By: Christopher Frerking (chris@ntknet.com)
`
`Reg. No. 42,557
`
`UNITED STATES PATENT AND TRADEMARK OFFICE
` _______________
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`_____________
`
`LAM RESEARCH CORP.,
`
`Petitioner
`
`v.
`
`DANIEL L. FLAMM,
`
`Patent Owner
`
`CASE IPR2016-0466
`U.S. Patent No. 5,711,849
`
`
`
`PATENT OWNER’S PRELIMINARY RESPONSE
`UNDER 37 C.F.R. § 42.107
`
`
`
`
`
`
`
`
`
`
`
`Mail Stop: PATENT BOARD
`Patent Trial and Appeal Board
`U.S. Patent & Trademark Office
`P.O. Box 1450
`Alexandria, VA 22313-1450
`
`

`
`Inter Partes Review of U.S. Patent No. 5,711,849
`IPR2016-0466
`
`
`
`
`
`Page(s)
`
`
`TABLE OF CONTENTS ...................................................................................... i
`
`TABLE OF AUTHORITIES ................................................................................ ii
`
`I.
`
`II.
`
`Introduction ........................................................................................... 1
`
`Background ............................................................................................. 1
`
`A. Dr. Flamm’s Invention ...................................................................... 1
`
`B. Battey ................................................................................................. 4
`
`C. Battey and the ‘849 Patent ................................................................ 5
`
`III. Ground 1, Claim 26 ............................................................................... 6
`
`IV. Ground 2 ................................................................................................ 7
`
`V. Ground 3 ................................................................................................ 11
`
`VI. Conclusion ............................................................................................. 13
`
`
`
`
`
`i
`
`

`
`Inter Partes Review of U.S. Patent No. 5,711,849
`IPR2016-0466
`
`
`
`TABLE OF AUTHORITIES
`
`Statutes Page(s)
`
`35 U.S.C. § 102 ................................................................................................... 6
`
`35 U.S.C. § 103 ................................................................................................... 6
`
`37 C.F.R. § 42.107 .............................................................................................. 1
`
`
`
`ii
`
`
`
`
`
`

`
`Inter Partes Review of U.S. Patent No. 5,711,849
`IPR2016-0466
`
`
`
`Daniel L. Flamm, Sc.D., the co-inventor and sole owner of the U.S. Patent
`
`No. 5,711,849 (“the ‘849 patent”), through his counsel, submits this preliminary
`
`response pursuant to 37 C.F.R. § 42.107 and asks that the Patent Trial and Appeals
`
`Board decline to institute inter partes review on the instant petition because the
`
`petition fails to show a reasonable likelihood that any challenged claim is
`
`unpatentable.
`
`I.
`
`Introduction
`
`Lam’s petition relies primarily on a paper written by James F. Battey in
`
`1976, some twenty years before Dr. Flamm applied for the ‘849 patent. In a field
`
`as fast moving and competitive as the semiconductor industry, one would think
`
`that if Battey’s teachings were as similar to the ‘849 patent as Lam now contends,
`
`that someone would have discovered Dr. Flamm’s invention long before Dr.
`
`Flamm did. The explanation for this conundrum is simple: Battey is not even close
`
`to Dr. Flamm’s invention and lacks teaching even the basic elements of Dr.
`
`Flamm’s invention.
`
`II. Background
`
`A. Dr. Flamm’s Invention
`
`The Background of the Invention in the ‘849 patent states the problems Dr.
`
`Flamm faced:
`
` A limitation with the conventional plasma etching technique is
`
`
`
`1
`
`

