DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Response to Amendment/Arguments
(Submitted on 11/14/2025)
Applicant’s arguments with respect to claim 1, 11 and 19 have been considered but are moot because the new ground of rejection does not rely on any reference applied in the prior rejection of record for any teaching or matter specifically challenged in the argument.
In regard to Claim Objection
As the applicant has clarified this on Page 8, the examiner REMOVES the objection.
In regard to 101 rejections
- On Page 9, the applicant argues that amended claims 1 and 19 reciting “depositing a material layer on a wafer using the hybrid machine learning model” which is a physical change and not an abstract idea.
Examiner’s Response:
Using a machine learning model is considered a computer function for performing prediction.
As machine learning model operate within a computer environment, processing data and making predictions based on complex algorithms, effectively simulating a physical change through calculations and analysis. Thus, the limitation is treated as a generic semiconductor processing tool that uses the output of a machine learning model. The applicant also amended to recite “experiments, wherein the experimental outputs include a pressure, deposition rate , and mole fraction of a radical oxidation ”. The examiner interprets that this a computer function (computer apparatus used for experimentation) for performing the experimentation.
In Conclusion, the applicant has not overcome the 101 rejection and the examiner hereby MAINTAINS the 101 rejections on claims 1 and 19 and on dependent claims 2-8 and 20.
In regard to 103 rejections
- On Pages 11-12, the applicant basically argues that the reference “Han” does not teach the amended claim limitation based “ a pressure, a deposition rate and mole fractions of a radical oxidation process”.
Examiner’s Response:
Reference “Han” does not call out all the process parameters ( deposition rate, pressure, and mole-fraction) and the neural network model in the reference. However, the examiner uses a new reference “King” to teach “pressure”, “deposition rate”, “mole fraction” and the “radical oxidation”.
In CONCLUSION, the examiner rejects claims 1-8, 11-17 and 19-20 under 103 and MOVE the application as FINAL REJECTION
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1-8 and 19-20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1:
According to the first part of the analysis, in the instant case, claims 1 and 19 are directed to a method claim and claim 11 are directed to a machine (tool chamber). Thus, each of the claims falls within one of the four statutory categories (i.e. process, machine, manufacture, or composition of matter).
In regard to claim 1: (Currently Amended)
Step 2A Prong 1:
“ identifying a first set of cases spanning a first range of process and/or hardware parameters”
is a mental step of data identification.
“ and correlating the model outputs with the experimental outputs (except machine learning
algorithm to provide [[the ]] a hybrid machine learning model) is a mental step of data
comparison.
Step 2A Prong 2:
“a method comprising” does not integrate the judicial exception into a practical application.
This additional element is merely using a computer as a tool to perform an abstract idea. See
MPEP 2106.05(h).
“ running experiments in a lab for the first set of cases” does not integrate
the judicial exception into a practical application. This additional element is merely using a
computer as a tool to perform an abstract idea. See MPEP 2106.05(h).
“ compiling experimental outputs from the experiments, wherein the experimental outputs
Include a pressure, a deposition rate, and mole fractions of a radical oxidation process;” does
not integrate the judicial exception into a practical application. This additional element is
merely using a computer as a tool to perform an abstract idea. See MPEP 2106.05(h).
“ running physics based simulations for the first set of cases” does not integrate
the judicial exception into a practical application. This additional element is merely using a
computer as a tool to perform an abstract idea. See MPEP 2106.05(h).
“ compiling model outputs from the experiments, wherein the experimental outputs
Include a pressure, a deposition rate, and mole fractions of a radical oxidation process;” does
not integrate the judicial exception into a practical application. This additional element is
merely using a computer as a tool to perform an abstract idea. See MPEP 2106.05(h).
This additional element is merely using a computer as a tool to perform an abstract idea. See
MPEP 2106.05(h).
“ using the hybrid machine learning model to provide a pressure, a deposition rate, and mole
fractions of a radical oxidation process to deposit a material layer on a wafer using radical
oxidation ” does not integrate the judicial exception into a practical application. This additional
element is merely using a computer as a tool to perform an abstract idea. See MPEP 2106.05(h).
“ and depositing the material layer on the substrate using the radical oxidation process, the
depositing comprising using the pressure, the deposition rate, and the mole fractions of
hybrid machine learning model” does not integrate the judicial exception into a practical
application. This additional element is merely using a computer as a tool to perform an abstract
idea. See MPEP 2106.05(h).
Step 2B:
“a method comprising” does not amount to significantly more than the judicial exception in the
claim. The additional element merely uses a computer as a tool to perform an abstract idea. See
MPEP 2106.05(h).
“ running experiments in a lab for the first set of cases” does not amount to
significantly more than the judicial exception in the claim. The additional element merely uses a
computer as a tool to perform an abstract idea. See MPEP 2106.05(h).
“ compiling experimental outputs from the experiments, wherein the experimental outputs
Include a pressure, a deposition rate, and mole fractions of a radical oxidation process;” does
not amount to significantly more than the judicial exception in the claim. The additional
element merely uses a computer as a tool to perform an abstract idea. See MPEP 2106.05(h).
“ running physics based simulations for the first set of cases” does not amount to
significantly more than the judicial exception in the claim. The additional element merely uses a
computer as a tool to perform an abstract idea. See MPEP 2106.05(h).
“ compiling model outputs from the experiments, wherein the experimental outputs
Include a pressure, a deposition rate, and mole fractions of a radical oxidation process;” does
not amount to significantly more than the judicial exception in the claim. The additional
element merely uses a computer as a tool to perform an abstract idea. See MPEP 2106.05(h).
“ using the hybrid machine learning model to provide a pressure, a deposition rate, and mole
fractions of a radical oxidation process to deposit a material layer on a wafer using radical
oxidation ” does not amount to significantly more than the judicial exception in the claim. The
additional element merely uses a computer as a tool to perform an abstract idea. See MPEP
2106.05(h).
“ and depositing the material layer on the substrate using the radical oxidation process, the
depositing comprising using the pressure, the deposition rate, and the mole fractions of
hybrid machine learning model” does not amount to significantly more than the judicial
exception in the claim. The additional element merely uses a computer as a tool to perform an
abstract idea. See MPEP 2106.05(h).
In regard to claim 2: (Original)
Step 2A Prong 2:
“ the physics based simulation is a reduced order physics simulation model” does not integrate
the judicial exception into a practical application. This additional element is merely using a
computer as a tool to perform an abstract idea. See MPEP 2106.05(h).
Step 2B:
“ the physics based simulation is a reduced order physics simulation model” does not amount to
significantly more than the judicial exception in the claim. The additional element merely uses a
computer as a tool to perform an abstract idea. See MPEP 2106.05(h).
In regard to claim 3: (Original)
Step 2A Prong 1:
“identifying a second set of cases spanning a second range of process and/or hardware parameters” is a mental step of data identification.
Additional Elements:
Step 2A Prong 2:
“running a physics based simulation for the second set of cases” does not integrate the judicial
exception into a practical application. This additional element is merely using a computer as a
tool to perform an abstract idea. See MPEP 2106.05(h).
“compiling outputs from the physics based simulation” does not integrate
the judicial exception into a practical application. This additional element is merely using a
computer as a tool to perform an abstract idea. See MPEP 2106.05(h).
“ the reduced order physics simulation model is generated by a method comprising” does not
integrate the judicial exception into a practical application. This additional element is merely
using a computer as a tool to perform an abstract idea. See MPEP 2106.05(h).
“ and using a second machine learning algorithm to generate the reduced order physics
simulation model” does not integrate the judicial exception into a practical application. This
additional element is merely using a computer as a tool to perform an abstract idea. See MPEP
2106.05(h).
Step 2B:
“running a physics based simulation for the second set of cases” does not amount to
significantly more than the judicial exception in the claim. This additional element is merely
using a computer as a tool to perform an abstract idea. See MPEP 2106.05(h).
“compiling outputs from the physics based simulation” does not amount to
significantly more than the judicial exception in the claim. This additional element is merely
using a computer as a tool to perform an abstract idea. See MPEP 2106.05(h).
“ the reduced order physics simulation model is generated by a method comprising” does not
amount to significantly more than the judicial exception in the claim. This additional element is
merely using a computer as a tool to perform an abstract idea. See MPEP 2106.05(h).
“ and using a second machine learning algorithm to generate the reduced order physics
simulation model” does not amount to significantly more than the judicial exception in the
claim. This additional element is merely using a computer as a tool to perform an abstract idea.
See MPEP 2106.05(h).
In regard to claim 4: (Original)
Step 2A Prong 1:
“the second set of cases is larger than the first set of cases” is a mental step of data comparison.
Step 2A Prong 2: no additional elements
Step 2B: no additional elements
In regard to claim 5: (Original)
Step 2A Prong 1:
“ the outputs from the physics based simulation comprise one or more of species concentrations, fluxes, and energies on wafer and/or additional quantities such as pressure, flow (velocity) and temperature at locations away from the wafer” is a mental step of data identification.
Step 2A Prong 2: no additional elements
Step 2B: no additional elements
In regard to claim 6: (Original)
Step 2A Prong 1:
“ selecting a new hardware and/or process condition” is a mental step of data selection.
Additional Elements:
Step 2A Prong 2:
“ evaluating the new hardware and/or process condition with the reduced order physics
simulation model” does not integrate the judicial exception into a practical application. This
additional element is merely using a computer as a tool to perform an abstract idea. See MPEP
2106.05(h).
“ evaluating the new hardware and/or process condition with the hybrid machine learning
model” does not integrate the judicial exception into a practical application. This additional
element is merely using a computer as a tool to perform an abstract idea. See MPEP 2106.05(h).
“ and predicting on-wafer results based on the evaluation of the reduced order physics
simulation model and the hybrid machine learning model” does not integrate
the judicial exception into a practical application. This additional element is merely using a
computer as a tool to perform an abstract idea. See MPEP 2106.05(h).
Step 2B:
“ evaluating the new hardware and/or process condition with the reduced order physics
simulation model” does not amount to significantly more than the judicial exception in the
claim. This additional element is merely using a computer as a tool to perform an abstract idea.
See MPEP 2106.05(h).
“ evaluating the new hardware and/or process condition with the hybrid machine learning
model” does not amount to significantly more than the judicial exception in the claim. This
additional element is merely using a computer as a tool to perform an abstract idea. See MPEP
2106.05(h).
“ and predicting on-wafer results based on the evaluation of the reduced order physics
simulation model and the hybrid machine learning model” does not amount to
significantly more than the judicial exception in the claim. This additional element is merely
using a computer as a tool to perform an abstract idea. See MPEP 2106.05(h).
In regard to claim 7: (Original)
Step 2A Prong 2:
“ the new hardware and/or process condition is on a tool different than the tool used to
generate the hybrid machine learning model” does not integrate the judicial exception into a
practical application. This additional element is merely using a computer as a tool to perform an
abstract idea. See MPEP 2106.05(h).
Step 2B:
“ the new hardware and/or process condition is on a tool different than the tool used to
generate the hybrid machine learning model” does not amount to significantly more than the
judicial exception in the claim. This additional element is merely using a computer as a tool to
perform an abstract idea. See MPEP 2106.05(h).
In regard to claim 8: (Original)
Step 2A Prong 1:
“ the model outputs comprise one or more of species concentrations, fluxes, and energies on wafer” is a mental step of data identification.
