CN114963979A - 3D NAND memory laminated structure critical dimension measuring method based on deep learning - Google Patents
3D NAND memory laminated structure critical dimension measuring method based on deep learning Download PDFInfo
- Publication number
- CN114963979A CN114963979A CN202210487425.4A CN202210487425A CN114963979A CN 114963979 A CN114963979 A CN 114963979A CN 202210487425 A CN202210487425 A CN 202210487425A CN 114963979 A CN114963979 A CN 114963979A
- Authority
- CN
- China
- Prior art keywords
- network model
- nand memory
- data
- critical dimension
- series
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 230000015654 memory Effects 0.000 title claims abstract description 73
- 238000013135 deep learning Methods 0.000 title claims abstract description 21
- 238000000034 method Methods 0.000 title claims abstract description 21
- 230000003595 spectral effect Effects 0.000 claims abstract description 43
- 238000001228 spectrum Methods 0.000 claims abstract description 25
- 238000012549 training Methods 0.000 claims description 26
- 238000012360 testing method Methods 0.000 claims description 23
- 230000006870 function Effects 0.000 claims description 10
- 238000000691 measurement method Methods 0.000 claims description 9
- 238000004088 simulation Methods 0.000 claims description 8
- 238000005259 measurement Methods 0.000 description 8
- 238000013528 artificial neural network Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 2
- 230000001066 destructive effect Effects 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000005265 energy consumption Methods 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 230000001681 protective effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
- G01B11/00—Measuring arrangements characterised by the use of optical techniques
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
- G01B11/00—Measuring arrangements characterised by the use of optical techniques
- G01B11/02—Measuring arrangements characterised by the use of optical techniques for measuring length, width or thickness
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
- G01B11/00—Measuring arrangements characterised by the use of optical techniques
- G01B11/02—Measuring arrangements characterised by the use of optical techniques for measuring length, width or thickness
- G01B11/06—Measuring arrangements characterised by the use of optical techniques for measuring length, width or thickness for measuring thickness ; e.g. of sheet material
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
- G01B11/00—Measuring arrangements characterised by the use of optical techniques
- G01B11/22—Measuring arrangements characterised by the use of optical techniques for measuring depth
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Molecular Biology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Length Measuring Devices By Optical Means (AREA)
Abstract
The invention provides a 3D NAND memory laminated structure key dimension measuring method based on deep learning, which comprises the following steps: establishing a matched forward network model and a matched series network model based on the key sizes of different laminated structures of the 3D NAND memory and a preset data set of the spectrum data; in the forward network model, the key size P is used as input, and the output is spectral data O; in the series network model, the spectral data R is used as the input of a reverse network model, and the key size P' output by the reverse network model is used as the input of a forward network model, so that the series network model outputs spectral data O; measuring to obtain actual spectral data of the target 3D NAND memory; after the actual spectrum data is input into the series network model, the key size output by the reverse network model is used as the key size of the laminated structure of the target 3D NAND memory, and the reverse network model is connected with the forward network model in series, so that the network is converged smoothly, and the model prediction accuracy is improved.
Description
Technical Field
The invention belongs to the technical field of 3D NAN memories, and particularly relates to a 3D NAND memory laminated structure key dimension measuring method based on deep learning.
Background
Because the 2D NAND memory structure faces practical expansion limits, the 3D NAND technology is now adopted, and by vertically stacking multiple layers of memory cells, a smaller space has higher storage capacity, and this technology can greatly save cost, reduce energy consumption, and improve performance. Since the critical dimensions of the 3D NAND memory, such as the film thickness of each layer, the step length uniformity, and the hole depth, directly affect the device performance of the memory, the critical dimensions of the memory need to be measured accurately during the manufacturing process.
At present, for the measurement of the critical dimension of a 3D NAND memory, a scanning Transmission Electron Microscope (TEM) is mostly used for measurement after slicing, and the measurement method is destructive, slow in detection speed and incapable of on-line detection. Optical scatterometry and Optical Critical Dimension (OCD) are one of the mainstream in-line inspection techniques in semiconductor manufacturing, and have the advantages of rapidness, accuracy and non-destructive property. However, the OCD has a crosstalk problem, that is, structures with different parameters have similar spectrograms, which also restricts effective measurement of the critical dimension of the 3D NAND memory.
Deep learning is a new research direction in the field of machine learning, and results are predicted by training certain data samples by using a neural network. Therefore, the deep learning technology is used for training the one-to-one correspondence relationship between the spectrum data and the key size of the 3D NAND memory, and the key size of the 3D NAND memory can be predicted efficiently. However, since two sets of similar spectral data may exist in two 3D NAND memories with different critical dimensions, the training of the neural network is difficult to converge, and the model prediction accuracy is greatly limited.
Disclosure of Invention
The invention aims to provide a method for measuring the key dimension of a 3DNAND memory laminated structure based on deep learning aiming at the defects of the prior art, wherein a series network model formed by connecting a forward network model and a reverse network model in series is adopted, so that the network is converged smoothly, parameters can be predicted accurately, and the model prediction accuracy is improved.
In order to solve the technical problems, the invention adopts the following technical scheme:
A3D NAND memory stack structure critical dimension measurement method based on deep learning, the measurement method comprises the following steps:
establishing a matched forward network model and a series network model comprising the forward network model and a reverse network model based on the key sizes of different laminated structures of the 3D NAND memory and a preset data set of spectrum data; in the forward network model, the key size P is used as network input, and the network output is spectral data O; in the series network model, spectral data R is used as the network input of the reverse network model, and then the key size P' output by the reverse network model is used as the input of the forward network model, so that the series network model outputs spectral data O;
measuring the laminated structure of the target 3D NAND memory to obtain actual spectrum data;
and inputting the measured actual spectrum data into the series network model, and taking the critical dimension output by the reverse network model as the critical dimension of the target 3D NAND memory laminated structure.
Further, the establishing process of the forward network model comprises the following steps:
constructing a first data set based on the preset data set, wherein the first data set comprises a plurality of data groups, and each data group comprises spectral data R and corresponding critical dimension P of different laminated structures of the 3D NAND memory;
constructing a primary forward network model;
and dividing the first data set into a training set and a testing set, and training and testing the preliminary forward network model so as to obtain the forward network model meeting preset conditions.
Further, the establishing process of the tandem network model comprises the following steps:
constructing a second data set based on the preset data set, wherein the second data set comprises a plurality of data groups, and each data group comprises spectral data R and corresponding critical dimension P of different laminated structures of the 3D NAND memory;
constructing a preliminary series network model;
and dividing the second data set into a training set and a testing set, and training and testing the preliminary tandem network model so as to obtain the tandem network model meeting preset conditions.
Further, the critical dimension P may be at least any of the following parameters: the film thickness, the step length uniformity and the hole depth of each layer of the 3D NAND memory laminated structure,
the spectral data R comprises n discrete points, and the array R ═ R 1 ，r 2 ，…，r n ]It is shown that there are m critical dimensions of the stacked structure, which are represented by the array P ═ P 1 ，p 2 ，…，p m ]Is represented by the formula, wherein p i Is the critical dimension of the ith die,
in the forward network model, the key sizes P of different laminated structures are used as network input, and the network output is spectral data O ═ O 1 ，o 2 ，…，o n ]，
In the preliminary model of the series network, taking spectral data R as the input of the reverse network model to obtain the critical dimension P' ═ P output by the reverse network model 1 ′，p 2 ′，…，p m ′]And taking P' as the input of the forward network model, and enabling the tandem network model to output spectral data O ═ O 1 ，o 2 ，…，o n ]。
Further, in the process of establishing the forward network model and the series network model, the loss functions of the corresponding models are respectively defined as
Training and testing the corresponding network model by minimizing the loss function, and obtaining the forward network model and the series network model when the spectral data output by the series network model meets the preset precision requirement.
Further, the preset data set is a simulation data set obtained by simulating the critical dimension and the spectrum data of different laminated structures of the 3D NAND memory through simulation equipment.
Compared with the prior art, the invention has the beneficial effects that: the invention discloses a method for measuring the key size of a 3D NAND memory laminated structure based on deep learning, which constructs a deep-learning series network framework, establishes a forward network model of a corresponding relation between the key size of different laminated structures of a 3D NAND memory and a large amount of data of spectrum data based on the key size of the different laminated structures of the 3D NAND memory included in a preset data set, and further establishes a series network model based on the forward network model and a reverse network model, wherein the reverse network model is in front of the forward network model, the two models are trained and tested through a large amount of data after the forward network model, so that the models meet enough precision requirements, namely the key size of the laminated structure of any 3D NAND memory can be predicted, namely, the actual spectrum data obtained by measuring the laminated structure of a target 3D NAND memory is input into the series network model, wherein the key size output by the reverse network model is the relation of the laminated structure of the target 3D NAND memory, and the key size output by the reverse network model is the relation of the laminated structure of the target 3D NAND memory The key size improves the measurement efficiency, and meanwhile, the prediction accuracy of the critical dimension of the laminated structure of the 3D NAND memory is ensured by reconstructing the critical dimension of the 3D NAND memory.
Drawings
FIG. 1 is a flowchart of a method for deep learning based critical dimension measurement of a 3D NAND memory stack structure in an embodiment of the invention.
FIG. 2 is a schematic diagram of a stacked structure of a 3D NAND memory according to an embodiment of the present invention.
Fig. 3 is a functional diagram of a tandem network model according to an embodiment of the present invention.
Detailed Description
The invention will be further described with reference to examples of embodiments shown in the drawings.
As shown in fig. 1, the present embodiment provides a method for measuring a critical dimension of a stacked structure of a 3D NAND memory based on deep learning, which is used to measure a critical dimension of a stacked structure of a 3D NAND memory as shown in fig. 2. The measuring method comprises the following steps:
step S10: based on the key dimensions of different laminated structures of the 3D NAND memory and a preset data set of spectrum data, a matched forward network model and a series network model comprising the forward network model and a reverse network model are established. In the forward network model, the key size P is used as the network input, and the network output is the spectrum data O; in the series network model, the spectrum data R is used as the network input of the reverse network model, and the key dimension P' output by the reverse network model is used as the input of the forward network model, so that the series network model outputs the spectrum data O.
In this embodiment, the use of deep learning allows the predicted spectral data, i.e., the forward network, to be output with the critical dimensions of the stacked structure as input. Alternatively, the spectral data may be used as input and the predicted critical dimension of the stacked structure may be used as output, i.e., the inverse network.
In step S10, the forward network model and the tandem network model are created in this order.
As an initial step S11 of building the model, first, a first data set and a second data set are constructed based on a preset data set, the first data set and the second data set respectively include a plurality of data sets, each data set includes spectral data R and corresponding critical dimensions R of different stacked structures of the 3D NAND memory. Here, the critical dimension P may be at least any of the following parameters: the film thickness, the step length uniformity and the hole depth of each layer of the 3D NAND memory laminated structure. The spectral data R comprises n discrete points, represented by the array R ═ R 1 ，r 2 ，…，r n ]The number of the laminated structures is m, and the array P is P 1 ，p 2 ，…，p m ]Is shown byIn (c) p i Is the critical dimension of the ith.
In a specific embodiment, the preset data set may be a simulation data set obtained by simulating the critical dimension and the spectrum data of different stacked structures of the 3D NAND memory through a simulation device, and the model is trained and tested based on the simulation data set, so that the prediction accuracy of the model can be maximally improved. Of course, the solution of this embodiment does not limit to only using the simulation data set, and a model meeting the precision requirement can be established through the historical data or the measured data with higher accuracy.
Further, the establishing process of the forward network model comprises the following steps:
step S12: and constructing a preliminary forward network model. In the model, the key sizes P of different laminated structures are used as network input, and the network output is spectral data O ═ O 1 ，o 2 ，…，o n ]。
Step S13: and dividing the first data set into a training set and a testing set, training and testing the preliminary forward network model, training the forward network model through the training set, and testing the forward network model through the testing set, so as to obtain the forward network model meeting the preset conditions.
In step S13, during training, the key size P of each set of data in the corresponding training set is used as the input of the network, and the network output spectrum data O ═ O 1 ，o 2 ，…，o n ]The corresponding label is the actual spectral data R ═ R 1 ，r 2 ，…，r n ]And defining the loss function as:
and obtaining the mapping relation between the critical dimension P and the spectral data R through forward network model learning.
During testing, the key size P of each group of data in the corresponding test set is used as the input of the network, the spectrum data R is used as the output of the network, the precision of the forward network model is obtained through calculation, if the precision meets the preset precision requirement, the network is regarded as qualified, the weight parameters of the forward network model are fixed, and the trained forward network model is obtained.
Accordingly, the process of establishing the tandem network model comprises the following steps:
step S14: constructing a preliminary series network model; as shown in fig. 3, in the tandem network model, the forward network model and the reverse network model are connected in series, specifically: the reverse network model is before and the forward network model is after. In the model, the spectral data R is used as the input of the reverse network model, and the critical dimension P' ═ P of the reverse network model output is obtained 1 ′，p 2 ′，…，p m ′]And taking P' as the input of the forward network model, and enabling the series network model to output the spectrum data O ═ O 1 ，o 2 ，…，o n ]Therefore, the reconstruction of the key size of the 3D NAND memory is realized, and the prediction precision of the series network model is improved.
Step S15: and dividing the second data set into a training set and a testing set, training and testing the preliminary tandem network model, training the tandem network model through the training set, and testing the tandem network model through the testing set, so as to obtain the tandem network model meeting the preset condition.
In this step S15, during training, the predicted critical dimension P' ═ P of the layered structure output from the inverse network model is obtained using the spectral data R in the corresponding training set as the input of the tandem network model, that is, as the input of the inverse network model 1 ′，p 2 ′，…，p m ′]And finally obtaining output spectrum data O ═ O of the series network model by taking P' as the input of the forward network 1 1 ，o 2 ，…，o n ]. Wherein the loss function is defined as
The tandem network model is trained by minimizing the loss function.
During testing, the spectral data R in the corresponding test set is used as the input of the series network model, the predicted spectral data O is finally output, the precision of the series network model is obtained through calculation, if the precision meets the preset precision requirement, the network model is regarded as qualified, and the trained series network model is obtained.
After the establishment of the forward network model and the tandem network model is completed at step S10, step S20 is performed.
Step S20: measuring the laminated structure of the target 3D NAND memory to obtain actual spectrum data R [ R ] corresponding to the laminated structure of the target 3D NAND memory 1 ，r 2 ，…，r n ]。
As a specific form of measurement of the target 3D NAND memory, a specific measurement process may be performed by a scatterometer, OCD, or the like.
Step S30: the measured actual spectral data R ═ R 1 ，r 2 ，…，r n ]Inputting the critical dimension P ═ P output by the reverse network model into the series network model 1 ′，p 2 ′，…，p m ′]As a critical dimension of the target stacked structure of the 3D NAND memory.
In this step, after the actual spectral data is input into the tandem network model, the actual spectral data is first used as the input of the reverse network model, so that the reverse network model outputs the corresponding key size, and the key size is used as the input of the forward network model again, so that the forward network model outputs the spectral data, which is the output of the tandem network model. In this process, the prediction accuracy of the critical dimension of the stacked structure of the 3D NAND memory is ensured by reconstructing the critical dimension of the 3D NAND memory.
Two sets of similar spectral data may exist for two 3D NAND memories with different critical dimensions of existing stacked structures, which makes training of a neural network difficult to converge. If the inverse network is used directly, the spectral data R ═ R is used 1 ，r 2 ，…，r n ]For input, predict each layer critical dimension P ═ P 1 ′，p 2 ′，…，p m ′]The corresponding label is P ═ P 1 ，p 2 ，…，p m ]Then the corresponding loss function is
Because two groups of similar spectral data may correspond to two groups of different key size labels, the loss function is difficult to converge, and the network precision is greatly limited. According to the scheme of the embodiment, the reverse network model is connected with the forward network model in series, the network is converged smoothly through the conversion loss function, and parameters can be predicted accurately.
The method for measuring the key size of the laminated structure of the 3D NAND memory based on deep learning constructs a deep-learning series network frame, establishes a forward network model of the corresponding relationship between the key size of different laminated structures of the 3D NAND memory and a large amount of data of spectrum data based on the key size of different laminated structures of the 3D NAND memory included in a preset data set, further establishes a series network model based on the forward network model and a reverse network model, trains and tests the two models through a large amount of data after the reverse network model is in front of the forward network model, enables the models to meet enough precision requirements, can predict the key size of the laminated structure of any 3D NAND memory, namely, inputs actual spectrum data obtained by measuring the laminated structure of a target 3D NAND memory into the series network model, the critical dimension output by the reverse network model is the critical dimension of the laminated structure of the target 3D NAND memory, so that the measurement efficiency is improved, and meanwhile, the prediction precision of the critical dimension of the laminated structure of the 3D NAND memory is ensured by reconstructing the critical dimension of the 3D NAND memory.
The protective scope of the present invention is not limited to the above-described embodiments, and it is apparent that various modifications and variations can be made to the present invention by those skilled in the art without departing from the scope and spirit of the present invention. It is intended that the present invention cover the modifications and variations of this invention provided they come within the scope of the appended claims and their equivalents.
Claims (6)
1. A3D NAND memory laminated structure critical dimension measurement method based on deep learning is characterized by comprising the following steps:
establishing a matched forward network model and a series network model comprising the forward network model and a reverse network model based on the key sizes of different laminated structures of the 3D NAND memory and a preset data set of spectrum data; in the forward network model, the key size P is used as network input, and the network output is spectral data O; in the series network model, spectral data R is used as the network input of the reverse network model, and then the key size P' output by the reverse network model is used as the input of the forward network model, so that the series network model outputs spectral data O;
measuring the laminated structure of the target 3D NAND memory to obtain actual spectrum data;
and inputting the measured actual spectrum data into the series network model, and taking the critical dimension output by the reverse network model as the critical dimension of the target 3D NAND memory laminated structure.
2. The deep learning based 3D NAND memory stack structure critical dimension measurement method of claim 1, wherein:
the establishing process of the forward network model comprises the following steps:
constructing a first data set based on the preset data set, wherein the first data set comprises a plurality of data groups, and each data group comprises spectral data R and corresponding critical dimension P of different laminated structures of the 3D NAND memory;
constructing a primary forward network model;
and dividing the first data set into a training set and a testing set, and training and testing the preliminary forward network model so as to obtain the forward network model meeting preset conditions.
3. The deep learning based 3D NAND memory stack structure critical dimension measurement method of claim 2, wherein:
the establishing process of the series network model comprises the following steps:
constructing a second data set based on the preset data set, wherein the second data set comprises a plurality of data groups, and each data group comprises spectral data R and corresponding critical dimension P of different laminated structures of the 3D NAND memory;
constructing a preliminary series network model;
and dividing the second data set into a training set and a testing set, and training and testing the preliminary tandem network model so as to obtain the tandem network model meeting preset conditions.
4. The deep learning based 3D NAND memory stack structure critical dimension measurement method according to claim 3, wherein:
the critical dimension P may be at least any of the following parameters: film thickness, step length uniformity, and hole depth of each layer of the 3D NAND memory stack structure,
the spectral data R comprises n discrete points, and the array R ═ R 1 ，r 2 ，…，r n ]It is shown that there are m critical dimensions of the stacked structure, which are represented by the array P ═ P 1 ，p 2 ，…，p m ]Is represented by the formula, wherein p i Is the critical dimension of the ith die,
in the forward network model, the key sizes P of different laminated structures are used as network input, and the network output is spectral data O ═ O 1 ，o 2 ，…，o n ]，
In the preliminary model of the series network, taking spectral data R as the input of the reverse network model to obtain the critical dimension P' ═ P output by the reverse network model 1 ′，p 2 ′，…，p m ′]Taking P' as the input of the forward network model, and making the output spectrum data O of the series network model equal to [ O ] 1 ，o 2 ，…，o n ]。
5. The deep learning based 3D NAND memory stack structure critical dimension measurement method of claim 4, wherein:
in the positiveIn the process of establishing the network model and the series network model, respectively defining the loss functions of the corresponding models as
And training and testing the corresponding network model by minimizing the loss function, and obtaining the forward network model and the series network model when the spectral data output by the series network model meets the preset precision requirement.
6. The deep learning based 3D NAND memory stack structure critical dimension measurement method according to any one of claims 1 to 5, wherein:
the preset data set is a simulation data set obtained by simulating the critical dimension and the spectral data of different laminated structures of the 3D NAND memory through simulation equipment.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210487425.4A CN114963979A (en) | 2022-05-06 | 2022-05-06 | 3D NAND memory laminated structure critical dimension measuring method based on deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210487425.4A CN114963979A (en) | 2022-05-06 | 2022-05-06 | 3D NAND memory laminated structure critical dimension measuring method based on deep learning |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114963979A true CN114963979A (en) | 2022-08-30 |
Family
ID=82981469
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210487425.4A Pending CN114963979A (en) | 2022-05-06 | 2022-05-06 | 3D NAND memory laminated structure critical dimension measuring method based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114963979A (en) |
Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100094790A1 (en) * | 2008-10-10 | 2010-04-15 | Micron Technology, Inc. | Machine learning of dimensions using spectral intensity response of a reflectometer |
| US20120226644A1 (en) * | 2011-03-04 | 2012-09-06 | Wen Jin | Accurate and Fast Neural network Training for Library-Based Critical Dimension (CD) Metrology |
| CN107816949A (en) * | 2017-11-01 | 2018-03-20 | 长江存储科技有限责任公司 | A kind of accumulation layer measured film thickness method for 3D nand memories |
| US20180232630A1 (en) * | 2017-02-13 | 2018-08-16 | Kla-Tencor Corporation | Automated Accuracy-Oriented Model Optimization System for Critical Dimension Metrology |
| US20190148246A1 (en) * | 2017-11-15 | 2019-05-16 | Kla-Tencor Corporation | Automatic optimization of measurement accuracy through advanced machine learning techniques |
| CN110188780A (en) * | 2019-06-03 | 2019-08-30 | 电子科技大学中山学院 | Method and device for constructing deep learning model for positioning multi-target feature points |
| US20200003682A1 (en) * | 2018-07-02 | 2020-01-02 | The Research Foundation For The State University Of New York | System and method for structural characterization of materials by supervised machine learning-based analysis of their spectra |
| CN112241937A (en) * | 2020-07-22 | 2021-01-19 | 西安电子科技大学 | Hyperspectral image reconstruction method based on neural network |
| CN112469962A (en) * | 2018-11-02 | 2021-03-09 | 应用材料以色列公司 | Methods, systems, and computer program products for CDSEM metrology of 3D-NAND |
| CN112926157A (en) * | 2021-03-10 | 2021-06-08 | 中国计量大学 | Grating optical filter structure optimization method based on neural network |
| CN113111958A (en) * | 2021-04-23 | 2021-07-13 | 中南大学 | Spectrum model transfer method based on CNN-SVR model and transfer learning |
| CN113793656A (en) * | 2021-11-10 | 2021-12-14 | 山东大学 | Forward prediction and reverse design method and system for micro-nano particle system structural color |
-
2022
- 2022-05-06 CN CN202210487425.4A patent/CN114963979A/en active Pending
Patent Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100094790A1 (en) * | 2008-10-10 | 2010-04-15 | Micron Technology, Inc. | Machine learning of dimensions using spectral intensity response of a reflectometer |
| US20120226644A1 (en) * | 2011-03-04 | 2012-09-06 | Wen Jin | Accurate and Fast Neural network Training for Library-Based Critical Dimension (CD) Metrology |
| US20180232630A1 (en) * | 2017-02-13 | 2018-08-16 | Kla-Tencor Corporation | Automated Accuracy-Oriented Model Optimization System for Critical Dimension Metrology |
| CN107816949A (en) * | 2017-11-01 | 2018-03-20 | 长江存储科技有限责任公司 | A kind of accumulation layer measured film thickness method for 3D nand memories |
| US20190148246A1 (en) * | 2017-11-15 | 2019-05-16 | Kla-Tencor Corporation | Automatic optimization of measurement accuracy through advanced machine learning techniques |
| US20200003682A1 (en) * | 2018-07-02 | 2020-01-02 | The Research Foundation For The State University Of New York | System and method for structural characterization of materials by supervised machine learning-based analysis of their spectra |
| CN112469962A (en) * | 2018-11-02 | 2021-03-09 | 应用材料以色列公司 | Methods, systems, and computer program products for CDSEM metrology of 3D-NAND |
| CN110188780A (en) * | 2019-06-03 | 2019-08-30 | 电子科技大学中山学院 | Method and device for constructing deep learning model for positioning multi-target feature points |
| CN112241937A (en) * | 2020-07-22 | 2021-01-19 | 西安电子科技大学 | Hyperspectral image reconstruction method based on neural network |
| CN112926157A (en) * | 2021-03-10 | 2021-06-08 | 中国计量大学 | Grating optical filter structure optimization method based on neural network |
| CN113111958A (en) * | 2021-04-23 | 2021-07-13 | 中南大学 | Spectrum model transfer method based on CNN-SVR model and transfer learning |
| CN113793656A (en) * | 2021-11-10 | 2021-12-14 | 山东大学 | Forward prediction and reverse design method and system for micro-nano particle system structural color |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109491216B (en) | Method for optimizing photoetching process parameters | |
| CN108963356B (en) | Device and method for acquiring internal state distribution of lithium ion battery | |
| CN102680812A (en) | Method and device for evaluating life of product | |
| CN113391211B (en) | Method for predicting remaining life of lithium battery under small sample condition | |
| CN111336935A (en) | Nano-film parameter inversion calculation method based on improved hybrid optimization algorithm | |
| CN110348665A (en) | A kind of low-voltage platform area electric power system data quality evaluating method and device | |
| CN111276414A (en) | Detection method and device | |
| US20240005061A1 (en) | Method and device for predicting insulation aging life of high-voltage submarine cable | |
| CN114199945B (en) | Geopolymer mortar strength and chloride ion permeability resistance prediction method | |
| CN114963979A (en) | 3D NAND memory laminated structure critical dimension measuring method based on deep learning | |
| CN114596302A (en) | PCB defect detection method, system, medium, equipment and terminal | |
| CN105260521B (en) | A kind of load Fusion Modeling Method based on analytic hierarchy process (AHP) | |
| CN107341336A (en) | A kind of tank product geometric accuracy method for evaluating consistency | |
| KR102567843B1 (en) | System and method for measuring the thickness of a multilayer thin film structure | |
| CN115101149B (en) | Method for predicting total energy of microstructure of material | |
| CN103399974B (en) | Quantize the method comparing random vibration emulated data and experimental data | |
| CN105354737A (en) | Computing method suitable for big data value evaluation | |
| CN113466700B (en) | Lithium battery online life prediction method based on two-point impedance life characteristics | |
| CN105699043B (en) | A kind of wind tunnel sensors that improve measure stability and the method for precision | |
| CN114966409A (en) | Power lithium battery state of charge estimation method based on multi-layer perceptron algorithm | |
| CN114970601A (en) | Power equipment partial discharge type identification method, equipment and storage medium | |
| CN115965609B (en) | Intelligent detection method for flaws of ceramic substrate by utilizing knowledge distillation | |
| Chen et al. | Data-Driven Stress/Warpage Analyses Based on Stoney Equation for Packaging Applications | |
| Yun et al. | [Retracted] Quality Evaluation and Satisfaction Analysis of Online Learning of College Students Based on Artificial Intelligence | |
| CN110349127A (en) | The prediction technique and device of the equivalent elastic modulus of particle layer |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20220830 |
|
| RJ01 | Rejection of invention patent application after publication |