`
`
`
`Inter Partes Review of U.S. Patent No. 5,711,849
`IPR2016-0466
`
`obtaining and maintaining etching uniformity within selected
`predetermined limits. In fact, the conventional technique for
`obtaining and maintaining uniform etching relies upon a “trial and
`error” process. The trial and error process often cannot anticipate the
`effects of parameter changes necessary for actual wafer production.
`Accordingly,
`the conventional
`technique
`for obtaining and
`maintaining etching uniformity is often costly, laborious, and difficult
`to achieve.
`
` Another limitation with the conventional plasma etching technique
`is reaction rates between the etching species and the etched material
`are often not available. Accordingly, it is often impossible to
`anticipate actual etch rates from reaction rate constants since no
`accurate reaction rate constants are available. In fact, conventional
`techniques require the actual construction and operation of an etching
`apparatus to obtain accurate etch rates.
` Full-scale prototype
`equipment and the use of actual semiconductor wafers are often
`required, thereby being an expensive and time-consuming process.
`
`(Ex. 1001 at 1:26-:44.)
`
`Dr. Flamm’s solution to these problems is summarized in the first paragraph
`
`of the Summary of Invention:
`
`According to the present invention, a plasma etching method that
`includes determining a reaction rate coefficient based upon etch
`profile data is provided. The present plasma etching method provides
`for an easy and cost effective way to select appropriate etching
`parameters such as reactor dimensions, temperature, pressure, radio
`frequency (rf) power, flow rate and the like by way of the etch profile
`data.
`
`(Id. at 1:51-:57.)
`
`The “reaction rate coefficient” is a key factor in the “surface reaction rate
`
`constant,” which appears in all claims of the ‘849 patent.
`
`
`
`2
`
`

`
`Inter Partes Review of U.S. Patent No. 5,711,849
`IPR2016-0466
`
`
`
`B.
`
`Battey
`
`At the threshold, the analysis set forth in the Battey reference is
`
`fundamentally mathematically flawed and the edge rates he purports to calculate
`
`are obvious error, as demonstrated in the attached Appendix A.
`
`Setting that aside, Battey reports his theoretical calculations and experiments
`
`performed to determine the difference in plasma strip rates between the edge and
`
`center of small wafers. A photoresist strip rate at the edge of a silicon wafer was
`
`independent of wafer diameter and spacing, and later assumed using a constant
`
`oxygen concentration. (Ex. 1002 at Abstract, p. 2 Analysis.) A photoresist strip
`
`rate at the center of the silicon wafer was calculated using complex formulae
`
`incorporating a number of assumptions. (Ex. 1002 at Abstract.) The theory
`
`provided for having an array of wafers stacked parallel to one another. In contrast
`
`to his theory, however, Battey’s experiments measured a strip rate at the edge of an
`
`isolated single wafer on a block and at the center of the same wafer. His study and
`
`calculations were limited to these two points, including the center and edge, which
`
`assumed a constant oxygen concentration.
`
`Battey then solved for concentration in the z-direction, which is normal to
`
`the wafer. (Ex. 1002 at 2, Analysis.) Battey concludes that the “etch rate at the
`
`center increases linearly with increasing wafer spacing” among other factors. (Id.
`
`at Abstract.) Given that the concentration is assumed to be constant at the edge,
`
`
`
`3
`
`

`
`Inter Partes Review of U.S. Patent No. 5,711,849
`IPR2016-0466
`
`
`Battey’s change in concentration at the edge is zero. Battey never calculates an
`
`etch rate at the edge.
`
`Battey does not teach any process for performing optimization, let alone
`
`Lam’s conclusory, but unspecified, “well-known process.” (Pet. at 12.) The
`
`Battey publication merely concerns one explicit analytical model he has
`
`formulated (the various equations) based on a series of assumptions made to
`
`describe resist stripping in a particularized plasma reactor geometry. Battey asserts
`
`that his model “is also applicable to calculating the degree of inhomogeneity across
`
`the face of a wafer in an etching plasma or in a diffusion operation” (Ex. 1002 at 1),
`
`but offers no explanation or further discussion of any etching other than photoresist
`
`stripping.
`
`The most conspicuous failing of Battey as prior art to the ‘849 patent is the
`
`total absence of any teaching about a “surface reaction rate constant.” Even
`
`assuming arguendo that Battey had successfully modeled the manner in which the
`
`edge-to-center resist stripping rates vary with wafer size, which admittedly he did
`
`not, nothing in Battey describes extracting a surface reaction rate constant for a
`
`temperature from a relatively non-uniform etching rate profile on a film. In this
`
`regard, Battey merely teaches to compare experimental ratios of a center etching
`
`rate and an edge etching rate to each other and to respective values calculated
`
`using his transcendental formulae after making numerous further assumptions.
`
`
`
`4
`
`

