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 Preliminary to Amendment
The preliminary amendments filed 04/2022/2024 have been acknowledged
Claims 1, 3, 4, 6-8, 10-13 and 17 have been amended.
Claim 21 has been cancelled.
Information Disclosure Statement
The information disclosure statement (IDS) submitted on 05/06/2024 is in compliance with the provisions of 37 CFR 1.97. Accordingly, the information disclosure statement is being considered by the examiner.
The information disclosure statement (IDS) submitted on 04/22/2024 is in compliance with the provisions of 37 CFR 1.97. Accordingly, the information disclosure statement is being considered by the examiner.
Claim Rejections - 35 USC § 102
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of the appropriate paragraphs of 35 U.S.C. 102 that form the basis for the rejections under this section made in this Office action:
A person shall be entitled to a patent unless –
(a)(1) the claimed invention was patented, described in a printed publication, or in public use, on sale, or otherwise available to the public before the effective filing date of the claimed invention.
(a)(2) the claimed invention was described in a patent issued under section 151, or in an application for patent published or deemed published under section 122(b), in which the patent or application, as the case may be, names another inventor and was effectively filed before the effective filing date of the claimed invention.
Claim 1-15 and 17 are rejected under 35 U.S.C. 102 as being anticipated by Bastani et al (Bastani hereinafter EP 3767392 A1 “METHOD AND APPARATUS FOR DETERMINING FEATURE CONTRIBUTION TO PERFORMANCE” )
As per claim 1
Bastani teaches A non-transitory computer readable medium having instructions therein or thereon, the instructions, when executed by a computer system, configured to cause the computer system to at least (Summary: “processor-readable medium comprising instructions which, when executed by a processor, cause the processor to perform a method as described above.”) determine a physical trend of one or more imaging characteristics with respect to pattern variation on a substrate resulting from a patterning process the physical trend determined based on first data for a first set of patterns and/or the patterning process (Detailed Description: “Figure 4 depicts a flow diagram of steps in a method of determining the contribution of a process feature to the performance of a pattering process…The first model may be trained on first process data and first performance data. The first process data and first performance data may comprise historical data… Process data may comprise a plurality of data points for one or more process features. Process features may relate to properties of a substrate, for example overlay, alignment, critical dimension, levelling, etc…The first model will be referred to as a global model for the rest of the description. The first process data and first performance data may be referred to as global process data and global performance data, respectively. The global process data and global performance data may comprise data relating to a large number of substrates. Therefore, the global model may be trained on a large number of substrates. As a result, the global model may have been trained on process data representing many different process features creating many different effects on the performance of the patterning process. The global model may be used to predict a performance of the patterning process for a set of substrates” Examiner views “contribution” as physical trend.) and generate augmented data based on the physical trend (Detailed description: “generating augmented process data may comprise modelling a parametric distribution for a process feature, and generating a plurality of samples according to the modelled parametric distribution… the method may further comprise providing the augmented process data as input to the first model, and obtaining augmented predicted performance data as an output of the first model…Generated data associated with data related to a substrate may be referred to as augmented data or as simulated data…A first example method of generating augmented process data may comprise modelling a parametric distribution for a process feature such as overlay, alignment, critical dimension, etc. Augmented process data may then be obtained by generating a plurality of samples according to the modelled parametric distribution…In order to use the augmented process data for training, it may be paired with corresponding augmented performance data “) the augmented data comprising second data that conforms to the physical trend and is derived based on the first data (Detailed description: “the method may further comprise providing the augmented process data as input to the first model, and obtaining augmented predicted performance data as an output of the first model.
