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 .
DETAILED ACTION
Continued Examination Under 37 CFR 1.114
A request for continued examination under 37 CFR 1.114, including the fee set forth in 37 CFR 1.17(e), was filed in this application after final rejection. Since this application is eligible for continued examination under 37 CFR 1.114, and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114. Applicant's submission filed on 3/30/2026 has been entered.
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 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.
Claims 9-10 and 17-18 are rejected under 35 U.S.C. 103 as being unpatentable over Jeffrey Drue David et al. (US 20150120243), hereinafter ‘David’ in view of Yoonsung Bae et al. (US 20220121956), hereinafter ‘Bae’, in further view of TZONG-DAW WU et al. (US 20190120610), hereinafter ‘Wu’, in further view of Nachiketa CHAUHAN et al. (US 20220316863), hereinafter ‘Chauhan’.
With regards to Claim 9, David discloses
A control method (controlling polishing of the product substrate [0065]) comprising:
receiving a spectrum signal including thickness information of a wafer substrate from a spectroscopic monitoring device (the measured spectrum contains noise data that may affect the precision of the characterizing value or the thickness measurement [0027]; The light detector 164 can be a spectrometer [0042]; properties of the spectrum of the reflected light changes as a thickness of the film changes [0043]);
determining a noise parameter for a noise included in the spectrum signal (Such a small fraction can correspond to the undesirable noise in the spectra. Accordingly, these terms .SIGMA..sub.i>t.sigma..sub.iu.sub.iv.sub.i.sup.T can be discarded [0055]; The light intensity data may also contain other noise data that originate from sources, such as variation in structural dimensions of the device (e.g., critical dimensions, sidewall angle, etc.), pad window variations (e.g., changing absorption in the shorter wavelengths), other process influences (e.g., slurry pooling) [0046]),
performing a noise reduction process of reducing a noise from the spectrum signal based on the determined noise parameter (The noise data contained in the spectra represented by the matrix A can be reduced using singular value decomposition (SVD), CUR matrix approximation, or principal component analysis, each followed by dimension reductions. These techniques can identify similar components in a dataset, e.g., thickness in this example, while filtering out noise components in the dataset, e.g., due to variations in the thickness of an underlying layer. The techniques can be used in the in-situ systems or the in-line systems [0052]; Fig.7, Step 703);
determining an estimated thickness value of the wafer substrate using a thickness estimation model configured to receive the spectrum signal having the reduced noise as an input (Characterizing values, e.g., thicknesses, can be extracted (704) from the spectra with reduced noise. The characterizing values could be generated by identifying a matching reference spectrum from a library of reference spectra, or by fitting an optical model to the collected spectrum. The characterizing values are then used (706) in controlling polishing of the product substrate [0065]), and
controlling operations of a polishing apparatus for the wafer substrate based on the determined estimated thickness value (FIG. 7 shows an example process 700 of implementing the noise reduction discussed above in controlling polishing of a product substrate … Characterizing values, e.g., thicknesses, can be extracted (704) from the spectra with reduced noise … The characterizing values are then used (706) in controlling polishing of the product substrate [0065]).
David also discloses using an optical model [0065].
However, David does not specifically disclose using a noise reduction model configured to receive the spectrum signal as an input, wherein the noise reduction model is configured to receive the spectrum signal as an input and estimate the noise parameter as an output.
Bae discloses using a noise reduction model (Figs. 1 and 3) configured to receive the spectrum signal as an input, wherein the noise reduction model is configured to receive the spectrum signal as an input and estimate the noise parameter as an output (generating the augmented spectra by physically modeling (i.e., representing analytically) an optical system of the wafer inspection device 100 to compensate for noise parameters of the wafer inspection device 100 and arbitrarily changing correction coefficients for compensating for the noise parameters in the physical modeling [0048]; When a spectrum is input to the input layer IL, the second output layer OL2 may output noise of a spectrum according to a change in a noise parameter. Based on the input spectrum and the noise of the spectrum output from the second output layer OL2, the augmented spectra as shown in (b) of FIG. 4 may be provided [0061]).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify David in view of Bae to use a noise reduction model configured to receive the spectrum signal as an input to obtain a corrected optical spectrum obtained by excluding an influence of noise (the augmented spectra as shown in (b) of FIG. 4 may be provided, Bae [0061]).
David also discloses an optical model [0024, 0025].
However, David is silent on the noise reduction model is trained through the operations of: generating a training spectrum signal according to a thickness of a wafer substrate using an optical model; generating a training spectrum signal having a noise by applying a noise based on a noise parameter to the training spectrum signal; calculating a similarity between the training spectrum signal having the noise and an actually measured spectrum signal; and when the similarity satisfies a set condition, training the noise reduction model using the training spectrum signal having the noise.
Bae discloses generating a training spectrum signal using training data [0062, 0067, 0070-0072] and as discussed in Claim 1.
Wu discloses generating a training spectrum signal according to a thickness of a wafer substrate using an optical model (By adding the 3% noise parameter, a large number of different spectra would be generated. By having the modified spectra like T3 to train the artificial neural network, the corresponding trend of spectrum with respect to the same film thickness would be realized as well. Namely, by having the artificial neural network to be trained by the modified spectra like T3, while in measuring the instant spectrum, the artificial neural network can then estimate the corresponding film thickness by judging the trend of the spectrum [0046]).
