of time that we are able to measure, [I], can be expressed by equation (12), which makes use of linear diffusion rates: k b kb t e kc t e (12) ½I ¼ ½B0 kc kb In our experiments, we found that we needed two types of corrections to fit the experimental data with the theoretical model: the fraction of particles removed upon immersion and non-ideal CSTR behavior. During the immersion process of the surface into the cleaning liquid, a fraction (f) of the particles is immediately removed by the shear forces
99
Particles in Semiconductor Processing
of the liquid, causing intermixing of a part of the boundary layer and the bulk liquid. Unfortunately, in this experimental setup f is not a constant, as it depends on the speed and the angle at which the wafers are immersed into the liquid. This was done manually. A tank filled with 25 wafers does not behave as an ideal CSTR. ki is a time constant that is a measure of the rate of the system to come from a non-ideal CSTR state to the ideal state. Equation (13) describes the corrected particle concentration in the bulk of the liquid as measured with a liquid particle counter: kb k t kc t kc t b þ 1f e e 1 e ki t (13) ½I ¼ ½B0 f e kc kb It appears that many parameters are necessary to fit some experimental data, but some of the parameters (ki ¼ 0.0166 s 1, kc ¼ 0.0098 s 1) are determined only one time and are kept constant for all experiments. In Figure 3.8, an example is given with kb ¼ 0.00503 s 1, f ¼ 0.432, and [B]0 ¼ 18,135 particles mL 1. The need to use a non-ideality constant appears obvious from these results. 5000 Exptl. data non ideal CSTR
particles (n/mL)
4000
ideal CSTR 3000
2000
1000
0
0
200
400
600
800
1000
t (s) FIGURE 3.8 Example of experimental data fitted (least squares) by a model with ideal CSTR behavior and one that is corrected for non idealities; n is the number of particles.
ka. Similar to the previous derivation, the concentration of particles in the bulk could be modeled by equation (14) if the particles on the surface would immediately start to detach as a kind of first-order reaction rate: ka kb e ka t e kb t þ e kc t e ðka kb þkc Þt (14) ½I ¼ ½A0 ðkb ka Þðkc kb Þ Also, this can be corrected for the non-ideal CSTR model and for the fraction of particles this is immediately sheared off during the immersion of
100
Developments in Surface Contamination and Cleaning
the wafers into the cleaning liquid. The complete model that describes the process in Figure 3.7 is given by: ka kb ½I ¼ ½A0 f e kc t þ 1 f e ka t e kb t þ e kc t ðkb ka Þðkc kb Þ ðka kb þkc Þt (15) e 1 e ki t The results we have obtained so far for the removal of particles from a silicon surface cannot be fitted with equation (15). The reason is that the removal process (going from A to B) is not correctly modeled. In Figure 3.9, data from a typical removal experiment are depicted and the particles are removed in two steps. The best fit is obtained by a model that assumes that a fraction (f) is removed by shear forces and a second fraction (1 f) is removed according to a stochastic process. A Gaussian distribution with an average particle cleaning time and a standard deviation (s) describes the latter process, but the distribution becomes skewed by the process mentioned earlier. The correction for the skewness and determination of the correct average cleaning time is done with equation (13), but in this case [B]0 is replaced by the Gaussian error function that was approached by summation over time. The fitting parameters are given in the caption of Figure 3.9. It should be noted that ki and kc are identical (these were kept fixed) to what was previously measured, kb is slightly smaller (perhaps due to the slightly longer average diffusion path), and f is larger.
Exptl. data best fit
Particles (n/mL)
400
300
200
100
0
0
1000
2000
3000
4000
5000
t (s) FIGURE 3.9 Example of experimental data fitted (least squares) by a model with non ideal CSTR behavior whereby particles are detached after an average cleaning time of 2280 s with s 228 s. Other parameters: kb 0.0048 s 1; kc 0.0098 s 1; ki 0.0166 s 1; f 0.93; background offset 17 n/mL 1; [A]0 1290 n/mL 1; n number of particles.
Particles in Semiconductor Processing
101
2.4. Metal Ion Core Particles Removal of nanosized particle contamination from the surface of the substrate is an essential requirement for IC manufacturing with critical dimensions smaller than 100 nm. However, for an effective particle removal study, proper detection of the particles is important. The conventional and previously described method to detect the particles is based on light scattering. The problem with this method is that it has size limitations (>100 nm) and it cannot distinguish between the signals of nanoparticles and background noise or surfactant micelles. To circumvent this problem, particles for use with an alternative detection method were synthesized. In 1968, Sto¨ber et al. [44] developed a method known as the Sto¨ber synthesis capable of producing nanosized monodisperse silica particles using ammonia, ethanol, and tetraethoxysilane (TEOS). Van Blaaderen et al. [45] described the so-called seed technique to synthesize particles with a luminescent core. First, particles were formed together with the luminescent material and, second, the particles were grown further by adding clean TEOS to the suspension. The net effect is a silica particle with a pure silica shell and a slightly contaminated core. UV detection tools are not sensitive enough to measure the four to five orders of magnitude of concentration changes required for the particle removal study. The seeded technique was used to make metal-ion core particles. Instead of the luminescent material, a diluted scandium nitrate solution was added. Scandium was chosen because it is rarely used in the microelectronic environment (low background intervention) and this metal ion attaches very strongly to silica. The size of the particles is controlled by varying the reactant quantities [46]. Quantification of the particle concentration can be done by measuring the metal ion concentration, either in the solution or on the wafer surface with inductively coupled plasma mass spectroscopy (ICP-MS). It was found that the scandium does not leach out of the particles even at pH 1 [47].
3. PARTICLE REMOVAL STUDY 3.1. Tank Dynamics, Impact of Particle Counter, and Particle Composition In the previous section, a differential equation was derived to measure the tank dynamics in a CSTR (equation (11)). Solving this equation results in a function where the particle decay is exponential with time:
(16) I ¼ ½I0 e kc t To establish the parameters that describe this process, a tank was filled with particles and homogenized before the recirculating liquid was filtered. As the goal was to use HF-based cleaning solution, particles with a low etch rate in
102
Developments in Surface Contamination and Cleaning
HF were used for this study. They were composed either of a mixture of different size polystyrene latex spheres (PLS) or Si3N4 particles with a wide size distribution. It was found that the apparent removal rate of particles depends on the particle composition and the particle size. Figure 3.10 summarizes the results of particle removal experiments using PLS and Si3N4 particles. The removal rate is expressed in half-life time, t½, that relates to kc as t½ ¼ ln 2/kc. In this case, the theoretical half-life time, which is based on the measured tank volume and flow rate, would have been 49.6 s. In the data supplied by the liquid particle counter, this value is only measured for the large Si3N4 particles. After correction of the data for optical coincidence, all sizes of Si3N4 particles are removed at the rate expected on the basis of tank dynamics, while PLS is filtered out much faster than expected. This does not change much after a second correction for optical coincidence resulting in size promotion (two particles coincide, resulting in a measurement of one particle of the next larger category). As the filter used had a capture efficiency >99% for particles with sizes larger than 0.05 mm, this effect could not have been caused by the difference in filter efficiency. An explanation for the faster than theoretical removal of PLS out of the recirculating liquid and the filter tank is that there is an extra transport mechanism for the particles towards the surface of the liquid. Since the tank overflows and the liquid leaving the tank comes from the surface, a higher than average concentration of particle-contaminated water leaves the tank. A particle with an apparent weight smaller than the total pull of the meniscus
60 0.1−0.12 µm 0.12−0.15 µm 0.15−0.25 µm
55
t½ (s)
50
45
40
35
PLS raw
Si3N4 raw
PLS cor
Si3N4 cor
FIGURE 3.10 Half life time of the filtration process of particles out of a recirculated tank. Raw data represent data as given by the measurement tool; ‘‘cor’’ is the same data after correction for optical particle coincidence in the measurement tool. The solid black line is the theoretically expected t½ with 100% filter efficiency (49.6 s) [44].
Particles in Semiconductor Processing
103
around it will float, or it will have a high tendency to be located at gas liquid interfaces [48]. Since PLS is organic in nature, it will have low surface tension and its preference for location at the gas liquid interface will be much higher than that of the hydrophilic Si3N4 particles. Furthermore, small gas bubbles that are present due to agitation of the liquid will be the extra transport carrier for the particles to become submerged, as particles will attach to the gas liquid interface of the bubble. This property makes PLS unsuitable as model particles in particle removal studies in liquids. Additionally, if these particles would have deposited on the silicon surface, the lift forces induced by the advancing contact angle will significantly aid the detachment of PLS from the surface (equation (10)). Indeed, PLS deposited on silicon have been found to be removed in large part upon the immersion in water, while particles that are present on product surfaces are not removed in such a step. The disadvantage of Si3N4 is that it is commercially not available as a monodisperse suspension. Silica has similar surface properties as Si3N4 and is available as monodisperse suspensions [49]. However, besides the fact that silica is etched in HF solutions, experiments with the removal of silica ran into a completely different problem. As the refractive index of silica is so much closer to water than Si3N4, the scatter intensity becomes less efficient. Therefore, particles with a nominal mean diameter of 0.16 mm are invisible to a detector with a detection limit lower than 0.1 mm latex sphere equivalent (LSE), while 0.33 mm particles will appear at 0.15 mm and smaller, which is close to the detection limit. Still, silica will be the particle of choice for further particle removal studies, because they can be made with a marker that allows alternative detection methods. The preference for Si3N4 for particle removal studies is based on the more challenging conditions for its removal compared to that of silica [50]. This difference can be caused by the methods of detection, which are all based on light scattering. The scatter cross-section of Si3N4 is much larger than that of silica for a particle of the same size. In water the difference is almost three times (see Figure 3.11) and in a vacuum it is 1.3 times. Consequently, a study using a liquid particle counter in which 100 nm (LSE) Si3N4 and silica particles are being removed is actually removing 70 nm Si3N4 and 200 nm silica particles. It is easier to remove larger particles and therefore Si3N4 seems more difficult to remove [51]. It is important in these kind of experiments to check the presence of dead volume, which is areas in the tank where liquid is only refreshed by diffusion. If a dead volume becomes filled with a significant amount of particles, it will deliver particles to the bulk slower than the removal of particles out of the bath. Consequently, the determined ka value will be larger than it actually is. It was found that a full batch of 25 wafers placed in a specific way into the tank behaves as a semi-closed box. The flow through the wafers was limited. If the particle level in the tank was reduced to nearly zero and the wafers were taken out of the solution, a peak in particle concentration could be observed.
104
Developments in Surface Contamination and Cleaning
0.3 PLS 0.25 Si3N4
LSE (µm)
0.2
0.15 SiO2 0.1
0.05
0
0
0.05
0.1
0.15
0.2
0.25
0.3
Particle diameter (µm) FIGURE 3.11 The scatter cross section of particles in water (nliq approximation expressed as latex spheres equivalent (LSE) [44].
1.34) using the Rayleigh
Reducing the number of wafers to avoid flow restrictions between wafers was not an option, because a large contaminated surface area is required to enable a particle concentration 1000 times above the background level.
3.2. Particle Diffusion Out of the Boundary Layer The boundary layer is a mathematical concept and describes the thickness of the liquid layer on top of a solid surface where the liquid is not flowing, while the bulk of the liquid is flowing. However, the velocity of the flowing liquid does not decrease abruptly from maximum velocity to zero on approach to a solid surface. Therefore, the boundary layer is defined as the distance in the liquid from the solid surface where the velocity is smaller than 99% of the velocity at infinite distance. In the boundary layer approximation, the boundary layer thickness is assumed to be constant over the whole wafer surface, which is only true when the liquid velocity is constant over the whole surface. Since the boundary layer is a liquid layer with almost zero velocity, within the boundary layer the only mass transport mechanism will be diffusion. In the laminar flow regime, the boundary layer thickness depends reciprocally on the square root of the water velocity at infinite distance and is around 1 cm in most semiconductor applications. A second layer that is of interest is the carry-over layer. This is the layer that is left behind on the surface after taking the surface out of a liquid (Figure 3.12). The carry-over layer is much thinner than the boundary layer and, in most cases,
105
Particles in Semiconductor Processing
v (cm/s) Carry over layer ~ 0.0007 v 2/3
~ 20 µm FIGURE 3.12 of the liquid.
Carry over layer is the liquid that remains on the surface after removal of the bulk
is of the order of 20 mm. All the contaminants that are left in the carry-over layer will remain on the wafer surface when the liquid is evaporated. It is assumed that the concentration of contaminants in the boundary layer is uniform and, therefore, the same as in the carry-over layer. Therefore, by using an LPC that enables the measurement of the cleaning rate of the boundary layer, it also measures the cleaning rates of the carry-over layer. Besides the distance a particle has to travel, its velocity also determines the residence time in the boundary layer. Particles are placed in motion due to random collisions with liquid molecules, which results in so-called Brownian motion. The average kinetic energy of the particles, {1/2}mv2 with v ¼ average velocity and m ¼ particle mass, is discrete values of {1/2}kT (here k is the Boltzmann constant). The resulting diffusion rate of particles is described by Einstein’s law of diffusion. This law, combined with the Stokes friction factor, results in: D ¼
kT 3phd
(17)
This equation shows that the diffusion coefficient (D) depends linearly on temperature (T) and relates reciprocally to the viscosity (h) and the particle diameter (d); the latter can also be used to convert to the mass of the particle. To determine the effect of all variables on diffusion rates of particles out of the boundary layer, the boundary layer is filled with particles by making use of the carry-over layer. Immersing a batch of dry silicon wafers in a slightly alkaline suspension of silica particles, and subsequently taking them out, will result in a carry-over layer on the wafers filled with particles. Alkalinity stabilizes the suspension and prevents the particles from depositing on the wafer surface. During reimmersion of these wafers in a clean process tank, a fraction (f) of the carry-over layer with particles will be rinsed off immediately, but a significant fraction will remain in the boundary layer from which particles slowly diffuse into the bulk of the liquid. Measuring the particle concentration in the bulk results in a concentration profile ([I]) described by equation (15). An example of such a measurement was previously given in Figure 3.8.
106
Developments in Surface Contamination and Cleaning
3.2.1. Effect of pH on kb The effect of pH on the linear diffusion rate of particles leaving the boundary layer was measured for 0.33 and 0.43 mm diameter silica particles [41]. The pH of the cleaning solution was adjusted with ammonia and nitric acid and the diffusion rate constants were calculated from the average of three to five particle concentration measurements. The method described in Section 2.3 has been used to fit the data and establish the one-dimensional diffusion rate constants. Figure 3.13 summarizes the results. The smaller silica particles indeed diffuse faster than the larger silica particles. Surprisingly, there was a strong effect of the pH on the linear diffusion rate. For both particle sizes, the diffusion rate is highest at a pH of 1.8, which is also known to be the point of zero charge for silica surfaces. This leads to the conclusion that the thickness of the electrostatic double layer (EDL) around the particles affects the effective radius of the particle and effectively slows down the diffusion rate. In this case, the EDL thickness would be thinnest at the point of zero charge unless the EDL is compressed by a high salt concentration. 3.2.2. Effect of Salt Concentration on kb The hypothesis of the impact of the compressed EDL can be tested by adding extra salt to the aqueous solution while keeping the pH fixed and subsequently determining kb. The results of such an experiment are summarized in Figure 3.14. The pH of the solution is fixed with ammonia at 9.5. By adding extra NH4NO3 the salt concentration increases. Particles diffuse faster out of the boundary layer with high salt concentration than with low salt concentration. This confirms that 20 330 nm
430 nm
kb (ms-1)
15
10
5
0
1
1.8
3
67
10
11.5
12
pH FIGURE 3.13 Linear diffusion rate constant for 0.33 and 0.43 mm silica particles in aqueous solution at different pH values [43].
107
Particles in Semiconductor Processing
18 15
kb (ms-1)
12 9 6 330 nm
3
430 nm 0 0
10
20
30
40
50
[NH4+] (mmol/L) FIGURE 3.14 Linear diffusion rate constant of 0.33 and 0.43 mm diameter silica particles in aqueous solution at pH 9.5 and various salt concentrations (NH4NO3). The dashed line is the maximum measured diffusion rate at pH 1.8.
the thickness of the EDL has an effect on the diffusion of particles out of the boundary layer. However, the maximum diffusion rate at pH 9.5 with increased salt concentration is still lower than the diffusion rate at a pH of 1.8. Several explanations can be proposed for this phenomenon.
3.2.3. Effect of Temperature and Viscosity on kb According to equation (17), increasing temperature will accelerate the diffusion process and thus the removal of the particles out of the boundary layer. This is due to the higher velocity of the particles, but also due to the reduced viscosity of the liquid resulting in thinning of the boundary layer. Consequently, the distance through which the particle has to migrate is shorter. The measured diffusion rate in this experimental setup is a combined effect. To study the effect separately, a constant viscosity at different temperatures is obtained by the addition of polyethylene glycol (PEG) to the cleaning solution. The side-effect of adding PEG, although concentrations are kept low (0.005 0.04 mmol L 1), is that the dielectric constant of the liquid medium changes as well, which in turn influences the EDL. The results for 0.43 mm particles are summarized in Figure 3.15. It was no surprise that the diffusion rate increased with increasing temperature and decreased as the viscosity increased by the addition of PEG. However, looking at the data points marked with an asterisk in Figure 3.15, which have an identical viscosity at different temperatures, the diffusion rate does not increase with increasing temperature, instead it decreases by 24%. This means that the PEG, which is also a non-ionic surfactant, chemically or physically interacts with the particles, which results in an increase in the effective radius/mass of the particles and thereby decelerates the diffusion process.
108
Developments in Surface Contamination and Cleaning
18
15
kb (ms-1)
12 PEG (g/L) 0.00
9
1.62 4.63
6
13.89 3
0
*
* 20
25
30
35
T (ºC) FIGURE 3.15 Linear diffusion rate constant of 0.43 mm silica particles as functions of temperature and PEG 8000 concentration. Bars marked with an asterisk have identical viscosity at a fixed pH of 3.5 [43].
3.2.4. Effect of Surfactant on kb Surfactants are considered to aid the particle removal process, while in the previous experiments it was shown that the surfactant properties of PEG reduce the diffusion of the particles out of the boundary layer. Therefore, smaller surfactant molecules were used to study the specific effect of surfactants on kb. The selected surfactants were a cationic surfactant, benzalkonium chloride (BAC; 0.02 mmol L 1), and a non-ionic surfactant, polyoxyethylene(12) tridecyl ether (0.006 mmol L 1), which is similar to PEG but with a much lower molecular weight. The same experiments as with PEG were repeated, but now in the presence of a surfactant (BAC) whose concentration was kept well below the critical micelle concentration (CMC). At pH 3.5 the diffusion rate is considerably reduced by the addition of a surfactant (compare Figure 3.16 with Figure 3.15). The effect of PEG in the solution containing BAC has become insignificant and the only remaining effect is the impact on viscosity. In solutions with constant viscosity, the diffusion rate now tends to increase slightly with increasing temperature. In the solution without any PEG, the effect of surfactant is greater than if PEG was present. To visualize this effect, the change in diffusion rate constant has been plotted as a ratio of kb with surfactant to kb without surfactant (Figure 3.17). The effect of temperature was removed by taking the average value of the four temperatures. At pH 3.5, both the cationic and non-ionic surfactants decrease the diffusion rate in solutions without PEG, while at pH 10 the addition of the surfactants seems to have no effect. Increasing the PEG concentration, the effect of
109
Particles in Semiconductor Processing
18
15
kb (ms-1)
12
PEG (g/L) 0.00
9
1.62 4.63 13.89
6
3
* 0
* 20
*
25
30
35
T (ºC) FIGURE 3.16 added [43].
The results of the same experiments as in Figure 3.15, but with 7 mmol L
1
BAC
additional surfactant at pH 3.5 is canceled out, whereas at pH 10 the diffusion rate initially is accelerated with increasing PEG concentration, but this effect subsequently also fades. All these observations are explained by a mechanism that impacts either the effective particle radius, i.e. the particle diameter and some part of the electrostatic double layer, or the effective particle mass, i.e. the particle mass plus the material adsorbed on it. No surfactant and no PEG. The pH impacts the effective radius of the particle (see Figure 3.13). At pH 3.5, the zeta potential of the silica particle is smaller than at pH 10 and, therefore, the electrostatic double layer at pH 10 (R5 in Figure 3.18) is thicker than R1. With surfactant and no PEG. If only a cationic surfactant is adsorbed on the surface, the surface charge is shielded, resulting in a thinner electrostatic double layer compared to no adsorption (R1 and R5 versus R2 and R6 respectively). Also, the effective mass of the particle slightly increases. The net effect on the diffusion rate is that at pH 3.5 it is significantly reduced, and at pH 10 there is a balance and no effect is measured (Figure 3.17 at zero PEG concentration). With surfactant and with PEG. Surfactants and PEG have different impacts. These molecules are in a competition for adsorption sites on the silica particle. Replacing a surfactant molecule with PEG implies less charge shielding and thus an increased effective radius, but the total mass of the particle will increase. For this reason, the diffusion rate constant increases when the PEG concentration is not too large (between 0 and 10 g L 1 in Figure 3.17).
110
Developments in Surface Contamination and Cleaning
1.2 cationic
kb + surf / kb no surf
non ionic 1.0
0.8
0.6 pH 3.5 0.4
0
5
10
15
PEG (g/L) 1.6
kb + surf / kb no surf
cationic non ionic 1.4
1.2
1.0 pH 10 0.8
0
5
10
15
PEG (g/L) FIGURE 3.17 Ratio of kb with surfactant and kb without surfactant at different PEG concen trations (average overall temperatures). Left figure is at pH 3.5. Right figure is at pH 10 [43].
No surfactant and with PEG. PEG adsorbed on the particle surface will have only a minor effect on the effective radius, but it will increase the effective mass. Besides the increased viscosity, the higher mass also reduces the diffusion rate (Figure 3.15). In the presence of the surfactant, PEG competes for surface sites. If the PEG concentration is high enough, the surfactant molecules are expelled from the surface by PEG. Therefore, the effect of the combination of surfactant and PEG at 14 g L 1 is essentially negligible (Figure 3.17). One of the contributing actions of megasonics, spinning, or droplet bombardment is that it reduces the thickness of the boundary layer. A thinner boundary layer means that the particles are removed away faster from the wafer surface. This results in increased particle removal efficiency.
no PEG, + surfactant
+ PEG, + surfactant
+
+
+
pH = 3.5
+
+
R1
+ +
+
+
+ +
m2
m1
+
+ +
+
R5
+ +
+
+
m5
+
R7
+
+
R8 +
+
m6
+ = cationic surfactant
+
+
-
+
+ +
+ +
+
+ m7
+
m8
= PEG
Impact of pH, surfactant, and PEG on the thickness of the electrostatic double layer and the effective mass of the particle. Sizes are relative and
111
FIGURE 3.18 notional.
m4 +
+
R6
+
+ +
+ +
+ +
+
m3 +
+
+
R4
+
+ +
+
+
R3
+
R2
+
pH = 10
+ PEG, no surfactant
Particles in Semiconductor Processing
no PEG, no surfactant
112
Developments in Surface Contamination and Cleaning
3.3. Particle Detachment Once all rate constants have been determined, the first step of the removal process can be studied, i.e. the detachment of particles from the surface by chemical processes. To this end, monodisperse silica particles (average size 0.33 mm) were deposited from an aqueous suspension on to wafers. Using a diluted APM solution at 35 C as a slow-etching cleaning solution [52], the freshly deposited particles came from the dried wafer surface in two waves (Figure 3.9). In the first wave, the loosely bonded particles come off upon the immersion of the wafers in the cleaning solution. This is probably due to hydrodynamic forces (either surface tension or drag forces) acting on the particles at the gas solid liquid interface. The second wave started much later. This peak is almost Gaussian in shape, but is slightly skewed because of the earlier-determined diffusion (kb) and filtration (kc) processes (Figure 3.19). The time scales of the detachment process and the subsequent diffusion and filtration are so much different that the integration of the respective time function (equation (15)) is not required. This can change if the detachment process is accelerated by higher etch rates or more physical power.
3.3.1. Impact of Deposition Condition on ka The measurement system to determine the particle removal rate is strongly impacted by the storage conditions of the particle-contaminated wafers prior to the cleaning process [53]. The effect on the process is observed in the height of the first peak and the average removal time of the second peak. If the
Hydrodynamic Forces
4000
delays caused by kb(diffusion), kc (filtration)
3000 height
Number of particles (/mL)
5000
2000
Etching of SiO2 1000
0 0
1000
2000
3000
time (s) average removal time FIGURE 3.19 Two key parameters, height of first peak and average removal time, used to described the particle aging process.
113
Height 1st peak (1000 particles/mL)
Particles in Semiconductor Processing
6 40% 70% 100%
5 4 3 2 1 0 0.1
1
10
100
1000
Storage time (h) FIGURE 3.20 Intensity of first peak as a function of storage time of particle contaminated wafers (RH 40%, 70%, and 100%).
Average removal time (s)
particle-contaminated wafers were stored for longer times, the height of the first peak decreases and the average removal time of the second peak changes. Under normal cleanroom conditions (relative humidity (RH) between 40% and 70%), the first peak disappears within 10 hours of storage (Figure 3.20). If RH during storage was set at 100%, the intensity of the first peak remained constant for the first 24 hours, after which the peak slowly decreased. The dependence of the average removal time on the storage conditions is depicted in Figure 3.21. The time to maximum of the peak increased at RH 40 70% from 2000 to 3000 seconds within the first 6 hours. After 6 hours the average removal time remained constant at around 3000 seconds. This means that more etching is required to remove all particles. During the first 24 hours of storage at RH of 100% (Figure 3.22), the removal time remained constant at around 1700 seconds; thereafter it increased to around 2300 seconds from 36 hours’’ storage onwards. Although in the experiments under dry conditions 4000 3000 2000 1000 0
0
2
4
6
8
Storage time (hours) FIGURE 3.21 Average removal time as a function of storage time of particle contaminated wafers (RH 40%, 70%). The dots are 70% RH and the triangles are 40% RH measurement points.
114
Developments in Surface Contamination and Cleaning
Average removal time (s)
3000
2000
1000
0
1
10
100
1000
Storage time (hours) FIGURE 3.22 Average removal time as a function of storage time of particle contaminated wafers (RH 100%).
(RH < 6%) the humidity conditions decreased during the first few hours, the average removal time was high and remained high (between 3000 and 4000 seconds) right from the start. With the assumption that longer etching means stronger particle substrate interaction, the experiments seem to indicate that removal of particles becomes more difficult upon storage. After the application and dry-in, particles are attached to the surface with certain strength, which will not be equal for all particles. On the basis of Figure 3.9, this strength is considered to be normally distributed over the particles (Figure 3.23). If external forces are applied during the immersion, the more weakly bonded particles will yield and detach from the surface, resulting in the first particle wave. During the subsequent etching process, the remainder of the particles will be removed, giving rise to the second particle wave. The effect of aging/storage is that the average adhesion
Number of particles
aging Immersion forces 2nd peak particles 1st peak particles
Particle adhesion force FIGURE 3.23
Adhesion force distribution of particles attached to a surface.
115
Particles in Semiconductor Processing
strength increases and that the immersion forces are not sufficient to remove particles. As a result, the first peak will vanish and the second peak that is fed by the slightly weaker bonded particles could show up initially a little earlier in the etch process, but eventually it will shift to longer etch times for complete removal. The effect of moisture on aging can be understood by the capillary forces causing a thin condensed water layer to exist between the particle and the wafer surface (Figure 3.24). Dry-in of such a condensation layer results in a shorter particle substrate distance up to an extent that the two interfaces are in the range of the van der Waals force. Subsequently, capillary forces acting on the particles cause the particles to deform, whereby a larger contact area is created and thus stronger adhesion is established by van der Waals forces. In the case of 100% humidity, the condensed layer is too thick to cause capillary forces to act on and deform the particles. However, in the first step, evaporation/condensation is likely to occur at 100% RH and, therefore, the first particle wave also disappears under these conditions (Figure 3.24). These results can be explained by assuming the first particle wave is due to particles floating on a thin water layer between the particle and the substrate. In the second wave, particles are detached from the surface, which are in the range of van der Waals forces. The increase in the adhesion force of the particles in the second wave is due to the particle deformation induced by capillary forces. Additionally, as explained later, the increase in adhesion can be a result of dissolved and subsequently precipitated silicates, which create actual chemical bonds between the particle and the substrate during the aging process. We have shown that the storage conditions during the time the particles are applied and removed have a great impact on the adhesion strength of particles on a silicon surface. This implies that particle removal studies using this kind of model system will have a problem with reproducibility and repeatability. This will depend on the laboratory, or even on the researcher, as to what the absolute outcome will be of a particle removal study. However, general trends will remain evident if the contamination part of the cleaning experiment remains exactly the same.
evaporation / condensation
van der Waals forces
capillary forces aging FIGURE 3.24 Evaporation of the condensation layer will result in stronger particle adhesion and eventually lead to a deformed particle having a larger contact area.
116
Average removal time (s)
Developments in Surface Contamination and Cleaning
2000
1500
1000
500
0
1
3
5
7
9
11
13
pH FIGURE 3.25 Effect of pH of the particle suspension used to contaminate the wafers on the removal process.
3.3.2. Impact of Deposition Conditions Particles that are deposited from a dry aerosol are relatively easy to remove. This is disadvantageous for particle removal studies because the method does not discriminate between a good and a better clean [54]. To make this more challenging, these particle-contaminated wafers can be immersed in a liquid, dried, and subsequently used for a cleaning experiment. The particles become more difficult to remove, which means that the deposition conditions impact upon the adhesion of the particles. In the previous experiments, silica particles were deposited from a neutral or slightly alkaline solution. If the deposition is done from a particle suspension at pH 2, the particles are very easy to remove. All the particles were removed or detached by immersion of the wafers into the liquid, i.e. there was no second peak. Increasing the pH of the particle suspension, the second peak appeared again and the average removal time increased with increasing pH (Figure 3.25). The first peak from the experiment with particles deposited at pH 2 was broader than expected for a process controlled only by ka. This indicates that a fraction of the particles ends up in the boundary layer upon immersion, which causes kb to become dominant in the removal rate. An explanation for the effect of pH on the adhesion strength of the deposited particles is that particles deposited at higher pH could be ‘‘glued’’ on the surface [55]. By dissolution and redeposition of silica at the particle substrate interface (Figure 3.26), the particles become chemically bound to the surface.
SiO2 + OH-
FIGURE 3.26
SiO3H-
Particles are glued to the surface.
Particles in Semiconductor Processing
117
FIGURE 3.27 SEM picture (left) of a surface formally covered with 1.5 mm sized silica particles leaving residues that have also been measured with AFM (right). (see colour plate section at end for coloured version) Courtesy of Frank Holsteyns [57].
This would not happen at low pH, because the solubility of SiO2 is much lower. Holsteyns [55] has obtained SEM images of the residue of such ‘‘glue’’ after the particles were removed (Figure 3.27).
4. CONCLUSIONS Particle research in the semiconductor industry is very pragmatic, but also very challenging. A state-of-the-art complementary metal oxide semiconductor (CMOS) manufacturer is interested in the removal of particles with no or minimal substrate loss, while in mature manufacturing the goal is to have ‘‘zero defects’’, i.e. less than one out of a million products that are allowed to fail. Due to this competitiveness, advances in fundamental research on particle removal are commercially implemented within a couple of years. Research is focused on mechanical and/or chemical enhancement of particle removal, but with minimal substrate etching. Addition of surfactants to cleaning solutions to aid particle removal might inhibit efficient particle removal if surface tension forces are required to lift the particle, or if particle diffusion away from the surface is a rate-limiting step. Prevention of particle deposition is not yet a major research topic. Particles in the wafer environment causing random yield loss are not part of current yield models. It is our intention to do so in future.
ACKNOWLEDGEMENTS The authors would like to acknowledge all the students who have worked on this particle project throughout the years: Yolaine Dumesnil, Sander Wolters, Roy te Brake, Michiel Enkelaar, Melvin Kasanrokijat, Romuald Roucou, Remi Peyrin, Florian le Goupil, Federic Michel, Adrien Maurel, Michel van Straten, Wybe Roodhuizen, and Clement Sieutat.
118
Developments in Surface Contamination and Cleaning
REFERENCES [1] The International Technology Roadmap for Semiconductors, 2008 edition, International Sematech, San Jose, CA; http://www.itrs.net/, 2009. [2] T. Kim, W. Kuo, Modeling manufacturing yield and reliability, IEEE Trans. Semicond. Manuf. 12 (1999) 485. [3] J.T. Wallmark, Design considerations for integrated electronic devices, Proc. IRE 48 (1960) 293. [4] S.R. Hofstein, F.P. Heiman, The insulated gate field effect transistor, Proc. IEEE 51 (1963) 1190. [5] B.T. Murphy, Cost size optima of monolithic integrated circuits, Proc. IEEE 52 (1964) 1537. [6] C.H. Stapper, Defect Density Distribution for LSI Yield Calculations, IEEE Trans. Electron Dev. 655 (1973) 20. [7] W. Kuo, T. Kim, An Overview of Manufacturing Yield and Reliability Modeling for Semiconductor Products, Proc. IEEE 1329 (1999) 87. [8] R.S. Hemmert, Poisson Process and Integrated Circuit Yield Prediction, Solid State Electr. 511 (1981) 24. [9] C.H. Stapper, Integrated Circuit Yield Statistics, IEEE Trans. Semicond. Manuf. 294 (1991) 4. [10] F. Wali, D.M. Knotter, F.G. Kuper, Impact of Particles in Ultra Pure Water on the Random Yield Loss in IC Production, Microelectronic Eng. 140 (2009) 86. [11] F. Wali, D.M. Knotter, T. Bearda, P.W. Mertens, Local Distribution of Particles Deposited on the Structured Surfaces, Solid State Phenom. 65 (2009) 145 146. [12] L. Mouche, F. Tardif, Mechanisms of Particle Removal from Silicon Wafer Surface in Wet Chemical Cleaning Process, J. Electrochem. Soc. 1684 (1995) 141. [13] W. Fyen, K. Xu, R. Vos, G. Vereecke, P.W. Mertens, M. Heyns, Particle Deposition from a Carry Over Layer During Immersion Rinsing, in: K.L. Mittal (Ed.), Particles on Surfaces 8: Detection, Adhesion and Removal, VSP, Utrecht, The Netherlands, 2003, pp. 77 128. [14] W. Fyen, P.W. Mertens, Study of the Carry Over Layer and its Importance for Wet Processing of Semiconductor Substrates, J. Electrochem. Soc. 443 (2006) 153. [15] D.M. Knotter, Cleaning for Sub 0.1 mm Technology: a Particular Challenge, in: C. Claeys, F. Gonzalez, R. Singh, J. Murota, P. Fazan (Eds.), Proc. ULSI Process Integration III, PV2003 6, The Electrochemical Society, Pennington, NJ, 2003, p. 349. [16] T. Hattori, Detection and Analysis of Particles in Production Lines, in: T. Hattori (Ed.), Ultra Clean Surface Processing of Silicon Wafers, Springer Verlag, New York, 1998, pp. 243 258. [17] A. Mayer, S. Schwartzman, Post CMP Cleaning Using Acoustic Streaming, J. Electronic Mater. 855 (1979) 6. [18] K. Xu, Nano Sized Particles: Quantification and Removal by Brush Scrubber Cleaning, PhD Thesis, K.U. Leuven, Leuven, Belgium (2004). [19] J. Visser, Particle Adhesion and Removal: A Review, Particulate Sci. Technol. 169 (1995) 13. [20] I. Kanno, N. Yokoi, K. Sato, Wafer Cleaning by Water and Gas Mixture with High Velocity, in: J. Ruzyllo, R. Novak (Eds.), Proc. Cleaning Technology in Semiconductor Device Manufacturing V, The Electrochemical Society, Pennington, NJ, 1998, p. 54. PV 97 35. [21] H. Hirano, K. Sato, T. Osaka, H. Kuniyasu, T. Hattori, Damage Free Ultradiluted HF/ Nitrogen Jet Spray Cleaning for Particle Removal with Minimal Silicon and Oxide Loss, Electrochem. Solid State Lett. 62 (2006) 9. [22] N. Narayanswami, J. Heitzinger, J. Patrin, Development and Optimization of a Cryogenic Aerosol Based Wafer Cleaning System, in: K.L. Mittal (Ed.), Particles on Surfaces 5&6: Detection, Adhesion and Removal, VSP, Utrecht, The Netherlands, 1999, pp. 251 266. [23] C. Toscano, G. Ahmadi, Particle Removal Mechanism in Cryogenic Surface Cleaning, J. Adhesion 175 (2003) 79. [24] G. Vereecke, E. Rohr, M.M. Heyns, Laser Assisted Removal of Particles on Silicon Wafers, J. Appl. Phys. 3837 (1999) 85. [25] I. Varghese, C. Cetinkaya, Non Contact Removal of 60 nm Latex Particles from Silicon Wafers with Laser Induced Plasma, J. Adhesion Sci. Technol. 795 (2004) 18.
Particles in Semiconductor Processing
119
[26] D.M. Knotter, S. de Gendt, P.W. Mertens, M.M. Heyns, Silicon Surface Roughening Mecha nisms in Ammonia Hydrogen Peroxide Mixtures, J. Electrochem. Soc. 736 (2000) 147. [27] G. Briend, P. Besson, T. Salvetat, S. Petitdidier, Impact of Megasonic Activation with Different Chemistries on Silicon Surfaces in Single Wafer Tool, Solid State Phenom. 15 (2009) 145 146. [28] W. Kern, D.A. Puotinen, Cleaning Solutions Based on Hydrogen Peroxide for Use in Silicon Semiconductor Technology, RCA Review 187 (1970) 31. [29] M. Meuris, S. Verhaverbeke, P.W. Mertens, H.F. Schmidt, A.L.P. Rotondaro, M.M. Heyns, A. Philipossian, A New Cleaning Concept for Particle and Metal Removal on Si Surfaces, in: J. Ruzyllo, R.E. Novak (Eds.), Proc. Cleaning Technology in Semiconductor Device Manufacturing III, The Electrochemical Society, Pennington, NJ, 1994, pp. 15 25. PV 94 07. [30] A. Mayer, S. Schwartzman, Megasonic Cleaning: A New Cleaning and Drying System for Use in Semiconductor Processing, J. Electronic Mater. 855 (1979) 8. [31] G. Vereecke, E. Parton, F. Holsteyns, K. Xu, R. Vos, P.W. Mertens, M. Schmidt, T. Bauer, Evaluation of Megasonic Cleaning for Sub 90 nm Technologies, Solid State Phenom. 146 (2004) 14. [32] A. Otto, T. Nowak, R. Mettin, F. Holsteyns, A. Lippert, Characterization of a Cavitation Bubble Structure at 230 kHz: Bubble Population, Sonoluminescence and Cleaning Potential, Solid State Phenom. 11 (2009) 145 146. [33] A. Phillipp, W. Lauterborn, Cavitation Erosion by Single Laser Produced Bubbles, J. Fluid. Mech. 75 (1998) 361. [34] K. Xu, S. Pichler, K. Wostyn, G. Cado, C. Springer, G. Gale, M. Dalmer, P.W. Mertens, T. Bearda, E. Gaulhofer, D. Chen, Removal of Nano Particles by Aerosol Spray: Effect of Droplet Size and Velocity on Cleaning Performance, Solid State Phenom. 31 (2009) 145 146. [35] A.C. Engelsberg, Removal of Surface Contaminants by Irradiation from a High Energy Source, US Patent (1991) 5,024,968. [36] K.L. Mittal (Ed.), Particles on Surfaces 8: Detection, Adhesion and Removal, VSP/Brill, Leiden, The Netherlands, 2003. [37] K.L. Mittal (Ed.), Particles on Surfaces 9: Detection, Adhesion and Removal, VSP/Brill, Leiden, The Netherlands, 2006. [38] J. Visser, in: E. Matijevic (Ed.), Surface and Colloid Science, Adhesion of Colloidal Particles, Vol. 8, John Wiley, New York, 1976, pp. 3 84. [39] K. Bakhtari, A.O. Guldiken, A.A. Busnaina, J. Park, Experimental and Analytical Study of Submicrometer Particle Removal from Deep Trenches, J. Electrochem. Soc. 153 (2006) 9. [40] A.F.M. Leenaars, Methods of Removing Undesired Particles from a Surface of a Substrate, US Patent (1988) 4,781,764. [41] P.A. Kittle, Surface Treatment of Semiconductor Substrates, US Patent (2000) 6, 090, 217. [42] D.M. Knotter, R. te Brake, M. Enkelaar, In Situ Particle Removal Studies Using an Optical Particle Counter, Solid State Phenom. 189 (2008) 134. [43] D.M. Knotter, S.A.M. Wolters, M.A. Kasanroijat, Particle Concentration Measurements in Process Liquids Using Light Scattering Techniques, Particulate Sci. Technol. 435 (2007) 25. [44] W. Stober, A. Fink, E. Bohn, Controlled Growth of Monodisperse Silica Spheres in the Micron Size Range, J. Colloid Interface Sci. 62 (1968) 26. [45] A. Van Blaaderen, J. Van Geest, A. Vrij, Monodisperse Colloidal Silica Spheres from Tetraoxysilanes: Particle Formation and Growth Mechanism, J. Colloid Interface Sci. 2 (1992) 152. [46] F. Wali, D.M. Knotter, J.J. Kelly, F.G. Kuper, Deposition and Detection of Particles During Integrated Circuit Manufacturing, in: Proc. 9th SAFE Workshop Semiconductor Advances for Future Electronics and Sensors, Technology Foundation STW, The Netherlands, 2006, pp. 483 487. [47] F. Wali, D.M. Knotter, J.J. Kelly, Preparation of Mono Disperse Silica Particles with Metal Ion Tracer, paper presented at 11th International Symposium on Particles on Substrate: Detection, Adhesion and Removal, Orono, ME (2008).
120
Developments in Surface Contamination and Cleaning
[48] D.J. Shaw, Introduction to Colloid and Surface Chemistry, third ed. Butterworth, London, 1980, Chapter 6. [49] M. Tokuse, R. Oyama, H. Nakagawa, I. Nakabayashi, Characterization of the Oxidized Si3N4 Whisker Surface Layer Using XPS and TOF SIMS, Japan Soc. Anal. Sci. 281 (2001) 17. [50] R. Vos, I. Cornelissen, M. Meuris, P. Mertens, M. Heyns, Optimisation of dHF based Cleaning Recipes: To Remove or Not to Remove a Particle? in: T. Hattori, R.E. Novak, J. Ruzyllo (Eds.), Proc. Cleaning Technology in Semiconductor Device Manufacturing VI The Electrochemical Society, Pennington, NJ, 1999, pp. 461 468. PV 99 36. [51] A.A. Busnaina, J. Taylor, I. Kashkoush, Measurements of Adhesion and Removal Forces of Submicron Particles on Silicon Surfaces, J. Adhes. Sci. Technol. 441 (1993) 7. [52] D.M. Knotter, R. Roucou, R. Peyrin, Reduced Particle Removal Efficiency upon Wafer Storage, Solid State Phenom. 61 (2009) 145 146. [53] G. Vereecke, F. Holsteyns, S. Arnauts, S. Beckx, P. Jaenen, K. Kenis, M. Lismont, M. Lux, R. Vos, J. Snow, P.W. Mertens, Using Megasonics for Particle and Residue Removal in Single Wafer Cleaning, Solid State Phenom. 141 (2005) 103. [54] N. Narayanswami, P.A. Ruether, G. Thomes, J.F. Wygand, N. Lee, K.K. Christenson, J.W. Butterbaugh, Method for Evaluation and Optimization of Particle Removal Processes, Electrochem. Soc. Proc. 469 (1999) 99. [55] F. Holsteyns, Removal of Nanoparticles Contaminants from Semiconductor Substrates by Megasonic Cleaning, PhD Thesis, K.U. Leuven, Leuven, Belgium (2008).
Chapter 4
Continuous Contamination Monitoring Systems Roger Welker R.W. Welker Associates, Northridge, CA, USA
2.5. Case Study 5. Extended 1. Introduction Duration Manual Monitoring 1.1. Background 2.6. Case Study 6. Diagnosing 1.2. Justifying a Continuous Problems Using Continuous Monitoring System Monitoring 1.3. Traditional Airborne Particle 2.7. Case Study 7. Continuous Measurements Electrostatic Charge 1.4. Why 100% Sampling? Monitoring 1.5. Critical and Busy Sampling 2.8. Case Study 8. Continuous Air 1.6. Modified Data Collection Flow Monitoring Protocol 3. In Situ Particle Monitoring 1.7. Ongoing Use of Critical and of Cleaning Equipment Busy Sampling 3.1. Background 2. Case Studies of Traditional Versus 3.2. ISPM for Monitoring Critical and Busy Sampling for Cleaning of a Wide Variety Airborne Particle Counts of Parts 2.1. Case Study 1. Workstations 4. What to Continuously Monitor in in an ISO Class 7 Cleanroom High-Technology Manufacturing 2.2. Case Study 2. ISO Class 5 4.1. Air Quality Unidirectional Flow 4.2. Process Fluid Purity Benches in an ISO Class 7 4.3. Cleanliness of Surfaces and Cleanroom Electrostatic Charge 2.3. Case Study 3. ISO Class 5 4.4. A Discussion of Antennas for Vertical Laminar Flow Hoods Electrostatic Charge in an ISO Class 6 Room Monitoring 2.4. Case Study 4. Trend, Cyclic, References and Burst Patterns of Particle Generation
Developments in Surface Contamination and Cleaning Copyright Ó 2010 Elsevier Inc. All rights of reproduction in any form reserved.
121
122
Developments in Surface Contamination and Cleaning
1. INTRODUCTION The first fully functional continuous contamination monitoring system was developed for IBM by TSI Incorporated in the mid-1980s [1]. Even though the technical advantages of using a continuous monitoring system were obvious [2], the system did not initially gain widespread acceptance. Therefore, any discussion of continuous monitoring should begin by exploring the issue of how to justify it. The published literature on this subject, continuous monitoring of contamination, is sparse. As a consequence there are a limited number of references available to cite.
1.1. Background Traditionally, a variety of approaches have been taken to measure contamination or electrostatic charge in manufacturing areas. For example, it is universally recognized that a cleanroom must be positively pressurized with respect to the general factory environment in order to prevent the intrusion of contamination from uncontrolled adjacent factory areas. In most cleanrooms, a pressure gage or inclined tube manometer would be permanently installed on an outside wall. Once per day or once per shift the pressure would be read and the reading would be noted. In this way, the cleanroom would be audited on a sampling basis. Other examples can be cited. Rotating-vane or hot-wire anemometer measurements are taken near the face of HEPA filters to verify linear discharge velocities, verifying that the room air recirculation system was functioning correctly. This would involve taking a large number of measurements and so would be done infrequently. Often such a survey would only be taken as part of an annual room certification. Room air velocities at workstation level would be measured for a smaller number of sampling points, and often manual survey would be done less than once per week. In cleanrooms with unreliable recirculation fan systems, airflow problems might go undetected for weeks at a time. Surveys for electrostatic discharge (ESD) compliance are similarly troubled by inadequate data from manual sampling. Most ESD protected work areas are surveyed manually. These surveys are time-consuming and occur infrequently. In addition, the duration of data collection in each survey can be quite short, so only a snapshot of charge generation and electrostatic discharge can be obtained. As with contamination manual sampling, activity in the area being surveyed often changes during the survey, so the data can further be distorted. Some environmental conditions were recognized to be more critical than room pressure or air velocity, and would be checked at least once per shift (e.g. relative humidity) or as often as once per batch (e.g. starting pH of a bath). For critical contamination parameters, often a continuous monitoring system would be built into the process equipment or its dedicated environmental enclosure (e.g. temperature in a stepper). The continuous monitoring system could be
Continuous Contamination Monitoring Systems
123
easily justified, because a clear link between the process parameter and yield can be made. Airborne particle contamination has long been considered an important factor to measure and control. As a consequence, many manufacturing processes would require that airborne particle measurements be taken every day or every shift. However, the traditional methods used for sampling airborne particle contamination often produce erroneously low particle count results. In addition, the infrequency of particle count measurement makes it difficult or impossible to correlate with yield. These erroneous data are often used to justify minimizing the frequency of the manual survey and have been cited as evidence that automated continuous contamination monitoring is not justifiable. Prior discussions of continuous contamination monitoring systems have tended to focus on the data management software [3] or make the tacit assumption that a system will be bought [4], without a discussion of how to justify acquisition of a system to skeptical management. Occasionally, clever methods have been developed for reducing the cost per sample point of a continuous monitor [5], but still omitted discussing a method of how to justify the cost of continuous monitoring.
