CN118261114B - Thin film data system design method, storage medium and terminal - Google Patents
Thin film data system design method, storage medium and terminal Download PDFInfo
- Publication number
- CN118261114B CN118261114B CN202410683124.8A CN202410683124A CN118261114B CN 118261114 B CN118261114 B CN 118261114B CN 202410683124 A CN202410683124 A CN 202410683124A CN 118261114 B CN118261114 B CN 118261114B
- Authority
- CN
- China
- Prior art keywords
- film
- machine learning
- learning model
- precursor
- model
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 65
- 239000010409 thin film Substances 0.000 title claims abstract description 35
- 238000013461 design Methods 0.000 title claims abstract description 27
- 238000003860 storage Methods 0.000 title claims abstract description 9
- 239000010408 film Substances 0.000 claims abstract description 168
- 238000010801 machine learning Methods 0.000 claims abstract description 76
- 238000001514 detection method Methods 0.000 claims abstract description 52
- 238000012549 training Methods 0.000 claims abstract description 32
- 238000005259 measurement Methods 0.000 claims abstract description 20
- 239000002243 precursor Substances 0.000 claims description 65
- 239000000758 substrate Substances 0.000 claims description 23
- 238000000231 atomic layer deposition Methods 0.000 claims description 19
- 238000006243 chemical reaction Methods 0.000 claims description 9
- 238000005530 etching Methods 0.000 claims description 8
- 238000002474 experimental method Methods 0.000 claims description 5
- 239000000203 mixture Substances 0.000 claims description 5
- 239000000463 material Substances 0.000 claims description 4
- 238000003062 neural network model Methods 0.000 claims description 3
- 238000007637 random forest analysis Methods 0.000 claims description 3
- 238000000427 thin-film deposition Methods 0.000 claims description 3
- 238000001459 lithography Methods 0.000 claims description 2
- 238000012360 testing method Methods 0.000 claims 2
- 238000004519 manufacturing process Methods 0.000 abstract description 7
- 238000004088 simulation Methods 0.000 description 5
- 238000005516 engineering process Methods 0.000 description 3
- 238000010926 purge Methods 0.000 description 3
- OAICVXFJPJFONN-UHFFFAOYSA-N Phosphorus Chemical compound [P] OAICVXFJPJFONN-UHFFFAOYSA-N 0.000 description 2
- 230000008033 biological extinction Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000009396 hybridization Methods 0.000 description 2
- 229910052698 phosphorus Inorganic materials 0.000 description 2
- 239000011574 phosphorus Substances 0.000 description 2
- 238000000206 photolithography Methods 0.000 description 2
- 229910021420 polycrystalline silicon Inorganic materials 0.000 description 2
- 229920005591 polysilicon Polymers 0.000 description 2
- 239000004065 semiconductor Substances 0.000 description 2
- 230000000903 blocking effect Effects 0.000 description 1
- 238000011960 computer-aided design Methods 0.000 description 1
- 239000013078 crystal Substances 0.000 description 1
- 238000000151 deposition Methods 0.000 description 1
- 230000008021 deposition Effects 0.000 description 1
- 239000007789 gas Substances 0.000 description 1
- 239000011261 inert gas Substances 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 229920002120 photoresistant polymer Polymers 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 229910052710 silicon Inorganic materials 0.000 description 1
- 239000010703 silicon Substances 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/30—Circuit design
- G06F30/39—Circuit design at the physical level
- G06F30/398—Design verification or optimisation, e.g. using design rule check [DRC], layout versus schematics [LVS] or finite element methods [FEM]
-
- C—CHEMISTRY; METALLURGY
- C23—COATING METALLIC MATERIAL; COATING MATERIAL WITH METALLIC MATERIAL; CHEMICAL SURFACE TREATMENT; DIFFUSION TREATMENT OF METALLIC MATERIAL; COATING BY VACUUM EVAPORATION, BY SPUTTERING, BY ION IMPLANTATION OR BY CHEMICAL VAPOUR DEPOSITION, IN GENERAL; INHIBITING CORROSION OF METALLIC MATERIAL OR INCRUSTATION IN GENERAL