Step 2A Prong 2: no additional elements
Step 2B: no additional elements
In regard to claim 19: (Currently Amended)
Step 2A Prong 1:
“ identifying a second set of cases spanning a second range of process and/or hardware
parameters wherein the second set of cases is smaller than the first set of cases” is a mental
step of data identification and comparison.
“and correlating the model outputs with the experimental outputs with a second machine learning algorithm to provide the hybrid machine learning model” (except second machine learning algorithm to provide the hybrid machine learning model) is a mental step of data comparison.
Step 2A Prong 2:
“ A method, comprising:” does not integrate the judicial exception into a practical application.
This additional element is merely using a computer as a tool to perform an abstract idea. See
MPEP 2106.05(h).
“ running a physics based simulation for the first set of cases” does not
integrate the judicial exception into a practical application. This additional element is merely
using a computer as a tool to perform an abstract idea. See MPEP 2106.05(h).
“ using a machine learning algorithm to generate a reduced order physics simulation model”
does not integrate the judicial exception into a practical application. This additional element is
merely using a computer as a tool to perform an abstract idea. See MPEP 2106.05(h).
“running experiments in a lab for the second set of cases” does not
integrate the judicial exception into a practical application. This additional element is merely
using a computer as a tool to perform an abstract idea. See MPEP 2106.05(h).
“ compiling experimental outputs from the experiments, wherein the experimental outputs
include a pressure, a deposition rate and mole fraction of a radical oxidation” does not
integrate the judicial exception into a practical application. This additional element is merely
using a computer as a tool to perform an abstract idea. See MPEP 2106.05(h).
“ compiling model outputs from the simulations wherein the experimental outputs
include a pressure, a deposition rate and mole fraction of a radical oxidation” does not
integrate the judicial exception into a practical application. This additional element is merely
using a computer as a tool to perform an abstract idea. See MPEP 2106.05(h).
“using the hybrid machine learning model to provide a pressure, a deposition rate and mole
fraction of a radical oxidation to deposit a material layer on a wafer using radical
oxidation process” does not integrate the judicial exception into a practical application. This
additional element is merely using a computer as a tool to perform an abstract idea. See MPEP
2106.05(h).
Step 2B:
“ A method, comprising:” does not amount to significantly more than the judicial exception in
the claim. This additional element is merely using a computer as a tool to perform an abstract
idea. See MPEP 2106.05(h).
“ running a physics based simulation for the first set of cases” does not amount to significantly
more than the judicial exception in the claim. This additional element is merely using a
computer as a tool to perform an abstract idea. See MPEP 2106.05(h).
“ using a machine learning algorithm to generate a reduced order physics simulation model”
does not amount to significantly more than the judicial exception in the
claim. This additional element is merely using a computer as a tool to perform an abstract idea.
See MPEP 2106.05(h).
“running experiments in a lab for the second set of cases” does not amount to significantly more
than the judicial exception in the claim. This additional element is merely using a computer as a
tool to perform an abstract idea. See MPEP 2106.05(h).
“ compiling experimental outputs from the experiments, wherein the experimental outputs
include a pressure, a deposition rate and mole fraction of a radical oxidation” does not amount
to significantly more than the judicial exception in the claim. This additional element is merely
using a computer as a tool to perform an abstract idea. See MPEP 2106.05(h).
“ compiling model outputs from the simulations wherein the experimental outputs
include a pressure, a deposition rate and mole fraction of a radical oxidation” does not amount
to significantly more than the judicial exception in the claim. This additional element is merely
using a computer as a tool to perform an abstract idea. See MPEP 2106.05(h).
“using the hybrid machine learning model to provide a pressure, a deposition rate and mole
fraction of a radical oxidation to deposit a material layer on a wafer using radical
oxidation process” does not amount to significantly more than the judicial exception in the
claim. This additional element is merely using a computer as a tool to perform an abstract idea.
See MPEP 2106.05(h).
In regard to claim 20: (Original)
Step 2A Prong 1:
“ selecting a new hardware and/or process condition” is a mental step of data selection.
Additional Elements:
Step 2A Prong 2:
“ evaluating the new hardware and/or process condition with the reduced order physics
simulation model” does not integrate the judicial exception into a practical application. This
additional element is merely using a computer as a tool to perform an abstract idea. See MPEP
2106.05(h).
“ evaluating the new hardware and/or process condition with the hybrid machine learning
model” does not integrate the judicial exception into a practical application. This additional
element is merely using a computer as a tool to perform an abstract idea. See MPEP 2106.05(h).
“ and predicting on-wafer results based on the evaluation of the reduced order physics
simulation model and the hybrid machine learning model” does not
integrate the judicial exception into a practical application. This additional element is merely
using a computer as a tool to perform an abstract idea. See MPEP 2106.05(h).
Step 2B:
“ evaluating the new hardware and/or process condition with the reduced order physics
simulation model” does not amount to significantly more than the judicial exception in the
claim. This additional element is merely using a computer as a tool to perform an abstract idea.
See MPEP 2106.05(h).
“ evaluating the new hardware and/or process condition with the hybrid machine learning
model”. does not amount to significantly more than the judicial exception in the claim. This
additional element is merely using a computer as a tool to perform an abstract idea. See MPEP
2106.05(h).
“ and predicting on-wafer results based on the evaluation of the reduced order physics
simulation model and the hybrid machine learning model” does not amount to
significantly more than the judicial exception in the claim. This additional element is merely
using a computer as a tool to perform an abstract idea. See MPEP 2106.05(h).
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claims 1-8, and 19-20 are rejected under 35 U.S.C. 103 as being unpatentable
over
Yohan Kim et.al (hereinafter Kim) US 2022/0121800 A1 [Foreign Priority: KR 10-2020-0135524 Filed: 2020-10-19]
in view of David KING et.al (hereinafter King) US 2019/0062914 A1.
In regard to claim 1: (Currently Amended)
Kim discloses:
- A method, comprising: identifying a first set of cases spanning a first range of process and/or hardware parameters;
In [Abstract]:
classifying feature data of a target semiconductor device according to measurement conditions, generating first target data and second target data by preprocessing the first feature element data and the second feature element data, respectively, generating a first machine learning model using the first target data and extracting a second machine learning model using the second target data,
(BRI: the initial (pre-processing and refinement classification are the first and second set of cases)
In [0074]:
Machine learning models may be created by training (or learning) based on massive sample data, and the physical rule-based model may be generated by at least one rule defined based on physical laws or the like. Machine learning models and physical rule-based models may have different characteristics, and thus, different advantages and disadvantages in addition to different application fields. Therefore, in relation to the method of manufacturing an integrated circuit according to the present disclosure, in modeling the properties of the target semiconductor device, because hybrid models including first to i-th machine learning models MM1 to MMi and first to m-th physical rule-based models PM1 to PMm are used,
In [0034]:
The method of manufacturing an integrated circuit according to the inventive concepts of the present disclosure may generate a circuit model used to simulate electrical properties of the integrated circuit through machine learning. A more accurate circuit model corresponding to a process condition range between different process conditions for manufacturing semiconductor devices may be provided, and as the consistency of the circuit model is improved, the electrical properties of semiconductor devices included in the integrated circuit may be more accurately predicted.
In [0028]:
the circuit model may include a part implemented by logic hardware designed by logic synthesis. In this specification, the processor may refer to any hardware-implemented data processing device that includes a physically structured circuit to execute predefined operations
- running experiments in a lab for the first set of cases
In [Abstract]:
classifying feature data of a target semiconductor device according to measurement conditions, generating first target data and second target data by preprocessing the first feature element data and the second feature element data, respectively, generating a first machine learning model using the first target data and extracting a second machine learning model using the second target data,
In [0026]:
FIG. 1 is a flowchart showing a method of manufacturing an integrated circuit, according to some exemplary embodiments of the present disclosure. Specifically, the flowchart of FIG. 1 shows a method of designing an integrated circuit for performing a simulation of an integrated circuit using a circuit model.
- running physics based simulations for the first set of cases;
In [0076]:
A model compensation operation 244 for compensating the equivalent circuit EC to satisfy the laws of physics may be performed.
In [0066]:
in operation S245, an operation of programming the machine learning model as a circuit model to correspond to the simulation program may be performed. For example, a machine learning model finally constructed in the form of an equivalent circuit may be programmed into a circuit model in the form of an equivalent circuit using the model Application Program Interface (API) provided by commercial EDA software, so that EDA circuit analysis may be performed.
Kim does not explicitly disclose:
- compiling experimental outputs from the experiments, wherein the experimental outputs include a pressure, deposition rate and mole fraction of a radical oxidation process
- compiling model outputs from the simulations, wherein the models outputs include a pressure, deposition rate and mole fraction of a radical oxidation process
- and correlating the model outputs with the experimental outputs with a machine learning algorithm to provide the hybrid machine learning model;
- using the hybrid machine learning model to provide [[the]]a pressure, a deposition rate, and mole fraction
- depositing the material layer on the substrate using the radical oxidation process, depositing comprising using the pressure, the deposition rate, and the mole fractions of the hybrid machine learning model.
However, King discloses:
- compiling experimental outputs from the experiments, wherein the experimental outputs include a pressure, deposition rate and mole fraction of a radical oxidation process
In [0090]:
The terms “substrate,” “articles” and “materials” are used interchangeably herein.
In [0096]:
Various embodiments of the present technology described herein relates to systems, apparatus and methods for processing articles.
In [0235]:
For each target loading, the experimental results for actual Al and P deposited are shown in the Table 2.
PNG
media_image1.png
198
411
media_image1.png
Greyscale
In [0157]:
FIG. 2 shows a process flow diagram for one embodiment of FIG. 1A, including a synthesis subsystem, a pre-treatment subsystem, a two-step ALD coating processes in series, a post-treatment subsystem and a unit operation for collecting optimized materials, with common computer control over all critical processes and operating parameters. The process is designed to produce a composite powder product that is tailored and designed to achieve a value proposition for a customer in its end-use environment. Synthesis subsystem 101 can represent one of an array of particle synthesis systems, in which one or more precursor feedstocks are effectively delivered into a system through inlet assembly 102 at known flow rates, concentrations, temperatures, pressures, periodicity, measured and controlled through control ports 103, and the system outputs the synthesized materials through outlet assembly 104. The control ports may comprise one or more of i) valve-regulated ports designed for mass or material flow into or out of the system with optional filtration unit; ii) diagnostic ports for in-situ measurement capabilities for process matter and/or product monitoring; ii
PNG
media_image2.png
457
663
media_image2.png
Greyscale
In [0104]:
physicochemical parameters are considered critical wafer process parameters and characteristics in semiconductor manufacturing. They represent the chemical and physical properties of the wafer surface and materials that directly influence the performance, yield, and quality of the final device.
In [0205]:
In vapor deposition processes such as CVD, the particles can be contacted with two or more different reactants concurrently, or by one or more reactants that do not exhibit the self-limiting behavior characteristic of ALD and MLD processes. For a traditional batch CVD process, the primary methods of controlling reactions are limited to reactant exposure time and operating conditions such as process temperatures and pressures
In [209]:
substrate media provides heat transfer and a greater surface area to enhance vaporization. In another embodiment the precursor may be ionized or energized to a plasma.
