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 .
This action is responsive to the filing on 04/28/2026. Claims 1, 10 and 15 have been amended. Claim 8 have been canceled. Claims 1-7 and 9-20 are pending in this case. Claims 1, 10 and 15 are independent claims.
Claim Rejections - 35 USC § 112
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claim(s) 1 and similar claim structure are rejected under 35 U.S.C. 112(b) as being indefinite for failing to particularly point out and distinctly claim the subject matter. The limitation “ (ii) first process tool data comprising at least one of a first number of substrates historically processed by one or more first processing tools relative to other processing tools of the manufacturing system, or a second number of substrates processed since a prior preventative maintenance procedure at the one or more first processing tools relative to the other processing tool”. The term “relative” is a term of degree that lacks an objective standard for determining its scope. The claim does not specify whether “relative” requires, for example, a numerical ratio, percentage, etc. Furthermore, the specification fails to define the term “relative”.
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 and 10 are rejected under 35 U.S.C. 103 as being unpatentable over JUNG et al. (Pub No.: 20190286983 A1), hereinafter referred to as JUNG, in view of Kang et al. (Pub No.: 20150147482 A1), hereinafter referred to as Kang.
With respect to claim 1, JUNG disclose:
A method, comprising: receiving, by a processing device, training data comprising: first sensor data indicating a first state of an environment of a first <equipment-related data> processing a first substrate at a manufacturing system (Under the broadest reasonable interpretation (BRI), "first substrate and second substrate" are simply training example substrate and a later substrate presented to the trained model during inference. From a patent examination perspective, those are typically the same type of data collected at different points in time, not two different technical concepts. In paragraphs [0031-0035], JUNG discloses a semiconductor manufacturing yield prediction system including a data extraction unit, learning unit, storage unit and prediction unit. The extraction unit extract and accumulates manufacturing data, and the learning unit performs machine learning on the extracted data. In paragraphs [0039 & 0046-0050], JUNG discloses training data from input data, explaining that training data are periodically extracted and accepts for machine learning. In paragraph [0041], JUNG discloses the equipment-related data measured through an internal/external sensor. Input data regarding the operation/state of equipment in which various physical/electrical state values are arranged in time series.
First process result data corresponding to the first substrate (In paragraph [0042-0044], JUNG discloses process result data produced after processing a semiconductor substrate. The measured characteristics (e.g., thickness, line width, misalignment, and defects) are the results of the manufacturing process for that substrate and can be used as outputs of targets when training the machine learning model. )
Training, by the processing device, a first model with input data comprising the first sensor data and the first process tool data and a target output comprising the process result data, wherein the trained first model is to receive a new input having second sensor data indicating a second state of an environment of a second processing chamber processing a second substrate and second process tool data associated with a second processing tool processing the second substrate to produce a second output based on the new input, the second process tool data comprising at least one of the first number of substrates historically processed by the second processing tool or the second number of substrates processed since the prior preventative maintenance procedure at the second processing tool, and the second output indicating a second process result data corresponding to the second substrate (In paragraphs [0058-0073], JUNG discloses the learning unit trains neural network models 131-134. These trained models are later used by the prediction unit. In paragraph [0075], JUNG discloses that external systems provide input data to the prediction system after the models have already been trained. This is the operational (inference) stage where the trained model receives new data. (In paragraph 0041, JUNG discloses equipment sensor data including operating time, pressure, temperature, etc. These measurements indicate the operating state of manufacturing equipment/chambers. When new equipment, sensor data is later supplied to the trainer.))