Optionally, the second performance data may comprise the augmented predicted performance data and the predicted performance data for the one or more substrates”) the augmented data derived for a second set of patterns that are different from the first set (Figure 5, Detailed Description “Augmented process data may be obtained by determining a range 604 around process data of one or more process features…Augmented process data may then be obtained by generating a plurality of samples according to the modelled parametric distribution…generating augmented process data may comprise determining a range of values for a process feature….the augmented data process may be obtained by generating a plurality of samples within the range of values for the feature.” Samples based on a range or distribution imply a variety of patterns. One or more process features implies different patterns ) wherein the augmented data is configured to be provided as input to a machine learning model to train the machine learning model to predict the one or more imaging characteristics according to the physical trend (Detailed Description: “using the second model to determine a contribution of a process feature of the second process data to the second performance data associated with the identified one or more substrates may comprise training the second model on the second process data and the second performance data, using the second model to estimate a relation between a process feature and second performance data, and determining the contribution of a process feature based on the estimated relation…augmented process data may be provided as input to the global model GM. The output produced by the global model GM may be augmented performance data 608 corresponding to the augmented process data input.” “At 406, a second model may be trained on second process data and second performance data. The second process data and second performance data may be associated with the one or more substrates. At 408, the trained second model may be used to determine a contribution of a feature of the second process data to the second performance data associated with the one or more substrates.”)
As per claim 2
Bastani teaches all claim limitations previously rejected under claim 1’s 102 rejection , See claim 1’s 102 rejection.
Bastani teaches wherein the physical trend is associated with pattern design variation on the substrate and/or patterning process variation. (Figure 5, Detailed Description: For example, in one layer, an issue with overlay may cause a yield loss in one region of the substrate. In another layer, a problem with critical dimension variation may cause yield loss in another region of the substrate… The present disclosure relates to a method and apparatus for determining one or more potential root causes for yield loss on one or more substrates. Alternatively, root causes for loss in another indicator of patterning process performance may be determined. The method may determine potential root causes by making an estimation of yield loss based on process data, and make an estimation of which features of the process data most affect the loss in yield….Figure 4 depicts a flow diagram of a method of determine the contribution of a process feature to the performance of a process of patterning substrates…The second process data and second performance data may be associated with the one or more substrates. At 408, the trained second model may be used to determine a contribution of a feature of the second process data to the second performance data associated with the one or more substrates”)
As per claim 3
Bastani teaches all claim limitations previously rejected under claim 1’s 102 rejection , See claim 1’s 102 rejection.
Bastani teaches wherein the instructions are further configured to cause the computer system to provide the augmented data as input to the machine learning model to train the machine learning model to conform predictions of the one or more imaging characteristics according to the physical trend. (Figure 5, Detailed Description: “ At 406, a second model may be trained on second process data and second performance data. The second process data and second performance data may be associated with the one or more substrates. At 408, the trained second model may be used to determine a contribution of a feature of the second process data to the second performance data associated with the one or more substrates…The process and performance data may form input-output training pairs for the local model LM. In order to form process-performance data pairs, the augmented process data may be used to generate augmented performance data. In one example implementation, augmented process data may be provided as input to the global model GM. The output produced by the global model GM may be augmented performance data 608 corresponding to the augmented process data input…The data used to train the local model LM may be referred to as local process data and local performance data 610. Local process data may comprise the augmented process data…local model LM may be based on either both augmented and subset data, or on augmented data alone, respectively. The local model LM may be referred to as a surrogate model, due to it being trained at least in part on augmented data…the local model LM may provide, once it has been trained, an estimation of the contribution of different process features to the performance data output by the model” )
As per claim 4
Bastani teaches all claim limitations previously rejected under claim 1’s 102 rejection , See claim 1’s 102 rejection.
Bastani teaches the first data comprises previously determined measurements of a pattern on the substrate, and/or information indicative of a physical behavior of the pattern on the substrate resulting from the patterning process (Detailed Description “the first process data and first performance data may comprise historical data. Historical data may relate to earlier runs of the patterning processes for which performance is analysed in the method. Process data may comprise a plurality of data points for one or more process features.” )
As per claim 5
Bastani teaches all claim limitations previously rejected under claim 1’s 102 rejection , See claim 1’s 102 rejection.
Bastani teaches wherein the one or more imaging characteristics comprise a critical dimension, an edge location, a curvature, a pitch, a symmetry, or a rotation. (Detailed Description “Historical data may relate to earlier runs of the patterning processes for which performance is analysed in the method. Process data may comprise a plurality of data points for one or more process features. Process features may relate to properties of a substrate, for example overlay, alignment, critical dimension, levelling,”)
As per claim 6
Bastani teaches all claim limitations previously rejected under claim 1’s 102 rejection , See claim 1’s 102 rejection.
