A crystal growth method, system, device and storage medium

Abstract

The embodiments of the present specification provide a crystal growth method, system, device and storage medium. The aforementioned crystal growth method includes: determining a parameter set of a crystal growth device in a target furnace, wherein the parameter set includes at least one target parameter, and the target parameter is a parameter in the crystal growth device that affects crystal growth; obtaining a first temperature detection value of a first temperature measuring point of the crystal growth device in a target time period of the target furnace; based on the target parameter and the first temperature detection value, determining a simulated crystal surface temperature corresponding to the first temperature detection value through a crystal surface temperature determination model.

Description

Technical Field

[0001] The present invention relates to the field of crystal preparation, and in particular to a crystal growth method, system, device and storage medium. Background Art

[0002] With the development of science and technology, crystals (for example, silicon carbide) are widely used in various optoelectronic devices and electronic devices. As the demand for crystals gradually increases, how to improve the quality of crystal growth has also become a focus of attention in the field.

[0003] During the crystal growth process, the crystal surface temperature is the core factor of the crystal growth quality, but it is difficult to observe the crystal surface temperature during the crystal growth process, resulting in the inability to accurately control the crystal surface temperature during the crystal growth process and the low crystal growth quality.

[0004] Therefore, how to improve the quality of crystal growth is a technical problem that needs to be solved urgently in this field. Summary of the invention

[0005] One of the embodiments of the present specification provides a crystal growth method, which includes: determining a parameter set of a crystal growth device in a target furnace, wherein the parameter set includes at least one target parameter, and the target parameter is a parameter in the crystal growth device that affects crystal growth; obtaining a first temperature detection value of a first temperature measurement point of the crystal growth device in a target time period of the target furnace; and determining a simulated crystal surface temperature corresponding to the first temperature detection value through a crystal surface temperature determination model based on the target parameter and the first temperature detection value.

[0006] In some embodiments, the method also includes: determining the temperature control parameters of the crystal growth equipment in the target time period of the target furnace based on the simulated crystal surface temperature; and automatically sending temperature adjustment instructions based on the temperature control parameters to adjust the crystal growth temperature of the crystal growth equipment in the target time period.

[0007] In some embodiments, the target parameters include a target first parameter and a target second parameter, the target first parameter includes a first parameter of a preset object in the crystal growth equipment in a target furnace, the first parameter reflects the physical properties of the material in the preset object related to crystal growth, the target second parameter includes a second parameter of the crystal growth equipment in the target furnace, and the second parameter includes parameters of related settings of the crystal growth equipment.

[0008] In some embodiments, determining the parameter set of the crystal growth equipment in the target furnace includes: obtaining an initial first parameter, wherein the initial first parameter reflects the first parameter of the preset object in the crystal growth equipment in the initial environment; processing the initial first parameter based on a reinforcement learning model to determine the target first parameter, and the reward value of the reinforcement learning model is related to the furnace information of the crystal growth equipment.

[0009] In some embodiments, determining the parameter set of the crystal growth equipment in the target furnace includes: obtaining a first correspondence between the furnace of the crystal growth equipment and the first parameter; determining the target first parameter based on the target furnace and the first correspondence; wherein the first correspondence is determined by the following steps: for each of a plurality of preset furnaces of the crystal growth equipment, processing the initial first parameter based on a reinforcement learning model, and determining a reference first parameter corresponding to the preset furnace, wherein the initial first parameter reflects the first parameter of the preset object in the crystal growth equipment in the initial environment, the reference first parameter is the first parameter of the crystal growth equipment in the preset furnace, and the reward value of the reinforcement learning model is related to the furnace information of the crystal growth equipment; determining the first correspondence between the furnace of the crystal growth equipment and the first parameter based on the reference first parameters corresponding to a plurality of the preset furnaces.

[0010] In some embodiments, determining the parameter set of the crystal growth equipment in the target furnace includes: obtaining the historical first parameters of the crystal growth equipment in multiple historical furnaces and the historical operating condition parameters corresponding to the historical first parameters, wherein the historical first parameters include the first parameters reflecting the preset object in the crystal growth equipment in the historical furnace; determining the target first parameters based on the historical first parameters of the multiple historical furnaces and the historical operating condition parameters.

[0011] In some embodiments, the historical operating condition parameters include a second temperature detection value of the crystal growth equipment at a second temperature measuring point of multiple historical furnaces; determining the first parameter based on the historical first parameters of multiple historical furnaces and the historical operating condition parameters includes: determining a second corresponding relationship based on the historical first parameters of multiple historical furnaces and the second temperature detection value, wherein the second corresponding relationship reflects the corresponding relationship between the temperature detection value of the second temperature measuring point of the crystal growth equipment and the first parameter; obtaining a third temperature detection value of the crystal growth equipment at the second temperature measuring point of the target furnace; and determining the target first parameter based on the third temperature detection value and the second corresponding relationship.

[0012] In some embodiments, the historical operating condition parameters include the historical growth time of the crystal growing equipment in the historical furnaces for growing crystals; the determining the target first parameter based on the historical first parameters of the multiple historical furnaces and the historical operating condition parameters includes: determining a third corresponding relationship based on the historical first parameters of the multiple historical furnaces and the historical growth time, the third corresponding relationship reflecting the corresponding relationship between the growth time of the crystal and the first parameter; obtaining the target growth time of the crystal growing equipment in the target furnace for growing crystals; and determining the target first parameter based on the target growth time and the third corresponding relationship.