1.2. Justifying a Continuous Monitoring System [6] It is often difficult to justify the large capital investment necessary to install a continuous contamination or electrostatic charge monitoring system. An exception is in the pharmaceutical industry, where revisions to regulations in 2003 [7] and 2004 [8] resulted in manufacturers being recommended to continuously monitor where product is exposed. The difficulty in justifying the investment in a continuous monitoring system is often the result of faulty historical data, where sampling practices precluded obtaining a true characterization of the workplace. There is a fear that hasty installation of a system will result in placement of sensors in locations where they are not needed. Finally, there often are questions about the types of sensors that should be used, the required resolution, and other technical concerns that make decision-making difficult. In order to overcome these difficulties, a method is needed that will permit one to objectively determine where and what kind of continuous monitoring system is needed. Several examples illustrate the method as applied to particle sampling. The first step in this method is installation of sampling hardware on workstations that conforms to the requirements for critical and busy sampling. Data are collected to determine if the traditional sampling method has determined an accurate measure of the conditions at the workstation. Thereafter, sampling may continue using the manual optical particle counter, electrostatic charge monitor or other workstation monitor, with a modified sampling protocol to
124
Developments in Surface Contamination and Cleaning
collect comparative data. Sampling may also continue using the previous protocol to provide control data. Data collected with the new protocol are then compared with the historical database using the old protocol and the historical database. Generally this uncovers a number of sample points where the old protocol grossly underestimates the particle concentrations or static charge levels present. The new data are used to identify workstations that are out of compliance with contamination or electrostatic charge acceptance limits. An attempt can then be made to isolate and correct those items found to be contributing to the unacceptable conditions. Workstations that can be brought under control and maintained using a reasonable manual sampling frequency do not need continuous monitoring. Workstations that repeatedly show unacceptable conditions under manual sampling are candidates for continuous monitoring. The manually collected data are examined for evidence of burst, trend, and periodic contamination or electrostatic charge behavior. In addition, the results of modified manual sampling allow for the selection of contamination or electrostatic charge sensors with the optimum resolution, avoiding unnecessary costs associated with selecting sensors with unneeded resolution. The final examples will illustrate the evaluation of the need for continuous horizontal flow monitoring in a vertical unidirectional flow cleanroom and electrostatic charge monitoring.
1.3. Traditional Airborne Particle Measurements In the traditional approach to monitoring airborne particle contamination, an operator moves a conventional, self-contained optical particle counter to the workstation and places an isokinetic sample probe at some convenient location on the workstation, usually held in place by a stand. The conventional self-contained optical particle counter usually contains a vacuum pump, power supply, display, and often a printer. Inclusion of these features often results in a large and heavy particle counter. As a consequence, the conventional particle counters are mounted on a lab cart to facilitate moving about the cleanroom. This is conspicuous to production personnel. In addition, the isokinetic probe and its stand are frequently bulky and difficult to locate close to the product or process. As a consequence, the probe is often arbitrarily placed on the workstation in a location driven more by convenience than other considerations. Production personnel at workstations almost invariably stop all activity and move away when the particle count is to be taken. This results in the elimination of actions that may be generating contamination during normal production, thus lowering the particle count in the sample. More often than not, sampling via the conventional approach is only able to obtain contamination associated with the cleanroom or clean bench. This is often described as cleanroom idle sampling sampling in which the contribution of equipment
Continuous Contamination Monitoring Systems
125
and personnel is not included in the total. Contamination associated with materials handling, load/unload operations, personnel-generated contamination, and other related activities is seldom included in such sampling. Improper censoring of data is often observed. The particle count operator observes the rate of particle count. So long as the counts arrive at a relatively steady rate, the count is allowed to proceed. However, the particle count operator will almost invariably terminate the count if a sudden burst of particles is observed to occur, especially if the particle count operator can associate the burst with some undesired activity, such as someone walking by. Improper data censoring is repeated as often as is necessary until an acceptable result is obtained. Quite often the only count reported is one that is below the class limit for the area being sampled. These two factors, sampling during workstation idle periods and improper data censoring, result in a historical particle count database that makes the cleanroom and its workstations appear to be well in control with respect to particle count. Adverse actions are generally taken using such data. First, there is a tendency to reduce the manual particle sampling frequency to reduce the labor cost associated with particle count sampling. It is difficult to justify sampling more often when the data indicate the areas are in control from a particle count perspective. It is not uncommon to see the facilities monitoring points divided into two or four subgroups, cutting in half or quartering the sampling frequency. In extreme cases of very large assembly operations, this may be carried to the extreme that each sample location is visited only once per month. Second, given the apparent compliance with airborne particle limits, all appears to be in control, and switching from a manual sampling protocol to a capital-intensive continuous monitoring system simply cannot be justified. In order to correct the historical database and develop a more accurate description of the work area, a new sampling strategy must be developed. In the early stages of implementation, this strategy should be designed to minimize cost. The strategy must also deal with the two chief factors affecting the accuracy of the particle count: sampling the wrong place at the wrong time and improper data censorship.
1.4. Why 100% Sampling? Often, choosing the sensor for a continuous monitoring system particle counter is confusing. This is because there are so many factors to consider. The factor most often considered pivotal in the choice is the lower detection limit. However, factors other than the lower detection limit of the sensor should be considered when selecting a sensor. For example, one should also consider the sampling rate of the sensor, which defines both the volume sampled and the detection probability for contaminants or charges.
126
Developments in Surface Contamination and Cleaning
The detection probability for a particle counter serves as an illustrative example. For a particle counter, the lower detection limit is defined as the probability that a particle that is equal to or larger than the sensor’s lower size detection limit will be counted by the particle counter. This is a function of the fraction of the volume that the particle must pass through in order to be detected. Thus, if a particle counter samples 100 mL per minute but only can sense particles passing through 50% of this volume, the detection probability is expected to be 50%. This should be distinguished from the particle size accuracy of the particle counter [9]. We will illustrate this using a comparison between two popular sensors: one features 90% detection at 0.05 micrometer (mm) size and the other features 50% detection at 0.1 mm size. For both particle counters, the detection efficiency is essentially 100% at 0.2 mm and larger sizes. The 0.05 mm lower detection limit sensor has a sampling rate of 100 mL per minute, but can only detect particles passing through 0.25% of that volume. The detection probability is 1/400th of the volume sampled. Thus, the detection probability is 0.25% or 0.0025. The 0.1 mm lower detection limit sensor also has a sampling rate of 100 mL per minute and can detect particles passing through 100% of that volume. In detection probability terms this means the sensor sees 100% of the volume sampled. Thus the detection probability is 100% or 1.0. Now, let us see how these two factors, detection efficiency and detection probability, combine to affect the data reported by the two different sensors. In the first case, the 0.05 mm resolution sensor has 90% detection efficiency at 0.05 mm, but only 0.25% detection probability. Multiplying these two factors together gives the result that the probability of detecting a 0.05 mm diameter particle is 0.9 0.0025 ¼ 0.00225, or 0.225%, at the lower detection limit. In the second case, the 0.1 mm resolution sensor has 50% detection efficiency at 0.1 mm, but 100% detection probability. Multiplying these two factors together gives the result that the probability of detecting a 0.1 mm diameter particle is 0.5 1.0 ¼ 0.5, or 50%, at the lower detection limit. Clearly, the 0.1 mm resolution particle counter is more likely to detect particles at its lower detection limit than the 0.05 mm resolution particle counter. This problem becomes more obvious when we consider particles that are above the 100% detection efficiency limit. The 0.05 mm lower detection limit particle counter has 100% detection efficiency for 0.5 mm diameter particles, but a 0.25% detection probability. Multiplying these two factors together gives the result that the probability of detecting a 0.5 mm diameter particle is 1.0 0.0025 ¼ 0.0025, or 0.25%. The 0.1 mm resolution sensor has 100% detection efficiency at 0.5 mm, and a 100% detection probability. Multiplying these two factors together gives the result that the probability of detecting a 0.5 mm diameter particle is 1.0 1.0 ¼ 100%.
Continuous Contamination Monitoring Systems
127
1.5. Critical and Busy Sampling Here we introduce and define critical and busy sampling. Critical location: Location as close to the product or process as possible, without physically interfering with the movement of product, the people or the process equipment. Busy periods: Periods during actual manufacturing operations, especially when product is exposed. Critical and busy sampling: Sampling that satisfies the requirements of critical locations and busy periods. The critical location often places the inlet to the particle counter in a place where unidirectional airflow does not exist. This works to great advantage, since the bulky isokinetic probe can be eliminated, allowing greater freedom in placement of the inlet near the product. The tubing for the inlet to the particle counter should then be fixed to the workstation with brackets, tie wraps, and other means. This ensures repeatability of the sample location and protects against the tubing getting loose to interfere with the process. Hardware needed to implement critical and busy sampling costs only a few dollars per workstation and takes only minutes to install. The particle counter outlet end of the tube should then be terminated at some point on the workstation, which allows the particle count operator to attach the conventional particle counter to the sample tube without disturbing the process. This allows for sampling without stopping the process, referred to as busy period sampling. Figure 4.1 shows a photograph of a critical and busy sampling tube installed on a workstation.
FIGURE 4.1 A critical and busy sampling tube installation on a workstation. (see colour plate section at end for coloured version)
128
Developments in Surface Contamination and Cleaning
1.6. Modified Data Collection Protocol Once this low-cost, critical, and busy sampling hardware is in place, a new data collection protocol must be adopted. In the new protocol, data censoring is not allowed. The operator observes and records the activity at the workstation during each sample. If no product is in production and the workstation is unoccupied, the sample is labeled as taken during stage 1 operation, or a cleanroom idle sample. If product is being processed, but no production personnel are present, the sample is labeled as taken during stage 2 operation, or cleanroom and process tooling, but no personnel. This rule needs further discussion, covered below. If the sample point is at the load/unload and/or materials handling location on the workstation, table or cart, but the product is processed inside a tool or enclosure and no personnel are present, then the sample is labeled a stage 2 sample. If product, people, and tooling are present, the sample is labeled as stage 3 operation, fully operational and fully populated. Again, this rule requires further discussion. If the process is within a tool or enclosure that effectively prevents contamination or electrostatic charge generated by the operator from getting on the product, the sample is labeled a stage 2 sample. If the inlet to the particle counter is scraped, the tubing is bumped or otherwise disturbed to invalidate the count, the sample is recorded and the invalidating event annotated. Similarly, if charge sensors are disturbed in a manner that alters their readings in a way not representative of actual use conditions, these disturbances are noted. These occurrences indicate the need to correct the installation of the critical and busy sampling hardware to assure the highest quality data. By eliminating the option for data to be censored, we eliminate rejection of otherwise valid data. In addition, by labeling the stage of operation for each sample, it is possible to diagnose possible sources of the contamination or electrostatic charge. For example, if stage 1 particle counts are a significant fraction of stage 3 counts and the stage 3 counts are out-of-specification, the facility probably would be a fruitful place to begin searching for the source of contamination.
1.7. Ongoing Use of Critical and Busy Sampling When a sample location is identified as out-of-specification with respect to contamination or electrostatic charge, a second stage of investigation is initiated. For example, a stand-alone particle counter may be used like a Geiger counter, sniffing out the individual particle generation points. If these can be
129
Continuous Contamination Monitoring Systems
located, fixed, and kept under control by manually sampling at some tolerably low frequency, then a continuous monitor is not justifiable. However, the critical and busy sampling hardware and protocol should continue to be used. What if workstations are identified that require continuous monitoring? In this case, the continuous monitoring system is connected to the same critical and busy sampling system. Whenever an alarm is signaled, the manual sampling equipment is brought back to the location and is again used in the Geiger counter mode.
2. CASE STUDIES OF TRADITIONAL VERSUS CRITICAL AND BUSY SAMPLING FOR AIRBORNE PARTICLE COUNTS 2.1. Case Study 1. Workstations in an ISO Class 7 Cleanroom Table 4.1 shows the results of sampling two sets of data collected at workstations in an ISO Class 7 (FED-STD-209 Class 10,000) ballroom-style cleanroom [10,11]. Data listed are the average and standard deviation (SD) of particle concentration, in particle per cubic foot (ppcf), for particles 0.5 mm in diameter and larger. All workstations previously had been found to comply with the airborne particle count requirements of an ISO Class 7 cleanroom,
TABLE 4.1 Traditional Versus Critical and Busy Sampling in an ISO Class 7 Mixed-Flow Cleanroom, in ppcf 0.5 mm
Workstation #, line letter
Traditional sampling
Critical and busy sampling
Average
SD
Average
SD
1A
325
79
456
84
2A
458
85
531
139
3A
325
45
357
38
4A
452
250
694
242
5A
675
201
628
165
1B
236
125
288
159
2B
601
322
908
404
3B
266
64
254
52
4B
301
102
321
125
5B
425
211
623
364
130
Developments in Surface Contamination and Cleaning
using the traditional manual sampling protocol. The particle count increases slightly using the critical and busy sampling protocol, but not enough to change the conclusion that all workstations are in compliance with ISO Class 7. There are only slight differences between stations on line A versus line B. Results similar to those in Table 4.1 are often found in mixed-flow cleanrooms. The general contamination in the room is dominant over the contamination generated at the individual workstation. Data like these indicate a continuous monitoring system would not be necessary for these workstations. Plotted in Figure 4.2 are the data from Table 4.1, showing airborne particle concentration in ppcf of air for 0.5 mm diameter and larger particles. The tick marks on the vertical bars represent the average particle concentration. The upper and lower ends of the bars represent the mean plus or minus three standard deviations respectively. The bars are labeled along the X-axis to indicate the sample location number and traditional versus critical and busy sampling protocol.
2.2. Case Study 2. ISO Class 5 Unidirectional Flow Benches in an ISO Class 7 Cleanroom Table 4.2 shows comparative data for two identical sets of ISO Class 5 (FEDSTD-209 Class 100) workstations in the same cleanroom as case study 1. These ISO Class 5 workstations were located under vertical unidirectional flow units, effectively isolating each from the others.
Concentration, ppcf ≥ 0.5 µm
2500
2000
1500
Mean + 3s Mean 3s Mean
1000
500
FIGURE 4.2 cleanroom.
2B 2B C &B 3B 3 B C &B 4B 4B C &B 5B 5B C &B
1B 1B C &B
5A 5A C &B
2A 2 A C &B 3A 3 C A &B 4A 4 A C &B
1A 1A C &B
0
Workstation identification Traditional versus critical and busy sampling in an ISO Class 7 mixed flow
131
Continuous Contamination Monitoring Systems
TABLE 4.2 Average and Standard Deviation of Particle Counts in ISO Class 5 Unidirectional-Flow Workstations Located in an ISO Class 7 Ballroom, in ppcf 0.5 mm
Workstation #, line letter
Traditional sampling
Critical and busy sampling
Average
SD
Average
SD
6A
2
2
27
39
7A
14
4
180
114
8A
12
5
238
169
9A
6
5
292
151
10A
2
3
28
40
11A
6
4
52
31
12A
2
2
11
9
13A
1
2
10
55
14A
3
2
31
28
6B
3
2
258
200
7B
10
4
293
88
8B
8
4
153
116
9B
5
4
223
47
10B
5
2
36
17
11B
4
3
56
52
12B
2
2
19
9
13B
1
2
44
20
14B
3
2
10
6
None of the nine workstations from either line A or line B exceeds ISO Class 5 air cleanliness levels when sampled using the traditional approach. Conversely, the average and standard deviation of particle count increase for all 18 workstations when sampled using the critical and busy protocol. In seven of the 18 cases, critical and busy sampling shows workstations far dirtier than ISO Class 5 air cleanliness levels. Also of interest is a comparison of workstation number 6 on line A versus the identical counterpart on line B. The line B station is almost 10 times dirtier.
Developments in Surface Contamination and Cleaning
1000 900 800 700 600
Mean 3 sigma
500
Mean + 3 sigma
400
Mean
300 200 100 9B C&B
9B
9A
9A C&B
8B
8B C&B
8A
8A C&B
7B
7B C&B
7A
7A C&B
6B
6B C&B
6A
0 6A C&B
Concentration, ppcf ≥ 0.5 µm
132
Workstation identification FIGURE 4.3 Comparison of traditional versus critical and busy sampling protocols for measured contamination in ISO Class 5 clean flow benches.
Plotted in Figure 4.3 are some of the data from Table 4.2 to illustrate the differences in results obtained using traditional versus critical and busy sampling protocols. The Y-axis lists the airborne particle concentration in ppcf of air for 0.5 mm diameter and larger particles. In Figure 4.4, the particle concentrations are plotted on a logarithmic scale, unlike Figure 4.3, to accommodate the broad data range. The tick marks on the vertical bars represent the average particle concentration. The upper and lower ends of the bars represent the mean plus three standard deviations and mean minus three standard deviations respectively. The bars are labeled along the X-axis to indicate the sample location number and traditional versus critical and busy sampling protocol.
Log concentration, ppcf ≥ 0.5 µm
3.5 3 2.5 2
Mean 3 sigma
1.5
Mean + 3 sigma Mean
1 0.5
7B 7B C &B 8A 8A C &B 8B 8B C &B 9A 9A C &B 9B 9B C &B
7A 7A C &B
6B 6B C &B
6A 6A C &B
0
Workstation identification FIGURE 4.4 Comparison of traditional versus critical and busy sampling protocols on measured contamination in ISO Class 5 clean flow benches plotted using a logarithmic particle concentration axis.
Continuous Contamination Monitoring Systems
133
Case study 2 illustrates two common results of using critical and busy sampling in unidirectional flow work areas: 1. The emissions from the individual workstations are evident, because the mixing effects of the non-unidirectional flow cleanroom are eliminated. 2. Differences between pairs of otherwise identical workstations can be detected.
2.3. Case Study 3. ISO Class 5 Vertical Laminar Flow Hoods in an ISO Class 6 Room Case study 3 examines operations under ISO Class 5 vertical unidirectional flow units located in an ISO Class 6 (FED-STD-209 Class 1000) cleanroom. Here average values are shown, omitting standard deviations, due to the limited sample size of the survey. All 20 workstations sampled using the traditional approach easily meet ISO Class 5. The claim in this facility was that most of the workstations would meet or be better than ISO Class 4 (FED-STD-209 Class 10). The critical and busy samples indicate that most do not even satisfy ISO Class 5 requirements, as shown in Table 4.3. The worst case discrepancy is found in location 18, where ISO Class 6 was exceeded. Note that when using the critical and busy sampling approach, there are sufficient particles available to allow use of a 0.5 mm resolution, 0.1 cubic feet per minute (cfm) optical particle counter at nearly every workstation. If the data collected using the historical approach were used, the particle counter chosen would probably have been either a 0.3 or 0.1 mm resolution particle counter, greatly increasing the cost of continuous monitoring. The data in Table 4.3 may also be plotted. Figure 4.5 is a plot of traditional versus critical and busy sample averages for the fully automated workstations. These data illustrate an important feature of the critical and busy sampling approach. In order to sample using the traditional protocol, the operator must open the doors to the work cell. The safety interlock would stop the machinery inside, eliminating their contribution to contamination. The critical and busy sampling hardware was mounted so the operator could connect to the sample tube without having to open the enclosure. Thus, the machinery would continue operating, allowing its contribution to be detected. This is most dramatically illustrated by workstation 7. Figure 4.6 is a plot of traditional versus critical and busy sampling in the hybrid workstations, where an operator and automated tooling work together. Comparison of Figures 4.5 and 4.6 illustrates a fairly widely held belief in a way seldom so clearly demonstrated: people are a major contributor to contamination.
134
Developments in Surface Contamination and Cleaning
TABLE 4.3 Average Obtained by Traditional Versus Critical and Busy Sampling in ISO Class 5 Vertical Laminar Flow Units Installed in an ISO Class 6 Cleanroom, in ppcf 0.5 mm Location
Traditional average
Critical and busy average
Location type
1
5
198
Hybrid automation
2
9
77
Hybrid automation
3
5
31
Fully automated
4
5
292
Hybrid automation
5
1
507
Hybrid automation
6
5
326
Hybrid automation
7
5
977
Fully automation
8
33
36
Fully automated
9
11
36
Fully automated
10
7
489
11
12
12
Fully automated
12
5
70
Fully automated
13
10
407
Hybrid automation
14
10
155
Hybrid automation
15
27
499
Hybrid automation
16
12
254
Hybrid automation
17
3
78
Fully automated
18
14
1258
Fully automated
19
26
224
Hybrid automation
20
5
56
Hybrid automation
Hybrid automation
2.4. Case Study 4. Trend, Cyclic, and Burst Patterns of Particle Generation In addition to the average particle concentration prevailing at a workstation, there must be concern with trend, cyclic, and burst patterns of particle generation [12]. Sampled over a long duration, the average particle concentration may appear to be within control limits. Looking at the data in more detail may reveal unwanted particle concentration behaviors.
135
Continuous Contamination Monitoring Systems
Log concentration, ppcf ./+ 0.5 µm
3.5 3 2.5 2 1.5 1 0.5
18 18 C &B
12 12 C &B
11 11 C &B
9 C &B 9
7 C &B
8 C &B 8
3
7
3 C &B
0
Fully automated workstation number FIGURE 4.5 Comparison of critical and busy sampling versus traditional sampling protocols for fully automated workstations.
3 2.5 2 1.5 1 0.5
15 1 C 5 &B 16 1 C 6 &B 19 1 C 9 &B 20 2 C 0 &B
14 1 C 4 &B
10 1 C 0 &B 13 1 C 3 &B
C 6 &B 6
C 5 &B 5
C 4 &B
C 2 &B 2
4
C 1 &B
0 1
Log concentration, ppcf ./+ 0.5 µm
Upward trends in particle counts are considered undesirable because they may, at some future moment, exceed the control limits. Examples of upward trend are observed where workstations gradually become dirty between deep cleaning intervals. (Deep cleaning differs from routine cleaning, which generally occurs one or more times per work shift. In deep cleaning, areas of
Hybrid (partially automated) workstation identification FIGURE 4.6 Comparison of critical and busy sampling protocols for workstations containing automation and the presence of people.
136
Developments in Surface Contamination and Cleaning
the workstation not normally cleaned during routine cleaning receive increased attention. Routine cleaning generally is limited to areas of the workstation that contact or come in close proximity to the product and generally takes only a few minutes. Deep cleaning addresses the entire workstation and may require several hours.) Since the rate at which workstations become contaminated is not constant, it is seldom easy to predict when the next deep cleaning should be scheduled. This is an example where a continuous monitoring system may provide a useful benefit. Cyclic patterns of particle generation are a special case of burst pattern, where the bursts have a repeatable pattern. It is usually easy to associate these patterns with specific activities on the workstation. If such associations can be established, it is often easy to develop and implement fixes. Experience has shown that cyclic patterns of particle generation can usually be adequately controlled using manual monitoring and the critical and busy sampling hardware. Random bursts of contamination are observed in nearly every cleanroom. These can be associated with sudden, catastrophic events. A good example is shedding from an electric motor. The electric motor was mounted in a horizontal, unidirectional clean bench for study. This stepper motor was continuously monitored for over a week. The results are shown in Figure 4.7. The counts downwind of the motor started out in the 15 30 ppcf range, but cleaned up within a short time to 1 3 ppcf. Two large bursts are seen. Each sample is the average over 10 minutes of sampling, collected at 0.1 cfm. Averaged over the seven plus days, the electric motor produces only 16 ppcf. The second burst exceeded ISO Class 5 air cleanliness levels for 25 hours. With a once per week manual sampling plan, the chance of detecting this burst is only 1 in 7. The first burst, with a duration over ISO Class 5 air cleanliness
Concentration, ppcf >/= 0.5 µm
2500
2000
1500
1000
500
0 0
FIGURE 4.7
200
400
600
800
1000
Sample number Burst pattern of particle generation behavior from a stepper motor.
137
Continuous Contamination Monitoring Systems
levels for 5 hours, has only a 1 in 37 chance of being detected by sampling once per week.
2.5. Case Study 5. Extended Duration Manual Monitoring Ten different ISO Class 5 workstations were monitored using critical and busy sampling hardware [13]. These data were also compared to traditional monitoring results. Sampling was of sufficient duration that the percentage compliance could be calculated. Percentage compliance is the percentage of time that a workstation is monitored that it is below its particle count limit. High percentage compliance is considered to be good. Workstations with very low percentage compliance are highly likely to be detected in a traditional, once per week particle sampling protocol. This is illustrated in Figure 4.8.
2.6. Case Study 6. Diagnosing Problems Using Continuous Monitoring In this study, the manual optical particle counter was used to sample a workstation using the critical and busy sampling hardware for several hours. The data were collected once per minute in an ISO Class 5 vertical laminar flow (VLF) clean bench located within an ISO Class 7 ballroom. The particle count operator observed and recorded the activities at the workstation, but did not interfere with the actions of the production operators in any way. Table 4.4 is a partial summary of the particle concentration data, averaged to the nearest 5 ppcf 0.5 mm and the particle count operator’s notes of the activities in the workstation. The results are also plotted in Figure 4.9.
Log concentration, ppcf >/= 0.5 µm
3 2.5 2 1.5 1 0.5 0 A
B
C
D
E
F
G
H
I
% Compliance
100 90 80 70 60 50 40 30 20 10 0
3.5
J
Workstation Traditional
C&B
% Compliance
FIGURE 4.8 Comparison of critical and busy sampling versus traditional sampling. In this example, we see that percentage compliance, the percentage of time that a location complies with particle count requirements, must be viewed separately from the average particle count perfor mance of workstations.
138
Developments in Surface Contamination and Cleaning
TABLE 4.4 Extended Duration Manual Particle Monitoring Data, in ppcf 0.5 mm Time
Count
Observations
Time
Count
Observations
1150
2
Lunch break, room empty
1326
125
Wipe down
1152
1
Lunch break, room empty
1328
325
Wipe down
1154
6
Lunch break, room empty
1330
125
Operator adjusts fume extractor
1156
3
Lunch break, room empty
1332
80
Setup
1158
4
Lunch break, room empty
1334
65
Setup
1200
2
Lunch break, room empty
1336
150
Assemble
1202
0
Lunch break, room empty
1338
225
Assemble
1204
1
Lunch break, room empty
1340
555
Solder
1206
20
Operators returning
1342
985
Solder
1208
15
Operators returning
1344
65
Setup
1210
135
Operator wiping down
1346
25
Setup
1212
255
Operator wiping down
1348
125
Assemble
1214
600
Operator wiping down
1350
225
Assemble
1216
125
Operator wiping down
1352
750
Solder
1218
90
Setup
1354
625
Solder
1220
125
Setup
1356
25
Setup
1222
360
Assemble
1358
15
Setup
1224
425
Solder
1400
125
Assemble
1226
250
Solder
1402
155
Assemble
1228
35
Setup
1404
455
Solder
1230
50
Setup
1406
625
Solder
1232
100
Assemble
1408
25
Setup
1234
255
Assemble
1410
55
Setup
1236
325
Solder
1412
125
Assemble
1238
385
Solder
1414
250
Assemble
1240
100
Setup
1416
1250
Solder
139
Continuous Contamination Monitoring Systems
TABLE 4.4 Extended Duration Manual Particle Monitoring Data, in ppcf 0.5 mmdcont’d Time
Count
Observations
Time
Count
1242
Observations
35
Setup
1418
955
Solder
1244
60
Setup
1420
50
Setup
1246
125
Assemble
1422
35
Setup
1248
175
Assemble
1424
225
Assemble
1250
325
Solder
1426
175
Assemble
1252
475
Solder
1428
655
Solder
1254
100
Setup
1430
475
Solder
1256
65
Setup
1432
25
Waiting for WIP
1258
25
Setup
1434
15
Waiting for WIP
1300
175
Assemble
1436
25
Waiting for WIP
1302
225
Assemble
1438
35
Waiting for WIP
1304
375
Solder
1440
65
Setup
1306
400
Solder
1442
55
Setup
1308
100
Setup
1444
225
Assemble
1310
65
Setup
1446
350
Assemble
1312
35
Supervisor interrupts
1448
875
Solder
1314
90
Supervisor interrupts
1450
1120
Solder
1316
55
Supervisor interrupts
1452
25
Waiting for WIP
1318
35
Supervisor interrupts
1454
15
Waiting for WIP
1320
25
Supervisor interrupts
1456
30
Waiting for WIP
1322
45
Second operator replaces first
1458
25
Waiting for WIP
1324
75
Wipe down
1500
10
Waiting for WIP
The particle count operator’s notes provide a very clear understanding of what is happening at the workstation. Wipe-down is a relatively messy process, since it stirs up large amounts of contamination. Setup or waiting for work in progress (WIP) generate only little contamination. Assembly and especially soldering generate large quantities of airborne contamination. The arrangement
140
Concentration, ppcf >/= 0.5 µm
Developments in Surface Contamination and Cleaning
1400 1200 1000 800 600 400 200 0 12:00
12:28
12:57
13:26
13:55
14:24
14:52
Time HH:MM FIGURE 4.9 An example of insights provided by extended manual sampling using critical and busy sampling hardware. Between hour 13:12 and 13:32, a new operator arrived at the workstation and changed the location of the soldering station. The operator did not change the position of the solder fume extractor, causing an increase in particle count.
of the items on the workstation is not fixed. For convenience, the second operator moved the solder fixture and fume extraction system, with disastrous results. During soldering, the first operator averaged 370 ppcf 0.5 mm; the second operator averaged 746 ppcf 0.5 mm. This case study illustrates one example where a continuous monitoring system may be justifiable. Some flexibility in workstation layout must be provided to accommodate the reach and comfort of the operator. A continuous monitoring system might be a useful tool to keep particle counts under control after such rearrangements.
2.7. Case Study 7. Continuous Electrostatic Charge Monitoring Magnetoresistive (MR) heads are among the most electrostatic discharge (ESD) sensitive devices in existence. Modern static-protected work areas thus require many ESD protection tools to allow for safe manufacture. Among the more important tools provided for these static-protected work areas are air ionizers. The performance of air ionizers traditionally has been measured using charged plate monitors. During weekly audits the charged plate is used to measure discharge times and float voltages. In this procedure, the ESD technician places the sensor of the charged plate as close to the intended product location as possible. It is occasionally found that the air ionizer has drifted out of balance and needs service. Many times this consists of merely cleaning the emitter points on the air ionizer. Occasionally, simply cleaning the emitter points is inadequate and the ionizer must be manually balanced. One of the less well-understood features of air ionizer performance is that they interact with their environment. That is, grounded objects on the workstation below the ionizer tend to drain charge to ground. The polarity and
Continuous Contamination Monitoring Systems
141
amount of charge drained to ground is a function of the distance to and position of the grounded item below the emitter points on the ionizer. Relocating objects on the workstation thus can change the balance of the ionizer. This can occur frequently in a development site, where tooling and workstations are often changed due to changes in products or process flow.
2.8. Case Study 8. Continuous Air Flow Monitoring In this example, a very large cleanroom was equipped with 54 modular unidirectional flow units. Once per week a cleanroom technician would do a velocity survey, measuring the linear air velocity discharged from the filters in each module. Approximately every other week, at least one of the modules would be found to have very low or no air velocity. The problem this creates in a unidirectional flow cleanroom is unwanted horizontal airflow. Clearly, discovering this once per week is undesirable, but how would one design a costeffective continuous monitoring system to monitor for the condition. The answer lies in the design of the cleanroom, as shown in the plan view in Figure 4.10. The design of the cleanroom lent itself to definition of four airflow zones, labeled A, B, C and D in Figure 4.10. These zones were supplied with air from 10 to 16 unidirectional flow modules. It was immediately recognized that if airflow from any module changed, then the horizontal airflows through the restricted areas defined by the return plenums would change as well. In order to provide a module flow monitoring system, five hot-wire anemometers were installed in the restricted locations numbered 1 5 in Figure 4.10. Hot-wire anemometers are frequently used to measure airflow in cleanrooms. In this application, though, they were mounted to monitor horizontal flow, rather than vertical flow. After installation of the horizontal flow monitor, no imbalance condition went unnoticed for longer than a single shift. Of course, the flow monitor would
FIGURE 4.10 monitoring.
Plan view of a vertical unidirectional flow cleanroom with horizontal flow
142
Developments in Surface Contamination and Cleaning
not tell the cleanroom technician which module had failed. But the monitor would tell which intersection between zones was out of control. The technician would then go to the out-of-control intersection and determine the direction of the horizontal flow. This would then determine in which zone a module had failed, allowing the technician to quickly survey and identify the failed unit.
3. IN SITU PARTICLE MONITORING OF CLEANING EQUIPMENT The dominant cleaning process used in the cleaning of individual piece parts or subassemblies in precision assembly industries is aqueous cleaning. This process often involves initial cleaning by immersion in an ultrasonically agitated, deionized (DI) water detergent mixture, followed by rinsing in multiple, consecutive, ultrasonic tanks of increasingly pure DI water. Of increasing recent importance is also solvent immersion cleaning, an outgrowth of concern over the high energy cost associated with drying material after aqueous cleaning. This section will explore the feasibility of the use of an in situ particle monitor (ISPM) to monitor aqueous and solvent immersion cleaning. Some of the variables affecting particle counts are explored. Preliminary correlation with offline measurement of piece part cleanliness, using two different liquid-borne particle counters, is also explored. Possible management strategies enabled by the use of the particle counter as an online, real-time, ISPM are discussed.
3.1. Background Several approaches can be taken to monitor and control particle contamination on high-technology products that are subject to cleaning. Among these are periodic sampling of liquids from the cleaning baths and periodic measurement of parts using direct or indirect particle measurement techniques. Periodic measurements on parts are well supported by work reported by Nagarajan and Welker [14], Gouk [15], and Welker [16]. The approach of periodic sampling of parts from production has historically satisfied the needs of the user. Having a well-established parts measurement in place provides the historical and comparative database, which enables the implementation of in situ monitoring. Periodic sampling of the bath liquids or parts from the bath suffers from several drawbacks, among which are the following: Manual bath sampling may interrupt production and can result in bath contamination. Parts sampled from ongoing production may undergo recontamination in handling or packaging prior to analysis. Both bath sampling and parts sampling are periodic and may involve a delay in obtaining results.
Continuous Contamination Monitoring Systems
143
Both techniques require offline laboratory analysis, which may introduce procedural errors. These drawbacks result in possible loss of data integrity and are unable to capture batches of parts that do not conform to particle cleanliness requirements on a real-time basis. Knowledge of particle cleanliness for individual samples is excellent, but knowledge of the statistical cleanliness is poor, since sampling is infrequent. This makes it difficult to implement statistical process control. These difficulties were well articulated by Vargason [17], who also described a multipoint ISPM for semiconductor acid processing baths. Hess et al. [18] described the application of an ISPM for rapid optimization of a semiconductor cleaning bath. Later, Hess and colleagues [19] described the application of the ISPM to a second semiconductor cleaning system, where an attempt was made to show correlation with direct surface inspection using a wafer scanner. The results showed an apparent negative correlation between the counts in the bath reported by the ISPM and surface counts of particles on the wafers. The implications of this result are beyond the scope of this discussion. While these studies address many of the issues surrounding the application of ISPMs to monitoring process baths, all are focused on baths for processing semiconductor wafers. Knollenberg and Edwards [20] reported on the use of an ISPM for monitoring particles in a cleaner for head stack assemblies, an important assembly for a hard-disk drive. In this study the authors were able to show a strong positive correlation between particles in the final rinse overflow (i.e. sampling after the weir) and the residual particles ultrasonically extracted from the assembly and measured using an optical liquid-borne particle counter (LPC). They were also able to show several applications of the ISPM for optimizing the performance of the cleaner and for monitoring equipment failure modes. For example, one way to manage the cleaner operation using the ISPM could be to rinse to an endpoint. In this management approach, parts stay in the final rinse only as long as the LPC count is above a target level. This study was simplified by the fact that only one part type was cleaned, leaving bath load (number of parts) and arrival interval as the only major variables. In a more recent similar study, Welker reported on solvent-based cleaning, with particle counts being monitored continuously in the final solvent rinse tank [21]. This study produced a rather interesting solution to managing the cleaning process using an ISPM. Figure 4.11 shows the cumulative LPC count in the final solvent rinse tank over a 24-hour period. During periods when baskets of parts arrive slowly, the recirculation filters allow the particle counts to recover to low concentrations in the tank. Conversely, when baskets arrive in rapid succession, the recirculation filters are unable to remove contamination fast enough to prevent particle concentration from rising in the final rinse tank. The LPC count in the final rinse tank was monitored over a period of several weeks. These data were then sorted into categories, based on the number of
144
Developments in Surface Contamination and Cleaning
SOLVENT 12 X Bar Graph
6000
900 2 baskets in 2 hours falling LPC count
3 baskets in 2.5 hours falling LPC count
4 baskets in 0.8 hours rising LPC count
0
FIGURE 4.11
Particle concentration in the final solvent rinse tank over a 24 hour period.
baskets arriving per hour, and the average LPC count when the basket was withdrawn from the tank was calculated. These data are shown in Figure 4.12. This shows that the particle count at the end of each rinse cycle increases as the number of baskets per hour increases. The maximum number of baskets that can be processed in an hour was five. There is a relatively steady rate of increase in particle concentration with increase in basket arrival rate up to and including four baskets per hour. At five baskets per hour there is a significantly higher increase that would be expected, indicating that greater than four baskets per hour has exceeded some ‘‘tipping point’’. The amount of contamination remaining on the parts when withdrawn from the final rinse tank should be proportional to the LPC count in the final rinse tank at the end of the rinse cycle. The management strategy described above,
LPC (0.2 µm and larger particles)
6000
5000
4000
3000
2000
1000
0 1 b/hr
2 b/hr
3 b/hr
4 b/hr
5 b/hr
Arrival rate (number of baskets entered/hr) FIGURE 4.12 The relationship between basket arrival rate and LPC count in the final rinse at the end of the rinse period.
Continuous Contamination Monitoring Systems
145
rinsing to an endpoint, would require a variable amount of rinse time. This was considered to be unacceptable to the manufacturing department for fear that excessive rinse times would seriously decrease capacity. In addition, a variable cycle time was not favored for fear it would disrupt production and would require major reprogramming for the control system for these cleaners. Instead, a compromise solution was found. Machine capacity would not be adversely affected if the interval between basket starts was to be 15 minutes. This was a minor programming change, which resulted in a 37% reduction in LPC count in the final rinse, and could be handled within the capacity of the machines that needed to be reprogrammed. Both of these studies clearly demonstrated the utility of using an ISPM as a process control tool for operating a cleaner. However, the results are somewhat limited in their applicability because, in each case, a single part number was being cleaned. Could an ISPM be used to monitor the performance of a cleaner where several different types of parts, with variable incoming cleanliness levels, are being cleaned? Further complicating this problem is the fact that the arrival rate, sequence of baskets, and number of parts per basket were variables. This is considered an extreme challenge for the application of an ISPM, but one worthy of evaluation.
3.2. ISPM for Monitoring Cleaning of a Wide Variety of Parts 3.2.1. The Cleaning Process The cleaner consisted of five consecutive tanks. The first two tanks were prewash and wash tanks, and the final three were rinse tanks. The tanks were made of stainless steel, with approximately 80-liter capacity and equipped with immersible ultrasonic transducers. All power generators were 1000-watt units, operating at 95% of full power level, in the sweep frequency mode. The system was robotically loaded, although for certain tests described below, the cleaner was operated in the manual mode. Table 4.5 describes the important operating statistics for the system. (The system also included two forced hot air dryers, whose particle performance was not measured in this study.) The system was operated in the following fashion. When baskets of parts were in each tank, the recirculation system was turned off, but the ultrasonic power was on for a duration of 175 seconds. When baskets were entering or leaving the tanks, ultrasonic power was turned off, but the recirculation system was on, causing a relatively large volume of water to overflow the weir. If the arrival rate of baskets was continuous, the recirculation was for a total of 35 seconds. If parts arrived at longer intervals, the recirculation system would remain on for longer periods of time. All baths were continuously fed, independently, with approximately 4 liters per minute of fresh water. Approximately 4 liters per minute were drained continuously from each tank. The filtration system for recirculation and makeup water was via 5 and 0.2 mm filters.
146
Developments in Surface Contamination and Cleaning
TABLE 4.5 Description of the Cleaning Process for ISPM Monitoring Tank Parameter
1
2
3
4
5
Tank name
Prewash
Wash
Rinse 1
Rinse 2
Rinse 3
Weir
Single sided Single sided
Fluid
DI water
Four sided Four sided Four sided
DI water þ 0.02% DI water non ionic surfactant
DI water
DI water
Temperature ( C) 45 þ7/ 3
45 þ7/ 3
45 þ7/ 3 45 þ7/ 3 45 þ7/ 3
Ultrasonic frequency (kHz)
40
75
40
75
75
The operating sequence, turning on and off the ultrasonic power and recirculation pumps, had a noticeable effect on the particle count signature, as will be seen in the results shown below.
3.2.2. The In Situ Particle Monitor and Installation The ISPM consisted of a PMS model 900 CLS sampler equipped with a 0.3 mm resolution sensor. The output from the instrument was collected on a portable computer using the PMS Pharmacy View software, although PMS Facility View software could have been used with the same results. The in situ particle monitor sampler draws a sample through the sample burette and an overflowing reservoir. In this installation approximately 2 meters of 3 mm inside diameter TeflonÔ tubing connected the tank to the particle counter. The internal volume of the tubing was approximately 13 cm3. The total sample collected was approximately four times the volume contained within the sample tubing, assuring adequate sample flushing. Particle counts were reported in particles per cubic centimeter (ppcc) equal to or larger than 0.3, 0.5, 0.7, 1, 2, and 3 mm diameter. Samples were collected in three different locations: In the overflow weir from tank 4 (rinse tank 2) In tank 4 (rinse tank 2) In tank 5 (rinse tank 3). The desire was to obtain a comparison between the results measured in the weir for tank 4 and within tank 4, and also to compare results for sampling within tank 4 and within tank 5.
147
Continuous Contamination Monitoring Systems
3.2.3. The Particle Measurement Process The operation of the ISPM was not triggered by the arrival of baskets in any particular tank. That is, the cleaner and ISPM operated independently. As a consequence, in the sequence of events for the ISPM, when a basket of parts entered the tank being monitored was uncontrolled. This may affect the peak value for any basket load. The sample process began by drawing liquid from the sample tube through the sample burette and its attached overflow reservoir until liquid reached a preset limit. Pressure was then applied to force the sample liquid through the sensor to suppress bubble formation. At completion of the preset sample time, the liquid remaining in the sample burette and overflow reservoir was expelled to a drain. Pressure in the apparatus was vented to the atmosphere to prepare for the next sample. 3.2.4. Parts Cleaned During this Study Twelve different parts were cleaned during this study. These parts ranged from: large surface area, electrophoretically painted, cast and machined aluminum parts; bare aluminum parts; cast plastic parts; stainless steel parts; and parts consisting of a combination of stainless steel and elastomeric plastic. The arrival frequency of these parts varied according to their size and consumption. Three generic descriptions can be used to combine these 12 different parts into sets: bare aluminum parts, plastic parts, and others. Table 4.6 summarizes the basket count for these three groups of generic part types for the three sampling conditions: sampling in the weir from tank 4, sampling within tank 4, and sampling within tank 5. Clearly, the ‘‘Other’’ category represents the vast majority of baskets evaluated in this study. Within this category, one part number represents 111 of the 180 basket loads. This might permit some detailed description of the behavior of the particle counts in the cleaner. The type and number of parts entering the cleaner were recorded. This allows unique association of each peak recorded by the ISPM with the type of part being cleaned.
TABLE 4.6 Basket Count for Parts Groupings and Sampling Condition Sample location
Bare aluminum
Plastic
Other
7
6
29
In tank 4
11
13
98
In tank 5
7
9
53
In tank 4 weir
148
Developments in Surface Contamination and Cleaning
Cumulative ppcc
100000 10000 1000 100 10
7
4 17
:1
2 17
:0
9
:5 16
7
:3 16
:2
5 16
3
:1 16
1 16
:0
9
:5 15
7 15
:3
5
:2 15
:1
2 15
:0 15
14
:4
8
1
Time of day >/= 0.5 µm
>/= 0.3 µm
FIGURE 4.13
>/= 0.7 µm
Basket in Tank 4
Particle counts sampled from the overflow fluid sampling in tank 4.
3.2.5. Effect of ISPM Sample Inlet Location The study began with the sample inlet positioned in the liquid overflowing the sides of tank 4 into its weir (rinse tank 2). Figure 4.13 shows a representative plot of the particle counts recorded during a period of production in which parts were cleaned (data listed in Table 4.7). This chart shows the cumulative particle counts in three size ranges versus time of day. The horizontal bars above the particle count traces indicate the approximate times for entry into and exit out of tank 4 for 11 different baskets of parts. One feature is striking. Each basket of parts is apparently reported as two separate peaks, or as a peak preceded by a plateau. This can be explained by understanding the operation of the cleaner. As parts enter each tank, the
TABLE 4.7 Comparison of ISPM Results for Sampling in the Weir from Tank 4 or Sampling Within Tank 4, ppcc 0.3 mm
Location
Stainless Painted Large bare steel aluminum Bare stainless plus casting aluminum steel elastomer
Small bare stainless steel
Tank 4 weir
Mean
16,433
34,789
12,506
7483
26,641
SD
3831
7959
4535
3998
5182
Within tank 4
Mean
18,400
57,903
14,037
17,229
14,200
SD
6184
13,196
7722
9905
5986
Continuous Contamination Monitoring Systems
149
recirculation/filtration system is operating. Thus, a large volume of liquid from the tank is overflowing the weir to the inlet to the particle counter. Thirty-five seconds after basket entry, the recirculation pump is turned off, so the flow rate overflowing the weir is that of the makeup flow rate, about 4 liters per minute. At the end of the cycle, the ultrasonic agitation is turned off and the recirculation pumps are turned on again, increasing the flow rate from the tank over the weir. Thus, we see an initial peak in particle counts when baskets first enter the tank (high overflow rate), a lower particle count during ultrasonic cleaning (low weir overflow rate) and, finally, a high particle count after ultrasonic agitation is turned off (high weir overflow). The rate of arrival of particles into the sampling location is thus affected by the operating cycle of the cleaning equipment. Relocation of the sample inlet to inside tank 4 eliminates this double-peak effect, as shown in Figure 4.14. Horizontal bars indicate the periods in which baskets were in tank 4. A less ambiguous peak is recognizable. This allows a clearer interpretation of the particle count history during the cleaning process. A statistical comparison can be made of the location of the sample inlet outside the weir of tank 4 versus inside tank 4. These results show that sampling outside the weir significantly reduces the peak particle counts versus sampling within tank 4 for four out of five parts. The opposite result occurred for the small bare stainless steel parts: the particle count obtained in the weir is greater than that obtained in the tank. This result may be due to the small number of batches of these parts in the sample. Using Student’s t-test, only in the case of the bare aluminum and the small bare stainless steel parts are the comparisons statistically meaningful at the 95% confidence level. Table 4.8 summarizes the mean particle counts obtained for comparable baskets measured inside tank 4 and tank 5 for four parts with a high number of baskets. These results show a surprise. In two cases, the outcome is as anticipated: the average particle count in tank 4 is greater than the particle count in tank 5.
FIGURE 4.14
Sampling inside tank 4 eliminates the double peak effect.