Examiner’s BRI
(physicochemical parameters are considered critical wafer process parameters and characteristics in semiconductor manufacturing. They represent the chemical and physical properties of the wafer surface and materials. Pressure, deposition rate, and mole fraction in a radical oxidation process (such as plasma-enhanced CVD or advanced oxidation processes) represent key physicochemical parameters. They are measurable quantities that define the physical environment and chemical composition, governing the rate and nature of the reaction)
In [0008]:
In at least one embodiment, the apparatus further comprises at least one transport unit having one or more actuation mechanisms and configured for controlling the temperature, pressure and composition of a gaseous environment while regulating the material flow rate of a plurality of flowable articles.
Examiner’s BRI
(controlling the temperature, pressure, and composition of a gaseous environment while regulating the material flow rate of a plurality of flowable articles directly represents the active manipulation of the pressure and chemical environment on the wafer during semiconductor fabrication processes.
In [0091]:
The substrate or article can be any material which is chemically and/or thermally stable under the conditions of the deposition reaction. By “chemically” stable, it is meant that no more than 15% of the surface of the article undergoes any undesirable chemical reaction during the deposition process, other than in some cases bonding to the applied coating.
In [0180]:
FIG. 14B) in a coating subsystem 301, followed by a post-treatment process that provides for diffusion of the surface coating species and creates a penetrated coating region of greater thickness than the starting coated material (FIG. 14C).
In [0008] :
In at least one embodiment, the apparatus further comprises at least one transport unit having one or more actuation mechanisms and configured for controlling the temperature, pressure and composition of a gaseous environment while regulating the material flow rate of a plurality of flowable articles.
In [0168]:
For processes such as oxidation, reduction and etching, both surface area and reaction penetration depth that can define a volumetric value may be utilized as a critical parameter.
In [0168]:
These processes and potential critical parameters are not intended to be limiting to the invention but rather are included as a representative subset of parameters that have been qualitatively or quantitatively identified and monitored at the input and output of each subsystem of interest and permit machine learning for process
(Within the context of CVD, perhaps it is known to the POSITA that controlling the precursor or material flow rate is a standard technique for adjusting the deposition rate and thus regulating the material flow rate of a plurality of flowable articles directly represents deposition rate)
In [0212]:
Many formation methods are possible including decomposition initiated by temperature or pressure in the gas phase inside or prior to the reactor, at the substrate, reactor or manifold surfaces or by passing over an incorporated decomposition element like a hot filament or wire; or reaction with other gas phase precursors, reaction with surfaces inside or prior to the reactor, reaction with charged species, radicals or plasma; from pass through or nearby a plasma source, electron beam or ion beam.
Examiner’s BRI
(reactions with surfaces, charged species, and radicals inside a reactor, particularly when influenced by plasma, electron beams, or ion beams, are central examples of radical-induced oxidation (or radical-controlled plasma processes).
In [0020]:
flowable articles to be processed into the control system of a surface treatment system, thereby defining a first total surface area target c) providing a reactive precursor with which to treat the surfaces of said plurality of flowable articles, and entering into said control system the provided, estimated, measured or known number of moles of a reactive precursor required to saturate, react with or treat the entirety of the first total surface area target using empirical or estimated process conditions, thereby defining a complete saturation quantity, and d) selecting a target saturation ratio, to obtain a process recipe for a batch, semi-batch, semi-continuous or continuous surface treatment process, wherein said process recipe comprises at least one target pressure level associated with said target saturation ratio.
Examiner’s BRI
(does target saturation ratio based on a known complete saturation quantity (moles of precursor per total surface area) represents a mole fraction)
- compiling model outputs from the simulations, wherein the models outputs include a pressure, deposition rate and mole fraction of a radical oxidation process
In [0240]:
50 kilotons of powder are simulated in a system configured for synthesizing and coating materials to produce composite articles,
In [0240]:
Machine learning algorithm 376 receives critical inputs 364 for this material set
In [0034]:
In at least one embodiment, a machine learning algorithm calculates a subprocess deviation from modeled or empirical data with information derived from one or more of a direct in-situ signal, an indirect in-situ signal, a direct ex-situ signal
Examiner’s BRI
(powder is used as a coating material applied onto a solid component, which is referred to as the substrate. in modern advanced manufacturing systems designed for synthesizing and coating materials to produce composite articles, coating powders and their interaction with substrates are simulated, and Machine Learning (ML) is frequently used to process, interpret, and compile the outputs of these simulations. This is a simulation based training. Input from a simulation to an ML algorithm can be considered an indirect in-situ signal. The simulated data acts as an "indirect" because it is a computational model of reality rather than a direct physical measurement (experimental or measured)
In [0104]:
physicochemical parameters are considered critical wafer process parameters and characteristics in semiconductor manufacturing. They represent the chemical and physical properties of the wafer surface and materials that directly influence the performance, yield, and quality of the final device.
In [0205]:
In vapor deposition processes such as CVD, the particles can be contacted with two or more different reactants concurrently, or by one or more reactants that do not exhibit the self-limiting behavior characteristic of ALD and MLD processes. For a traditional batch CVD process, the primary methods of controlling reactions are limited to reactant exposure time and operating conditions such as process temperatures and pressures
In [209]:
substrate media provides heat transfer and a greater surface area to enhance vaporization. In another embodiment the precursor may be ionized or energized to a plasma.
In [0008]:
In at least one embodiment, the apparatus further comprises at least one transport unit having one or more actuation mechanisms and configured for controlling the temperature, pressure and composition of a gaseous environment while regulating the material flow rate of a plurality of flowable articles.
In [0091]:
The substrate or article can be any material which is chemically and/or thermally stable under the conditions of the deposition reaction. By “chemically” stable, it is meant that no more than 15% of the surface of the article undergoes any undesirable chemical reaction during the deposition process, other than in some cases bonding to the applied coating.
In [0180]:
FIG. 14B) in a coating subsystem 301, followed by a post-treatment process that provides for diffusion of the surface coating species and creates a penetrated coating region of greater thickness than the starting coated material (FIG. 14C).
In [0008] :
In at least one embodiment, the apparatus further comprises at least one transport unit having one or more actuation mechanisms and configured for controlling the temperature, pressure and composition of a gaseous environment while regulating the material flow rate of a plurality of flowable articles.
In [0168]:
For processes such as oxidation, reduction and etching, both surface area and reaction penetration depth that can define a volumetric value may be utilized as a critical parameter.
In [0168]:
These processes and potential critical parameters are not intended to be limiting to the invention but rather are included as a representative subset of parameters that have been qualitatively or quantitatively identified and monitored at the input and output of each subsystem of interest and permit machine learning for process
(Within the context of CVD, perhaps it is known to the POSITA that controlling the precursor or material flow rate is a standard technique for adjusting the deposition rate and thus regulating the material flow rate of a plurality of flowable articles directly represents deposition rate)
In [0212]:
Many formation methods are possible including decomposition initiated by temperature or pressure in the gas phase inside or prior to the reactor, at the substrate, reactor or manifold surfaces or by passing over an incorporated decomposition element like a hot filament or wire; or reaction with other gas phase precursors, reaction with surfaces inside or prior to the reactor, reaction with charged species, radicals or plasma; from pass through or nearby a plasma source, electron beam or ion beam.
In [0020]:
flowable articles to be processed into the control system of a surface treatment system, thereby defining a first total surface area target c) providing a reactive precursor with which to treat the surfaces of said plurality of flowable articles, and entering into said control system the provided, estimated, measured or known number of moles of a reactive precursor required to saturate, react with or treat the entirety of the first total surface area target using empirical or estimated process conditions, thereby defining a complete saturation quantity, and d) selecting a target saturation ratio, to obtain a process recipe for a batch, semi-batch, semi-continuous or continuous surface treatment process, wherein said process recipe comprises at least one target pressure level associated with said target saturation ratio.
- and correlating the model outputs with the experimental outputs with a machine learning algorithm to provide the hybrid machine learning model;
In [0139]:
The apparatus and methods described herein are suitable for processing a plurality of composite articles
In [0139]:
the method further includes a machine learning algorithm which calculates a subprocess deviation from modeled or empirical data with information derived from one or more of a direct in-situ signal
Examiner’s BRI
(an in-situ signal is an experimental data)
- using the hybrid machine learning model to provide a pressure, a deposition rate, the mole fractions to deposit a material layer on a wafer using the radical oxidation process; and
In [0034]:
In at least one embodiment, a machine learning algorithm calculates a subprocess deviation from modeled or empirical data with information derived from one or more of a direct in-situ signal
Examiner’s BRI
(machine learning (ML) algorithm that calculates a subprocess deviation from modeled using information derived from one or more direct in-situ signals (measured) represents a hybrid ML model)
In [0104]:
physicochemical parameters are considered critical wafer process parameters and characteristics in semiconductor manufacturing. They represent the chemical and physical properties of the wafer surface and materials that directly influence the performance, yield, and quality of the final device.
In [0205]:
In vapor deposition processes such as CVD, the particles can be contacted with two or more different reactants concurrently, or by one or more reactants that do not exhibit the self-limiting behavior characteristic of ALD and MLD processes. For a traditional batch CVD process, the primary methods of controlling reactions are limited to reactant exposure time and operating conditions such as process temperatures and pressures
In [209]:
substrate media provides heat transfer and a greater surface area to enhance vaporization. In another embodiment the precursor may be ionized or energized to a plasma.
In [0008]:
In at least one embodiment, the apparatus further comprises at least one transport unit having one or more actuation mechanisms and configured for controlling the temperature, pressure and composition of a gaseous environment while regulating the material flow rate of a plurality of flowable articles.
In [0091]:
The substrate or article can be any material which is chemically and/or thermally stable under the conditions of the deposition reaction. By “chemically” stable, it is meant that no more than 15% of the surface of the article undergoes any undesirable chemical reaction during the deposition process, other than in some cases bonding to the applied coating.
In [0180]:
FIG. 14B) in a coating subsystem 301, followed by a post-treatment process that provides for diffusion of the surface coating species and creates a penetrated coating region of greater thickness than the starting coated material (FIG. 14C).
In [0008] :
In at least one embodiment, the apparatus further comprises at least one transport unit having one or more actuation mechanisms and configured for controlling the temperature, pressure and composition of a gaseous environment while regulating the material flow rate of a plurality of flowable articles.
In [0168]:
For processes such as oxidation, reduction and etching, both surface area and reaction penetration depth that can define a volumetric value may be utilized as a critical parameter.
In [0168]:
These processes and potential critical parameters are not intended to be limiting to the invention but rather are included as a representative subset of parameters that have been qualitatively or quantitatively identified and monitored at the input and output of each subsystem of interest and permit machine learning for process
Examiner’s BRI
(Within the context of CVD, perhaps it is known to the POSITA that controlling the precursor or material flow rate is a standard technique for adjusting the deposition rate and thus regulating the material flow rate of a plurality of flowable articles directly represents deposition rate)
In [0212]:
Many formation methods are possible including decomposition initiated by temperature or pressure in the gas phase inside or prior to the reactor, at the substrate, reactor or manifold surfaces or by passing over an incorporated decomposition element like a hot filament or wire; or reaction with other gas phase precursors, reaction with surfaces inside or prior to the reactor, reaction with charged species, radicals or plasma; from pass through or nearby a plasma source, electron beam or ion beam.
In [0020]:
flowable articles to be processed into the control system of a surface treatment system, thereby defining a first total surface area target c) providing a reactive precursor with which to treat the surfaces of said plurality of flowable articles, and entering into said control system the provided, estimated, measured or known number of moles of a reactive precursor required to saturate, react with or treat the entirety of the first total surface area target using empirical or estimated process conditions, thereby defining a complete saturation quantity, and d) selecting a target saturation ratio, to obtain a process recipe for a batch, semi-batch, semi-continuous or continuous surface treatment process, wherein said process recipe comprises at least one target pressure level associated with said target saturation ratio.