[0013] In some embodiments, the type of the first parameter is determined by the following steps: obtaining multiple sets of reference data, each set of the reference data includes multiple candidate first parameters and verification information corresponding to the multiple candidate first parameters, each of the candidate first parameters reflects a material physical property of the preset object in the crystal growth equipment; processing the multiple sets of the reference data through an importance analysis model to determine the type of the first parameter, and the importance analysis model is a random forest model.

[0014] One of the embodiments of the present specification also provides a crystal growth system, the system comprising: a parameter determination module, used to determine a parameter set of a crystal growth device in a target furnace, wherein the parameter set includes at least one target parameter, and the target parameter is a parameter in the crystal growth device that affects crystal growth; a temperature detection module, used to obtain a first temperature detection value of a first temperature measurement point of the crystal growth device in a target time period of the target furnace; and a temperature determination module, used to determine a simulated crystal surface temperature corresponding to the first temperature detection value through a crystal surface temperature determination model based on the target parameter and the first temperature detection value. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] This specification will be further described in the form of exemplary embodiments, which will be described in detail by the accompanying drawings. These embodiments are not restrictive, and in these embodiments, the same number represents the same structure, wherein:

[0016] Figure 1 is a schematic diagram of an application scenario of a crystal growth system according to some embodiments of this specification;

[0017] Figure 2 is an exemplary flow chart of a crystal growth method according to some embodiments of this specification;

[0018] Figure 3 is an exemplary flow chart of determining a first parameter according to some embodiments of this specification;

[0019] Figure 4is a schematic diagram of a reinforcement learning model according to some embodiments of this specification;

[0020] Figure 5 is another exemplary flow chart of determining a first target parameter according to some embodiments of this specification;

[0021] Figure 6 is another exemplary flow chart of determining a first target parameter according to some embodiments of this specification;

[0022] Figure 7 is an exemplary module diagram of a crystal growth system according to some embodiments of the present specification. DETAILED DESCRIPTION

[0023] In order to more clearly illustrate the technical solutions of the embodiments of this specification, the following is a brief introduction to the drawings required for the description of the embodiments. Obviously, the drawings described below are only some examples or embodiments of this specification. For ordinary technicians in this field, this specification can also be applied to other similar scenarios based on these drawings without creative work. Unless it is obvious from the language environment or otherwise explained, the same reference numerals in the figures represent the same structure or operation.

[0024] It should be understood that the "system", "device", "unit" and / or "module" used herein are a method for distinguishing different components, elements, parts, portions or assemblies at different levels. However, if other words can achieve the same purpose, the words can be replaced by other expressions.

[0025] As shown in this specification and claims, unless the context clearly indicates an exception, the words "a", "an", "an" and / or "the" do not refer to the singular and may also include the plural. Generally speaking, the terms "comprise" and "include" only indicate the inclusion of the steps and elements that have been clearly identified, and these steps and elements do not constitute an exclusive list. The method or device may also include other steps or elements.

[0026] Flowcharts are used in this specification to illustrate the operations performed by the system according to the embodiments of this specification. It should be understood that the preceding or following operations are not necessarily performed precisely in order. Instead, the steps may be processed in reverse order or simultaneously. At the same time, other operations may also be added to these processes, or one or more operations may be removed from these processes.

[0027] Figure 1 It is a schematic diagram of an application scenario of a crystal growth system according to some embodiments of this specification.

[0028] like Figure 1As shown, the application scenario 100 of the crystal growth system may include a crystal growth device 110 , a temperature detection device 120 , a network 130 , a storage device 140 , and a processor 150 .

[0029] The crystal growth device 110 is a device for crystal growth. The aforementioned crystal may include but is not limited to silicon carbide, germanium single crystal, etc. The crystal growth device 110 may include but is not limited to physical vapor transport crystal growth equipment, liquid phase epitaxy crystal growth equipment, Czochralski crystal growth equipment, etc. The crystal growth device 110 may include structures such as a furnace body, a crucible, a thermal insulation layer, and a seed crystal holder. It is understood that any of the aforementioned crystal growth methods can determine the crystal surface temperature during the crystal growth process through the crystal growth method shown in any embodiment of this specification to complete the crystal growth.

[0030] The temperature detection device 120 is a device for detecting temperature. The temperature detection device 120 may include, but is not limited to, an infrared thermal imager, an optical fiber temperature sensor, etc. The temperature detection device 120 may be disposed in the crystal growth device 110 to obtain a first temperature detection value of a first temperature measurement point of the crystal growth device 110 in a target time period of a target furnace. For more information about the target furnace, target time period, first temperature measurement point, and first temperature detection value, please refer to Figure 2 and its related description.

[0031] The network 130 can connect the components of the application scenario 100 of the crystal growth system and / or connect the external resource part. The network 130 enables communication between the components and with other parts to promote the exchange of data and / or information. For example, the processor 150 can obtain the first temperature detection value detected by the temperature detection device 120 through the network 130. For another example, the processor 150 can also control the crystal growth device 110 through the network 130. The network 130 can be any one or more of a wired network or a wireless network. For example, the network 130 can include a cable network, an optical fiber network, a telecommunications network, the Internet, a local area network, a wide area network, a wireless local area network, a metropolitan area network, a public switched telephone network, a Bluetooth network, a purple bee network, a near field communication, a bus within the device, a line within the device, a cable connection, etc. or any combination thereof. The network connection between the various parts can be in one of the above-mentioned ways, or in multiple ways. The network can be a point-to-point, shared, central, and other topological structures or a combination of multiple topological structures.

