Automated analysis system for chemical vapor deposition based on atomic-level precision manufacturing

Abstract

The present invention relates to the field of material analysis technology, and specifically to a chemical vapor deposition automated analysis system based on atomic-level precision manufacturing, including a process control module, a process optimization module, an optimization revision module, an instruction sending module, a quality monitoring module, a data storage module, and a user interface module, wherein: the process control module is used to set and adjust the process parameters in the chemical vapor deposition process according to the control instructions; the process optimization module is used to perform real-time analysis on the collected process data to generate a first process optimization suggestion; the optimization revision module analyzes the physical and chemical properties of the deposited material to obtain a revised second process optimization suggestion. The present invention uses a Gaussian process model to quantify uncertainty and efficiently search for the optimal process parameter combination. Through this innovation, the efficiency and accuracy of process optimization are significantly improved.

Description

Technical Field

[0001] The present invention relates to the technical field of material analysis, and in particular to a chemical vapor deposition automated analysis system manufactured based on atomic-level precision. Background Art

[0002] Chemical Vapor Deposition (CVD) is an important technology widely used in semiconductor, optoelectronic materials and other high-precision manufacturing fields. CVD converts gaseous reactants into solid thin film materials at high temperatures, covers the surface of the substrate, and forms the desired thin film structure. This technology has important applications in the fields of microelectronic device manufacturing, solar cells, and optical coatings.

[0003] Although CVD technology has been widely used in industrial production, there are still many challenges in high-precision control and automated analysis. First, the traditional CVD system lacks sufficient accuracy and real-time performance when controlling key process parameters such as reaction temperature, gas flow, pressure, substrate rotation speed, and precursor supply rate, resulting in large parameter fluctuations during the deposition process, affecting the uniformity and consistency of film quality. Secondly, existing process optimization methods usually rely on experience and trial and error processes, which are inefficient and difficult to efficiently search for the optimal parameter combination in a high-dimensional parameter space. In addition, existing technologies often lack real-time analysis and feedback on the physical and chemical properties of deposited materials, resulting in delayed adjustment of process parameters and inability to optimize the deposition process in a timely manner. Finally, the traditional CVD system has a low level of automation and intelligence in process control and adjustment, relies on manual intervention, and cannot achieve closed-loop control, which increases operational complexity and reduces production efficiency and product consistency.

[0004] In view of the limitations of the above-mentioned prior art, the present invention proposes a chemical vapor deposition automated analysis system based on atomic-level precision manufacturing. Summary of the invention

[0005] Based on the above objectives, the present invention provides a chemical vapor deposition automated analysis system based on atomic-level precision manufacturing.

[0006] The chemical vapor deposition automated analysis system based on atomic-level precision manufacturing includes a process control module, a process optimization module, an optimization revision module, an instruction sending module, a quality monitoring module, a data storage module and a user interface module, wherein;

[0007] The process control module is used to set and adjust the process parameters in the chemical vapor deposition process according to the control instructions, including reaction temperature, deposition time, gas flow rate, pressure, substrate rotation speed and precursor supply rate. The process control module is embedded with a data collector, which collects various process data in the chemical vapor deposition process in real time, including reaction temperature data, deposition time data, gas flow rate data, pressure data, substrate rotation speed data and precursor supply rate data;

[0008] The process optimization module is used to analyze the collected process data in real time and generate a first process optimization suggestion. The process optimization module specifically includes:

[0009] A data processing unit receives and pre-processes the reaction temperature data, deposition time data, gas flow data, pressure data, substrate rotation speed data, and precursor supply rate data transmitted by the data collector, performs data cleaning and filtering, and removes noise and abnormal values;

[0010] The parameter correlation analysis unit performs multi-dimensional correlation analysis on the pre-processed process data, establishes a relationship model between each process parameter and the deposition effect, and identifies the key parameters that have the greatest impact on the deposition quality;

[0011] The process optimization algorithm unit uses the Bayesian optimization algorithm to optimize the process parameters based on the correlation analysis results and generate the first process optimization suggestion. The Bayesian optimization algorithm builds a proxy model, quantifies uncertainty, and searches for the optimal process parameter combination;

[0012] The optimization revision module analyzes the physical and chemical properties of the deposited material, including the crystal structure, composition ratio and electrical properties, and revises the process optimization suggestions generated by the data analysis module based on the analysis results to obtain a revised second process optimization suggestion to ensure the performance and consistency of the final product;

[0013] The instruction sending module is connected to the optimization revision module and the process control module, converts the revised second process optimization suggestion into a control instruction, and automatically adjusts the process parameters in the chemical vapor deposition process;

[0014] The quality monitoring module monitors the quality of the film after chemical vapor deposition, including film thickness, surface roughness, and uniformity;

[0015] The data storage module is used to store various process data and analysis results to form a process database;

[0016] The user interface module is used to display various process parameters and analysis results, and accept process settings and adjustment instructions input by the user.