150
Developments in Surface Contamination and Cleaning
TABLE 4.8 Comparison of ISPM Results Sampling Within Tank 4 or Sampling Within Tank 5, ppcc 0.3 mm Painted aluminum casting
Bare aluminum
Large bare stainless Stainless steel steel plus elastomer
Mean
18,400
57,908
14,037
17,229
SD
6184
19,186
7722
9905
Mean
19,280
77,055
13,366
13,553
SD
5198
9332
3479
9020
Location Tank 4
Tank 5
For the other two parts, the reverse is evident: the particle count in tank 5 is higher than in tank 4. This may be an indication that the painted aluminum castings and bare aluminum parts may require longer rinse times. Again, using Student’s t-test the difference between particle counts sampled within tank 4 and within tank 5 is statistically meaningful only for the bare aluminum parts.
3.2.6. Effect of Part Arrival Rate and Sequence Note that the arrival rate and sequence of baskets of parts is uncontrolled. As a consequence, a relatively clean part, such as the large bare stainless steel parts, may be preceded by a relatively dirty bare aluminum part or by a relatively clean plastic part. If this occurs over an interval too short for the rinse tank to recover to baseline, the starting value for each basket will be highly variable. This unpredictability of the baseline when a basket enters a rinse tank has a significant influence on the peak particle count achieved by any given basket of parts. One possible way to correct for this is to subtract the lowest value of the baseline preceding entry of a basket from the peak value for that basket, regardless of whether the baseline has fully recovered or not. For the purpose of this analysis, peak values sampled in tanks 4 and 5 are combined and compared to peak values corrected for the lowest value in the preceding 5 minutes of cleaner operation: the results are shown in Table 4.9. In every case, the mean value for a type of part was reduced by correcting for the preceding baseline. However, the amount of correction remains variable, as the type of part arriving in the preceding basket was not controlled. Hence, the starting point for the correction was variable. As a consequence, the coefficient of variability (CoV), defined as the standard deviation 100 divided by the mean, was not always improved for this method of baseline correction. This indicated that subtraction of the lowest value preceding a peak was not a viable way to correct for variable baseline in this study. To determine if baseline correction might eventually be of value, uncorrected peak values were compared to peak values that occurred only after the
151
Continuous Contamination Monitoring Systems
TABLE 4.9 Peak Value for Each Part Type Sampled in Tanks 4 and 5 Compared to Peak Value Corrected for Preceding Baseline, ppcc 0.3 mm Painted Large bare Stainless Small bare aluminum Bare stainless steel plus stainless casting aluminum steel elastomer steel
Condition Peak value
Peak corrected for baseline
Mean
18,695
64,665
13,778
15,595
14,211
SD
5857
18,607
6378
9434
5986
CoV (%)
31
29
46
60
42
Mean
13,519
64,174
8758
6526
12,294
SD
5942
19,227
4247
3036
3745
CoV (%)
44
30
49
47
30
cleaner had an opportunity to essentially fully recover the baseline, i.e. to less than 1000 ppcc 0.3 mm diameter particles. This comparison is shown in Table 4.10 for four types of parts. Correcting the peak value for a fully recovered baseline improved the coefficient of variability of all four types of parts. The comparisons were statistically meaningful using Student’s t-test for bare aluminum, large bare stainless steel, and stainless steel plus elastomer parts. In the management of the cleaner using the ISPM, allowing the cleaner to recover to a lower baseline
TABLE 4.10 Peak Value for Each Part Type in Tanks 4 and 5 Compared to Peak Value Corrected for Preceding, Fully Recovered Baseline (i.e. Less Than 1000 ppcc 0.3 mm) Painted aluminum casting
Bare aluminum
Large bare Stainless stainless steel plus steel elastomer
Mean
18,695
64,665
13,778
15,595
SD
5857
18,607
6378
9434
CoV (%)
31
29
46
60
Mean
18,636
73,171
11,035
6073
SD
5170
9881
4413
1073
CoV (%)
27
14
40
18
Condition Peak
Peak corrected
152
Developments in Surface Contamination and Cleaning
would be beneficial. This could be accomplished by allowing a longer delay time between baskets, increasing the recirculation flow rate to allow more rapid recovery or a combination of the two.
3.2.7. Effect of Basket Loading (Fill Level) In general, each basket contained one or more inserts of parts and each insert was completely full. In a few cases, more than one type of part insert would enter the cleaner as a basket load. By noting the number and type of insert in each basket load, it was possible to determine if basket loading has a meaningful effect on the uncorrected peak value for the basket. The results are summarized in Table 4.11. The large painted aluminum castings do not appear in this analysis, as only one insert was cleaned at a time. These data show that basket fill level had a significant effect on uncorrected peak value. In the management of the cleaner using the ISPM, basket fill level is clearly a variable that should be controlled. 3.2.8. Correlation with Parts Measurements For selected baskets, parts were sampled and taken to the offline contamination lab for LPC analysis. The purpose of this portion of the study was to determine if correlation could be established between particle counts in the cleaner and particle counts using the offline extraction and liquid-borne particle counters in the laboratory. The offline extractions were done using a Branson DHA 1000 ultrasonic tank. The tank was always filled with approximately 4 liters of clean DI water,
TABLE 4.11 Effect of Basket Load, as Measured by Number of Inserts, in ppcc 0.3 mm Bare aluminum
Large bare stainless steel
Stainless steel plus elastomer
Number of inserts
Mean
SD
Mean
SD
Mean
SD
1
28,359
8747
12,845
4354
5440
*
2
39,099
*
17,152
10,371
11,255
8946
3
70,071
8578
NA
NA
17,180
6242
4
74,382
*
NA
NA
NA
NA
5
68,563
13,308
NA
NA
NA
NA
6
79,744
13,013
NA
NA
NA
NA
*Only a single basket contained this number of inserts.
153
Continuous Contamination Monitoring Systems
to which was added a few drops of dishwashing detergent. Parts were placed in Pyrex glassware for extraction, containing filtered DI water and 0.02% of a surfactant solution. Parts were extracted for 1 minute, after which the LPC counts were taken immediately. The DHA 1000 ultrasonic tank is rated at 150 watts and is filled with approximately 4 liters (1 gallon) of water. By contrast, the in-line cleaner has ultrasonic generators set to output 950 watts into approximately 80 liters (20 gallons) of water. Thus, the energy density in the DHA 1000 is nominally 150 watts per gallon, while the cleaner ultrasonic tanks are nominally 48 watts per gallon. Particle removal in the laboratory extraction should thus be greater than in the in-line cleaner. Two different particle counters were available for LPC analysis. All parts were measured using a HIAC/Royco model 8000A particle counter, sampling with an ASAP sampler and measured using an HRLD-150 sensor. Cumulative particle counts for greater than 2, 3, 5, 9, 15, 25, and 50 mm sizes were recorded. In addition, the remaining extract from some parts were counted using a PMS mLPS particle counter through a CLS600 sampler and measured using an IMOLV 0.5 sensor. Cumulative particle counts greater than 0.5, 0.7, 1.0, 2, 3, 5, 10, 15, 25, and 50 mm were recorded. This allowed a three-way comparison of the LPCs, as the particle sizes for the three instruments overlapped as shown in Table 4.12.
TABLE 4.12 Common Size Ranges for the Three LPCs Used in this Study Size (mm)
PMS ISPM
0.3
Yes
0.5
Yes
Yes
0.7
Yes
Yes
1
Yes
Yes
2
Yes
Yes
Yes
3
Yes
Yes
Yes
Yes
Yes
5
PMS mLPS
9
HIAC/Royco 8000
Yes
10
Yes
15
Yes
Yes
25
Yes
Yes
50
Yes
Yes
154
Developments in Surface Contamination and Cleaning
As can be seen from Table 4.12, the ISPM overlapped with the mLPS in five size channels ranging from 0.5 to 3 mm and with the 8000A in two size channels, 2 and 3 mm. The mLPS overlapped with the 8000A in six size channels ranging from 2 to 50 mm. The ISPM used in this test only overlapped with the HIAC/Royco 8000 in two size channels. This is unfortunate, as the offline parts cleanliness measurements are taken using the HIAC/Royco instrument, rather than the PMS mLPS.
3.2.9. Bare Aluminum Parts Two batches of bare aluminum parts were sampled for laboratory analysis. The particle size distributions are listed in Tables 4.13 and 4.14 in ppcc. A basket load of six inserts of bare aluminum parts contained 1920 parts in 80 liters of water. This corresponds to 24 parts per liter. Conversely, the offline measurements used only 5 parts per liter of water for measurement. Thus, the particle burden in the cleaner should be nearly five times higher than in the offline laboratory measurement. This may partially compensate for the fact that the nominal ultrasonic energy density in the offline extraction was approximately three times greater than that in the cleaner. The results are first compared for the particle count data obtained for the ISPM in the cleaner versus the mLPS in the offline laboratory measurements.
TABLE 4.13 Bare Aluminum Part Cumulative Particle Size Distributions, in ppcc the Stated Size as Measured by the ISPM, by the PMS mLPS, and the HIAC Royco 8000A Size (mm)
ISPM
PMS mLPS
0.3
76,010
0.5
21,156
7752
0.7
5214
3716
1
1193
2030
2
58
128
483
3
16
21
279
5
3
66
9 HIAC or 10 mLPS
1
7
15
0
1
25
0
0
50
0
0
HIAC Royco 8000A
155
Continuous Contamination Monitoring Systems
The particle counts for the ISPM in the 0.3 and 0.5 mm size channels exceed the 12,500 ppcc count coincidence limit. The counts in these channels thus underestimate the true particle concentrations in the rinse tank. In addition, exceeding the count coincidence limit can result in counting several small particles as a single larger particle, which can then appear in a larger size channel. This may explain the observation that the ISPM reports greater number of particles in the 0.5 and 0.7 mm size channels, but fewer in the 1, 2, and 3 mm size channels than the mLPS reported in the offline size measurements for the bare aluminum parts. Particles may have been shifted to higher size channels at 0.5 and 0.7 mm, distorting the size distribution. If comparison was limited to particles larger than 1 mm, it would appear that the ISPM reported lower particle counts than the mLPS, but the disagreement was small (less than a factor of 3). Considering the differences between the particle counters and the parts per liter loading and power density between the cleaner and the laboratory for this comparison, this is a reasonable agreement. Turning to a comparison between the laboratory measurements by the mLPS and the HIAC/Royco, the mLPS counted less than the HIAC/Royco instrument by a factor of 4 to greater than 30 (Table 4.14). This was somewhat surprising, in that an earlier study by Nagarajan and Welker [22] indicated good correlation between the mLPS and the HIAC/Royco 8000A using data with the same lower detection limit. TABLE 4.14 Bare Aluminum Part Cumulative Particle Size Distributions, in ppcc the Stated Size as Measured by the ISPM, by the PMS mLPS, and the HIAC Royco 8000A Size (mm)
ISPM
PMS mLPS
0.3
65,217
0.5
18,066
11,819
0.7
4054
7472
1
878
4526
2
47
298
1014
3
4
41
603
5
5
147
9 HIAC or 10 mLPS
2
12
15
1
1
25
1
0
50
0
0
HIAC Royco 8000A
156
Developments in Surface Contamination and Cleaning
Finally, comparing the ISPM to the HIAC/Royco, the counts measured using the HIAC/Royco were much greater than the ISPM counts over comparable size channels, 20 and 150 times larger than the HIAC/Royco in the 2 and 3 mm size channels respectively. The particle concentrations measured in the cleaner by the ISPM exceeded the count coincidence limit for the 0.3 and 0.5 mm size channels. Thus, particle concentrations in the 0.3 and 0.5 mm size channels were probably underestimated and those in the 0.7 and 1 mm channels were probably overestimated. Again, limiting comparison to particles larger than 1 mm, the ISPM reported 5 10 times fewer particles than the mLPS. Comparing the mLPS to the HIAC/ Royco laboratory measurements, the HIAC/Royco reported significantly larger numbers of particles than the mLPS. Since these ISPM measurements were compromised by grossly exceeding the coincidence count limit, comparison of particle size distributions may not be the most reliable method for establishing correlation with laboratory measurement of parts cleanliness.
3.2.10. Large Bare Stainless Steel Parts Four batches of large bare stainless steel parts were sampled and compared using the three techniques, as shown in Table 4.15. The basket loading in the cleaner
TABLE 4.15 Large Bare Stainless Steel Part Cumulative Particle Size Distributions, in ppcc the Stated Size as Measured by the ISPM, by the PMS mLPS, and the HIAC Royco 8000A Size (mm)
ISPM
PMS mLPS
0.3
13,381
0.5
3849
5216
0.7
921
1889
1
172
881
2
15
74
196
3
5
22
129
5
7
50
9 HIAC or 10 mLPS
3
13
15
1
3
25
0
0
50
0
0
HIAC Royco 8000A
Continuous Contamination Monitoring Systems
157
equaled approximately 0.75 parts per liter; that in the laboratory measurement of parts cleanliness equaled 2 parts per liter. Thus the particle burden in the cleaner was expected to be 2.67 times lower than in the offline parts measurement. The overloading of the ISPM now only occurs in the 0.3 mm size channel. The comparison among the three particle counters therefore is expected to be more reliable using the stainless steel part data than the aluminum parts data. The ISPM counts fewer particles than the mLPC. However, if one multiplies the ISPM data by 2.67, the disagreement between the two particle counters diminishes. The agreement between the mLPS and the HIAC/Royco was still unsatisfactory.
3.2.11. Cleaner Management Using the ISPM Several objectives can be defined for use of the ISPM to assist in the management of cleaners. One of the most important of these is identification and isolation of batches of parts that do not meet statistical process control criteria for quality control requirements. Another objective is to collect data describing the cleanliness of parts in an automated fashion. In this study, parts were cleaned that had a wide variety of surface materials and thus a wide range of surface cleanliness. In addition, the arrival rate and composition of baskets of parts were uncontrolled. This led to difficulty in interpretation of peak contamination values in rinse baths. It also led to difficulty in correlation of particle counts in the rinse bath to offline laboratory measurements of parts cleanliness. Several approaches can be suggested to improve the viability of the application of the ISPM evaluated for application with this cleaner: The cleaner recovery should be improved to get closer to the baseline for full operation. That is, the recirculation time and flow rate should be increased to allow the baseline particle counts to more nearly approach 1000 ppcc 0.3 mm between baskets. The loading of baskets should be more closely controlled. The cleanliness contribution of inserts needs to be controlled, to prevent this from influencing the outcome of tests. The sequence of basket loading should consider the influence of the part type. For stainless steel screws, recovery times are very long. These should be cleaned last in a series of baskets to take advantage of longer recovery times due to longer cleaner idle time. In this application, the operating cycle of the cleaner has a measurable influence on the results obtainable from the ISPM. First, the time for the cleaner to recover to baseline particle counts is significantly longer than the arrival interval for baskets of parts under fully loaded conditions. The arrival sequence of baskets is unregulated so baskets of relatively dirty parts may immediately precede arrival of relatively clean parts. This, in combination with the long bath recovery time, interferes with interpretation of cleaner performance.
158
Developments in Surface Contamination and Cleaning
As discussed above, an effective method has been developed to permit the assessment of the need for continuous contamination monitoring. Its use has been demonstrated for sampling and measurement of airborne contamination, for monitoring voltage balance in a static-protected work area, and for monitoring the function of unidirectional flow modules. This method optimizes the placement of sample points to allow correct characterization of the workplace to be made. The data obtained allow for the selection of a particle counter or other sensors using a lowest cost strategy. In situ particle monitoring is a well-accepted technique in the semiconductor industry for vacuum processes. It has been demonstrated for wet process baths in applications in the semiconductor industry. It has also been shown to be feasible where the identity of parts and contamination load in the cleaner are relatively tightly controlled. The use of the ISPM for monitoring liquids in cleaners where a large variety of parts are being cleaned is a more difficult application. The range of cleanliness of parts, basket loading, and sequence of arrival of parts in the cleaner are more or less random. As a consequence, control of cleaner performance requires modification of the management strategy for parts entry into the cleaner in order for the ISPM to provide full benefit as a process control tool.
4. WHAT TO CONTINUOUSLY MONITOR IN HIGHTECHNOLOGY MANUFACTURING Continuous monitoring in high-technology manufacturing must consider many factors. The factors monitored are primarily dictated by the requirements of the product and the process by which they are produced, or may reflect prior experience with the operation of a particular cleanroom or ESD protected work area. Each manufacturer will monitor different factors and, in many cases, to unique control limits. Thus, a discussion of what parameters to monitor for manufacture must necessarily be general. Factors that affect product and process yield and reliability include: 1. The quality of air 2. Process fluids (such as DI water or compressed gases) 3. The cleanliness of surfaces and electrostatic charge. The reasons for monitoring include both contamination and ESD control, and concern for occupational health and safety and environmental health and safety.
4.1. Air Quality Many parameters affect the quality of the air in a cleanroom: Airborne particle concentrations Temperature and relative humidity Room pressurization
Continuous Contamination Monitoring Systems
159
Air velocity and direction Airborne molecular contamination Other factors.
4.1.1. Airborne Particle Concentrations There are two generic approaches to monitoring airborne particles. One uses a network of individual particle counters connected to a central monitoring system. This is often referred to as an electronically multiplexed particle monitoring system, sometimes also referred to as a real-time particle monitoring system. The second approach uses a single particle counter that sequentially samples a number of different locations in the room using a manifold of sample tubes. The latter approach is often referred to as a pneumatically multiplexed particle monitoring system, and sometimes called a sequential particle monitoring system. An electronically multiplexed system is illustrated in Figure 4.15. In an electronically multiplexed system every sampling location is populated with a dedicated individual particle counter. These miniature counters require external sources of power and vacuum and an external data collection system. This is done so the counter can be made as small as possible. All sampling points are monitored simultaneously and continuously. Particle losses are kept to a minimum, since they do not have to be transported through long vacuum lines. The advantages of this type of monitoring system are: Continuous detection of all events Good for critical or sensitive monitoring at lower detection limits greater than 0.1 mm
Central data collection system
Data communication system (direct wired, intranet, internet, RF, etc)
To external source of vacuum, either a vacuum pump or air ejector
FIGURE 4.15 A typical electronically multiplexed particle monitoring system. Four sampling points, each equipped with a dedicated particle counter, are illustrated here. Much larger systems can be built because of the flexibility of the data communication systems in existence today.
160
Developments in Surface Contamination and Cleaning
Good for equipment monitoring for failure and preventive maintenance (PM) at higher detection limits Immediate notification or alarming of yield destroying excursions It allows for emergency reaction to undesired events Immediate feedback to operators and technicians when procedures are not being followed Essential to observe during shift changes and evacuations where traffic may be higher Immediate ability to recertify an area after shutdown/evacuation. Particle counters with lower size resolution equal to 0.5, 0.3, 0.2, and 0.1 mm resolution, in 0.1 or 1.0 cfm sampling rates, are available. This wide range of particle counter size resolution and sampling rate means that counters at each sampling location can be chosen on the basis of a statistically meaningful number of particles to count, or chosen on the basis of the class of any given sampling location. The selection of particle lower size resolution and volume flow rate for the airborne optical particle counter is a function of the class of cleanroom. Following the statistical requirement of the consensus cleanliness standards [10,11], a minimum of 20 particles should be counted to obtain a 95% confidence level in the particle concentration. Thus, the resolution and flow rate of the particle counter are selected to ensure that 20 particles would be sampled if the room operates exactly at its class limit. Two examples serve to illustrate this principle: Example 1 Assume an ISO Class 5 cleanroom, with the average process time at each assembly operation of two minutes. For the fully operational cleanroom, we expect 100 particles per cubic foot of air, 0.5 mm in diameter and larger. Thus, we want to use a particle counter where we expect to sample 20 particles in 2 minutes. Twenty particles are expected to be found in two-tenths of a cubic foot of air. In this case, a 0.5 mm resolution, 0.1 cubic feet per minute particle counter is adequate. Example 2 Assume an ISO Class 4 mini-environment, with a process time of 4 minutes. For the fully operational state, we expect 10 ppcf, 0.5 mm in diameter, 30 ppcf, 0.3 mm in diameter and larger, and 75 ppcf, 0.2 mm in diameter and larger. For this case, a 0.5 mm resolution, 0.1 cfm particle counter would require 20 minutes to sample 20 particles and thus is inadequate. A 0.3 mm resolution, 0.1 cfm particle counter would sample 20 particles in 6.6 minutes and might be considered marginally adequate. A better choice would be to use either a 0.5 mm resolution, 1.0 cfm particle counter (20 particles in 2 minutes), or a 0.2 mm resolution, 0.1 cfm counter (20 particles in 2.66 minutes).
Continuous Contamination Monitoring Systems
161
A pneumatically multiplexed particle monitoring system uses a single optical particle counter. A sampling manifold is used to continuously sample from a large number of tubes. Inside the sampling manifold a sampling probe moves from tube to tube, capturing a sample of the particles from each and conveying them to the single particle counter. The tubes snake out through the facility and terminate at the desired sampling locations. The advantage of this type of system is reduced cost per sampling location over an electronically multiplexed monitoring system. The disadvantage is that all locations are seen sequentially. Short-term events might not be detected. The choice between the two different systems depends on the type of cleanroom (mixed flow versus unidirectional flow) and the probability that unpredictable bursts of excessive contamination will adversely affect the product or process. In a mixed-flow cleanroom, sample locations nearby one another tend to exhibit relatively uniform and constant particle concentrations. The gaps in time between each count are tolerable under such conditions. In addition, because airborne particle concentration tends to be relatively constant compared to airborne particle concentrations in unidirectional flow environments, sequential sampling particle counters are often recommended for mixed-flow cleanrooms. Conversely, in unidirectional-flow environments particle concentrations are highly dependent on the activity at each sample location and the contributions from nearby workstations are often completely undetectable. In addition, the activity in the individual workstation results in particle counts that are highly variable in time. Here, an electronically multiplexed continuous monitoring system is favored.
4.1.2. Temperature and Relative Humidity The temperature and relative humidity in a cleanroom are dictated by the process or, in the absence of process factors, by the need to provide a comfortable workplace for the operators. Cleanrooms are generally considered comfortable when the temperature is in the range from 20 to 22 C (approximately 68 72 F) and the relative humidity is in the range from approximately 35% to 55%. Temperature probes with accuracy of 0.2 C and relative humidity sensors with accuracy of 2% are generally considered adequate for monitoring cleanroom conditions that are dictated by comfort. Some processes require tighter control over temperature and humidity than others. An example is photolithography. Often, the photolithography tools are housed in a special enclosure, called a mini-environment, so that the tighter temperature and humidity controls need not be applied over the entire cleanroom. Sensors for these more highly controlled areas have correspondingly greater precision and accuracy than those used in the cleanroom in general.
162
Developments in Surface Contamination and Cleaning
4.1.3. Room Pressurization Cleanrooms are pressurized with respect to the ambient factory environment to ensure that contaminated air from the factory does not enter the cleanroom. Indeed, within a given cleanroom there are likely to be several areas where differential pressure may be monitored. For example, the change room will be at a slightly lower pressure than the cleanroom, but still positively pressurized with respect to the factory. In addition, service cores in the cleanroom should be negatively pressurized with respect to the process areas. Mini-environments are usually at a positive pressure with respect to the cleanroom. Equipment contamination enclosures are often at a negative pressure with respect to the cleanroom. A typical cleanroom will operate at between 25 and 50 Pa (0.05 0.1 inches of positive water pressure) with respect to the factory, service cores, and the interiors of evacuated enclosure. Mini-environments are often pressurized to as much as 100 Pa (0.2 inches of positive water pressure) with respect to the cleanroom. Mini-environments can be equipped with intelligent differential pressure controllers, which adjust fan speeds to maintain the positive pressure regardless of door positions. 4.1.4. Air Velocity and Direction Many cleanrooms can have multiple fan centers or may be designed to use fanfilter units. Occasionally it is found that undesired horizontal airflows exist, due to imbalance in the cleanroom air system. The flow imbalance can be caused by failure of individual fans, adjustment of fans, improper opening or closure of dampers, collapse of ductwork, and other facilities problems. The effect of these airflow problems can be monitored using hot-wire anemometers, mounted on the walls in openings that tend to amplify horizontal flows. A few strategically placed anemometers allow one to monitor the balance of even relatively large facilities. Typical horizontal flows in unidirectional flow cleanrooms are limited to no more than about 0.15 m s 1 (about 35 feet per minute). When the rooms go out of balance, it is not unusual to encounter horizontal flows of several hundred feet per minute. 4.1.5. Airborne Molecular Contamination Airborne molecular contamination (AMC) can roughly be categorized as inorganic acids and bases and organic compounds. The Semiconductor Equipment and Materials International (SEMI) classifies AMC as acids, bases, condensables, and dopants [23]. These compounds, if present at high enough concentrations, can severely corrode or chemically react with sensitive product surfaces. Types of failures attributed to AMC include t-topping of photoresist, unintentional doping, corrosion, adhesion failure of films, and uneven film growth in epitaxial deposition or oxidation [24,25]. Where AMC is a potential concern, its control must be anticipated before the facility and its contents are
Continuous Contamination Monitoring Systems
FIGURE 4.16
163
Electrode geometry in a quartz crystal microbalance (QCM).
constructed [26]. Fortunately, real-time and near-real-time monitoring systems are available that can monitor down to low part per billion levels. Even lower concentrations can be detected using sample concentration techniques, specialized detectors, and other methods, within 10 seconds to a few minutes [27,28]. Two types of continuous monitors used for monitoring airborne molecular contamination are the quartz crystal microbalance (QCM) [28,29] and the surface acoustic wave (SAW) sensor [30,31]. QCMs have been used since the early 1970s as particle detectors. If the QCM or SAW sensor is coated with an absorbent material it can be used as a gas detector. A typical design for a QCM is illustrated in Figure 4.16. Electrical charge is applied to electrodes on either side of the quartz crystal. The thickness of the quartz crystal changes in response to the applied electrical potential because of the piezoelectric effect. Typical quartz crystals will resonate at a few megahertz. In operation, a pair of well-matched crystals is used: one is exposed to a chemical-laden flow, while the other is kept clean. The frequencies of oscillation of the two crystals are compared, producing a difference signal oscillating at a few thousand hertz. The change in frequency is then a measure of the change in resonance of the exposed crystal imposed by the mass loading of the exposed crystal. The sensitivity of these quartz crystal microbalances is tens of micrograms per square centimeter. Coating the electrodes with selective absorbents allows for detection of specific classes of compound, such as water vapor. In a surface acoustic wave sensor, the electrodes are arranged on the same side of the crystal, as illustrated in Figure 4.17. The oscillations are across the
FIGURE 4.17
Electrode geometry in a surface acoustic wave (SAW) device.
164
Developments in Surface Contamination and Cleaning
surface as opposed to through the thickness of the crystal. Again, as mass is applied to the crystal, the resonant frequency changes. However, in the SAW device the resonant frequency is hundreds of megahertz. The SAW device has a correspondingly greater sensitivity than a QCM and is capable of typically measuring 0.2 ng cm 2. One of the more interesting features of these SAW devices is that absorbent coatings can be applied to their surfaces. The chemistry of the coatings can then impart some chemical selectivity to the sensor.
4.1.6. Other Factors in Air Quality There are other factors that can have an effect on air quality in the cleanroom. Among these are: Plenum pressures. Steady plenum pressures are important to air quality, indicating continued fan operation and ensuring that pressure surges do not damage filters and other vital components. Plenum pressure monitors can be used in place of or in combination with horizontal flow monitors to monitor room performance. Fan filter units. Many modern facilities are equipped with fan filter units. There may be hundreds or thousands of these units in mini-environments and the facility. Their continuing operation is essential to air quality. Several approaches to monitoring fan filter units are currently under development. Organic, acid, and general gas exhaust. Monitoring the pressure in the exhaust systems is a convenient and low-cost way to ensure the gases they are intended to remove from the process do not escape into the cleanroom to degrade air quality. In many cases, the materials being exhausted, if released into the air in the cleanroom, can pose potential health and safety risks. In some cases, outside air quality has a perceptible effect on air quality in the cleanroom. In these cases, weather stations can be installed outside the building to assist in management of the makeup air system.
4.2. Process Fluid Purity Process fluids used in high-technology manufacture include deionized water, compressed gases, detergents, solvents, and wipe-down chemicals. In the case of detergents and wipe-down chemicals, these are usually handled as bottled chemicals. The best way to control cleanliness of these types of chemicals is to periodically sample them using a qualified test lab and appropriate analysis methods to ensure batch purity. On the other hand, DI water, solvents, and compressed gases are usually distributed from a central bulk storage location using a delivery system. The purity of the fluid is thus not only source dependent: the purity is also delivery system dependent.
Continuous Contamination Monitoring Systems
165
Process purities of liquids and gases are dependent upon the performance of the delivery system. Flow variables are then essential to monitoring a process fluid delivery system:
Fluid-borne particle concentrations Flow rates Filter differential pressure Conductivity and specific ions Non-volatile residue Moisture content and dew point Total organic carbon Composition.
4.2.1. Fluid-borne Particle Concentrations There are two primary cases where process fluids can be monitored for fluidborne particles: in-line and in the process. In-line monitors are available for both DI water and organic solvent, and for compressed gases [32]. In-line particle analyzers for liquids generally depend on line pressure to force the liquid through the sensor, suppressing bubble formation. In-line particle analyzers for particles in compressed gases generally make use of pressure diffusers to lower the gas pressure to the operating pressure of the particle counter. Particle counter lower resolution limits range from 0.05 to 2 mm at flow rates capable of measuring tens to hundreds of particles per liter, adequate resolution for most applications. Grab samples can also be taken from within process tanks, drains and other chambers, where fluids are no longer at line pressure. In these cases, sampling must rely on sampling pumps to pass the samples through the particle counter. For particles in air, this reverts to standard airborne optical particle counters. However, for particles in liquids, syringe pumps or samplers that push the sample through the sensors under gas pressure are needed to avoid bubble formation. 4.2.2. Flow Rates For many processes, flow rate is a critical parameter. For example, flow rates through the recirculation filters and makeup DI water filters are critical to the operation of DI water cleaners. Often these are set using a rotating ball flow meter, which must be periodically visually checked, but that is otherwise not continuously monitored. In cases where visual checks prove inadequate, flow rate can be continuously monitored. 4.2.3. Filter Differential Pressure Pressure gages are often installed across the media of in-line filters. These are used to check for filter leakage (low differential pressure) and for filter clogging (high differential pressure). As with in-line flow meters, differential filtration
166
Developments in Surface Contamination and Cleaning
pressure gages are seldom continuously monitored. Where experience shows visual inspection is inadequate, continuous differential pressure monitoring is called for.
4.2.4. Conductivity and Specific Ions Deionized water, alcohol, and other fluids can become contaminated with ionizable contamination, which increases the conductivity of the fluid. In these cases, fluid conductivity meters or in-line ion-specific electrodes prove to be cost-effective continuous monitoring methods. 4.2.5. Non-Volatile Residue In some cases, DI water and other solvents can become contaminated by materials that do not ionize so they are not detectable using conductivity sensors or do not scatter sufficient light to be detected by liquid-borne particle counters. In many of these cases, an in-line non-volatile residue (NVR) monitor can provide the desired protection, where LPCs or conductivity monitors fail to detect the contaminants. The NVR monitor samples fluid under pressure and atomizes the fluid using a nebulizer. The volatile solvent is evaporated away, leaving condensation nuclei, which are too small to be seen with a conventional optical particle counter (OPC). The aerosol is first passed through a particle-conditioning chamber (an evaporation condensation apparatus) to grow the condensation nuclei to a size detectable by the OPC. This detector is capable of reporting NVR down to 1 part per billion in near real time (tens of seconds) [33 35]. Unfortunately, these sensors are quite expensive per sample point. A practical multiple-point sampling system is needed in order for the NVR monitor to provide an affordable solution for monitoring process fluids in high-volume manufacturing. Conversely, where NVR is measured using conventional laboratory methods [36] the NVR monitor offers the advantage of speed and can avoid problems with loss of high-volatility contaminants using conventional methods [37]. 4.2.6. Moisture Content and Dew Point Compressed air and other gases often must be monitored for condensable vapors, especially water vapor. The requirement for compressed air is a water vapor content so low that it requires a dew point analyzer. The dew point of a gas is the temperature at which liquid water condenses out of the compressed gas. A common dew point requirement is 40 C (40 F). 4.2.7. Total Organic Carbon Total organic carbon (TOC) is a method that can be used to monitor the purity of DI water for the presence of organic molecules that do not ionize, are not
Continuous Contamination Monitoring Systems
167
soluble, and present above their critical micelle concentration. If the organic contaminant does not ionize, it is undetectable by a conductivity monitor. If it is insoluble and below its critical micelle concentration, it will not form droplets and so will be undetectable by LPCs. TOC analyzers provide near-real-time, low part per billion sensitivity detection and monitoring for organic molecules.
4.2.8. Fluid Composition Often cleaning equipment is designed to automatically meter detergent into the fresh DI water put into the cleaning tanks. Metering pumps often perform unreliably. In-line UV absorbance detectors can be added to a monitoring system to make certain the detergent solution concentration entering the wash tanks are correct. Composition of other fluid streams can be monitored using ion-specific electrodes, electrochemical analyzers, and a variety of other sensors [38,39]. Perhaps the ultimate application of continuous monitoring is to use the data for feedback control [40]. 4.2.9. Ultrasonic and Megasonic Tank Performance Two processes take place to remove contamination in ultrasonic cleaning: cavitation and acoustic streaming. The primary cleaning mechanism at low frequencies (i.e. below about 150 MHz) is cavitation. Cavitation is the formation of tiny rarified bubbles in the cleaning fluid due to the constructive interference of ultrasonic pressure waves. When the cavitation bubbles collapse they form a tiny high-velocity jet of liquid that removes contamination from the surface. The contaminants then are carried away from the surface by acoustic streaming, the secondary and weaker cleaning mechanism at lower frequencies. Above about 150 MHz acoustic streaming becomes the dominant contamination removal mechanism. If the sound pressure level in the liquid in the tank is too low, no cavitation bubbles are formed and the cleaning efficiency of the tank is low. Conversely, if the sound pressure level in the liquid is too high, significant damage to the surface of the parts can occur, especially if the parts are made of erosionsensitive materials, such as soft aluminum alloys and many polymers. A sensor is available for monitoring the performance of ultrasonic baths [41,42].
4.3. Cleanliness of Surfaces and Electrostatic Charge Several items are of interest here:
Surface contamination rate increases due to electrostatic charge Workstation, tool, and operator grounding Mini-environments and standard machine interface enclosures In situ contamination monitoring in vacuum processing equipment Air ionizer status.
168
Developments in Surface Contamination and Cleaning
4.3.1. Surface Contamination Rate Increases due to Electrostatic Charge Surface contamination rates increase due to electrostatic charge. Surface contamination rates can be dominated by contact transfer. In contact transfer the contamination comes from the hands and gloves of people working in the area, and from contact with work surfaces and with packaging materials. Surface contamination rates can be indirectly monitored using several realtime airborne contamination monitors, such as optical particle counters and airborne molecular contamination monitors, as described previously. In addition, particle contamination rates to surfaces are strongly affected by the electrostatic charge levels on surfaces. Therefore, one way to indirectly monitor surface contamination rates is to monitor air ionization performance and surface charge levels, as well as airborne particle contamination monitoring. The literature on this topic is voluminous. Surface contamination rates vary in proportion to the average charge level on a surface for small particles. In general, the surface contamination rate increases in direct proportion to the average charge level on surfaces. At low surface charge levels, surface contamination rates by particles larger than about 3 mm are largely unaffected by surface contamination levels. However, as surface charge levels increase, the maximum particle size at which there is no effect on surface contamination rate shifts to larger particle sizes. Two factors are used to describe the charge levels on surfaces: the rate at which static electric charge is dissipated or shunted to ground and the average charge level on the surfaces. Three types of materials are commonly used in cleanrooms. These are: Conductive materials with surface resistivities less than 106 ohm per square Dissipative materials with surface resistivities in the range from 106 to 1012 ohms per square Insulative materials with surface resistivities greater than 1012 ohm per square. Conductors and static dissipative materials, if grounded, will very rapidly discharge to near zero volts. Insulative materials cannot be grounded and, in a cleanroom void of air ions, will only discharge very slowly. Conductive and dissipative materials will remain charged for a long time if they are not properly grounded. The charge levels on surfaces can be characterized using a field potential meter. If high charge levels are found, it may be necessary to reionize the air in the cleanroom to help neutralize these charges. This will require the use of air ionizers: this opens another avenue for continuous monitoring.
Continuous Contamination Monitoring Systems
169
The discharge time and float potential (average voltage) in a cleanroom equipped with air ionizers can be monitored with a charged plate monitor. Miniature charged plate sensors and low-cost monitoring systems are available to provide monitoring systems for cleanrooms and ESD protected work areas. In addition, these real-time monitors can be equipped with specialized sensors that allow monitoring of charge levels on tools, work surfaces, products, peoples, and other surfaces. Thus, one can monitor factors that directly affect surface contamination rates.
4.3.2. Workstation, Tool, and Operator Grounding In addition to monitoring charge levels, monitors are available to monitor the grounding status of many items in the factory. For example, workstation mats and floors can be monitored for grounding. The performance of the wrist strap grounding system may also be monitored by measuring the resistance of the skin of the operator. Several designs of air ionizers have self-balancing circuits. These have limited ability to compensate for emitter wear, contamination accumulation, and changes in workstation layout. When any of these factors drive the selfbalancing ionizer out of control, they can cause it to go into alarm. These alarms can be monitored continuously for most types of air ionizers. With the majority of self-balancing ionizers, the sensors used to detect ion output are built into the ionizer itself. There are also a few air ionizers on the market that use a workstation feedback system to adjust their performance. In these ionizers, a small sensor is placed in a critical and busy location. Conditions in the critical and busy location can change as the process continues. Robotic arms move, the capacitance of the product location can change as the assembly is built and operators move about. All of these can affect the balance of ions in the critical and busy location. The charge state on the sensor then is a more accurate measurement of the effect the ionizer is having on the product or process than a self-balancing ionizer with internal sensors can provide. Alpha air ionizers avoid the problems of emitter wear and contamination found with corona discharge air ionizers. 4.3.3. Mini-Environments and Standard Machine Interface Enclosures Mini-environments and standard machine interface (SMIF) enclosures are far more common in the integrated circuit (IC) factory than in precision assembly factories. Both are used to isolate the process equipment from the ambient cleanroom environment. This isolation can achieve several airborne contamination requirements that are more restrictive than can be economically achieved throughout a cleanroom, i.e. Class 1 around a specific tool,
170
Developments in Surface Contamination and Cleaning
versus the entire room surrounding the tool. These requirements can make the specification of continuous monitoring devices different from those that would be selected for an ambient cleanroom [43]. Among the most important of these are: Tighter control over airborne particle concentration. This can result in the need for selecting particle counters with significantly smaller size resolution and higher volume sampling rate. Tighter control over temperature and relative humidity. Enclosure pressurization. Generally the mini-environment or SMIF enclosure will be at a higher pressure than the ambient cleanroom. Periodically the doors to the mini-environment must be opened for maintenance or adjustment of the equipment inside. Fan speed can be increased during these intrusions to help maintain pressure differential with respect to the ambient cleanroom. The pressure sensor in the enclosure is very useful for this purpose. Airborne molecular contamination. Many times mini-environments and SMIF enclosures are used to eliminate exposure to AMC. These will be equipped with self-contained AMC filters. Two approaches to monitoring can be taken. The air within the enclosure can be continuously monitored or the air within the AMC filter can be monitored. AMC filters often consist of several layers of absorbent material separated by air gaps. The sampling point for the AMC monitor placed within the AMC filter will generally be between the next to last and last layers of absorbent. When the next to last layer of absorbent is saturated, AMC begins to be detected in the air downwind. 4.3.3.1. Cautions about Enclosures Early in the development of mini-environments and SMIF enclosures there was a widespread belief that continuous monitoring could be relaxed within them. This was the result of the widespread belief that the improved isolation with these enclosures made them less vulnerable to contamination than the ambient cleanroom. In fact, the high cost of enclosures was partially justified by the argument that cost of continuous monitoring can be avoided. The isolation created by the mini-environment can have the opposite effect. The enclosure can actually act as a contamination concentrator. In addition, the limited access to the enclosure for conventional contamination testing can allow this condition to go unnoticed for a long time.
4.3.4. In Situ Contamination Monitoring in Vacuum Processing Equipment Vacuum processing equipment is far more common in IC manufacturing than in precision assembly manufacturing. Access to the interior of the vacuum system
Continuous Contamination Monitoring Systems
171
has always presented a problem for contamination monitoring. As a consequence the IC industry adopted the witness plate sampling protocol. Early applications of the witness plate focused on tool qualification testing. These tests were usually done after tool maintenance or adjustment. Witness plates would be loaded into the process tool with no product wafers. The condition of the witness plate would then be used to verify that process settings were correct and that the tool was free of defects, especially particles. As products evolved, the control of the vacuum tools became more rigorous and product tolerance of defects became less. As a consequence, qualification runs using witness plates but no product were taking up an ever increasing fraction of run time. It was becoming obvious that a method for sampling contamination during regular production was needed in order to recover the lost production time in process tools. This was achieved in the early 1990s with the development of in situ particle monitors (ISPM) for vacuum processing equipment. Some of the earliest descriptions of these instruments focused on their generic application [44 48]. These applications primarily focused on monitoring vacuum lines for the load-lock chambers and process chambers. The range of processes for which the ISPM for vacuum equipment was developed quickly expanded and now includes tungsten chemical vapor deposition [49], ion implantation [50], etching [51], diffusion furnaces [52], and polysilicon low-pressure chemical vapor deposition [53]. One drawback that always limited the early ISPM was the limited sample volume and remote position (in the vacuum lines) of the early design. Recent developments place the sensor-detected volume in close proximity to the wafer surface and use a laser scanning technique to increase the sampled volume [54].
4.3.5. Air Ionizer Status Several designs of air ionizers have self-balancing circuits. These have limited ability to compensate for emitter wear, contamination accumulation, and changes in workstation layout. When any of these factors drives the selfbalancing ionizer out of control, it goes into alarm. These alarms can be monitored continuously.
4.4. A Discussion of Antennas for Electrostatic Charge Monitoring Several types of antennas are available for electrostatic charge monitoring. The choice of antenna is dependent upon the purpose of charge monitoring. There are three types of antennas and two ways of connecting them to the charge monitor. The three types of antennas are a disk-shaped antenna, a telescoping antenna, and a 10-picofarad charged plate antenna. Antennas can be connected to the charge monitor using a single- or double-banana jack, which places the impedance-matching resistor inside the charge monitor, or using
172
Developments in Surface Contamination and Cleaning
a microdot connector, which places the impedance-matching resistor in the antenna mount. The telescoping antenna is a good general area sensor. The disk antenna is intended for monitoring more localized charge, such as on a product. The 10-picofarad charged plate is intended specifically for monitoring air ionizer float voltage. Connecting the telescoping antenna or the disk antenna to the charge monitor using a single-banana plug allows the wire to act as part of the sensor, making it possible to monitor electrostatic charge over a relatively large area of the workstation. The double-banana plug is used to connect the charged plate to the monitor, thus providing a ground plane for the floating surface of the charged plate. Connecting to the microdot connector eliminates the ability of the wire to detect electrostatic charge, providing for highly localized charge detection. Many insulators are found in static safe work areas, in addition to the TeflonÔ used on the bases of the antennas. How does one determine if an insulator is suitable for use in a static safe work area? To answer this question, one must recognize that the majority of static safe work areas will be equipped with air ionization to control tribocharging of insulators and floating (i.e. ungrounded) conductors. These ionizers are installed and balanced to achieve discharge times and float potentials at critical product locations on the workstations. For example, MR critical workstations, in which the non-shunted MR head has head wires attached, but they are not yet bonded to the bond pads on the flexible circuit assembly, will typically require discharge from 1000 V to less than 20 V in under 10 seconds, with float potential limited to less than 20 V. If a surface becomes charged under normal conditions of use, it is considered acceptable if the ionization is capable of discharging it to below the float potential in an acceptable time. It is for this reason that TeflonÔ is considered to be acceptable for nearly all ESD protected work areas, including MR critical ones, when capable air ionizers are present.
REFERENCES [1] G.J. Sem, R.J. Ramierez, F.R. Quant, K.J. Weyrauch, P.P. Hairston, J.K. Agarwal, P.A. Nelson, New System for Continuous Clean Room Particle Monitoring, Proc. 31st Annual Technical Meeting of the Institute of Environmental Sciences (IES), 1985, p. 18. [2] G.J. Sem, A Case for Continuous Multipoint Particle Monitoring in Semiconductor Clean rooms, Proc. 32nd Annual Technical Meeting of IES, May 1986, p. 432. [3] D. Pariseau, The Future of Cleanroom Monitoring Systems, Cleanrooms Magazine (January 1995). [4] J. Livingston, R. Bower, R. Pochy, L. Branst, Using an Automated Cleanroom Monitoring System to Maximize Contamination Control, Microcontamination 15 (October 1997) 113. [5] B. Fardi, An Evaluation of a Cost Effective and Efficient Airborne Particle Monitoring System, Proc. 38th Annual Technical Meeting of IES, May 1992, p. 38. [6] R.W. Welker, Justifying a Continuous Contamination Monitoring System, Micro contamination 16 (1998) 51.
Continuous Contamination Monitoring Systems
173
[7] EC Guide to Good Manufacturing Practice, Revision to Annex 1, May 2003. [8] Sterile FDA, Drug Products Produced by Aseptic Processing Current Good Manufacturing Practice, US Food and Drug Agency (August 2004). [9] ISO 21501 4, Determination of Particle Size Distribution Single Particle Light Interaction Methods, Part 4: Light Scattering Airborne Particle Counter for Clean Spaces (2007). [10] ISO 14644 1, Cleanrooms and Associated Controlled Environments. Part 1: Classification of Air Cleanliness, International Organization for Standardization, Geneva, Switzerland, 1999. [11] FED STD 209, Federal Standard Airborne Particulate Cleanliness Classes in Cleanrooms and Clean Zones, General Services Administration, Washington, DC, 1992. [12] T.J. Bzik, Statistical Management and Analysis of Particle Count Data in Ultraclean Envi ronments, Proc. Microcontamination Conference, San Jose, CA (1985) pp. 93 118. [13] C.F. Query, Continuous Monitoring in Cleanrooms: A Guide for the First Time User, Proc. Asia Pacific Magnetic Recording Conference (July 2000) p. 91. [14] R. Nagarajan, R.W. Welker, Size Distributions of Particles Extracted from DiskDrive Parts, J. Inst. Environ. Sci. 36 (1993) 43. [15] R. Gouk, Optimizing Ultrasonic Cleaning for Disk Drive Components, Precision Cleaning Magazine (August 1997). [16] R.W. Welker, Size Distributions of Particles Extracted from Different Materials Compared with the MIL STD 1246 Particle Size Distribution, J. Inst. Environ. Sci. 43 (2000) 25. [17] R. Vargason, Liquid Multiport System Provides Automatic Real Time Monitoring of Wet Process Station Liquids, Microcontamination 8 (1990) 39. [18] D. Hess, S. Klem, J.M. Grobelny, Using In Situ Particle Monitoring to Optimize Cleaning Bath Performance, Microcontamination 14 (1996) 39. [19] D. Hess, K. Dillenbeck, P. Dryer, Comparison of Surface Monitoring and Liquid In Situ Particle Monitoring (for an HF and DI Water Rinse Hood), Proc. 43rd Annual Technical Meeting of IES (May 1997) pp. 321. [20] B. Knollenberg, K. Edwards, Use of In Situ Particle Monitors in HSA Aqueous Cleaning Processes, Proc. of the IDEMA Microcontamination Symposium (March 1998) p. 39. [21] R.W. Welker, Justifying a Continuous Monitoring System, Paper ID 58 in Proc. DST CON 2008 1st Int. Data Storage Conference, Bangkok, Thailand (in press). [22] Unpublished laboratory findings. [23] SEMI F21 1102, Classification of Airborne Molecular Contaminant Levels in Clean Envi ronments, Semiconductor Equipment and Materials International, San Jose, CA, November 2002. [24] M.J. Camenzind, H. Liang, J. Fucsko, M.K. Balazs, How Clean is Your Clean Room Air? Microcontamination 13 (1995) 49. [25] M.J. Camenzind, Airborne Molecular Contamination in Cleanrooms, Cleanrooms Magazine (January 1998). [26] S. Yellin, K. Kibbee, Nanotechnology Minimizing Airborne Molecular Contamination (AMC) Prior to Facility Construction, Controlled Environments Magazine (September 2005). [27] C. King, The Importance of SO2 Monitoring, Controlled Environments Magazine (October 2005). [28] S. Rowley, J. Yang, A. Wei, Reducing Capital and Labor Costs of 193nm Lithography Monitoring of Airborne Molecular Contamination (AMC) Through Proactive Assessment and Implementation of AMC Monitoring Techniques and Strategies, Lithography Asia 2008, SPIE 7140 (2008), 7140U1 7140U9. [29] R. Takasu, A. Akbar, Y. Takigawa, M. Miyajima, Y. Kataoka, New QCM Sensor for Real time AMC Detection in SMIF Pods, Proc. ISSM 2006 International Symposium on Semi conductor Manufacturing (September 2006) 421. [30] D.A. Hope, W.D. Bowers, Paper presented at Productronica 97, Measurement of Molecular Contamination in a Semiconductor Manufacturing Environment Using Surface Acoustic Wave Sensor, Munich, Germany, November 1997.