[0032] The storage device 140 can be used to store data and / or instructions. For example, the storage device 140 can be used to store historical data, device information, etc. of the crystal growth device 110. The storage device 140 may include one or more storage components, each of which may be an independent device or a part of other devices. The storage device 140 may include a random access memory, a read-only memory, a large capacity memory, a removable memory, a volatile read-write memory, etc. or any combination thereof. Exemplarily, the large capacity storage may include a magnetic disk, an optical disk, a solid-state disk, etc. In some embodiments, the storage device 140 may be implemented on a cloud platform.

[0033] The processor 150 can process data and / or information obtained from the components of the application scenario 100 of the external resource or crystal growth system. For example, the processor 150 can determine the parameter set of the crystal growth device 110 in the target furnace. The processor can execute program instructions based on these data, information and / or processing results to perform one or more functions described in this specification. For example, the processor 150 can determine the simulated crystal surface temperature corresponding to the first temperature detection value through the crystal surface temperature determination model based on the target parameter and the first temperature detection value. For another example, the processor 150 can also determine the temperature control parameters of the crystal growth device 110 in the target time period of the target furnace based on the simulated crystal surface temperature. For more information about the above examples, please refer to the following of this specification. The processor 150 may include one or more sub-processing devices (for example, a single-core processing device or a multi-core multi-core processing device). As an example only, the processor 150 may include a central processing unit, an application-specific integrated circuit, a dedicated instruction processor, a graphics processor, a physical processor, a digital signal processor, a field programmable gate array, an editable logic circuit, a controller, a microcontroller unit, a reduced instruction set computer, a microprocessor, etc. or any combination thereof.

[0034] It should be noted that the application scenario 100 of the crystal growth system is provided only for the purpose of illustration and is not intended to limit the scope of this specification. For those of ordinary skill in the art, various modifications or changes can be made according to the description of this specification. For example, the application scenario 100 of the crystal growth system can also include a database, an information source, etc. For another example, the application scenario 100 of the crystal growth system can be implemented on other devices to achieve similar or different functions. However, changes and modifications will not deviate from the scope of this specification.

[0035] Figure 2 200 is an exemplary flow chart of a crystal growth method according to some embodiments of the present specification. In some embodiments, process 200 may be performed by a crystal growth system 700. Figure 2 As shown, process 200 includes the following steps:

[0036] Step 210 , determining a parameter set of the crystal growth equipment in the target heat. In some embodiments, step 210 may be performed by parameter determination module 710 .

[0037] The target heat refers to the crystal growth equipment (e.g. Figure 1 The target crystal growth device 110 is currently performing a crystal growth furnace. For example, the target furnace may be the 20th furnace, indicating that the crystal growth device is currently in the process of growing the 20th furnace of crystals.

[0038] The parameter set is a collection of target parameters of the crystal growth equipment in a target furnace, and the parameter set may include at least one target parameter.

[0039] The target parameter is a parameter in the crystal growth device that affects crystal growth. Specifically, the target parameter may affect the speed of crystal growth, the quality of the grown crystal, etc. In some embodiments, the target parameter may include a target first parameter and a target second parameter.

[0040] The target first parameter may include a first parameter of a preset object in the crystal growth device at a target furnace. The preset object may include at least one of a crucible, a heat-insulating felt (e.g., soft felt and / or hard felt), a raw material melt, and a protective gas (e.g., argon). The first parameter may reflect the physical properties of the material in the preset object related to crystal growth.

[0041] In some embodiments, the parameter determination module 710 can determine the type of the first parameter. It is understandable that the material physical properties of the preset object include parameters of various different types. Exemplarily, the material physical properties of the preset object may include parameters in terms of density, optical properties, mechanical properties, thermal properties, electrical properties, and magnetic properties of the preset object. However, when some material physical properties (e.g., mechanical properties) of the preset object change, the influence on the crystal surface temperature is greater; when some material physical properties (e.g., thermal properties) of the preset object change, the influence on the crystal surface temperature is smaller. In some embodiments, the parameter determination module 710 can determine the parameters of the material physical properties of the preset object that have a greater influence on the crystal surface temperature when the material physical properties change as the first parameter, and detect the target first parameter in the target furnace of the crystal growth equipment through the aforementioned multiple embodiments, so as to reduce the amount of calculation and avoid the waste of calculation resources. In some embodiments, the first parameter may include at least one of the parameters in terms of thermal properties, electrical properties, and magnetic properties of the preset object. Exemplarily, the first parameter may include at least one of thermal conductivity, electrical conductivity, surface emissivity, thermal conductivity, constant pressure heat capacity, and relative magnetic permeability.

[0042] In some embodiments, the parameter determination module 710 can obtain multiple groups of reference data. The reference data can be data used to determine the type of the first parameter. Each group of reference data may include multiple candidate first parameters and verification information corresponding to the aforementioned multiple candidate first parameters. The candidate first parameter can be a parameter of the first parameter to be evaluated for determination, and the candidate first parameter can include all material physical properties of the preset object or some user-specified material physical properties. Each candidate first parameter can reflect a material physical property of the preset object in the crystal growth device. Verification information is information used to verify the importance of the first parameter. For example, verification information may include but is not limited to furnace temperature, furnace pressure, crystal growth quality, etc. Reference data can be obtained through historical data of the crystal growth device. For example, multiple candidate first parameters and verification information in a certain group of reference data can be obtained by measuring the preset object of the crystal growth device in historical furnaces.