[0017] Optionally, the process control module specifically includes:

[0018] The reaction temperature control unit sets and adjusts the temperature in the reaction chamber according to the control instructions, adjusts and monitors the reaction temperature in real time through the heater and temperature sensor, and the data collector collects the reaction temperature data in real time;

[0019] A deposition time control unit is used to set and adjust the duration of the deposition process according to control instructions, and a data collector records deposition time data;

[0020] The gas flow control unit is used to set and adjust the flow of the reaction gas and the carrier gas according to the control instructions, and the gas flow is controlled by the flow meter and the regulating valve. The data collector collects the gas flow data in real time;

[0021] The pressure control unit is used to set and adjust the pressure in the reaction chamber according to the control instructions, adjust and monitor the pressure through the vacuum pump and pressure sensor, and the data collector collects pressure data in real time;

[0022] A substrate rotation control unit is used to set and adjust the rotation speed of the substrate according to the control instruction, control and monitor the substrate rotation speed through the rotation motor and the speed sensor, and the data collector collects the substrate rotation speed data in real time;

[0023] The precursor supply control unit is used to set and adjust the supply rate of the precursor compound according to the control instructions, control and monitor the precursor supply rate through the supply drive device and the flow sensor, and the data collector collects the precursor supply rate data in real time.

[0024] Optionally, the parameter correlation analysis unit establishes a relationship model between each process parameter and the deposition effect through multi-dimensional correlation analysis, specifically including:

[0025] Multiple linear regression model: used to preliminarily establish a relationship model between parameters. The regression model is expressed as: y = β 0 +β 1 x 1 +β 2 x 2 +…+β n x n +∈, where y is the deposition effect, x i is the process parameter, β is the regression coefficient, and ∈ is the error term;

[0026] Correlation analysis: Calculate the correlation coefficient between each parameter and the deposition effect:

[0027] Among them, x i and i are the observed values ​​of process parameters and deposition effect, respectively. and are the means of process parameters and deposition effects, respectively;

[0028] Based on the correlation analysis results, the key parameters with the greatest impact on the deposition quality were identified.

[0029] Optionally, the process optimization algorithm unit uses a Bayesian optimization algorithm to optimize and calculate the process parameters, and the key parameters identified from the parameter association analysis unit are used as input parameters of the Bayesian optimization. These key parameters play a decisive role in process optimization, ensuring that the optimization process focuses on the most important parameters, specifically including:

[0030] Constructing a surrogate model: Using Gaussian process as the surrogate model, approximate the target function. The mean function and covariance function of Gaussian process are defined as:

[0031] Mean function:

[0032] Covariance function:

[0033] Gaussian Process Model:

[0034] Select the acquisition function, which includes the expected improvement EI and the upper confidence limit UCB;

[0035] Optimize the acquisition function: By optimizing the acquisition function, select the next evaluation point. The optimization of the acquisition function is expressed as:

[0036] x next = argmax x EI(x) or x next = argmax x UCB(x);

[0037] Update the model: Conduct experiments at new evaluation points, obtain new data, and update the Gaussian process model, including the update of the mean and covariance;

[0038] Iterative optimization: Repeatedly update the model and iteratively optimize until the preset maximum number of iterations is reached or the optimization effect converges.

[0039] Optionally, the expected improvement function is expressed as: EI(x)=E[max(f(x)-f(x + ),0)], where f(x) + ) is the function value of the current optimal solution;

[0040] The upper confidence limit function is expressed as: UCB(x)=μ(x)+κσ(x), wherein μ(x) and σ(x) are the mean and standard deviation of the Gaussian process model, respectively, and κ is a control parameter.

[0041] Optionally, in the update model:

[0042] The updated mean is expressed as:

[0043] m new (x) = K(x,X) T (K(X,X)+σ 2 I) -1 Y;

[0044] The updated covariance is expressed as:

[0045] k new (x,x′)=k(x,x′)-K(x,X) T (K(X,X)+σ 2 I) -1 K(X,x′), where K is the covariance matrix, X and T are the training data and target values, respectively.

[0046] Optionally, the optimization and revision module specifically includes:

[0047] A crystal structure analysis unit is used to analyze the crystal structure of the deposited material and obtain crystal structure data by X-ray diffraction method;

[0048] A component ratio analysis unit is used to analyze the chemical composition of the deposited material and obtain component ratio data by energy dispersive X-ray spectroscopy;

[0049] An electrical performance analysis unit, used to analyze the electrical performance of the deposited material and obtain electrical performance data through a Hall effect test method;

[0050] The optimization and revision unit integrates the data from crystal structure analysis, component ratio analysis and electrical performance analysis to form a comprehensive data set, and obtains the crystal structure optimization results, component ratio optimization results and electrical performance optimization results based on the comprehensive data set;

[0051] Based on the crystal structure optimization results, the component ratio optimization results and the electrical performance optimization results, the generated first process optimization suggestion is revised to obtain a revised second process optimization suggestion.

[0052] Optionally, the comprehensive data set includes:

[0053] Crystal structure data: grain size, interplanar spacing, grain orientation;

[0054] Composition ratio data: the ratio of each element in the material;

[0055] Electrical performance data: carrier concentration, mobility, resistivity.

[0056] Optionally, the optimization results of the crystal structure, component ratio and electrical performance are integrated into the first process optimization suggestion to form an updated process parameter space, and the Gaussian process model in the Bayesian optimization is applied again to optimize the integrated data to generate a revised second process optimization suggestion;

[0057] The mean function and covariance function of the updated Gaussian process model are:

[0058] Mean function:

[0059] m updated (x) = K(x,X updated) T (K(X updated ,X updated )+σ 2 I) -1 Y updated ;

[0060] Covariance function:

[0061] k updated (x,x′)=k(x,x′)-K(x,X updated ) T (K(X updated ,X updated )+σ 2 I) -1 K(X updated ,x′);

[0062] Among them, X updated and Y updated is the integrated data set, representing the updated process parameters and the corresponding deposition effects.