- depositing the material layer on the substrate using the radical oxidation process, depositing comprising using the pressure, the deposition rate, and the mole fractions of the hybrid machine learning model.
In [0034]:
In at least one embodiment, a machine learning algorithm calculates a subprocess deviation from modeled or empirical data with information derived from one or more of a direct in-situ signal
Examiner’s BRI
(machine learning (ML) algorithm that calculates a subprocess deviation from modeled using information derived from one or more direct in-situ signals (measured) represents a hybrid ML model)
The theme of the invention is to predict using neural network model a pressure, a deposition rate and mole fraction of a radical oxidation and correlate it with an experimental data.
It would have obvious to one of ordinary skill in the art before the effective filing date of the present application to combine Kim, and King.
Kim teaches cases for experiments and simulation.
King teaches a radical oxidation and the deposition rate and pressure, pressure and
mole fraction of oxidation, the machine learning to correlate the model outputs to the experimental data.
One of ordinary skill would have motivation to combine Kim, and King that can provide machine learning to optimize the process conditions (King [0171]).
In regard to claim 2: (Original)
Kim discloses:
- the physics based simulation is a reduced order physics simulation model.
In [0020]:
FIG. 11 is a diagram for explaining an example of a model compression operation of FIG. 10, and a diagram for explaining a case where a machine learning model is a neural network model;
In [0064]:
A compression operation capable of reducing the complexity of the machine learning model may be performed while maintaining the consistency of the machine learning model.
In [0068]:
the model compression operation 242 may include an operation of reducing the number of model parameters of at least some of the first to i-th machine learning models MM1 to MMi.
In [0074]:
in relation to the method of manufacturing an integrated circuit according to the present disclosure, in modeling the properties of the target semiconductor device, because hybrid models including first to i-th machine learning models MM1 to MMi and first to m-th physical rule-based models PM1 to PMm are used.
In regard to claim 3: (Original)
Kim discloses:
- the reduced order physics simulation model is generated by a method comprising:
identifying a second set of cases spanning a second range of process and/or hardware parameters;
In [Abstract]:
classifying feature data of a target semiconductor device according to measurement conditions, generating first target data and second target data by preprocessing the first feature element data and the second feature element data, respectively, generating a first machine learning model using the first target data and extracting a second machine learning model using the second target data,
(BRI: the initial (pre-processing and refinement classification are the first and second set of cases)
In [0034]:
The method of manufacturing an integrated circuit according to the inventive concepts of the present disclosure may generate a circuit model used to simulate electrical properties of the integrated circuit through machine learning. A more accurate circuit model corresponding to a process condition range between different process conditions for manufacturing semiconductor devices may be provided, and as the consistency of the circuit model is improved, the electrical properties of semiconductor devices included in the integrated circuit may be more accurately predicted.
(BRI: the second range in context of second machine learning model)
- running a physics based simulation for the second set of cases;
In [0074]:
in relation to the method of manufacturing an integrated circuit according to the present disclosure, in modeling the properties of the target semiconductor device, because hybrid models including first to i-th machine learning models MM1 to MMi and first to m-th physical rule-based models PM1 to PMm are used,
in [0091]:
the simulation operation may be performed using the generated circuit model by performing the modeling operation (e.g., S20 of FIG. 1) of the target semiconductor device described with reference to FIGS. 1 to 13. A circuit model with improved consistency may be generated through machine learning operations, and because the consistency is improved and/or further improved by merging the machine learning model with the physical rule-based model according to the existing physical equation, the accuracy of the simulation result for the properties of the integrated circuit may be improved.
- compiling outputs from the physics based simulation;
In [0019]:
FIG. 10 is a diagram illustrating an example of an operation of mounting the machine learning model according to operations S242 to S244 of FIG. 9 as a circuit model in a simulation tool;
In [0037]:
feature data Y is classified into first to n-th feature element data Y1 to Yn, each of the first to n-th feature elements may be individually modeled.
- and using a second machine learning algorithm to generate the reduced order physics simulation model.
In [Abstract]:
classifying feature data of a target semiconductor device according to measurement conditions, generating first target data and second target data by preprocessing the first feature element data and the second feature element data, respectively, generating a first machine learning model using the first target data and extracting a second machine learning model using the second target data,
(BRI: the second machine learning is the second ML algorithm.
In [0064]:
A compression operation capable of reducing the complexity of the machine learning model may be performed while maintaining the consistency of the machine learning model.
In [0074]:
in relation to the method of manufacturing an integrated circuit according to the present disclosure, in modeling the properties of the target semiconductor device, because hybrid models including first to i-th machine learning models MM1 to MMi and first to m-th physical rule-based models PM1 to PMm are used.
In regard to claim 4: (Original)
Kim discloses:
- the second set of cases is larger than the first set of cases.
In [Abstract]:
classifying feature data of a target semiconductor device according to measurement conditions, generating first target data and second target data by preprocessing the first feature element data and the second feature element data, respectively, generating a first machine learning model using the first target data and extracting a second machine learning model using the second target data,
in [0036]:
One feature of the target semiconductor may be classified into first to n-th feature elements, and feature data Y may be classified into first to n-th feature element data Y1 to Yn corresponding to the respective first to n-th feature elements. Each of the first to n-th feature elements may be independent and may not affect each other.
in [0037] :
In some embodiments, n may be a natural number of 2 or more, but n may be 1 when one feature of the target semiconductor is difficult to be classified into a plurality of independent feature elements.
(BRI: when n is natural number of 2 representing the second case a case that can be classified independently (larger than the first set of cases)). It the number is 1 (smaller) than, then the classification is not independent)
In regard to claim 19: (Currently Amended)
Kim discloses:
- A method comprising: identifying a first set of cases spanning a first range of process and/or hardware parameters;
In [Abstract]:
classifying feature data of a target semiconductor device according to measurement conditions, generating first target data and second target data by preprocessing the first feature element data and the second feature element data, respectively, generating a first machine learning model using the first target data and extracting a second machine learning model using the second target data,
(BRI: the initial (pre-processing and refinement classification are the first and second set of cases)
In [0026]:
FIG. 1 is a flowchart showing a method of manufacturing an integrated circuit, according to some exemplary embodiments of the present disclosure. Specifically, the flowchart of FIG. 1 shows a method of designing an integrated circuit for performing a simulation of an integrated circuit using a circuit model.
- running a physics based simulation for the first set of cases;
In [0076]:
A model compensation operation 244 for compensating the equivalent circuit EC to satisfy the laws of physics may be performed.
In [0066]:
in operation S245, an operation of programming the machine learning model as a circuit model to correspond to the simulation program may be performed. For example, a machine learning model finally constructed in the form of an equivalent circuit may be programmed into a circuit model in the form of an equivalent circuit using the model Application Program Interface (API) provided by commercial EDA software, so that EDA circuit analysis may be performed.
- compiling outputs from the physics based simulation;
In [0029]:
In operation S21, an operation of classifying feature data of a semiconductor device based on a plurality of feature elements independent from each other may be performed, and accordingly, feature element data may be generated. One feature of a semiconductor device may be composed of a plurality of feature elements, and feature data for the one feature may be classified into different feature element data corresponding to each of the plurality of feature elements.
(BRI: classification of feature element is a model output)
- using a first machine learning algorithm to generate a reduced order physics simulation model;
In [Abstract]:
classifying feature data of a target semiconductor device according to measurement conditions, generating first target data and second target data by preprocessing the first feature element data and the second feature element data, respectively, generating a first machine learning model using the first target data and extracting a second machine learning model using the second target data,
(BRI: the first ML algorithm is context for first machine learning model)
In [0034]:
The method of manufacturing an integrated circuit according to the inventive concepts of the present disclosure may generate a circuit model used to simulate electrical properties of the integrated circuit through machine learning. A more accurate circuit model corresponding to a process condition range between different process conditions for manufacturing semiconductor devices may be provided, and as the consistency of the circuit model is improved, the electrical properties of semiconductor devices included in the integrated circuit may be more accurately predicted.
(BRI: the second range in context of second machine learning model)
In [0020]:
FIG. 11 is a diagram for explaining an example of a model compression operation of FIG. 10, and a diagram for explaining a case where a machine learning model is a neural network model;
In [0064]:
A compression operation capable of reducing the complexity of the machine learning model may be performed while maintaining the consistency of the machine learning model.
In [0068]:
the model compression operation 242 may include an operation of reducing the number of model parameters of at least some of the first to i-th machine learning models MM1 to MMi.
In [0074]:
in relation to the method of manufacturing an integrated circuit according to the present disclosure, in modeling the properties of the target semiconductor device, because hybrid models including first to i-th machine learning models MM1 to MMi and first to m-th physical rule-based models PM1 to PMm are used,
- identifying a second set of cases spanning a second range of process and/or hardware parameters
In [Abstract]:
classifying feature data of a target semiconductor device according to measurement conditions, generating first target data and second target data by preprocessing the first feature element data and the second feature element data, respectively, generating a first machine learning model using the first target data and extracting a second machine learning model using the second target data,
(BRI: the initial (pre-processing and refinement classification are the first and second set of cases)
In [0034]:
The method of manufacturing an integrated circuit according to the inventive concepts of the present disclosure may generate a circuit model used to simulate electrical properties of the integrated circuit through machine learning. A more accurate circuit model corresponding to a process condition range between different process conditions for manufacturing semiconductor devices may be provided, and as the consistency of the circuit model is improved, the electrical properties of semiconductor devices included in the integrated circuit may be more accurately predicted.
In [0028]:
the circuit model may include a part implemented by logic hardware designed by logic synthesis. In this specification, the processor may refer to any hardware-implemented data processing device that includes a physically structured circuit to execute predefined operations
- wherein the second set of cases is smaller than the first set of cases;
In [Abstract]:
classifying feature data of a target semiconductor device according to measurement conditions, generating first target data and second target data by preprocessing the first feature element data and the second feature element data, respectively, generating a first machine learning model using the first target data and extracting a second machine learning model using the second target data,
in [0036]:
One feature of the target semiconductor may be classified into first to n-th feature elements, and feature data Y may be classified into first to n-th feature element data Y1 to Yn corresponding to the respective first to n-th feature elements. Each of the first to n-th feature elements may be independent and may not affect each other.
in [0037] :
In some embodiments, n may be a natural number of 2 or more, but n may be 1 when one feature of the target semiconductor is difficult to be classified into a plurality of independent feature elements.
(BRI: It the number is 1, then the classification for the second set of cases are not independent)
- running experiments in a lab for the second set of cases;
In [Abstract]:
classifying feature data of a target semiconductor device according to measurement conditions, generating first target data and second target data by preprocessing the first feature element data and the second feature element data, respectively, generating a first machine learning model using the first target data and extracting a second machine learning model using the second target data,
(BRI: the initial (pre-processing and refinement classification using the second machine learning model are the first and second set of cases respectively)
- running physics based simulations for the second set of cases, wherein the physics based simulations use the reduced order physics simulation model;
In [Abstract]:
classifying feature data of a target semiconductor device according to measurement conditions, generating first target data and second target data by preprocessing the first feature element data and the second feature element data, respectively, generating a first machine learning model using the first target data and extracting a second machine learning model using the second target data,
In [0020]:
FIG. 11 is a diagram for explaining an example of a model compression operation of FIG. 10, and a diagram for explaining a case where a machine learning model is a neural network model;
In [0064]:
A compression operation capable of reducing the complexity of the machine learning model may be performed while maintaining the consistency of the machine learning model.