[0043] In some embodiments, the parameter determination module 710 can process multiple sets of reference data through an importance analysis model to determine the type of the first parameter, and the importance analysis model is a random forest model. For example, the parameter determination module 710 can process multiple sets of reference data through an importance analysis model to determine four types of first parameters from the material physical properties of 20 preset objects, namely surface emissivity, thermal conductivity, constant pressure heat capacity, and relative magnetic permeability.

[0044] The importance analysis model can analyze the multiple candidate first parameters input through the input verification information, determine the importance of each candidate first parameter, and thus determine the type of the first parameter. The importance of each of the aforementioned candidate first parameters can be expressed in a variety of ways, such as numerical values, ratios, etc. Exemplarily, the parameter determination module 710 can determine the type of the candidate first parameter whose importance exceeds a preset importance threshold as the type of the first parameter. Some embodiments of the present specification screen multiple candidate first parameters through a random forest model to determine the type of the first parameter, avoid interference from the physical properties of the material in the preset object that has little effect on crystal growth, and reduce the amount of calculation of the crystal growth system 700.

[0045] In some embodiments, the parameter determination module 710 may determine the target first parameter of the crystal growth device in the target batch based on the initial first parameter corresponding to the preset object. For example, the parameter determination module 710 may determine the target first parameter of the crystal growth device in the target batch based on the device information of the crystal growth device. The aforementioned device information may include relevant information of the crystal growth device, and the device information may include but is not limited to the type of crystal growth device, the overall size, the size of the preset object, the initial first parameter, etc. The device information may be provided by the manufacturer of the crystal growth device. For more information about the initial first parameter, see Figure 4and its related description.

[0046] In some embodiments, the parameter determination module 710 can determine the specific value of the target first parameter. It is worth noting that when in a room temperature state, the first parameter is relatively stable, for example, the conductivity can be a fixed value. However, when the crystal grows in a crystal growth device, the crystal growth device will induction heat the crucible through an induction coil, so the temperature of the crystal growth environment is relatively high (for example, about 900 to 7200°C), and is located in an electric field and a magnetic field. At this time, the first parameter may change according to the different environments in the furnace of the crystal growth device. In addition, when the crystal growth device is in use, the material properties of the preset object may also change with use, resulting in changes in the first parameter. Based on this, in order to more accurately determine the crystal surface temperature of the crystal during the growth process, the parameter determination module 710 can determine the target first parameter of the crystal growth device in the target furnace by a variety of methods, that is, the specific value of the first parameter in the target furnace.

[0047] In some embodiments, the parameter determination module 710 can obtain an initial first parameter, wherein the initial first parameter reflects the first parameter of the preset object in the crystal growth device in the initial environment; the initial first parameter is processed based on the reinforcement learning model to determine the target first parameter, and the reward value of the reinforcement learning model is related to the furnace information of the crystal growth device. For more information about the above embodiments, please refer to Figure 3 and its related description.

[0048] In some embodiments, the parameter determination module 710 may also obtain a first correspondence between the furnace of the crystal growth device and the first parameter; and determine the target first parameter based on the target furnace and the first correspondence. Figure 4 and its related description.

[0049] In some embodiments, the parameter determination module 710 can also obtain the historical first parameters of the crystal growth equipment in multiple historical furnaces and the historical operating parameters corresponding to the historical first parameters, wherein the historical first parameters include the first parameters reflecting the preset objects in the crystal growth equipment in the historical furnaces; based on the historical first parameters of the multiple historical furnaces and the historical operating parameters, determine the target first parameter. For more information about the above embodiments, please refer to Figure 5 and its related description.

[0050] The target second parameter may include a second parameter of the crystal growth equipment in the target heat. The second parameter may include a parameter of the crystal growth equipment related to the target heat.

[0051] In some embodiments, the parameter determination module 710 can determine the type of the second parameter by preset, and further determine the number of the target second parameter in the target furnace. The second parameter may include furnace equipment parameters and furnace growth parameters. The aforementioned furnace equipment parameters may be parameters related to crystal growth equipment. Exemplarily, furnace equipment parameters may include but are not limited to the number of top insulation layers, the distance between the top insulation bottom and the top of the crucible, the number of soft felts, the minimum inner diameter, the maximum outer diameter, the cover thickness, the material distance, the inner diameter of the gasket, the material height, the bottom insulation soft felt, the insulation inner tube hard felt, the insulation inner tube soft felt, the insulation outer tube hard felt, the insulation outer tube soft felt, the outermost layer of soft felt, the ring height, etc. The furnace growth parameters may be parameters related to the crystal growth setting. For example, the furnace growth parameters may include growth pressure, growth power, high temperature temperature, high temperature line, crystal growth time, center thickness, edge thickness, etc. Different furnaces of crystal growth equipment may cause the corresponding second parameters to be different due to differences in the generated crystal type, crystal demand, etc.

[0052] In some embodiments, the target second parameter can be obtained by presetting the target furnace. For example, the target second parameter is determined by inputting furnace equipment parameters and furnace growth parameters of the target furnace by a user (e.g., a manager of a crystal growth device). For another example, the furnace equipment parameters of the crystal growth device in the target furnace can be determined based on a pre-set crystal growth device type. For another example, the furnace growth parameters of the crystal growth device in the target furnace can be determined based on a pre-set crystal type, crystal demand, etc.

[0053] Step 220 , obtaining a first temperature detection value of a first temperature measurement point of the crystal growth device in a target time period of the target furnace. In some embodiments, step 220 may be performed by the temperature detection module 720 .

[0054] The target time period refers to the time period for measuring the temperature of the first temperature measurement point. The target time period can be preset by a user (e.g., an operator of a crystal growth device). The target heat can include multiple time periods, and the target time period can be the current time period.