- C23C—COATING METALLIC MATERIAL; COATING MATERIAL WITH METALLIC MATERIAL; SURFACE TREATMENT OF METALLIC MATERIAL BY DIFFUSION INTO THE SURFACE, BY CHEMICAL CONVERSION OR SUBSTITUTION; COATING BY VACUUM EVAPORATION, BY SPUTTERING, BY ION IMPLANTATION OR BY CHEMICAL VAPOUR DEPOSITION, IN GENERAL
- C23C16/00—Chemical coating by decomposition of gaseous compounds, without leaving reaction products of surface material in the coating, i.e. chemical vapour deposition [CVD] processes
- C23C16/44—Chemical coating by decomposition of gaseous compounds, without leaving reaction products of surface material in the coating, i.e. chemical vapour deposition [CVD] processes characterised by the method of coating
- C23C16/455—Chemical coating by decomposition of gaseous compounds, without leaving reaction products of surface material in the coating, i.e. chemical vapour deposition [CVD] processes characterised by the method of coating characterised by the method used for introducing gases into reaction chamber or for modifying gas flows in reaction chamber
- C23C16/45523—Pulsed gas flow or change of composition over time
- C23C16/45525—Atomic layer deposition [ALD]
-
- C—CHEMISTRY; METALLURGY
- C23—COATING METALLIC MATERIAL; COATING MATERIAL WITH METALLIC MATERIAL; CHEMICAL SURFACE TREATMENT; DIFFUSION TREATMENT OF METALLIC MATERIAL; COATING BY VACUUM EVAPORATION, BY SPUTTERING, BY ION IMPLANTATION OR BY CHEMICAL VAPOUR DEPOSITION, IN GENERAL; INHIBITING CORROSION OF METALLIC MATERIAL OR INCRUSTATION IN GENERAL
- C23C—COATING METALLIC MATERIAL; COATING MATERIAL WITH METALLIC MATERIAL; SURFACE TREATMENT OF METALLIC MATERIAL BY DIFFUSION INTO THE SURFACE, BY CHEMICAL CONVERSION OR SUBSTITUTION; COATING BY VACUUM EVAPORATION, BY SPUTTERING, BY ION IMPLANTATION OR BY CHEMICAL VAPOUR DEPOSITION, IN GENERAL
- C23C16/00—Chemical coating by decomposition of gaseous compounds, without leaving reaction products of surface material in the coating, i.e. chemical vapour deposition [CVD] processes
- C23C16/44—Chemical coating by decomposition of gaseous compounds, without leaving reaction products of surface material in the coating, i.e. chemical vapour deposition [CVD] processes characterised by the method of coating
- C23C16/52—Controlling or regulating the coating process
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
- G06N20/20—Ensemble learning
-
- 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
- Engineering & Computer Science (AREA)
- Chemical & Material Sciences (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Organic Chemistry (AREA)
- General Chemical & Material Sciences (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- Computer Hardware Design (AREA)
- Metallurgy (AREA)
- Mechanical Engineering (AREA)
- Computing Systems (AREA)
- Materials Engineering (AREA)
- Chemical Kinetics & Catalysis (AREA)
- Mathematical Physics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Medical Informatics (AREA)
- Geometry (AREA)
- Chemical Vapour Deposition (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides a design method of a thin film data system, a storage medium and a terminal, comprising the following steps: providing a plurality of film groups with different application scenes, wherein each film group is provided with a plurality of historical films, and each historical film is provided with a corresponding measurement parameter value; providing a machine learning model; inputting the characteristic data and corresponding measurement parameter values associated with the historical film into a machine learning model for model training; after model training, inputting the characteristic data associated with the historical film into a trained machine learning model for model detection, and outputting a prediction parameter value based on the machine learning model; comparing the predicted parameter value with the corresponding measured parameter value until the detection accuracy reaches the preset accuracy, and completing the model detection of the machine learning model; providing design requirement data of a film to be formed; inputting design demand data into a machine learning model for completing model detection, and generating a film to be formed based on the machine learning model; intelligent manufacturing is realized.