[0063] Furthermore, the instruction sending module specifically includes:

[0064] a communication unit connected to the optimization revision module and the process control module, configured to receive the revised second process optimization suggestion and transmit it to the process control module;

[0065] An instruction conversion unit, used to convert the revised second process optimization suggestion into specific control instructions, including set values ​​of reaction temperature, deposition time, gas flow rate, pressure, substrate rotation speed and precursor supply rate;

[0066] The parameter adjustment unit is used to send the generated control instructions to the process control module to automatically adjust various process parameters in the chemical vapor deposition process.

[0067] Beneficial effects of the present invention:

[0068] In the present invention, the process optimization module analyzes the collected process data in real time through multi-dimensional correlation analysis and Bayesian optimization algorithm to generate a first process optimization suggestion. The optimization revision unit revises the first process optimization suggestion based on the physical and chemical property analysis results of the deposited material (including crystal structure, component ratio and electrical performance) to generate a more accurate and effective second process optimization suggestion. The Bayesian optimization algorithm uses a Gaussian process model to quantify uncertainty and efficiently searches for the optimal process parameter combination. Through this innovation, the efficiency and accuracy of process optimization are significantly improved.

[0069] In the present invention, the process optimization module performs multi-dimensional correlation analysis on the collected process data, identifies the key parameters that have the greatest impact on the deposition quality, and applies the Bayesian optimization algorithm to generate the first process optimization suggestion. The Bayesian optimization algorithm uses the Gaussian process model to quantify the uncertainty of each prediction by constructing a proxy model, efficiently searches for the optimal process parameter combination, and ensures that the best parameter combination is found in the high-dimensional parameter space. This innovation significantly improves the efficiency and accuracy of process optimization, and can achieve efficient optimization with less experimental data, thereby improving deposition quality and process consistency.

[0070] In the present invention, the optimization and revision module makes a second revision to the first process optimization suggestion based on the results of the physical and chemical property analysis (including crystal structure, component ratio and electrical properties) of the deposited material, analyzes the grain size and orientation through XRD, determines the optimal reaction temperature and substrate rotation speed, and integrates these optimization results into the first process optimization suggestion. The material component ratio is analyzed through EDS to determine the optimal gas flow rate and precursor supply rate, and these optimization results are integrated into the first process optimization suggestion. The carrier concentration and conductivity of the material are analyzed through the Hall effect test to determine the optimal deposition time and pressure, and these optimization results are integrated into the first process optimization suggestion. The optimization and revision unit regenerates the revised second process optimization suggestion by updating the Gaussian process model with the above analysis results. Through this second revision, it is ensured that the final process parameters are more in line with the physical and chemical properties of the material, significantly improving the deposition quality and consistency. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only for the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0072] Figure 1 A schematic diagram of a process optimization module according to an embodiment of the present invention;

[0073] Figure 2A schematic diagram of an optimization and revision module according to an embodiment of the present invention;

[0074] Figure 3 Schematic diagram of functional modules of an analysis system according to an embodiment of the present invention. DETAILED DESCRIPTION

[0075] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. At the same time, it is explained here that in order to make the embodiments more detailed, the following embodiments are the best and preferred embodiments, and those skilled in the art may also adopt other alternatives to implement some known technologies; and the accompanying drawings are only for more specific description of the embodiments, and are not intended to specifically limit the present invention.

[0076] It should be noted that the references to "one embodiment", "an embodiment", "an exemplary embodiment", "some embodiments" and the like in the specification indicate that the embodiments described may include specific features, structures or characteristics, but not every embodiment may include the specific features, structures or characteristics. In addition, when a specific feature, structure or characteristic is described in conjunction with an embodiment, it should be within the knowledge of a person skilled in the art to implement such feature, structure or characteristic in conjunction with other embodiments (whether or not explicitly described).

[0077] In general, a term can be understood, at least in part, from its use in context. For example, depending, at least in part, on the context, the term "one or more" as used herein can be used to describe any feature, structure, or characteristic in the singular sense, or can be used to describe a combination of features, structures, or characteristics in the plural sense. Additionally, the term "based on" can be understood as not necessarily intended to convey an exclusive set of factors, but can instead, depending, at least in part, on the context, allow for the presence of other factors that are not necessarily explicitly described.

[0078] like Figure 1-Figure 3 As shown, the chemical vapor deposition automated analysis system based on atomic-level precision manufacturing includes a process control module, a process optimization module, an optimization revision module, an instruction sending module, a quality monitoring module, a data storage module and a user interface module, wherein;

[0079] A process control module is used to set and adjust the process parameters in the chemical vapor deposition process according to the control instructions (which can be manual instruction input or instructions received from the instruction sending module), including reaction temperature, deposition time, gas flow rate, pressure, substrate rotation speed and precursor supply rate. Controlling the substrate rotation speed affects the uniformity of deposition, and controlling the supply rate of the precursor compound to adjust the reaction rate and the growth rate of the film. The process control module is embedded with a data collector, which collects various process data in the chemical vapor deposition process in real time, including reaction temperature data, deposition time data, gas flow data, pressure data, substrate rotation speed data and precursor supply rate data;