[0055] The temperature detection module 720 can perform at least one temperature detection within the target time period to obtain a first temperature detection value of a first temperature measurement point of the crystal growth device. The temperature detection module 720 can perform at least one temperature detection within the target time period to obtain a first temperature detection value of a first temperature measurement point of the crystal growth device. Figure 1 The temperature detection device 120 shown is implemented.

[0056] The first temperature measurement point is a temperature detection point in the crystal growth device (for example, in the furnace of the crystal growth device), and the first temperature detection value is the temperature of the first temperature measurement point obtained by detection.

[0057] The crystal growth device may include one or more first temperature measurement points. The first temperature measurement point may be preset by the user. The user may set a temperature detection device at one or more locations, and the temperature detection module 720 may obtain a first temperature detection value of the first temperature measurement point through the aforementioned temperature detection device during a target time period of a target furnace.

[0058] Step 230 , based on the target parameter and the first temperature detection value, determine the simulated crystal surface temperature corresponding to the first temperature detection value through the crystal surface temperature determination model. In some embodiments, step 230 may be performed by the temperature determination module 730 .

[0059] The simulated crystal surface temperature refers to a crystal surface temperature corresponding to the first temperature detection value determined through simulation.

[0060] The crystal surface temperature determination model can be a numerical simulation model. The processor 150 can establish multiple mathematical equations based on the historical target parameters of multiple historical furnaces, the historical first temperature detection chamber, and the historical crystal surface temperature, and discretize the multiple mathematical equations, convert them into discrete forms for processing, and then use numerical calculation methods, such as finite difference method, finite element method, etc., to solve the aforementioned discretized mathematical equations to obtain the crystal surface temperature determination model. The temperature determination module 730 can input the target parameters and the first temperature detection value into the crystal surface temperature determination model, and the crystal surface temperature determination model performs simulation based on the input data to determine the simulated crystal surface temperature.

[0061] Some embodiments of the present specification can use a crystal surface temperature determination model to simulate and determine the simulated crystal surface temperature, thereby solving the problem that the crystal surface temperature is difficult to observe.

[0062] In some embodiments, the crystal plane temperature determination model can also be a machine learning model. For example, the crystal plane temperature determination model can include but is not limited to one or a combination of a logistic regression model, a decision tree model, a stochastic gradient descent model, etc. The input of the crystal plane temperature determination model can include a target parameter and a first temperature detection value, and the output can include a simulated crystal plane temperature corresponding to the first temperature detection value. The aforementioned crystal plane temperature determination model can be obtained by training after modeling the crystal growth equipment through numerical simulation software (for example, Virtual reactor, COMSOL, etc.) or a designed heat transfer calculation program to obtain an initial crystal plane temperature determination model. The first training sample can include a sample target parameter and a sample first temperature detection value, and the first training label can include a sample crystal plane temperature. The aforementioned first training sample and the first training label can be obtained by heating the crystal growth equipment.

[0063] In some embodiments, the temperature determination module 730 may also determine a corresponding crystal surface temperature determination model based on the crystal growth equipment type. The crystal surface temperature determination model corresponding to each crystal growth equipment type may be obtained by training based on the first training sample and the first training label corresponding to each crystal growth equipment type.

[0064] In some embodiments, the input of the crystal surface temperature determination model may also include the position information of the first temperature measurement point. Correspondingly, the first training sample during training may also include the sample position information of the first temperature measurement point. By also using the position information of the first temperature measurement point as the input of the crystal surface temperature determination model, the error caused by the position difference of the first temperature measurement point can be avoided, and the accuracy of the output of the crystal surface temperature determination model can be improved.

[0065] In some embodiments, the crystal surface temperature determination model may also include a numerical simulation model and a machine learning model. For example, the temperature determination module 730 may process the target parameter and the first temperature detection value based on the numerical simulation model and the machine learning model, and perform weighted summation on the results output by the two, and determine the result of the weighted summation as the simulated crystal surface temperature. Among them, the aforementioned numerical simulation model and machine learning model can be obtained by preset. Numerical simulation models are usually affected by uncertainties from different sources, such as model parameter uncertainty, noise in input data, etc. Machine learning models can be applied to estimate and model uncertainty to provide more accurate prediction results and effective uncertainty quantification. By combining the output results of the two, a more accurate simulated crystal surface temperature can be obtained.

[0066] Through some embodiments of this specification, the parameter set of the crystal growth equipment in the target furnace can be determined more accurately, so as to more accurately determine the crystal surface temperature during the crystal growth process and improve the growth quality of the crystal. In addition, by determining the crystal surface temperature through the crystal surface temperature determination model, the judgment efficiency can be improved and the cost of manual judgment can be reduced.

[0067] In some embodiments, after determining the simulated crystal surface temperature, process 200 may also determine a temperature control parameter for adjusting the temperature of the crystal growth device in the target time period. Optionally, process 200 may also include the following steps:

[0068] Step 240 , based on the simulated crystal surface temperature, determine the temperature control parameters of the crystal growth equipment in the target time period of the target furnace. In some embodiments, step 240 may be performed by the temperature control module 740 .

[0069] The temperature control parameter is a parameter used by the crystal growth device to adjust the temperature in the furnace. The temperature control parameter may include the power of the furnace in the crystal growth device. For example, the temperature control parameter may be +5KW, indicating that the current power of the furnace needs to be increased by 5KW within the target time period. The temperature control parameter may also include the adjustment value of the temperature in the furnace in the crystal growth device. For example, the temperature control parameter may be +100°C, indicating that the current temperature in the furnace needs to be increased by 100°C within the target time period.