`
`Inter Partes Review of U.S. Patent No. 5,711,849
`IPR2016-0466
`
`
`
`It is very clear that Battey teaches away from using his theoretical model to
`
`predict the ratio, because he concludes that the model does not agree with
`
`experimental trends in his edge-to-center data.
`
`Since the true edge etch rates arising from Battey’s model are obviously
`
`wrong (e.g., the model predicts there will be no etching at the edge of a wafer as
`
`demonstrated in Appendix A), and Battey provides no method operable to solve
`
`the transcendental equations, the reference is useless and would have been ignored
`
`by those skilled in the art.
`
`Most particularly, given that Battey never calculates a surface reaction rate,
`
`he has no ability to use such surface reaction rate in the manufacture of
`
`semiconductor devices.
`
`C. Battey and the ‘849 Patent
`
`Unlike Battey, the ‘849 invention comprises a process for fabricating a
`
`device using a surface reaction rate that is derived from the disclosed method. The
`
`‘849 method comprises an embodiment for defining an entire relatively non-
`
`uniform etching profile, and defining a set of etching and spatial coordinate data,
`
`and using this data to extract a rate constant. The spatial coordinate data were
`
`either in the spatial x-y coordinates, or r-theta coordinates while the z-direction is
`
`normal to the wafer. By contrast, Battey solved for oxygen concentration in the z-
`
`direction at the center and assumed a constant at the edge, which is a flawed
`
`
`
`5
`
`

`
`Inter Partes Review of U.S. Patent No. 5,711,849
`IPR2016-0466
`
`
`approximation to his theory, and he finds that his hypothesis fails.
`
`The ‘849 specification further includes complete and internally consistent
`
`enabling methods to acquire and evaluate an array of etch rate data along a spatial
`
`region of the wafer as a function of temperature. The specification discloses how
`
`to perform these methods using production-worthy commercial equipment to
`
`process a batch of substrates, or a single substrate. The methods enable finding a
`
`pre-exponential and activation energy for an etching reaction and for finding an
`
`etchant concentration at an edge of a wafer from the data. It also includes a
`
`method to find the effective etchant-consuming area of a substrate based on the
`
`etching profile data and to use that effective area to choose plasma source
`
`parameters for processing (Ex. 1001 at Figs. 3 & 4), and to use these things to
`
`determine the subspace of parameters where a predetermined degree of uniformity
`
`can be obtained (id. at Fig. 5). Battey discloses none of this and teaches away from
`
`using etch rate data in the spatial region of the wafer, but focusing his calculations
`
`in the z-direction. In fact, Battey admits that his experimental method was not
`
`useful in a batch processing configuration, and he could not obtain reproducible
`
`data even for a single wafer without modifying the apparatus and insuring that
`
`three different simultaneous temperature readings were maintained. Lam’s
`
`allegations, to the extent that they contend that Battey teaches any of these things,
`
`are merely conclusory and, indeed, illogical.
`
`
`
`6
`
`

`
`Inter Partes Review of U.S. Patent No. 5,711,849
`IPR2016-0466
`
`
`III. Ground 1, Claim 26
`
`Lam makes a peculiar assertion that “Battey Teaches All the Limitations of
`
`Independent Claim 26,” but then argues based under 35 U.S.C. § 103(a) rather than
`
`35 U.S.C. § 102. (Pet. at 13.) It is of no moment, however, since Battey does not
`
`teach all the limitations of claim 26—far from it.
`
`Contrary to Lam’s assertion, Battey does not teach the limitation of “a
`
`process for fabricating a device,” Battey merely ran experiments concerned with
`
`photoresist stripping. Battey found it necessary to modify an apparatus to accept a
`
`single wafer on an aluminum block in a certain position, drill a hole in the block to
`
`accept a thermometer, and fasten at least two external thermometers to his chamber
`
`to even obtain any reproducible data—a Rube Goldberg type of configuration.
`
`Plus, Battey found it necessary to perform downstream NO2 titrations to obtain a
`
`value of etchant concentration. Downstream titration is strictly a laboratory
`
`technique that was not, and still is not, compatible with production processing
`
`equipment.
`
`Claim element [26.a] requires a “surface reaction rate constant.” (Pet. at 18-
`
`19.) As discussed above, Battey neither extracted a surface reaction rate constant
`
`nor could he have done so given the paucity of his data: e.g., there is no data
`
`beyond two points, where one of the points was a constant; there is no temperature
`
`dependent analysis; the effect of the effective etchable area on oxygen
`
`
`
`7
`
`

`
`Inter Partes Review of U.S. Patent No. 5,711,849
`IPR2016-0466
`
`
`concentration is not considered, etc. Despite that, Lam blithely states: “Forming a
`
`quantity, h, that is the ratio of the surface reaction rate constant to the diffusion
`
`coefficient, of the wafer and at the center were calculated for values of h of 25, 2.5,
`
`0.25, 0.025, and 0.0025” (Pet. at 14 (emphasis added)), but provides no support for
`
`its assertion that “h” is the ratio of the surface reaction rate constant to the
`
`diffusion coefficient. Nor does Lam ever identify where in Battey there is a
`
`purported teaching of a surface reaction rate constant. That omission is
`
`understandable, since there is none.
`
`Claim element [26.d] expresses the need for a “relatively non-uniform
`
`etching profile.” Battey did not use a profile; his data consisted of a single value
`
`center-to-edge thickness ratio. He deliberately focused on measuring only two
`
`points on the wafer (“the edge and the center” (Ex. 1002 at Abstract)), which are
`
`only enough for one relative value that might define a shape—assuming, arguendo,
`
`his theory was correct and the single datum was not noisy. In fact, Battey never
`
`obtained or examined any etch profiles from his theory, never obtained any
`
`experimental etch profile data, and mentions nothing about any need or reason to
`
`do so.
`
`To obtain a profile, an array of etch-rate points distributed along a surface
`
`coordinate is necessary. The Battey article says nothing about this kind of data and
`
`Battey did not acquire any such data.
`
`
`
`8
`
`

`
`Inter Partes Review of U.S. Patent No. 5,711,849
`IPR2016-0466
`
`
`
`Claim element [26.e] requires a “relatively non-uniform etching profile” and
`
`claim element [26.f] requires a “surface reaction rate constant,” which similarly
`
`cannot be met by Battey.
`
`IV. Ground 2
`
`Accepting that Battey “may lack explicit discussion of using the surface
`
`reaction rate constant” (Pet. at 25), Lam introduces Galewski (Ex. 1003) to argue
`
`that it, in combination with Battey, renders obvious claims 1-3, 5, 7-12, 14, 16-21,
`
`and 29. There is, however, no basis to combine Battey with Galewski.
`
`The Galewski article does not relate to plasmas or plasma etching. Instead,
`
`Galewski concerns high-temperature epitaxial growth of silicon. As a matter of
`
`fact, Galewski only mentions the term “plasma etching” at one place (Ex. 1003 at
`
`9) to note that an oxide pattern having the vertical walls needed for selective
`
`epitaxy can be made using plasma etching.
`
`Epitaxy is a specialized type of chemical vapor deposition of a crystalline
`
`overlayer that matches to the lattice of a crystalline substrate. It amounts to the
`
`growth of a single crystal film on top of a preexisting crystalline substrate, such as
`
`a crystalline semiconductor layer. Generally, a tightly controlled physical and
`
`chemical environment, scrupulously free of contamination, high temperature
`
`(800°-900°C in Galewski), and relatively low growth rates are required to realize a
`
`high-quality epitaxial film suitable for semiconductor device applications.
`
`
`
`9
`
`

`
`Inter Partes Review of U.S. Patent No. 5,711,849
`IPR2016-0466
`
`
`
`Because a plasma environment tends to bombard surfaces with various types
`
`of disruptive radiation, and because plasma-surface interactions tend to release
`
`material from reactor walls into the gas phase, even plasma-enhanced chemical
`
`vapor deposition has not been useful for the type of high-quality epitaxy discussed
`
`in the Galewski publication. As for plasma etching, exposing crystalline silicon to
`
`an etch plasma inevitably causes lattice disruption, and implants species from the
`
`gas phase that make it difficult to achieve high-quality epitaxy afterwards.
`
`Accordingly, when plasma etching is first performed before epitaxial growth, such
`
`as to pattern a mask, cleaning and purely thermal processing steps are generally
`
`performed afterwards to reconstruct the surface substrate lattice as a prerequisite
`
`for the growth of a useful epitaxial film.
`
`In short, epitaxy is practically the antithesis of plasma etching. Despite that,
`
`Lam argues that it would have been obvious to a person of skill in the art to
`
`combine Galewski’s reactor design, deposition conditions, and wafer cleaning, etc.
`
`for defect-free epitaxy to Battey’s plasma stripping of photoresist from a single
`
`wafer in an oxygen plasma. The operations are neither analogous nor compatible.
`
`A person of ordinary skill in the art of plasma etching would not have known how
`
`to design or operate an epitaxy reactor or the detailed physics and chemistry of
`
`epitaxial growth. These were very different things, having disparate purposes
`
`(other than being individual steps in the very complex process of producing a
`
`
`
`10
`
`