In [0068]:
the model compression operation 242 may include an operation of reducing the number of model parameters of at least some of the first to i-th machine learning models MM1 to MMi.
In [0074]:
in relation to the method of manufacturing an integrated circuit according to the present disclosure, in modeling the properties of the target semiconductor device, because hybrid models including first to i-th machine learning models MM1 to MMi and first to m-th physical rule-based models PM1 to PMm are used,
in [0091]:
the simulation operation may be performed using the generated circuit model by performing the modeling operation (e.g., S20 of FIG. 1) of the target semiconductor device described with reference to FIGS. 1 to 13. A circuit model with improved consistency may be generated through machine learning operations, and because the consistency is improved and/or further improved by merging the machine learning model with the physical rule-based model according to the existing physical equation, the accuracy of the simulation result for the properties of the integrated circuit may be improved.
Kim does not explicitly disclose:
- compiling experimental outputs from the experiment, wherein the experimental outputs include a pressure, a deposition and mole fractions of a radical oxidation
- compiling model outputs with the experimental outputs include a pressure, a deposition and mole fractions of a radical oxidation
- correlating the model outputs with the experimental outputs with a second machine learning algorithm to provide [[the]] a hybrid machine learning model;
- using the hybrid machine learning model to a pressure, a deposition and mole fractions of a radical oxidation to deposit a material layer on a wafer using radical oxidation
- depositing the material layer on the substrate using the radical oxidation process, the depositing comprising using the pressure, the deposition rate, and the mole fractions of the hybrid machine learning model.
However, King discloses:
- compiling experimental outputs from the experiments, wherein the experimental outputs include a pressure, deposition rate and mole fraction of a radical oxidation process
In [0090]:
The terms “substrate,” “articles” and “materials” are used interchangeably herein.
In [0096]:
Various embodiments of the present technology described herein relates to systems, apparatus and methods for processing articles.
In [0235]:
For each target loading, the experimental results for actual Al and P deposited are shown in the Table 2.
PNG
media_image1.png
198
411
media_image1.png
Greyscale
In [0157]:
FIG. 2 shows a process flow diagram for one embodiment of FIG. 1A, including a synthesis subsystem, a pre-treatment subsystem, a two-step ALD coating processes in series, a post-treatment subsystem and a unit operation for collecting optimized materials, with common computer control over all critical processes and operating parameters. The process is designed to produce a composite powder product that is tailored and designed to achieve a value proposition for a customer in its end-use environment. Synthesis subsystem 101 can represent one of an array of particle synthesis systems, in which one or more precursor feedstocks are effectively delivered into a system through inlet assembly 102 at known flow rates, concentrations, temperatures, pressures, periodicity, measured and controlled through control ports 103, and the system outputs the synthesized materials through outlet assembly 104. The control ports may comprise one or more of i) valve-regulated ports designed for mass or material flow into or out of the system with optional filtration unit; ii) diagnostic ports for in-situ measurement capabilities for process matter and/or product monitoring; ii
PNG
media_image2.png
457
663
media_image2.png
Greyscale
In [0104]:
physicochemical parameters are considered critical wafer process parameters and characteristics in semiconductor manufacturing. They represent the chemical and physical properties of the wafer surface and materials that directly influence the performance, yield, and quality of the final device.
In [0205]:
In vapor deposition processes such as CVD, the particles can be contacted with two or more different reactants concurrently, or by one or more reactants that do not exhibit the self-limiting behavior characteristic of ALD and MLD processes. For a traditional batch CVD process, the primary methods of controlling reactions are limited to reactant exposure time and operating conditions such as process temperatures and pressures
In [209]:
substrate media provides heat transfer and a greater surface area to enhance vaporization. In another embodiment the precursor may be ionized or energized to a plasma.
Examiner’s BRI
(physicochemical parameters are considered critical wafer process parameters and characteristics in semiconductor manufacturing. They represent the chemical and physical properties of the wafer surface and materials. Pressure, deposition rate, and mole fraction in a radical oxidation process (such as plasma-enhanced CVD or advanced oxidation processes) represent key physicochemical parameters. They are measurable quantities that define the physical environment and chemical composition, governing the rate and nature of the reaction)
In [0008]:
In at least one embodiment, the apparatus further comprises at least one transport unit having one or more actuation mechanisms and configured for controlling the temperature, pressure and composition of a gaseous environment while regulating the material flow rate of a plurality of flowable articles.
Examiner’s BRI
(controlling the temperature, pressure, and composition of a gaseous environment while regulating the material flow rate of a plurality of flowable articles directly represents the active manipulation of the pressure and chemical environment on the wafer during semiconductor fabrication processes.
In [0091]:
The substrate or article can be any material which is chemically and/or thermally stable under the conditions of the deposition reaction. By “chemically” stable, it is meant that no more than 15% of the surface of the article undergoes any undesirable chemical reaction during the deposition process, other than in some cases bonding to the applied coating.
In [0180]:
FIG. 14B) in a coating subsystem 301, followed by a post-treatment process that provides for diffusion of the surface coating species and creates a penetrated coating region of greater thickness than the starting coated material (FIG. 14C).
In [0008] :
In at least one embodiment, the apparatus further comprises at least one transport unit having one or more actuation mechanisms and configured for controlling the temperature, pressure and composition of a gaseous environment while regulating the material flow rate of a plurality of flowable articles.
In [0168]:
For processes such as oxidation, reduction and etching, both surface area and reaction penetration depth that can define a volumetric value may be utilized as a critical parameter.
In [0168]:
These processes and potential critical parameters are not intended to be limiting to the invention but rather are included as a representative subset of parameters that have been qualitatively or quantitatively identified and monitored at the input and output of each subsystem of interest and permit machine learning for process
(Within the context of CVD, perhaps it is known to the POSITA that controlling the precursor or material flow rate is a standard technique for adjusting the deposition rate and thus regulating the material flow rate of a plurality of flowable articles directly represents deposition rate)
In [0212]:
Many formation methods are possible including decomposition initiated by temperature or pressure in the gas phase inside or prior to the reactor, at the substrate, reactor or manifold surfaces or by passing over an incorporated decomposition element like a hot filament or wire; or reaction with other gas phase precursors, reaction with surfaces inside or prior to the reactor, reaction with charged species, radicals or plasma; from pass through or nearby a plasma source, electron beam or ion beam.
Examiner’s BRI
(reactions with surfaces, charged species, and radicals inside a reactor, particularly when influenced by plasma, electron beams, or ion beams, are central examples of radical-induced oxidation (or radical-controlled plasma processes).
In [0020]:
flowable articles to be processed into the control system of a surface treatment system, thereby defining a first total surface area target c) providing a reactive precursor with which to treat the surfaces of said plurality of flowable articles, and entering into said control system the provided, estimated, measured or known number of moles of a reactive precursor required to saturate, react with or treat the entirety of the first total surface area target using empirical or estimated process conditions, thereby defining a complete saturation quantity, and d) selecting a target saturation ratio, to obtain a process recipe for a batch, semi-batch, semi-continuous or continuous surface treatment process, wherein said process recipe comprises at least one target pressure level associated with said target saturation ratio.
Examiner’s BRI
(target saturation ratio based on a known complete saturation quantity (moles of precursor per total surface area) represents a mole fraction)
- compiling model outputs from the simulations, wherein the models outputs include a pressure, deposition rate and mole fraction of a radical oxidation process
In [0240]:
50 kilotons of powder are simulated in a system configured for synthesizing and coating materials to produce composite articles,
In [0240]:
Machine learning algorithm 376 receives critical inputs 364 for this material set
In [0034]:
In at least one embodiment, a machine learning algorithm calculates a subprocess deviation from modeled or empirical data with information derived from one or more of a direct in-situ signal, an indirect in-situ signal, a direct ex-situ signal
Examiner’s BRI
(powder is used as a coating material applied onto a solid component, which is referred to as the substrate. in modern advanced manufacturing systems designed for synthesizing and coating materials to produce composite articles, coating powders and their interaction with substrates are simulated, and Machine Learning (ML) is frequently used to process, interpret, and compile the outputs of these simulations. This is a simulation based training. Input from a simulation to an ML algorithm can be considered an indirect in-situ signal. The simulated data acts as an "indirect" because it is a computational model of reality rather than a direct physical measurement (experimental or measured)
In [0104]:
physicochemical parameters are considered critical wafer process parameters and characteristics in semiconductor manufacturing. They represent the chemical and physical properties of the wafer surface and materials that directly influence the performance, yield, and quality of the final device.
In [0205]:
In vapor deposition processes such as CVD, the particles can be contacted with two or more different reactants concurrently, or by one or more reactants that do not exhibit the self-limiting behavior characteristic of ALD and MLD processes. For a traditional batch CVD process, the primary methods of controlling reactions are limited to reactant exposure time and operating conditions such as process temperatures and pressures
In [209]:
substrate media provides heat transfer and a greater surface area to enhance vaporization. In another embodiment the precursor may be ionized or energized to a plasma.
In [0008]:
In at least one embodiment, the apparatus further comprises at least one transport unit having one or more actuation mechanisms and configured for controlling the temperature, pressure and composition of a gaseous environment while regulating the material flow rate of a plurality of flowable articles.
In [0091]:
The substrate or article can be any material which is chemically and/or thermally stable under the conditions of the deposition reaction. By “chemically” stable, it is meant that no more than 15% of the surface of the article undergoes any undesirable chemical reaction during the deposition process, other than in some cases bonding to the applied coating.
In [0180]:
FIG. 14B) in a coating subsystem 301, followed by a post-treatment process that provides for diffusion of the surface coating species and creates a penetrated coating region of greater thickness than the starting coated material (FIG. 14C).
In [0008] :
In at least one embodiment, the apparatus further comprises at least one transport unit having one or more actuation mechanisms and configured for controlling the temperature, pressure and composition of a gaseous environment while regulating the material flow rate of a plurality of flowable articles.
In [0168]:
For processes such as oxidation, reduction and etching, both surface area and reaction penetration depth that can define a volumetric value may be utilized as a critical parameter.
In [0168]:
These processes and potential critical parameters are not intended to be limiting to the invention but rather are included as a representative subset of parameters that have been qualitatively or quantitatively identified and monitored at the input and output of each subsystem of interest and permit machine learning for process
(Within the context of CVD, perhaps it is known to the POSITA that controlling the precursor or material flow rate is a standard technique for adjusting the deposition rate and thus regulating the material flow rate of a plurality of flowable articles directly represents deposition rate)
In [0212]:
Many formation methods are possible including decomposition initiated by temperature or pressure in the gas phase inside or prior to the reactor, at the substrate, reactor or manifold surfaces or by passing over an incorporated decomposition element like a hot filament or wire; or reaction with other gas phase precursors, reaction with surfaces inside or prior to the reactor, reaction with charged species, radicals or plasma; from pass through or nearby a plasma source, electron beam or ion beam.
In [0020]:
flowable articles to be processed into the control system of a surface treatment system, thereby defining a first total surface area target c) providing a reactive precursor with which to treat the surfaces of said plurality of flowable articles, and entering into said control system the provided, estimated, measured or known number of moles of a reactive precursor required to saturate, react with or treat the entirety of the first total surface area target using empirical or estimated process conditions, thereby defining a complete saturation quantity, and d) selecting a target saturation ratio, to obtain a process recipe for a batch, semi-batch, semi-continuous or continuous surface treatment process, wherein said process recipe comprises at least one target pressure level associated with said target saturation ratio.