[0080] The process optimization module is used to analyze the collected process data in real time and generate the first process optimization suggestion. In the chemical vapor deposition process, the relationship between process parameters and deposition quality may be complex and nonlinear. The Bayesian optimization algorithm is used to approximate the objective function by constructing a proxy model (Gaussian process), which can efficiently handle complex and nonlinear objective functions and gradually approach the global optimal solution. The Bayesian optimization algorithm performs well in sample efficiency. By using the results of parameter association analysis, the proxy model is continuously updated to guide the next optimization direction. Compared with other optimization algorithms, Bayesian optimization requires fewer experiments and can find the optimal solution with less experimental data. Bayesian optimization can quantify the uncertainty of model prediction, which is very important for decision-making in the optimization process. Through uncertainty quantification, it is possible to identify which parameter areas require more data, thereby effectively guiding the experimental design and improving the optimization efficiency. It can flexibly respond to different process requirements and changes. Through the continuously updated proxy model, Bayesian optimization can adapt to changes that may occur in the process and maintain the stability and effectiveness of the optimization process; the process optimization module specifically includes:

[0081] The data processing unit receives and pre-processes the reaction temperature data, deposition time data, gas flow data, pressure data, substrate rotation speed data, and precursor supply rate data transmitted by the data collector, performs data cleaning and filtering, removes noise and outliers, and uses statistical methods such as the ZScore method to identify and remove outliers: Where X is the data point, μ is the mean of the data, σ is the standard deviation of the data, and |Z|>3 is considered an outlier; data filtering applies a low-pass filter: y[t]=α·x[t]+(1-α)·y[t-1], where y[t] is the filtered data, x[t] is the original data, and α is the filter coefficient, which ranges from 0 to 1;

[0082] The parameter correlation analysis unit performs multi-dimensional correlation analysis on the pre-processed process data, establishes a relationship model between each process parameter and the deposition effect, and identifies the key parameters that have the greatest impact on the deposition quality;

[0083] The process optimization algorithm unit uses the Bayesian optimization algorithm to optimize the process parameters based on the correlation analysis results and generate the first process optimization suggestion. The Bayesian optimization algorithm builds a proxy model to quantify uncertainty and search for the optimal process parameter combination. The covariance function and standard deviation of the proxy model are used to quantify uncertainty, and the expected improvement function is used to efficiently search for the optimal parameters. The next evaluation point is selected by maximizing EI, and the next evaluation point is selected by optimizing the acquisition function to achieve efficient search for the optimal process parameter combination.

[0084] The optimization and revision module analyzes the physical and chemical properties of the deposited materials, including crystal structure, composition ratio and electrical properties, and revises the process optimization suggestions generated by the data analysis module based on the analysis results to obtain the revised second process optimization suggestions to ensure the performance and consistency of the final product;

[0085] An instruction sending module is connected to the optimization revision module and the process control module to convert the revised second process optimization suggestion into a control instruction to automatically adjust the process parameters in the chemical vapor deposition process;

[0086] The quality monitoring module monitors the quality of the film after chemical vapor deposition, including film thickness, surface roughness, and uniformity;

[0087] Data storage module, used to store various process data and analysis results to form a process database;

[0088] The user interface module is used to display various process parameters, analysis results, and accept process settings and adjustment instructions input by the user.

[0089] The process control module specifically includes:

[0090] The reaction temperature control unit sets and adjusts the temperature in the reaction chamber according to the control instructions, adjusts and monitors the reaction temperature in real time through the heater and temperature sensor, and the data collector collects the reaction temperature data in real time; the reaction temperature control unit accurately controls the temperature in the reaction chamber by setting the power of the heater. The temperature sensor monitors the temperature of the reaction zone in real time and feeds the monitoring data back to the control system to ensure that the temperature is maintained at the set value. This unit can not only adjust the power of the heater, but also enable the cooling system as needed to deal with over-high temperatures;

[0091] A deposition time control unit is used to set and adjust the duration of the deposition process according to control instructions, and a data collector records deposition time data; the deposition time control unit accurately controls the length of the deposition time by setting the start and end time of the deposition process. The unit can preset the deposition cycle and automatically stop the reaction when the deposition time is reached. The data collector records the start and end time points of the deposition and calculates the total deposition time. These time data are used to evaluate the stability and repeatability of the deposition process and provide basic data for subsequent optimization;

[0092] The gas flow control unit is used to set and adjust the flow of reaction gas and carrier gas according to the control instructions. The gas flow is controlled by the flow meter and the regulating valve. The data collector collects the gas flow data in real time. The gas flow control unit controls the flow of reaction gas and carrier gas by adjusting the opening of the valve. The flow meter measures the flow of each gas in real time and feeds the data back to the control system to ensure that the gas flow is accurately controlled within the set value range. The unit can independently adjust and monitor multiple gases to ensure the accuracy of the mixed gas ratio.

[0093] The pressure control unit is used to set and adjust the pressure in the reaction chamber according to the control instructions. The pressure is adjusted and monitored by the vacuum pump and pressure sensor. The data collector collects pressure data in real time. The pressure control unit controls the pressure in the reaction chamber by adjusting the working state of the vacuum pump and the sealing of the reaction chamber. The pressure sensor monitors the pressure in the chamber in real time and feeds the data back to the control system to maintain a stable reaction pressure. The unit can accurately adjust the reaction pressure to meet different material and process requirements.

[0094] The substrate rotation control unit is used to set and adjust the rotation speed of the substrate according to the control instructions. The substrate rotation speed is controlled and monitored by the rotation motor and the speed sensor. The data collector collects the substrate rotation speed data in real time. The substrate rotation control unit adjusts the rotation speed of the substrate by controlling the rotation speed of the rotation motor. The speed sensor monitors the rotation speed of the substrate in real time and feeds the data back to the control system to ensure the stability and uniformity of the rotation speed. The unit can realize uniform rotation of the substrate or periodic variable speed rotation to optimize the uniformity of deposition and the quality of the film.

[0095] The precursor supply control unit is used to set and adjust the supply rate of the precursor compound according to the control instructions. The precursor supply rate is controlled and monitored by the supply drive device and the flow sensor. The data collector collects the precursor supply rate data in real time. The precursor supply control unit controls the supply rate of the precursor compound by adjusting the flow rate of the supply device. The flow sensor monitors the supply rate of the precursor in real time and feeds the data back to the control system to ensure the stability and accuracy of the supply rate. The unit can accurately control the supply rate according to different precursor compounds and process requirements to optimize the reaction rate and film growth rate.

[0096] The parameter correlation analysis unit establishes a relationship model between each process parameter and the deposition effect through multi-dimensional correlation analysis, including:

[0097] Multiple linear regression model: used to preliminarily establish the relationship model between parameters. The regression model is expressed as: y = β 0 +β 1 x 1 +β 2 x 2 +…+β n x n +∈, where y is the deposition effect, x i is the process parameter, β is the regression coefficient, and ∈ is the error term;

[0098] The regression coefficient β is determined by the least squares method as follows:

[0099] 1. Construct the design matrix: Let X be the design matrix, containing all observations, and y be the target vector, containing all observations of the deposition effect.

[0100]

[0101] 2. Least Squares Formula: Estimated value of regression coefficient β Determined by minimizing the sum of squared errors: Among them, X T is the transposed matrix of X, (X T X) -1 Yes X T The inverse matrix of X.

[0102] The error term ∈ is estimated by the residuals of the regression model: The mean of the residuals should be close to zero, and the standard deviation is used to assess the fit of the regression model;

[0103] Correlation analysis: Calculate the correlation coefficient between each parameter and the deposition effect:

[0104] Among them, xi and i are the observed values ​​of process parameters and deposition effect, respectively. and are the means of process parameters and deposition effects, respectively;

[0105] Based on the correlation analysis results, the key parameters that have the greatest impact on deposition quality are identified. They are as follows:

[0106] 1. Calculate the correlation coefficient: Calculate the correlation coefficient r between each process parameter and the deposition effect.

[0107] 2. Evaluate the strength of correlation: The strength of correlation is determined based on the absolute value of the correlation coefficient. |r|>0.5 indicates a strong correlation, |r|>0.3 indicates a moderate correlation, and |r|≤0.3 indicates a weak correlation. Select process parameters that have a strong or moderate correlation with the deposition effect as key parameters. These parameters will serve as input variables for Bayesian optimization.

[0108] The process optimization algorithm unit uses the Bayesian optimization algorithm to optimize the process parameters. The key parameters identified from the parameter association analysis unit are used as the input parameters of the Bayesian optimization. These key parameters play a decisive role in process optimization, ensuring that the optimization process focuses on the most important parameters, including:

[0109] Constructing a surrogate model: Using Gaussian process as the surrogate model, approximate the target function. The mean function and covariance function of Gaussian process are defined as:

[0110] Mean function:

[0111] Covariance function:

[0112] Gaussian Process Model:

[0113] Select the acquisition function, which includes the expected improvement EI and the upper confidence limit UCB;

[0114] Optimize the acquisition function: By optimizing the acquisition function, select the next evaluation point. The optimization of the acquisition function is expressed as:

[0115] x next = argmax x EI(x) or x next = argmax x UCB(x);

[0116] Update the model: Conduct experiments at new evaluation points, obtain new data, and update the Gaussian process model, including the update of the mean and covariance;

[0117] Iterative optimization: Repeatedly update the model and iteratively optimize until the preset maximum number of iterations is reached or the optimization effect converges.

[0118] The expected improvement function is expressed as: EI(x) = E[max(f(x)-f(x + ),0)], where f(x + ) is the function value of the current optimal solution;

[0119] The upper confidence limit function is expressed as: UCB(x)=μ(x)+κσ(x), where μ(x) and σ(x) are the mean and standard deviation of the Gaussian process model, respectively, and κ is a control parameter.

[0120] Update the model:

[0121] The updated mean is expressed as:

[0122] m new (x) = K(x,X) T (K(X,X)+σ 2 I) -1 Y;

[0123] The updated covariance is expressed as:

[0124] k new (x,x′)=k(x,x′)-K(x,X) T (K(X,X)+σ 2 I) -1 K(X,x′), where J is the covariance matrix, X and Y are the training data and target values, respectively.

[0125] The optimized revision module specifically includes:

[0126] The crystal structure analysis unit is used to analyze the crystal structure of the deposited material and obtain the crystal structure data by X-ray diffraction method; X-ray diffraction (XRD) uses XRD equipment to measure the crystal structure of the deposited material and determines the crystal structure by the position and intensity of the diffraction peak. The processing formula is: Bragg's law nλ=2dsinθ, where λ is the X-ray wavelength, d is the crystal plane spacing, and θ is the diffraction angle;

[0127] The composition ratio analysis unit is used to analyze the chemical composition of the deposited material and obtain the composition ratio data through the energy dispersive X-ray spectroscopy method; Energy dispersive X-ray spectroscopy (EDS) uses EDS equipment to detect the elements in the deposited material and determines the composition ratio of the material by analyzing the energy and intensity of the characteristic X-rays. The processing formula is: quantitative analysis formula Among them, C i is the mass fraction of element i, I i is the characteristic X-ray intensity of element i;

[0128] The electrical performance analysis unit is used to analyze the electrical properties of the deposited material and obtain electrical performance data through the Hall effect test method; by measuring the Hall voltage, the carrier concentration and mobility of the material are determined, and the processing formula is: Hall coefficient Among them, V H is the Hall voltage, t is the sample thickness, I is the current, and B is the magnetic field strength;

[0129] The optimization and revision unit integrates the data from crystal structure analysis, component ratio analysis and electrical performance analysis to form a comprehensive data set, and obtains the crystal structure optimization results, component ratio optimization results and electrical performance optimization results based on the comprehensive data set;

[0130] Based on the crystal structure optimization results, the component ratio optimization results and the electrical performance optimization results, the generated first process optimization suggestion is revised to obtain a revised second process optimization suggestion.

[0131] Based on the integrated data set, the optimization revision unit updates the process parameter space. The specific steps are as follows:

[0132] Crystal structure optimization:

[0133] Crystal structure data were analyzed to determine the optimal reaction temperature and substrate rotation speed to optimize grain size and crystal orientation.

[0134] These optimization results are integrated into the first process optimization suggestion. For example, the reaction temperature parameter in the first process optimization suggestion is adjusted to the temperature range with the most uniform grain size, and the substrate rotation speed parameter is adjusted to the speed range with the best grain orientation.

[0135] Optimized ingredient ratio:

[0136] Analyze composition ratio data to determine optimal gas flow and precursor feed rates to ensure uniformity and consistency of material composition.

[0137] Integrate these optimization results into the first process optimization suggestion. For example, adjust the gas flow rate parameters in the first process optimization suggestion to the flow rate range with the best component ratio, and adjust the precursor supply rate parameters to the supply rate range with the best component ratio.

[0138] Electrical performance optimization:

[0139] Analyze electrical performance data to determine the best deposition time and pressure to optimize the material's carrier concentration and conductivity.

[0140] These optimization results are integrated into the first process optimization suggestion. For example, the deposition time parameter in the first process optimization suggestion is adjusted to a time range with optimal conductivity, and the pressure parameter is adjusted to a pressure range with optimal carrier concentration.

[0141] The integrated optimal parameter range is substituted into the first process optimization suggestion, and the updated process parameter space will be used for further Bayesian optimization to generate a more accurate second process optimization suggestion.

[0142] When it comes to specific implementation, the further plan is:

[0143] 1. Crystal structure optimization

[0144] Analytical methods:

[0145] X-ray diffraction (XRD): used to measure grain size and interplanar spacing, and analyze grain size and orientation through the position and intensity of diffraction peaks.

[0146] Specific steps:

[0147] Determine the optimal reaction temperature:

[0148] Through XRD analysis, the grain size and interplanar spacing data at different temperatures were recorded.

[0149] Compare the grain uniformity at each temperature and select the temperature range with the most uniform grain size.

[0150] For example, if XRD analysis shows that the grain size is most uniform in the range of 450°C to 500°C, the reaction temperature is adjusted to 450°C to 500°C.

[0151] Determine the optimal substrate rotation speed:

[0152] By EBSD analysis, the grain orientation and grain boundary distribution at different rotation speeds were recorded.

[0153] Compare the grain orientation consistency at each rotation speed and select the speed range with the best grain orientation.

[0154] For example, if the EBSD analysis shows that the grain orientation is most consistent in the range of 100 rpm to 150 rpm, the substrate rotation speed is adjusted to 100 rpm to 150 rpm.

[0155] 2. Optimization of ingredient ratio

[0156] Analytical methods:

[0157] Energy dispersive X-ray spectroscopy (EDS): used to analyze the elemental composition of the deposited material and record the composition ratio under different process parameters.

[0158] Specific steps:

[0159] Determine the optimal gas flow rate:

[0160] Through EDS analysis, the material composition ratio data under different gas flow rates are recorded.

[0161] Compare the composition uniformity of each gas flow meter and select the flow range with the most uniform composition ratio.

[0162] For example, EDS analysis shows that the composition ratio is optimal within the range of 200 sccm to 300 sccm, and the gas flow rate is adjusted to 200 sccm to 300 sccm.

[0163] Determine the optimal precursor feed rate:

[0164] Through XRF analysis, the element content data at different supply rates were recorded.

[0165] Compare the consistency of ingredients at each supply rate and select the supply rate range with the optimal ingredient ratio.

[0166] For example, if XRF analysis shows that the optimal component ratio is in the range of 2 g / min to 3 g / min, the precursor supply rate is adjusted to 2 g / min to 3 g / min.

[0167] 3. Electrical performance optimization

[0168] Analytical methods:

[0169] Hall effect measurement: used to measure the carrier concentration and mobility of materials and record the electrical properties under different process parameters.

[0170] Specific steps:

[0171] Determine the optimal deposition time:

[0172] The carrier concentration and mobility at different deposition times were recorded by Hall effect testing.

[0173] The electrical properties at different deposition times were compared, and the time range with the best carrier concentration and mobility was selected.

[0174] For example, if the Hall effect test shows that the carrier concentration and mobility are optimal in the range of 20 minutes to 40 minutes, the deposition time is adjusted to 20 minutes to 40 minutes.

[0175] Determine the optimal pressure:

[0176] Through four-probe testing, the resistivity data under different pressures are recorded.

[0177] Compare the conductivity at various pressures and select the pressure range with the best conductivity.

[0178] For example, if the four-probe test shows that the conductivity is best in the range of 20 Torr to 30 Torr, the pressure is adjusted to 20 Torr to 30 Torr.

[0179] The comprehensive dataset includes:

[0180] Crystal structure data: grain size, interplanar spacing, grain orientation;

[0181] Composition ratio data: the ratio of each element in the material;

[0182] Electrical performance data: carrier concentration, mobility, resistivity.

[0183] The optimization results of crystal structure, component ratio and electrical performance are integrated into the first process optimization suggestion to form an updated process parameter space. The Gaussian process model in Bayesian optimization is applied again to optimize the integrated data to generate a revised second process optimization suggestion.

[0184] The mean function and covariance function of the updated Gaussian process model are:

[0185] Mean function:

[0186] m updated (x) = K(x,X updated ) T (K(X updated ,X updated )+σ 2 I) -1 Y updated ;

[0187] Covariance function:

[0188] k updated (x,x′)=k(x,x′)-K(x,X updated ) T (K(X updated ,X updated )+σ 2 I) -1 K(X updated ,x′);

[0189] Among them, X updated and Y updated is the integrated data set, representing the updated process parameters and the corresponding deposition effects.

[0190] The instruction sending module specifically includes:

[0191] a communication unit connected to the optimization revision module and the process control module, for receiving the revised second process optimization suggestion and transmitting it to the process control module;

[0192] An instruction conversion unit, used to convert the revised second process optimization suggestion into specific control instructions, including set values ​​of reaction temperature, deposition time, gas flow rate, pressure, substrate rotation speed and precursor supply rate;

[0193] The parameter adjustment unit is used to send the generated control instructions to the process control module to automatically adjust various process parameters in the chemical vapor deposition process.

[0194] The instruction conversion unit is responsible for converting the revised second process optimization suggestion into specific control instructions, parsing the various process parameters in the revised second process optimization suggestion, including reaction temperature, deposition time, gas flow rate, pressure, substrate rotation speed and precursor supply rate, and generating corresponding control instructions according to the parsed process parameters. For example, the reaction temperature is adjusted to 450°C to 500°C, the deposition time is adjusted to 20 minutes to 40 minutes, the gas flow rate is adjusted to 200sccm to 300sccm, the pressure is adjusted to 20Torr to 30Torr, the substrate rotation speed is adjusted to 100rpm to 150rpm, and the precursor supply rate is adjusted to 2g / min to 3g / min.

[0195] The parameter adjustment unit is responsible for sending the generated control instructions to the process control module and automatically adjusting various process parameters in the chemical vapor deposition process. The generated control instructions are sent to the process control module through the communication unit. The process control module automatically adjusts various process parameters in the chemical vapor deposition process according to the received control instructions, for example, adjusting the heater power to control the reaction temperature, adjusting the valve opening to control the gas flow, adjusting the vacuum pump to control the pressure, controlling the rotating motor to adjust the substrate rotation speed, and adjusting the supply device to control the precursor supply rate.

[0196] The present invention covers any substitution, modification, equivalent method and scheme made on the essence and scope of the present invention. In order to make the public have a thorough understanding of the present invention, specific details are described in detail in the following preferred embodiments of the present invention, but those skilled in the art can fully understand the present invention without the description of these details. In addition, in order to avoid unnecessary confusion about the essence of the present invention, well-known methods, processes, procedures, components and circuits are not described in detail.

[0197] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A chemical vapor deposition automated analysis system based on atomic-level precision manufacturing, characterized in that: It includes a process control module, a process optimization module, an optimization revision module, an instruction sending module, a quality monitoring module, a data storage module and a user interface module, wherein; The process control module is used to set and adjust the process parameters in the chemical vapor deposition process according to the control instructions, including reaction temperature, deposition time, gas flow rate, pressure, substrate rotation speed and precursor supply rate. The process control module is embedded with a data collector, which collects various process data in the chemical vapor deposition process in real time, including reaction temperature data, deposition time data, gas flow rate data, pressure data, substrate rotation speed data and precursor supply rate data; The process optimization module is used to analyze the collected process data in real time and generate a first process optimization suggestion. The process optimization module specifically includes: A data processing unit receives and pre-processes the reaction temperature data, deposition time data, gas flow data, pressure data, substrate rotation speed data, and precursor supply rate data transmitted by the data collector, performs data cleaning and filtering, and removes noise and abnormal values; The parameter correlation analysis unit performs multi-dimensional correlation analysis on the pre-processed process data, establishes a relationship model between each process parameter and the deposition effect, and identifies the key parameters that have the greatest impact on the deposition quality; The process optimization algorithm unit uses the Bayesian optimization algorithm to optimize the process parameters based on the correlation analysis results and generate the first process optimization suggestion. The Bayesian optimization algorithm builds a proxy model, quantifies uncertainty, and searches for the optimal process parameter combination; The optimization revision module analyzes the physical and chemical properties of the deposited material, including the crystal structure, component ratio and electrical properties, and revises the process optimization suggestion generated by the data analysis module for a second time according to the analysis results to obtain a revised second process optimization suggestion; The parameter correlation analysis unit establishes a relationship model between each process parameter and the deposition effect through multi-dimensional correlation analysis, specifically including: Multiple linear regression model: used to preliminarily establish the relationship model between parameters. The regression model is expressed as: ,in, is the sedimentation effect, is the process parameter, is the regression coefficient, is the error term; Correlation analysis: Calculate the correlation coefficient between each parameter and the deposition effect: ,in, and are the observed values ​​of process parameters and deposition effect, respectively. and are the means of process parameters and deposition effects, respectively; Based on the correlation analysis results, the key parameters that have the greatest impact on the deposition quality are identified; The process optimization algorithm unit uses the Bayesian optimization algorithm to optimize and calculate the process parameters. The key parameters identified from the parameter association analysis unit are used as input parameters of the Bayesian optimization, specifically including: Constructing a surrogate model: Using Gaussian process as the surrogate model, approximate the target function. The mean function and covariance function of Gaussian process are defined as: Mean function: ; Covariance function: ; Gaussian Process Model: ; Select the acquisition function, which includes the expected improvement and the upper confidence limit; Optimize the acquisition function: By optimizing the acquisition function, select the next evaluation point. The optimization of the acquisition function is expressed as: ； Update the model: Conduct experiments at new evaluation points, obtain new data, and update the Gaussian process model, including the update of the mean and covariance; Iterative optimization: Repeatedly update the model and iteratively optimize until the preset maximum number of iterations is reached or the optimization effect converges; The expected improvement function is expressed as: ,in, is the function value of the current optimal solution; The upper confidence limit function is expressed as: ,in, and are the mean and standard deviation of the Gaussian process model, is the control parameter; In the update model: The updated mean is expressed as: ； The updated covariance is expressed as: ,in, is the covariance matrix, and are training data and target values ​​respectively; The optimization and revision module specifically includes: A crystal structure analysis unit is used to analyze the crystal structure of the deposited material and obtain crystal structure data by X-ray diffraction method; A component ratio analysis unit is used to analyze the chemical composition of the deposited material and obtain component ratio data by energy dispersive X-ray spectroscopy; An electrical performance analysis unit, used to analyze the electrical performance of the deposited material and obtain electrical performance data through a Hall effect test method; The optimization and revision unit integrates the data from crystal structure analysis, component ratio analysis and electrical performance analysis to form a comprehensive data set, and obtains the crystal structure optimization results, component ratio optimization results and electrical performance optimization results based on the comprehensive data set; Based on the crystal structure optimization results, the component ratio optimization results, and the electrical performance optimization results, the generated first process optimization suggestion is revised to obtain a revised second process optimization suggestion; The instruction sending module is connected to the optimization revision module and the process control module, converts the revised second process optimization suggestion into a control instruction, and automatically adjusts the process parameters in the chemical vapor deposition process; The quality monitoring module monitors the quality of the film after chemical vapor deposition, including film thickness, surface roughness, and uniformity; The data storage module is used to store various process data and analysis results to form a process database; The user interface module is used to display various process parameters and analysis results, and accept process settings and adjustment instructions input by the user.

2. The chemical vapor deposition automated analysis system based on atomic level precision manufacturing according to claim 1, characterized in that: The process control module specifically includes: The reaction temperature control unit sets and adjusts the temperature in the reaction chamber according to the control instructions, adjusts and monitors the reaction temperature in real time through the heater and temperature sensor, and the data collector collects the reaction temperature data in real time; A deposition time control unit is used to set and adjust the duration of the deposition process according to control instructions, and a data collector records deposition time data; The gas flow control unit is used to set and adjust the flow of the reaction gas and the carrier gas according to the control instructions, and the gas flow is controlled by the flow meter and the regulating valve. The data collector collects the gas flow data in real time; The pressure control unit is used to set and adjust the pressure in the reaction chamber according to the control instructions, adjust and monitor the pressure through the vacuum pump and pressure sensor, and the data collector collects pressure data in real time; A substrate rotation control unit is used to set and adjust the rotation speed of the substrate according to the control instruction, control and monitor the substrate rotation speed through the rotation motor and the speed sensor, and the data collector collects the substrate rotation speed data in real time; The precursor supply control unit is used to set and adjust the supply rate of the precursor compound according to the control instructions, control and monitor the precursor supply rate through the supply drive device and the flow sensor, and the data collector collects the precursor supply rate data in real time.

3. The chemical vapor deposition automated analysis system based on atomic level precision manufacturing according to claim 1, characterized in that: The comprehensive dataset includes: Crystal structure data: grain size, interplanar spacing, grain orientation; Composition ratio data: the ratio of each element in the material; Electrical performance data: carrier concentration, mobility, resistivity.

4. The chemical vapor deposition automated analysis system based on atomic level precision manufacturing according to claim 1, characterized in that: The optimization results of the crystal structure, component ratio and electrical performance are integrated into the first process optimization suggestion to form an updated process parameter space, and the Gaussian process model in the Bayesian optimization is applied again to optimize the integrated data to generate a revised second process optimization suggestion; The mean function and covariance function of the updated Gaussian process model are: Mean function: ； Covariance function: ； in, and is the integrated data set, representing the updated process parameters and the corresponding deposition effects.

5. The chemical vapor deposition automated analysis system based on atomic level precision manufacturing according to claim 4, characterized in that: The instruction sending module specifically includes: a communication unit connected to the optimization revision module and the process control module, configured to receive the revised second process optimization suggestion and transmit it to the process control module; An instruction conversion unit, used to convert the revised second process optimization suggestion into specific control instructions, including set values ​​of reaction temperature, deposition time, gas flow rate, pressure, substrate rotation speed and precursor supply rate; The parameter adjustment unit is used to send the generated control instructions to the process control module to automatically adjust various process parameters in the chemical vapor deposition process.

Images