Description
Technical Field
The present invention relates to the field of semiconductor manufacturing technologies, and in particular, to a method for designing a thin film data system, a storage medium, and a terminal.
Background
In semiconductor processing, various film growth is required to achieve some process objectives, such as a hard mask for blocking etching, an anti-reflection film for reducing reflection, a stress layer for improving stress, a low dielectric constant dielectric layer for reducing capacitance, and the like.
The doping or hybridization technology has significant influence on the structures such as crystal lattice, energy band and the like of the film, and can cause the changes of parameters such as etching rate, refractive index, extinction coefficient, stress, dielectric constant, resistance and the like.
Atomic layer deposition (Atomic Layer Deposition, ALD) is a process that utilizes a self-limiting reaction that utilizes limited reactive sites on the growth substrate to allow the growth rate to become controllable from the atomic level, controlling the growth thickness by adjusting the number of cycles of growth.
However, there is a lack of a thin film data system that provides thin film selection guidance to process engineers, and the system is enriched in its database by some technique for controllably growing doped or hybrid thin films.
Disclosure of Invention
The invention solves the technical problem of providing a design method, a storage medium and a terminal of a film data system, which can provide films with required properties for process engineers in different directions to fulfill the process aim and realize intelligent manufacturing.
In order to solve the above problems, the present invention provides a method for designing a thin film data system, including: providing a plurality of film groups with different application scenes, wherein each film group is provided with a plurality of historical films, and each historical film is provided with a corresponding measurement parameter value; providing a machine learning model; inputting the characteristic data associated with the historical film and the corresponding measurement parameter values into the machine learning model for model training; after model training, inputting the characteristic data associated with the historical film into the trained machine learning model for model detection, and outputting a prediction parameter value based on the machine learning model; comparing the predicted parameter value with the corresponding measured parameter value to obtain detection accuracy rate until the detection accuracy rate reaches a preset accuracy rate, and completing model detection of the machine learning model; providing design requirement data of a film to be formed; inputting the design requirement data into the machine learning model for completing model detection, and generating the film to be formed based on the machine learning model.
Optionally, the history film and the film to be formed are formed by a multi-source atomic layer deposition process.
Optionally, the steps of the multi-source atomic layer deposition process include: pulsing a first type of precursor to be exposed to a surface of a substrate while chemisorbing the first type of precursor to the surface of the substrate; pulsing a second type of precursor to perform a chemical reaction on the surface of the substrate; repeating the pulses of the first type of precursor and the second type of precursor until the historical film or the film to be formed is of a required thickness.
Optionally, the first type of precursor includes a plurality of precursors, and the second type of precursor includes one precursor.
Optionally, the steps of the multi-source atomic layer deposition process include: pulsing a first precursor to be exposed to a surface of a substrate while chemisorbing the first precursor to the surface of the substrate; pulsing a second precursor to form a chemical reaction on the surface of the substrate; repeating pulsing the first precursor and the second precursor to a desired thickness of the first film; pulsing a third precursor to expose a surface of the first film while the first film chemisorbs the third precursor; pulsing a fourth precursor to form a chemical reaction on the surface of the first film, and repeatedly pulsing the third precursor and the fourth precursor to form a second film on the surface of the first film; and sequentially stacking the first film and the second film until the history film or the film to be formed with the required thickness is formed.
Optionally, the different application scenarios include one or more of a lithography application scenario, a thin film deposition application scenario, and an etching application scenario.
Optionally, the characteristic parameters associated with the historical film include: the method comprises the steps of working branches, design rules, film action, films, corresponding information and performance in application scenes, wherein the corresponding information comprises doping amounts of doping elements in the films and material compositions of the films.
Optionally, the machine learning model includes: neural network models or random forest models.
Optionally, the design requirement data of the film to be formed includes: the method comprises the steps of application scenes, film action, films and corresponding information, wherein the corresponding information comprises doping amounts of doping elements in the films and material compositions of the films.
Optionally, the plurality of history films in each film group are divided into a plurality of training history films and a plurality of detection history films.
Optionally, feature data associated with a plurality of training history films in each film group and corresponding measurement parameter values are input into the machine learning model for model training.