Bastani teaches wherein the instructions configured to cause the computer system to generate the augmented data are further configured to cause the computer system to generate the augmented data based on measurements of the one or more imaging characteristics included in the first data (Figure 5, Detailed Description: “ Augmented process data may have the same type of content as the process data of a substrate in the subset…Local process data may comprise the augmented process data…Optionally, the local process data may also comprise the process data of the subset 170. Similarly, local performance data may comprise augmented performance data, and optionally performance data of the subset 170”)
As per claim 7
Bastani teaches all claim limitations previously rejected under claim 6’s 102 rejection , See claim 6’s 102 rejection.
Bastani teaches wherein the instructions configured to cause the computer system to generate the augmented data are further configured to cause the computer system to mathematically interpolate between the measurements of a given imaging characteristic to determine additional measurements of the given imaging characteristic (Detailed description: “The output produced by the global model GM may be augmented performance data 608 corresponding to the augmented process data input… The local model LM may provide, once it has been trained, an estimation of the contribution of different process features to the performance data output by the model.” Both local and global models interpolate, as is the nature of neural networks. Bastani says that “generating augmented process data may comprise determining a range of values for a process feature, and generating a plurality of samples within the range of values for the process feature, wherein the augmented process data comprises the plurality of samples.” The measurements being interpolated are those within each sample. The additional measurements of the given imaging characteristic (which Bastani states are process features such as overlay, levelling, alignment, critical dimension, ) can be seen as the estimation of contribution of different process features. Bastani also states that “The method of determining a predicted yield may be repeated for each substrate in a set of substrates 160. The resulting prediction qualities 150 are illustrated in the graph 152 for each substrate in set 160. Based on the prediction qualities for the substrates in the set, a selection of one or more substrates may be identified to form a subset 170.” The predicted yield interpolated from the GM model is yet another measurement of the imaging characteristic)
As per claim 8
Bastani teaches all claim limitations previously rejected under claim 6’s 102 rejection , See claim 6’s 102 rejection.
Bastani teaches wherein the instructions configured to cause the computer system to generate the augmented data are further configured to cause the computer system to calibrate a physical model associated with the physical trend using the measurements, (Figure 5, Detailed description “ The method of determining a predicted yield may be repeated for each substrate in a set of substrates 160. The resulting prediction qualities 150 are illustrated in the graph 152 for each substrate in set 160. Based on the prediction qualities for the substrates in the set, a selection of one or more substrates may be identified to form a subset 170. …If a local model is trained on a small amount of data, this may cause overfitting. In order to avoid overfitting, the method may generate data based on the subset process data. The generated data may be used to train the local model LM.”) the calibrated physical model to predict additional measurements that conform to the physical trend (Detailed Description: “The local model LM may provide, once it has been trained, an estimation of the contribution of different process features to the performance data output by the model”)
As per claim 9
Bastani teaches all claim limitations previously rejected in claim 8’s 102 rejection. See claim 8’s 102 rejection.
Bastani teaches wherein the physical model comprises a multi- dimensional algorithm having terms that collectively simulate the physical trend and/or the patterning process. (Figure 5)
As per claim 10
Bastani teaches all claim limitations previously rejected in claim 8’s 102 rejection. See claim 8’s 102 rejection.