With respect to claim 1, JUNG does not explicitly disclose:
First process tool data comprising at least one of a first number of substrates historically processed by one or more first processing tools relative to other processing tools of the manufacturing system, or a second number of substrates processed since a prior preventative maintenance procedure at the one or more first processing tools relative to the other processing tools
<Processing chamber>
However, it is known by Kang to disclose
First process tool data comprising at least one of a first number of substrates historically processed by one or more first processing tools relative to other processing tools of the manufacturing system, or a second number of substrates processed since a prior preventative maintenance procedure at the one or more first processing tools relative to the other processing tools (Examiner selects the first portion of (ii): In Fig. 1-2 and paragraph [0037], Kang discloses a semiconductor reaction chamber that processes semiconductor substrates (wafers), processing many substrates through the same chamber. In paragraph [0131], Kang discloses evaluating wafer-to-wafer repeatability. The count simply identifies the order in which each wafer was processed, where wafer count 1 is the first processed wafer and wafer count 25 is the last. (JUNG is known to teach collecting data from multiple pieces of semiconductor manufacturing equipment IDs, historical production data, etc.) So, using Kang's known substrate count metric across the manufacturing system would have been a predictable implementation.)
<Processing chamber> (In paragraph [0028], Kang disclose a processing chamber (referred to as a reaction chamber), the reaction chamber is processing semiconductor substrates (wafers) inside the reaction chamber.
JUNG in view of Kang are analogous pieces of art because both references concern the field of semiconductor manufacturing, both concern semiconductor wafer fabrication and operation of semiconductor processing tools. Accordingly, it would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, to modify JUNG, with a machine learning-based semiconductor manufacturing yield prediction system as taught by JUNG, with a reaction chamber for processing substrates as taught by Kang. The motivation for doing so would have been to accurately predict a final yield based on data generated during a semiconductor manufacturing process (See [0001] of JUNG.)
With respect to claim 10, JUNG disclose:
A method, comprising: receiving, by a processing device[[,]]:(i) sensor data indicating a state of an environment of a < equipment-related data > processing a first substrate according to a substrate processing procedure at a manufacturing system (Under the broadest reasonable interpretation (BRI), "first substrate and second substrate" are simply training example substrate and a later substrate presented to the trained model during inference. From a patent examination perspective, those are typically the same type of data collected at different points in time, not two different technical concepts. In paragraphs [0031-0035], JUNG discloses a semiconductor manufacturing yield prediction system including a data extraction unit, learning unit, storage unit and prediction unit. The extraction unit extract and accumulates manufacturing data, and the learning unit performs machine learning on the extracted data. In paragraphs [0039 & 0046-0050], JUNG discloses training data from input data, explaining that training data are periodically extracted and accepts for machine learning. In paragraph [0041], JUNG discloses the equipment-related data measured through an internal/external sensor. Input data regarding the operation/state of equipment in which various physical/electrical state values are arranged in time series.
Processing the sensor data and the process tool data using one or more machine- learning models (MLMs) to determine a prediction of a process result measurement of the first substrate (In paragraph [0029], JUNG disclose the semiconductor manufacturing yield prediction system according to an embodiment of the present disclosure predicts a semiconductor manufacturing yield by using a neural network model which is generated by machine-learning data generated during a semiconductor manufacturing process.)
Performing, by the processing device, at least one of a) preparing the prediction for presentation on a graphical user interface (GUI) or b) altering an operation of at least one of the processing chamber or the processing tool based on the prediction (Examiner selects: B. In paragraph [0058], JUNG disclose a machine learning by the learning unit 120 is separately performed according to types of data, and separate neural network models are generated/updated according to types of resulting data. Hereinafter, a method for generating/updating neutral network models will be described according to types of data)
With respect to claim 10, JUNG does not explicitly disclose:
Process tool data comprising at least one of a first number of substrates historically processed by processing tool processing the first substrate relative to other process tools of the manufacturing system or a second number of substrates processed since a prior preventative maintenance procedure at the processing tool relative to the other process tools
<Processing chamber>
However, it is known by Kang to disclose:
Process tool data comprising at least one of a first number of substrates historically processed by processing tool processing the first substrate relative to other process tools of the manufacturing system or a second number of substrates processed since a prior preventative maintenance procedure at the processing tool relative to the other process tools (Examiner selects the first portion of (ii): In Fig. 1-2 and paragraph [0037], Kang discloses a semiconductor reaction chamber that processes semiconductor substrates (wafers), processing many substrates through the same chamber. In paragraph [0131], Kang discloses evaluating wafer-to-wafer repeatability. The count simply identifies the order in which each wafer was processed, where wafer count 1 is the first processed wafer and wafer count 25 is the last. (JUNG is known to teach collecting data from multiple pieces of semiconductor manufacturing equipment IDs, historical production data, etc.) So, using Kang's known substrate count metric across the manufacturing system would have been a predictable implementation.)