Optionally, feature data associated with a plurality of detection history films in each film group is input into the trained machine learning model for model detection.
Optionally, the method for obtaining the detection accuracy includes comparing the predicted parameter value with a corresponding measured parameter value: obtaining the matching degree between the predicted parameter value and the corresponding measured parameter value; when the matching degree reaches a preset matching degree, judging that the predicted parameter value passes the matching; and taking the ratio of the number of the preset parameter values passing through the matching to the total number of the preset parameter values as the detection accuracy.
Optionally, the method until the detection accuracy reaches the preset accuracy comprises the following steps: and when the detection accuracy of the machine learning model does not reach the preset accuracy, increasing the model training times of the machine learning model until the detection accuracy of the machine learning model reaches the preset accuracy.
Correspondingly, the invention also provides a storage medium, on which computer instructions are stored, which when run perform the steps of the above method.
Correspondingly, the invention further provides a terminal which comprises a memory and a processor, wherein the memory stores computer instructions capable of being operated on the processor, and the processor executes the steps of the method when the processor operates the computer instructions.
Compared with the prior art, the technical scheme of the invention has the following advantages:
According to the design method of the film data system, the machine learning model is trained and detected through the characteristic data related to the historical film and the corresponding measurement parameter values, after the machine learning model passes the detection, the film to be formed can be rapidly provided for the design requirement based on the machine learning model, so that manpower is effectively reduced, and intelligent manufacturing is realized.
Furthermore, the historical film and the film to be formed are formed by adopting a multi-source atomic layer deposition process, film doping or hybridization is realized based on the multi-source atomic layer deposition process, and a film data system is greatly enriched.
Drawings
FIG. 1 is a flow chart of a method of designing a thin film data system in accordance with an embodiment of the present invention.
FIG. 2 is a schematic diagram illustrating steps of a multi-source ALD process according to one embodiment of the present invention;
FIG. 3 is a schematic diagram illustrating steps of a multi-source ALD process according to another embodiment of the present invention.
Detailed Description
As the background art shows, there is a lack of a thin film data system that provides thin film selection guidance to process engineers, while enriching its database by some technique of controllably growing doped or hybrid thin films.
On the basis, the invention provides a design method, a storage medium and a terminal of a film data system, wherein a machine learning model is trained and detected through characteristic data related to a historical film and corresponding measurement parameter values, and after the machine learning model passes the detection, a film to be formed can be rapidly provided for design requirements based on the machine learning model, so that manpower is effectively reduced, and intelligent manufacturing is realized.
In order that the above-recited objects, features and advantages of the present invention will become more readily apparent, a more particular description of the invention briefly described above will be rendered by reference to specific embodiments thereof which are illustrated in the appended drawings.
FIG. 1 is a flow chart of a method of designing a thin film data system according to an embodiment of the present invention, including:
Step S101, providing a plurality of film groups with different application scenes, wherein each film group is provided with a plurality of historical films, and each historical film is provided with a corresponding measurement parameter value;
Step S102, providing a machine learning model;
Step S103, inputting the characteristic data related to the historical film and the corresponding measurement parameter values into the machine learning model for model training;
step S104, after model training, inputting the characteristic data associated with the historical film into the trained machine learning model for model detection, and outputting a prediction parameter value based on the machine learning model;
Step S105, comparing the predicted parameter value with the corresponding measured parameter value to obtain detection accuracy rate until the detection accuracy rate reaches a preset accuracy rate, and completing model detection of the machine learning model;
Step S106, providing design requirement data of the film to be formed;
step S107, inputting the design requirement data into the machine learning model for completing model detection, and generating the film to be formed based on the machine learning model.
The steps of the design method of the thin film data system will be described in detail below.
In step S101, a plurality of film sets with different application scenarios are provided, each film set having a plurality of history films, each history film having a corresponding measurement parameter value.
In this embodiment, the history film is a doped or hybridized film formed on the substrate by a multi-source atomic layer deposition process under a certain application scenario, and the measurement of the properties of the history film is completed through experiments, so as to obtain the measurement parameter values corresponding to each history film. These data will provide support for the training and detection of subsequent machine learning models.
