Case 1:14-cv-01432-LPS Document 238-1 Filed 12/12/19 Page 1 of 61 PageID #: 16024
`
`Exhibit A
`
`

`

`Case 1:14-cv-01432-LPS Document 238-1 Filed 12/12/19 Page 2 of 61 PageID #: 16025
`
`IN THE UNITED STATES DISTRICT COURT
`FOR THE DISTRICT OF DELAWARE
`
`ELM 3DS INNOVATIONS, LLC,
`
`
`
`Plaintiff,
`
`
`v.
`SAMSUNG ELECTRONICS CO., LTD., et
`al.,
`Defendants.
`
`
`
`ELM 3DS INNOVATIONS, LLC,
`
`
`
`Plaintiff,
`
`
`v.
`MICRON TECHNOLOGY, INC., et al.,
`
`
`
`Defendants.
`
`ELM 3DS INNOVATIONS, LLC,
`
`
`
`Plaintiff,
`
`
`v.
`SK HYNIX INC., et al.,
`
`
`
`Defendants.
`
`
`
`
`
`
`
`
`
`
`
`C.A. No. 14-cv-1430-LPS-CJB
`
`JURY TRIAL DEMANDED
`
`
`C.A. No. 14-cv-1431-LPS-CJB
`
`JURY TRIAL DEMANDED
`
`
`
`
`C.A. No. 14-cv-1432-LPS-CJB
`
`JURY TRIAL DEMANDED
`
`
`
`DECLARATION OF SHEFFORD BAKER
`
`My name is Shefford P. Baker. I am an Associate Professor in the Department of
`
`
`
`
`
`
`
`Materials Science and Engineering at Cornell University. I received my undergraduate degree in
`
`Music from the University of New Mexico before earning my M.S. and PhD (1992) in Materials
`
`Science and Engineering at Stanford University. My PhD work focuses on stresses and
`
`mechanical properties of thin metal/metal multilayer films. I also developed methods for
`
`measuring mechanical properties of thin films using nanoindentation.
`
`Following Stanford, I worked at the Max-Planck-Institut für Metallforschung in Stuttgart
`
`Germany for five years as a member of the research staff. I supervised PhD students and conducted
`
`research. My work there focused on projects related to thin film metallizations for use in integrated
`
`circuits. In one project, I studied electromigration phenomena and developed an experiment to
`
`
`
`
` 1
`
`

`

`Case 1:14-cv-01432-LPS Document 238-1 Filed 12/12/19 Page 3 of 61 PageID #: 16026
`
`relate conductor
`
`line microstructure
`
`to electromigration failure and correlated failure
`
`characteristics to line texture. In several projects, we investigated stresses and thermomechanical
`
`behavior of thin copper metallizations. The semiconductor industry was gearing up to transition to
`
`copper metallizations and did not know much about it. I supervised the design and construction of
`
`two ultra-high vacuum sputter deposition systems and a substrate curvature stress measurement
`
`system. My students and I also conducted thin film stress measurements using x-ray diffraction
`
`and mechanical property measurements using nanoindentation.
`
`I joined the faculty at Cornell in 1998 in the Department of Material Science and
`
`Engineering (MSE). During my twenty years at Cornell, my research has focused on structure and
`
`mechanical properties in a range of materials including metal and ceramic thin films,
`
`biomineralized tissues and biogenic, geologic, and synthetic mineral crystals, silicate glasses,
`
`metallic glasses, and a number of other materials.
`
`My research group at Cornell develops sophisticated machinery and equipment to produce
`
`and study thin films. For example, in our thin film lab, we have built (and rebuilt) a high vacuum
`
`(≈10-7 Torr) evaporator system with thermal and e-beam sources, complete source and substrate
`
`shuttering, a heated and cooled sample stage and an ion gun for ion beam assisted deposition. We
`
`also designed and built an ultra-high vacuum sputter deposition system (< 10-9 Torr) with three
`
`confocal sputter guns, a rotating heated (500˚C) sample stage, RF and DC power supplies,
`
`substrate bias. This system includes a substrate curvature stress measurement system that can
`
`detect a radius of curvature to about 60 km on a 100 mm substrate (very high stress resolution) at
`
`temperatures from liquid nitrogen to over 800˚C. In addition, we outfitted the G-2 beamline at the
`
`Cornell High Energy Synchrotron Source (CHESS), designing and building a 6-circle kappa
`
`geometry goniometer, a heated environmentally controlled stage, and other features dedicated to
`
`thin film structure and stress measurements. My group has used this machinery to study
`
`
`
`
` 2
`
`

`

`Case 1:14-cv-01432-LPS Document 238-1 Filed 12/12/19 Page 4 of 61 PageID #: 16027
`
`electromigration and adhesion in thin copper films, texture and texture transformations in a variety
`
`of FCC metals films (primarily silver), phase formation, phase transformation and texture
`
`patterning in thin tantalum films, stresses in thin tungsten films, and many others. We also have a
`
`nanomechanics lab that has included several nanoindenters, an AFM and several homemade
`
`fracture and adhesion test setups. We also operate the MSE department’s tensile tester where we
`
`have conducted tests on the mechanical properties of a number of samples from brazed lap joints
`
`for stainless steel heat exchangers to grafted joints in wine grape plants. Our development of and
`
`access to this equipment allows us to conduct a broad range of experiments. In particular our thin
`
`film lab allows us to produce extremely pure and clean films for model studies.
`
`I have published extensively in the area of thin films and semiconductors. My publications
`
`have examined issues relating to stress, creep, strain hardening, structure, texture and texture
`
`transformations, phase formation and phase transformations, and many other features in thin films.
`
`A full list of my publications is attached as Exhibit A.
`
`Much of the thin film work was motivated by the needs of the semiconductor
`
`manufacturing industry. Starting in Germany my students and I worked to understand the
`
`mechanical properties of the copper metallizations that were eventually adopted by the industry.
`
`For example, we studied the effect of tantalum barrier layers on structure and properties of copper
`
`films and worked out the relationship between interfacial oxygen concentration and adhesion
`
`between Cu films and adjacent SiO2 layers. In another project, we studied tantalum films that were
`
`used as thin film resistors and that are now under development for Giant Spin Hall Effect devices.
`
`During my time at Cornell, I have received several awards, including Excellence in
`
`Teaching Awards and the CAREER Award from the National Science Foundation. In addition to
`
`my research and teaching, I have been involved in developing the engineering curriculum for
`
`undergraduates, serving as the Director of Undergraduate Studies for the Department of Materials
`
`
`
`
` 3
`
`

`

`Case 1:14-cv-01432-LPS Document 238-1 Filed 12/12/19 Page 5 of 61 PageID #: 16028
`
`Science and Engineering for several years. I am currently the Director of the Master of Engineering
`
`program in MSE, a program that I and several colleagues created 4 years ago to prepare MSE
`
`students for careers in industry. I am also a member of the Fields of Theoretical and Applied
`
`Mechanics, Mechanical Engineering, and Aerospace Engineering at Cornell.
`
`Outside of Cornell, I have held many roles in the Materials Research Society, which is an
`
`international organization that promotes interdisciplinary materials research among professionals
`
`worldwide. I was the president of that organization in 2009 and am now the chair of the
`
`Publications Committee. I was also involved in the formation of the Nanoscale Informal Science
`
`Education network. The NISE is funded by the National Science Foundation and promotes public
`
`education of science (including nanoscale science) in the United States. I am also currently a
`
`member of the American Ceramics Society and TMS.
`
`I. ASSIGNMENT, LEGAL STANDARDS, AND MATERIALS CONSIDERED
`
`I have been asked to provide opinions about the reports offered by the experts retained by
`
`the defendants in this matter, Drs. Steven Murray and Richard B. Fair.
`
`I understand from Elm’s counsel that the terms of a patent should have their plain and
`
`ordinary meaning in the field of the invention as understood by a person having ordinary skill in
`
`the art. I also understand that the defendants argue that certain claims are indefinite. I understand
`
`that a claim is indefinite when it does not point out and distinctly claim the subject matter of the
`
`invention, which means that the claims fail to inform a person of skill in the art, with reasonable
`
`certainty, about the scope of the invention. I have conducted both inquiries as of the date of the
`
`patent, which is April 1997.
`
`Given that understanding, my analysis has focused on how a person of skill in the art
`
`would have understood the claims and the other sections of the patent given my background with
`
`semiconductor technology and thin films, which include dielectrics. I have also focused on the
`
`
`
`
` 4
`
`

`

`Case 1:14-cv-01432-LPS Document 238-1 Filed 12/12/19 Page 6 of 61 PageID #: 16029
`
`assertions regarding the patents in this lawsuit in Murray’s and Fair’s reports based on my
`
`knowledge of the art.
`
`I am being compensated at my ordinary and customary consulting rate of $320 per hour
`
`for my work. My compensation is in no way contingent on the nature of my findings, the
`
`presentation of my findings in testimony, or the outcome of this or any other proceeding. I have
`
`no other interest in this proceeding.
`
`I had one week to complete this report. So in the interests of time, I have not addressed
`
`all of those portions of the reports where Murray and Fair just quote portions of the patent,
`
`specification, or prosecution history. Instead, I discuss the technology and how a person of
`
`ordinary skill’s technical knowledge would inform the reading of the patent in certain areas,
`
`including how the claims provide a reasonably certain scope of the invention.
`
`In forming my opinions, I have read the patents at issue in this lawsuit, the reports of the
`
`defendants’ experts, the articles and other materials cited in this report, and the materials cited by
`
`the defendants’ experts. I have also used my background in the technology and general
`
`knowledge that I have gained in my career as a professor and practitioner. I have read Murray’s
`
`and Fair’s description of the level of ordinary skill in the art and use the same understanding
`
`here.
`
`II. DISCUSSION
`
`In preparing this report, I came to the conclusion that a person of ordinary skill in the art
`
`would find the descriptions in the patents quite straightforward to follow, and so have taken the
`
`tactic of making a tutorial presentation of the relevant state of the art at the time of the patents,
`
`with a focus on what is needed to understand both the patents, and the arguments of Defendants’
`
`expert witnesses, Fair and Murry. This information is included in section A below. Following
`
`that, relatively short technical discussions of how I would interpret two of the topics of the Claim
`
`
`
`
` 5
`
`

`

`Case 1:14-cv-01432-LPS Document 238-1 Filed 12/12/19 Page 7 of 61 PageID #: 16030
`
`Construction process, namely “Low Stress Dielectrics” (and similar terms) and “Substantially
`
`Flexible Substrate” and similar terms in Sections B, C, and D.
`
`A. STRESSES IN INTEGRATED CIRCUIT MANUFACTURING
`
`A key feature in the patents has to do with the stresses that arise in a thin dielectric layer
`
`on a silicon substrate. The defendants’ expert witnesses made a series of arguments in their
`
`explanation of their claim constructions that are based on stresses. In short, they claim that
`
`because stresses are tensor quantities, inhomogeneous, of different “types,” and difficult to
`
`measure, that a person of ordinary skill in the art would not be able to ascertain whether stresses
`
`in a dielectric were low and that therefore claims having to do with low stresses in a layer are
`
`indefinite.
`
`In sharp contrast, I discuss that, while it is true that stress is a tensor quantity, that stress
`
`distributions are inhomogeneous, and that stresses may arise from different origins (the meaning
`
`of stress “types”), none of these features have any effect on the understanding or application of
`
`the patents. In addition, I explain that stresses in thin semiconductor layers can readily be
`
`measured with sufficient accuracy using methods widely known to persons of ordinary skill in
`
`the art at the time of the patents.
`
`To make these arguments, I begin with a brief tutorial on stresses in integrated circuit
`
`manufacturing.
`
`A.1. Stress as a tensor
`
``
`
`To understand what is being discussed in the patents and by the defendants’ expert
`
`witnesses, it is helpful to understand the concept of “stress.” (Reference for this entire section:
`
`Fung, Y.C., A First Course in Continuum Mechanics, (2nd ed.), Englewood Cliffs, New Jersey:
`
`Prentice-Hall Inc. 1997.) In general terms, stress describes the forces acting at a point in a
`
`
`
`
` 6
`
`

`

`Case 1:14-cv-01432-LPS Document 238-1 Filed 12/12/19 Page 8 of 61 PageID #: 16031
`
`material in terms of the force per unit area. Imagine a piece of material upon which different
`
`forces are acting including pushing, pulling, twisting, etc. as shown in Figure 1a.
`
`
`
`
`At any point P inside the material the net force is going in some direction, indicated by the arrow
`
`Figure 1: Forces in a body.
`
`
`
`marked T in Figure 1b. We can pick any plane that goes through P. For our bookkeeping, we
`
`assign a coordinate system such that the x and y directions are perpendicular to each other and lie
`
`in the plane, while z is perpendicular to the plane. The force T can then be broken down into
`
`three components: the force that is perpendicular to the plane is the “normal” force component,
`
`Tx, and the parts that act in the plane are the shear components, Ty and Tz. To turn these into
`
`stresses, we divide each of these force components by the area over which it acts to get stress
`
`components sxx (normal stress), and sxy, and sxz (shear stresses). (Stress is indicated by the
`
`Greek letter s or sigma.) To get the complete stress state at point P, we do the same calculations
`
`on two additional planes that are perpendicular to our first plane and to each other. It is common
`
`to show these three planes as the faces of a cube as shown in Figure 2. We see that the full stress
`
`state has 9 components, sxx, sxy, sxz, syx, syy, syz, szx, szy, and szz. But as it turns out, some of
`
`the shear stresses have the same values (sxy = syx, sxz = szx, and syz = szy). So only 3 values are
`
`needed to represent the 6 shear stress components acting at any given point P.
`
`
`
`
` 7
`
`

`

`Case 1:14-cv-01432-LPS Document 238-1 Filed 12/12/19 Page 9 of 61 PageID #: 16032
`
`
`Figure 2: The 9 components of stress in an x-y-z coordinate system.
`
`
`This means that, in general, 6 numbers are needed to define the stress at any point. There are
`
`three normal stress components sxx, syy, and szz, and there are three shear stress components sxy,
`
`sxz, and syz. The particular set of numbers depends on the coordinate system we picked (i.e., if
`
`you changed the orientation of the cube in Figure 2, you would need 6 different numbers to
`
`describe the same stress state at the same point).The fact that 6 numbers are needed at every
`
`point arises because stress is what is known as a tensor quantity (2nd rank tensor to be specific).
`
`In a loaded body such as that shown above, the stress varies from point to point and the “stress
`
`field” is the 3-D map of those values (6 values at every point).
`
`The key facts to remember here are:
`
`In theory and at its most basic, stress is a tensor quantity that requires 6 different
`
`numbers (components) to fully specify it at a point.
`
` •
`
`
`
`•
`
`In a body with irregular loads (Fig. 1a) or a body that is uniformly loaded but is
`
`inhomogeneous (different properties at different points), the stress field will be
`
`inhomogeneous.
`
`A.2. What are stress and strain?
`
`So what is stress? Atoms in a material are connected to each other by bonds, which, for
`
`our purposes here, may be thought of as tiny springs connecting the atoms. At a given set of
`
`
`
`
` 8
`
`

`

`Case 1:14-cv-01432-LPS Document 238-1 Filed 12/12/19 Page 10 of 61 PageID #: 16033
`
`conditions (pressure, temperature) each bond has an equilibrium length. This equilibrium length
`
`is the length the bond adopts if no external forces act on it. If the atoms are pushed closer
`
`together, the bond responds—just as a spring—with a compressive force that tries to return the
`
`atoms to their equilibrium separation. Similarly, if the atoms are pulled farther apart, the bond
`
`exerts a tensile force to return the springs to their equilibrium positions. Stress is just a measure
`
`of the forces in the bond “springs”, per unit area.
`
`The stretching of bonds (springs) discussed above, is described in terms of strain. A
`
`normal strain is just a change in length per unit length. Since materials may be both stretched and
`
`sheared, a complete description of the state of strain at any given point also requires 6 numbers
`
`(i.e. strain is also a second-rank tensor).
`
`A.3 Stress in real engineering applications (not a tensor!)
`
`While a tensor description is complicated, that is not what is being discussed in the
`
`patents. In most engineering applications, including semiconductor manufacturing, it is possible
`
`to pick coordinate systems and use well-accepted simplifying assumptions so that only a few
`
`stress and strain components are needed. Most scientists, engineers, and technologists never
`
`think of stress in its tensor form at all. For example, the most common test used to understand the
`
`mechanical properties of materials is a “uniaxial tension test.” In this test a cylinder of material is
`
`stretched along its axis. If the material can be thought of as homogenous (properties are the same
`
`at all points) and isotropic (properties are the same in all directions at one point), then there is
`
`only one stress component, s, which is just the stretching force divided by the cross sectional
`
`area of the cylinder, and there are two strains, e, which is the change in length per unit length
`
`along the axis, and eT, which represents the change in diameter as the cylinder is stretched.
`
`Deformation can be elastic or inelastic. Deformation is elastic if the deformed body
`
`returns to its original shape when the forces that caused the deformation are removed. For elastic
`
`
`
`
` 9
`
`

`

`Case 1:14-cv-01432-LPS Document 238-1 Filed 12/12/19 Page 11 of 61 PageID #: 16034
`
`deformation, stress and strain are simply related by Hooke’s law, s = Ee, where E is a numerical
`
`constant called Young’s modulus, and the axial and transverse strains are related by eT
`
`= -ne, where n is Poisson’s ratio. This formulation would be well known by persons with
`
`ordinary skill in almost all technologies where stress is a concern, including the art of
`
`semiconductor manufacturing.
`
`The key fact to remember here is
`
`•
`
`In most situations we do not need the full stress (or strain) tensor to assess the stress
`
`in a given application. A sufficient description of the stress can be obtained with a
`
`single number.
`
`A.4 Stress in a thin layer on a thick substrate
`
`A similarly common and simple formulation arises in fields that involve the use of thin
`
`layers on thick substrates. This includes thin films and coatings used in optical devices, wear-
`
`resistant coatings in tools, decorative coatings, catalytic thin films, biocompatible coatings and
`
`many other applications, including dielectric layers in semiconductor manufacturing. In all of
`
`these fields of application, it is well known that the interaction between the layer and the
`
`substrate can lead to stresses in the layer, and a simple way to measure this “film stress” or “layer
`
`stress” is very widely understood (Nix, W.D., Mechanical Properties of Thin Films.
`
`Metallurgical Transactions A, 1989. 20A: p. 2217-2245, Ohring, M., Materials Science of Thin
`
`Films. 2nd ed., 2002: Academic press., Freund, L. B., and Suresh, S. Thin Film Materials: Stress,
`
`Defect Formation and Surface Evolution. Cambridge University Press New York, NY, 2003.).
`
`We can understand this layer stress as follows: Imagine that a thin film is attached to a
`
`thick substrate as indicated in Figure 3a. Suppose that both the film and the substrate have
`
`uniform thickness, are homogeneous and isotropic, and are stress-free and flat. We conduct a
`
`thought experiment to see how film stresses arise (Nix, W.D., Mechanical Properties of Thin
`
`
`10
`
`
`

`

`Case 1:14-cv-01432-LPS Document 238-1 Filed 12/12/19 Page 12 of 61 PageID #: 16035
`
`Films. Metallurgical Transactions A, 1989. 20A: p. 2217-2245): First, we imagine removing the
`
`film from the substrate (b), and then changing the film dimensions relative to the substrate (c),
`
`meaning that the width of the substrate and film are not the same. There are many ways that such
`
`a relative dimension change can occur. We can then imagine applying external forces to stretch
`
`or compress the film so that it once again fits on the substrate (d), attaching it to the substrate (e),
`
`and releasing the external forces that we used to stretch or compress the film, (f). The substrate
`
`now carries the load needed to stretch or compress the film and, because that load is applied at
`
`the film/substrate interface, it causes the substrate to bend (f). In other words, the film wants to
`
`return to its original dimensions after it is stretched and attached to the substrate, but it cannot do
`
`so because the substrate is keeping it stretched to its new dimension. The film is held in a
`
`stressed state, bending the substrate with it. The force exerted on the film must be equal in
`
`magnitude and opposite in sign to the force on the substrate (no external forces are applied) but
`
`the film is thin so that force is spread over a small area leading to a high stress while the
`
`substrate is thick so the force there is spread over a large area leading to small stresses.
`
`In reality, the film remains on the substrate and something happens to make it want to
`
`shrink or expand relative to the substrate, but because it is forced to fit the substrate it gets
`
`stretched or compressed, i.e. stresses arise! We can use the thought experiment to imagine what
`
`would happen if we could take the film off of the substrate to see the relative dimension changes.
`
`This helps us to calculate the sign and the magnitude of the stresses.
`
`
`11
`
`
`

`

`Case 1:14-cv-01432-LPS Document 238-1 Filed 12/12/19 Page 13 of 61 PageID #: 16036
`Substrate Interaction Stress
`
`Film
`
`Stress-Free Film on Substrate
`
`Apply Stress to Return Film to Substrate Dimensions
`
`Reattach Film to Substrate
`
`Remove External Forces
`
`
`
`Substrate
`
`Remove Film from Substrate
`
`a
`
`b
`
`d
`
`e
`
`hange Film Dimensions Relative to Substrate
`
`cC
`
`f
`Fig. 3: Origin of film stresses
`This type of stress arises from an interaction between the film
`and the substrate ⇒ "Substrate Interaction Stresses"
`So we see that any process that changes the equilibrium in-plane dimension of the film
`(not common terminology)
`relative to the substrate (Fig. 3b-c) leads to a stress in the film (Fig. 3e) and that stress leads to
`
`
`
`curvature of the film/substrate package (Fig. 3f. Note that the relaxation of the stress in the film
`
`due to curvature, Fig. 3e to Fig. 3f is very small—we’ll come back to this shortly.). Two of the
`
`most common ways that these relative dimensional changes can occur are differential thermal
`
`expansion and structure evolution in the film (Nix, W.D., Mechanical Properties of Thin Films.
`
`Metallurgical Transactions A, 1989. 20A: p. 2217-2245, Ohring, M., Materials Science of Thin
`
`Films. 2nd ed., 2002: Academic press., Freund, L. B., and Suresh, S. Thin Film Materials: Stress,
`
`Defect Formation and Surface Evolution. Cambridge University Press New York, NY, 2003.).
`
`Differential thermal expansion occurs when the film and substrate have different thermal
`
`expansion coefficients and are subjected to a temperature change. The associated stresses are
`
`often called “thermal stresses.” Structure evolution occurs when the film becomes more or less
`
`dense due to thermal processing, ion implantation, or other processes. For example, atoms
`
`arriving at the film surface with enough energy to be implanted into the film, cause the film to be
`
`denser than equilibrium, it will want to expand but will be constrained by the substrate and will
`
`be in compression. The stresses associated with structure evolution during deposition are often
`
`
`12
`
`
`

`

`Case 1:14-cv-01432-LPS Document 238-1 Filed 12/12/19 Page 14 of 61 PageID #: 16037
`
`called “growth stresses” (or “intrinsic stresses”). These two sources of stress are by far the most
`
`important in thin film technology. After deposition, stresses may also change due to plastic
`
`deformation or further structure evolution.
`
`Fortunately, the stress state in the thin layer shown in Fig. 3 is not complicated. If we
`
`choose a coordinate system in which the x and y directions are in the plane of the film and the z
`
`direction is perpendicular to the film, the stress state in the film everywhere except very near the
`
`edges (≈ 1 or 2 film thicknesses from the edge) can be characterized by a single value, sxx = syy,
`
`which is commonly referred to as the “film stress” sf, which is a normal stress acting in the plane
`
`of the film. This stress is also known as a “layer stress.” For a thin film on a thick substrate it is
`
`easy to show that the film stress is simply related to the radius of curvature R that this stress
`
`𝜎"= $%&%’
`()*+%)&-). , (1)
`
`induces in the substrate as follows,
`
`
`
`where Es and ns are Young’s modulus and Poisson’s ratio (elastic constants) of the substrate,
`
`respectively, and ts and tf are the thicknesses of the substrate and film, respectively (Nix, W.D.,
`
`Mechanical Properties of Thin Films. Metallurgical Transactions A, 1989. 20A: p. 2217-2245,
`
`Ohring, M., Materials Science of Thin Films. 2nd ed., 2002: Academic press., Freund, L. B., and
`
`Suresh, S. Thin Film Materials: Stress, Defect Formation and Surface Evolution. Cambridge
`
`University Press New York, NY, 2003.). Equation 1 is also known as the “Stoney Equation”
`
`(Stoney, G. G., "The Tension of Metallic Films Deposited by Electrolysis". Proceedings of the
`
`Royal Society A: Mathematical, Physical and Engineering Sciences (1364-5021), 82 (553), p.
`
`172 (1909))
`
`The formulation shown in Equation 1 is very important to us for three reasons:
`
`
`13
`
`
`

`

`Case 1:14-cv-01432-LPS Document 238-1 Filed 12/12/19 Page 15 of 61 PageID #: 16038
`
`•
`
`(1) This situation is very common in semiconductor manufacturing. As we will see, an
`
`integrated circuit is formed from a series of thin layers deposited on a substrate. Each
`
`of these layers individually and all of them collectively interact with the substrate
`
`giving rise to layer stresses and substrate curvature according to Eq. 1. (Nix, W.D.,
`
`Mechanical Properties of Thin Films. Metallurgical Transactions A, 1989. 20A: p.
`
`2217-2245, Ohring, M., Materials Science of Thin Films. 2nd ed., 2002: Academic
`
`press., Freund, L. B., and Suresh, S. Thin Film Materials: Stress, Defect Formation
`
`and Surface Evolution. Cambridge University Press New York, NY, 2003.)
`
`•
`
`(2) Like the uniaxial tension test described above, we again have a simple
`
`relationship that does not require the use of tensors. The layer stress is represented by
`
`a single number. (Nix, W.D., Mechanical Properties of Thin Films. Metallurgical
`
`Transactions A, 1989. 20A: p. 2217-2245, Ohring, M., Materials Science of Thin
`
`Films. 2nd ed., 2002: Academic press., Freund, L. B., and Suresh, S. Thin Film
`
`Materials: Stress, Defect Formation and Surface Evolution. Cambridge University
`
`Press New York, NY, 2003.)
`
`•
`
`(3) Eq. 1 gives us a very simple way to determine the layer stress. If we can measure
`
`the radius of curvature R and know the thicknesses of the layer and the substrate, as
`
`well as Young’s modulus and Poisson’s ratio for the substrate, we can calculate the
`
`layer stress. In fact, this kind of measurement is very commonly used in all industries
`
`that depend on thin films (or “layers” or “coatings”) (Ohring, M., Materials Science
`
`of Thin Films. 2nd ed., 2002: Academic press., Freund, L. B., and Suresh, S. Thin
`
`Film Materials: Stress, Defect Formation and Surface Evolution. Cambridge
`
`University Press New York, NY, 2003.).
`
`And here are four additional important things to know about these substrate interaction stresses:
`
`
`14
`
`
`