Bastani teaches wherein the instructions configured to cause the computer system to generate the augmented data are further configured to cause the computer system to calibrate using the measurements, a residue model configured to compensate for error in the predicted additional measurements (Figure 5 Detailed Description “The identification step may comprise a selection of one or more substrates in the set for which the quality of prediction by the global model is above a predetermined threshold and/or a predetermined loss metric…the threshold may depend on the expected use case of the pattern of the dies on the substrate. The use case may relate to a use of the analysis data of the feature contributions, and/or may relate to an intended use of the dies themselves. An example threshold may be above 75% correspondence between the prediction and measured data. Example loss metrics include Mean Square Error (MSE) loss, logarithmic (log) loss, cross entropy, classification accuracy …Additionally to selection based on a quality of prediction, the identification of a subset 170 of a set of substrates may include a selection of substrates that have a yield within a predetermined range Whether the yield falls within the predetermined range may be assessed based on a predicted yield or a measured yield. The predetermined range may be set to exclude substrates which have a very high yield…the predetermined range may also exclude substrates which have a very low yield. The threshold for very low yield may, for example, be set in comparison to other substrates in the set. “ Furthermore, the relationship between LM and GM models is that of a model that refines the other’s contribution/quality performance estimation) and generate the augmented data using the physical model and the residue model, (Figure 5 , Figure 6 , Detailed description “As described above, the global model GM may be used to select a subset of substrates based on a quality of predicted performance data 120 output by the global model GM. Once the one or more substrates have been selected, the process data for the subset may be analysed to determine a contribution of process features to the yield loss on the substrates. The analysis may use a local model LM, as shown in Figure 6….Augmented process data may be obtained by determining a range 604 around process data of one or more process features, and generating 606 data points within that range.) wherein the residue model comprises a purely mathematical model calibrated by fitting errors in predictions from the physical model to the measurements. (Detailed Description “the method includes the identification of a subset of the set of substrates that have been provided as input to the global model. The identification step may comprise a selection of one or more substrates in the set for which the quality of prediction by the global model is above a predetermined threshold and/or a predetermined loss metric…An example threshold may be above 75% correspondence between the prediction and measured data. Example loss metrics include Mean Square Error (MSE) loss, logarithmic (log) loss, cross entropy, classification accuracy, )
As per claim 11
Bastani teaches all claim limitations previously rejected under claim 6’s 102 rejection , See claim 6’s 102 rejection.
Bastani teaches wherein the instructions configured to cause the computer system to generate the augmented data are further configured to cause the computer system to generate the augmented data is based on the measurements and symmetry in the pattern resulting from the patterning process. (Figure 5, Detailed description: Process features may for example include overlay OVL, levelling LVL, alignment AL, critical dimension CD, and context data CXT” Overlay directly deals with pattern symmetry in wafer metrology by analyzing the centers of symmetry in geometric shapes within the wafers pattern. Alignment also centers and matches a wafer pattern to a tools axis which would constitute dealing with symmetry in the pattern)
As per claim 12
Bastani teaches all claim limitations previously rejected under claim 1’s 102 rejection , See claim 1’s 102 rejection.
Bastani teaches wherein the instructions configured to cause the computer system to generate the augmented data are further configured to cause the computer system to determine the physical trend based on outputs from a trained physical model, (Figure 5, Figure 6 Detailed Description: “The performance data may comprise predicted yield data 120 for the substrate. The global model may also provide an estimation of the contribution 140 of a plurality of features to the predicted performance data…The local model LM may provide, once it has been trained, an estimation of the contribution of different process features to the performance data output by the model…The models GM and LM may determine a contribution of the features based on calculations, without generating a graphic representation.”) the trend described by relative relationships between values of imaging characteristics predicted by the physical model, rather than absolute values of the imaging characteristics themselves. (Figure 5 label 180 and 140, Detailed Description; “The local model may be an interpretable model. An interpretable model may be a type of model for which one or more relationships between the model input and the model can be explained intuitively…The global model may also provide an estimation of the contribution 140 of a plurality of features to the predicted performance data…Once the local model LM has been trained, it may be used to determine a predicted feature importance 180 for the subset 170…The trained local model LM may be interpretable, so that the trained local model LM may be used to estimate a relation between the one or more features and the local performance data. Based on the interpreted estimated relation obtained from the trained model, the contribution one or more features to the local performance data”)
As per claim 13
Bastani teaches all claim limitations previously rejected under claim 1’s 102 rejection , See claim 1’s 102 rejection.