In this embodiment, the different application scenarios include one or more of a photolithography application scenario, a film deposition application scenario, and an etching application scenario, where the films corresponding to the different application scenarios are different, for example, the films required for the etching application scenario are different from the films required for the photolithography application scenario.
In this embodiment, the measured parameter value includes one or more of an etching rate, a refractive index, an extinction coefficient, a dielectric constant, and a resistance.
In this embodiment, since a large amount of data is required for support in both the training and detection phases of the machine learning model, the number of history films in each of the film groups cannot be only 1 either. And it is necessary to divide the plurality of the history films in each of the film groups into a plurality of training history films and a plurality of detection history films. Wherein a number of training history films are to be used for training phases of the machine learning model and a number of detection history films are to be used for detection phases of the machine learning model.
It should be noted that the ratio of the number of the training history films to the number of the detection history films in each of the film groups is not fixed, and is usually 1:1. However, in the case of more various application scenes, the proportion of the training history film needs to be increased.
In step S102, a machine learning model is provided.
In this embodiment, the machine learning model includes: neural network models or random forest models.
In step S103, the feature data associated with the history film and the corresponding measurement parameter values are input into the machine learning model for model training.
In this embodiment, feature data associated with a plurality of training history films in each of the film sets and corresponding measurement parameter values are input into the machine learning model for model training.
In this embodiment, in the model training stage, the corresponding measurement parameter values are input into the machine learning model at the same time by taking the feature data associated with the history film as a condition, and the mapping relationship is automatically learned and established inside the machine learning model.
In this embodiment, the characteristic parameters associated with the history film include: application scene, working branches, design rules, film action, films, corresponding information and expression in the application scene, wherein the corresponding information comprises doping amount of doping elements.
Taking etching polysilicon as an example, a film structure of 'silicon substrate + gate oxide + polysilicon + hard mask + photoresist' is needed for a high probability, and then the associated characteristic parameters of the hard mask include working branches, design rules, film action, film and corresponding information, and expression in application scenarios.
It should be noted that, in the training stage of the machine learning model, the characteristic parameters associated with the historical films of different application scenarios are all indefinite, so that the machine model needs to be optimized by itself.
In step S104, after model training, the feature data associated with the history film is input into the trained machine learning model to perform model detection, and a prediction parameter value is output based on the machine learning model.
In this embodiment, feature data associated with a plurality of detection history films in each of the film groups is input into the trained machine learning model to perform model detection.
In this embodiment, a method for outputting a prediction parameter value based on a machine learning model includes: associating a machine learning model with the simulation tool; the predicted parameter values are output by the simulation tool.
In this embodiment, the simulation process may employ a TCAD (Technology Computer AIDED DESIGN) simulation tool.
After model training, for example, a film with a phosphorus doping level of 5% or 10% in a film data system, the film data system can instruct the multisource atomic layer deposition process to collect films with higher doping levels if the film data system finds that the phosphorus doping level is greater.
In step S105, the predicted parameter value is compared with the corresponding measured parameter value until the detection accuracy reaches a preset accuracy, so as to complete the model detection of the machine learning model.
In this embodiment, the method for comparing the predicted parameter value with the corresponding measured parameter value to obtain the detection accuracy includes: obtaining the matching degree between the predicted parameter value and the corresponding measured parameter value; when the matching degree reaches a preset matching degree, judging that the predicted parameter value passes the matching; and taking the ratio of the number of the preset parameter values passing through the matching to the total number of the preset parameter values as the detection accuracy.
In this embodiment, the method for obtaining the matching degree between the prediction parameter value and the corresponding measurement parameter value includes: root mean square error (Root Mean Squared Error, RMSE) or R-square index (R-squared, R2). The root mean square error or the R-square index is used for measuring the index of the difference between the predicted parameter value and the measured parameter value, and when the difference index of the two is in a preset range, the two indexes are considered to be matched.
In this embodiment, the method until the detection accuracy reaches the preset accuracy includes: and when the detection accuracy of the machine learning model does not reach the preset accuracy, increasing the model training times of the machine learning model until the detection accuracy of the machine learning model reaches the preset accuracy.