[0070] In some embodiments, the temperature control module 740 may feed back the simulated crystal surface temperature to the operator of the target furnace, and the operator may determine the temperature control parameters of the crystal growth equipment in the target time period of the target furnace according to the simulated crystal surface temperature.

[0071] In some embodiments, the temperature control module 740 can determine the temperature control parameters of the crystal growth equipment in the target time period of the target furnace based on the crystal growth information of the target furnace and the simulated crystal surface temperature through the temperature control relationship. Crystal growth information can refer to relevant information of crystal growth of the target furnace. For example, the crystal growth information can include but is not limited to crystal type, current crystal growth situation (such as size), current crystal growth time, target growth time, etc. The aforementioned target growth time refers to the estimated time required for crystal growth in the target furnace. The crystal growth information can be directly determined by user input, or it can be obtained in real time during the crystal growth process. For example, the crystal type, target growth time, etc. can be directly determined by user input; for another example, the current crystal growth situation (such as crystal size) and the current crystal growth time can be obtained in real time during the crystal growth process.

[0072] The temperature control module 740 can determine the crystal surface temperature required by the crystal in the target time period according to the crystal growth information, and determine the temperature control parameters of the crystal growth device in the target time period in combination with the simulated crystal surface temperature and the preset control rule relationship. For more information about the temperature control relationship, please refer to the relevant description below in this specification.

[0073] The temperature control module 740 can obtain multiple reference crystal surface temperatures of each reference device in a reference furnace among multiple reference devices and multiple reference control parameters corresponding to the multiple reference crystal surface temperatures. Among them, the reference device is a device of the same type as the crystal growth device, and the reference furnace is a crystal growth furnace performed by the reference device. The reference crystal surface temperature can refer to the crystal surface temperature of the reference device in the reference furnace, and the aforementioned reference crystal surface temperature can be a temperature sequence, characterizing the different crystal surface temperatures of the reference device in multiple time periods of the reference furnace. The reference control parameter can refer to the temperature control parameter of the reference device in the reference furnace, and the aforementioned reference crystal surface temperature can be a temperature control parameter sequence, characterizing the different temperature control parameters of the reference device in multiple time periods of the reference furnace.

[0074] In some embodiments, the temperature control module 740 may determine the temperature control parameters based on reference crystal surface temperatures of a plurality of reference devices in reference furnaces, reference control parameters, and simulated crystal surface temperatures.

[0075] For example, the temperature control module 740 can determine at least one reference furnace based on the simulated crystal surface temperature and the reference crystal surface temperature, and then determine the temperature control parameter of the crystal growth device in the target time period of the target furnace based on the reference control parameter corresponding to the aforementioned at least one reference furnace. When the number of reference furnaces is one, the temperature control module 740 can directly determine the reference control parameter corresponding to the reference furnace as the temperature control parameter of the crystal growth device in the target time period of the target furnace. It can be understood that in different reference furnaces, the reference control parameter corresponding to the same reference crystal surface temperature may be different. Therefore, when the number of reference furnaces is multiple, the temperature control module 740 can comprehensively evaluate the reference control parameters of multiple reference furnaces to determine the temperature control parameter of the crystal growth device in the target time period of the target furnace. For example, the average value or median value of the reference control parameters corresponding to multiple reference furnaces is determined, and it is determined as the temperature control parameter of the crystal growth device in the target time period of the target furnace.

[0076] For example, the temperature control module 740 can determine the temperature control relationship between the temperature control parameters in the crystal growth equipment and the crystal surface temperature based on the reference crystal surface temperature and reference control parameters of multiple reference equipment in the reference furnace. The temperature control relationship can characterize the effect of the adjustment of the temperature control parameters on the crystal surface temperature.

[0077] In some embodiments, the temperature control relationship can be characterized by a temperature control model. The input of the temperature control model can include the simulated crystal surface temperature from the start of crystal growth to the target time period, and the output is the temperature control parameter from the target time period to the completion of crystal growth.

[0078] The temperature control model can be obtained by training the reference crystal surface temperature and reference control parameters of the reference furnace by multiple reference devices. The temperature control module 740 can block or remove the crystal surface temperature of a part of the time period in the reference crystal surface temperature to obtain a second training sample. The second training label can be the reference control parameter corresponding to the aforementioned blocked or removed time period. The temperature control module 740 can train the initial temperature control model based on the second training sample and the second training label to obtain the temperature control model.

[0079] The temperature control module 740 can determine the crystal surface temperature required by the crystal growth equipment in the target time period of the target furnace based on the crystal growth information of the target furnace, and determine the temperature control parameters of the crystal growth equipment in the target time period of the target furnace through the temperature control relationship based on the crystal surface temperature required by the crystal growth equipment in the target time period of the target furnace and the simulated crystal surface temperature, so that the grown crystals meet the requirements.

[0080] The temperature control module 740 can also obtain crystal detection data of crystals generated by multiple reference devices in reference furnaces. The crystal detection data can refer to relevant detection data of the crystal. For example, the crystal detection data can include but is not limited to crystal size, crystal quantity, whether there is a crystal, etc.