Bastani teaches wherein the physical trend comprises a symmetry trend, an orientation trend, a focus trend, a dose trend, a through pitch trend, a linearity trend, and/or a through critical dimension trend (Figure 5, Detailed Description : “A method for determining a contribution of a process feature to the performance of a patterning process will now be described in more detail in relation to Figure 5…Process features may comprise any feature of interest from a patterning process. Process features may for example include overlay OVL, levelling LVL, alignment AL, critical dimension CD, and context data CXT.”) and wherein the augmented data is associated with a normalized standard deviation based loss function. (Figure 5, Detailed Description: “The identification step may comprise a selection of one or more substrates in the set for which the quality of prediction by the global model is above a predetermined threshold and/or a predetermined loss metric…Example loss metrics include Mean Square Error (MSE) loss, logarithmic (log) loss, cross entropy, classification accuracy” MSE can be interpreted as a normalized standard deviation for example. This loss measures between predictions and targets. It averages the squared errors, which directly mirrors the variance (the square of standard deviation) of the prediction errors when the mean error is zero.)
As per claim 14
Bastani teaches all claim limitations previously rejected under claim 12’s 102 rejection , See claim 12’s 102 rejection.
Bastani teaches wherein the machine learning model is trained to predict the one or more imaging characteristics according to the physical trend based on a loss function (Figure 5, Detailed Description “The identification step may comprise a selection of one or more substrates in the set for which the quality of prediction by the global model is above a predetermined threshold and/or a predetermined loss metric”) configured to cause the machine learning model to fit the relative relationships between values of imaging characteristics predicted by the physical model, rather than the absolute values of the imaging characteristics themselves. (Figure 5 label 180 and 140, Detailed Description; “The local model may be an interpretable model. An interpretable model may be a type of model for which one or more relationships between the model input and the model can be explained intuitively…The global model may also provide an estimation of the contribution 140 of a plurality of features to the predicted performance data…Once the local model LM has been trained, it may be used to determine a predicted feature importance 180 for the subset 170…The trained local model LM may be interpretable, so that the trained local model LM may be used to estimate a relation between the one or more features and the local performance data. Based on the interpreted estimated relation obtained from the trained model, the contribution one or more features to the local performance data”)
As per claim 15
Bastani teaches all claim limitations previously rejected under claim 1’s 102 rejection , See claim 1’s 102 rejection.
Bastani wherein the physical trend is known based on prior pattern variation on the substrate and/or prior patterning process variation. (Detailed Description: Figure 4 depicts a flow diagram of a method of determine the contribution of a process feature to the performance of a process of patterning substrates…The first process data and first performance data may comprise historical data. Historical data may relate to earlier runs of the patterning processes for which performance is analysed in the method. Process data may comprise a plurality of data points for one or more process features. Process features may relate to properties of a substrate, for example overlay, alignment, critical dimension, levelling, etc.)
As per claim 17
Bastani teaches all claim limitations previously rejected under claim 1’s 102 rejection , See claim 1’s 102 rejection.
Bastani teaches wherein the first data comprises previously determined measurements of a first pattern or set of patterns on a first substrate produced by a first patterning process (Detailed Description: “The first process data and first performance data may comprise historical data. Historical data may relate to earlier runs of the patterning processes for which performance is analysed in the method”) and wherein the physical trend is based on measurements of the first pattern or set of patterns on the first substrate associated with the physical trend (Figure 5, Detailed Description: “The method described herein uses two models, in order to distinguish between the contributions of different process features to the yield of one or more substrates. Based on the different contributions made by different process features to the yield loss, potential root causes for yield loss may be identified… Figure 4 depicts a flow diagram of a method of determine the contribution of a process feature to the performance of a process of patterning substrates…A method for determining a contribution of a process feature to the performance of a patterning process…The process data may relate to one or more process features. The process data may comprise measured data of a process feature. For example, a process feature may be overlay, and the process data may comprise overlay data. Alternatively, or in addition to measured data, the process data may comprise estimated and/or inferred data….The process data may comprise data relating to process features for different layers of the substrate. Process features may comprise any feature of interest from a patterning process. Process features may for example include overlay OVL, levelling LVL, alignment AL, critical dimension CD, and context data CXT.”)
As per claim 18
Bastani teaches all claim limitations previously rejected under claim 17’s 102 rejection , See claim 17’s 102 rejection.
Bastani teaches wherein the first patterning process comprises one or more semiconductor manufacturing patterning processes that have substantially similar process conditions with a second patterning process (Figure 5, Detailed Description: “Process data may also comprise manufacturing process data related to lithographic patterning and/or other manufacturing processes for the substrates The metrology data and/or process data provided to the method may relate to process features. The method may also use yield data obtained for the one or more substrates. Examples of process features may include metrology data such as overlay, levelling, alignment, critical dimension, or lithographic apparatus settings, recipe settings for patterning steps, context data, etc. .” ) wherein the second patterning process comprises a target patterning process which is simulated and/or for which a model is constructed (Figure 5) wherein the augmented data comprises new data generated based on the physical trend and/or by the physical model (Detailed Description “Augmented process data may be obtained by determining a range 604 around process data of one or more process features, and generating 606 data points within that range. A first example method of generating augmented process data may comprise modelling a parametric distribution for a process feature such as overlay, alignment, critical dimension, etc. Augmented process data may then be obtained by generating a plurality of samples according to the modelled parametric distribution. A second example method of generating augmented process data may comprise determining a range of values for a process feature. “) and/or a subset of the previously determined measurements from the first data that conforms to the physical trend (Detailed description “ Augmented process data may have the same type of content as the process data of a substrate in the subset. For example, augmented and substrate process data may comprise one or more of the same process features, the same number of layers, the same amount of data points, “)
As per claim 19
Bastani teaches all claim limitations previously rejected under claim 17’s 102 rejection , See claim 17’s 102 rejection.
Bastani teaches wherein the augmented data is combinable with second measurements from a second pattern or set of patterns on a second substrate produced by the second patterning process (Figure 6, Detailed Description: “The analysis may use a local model LM, as shown in Figure 6. In order to determine a contribution of features to yield loss for a subset of substrates 170…Local process data may comprise the augmented process data. Optionally, the local process data may also comprise the process data of the subset 170 Similarly, local performance data may comprise augmented performance data, and optionally performance data of the subset 170. Depending on whether the process data and performance data of the subset are included in the local data or not, training of the local model LM may be based on either both augmented and subset data,” ) wherein the subset of the previously determined measurements from the first data that conforms to the physical trend comprises data associated with a previous mask when the first patterning process and the second patterning process are similar. (Figure 5, Figure 6, Detailed description “The first model may be trained on first process data and first performance data. The first process data and first performance data may comprise historical data. Historical data may relate to earlier runs of the patterning processes for which performance is analysed in the method….Process data may comprise a plurality of data points for one or more process features. Process features may relate to properties of a substrate, for example overlay, alignment,…The first model may have been applied to a set of substrates for making the prediction. The one or more substrates may be a subset of the set of substrates.” Overlay and alignment are similar patterning processes that are all geometric positioning process with a focus on spatial orientation and positioning of the wafer during photolithography. Furthermore, Bastani states in the Detailed description that “The term "reticle", "mask" or "patterning device" as employed in this text may be broadly interpreted as referring to a generic patterning device that can be used to endow an incoming radiation beam with a patterned cross-section” A person of ordinary skill in the art is fully aware that lithographic patterning processes use a mask to define the overlay and alignment features. )
As per claim 20
Bastani teaches all claim limitations previously rejected under claim 19’s 102 rejection , See claim 19’s 102 rejection.
Bastani teaches wherein the subset of the previously determined measurements from the first data that conforms to the physical trend comprises data associated with different locations on a current mask when the second patterning process is adjusted relative to the first patterning process. (Detailed description Figure 5 “The first model may be trained on first process data and first performance data. The first process data and first performance data may comprise historical data…The process data may comprise data relating to process features for different layers of the substrate…Process features may comprise any feature of interest from a patterning process. Process features may for example include overlay OVL, levelling LVL, alignment AL, critical dimension CD, and context data CXT. Process data may be provided for one or more process features for each of a plurality of layers, labelled in Figure 5 by a letter A, ... X.” The “current mask” would be the mask used for whichever specific layer is in question. The locations of the current mask associated with each layer will be of different locations within the layer its is associated with. The patterning process of preceding or subsequent layers would necessarily have to be adjusted between layers as well as patterning processes i.e. “adjusted relative to the first patterning process. “ )
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
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.