In step S106, design requirement data of a thin film to be formed is provided.
In this embodiment, the design requirement data of the film to be formed includes: application scene, film action, film and corresponding information, wherein the corresponding information comprises doping amount of doping elements. For example, taking a thin film deposition application scenario as an example, after the machine learning model is detected and put into practical application, only the doping amount of the doping element in the doping thin film needs to be input for the doping thin film.
In step S107, the design requirement data is input into the machine learning model that completes model detection, and the thin film to be formed is generated based on the machine learning model.
In this embodiment, the machine learning model is trained and detected by the feature data and the corresponding measurement parameter values associated with the historical thin film, and after the machine learning model passes the detection, the thin film to be formed can be rapidly provided for the design requirement based on the machine learning model, so that manpower is effectively reduced, and intelligent manufacturing is realized.
In this embodiment, the thin film data system provides thin film simulation data and composition information according to the needs of engineers, and the multi-source atomic layer deposition system performs production according to the information provided by the thin film data system.
In this embodiment, the method of generating the thin film to be formed based on the machine learning model includes: and checking the generated film to be formed through experiments by using a machine learning model, and verifying whether the information provided by the system is correct through the experiments.
In this embodiment, the history film and the film to be formed are formed by a multi-source atomic layer deposition process.
Referring to fig. 2, the steps of the multi-source atomic layer deposition process include: pulsing a first type precursor a/B/… to be exposed to a surface of a substrate 100 while chemisorbing the first type precursor to the surface of the substrate 100; pulsing a second type of precursor G to chemically react on the surface of the substrate 100; repeating (cycle) pulses of a first type of precursor and the second type of precursor to the desired thickness of the history film 101 or the film 102 to be formed.
In this embodiment, the first type of precursor includes a plurality of precursors, and the second type of precursor includes one precursor.
In this embodiment, purging the first type precursor is also included prior to exposing the first type precursor a/B/… to the surface of the substrate 100.
In this embodiment, before the second type precursor G is chemically reacted on the surface of the substrate 100, a gas purging step is further performed, specifically, inert gas is used to blow away the remaining unreacted first type precursor.
In another embodiment, please refer to fig. 3, wherein the steps of the multi-source atomic layer deposition process are as follows: pulsing a first precursor a to be exposed to a surface of a substrate 200 while the surface of the substrate 200 chemisorbs the first precursor a; pulsing the second precursor G to form a chemical reaction on the surface of the substrate 200; repeating (cycle) pulses of the first precursor a and the second precursor G to a first film of a desired thickness; pulsing a third precursor C to expose the surface of the first film while the first film chemisorbs the third precursor C; pulsing a fourth precursor G to form a chemical reaction on the surface of the first film; repeating (cycle) the third precursor C and the fourth precursor G to form a second film on the surface of the first film, and sequentially stacking (cycle) the first film and the second film to the desired thickness of the history film 201 or the film 202 to be formed.
In this embodiment, a purge step is further included before the first precursor a, the second precursor G, the third precursor C, and the fourth precursor G are pulsed.
Correspondingly, the invention also provides a storage medium, on which computer instructions are stored, which when run perform the steps of the method of any one of the embodiments described above.
Correspondingly, the embodiment of the invention also provides a terminal which comprises a memory and a processor, wherein the memory stores computer instructions capable of running on the processor, and the processor executes the steps of the method of any embodiment when running the computer instructions.
Although the present invention is disclosed above, the present invention is not limited thereto. Various changes and modifications may be made by one skilled in the art without departing from the spirit and scope of the invention, and the scope of the invention should be assessed accordingly to that of the appended claims.