- correlating the model outputs with the experimental outputs with a second machine learning algorithm to provide [[the]] a hybrid machine learning model;
In [0139]:
The apparatus and methods described herein are suitable for processing a plurality of composite articles
In [0139]:
the method further includes a machine learning algorithm which calculates a subprocess deviation from modeled or empirical data with information derived from one or more of a direct in-situ signal
Examiner’s BRI
(an in-situ signal is an experimental data)
- using the hybrid machine learning model to provide a pressure, a deposition rate, and mole fraction to deposit a material layer on a wafer using a radical oxidation process
In [0034]:
In at least one embodiment, a machine learning algorithm calculates a subprocess deviation from modeled or empirical data with information derived from one or more of a direct in-situ signal
Examiner’s BRI
(machine learning (ML) algorithm that calculates a subprocess deviation from modeled using information derived from one or more direct in-situ signals (measured) represents a hybrid ML model)
In [0104]:
physicochemical parameters are considered critical wafer process parameters and characteristics in semiconductor manufacturing. They represent the chemical and physical properties of the wafer surface and materials that directly influence the performance, yield, and quality of the final device.
In [0205]:
In vapor deposition processes such as CVD, the particles can be contacted with two or more different reactants concurrently, or by one or more reactants that do not exhibit the self-limiting behavior characteristic of ALD and MLD processes. For a traditional batch CVD process, the primary methods of controlling reactions are limited to reactant exposure time and operating conditions such as process temperatures and pressures
In [209]:
substrate media provides heat transfer and a greater surface area to enhance vaporization. In another embodiment the precursor may be ionized or energized to a plasma.
In [0008]:
In at least one embodiment, the apparatus further comprises at least one transport unit having one or more actuation mechanisms and configured for controlling the temperature, pressure and composition of a gaseous environment while regulating the material flow rate of a plurality of flowable articles.
In [0091]:
The substrate or article can be any material which is chemically and/or thermally stable under the conditions of the deposition reaction. By “chemically” stable, it is meant that no more than 15% of the surface of the article undergoes any undesirable chemical reaction during the deposition process, other than in some cases bonding to the applied coating.
In [0180]:
FIG. 14B) in a coating subsystem 301, followed by a post-treatment process that provides for diffusion of the surface coating species and creates a penetrated coating region of greater thickness than the starting coated material (FIG. 14C).
In [0008] :
In at least one embodiment, the apparatus further comprises at least one transport unit having one or more actuation mechanisms and configured for controlling the temperature, pressure and composition of a gaseous environment while regulating the material flow rate of a plurality of flowable articles.
In [0168]:
For processes such as oxidation, reduction and etching, both surface area and reaction penetration depth that can define a volumetric value may be utilized as a critical parameter.
In [0168]:
These processes and potential critical parameters are not intended to be limiting to the invention but rather are included as a representative subset of parameters that have been qualitatively or quantitatively identified and monitored at the input and output of each subsystem of interest and permit machine learning for process
Examiner’s BRI
(Within the context of CVD, perhaps it is known to the POSITA that controlling the precursor or material flow rate is a standard technique for adjusting the deposition rate and thus regulating the material flow rate of a plurality of flowable articles directly represents deposition rate)
In [0212]:
Many formation methods are possible including decomposition initiated by temperature or pressure in the gas phase inside or prior to the reactor, at the substrate, reactor or manifold surfaces or by passing over an incorporated decomposition element like a hot filament or wire; or reaction with other gas phase precursors, reaction with surfaces inside or prior to the reactor, reaction with charged species, radicals or plasma; from pass through or nearby a plasma source, electron beam or ion beam.
In [0020]:
flowable articles to be processed into the control system of a surface treatment system, thereby defining a first total surface area target c) providing a reactive precursor with which to treat the surfaces of said plurality of flowable articles, and entering into said control system the provided, estimated, measured or known number of moles of a reactive precursor required to saturate, react with or treat the entirety of the first total surface area target using empirical or estimated process conditions, thereby defining a complete saturation quantity, and d) selecting a target saturation ratio, to obtain a process recipe for a batch, semi-batch, semi-continuous or continuous surface treatment process, wherein said process recipe comprises at least one target pressure level associated with said target saturation ratio.
- depositing the material layer on the substrate using the radical oxidation process, the depositing comprising using the pressure, the deposition rate, and the mole fractions of the hybrid machine learning model.
In [0034]:
In at least one embodiment, a machine learning algorithm calculates a subprocess deviation from modeled or empirical data with information derived from one or more of a direct in-situ signal
Examiner’s BRI
(machine learning (ML) algorithm that calculates a subprocess deviation from modeled using information derived from one or more direct in-situ signals (measured) represents a hybrid ML model)
The theme of the invention is to predict using neural network model a pressure, a deposition rate and mole fraction of a radical oxidation and correlate it with an experimental data.
It would have obvious to one of ordinary skill in the art before the effective filing date of the present application to combine Kim, and King.
Kim teaches cases for experiments and simulation.
King teaches a radical oxidation and the deposition rate and pressure, pressure and
mole fraction of oxidation, the machine learning to correlate the model outputs to the experimental data.
One of ordinary skill would have motivation to combine Kim, and King that can provide machine learning to optimize the process conditions (King [0171]).
The theme of the invention is to predict using neural network model a pressure, a deposition rate and mole fraction of a radical oxidation and correlate it with an experimental data.
It would have obvious to one of ordinary skill in the art before the effective filing date of the present application to combine Kim, and King.
Kim teaches cases for experiments and simulation.
King teaches a radical oxidation and the deposition rate and pressure, pressure and
mole fraction of oxidation, the machine learning to correlate the model outputs to the experimental data.
One of ordinary skill would have motivation to combine Kim, and King that can provide machine learning to optimize the process conditions (King [0171]).
In regard to claim 20: (Original)
Kim discloses:
- selecting a new hardware and/or process condition;
In [0092]:
In operation S50, based on the simulation result, an operation of manufacturing an integrated circuit including the target semiconductor device by a semiconductor process may be performed. For example, the integrated circuit may be manufactured by a semiconductor process to which process parameters finally adjusted in operation S50 are applied.
- evaluating the new hardware and/or process condition with the reduced order physics simulation model;
In [0076]:
A model compensation operation 244 for compensating the equivalent circuit EC to satisfy the laws of physics may be performed. The model compensation operation 244 may include an operation of correcting at least one of the first to i-th compressed models CM1 to CMi and the first to m-th physical rule-based models PM1 to PMm constituting the equivalent circuit EC
In [0076]:
A modified circuit model RSCM may be generated by the model compensation operation 244.
In [0065]:
After the machine learning model satisfies the reference condition or the model parameters are compressed, in operation S243, an operation of configuring an equivalent circuit corresponding to the target semiconductor device may be performed using a machine learning model. In operation S244, an operation of modifying the machine learning model may be performed so that the equivalent circuit satisfies the laws of physics.
BRI: compensating an Equivalent Circuit (EC) to satisfy the laws of physics using compressed models is a form of reduced-order physics modeling
- evaluating the new hardware and/or process condition with the hybrid machine learning model;
In [0065]:
After the machine learning model satisfies the reference condition or the model parameters are compressed, in operation S243, an operation of configuring an equivalent circuit corresponding to the target semiconductor device may be performed using a machine learning model. In operation S244, an operation of modifying the machine learning model may be performed so that the equivalent circuit satisfies the laws of physics..
In [0068]:
Referring to FIG. 10, a model compression operation 242 for compressing at least a portion of the first to i-th machine learning models MM1 to MMi may be performed. For example, an operation 242 of compressing a machine learning model that does not satisfy a reference condition among the first to i-th machine learning models MM1 to MMi may be performed.
- and predicting on-wafer results based on the evaluation of the reduced order physics simulation model and the hybrid machine learning model.
In [0034]:
A more accurate circuit model corresponding to a process condition range between different process conditions for manufacturing semiconductor devices may be provided, and as the consistency of the circuit model is improved, the electrical properties of semiconductor devices included in the integrated circuit may be more accurately predicted.
Claims 5-8 are rejected under 35 U.S.C. 103 as being unpatentable
Yohan Kim et.al (hereinafter Kim) US 2022/0121800 A1 [Foreign Priority: KR 10-2020-0135524 Filed: 2020-10-19]
in view of David KING et.al (hereinafter King) US 2019/0062914 A1.
further in view of Jen-Shiang WANG et.al (hereinafter WANG) US 2022/0260921 A1,
In regard to claim 5: (Original)
Kim, and King do not explicitly disclose:
- the outputs from the physics based simulation comprise one or more of species concentrations, fluxes, and energies on wafer and/or additional quantities such as pressure, flow (velocity) and temperature at locations away from the wafer.
However, WANG discloses:
- the outputs from the physics based simulation comprise one or more of species concentrations, fluxes, and energies on wafer and/or additional quantities such as pressure, flow (velocity) and temperature at locations away from the wafer.
In [0115]:
Returning to FIG. 5, some simulators exist that are able to numerically model a complex operation within a patterning process flow such as etching.
In [0115]:
Rigorous modelling of an operation within a patterning process flow (e.g., etching) typically requires simultaneously solving very different physical equations (plasma physics in the reactor, particle transport, surface chemistry, electromagnetic equations, surface time evolution, etc.). These often require different discretization schemes. As a consequence, there is not a unified framework that can accommodate a full modelling flow. This may cause a myriad of parameters that need to be modeled and/or account for, which causes prohibitive simulation times. In addition, the specific parameters of a simulation are typically unknown a priori, and many trials may be required to tune model parameters. For example, a set of ions possibly present in one particular plasma gas experiment is not known, and may be guessed until results match observations. Typically, a developed gas-phase chemistry includes tens of species and hundreds of chemical reactions.
In [0122]:
In some embodiments, front end 502 (and/or front end prep portion 506) is configured to determine incoming flux at individual points on a modeled surface of the wafer, at an interface between the wafer and an environment around the wafer
In [0124]:
topography simulations performed as a moving surface problem (e.g., as described herein), where surface velocities may be determined from a particle flux of particles arriving at a simulated surface (e.g., etch, deposition, and/or other simulations), a representation of the simulated topography may be a (discretization of a) level set function, for example.
in [0156]:
The concepts disclosed herein may simulate or mathematically model any generic imaging system for imaging sub wavelength features, and may be especially useful with emerging imaging technologies capable of producing increasingly shorter wavelengths.
In [0156]:
EUV lithography is capable of producing wavelengths within a range of 20-5 nm by using a synchrotron or by hitting a material (either solid or a plasma) with high energy electrons
The theme of the invention is to predict using neural network model a pressure, a deposition rate and mole fraction of a radical oxidation and correlate it with an experimental data.
It would have obvious to one of ordinary skill in the art before the effective filing date of the present application to combine Kim, , King and WANG.
Kim teaches cases for experiments and simulation.
King teaches a radical oxidation and the deposition rate and pressure, pressure and
mole fraction of oxidation, the machine learning to correlate the model outputs to the experimental data.
WANG teaches simulation species.
One of ordinary skill would have motivation to combine Kim, King, and WANG that can productivity improvement for manufacturing processes (WANG [0046].