`

`Case 1:14-cv-01432-LPS Document 238-1 Filed 12/12/19 Page 16 of 61 PageID #: 16039
`
`•
`
`(4) Although a number of terms (like “thermal stress” and “intrinsic stress”) are
`
`commonly used to distinguish the origins of stress, there are no physical distinctions
`
`among these stresses. Stress is stress; there are no actual different “types” of stress,
`
`these terms are just shorthand to refer to the stress’s origin. It is more correct to think
`
`of the origin of the relative change in dimensions shown in Fig. 3c as the
`
`distinguishing feature. Thus, one would describe the stretching/compressing of the
`
`film relative to the substrate in terms of “thermal strain” or other appropriate terms
`
`(Nix, W.D., Mechanical Properties of Thin Films. Metallurgical Transactions A,
`
`1989. 20A: p. 2217-2245), and many in the industry do. The sources of strain can be
`
`identified and summed up to get the film strain ef. The film stress is simply related to
`
`the film strain by the appropriate version of Hooke’s law, which is in this case sf =
`
`(Ef/(1-nf))ef.
`
`•
`
`(5) When the forces acting to compress or stretch an atomic bond are removed, it
`
`returns to its equilibrium length. This is the source of elastic behavior. Because
`
`stresses are associated with bond stretching, stresses are associated with elastic
`
`strains only. Plastic deformation occurs when atomic bonds are broken and reformed
`
`with different neighbors, permanently changing the shape of the material. The total
`
`strain is the sum of the elastic strain and the plastic strain, but the stress arises only
`
`from the elastic part.
`
`•
`
`(6) For cases where Eq. 1 (the Stoney Equation) applies, the stress in a layer needed
`
`to produce a certain curvature (curvature k = 1/R) is much greater than the stress that
`
`would arise in that layer if you simply took a flat stress-free film/substrate
`
`combination and bent it to the same curvature using external forces. The result of this
`
`is that a change in curvature induced by adding a second film does not significantly
`
`
`15
`
`
`

`

`Case 1:14-cv-01432-LPS Document 238-1 Filed 12/12/19 Page 17 of 61 PageID #: 16040
`
`affect the stress in a first film. Thus, to a good first approximation, all films interact
`
`with the substrate independently. (Freund, L. B., and Suresh, S. Thin Film Materials:
`
`Stress, Defect Formation and Surface Evolution. Cambridge University Press New
`
`York, NY, 2003.)
`
`•
`
`(7) Real films may not be homogeneous, but, as I will demonstrate below,
`
`inhomogeneity does not necessarily affect the technologist’s use of Eq. 1.
`
`Everything in this section would have been well known to a person of ordinary skill in the art of
`
`semiconductor manufacturing at the time of the patents.
`
`A.5 Layer stresses in integrated circuits
`
`Film stresses are a major concern in semiconductor processing. As indicated in Equation
`
`1, stresses in an attached layer can cause the substrate to curve. If the curvature of the substrate,
`
`k = 1/R, becomes too great, it can become impossible to align photolithography masks so that the
`
`next layer can be made in registry with previous layers, impossible to successfully planarize by
`
`chemical mechanical processing, and impossible to attach a (curved) chip die to a (flat) substrate
`
`or package. (Garrou, P., C. Bower, and P. Ramm, Handbook of 3-D Integration: Technology and
`
`Applications of 3D Integrated Circuits, Weinheim: Wiley-VCH Verlag; 2008., Nix, W.D.,
`
`Mechanical Properties of Thin Films. Metallurgical Transactions A, 1989. 20A: p. 2217-2245,
`
`Clemens, B.M. and J.A. Bain, Stress Determination in Textured Thin Films Using X-ray
`
`Diffraction, Materials Research Society Bulletin, XVII(7): p. 46-51 (1992)) For these reasons
`
`(and others), a great deal of effort has been spent over the past five decades to reduce this
`
`curvature. One way to reduce the curvature is to make the substrate thicker. Indeed, as substrate
`
`wafer diameters have increased, the specified thickness also increased (e.g. in going from 100
`
`mm to 200 mm diameter Si wafers, the standard thickness was increased from 525 µm to 725
`
`
`16
`
`
`

`

`Case 1:14-cv-01432-LPS Document 238-1 Filed 12/12/19 Page 18 of 61 PageID #: 16041
`
`µm). The patents at issue are concerned with a diffe