Claims (14)
1. A method of designing a thin film data system, comprising:
Providing a plurality of film groups with different application scenes, wherein each film group is provided with a plurality of historical films, and each historical film is provided with a corresponding measurement parameter value;
providing a machine learning model;
Inputting the characteristic data associated with the historical film and the corresponding measurement parameter values into the machine learning model for model training;
After model training, inputting the characteristic data associated with the historical film into the trained machine learning model for model detection, and outputting a prediction parameter value based on the machine learning model;
comparing the predicted parameter value with the corresponding measured parameter value until the detection accuracy reaches a preset accuracy position, and completing model detection of the machine learning model;
Providing design requirement data of a film to be formed;
Inputting the design requirement data into the machine learning model for completing model detection, generating the film to be formed based on the machine learning model, and generating the film to be formed based on the machine learning model, wherein the method comprises the following steps: checking the generated film to be formed through experiments by using a machine learning model, and verifying whether the information provided by the system is correct through the experiments;
wherein, the characteristic parameters associated with the history film comprise: working branches, design rules, film action, corresponding information and expression in application scenes, wherein the corresponding information comprises doping amount of doping elements in the film and material composition of the film;
The design requirement data of the film to be formed comprises: the method comprises the steps of application scenes, film action and corresponding information, wherein the corresponding information comprises doping amount of doping elements in the film and material composition of the film.
2. The method of claim 1, wherein the history film and the film to be formed are formed using a multi-source atomic layer deposition process.
3. The method of designing a thin film data system of claim 2, wherein the step of the multi-source atomic layer deposition process comprises: pulsing a first type of precursor to be exposed to a surface of a substrate while chemisorbing the first type of precursor to the surface of the substrate; pulsing a second type of precursor to perform a chemical reaction on the surface of the substrate; repeating the pulses of the first type of precursor and the second type of precursor until the historical film or the film to be formed is of a required thickness.
4. The method of designing a thin film data system as claimed in claim 3, wherein the first type of precursor comprises a plurality of precursors and the second type of precursor comprises a single precursor.
5. The method of designing a thin film data system of claim 2, wherein the step of the multi-source atomic layer deposition process comprises: pulsing a first precursor to be exposed to a surface of a substrate while chemisorbing the first precursor to the surface of the substrate; pulsing a second precursor to form a chemical reaction on the surface of the substrate; repeating pulsing the first precursor and the second precursor to a desired thickness of the first film; pulsing a third precursor to expose a surface of the first film while the first film chemisorbs the third precursor; pulsing a fourth precursor to form a chemical reaction on the surface of the first film, and repeatedly pulsing the third precursor and the fourth precursor to form a second film on the surface of the first film; and sequentially stacking the first film and the second film until the history film or the film to be formed with the required thickness is formed.
6. The method of claim 1, wherein the different application scenarios include one or more of a lithography application scenario, a thin film deposition application scenario, and an etching application scenario.
7. The method of designing a thin film data system of claim 1, wherein the machine learning model comprises: neural network models or random forest models.
8. The method of designing a thin film data system as claimed in claim 1, wherein a plurality of the history thin films in each of the thin film groups are divided into a plurality of training history thin films and a plurality of detection history thin films.
9. The method of claim 8, wherein the feature data associated with the plurality of training history films in each of the film sets and corresponding measurement parameter values are input into the machine learning model for model training.
10. The method of claim 8, wherein the feature data associated with a plurality of said test history films in each of said film sets is input to said machine learning model after training for model testing.
11. The method of claim 1, wherein comparing the predicted parameter values with corresponding measured parameter values to obtain the detection accuracy comprises: obtaining the matching degree between the predicted parameter value and the corresponding measured parameter value; when the matching degree reaches a preset matching degree, judging that the predicted parameter value passes the matching; and taking the ratio of the number of the preset parameter values passing through the matching to the total number of the preset parameter values as the detection accuracy.
12. The method of designing a thin film data system as claimed in claim 11, wherein the method until the detection accuracy reaches a preset accuracy comprises: and when the detection accuracy of the machine learning model does not reach the preset accuracy, increasing the model training times of the machine learning model until the detection accuracy of the machine learning model reaches the preset accuracy.
13. A storage medium having stored thereon computer instructions which, when run, perform the steps of the method of any of claims 1 to 12.