Claim 16 is/are rejected under 35 U.S.C. 103 as being unpatentable over Bastani et al (Bastani hereinafter EP 3767392 A1 “METHOD AND APPARATUS FOR DETERMINING FEATURE CONTRIBUTION TO PERFORMANCE” ) in view of Cao et al (Cao hereinafter US-20200380362-A1)
As per claim 16
Bastani teaches all claim limitations previously rejected under claim 1’s 102 rejection , See claim 1’s 102 rejection.
Bastani does not teach wherein output from the machine learning model is configured to be provided to a cost function to facilitate determination of costs associated with individual patterning process variables.
Cao teaches wherein output from the machine learning model is configured to be provided to a cost function to facilitate determination of costs associated with individual patterning process variables. (Paragraph [0098] “the machine learning architecture may be divided into several parts: (i) training of individual process model (e.g., 8004, 8006, and 8008), further discussed later in the disclosure, (ii) coupling the individual process models and further training and/or fine-tuning the trained process models based on a first training data set (e.g., printed patterns) and a first cost function (e.g., difference between printed patterns and predicted patterns), further discussed in FIG. 9, and (iii) using the trained process models to train another machine learning model (e.g., 8002) configured to predict mask pattern (e.g., including OPC) based on a second training data set (e.g., a target pattern) and a second cost function (e.g., EPE between the target pattern and the predicted pattern…using the trained process model may be considered as an unsupervised learning, where target patterns are compared with the predicted patterns based on a cost function” Paragraph [0101] “The cost function (e.g., the first cost function) may be defined based on a difference between the experimental data (i.e., printed patterns on a substrate) and the output of the third model 8008. For example, the cost function may be a metric (e.g., RMS, MSE, MXE etc.) based on a parameter (e.g., CD, overlay) of the patterning process determined based on the output of the third trained model” Paragraph [0105] “The cost function can be based on image pixel intensity difference, contour to contour difference, or CD difference,’ Paragraph [0177] “in process, P1449, a gradient map may be generated based on the cost function which is further used to determine values to the mask variables corresponding to the initial image (e.g., pixel values of the mask image) such that the cost function is reduced.” )
Accordingly, a person of ordinary skill in the art, at the time this invention was effectively filed would have found it obvious to modify Bastani’s methodology with Cao’s concept of feeding a machine learning model’s output into a cost function to facilitate determination of costs associated with individual patterning process variables. Both Bastani and Cao are in the same realm of endeavor which comprises predicting wafer via physical trend interpretation. A person of ordinary skill in the art is aware that a model trained to predict imagining characteristics according to known physical trends produces outputs that are directly correlated with manufacturing quality. Knowing this, a person of ordinary skill in the art would use those predictions in a cost function to allow the optimization algorithm to determine which process variable minimize predicted defects. A person of ordinary skill in the art is also aware that putting the machine learning output into the cost function prioritizes process variables that affect wafer quality the most. Including the prediction into the cost function penalizes variables that degrade the desired product the most and steer optimization toward the more suited or robust option. The predicted imagining characteristics provide quantifiable estimates of process quality that are directly related to lithography performance. Doing so enables optimization of patterning parameters using physically meaningful predictions and facilitates selection of variables that minimize predicted imagining defects. A person of ordinary skill in the art would see these outcomes as common goals/practice.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to SHANE WRENSFORD CODRINGTON whose telephone number is (571)272-8130. The examiner can normally be reached 8:00am-5pm.
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, Matthew Bella can be reached at (571) 272-7778. 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.
/SHANE WRENSFORD CODRINGTON/Examiner, Art Unit 2667
/MATTHEW C BELLA/Supervisory Patent Examiner, Art Unit 2667
Prosecution Insights