<Processing chamber> (In paragraph [0028], Kang disclose a processing chamber (referred to as a reaction chamber), the reaction chamber is processing semiconductor substrates (wafers) inside the reaction chamber.
JUNG in view of Kang are analogous pieces of art because both references concern the field of semiconductor manufacturing, both concern semiconductor wafer fabrication and operation of semiconductor processing tools. Accordingly, it would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, to modify JUNG, with a machine learning-based semiconductor manufacturing yield prediction system as taught by JUNG, with a reaction chamber for processing substrates as taught by Kang. The motivation for doing so would have been to accurately predict a final yield based on data generated during a semiconductor manufacturing process (See [0001] of JUNG.)
Claims 2-5, 7, 11 are rejected under 35 U.S.C. 103 as being unpatentable over JUNG, in view of Kang and further in view of DAVID et al. (Pub No.: 20170109646 A1), hereinafter referred to as DAVID.
Regarding claim 2, JUNG in view of Kang disclose the elements of claim 1. JUNG in view of Kang do not disclose:
The method of claim 1, wherein training the first model further comprises: processing the first process result data using the first process tool data to generate time-independent process result data
causing a first regression to be performed using the time-independent process result data and the first sensor data
However, DAVID disclose the limitations:
The method of claim 1, wherein training the first model further comprises: processing the first process result data using the first process tool data to generate time-independent process result data (In paragraph [0050], DAVID explains that machine learning can be used to control semiconductor manufacturing processes by first predicting process results, such as critical dimension or file thickness, through virtual metrology. The prediction is generated before or during a manufacturing step and is then used to adjust one or more processing parameters (e.g., process run time).)
causing a first regression to be performed using the time-independent process result data and the first sensor data (In paragraph [0085], DAVID discloses examples of machine learning algorithms include Decision Trees, such as CART (Classification and Regression Trees),)
Accordingly, it would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, having the teaching of JUNG in view of Kang to include DAVID, with training and re-training the model, and deploying the model, are applicable to yield prediction and evaluation of other target as taught by DAVID. The motivation for doing so would have been to improve process control techniques for lithography, yield prediction, and other aspects of semiconductor manufacturing processes. (See [0002] of DAVID.)
Regarding claim 3, JUNG in view of Kang and DAVID disclose the elements of claim 2. In addition, DAVID disclose:
The method of claim 2, wherein training the first model further comprises: determining a residual between the first process result data and the time- independent process result data (In paragraph [0125], DAVID discloses predicting semiconductor metrology/process results from equipment sensor data, comparing predicted values to measured metrology values to evaluate model performance.)
causing a second regression to be performed using the residual and the first sensor data (In paragraph [0125], DAVID discloses a refining regression model based on prediction accuracy.)
Regarding claim 4, JUNG in view of Kang and DAVID disclose the elements of claim 3. In addition, DAVID disclose:
The method of claim 3, wherein at least one of the first regression or the second regression is performed using a partial least squares (PLS) algorithm (In paragraph [0121], DAVID discloses common dimensionality reduction techniques including partial least squares and principal component analysis.)
Regarding claim 5, JUNG in view of Kang and DAVID disclose the elements of claim 3. In addition, DAVID disclose:
The method of claim 3, wherein at least one of the first regression or the second regression is performed as part of a gradient boosting regression (GBR) algorithm (In paragraph [0086], DAVID discloses that the GBM (Gradient Boosting Machine) and Random Forests algorithms can produce the best results.)
Regarding claim 7, JUNG in view of Kang and DAVID disclose the elements of claim 1. In addition, DAVID disclose:
The method of claim 1, wherein the first process result data comprises a value corresponding to an etch bias of the first substrate (In paragraph [0122], DAVID discloses an etch depth may be predicted given certain upstream variables such as etch tool process parameters, previous step thickness and process variables such as deposition tool process parameters, CMP process parameters, and optical n and k values of the film.)
Regarding claim 11, JUNG in view of Kang disclose the elements of claim 10. JUNG in view of Kang do not disclose:
The method of claim 10, wherein the prediction of the process result measurement comprises a value corresponding to an etch bias of the first substrate
However, DAVID disclose the limitation (In paragraph [0122], DAVID discloses an etch depth may be predicted given certain upstream variables such as etch tool process parameters, previous step thickness and process variables such as deposition tool process parameters, CMP process parameters, and optical n and k values of the film.)
Accordingly, it would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, having the teaching of JUNG in view of Kang to include DAVID, with training and re-training the model, and deploying the model, are applicable to yield prediction and evaluation of other target as taught by DAVID. The motivation for doing so would have been to improve process control techniques for lithography, yield prediction, and other aspects of semiconductor manufacturing processes. (See [0002] of DAVID.)
Claim 6 is rejected under 35 U.S.C. 103 as being unpatentable over JUNG, in view of Kang and further in view of Achin et al. (Pub No.: 20180060738 A1), hereinafter referred to as Achin.
Regarding claim 6, JUNG in view of Kang disclose the elements of claim 1. JUNG in view of Kang do not disclose:
The method of claim 1, wherein training the first model further comprises: causing a first regression to be performed using a first subset of training data to generate a first regression model
causing a second regression to be performed using a second subset of training data to generate a second regression model
determining a first accuracy of the first regression model is greater than a second accuracy of the second regression model based on a comparison of the first regression model, the second regression model, and the training data
However, Achin disclose the limitations:
The method of claim 1, wherein training the first model further comprises: causing a first regression to be performed using a first subset of training data to generate a first regression model (In Fig. 9 and paragraph [0311], Achin discloses the first accuracy score of each of the fitted predictive models. The first accuracy score of a fitted model represents the accuracy with which the fitted model predicts one or more outcomes of the initial prediction problem.)
Causing a second regression to be performed using a second subset of training data to generate a second regression model (In Fig. 9 and paragraph [0313], Achin discloses a second accuracy score of the fitted predictive model for the modified prediction problem. The second accuracy score represents an accuracy with which the fitted model predicts one or more outcomes of the modified prediction problem.)
Determining a first accuracy of the first regression model is greater than a second accuracy of the second regression model based on a comparison of the first regression model, the second regression model, and the training data (In Fig. 9 and paragraph [0314], Achin discloses that the system 100 calculates the predictive value of the feature F. In some embodiments, the predictive value of the feature F for a modeling procedure or model is calculated based on the change in accuracy (e.g., based on the difference between the first and second accuracy scores for model).)
Accordingly, it would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, having the teaching of JUNG in view of Kang to include Achin, with determining a first and second accuracy score of each of the fitted models as taught by Achin. The motivation for doing so would have been to measure the effectiveness of efforts to improve processes, or to decide how to adjust processes (See [0003] of Achin.)
Claims 9 and 12 are rejected under 35 U.S.C. 103 as being unpatentable over JUNG, in view of Kang and further in view of Lu et al. (US Patent No. 9,255,780 B2), hereinafter referred to as Lu.
Regarding claim 9, JUNG in view of Kang disclose the elements of claim 1. JUNG in view of Kang do not disclose:
The method of claim 1, wherein the first process result data indicates a first average thickness associated with a central region of the first substrate and a second average thickness associated with an edge region of the first substrate
However, it is known by Lu to disclose (In Col. 4, lines 46-67, Lu discloses measuring file thickness at locations extending from the center toward the edge of a wafer; determining the actual thickness at those locations; specifically distinguishing the edge region from the remainder of the wafer; and generating thickness values used to evaluate wafer thickness profiles.)
Accordingly, it would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, having the teaching of JUNG in view of Kang to include Lu, with determining a film thickness measuring correction factor according to the real film thickness as taught by Lu. The motivation for doing so would have been to measure the thickness of the film on the edge of the wafer (i.e. the film thickness of the edge of the film of the wafer) during a chemical mechanical polishing process (See (Col. 1, lines 23-26) of Lu.)
Regarding claim 12, JUNG in view of Kang disclose the elements of claim 10. JUNG in view of Kang do not disclose:
The method of claim 10, wherein the prediction of the process result measurement comprises indicates a first average thickness associated with a central region of the first substrate and a second average thickness associated with an edge region of the first substrate
However, it is known by Lu to disclose (In Col. 4, lines 46-67, Lu disclose measuring film thickness at locations extending from the center toward the edge of a wafer; determining the actual thickness at those location; specifically distinguishing the edge region from the remainder of the wafer; and generating thickness values used evaluate wafer thickness profiles.)
Accordingly, it would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, having the teaching of JUNG in view of Kang to include Lu, with determining a film thickness measuring correction factor according to the real film thickness as taught by Lu. The motivation for doing so would have been to measure the thickness of the film on the edge of the wafer (i.e. the film thickness of the edge of the film of the wafer) during a chemical mechanical polishing process (See (Col. 1, lines 23-26) of Lu.)
Claim 13 is rejected under 35 U.S.C. 103 as being unpatentable over JUNG, in view of Kang and further in view of EL-Shaer et al. (US Pub No.: 20230219561 A1), hereinafter referred to as EL-Shaer.
Regarding claim 13, JUNG in view of Kang disclose the elements of claim 10. JUNG in view of Kang do not explicitly disclose:
The method of claim 10, wherein processing the sensor data and the process tool data further comprises processing the sensor data using the process tool data to generate modified sensor data, wherein the modified sensor data comprises sensor data weighted according to the process tool data, wherein the prediction is determined based on the modified sensor data
However, EL-Shaer disclose the limitation (In paragraph [0076], EL-Shaer discloses that the signal processing system 502 may obtain the sensor data 506 from a different component other than the sensor 504. Further, one or more sensors 504 and/or a different component can perform preliminary signal processing to modify the sensor data 506 prior to the signal processing system 502 obtaining the sensor data 506).
Accordingly, it would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention having the teachings of JUNG in view of Kang before them, to include EL-Shaer’s state estimation to improve the accuracy of the vehicle state variables outputted to one or more other systems in the environment (e.g., environment 100) as processed sensor data 510 as taught by EL-Shaer (see[0080]).
Claim 14 are rejected under 35 U.S.C. 103 as being unpatentable over JUNG, in view of Kang and further in view of ARMITAGE et al. (US Pub No.: 20220254492 A1), hereinafter referred to as ARMITAGE.
Regarding claim 14, JUNG in view of Kang disclose the elements of claim 10. JUNG in view of Kang do not explicitly disclose:
The method of claim 10, wherein processing the sensor data and the process tool data further comprises: processing, using a first MLM of the one or more MLMs, the sensor data to obtain a first process result prediction
processing, using a second MLM of the one or more MLMs, the first process result prediction to obtain a second process result prediction
determining the prediction based on a combination of at least the first process result prediction and the second process result prediction
However, ARMITAGE disclose the limitation:
The method of claim 10, wherein processing the sensor data and the process tool data further comprises: processing, using a first MLM of the one or more MLMs, the sensor data to obtain a first process result prediction (In paragraph [0143], ARMITAGE discloses the training sensor dataset used for training the first set of ML model(s).)
processing, using a second MLM of the one or more MLMs, the first process result prediction to obtain a second process result prediction (In paragraph [0143], ARMITAGE discloses the second set of ML model(s), corresponding to estimates or calculations of the second sensor.)
determining the prediction based on a combination of at least the first process result prediction and the second process result prediction (In paragraph [0143], ARMITAGE discloses calculating/determining first and second process results.)
Accordingly, it would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention having the teachings of JUNG in view of Kang before them, to include ARMITAGE’s automated detection to estimate the one or more clinical biomarker(s) of the subject based on the extracted and classified segments of sensor data as taught by ARMITAGE (see[0001].)
Claim 15 is rejected under 35 U.S.C. 103 as being unpatentable over JUNG et al. (Pub No.: 20190286983 A1), hereinafter referred to as JUNG, in view of Kang et al. (Pub No.: 20150147482 A1), hereinafter referred to as Kang and further in view of Grill et al. (Pub No.: 20210383225 A1), hereinafter referred to as Grill
With respect to claim 15, JUNG disclose:
A method, comprising: training a machine learning model (MLM) comprising: receiving training data comprising (i) first sensor data indicating a first state of an environment of a first process < equipment-related data > a first substrate (Under the broadest reasonable interpretation (BRI), "first substrate and second substrate" are simply training example substrate and a later substrate presented to the trained model during inference. From a patent examination perspective, those are typically the same type of data collected at different points in time, not two different technical concepts. In paragraphs [0031-0035], JUNG discloses a semiconductor manufacturing yield prediction system including a data extraction unit, learning unit, storage unit and prediction unit. The extraction unit extract and accumulates manufacturing data, and the learning unit performs machine learning on the extracted data. In paragraphs [0039 & 0046-0050], JUNG discloses training data from input data, explaining that training data are periodically extracted and accepts for machine learning. In paragraph [0041], JUNG discloses the equipment-related data measured through an internal/external sensor. Input data regarding the operation/state of equipment in which various physical/electrical state values are arranged in time series.
Metrology data comprising process result measurements and location data indicating first locations across a surface of the first substrate corresponding to the process result measurements (In paragraph [0042], JUNG disclose quality data as data obtained by measuring the quality of a semiconductor product (or intermediate product). It further explains that quality data includes: fault data, and measurement data. This is essentially metrology data because it consists of measurements made on the processed wafer.)
With respect to claim 15, JUNG does not explicitly disclose:
First process tool data comprising at least one of a first number of substrates historically processed by one or more first processing tools of the first process chamber relative to other processing tools, or a second number of substrates processed since a prior preventative maintenance procedure at the one or more first processing tools relative to the other processing tools
Encoding the training data to generate encoded training data
Causing a regression to be performed using the encoded training data
However, it is known by Kang to disclose:
First process tool data comprising at least one of a first number of substrates historically processed by one or more first processing tools of the first process chamber relative to other processing tools, or a second number of substrates processed since a prior preventative maintenance procedure at the one or more first processing tools relative to the other processing tools (Examiner selects the first portion of (ii): In Fig. 1-2 and paragraph [0037], Kang discloses a semiconductor reaction chamber that processes semiconductor substrates (wafers), processing many substrates through the same chamber. In paragraph [0131], Kang discloses evaluating wafer-to-wafer repeatability. The count simply identifies the order in which each wafer was processed, where wafer count 1 is the first processed wafer and wafer count 25 is the last. (JUNG is known to teach collecting data from multiple pieces of semiconductor manufacturing equipment IDs, historical production data, etc.) So, using Kang's known substrate count metric across the manufacturing system would have been a predictable implementation.)
JUNG in view of Kang are analogous pieces of art because both references concern the field of semiconductor manufacturing, both concern semiconductor wafer fabrication and operation of semiconductor processing tools. Accordingly, it would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, to modify JUNG, with a machine learning-based semiconductor manufacturing yield prediction system as taught by JUNG, with a reaction chamber for processing substrates as taught by Kang. The motivation for doing so would have been to accurately predict a final yield based on data generated during a semiconductor manufacturing process (See [0001] of JUNG.)
With respect to claim 15, JUNG in view of Kang does not explicitly disclose:
Encoding the training data to generate encoded training data
Causing a regression to be performed using the encoded training data
However, Grill is known to disclose:
Encoding the training data to generate encoded training data (In paragraph [0045], Grill disclose training an encoder neural network to generate representations of data items, without using labelled training data items, and without using a contrastive loss)
Causing a regression to be performed using the encoded training data (In paragraph [0016], Grill disclose using a regression model embodied by parameters of the prediction neural network, to generate the prediction of the target output.)
JUNG in view of Kang are analogous pieces of art because both references concern the field of semiconductor manufacturing, both concern semiconductor wafer fabrication and operation of semiconductor processing tools. Accordingly, it would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, to modify JUNG, with a machine learning-based semiconductor manufacturing yield prediction system as taught by JUNG, with a reaction chamber for processing substrates as taught by Kang, with using a regression model embodied by parameters of the prediction neural network, to generate the prediction of the target output as taught by Grill. The motivation for doing so would have been to improve wafer-to-wafer thickness uniformity and within-wafer thickness uniformity (See [0001] of Kang.)
Claim 16 is rejected under 35 U.S.C. 103 as being unpatentable over JUNG in view of Kang, Grill and further in view of Hester et al. (US Patent No.11,921,824 B1), hereinafter referred to as Hester.
Regarding claim 16, JUNG in view of Kang and Grill disclose the elements of claim 15. JUNG in view of Kang and Grill do not explicitly disclose:
The method of claim 15, further comprising: receiving second sensor data indicating a second state of an environment of a second process chamber processing a second substrate
encoding the second sensor data to generate encoded sensor data
using the encoded sensor data as input to the trained MLM
receiving one or more outputs from the trained MLM, the one or more outputs comprising encoded prediction data
and decoding the encoded prediction data to generate prediction data comprising values indicating process results of the second substrate in second locations across a surface of the second substrate, the second locations corresponding to the first locations of the first substrate
However, Hester disclose the limitation:
The method of claim 15, further comprising: receiving second sensor data indicating a second state of an environment of a second process chamber processing a second substrate (In Col. 11, lines 39-47, Hester disclose receiving second sensor data from a second sensor for fusing sensor data of different modalities using a transformer encoder.)
encoding the second sensor data to generate encoded sensor data ( In Col. 11, lines 39-47, Hester disclose the transformer encoder may be used to generate a second modified plurality of tokens representing the point cloud lidar sensor data .)
using the encoded sensor data as input to the trained MLM (In Col. 2, lines 27-29, Hester disclose transformer models are machine learning models that include an encoder network and a decoder network.)
receiving one or more outputs from the trained MLM, the one or more outputs comprising encoded prediction data (In Col. 2, lines 27-34, Hester disclose transformer models are machine learning models that include an encoder network and a decoder network. The encoder takes an input and generates feature representations (e.g., feature vectors, feature maps, etc.) from the input. The feature representation is then fed into a decoder that may generate an output based on the encodings.)
and decoding the encoded prediction data to generate prediction data comprising values indicating process results of the second substrate in second locations across a surface of the second substrate, the second locations corresponding to the first locations of the first substrate (In Col. 11, lines 48-54, Hester disclose processing may continue at action 418, at which a computer vision (CV) operation may be performed using the first modified plurality of tokens and the second modified plurality of tokens. The particular computer vision operation and/or other predictive task may be dependent on the tasks for which the particular transformer decoder and the prediction heads have been trained.)
Accordingly, it would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention having the teachings of JUNG in view of Kang and Grill before them, to include Hester’s sensor data to improve the machine learning models over time by retraining the models as more and more data becomes available as taught by Hester (see(Col. 1, lines 65-67)).
Claim 17 is rejected under 35 U.S.C. 103 as being unpatentable over JUNG in view of Kang, Grill, Hester and further in view of DAVID et al. (Pub No.: 20170109646 A1), hereinafter referred to as DAVID.
Regarding claim 17, JUNG in view of Kang, Grill and Hester disclose the elements of claim 16. JUNG in view of Kang, Grill and Hester do not explicitly disclose:
The method of claim 16, wherein at least one of encoding the sensor data or decoding the encoded prediction data is performed using principal component analysis (PCA).
However, In paragraph [0084], DAVID discloses dimensionality reduction techniques are generally known, for example, principal component analysis (PCA).