In regard to claim 6: (Original)
Kim discloses:
- selecting a new hardware and/or process condition;
In [0092]:
In operation S50, based on the simulation result, an operation of manufacturing an integrated circuit including the target semiconductor device by a semiconductor process may be performed. For example, the integrated circuit may be manufactured by a semiconductor process to which process parameters finally adjusted in operation S50 are applied.
- evaluating the new hardware and/or process condition with the reduced order physics simulation model;
In [0076]:
A model compensation operation 244 for compensating the equivalent circuit EC to satisfy the laws of physics may be performed. The model compensation operation 244 may include an operation of correcting at least one of the first to i-th compressed models CM1 to CMi and the first to m-th physical rule-based models PM1 to PMm constituting the equivalent circuit EC
In [0076]:
A modified circuit model RSCM may be generated by the model compensation operation 244.
In [0065]:
After the machine learning model satisfies the reference condition or the model parameters are compressed, in operation S243, an operation of configuring an equivalent circuit corresponding to the target semiconductor device may be performed using a machine learning model. In operation S244, an operation of modifying the machine learning model may be performed so that the equivalent circuit satisfies the laws of physics.
BRI: compensating an Equivalent Circuit (EC) to satisfy the laws of physics using compressed models is a form of reduced-order physics modeling
- evaluating the new hardware and/or process condition with the hybrid machine learning model;
In [0065]:
After the machine learning model satisfies the reference condition or the model parameters are compressed, in operation S243, an operation of configuring an equivalent circuit corresponding to the target semiconductor device may be performed using a machine learning model. In operation S244, an operation of modifying the machine learning model may be performed so that the equivalent circuit satisfies the laws of physics..
In [0068]:
Referring to FIG. 10, a model compression operation 242 for compressing at least a portion of the first to i-th machine learning models MM1 to MMi may be performed. For example, an operation 242 of compressing a machine learning model that does not satisfy a reference condition among the first to i-th machine learning models MM1 to MMi may be performed.
- and predicting on-wafer results based on the evaluation of the reduced order physics simulation model and the hybrid machine learning model.
In [0034]:
A more accurate circuit model corresponding to a process condition range between different process conditions for manufacturing semiconductor devices may be provided, and as the consistency of the circuit model is improved, the electrical properties of semiconductor devices included in the integrated circuit may be more accurately predicted.
In regard to claim 7: (Original)
Kim discloses:
- the new hardware and/or process condition is on a tool different than the tool used to generate the hybrid machine learning model.
In [0019]:
FIG. 10 is a diagram illustrating an example of an operation of mounting the machine learning model according to operations S242 to S244 of FIG. 9 as a circuit model in a simulation tool;
In [0032]:
In operation S24, an operation of generating a circuit model may be performed using the extracted machine learning models. The circuit model may be programmed to be mounted on a simulation tool for simulating an integrated circuit. The simulation tool may be an Electronic Design Automation (EDA) tool. For example, the EDA tool may be a Simulation Program with Integrated Circuit Emphasis (SPICE) tool. The extracted machine learning model may be included in a Process Design Kit (PDK) used in the tool.
In regard to claim 8: (Original)
Kim, and King do not explicitly disclose:
- the model outputs comprise one or more of species concentrations, fluxes, and energies on wafer.
However, WANG discloses:
- the model outputs comprise one or more of species concentrations, fluxes, and energies on wafer.
In [0115]:
Returning to FIG. 5, some simulators exist that are able to numerically model a complex operation within a patterning process flow such as etching.
In [0115]:
Rigorous modelling of an operation within a patterning process flow (e.g., etching) typically requires simultaneously solving very different physical equations (plasma physics in the reactor, particle transport, surface chemistry, electromagnetic equations, surface time evolution, etc.). These often require different discretization schemes. As a consequence, there is not a unified framework that can accommodate a full modelling flow. This may cause a myriad of parameters that need to be modeled and/or account for, which causes prohibitive simulation times. In addition, the specific parameters of a simulation are typically unknown a priori, and many trials may be required to tune model parameters. For example, a set of ions possibly present in one particular plasma gas experiment is not known, and may be guessed until results match observations. Typically, a developed gas-phase chemistry includes tens of species and hundreds of chemical reactions.
In [0122]:
In some embodiments, front end 502 (and/or front end prep portion 506) is configured to determine incoming flux at individual points on a modeled surface of the wafer, at an interface between the wafer and an environment around the wafer
In [0124]:
topography simulations performed as a moving surface problem (e.g., as described herein), where surface velocities may be determined from a particle flux of particles arriving at a simulated surface (e.g., etch, deposition, and/or other simulations), a representation of the simulated topography may be a (discretization of a) level set function, for example.
in [0156]:
The concepts disclosed herein may simulate or mathematically model any generic imaging system for imaging sub wavelength features, and may be especially useful with emerging imaging technologies capable of producing increasingly shorter wavelengths.
In [0156]:
EUV lithography is capable of producing wavelengths within a range of 20-5 nm by using a synchrotron or by hitting a material (either solid or a plasma) with high energy electrons in order to produce photons within this range.
It would have obvious to one of ordinary skill in the art before the effective filing date of the present application to combine Kim, , King and WANG.
Kim teaches cases for experiments and simulation.
King teaches a radical oxidation and the deposition rate and pressure, pressure and
mole fraction of oxidation, the machine learning to correlate the model outputs to the experimental data.
WANG teaches simulation species.
One of ordinary skill would have motivation to combine Kim, King, and WANG that can productivity improvement for manufacturing processes (WANG [0046].
Claims 11-17 are rejected under 35 U.S.C. 103 as being unpatentable over
in view of David KING et.al (hereinafter King) US 2019/0062914 A1,
further in view Sawlani et.al (hereinafter Sawlani ) US 2023/0049157 A1,
further in view of TSUTSUI (hereinafter TSUT ) US 2023/0004837 A1.
In regard to claim 11: (Currently Amended)
King discloses:
- a chamber for depositing a material layer on a wafer using process parameters
provided by a hybrid machine learning model based on experimental outputs that include a pressure, a deposition rate, and mole fractions of a radical oxidation process, wherein depositing the material layer comprises using the radical oxidation process;
In [0048]:
a precursor volume controller configurable to the specific articles and processes being carried out in said first chamber
in [00966]:
the system, apparatus or method is configured for the application of layers to articles or substrates by various vapor deposition techniques. Examples of vapor deposition techniques can include molecular layering (ML), chemical vapor deposition (CVD), physical vapor deposition (PVD), atomic layer deposition (ALD), molecular layer deposition (MLD), vapor phase epitaxy (VPE), atomic layer chemical vapor deposition (ALCVD), ion implantation or similar techniques
In [0104]:
physicochemical parameters are considered critical wafer process parameters and characteristics in semiconductor manufacturing. They represent the chemical and physical properties of the wafer surface and materials that directly influence the performance, yield, and quality of the final device.
In [0205]:
In vapor deposition processes such as CVD, the particles can be contacted with two or more different reactants concurrently, or by one or more reactants that do not exhibit the self-limiting behavior characteristic of ALD and MLD processes. For a traditional batch CVD process, the primary methods of controlling reactions are limited to reactant exposure time and operating conditions such as process temperatures and pressures
In [209]:
substrate media provides heat transfer and a greater surface area to enhance vaporization. In another embodiment the precursor may be ionized or energized to a plasma.
Examiner’s BRI
(physicochemical parameters are considered critical wafer process parameters and characteristics in semiconductor manufacturing. They represent the chemical and physical properties of the wafer surface and materials. Pressure, deposition rate, and mole fraction in a radical oxidation process (such as plasma-enhanced CVD or advanced oxidation processes) represent key physicochemical parameters. They are measurable quantities that define the physical environment and chemical composition, governing the rate and nature of the reaction)
In [0008]:
In at least one embodiment, the apparatus further comprises at least one transport unit having one or more actuation mechanisms and configured for controlling the temperature, pressure and composition of a gaseous environment while regulating the material flow rate of a plurality of flowable articles.
Examiner’s BRI
(controlling the temperature, pressure, and composition of a gaseous environment while regulating the material flow rate of a plurality of flowable articles directly represents the active manipulation of the pressure and chemical environment on the wafer during semiconductor fabrication processes.
In [0091]:
The substrate or article can be any material which is chemically and/or thermally stable under the conditions of the deposition reaction. By “chemically” stable, it is meant that no more than 15% of the surface of the article undergoes any undesirable chemical reaction during the deposition process, other than in some cases bonding to the applied coating.
In [0180]:
FIG. 14B) in a coating subsystem 301, followed by a post-treatment process that provides for diffusion of the surface coating species and creates a penetrated coating region of greater thickness than the starting coated material (FIG. 14C).
In [0008] :
In at least one embodiment, the apparatus further comprises at least one transport unit having one or more actuation mechanisms and configured for controlling the temperature, pressure and composition of a gaseous environment while regulating the material flow rate of a plurality of flowable articles.
In [0168]:
For processes such as oxidation, reduction and etching, both surface area and reaction penetration depth that can define a volumetric value may be utilized as a critical parameter.
In [0168]:
These processes and potential critical parameters are not intended to be limiting to the invention but rather are included as a representative subset of parameters that have been qualitatively or quantitatively identified and monitored at the input and output of each subsystem of interest and permit machine learning for process
(Within the context of CVD, perhaps it is known to the POSITA that controlling the precursor or material flow rate is a standard technique for adjusting the deposition rate and thus regulating the material flow rate of a plurality of flowable articles directly represents deposition rate)
In [0212]:
Many formation methods are possible including decomposition initiated by temperature or pressure in the gas phase inside or prior to the reactor, at the substrate, reactor or manifold surfaces or by passing over an incorporated decomposition element like a hot filament or wire; or reaction with other gas phase precursors, reaction with surfaces inside or prior to the reactor, reaction with charged species, radicals or plasma; from pass through or nearby a plasma source, electron beam or ion beam.
Examiner’s BRI
(reactions with surfaces, charged species, and radicals inside a reactor, particularly when influenced by plasma, electron beams, or ion beams, are central examples of radical-induced oxidation (or radical-controlled plasma processes).
In [0020]:
flowable articles to be processed into the control system of a surface treatment system, thereby defining a first total surface area target c) providing a reactive precursor with which to treat the surfaces of said plurality of flowable articles, and entering into said control system the provided, estimated, measured or known number of moles of a reactive precursor required to saturate, react with or treat the entirety of the first total surface area target using empirical or estimated process conditions, thereby defining a complete saturation quantity, and d) selecting a target saturation ratio, to obtain a process recipe for a batch, semi-batch, semi-continuous or continuous surface treatment process, wherein said process recipe comprises at least one target pressure level associated with said target saturation ratio.