174
Developments in Surface Contamination and Cleaning
[31] B. Kanegsberg, E. Kanegsberg, Surface Acoustic Wave (SAW) Detectors for Real Time AMC Detection, Controlled Environments Magazine (November 2003). [32] O. Schmid, M.B. Trueblood, D.E. Hagen, P.D. Whitefield, Online Monitoring of Nano particle Contamination in Reactive Gases, J. Inst. Environ. Sci. 47 (2004) 111. [33] Y. Kousaka, T. Niida, Y. Tanaka, Y. Sato, H. Kano, N. Fukushima, H. Sato, Development of a New Continuous Monitor for Nonvolatile Solute in Ultrapure Water by Atomization, J. Inst. Environ. Sci. 30 (1987) 39. [34] D.B. Blackford, K.J. Belling, G. Sem, A New Method for Measuring Nonvolatile Residue for Ultrapure Solvents, J. Inst. Environ. Sci. 30 (1987) 43. [35] M. Xu, S.Y. Chang Chien, H.C. Wang, Measurements of Impurities in Liquids with a Nonvolatile Residue Monitor, J. Inst. Environ. Sci. 39 (1996) 21. [36] ASTME1235 01, Standard Test Method for Gravimetric Determination of Nonvolatile Residue (NVR) in Environmentally Controlled Areas for Spacecraft, ASTM International, West Conshohocken, PA (January 2002). [37] S.D. Hornung, D.D. David, A.M Randall, H.D. Beeson, Improved Detection Technique for Solvent Rinse Cleanliness Verification, NASA Investigative Report WSTF IR 0125 (April 2000). [38] J. Wei, J. Pillion, C. Hoang, In line Moisture Monitoring in Semiconductor Process Gases by a Reactive Metal Coated Quartz Crystal Microbalance, J. Inst. Environ. Sci. 40 (1997) 43. [39] L.W. Shive, K. Ruth, P. Schmidt, Using ICP MS for In Line Monitoring of Metallics in Silicon Wafer Cleaning Baths, Microcontamination 17 (February 1999) 27. [40] G.C. Frye, D.S. Blair, T.W. Schneider, C.D. Mowry, C.W. Colburn, R.P. Donovan, Devel opment and Evaluation of On Line Detection Techniques for Polar Organics in Ultrapure Water, J. Inst. Environ. Sci. 39 (1996) 30. [41] I. Kaskkoush, E. Brause, In Situ Chemical Concentration Control for Wafer Wet Cleaning, J. Inst. Environ. Sci. 41 (1998) 24. [42] Y. Wu, C. Franklin, M. Bran, B. Fraser, Acoustic Property Characterization of a Single Wafer Megasonic Cleaner, in Proc. Cleaning Technology in Semiconductor Device Manufacturing, in: J. Ruzyllo, R. Novak (Eds.), The Electrochemical Society, Pennington, NJ, October 1999, pp. 360 368. PV 99 36. [43] B. Belew, R.B. Lucero, S.D. Kochevar, S. Jorgensen, Continuous Particle Monitoring Inside Minienvironments Improves Wafer Yield but Requires New Monitoring Tactics, Cleanrooms Magazine (April 2004). [44] P.G. Borden, Monitoring Particles in Production Vacuum Process Equipment, Part 1: The Nature of Particle Generation, Microcontamination 8 (January 1990) 21. [45] P.G. Borden, Monitoring Particles in Production Vacuum Process Equipment, Part 2: Imple menting a Continuous Real Time Program, Microcontamination 8 (February 1990) 23. [46] P.G. Borden, Monitoring Particles in Production Vacuum Process Equipment: Data Collec tion, Analysis, and Subsequent Action for Process Optimization, Microcontamination 8 (March 1990) 47. [47] P.G. Borden, Monitoring Vacuum Process Equipment: In Situ Monitors Design and Specification, Microcontamination 9 (January 1991) 43. [48] P.G. Borden, Meeting the Challenges of Monitoring Particles in a Tungsten CVD System, Microcontamination 9 (March 1991) 39. [49] B. Fardi, B.S. MacGibbon, S. Tripathi, F. Moghadam, Feasibility of an In Situ Particle Monitor on a Tungsten LPCVD Reactor, J. Inst. Environ. Sci. 39 (1996) 25. [50] R. Burghard, D. Dance, R. Markle, Reducing Ion Implant Equipment Cost of Ownership Through In Situ Contamination Prevention and Reduction, Microcontamination 10 (September 1992) 27. [51] T. Scharnagl, Application of an HYT In Situ Particle Monitor for Selective Nitride Etching on a LAM 4420 (Rainbow), Proc. 42nd Annual Technical Meeting of IES (1996) pp. 317 322.
Continuous Contamination Monitoring Systems
175
[52] B.P. Bosch, D. Hess, C. Smith, et al., In Situ Particle Monitoring in a Poly LPCVD Diffusion Furnace, Proc. 43rd Annual Technical Meeting of IES (1997) pp. 315 320. [53] A.M. Haider, S. Paul, Implementation on an In Situ Particle Monitor on a LPCVD Polysilicon Furnace, Proc. 42nd Annual Technical Meeting of IES (1996) pp. 323 333. [54] J. Allinger, R. Burghard, In Situ Particle Detection for Pre Metal Sputter Etch, Semiconductor International (September 2005) p. 67.
Chapter 5
Strippable Coatings for Removal of Surface Contaminants Rajiv Kohli The Aerospace Corporation, NASA Johnson Space Center, Houston, TX, USA
1. Introduction 2. Coating Description 2.1. Coating Ingredients 2.2. Coating Properties 3. Types of Strippable Coatings 3.1. Solvent Based Coatings 3.2. Water Based Coatings 3.3. UV Curable Coatings 4. Issues with Strippable Coatings
5. Precision Cleaning Applications 5.1. Optical Surfaces 5.2. Other Applications 5.3. Non Optical Cleaning Applications 6. Summary Acknowledgements Disclaimer References
1. INTRODUCTION Strippable coatings have been used for more than 50 years for various applications for protection of surfaces from external contaminants. For these applications, the coatings have to be non-penetrable and resistant to water and other external contaminants. More recently, strippable coatings that are also peelable have been applied to remove microsize contaminants from surfaces such as glass and optics, as well as metals, ceramics, and polymers. For such cleaning applications, the coating must also mechanically or chemically trap the surface contaminants, and it must easily release from the surface after it has dried. These coatings are used to remove surface contaminants from high-quality parts and to protect the cleaned parts from surface damage. One successful application has been removal of dust and debris from precision optical surfaces, such as coated lenses and mirrors. The coating can be applied by simply pouring it over the surface to be cleaned. After drying (5 minutes to Developments in Surface Contamination and Cleaning Copyright Ó 2010 Elsevier Inc. All rights of reproduction in any form reserved.
177
178
Developments in Surface Contamination and Cleaning
4 hours at room temperature in air), the coating is removed by peeling. Contaminant particles (20 30 mm) on the surface can be completely removed. This is a low-cost, effective method for protecting and cleaning high-quality surfaces. However, care has to be taken to ensure that the coating itself does not leave any residue on the surface. The strippable coating can be effectively applied to all solid surfaces for particle removal, regardless of the roughness of the surface. As discussed in this chapter, the reference to strippable coatings will concern coatings that are physically removable after application by peeling or other mechanical means. Coatings that are chemically stripped will not be addressed. Also, the primary emphasis in this chapter will be on coatings for removal of contaminants rather than for temporary or permanent protection of surfaces. This is an important distinction because of the differences in the formulation of the coatings to meet the very different requirements of each of these functions, although some of these coatings serve both functions. We will also discuss the use of strippable coatings to remove both particles and films. In general, most strippable coatings are formulated to remove both types of contaminants. It is rare that a coating is formulated specifically to remove either films or particles but not both.
2. COATING DESCRIPTION Strippable coatings typically contain non-volatile components comprising about 41 71% and volatile components comprising about 29 59% by weight of the coating composition. The non-volatile components consist of a resin that is incorporated into a solvent or aqueous carrier, together with wetting agents and defoamers. The mixture is dispersed by means of additives to produce a stable polymeric emulsion or dispersion, which can be easily applied by spraying, brushing or rolling. The solvent and the water in the carrier evaporate upon application of the coating and leave behind a clear coating. A combination of release aids and plasticizers incorporated into the composition allows the coating to maintain its cohesiveness when it is subsequently peeled off.
2.1. Coating Ingredients The resin component of the coating composition includes a polymeric material that is compatible with the aqueous vehicle or carrier. Three general types of polymers are commonly employed in aqueous-based systems. These include soluble, semi-soluble, and latex polymers. The soluble polymers are characterized by clear solutions, whose viscosity depends on the molecular weight of the resins. They are generally made in solution and diluted with water to achieve the desired viscosity for proper application. The latexes are opaque suspensions of polymer particles, generally < 1 mm in size. The viscosity of the
Strippable Coatings for Removal of Surface Contaminants
179
latex is dependent on particle packing rather than on molecular weight and, as a consequence, a combination of high molecular weight and low viscosity at relatively high solids content can readily be obtained. The semi-soluble dispersions, also called colloidal dispersions, are translucent in appearance and are generally characterized by properties intermediate between the soluble components and the latexes. Polymers that fall within the above general description include poly(vinyl acetate), poly(vinyl formal), poly(vinyl butyral), poly(vinyl chloride), poly(vinyl chloride vinyl acetate) copolymers, poly(vinylidene) copolymers and poly(vinylidene chloride vinyl chloride) copolymers. The polymers can be produced as latexes by emulsion polymerization processes using known catalysts and chain transfer agents. The resin component is blended with dispersing aids and wetting aids and with plasticizers and coalescents in order to produce a stable polymeric emulsion. Surfactants of this general type include anionic and/or non-ionic and/ or cationic surfactants, depending on the particular resin selected for use in the aqueous carrier vehicle. Anionic surfactants include, for example, sodium oleate, potassium oleate and other metal salts of fatty acids, alkyl benzene sulfonates, sulfuric acid esters of higher alcohols, sodium alkyl sulfates and sulfonates. Non-ionic surfactants include polyoxyethylene alkylaryl ethers. Typical wetting agents include the nonylphenol polyethylene, phosphate ester, and the sodium salt of polymeric carboxylic acid. The coating composition often also include antifoaming agents that are compatible with the selected surfactants. The antifoaming agents include a wide variety of commercially available materials, including various dispersions such as reacted silica in mineral oil or silicone-based materials. Plasticizers and coalescents include, for example, glycerin, dibutyl phthalate, and 2,2,4-trimethyl-1,3-pentanediol monoisobutyrate. Release agents are added to the coating composition in an amount to insure easy removal of the dried coating from the substrate. Typical release agents include silicone polymers, oleic acid monoamide, fatty bisamides, polyethylene glycol monostearate, petrolatum, sodium alkyl benzene sulfonates, and synthetic waxes including polyethylene waxes. The addition of poly(vinyl alcohol) to the paraffin wax emulsion has proved to be particularly effective in forming the stable emulsion and subsequently acting as a release agent. The dried coating itself acts as a barrier to protect the release agents. Many coating formulations applied to metal test panels and dried at 120 C for 30 minutes form a film that is mechanically strippable even more than a year after application [1]. Various thickeners, such as hydroxyethylcellulose, sodium polyacrylate, poly(acrylic acid), polysaccharides, fumed silicas and clays, are often used to control the application properties of the resin component. A microbiocide such as 1,2-dibromo-2,4-dicyanobutane may also be employed with the aqueous emulsion polymer to extend the shelf life of the coating.
180
Developments in Surface Contamination and Cleaning
2.2. Coating Properties The desirable properties of a strippable coating are listed below:
Low-temperature curing Low viscosity for air-assisted spray application Excellent water and acid resistance at temperatures to 100 C No residue after removal from the applied surface No sag Excellent cohesion when pulled, it stretches and peels off in large sheets Retains its flexibility and does not become brittle with age Excellent adhesion to most surfaces including metals, ceramics, concrete, and other materials Nonflammable Meets United States Environmental Protection Agency (EPA) and Occupational Safety and Health Administration (OSHA) requirements for low solvent emission and safety in the workplace Easy applicability by spray, brush or roller Fast drying or curing Ability to lock in the surface contaminants.
3. TYPES OF STRIPPABLE COATINGS There are two primary types of strippable coatings in commercial use. These are either solvent-based coatings or water-based coatings. Due to concerns for the environment, the use of water-based coatings is finding increasing application. Other coating formulations have been developed for special applications.
3.1. Solvent-Based Coatings One of the most common solvent-based strippable coatings that has proven to be very effective in precision cleaning of optical components is collodion.
3.1.1. Collodion Collodion is a nitrocellulose, ethyl alcohol, and ether solution that has been around since 1846 [2,3]. It was first discovered by Louis Me´nard in 1846, who was then working in the laboratory of the French chemist Theophile Pelouze, the discoverer of cellulose nitrate [4 6]. Me´nard developed collodion while working on some aspects of smokeless gunpowder. This substance consisted of nitrated cellulose dissolved in a solvent. When the solvent evaporated, it left a tough, clear, transparent film. However, its immediate application was not apparent and Me´nard did not profit from its invention or subsequent
Strippable Coatings for Removal of Surface Contaminants
181
applications. It was not until 1848, when John Maynard created collodion as a liquid bandage for surgical application, which would harden on skin, seal the wound, and provide a protective covering, that collodion became a highly desirable commercial commodity [7]. Collodion is currently manufactured by several companies as an adhesive for EKG/EEG electrodes. Its structure and properties have been extensively investigated [8 21]. Subsequently, collodion also found application in photography. In 1851 the ‘‘wet collodion process’’ was introduced by Frederick Scott Archer [22 24]. In this process a glass plate was coated with wet collodion and then sensitized by dipping it into a bath of silver nitrate, then while still wet it was placed in the camera and an exposure was made. The plate was then immediately developed, while still wet, to form the negative. There have always been efforts to use the collodion process in a dry condition to make it easier to handle [25,26]. Most of the formulations developed for the ‘‘dry collodion process’’ were without success. Only a few processes became accepted for a short while. Dry collodion processes generally need two to six times longer exposure times than the wet process (this was improved by the introduction of alkaline developers). Dry plates have to be developed slowly; therefore, the adhesion between glass and emulsion is increased by an intermediate coating made of albumen or natural rubber. Collodion in which there is a great excess of ether forms a very tough film upon evaporation. The film left by collodion containing a large quantity of alcohol is soft and easily torn, but in hot climates the presence of an excess of alcohol is an advantage, as it prevents the rapid evaporation of the ether. Collodion has a wide range of uses in industry, including applications in the manufacture of photographic film, in fibers, in lacquers, and in engraving and lithography. Collodion is also widely used for manipulating and removing small particles (0.5 mm and larger) in preparation for microanalysis [27,28]. Celloidin is a pure type of pyroxylin used to embed specimens that will be examined under a microscope [29,30]. Collodion is widely used to glue electrodes to the head for electroencephalography. Pyroxylin with added pigments is used as a nitrocellulose lacquer. It is also added to nitroglycerin to stabilize it as blasting gelatin. Collodion is also used in theatrical makeup for various effects, such as simulating old-age wrinkles or scars. In medicine it is used as a drug solvent and a wound sealant, such as in boxing and other applications to cover up cuts. However, it has been generally replaced by liquid bandage, which includes pyroxylin as an ingredient. Collodion was also used by pathologist Thomas Stoltz Harvey to preserve Einstein’s brain in 1955. Collodion membranes have been used extensively in medical and biological applications [31], while composite membranes with a fluorocarbon polymer, such as Nafion, have also been used for desalination of water [32].
182
Developments in Surface Contamination and Cleaning
3.1.1.1. Types of Collodion From its first discovery, collodion has been available in many different grades and compositions [23,33 35]. Normal or plain collodion is a solution of pyroxylin in a mixture of ether and alcohol. For cleaning applications, only collodion USP should be used. This grade of collodion is a mixture of cellulose nitrate (3 7 wt %) in an ethyl ether (65 75 wt %) and ethanol (20 30 wt %) solution [36]. For photographic use, plain collodion is iodized or bromo-iodized to form sensitized, negative or positive collodion. Flexible collodion is a mixture of camphor plasticizer (~2% by weight), castor oil, and collodion, or a mixture of castor oil, Canada turpentine, and collodion. It is used for the same purposes as collodion, but its film possesses the advantage, for certain conditions, of not contracting. The addition of a small proportion of castor oil (~3% by weight) makes the resulting film elastic and more tenacious. Flexible collodion, which is used in medical applications, is unsuitable for cleaning applications since it leaves a residue on the surface. Other collodion formulations have been found to be suitable for medical applications, including aconite, carbolic belladonna, hemostatic, medicated, morphia, styptic, and vesicating or blistering collodion. Wood-tar and coal-tar collodion formulations, in which one part by weight of wood tar or coal tar is mixed with four parts by weight of collodion, have been used to treat dermatological conditions with limited success [37 39].
3.1.2. Other Solvent-Based Coatings Disccoat 4210 (General Chemical Corporation, Brighton, MI) is a clear, solvent-based, water-resistant coating for protection of optical media such as CD and video discs from scratching and marring during the manufacturing process [40]. It is stabilized against brittleness and is not softened or penetrated by most water-based compounds. It can be applied by spinning or dipping and it does not leave a residue. The coating is impregnated with transparent blue dye for easy visual inspection as well as identification, and is non-staining and stable to 100 C. Disccoat 4210 offers some real performance advantages, especially its ability to remove water stains and moisture-related defects. Moisture-related defects, as measured by loss of electrical signal read-back of nickel electroforms, are reduced by as much as 50%. This is significant since an optical nickel electroform can cost between $50 and $75 to produce. Carbicote 946 (Carbit Paint Company, Chicago, IL) [41] is a white solventbased booth coating that protects spray booth surfaces from the overspray of paints and coatings applied within the spray booth. The coating applies easily by spray, brush or roller, and forms a continuous film membrane, which acts as a barrier between the wet paint overspray and the clean metal spray booth surface. It is a temporary coating designed to be stripped during booth maintenance. The product dries quickly; after 15 minutes it can be handled and it is
Strippable Coatings for Removal of Surface Contaminants
183
hard in 30 minutes for a typical thickness of 0.05 mm. Drying times will be extended by high humidity, cold temperatures, and increased film thickness. The thickness can be built quite rapidly with minimum sagging and running. Typical coverage is 4 m2 per liter at 0.05 mm dry thickness. Chemco 791 (Chemco Manufacturing Company, Chicago, IL) booth coating is a solvent-based, fast-drying, removable coating [42]. It is available in white or clear transparent versions that are easy to apply by brushing, rolling or spraying. The transparent coating is suitable for windows, lights, and precoated spray booths.
3.2. Water-Based Coatings Several water-based coatings have been developed and are available commercially.
3.2.1. Opticlean Opticlean (Photonic Cleaning Technologies, Platteville, WI) is a molecular polymer cleaning system [43,44]. It was developed to clean and protect silicon wafers during manufacturing. The solvents used to hold the polymer in solution are acetone, ethanol, and ethyl acetate. As the polymer cures it shrinks and absorbs any contamination on the part surface into the polymer molecule. This includes particulates as well as fingerprints, grease, oil, and atmospheric pollution. When the cured film is peeled away, the polymer removes all the absorbed contaminants with it, leaving a molecularly clean surface (Figure 5.1). The polymer is hydrophilic and can be softened with the application of a touch of water. In many industrial applications the polymer is left on for long periods of time, sometimes for many months. The Opticlean cured film is rugged and slightly pliable, giving it good protection against fingerprints,
FIGURE 5.1 A protected aluminum mirror before and after cleaning with Opticlean. Courtesy of Photonic Cleaning Technologies, Platteville, WI. (see colour plate section at end for coloured version)
184
Developments in Surface Contamination and Cleaning
atmospheric contamination, and any other minor hazards. The film can be removed just as easily after 2 weeks as after 2 minutes. This type of coating is difficult to see as it forms a thin uniform layer. The ‘‘D’’ test shows the effect very well. By cleaning a ‘‘D’’-shaped section on one half of the lens, it is easy to see how much contamination has been removed from the surface by the coating. The high cost of Opticlean can be a deterrent to more common use of the coating for cleaning lenses. However, so little is used to clean camera lenses that one standard package can clean both the front and back elements of more than 12 typical standard lenses. This compares favorably to the cost of conventional cleaning methods. Furthermore, Opticlean is non-abrasive and leaves a perfectly clean surface, which is the best way to clean mirrors and lenses. Some plastic materials are soluble in one or more of these solvents. However, plastic spectacle lenses and most plastic camera lenses are normally impervious to the solvents. Lens mounts are not normally plastic and are unlikely to be damaged in any way by the polymer.
3.2.2. Coatings from Universal Photonics Several coatings are offered by Universal Photonics (Hicksville, NY) [45]. The Strippable Black Coating is a fully strippable material that leaves no residue or film whatsoever and will, in fact, pick up foreign particles that are removed when the coating is stripped off. It is well suited for cleaning mirrors. The Peelable Blue Coating removes easily with cellophane tape or rinsing in hot water and is suitable for plastic or glass lenses and mirrors. This product will remove dust and loose particles on the surface of the mirror, but it will not remove spots or stains caused by liquid or moisture on the surface. The Ultra Red Stripcoat is a transparent, brushable or sprayable coating that prevents marring of bright metal and optical surfaces from dirt and dust. It is soluble in toluene or methyl ethyl ketone. These coatings are available in liquid form or in aerosol spray cans and may be easily peeled off after use. They dry to the touch in 15 20 minutes and one can of the coating will cover approximately 0.19 m2 surface area or approximately 5.4 m2 per liter. 3.2.3. Chemco Strippable Coatings Chemco offers white and clear strippable coatings for ambient and higher temperature applications [42]. The high solids content of the bright white coating and its high tensile strength allow for maximum coverage and a good choice for safety-conscious spray booth applications. The high-temperature versions of the coatings will withstand 150 bake cycles to 260 C and still peel easily from any clean surface. The coatings can be applied to the widest range of surfaces.
Strippable Coatings for Removal of Surface Contaminants
185
In addition to strippable coatings, Chemco also offers Chem-Guard protective film for windows and lights. This highly durable, self-releasing film allows for easy application and removal.
3.2.4. CarbicoteÒ CarbicoteÒ is a water-based strippable coating that can be easily removed from many surfaces [46]. It is commonly used as safe peel-off protection for spray booth walls, lights, and windows. The coating is available as a reflective or a transparent coating (Figure 5.2). The coating offers several advantages. It is easy to apply by spray, brush or roller. It dries quickly, forming a removable barrier in minutes. The coating has excellent cohesion and peels off in large sheets. It adheres well to most surfaces including stainless steel and it does not become brittle with age. The coating is freeze thaw stable with a built-in safety factor that protects it from accidental freezing. 3.2.5. CorShieldÒ VpCIÔ CorShieldÒ VpCIÔ Strippable Coating (Cortec Corporation, St Paul, MN) is a water-based, nonflammable, environmentally friendly temporary coating that can be easily removed without the use of paint strippers or cleaners [47]. It does not leave residual contamination on stripping. The unique combination of water-based acrylic polymers, Vapor phase Corrosion Inhibitors (VpCIÔ) and a thixotropic thickener provides excellent barrier, surface, and corrosion protection. Designed for use as a fast-drying, temporary coating for parts and equipment, this product is resistant to sagging and running. It can be applied by spraying, brushing, rolling or dipping. The clear coating can be tinted in a variety of custom colors.
FIGURE 5.2 Reflective and transparent Carbicote strippable coatings. Courtesy of Carbit Paint Company, Chicago, IL.
186
Developments in Surface Contamination and Cleaning
Typical uses of the coating are protection from corrosion of clean metal surfaces for storage or shipment, protection of painted surfaces from physical damage during storage or shipping, and coverage of paint booths, tooling and highly polished metal surfaces. The coating can be used in one or two coats. Two coats should be applied on porous surfaces and for medium- to long-term outdoor protection. To achieve good strippability, the recommended thickness per coat should not be less than 0.05 mm.
3.2.6. Stripcoat TLC FreeÔ Stripcoat TLC FreeÔ (Bartlett Services, Inc., Plymouth, MA) is a nonhazardous, non-toxic solution designed as a simple, cost-effective method for safely removing and preventing the spread of radioactive contamination [48]. It can be used for nuclear reactor cavity, area or equipment decontamination, including glove boxes and hot cells. The coating can also be used as a barrier to prevent areas and equipment from becoming contaminated during maintenance activities, or as a covering to contain contamination. For decontamination purposes, the coating is applied and allowed to cure. Surface contaminants are mechanically and chemically entrapped in the coating. When the cured coating is removed, the surface contaminants are removed with the coating, yielding a solid waste product that is fully approved for disposal at low-level radioactive waste facilities or by incineration. Thirty square meters of coated surface generate approximately 0.03 m3 of uncompacted radioactive waste. Depending on the substrate, decontamination factors (DFs)1 of several hundred can be achieved with Stripcoat. 3.2.7. ALARA 1146Ô Strippable Coating ALARA 1146Ô Strippable Coating (Carboline Company, St Louis, MO) is a single-component, water-borne vinyl coating that contains no solvents or toxic materials [49]. It has a solids content of 41 2% by volume and volatile organic content of 14 g L 1. The coating migrates into microvoids on the surface to contact contaminants. It attracts and mechanically binds the surface contaminants into the polymer matrix. Drying times are approximately 18 hours for handling and foot traffic, and 24 hours for removal by peeling. These times are based on recommended dry film thickness per coat of 0.5 0.75 mm at 24 C and 75% relative humidity. Strippability is thickness dependent. Low film thicknesses make it difficult to remove the coating. Coating thicknesses above the recommended maximum result in slower drying times. Removal of the film decontaminates the surface and produces a solid waste. ALARA 1146 can achieve decontamination factors of 30 100 depending on the substrate. 1 DF is a measure of the effectiveness of a decontamination process. It is the ratio of the original radioactivity to the remaining activity after decontamination.
Strippable Coatings for Removal of Surface Contaminants
187
3.2.8. DeconGelÔ DeconGelÔ (Cellular Bioengineering, Inc., Honolulu, HI) is a one-component, water-based, broad application peelable decontamination hydrogel that lifts, binds, and encapsulates surface contaminants into the rehydratable matrix [50]. Safe and user-friendly DeconGelÔ can be used for radiological and nonradiological decontamination, including radioisotopes, particulates, and heavy metals, as well as water-soluble and insoluble organic compounds (including tritiated compounds). The product can be easily applied to a wide variety of horizontal, vertical, and inverted metallic and non-metallic surfaces (Figure 5.3). It is already in broad use at many industrial and nuclear sites. Further testing has shown the efficacy of variations of this gel against chemical and biological contamination including spore-forming organisms, raising the possibility of its use as a single chemical, biological, and nuclear decontamination agent. 3.2.9. InstaCote CC Wet and CC Strip Contamination Control Strippable (CC Strip) and CC Wet (InstaCote Inc, Erie, MI) are used in a dual-step process to decontaminate surfaces contaminated with plutonium and uranium [51,52]. Both materials are water-based, nontoxic, non-hazardous, and can be applied by spraying or brushing. The polyurea elastomer is prepared by mixing and reacting an isocyanate-terminated
FIGURE 5.3 Application of DeconGelÔ coating on flat surfaces. (a) One end of the contami nated area is taped. (b) The coating is applied to the tape and the contaminated area. (c, d) The coating is spread by a trowel held at 90 to the surface to cover the contaminated area. (e, f) Once the coating has dried, it is peeled off using the taped end. Courtesy of Cellular Bioengineering Inc., Honolulu, HI. (see colour plate section at end for col oured version)
188
Developments in Surface Contamination and Cleaning
compound or polymer with an amine compound containing at least two reactive amine groups. The isocyanate-terminated compounds or polymers have at least two terminal isocyanate groups and may be either aliphatic or aromatic in nature. The material sets in less than about 1 hour and is substantially unaffected by exposure to radiation. CC Wet is the first step in controlling removable contamination. CC Strip is then applied over CC Wet. CC Strip causes the CC Wet to be rehydrated and absorbed by the CC Strip. The contamination originally captured by the CC Wet is now removable with the CC Strip. CC Wet contains a UV-sensitive dye that allows feedback to the applicator during application. After CC Wet has cured, an ultraviolet light will insure verification of coverage to all surfaces. DFs as high as 100 can be achieved using CC Strip in plutonium glove boxes.
3.2.10. Capture Coating Encapsulation Technologies (Los Angeles, CA) have developed a patented process to eliminate airborne radioactivity and fix contamination in place remotely without the need for people or equipment to enter the process area [53]. The process employs a passive aerosol generator (PAG), which creates an aerosol by submerging parabolic-shaped piezoelectric ultrasonic transducers in the solution of the capture coating. Capture coatings consist of a water-based polyurethane suspended in a two part organic solution. They are formulated for various applications, from tacky, glycerin-based coatings that allow the application of a strippable coating for convenient disposal of the surface contaminants, to formulations with special adhesive promoters to enable the capture coating to more easily form on difficult hard surfaces. Using ultrasonic technology creates an aerosol that has the same chemical properties as the liquid coating. The droplet size (mean size ~2 mm) is such that the coating material assumes the properties of a gas, making it possible to slowly and evenly coat inaccessible areas and crevices. The aerosol droplets formed have the same chemical properties as the liquid. The droplet size of the aerosol is controlled by varying the frequency of the transducers. The aerosol or ‘‘fog’’ is introduced into the process area, where it condenses on surfaces. The small organic molecules begin to coalesce and encapsulate the contaminants in place. The small aerosol droplet size allows the capture coating to effectively ‘‘scrub’’ airborne contaminants from the air. 3.2.11. Isotron RadblockÔ Isotron RadblockÔ (Isotron Corporation, Seattle, WA) is a fixative system using a rubber-based coating that attracts and binds particulate contamination through adhesive bonding or ionic radionuclides through chemical bonding or ionic exchange [54 56]. It is applied by a variety of brush, roll or spray equipment. Upon curing, the coating mechanically and chemically traps contaminants and can be peeled to remove contamination in a solid waste form.
Strippable Coatings for Removal of Surface Contaminants
189
The polymeric membrane provides an impermeable barrier, which acts as a secondary sealant and prevents the spread, deposition or migration of contaminants, such as ionic forms of cesium that are known to be easily transported into building and soil subsurfaces. The coating provides theoretical coverage of approximately 3 m2 surface area and the recommended optimal coating thickness is 0.5 1.0 mm.
3.2.12. Argonne Supergel The Argonne Supergel, developed by Argonne National Laboratory, is superabsorbent gel with engineered nanoparticles designed to clean up radioactive materials from brick or concrete structures [57,58]. The porous structure of brick and concrete can trap radioactive contaminants. A wetting agent and the super-absorbent gel are applied on to the contaminated surface. When exposed to a wetting agent, the polymers start to cross-link, which allows the gel to absorb significant amounts of liquid. The wetting agent causes the bound radioactivity to resuspend in the pores. The super-absorbent polymer gel then draws the radioactivity out of the pores and fixes it in the engineered nanoparticles in the gel. The gel can be vacuumed and recycled, leaving only a small amount of radioactive waste. Laboratory testing has shown that a single application of the gel can remove better than 98% of the targeted radioactive element from the cement component of the concrete and more than 80% can be removed from concrete overall. 3.2.13. Poly(vinyl acetate) Coatings Poly(vinyl acetate) (PVA)-based strippable coatings have been developed for surface decontamination [59,60]. The PVA matrix coating serves to initially ‘‘fix’’ the contaminants in place for contaminant detection and ultimate removal. The PVA coating not only functions to mechanically entrap radioactive contaminants, but also contains one or more compounds that selectively entrap or bind certain radioactive isotopes or species to the PVA matrix. The coating can be applied to various surfaces such as stainless steel, mild steel, poly(vinyl chloride), and rubber. After use, the PVA coating containing one or more contaminants may be disposed of by dissolving the PVA coating and, if necessary, such as in the case of radioactive contaminants, by separating one or more contaminants from the dissolved PVA. The DF obtained from this process is primarily dependent on the type of the surface and also on the surface finish. 3.2.14. Poly(vinyl alcohol) Coatings and Pastes Several new water-based strippable coatings of poly(vinyl alcohol) and active additives have been investigated for decontamination of industrial equipment contaminated as a result of the Chernobyl accident and for general decontamination at nuclear power plants in eastern Europe and Russia [61 68].
190
Developments in Surface Contamination and Cleaning
The application of strippable coatings makes it possible to minimize the quantity of liquid radioactive waste. The proposed method offers the advantage that contaminated surfaces can be decontaminated in situ, with the contaminated waste generated only in compact and solid form. No washing or flushing of the decontaminated surfaces is required and hence there is no further propagation of radioactivity or contamination. Unique film-forming compositions consist of an aqueous polymer solution doped with plasticizers (water-soluble organic compounds with hydroxyl groups, such as glycerol, poly(ethylene glycol), ethylene glycol), mineral and organic acids, and oxidizers. Such compounds as H3PO4, H2SO4, HNO3, KOH, NaOH, organic acids and their salts, oxidants and EDTA, 1-hydroxyethylidene diphosphonic acid (HEDPA) and/or its salts were investigated as pickling or decontaminating agents. The advantage of organic acids is the absence of corrosion effects on the underlying metal surface. The combined action of organic acid and complexing agent increases the decontaminating ability as compared with each of these components separately. A reason for this is that the organic acids react with cations of the corrosion deposit and transfer them into solution. The complexing agent reacts with the cations by forming the complex compound and releasing anions of the acid for the next reaction. The film mixtures were applied to samples of stainless steel, carbon steel, Al Mg alloys, brass, and brick by brushing and spraying. The results showed decontamination factors of 10 40 were achieved with one treatment cycle. The decontamination ability of pastes was also investigated with the aim of optimizing decontamination compositions and technologies [61 68]. Decontamination pastes combine the advantages of strippable films and leachingdesorbing pastes. Decontaminating pastes are highly concentrated dispersed systems possessing structural properties. They can be applied on the surfaces by equipment used to dispense viscous substances and can be removed as a dry, hard mass. Application of decontaminating pastes reduces the consumption of chemicals and the volume of generated liquid radioactive wastes. The tests were performed with the aim of evaluating the sorption of Cs-137 and Sr-90 radionuclides by lignin, as well as clinoptilolite. Lignin is a natural polymer that is contained in timber (~30%). It is obtained as a by-product during the hydrolysis of wood polysaccharides. The lignin molecule is a large and complex polymer, which is very efficient in forming chelates with cations, even with monovalent ions like cesium. In an experiment to evaluate decontamination of rusted parts, rusty painted agriculture tools were decontaminated. The first decontamination cycle was performed in order to remove the rust and part of the contamination included in corrosion deposits. The first paste composition was oxalic acid, HEDPA, ammonium thiocarbamate, poly(vinyl alcohol), and lignin, which gave a DF of 2. The second cycle was aimed at removal of the paint and the contamination in the paint. The composition of the decontamination paste was NaOH, EDTA-Na, lignin, and poly(vinyl alcohol), which resulted in a DF of 4 7.
Strippable Coatings for Removal of Surface Contaminants
191
On oily or soil-contaminated metal surfaces, the DF could be as high as 30. The data showed new strippable compositions based on poly(vinyl alcohol) and active additives to be sufficiently effective for radioactive decontamination of equipment. Direct comparison of the total costs for the complete decontamination cycle (consumables, decontamination, waste volume and treatment, and waste transportation) for a 1 m2 metal surface showed that decontamination with these strippable films or pastes is nearly three times as cost-effective as decontamination by chemical solutions.
3.2.15. Polymeric Films New polymeric films have been developed based on interpolyelectrolyte complexes (IPECs) that contain glycerin as a softener and include silica nanoparticles (10 30 weight percent) as filler [69]. The film is applied in multiple layers (five to seven layers) and peeled off after drying for 12 24 hours. Tests on concrete, rusted metals, plastics, and other contaminated surfaces gave DFs between 60 and 280. 3.2.16. Coatings from Adhesives Research ARcleanÒ and ARclearÒ (Adhesives Research Inc., Glen Rock, PA) [70] are acid-free, electronically clean, low-outgassing acrylic adhesive products for controlling chemical contamination in the hard-disk drive industry and the assembly of touch screens, flat panel displays, and liquid crystal display (LCD) screens. They have low extractable ions and offer resistance to environmental aging. The products minimize corrosion and reduce fogging and potential oxidation of oxide surfaces or conductive circuitry in touch-screen devices. These products eliminate contamination and minimize labor costs associated with adhesive residue during rework. 3.2.17. Adhesive Tape An adhesive tape can be used to remove submicrometer particles from silicon wafers [71 74]. Typically, the tape has a carrier film (polypropylene) and an adhesive layer (acrylic-based adhesive), which is applied to the polished wafer surface. UV radiation is used for curing, which breaks down the adhesive layer. Since the cohesive strength of the adhesive layer is greater than the adhesion force between the contaminant particles and adhesive layer, the particles are removed when the tape is peeled off (Figure 5.4). One commercially available tape from Nitto Denko Corporation (Toyohashi, Aichi, Japan) contains a microcapsule foaming agent in the adhesive that is activated when a certain temperature is reached during UV curing [75]. The heat energy is converted to mechanical energy for adhesive release. The foam adhesive is nearly three times as effective in removing contaminant particles as is conventional tape that contains flat adhesive [76].
192
Developments in Surface Contamination and Cleaning
FIGURE 5.4 Process flow for removal of resist from a wafer surface using adhesive tape. (A) The initial lamination step in which the tape is applied to the wafer on a heated stage. (B) UV radiation exposure to cure the adhesive. (C) The delamination step in which the tape with the resist is delaminated and peeled off the wafer surface [74].
One disadvantage of conventional adhesive tape is that it leaves residual contamination on the wafer or other substrate when the tape is removed [75,77]. However, cleanroom-compatible tapes used for cleaning and protection of surfaces contain very low amounts of additives of low cohesive strength, which help reduce the compatibility of the adhesive and the substrate [75,76]. Any adhesive residue remaining on the wafer surface can be removed by rinsing the wafer in a solvent.
3.2.18. Coatings for Removal of Beryllium For solid beryllium remediation, the common preferred method is to wash, soak or rinse the contaminated area with a-aminobenzyl-a,a-diphosphoric acid (APMDP) and collecting the washes. This is ideal for machine tools or work surfaces, but not practical for larger environmental areas such as fields or roads. Sutton et al. [78] proposed a more appropriate method that uses a gel, foam or strippable coating that could be pulled away from the surface by hand, by mechanical means, or by vacuum. Similarly, for liquid contamination, solid material containing APMDP such as gel, foam or strippable coating would trap the beryllium contamination, preventing it from spreading further. The gel, foam, and/or coating can be mechanically removed. Another option is to incorporate APMDP into a solid support matrix (e.g. styrene beads, silica beads or even silica aerogel) and beryllium-contaminated liquids could be flowed over the surface to remove the beryllium. This latter example would be useful for removing liquid beryllium from liquid environments such as drinking water, sewer water or seawater. The cleaned water may then be recycled. 3.2.19. Coatings for Hazardous Material Cleaning A novel method and process has been proposed for immobilizing and decontaminating hazardous chemicals from a metal surface [79]. The method consists of applying a customized polymer film that is capable of being crosslinked to the contaminated surface that will take up the undesirable materials by
Strippable Coatings for Removal of Surface Contaminants
193
solution, absorption or adsorption and keep the contaminants in solid suspension with subsequent stripping of the polymeric material. Stripping the coating from the surface can be accomplished by physical means or by the use of a material that causes de-cross-linking.
3.2.20. Smart Coatings Polymeric smart coatings have been developed that are capable of both detecting and removing hazardous nuclear contaminants [80,81]. The coatings are commonly based on polymers and copolymers in a water base modified through the use of organic or inorganic additives as plasticizers, chelating agents, and indicators. A typical example of such a smart coating is SensorCoat developed by Los Alamos National Laboratory for the detection and removal of both uranium and plutonium from contaminated surfaces [82]. This coating consists of a blend of a low-viscosity, partially hydrolyzed poly(vinyl alcohol) and poly(vinyl pyrrolidone) in water. The coating also contains a glycerin plasticizer (4 12%), a chelating masking agent, and a colorimetric indicator. The coating exhibits color changes from orange to purple for uranium and orange to red for plutonium. Tests on uranium-contaminated surfaces (Al, Ni, stainless steel, and painted cement) showed DFs from 490 to 1540, illustrating the effectiveness of this coating. The coating was less effective on plutonium-contaminated stainless steel, where a DF of only 146 was achieved. 3.2.21. HaloShieldÒ Coatings HaloShieldÒ (HaloSource Inc., Bothell, WA) coatings are being developed as strippable barriers and reactive coatings for the US Department of Defense (DoD) and US Department of Homeland Security (DHS) [83 96]. These coatings are based on patented technology that binds chlorine molecules to virtually any textile, hard surface or paint, thereby extending chlorine’s efficacy for the life of the material (Figure 5.5). Research shows that 99.99% of odorcausing bacteria are killed in 30 seconds to an hour after they come into contact with a surface or textile. 3.2.22. Electrodecontamination System Veatch et al. [97] proposed a decontamination system that combines the advantages of strippable coatings and electrochemical stripping to remove both smearable and fixed radioactive contaminants. The method consists of applying a gel-like strippable coating to a contaminated surface of the object and passing an electric current through the applied gel, which will drive the contaminants into the coating material. The applied coating is cured and removed by peeling. The gel-like material includes an electrolytic agent, a latex formulation, and a chelating agent. This system offers several advantages. It can remove both fixed and smearable contaminants from large stationary surfaces, regardless of
194
Developments in Surface Contamination and Cleaning
FIGURE 5.5 Principle of HaloShieldÒ technology. (a) Untreated fiber. (b) The fiber surface is treated with HaloShieldÒ N halamine technology. (c) The treated fiber is washed in chlorine bleach. (d) The chlorine molecules are bound by the HaloShieldÒ coating and are anchored to the fiber. Courtesy of HaloSource Inc., Bothell, WA. (see colour plate section at end for coloured version)
orientation, by capturing the contaminants in the gel, and preventing uncontrolled transport of the radioactive materials to other locations. The cured gel contains the radioactive materials in a solid form that can be handled by existing radioactive waste-handling processes. The system is suitable for use in tight and confined spaces, such as under glove boxes, inside tanks and vessels, or in overhead ceiling spaces and pipe chases. This electrochemical strippable coating system is available commercially as the ElectroDeconÔ system with the electro-active strippable coating RedOxy PeelÔ [98].
3.2.23. Thick Film Etching Fluid Lierke et al. [99] have reported an effective process for removal of surface contaminants on metals such as copper and stainless steel, using a gel-like thick
Strippable Coatings for Removal of Surface Contaminants
195
etching fluid. The fluid consists of nitric and hydrofluoric acids and methyl hydroxyl ethyl cellulose (MHEC) that forms a gel. The gel is atomized by ultrasonic atomization and is deposited on the surface in layers up to a few millimeters. After processing, the gel can be removed by a scanning suction tube or it can be refluidized and skimmed off the surface. Removal rates as high as 40 mm per hour have been achieved for these metals.
3.2.24. Other Water-Based Coatings Several other aqueous-based coatings and systems for their application have been developed. Okano [100] developed a method of removing smears or stains from a coated surface with an aqueous dispersion of a highly absorbent cross-linked polymer. The gel-like surface layer absorbs the contaminants and can be stripped by a water spray. Muller et al. [101,102] reported the development of cosolvent-free aqueous, anionic polyurethane dispersions and their use as peelable coatings. The stripped coatings can be used as recycled material by mechanically comminuting the stripped coatings after cleaning, and then either pressing them in heated presses to form sheets, or extruding them in an extruder to form endless thermoplastic threads. The resulting threads are processed by known granulating methods to form cylindrical, spherical, lenticular or rhombic granules, which can serve as feedstock for known processes such as injection molding, blow molding, deep-drawing, slush molding or flat extrusion. Yamashita et al. [103] have described an aqueous dispersion of a peelable coating composition consisting of two acrylic polymers with different predetermined glass transition and a reactive surfactant. The presence of the surfactant increases the resistance to water of the coating, enhances film strength, and improves its peelability.
3.3. UV-Curable Coatings In UV curing, a reactive, low-viscosity, and usually solvent-free coating material is applied to the substrate and then polymerized by exposure to UV light [40]. The UV-curing mechanism involves chemical reactions of polymerization of difunctional and polyfunctional compounds in the formulation, resulting in cross-linking of the cured films. UV-curable peelable coatings are one-component systems that typically consist of: oligomeric acrylates, acrylic monomers, reactive diluents, photoinitiators, additives, and modifiers. The main types of acrylic oligomers used are: epoxy acrylates, polyester acrylates, and polyurethane acrylates. These oligomers provide the basic functional properties of the resulting coating. Photoinitiators are a very important component of coating composition, which initiate the polymerization (curing)
196
Developments in Surface Contamination and Cleaning
process when the coating composition is exposed to UV light at a certain wavelength. Different additives can be used to enhance coating properties such as wetting, surface leveling, flow rate, and color. Modifiers can increase durability (impact and crack resistance) of the UV peelable coatings. Through careful raw material selection, UV-curable coating formulators can more easily manipulate physical properties, such as chemical and weather resistance, as well as mechanical properties, such as tensile strength and elongation. Controlling these properties can lead to customized peelable coatings that meet specific performance criteria for each application. The properties of UV-cured peelable coatings are often superior to those of other systems. In addition to performance benefits, UV-curable strippable coatings provide several processing advantages over traditional solvent-borne or aqueous-based technologies: Since UV-curable coatings contain no solvents, volatile organic compound (VOC) emissions are in compliance with US EPA regulations. UV-curable, peelable coatings cure (air dry) much faster than solventborne/water-borne counterparts. In fact, curing takes only a few seconds, which makes UV-curable coatings particularly economical by speeding application (masking) times and increasing production cycles (profitability). Superior film properties (thermosetting, mechanical, and chemical properties) can be achieved. Good cost/performance ratios. Worker and environmentally friendly no solvents and little to no VOC emission. Minimal capital investment. Space-saving installations. Coating dispensing can be automated. The coating may be applied by dipping, spraying, screen printing or pad printing. Only very little heat is applied to the substrate. Curing is not dependent on temperature or humidity conditions of the environment. Probably the most widely used UV-cured protective coating is the peelable-type coating. UV curing, which cures the film in a few seconds, provides durable and cross-linked films with sufficient adhesion for surface protection of objects during handling, storage, transportation, and in manufacturing operations such as machining, acid stripping, and solvent cleaning. UV-curable peelable coatings are easily removed manually or with the help of a non-abrasive tool. Typically, removal of the coating starts by lifting up the edge and then peeling off the complete coating in one piece rather than in fragments. The peeled coating is basically cross-linked plastic and non-hazardous film, which can be
Strippable Coatings for Removal of Surface Contaminants
197
disposed of according to local regulations for waste plastic materials. These films exhibit elastic properties at room temperature and several films are available that retain their elastic properties at temperatures above 100 C [104 107]. These water-based and solvent-based peelable optical media protective coatings are used for silicon wafers, glass, photomasks, magnetic media platter, and optical media discs. The coatings are fast-drying peelable coatings to protect optical media temporarily from scratching and marring during industrial processing or in material handling and long-term storage.
4. ISSUES WITH STRIPPABLE COATINGS There are several issues with strippable coatings that need to be considered for their use in surface cleaning [108]. These coatings require surface preparation similar to conventional coatings to achieve optimum results. Contaminants, such as grease and oil in large concentrations, result in poor adhesion to the surface, while dirt and paint affect the release of the film from the surface due to increased porosity on the surface. The coating tends to cling to the irregular surface, loses its coherence, and breaks off in pieces when peeled. The thickness of the applied film is also of concern, particularly for lowweight solid coatings. If the film is applied without sufficient thickness, it will tend to strip poorly from the surface. The application of additional film material helps the film to release properly. Films with high-weight solids do not exhibit this poor release characteristic. Water-based strippable coatings should not be excessively stirred or agitated, which can cause bubbling. These air bubbles create irregularities, such as craters and depressions, in the surface of the dried film, which can result in poor cleaning of high-quality surfaces. Water-based coatings that are not freeze thaw stable are also affected by freezing conditions. They tend to lose their adhesion and release properties when frozen and thawed. Most strippable coatings soften when exposed to strong solvents such as acetone and methyl ethyl ketone, making them difficult to remove. However, this problem can be remedied by building sufficient thickness of coating to prevent the solvent from penetrating the coating before it evaporates. Alternatively, a solvent-resistant strippable coating can be used. Water-based coatings with excellent cohesion are available that resist solvent penetration and strip very well. If used for removal of hazardous and radioactive contaminants, strippable coatings can also generate hazardous or radioactive mixed wastes that must be disposed of. The regulations for disposal of such wastes are stringent and the costs to meet the regulations can be very high.
198
Developments in Surface Contamination and Cleaning
5. PRECISION CLEANING APPLICATIONS The use of strippable coatings for precision and other cleaning applications is discussed in the following sections using examples from different industries and applications.
5.1. Optical Surfaces Modern optical instruments have very sophisticated high-quality antireflection coated optic surfaces that sometimes require cleaning [109,110]. Even with a filter permanently in place there are many sources of contamination that form a coating on the lens surface. These come from the lubricant in the filter threads, the atmosphere, and even from within the lens assembly itself. Over time they form a surface layer, which negates the antireflection coating and softens the focus of the lens and contrast of the image, thus reducing the efficiency of the lens coating. Dry-cleaning systems for optical applications require wiping the lens surface with the risk of scratching if there are any small, hard particles of contamination on the surface. Wet-cleaning systems leave a residual smear. This smear acts as a surface layer, effectively softening the focus and contrast of the image.
5.1.1. Cleaning Optics with Collodion The technique for cleaning optical mirrors with collodion was first reported by Purcell in 1953 [111]. He measured the reflectance of mirror samples with an aluminum surface coating that had been cleaned with collodion. The collodion was poured over the mirror surface, allowed to dry, and peeled off. As shown in Figure 5.6, cleaning with collodion increased the reflectance, although the maximum reflectance was achieved only after cleaning multiple times. Even for a fresh aluminum surface that had been protected in a closed drawer, the reflectance increased from 14% after 8 weeks to 33% after cleaning six times with collodion. Kaye [112] also used collodion to successfully clean aluminum mirrors without removing them from their mounts. However, the widespread use of collodion to clean mirrors and other high-quality optical surfaces was established in 1964 after McDaniel described two methods for cleaning mirrors that had been removed from the instrument and for cleaning mirrors remaining in the instrument [113]. Cleaning of mirrors in double-beam spectrophotometers through the wavelength range 0.16 200 mm resulted in no mismatch in mirrors. Another mirror that had been exposed to the atmosphere for 6 months was cleaned on one half of the surface with a solution of distilled water and methanol, and the other half with collodion. The results showed that the side cleaned with collodion was returned to almost its original reflectance, while the other half showed a coarse surface appearance with heavy scattering. In addition, fresh fingerprints also could be easily removed. Surfaces cleaned by
199
Strippable Coatings for Removal of Surface Contaminants
0.35
8 weeks old (protected in closed drawer)
0.30
Reflectance (30º)
0.25 2 years old (very dirty) 0.20
0.15 1 year old (appeared clean)
0.10
0.05
1
2
3
4
5
6
7
8
9
Number of times cleaned FIGURE 5.6 Effect of collodion cleaning on the reflectance of two polished aluminum samples ˚ . The upper curve is for a sample 2 years old that was kept in the wavelength range 1140 1190 A unwrapped in a drawer. The lower curve is for a 1 year old sample that had been kept wrapped in cotton [111].
this method have perhaps 10 times less residual contaminants as compared to a methanol distilled water cleaned surface. In 1966, Canfield et al. [114] performed tests on the feasibility of cleaning MgF2-coated aluminum mirrors whose surface had been contaminated with oil. The best results were obtained when the contaminated surface was rinsed with ether, acetone or Freon followed by collodion cleaning. As an example, ˚ due to oil a mirror whose reflectance was decreased to 9.5% at 1216 A contamination, was restored to 80.7% reflectance after rinsing with Freon, and further increased to 81.8% reflectance after further cleaning with collodion. No damage to the coatings was observed, even after cleaning several of the mirrors more than 10 times. By contrast, Gillette and Kenyon [115] were unsuccessful in removing the proton-irradiated contaminant film on MgF2, LiF or platinumcoated mirrors using Freon and collodion cleaning. There was no change in reflectance in the far ultraviolet wavelength region (100 400 nm). Positive cleaning results were reported by Bloch and Rice [116] on frontfaced aluminum reflector mirrors using McDaniel’s procedure without removing the mirrors from their mounts. Cox et al. [117] could restore the reflectance of contaminated evaporated rhenium films from 29% to 33% after collodion cleaning. In 1970, Tyndall [118] produced a concise report with improvements on the procedures suggested by McDaniel [113]. Essentially, Tyndall introduced the
200
Developments in Surface Contamination and Cleaning
addition of multiple layers of cheesecloth embedded in the collodion layers. This allowed the collodion to seep through and form a solid seal of cheesecloth and collodion over the entire mirror surface. The cheesecloth also aided in peeling the dry collodion from the mirror. Miller [119] has suggested collodion cleaning to maintain consistent reflecting qualities of the surface of a variety of semitransparent materials used as a light reflection standard. These materials include quartz, IR glass, AgCl, ZnS, AsS, Si, and Ge. More recently, Kienzle et al. [120] reported very successfully removing particle contamination from the surface of high-quality multilayer-coated optics and even uncoated transparent ‘superpolished’ fused quartz substrates with collodion. The scattering losses of the cleaned substrates are nearly constant at the lowest scattering losses of the substrate before cleaning. A previous investigation measuring substrates before and after the cleaning process showed no influence on the lowest scatter losses. Thus, the cleaning process removes particles without degradation of the surface. A surface layer of collodion possibly remaining after cleaning does not affect the measured scatter loss. Experiments with collodion to clean large primary mirrors have shown promising results, but there have been objections on safety grounds [121]. One of the largest lenses ever cleaned with collodion is the 0.58-meter Alvan Clark objective at the Charles E. Daniel Observatory in Greenville, South Carolina [122]. The lens was built more than 50 years before the invention in 1935 of modern antireflection coating technology by Carl Zeiss in 1882. Other success stories have been cleaning of the 0.76-meter telescope of the Manastash Ridge Observatory at the University of Washington, which needed cleaning once a year. Again, the results were better than could have been obtained by any other method [122]. Similar results were obtained after cleaning the 0.6-meter telescope at the University of California Los Angeles [123]. All the grit and dust was trapped in the coating. After peeling off the coating, the mirror was left dust and oil free, with no new scratches or damage on its surface.
5.1.2. Cleaning Laser Windows Custom lasers often require Brewster windows free of films and particles. Laser windows in operation show even small numbers of particles. The University of California Berkeley tried other cleaning techniques, but found the collodion method of window preparation to be best suited for their application [124]. The final step in the window preparation protocol that yielded laser-quality windows consisted of coating the optical surfaces of the Brewster windows with collodion and then peeling away the resulting film. A cotton swab was used to evenly apply a layer of collodion over the optical surface. This coating was allowed to dry for 30 60 seconds and then peeled off. Fine contaminant particles were bound in the collodion layer and removed with it. A radioactive,
Strippable Coatings for Removal of Surface Contaminants
201
Po-210 alpha source in the shape of a bar, mounted on a gooseneck, was used to neutralize static charges created by the removal of the collodion. This alpha source may not be necessary for success. The collodion method of particle removal was used in the late 1970s to bond clean, thinned Si wafers to low sodium glass for solar cell research. This process was originally termed field-assisted bonding and is now known as anodic bonding. The collodion method could be applicable to eliminating particles on microelectromechanical systems.
5.1.3. Cleaning Guidelines In using collodion for cleaning, these guidelines should be followed. The general cleaning procedure is outlined on the Antique Telescope Society website [122]. 1. Only USP collodion should be used. Flexible collodion should not be used. 2. Collodion should be applied by simply pouring, spraying or laying it on with a camel’s hair brush. 3. If the mirror or lens is in its cell, it may be necessary to make a dam of masking tape, cardboard or other material to prevent the collodion from seeping between the glass and the cell. If applied with a brush in thin coats, the dam may not be necessary. Any collodion that manages to creep under the tape will form a hard, thick layer that will have to be cleaned by dissolving in acetone. 4. Adding a layer of cheesecloth or surgical gauze while the collodion is still wet will make it easier and simpler to peel the substance off in a single sheet when dried. For small optics, the gauze may not be necessary or desirable. A second coat of collodion should be applied over the cheesecloth. 5. After application, the collodion should be allowed to dry before removal. Generally, telltale signs that it is ready for removal are slight shrinkage, and curling or lifting of the leading edges. At this stage, it can be simply peeled off slowly and carefully in a single sheet, if possible. If patches remain, they can be mopped up carefully with masking tape. 6. Collodion is extremely flammable and emits ether in concentrations high enough to cause dizziness and eye irritation. All work should be performed in a well-ventilated area with no possible ignition sources anywhere in the area. Also, gloves may be desirable for some. The cleaning steps are illustrated below for cleaning a small mirror (Figure 5.7). On small to mid-sized mirrors, collodion is very effective in removing surface contaminants. Another advantage is that the optics can be cleaned in their holder without the need for removing them. One disadvantage is that collodion tends to part and leave small residual fragments on the surface. These
202
Developments in Surface Contamination and Cleaning
FIGURE 5.7 Cleaning a small mirror with collodion. (a) Visibly contaminated mirror. (b) The mirror is coated with collodion. (c) Cheesecloth is placed on the mirror to hold the collodion film together when it is peeled. (d) The edges of the collodion film lift up when the film is dry. (e) The cheesecloth and collodion coating is peeled off, leaving a clean mirror surface [122].
fragments can be removed with masking or adhesive tape. However, collodion generally adheres more efficiently to modern antireflection coatings than to uncoated glass, which could result in removal of the coating [110]. Back [125] had noted the possibility of damaging the lens or mirror coating on cleaning with collodion if it pulled off a part of the coating, as happened before with both mirrors and coated lenses. He did not recommend this technique for the general cleaning of optics. A similar experience was reported at the National Optical Astronomy Observatories, where the collodion coating appeared to trap water under the aluminum mirror coating, which can cause flaking of the coating from the substrate [126].
Strippable Coatings for Removal of Surface Contaminants
203
5.1.4. Non-Collodion Cleaning of Optics In early 1987, Fine and Pernick [127] tested two commercially available noncollodion strippable coatings from Universal Photonics for their efficacy to remove dust and debris from an optical surface. In addition, the tests were designed to determine if the coating itself left any contamination on the surface and to determine the ease of application and removal of the coating. The optical parts included surface mirrors and glass flats with various thin-film coatings (antireflective (AR), tin oxide, or indium tin oxide), as well as blank glass plates for comparison. The coatings were applied by simply pouring the coating over the surface at a slow rate to achieve a thick layer. The high viscosity and surface tension of the coatings made it easier to spread the coating on the part and also prevented it from spilling over the edge. The coatings were dried at room temperature for 24 hours and then easily removed in one piece by pulling a strip of adhesive duct tape attached to the dry coating. Both coatings were capable of removing dust particles from the surface, but one of the materials left a visible film on the sample. Interferometry measurements on the thin-film coated samples and the blank plates did not show any evidence of a residual film after removal of the strippable material. The coating materials could also be used to form a uniform layer on concave or convex parts, by either rotating the part or by a barrier to prevent the material from flowing over the edges. An aerosol spray of plastic base (polyurethane) dissolved in solvent has been used for cleaning the 1-meter primary mirror of the Royal Greenwich Observatory telescope. When sprayed on to a dusty or greasy mirror, the solvent lifts the dirt from the coated surface and suspends it in the plastic film. The contamination is removed when the film is peeled off. Provided the coating has not been attacked chemically, the original performance should be restored [121]. The peelable blue coating from Universal Photonics was used to successfully clean several small mirrors [128]. The coating comes in a spray can and was applied by spraying on the surface of the mirror. The coating dries quickly, leaving a soft, pliable blue membrane on the mirror. It came off easily in a sheet, taking the dust and grime with it from the face of the mirror. There was no apparent residue and the mirrors were clean (Figure 5.8). Davison et al. [129] have used a polymer-based strippable coating to protect and clean the contamination from vacuum pads used to lift the 6.5 and 8.4 m mirrors after polishing, and to install them in the telescope. The pad was placed on polished glass and a vacuum applied several times. The neoprene pad material left contamination that was visibly obvious but was easily cleaned. After the glass was silver coated, the pad was placed on the silver mirrored glass surface coated with the strippable coating. When the strippable coating was removed, absolutely no damage to the mirror coating was observed. Many optics users in the world are using Opticlean routinely to clean astronomical and laser optics with no problems in several major observatories
204
Developments in Surface Contamination and Cleaning
FIGURE 5.8 Sequence of steps for cleaning an optical mirror with peelable blue coating. (a) The mount is wrapped to protect it from the coating. The coating is applied by spraying a thick layer on the surface. (b) The coating is starting to dry. (c) The dried coating is removed by peeling. The mirror surface is clean. (d) Any remaining coating around the bevel is removed with fresh tape [128]. (see colour plate section at end for coloured version)
around the world. Many laser laboratories also use the polymer to clean and protect multi-coated and silvered optics. These applications are exceptionally sensitive to damage and there are no reports of damage yet on record. Many of the major camera and lens manufacturers have tested and approved Opticlean for use on their lenses [43,44]. Opticlean is used to clean almost every optical surface available. Laser gyro mirrors are perhaps the most perfect surface made by man. Flatness of 1/ 10 wave or better is achieved regularly, along with parallelism of 0.762 mm over a 230 mm substrate. Parts with angular tolerances better than 3 arcminutes are provided to customers with NIST traceable certificates accurate to within 5 arc-seconds for the optical angle. Opticlean is the cleaning method of choice for these mirrors. Many manufacturers of coated surfaces are now using Opticlean as the cleaning system to prepare the surface. This has been shown to be equal to or better than the automated ultrasonic CFC-based system commonly used to produce the level of cleanliness needed for this application. Cleaning large mirrors has also been successfully performed with Opticlean Spray Polymer Solution [43,44]. Spray application of Opticlean polymer using manual pumps or pressure sprayers allows complete coverage of any size optic
Strippable Coatings for Removal of Surface Contaminants
205
FIGURE 5.9 A large 1 meter unprotected gold first surface mirror before cleaning (left photograph) and after cleaning (right photograph) with Opticlean. Courtesy of Photonic Cleaning Technologies, Platteville, WI. (see colour plate section at end for coloured version)
surface using a completely touch-free process (Figure 5.9). No wipes or brushes are required, thus eliminating any danger of scratching the surface. Examples of large, unprotected mirrors cleaned with Opticlean include mirrors at the W.M. Keck Observatory in Hawaii (1.8-meter-diameter hexagonal mirrors), the La Silla Paranal Observatory (0.875- to 3.6-meter-diameter mirrors) in La Silla, Chile, and the LuLin Observatory (76-cm-diameter honeycomb mirror) in Taiwan. Other applications include several military optics depots with responsibility for cleanliness of telescope mirrors from 20 cm to 3.6 m diameter.
5.1.5. Cleaning Phase Masks Traditional methods for cleaning phase masks are time-consuming and require hazardous chemicals. Opticlean simplifies cleaning phase masks. If the contaminants are only particulates the application amount may be as little as 1 mL for an area of 22.5 cm2, but if there are oils or fingerprints on the surface, 1 mL may be required for 13 cm2 surface area. When lifting oils from a contaminated phase mask, placing a cover over the treated optical part will help to slow down the drying process and allow it to dry overnight. If the oily contamination is heavy, rinsing with an organic solvent will help prior to using the coating. The coating is safe on fused silica, glasses of all kinds, and crystals. 5.1.6. Cleaning Silicon Wafers For extreme ultraviolet (EUV) applications in space-based astronomy and future lithography for integrated circuit computer chips, the main impediment to further development of efficient mirrors is the lack of reliable optical constants for various materials in this region of the electromagnetic spectrum. One reason for the unreliability of the optical constants is that the sample
206
Developments in Surface Contamination and Cleaning
surfaces are contaminated with organic material when they are exposed to laboratory air. Robinson [130] evaluated the efficacy of Opticlean, oxygen plasma etching, high-intensity UV light, and Opticlean followed with oxygen plasma etching for removing the organic contaminants from the surface of silicon wafers. Oxygen plasma effectively removes any Opticlean residue left on the surface. Bennett et al. [131,132] have found that Opticlean can remove 1- to 5-mm-diameter particles as well as contamination remaining from previous drag wipe cleaning on a used silicon wafer. In addition, no residue that produced scattering was found on a fresh silicon wafer when Opticlean was applied and then stripped off. The total integrated scattering technique used for the measurements could measure scattering levels of He Ne laser light as low as a few ppm, corresponding to an rms (root mean square) surface roughness ˚. of <1 A
5.1.7. Cleaning Stampers for Compact Discs Stampers used for compact disc manufacturing often have foreign debris defects from grease or oil, water marks, and residue from solvents used in stamper manufacturing, coating and drying processes, as well as environmental contamination such as fingerprints and saliva. These defects are transferred on to the discs during fabrication. Removal techniques such as electrolysis or plasma etching techniques, or solvent cleaning are generally time-consuming. Moynagh [133] reported a novel, improved method for cleaning highprecision optical components by removing smudges, oil, dust, saliva, and other contaminants from cleaning molds utilized in the production of compact discs (CDs). The method involves applying a curable coating to an optical or optical forming surface that has been contaminated with undesirable foreign matter, allowing the coating to dry to form a resilient dried film, and then removing the dried film together with the undesirable foreign debris that has contaminated the optical surface. The curable coating is formed from an aqueous polyurethane emulsion. It also includes a butyl cellusolve (2-butoxyethanol or C4H9OCH2CH2OH) solvent and a trace amount of surfactant/thickener as wetting agent and for viscosity control. The use of a dual-solvent system (water and butyl cellusolve) is advantageous since these solvents are mutually soluble and have the ability to dissolve a variety of solutes that may be present as foreign debris on the surface being decontaminated. In practice, the stamper is placed in a spin applicator unit and rotated at about 20 30 rpm, while applying a curable coating to the optical forming surface of the stamper. The coating is applied at or near the inner edge or inner diameter of the annulus-shaped compact disc. After application of the coating, the stamper is rotated at high speed, up to about 300 rpm, to insure the coating is applied evenly. Then the stamper is removed from the spin applicator before curing the coating. The curing step can be accelerated by heating the stamper with hot air after the
Strippable Coatings for Removal of Surface Contaminants
207
stamper has been removed from the spin applicator unit. The coating forms a resilient, dry film on the optical forming surface, which is then removed from the surface. The method allows foreign debris to be removed with a very high degree of reliability and completeness, and the stampers are restored to a like new condition.
5.2. Other Applications Several non-optical applications of strippable coatings are described below.
5.2.1. Particle Sample Preparation Collodion film is a very popular sample preparation medium. It is conveniently available on a microscope slide and can be reconstituted instantly (made lightly viscous) with a single drop of amyl acetate [27,28]. Small drops of the reconstituted sample on the end of a tungsten needle are used for isolating, characterizing, and preparing small particles for further analysis. The reconstituted sample will remain lightly viscous for up to 15 minutes. The slide can be reused many times and stored indefinitely (Figure 5.10). This extraction replication technique is especially useful when the surface is rough and soft at the same time. The collodion film will separate the particles lightly held in crevices without removing any of the substrate. Particles can also be collected directly on collodion films for subsequent analysis [134,135]. Suratwala et al. [136] used Opticlean to collect the residual slurry debris from polishing the surface of meta-phosphate laser glass to study the effects of the slurry on the surface topography of the glass. This was done by applying Opticlean on the surface of full-size samples. The coating was peeled off and portions of the coating were treated in aqua regia to analyze the solution by inductively coupled plasma mass spectrometry (ICP-MS) for cerium polishing residue. The coating was effective in collecting all the residue from the surface.
FIGURE 5.10 Collodion extraction replication technique for particle sample preparation. (a) In Step 1, the collodion solution is spread over a small area of the contaminated surface. (b) The dry collodion film with the particles is pulled off the substrate in Step 2 and placed, particle side down, on a glass slide for analysis [27,28].
208
Developments in Surface Contamination and Cleaning
5.2.2. Skin Wart Remover For special medical applications, a solution containing pyroxylin (nitrocellulose) in an appropriate non-aqueous solvent can be used to form a transparent cohesive film when applied to the skin in a thin layer. This collodion-like film has been shown to be very effective in removing warts on skin [137]. 5.2.3. Magnetic Particle Testing Molina [138] has employed a clear vinyl strippable coating to obtain a permanent record of the location and size of the defects and discontinuities in the surface of an object inspected by magnetic particles. Following magnetization and the agglomeration of the magnetic particles according to the pattern of defects, the coating is sprayed on the surface. The magnetic indications are ‘‘frozen’’ in place on the coating when it is stripped off. The defect pattern can be recorded photographically from the coating.
5.3. Non-Optical Cleaning Applications Strippable coatings are increasingly employed in non-optical cleaning applications, such as radioactive decontamination and other hazardous contaminant removal.
5.3.1. Radioactive Decontamination Applications The conventional treatment method for radioactively contaminated components is to apply a prevulcanized rubber latex compound to the surface, allow it to dry, and then pull it off the surface to remove the contaminants. Unfortunately, the conventional method suffers from several drawbacks. The latex compound can take days to dry and can thus prolong the shutdown time. The latex compound is also sometimes difficult to pull off the surface after it dries. Furthermore, the latex compound is not totally effective in removing the contaminants and reducing radioactive exposure to an industry-accepted level. The latex rubber coating has been modified by incorporating a complexing alkali detergent and a fine pumice in the latex solution to enhance its decontaminating efficiency [139]. DFs of 100 200 were achieved with the modified coating when the coating was used to remove fission product contaminants from a variety of substrates, including stainless steel, aluminum, copper, painted wood, and plastic. The disadvantages of conventional treatments for radioactive surface decontamination can be overcome by improved strippable coatings [140 145]. However, strippable coatings are not very effective in removing contaminants that have seeped below the surface of the part. The environmental impact of strippable coatings is less than competing decontamination technologies. A recent life cycle assessment (LCA) of strippable coating compared with a competing baseline steam vacuum cleaning technology confirmed better
Strippable Coatings for Removal of Surface Contaminants
209
environmental performance (impact on human health, ecosystem quality, and resources expended for production and waste disposal) of strippable coatings in all phases of application of these technologies [146]. The LCA was extended to include a simplified economic analysis, which showed strippable coatings to be five times more cost-effective than the baseline steam vacuum cleaning technology due to the need for disposal of much larger volumes of waste generated by steam vacuum cleaning [147,148]. In effect, the waste disposal phase becomes the most important LCA discriminator between these technologies. Vogel and Cugini [149] reported experimental work performed with a selfstripping copolymer compound (PENTEK-6002) to evaluate its ability to remove oxides and radioactive contamination from oxidized ferrous and nonferrous surfaces and non-porous surfaces. PENTEK 604Ô was a coating that incorporated organic acids and chelants and also was self-stripping. It was designed to release from the substrate in about 24 hours and flake off the surface. As a decontamination agent, the copolymer produced DFs of up to 245 [150]. Based on the experimental data, the underlying carbon steel surface condition after oxide removal using the copolymer was comparable to that achieved by commercial blast cleaning [150]. Similar high decontamination factors were obtained from application of the PENTEK coating on various radioactively contaminated bare and coated concrete surfaces, as well as stainless steel [151]. The coating has also been found to be effective on incinerated graphite fuel residue, cesium capsule, and lead bricks. DFs of several hundred have been achieved [143,152,153]. Yin et al. [154] have combined film coating and electrolysis methods to decontaminate equipment contaminated with uranium. The strippable film consisted of polyaniline doped with SO24 /TiO2 superacid that replaced the more commonly used HCl to increase the conductivity of the film. The decontamination methods evaluated were as follows: Coating the contaminated metal surface with the film and peeling off the dried film. Coating the surface with the film and electrolyzing in a phosphoric acid solution. The film was peeled off the dried metal surface after electrolysis and analyzed. The contaminated metal surface was coated with the film and electrolyzed in a dry medium. The film was peeled off after electrolysis and analyzed. Decontamination ratios from 73% to 98% were achieved, with the best results obtained from dry electrolysis in the third method. Reactor cavity decontamination at the Nine Mile Point Nuclear Station in Oswego, New York, was performed using the Isotron coating Isolock 300 [155].
2
The PENTEK series of strippable coatings were developed and commercially offered by Pentek, Inc. (Coraopolis, PA). The coatings have been discontinued.
210
Developments in Surface Contamination and Cleaning
The application and removal of the coating required a total of 148 hours to complete. Radiological surveys performed after the removal of the coating indicated contamination levels of 10,000 and 50,000 dpm/100 cm2 (dpm ¼ disintegrations per minute) on the walls and floors respectively. The cavity decontamination was completed with 65 mman-Sv of exposure, a reduction of approximately 38% as compared to the previous outage. This was attributed to an increase in worker efficiency due to a decline in the protective clothing and respiratory protection required to perform the work. The calculated dose equivalent to the skin as a result of these contaminations was down from 312 to 120 mSv. At the Savannah River Site (SRS), approximately 264 m2 of the wall and floor areas of a fuel fabrication facility were decontaminated with ALARA 1146Ô strippable coating [156 158]. This area was contaminated with highly enriched uranium and designated a high contamination area. The intent was to remove enough of the loose radioactive material to drive transferable contamination levels below the free release limits. The coatings were sprayed on to painted and unpainted carbon steel and concrete surfaces in the process area of the facility. The results showed that the strippable coating was effective in removing contamination from all surfaces that were tested. For the surfaces that had the highest contamination levels, DFs up to 7.2 for alpha activity and up to 3.9 for beta gamma activity were achieved. The DF was lower for the less contaminated surfaces, as would be expected. The Capture Coating has been used in conjunction with InstaCote CC Strip [52] in a two-step process that removes the plutonium contamination from the air by the aerosolized Capture Coating and then seals it in place on the floor or walls of a room or other surfaces with InstaCote. At the Rocky Flats plant in Colorado, implementation of this process reduced the contamination in a room containing plutonium nitrate solution processing tanks from more than 90,000 derived air concentrations (DAC), a measure for plutonium particles in the air, to less than 100 DAC. For this unique application, the Capture Coating was modified by adding a blue fluorescent tracer that is visible in UV light (‘‘black light’’). This helps in tracking potentially contaminated fog residues that may adhere to worker clothing and equipment. Capture Coating and InstaCote, either together or separately, have been used to decontaminate a range of items at the Rocky Flats plant from tanks to glove boxes, to ventilation ducts, process piping, and other highly contaminated rooms, as well as at other sites [53,159]. Recently, the US Environmental Protection Agency (EPA) evaluated Stripcoat TLC FreeÔ and Isotron coatings for their ability to remove Cs-137 isotope from concrete surfaces [160,161]. The coatings were applied with a paint sprayer to concrete coupons (Type I and II Portland cements) contaminated with Cs-137 (Figure 5.11). The treated surfaces were allowed to cure overnight into a solid coating and the coating was then removed from the concrete surface. Isotron binds radiological material through chemical and physical interactions when it is cured.
211
Strippable Coatings for Removal of Surface Contaminants
Stripcoat Coating
Preparation
Application
Removal
Isotron Coating
Preparation
Application
Removal
FIGURE 5.11 Application of commercial strippable coatings for removal of radioactive contaminants from a concrete surface. The upper series of pictures shows the preparation, application, and removal of Stripcoat TLC FreeÔ coating from the surface. The lower series of pictures show the preparation, application, and removal of Isotron coating from the surface [160,161]. (see colour plate section at end for coloured version)
Stripcoat TLC FreeÔ binds radiological material only through physical interaction between the radiological material and the cured coating. Each technology was evaluated for the following operational factors: Decontamination efficacy (percentage removal) from vertical and horizontal surfaces, 7 and 30 days after contamination of the coupons Application and removal times Ease of use on irregular surfaces Labor requirements Utility requirements Portability Secondary waste Surface damage Preparation and cleanup Cost.
212
Developments in Surface Contamination and Cleaning
The decontamination efficacy calculated for each of the contaminated coupons is expressed in terms of percentage removal (%R). For Stripcoat TLC FreeÔ, the overall average %R for both 7- and 30-day tests was 32.0 9.9. For Isotron, the overall average %R for both 7- and 30-day tests was 76.2 7.4. Operational factors of using these strippable coatings are identified and summarized in Table 5.1 [160,161]. The removal rate of Isotron will likely depend on the characteristics of the surface being decontaminated, as some scraping is required for removal. The Stripcoat TLC FreeÔ can be applied to irregular surfaces and easily removed from across the borders of the coupons. The decontamination of concrete by an Isotron hyper-accumulating strippable polymer (HASP) coating has been modeled using a finite element method [162]. The basic concept of decontamination by HASP coating is to include tight-binding sites or ‘‘traps’’ within the coating that accumulate the target contaminants. The overall decontamination process involves the solvation and transport of the contamination, diffusion of the contaminants out of the
TABLE 5.1 Strippable Coating Operational Factors Factors
Stripcoat TLC FreeÔ
Isotron
Application and removal
Application: 12 m2 h Removal: 4.9 m2 h 1
Application: 4.6 m2 h Removal: 1.6 m2 h 1
Ease of use on irregular surfaces
Elastic coating readily peels off surface
Some scraping might be required
Labor requirements
No specialized training
No specialized training
Utility requirements
If sprayer used, 110 V; otherwise none
If sprayer used, 110 V; otherwise none
Portability
Portable
Portable
Secondary waste
Solid waste production: ~0.26 kg m 2 Solid waste density: ~0.145 g cm 3
Solid waste production: ~0.5 kg m 2 Solid waste density: ~0.188 g cm 3
Surface damage
Minimal, only loose particles removed
Minimal, only loose particles removed
Preparation and cleanup
Product used ‘‘as is’’ pump rinsed with mineral spirits between applications to avoid clogging
Product requires mixing; pump rinsed with water between applications
Cost
$16.66/m2 for one application
$58.84/m2 for one application
1
1
Strippable Coatings for Removal of Surface Contaminants
213
substrate due to the very large equilibrium distribution coefficient values of the polymer, and removal and transportation of the contamination. The concentration profile of the contaminant within the porous coating can be obtained by solving the governing equation describing species transport driven by capillary flow, diffusion, and adsorption, as well as the governing equation for unsaturated flow. The simulation results showed that about 90% of the contamination was removed when the coating was applied within 30 minutes of the contamination event, but this was reduced to less than 50% removal if the contaminated surface was allowed to dry for 151 days before the coating was applied. Prompt application of the coating results in higher decontamination efficacies. DeconGelÔ [50] has been evaluated on a variety of materials (concrete, carbon steel, stainless steel, Plexiglas, painted concrete, porcelain tile (with and without grout), granite (with grout), vinyl composite tile) and surfaces (smooth, rough, painted, porous) for its efficacy to remove radionuclides. In general, the gel was poured on to the surface of the contaminated coupons and allowed to dry overnight. All the coupons were easy to peel off the coating, the most difficult being concrete. The coatings were all removed in a single sheet. None of the coatings fractured during removal. Testing on samples contaminated with isotopes Cs-137, Co-60, Eu-154, Np-237, U-238, Pu-238, Pu-239 and Pu-241, Am-241, I-125 and I-131, Tc-99, Tl-201, C-14, and tritium has shown removal efficiencies ranging from 82% to as high as 100% [163 167]. Lower values were attributed to particular material and surface chemistry combinations, as well as 90% humidity conditions, damp porous concrete, and incomplete cure time, illustrating the importance of these factors when planning any decontamination procedure. A second gel application performed 2 months later on the same test coupons gave results that were as good as or better than first application results with an average removal efficiency of 87% and a median of 99%. The results indicated that the majority of contamination removed by the gel with high DFs was not simply loose particulate, but fixed in the surface. Davison et al. [168] have developed a process in which an aqueous biopolymer solution is used to coat a contaminated metal surface (such as steel), solubilize the heavy metals from the surface, and bind the heavy metals into the biopolymer. The biopolymer coating containing the immobilized hazardous metal contaminants can be stripped as a viscous film, as a dry powder, or by washing. The soluble polysaccharide (SPS) fraction of biopolymers, produced by the cyanobacterium Nostoc muscorum strain HPDP22, was tested for sorption properties with regard to barium, cadmium, iron, copper, and uranium. The results showed that the biopolymer removes >80% of the fixed uranium(VI) from contaminated steel coupons. Andersson and Roed [169] evaluated the efficiency and cost-effectiveness of several strippable coatings for removal of cesium isotopes deposited on soil and
214
Developments in Surface Contamination and Cleaning
grass from radioactive fallout from nuclear accidents such as Chernobyl. A modified poly(vinyl alcohol) coating, a commercial lignin coating Lignosol BD, and a liquid plastic coating ‘‘Liquid Envelope’’ were tested on small (0.03 m3) soil core samples from the Chernobyl area. All three coatings removed 94 95% of the cesium from the bare soil samples, while Lignosol BD could remove nearly 72% cesium from a cut grassed surface. The other coatings were not tested on the grass samples. A comparison of the total cost of decontamination (materials, equipment, labor, and waste management) showed that Liquid Envelope was nearly three to four times higher than the other two coatings.
5.3.2. Non-Radioactive Decontamination Applications Lumia and Gentile [170] successfully decontaminated three floor areas of the Tokamak Fusion Test Reactor at Princeton University using Stripcoat TLC FreeÔ. The floor areas were contaminated with lead dust. A thick coating was applied to the floor and after approximately 24 hours the coating was peeled and the floor was sampled. The results were below the release guidelines. Many surfaces exposed to the environment are continuously being contaminated by undesirable deposits, such as soot, grease, dust, traffic pollution, and other contaminants. Furthermore, accidental spills can make stains that can be extremely difficult or even impossible to remove. Cleaning such contaminated surfaces generally requires strong alkaline solutions or organic solvents that cause health hazards and are not environmentally friendly. Masking procedures are often difficult to perform on curved or irregular surfaces. Svensson [171] has developed a process for facilitating the removal of undesired contamination from a surface that overcomes these disadvantages. A solution containing a polysaccharide and a solvent is applied to the contaminated surface and the solution is allowed to dry to form a solid polysaccharide film. The solid polysaccharide is capable of redissolving or swelling in water or other liquids. The swollen or dissolved layer can be removed by any suitable technique, carrying the contaminants with the film. The polysaccharide can be selected from the groups consisting of celluloses and derivatives, starch and starch derivatives, plant gums, capsular microbial polysaccharides, pectins, inulins, and algal polysaccharides. To obtain maximum mechanical strength about 5 20% by weight of plasticizer is usually added to the solution. Although dry polysaccharide films are hard and brittle due to the multiplicity of hydrogen bonding sites, under normal conditions water is always present, making the films soft and pliant. The polysaccharide films have been applied to contaminated concrete, galvanized steel, and other surfaces. After removal of the dried film, no contamination was visible. Fox et al. [172] proposed an alternative method for decontaminating porous materials. In this method, a polyphosphazine-based polymer coating is applied
Strippable Coatings for Removal of Surface Contaminants
215
to the surface of the contaminated porous material. The surface of the porous material is irradiated with a beam of coherent electromagnetic radiation with a wavelength in the range that does not cause ablation of the structure. Irradiation of the surface causes the contaminants within the cracks and pores to redistribute towards the surface of the structure. The contaminants are physically or chemically bound to the coating and are removed when the cured coating is peeled off the surface. To prove the effectiveness of the chelator on environmental samples, and to prove that chelation could in fact dissolve and bind beryllium oxide, Sutton et al. [78] investigated the effect of APMDP on BeO debris in the Contained Firing Facility at Lawrence Livermore National Laboratory. Varying concentrations of APMDP chelator (pH adjusted to 7) were added to vials containing BeO debris and left to stand for 3 days, with manual shaking performed for 2 minutes for each vial, twice a day. Samples were then filtered through a 0.2 mm membrane and filtrates were analyzed by ICP-MS. The results clearly demonstrated a linear concentration profile, indicating that APMDP dissolves insoluble BeO fines and binds beryllium. Hand-peelable commercial coatings have been employed for cleaning and protecting the surfaces of military aircraft prior to and during application of the final epoxy-polyamide paint [173 175]. The coatings provided good protection for the aircraft panels during chemical cleaning and during drilling, countersinking, and riveting operations. The coatings also provide long-term durability as evidenced by laboratory testing and field inspection of coated surfaces after exposure up to 12 months.
6. SUMMARY A wide variety of solvent- and water-based strippable coatings are available for precision cleaning of surfaces and for protecting surfaces that have been cleaned. This is a low-cost, effective method for precision cleaning of surfaces, in which the coating is applied to the surface by spraying, rolling or brushing. The coating is allowed to cure and then removed by peeling. The coating formulations are designed to entrain the contaminants by physical or chemical means. When the coating is peeled, the contaminants are carried with the coating. The types of surfaces cleaned include coated and uncoated optical lenses and mirrors, silicon wafers, metals, plastics, and concrete. The contaminants that have been successfully removed include dust, hydrocarbon films, radioactive materials, beryllium, and other hazardous materials.
ACKNOWLEDGEMENTS The author is very grateful to Jody Mantell of the University of Houston at Clear Lake for help with locating obscure reference articles.
216
Developments in Surface Contamination and Cleaning
DISCLAIMER Mention of commercial products in this chapter is for information only and does not imply recommendation or endorsement by The Aerospace Corporation. All trademarks, service marks, and trade names are the property of their respective owners.
REFERENCES [1] G.W. Grogan, R.H. Boyd, Aqueous Based, Strippable Coating Composition and Method, US Patent 5,143,949 (1992). [2] M.R. Fisch, Liquid Crystals, Laptops and Life, World Scientific Publishing, Singapore, 2004. [3] J.J. Meister, Polymer Modification: Principles, Techniques, and Applications, CRC Press, Boca Raton, FL, 2000. [4] T. J. Pelouze, E. Fre´my, Traite´ de chimie ge´ne´rale, analytique, industrielle et agricole, Masson, Paris, France, 1865. [5] T. J. Pelouze, E. Fre´my, Me´taux et me´tallurgie, Masson, Paris, France, 1878. [6] P. Valkhoff, Louis Me´nard (1822 1901), Neophilologus 1 (1916) 88. [7] J.P. Maynard, Communication to the Boston Society for Medical Improvement, Boston, MA, March 1848. See New York Medical Journal IV, (1867) 390 and Annual Cyclopaedia and Register of Important Events of the Year 1898, D. Appleton, New York, 1899, p. 558. [8] M.E. Gripon, Proprie´te´s physiques des lames de collodion [Physical Properties of Collodion Layers], Compt. Rend 80 (1875) 882. [9] K. Sollner, I. Abrams, C.W. Carr, The Structure of the Collodion membrane and its Elec trical Behavior. I. The Behavior and Properties of Commercial Collodion, J. Gen. Physiol. 24 (1941) 467. [10] K. Sollner, I. Abrams, C.W. Carr, The Structure of the Collodion Membrane and its Electrical Behavior. II. The Activated Collodion Membrane, J. Gen. Physiol. 25 (1941) 7. [11] K. Sollner, C.W. Carr, I. Abrams, The Structure of the Collodion Membrane and its Electrical Behavior. III. The Base Exchange Properties of Collodion, J. Gen. Physiol. 25 (1942) 411. [12] K. Sollner, C.W. Carr, The Structure of the Collodion Membrane and its Electrical Behavior. IV. The Relative Merits of the Homogeneous Phase Theory and the Micellar Structural Theory as Applied to the Dried Collodion Membrane, J. Gen. Physiol. 26 (1942) 17. [13] K. Sollner, C.W. Carr, The Structure of the Collodion Membrane and its Electrical Behavior. V. The Influence of the Thickness of Dried Collodion Membranes upon Their Electromotive Behavior, J. Gen. Physiol. 26 (1943) 309. [14] I. Abrams, K. Sollner, The Structure of the Collodion Membrane and its Electrical Behavior. VI. The Protamine Collodion Membrane, A New Electropositive Membrane, J. Gen. Physiol. 26 (1943) 369. [15] C.W. Carr, K. Sollner, The Structure of the Collodion Membrane and its Electrical Behavior. VII. Water Uptake and Swelling of Collodion Membranes in Water and Solutions of Strong Inorganic Electrolytes, J. Gen. Physiol. 27 (1943) 77. [16] K. Sollner, J. Anderman, The Structure of the Collodion Membrane and its Electrical Behavior. VIII. Quantitative Studies Concerning the Acidic Properties of Collodion and Their Correlation with Membrane Structure and Activity, J. Gen. Physiol. 27 (1944) 433. [17] K. Sollner, P.W. Beck, The Structure of the Collodion Membrane and its Electrical Behavior. IX. Water Uptake and Swelling of Collodion Membranes in Aqueous Solutions of Organic Electrolytes and Non Electrolytes, J. Gen. Physiol. 27 (1944) 451.
Strippable Coatings for Removal of Surface Contaminants
217
[18] K. Sollner, C.W. Carr, The Structure of the Collodion Membrane and its Electrical Behavior. X. An Experimental Test of Some Aspects of the Teorell and Meyer Sievers Theories of Electrical Membrane Behavior, J. Gen. Physiol. 28 (1944) 1. [19] C.W. Carr, K. Sollner, The Structure of the Collodion Membrane and its Electrical Behavior. XI. The Preparation and Properties of Megapermselective Collodion Membranes Combining Extreme Ionic Selectivity with High Permeability, J. Gen. Physiol. 28 (1944) 119. [20] C.W. Carr, H.P. Gregor, K. Sollner, The Structure of the Collodion Membrane and its Electrical Behavior. XII. The Preparation and Properties of Megapermselective Protamine Collodion Membranes Combining High Ionic Selectivity with High Permeability, J. Gen. Physiol. 28 (1945) 179. [21] J.F. Marvin, Studies of the Structure of Collodion and Collodion Membranes Utilizing the Electron Microscope and the High Speed Microtome, Ph.D. Thesis, University of Minnesota, Minneapolis, MN, 1950. [22] F.S. Archer, On the Use of Collodion in Photography, The Chemist 2 (1851) 257. [23] J. Towler, The Silver Sunbeam, Joseph H. Ladd, New York, 1864. [24] R. Leggat, A History of Photography from its Beginnings till the 1920s, online publication. Website: http://www.rleggat.com/photohistory, 1995. [25] C.A. Long, The Dry Collodion Process, second ed., Bland & Long, London, 1857. [26] H. Gernsheim, Geschichte der Photographie: Die ersten 100 Jahre, Ullstein, Frankfurt/Main, Germany, 1983. [27] A.S. Teetsov, Unique Preparation Techniques for Nanogram Samples, in: H.J. Humecki (Ed.), Practical Guide to Infrared Microspectroscopy, CRC Press, Boca Raton, FL, 1995. [28] A.S. Teetsov, Uses for Flexible Collodion in the Analysis of Small Particles, White Paper, McCrone Associates, Westmont, IL, 2004. [29] R. Jagels, Celloidin Embedding Under Alternating Pressure and Vacuum, Trans. Am. Microsc. Soc. 87 (1968) 263. [30] Dorland’s Illustrated Medical Dictionary, 31st ed., Saunders Imprint, Elsevier, UK, 2007. [31] S.L. Bigelow, A. Gemberling, Collodion Membranes, J. Membrane Sci. 100 (1995) 57. [32] A. Yamauchi, A.M. El Sayed, New Approach to a Collodion Membrane Composite via Fluorocarbon Polymer (Nafion) Blending in Terms of a Diffusion Coefficient of Redox Substances and Transport Properties, Desalination 192 (2006) 364. [33] F.A. Castle, C. Rice (Eds.), Collodion Combinations, 13, American Druggist 13 (1884) 5. [34] H.W. North, Cooley’s Cyclopaedia of Practical Receipts, seventh ed., J&A Churchill, London, vol. 1. 1892. [35] The Columbia Encyclopedia, sixth ed., Columbia University Press, New York, 2009. [36] Collodion USP, Material Safety Data Sheet (2001). [37] H.C. Semon, Some Recent Experiences with Pure Coal Tar (Pix Carbonis Preparata, B.P.) at a Base Hospital in France, Br. J. Dermatol. 29 (1917) 83. [38] A. Di Landro, R. Valsecchi, T. Cainelli, Contact Allergy to Dithranol, Contact Dermatitis. 26 (1992) 49. [39] M. Kucharekova, L. Lieffers, P.C.M. van de Kerkhof, P.G.M. van der Valk, Dithranol Irritation in Psoriasis Treatment: A Study of 68 Inpatients, J. Eur. Acad. Dermatol. Venereol. 19 (2005) 176. [40] UV Curable Peelable Coatings, General Chemical Corporation, Brighton, MI. Website: http://www.strippablecoating.com/OpticalMedia.aspx, 2009. [41] Carbicote 946 White Booth Coating, Product Data Sheet, Carbit Paint Company, Chicago, IL. Website: http://www.carbit.com, 2009. [42] Removable Spray Booth Coatings, Product Data Sheet, Chemco Manufacturing Co., Chicago, IL. Website: http://www.chemcomfg.net/coatings.htm, 2009. [43] P. Jackson, Große Spiegel reinigen und schutzen, Photonik (May 2007) 64 67. [44] Opticlean First Contact Cleaning Coating, Dantronix Research, Platteville, WI. Website: http://www.dantronix.com, 2009.