[0081] In some embodiments, the temperature control module 740 can determine the temperature control parameters based on the reference crystal surface temperature, reference control parameters, crystal detection data and simulated crystal surface temperature of multiple reference devices in the reference furnace. The temperature control module 740 can screen the reference crystal surface temperature and reference control parameters of multiple reference devices in the reference furnace based on the crystal growth information of the target furnace and the crystal detection data of multiple reference devices in the reference furnace, obtain the screened reference crystal surface temperature and reference control parameters, and determine the temperature control parameters of the crystal growth device in the target time period of the target furnace based on the screened reference crystal surface temperature, reference control parameters and simulated crystal surface temperature. For example, the temperature control module 740 can only retain the reference crystal surface temperature and reference control parameters corresponding to the reference furnace whose crystal size in the crystal detection data is equal to or greater than the preset size threshold. By screening the reference furnaces of the reference equipment through the crystal detection data, data that does not meet the requirements can be screened out, so that the temperature control parameters of the target furnace in the target time period are more in line with the requirements of the crystal growth information.

[0082] It is worth noting that if the simulated crystal surface temperature is only fed back to the operator as a reference, the operator is still required to set the temperature of the crystal growth furnace based on experience, which makes the monitoring of the crystal surface temperature still not comprehensive enough, and cannot ensure that the temperature of the crystal surface during the crystal growth process is constant, which may reduce the quality of crystal growth. Some embodiments of this specification determine the temperature control parameters through the temperature control module 740, which can make the determined temperature control parameters more suitable for crystal growth, realize automatic control, reduce labor costs, and reduce the operation error rate.

[0083] In some embodiments, the crystal growth system 200 may also adjust the crystal growth temperature of the crystal growth device in the target time period based on the temperature control parameter. Optionally, the process 200 may also include the following steps:

[0084] Step 250 , based on the temperature control parameters, automatically sends a temperature adjustment instruction to adjust the crystal growth temperature of the crystal growth equipment in the target time period of the target furnace. In some embodiments, step 250 may be performed by the temperature adjustment module 750 .

[0085] The temperature adjustment instruction is a control instruction for each component in the crystal growth device to adjust the temperature. The temperature adjustment module 750 can generate a corresponding temperature adjustment instruction based on the temperature control parameter, and send the temperature adjustment instruction to the corresponding component of the crystal growth device to adjust the crystal growth temperature of the crystal growth device in the target time period of the target furnace, so that the crystal growth temperature meets the requirements of crystal growth and improves the quality of crystal growth.

[0086] Exemplarily, when the temperature control parameter includes the power of the furnace in the crystal growth equipment +5KW, the temperature adjustment module 750 can convert it into a corresponding temperature adjustment instruction, and the aforementioned temperature adjustment instruction can be sent to the furnace; the furnace can increase the current power by 5KW in response to the temperature adjustment instruction.

[0087] Some embodiments of the present specification can automatically control the temperature of crystal growth equipment through temperature control parameters, thereby realizing intelligent control of crystal growth and reducing the cost caused by manual operation.

[0088] It is worth noting that in the target furnace of the crystal growth equipment, the crystal growth system 700 can continuously execute process 200 to determine the crystal surface temperature in each time period, thereby realizing intelligent control of the crystal growth equipment based on the aforementioned crystal surface temperature and improving the quality of crystal growth.

[0089] It should be noted that the above description of process 200 is only for example and illustration, and does not limit the scope of application of this specification. For those skilled in the art, various modifications and changes can be made to process 200 under the guidance of this specification. However, these modifications and changes are still within the scope of this specification. For example, the crystal growth system 700 can first perform step 220 and then perform step 210 to achieve the crystal growth method described in the aforementioned embodiment of this specification.

[0090] Figure 3 is an exemplary flow chart of determining the first parameter according to some embodiments of this specification. In some embodiments, process 300 may be executed by parameter determination module 710. Figure 3 As shown, process 300 may include the following steps:

[0091] Step 310, obtaining an initial first parameter.

[0092] The initial first parameter may reflect the first parameter of the preset object in the crystal growth device in the initial environment. The initial environment may refer to the environment when the manufacturer of the crystal growth device measures the first parameter of the preset object. At this time, the crystal growth device has not yet performed crystal growth. One or more of the temperature, magnetic field, electric field, etc. in the initial environment may be determined by preset. The initial first parameter may be determined by the device information of the crystal growth device. For example, the designer of the crystal growth device may test the preset object in the initial environment (e.g., the temperature is 20 to 30°C and the magnetic field is 1 to 2 milligauss) when producing the crystal growth device to determine the first parameter of the preset object under the initial conditions.

[0093] It is understandable that, as the crystal is generated, the environment in which the preset object is located will change relative to the initial environment (for example, the temperature in the initial environment is 20-30°C, and the temperature during crystal growth is 7400-7500°C). In addition, as the crystal growth equipment is used, the internal structure of the preset object may also change, thereby causing the first parameter of the preset object in the crystal growth equipment to continue to change.

[0094] Step 320: Process the initial first parameter based on the reinforcement learning model to determine the target first parameter.

[0095] For each initial first parameter, the parameter determination module 710 can input the initial first parameter into the reinforcement learning model, and the output of the reinforcement learning model is the corresponding target first parameter. Among them, the reward value of the aforementioned reinforcement learning model can be related to the furnace information of the crystal growth equipment. The aforementioned furnace information can be relevant information of the crystal growth equipment during the crystal growth process, including but not limited to the furnace temperature, furnace pressure, exhaust rate, etc. The aforementioned furnace information can be determined based on a variety of methods. The aforementioned furnace information can be determined based on detection. For example, the parameter determination module 710 can detect through a related detection device to obtain the furnace information. The aforementioned furnace information can also be determined based on calculation. For example, the parameter determination module 710 can also determine the furnace temperature based on the first temperature detection value, through the preset furnace temperature and the temperature correspondence between the first temperature detection value. For another example, the parameter determination module 710 can also determine the furnace temperature through calculation based on the target furnace and the power of the crystal growth equipment.