Accordingly, it would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention having the teachings of JUNG in view of Kang, Grill and Hester before them, to include DAVID. The motivation for doing so would have been to improve process control techniques for lithography, yield prediction, and other aspects of semiconductor manufacturing processes. (See [0002] of DAVID.)
Claim 18 is rejected under 35 U.S.C. 103 as being unpatentable over JUNG in view of Kang, Grill, Hester, DAVID and further in view of Lu et al. (US Patent No. 9,255,780 B2), hereinafter referred to as Lu.
Regarding claim 18, JUNG in view of Kang, Grill, Hester and DAVID disclose the elements of claim 16. JUNG in view of Kang, Grill, Hester and DAVID do not explicitly disclose:
The method of claim 18, wherein the predication data indicates a first average thickness associated with a central region of the second substrate and a second average thickness associated with an edge region of the second substrate
However, Lu disclose the limitation (In Col. 4, lines 46-67, Lu discloses measuring file thickness at locations extending from the center toward the edge of a wafer; determining the actual thickness at those locations; specifically distinguishing the edge region from the remainder of the wafer; and generating thickness values used to evaluate wafer thickness profiles.)
Accordingly, it would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention having the teachings of JUNG in view of Kang, Bramley, Hester and DAVID before them to include to include Lu, with determining a film thickness measuring correction factor according to the real film thickness as taught by Lu. The motivation for doing so would have been to measure the thickness of the film on the edge of the wafer (i.e. the film thickness of the edge of the film of the wafer) during a chemical mechanical polishing process (See (Col. 1, lines 23-26) of Lu.)
Claims 19-20 are rejected under 35 U.S.C. 103 as being unpatentable over JUNG in view of Kang, Grill and further in view of DAVID et al. (Pub No.: 20170109646 A1), hereinafter referred to as DAVID.
Regarding claim 19, JUNG in view of Kang and Grill disclose the elements of claim 15. JUNG in view of Kang and Grill do not explicitly disclose:
The method of claim 15, wherein the process result measurements comprise a value indicating an etch bias of the first substrate
However, DAVID disclose the limitation (In paragraph [0122], DAVID discloses an etch depth may be predicted given certain upstream variables such as etch tool process parameters, previous step thickness and process variables such as deposition tool process parameters, CMP process parameters, and optical n and k values of the film.)
Accordingly, it would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention having the teachings of JUNG in view of Kang and Grill before them, to include DAVID. The motivation for doing so would have been to improve process control techniques for lithography, yield prediction, and other aspects of semiconductor manufacturing processes. (See [0002] of DAVID.)
Regarding claim 20, JUNG in view of Kang and Grill disclose the elements of claim 15. JUNG in view of Kang and Grill do not explicitly disclose:
The method of claim 15, wherein the regression is performed as a part of a gradient boosting regression (GBR)
However, DAVID disclose the limitation (In paragraph [0086], DAVID discloses that the GBM (Gradient Boosting Machine) and Random Forests algorithms can produce the best results.)
Accordingly, it would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention having the teachings of JUNG in view of Kang and Grill before them, to include DAVID. The motivation for doing so would have been to improve process control techniques for lithography, yield prediction, and other aspects of semiconductor manufacturing processes. (See [0002] of DAVID.)
Response to Arguments
Applicant's arguments filed on 04/28/2026 have been fully considered, and in part are persuasive.
Pertaining to Rejection under 101
Rejections for claims 1-7 and 9-20 are withdrawn under 35 USC § 101
Pertaining to Rejection under 103
Applicant’s arguments in regard to the examiner’s rejections under 35 USC 103 are moot in view of the new grounds of rejection
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to EVEL HONORE whose telephone number is (703)756-1179. The examiner can normally be reached Monday-Friday 8 a.m. -5:30 p.m.
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, Mariela D Reyes can be reached at (571) 270-1006. 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.
EVEL HONORE
Examiner
Art Unit 2142
/HAIMEI JIANG/Primary Examiner, Art Unit 2142
Prosecution Insights