218
Developments in Surface Contamination and Cleaning
[45] Universal Photonics, Hicksville, NY. Website: http://www.universalphotonics.com, 2009. [46] CarbicoteÒ Strippable Coatings, Carbit Paint Company, Chicago, IL, 2009. Website: www. carbit.com. [47] CorShieldÒ VpCIÔ Strippable Coating, Cortec Corporation, St Paul, MN, 2009. Website: http://www.cortecvci.com. [48] Stripcoat TLC FreeÔ, Product Data Sheet, Bartlett Services, Inc, Plymouth, MA, 2009. Website: http://www.bartlettinc.com. [49] ALARA 1146 Strippable Coating, Product Data Sheet, Carboline Company, St Louis, MO, 2009. Website: http://www.carboline.com. [50] DeconGelÔ, Decontamination Solutions for Chemical, Biological and Radioactive Threats, Cellular Bioengineering, Inc, Honolulu, HI, 2009. Website: http://www.decongel.com. [51] T.J. Nachtman, J.H. Hull, Method for Containing or Removing Contaminants from a Substrate, US Patent 5,763,734 (1998). [52] Contamination Control Strippable (CC Strip) Coating, Product Data Sheet, InstaCote Inc, Erie, MI, 2009. Website: http://instacote.com/products.htm. [53] Capture Coating, Product Data Sheet, Encapsulation Technologies, Los Angeles, CA, 2009. Website: http://www.fogging.com. [54] H.L. Lomansey, In Situ Polymeric Membrane for Cavity Sealing and Mitigating Transport of Liquid Hazardous Materials Based on Aqueous Epoxy Rubber Alloys, US Patent 5,091,447 (1992). [55] J. Shelton, Polymer Technologies for Lockdown and Removal of Radioactive Contamina tion, in Proc. United States Special Operations Command (USSOCOM) Chemical, Bio logical, Radiological, and Nuclear (CBRN) Conference & Exhibition, Tampa, FL (2005). Website: www.dtic.mil/ndia/2005ussocom/wednesday/shelton.pdf. [56] Isotron Radblock, Product Data Sheet, Isotron Corporation, Seattle, WA, 2009. Website: http://www.isotron.net. [57] M.D. Kaminski, M.R. Finck, C.J. Mertz, M. Kalensky, N. Kivenas, J.L. Jerden, Nano particles, Super Absorbent Gel Used to Clean Radioactivity from Porous Structure Nondestructively, Chemical Engineering Division Report, Argonne National Laboratory, Argonne, IL, 2005. [58] L. Nunez, M.D. Kaminski, Foam and Gel Methods for the Decontamination of Metallic Surfaces, US Patent 7,166,758 (2007). [59] S.V.S. Rao, K.B. Lal, Surface Decontamination Studies Using Polyvinyl Acetate Based Strippable Polymers, J. Radioanal. Nucl. Chem. 260 (2004) 35. [60] J.B. Steward, J.M. Johnston, Strippable PVA Coatings and Methods of Making and Using the Same, International Patent WO/2005/031757 (2005). [61] J. Roed, K.G. Andersson, H. Prip, Practical Means for Decontamination 9 Years after a Nuclear Accident, Report Risø R 828(EN), Risø National Laboratory, Roskilde, Denmark, 1995. [62] P. Hubert, L. Annisomova, G. Antsipov, V. Ramsaev, V. Sobotovitch, Strategies of Decontamination. Final Report of ECP4 Project, Report EUR 16530, Commission of the European Communities, Luxembourg, 1996. [63] J.M. Kuperberg, M. Khankhasayev, J. Moerlins, R. Herndon, Overview of International Programs for Identification and Evaluation of Technologies for DOE EM, in Proc. Industry Partnerships for Environmental Science and Technology, Sect. 4.3, DOE NETL, Morgan town, WV, 2001. [64] K. Eged, Z. Kis, G. Voigt, K.G. Andersson, J. Roed, K. Varga, Guidelines for Planning Interventions Against External Exposure in Industrial Areas after a Nuclear Accident. Part 1: A Holistic Approach of Countermeasure Implementation, Report GSF Bericht 01/2003, Institut fur Strahlenschutz, Gesellschaft fur Strahlenforschung (GSF), Germany, 2003.
Strippable Coatings for Removal of Surface Contaminants
219
[65] M. Prazska´, J. Rezba´rik, M. Solca´nyi, R. Trtilek, Actual Situation on the Field of Decon tamination in Slovak and Czech NPPs, Czech. J. Phys. 53 (2003) A687. [66] Technologies and Strippable Coats for Radiation Situation Normalization, NIKIMT Research and Design Institute of Assembly Technology, Moscow, Russia, 2003. Website: http://www.fita.org/prbc/event/techexpo proposal.html. [67] N. Voronik, Yu. P. Davydov, Decontamination of Industrial Equipment Contaminated as a Result of Nuclear Accident. Website: http://130.226.56.167/nordisk/publikationer/1994 2004/Contamination Urban Areas/Voronik.ppt, 2004. [68] J. Brown, K. Mortimer, K.G. Andersson, T. Du´ranova´, A. Mrsˇkova´, R. Hanninen, T. Ikaheimonen, G. Kirchner, V. Bertsch, F. Gallay, N. Reales, Generic Handbook for Assisting in the Management of Contaminated Inhabited Areas in Europe Following a Radiological Emergency. Part II: Compendium of Information on Countermeasure Options, EURANOS (CAT1) TN(07) 02, Health Protection Agency, Oxfordshire, UK, 2007. [69] S. Mikheykin, A. Barinov, V. Safronov, V. Salikov, D&D Experience in Mos SIA Radon, in 2008 IAEA Annual Forum for Regulators and Operators in the Field of Decommissioning: International Decommissioning Network (IDN), International Atomic Energy Agency, Vienna, Austria. Website: http://www.iaea.org/OurWork/ST/NE/NEFW/wts IDN peerrvw. html, November 2008. [70] Adhesives Research, Inc, Glen Rock, PA. Website: http://www.adhesivesresearch.com, 2009. [71] R. Sugino, H. Mori, Removing Particles from Silicon Wafer Surfaces with Adhesive Tape, Microcontamination 14 (1996) 43. [72] T. Kubozono, Y. Moroshi, Y. Ohta, H. Shimodan, N. Moriuchi, A New Process for Resist Removal after Photolithography Process Using Adhesive Tape, Proc. SPIE 2724 (1996) 677. [73] C.C. Lee, Method for Removing Sub Micron Particles from a Semiconductor Wafer Surface by Exposing the Wafer Surface to Clean Room Adhesive Tape Material, US Patent 5,690,749 (1997). [74] Y. Terada, E. Toyoda, M. Namikawa, A. Maekawa, T. Tokunaga, A Novel Method for Resist Removal after Etching of the Organic SOG Layer with the Use of Adhesive Tape, in Proc. IEEE Int. Symp. Semiconductor Manufacturing, pp. 295 298, IEEE, Piscataway, NJ (1999). [75] K. Sano, Future Evolution of Pressure Sensitive Adhesive Tapes, Nitto Technical Report 38 (December 2000) 11. [76] High performance Dust Preventive Adhesive Roll Cleaner SDR/ELEP F Series, Product Information, Nitto Denko Corporation, Toyohashi, Japan, 2009. Website: http://www.nitto. com/produtc/datasheet/process/011. [77] P. Lazzeri, G. Franco, M. Garozzo, C. Gerardi, E. Iacob, A. Lo Faro, A. Privitera, L. Vanzetti, M. Bersani, TOF SIMS Study of Adhesive Residuals on Device Contact Pads after Wafer Taping and Backgrinding, Appl. Surf. Sci. 203 204 (2003) 445. [78] M. Sutton, S.R. Burastero, J. Perkins, M.L. Chiarappa Zucca, B.D. Andresen, Alpha Aminobenzyl Alpha, Alpha, Diphosphotric Acid Selective Chelation of Beryllium, US Patent Application 20070254827 (November 2007). [79] L.O. Malotky, Hazardous Material Removal Using Strippable Coatings, US Patent 5,120,369 (1992). [80] H.N. Gray, D.E. Bergbreiter, Applications of Polymeric Smart Materials to Environmental Problems, Environ. Health Perspect. 105 (Suppl. 1) (1997) 55. [81] W. Feng, S.H. Patel, M. Y. Young, J.L. Zunino III, M. Xanthos, Smart Polymeric Coatings Recent Advances, Adv. Polym. Technol. 26 (2007) 1. [82] P. LaFrate Jr., J. Elliott, D. Siddoway, M. Valasquez, Decontamination and Size Reduction of Plutonium Contaminated Process Exhaust Ductwork and Glove Boxes, Report LA UR 97 254, Los Alamos National Laboratory, Los Alamos, NM, 1996.
220
Developments in Surface Contamination and Cleaning
[83] G. Sun, X.J. Xu, Durable and Regenerable Antibacterial Finishing of Fabrics: Biocidal Properties, Textile Chem. Colorist 30 (1998) 26. [84] G. Sun, X.J. Xu, Durable and Regenerable Antibacterial Finishing of Fabrics: Fabric Properties, Textile Chem. Colorist 31 (1999) 21. [85] G. Sun, X.J. Xu, Durable and Regenerable Antibacterial Finishing of Fabrics: Chemical Structures, Textile Chem. Colorist 31 (1999) 31. [86] G. Sun, X.J. Xu, J.R. Bickett, J.F. Williams, Durable and Regenerable Antibacterial Finishing of Fabrics with a New Hydantoin Derivative, Ind. Eng. Chem. Res. 40 (2001) 1016. [87] Y.Y. Sun, G. Sun, Durable and Regenerable Antimicrobial Textile Materials Prepared by a Continuous Grafting Process, J. Appl. Polym. Sci. 84 (2002) 1592. [88] S.D. Worley, J. Kim, C.I. Wei, J.F. Williams, J.R. Owens, J.D. Wander, A.M. Bargmeyer, M.E. Shirtliff, A Novel N halamine Monomer for Preparing Biocidal Polyurethane Coatings, Surf. Coatings Int. Part B: Coatings Trans. 86 (2003) 273. [89] M. Braun, Y.Y. Sun, Antimicrobial Polymers Containing Melamine Derivatives. I. Prepa ration and Characterization of Chloromelamine Based Cellulose, J. Polym. Sci. Part A: Polym. Chem. 42 (2004) 3818. [90] J.F. Williams, J.C. Suess, J.I. Santiago, T. Y. Chen, J. Wang, R. Wu, S.D. Worley, Anti microbial Properties of Novel N halamine Siloxane Coatings, Surf. Coatings Int. Part B: Coatings Trans. 88 (2005) 35. [91] J.F. Williams, U. Cho, Antimicrobial Functions for Synthetic Fibers: Recent Developments, American Association of Textile Chemists and Colorists (AATCC) Review 5 (2005) 17. [92] J.F. Williams, J.C. Suess, M.M. Cooper, J.I. Santiago, T. Y. Chen, C.D. Mackenzie, C. Fleiger, Antimicrobial Functionality of Healthcare Textiles: Current Needs, Options, and Characterization of N halamine Based Finishes, Res. J. Textiles and Apparel 10 (2006) 1. [93] L. Quian, T.Y. Chen, G. Sun, J.F. Williams, Durable and Regenerable Antimicrobial Textiles: Thermal Stability of Halamine Structures, American Association of Textile Chemists and Colorists (AATCC) Review 6 (2006) 55. [94] J. Liang, Y. Chen, K. Barnes, R. Wu, S.D. Worley, T. S. Huang, N halamine/quat Siloxane Copolymers for Use in Biocidal Coatings, Biomaterials 27 (2006) 2495. [95] G. Sun, S.D. Worley, Halamine Chemistry and its Application in Biocidal Textiles and Polymers, in: J.V. Edwards, G. Buschle Diller, S.C. Goheen (Eds.), Modified Fibers with Medical and Specialty Applications, Springer Verlag, New York, pp. 81 89. Chap. 6, 2006. [96] HaloSource, Inc, Bothell, WA, 2009. Website: http://www.halosource.com. [97] B.D. Veatch, Z. Ibrisagic, A.K. Kimball, T.E. Broderick, Electro Decontamination of Contaminated Surfaces, International Patent WO/2007/001263 (2007). [98] ElectroDeconÔ, The Electrochemical Strippable Coating System, ADA Technologies, Inc., Littleton, CO, 2009. Website: http://www.adatech.com. [99] E.G. Lierke, J. Bellenberg, G. Poß, Decontamination of Metal Surfaces by Thin Film Etching with Aerosols, Research Report, Battelle Institut eV, Frankfurt, Germany, April 1992. [100] I. Okano, Method of Removing Smear or Stain from a Coated Surface with an Aqueous Dispersion of a Highly Absorbent Polymer, US Patent 5,681,399 (1997). [101] H. P. Muller, H. Gruttmann, H. Casselmann, H. Muller, J. Petzoldt, M. Bock, Cosolvent Free Aqueous, Anionic Polyurethane Dispersions and Their Use as Peelable Coatings, US Patent 5,965,195 (1999). [102] H. P. Muller, H. Gruttmann, J. Petzoldt, H. Muller, Ch. Irle, Strippable Coating Composi tions, US Patent 6,903,156 (2005). [103] K. Yamashita, M. Matsuki, H. Asai, N. Matsuyama, H. Tojo, H. Kurota, K. Akasaka, H. Obara, Aqueous Dispersion of a Peelable Coating Composition, US Patent 6,344,236 (2002).
Strippable Coatings for Removal of Surface Contaminants
221
[104] M. Gotoh, S. Kobayashi, K. Kawabata, Pressure Sensitive Adhesive Composition and Pressure Sensitive Adhesive Tape or Sheet Making use of the Same, US Patent 5,286,781 (1994). [105] J.G. Southwick, K.S. Kiibler, J.R. Erickson, Non Aqueous Solvent Free Process for Making UV Curable Adhesives and Sealants from Epoxidized Monohydroxylated Diene Polymers, US Patent 5,776,998 (1998). [106] J.R. Erickson, J. Crivello, Non Aqueous Solvent Free Process for Making UV Curable Adhesives and Sealants from Epoxidized Monohydroxylated Diene Polymers (III), US Patent 5,837,749 (1998). [107] J.R. Erickson, D.R. Hansen, Formulation for Strippable Adhesive and Coating Films and High Performance Adhesives, US Patent 6,541,553 (2003). [108] D.L. Westerman, Strippable Coatings, Metal Finishing 98 (June 2000) 155. [109] M.B. Pepin, Care of Astronomical Telescopes and Accessories: A Manual for the Astro nomical Observer and Amateur Telescope Maker, Springer, London, 2004. [110] J. Cipriano, Cleaning Modern Antireflection Coated Optics with Collodion, Cloudy Nights Telescope Reviews (November 2005). Website: http://216.92.113.163/item.php?item id 1304. ˚ to 1190A ˚ by a Simple Open Air [111] J.D. Purcell, Reflectance Measurements from 1140A Method, J. Opt. Soc. Am. 43 (1953) 1166. [112] W.I. Kaye, Far Ultraviolet Spectroscopy. 1. Modification of the Beckman DK Spectro ˚ , Appl. photometer for Automatic Transmittance Recording in the Region 1700 2000 A Spectroscopy 15 (1961) 89. [113] J.B. McDaniel, Collodion Technique of Mirror Cleaning, Appl. Opt. 3 (1964) 152. [114] L.R. Canfield, G. Hass, J.E. Waylonis, Further Studies on MgF2 Overcoated Aluminum Mirrors with Highest Reflectance in the Vacuum Ultraviolet, Appl. Opt. 5 (1966) 45. [115] R.B. Gillette, B.A. Kenyon, Proton Induced Contaminant Film Effects on Ultraviolet Reflecting Mirrors, Appl. Opt. 10 (1971) 545. [116] A.N. Bloch, S.A. Rice, Reflections in a Pool of Mercury: An Experimental and Theoretical Study of the Interaction between Electromagnetic Radiation and a Liquid Metal, Phys. Rev. 185 (1969) 933. [117] J.T. Cox, G. Haas, J.B. Ramsey, W.R. Hunter, Reflectance of Evaporated Rhenium and ˚ , J. Opt. Soc. Am. 62 (1972) Tungsten Films in the Vacuum Ultraviolet from 300 to 2000 A 781. [118] J.B. Tyndall, Collodion Technique of Mirror Cleaning, Report LAR 10675, NASA Langley Research Center, Langley, VA, 1970. See also Brief 70 10463, NASA Tech Briefs (August 1970). [119] C.E. Miller, Standard Reflective Device, US Patent 4,346,996 (1982). [120] O. Kienzle, J. Staub, T. Tschudi, Description of an Integrated Scatter Instrument for Measuring Scatter Losses of ‘Superpolished’ Optical Surfaces, Measurement Sci. Technol. 5 (1994) 747. [121] J.R. Powell, D.M. Jackson, Cleaning of Optical Components, RGO/La Palma Technical Note No. 2, Royal Greenwich Observatory, Greenwich, London, UK, June 1984. [122] R. Ariail, The Artistry of Telescope Restoration A Professional Method for Cleaning Optics, J. Antique Telescope Soc. 6 (1995) 1008. See also Antique Telescope Society, website: http://www.webari.com/oldscope, 2009. [123] F. Henriquez, Cleaning the Optics. The UCLA 24 Inch Telescope. Website: http://frank.bol. ucla.edu/24inch.html#Cleaning the optics, October 2006. [124] R.M. Hamilton, Fast, Cost Effective Particle Removal for Anodic Bonding, University of California Berkeley, Berkeley, CA, October 1995. See MEMS website: http://mail.mems exchange.org/pipermail/mems talk/1995 October/000212.html. [125] T.M. Back, Cleaning Telescope Optics, NightTimes, Lake County Astronomical Society, Ingleside, IL, April 2004. Website: http://www.bpccs.com/lcas/Articles/cleanoptics.htm.
222
Developments in Surface Contamination and Cleaning
[126] W.D. Kimura, G.H. Kim, B. Balick, Comparison of Laser and CO2 Snow Cleaning of Astronomical Mirror Samples, in Proc. SPIE 2199 (1994) 1164 1171. [127] J. Fine, J.B. Pernick, Use of Strippable Coatings to Protect and Clean Optical Surfaces, Appl. Opt. 26 (1987) 3172. [128] W. Prewitt, How to Remove Dust from Your Mirror without Scratching. Website: http:// www.astrocaver.com/blue clean.html. [129] W. Davison, D. Ketelsen, R. Cordova, J. Williams, W. Omann, W. Kindred, Vacuum Lifting Equipment to Handle Polished Mirrors, Technical Memo UA 93 09, Large Binocular Telescope Project, University of Arizona, Tempe, AZ, 1993. [130] R. Robinson, Removing Surface Contaminants from Silicon Wafers to Facilitate EUV Optical Characterization, B.S. Thesis, Brigham Young University, Provo, UT, 2003. [131] J.M. Bennett, L. Mattsson, M.P. Keane, L. Karlsson, Test of Strip Coating Materials for Protecting Optics, Appl. Opt. 28 (1989) 1018. [132] J.M. Bennett, D. Ronnow, Test of Opticlean Strip Coating Material for Removing Surface Contamination, Appl. Opt. 39 (2000) 2737. [133] T.K. Moynagh, Cleaning Method for High Precision Molding Components, US Patent 5,810,941 (1998). [134] K. Okada, K. Kai, Atmospheric Mineral Particles Collected at Qira in the Taklamakan Desert, China, Atmos. Env. 38 (2004) 6927. [135] Air Particle Analyzer. QCM Cascade Particle Impactor, Technical Note, California Measurements, Inc., Sierra Madre, CA, 2009. [136] T.I. Suratwala, P.E. Miller, P.R. Ehrmann, R.A. Steele, Polishing Slurry Induced Surface Haze on Phosphate Laser Glasses, J. Noncryst. Solids 351 (2004) 2091. [137] Food and Drug Administration, US Department of Health and Human Services Part 358 Miscellaneous External Drug Products for Over the Counter Human Use Code of Federal Regulations, 21CFR358, Title 21, vol. 5 (1 April 2002). [138] O.G. Molina, Method of Magnetic Particle Testing Using Strippable Coatings, US Patent 3,951,881 (1976). [139] B.W. Ariss, C.R. Thomas, The Use of Coatings to Facilitate Decontamination, in: H.J. Blythe, A. Catherall, A. Cook, H. Well (Eds.), Proc. 1st Int. Symp. Decontamination of Nuclear Installations, Cambridge University Press, Cambridge, UK, pp. 55 64, 1967. [140] A.D. Turner, G. Worrall, J.T. Dalton, A Survey of Strippable and Tie Down Coatings for Use in the Decommissioning of Alpha Active Facilities, Report AERE R 12474, United Kingdom Atomic Energy Authority, Harwell Laboratory, Oxfordshire, UK, 1987. [141] H. Weichselgartner, Decontamination with Pasty Pickling Agents Forming a Strippable Coating, Report EUR 13498, Commission of the European Communities, Luxembourg, 1991. [142] Decommissioning Handbook, DOE/EM 0142P, Office of Environmental Restoration, US Department of Energy, Washington, DC, 1994. [143] IAEA Technical Report Series, State of the Art for Decontamination and Dismantling of Nuclear Facilities, TRS No. 395, International Atomic Energy Agency, Vienna, Austria, 1999. [144] J.L. Tripp, R.H. Demmer, R.L. Merservey, Decontamination, in: C.H. Oh (Ed.), Hazardous and Radioactive Waste Treatment Technologies Handbook, CRC Press, Boca Raton, FL, 2001. Chapter 8. [145] EPA Technology Reference Guide for Radiologically Contaminated Surfaces, EPA 402 R 06 003, US Environmental Protection Agency, Washington, DC, 2006. [146] F. Cumo, L. de Santoli, G. Guidi, LCA of Strippable Coatings and of Principal Competing Technology used in Nuclear Decommissioning, Chem. Eng. Trans. 7 (2005) 601. [147] Radioactive Waste Management and Advanced Nuclear Fuel Cycle Technologies, Progress Report 2006, FPN Radwaste Division, ENEA, Rome Frascati, Italy, 2006. [148] G. Guidi, F. Cumo, L. de Santoli, LCA of Strippable Coatings and of Steam Vacuum Technology Used for Nuclear Plants Decontamination, Clean Technols. Environ. Policy, Published online DOI: 10.1007/s10098 009 0208 5 (24 March 2009).
Strippable Coatings for Removal of Surface Contaminants
223
[149] D.M. Vogel, A.V. Cugini, Evaluation of a Polymer Based Chemical for Oxide/Scale Removal, Final Report EPRI NP 5278, Electric Power Research Institute, Palo Alto, CA, 1987. [150] A Strippable Coating to Reduce Surface Contamination, EPRI First Use Brochure RP2012 11, Electric Power Research Institute, Palo Alto, CA, 1989. [151] E.H. Sands, R. Kohli, Pentek Self Stripping Polymers for Decontamination of Radioactively Contaminated Surfaces, Report 5012 1166, Battelle Memorial Institute, Columbus, OH, 1991. [152] K.E. Archibald, R.L. Demmer, Tests Conducted with Strippable Coatings, Report INEEL/ EXT 99 00791, Idaho National Engineering Laboratory, Idaho Falls, ID, 1999. [153] R.L. Demmer, K.E. Archibald, J.H. Pao, M.D. Argyle, B.D. Veatch, A. Kimball, Modern Strippable Coating Methods, in WM’05, Proc. Waste Management Conference, R. Post (Ed.), Waste Management Society, Phoenix, AZ (2005). [154] H. L. Yin, Z. Y. Tan, Y. T. Liao, Yi. J. Feng, Application of SO24 /TiO2 Solid Superacid in Decontaminating Radioactive Pollutants, J. Env. Radioactivity 87 (2006) 227. [155] C. Laney, Decontamination of Reactor Cavity using Isolock 300, ALARA Note #8, Brookhaven National Laboratory, Brookhaven, NY, 2005. [156] J.W. Lee, M. Ahlen, M. Bruns, V. Fricke, C. May, J. Pickett, S. Salaymeh, Application of New and Innovative Technologies on the 321 M Fuel Fabrication Facility Large Scale Demonstration and Deployment Project, Report WSRC MS 98 00122, Office of Science and Technology, US Department of Energy, Savannah River, SC, 1998. [157] M.A. Ebadian, Assessment of Strippable Coatings for Decontamination and Decom missioning, Report DE FG21 95EW55094 32, Florida International University, Miami, FL, 1998. [158] V. Fricke, S. Madaris, C. May, ALARAÔ Strippable Coating, Report WSRC TR 99 00458, Office of Science and Technology, US Department of Energy, Savannah River, SC, 1999. [159] J. McFee, E. Stallings, M. Romero, K. Barbour, Improved Technologies for Decontami nation of Crated Large Metal Objects, in WM’02, Proc. Waste Management Conference 2002, R. Post (Ed.), Waste Management Society, Phoenix, AZ (2002). [160] R.R. James, Z.J. Willenberg, R.V. Fox, J. Drake, Bartlett Services Inc. Stripcoat TLC FreeÔ Radiological Decontamination Strippable Coating, Technology Evaluation Report EPA/600/ R 08/099, US Environmental Protection Agency, Cincinnati, OH, September 2008. [161] R.R. James, Z.J. Willenberg, R.V. Fox, J. Drake, Isotron Corp. OrionÔ Radiological Decontamination Strippable Coating, Technology Evaluation Report EPA/600/R 08/100, US Environmental Protection Agency, Cincinnati, OH, September 2008. [162] M. Tan, J.D. Whitaker, D.T. Schwartz, Simulation Study on the Use of Strippable Coatings for Radiocesium Decontamination of Concrete, J. Hazardous Mater. 162 (2009) 1111. [163] K.C. Holt, Testing for Radiological Decontamination Strippable Coating for Cellular Bioengineering, Inc. (Cs 137, Pu 239, Am 241), Letter Reports, Sandia National Labora tories, Albuquerque, NM, October 2007 and December 2007. [164] K.J. Walter, A.E. Draine, T.E. Johnson, Decontamination of a Fume Hood Contaminated with Tritiated Thymidine, Paper THAM C.4, Proc., 52nd Ann. Mtg. Health Physics Society, Portland, OR, 2007. [165] J.D. VanHorne Sealy, Evaluating the Efficiency of Decon Gel 1101 for Removal of Cs 137, Co 60, and Eu 154 on Common Commercial Construction Materials, M.S. Thesis, Oregon State University, Corvallis, OR, May 2008. [166] M. Sutton, R.P. Fischer, M.M. Thoet, M.P. O’Neill, G.J. Edgington, Plutonium Decon tamination Using CBI Decon Gel 1101 in Highly Contaminated and Unique Areas at LLNL, Report LLNL TR 404723, Lawrence Livermore National Laboratory, Livermore, CA, June 2008.
224
Developments in Surface Contamination and Cleaning
[167] A.E. Draine, T.E. Johnson, M.P. O’Neill, G.J. Edgington, K.J. Walter, Decontamination of Medical Isotopes from Hard Surfaces Using Peelable Polymer Based Decontamination Agents, Proc. 53rd Ann. Mtg. Health Physics Society, Philadelphia, PA (2008). [168] B.H. Davison, T. Kuritz, K.P. Vercruysse, C.K. McKeown, Green Biopolymers for Improved Decontamination of Metals from Surfaces: Sorptive Characterization and Coating Proper ties, Final Report, Project Number: EMSP 64907, Oak Ridge National Laboratory, Oak Ridge, TN, 2002. [169] K.G. Andersson, J. Roed, Removal of Radioactive Fallout from Surface of Soil and Grassed Surfaces Using Peelable Coatings, J. Env. Rad. 22 (1994) 197. [170] M.E. Lumia, C.A. Gentile, Industrial Hygiene Concerns During the Decontamination and Decommissioning of the Tokamak Fusion Test Reactor, Report PPPL 3647, Princeton Plasma Physics Laboratory, Princeton University, Princeton, NJ, 2002. [171] S. Svensson, Contamination Removal Process, US Patent 5,017,237 (1991). [172] R.V. Fox, R. Avci, G.S. Groenewold, Decontamination Materials, Methods for Removing Contaminant Matter from a Porous Material, and Systems and Strippable Coatings for Decontaminating Structures that Include Porous Material, US Patent Application 200702400733 (October 2007). [173] R.N. Miller, F.T. Humphrey, A. Bleich, Cleaning and Chemical Treatment of Aircraft Surfaces to Provide Optimum Cleaning Properties, Report ER 9703 8, Lockheed Georgia Company, Marietta, GA, 1970. [174] W.R. Drake, MIL C 6799, Strippable Coating Evaluation, Report AD 758184, Air Force Packaging Evaluation Agency, Wright Patterson Air Force Base, Dayton, OH, 1973. [175] G.P. Bierwagen, Next Generation of Aircraft Coatings Systems, J. Coatings Technol. 73 (April 2001) 45.
Chapter 6
Ultrasonic Cleaning S.B. Awad * and R. Nagarajan y *
Crest Ultrasonics Corporation, Trenton, NJ, USA
y
Indian Institute of Technology Madras, Chennai, India
1. Introduction 1.1. Particulate Contamination Effects in Microelectronic Manufacturing 2. Ultrasonic Cleaning 2.1. Ultrasonic Cleaning Equipment 2.2. Process Equipment Design 3. Principles of Ultrasonic Cleaning 3.1. Mechanism of Cavitation Erosion 3.2. Need for Measuring Surface Roughness 3.3. Structural Changes of Metallic Surfaces Induced by Ultrasound 4. Surface Cleanliness Measurement 4.1. Liquid Particle Counters 4.2. Precision Turbidity Meter 4.3. Microbalance 4.4. Cavitation Meter
5. Theory of Ultrasonic Cleaning 5.1. Theory of Ultrasonic Fields and their Effects on Immersed Surfaces 5.2. Contribution of Acoustic Streaming to Particle Removal from Immersed Surfaces 6. Experiments in Sonic Cleaning 6.1. Study of Ultrasonic Cleaning with Dual Frequencies 7. Cleaning Optimization 7.1. Effect of Substrate Material on Ultrasonic Cleanability and Erodibility 7.2. Erodibility Measurements Using a Microbalance 7.3. Cleanability and Erodibility Measurements Using an LPC 7.4. Effect of Frequency on Size of Particle Removed 7.5. Frequency Ranges Acknowledgements References
Developments in Surface Contamination and Cleaning Copyright Ó 2010 Elsevier Inc. All rights of reproduction in any form reserved.
225
226
Developments in Surface Contamination and Cleaning
1. INTRODUCTION Surface cleaning, by general definition, is freeing the surface from contaminants that are adhered chemically, physically or mechanically to that surface. Contaminants are soils or impurities either generated during the forming process of new surfaces or deposited as foreign matter from surrounding environments. Contaminants adhered to the surface under high mechanical pressure, by-products of chemical additives, and chemical protective films are common in metal-forming processes and are difficult to remove. Particles are insoluble individual or aggregates of micro-solid contaminants, which tenaciously adhere to the surface with various physical forces. Obviously, smaller sized particles are the most difficult to remove. Ultrasonic cleaning has been around for decades and is being used in applications ranging from microelectronics, hard-disk drives, medical devices, biomedical, automotive components, and optics to jewelry cleaning. Basically, ultrasonic energy is applied to a cleaning liquid or solution causing cavitation, which in turn scrubs the surface free from contaminants. However, there is a lot more involved beyond the basics of ultrasonic cleaning. Special attention must be given to a whole set of parameters in designing an effective cleaning process in order to achieve the high level of cleanliness, at the nano level, as demanded by continuous advancement in technology. This includes machine design and construction, sonic frequency, sonic power, proper and compatible cleaning chemistry, whether aqueous or solvent. Some other factors such as parts handling, process parameters, packaging, and maintenance are equally important and need to be included in the design of a cleaning process. Contaminants may be categorized as follows: a.
Organic contaminants. Examples include lubricating oils, cutting and machining fluids and oils, fingerprints, carbon, organic media in buffing and lapping compounds, waxes, silicone oils, mold release compounds, coolants, polymers, adhesives, photoresist compounds, lacquers, paints, inks, antifoam additives, and residual biocides. b. Inorganic contaminants. Examples include various metal oxides in buffing and lapping compounds, polishing compounds, inorganic salts, dust, metal fumes, slivers, and other metal oxides. c. Particulate contaminants. Examples include environmental debris, skin flakes, cosmetics, hair, in addition to others from the previous examples.
1.1. Particulate Contamination Effects in Microelectronic Manufacturing Particulate contamination has a significant effect on in-process yields, on final outgoing parts quality, and on reliability in customer applications of many microelectronic devices. In integrated circuit manufacture, particulate-related
Ultrasonic Cleaning
227
defects include circuit damage, pattern disruption, bridging of separated features, shorting of conductors, masking of areas that are supposed to be etched, and unmasking of areas that are not supposed to be etched [1]. In order to cause physical and mechanical defects, a common rule of thumb is that the particle must be at least as large as one-tenth the minimum circuit feature size 0.1 mm in the case of 1 mm feature sizes [2]. However, even smaller particles can cause ‘‘killer’’ defects via chemical interactions, or interactions with thin layers. Crystal-originated particles are known to cause a local reduction of the dielectric breakdown voltage of the insulating SiO2 layer in metal-oxide semiconductor (MOS) structures [3]. Complementary metal-oxide semiconductor (CMOS) on silicon-on-insulator (SOI) offers the potential for significant reduction in chip power without sacrificing performance, but is severely affected by many types of particle defects (metallic, oxide, and organic) in the 0.2 2 mm size range [4]. In thin film deposition processes such as sputtering and plasma-enhanced chemical vapor deposition, particle contamination causes direct yield loss or scrap due to point defects, as well as reliability failures by film failure during product use [5]. In hard-disk drives (HDDs), where head flying heights over the disk have dropped to the order of a few nanometers, particulate contamination has always played a central role in determining the quality and reliability of the drive [6]. Particles trapped between the head and disk can cause irrecoverable data losses in several ways head crash, circumferential scratching, embedding on the disk, thermal asperities, etc. Particles captured on the air-bearing surfaces of the slider can alter the trajectory of the flying head, causing it to move closer to the disk than designed (potential for hard contact), or farther from it (potential for loss of read/write signal) depending on how the flying dynamics are affected. Larger particles entrained in the airflow within the disk enclosure (DE) can damage the disks and/or the head by high-velocity impact. As particle concentration within the DE increases, the durability of the head disk interface decreases [7], with harder and larger particles causing earlier fails. Hence, the HDD industry has historically emphasized thorough cleaning of the technology components (such as disk and slider) and of the myriad mechanical components comprising a drive [8]. As a survey by Nagarajan [8] shows, sonication is the predominant mode of cleaning in the HDD industry, although it is used far more aggressively for the mechanical components than for the head and disk. The ceramic that the magnetorestrictive (MR) head is made of, and the glass substrate that is increasingly used for the data disk, are both fragile materials that can easily be damaged by high-energy ultrasonics. However, functional cleanliness requirements still render it imperative that submicron particles be removed completely from these surfaces, thus necessitating the use of high-frequency ultrasonics or >1 MHz ‘‘megasonics’’. Megasonics, in combination with strong chemistry, is also the method of
228
Developments in Surface Contamination and Cleaning
choice in cleaning semiconductor wafers [9]. Ultrasonic cleaners are used to remove large particles from wafers after sawing, and to remove lapping and polishing compounds after these operations [10]; however, their use continues to be limited by surface-damage considerations. From the above discussion and the references cited here, it is clear that particulate contamination is a major impactor of yield and reliability in microelectronic device manufacturing, and that ultrasonics is the preferred cleaning method in a majority of these applications. It is also apparent that the application of sonic fields to clean surfaces is not fully optimized due to legitimate concerns about the potential for surface and critical-feature damage.
2. ULTRASONIC CLEANING One of the most popular precision cleaning processes is ultrasonic cleaning [11]. Ultrasonic cleaning can be very effective, but a number of performance features of ultrasonic cleaning need consideration. The tendency of the ultrasonic energy to damage parts is a consideration in selection of frequency and power density in the cleaning tank, and can influence design of the equipment and design of the process [12]. Figure 6.1 shows how electric power is used in the cleaning process. Frequency and amplitude are properties of sound waves [13]. Figure 6.2 demonstrates frequency and amplitude using a spring model [13]. If A is the base sound wave, B, with less displacement of the media (less intense compression and rarefaction) as the wave front passes, represents a sound wave of less amplitude or ‘‘loudness’’. C represents a sound wave of higher frequency indicated by more wave fronts passing a given point within a given period of time.
FIGURE 6.1
Ultrasonic cleaning system (from www.crest ultrasonics.com).
229
Ultrasonic Cleaning
FIGURE 6.2
Frequency and amplitude in a spring model [13].
Frequency impacts include the following: Low-frequency ultrasonic cleaning (less than about 200 kHz) is relatively omnidirectional. As frequency increases above 200 kHz, ultrasonic cleaners tend to become more unidirectional. Very-low-frequency ultrasonic cleaning (less than 30 kHz) can produce subharmonics that can be heard by workers and can be a source of irritation. Very-low-frequency ultrasonic cleaners can result in severe damage to surfaces. Ultrasonic cleaners with frequencies in the 30 70 kHz frequency range tend to produce less mechanical damage to parts and operate more quietly than low-frequency ultrasonic cleaners. Ultrasonic cleaning involves the use of high-frequency sound waves in the range between 40 and 200 kHz (above the upper range of human hearing, or about 18 kHz) [14] to remove a variety of contaminants from parts immersed in aqueous media. The megasonic range encompasses frequencies above 800 kHz. Many industries require precision cleanliness in the micrometer to submicron particle size range e.g. semiconductor wafer fabrication [15], hard-disk drive manufacturing [6,8], and integrated circuit assembly. The contaminants can be dirt, oil, grease, buffing/polishing compounds, and mold release agents, to name just a few. Materials that can be cleaned include metals [16], glass, ceramics, etc. Ultrasonic agitation can be used with a variety of cleaning agents. Typical applications include removing chips and cutting oils from cutting and machining operations, removing buffing and polishing compounds prior to plating operations, and cleaning greases and sludge from rebuilt components for automotive and aircraft applications.
230
Developments in Surface Contamination and Cleaning
The use of ultrasonics in cleaning has become increasingly popular due to the restrictions on the use of chlorofluorocarbons such as 1,1,1-trichloroethane. Because of these restrictions, many manufacturers and surface finishers are using immersion cleaning techniques rather than solvent-based vapor degreasing. The use of ultrasonics enables the cleaning of intricately shaped parts with an effectiveness that exceeds that achieved by vapor degreasing. The efficacy of ultrasonic cleaning has been recognized in many industries; more recently, specially designed plants and tanks have been installed. Many companies have rejected conventional methods and have successfully adopted ultrasonic cleaning. They have discovered numerous advantages: 1. Costs of cleaning per component are greatly reduced since, although capital costs, solvent consumption and service charges rise, they are more than countered by lower labor costs. Figure 6.3 shows the savings achieved in cleaning printed circuit boards (PCBs) in one typical plant [17]. 2. Less time is needed to clean a component to any given standard. 3. Ultrasonic cleaning is more efficient than conventional methods in removal of contaminants (‘‘dirt’’). Figure 6.4 shows the results of a previous comparison [17]. 4. Components cleaned ultrasonically are more reliable and have a greater lifespan than those cleaned by other methods. The improved surface finish reduces wear and damaging friction. 5. Cost of materials is reduced since fewer parts have to be rejected through inefficient or damaging cleaning. 6. Productivity can be increased by the combination of these advantages.
FIGURE 6.3 The cost of cleaning printed circuit boards by hand using alcohol is compared with that of cleaning by ultrasonics [17].
Ultrasonic Cleaning
231
FIGURE 6.4 The efficiency of cleaning by ultrasonics is compared with the efficiency of cleaning by more conventional methods. (a) Ultrasonics. (b) Brush cleaning by hand. (c) Steam degreasing. (d) Forcible motion in fluid. (e) Spraying [17].
Ultrasonic cleaning has been used to great advantage for extremely tenacious deposits, such as corrosion deposits on metals. Cavitation forces can be controlled; thus, given proper selection of critical parameters, ultrasonics can be used successfully in virtually any cleaning application that requires removal of small particulates. High-intensity ultrasonic fields are known to exert powerful forces that are capable of eroding even the hardest surfaces. Quartz, silicon, and alumina, for example, can be etched by prolonged exposure to ultrasonic cavitation, and ‘‘cavitation burn’’ has been encountered following repeated cleaning of glass surfaces. The solid material itself does not affect the existence of cavitation. When cavitation exists, the resultant cavitation erosion is dependent on material properties such as hardness, work hardening capability, and grain size. Also, the stress state and corrosion resistance of a material affect the degree of cavitation erosion. Ultrasonic cleaning is powerful enough to remove strongly adhered (e.g. embedded from chemical mechanical planarization (CMP) process) contaminants yet gentle enough not to damage most substrates. It provides excellent penetration and cleaning in the smallest crevices and between tightly spaced parts in a cleaning tank. One of the major concerns with the usage of ultrasonics for cleaning purposes is cavitation erosion loss of surface material due to microscopic bubble implosion. The simultaneous processes of surface cleaning and of surface erosion in the presence of a high-frequency ultrasonic field (58 kHz) allow conceptual optimization of parametric settings to maximize the cleaning efficiency, even while minimizing the level of erosion damage [18]. Different frequency choices are available for cleaning: 20 40 kHz heavy-duty cleaning for items such as engine blocks and heavy metal parts, and for removal of heavy greasy soils.
232
Developments in Surface Contamination and Cleaning
40 70 kHz general cleaning of machine parts, optics, and other components. This frequency range is very good at removing small particles. 70 190 kHz gentle cleaning of optics, disk drive components, and other sensitive parts. 190 500 kHz ultrafine cleaning of semiconductor wafers, ultrathin ceramics, optics, and highly polished metallic mirrors or reflectors.
2.1. Ultrasonic Cleaning Equipment Ultrasonic cleaning systems are, in general, composed of ultrasonic transducers mounted on a radiating diaphragm, an electrical generator, and a tank filled with cleaning chemistry.
2.1.1. Ultrasonic Generators An ultrasonic generator energizes the transducers. The generator transforms the electrical energy from the power source into a suitable form for efficiently energizing the transducers at the desired frequencies. The generator produces an electronic signal of high voltage and sends it to the transducers. When the transducers receive the signal, they respond by changing shape as long as the signal is applied. The response range of the transducer determines the frequency of the generator. Since the response range of the transducer is narrow, the signal from the generator must be close to the response range of the transducer. The generator is designed to power a specific number of transducers. Each transducer requires a minimum amount of voltage to activate, usually about 75% of the maximum voltage of the transducer. A typical generator usually has controls for varying the amount of power to the transducers (power intensity control), and a built-in frequency sweep that will vary the frequency sweep rate (sweep frequency rate control) over the operation range of the transducer. A control that automatically turns the signal on and off very rapidly is sometimes provided to help degas the cleaning solution. The ultrasonic generator (Figure 6.5) converts a standard electrical frequency of 60 Hz into the high frequencies required in ultrasonic transmission, generally in the range of 20 80 kHz. 2.1.2. Transducer The transducer (Figure 6.6) is a piezo-ceramic material that changes shape instantly when excited by an electric signal. When excited by a high-frequency electronic generator, the transducer vibrates at resonant frequency and induces amplified vibrations of the diaphragm. This causes a series of compressions and rarefactions in the liquid in the tank. There are two types of ultrasonic transducers used in the industry: piezoelectric and magnetostrictive. Both
233
Ultrasonic Cleaning
FIGURE 6.5 Ultrasonic generators (single and dual frequency). Courtesy of Martin Walter/Crest Ultrasonics.
accomplish the same task of converting alternating electrical energy to vibratory mechanical energy, but do it by different means. In order to produce the positive and negative pressure waves in the aqueous medium, a mechanical vibrating device is required. Ultrasonic manufacturers have made use of a diaphragm attached to high-frequency transducers. The transducers, which vibrate at their resonant frequency due to a high-frequency electronic generator source, induce amplified vibration of the diaphragm. This amplified vibration is the source of positive and negative pressure waves that
FIGURE 6.6
Different types of ultrasonic transducers.
234
FIGURE 6.7 piezo.php).
Developments in Surface Contamination and Cleaning
Magnetostrictive
transducer
(http://www.blackstoneney.com/04.TP mag vs
propagate through the solution in the tank. When transmitted through water, these pressure waves create the cavitation process. The resonant frequency of the transducer determines the size and magnitude of the resonant bubbles. The lower frequency produces larger bubbles with more energy. The more powerful the cavitation process, the larger are the imploding bubbles. The higher the frequency, the less aggressive is cavitation and the smaller are the implosions. The lower frequencies are stronger because they concentrate the available power in localized regions. The higher the frequencies, the more evenly the available power is distributed throughout the tank volume. The tanks can be made to work at multiple frequencies by varying the frequency to the transducers, which shift them from their natural frequency. This only works to a limited degree as every transducer has a natural frequency at which it will resonate best. If they are shifted too far from the natural frequency, they will dissipate the power in the form of heat and will not generate cavitation bubbles. Any ultrasonic unit will generate more than one frequency; most units will generate sufficient high frequencies in addition to the fundamental low frequency. Higher frequencies, such as 130 200 kHz, generate much smaller cavitation bubbles. 2.1.2.1. Magnetostrictive Transducers Magnetostrictive transducers (Figure 6.7) utilize the principle of magnetostriction in which certain materials expand and contract when placed in an alternating magnetic field. Alternating electrical energy from the ultrasonic generator is first converted into an alternating magnetic field through the use of a coil of wire. The alternating magnetic field is then used to induce mechanical vibrations at the ultrasonic frequency in resonant strips of nickel or other ferromagnetic material, which are attached to the surface to be vibrated. Because magnetostrictive materials behave identically to a magnetic field of either polarity, the frequency of the electrical energy applied to the transducer is half that of the desired output frequency. Magnetostrictive transducers were the first to supply a robust source of ultrasonic vibrations for high-power applications such as ultrasonic cleaning. Because of inherent mechanical constraints on the physical size of the hardware, as well as electrical and magnetic complications, high-power
Ultrasonic Cleaning
235
magnetostrictive transducers seldom operate at frequencies much above 30 kHz. Piezoelectric transducers, on the other hand, can easily operate well into the megahertz range. Magnetostrictive transducers are generally less efficient than their piezoelectric counterparts. This is due primarily to the fact that the magnetostrictive transducer requires a dual energy conversion from electrical to magnetic and then from magnetic to mechanical. Some efficiency is lost in each conversion. Magnetic hysteresis effects also detract from the efficiency of the magnetostrictive transducer. 2.1.2.2. Piezoelectric Transducers Piezoelectric transducers (Figure 6.8) convert alternating electrical energy directly to mechanical energy through the use of the piezoelectric effect, in which certain materials change dimension when an electrical charge is applied to them. This phenomenon was discovered by Pierre Curie in 1883. The change is linear and proportional to the applied electrical energy. Electrical energy at the ultrasonic frequency is supplied to the transducer by the ultrasonic generator. This electrical energy is applied to piezoelectric element(s) in the transducer, which vibrate. These vibrations are amplified by the resonant masses of the transducer and directed into the liquid through the radiating plate. Early piezoelectric transducers utilized such piezoelectric materials as naturally occurring quartz crystals and barium titanate, which were fragile and unstable. Early piezoelectric transducers were therefore unreliable. Today’s transducers incorporate stronger, more efficient, and highly stable ceramic piezoelectric materials. The vast majority of transducers used today for ultrasonic cleaning utilize the piezoelectric effect. 2.1.2.3. Comparison of Piezoelectric and Magnetostrictive Transducers [19] In its simplest form, the control system for a sonic transducer of either type applies a tiny current for a very short duration, and then stops that current flow for an equivalent short duration so that the material can return to its original shape or dimension and the diaphragm can return to its original position.
FIGURE 6.8 php).
Piezoelectric transducer (http://www.blackstone ney.com/04.TP mag vs piezo.
236
Developments in Surface Contamination and Cleaning
The current causes the piezoelectric material to deform and changes the position of the attached diaphragm. The current passes through a wire coil, which generates a magnetic field that magnetizes the magnetostrictive material and changes the position of the attached diaphragm. Both types of materials cause diaphragms to vibrate. The surface to which the diaphragm is attached is also caused to vibrate. This is the container (housing) wall in which the transducer is housed. In use, the container is immersed into liquid. In other words, application of an electric current causes the housing wall immersed in liquid to vibrate so that pressure waves are spread within the liquid. 2.1.2.4. Immersible Transducers An immersible transducer is a radiating device sealed in a housing (usually stainless steel), the forward or front surface of which is the radiating surface, and which can be submerged under the surface of a liquid bath to energize the liquid to produce cavitation. An immersible transducer placed in a still tank turns that tank into an ultrasonic cleaner. The immersible transducer is, in effect, a standard tank inverted (turned inside out) with the radiating surface on the outside and the transducers on the inside, as can be seen from the diagrammatic sketch in Figure 6.9. Immersible transducers can have one transducer stack (radiator), or two or more stacks in a row or an array, depending on the size of tank to be energized and the power to be transmitted. The sketch does not show the method by which electrical energy is transmitted to the electrodes; it must go through or over the tank wall or up through the tank bottom and then through the housing (box), with all interfaces completely water/liquid-tight. A common way to simplify this is to use a piece of stainless steel conduit hanging over the rim of the tank; in that way only the entry into the housing (box) need be liquid-tight. The cable can also be in a flexible conduit, provided usually at additional cost.