`
`Inter Partes Review of U.S. Patent No. 5,711,849
`IPR2016-0466
`
`
`semiconductor device), and are conceptually incompatible with one another.
`
`Because the phenomena and requirements and chemistry of relatively “cold”
`
`non-equilibrium plasma etching processes were known to be inconsistent with the
`
`purity, precision, and conditions of high-temperature purely chemical thermal
`
`silicon epitaxy, and because Battey’s solution of a diffusion/chemical reaction
`
`equation was inconsistent with his data, a person of skill in the art would have not
`
`thought to combine, let alone known how to combine, Battey with Galewski.
`
`V. Ground 3
`
`For Ground 3, Lam’s argument grows even more attenuated, adding Sawin
`
`to Battey and Galewski. As there was no basis to combine Battey and Galewski,
`
`there is especially no basis to add Sawin to the mix.
`
`Sawin describes a method to measure and display changes in the thickness
`
`of a film over time and displays such data using an X-Y-Z Cartesian coordinate
`
`system. (Pet. at 48.) In fact, transformation of coordinates on a cylindrical
`
`coordinate system to a Cartesian coordinate system has been known for at least
`
`hundreds of years. That much would have been obvious without any reference to
`
`Sawin.
`
`But what Sawin did not do is extract a surface rate constant. The range of
`
`the X-direction and Y-direction spatial coordinates defining etch rates in Sawin
`
`extends beyond substrate area having any profile. The reason is that Sawin
`
`
`
`11
`
`

`
`Inter Partes Review of U.S. Patent No. 5,711,849
`IPR2016-0466
`
`
`discloses profile data for circular substrates. The X-range and Y-range of this data,
`
`however, define a rectangular region. Hence, the Sawin X-Y direction covers
`
`regions where there is no substrate, no etching, and no profile.
`
`Said another way, any X-value (spatial coordinate) within a profile of the
`
`‘849 patent, for example according to claim 4, combined with any Y-value (spatial
`
`coordinate) within said profile, is within the etching profile—e.g., no combination
`
`of X and Y lies outside of the profile. By contrast, since Sawin has round
`
`substrates, some of the Sawin X-values in combination with some of the Sawin Y-
`
`values are beyond the bounds of the profile.
`
`More importantly, Sawin teaches an entirely different thing, e.g., using an
`
`X-Y coordinate system for displaying data. Claim 4 of the ‘849 patent is directed
`
`toward extracting a rate constant based on an etch rate data profile that comprises
`
`an inherent rectangular X-Y coordinate range. The extraction of claim 4 is what
`
`inherently depends on X-Y coordinates. For example, an X-Y embodiment
`
`depends on the explicit closed form profile solution disclosed in the ‘849 patent at
`
`col. 15 lines 10-20.
`
`The explicit closed form solution is to extract a rate for a substrate covering
`
`inherent X-Y coordinates, e.g., rectangular. This operation requires that the
`
`substrate etching spans an inherent X-Y coordinate range.
`
`Sawin teaches nothing about extracting a surface rate constant. Battey
`
`
`
`12
`
`

`
`Inter Partes Review of U.S. Patent No. 5,711,849
`IPR2016-0466
`
`
`teaches nothing about displaying a graph based on X-Y coordinates because his
`
`theoretical equations inherently depend on cylindrical coordinates and etching all
`
`area covered by the cylindrical coordinate range (r,z), which are basic for his
`
`formulae and his computations. Therefore there would have been no motive or
`
`even any way to apply the Sawin coordinates to Battey’s expressions, and Lam
`
`identifies none.
`
`The ‘849 patent, however, discloses methods including formulae that are
`
`adapted to an X-Y coordinate range, as expressly claimed in Claims 4 and 13. The
`
`method of extracting a rate constant disclosed by the ‘849 patent operates in
`
`cooperation with selection of a coordinate system based on a substrate shape.
`
`VI. CONCLUSION
`
`For the foregoing reasons, the instant petition should be denied.
`
`
`Date: April 27, 2016
`
`
`
`
`
`
`
`
`
`
`
`
`
`Respectfully Submitted,
`
`By: /Christopher Frerking, reg. no. 42,557/
` Christopher Frerking, reg. no. 42,557
`
`174 Rumford Street
`Concord, New Hampshire 03301
`Telephone: (603) 706-3127
`Email: chris@ntknet.com
`
`
`
`Counsel for Daniel L. Flamm
`
`
`
`
`
`
`
`13
`
`