14. A terminal comprising a memory and a processor, the memory having stored thereon computer instructions executable on the processor, wherein the processor, when executing the computer instructions, performs the steps of the method of any of claims 1 to 12.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410683124.8A CN118261114B (en) | 2024-05-29 | 2024-05-29 | Thin film data system design method, storage medium and terminal |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410683124.8A CN118261114B (en) | 2024-05-29 | 2024-05-29 | Thin film data system design method, storage medium and terminal |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN118261114A CN118261114A (en) | 2024-06-28 |
| CN118261114B true CN118261114B (en) | 2024-08-06 |
Family
ID=91613332
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410683124.8A Active CN118261114B (en) | 2024-05-29 | 2024-05-29 | Thin film data system design method, storage medium and terminal |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN118261114B (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113822495A (en) * | 2021-10-20 | 2021-12-21 | 平安科技（深圳）有限公司 | Growth prediction method, device, equipment and medium based on machine learning model |
| CN117436355A (en) * | 2023-12-21 | 2024-01-23 | 东莞市钜欣电子有限公司 | Method for establishing optical film thickness model and related equipment |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20220228265A1 (en) * | 2021-01-15 | 2022-07-21 | Taiwan Semiconductor Manufacturing Co., Ltd. | System and method for dynamically adjusting thin-film deposition parameters |
| CN117313640B (en) * | 2022-06-14 | 2024-09-20 | 腾讯科技（深圳）有限公司 | Training method, device, equipment and storage medium for lithography mask generation model |
-
2024
- 2024-05-29 CN CN202410683124.8A patent/CN118261114B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113822495A (en) * | 2021-10-20 | 2021-12-21 | 平安科技（深圳）有限公司 | Growth prediction method, device, equipment and medium based on machine learning model |
| CN117436355A (en) * | 2023-12-21 | 2024-01-23 | 东莞市钜欣电子有限公司 | Method for establishing optical film thickness model and related equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN118261114A (en) | 2024-06-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US6915177B2 (en) | Comprehensive integrated lithographic process control system based on product design and yield feedback system | |
| US11886783B2 (en) | Simulation system for semiconductor process and simulation method thereof | |
| KR102586405B1 (en) | Metrology and process control for semiconductor manufacturing | |
| US9158301B2 (en) | Semiconductor processing dispatch control | |
| US20110060552A1 (en) | Measuring Apparatus, Measuring Coordinate Setting Method and Measuring Coordinate Number Calculation Method | |
| CN102037550B (en) | Substrate matrix to decouple tool and process effects | |
| JP3897922B2 (en) | Semiconductor device manufacturing method and computer-readable recording medium | |
| US6834212B1 (en) | Method and apparatus for APC solver engine and heuristic | |
| US7962234B2 (en) | Multidimensional process window optimization in semiconductor manufacturing | |
| CN118261114B (en) | Thin film data system design method, storage medium and terminal | |
| EP1290727A2 (en) | Method and apparatus for controlling deposition parameters based on polysilicon grain size feedback | |
| CN106599336B (en) | Semiconductor process simulation device and method and computing device | |
| CN116053164A (en) | Method and system for controlling critical dimension | |
| US7565254B2 (en) | Method and apparatus for metrology sampling using combination sampling rules | |
| CN112685958B (en) | SiC MOSFET blocking voltage determination method based on neural network | |
| US7155301B2 (en) | Method and system for dynamic modeling and recipe optimization of semiconductor etch processes | |
| Han et al. | Optimization of neural network structure and learning parameters using genetic algorithms | |
| US20010041971A1 (en) | Method of generating mesh and storage medium storing mesh generating program | |
| CN101944560A (en) | The manufacture method of semiconductor laser substrate | |
| CN113471093A (en) | Method and device for predicting film morphology of semiconductor device | |
| KR20100048186A (en) | System of executing unified process-device-circuit simulation | |
| Leu et al. | Reducing Thickness Deviation of W-Shaped Structures in Manufacturing DRAM Products Using RSM and ANN_GA | |
| CN116149146B (en) | Process control method, system and computer equipment | |
| CN118260975B (en) | Engine combustion chamber numerical simulation method based on self-adaptive pre-zoning mechanism | |
| US20220406798A1 (en) | Neuromorphic ferroelectric field effect transistor (fefet) device with anti-ferroelectric buffer 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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |