Superlattice material wavelength calculation method and system and computer equipment

Abstract

The invention discloses a superlattice material wavelength calculation method and system and computer equipment, and relates to the technical field of superlattice materials. The superlattice material wavelength calculation method comprises the following steps: acquiring theoretical wavelengths, material layer numbers, process parameters and actual wavelengths of historical samples to obtain training set data; a neural network model comprising an input layer, a hidden layer and an output layer is created, the input layer is composed of theoretical wavelengths, the number of material layers and technological parameters, and the output layer is composed of actual wavelengths; training a neural network model by using the training set data, and stopping training when the error between the calculated wavelength and the actual wavelength is smaller than a preset value to obtain a trained neural network model; and inputting the theoretical wavelength, the material layer number and the process parameters of the current sample into the trained neural network model to obtain the calculation wavelength of the current sample. According to the method for calculating the wavelength of the superlattice material, the accuracy of calculating the wavelength of the superlattice material is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of superlattice materials, and in particular to a method, system and computer equipment for calculating the wavelength of a superlattice material. Background Art

[0002] Superlattice materials are periodic multilayer structures formed by the alternating growth of two or more different materials. The thickness of these material layers is usually at the nanoscale, which makes superlattice materials exhibit many unique physical and chemical properties. They are widely used in detectors, lasers, and modulators. Their band structure can be easily tailored by adjusting the material structure, and the wavelength coverage range is wide, which can meet the needs of shortwave, medium wave and long wave bands.

[0003] There are many theoretical calculation methods for the band gap of superlattice materials, such as method, tight binding approximation method, pseudopotential method, etc. The method has the smallest amount of calculation and is the most widely used among these methods. This method takes into account the coupling of the conduction band and the valence band, as well as the influence of strain, etc. By assuming that electrons move in the periodically arranged and fixed nuclear potential field and the average potential field of other electrons, the band gap value under ideal conditions is calculated; however, in actual situations, superlattice materials do not grow strictly according to the desired structure. Their growth modes include two-dimensional growth, island growth, etc., so that the atomic arrangement of the material is not strictly arranged according to periodic repetition, and different process parameters in the growth process will have different effects on the background atmosphere, leading to segregation, relaxation and other phenomena, causing changes in the thickness component of the material, and the stress it is subjected to is also different from the theoretical value; these reasons may cause the actual wavelength of the final material to deviate from the target, and the longer the wavelength, the more obvious the impact, which has an adverse effect on the accuracy and efficiency of research and development in process production. Summary of the invention

[0004] The present invention provides a method, system and computer equipment for calculating the wavelength of a superlattice material, aiming to improve the accuracy of calculating the wavelength of a superlattice material.

[0005] The method for calculating the wavelength of a superlattice material provided by the present invention comprises:

[0006] S1, obtain the theoretical wavelength, number of material layers, process parameters and actual wavelength of historical samples to obtain training set data;

[0007] S2, creating a neural network model, the neural network model comprising an input layer, a hidden layer and an output layer, the input layer is composed of a theoretical wavelength, a number of material layers and process parameters, and the output layer is composed of an actual wavelength;

[0008] S3, using the training set data to train the neural network model, the data of the input layer is transmitted to the output layer through the hidden layer calculation to obtain the calculated wavelength; when the error between the calculated wavelength and the actual wavelength is less than a preset value, the training is stopped to obtain a trained neural network model;

[0009] S4, inputting the theoretical wavelength, number of material layers and process parameters of the current sample into the trained neural network model to obtain the calculated wavelength of the current sample.

[0010] Optionally, the process parameters include at least one of valve position, source furnace temperature, and growth time.

[0011] Optionally, in step S1 or step S4, the theoretical wavelength is The method is calculated, which specifically includes the following steps:

[0012] Construct a Hamiltonian matrix, which is:

[0013]

[0014] In the formula, represents the 8×8 diagonalized Hamiltonian matrix based on Luttinger-Kohn theory, and denote the decoupled 4×4 upper Hamiltonian matrix and lower Hamiltonian matrix respectively;

[0015] The electronic Schrödinger equation under the periodic potential field is solved according to the Hamiltonian matrix and the material parameters of the sample to obtain the theoretical band gap and theoretical wavelength of the sample; the electronic Schrödinger equation is:

[0016]

[0017] Where m0 is the mass of free electrons, is the Hamiltonian operator, E is the energy eigenvalue, is the wave function that describes the motion state of the electron, is the reduced Planck constant, is a periodic potential field.

[0018] Optionally, the step S2 further includes:

[0019] The number of neuron nodes in the input layer, hidden layer and output layer is set, and the initial weight and threshold of each neuron node are determined.

[0020] Optionally, step S3 further includes:

[0021] When the error between the calculated wavelength and the actual wavelength is greater than or equal to the preset value, the weight and threshold of each neuron node are adjusted, and the data of the input layer is recalculated.

[0022] Optionally, the step S1 further includes: dividing historical samples into a training set and a test set to obtain training set data and test set data; and after the step S3, further including the following steps: using the test set data to verify the trained neural network model.

[0023] Optionally, the neural network model is a BP neural network model.

[0024] The superlattice material wavelength calculation system provided by the present invention comprises:

[0025] The acquisition module is used to obtain the theoretical wavelength, number of material layers, process parameters and actual wavelength of historical samples;

[0026] A partitioning module, used for partitioning the data of the historical samples into a training set and a test set;

[0027] A determination module, used to determine the topological structure of the neural network model according to the data of the training set; the neural network model includes an input layer, a hidden layer and an output layer, the input layer is composed of the theoretical wavelength, the number of material layers and process parameters, and the output layer is composed of the actual wavelength;

[0028] A training module, used to train the neural network model using the data of the training set to obtain a trained neural network model;

[0029] A verification module, used to verify the trained neural network model according to the data of the test set;

[0030] The calculation module is used to use the trained neural network model to calculate the calculated wavelength of the current sample according to the input theoretical wavelength, number of material layers and process parameters of the current sample.

[0031] Optionally, the superlattice material wavelength calculation system further includes:

[0032] Theoretical wavelength calculation module is used to calculate the wavelength according to the material parameters of the sample. The theoretical wavelength of the sample can be calculated by this method.

[0033] The computer device provided by the present invention includes a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the above-mentioned method for calculating the wavelength of superlattice materials when executing the computer program.

[0034] The present invention has the following beneficial effects:

[0035] The present invention takes process parameters as influencing factors of the actual output wavelength of superlattice materials, and uses a neural network model to optimize the calculation results of the wavelength. Compared with the traditional theoretical wavelength calculation method, the accuracy of superlattice material wavelength calculation can be greatly improved, and an adaptive neural network model can be trained according to the process conditions of different equipment and machines, so as to calculate the wavelength of the produced superlattice materials in real time, efficiently and accurately, which can improve the research and development accuracy and efficiency in the process production of superlattice materials. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in 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 described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0037] Figure 1 A flowchart of some embodiments of the method for calculating the wavelength of superlattice materials of the present invention;

[0038] Figure 2 A schematic diagram of a neural network structure of some embodiments of the method for calculating the wavelength of a superlattice material of the present invention;

[0039] Figure 3 A schematic diagram of a neural network training process for some embodiments of the method for calculating the wavelength of a superlattice material of the present invention;

[0040] Figure 4 A schematic diagram of the calculation process of some embodiments of the method for calculating the wavelength of superlattice materials of the present invention;

[0041] Figure 5 This is a graph of the actual wavelength, theoretical wavelength and calculated wavelength of an InAs / GaSb superlattice material sample in a specific embodiment of the present invention. DETAILED DESCRIPTION

[0042] In order to make the purpose, features and advantages of the present invention more obvious and easy to understand, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described below are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0043] The present invention proposes a method for calculating the wavelength of superlattice materials, aiming to consider the influence of process parameters on the strain of superlattice materials and the actual material structure, and utilize the autonomous learning and optimization capabilities of a neural network model to improve the accuracy of superlattice material wavelength calculation.

[0044] See also Figure 1 In some embodiments, the method for calculating the wavelength of a superlattice material comprises the following steps:

[0045] S1, obtain the theoretical wavelength, number of material layers, process parameters and actual wavelength of historical samples to obtain training set data, which is used to train the neural network model.

[0046] Specifically, the actual wavelength of the superlattice material is the wavelength obtained by an actual PL test (photoluminescence test) of the sample.

[0047] The theoretical wavelength of superlattice materials can be Method calculation.

[0048] The method is based on the Luttinger-Kohn theory, considering the coupling between the conduction band and the valence band and the influence of strain, and using the wave function approximated by the envelope function to solve the Hamiltonian equation. After block diagonalization, the decoupled 8×8 block diagonal Hamiltonian matrix is ​​obtained; after determining the absorption layer structure, the matrix is ​​solved to obtain the energy level distribution at the center of the Brillouin zone, and then the energy level distribution at the center of the Brillouin zone can be obtained. Theoretical band gap and wavelength.

[0049] Although The 8-band theoretical model takes into account the influence of strain, but in some scenarios, for long-wave materials, the number of InAs layers is thick, the strain is difficult to balance, and relaxation and dislocation may occur, resulting in The numerical value of the strain influence part in the theoretical model has changed, which is inconsistent with the actual situation.

[0050] Since the strain value is related to the thickness of each layer, in order to facilitate subsequent input, in an embodiment of the present invention, the number of atomic layers of each material in a superlattice period (ie, the number of material layers) is directly input into the neural network model as an influencing factor.

[0051] In actual production, when superlattice materials are grown using MBE (molecular beam epitaxy) equipment, the set process parameters will affect the final growth quality of the material; the source furnace temperature, beam size, and even the growth time set during growth may have an impact on the material; for example, some materials have segregation diffusion, which can cause the material structure to differ from the target design, making the originally designed binary alloy actually a ternary alloy or even a quaternary alloy; therefore, it is necessary to select process parameters that may affect the above situation based on the actual material structure.

[0052] In some embodiments, the process parameter selects at least one of a valve position, a source furnace temperature, and a growth time.

[0053] S2, creating a neural network model and determining the topological structure of the neural network model based on the data of the training set.

[0054] See also Figure 2 ,The neural network model includes input layer, hidden layer and output layer.

[0055] The input layer is composed of theoretical wavelength, number of material layers and process parameters, the output layer is composed of actual wavelength, and the appropriate number of neuron nodes in the input layer, hidden layer and output layer is set.

[0056] The neural network model adopted in the embodiment of the present invention is a BP neural network model.

[0057] The BP neural network model is a multi-layer feedforward neural network widely used in the field of machine learning. Its core is the error back propagation algorithm. Its basic idea is to minimize the sum of squared errors between the actual calculated output value of the neural network and the expected output value by continuously adjusting the weights and thresholds of each neuron node in the neural network.

[0058] Those skilled in the art will appreciate that the neural network model used in the embodiment of the present invention can be implemented using a BP neural network tool in an existing software application environment, such as a BP neural network toolbox of MATLAB.

[0059] S3, train the neural network model using the data in the training set.

[0060] See also Figure 3 In the environment of the BP neural network model, the weights are first initialized. Given the input vector and the target output, the input vector is calculated layer by layer through the hidden layer to obtain the actual output. The error e between the actual output and the target output is calculated, and the local gradient is calculated to update the weights of the output layer and the hidden layer.

[0061] Determine whether the error e meets the conditions or reaches the maximum number of iterations. If so, stop training. If not, recalculate the input vector until the conditions are met or the maximum number of iterations is reached.

[0062] Specifically, in the embodiment of the present invention, first use The theoretical wavelength of the sample is calculated by the method, and the number of layers of each material in the superlattice material structure is then used as the strain influencing factor, the process parameters are used as the structural influencing factors, and the actual wavelength obtained by the actual PL test is used as the target output to train the neural network model.

[0063] First, the initial weights and thresholds of each neuron node are determined to obtain the initial neural network model; the training set data is input into the initial neural network model, and the theoretical wavelength, number of material layers, process parameters of the input layer, and the actual wavelength of the output layer are given; the data of the input layer is calculated layer by layer through the hidden layer and transmitted to the output layer. The state of each layer of neurons will affect the state of the next layer, and the actual output (i.e., the calculated wavelength) is calculated.

[0064] When the error between the calculated wavelength and the actual wavelength is less than a preset value (the preset value can be set according to the actual process production requirements), the training is stopped to obtain a trained neural network model; when the error is greater than or equal to the preset value, the weights and thresholds of each neuron node are automatically adjusted, and the data of the input layer is recalculated until the error meets the requirements.

[0065] If the training process reaches the maximum number of iterations and the error still does not meet the requirements, the training is stopped, the topology of the neural network model is adjusted, and the neural network model is trained again.

[0066] In addition, after training the neural network model, it is usually necessary to verify it using the data of the test set.

[0067] Specifically, the historical samples are divided into a training set and a test set. The training set data is used to train the neural network model in the above steps, and the test set data is used to verify the trained network model. The neural network model is used to calculate the samples in the test set. If the error between the calculated wavelength and the actual wavelength meets the requirements, the neural network model can be used to calculate new samples. If the error is too large, the above training process is repeated until the error meets the requirements.

[0068] S4, see Figure 4 , according to the structure and process parameters of the current sample The theoretical wavelength of the current sample is calculated by this method. The theoretical wavelength of the current sample, the number of material layers and the process parameters are input into the trained neural network model to obtain the calculated wavelength of the current sample.

[0069] The embodiment of the present invention uses process parameters as factors affecting the wavelength of superlattice materials, and uses a neural network model to optimize the wavelength calculation results of the sample. The calculated wavelength is closer to the actual situation than the theoretical wavelength, and the model can be continuously adjusted and optimized according to the calculation results to achieve real-time, efficient and accurate results.

[0070] In addition, the neural network model is trained with data on the theoretical wavelength, number of material layers, process parameters and actual wavelength of superlattice material samples grown on the same machine. A unique neural network model can be trained for different equipment and machines, eliminating the impact of machine differences on sample wavelength calculations, and achieving one model for one machine.

[0071] Based on the above-mentioned embodiment of the method for calculating the wavelength of superlattice materials, in order to further illustrate the technical solution of the present invention in detail, the present invention takes InAs / GaSb superlattice material as an example and proposes the following specific embodiments.

[0072] It should be noted that the following specific embodiments are not limitations of the present invention. Those skilled in the art can use the embodiments of the present invention to calculate the wavelength of other superlattice materials according to actual conditions, such as GaAs / AlGaAs, GaAs / AlAs and other III-V semiconductor compound superlattice materials; they can also select other growth process parameters of the superlattice material as influencing factors to train the neural network model according to actual process conditions; however, these more embodiments obtained based on the embodiments of the present invention do not deviate from the spirit and scope of the technical solution of the present invention.

[0073] Example 1

[0074] This embodiment considers the influence of process parameters and calculates the wavelength of InAs / GaSb superlattice material, including the following steps:

[0075] Step 1, based on The Hamiltonian matrix is ​​constructed by the method. The Hamiltonian matrix used in this embodiment is as follows:

[0076]

[0077] The diagonal matrix is ​​as follows:

[0078]

[0079] Among them, the parameters involved in the diagonal matrix can be calculated by the following formula:

[0080]

[0081] The InAs / GaSb superlattice material parameters involved in the above formula are shown in Table 1.

[0082] In Table 1, a is the lattice constant, , , is the deformation potential parameter, , is the elastic stiffness constant, is the band gap, is the corrected electron effective mass, , , is the Luttinger parameter, , , is the modified Luttinger parameter, is the optical transmission matrix parameter, is the spin-orbit separation energy, and VBO is the valence band offset.

[0083]

[0084] Table 1 InAs / GaSb superlattice material parameters

[0085] Step 2, solving the electronic Schrödinger equation under the periodic potential field to obtain the theoretical band gap and theoretical wavelength of the sample material; wherein the electronic Schrödinger equation is:

[0086]

[0087] Where m0 is the mass of free electrons, is the Hamiltonian operator, E is the energy eigenvalue, is the wave function that describes the motion state of the electron, is the reduced Planck constant, is a periodic potential field.

[0088] Step 3: Use the BP neural network toolbox in MATLAB to build a neural network model.

[0089] The input layer has 7 nodes, namely Theoretical calculated values, number of InAs layers, number of InSb (IF1) layers, number of GaSb layers, number of InSb (IF2) layers, As valve position, and Sb valve position; the hidden layer is set to 10 nodes; the output layer is set to 1 node, which is the actual wavelength value obtained by PL test; the rest of the model settings are default settings.

[0090] Step 4: Calculate the existing 5 samples according to steps S1 and S2 Theoretical wavelength value.

[0091] The material layer structure of samples 1-5 and The theoretical wavelength calculation results are shown in Table 2.

[0092]

[0093] Table 2 Structure (number of material layers) of samples 1-5 and Theoretical wavelength

[0094] Step 5, combined with the calculation results of step S4, the relevant parameters of samples 1-5 are used as input, and the actual wavelength of samples 1-5 is used as the target output to configure the neural network model.

[0095] The specific parameter configurations of samples 1-5 are shown in Table 3.

[0096]

[0097] Table 3 Input and target output data for samples 1-5

[0098] Step 6: Train the neural network model configured in step 5.

[0099] The error between the training value (calculated wavelength) and the actual value (actual wavelength) output by the neural network model was calculated. If the average error of the above five samples reached a level less than 0.5%, the training was stopped.

[0100] The training results of samples 1-5 are shown in Table 4.

[0101]

[0102] Table 4. Training results of neural network model

[0103] From Table 4, it can be calculated that the average error of the model training value is 0.456%, which meets the requirement that the error is less than 0.5%. After the training is completed, the neural network model can be used to predict the wavelength of new samples.

[0104] Step 7, predict the wavelength of the new sample 6 according to the above steps 1-6.

[0105] First, the traditional calculation of sample 6 is performed according to steps 1 and 2. Theoretical wavelength: The calculated theoretical wavelength is 11.11μm.

[0106] The number of material layers of sample 6 is InAs / InSb / GaSb / InSb=13.2 / 0.65 / 7 / 0.65.

[0107] Referring to Table 5, the relevant parameters of sample 6 are used as input, and the wavelength is predicted using the neural network model trained in step 6, and the final calculated wavelength is 10.158 μm.

[0108] The measured PL result of sample 6 is 10.15 μm, so the calculated error of sample 6 in this embodiment is only 0.08%, which is better than that of the conventional The theoretical method error was reduced by 9.38%.

[0109]

[0110] Table 5 Input data of sample 6

[0111] The results of the training samples 1-5 and the new sample 6 are summarized in Table 6 and Figure 5 shown.

[0112] in, Figure 5The red line (actual value) is the actual wavelength of the sample, the blue line (kp) is the theoretical wavelength of the sample, and the black line (bpkp) is the calculated wavelength of the sample.

[0113]

[0114] Table 6 Calculation results of samples 1-6

[0115] From Table 6 and Figure 5 It can be seen that the wavelength error calculated by this embodiment is less than 0.5%, while the traditional The error of the theoretical wavelength calculated by the method is about 9%. This embodiment greatly improves the accuracy of wavelength calculation of InAs / GaSb superlattice materials.

[0116] It can be seen from the above embodiments that the wavelength calculation method of superlattice materials proposed in the embodiments of the present invention takes into account the influence of process parameters on material strain and actual material structure, and effectively improves the traditional The problem of large error in theoretical wavelength calculated by the method has been solved, which greatly improves the accuracy of wavelength calculation of superlattice materials and provides strong support for the research and development accuracy and efficiency in the process production of superlattice materials.

[0117] The present invention also provides a superlattice material wavelength calculation system.

[0118] In some embodiments, the superlattice material wavelength calculation system includes an acquisition module for acquiring the theoretical wavelength, the number of material layers, the process parameters and the actual wavelength of the historical samples.

[0119] Specifically, the actual wavelength of the superlattice material is the wavelength obtained by an actual PL test (photoluminescence test) of the sample.

[0120] The number of material layers is the number of atomic layers of each material within one superlattice period.

[0121] In some embodiments, the process parameter selects at least one of a valve position, a source furnace temperature, and a growth time.

[0122] In some embodiments, the superlattice material wavelength calculation system also includes a partitioning module for partitioning historical samples into a training set and a test set, wherein the training set data is used to train the neural network model, and the test set data is used to verify the trained neural network model.

[0123] In some embodiments, the superlattice material wavelength calculation system also includes a determination module for determining the topological structure of the neural network model based on the data of the training set.

[0124] The neural network model includes an input layer, a hidden layer and an output layer. The input layer is composed of theoretical wavelength, number of material layers and process parameters, and the output layer is composed of actual wavelength.

[0125] In some embodiments, the superlattice material wavelength calculation system also includes a training module for training the neural network model using training set data to obtain a trained neural network model. The specific training process refers to the embodiment of the superlattice material wavelength calculation method described above.

[0126] In some embodiments, the superlattice material wavelength calculation system also includes a verification module for verifying the trained neural network model based on the data of the test set. The specific verification process refers to the embodiment of the aforementioned superlattice material wavelength calculation method.

[0127] In some embodiments, the superlattice material wavelength calculation system also includes a calculation module for using a trained neural network model to calculate the calculated wavelength of the current sample based on the input theoretical wavelength, number of material layers and process parameters of the current sample.

[0128] In some embodiments, the superlattice material wavelength calculation system further includes a theoretical wavelength calculation module for calculating the wavelength of the sample according to the material parameters of the sample. The method is used to calculate the theoretical wavelength of the sample. The specific method for calculating the theoretical wavelength refers to the embodiment of the above-mentioned method for calculating the wavelength of the superlattice material.

[0129] The present invention also provides a computer device.

[0130] The computer device includes a memory and a processor. The memory stores a computer program. When the processor executes the computer program, the steps of each embodiment of the above-mentioned superlattice material wavelength calculation method are implemented.

[0131] The embodiments described above are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for calculating the wavelength of a superlattice material, characterized in that: The following steps are involved: S1, obtain the theoretical wavelength, number of material layers, process parameters and actual wavelength of historical samples to obtain training set data; S2, creating a neural network model, the neural network model comprising an input layer, a hidden layer and an output layer, the input layer is composed of a theoretical wavelength, a number of material layers and process parameters, and the output layer is composed of an actual wavelength; S3, using the training set data to train the neural network model, the data of the input layer is transmitted to the output layer through the hidden layer calculation to obtain the calculated wavelength; When the error between the calculated wavelength and the actual wavelength is less than a preset value, the training is stopped to obtain a trained neural network model; S4, inputting the theoretical wavelength, number of material layers and process parameters of the current sample into the trained neural network model to obtain the calculated wavelength of the current sample.

2. The method for calculating the wavelength of superlattice materials according to claim 1, characterized in that: The process parameters include at least one of a valve position, a source furnace temperature, and a growth time.

3. The method for calculating the wavelength of superlattice materials according to claim 1, characterized in that: In step S1 or step S4, the theoretical wavelength is The method is calculated, which specifically includes the following steps: Construct a Hamiltonian matrix, which is: In the formula, represents the 8×8 diagonalized Hamiltonian matrix based on Luttinger-Kohn theory, and denote the decoupled 4×4 upper Hamiltonian matrix and lower Hamiltonian matrix respectively; The electronic Schrödinger equation under the periodic potential field is solved according to the Hamiltonian matrix and the material parameters of the sample to obtain the theoretical band gap and theoretical wavelength of the sample; the electronic Schrödinger equation is: Where m0 is the mass of free electrons, is the Hamiltonian operator, E is the energy eigenvalue, is the wave function that describes the motion state of the electron, is the reduced Planck constant, is a periodic potential field.

4. The method for calculating the wavelength of superlattice materials according to claim 1, characterized in that: The step S2 further comprises: The number of neuron nodes in the input layer, hidden layer and output layer is set, and the initial weight and threshold of each neuron node are determined.

5. The method for calculating the wavelength of superlattice materials according to claim 4, characterized in that: The step S3 further comprises: When the error between the calculated wavelength and the actual wavelength is greater than or equal to the preset value, the weight and threshold of each neuron node are adjusted, and the data of the input layer is recalculated.

6. The method for calculating the wavelength of superlattice materials according to claim 5, characterized in that: The step S1 further includes: dividing the historical samples into a training set and a test set to obtain training set data and test set data; The method further includes the step of verifying the trained neural network model using the test set data after step S3.

7. The method for calculating the wavelength of a superlattice material according to any one of claims 1 to 6, characterized in that: The neural network model is a BP neural network model.

8. A superlattice material wavelength calculation system, characterized in that: include: The acquisition module is used to obtain the theoretical wavelength, number of material layers, process parameters and actual wavelength of historical samples; A partitioning module, used for partitioning the data of the historical samples into a training set and a test set; A determination module, used to determine the topological structure of the neural network model according to the data of the training set; the neural network model includes an input layer, a hidden layer and an output layer, the input layer is composed of a theoretical wavelength, a number of material layers and process parameters, and the output layer is composed of an actual wavelength; A training module, used to train the neural network model using the data of the training set to obtain a trained neural network model; A verification module, used to verify the trained neural network model according to the data of the test set; The calculation module is used to use the trained neural network model to calculate the calculated wavelength of the current sample according to the input theoretical wavelength, number of material layers and process parameters of the current sample.

9. The superlattice material wavelength calculation system according to claim 8, characterized in that: Also includes: Theoretical wavelength calculation module is used to calculate the wavelength according to the material parameters of the sample. The theoretical wavelength of the sample can be calculated by this method.

10. A computer device, characterized in that: The computer device includes a memory and a processor, the memory stores a computer program, and the processor implements the steps of the superlattice material wavelength calculation method according to any one of claims 1 to 7 when executing the computer program.

Images