Examiner’s BRI
(target saturation ratio based on a known complete saturation quantity (moles of precursor per total surface area) represents a mole fraction)
King does not explicitly disclose:
- A semiconductor processing tool comprising:
However, Sawlani discloses:
- A semiconductor processing tool comprising:
In [0008] :
One method includes an operation for obtaining machine-learning (ML) models, each model related to predicting a performance metric for an operation of a semiconductor manufacturing tool
King does not explicitly disclose:
- A semiconductor processing tool comprising:
- a controller for changing a control variable of the semiconductor processing tool, wherein the controller receives, as an input, a difference between a measured output variable from the chamber and an output variable set-point;
- and a virtual sensor for generating a controller for changing a control variable of the semiconductor processing tool, wherein the controller receives, as an input
However, Sawlani discloses:
- A semiconductor processing tool comprising:
In [0008] :
One method includes an operation for obtaining machine-learning (ML) models, each model related to predicting a performance metric for an operation of a semiconductor manufacturing tool
- a controller for changing a control variable of the semiconductor processing tool, wherein the controller receives, as an input,
in [0053]:
The digital twin learns and updates itself from multiple sources to represent its near real-time status, working condition, or position. This learning system learns from itself (e.g., using sensor data that conveys various aspects of its operating condition),
in [0068]:
the metrology includes time-series data, which includes sensor measurements taken over time for a given parameter, such as how the pressure in the chamber evolves over time during the manufacturing process,
in [0045]:
A controller 216 manages the operation of the chamber 200 by controlling the different elements in the chamber 200, such as RF generator 218, gas sources 222, and gas pump 220. In one embodiment, fluorocarbon gases, such as CF.sub.4 and C—C.sub.4F.sub.8, are used in a dielectric etch process for their anisotropic and selective etching capabilities
in [0045]:
The fluorocarbon gases are readily dissociated into chemically reactive by-products that include smaller molecular and atomic radicals.
It would have obvious to one of ordinary skill in the art before the effective filing date of the present application to combine King and Sawlani.
King teaches chamber and hybrid machine learning model outputs.
Sawlani teaches process tool and controller.
One of ordinary skill would have motivation to combine King and Sawlani that can use experimental data to improve the quality of prediction (Sawlani [0105]).
King and Sawlani do not explicitly disclose:
- a virtual sensor for generating an estimated system state variable that is used to determine the output variable set-point.
- a difference between a measured output variable from the chamber and an output variable set-point;
However, TSUT discloses:
- virtual sensor for generating an estimated system state variable that is used to determine the output variable set-point.
in [0031]:
a virtual measurement model that infers a state of a wafer after processing;
in [0090]:
The display device 506 is a display device that displays an internal state of the virtual measurement devices 160A and 160B.
- a difference between a measured output variable from the chamber and an output variable set-point;
in [0061]:
As illustrated in FIG. 2, a semiconductor manufacturing device 200, which is an example of a substrate processing device, includes a plurality of chambers (one example of a plurality of processing spaces, in the example of FIG. 2, chamber A to chamber C), in which a wafer is processed in each chamber,
in [0035]:
In the first embodiment of each of the embodiments, a case will be described in which a virtual measurement model is generated as a model based on a time series data group, and a correction matrix is used as a fine-tuning function. In the second embodiment, a case will be described in which a neural network is used instead of the correction matrix as a fine-tuning function,
in [0055]:
To the inference section 162B with a fine-tuning function, while the virtual measurement model (the inference section 162A) generated in the virtual measurement device 160A is applied (see the dashed line 170), a fine-tuning function is added to reduce an error caused by an individual difference (error included in the inference result),
in [0041]:
In the system 100A, the semiconductor manufacturing process A processes a target object (a wafer 110A before processing) in a predetermined processing unit 120A to generate a result (a wafer 130A after processing)
(BRI: Target is the output set-point)
in [0056]:
Specifically, the inference section 162B with a fine-tuning function updates correction parameters (parameters included in a correction matrix used when tuning respective output data,
in [0033]:
by adding a fine-tuning function to the generated model, when the model is applied to other semiconductor manufacturing processes of the same type, errors (errors included in the inference result) caused by individual differences between processes are reduced by using the fine-tuning function,
in [0155]:
In step S1504, the inference section 162B with a fine-tuning function acquires inspection data for the post-processed wafer 130B and sends a notification to the comparison section 1440. Also, the comparison section 1440 compares the virtual measurement data output from the fine-tuning unit 1430 with the reported inspection data and calculates the difference,
in [0177]:
the fine-tuning network section 16100 of the inference section 1600B with a fine-tuning function continues to update the correction parameters until the difference between the virtual measurement data and the inspection data is equal to or less than a predetermined threshold value.
It would have obvious to one of ordinary skill in the art before the effective filing date of the present application to combine King, Sawlani and TSUT.
King teaches chamber and hybrid machine learning model outputs.
Sawlani teaches process tool and controller.
TSUT teaches estimating a state variable.
One of ordinary skill would have motivation to combine King, Sawlani, and TSUT to provide an optimized process using virtual measurement model(TSUT [0052].
In regard to claim 12: (Original)
King does not explicitly disclose:
- a second controller for changing the output variable setpoint, wherein the second controller receives, as an input,
However, Sawlani discloses:
- a second controller for changing the output variable setpoint, wherein the second controller receives, as an input,
in [0045]:
A controller 216 manages the operation of the chamber 200 by controlling the different elements in the chamber 200, such as RF generator 218, gas sources 222, and gas pump 220. In one embodiment, fluorocarbon gases, such as CF.sub.4 and C—C.sub.4F.sub.8, are used in a dielectric etch process for their anisotropic and selective etching capabilities
(in [0045] The fluorocarbon gases are readily dissociated into chemically reactive by-products that include smaller molecular and atomic radicals.
King and Sawlani do not explicitly disclose:
- a difference between the estimated system state variable and a system state variable set-point.
However, TSUT discloses:
- a difference between the estimated system state variable and a system state variable set-point.
in [0031]:
a virtual measurement model that infers a state of a wafer after processing;
in [0090]:
The display device 506 is a display device that displays an internal state of the virtual measurement devices 160A and 160B.
In regard to claim 13: (Original)
King does not explicitly disclose:
- a first model, wherein the first model receives the control variable as an input
However, Sawlani discloses:
- a first model, wherein the first model receives the control variable as an input
in [0064]:
The system model 402 comprises a plurality of ML models ML1-ML13 at the different levels of predictions. Each level includes one or more ML models that can interact with each other and with ML models at other layers. In some example embodiments, the models include behavioral models and physics-based models,
(BRI: the first model is drawn from plurality of models)
- and outputs the estimated system state variable that is provided to the virtual sensor.
in [0031]:
a virtual measurement model that infers a state of a wafer after processing;
in [0090]:
The display device 506 is a display device that displays an internal state of the virtual measurement devices 160A and 160B.
It would have obvious to one of ordinary skill in the art before the effective filing date of the present application to combine King and Sawlani.
King teaches chamber and hybrid machine learning model outputs.
Sawlani teaches process tool and controller.
One of ordinary skill would have motivation to combine King and Sawlani that can use experimental data to improve the quality of prediction (Sawlani [0105]).
In regard to claim 14: (Original)
King does not explicitly disclose:
- a second model, wherein the second model receives the control variable as an input
Sawlani discloses:
- a second model, wherein the first model receives the control variable as an input
in [0064]:
The system model 402 comprises a plurality of ML models ML1-ML13 at the different levels of predictions. Each level includes one or more ML models that can interact with each other and with ML models at other layers. In some example embodiments, the models include behavioral models and physics-based models,
(BRI: the second model is drawn from plurality of models)
- and outputs the estimated system state variable that is provided to the virtual sensor.
in [0031]:
a virtual measurement model that infers a state of a wafer after processing;
in [0090]:
The display device 506 is a display device that displays an internal state of the virtual measurement devices 160A and 160B.
In regard to claim 15: (Original)
King, and Sawlani do not explicitly disclose:
- a machine learning algorithm, wherein the machine learning algorithm receives as an input a difference between the output variable and the estimate of the output variable, and wherein the machine learning algorithm updates the first model.
However, TSUT discloses:
- a machine learning algorithm, wherein the machine learning algorithm receives as an input a difference between the output variable and the estimate of the output variable, and wherein the machine learning algorithm updates the first model.
in [0035]:
In the first embodiment of each of the embodiments, a case will be described in which a virtual measurement model is generated as a model based on a time series data group, and a correction matrix is used as a fine-tuning function. In the second embodiment, a case will be described in which a neural network is used instead of the correction matrix as a fine-tuning function
in [0056]:
Specifically, the inference section 162B with a fine-tuning function updates correction parameters (parameters included in a correction matrix used when tuning respective output data,
in [0033]:
by adding a fine-tuning function to the generated model, when the model is applied to other semiconductor manufacturing processes of the same type, errors (errors included in the inference result) caused by individual differences between processes are reduced by using the fine-tuning function,
in [0155]:
In step S1504, the inference section 162B with a fine-tuning function acquires inspection data for the post-processed wafer 130B and sends a notification to the comparison section 1440. Also, the comparison section 1440 compares the virtual measurement data output from the fine-tuning unit 1430 with the reported inspection data and calculates the difference,
in [0177]:
the fine-tuning network section 16100 of the inference section 1600B with a fine-tuning function continues to update the correction parameters until the difference between the virtual measurement data and the inspection data is equal to or less than a predetermined threshold value.
It would have obvious to one of ordinary skill in the art before the effective filing date of the present application to combine King, Sawlani and TSUT.
King teaches chamber and hybrid machine learning model outputs.
Sawlani teaches process tool and controller.
TSUT teaches estimating a state variable.
One of ordinary skill would have motivation to combine King, Sawlani, and TSUT to provide an optimized process using virtual measurement model(TSUT [0052].
In regard to claim 16: (Original)
King and Sawlani do not explicitly disclose:
- the machine learning algorithm utilizes a Kalman filter.
However, TSUT discloses:
- the machine learning algorithm utilizes a Kalman filter.
in [0203]:
a case is described in which an individual sensitivity and a correction matrix or a network section for fine-tuning are used as a method of fine-tuning each output data. However, the method of fine-tuning respective output data is not limited thereto, and, for example, a generalized linear mixed model, Gaussian process regression analysis, Kalman filter, or the like may be used.
In regard to claim 17: (Original)
King and Sawlani do not explicitly disclose
- the estimated system state variable is a wafer temperature.
However, TSUT discloses:
- the estimated system state variable is a wafer temperature.
in [0045]:
in the system 100A, a virtual measurement program including a learning program and an inference program is installed in the virtual measurement device 160A. When the virtual measurement program is executed, the virtual measurement device 160A functions as a learning section 161A and an inference section 162A.
in [0031]:
a virtual measurement model that infers a state of a wafer after processing
in [0080]:
The time series data acquisition devices 140A_1 to 140A_n and 140E_1 to 140E_n may acquire, as the time series data group 2, a plurality of sets of time series data measured during executing the wafer processing.
in [0185]:
a process data acquisition device that outputs process data such as temperature data or pressure data, which is a time series data group;
It would have obvious to one of ordinary skill in the art before the effective filing date of the present application to combine King, Sawlani and TSUT.
King teaches chamber and hybrid machine learning model outputs.
Sawlani teaches process tool and controller.
TSUT teaches estimating a state variable.
One of ordinary skill would have motivation to combine King, Sawlani, and TSUT to provide an optimized process using virtual measurement model(TSUT [0052].
Conclusion
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the
examiner should be directed to TIRUMALE KRISHNASWAMY RAMESH whose telephone number is (571)272-4605. The examiner can normally be reached by phone.
Examiner interviews are available via telephone, in-person, and video conferencing
using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at
http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Li B Zhen can be reached on phone (571-272-3768). The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be
obtained from Patent Center. Unpublished application information in Patent Center is available to registered users.
To file and manage patent submissions in Patent Center, visit:
https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for
information about filing in DOCX format. For additional questions, contact the Electronic
Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO
Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/TIRUMALE K RAMESH/Examiner, Art Unit 2121
/Li B. Zhen/Supervisory Patent Examiner, Art Unit 2121
Prosecution Insights