TANK Base Ceramic Metal
PZT
PZT Metal or Ceramic
CONVENTIONAL FIGURE 6.9
CERAMIC (Patented) Immersible transducer (not to scale).
Ultrasonic Cleaning
237
2.1.3. Ultrasonic Tank The size of the ultrasonic tank (Figure 6.10) is dependent upon the parts being cleaned. An ultrasonic cleaner is simply a metal tank (stainless steel; Figure 6.10A) that has piezo-ceramic transducers bonded to the bottom (Figure 6.10B) or side. The transducers are coupled to an ultrasonic generator. Water is typically used as a cleaning liquid in these tanks. The overall effectiveness of the cleaning is dependent upon the cleaning liquid. A tank is considered active if it is fitted with transducers and can be activated to produce cavitation; it is considered a still tank if it has not (yet) been activated to produce cavitation or has not been fitted with ultrasonic transducers. A still tank can be activated by insertion of an immersible transducer into the bath. The more common method of inducing cavitation in a cleaning tank is to fasten transducers to the outer surface of the bottom or sides (or both) of the tank and thus to energize the inner surfaces of the tank, thereby transmitting ultrasonic vibrations into a liquid bath contained in the tank. Cavitation in the tank creates shock waves and cleans surfaces of parts and assemblies by accelerating detergency of cleaning agents in the bath and by mechanically blasting contaminants off the surfaces. There are generally two styles of such tanks: small, self-contained models primarily for home and laboratory use, and larger units consisting of two (or more) modules, a tank, and one or more generators, intended for industrial uses. A typical low-intensity laboratory-type ultrasonic cleaning tank will usually have a deep-drawn tank and may have one or more transducers bonded to the bottom or side wall of the tank to energize the wall or bottom as a diaphragm, passing vibrational energy through virtually unimpeded and cavitating the water or other liquid inside (Figure 6.11).
FIGURE 6.10 Ultrasonic tank. Courtesy of Crest Ultrasonics.
238
Developments in Surface Contamination and Cleaning
TANK WALL
LIQUID
ENERGY Bonding
ANTINODE (Front Face)
RADIATING SURFACE FRONT DRIVER
ELECTRODES NODE
IMMERSIBLE TRANSDUCER HOUSING (BOX)
CRYSTAL CRYSTAL BACK DRIVER
ANTINODE (Back Face) PRESTRESSING BOLT
FIGURE 6.11
TANK BOTTOM
Basic ultrasonic transducer assemblies.
2.2. Process Equipment Design Ultrasonic aqueous batch cleaning equipment consists of at least four stations: ultrasonic wash tank, minimum of two ultrasonic separate (or reverse cascading) water rinse tanks, and heated recirculated clean air for drying. Typically, ultrasonic transducers are bonded to the outside of the bottom surface or to the outside of the side walls, or both. For large-sized tanks, two types of transducers are provided as immersibles installed inside the tanks. The two types of immersibles are commercially available in various sizes and frequencies. The traditional sealed metal box contains a multi-transducer system and a cylindrical immersible that is powered by two transducers at both ends of a cylindrical metal rod (known as push pull) (Figure 6.12). The cleaning system is designed to fit the desired cleaning process and to deliver the required production throughput. Systems may include more stations, a full cleanroom enclosure, automation, operation safeguards, constant vertical oscillation, and sometimes a vacuum dryer (Figure 6.13). Tanks are normally made of 316L stainless steel. However, some tanks are made of special alloys or polymers. The tank surface may be electropolished for reduced surface roughness or physical vapor deposition (PVD) coated with TiN to preserve against erosion. Special coved corner tanks are made for semiconductor, pharmaceutical, and medical applications. For quick and
Ultrasonic Cleaning
239
FIGURE 6.12 Push pull transducer system: metal rods and mounted assembly. Courtesy of Martin Walter/Crest Ultrasonics.
FIGURE 6.13 Aqueous ultrasonic cleaning system. Courtesy of Crest Ultrasonics.
efficient particle removal, the tank is electropolished and has four-sided overflow (Figure 6.14). An effective cleaning process must first be developed, and then the number and size of the stations determined based on the required yield, total process time, and space limitation. Automation of the ultrasonic cleaning system is well established. Automation includes a computerized transport system able to run different processes for various parts simultaneously, with in situ data monitoring and acquisition. Advantages of automation are numerous, including consistency, achieving desired throughputs, and full control on process parameters. Typical tank size ranges from 10 to 2500 liters, based on the size of the parts, production throughput, and the required drying time. The whole machine can be enclosed to provide a cleanroom environment meeting ISO Class 7 down to
240
Developments in Surface Contamination and Cleaning
FIGURE 6.14 Electropolished stainless steel ultrasonic cleaning tank. Courtesy of Crest Ultrasonics.
ISO Class 5 cleanroom specifications. Process control and monitoring equipment consists of flow controls, chemical feed-pumps, in-line particle count, total organic carbon (TOC) measurement, pH, turbidity, conductivity, refractive index, and other parameters. The tanks are typically made of corrosion-resistant stainless steel or electropolished stainless steel. Titanium nitride or similar coating is used to extend the lifetime of the radiating surface in the tanks or the immersible transducers.
3. PRINCIPLES OF ULTRASONIC CLEANING Ultrasonic cleaners work by two principal mechanisms: 1. Cavitation 2. Acoustic streaming. Cavitation is a process where the constructive interference of sonic energy causes the formation of rarefiable bubbles in the cleaning liquid. When these microscopic bubbles implode (due to the passage of the rarefaction energy, the moving sound wave), they produce microscopic jets of liquid that can impinge on the surface of parts to be cleaned. These high-velocity jets remove particles from surfaces and convey cleaning chemicals to organic and inorganic chemical contamination on the surface. When acoustic waves constructively
241
Ultrasonic Cleaning
FIGURE 6.15
Ultrasonic cavitations and cleaning.
combine, the resulting decrease in pressure creates a localized bubble. In properly degassed solutions, this bubble almost entirely consists of solvent vapors. When the ultrasonic pressure wave goes through a compression cycle, the localized pressure drops and the bubble collapses. When this occurs, a microscopic jet of liquid is formed, jetting from the bubble wall into the volume of the bubble. This high-velocity jet scours the surface of parts it comes in contact with, knocking loose material from the surface (Figures 6.15 and 6.16). This cavitation action preferentially takes place at discontinuities on surfaces. Discontinuities can be scratches, pinholes in paint and pre-existent pits, among other features. For this reason, ultrasonic erosion, which is so common below 70 kHz, is almost always associated with these kinds of surface features. Cavitation is generated through at least three steps: nucleation, growth, and violent collapse or implosion [20]. The transient cavities (or vacuum bubbles or vapor voids), ranging from 50 to 150 mm in diameter at 25 kHz, are produced during the sound waves’ half-cycles. During the rarefaction phase of the sound wave, the liquid molecules are extended outward against and beyond the liquid
NUCLEUS
FIGURE 6.16
VIOLENT IMPLOSION
Growth and collapse (implosion) of a cavitation bubble.
242
Developments in Surface Contamination and Cleaning
natural physical elasticity/bonding/attraction forces, generating a vacuum nucleus that continues to grow. A violent collapse occurs during the compression phase of the wave. It is believed that the latter phase is augmented by the enthalpy of the fluid medium and the degree of mobility of the molecules, as well as by the hydrostatic pressure of the medium. Cavitations are generated in the order of microseconds. At the 20 kHz frequency, it is estimated that the pressure is about 35 70 MPa and the transient localized temperatures are about 5000 C, with the velocity of micro-streaming around 400 km h 1. In acoustic streaming, bulk movement of the liquid occurs. Contaminants that get removed from the surface are carried away by acoustic streaming, and hence are prevented from reattaching to the surface. Acoustic streaming can penetrate through the boundary layer of motionless liquid that surrounds all of the surfaces in the ultrasonic tank. Particles dislodged from the surface by cavitation action are not swept away from the surface and may get reattached. At high frequencies (>200 kHz), the acoustic streaming is highly directional, so the orientation of the part to be cleaned becomes critical. At low ultrasonic frequencies, the acoustic streaming is randomized and not highly directional. Cavitation and acoustic streaming work together in all forms of ultrasonic cleaning, but the relative contribution of each is a function of frequency. At low ultrasonic frequencies, cavitation is very strong and dominates the cleaning process. At high ultrasonic frequencies, cavitation bubbles are very small, but acoustic streaming velocities can be very high. Thus, at high frequencies acoustic streaming dominates the cleaning process and less cleaning occurs due to cavitation. Figure 6.17 shows a plot of cavitation strength versus frequency. It should be noted that the cavitation strength increases rapidly with decreasing frequency. Also, cavitation abundance (bubble density) with frequency is
FIGURE 6.17
Cavitation strength as a function of frequency.
243
Ultrasonic Cleaning
pictured in Figure 6.18 (note the decrease in bubble size with increase in bubble density). High-intensity ultrasonic waves create micro-vapor/vacuum bubbles in the liquid medium, which grow to maximum sizes proportional to the applied ultrasonic frequency and then implode, releasing their energies. The higher the frequency, the smaller is the cavitation bubble size. At 20 kHz the bubble size is roughly 170 mm in diameter (Figure 6.18). At a higher frequency of 68 kHz, the total time from nucleation to implosion is estimated to be about one-third of that at 25 kHz. At different frequencies, the minimum amount of energy required to produce ultrasonic cavities must be above the cavitation threshold. In other words, the ultrasonic waves must have enough pressure amplitude to overcome the natural molecular bonding forces and the natural elasticity of the liquid medium in order to grow the cavities. For water, at ambient, the minimum amounts of energy needed to be above the threshold were found to be about 0.3 and 0.5 W cm 2 (per transducer radiating surface) for 20 and 40 kHz respectively. The energy released from an implosion in close vicinity to the surface collides with and fragments or disintegrates the contaminants, allowing the detergent or the cleaning solvent to displace them at a very fast rate. The implosion also produces dynamic pressure waves, which carry the fragments away from the surface. The implosion is also accompanied by high-speed micro-streaming currents of the liquid molecules. The cumulative effect of millions of continuous tiny implosions in the liquid medium is what provides the necessary mechanical energy to break physically bonded contaminants, speed up the hydrolysis of chemically bonded contaminants, and enhance the
FIGURE 6.18
Cavitation abundance varies with frequency.
244
Developments in Surface Contamination and Cleaning
solubilization of ionic contaminants. The chemical composition of the medium is an important factor in enhancing the removal rate of various contaminants. The thickness of the boundary layer surrounding the parts is a function of the ultrasonic frequency in the tank. The higher the ultrasonic frequency, the thinner the boundary layer. This is illustrated in Figure 6.19, where the boundary layer thickness is plotted as a function of frequency. The boundary layer, next to the substrate surface where the sound does not penetrate, is essentially motionless. At 40 kHz, it is fairly thick at 2.8 mm where smaller particles can hide out. As frequency increases, the boundary layer is reduced, permitting the liquid to get closer to the surface and therefore the contaminants. For example, at 400 kHz the boundary layer is reduced to 0.98 mm. A typical cavitation bubble with the liquid jet a jet of liquid moving at extreme velocity, resulting from the asymmetrical implosion of the bubble in close proximity to the surface to be cleaned is clearly shown in the dramatic high-speed motion micrograph in Figure 6.20 [21]. Cavitation initiates most readily at, and proceeds radially outward from, discontinuities (voids, contaminant particles, etc.) in the liquid, where bonds between adjacent particles are weakest. Theoretically, a completely pure liquid (an unlikely happenstance) would be virtually impossible to cavitate. Pressure waves generated by high-frequency sound waves create micrometer-size bubbles in the aqueous solution. These micrometer-size bubbles form and grow due to the alternating positive and negative pressure waves in the solution. The bubbles subjected to these alternating pressure waves continue to grow until they reach resonant size (at which maximum or peak response amplitude occurs). Just prior to the bubble implosion, there is a tremendous amount of energy stored inside the bubble. Temperatures inside a cavitating bubble can reach 10,000 K, with pressures up to 50.7 MPa (~500 atm) [14]. The implosion event, when it occurs near a hard surface, changes the bubble into a jet
Thickness, micrometers
Boundary Layer Thickness as a Function of Frequency 5 4.5 4 3.5 3 2.5 2 1.5 1 0.5 0 0
200
400
600
800
Frequency, kilohertz FIGURE 6.19 The relationship between frequency and boundary layer thickness for room temperature water (theoretical simulation).
245
Ultrasonic Cleaning
FIGURE 6.20
Cavitation bubble (bubble diameter approximately 1 mm) [21].
one-tenth the bubble size, which travels at speeds up to 400 km h 1 toward the hard surface. These micro-jets can be harnessed for efficient heat and mass transfer. At very high frequencies approaching 1 MHz, cavitation becomes a secondary phenomenon compared to ‘‘acoustic streaming’’, which is the time-independent liquid motion generated by the sound field [22]. The associated streamlines are shown in Figure 6.21. Principally, this flow is categorized into two main types [22 26]: (1) streaming caused by spatial attenuation of the wave in free space, and (2) streaming caused by friction between the vibrating medium and the solid wall. The second mechanism is further classified as inner streaming, which is induced within the acoustic boundary layer, and outer streaming, which is the steady vortex flow developed outside the acoustic boundary layer. In order to
FIGURE 6.21
Liquid motion generated by a sound field [22].
246
Developments in Surface Contamination and Cleaning
produce the positive and negative pressure waves in the aqueous medium, a mechanical vibrating device is required. Ultrasonic manufacturers have made use of a diaphragm attached to high-frequency transducers. The transducers, which vibrate at their resonant frequencies due to a high-frequency electronic generator source, induce amplified vibration of the diaphragm. This amplified vibration is the source of positive and negative pressure waves that propagate through the solution in the tank. When transmitted through water, these pressure waves create the cavitation process. The resonant frequency of the transducer determines the size and magnitude of the resonant bubbles. Typically, ultrasonic transducers used in the cleaning industry range in frequency from 20 to 200 kHz. The lower frequency cleaners create larger bubbles with more energy, and will show a more aggressive dimple pattern during a foil test, while higher frequency cleaners will show less, if any at all. The more powerful the cavitation process, the larger will be the imploding bubbles. The higher the frequency, the less aggressive the cavitation and the smaller the implosions. It should follow that for large parts and areas with heavy contamination levels, lower frequency is better for cleaning. And, for small parts with tiny blind holes and intricacies that contain small particles, the higher the frequency, the better the cleaning result. However, this is not always the case. Removing a contaminant from a substrate requires a certain amount of power, and a higher frequency system may not be able to create this level of energy. Cavitation and its power levels are affected by several variables: ultrasonic frequency; time; chemistry and concentration; load size; contamination level; power density; detergent or solvent type; part geometry; basket/rack configuration; cavitation uniformity; bath temperature; part material; contamination type; and bath filtration [27]. At lower frequencies, such as 20 kHz, cavitation erosion will eventually, over a long period of time, eat though the bottom of the tank. If the part has a smooth soft surface, it can also be eroded by this effect. These effects are most prominent with lower frequencies. Higher frequencies will also cause cavitation erosion, but it will take a relatively long cleaning time to see the effects. By adjusting the power and frequency of an ultrasonic system to an optimum level, one can avoid damage to the part where it cleans without eroding it. The higher the frequency, the more evenly spread out is the power. This produces more uniform cleaning on the parts. Higher frequencies also produce smaller cavitation bubbles and thus can remove smaller particles. Damage can also occur when the part is extremely fragile and it is placed in a position in the tank in such a way that part of the object is in an area of compression and part is in an area of rarefaction. This is more evident at the lower frequencies (20 40 kHz). For this reason, most delicate parts are cleaned in a high-frequency ultrasonic tank (70 200 kHz). By distributing the total energy of the tank over a greater number of energy peaks, the overall effect is to create a very homogeneous power distribution and subject the part to an even distribution of energy.
Ultrasonic Cleaning
247
3.1. Mechanism of Cavitation Erosion Visual observations of ultrasonic fields indicate that cavities rarely exist in isolation rather, they are present in the form of clusters; thus, the dynamics of a collapsing cavity are influenced by the dynamics of surrounding cavities. These cavitation forces lead to surface damages and deformations. During the last few years, the idea of an effect from the concerted collapse of cavity clusters has been put forward to explain the damage capability of the cluster. Two models for the effect of concerted collapse have been proposed: one assumes that the shock waves from individual cavity collapses are superposed to form a single high-intensity damaging shock wave; the other assumes that the collective collapse increases the pressure at which the last cavities collapse and that the liquid jets formed by asymmetrical collapse of the cavities close to the solid are the source of damage. According to the first idea [28], the concerted collapse is expected to develop as follows. When the first cavities in a cluster collapse, the emitted shock waves trigger the collapse of the other cavities. The shock waves from individual cavity collapses are expected to catch up to form a single highintensity shock wave, which preferably is directed towards the specimen surface. The directionality is ascribed to the increase of ambient pressure that starts at the outer cavities and initiates the collapse of the cluster. This model is based on measurements in vibratory cavitation of pressure pulses on a surface directly exposed to cavitation and on comparison of the measured peak pressures with the corresponding values of incubation period and mass loss [29,30]. In cavitation erosion studies, large-scale surface deformations (compared with the cavity diameters) are observed and the idea of an extended high-intensity shock wave as the source of erosion has its main advantage in being able to explain this large-scale deformation topography (craters and undulation). Meanwhile the formation of a single regular shock wave is questionable. It is well known that each cavity emits a shock wave at collapse, but the formation of a single shock wave by collapse of the whole cluster of cavities would require that the cavities collapse synchronously throughout the cluster. Further, even a small volume fraction of non-collapsed cavities would attenuate a shock wave very strongly. Experiments done by Brunton in 1979 [31] support this view that no single-intensity shock wave results from the collapse of a cavity cluster. The other idea [31 34] is based on transfer of energy from the cavities that collapse first, to the cavities not yet collapsed. Physically, the collapse of a cavity cluster is initiated at the cluster boundary by the hydrostatic pressure in the ambient liquid, and the collapse proceeds from the boundary towards the center of the cluster. The collapse of the outer cavities creates a field of increased pressure around the remaining part of the cluster, the inward radiated energy from the collapsed cavities being transferred into collapse energy of the other cavities. Thus the damage potential of the individual cavities increases
248
Developments in Surface Contamination and Cleaning
toward the center of the cluster, where the pressure increases far above the ambient pressure, and the jet velocities and the corresponding jet impact pressures become very high. This model then predicts stronger erosion close to the center of the cavity clusters with the damaging mechanism being the jet impacts (and shock waves) from single cavity collapses. The idea of energy transfer by concerted collapse explains directly the smaller (10 100 mm) indentations observed at the beginning of exposure to cavitation. The above-mentioned larger undulations are not developed until later in the erosion process. Though they may result from the buildup in time and space of effects of single cavity collapses, a main drawback of this model is that it does not comprehensively explain these larger deformations. However, the focusing of the erosion to the area at the center of the cluster is in direct support of this theory. In another theory, the concept of cluster collapse (versus collapse of single cavities) has been presented on the basis of the authors’ assertion [35,36] that experimental evidence of erosion patterns and direct visual observations favor the cluster hypothesis. In their hypothesis, the cluster collapse pressure depends weakly on frequency and is strongly dependent on the initial and final cluster radii, as well as on the cubic root of ultrasound intensity. In the more conventional single-bubble collapse framework, the energy contained within the bubble just before its collapse, WCav, is given by: WCav ¼ 4p=3$P0 $r3
(1)
where P0 is the hydrostatic liquid pressure outside the bubble and r is the maximum bubble radius before collapse, related to the frequency of ultrasound by the expression: r ¼ rresonance ¼ ½1=ð2pf Þ$½3kP0 =r1=2
(2)
where k is the polytropic index [37], f is the frequency of ultrasound, and r is the medium density.
3.2. Need for Measuring Surface Roughness Surface roughness is one of the parameters that affects the lifespan of integrated circuit (IC) chips in precision engineering and the semiconductor industry. Hence, it is important to monitor and control the surface roughness of a semiconductor wafer, as it is a basic component used in IC fabrication. Another need for measuring surface roughness is its use in surface polishing in silicon wafer planarization. Chemical mechanical polishing (CMP) is currently the most popular method for IC wafer planarization.
3.3. Structural Changes of Metallic Surfaces Induced by Ultrasound Cavitation damages on metallic plates oscillating at 20 kHz were studied with aluminum plates [38] directly bonded to the end of the ultrasonic horn. In this
Ultrasonic Cleaning
249
paper, the effects of ultrasound on several metallic surfaces have been reported. Cubic face-centered metals (Al, Ag, and Cu) and a hexagonal close-packed metal (Zn) were studied. The collapse of the cavitation bubbles induced structural and morphological surface changes studied by X-ray diffraction (XRD), scanning electron microscopy (SEM), and roughness measurements. The energy brought by the ultrasounds allows a modification of the general orientation of aluminum crystallites at the metal surface. The structural reorganization within crystallites causes small dimensional variations and weakens grains cohesion [38]. This transformation is due to the mechanical effect of ultrasounds that selectively tear off the metal grains whose orientation is such that the maximum density planes are not perpendicular to ultrasonic flow. The susceptibility of materials to ultrasonic damage is complex. Mechanical damage to delicate structures has been studied earlier [38]. Examples include breakage of wire bonds, delamination of adhesive bonds, and changes to shapes of fine metal parts. Some strategies have been implemented to try to minimize the damage from ultrasonic cleaners. The theory behind these is that the damage occurs because of standing waves in the ultrasonic tank that stay in fixed locations with respect to the parts being cleaned. So, the following techniques are followed: One approach is to vary the frequency of the ultrasonic field. The frequency cannot be varied far from the center of resonance, otherwise the power level in the ultrasonic tank drops markedly. For example, in the ‘‘sweep’’ frequency mode of operation, when ultrasonic energy is operated at a resonant frequency of 47 kHz, the sweep is less than 2 kHz. A second technique has been to move the parts slowly in the tank during the cleaning process, a procedure called ‘‘undulation’’. Undulation is often used together with low sweep frequency to reduce damage. A final consideration is the damage that occurs when parts are drawn through the liquid air interface. The energy density at the interface is higher than within the bulk of the cleaning liquid. Drawing a damage-sensitive part through the liquid air interface while the ultrasonic field is activated can result in severe part damage. Thus, many processes are designed to turn off the ultrasonic energy as parts are passing through the interface: this is often referred to as employing the ‘‘quiet interface’’.
4. SURFACE CLEANLINESS MEASUREMENT 4.1. Liquid Particle Counters Liquid particle counters (LPCs) are used to determine the size and number of particles suspended in liquids. Liquid particle counters can be used to test the quality of drinking water or cleaning solutions, or the cleanliness of power generation equipment, manufactured parts or injectable drugs.
250
Developments in Surface Contamination and Cleaning
This instrument utilizes the principle of ‘‘near angle light scattering’’, as shown in Figure 6.22, and consists of a basic light source, such as a laser diode (wavelength 670.8 nm). The beam from the laser diode is spatially filtered and focused by a lens assembly to form a small and well-defined illuminated volume within the liquid being inspected. As the illuminated volume moves across a particle suspended in the liquid, some light from the beam will be scattered. Much of this scattered light is in the near-forward direction and is collected by the optical system of the photodetector assembly. The amplitude and width of this pulse are a function of the size of the particles. The amount of light scattered by a particle in the sensitive zone of the optical system is a function of the scattering angle and the relative index of refraction of the particle. This instrument collects and averages light that has been scattered in a near-forward direction over a solid angle ranging from 4 to 19 . Variation of collected light is 15% on a single reading of one particle count. This instrument typically detects particles from 0.5 to 100 mm. Bottle Absorbing Scattered Light from Particle Co ector Lens
Unscattered Laser Beam Particle Sensitive Zone
1.5 cm
FIGURE 6.22
2 cm
Scanning Laser Beam
Optical principle of a liquid particle counter.
When particles are extracted ultrasonically from immersed component surfaces into a liquid medium, the LPC can be used to indirectly quantify surface contamination levels by counting and sizing the extracted particles suspended in the liquid. In the case of complex substrates, such as hard-disk drive components, this may be the only practical option to measure surface cleanliness [39].
4.2. Precision Turbidity Meter An important water quality indicator for almost any use is the presence of dispersed, suspended solids particles not in true solution and often including silt, clay, algae and other microorganisms, organic matter, and other minute
Ultrasonic Cleaning
251
particles. Suspended solids obstruct the transmittance of light through a water sample and impart a qualitative characteristic, known as turbidity, to water. The American Public Health Association (APHA) defines turbidity as an ‘‘expression of the optical property that causes light to be scattered and absorbed rather than transmitted in straight lines through the sample’’. Turbidity can be interpreted as a measure of the relative clarity of water. Turbidity is not a direct measure of suspended particles in water but, instead, a measure of the scattering effect such particles have on light. Very simply, the optical property expressed as turbidity is due to the interaction between light and suspended particles in water. A directed beam of light remains relatively undisturbed when transmitted through absolutely pure water, but even the molecules in a pure liquid will scatter light to a certain degree. Therefore, no solution will have zero turbidity. In samples containing suspended solids, the manner in which the sample interferes with light transmittance is related to the size, shape, and composition of the particles in the solution, as well as to the wavelength (color) of the incident light. A minute particle interacts with incident light by absorbing the light energy and then, as if acting as a point light source itself, re-radiates the light energy in all directions. This omnidirectional re-radiation constitutes the ‘‘scattering’’ of the incident light. The spatial distribution of scattered light depends on the ratio of particle size to wavelength of incident light. Particles much smaller than the wavelength of incident light exhibit a fairly symmetrical scattering distribution with approximately equal amounts of light scattered both forwards and backwards (Figure 6.23A) [40]. As particle sizes increase in relation to wavelength, light scattered from different points of the sample particle create interference patterns that are additive in the forward direction. This constructive interference results in forward-scattered light of a higher intensity than light scattered in other directions (Figure 6.23B, C). In addition, smaller particles scatter shorter (blue) wavelengths more intensely while having little effect on longer (red) wavelengths. Conversely, larger particles scatter long wavelengths more readily than they scatter short wavelengths of light. While most earlier turbidimetric methods measured the transmitted light, turbidity measurement standards changed in the 1970s, when the nephelometric turbidimeter was developed, which determines turbidity by the light scattered at an angle of 90 from the incident beam (Figure 6.24). A 90 detection angle is considered to be the least sensitive to variations in particle size. Nephelometry is a preferred means for measuring turbidity because of the method’s sensitivity, precision, and applicability over a wide range of particle size and concentration. Turbidity is commonly expressed in nephelometric turbidity units (NTU). As in the case of surface cleanliness measurement using extraction followed by LPC, here too extraction followed by turbidimetry can provide quantification of surface particulate levels.
252
Developments in Surface Contamination and Cleaning
FIGURE 6.23
Angular patterns of scattered intensity from particles of three sizes [40].
FIGURE 6.24
Schematic diagram of a nephelometer.
4.3. Microbalance A microbalance may be used to quantify mass loss from a coupon by cavitation erosion. The Cahn C-34/C-35Ô microbalance is one such sensitive weight and force measurement instrument. It is designed for weights and forces up to 3.5 g and is sensitive to changes as small as 0.1 mg. The balance may be described as a force-to-current converter. It consists of: (1) a balance beam mounted to, supported by, and pivoting about the center of a taut ribbon; (2) a torque motor
253
Ultrasonic Cleaning
coil located in a permanent magnetic field and also mounted to the taut ribbon; (3) sample suspension fixtures; (4) a beam position sensor system; and (5) controls, circuitry, and indicators (Figure 6.25). Weights or forces to be measured are applied to the sample (left) side of the beam, which produces a force about the axis of rotation. An electric current flowing in the torque motor also produces a force about the same axis that is equal and opposite to the force from the beam, if the beam is at the beam reference position. This reference position is detected by the beam position sensing system. A greater force on the beam will require a greater opposite force from the torque motor in order to keep the beam at its reference position. Therefore, the current necessary to produce the required torque motor force is a direct measure of the force on the beam. The process of calibration allows this current to be measured in units of weight (grams). In order to measure mass loss due to ultrasonic cleaning, the material coupon is first cleaned with pure water and dried in the oven so that it loses all its moisture and is weighed using the microbalance before immersing in a beaker full of water. This beaker is then suspended at the center of the tank using a fixture to hold it. The ultrasonic generator is switched on and the power level is adjusted. Experiments may be done to simulate even mild cavitation conditions (132 192 kHz) and low power inputs. After a certain time (say, 2 minutes) the generator is turned off. The specimen is then taken out and dried in the oven. When the specimen is dry, it is taken out and reweighed using the microbalance. This step is repeated in a multiple-extraction mode.
Controls Circuits Indicators
Photocell
LED
Magnet
Flag
Tare loop
Balance Beam
“B” loop “A” loop
Magnet
Sample
FIGURE 6.25
Torque Motor Coil
Rotation at Axis of Beam & Coil Assembly
Taut Band Suspension and Torque Motor Coil Connection
Schematic diagram of the Cahn microbalance.
254
Developments in Surface Contamination and Cleaning
4.4. Cavitation Meter The ultrasonic cavitation meter (ppb-500Ô) is an instrument used to measure the energy density (in watts per unit area) of cavitation in liquids (Figure 6.26). It is not a sound meter or hydrophone. The main difference is that it measures cavitation or the collapse of water as it implodes on a surface, instead of sound waves produced by a pressure transducer. The ultrasonic meter is simple and easy to use, yet it contains sophisticated electronics and options for data storage, retrieval, and analysis. The meter measures the instantaneous energy in a given direction. The probe is a 51 cm (20 inches) long stainless steel tube with an ethylene propylene diene monomer (EPDM) half sphere on one end and a cable on the other. The black half-sphere is made of an elastic material to isolate the filter lens mounted on it from the holding rod. The lens is a thick quartz crystal.
FIGURE 6.26
Cavitation meter.
Cavitation generated by the sound pressure waves is produced in the form of bubbles that grow and implode with micro-streaming water jets hitting the filter surface. The sensor mounted behind the lens detects these impacts and the signal is sent via cable to the electronic case.
5. THEORY OF ULTRASONIC CLEANING 5.1. Theory of Ultrasonic Fields and their Effects on Immersed Surfaces When high-intensity ultrasonic waves are propagated in a liquid, cavitation bubbles form at sites of rarefaction (where local pressure is negative with
255
Ultrasonic Cleaning
respect to vapor pressure). The cavity is filled with vapor from the surrounding liquid, and with dissolved gases. The bubbles collapse readily at the next compression, creating shock waves. Visual observations of ultrasonic fields indicate that cavities rarely exist in isolation, rather they are present in the form of clusters; thus, the dynamics of a collapsing cavity are influenced by the dynamics of surrounding cavities [35]. Based on this concept, Gogate and Pandit [36] have developed a correlation for the collapse pressure, PC (N m 2): Pc ¼ C1 ðIÞa ðr0 Þb ðf Þc ðgÞd ðr=r0 Þe
(3)
where C1 is a proportionality constant, I is the intensity of ultrasound (W m 2), r0 is the initial radius of a cluster (m), f is the frequency of ultrasound (Hz), r is the radius of cluster at collapse conditions (m), and g is the fraction of energy transferred into the cluster, which is generated due to the collapse of the individual cavities on the outer boundary. The values of the constant and exponents have been obtained by Gogate and Pandit [36] by fitting the data obtained from simulations, and are reproduced in Table 6.1. These values are stated by the authors to be valid over the following range of parameters: intensity of irradiation ¼ 50 250 104 W m 2, initial cluster radius ¼ 0.5 5 mm, frequency of ultrasound ¼ 20 300 kHz, fraction of energy transfer ¼ 0.25 0.45, and individual cavity radius ¼ 1 mm. Taking typical values to be I ¼ 100 104 W m 2, r0 ¼ 1 mm, f ¼ 58 kHz, g ¼ 0.35, (r/r0) ¼ 0.001 yields a value of PC of 1013 MPa (approximately 10,000 atm). Instantaneous pressures of this magnitude have indeed been confirmed via the use of sonochemistry and sonoluminescence [37]. It may be reasonably inferred that the collapse pressure directly affects the extent of surface impact by cavitation forces, and thereby influences any induced erosive effects. The instantaneous cluster wall velocity at the time of cluster collapse, SC (m s 1), is calculated by Kanthale et al. [35] as: SC ¼ ½PC =ðrbð1 bÞÞ1=2
(4)
TABLE 6.1 Values of Constant and Exponents in Equation (3) [36] Parameter
Value
C1
0.0159
a
0.33
b
0.134
c
0.042
d
0.634
e
2.969
256
Developments in Surface Contamination and Cleaning
where r is medium density and b is the void fraction in the cluster. A stated limitation of this model is that ‘‘acoustic streaming’’ i.e. medium streaming induced by pressure gradient along the ultrasonic beam has been neglected (in order to simplify the model by neglecting frictional forces between the liquid and cavity cluster); this aspect will be dealt with in the next section. The velocity, SC, is a direct contributor to contaminant particle dislodgment from the surface, via the mechanism of imparting a shear velocity to the particle. The shear stress aspect is dealt with extensively by Maisonhaute et al. [41]. Acoustic bubbles typically oscillate at distances of only a few tens of nanometers from the immersed surface. The flow velocities resulting from bubble-cluster collapse lead to drag and shear forces on surfaces and adhered particles. The surface tangential stress, stan,C, is given by: stan;C ¼ h dv=dx
(5)
where h is the liquid viscosity, v is the liquid velocity, and x is the coordinate perpendicular to the surface. For a thin layer of thickness x0, the maximum value of s (associated with cluster collapse in the immediate vicinity of the surface) can be approximated as: stan;C ¼ h SC =x0
(6)
where SC, as defined earlier, is the (unattenuated) cluster collapse velocity. This can be of the order of hundreds of meters per second in a high-intensity ultrasonic field. This velocity and the associated velocity gradient give rise to lift forces and drag forces of particle dislodgment that are countered by particle adhesion forces. Assuming that the global removal stress must exceed the attractive force in order to remove a spherical particle from a surface, a predictive parameter, aC, for cavitational particle removability has been defined as [41]: aC ¼ stan;C RP =ð1:5 WA Þ
(7)
where RP is the radius of the particle and WA is the work of adhesion, defined as the sum of the surface energies of the two contacting materials minus the interfacial energy. Combining equation (7) with equations (4) and (6) yields: aC ¼ hx0 RP ½PC =ðrbð1 bÞÞ1=2 =ð1:5 WA Þ
(8)
where PC can again be estimated from equation (3). The pressure exerted by the collapsing cavity cluster can also lead to cavitational erosion of the surface, with the onset and magnitude of surface erosion being primarily dependent on the material and on surface finish/quality. Bedkowski et al. [42] have shown a correlation between fatigue strength of the material under random loading and its hydrodynamic cavitation erosion resistance, based on results of tests on three steels; properties of eroded surfaces
257
Ultrasonic Cleaning
have been observed to be similar to characteristics of fatigue fracture. By extension to ultrasonic cavitation involving bubble clusters, the following relationship may be proposed: sEROS ¼ B$PC
(9)
where sEROS is the erosion stress associated with the sonic field and B is an ‘‘erosion susceptibility’’ constant for the specific immersed surface. The constant, B, incorporates material and surface properties that influence cavitation erosion in an ultrasonic field. Among these, the hardness of the material, its fragility, and the roughness of the surface are considered to be the most influential. For a soft, fragile, rough surface, B would be nearly unity; for a hard, ductile, smooth surface, B would be a very small value. Taking diamond and glass as reference materials for hardness and fragility respectively, and assuming that a surface with arithmetic roughness value Ra of 0.1 mm may be considered smooth for cavitation erosion purposes, an expression for B may be derived as follows: B ¼ 1 ½ðHm =HD ÞðFG;min =Fm Þð1=ð10Ra ÞÞ
(10)
where the subscript ‘‘m’’ represents the material of the surface being cleaned, the subscripts D and G refer to diamond and glass respectively, H denotes hardness (measured on the Knoop scale; on this scale, diamond has a value of 7000), F stands for fragility, and Ra is expressed in mm. The definition of a fragility parameter and a table of fragility parameters for several metallic alloys and glasses are provided by Zhao et al. [43]; the values range from 44 for Zr-alloy glass formers to 238 for Al-alloy glass formers. The lowest measured value (FG,min) is about 16. A higher parameter indicates greater fragility, and a reference maximum value of about 250 may be estimated for the most fragile glass. Since surface erosion is a mechanism that generates particles and cleaning is a process that removes particles, the net rate of particle removal from a sonicated surface may be represented as the difference between the erosive and surface-cleaning contributions due to the sonic field. Alternatively, the difference between the surface-erosion and particle-shearing stresses associated with cavitation, i.e. stan,C sEROS, may be treated as an index representing net cleanability of the surface associated with the collapsing cavity cluster. Such a net cleaning force exerted on the surface per unit area may then be written as: sClean;Cav ¼ ðstan;C sEROS Þ
(11)
sClean;Cav ¼ h=x0 ½PC =ðrbð1 bÞÞ1=2 BPC
(12)
Or:
258
Developments in Surface Contamination and Cleaning
This can be rewritten in the form: 1=2
sClean;Cav ¼ APC BPC
(13)
A ¼ ðh=x0 Þ=½ðrbð1 bÞÞ1=2
(14)
where:
This leads to the interesting result that the net zero cleaning condition, where particle removal and particle generation phenomena are in perfect balance, would prevail when the cluster collapse pressure, PC, satisfies the equality: PC ¼ C1 ðIÞa ðr0 Þb ðf Þc ðgÞd ðr=r0 Þe ¼ ðA=BÞ2
(15)
For a given combination of ultrasonic field, liquid properties, and surface/ material properties, equation (15) provides a mathematical criterion for zero net cleaning conditions. Equation (13) may also be used to determine optimum cleaning conditions, with respect to cluster collapse pressure, that maximize the difference between particle-removal and particle-generation stresses. Mathematically, this may be done by taking the differential (dsClean,Cav/dPC) and setting the resulting expression equal to zero; the maximum is proven by then taking the second differential, and verifying that it is less than zero (concavity condition in optimization of a function). PC,opt obtained in this manner is given by: PC;opt ¼ ½A=ð2BÞ2
(16)
The corresponding optimal cavitation-based cleaning stress is given by: sClean;Cav;opt ¼ A2 =4B
(17)
Equations (16) and (17) clearly do not apply to the special case of B ¼ 0, i.e. a hypothetical case of an absolutely non-erodible material and surface; this special case may be treated by reverting to equation (13), which would yield 1=2 a cleaning stress that scales as PC . In general, according to the predictions of this integrated model of surface erosion and surface particle removal, the optimum cluster collapse pressure increases rapidly (square function) with decreasing B, i.e. decreasing erosion sensitivity, and less rapidly (linear relationship) with increasing A. Table 6.2 provides the dependence of the optimum cluster collapse pressure, and the optimal cleaning stress on critical parameters of the liquid/surface system.
5.2. Contribution of Acoustic Streaming to Particle Removal from Immersed Surfaces Acoustic waves that propagate in liquids obey the general laws of hydrodynamics. Nyborg [22] solved the Navier Stokes equation for a Newtonian liquid using second-order approximations. Markham [23] showed that streaming was due to sound absorption and relaxation processes. In principle [21], constant
259
Ultrasonic Cleaning
TABLE 6.2 Parametric Dependencies of Optimal Cavitational Cleaning Conditions Optimized Liquid parameter viscosity Cluster collapse pressure
h2
Cleaning force
h2
(H0
Hm/HD; F0
FG,min/Fm; R0a
Liquid density
Boundary layer thickness
Void fraction in cluster
1/r
1=x20
1/[b(1
1=x20
1/[b(1
1/r
Hardness (H), fragility (F), roughness (R)
b)] 1=ð1 þ H0 Þ2 ðF 0 Þ2 ðR0a Þ2 2H 0 F 0 R0a
b)] 1=½1
H 0 F 0 R0a
1=ð10Ra Þ); see equation (8).
streaming occurs at all acoustic radiation fields, and increases until the intensity is reduced by beam divergence and/or attenuation in the medium. Tjotta [25] introduced a simple formula in which the streaming velocity, SAc, was proportional to the absorption coefficient, a, the beam width 2a, and acoustic intensity I, as well as inversely proportional to medium viscosity h and sound velocity in the medium c: SAc ¼ 8aIa2 =ðhcÞ
(18)
The absorption coefficient, a, varies as the square of frequency, f, according to the following expression: a ¼ 2hf 2 =ð3rc3 Þ
(19)
Combining equations (18) and (19) eliminates the dependence of acoustic streaming velocity on liquid viscosity: SAc ¼ 16Ia2 f 2 =ð3rc4 Þ
(20)
This acoustic streaming velocity is a significant contributor to the shearing force that acts to dislodge particles from exposed surfaces. The total shearing velocity, S, that the particle experiences may be written as: S ¼ SC þ SAc
(21)
(assuming the contributions to particle shear from cavitational cluster collapse and acoustic streaming). Thus, the dependence of total particle removal velocity on the relevant acoustic force-field parameters may be stated as: S ¼ ðC1 Þ1=2 ðIÞa=2 ðr0 Þb=2 ðf Þc=2 ðgÞd=2 ðr=r0 Þe=2 ½ðrbð1 bÞÞ1=2 þ 8aIa2 =ðhcÞ (22) The estimated values for C1, a, b, c, d, and e are presented in Table 6.1.
260
Developments in Surface Contamination and Cleaning
The corresponding total acoustically induced tangential particle removal shear, stan, may then also be written as a sum of the contributions due to cluster collapse, stan,C, and due to acoustic streaming, stan,Ac: stan ¼ ðh=x0 ÞðSÞ
(23)
Substituting for S from equation (22) into equation (23) enables the determination of parametric dependencies of stan. It is important to note that, unlike the shear stress due to a collapsing cavity cluster, the shear stress imparted by acoustic streaming is non-erosive, and only serves to enhance particle depletion from surfaces. Thus, the net cleaning force (per unit area) may now be written as: sClean ¼ ðstan sEROS Þ
(24)
Or, following equations (13) and (14): 1=2
sClean ¼ APC BPC þ 8aIa2 =ðx0 cÞ
(25)
Comparing this to equation (13), it is clear that optimal value of cluster collapse pressure is unaffected by the addition of the acoustic streaming term; however, the optimal cleaning stress is now augmented by the contribution of streaming shear stress according to equation (25). The concept of cluster collapse (namely, collapse of single cavities) results in a correlation for cluster collapse pressure that is weakly dependent on frequency, and strongly dependent on the initial and final cluster radii, as well as on the cubic root of ultrasound intensity. In the more conventional singlebubble collapse framework, the energy contained within the bubble just before its collapse was given by equations (1) and (2). For an aqueous solution exposed to 40 kHz ultrasonics, this gives a bubble radius of 0.08 mm; the smallest cluster size in the cavity cluster model is taken to be 0.5 mm, i.e. approximately the combined size of six to eight bubbles. Single-bubble collapse pressure is clearly dominated by its radius, and hence by the frequency of ultrasound (inverse cubic dependence); conversely, the cluster collapse model predicts a very weak dependence on the frequency. Systematic experimentation is required, and planned, to assess the role of frequency in microelectronic component surface cleaning applications. The stress exerted on an immersed surface by a collapsing cavity or cluster provides a shear force to remove surface-adhered particles, as well as a direct erosive stress on the surface that generates particles from parent material. While particle-removal shear stress scales as the square root of PC, the erosive stress is taken to scale linearly with PC. While the validity of the linear model is also amenable to further exploration, the current formulation does indicate that increasing cavity-collapse pressure will accelerate surface-erosion rates faster than particle-removal rates, except for highly non-erodible materials and surface coatings, plating or finishes. Hence, it is recommended that PC not be allowed to exceed the optimum value obtained from equation (16). As per
Ultrasonic Cleaning
261
Table 6.2, this optimum value increases as the square of liquid viscosity, and decreases with increasing liquid density (inverse linear dependence), boundary layer thickness (inverse squared), void fraction in the cluster, and ‘‘erosion susceptibility’’ parameter (squared). When trying to ultrasonically clean a soft, fragile, rough surface, it is advisable to keep PC well below its optimum value by careful selection of parameters such as ultrasonic intensity and frequency, the liquid medium and its viscoelastic properties, directness of coupling between the acoustic field and the surface being cleaned. Acoustic streaming, which gains in importance as cavitation intensity is reduced, has few detrimental effects on the immersed surface. Thus, its incorporation in the surface cleaning/erosion model does not alter the setting of optimum cavitation pressure, but does enhance the particle-removal velocity at the surface. Since the streaming velocity depends on the square of frequency, it is particularly significant in high-frequency ultrasonic fields (170 kHz), and even more so in the megasonic range (1 MHz). However, it should be borne in mind that in the case of particles that are strongly adhered to surfaces, acoustic streaming alone cannot be relied upon to dislodge and remove the particles. Typically, a staged cleaning system with lower-frequency ultrasonics in the front end to loosen the particles, followed by higher-frequency ultrasonics to flush the loosened particles away from the surface would be required in order to optimize surface cleaning.
6. EXPERIMENTS IN SONIC CLEANING The cleanability and erodibility associated with an acoustic field may be evaluated via a ‘‘multiple-extraction’’ procedure [12], i.e. the substrate to be cleaned is immersed in a liquid medium (e.g. pure water, or water with a small volume percent of surfactant), and subjected to the ultrasonic field conditions under study. Surface-adhered particles will be removed and suspended in the liquid due to the cleaning action, and parent-material particles will be dislodged by erosion and also suspended in the liquid. A turbidimeter or liquid-borne particle counter may then be used to measure extracted particle concentration in the liquid. This is an indirect measure of the prevailing particle population on the surface. The same substrate is then re-immersed in a fresh liquid medium, and re-extracted. This ‘‘multiple-extraction’’ procedure is repeated until an asymptotic particle concentration level in liquid (representing surface erodibility) is reached. The initial slope of this multiple-extraction curve is indicative of the cleanability associated with the acoustic field. The ratio of these two parameters cleanability/erodibility is defined as the cleanability index, and when it is at a maximum, the cleaning process is fully optimized. Alternatively, cleaning efficiency may be evaluated by quantifying particles on a surface prior to and after a cleaning step. A well-behaved (asymptotic) turbidity time plot for multiple ultrasonic extraction is displayed in Figure 6.27. Initially, the surface carries many loose
262
Developments in Surface Contamination and Cleaning
Turbidity (NTU)
1.2 1 0.8 0.6 0.4 0.2 0 30
60
90
120
150
180
210
360
600
900 1200
Time (s) FIGURE 6.27
Typical turbidity time plot [12].
0.0075
0.025
0.007
0.02
0.0065
0.015
0.006
0.01
0.0055
0.005
0.005
0 40
60
80
Cleanability (per sec)
Erodibility (NTU/cm2)
as well as bonded foreign particles. The high initial turbidity value in Figure 6.27, expressed in nephelometric turbidity units (NTU), reflects the same. The other foreign particles (relatively strongly attached to the surface) are removed in subsequent runs. Turbidity values gradually decrease after first extraction, which shows that surface is depleted in easily removable particles. Figure 6.28 shows a plot of erodibility (defined as turbidity per unit surface area) and cleanability (defined as reduction in turbidity per second) of a material at a given frequency for various power levels. As input power decreases, erodibility or erosion of material surface also decreases. So, to achieve low erosion, low power may seem a plausible solution. But it also shows that low power lowers the rate of removal of particles, i.e. cleanability decreases. Thus, a ‘‘cleanability index’’, defined as the ratio of cleanability to erodibility, may be used for sonocleaning process optimization. Figure 6.29 shows how the cleanability index varies with power level. It shows a maximum at 80%. Hence, one should operate at 80% power (i.e. 400 W) when cleaning aluminum at 58 kHz. Figure 6.30 shows that lower frequency results in higher erosion but also offers higher cleanability. Erodibility and cleanability are affected by power and frequency to different extents. Figure 6.31 shows the cleanability index
erodibility cleanability
100
Power level (% of total=500W) FIGURE 6.28 Erodibility cleanability power plot for aluminum sheet at 58 kHz (laboratory data from the Indian Institute of Technology (IIT) Madras).