Wu also discloses using a similarity between the training spectrum signal having the noise and an actually measured spectrum signal; and when the similarity satisfies a set condition, training the noise reduction model using the training spectrum signal having the noise (by adding the noise parameter into the simulated spectrum so as to generate the corresponding modified spectrum for the input layer of the artificial neural network, then a large amount of different spectra can be generated. By having the modified spectra to train the artificial neural network, the corresponding trend of spectrum with respect to the same film thickness would be realized as well [0050]).
CHAUHAN discloses calculating a similarity between the training spectrum signal having the noise and an actually measured spectrum signal; and when the similarity satisfies a set condition, training the noise reduction model using the training spectrum signal having the noise (obtaining similarities by calculating a similarity between each of the sample spectra and a representative spectrum, and producing a film-thickness estimation model by performing machine learning using training data including the sample features, the similarities, and film thicknesses corresponding to the sample spectra [0010]).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify David in view Bae, Wu, and Chauhan to train the noise reduction model is trained through the claimed steps to closely resemble a real measured spectrum that is known to comprise noise (it is found that, in comparison with the examples having no noise added, the examples with appropriate noise addition would be positive for the artificial neural network to estimate a precise film thickness, Wu [0051]) for accuracy of machine learning in thickness estimation (As a result, the accuracy of machine learning is improved, and the film-thickness estimation accuracy of the film-thickness estimation model is also improved (Chauhan [0132]).
With regards to Claim 10, David additionally discloses determining a target spectrum signal that is most similar to the spectrum signal having the reduced noise among theoretically generated fake spectrum signals according to a thickness of the wafer substrate; and determining a thickness of the wafer substrate corresponding to the target spectrum signal as the estimated thickness value of the wafer substrate (compare the measured spectrum to a plurality of reference spectra from a library, and identify a best-matching reference spectrum [0024]; The one or more spectra on which an endpoint determination is based can include a target spectrum, a reference spectrum, or both. [0043]; FIG. 5D illustrates a schematic plot of a part of a measured spectrum and a modified spectrum with reduced noise [0018]; For finding a match, the thickness value associated with the reference spectrum can be identified [0025]; The methods and systems of this disclosure reduce the noise in the spectrum to improve the precision of thickness measurement [0027]; The constructed matrix A can have all the major components that represent the thickness of the substrate, while a large portion of the noise data is removed as compared to the original matrix A.[0060]).
With regards to Claims 17-18, David in view of Bae discloses the claimed limitations a s discussed in Claims 9-10, respectively.
Response to Arguments
35 U.S.C. 101
Applicant’s arguments, see Applicant Arguments/Remarks, filed 3/4/2026, with respect to Claim 9 have been fully considered and are persuasive in view of the amendments. The 35 U.S.C. 101 rejection of Claim 9 has been withdrawn.
35 U.S.C. 103
With regards to former Claim 12, the Applicant's arguments filed 3/4/2026 have been fully considered but they are not persuasive.
The Applicant argues (p.9): However, Wu and Chauhan do not disclose a criterion (i.e., a gate) for training a noise-reduction model. Furthermore, the office action fails to identify any disclosure where a model is trained to output a quantitative noise parameter via supervised labels, or where inclusion of a training example is gated by a similarity to an actually measured spectrum. Wu and Chauhan teaches experimental spectra, which is different from the actually measured spectra as a filtering criterion to train the noise reduction model. Accordingly, the cited references including Wu and Chauhan do not teach the operation of training a noise-parameter regressor gated by similarity to actually measured spectra.
Further, in the present application, the training operation of the noise reduction model is configured to narrow the population of training sample and impose a process- linked criterion. During the training process of the noise reduction model, the training spectrum signal having a noise is discarded and excluded from the training via the criteria, which is different from the cited references including Wu and Chauhan.
The Examiner respectfully disagrees. With regards to “actually measured spectra as a filtering criterion to train the noise reduction model”, i.e. “inclusion of a training example is gated by a similarity to an actually measured spectrum”, the Examiner submits that training database in Wu is based on modified actually measured data [0046]. Additionally, the “three spectra” in Chauhan (see below) are also based on measured data (Fig.12).
The claimed “set criteria” used in similarity assessment, under the BRI, is also disclosed by the references (“corresponding trend” in Wu and/or similarity to “the three representative spectra” in Chauhan). These criteria is characteristic of the argued “process- linked” criterion.
The Examiner also notes that “The training spectrum signal having a noise is discarded and excluded from the training via the criteria” is not claimed. In fact, the opposite is true: the claim requires training on a spectrum having the noise.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to ALEXANDER SATANOVSKY whose telephone number is (571)270-5819. The examiner can normally be reached on M-F: 9 am-5 pm.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Catherine Rastovski can be reached on (571) 270-0349. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of an application may be obtained from the Patent Application Information Retrieval (PAIR) system. Status information for published applications may be obtained from either Private PAIR or Public PAIR. Status information for unpublished applications is available through Private PAIR only. For more information about the PAIR system, see http://pair-direct.uspto.gov. Should you have questions on access to the Private PAIR system, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative or access to the automated information system, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/ALEXANDER SATANOVSKY/
Primary Examiner, Art Unit 2857
Prosecution Insights