`
`APPENDIX A
`
`Battey’s Analysis and Calculated Values of Edge Rates are Error
`
`The Battey Relationship for Oxygen Atom Concentration
`
`Battey provides three expressions he purports to be useful to obtain a value of “v” the
`oxygen etchant concentration, in a position at height z above a circular substrate surface and at a
`radius r shown in (I) below:
`
`(Ex. 1002 at 2.)
`
`The locus of coordinate positions to where equations I apply is strictly limited to points
`within a
`right circular cylinder defined by a
`radius
`r=a and a height z=S:
`
`(I)
`
`
`This transcendental relationship does not provide an explicit relation for v(r,z) because the
`equations require solving for values of n. Values of n must be extracted from the second
`equation. Furthermore, both the first and second equations require numerical values of h before
`they can be evaluated and, still further, the evaluation of I0(nr) depends on having numerical
`values for these quantities.
`
`Battey discloses that this solution for v(r,z) was taken from p.220 of reference 3, a book
`concerning temperature profiles for conduction heat transfer authored by Carslaw and Jaeger,
`
`
`
`1
`
`

`
`“Conduction of Heat in Solids,” Oxford University Press 1959 (hereinafter “Carslaw”).
`According to Battey, the use of these equations is proper because:
`
`the oxygen concentration] becomes
`the calculating
`[of
`the problem
`mathematically the same as the conduction of heat in a solid cylinder with a
`radiation boundary condition at the photoresist-covered wafer surfaces and a
`constant oxygen concentration or temperature at the side of the cylinder formed
`by drawing lines between the edges of the wafers.
`
`(Id.)
`
`In comparing the Battey equations (I) with the Carslaw relationship, shown below:
`
`
`
`it can be seen that Battey retained the notation Carslaw associated with a temperature profile and
`related quantities (e.g., v, h, etc.). The Carslaw relationship admits specifying a temperature
`profile function f(z) boundary condition along the surface of a right cylinder at r=a, at the edge of
`the medium (z corresponds to the-height above the substrate at the edge of a substrate in Battey).
`This function appears in the integrand to be integrated between 0 and l in the right hand side of
`Carslaw’s equation 17, above.
`
`However, Battey fixes f(z) to be a constant value (corresponding to an oxygen
`concentration that would surround a region between adjacent wafers) independent of z, (in
`Battey f(z)= n0 = constant). This constant can be taken outside of the integrand such that
`integration of the remaining sine and cosine terms results in the expression within the curly
`brackets shown in the Battey equation, e.g., (I). Accordingly, the Battey oxygen concentration
`function, v(r,z), relates to a constant oxygen-concentration, (temperature in Carslaw) at the edge
`of the cylindrical region, (r=a, z=[0,S]).
`
`Battey’s Relationship Cannot Predict a Reaction Rate at the Edge of a Substrate
`
`As explained above, the Battey expression for v(r,z) defines a constant value,
`independent of z, on the right circular cylinder wall extending upward from the edge of a
`substrate, e.g., v(r,z=a) = Constant. Battey selected this boundary condition to fix the oxygen
`
`
`
`2
`
`

`
`atom concentration surrounding a substrate to be the bulk oxygen concentration being used to
`ask the photoresist.
`
`It is elementary that the derivative (or partial derivative) of a constant value that does not
`change with the independent derivative variable is 0 (naught). Battey overlooked this elementary
`mathematical fact when attempting to evaluate the derivative of v(r,z) at the edge of a substrate
`(r=a) in the z-direction.
`
`Battey attempts to use his theory to evaluate a diffusion current arriving at the edge of a
`circular substrate at r=a by assuming it will be proportional to the term dv(r=a, z)/dz):
`
` (II)
`
`This is an obvious error because the value of the left hand side of the transcendental
`Battey equations (II) above is zero owing to the constant value boundary condition at r=a. The
`derivative in the z-direction at r=0 is identically 0. Obviously, the right hand side of these
`equations (II) are zero as well.
`
`The expressions on the right hand side of Battey’s equations are an infinite series
`comprising transcendental equations and the roots n of another transcendental relation presented
`in Battey’s (I). Hence, evaluating v(r,z) in equations (I), and dv(r,z)/dz in from equation (II) is a
`relatively difficult process at best.
`
`Battey has no explicit solutions for v(r,z) or dv(r,z)/dz in terms of known values. Instead,
`the paper offers a convoluted series of arguments and approximations (Ex. 1002 at 3) that
`purport to circumvent the transcendental nature of the exact formulae shown in (II). These
`assumptions purport to allow the equations (II) to be closely approximated and provide
`theoretical values of dv/dz at the edge position r=a, z=0. The approximation of dv(r=a,z)/dz is
`used to obtain a flux of atoms to the edge of a substrate (the flux is proportional to dv/dz at the
`edge position r=a, z=0). These complex assumptions and approximations were unnecessary and
`are obviously wrong. A PHOSITA would have recognized that the derivative of a constant is
`
`
`
`3
`
`

`
`naught, and therefore that all of Battey’s calculated relative edge to center strip rates should be
`zero (the nonzero edge etch rates such as shown in Table I of Battey) are obviously an artifact of
`approximation errors.
`
`Battey’s discussion of trends in center to edge strip rate ratios are all based on these
`faulty approximations providing nonzero values of dv/dz at the edge (r=a, z=0).
`
`In summary, because the equations (I) satisfy the given constant boundary condition,
`dv(r=a,z)/dz = 0 and the correct value of the left hand side of the above equations (II) is 0
`(naught). Accordingly, Battey’s entire discussion concerning calculations of edge to center etch
`ratios are based on plain error and the analysis must be rejected.
`
`Physical Discussion
`
`The Battey/Carslaw solution for v(r,z) pertains to an isoconcentration boundary condition
`of oxygen atoms on a right circular cylinder r=a defined by the edge of a substrate. This solution
`can be used to provide values of a temperature field in a medium defined by a right circular
`cylinder by r=a, extending in the z-direction. However the directional derivative of this solution
`does not define a flux of atoms to the edge of the substrate.
`
`Battey’s attempt to use this solution to find a directional derivative at r=a amounts to
`circular reasoning; the isoconcentration (constant) boundary condition at r=a defines the given
`solution equations in the first place. In reality, the condition of isoconcentration is an
`approximation in which the vertical z-direction diffusion current is suppressed at the edge.
`
`Having a diffusive flux of oxygen atoms directed toward the substrate surface at (r=a,
`z=0) requires a directional derivative of oxygen atom concentration having a non-zero
`component in the z-direction. Physical continuity necessitates a continuous concentration
`gradient extending from the edge of the substrate into the external surrounding plasma medium
`beyond the edge of the cylinder at r=a, which is a boundary for the solution. In other words, in
`the true physical system, there is a vertical concentration gradient extending upward and outward
`from the edge of a substrate. The Battey isoconcentration boundary condition at r=a is
`inconsistent with this requirement, and is a fundamental reason why equation II cannot be used
`to evaluate an oxygen atom flux at the edge r=a of a wafer.
`
`Stated in yet another way, the isoconcentration boundary condition can be useful to
`approximate the concentration within the cylinder, but this approximation eliminates “shape”
`information about the derivative of concentration at the edge and beyond (and corresponding
`flux to edge of the substrate).
`
`
`
`
`
`4
`
`

`
`Inter Partes Review of U.S. Patent No. 5,711,849
`IPR2016-0466
`
`
`CERTIFICATE OF SERVICE
`
`The undersigned hereby certifies that the foregoing PATENT OWNER’S
`
`
`
`PRELIMINARY RESPONSE UNDER 37 C.F.R. § 42.107 was served by
`
`electronic mail on this day, April 27, 2016, on the following individuals:
`
`Michael Fleming
`(mfleming@irell.com)
`Samuel K. Lu
`(slu@irell.com)
`Kamran Vakili
`(kvakili@irell.com)
`IRELL & MANELLA LLP
`1800 Avenue of the Stars, Suite 900
`Los Angeles, CA 90067-4276
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`/Beata Ichou/
`Beata Ichou