263
Cleanability index (cm2/NTU/s)
Ultrasonic Cleaning
3.2 3 2.8 2.6 2.4 2.2 40
60
80
100
Power level (% of total=500W) FIGURE 6.29 Cleanability index power plot for aluminum sheet at 58 kHz (laboratory data from IIT Madras).
Cleanability index (cm2/NTU/s)
FIGURE 6.30 Erodibility cleanability frequency plot for aluminum sheet at 100% power level (laboratory data from IIT Madras).
8 7 6 5 4 3 2 1 0 58
132
172
192
Frequency (kHz) FIGURE 6.31 Cleanability index frequency plot for aluminum sheet at 100% power level (laboratory data from IIT Madras).
264
Developments in Surface Contamination and Cleaning
increases monotonically with increase in frequency. Hence, it is better to clean aluminum parts, in general, at higher frequencies. This would be true for most soft metals that have a tendency to erode.
6.1. Study of Ultrasonic Cleaning with Dual Frequencies When two frequencies are used simultaneously to sonicate a sample, it is termed as ‘‘dual-frequency sonication’’. It has been found in theory [41] that dual frequency effects are more than the sum of individual frequency effects. The effect of dual frequency mode of operation on the cleanability index (CI) is discussed here.
6.1.1. 172/192 kHz Combination Experiments were conducted on aluminum sheet to study the effect of using dual frequency on ultrasonic cleaning, and the results thus obtained were contrasted with those of single frequencies. The charts given below show the effect of dual frequency on different cleaning parameters like cleanability, erodibility, and CI. Figure 6.32 shows that erodibility value is higher for the combination (172/192 kHz) compared to either of two frequencies, which is understandable because two frequencies are simultaneously at work. Figure 6.33 shows the same effect for cleanability. Cleanability is highly improved in a combinational run. Removal of particles occurs at a higher rate in dual-frequency mode. Erodibility value in the combinational run is higher than the algebraic sum of the erodibilities from individual frequencies (Figure 6.34). Furthermore, CI is highest for 192 kHz (Figure 6.35). The following are the conclusions from the study of ultrasonic cleaning with dual frequency that combines two high frequencies (172 and 192 kHz): 1. Ultrasonic cleaning with dual frequency gives a higher value of both cleanability and erodibility for the substrate. 0.0018
Erodibility (NTU/cm2)
0.0016 0.0014 0.0012 0.0010 0.0008 0.0006 0.0004 0.0002 0.0000 192/172
172
192
Frequency (kHz) FIGURE 6.32 Erodibility frequency plot for aluminum sheet at 100% power level (laboratory data from IIT Madras).
265
Ultrasonic Cleaning
0.0090
Cleanability (per sec)
0.0080 0.0070 0.0060 0.0050 0.0040 0.0030 0.0020 0.0010 0.0000 192/172
172
192
Frequency (kHz) FIGURE 6.33 Cleanability frequency plot for aluminum sheet at 100% power level (laboratory data from IIT Madras).
2. The increase in erodibility is more than the sum of the erodibilities for the single frequencies. 3. CI for 192 kHz is higher than the other two frequencies. 4. Dual high frequency does not perform better (in terms of CI) than individual frequencies.
6.1.2. 192/58 kHz Combination These particular results were achieved with a combination of high frequencies. In order to further explore the dual-frequency effect, a different combination was studied. A combination of low and high frequency was taken, i.e. 192 and 58 kHz, and the above analysis was repeated. 0.0018
Erodibility (NTU/cm2)
0.0016 0.0014 0.0012 0.0010 0.0008 0.0006 0.0004 0.0002 0.0000 192+172
192/172
Frequency (kHz) FIGURE 6.34 Erodibility plot for aluminum sheet at 192/172 kHz combination vs. 192 þ 172 kHz and 100% power level (laboratory data from IIT Madras).
266
Developments in Surface Contamination and Cleaning
Cleanability index (cm2/NTU/s)
8 7 6 5 4 3 2 1 0 192/172
172
192
Frequency (kHz) FIGURE 6.35 Cleanability index frequency plot for aluminum sheet at 100% power level (laboratory data from IIT Madras).
The following observations are made from Figures 6.36 6.39: Cleanability is enhanced in a combinational run. Erodibility in the combinational run is lesser than that in the 58 kHz case (192 kHz provides a gentle effect to 58 kHz). CI is greatly improved for the 192/58 kHz combination.
7. CLEANING OPTIMIZATION 7.1. Effect of Substrate Material on Ultrasonic Cleanability and Erodibility Cleanability and erodibility both appear to increase when sonication is done at low frequencies. But the ratio of cleanability to erodibility does not necessarily 0.0080
Erodibility (NTU/cm2)
0.0070 0.0060 0.0050 0.0040 0.0030 0.0020 0.0010 0.0000 192/58
58
192
Frequency (kHz) FIGURE 6.36 Erodibility frequency plot for aluminum sheet at 192/58 kHz combination and 100% power level (laboratory data from IIT Madras).
267
Ultrasonic Cleaning
0.060
Cleanability (per sec)
0.050 0.040 0.030 0.020 0.010 0.000 192
58
192/58
Frequency (kHz) FIGURE 6.37 Cleanability frequency plot for aluminum sheet at 192/58 kHz combination and 100% power level (laboratory data from IIT Madras). 0.0080
Erodibility (NTU/cm2)
0.0070 0.0060 0.0050 0.0040 0.0030 0.0020 0.0010 0.0000 192+58
192/58
Frequency (kHz) FIGURE 6.38 Erodibility plot for aluminum sheet at 192/58 kHz combination vs. 192 þ 58 kHz and 100% power level (laboratory data from IIT Madras).
follow suit. This is due to the fact that erodibility does not increase by the same factor as does cleanability. Thus, there presumably exists an optimum frequency for cleaning a particular material in an ultrasonic field. This optimum corresponds to a maximum in the CI.
7.1.1. Aluminum Aluminum metal was taken in the form of a rectangular sheet (5 cm 6 cm). It can be seen that both cleanability and erodibility decrease with higher frequencies with the exception of the 192/58 kHz case (Figures 6.40 and 6.41). This dual frequency was also seen as an exception in the experiments
268
Developments in Surface Contamination and Cleaning
Cleanability index (cm2/NTU/s)
12 10 8 6 4 2 0 192/58
58
192
Frequency (kHz) FIGURE 6.39 Cleanability index frequency plot for aluminum sheet at 192/58 kHz combina tion vs. 192 þ 58 kHz and 100% power level (laboratory data from IIT Madras).
conducted with alumina and silicon. This reduction of erodibility may be attributed to the moderating effect of the 192 kHz frequency on the more erosive 58 kHz. Overall, for aluminum, lowest CI occurs at around 58 kHz (Figure 6.42).
7.1.2. Alumina Alumina, which is a ceramic, was taken in the form of a slab (2 cm 2 cm 0.55 cm). Alumina shows similar trends to aluminum in terms of erodibility and 0.008
Erodibility (NTU/cm2)
0.007 0.006 0.005 0.004 0.003 0.002 0.001 0.000 192/58
58
132
172/192
172
192
Frequency (kHz) FIGURE 6.40 Erodibility frequency plot for aluminum sheet at 100% power level (laboratory data from IIT Madras).
269
Ultrasonic Cleaning
Cleanability (per sec)
0.060 0.050 0.040 0.030 0.020 0.010 0.000 192/58
58
132
172/192
172
192
Frequency (kHz)
Cleanability index (cm2/NTU/s)
FIGURE 6.41 Cleanability frequency plot for aluminum sheet at 100% power level (laboratory data from IIT Madras). 12.00 10.00 8.00 6.00 4.00 2.00 0.00 192/58
58
132
172/192
172
192
Frequency (kHz) FIGURE 6.42 Cleanability index frequency plot for aluminum sheet at 100% power level (laboratory data from IIT Madras).
Erodibility (NTU/cm2)
0.300 0.250 0.200 0.150 0.100 0.050 0.000 192/58
58
132
172/192
172
192
Frequency (kHz) FIGURE 6.43 IIT Madras).
Erodibility frequency plot for alumina at 100% power level (laboratory data from
270
Developments in Surface Contamination and Cleaning
0.450
Cleanability (per sec)
0.400 0.350 0.300 0.250 0.200 0.150 0.100 0.050 0.000 192/58
58
132
172/192
172
192
Frequency (kHz) FIGURE 6.44 Cleanability frequency plot for alumina at 100% power level (laboratory data from IIT Madras).
cleanability (Figures 6.43 and 6.44), but the CI (Figure 6.45) is higher for more intense cleaning conditions, i.e. 192/58 kHz combination.
7.1.3. Silicone Rubber Silicone rubber was available in a tablet form (12 mm diameter, 1 mm thick). Since the size was small, two tablets were used at a time. Silicone rubber too has decreasing erodibility (Figure 6.46) and cleanability (Figure 6.47) with increasing frequencies as observed with the other substrates. It has high CI (Figure 6.48) at 192/58 and 172 kHz.
Cleanability index (cm2/NTU/s)
2.00 1.80 1.60 1.40 1.20 1.00 0.80 0.60 0.40 0.20 0.00 192/58
58
132
172/192
172
192
Frequency (kHz) FIGURE 6.45 Cleanability index frequency plot for alumina at 100% power level (laboratory data from IIT Madras).
271
Ultrasonic Cleaning
0.009
Erodibility (NTU/cm2)
0.008 0.007 0.006 0.005 0.004 0.003 0.002 0.001 0.000 192/58
58
132
172/192
172
192
Frequency (kHz) FIGURE 6.46 Erodibility frequency plot for silicone rubber at 100% power level (laboratory data from IIT Madras).
0.01 0.009
Cleanability (per sec)
0.008 0.007 0.006 0.005 0.004 0.003 0.002 0.001 0 192/58
58
132
172/192
172
192
Frequency (kHz) FIGURE 6.47 Cleanability frequency plot for silicone rubber at 100% power level (laboratory data from IIT Madras).
7.1.4. Polished Silicon Wafer The silicon wafer used in the study, a circular disc of 10 cm diameter, was polished on one side and etched on the other side. The polished side was manually contaminated with road dust and was kept to dry overnight. Experiments were then conducted with the contaminated side (polished) facing the transducers.
272
Cleanability index (cm2/NTU/s)
Developments in Surface Contamination and Cleaning
1.80 1.60 1.40 1.20 1.00 0.80 0.60 0.40 0.20 0.00 192/58
58
132
172/192
172
192
Frequency (kHz) FIGURE 6.48 Cleanability index frequency plot for silicone rubber at 100% power level (laboratory data from IIT Madras).
0.00016
Erodibility
(NTU/cm2)
0.00014 0.00012 0.00010 0.00008 0.00006 0.00004 0.00002 0.00000 192/58
58
132
192
Frequency (kHz) FIGURE 6.49 Erodibility frequency plot for silicon wafer at 100% power level (laboratory data from IIT Madras).
The cleanability and erodibility results obtained are qualitatively similar to those obtained for the other three substrates (Figures 6.49 and 6.50), but the CI (Figure 6.51) for polished silicon wafer reaches a maximum at around 132 kHz. A correlation of the form given below may be formulated from the above data: CI ¼ kðCaIÞa ðHÞb ðFracture toughnessÞg ðRa Þd
(26)
where the quantities are defined below: Cleanability index (CI) is the dimensionless ratio of cleanability and erodibility. Cavitation intensity (CaI) is a measure of the intensity of cavitation erosion occurring at a point in an ultrasonic field. It is measured in watts per square inch by a ppbÔ cavitation probe.
273
Ultrasonic Cleaning
Cleanability (per sec)
0.0025
0.002
0.0015
0.001
0.0005
0 192/58
58
132
192
Frequency (kHz)
Cleanability index (cm2/NTU/s)
FIGURE 6.50 Cleanability frequency plot for silicon wafer at 100% power level (laboratory data from IIT Madras).
25 20 15 10 5 0 192/58
58
132
192
Frequency (kHz) FIGURE 6.51 Cleanability index frequency plot for silicon wafer (100% power) (laboratory data from IIT Madras).
Hardness (H) is the characteristic of a solid material expressing its resistance to permanent deformation. It can be expressed on the Mhos scale. Fracture toughness is a property (measured in MPa m1/2), which describes the ability of a material to resist mechanical failure by fracture. Roughness (Ra) is the average height of the bumps on a surface, measured in micrometers. MatlabÔ software was used for curve fitting to determine the parameters, and the computed values are given below: k ¼ 9.287 10 a ¼ 0.383
4
274
Developments in Surface Contamination and Cleaning
Erosion vs. time curve Weight loss from unpolished silicon wafer (mg)
0.08 0.07 0.06 0.05 0.04 0.03 0.02
192 kHz 172 kHz 132 kHz
0.01 0 0
2
4
6
8
10
12
14
Time (minutes) FIGURE 6.52 Erosion vs. time curve for unpolished silicon wafer at 132, 172, and 192 kHz. (see colour plate section at end for coloured version)
b ¼ 2.630 g ¼ 1.407 d ¼ 3.9123. The cleanability index increases with cavitation intensity and fracture toughness, and decreases with increasing hardness and roughness.
7.2. Erodibility Measurements Using a Microbalance As described earlier, a precision microbalance with 0.1 mg sensitivity may be used to quantify erosive mass loss from materials exposed to high-frequency (mildly cavitational) ultrasonic cleaning, as evident from Figure 6.52. Such data can be collected for various materials and used to establish correlation between material properties (such as tensile strength, hardness, and Young’s modulus) and their erosion susceptibility (Figures 6.53 6.55). Based on these data, the following relationships may be established for erodibility (E): Ef1=T 0:5 Ef1=H 2 Ef1=Y where T is the tensile strength and Y is Young’s modulus. Such parametric dependencies can be valuable in determining a priori optimum selection of materials for a given ultrasonic cleaning process or, conversely, designing a non-erosive cleaning process for the material set at hand.
275
Ultrasonic Cleaning
Erosion vs. tensile strength
Log (erosion in milligrams per gram of material)
4.5 4
Rubber
3.5 3 2.5 Silicon
2 Aluminum
1.5
y = 0.5394 + 7.6795
1
Glass
0.5 0 0
2
4
6
8
10
12
14
Log (tensile strength in psi) FIGURE 6.53
Log log plot of erosion vs. tensile strength for different materials.
Erosion vs. hardness Log (erosion in milligrams per gram of material)
4.5 4
Rubber
3.5 3 2.5 Silicon
2
Aluminum
1.5
Glass
y = 1.8688x + 6.7067
1 0.5 0 0
0.5
1
1.5
2
2.5
3
3.5
Log (hardness) FIGURE 6.54
Log log plot of erosion vs. hardness for different materials.
Erosion vs. Young’s modulus Log (erosion in milligram per gram of material)
4.5 Rubber
4 3.5 3 2.5 2
Glass Aluminum y = 0.9943x + 6.382
1.5 1
Silicon
0.5 0 0
1
2
3
4
5
6
Log (Young’s modulus in GPa) FIGURE 6.55
Log log plot of erosion vs. Young’s modulus for different materials.
276
Developments in Surface Contamination and Cleaning
33
Cleanability (/min)
31 29 27 25 23 21 19 17 15 50
60
70
80
90
100
110
Power level (%) 40 kHz
58 kHz
192 kHz
430 kHz
58+192 kHz
FIGURE 6.56 Cleanability power plot for silicon at various frequencies. (see colour plate section at end for coloured version)
7.3. Cleanability and Erodibility Measurements Using an LPC The following data (Figures 6.56 6.58; data obtained in the laboratories at IIT Madras) were obtained using a 0.5 mm Spectrex PC-2200Ô LPC, by exposing a polished silicon wafer substrate to various cleaning conditions. The 430 kHz frequency has the lowest cleanability and erodibility values (Figures 6.56 and 6.57). But it has the highest cleanability index value at all power levels (Figure 6.58). The dual-frequency combination showed the highest cleanability value (Figure 6.56). Its erodibility is intermediate between that of 58 and 192 kHz. The optimal choice for silion wafer cleaning is to operate at
Erodibility (per cm2)
240 220 200 180 160 140 120 100 50
60
70
80
90
100
110
Power level (%) 40 kHz
58 kHz
192 kHz
430 kHz
58+192 kHz
FIGURE 6.57 Erodibility power plot for silicon at various frequencies. (see colour plate section at end for coloured version)
277
Ultrasonic Cleaning
Cleanability index (cm2/min)
15 14 13 60% 70% 80% 90% 100%
12 11 10 40
58
192
430
58+192
Frequency (kHz) FIGURE 6.58 Cleanability index frequency plots for silicon at various power levels. (see colour plate section at end for coloured version)
100% power level with frequencies in the 430 470 kHz range (Figure 6.58). The CI values either decrease or remain constant with power level.
7.4. Effect of Frequency on Size of Particle Removed In order to assess the effect of frequency on the size of particles removed, silicon wafer substrates were contaminated in an ISO Class 5 cleanroom using 0.9-mm-sized polystyrene latex (PSL) spheres. The substrates were left for one day in the cleanroom. Multiple extractions were carried out using different ultrasonic frequencies 40, 58, 172, 192, and 470 kHz at varying power levels from 300 to 500 W for six time intervals. The particle size distributions were collected with the LPC for each sample. Mean size of each sample was obtained from 0.5 mm LPC data. From Figure 6.59, it can be seen that the mean size of surface-residual particle decreases with frequency, which agrees with the theoretical explanation that Multiple extractions at 100% power level Mean size in microns
5 At 1min At 2min
4
At 3min At 4min 3
At 5min At 6min
2
58
172
192
470
Frequency in kHz FIGURE 6.59 Effect of ultrasonic frequency on the mean size of residual particles. (see colour plate section at end for coloured version)
278
Developments in Surface Contamination and Cleaning
boundary layer thickness decreases with frequency. Cleanability increases with frequency at all power levels. When frequency increases, the bubble size decreases and hence the effect of cavitation decreases. At higher frequencies, the acoustic streaming effect is more significant than the cavitational effect. In cavitation-effective frequencies, there is a chance for substrate erosion; in the case of acoustic streaming, erosion is negligible. As time increases, the removed particle size from the substrate decreases at all power levels. The removed particle size at initial time intervals is large in the case of low frequencies (as eroded particles tend to be larger in size compared to naturally occurring contamination) and small in the case of higher frequencies. From the graphs, it can be seen that at 470 kHz the particle removal is high. When the mean size obtained at 470 kHz is considered, at the initial time itself, the particles are very small, whereas for other frequencies it takes longer for finer particles to be removed.
7.5. Frequency Ranges The current commercially available high-power ultrasonics (HPUS) cover a frequency range from 20 kHz to 1 MHz. There is no unified agreement on nomenclature and a general classification is definitely needed. We propose the following general classification for sonic frequencies in the low and high ranges (ultrasonic and megasonic ranges): 1. 2. 3. 4.
Conventional range ultrasonics (CRUS): 20 90 kHz Extended range ultrasonics (ERUS): 100 350 kHz Submegasonic range (SMGS): 400 900 kHz Megasonic range (MGS): >1 MHz.
ACKNOWLEDGEMENTS The contributions to this study of several undergraduate and graduate students (Moiz Diwan, Aviral Shukla, Vinay Raman, Anish Gupta, V. Sudheer Kumar) at IIT Madras are gratefully acknowledged.
REFERENCES [1] D.W. Cooper, Particulate Contamination and Microelectronics Manufacturing: an Introduc tion, Aerosol Sci. Technol. 5 (1986) 287. [2] W.G. Fisher, Particle Interaction with Integrated Circuits, in: R.P. Donovan (Ed.), Particle Control for Semiconductor Manufacturing, Marcel Dekker, New York, 2000, pp. 1 8. [3] S. Huth, O. Breitenstein, A. Huber, D. Dantz, U. Lambert, Localization and Detailed Investi gation of Gate Oxide Integrity Defects in Silicon MOS Structures, Microelectronic Eng. 59 (2001) 109. [4] M.A. Mendicino, P.K. Vasudev, P. Maillot, C. Hoener, J. Baylis, J. Bennett, T. Boden, S. Jackett, K. Huffman, M. Goodwin, Silicon on Insulator Material Qualification for Low Power Complementary Metal Oxide Semiconductor Application, Thin Solid Films 270 (1995) 578.
Ultrasonic Cleaning
279
[5] G.S. Selwyn, C.A. Weiss, F. Sequeda, C. Huang, In Situ Analysis of Particle Contamination in Magnetron Sputtering Processes, Thin Solid Films 317 (1998) 85. [6] L. Nebenzahl, R. Nagarajan, L. Volpe, J.S. Wong, O. Melroy, Chemical Integration and Contamination Control in Hard Disk Drive Manufacturing, J. Inst. Environ. Sci. 41 (1998) 31. [7] E. S. Yoon, B. Bhushan, Effect of Particulate Concentration, Materials and Size on the Friction and Wear of a Negative Pressure Picoslider Flying on a Laser Textured Disk, Wear 247 (2001) 180. [8] R. Nagarajan, Survey of Cleaning and Cleanliness Measurement in Disk Drive Manufacture, Precision Cleaning Magazine (Feb 1997) 13 22. [9] T.H. Kuehn, D.B. Kittelson, Y. Wu, R. Gouk, Particle Removal from Semiconductor Surfaces by Megasonic Cleaning, J. Aerosol Sci. 27 (Supp. 1) (1996) S427. [10] V.B. Menon, Particle Adhesion to Surfaces: Theory of Cleaning, in: R.P. Donovan (Ed.), Particle Control for Semiconductor Manufacturing, Marcel Dekker, New York, NY, 1990, pp. 359 382. [11] J.R. Hesson, Fundamentals of Ultrasonic Cleaning, Hessonic Ultrasonic, Washington, UT. Website: http://www.hessonic.com/guide/index.html, 2003. [12] S.B. Awad, Aqueous Ultrasonic Cleaning and Corrosion Protection of Steel Components, Metal Finishing 102 (2004) 56. [13] J. Fuchs, Ultrasonic Cleaning: Fundamental Theory and Application, Blackstone Ney Ultra sonics, Jamestown, NY. Website: http://www.blackstone ney.com/04.tech papers.php, 2002. [14] K.S. Suslick, Ultrasound: Its Chemical, Physical, and Biological Effects, VCH, New York, 1988. [15] W. Kern, Handbook of Semiconductor Wafer Cleaning Technology: Science, Technology, and Applications, Noyes Publications, Park Ridge, NJ (1993); second ed., William Andrew Publishing, Norwich, NY (2008). [16] B.J. Goode, R.D. Jones, J.N.H. Howells, Ultrasonic Pickling of Steel Strip, Ultrasonics 36 (1998) 79. [17] A.H. Crawford, Large Scale Ultrasonic Cleaning, Ultrasonics 6 (1968) 211. [18] R. Nagarajan, Use of Ultrasonic Cavitation in Surface Cleaning: A Mathematical Model to Relate Cleaning Efficiency and Surface Erosion Rate, J. Inst. Environ. Sci. 49 (2006) 40. [19] J.B. Durkee II, Management of Industrial Cleaning Technology and Processes, Elsevier, Oxford, UK, 2006. [20] S.B. Awad, Ultrasonic Cavitation and Precision Cleaning, Crest Ultrasonics, Trenton, NJ. Website: http://www.crest ultrasonics.com/process methodology.htm, 2009. [21] E.W. Lamm, The Development of Ultrasonic Cleaning, Controlled Environments Magazine (October 2003). [22] W.L. Nyborg, Acoustic Streaming, in: W.P. Mason (Ed.), Physical Acoustics II, Academic Press, New York, 1965, pp. 265 331. [23] J.J. Markham, Second Order Acoustic Field: Streaming and Viscosity and Relaxation, Phys. Rev. 86 (1952) 497. [24] A. Nowicki, W. Secomski, J. Wojcik, Acoustic Streaming: Comparison of Low Amplitude Linear Model with Streaming Velocities Measured by 32 MHz Doppler, Ultrasound Med. Biol. 23 (1997) 783. [25] S. Tjotta, On Some Non Linear Effects in Sound Fields, with Special Emphasis on the Generation of Vorticity and the Formation of Streaming Patterns, Arch. Math. Naturvidensk 55 (1959) 1. [26] C.E. Brennen, Cavitation and Bubble Dynamics, Oxford University Press, Oxford, UK, 1995. Chapter 4. [27] J. Halbert, Surface Cleaning: Using Ultrasonic Techniques for Wet Process Cleaning, Microcontamination 6 (1988) 221. [28] B. Vyas, C.M. Preece, Cavitation Induced Deformation of Aluminum, in Proc. ASTM Sympo sium on Erosion, Wear, and Interfaces with Corrosion, ASTM STP 567 (1974) pp. 77 102. [29] B. Vyas, C.M. Preece, Stress Produced in a Solid by Cavitation, J. Appl. Phys. 47 (1976) 5133.
280
Developments in Surface Contamination and Cleaning
[30] B. Vyas, C.M. Preece, Cavitation Erosion of Face Centered Cubic Metals, Met. Trans. Ser. A 8 (1977) 915. [31] J.H. Brunton, Some Mechanisms of Cavitation Damage in Vibratory Systems, Proc. 5th Int. Conf. Erosion by Solid and Liquid Impact, Cambridge, UK, paper 59 (1979) pp. 1 7. [32] K.A. Mørch, Concerted Collapse of Cavity Clouds in Ultrasonic Cavitation, in Proc. Acoustic Cavitation Meeting, Poole, Dorset (Institute of Acoustics, London) (1977) p. 62. [33] K.A. Mørch, in: W. Lauterborn (Ed.), Springer Series in Electrophysics, On the Collapse of Cavity Cluster in Flow Cavitation, in Proc. 1st Int. Conf. Cavitation and Inhomogeneities in Underwater Acoustics, vol. 4, Springer Verlag, Berlin, 1980, pp. 95 107. [34] I. Hansson, K.A. Mørch, Some Aspects on the Initial Stage of Ultrasonically Induced Cavitation Erosion, Proc. Ultrasonics Int. Conference, IPC Science and Technology, UK (1979) pp. 221 226. [35] P.M. Kanthale, P.R. Gogate, A.B. Pandit, A.M. Wilhelm, Cavity Cluster Approach for Quantification of Cavitational Intensity in Sonochemical Reactors, Ultrasonics Sonochem istry 10 (2003) 181. [36] P.R. Gogate, A.B. Pandit, Sonochemical Reactors: Scale Up Aspects, Ultrasonics Sono chemistry 11 (2004) 105. [37] R.T. Lahey Jr., R.P. Taleyarkhan, R.I. Nigmatulin, I.S. Akhatov, Sonoluminescence and the Search for Sono Fusion, Elsevier, Oxford, UK, 2006. [38] S. Verdan, G. Burato, M. Comet, L. Reinert, H. Fuzellier, Structural Changes of Metallic Surfaces Induced by Ultrasound, Ultrasonics Sonochemistry 10 (2003) 291. [39] R. Nagarajan, Cavitation Erosion of Substrates in Disk Drive Component Cleaning: An Exploratory Study, Wear 152 (1992) 75. [40] H. Brumberger, R.S. Stein, R. Powell, R., Light. Scattering, Science and Technology, pp. 34 42 (November 1968). [41] E. Maisonhaute, C. Prado, P.C. White, R.G. Compton, Surface Acoustic Cavitation Under stood Via Nanosecond Electrochemistry. Part III: Shear Stress in Ultrasonic Cleaning, Ultrasonics Sonochemistry 9 (2002) 297. [42] W. Bedkowski, G. Gasiak, C. Lachowicz, A. Lichtarowicz, T. Lagoda, E. Macha, Relations between Cavitation Erosion Resistance of Materials and Their Fatigue Strength under Random Loading, Wear 230 (1999) 201. [43] Y. Zhao, X. Bian, K. Yin, J. Zhou, J. Zhang, X. Hou, Relations of the Characteristic Temperatures and Fragility Parameters in Glass Forming Metallic System, Physica B 349 (2004) 327.
Index
Acoustic streaming and ultrasonics, 240 6 Adhesive Research, 191 Adhesive tape, 191 2 Aerosol removal rate see Particle deposition rate Aerosols, flow, 2 Airborne molecular contamination (AMC), 58, 64 5, 162 4, 170 Airborne particles: contamination monitoring, 124 5 traditional v. critical/busy sampling, 129 42 Airflow: near surface characteristics, 4 6 particles, deposition, 21 3 scenarios, 5 Alara 1146Ô strippable coating, 186, 210 Alumina and ultrasonics, 268 9 Aluminium and ultrasonics, 267 8 AMC see Airborne molecular contamination Ammonia, ethanol and tetraethoxysilane (TEOS), 101 Ammonium hydroxide hydrogen peroxide mixture (APM), 93 4 Antique Telescope Society, 201 Archer, Frederick Scott, 181 ARclearÒ, 191 ARcleanÒ, 191 Argonne Supergel, 189 Benzylalkonium chloride (BAC), 108 Beryllium removal, 192, 215 Boundary layer concept, 4 7, 105 Brewster windows, 200 Brown, Robert, 7 Brownian diffusion: boundary layer, 5, 6, 105 particles, transport, 7 8 particle deposition, 6, 20 turbulent diffusion, 8 Brush cleaning, 94 Cahn C 34/C 35Ô microbalance, 252 3 Capture Coating, 188, 210
CarbicoteÒ: 946 coating, 182 3 coating, 185 Carbon contamination of EUV optics, 73 5 Cavitation and ultrasonics, 240 6, 254 CDs see Compact discs Chemco 791 coating, 183 Chemical mechanical planarization (CMP), 231 Chernobyl nuclear accident, 214 Cleanroom, 4 CMOS see complementary metal oxide semiconductor CMP see chemical mechanical planarization Collodion: cleaning guidelines, 201 3 cleaning optics, 198 200 particle sample preparation, 207 8 solvent based coatings, 180 2 Compact discs (CDs), 206 Complementary metal oxide semiconductor (CMOS), 117, 227 Contamination control: conclusions, 75 6 pitfalls, 75 RPN scores, 66 74 space applications, 3 4 systems approach, 59 66 Contamination monitoring systems (continuous), 121 75 Continuous contamination monitoring: 100% sampling, 125 6 airborne particle counts: case studies, 129 42 continuous air flow, 141 2 continuous electrostatic charge, 140 1 critical and busy sampling, 127, 128 9 data collection, 128 description, 124 5 diagnosing problems, 137 40 extended duration manual monitoring, 137 ISO class 5 unidirectional flow benches in Class 7 ballroom, 130 3
281
282
Continuous contamination monitoring: (Continued ) ISO class 5 vertical laminar flow hoods in class 6 room, 133 4 ISO Class 7 ballroom, 129 8 trend, cyclic and burst patterns of particle generation, 134 7 cleaning: bare aluminium parts, 154 6 bask loading (fill level), 152 cleaner management using ISPM, 157 8 correlation with parts measurement, 152 4 installation, 146 7 ISPM for monitoring cleaning, 145 6 ISPM and sample inlet location, 148 50 large bare stainless steel parts, 156 7 part arrival rate and sequence, 150 2 parts cleaned, 147 critical sampling, 127, 128 9 data collection, 128 high technology manufacturing: air quality, 158 9 air velocity and direction, 162 airborne molecular contamination, 162 4 airborne particle concentrations, 159 61 antennas for electrostatic charge monitoring, 171 2 cleanliness of surfaces, 167 71 electrostatic charge, 167 71 process fluid purity, 164 7 relative humidity, 161 room pressurization, 162 temperature, 161 justification, 123 4 CorShieldÒ VpCIÔ coating, 185 6 DEC see Derived air concentrations DeconGelÔ coating, 187, 213 Deposition condition and ka, 112 15 Deposition flux, particles, 6 7 Deposition velocity see Particle deposition velocity Derived air concentrations (DEC), 210 Detachment of particles, 112 17 ka, 112 16 Discoat 4210 coating, 182 3 Disk enclosure (DK), 227 Drag force and particle transport, 9 10 Dual frequencies and ultrasonics, 264 6 Einstein, Albert, 181 Electric field and surfaces, 30 2
Index
Electrodecontamination system, 193 4 Electrostatic discharge (ESD), 122, 140, 172 Electrostatic force and particles deposition, 12 13 Enclosure surfaces and particles deposition, 1 56 Environmental Protection Agency, US, 180, 210 ESD see Electrostatic discharge Extreme ultraviolet (EUV) lithographic masks, 68 73, 73 5, 205 6 Failure mode effect analysis (FMEA) carbon contamination of EUV optics, 73 4 cleaning, 64 monitoring, 66 occurrence, severity, detection, 76 particle contamination on lithographic reticules, 66 8 ranking of contamination steps, 59 RPN reduction, 62 source control, 63 systems approach, 60 yield optimization, 61 2 Fick’s first law of diffusion, 6 Flow, aerosols, 2 FMEA see failure mode effect analysis Forced convection, 21 2 Gravitational settling and particle transport, 10 11, 20 HaloshieldÒ coatings, 193 4 Hand peeled coatings, 215 Hard disk drives (HDDs), 227 HASP see Hyper accumulating strippable polymer HDDs see Hard disk drives HEPA filters, 122 High frequency sound waves, 229 Holtz, Thomas Stoltz, 181 Homogeneous turbulence model for particles deposition, 32 7 HPM see Hydrochloric acid hydrogen peroxide mixture Hubble Space Telescope, 3 Hydrochloric acid hydrogen peroxide mixture (HPM), 93 4 Hyper accumulating strippable polymer (HASP), 212
283
Index
ICP MS see Inductively coupled plasma mass spectrometry ICs see Integrated circuits Immersible transducers, 236 In situ particle monitors see ISPM In film particles, 83 Inductively coupled plasma mass spectrometry (ICP MS), 101, 207, 215 Instacote CC Wet and CC Strip, 187 8 ‘‘Insulated gate field effect transistor’’, 85 6 Integrated circuits (ICs), 58, 81 5, 248 Internal Reticle Inspection System (IRIS), 68 International Technology Roadmap for Semiconductors (ITRS), 68 70 interpolyelectrolyte complexes (IPECs), 191 IPECs see interpolyelectrolyte complexes IRIS see Internal Reticle Inspection System Isotron coatings, 188 9, 210 12 ISPM (in situ particle monitors), 145 6, 148 50, 171 ITRS see International Technology Roadmap for Semiconductors ka and deposition condition, 112 15 kb: relative humidity, 113 15 salt concentration, 106 7 surfactant, 108 11 temperature, 107 8 viscosity, 107 8 Laser cleaning, 94 Laser windows cleaning, 200 1 Life cycle assessment (LCA) of strippable coatings, 208 9 Lignosol BD coating, 214 Linear diffusion rate of particles, pH, 106 ‘‘Liquid Envelope’’ coating, 214 Liquid Particle Counters (LPCs), 98 100, 103, 105, 249 50 Liquid turbidity meters (ultrasonics), 250 2 Lithographic reticles and particle contamina tion, 66 73 Magnetic particle testing, 208 Magnetoresistive (MR) heads, 140, 172, 227 Magnetostrictive transducers, 234 5, 235 6 Mars Exploration Rover, 57 8, 59 Maynard, John, 181 Megasonics, 94, 110, 227 Me´nard, Louis, 180
Metal ion core particles in semiconductor processing, 101 Metal oxide semiconductor (MOS), 227 Metallic surfaces and ultrasonics, 248 9 Microbalances and ultrasonics, 252 3 Momentum boundary layer, 4 MOS see Metal oxide semiconductor (MOS) MR see Magnetoresistive ‘‘Multiple extraction’’ procedure (ultrasonics), 261 3 Murphy model for particles in semiconductor processing, 87 9 Nafion (polymer), 181 National Optical Astronomy Observatories, 202 Natural convection, 22 3 Negative binomial model for particles in semiconductor processing, 88 9 Nephelometric turbidity unit (NTU), 251, 262 Nitto Denko Corporation, 191 Non Collodion cleaning of optics, 203 5 Non optical cleaning applications, 207 8 non radioactive decontamination, 214 15 radioactive decontamination, 208 14 Non volatile residue monitor (NVR), 166 Nostoc muscorum bacterium, 213 NTU see Nephelometric turbidity unit NVR see Non volatile residue monitor Opticlean, 183 4, 203 7 Optics and non Collodion cleaning, 203 5 Orientation of surfaces, 24 6 Origin of particles: cleanroom air, 89 90 process liquids, 90 wafer edge, 90 1 PAG see passive aerosol generator Particle deposition rate, ß, 15 18 Particle deposition velocity, vd, 14 15, 18 19 Particle removal efficiency (PRE), 91 3 Particle sample preparation and Collodion, 207 8 Particles: Airborne, 124 5, 129 42 contamination and lithographic reticles, 66 73 deposition, 24 32 airflow, 21 3 Brownian diffusion, 6, 20
284
Particles: (Continued ) deposition flux, 6 7 enclosure surfaces, 1 56 experiments in enclosed environments, 48 56 measurement, 17 19 models, 32 7, 38 42, 42 3 parameters, 14 17 particle characteristics, 19 21 pH, 116 7 size of particle, 19 20 summary, 43 4 surfaces, 24 32 turbophoresis, 13 14 generation, TNO method, 71 removal from immersed surfaces (ultrasonics), 258 61 semiconductor processing, 81 120 detachment, 112 17 diffusion out of boundary layer, 104 11 holes/trenches, 83 integrated circuits manufacturing, 81 4 Liquid Particle Counter, 98 100, 103, 105 as mask, 84 metal ion core, 101 Murphy model (yield calculation), 87 9 negative binomial model (yield calculation), 88 9 origin, 89 91 Poisson model (yield calculation), 85 7 removal, 91 6, 96 8, 101 4 Wallmark model (yield calculation), 85 size removed and frequency, 276 8 transport: Brownian diffusion, 7 8 drag force, 9 10 electrostatic force, 12 13 gravitational settling, 12 11, 20 thermophoresis, 11 12, 29 turbulent diffusion, 8 Particulate contamination effects in microelectronics manufacturing, 226 8 Passive aerosol generator (PAG), 188 Patterning deviations of particles in masks, 84 Payload cavity, 4 Payload fairing, 4 PCBs see printed circuit boards PEG see polyethylene glycol Pelouze, Theophile, 180 PENTEK coating, 209
Index
pH: linear diffusion rate of particles, 106 particle deposition, 116 7 surfactant and kb, 108 9 Phase masks cleaning, 205 Physical vapor deposition (PVD), 238 Piezoelectric transducers, 235 PLS see Polystyrene latex spheres Poisson model for particles in semiconductor processing, 85 7 Polished silicon wafers and ultrasonics, 270 4 Polyethylene glycol (PEG), 107, 108 11 Polymeric films, 191 Polymers for coatings, 179 Polystyrene latex spheres (PLS), 102 3 Poly(vinyl acetate) (PVA) coatings, 189 Poly(vinyl alcohol) coatings/pastes, 189 91, 214 PRE see Particle removal efficiency Printed circuit boards (PCBs), 230 Priorities for contaminating steps, 61 2 PVA see poly(vinyl acetate) PVD see Physical vapor deposition Removal of particles in semiconductor processing, 93 4, 96 8, 101 4 Reticle handling (RH), 70 3, 113 15 Risk priority number see RPN Royal Greenwich Observatory Telescope, 203 RPN (risk priority number) reduction: cleaning, 64 5 contamination controls, 66 75 occurrence, 68 73 severity, 67 8 monitoring, 65 6 prevention, 62 4 adhesion prevention, 64 source control, 63 transport, 63 4 Salt concentration, and kb, 106 7 Savannah River Site (SRS), 210 SAW see Surface acoustic wave SC see Standard clean Semiconductors: cleaning, 94 production, 67 See also particles, semiconductor processing SensorCoat smart coating, 193 Silicon wafers cleaning, 205 6 Silicone rubber and ultrasonics, 270 1
285
Index
Skin wart remover, 208 Smart coatings, 193 SMIF see Standard machine interface Soluble polysaccharide (SPS), 213 Solvent based strippable coatings, 180 3 Space and contamination control, 3 4 SPS see Soluble polysaccharide SRS see Savannah River Site Stampers for compact discs cleaning, 206 7 Standard clean (SC), 93 Standard machine interface (SMIF), 169 70 Stober synthesis, 101 Stokes Einstein expression, 7 Stripcoat TLC FreeÔ coating, 186, 210 11, 214 Strippable coatings: Ingredients, 178 9 Issues, 197 non optical applications, 207 8 hand peelable, 215 life cycle assessment, 208 9 non radioactive decontamination, 214 15 radioactive decontamination, 208 14 precision cleaning: non optical applications, 207 215 optical surfaces, 198 207 properties, 180 solvent based: Collodion, 180 2 Others, 182 3 surface contaminants, 177 224 UV curable, 195 7 water based: Adhesive Research, 191 adhesive tape, 191 2 ALARA 1146Ô strippable coating, 186, 210 Argonne Supergel, 189 beryllium removal, 192, 215 Capture Coating, 188, 210 CarbicoteÒ, 185 Chemco, 184 5, 185 6 CorShieldÒ VpCIÔ, 185 6 DeconGelÔ, 187, 213 electrodecontamination system, 193 4 HaloshieldÒ Coatings, 193 4 Instacote CC Wet and CC Strip, 187 8 Isotron coatings, 188 9, 210 12 Opticlean, 183 4, 203 7 polymeric films, 191 poly(vinyl acetate), 189 poly(vinyl alcohol), 189 91, 214
smart, 193 Stripcoat TLC FreeÔ, 186, 210 11, 214 thick film etching fluid, 194 5 Universal Photonics, 184, 203 Surface acoustic wave (SAW) sensors, 163 4 Surfaces: contaminants and strippable coatings, 177 224 particle deposition: electric field, 30 2 orientation: horizontal vs. vertical, 24 6 temperature: warm vs. cold, 28 30 texture: smooth vs. rough, 26 8 ultrasonics: measurement, 249 54 roughness, 248 Surfactant and kb, 108 11 System definition for effect of contamination, 60 1 Systems approach: contamination control, 59 66 definition and effect of contamination, 60 1 priorities for contaminating steps, 61 2 RPN reduction by severity, occurrence and detection, 62 6 Temperature: kb, 107 8 Surfaces, 28 30 TEOS see Ammonia, ethanol and tetraethoxysilane Texture of surfaces, 26 8 Thermophoresis and particle transport, 11 12, 29 Thick film etching fluid, 194 5 Thickeners for coatings, 179 Three layer model for particle deposition, 38 42 Ticker tool tests, 71 TNO method and particle generation, 71 Total organic carbon (TOC) measurement, 240 Transducers for ultrasonics, 232 4, 234 6, 236 7 TSI Incorporated, 122 Turbophoresis and particles deposition, 13 14 Turbulent diffusion and particle transport, 8 Ultrasonic cleaning: acoustic streaming, 240 6, 258 61 cavitation, 240 6, 247 8
286
Ultrasonic cleaning: (Continued ) cleanability/erodibility: alumina, 268 9 aluminium, 267 8 polished silicon wafer, 270 4 silicone rubber, 269 270 substrate material, 266 7 dual frequencies, 264 6 erodibility measurement using a microbalance, 274 5 experiments, 261 6 frequency and size of particle removed, 277 8 introduction, 226 8 particle removal, 94 process equipment design, 238 40 structural changes of metallic surfaces, 248 9 substrate material: alumina, 268 70 aluminium, 267 8 cleanability/erodibility, 266 7 polished silicon wafer, 270 4 silicone rubber, 270 1 surface cleanliness: cavitation meter, 254 Liquid Particle Counters, 249 50 liquid turbidity meters, 250 2
Index
microbalances, 252 3 roughness, 248 tanks, 237 theory, 254 8 transducers: immersible, 236 7 magnetorestrictive, 234 5, 235 6 piezoelectric, 235 6 United States (US) Environmental Protection Agency, 180, 210 Universal Photonics coatings, 184, 203 US see United States UV (ultra violet): curable coatings, 195 7 sensors and Mars Exploration Rover, 59 Vapor Phased Corrosion Inhibitors (VpCI), 185 6 Viscosity and kb, 107 8 Volatile organic compounds (VOCs), 196 Wall loss rate see Particle deposition rate Wallmark model for particles in semiconductor processing, 85 Water based coatings, 183 95 ‘‘Wet collodion process’’, 181
FIGURE 2.3 Contamination of UV sensors by Mars dust, which reduces the transparency of the window: sensors with dust (left), detail of the sensors (middle), and the cleaning solution with a magnetic brush (right) [21].
FIGURE 2.6 Microscopy images of two materials A and B before (left) and after 100 contacts: middle: light microscopy (scale bar is 20 mm); right: scanning electron microscopy (scale bar for material A is 50 mm; scale bar for material B is 100 mm) [67].
FIGURE 2.7 Storage box for lithographic (EUV) reticles: hardware (left) [68] and drawing [28] (see text for explanation of the colors).
I
FIGURE 3.27 SEM picture (left) of a surface formally covered with 1.5 mm sized silica particles leaving residues that have also been measured with AFM (right). Courtesy of Frank Holsteyns [57].
FIGURE 4.1
A critical and busy sampling tube installation on a workstation.
FIGURE 5.1 A protected aluminum mirror before and after cleaning with Opticlean. Courtesy of Photonic Cleaning Technologies, Platteville, WI.
II
FIGURE 5.3 Application of DeconGelÔ coating on flat surfaces. (a) One end of the contami nated area is taped. (b) The coating is applied to the tape and the contaminated area. (c, d) The coating is spread by a trowel held at 90 to the surface to cover the contaminated area. (e, f) Once the coating has dried, it is peeled off using the taped end. Courtesy of Cellular Bioengineering Inc., Honolulu, HI.
III
FIGURE 5.5 Principle of HaloShieldÒ technology. (a) Untreated fiber. (b) The fiber surface is treated with HaloShieldÒ N halamine technology. (c) The treated fiber is washed in chlorine bleach. (d) The chlorine molecules are bound by the HaloShieldÒ coating and are anchored to the fiber. Courtesy of HaloSource Inc., Bothell, WA.
IV
FIGURE 5.8 Sequence of steps for cleaning an optical mirror with peelable blue coating. (a) The mount is wrapped to protect it from the coating. The coating is applied by spraying a thick layer on the surface. (b) The coating is starting to dry. (c) The dried coating is removed by peeling. The mirror surface is clean. (d) Any remaining coating around the bevel is removed with fresh tape [128].
FIGURE 5.9 A large 1 meter unprotected gold first surface mirror before cleaning (left photograph) and after cleaning (right photograph) with Opticlean. Courtesy of Photonic Cleaning Technologies, Platteville, WI.
V
Stripcoat Coating
Preparation
Application
Removal
Isotron Coating
Preparation
Application
Removal
FIGURE 5.11 Application of commercial strippable coatings for removal of radioactive contaminants from a concrete surface. The upper series of pictures shows the preparation, application, and removal of Stripcoat TLC FreeÔ coating from the surface. The lower series of pictures show the preparation, application, and removal of Isotron coating from the surface [160,161].
VI
Erosion vs. time curve Weight loss from unpolished silicon wafer (mg)
0.08 0.07 0.06 0.05 0.04 0.03 0.02
192 kHz 172 kHz 132 kHz
0.01 0 0
2
4
6
8
10
12
14
Time (minutes) FIGURE 6.52
Erosion vs. time curve for unpolished silicon wafer at 132, 172, and 192 kHz.
33
Cleanability (/min)
31 29 27 25 23 21 19 17 15 50
60
70
80
90
100
110
Power level (%) 40 kHz
FIGURE 6.56
58 kHz
192 kHz
430 kHz
58+192 kHz
Cleanability power plot for silicon at various frequencies.
VII
Erodibility (per cm2)
240 220 200 180 160 140 120 100 50
60
70
80
90
100
110
Power level (%) 58 kHz
40 kHz
FIGURE 6.57
430 kHz
192 kHz
58+192 kHz
Erodibility power plot for silicon at various frequencies.
C eanab ty ndex (cm2/m n)
15 14 13 60% 70% 80% 90% 100%
12 11 10 40
58
192
430
58+192
Frequency (kHz) FIGURE 6.58
Cleanability index frequency plots for silicon at various power levels.
Multiple extractions at 100% power level Mean size in microns
5 At 1min At 2min
4
At 3min At 4min 3
At 5min At 6min
2
58
172
192
470
Frequency in kHz FIGURE 6.59
VIII
Effect of ultrasonic frequency on the mean size of residual particles.