[0096] The reinforcement learning model can be used to modify the initial first parameter to obtain the modified target first parameter. The reinforcement learning model 420 includes an adjustment module 421 and an optimal action determination module 422 .

[0097] like Figure 4As shown, the parameter determination module 710 can input the initial first parameter 410 into the reinforcement learning model 420, and the output of the reinforcement learning model 420 is the target first parameter 430. In the reinforcement learning model 420, the initial first parameter 410 is input into the adjustment module 421, and the adjustment module 421 outputs the optional action set; then the initial first parameter 410 and the optional action set are input into the optimal action determination module 422, and the optimal action determination module 422 can output the optimal optional action 423. The parameter determination module 710 can use the value corresponding to the optimal optional action 423 output by the optimal action determination module 422 as the output of the reinforcement learning model 420, that is, the target first parameter 430.

[0098] The adjustment module 421 may include an optional action determination submodule 421-1, a state determination submodule 421-2, and a reward determination submodule 421-3. In the prediction process of the reinforcement learning model 420, the adjustment module may determine the optional action set through the optional action determination submodule 421-1 based on the initial first parameter 410. It is worth noting that the target first parameter 430 finally determined by adjusting the initial first parameter 410 of the optional action set should not exceed the correction range of the initial first parameter 410. The aforementioned correction range refers to the range for correcting the initial first parameter. It is understandable that because the types of the various initial first parameters are different, the correction range corresponding to each initial first parameter may be different. The aforementioned correction range can be preset based on experience.

[0099] During the training process of the reinforcement learning model 420, the state determination submodule 421-2 and the reward determination submodule 421-3 in the adjustment module 421 can be used to determine the first parameter of the next state and the reward value, respectively.

[0100] The optional action determination submodule 421-1 can determine the optional action set in the current state based on the first parameter of the current state (for example, during initial execution, the first parameter of the current state can be the initial first parameter 410). The optional action set refers to a set of actions that can be performed on the first parameter of the current state in a certain state. It is worth noting that the adjusted first parameter should not exceed the correction range corresponding to the initial first parameter.

[0101] The state determination submodule 421 - 2 may determine the first parameter of the next state based on the first parameter of the current state and the optimal optional action 423 output by the optimal action determination module 422 .

[0102] The reward determination submodule 421-3 can be used to determine the reward value. The reward value can be used to determine the accuracy of the adjusted first parameter. For example, for actions with high accuracy, the reward value can be higher; for actions with low accuracy or negative improvement, the reward value can be lower. The reward value can be expressed in numerical or other ways. In some embodiments, the reward value of the reinforcement learning model can be related to the furnace information of the crystal growth equipment. Taking the first parameter as conductivity and the furnace information as the furnace temperature as an example, the parameter determination module 710 can input the adjusted first parameter into the simulation model, and the output of the simulation model can be the simulation temperature of the crystal growth equipment corresponding to the current adjusted first parameter, and the aforementioned simulation temperature is compared with the furnace temperature of the crystal growth equipment to determine the reward value of the adjusted first parameter. Among them, the aforementioned reward value can be determined based on the temperature difference between the simulation temperature and the furnace temperature of the crystal growth equipment and the preset reward rule. The smaller the aforementioned temperature difference, the greater the reward value. The simulation model can be obtained by training with historical data of the crystal growth equipment.

[0103] The optimal action determination module 422 can determine the optimal optional action 423 based on the first parameter of the current state and the optional action set. In some embodiments, the optimal action determination module 422 can be a machine learning model, which can be implemented in various ways, such as a deep neural network (DNN), a convolutional neural network (CNN), a recurrent neural network (RNN), etc.

[0104] In some embodiments, the optimal action determination module 422 can be obtained by training multiple groups of third training samples with third training labels based on a reinforcement learning method, for example, a deep Q-Learning Network (DQN), a double deep Q-Learning Network (DDQN), etc. The third training sample can be a historical first parameter of the crystal growth device, and the third training label is the optimal optional action corresponding to the historical first parameter. The third training sample can be obtained based on the historical data of the crystal growth device, and the third training label can be obtained by a reinforcement learning method.

[0105] In some embodiments, the parameter determination module 710 may periodically execute the reinforcement learning model 420 based on a preset trigger condition and output the optimal optional action 423. For example, the preset trigger condition is that the crystal growth device is executed once per furnace, and the parameter determination module 710 may determine the optimal optional action 423 output by the reinforcement learning model 420 in the target furnace as the target first parameter of the target furnace.

[0106] Some embodiments of the present specification process the initial first parameter through a reinforcement learning model, so as to more accurately determine the target first parameter and improve the accuracy of the determined simulated crystal surface temperature.

[0107] Figure 5 FIG. 5 is another exemplary flow chart of determining a target first parameter according to some embodiments of the present specification. In some embodiments, process 500 may be executed by parameter determination module 710. Figure 5 As shown, process 500 may include the following steps:

[0108] Step 510, obtaining historical first parameters of the crystal growth equipment in multiple historical furnaces and historical operating condition parameters corresponding to the historical first parameters.

[0109] The historical heat refers to the heat of crystal growth performed by the crystal growth device before the target heat. Correspondingly, the historical first parameter refers to the first parameter of the crystal growth device in the historical heat.

[0110] Historical operating parameters refer to parameters related to the operating conditions of the crystal growth equipment in historical furnaces.

[0111] In some embodiments, the historical operating condition parameter may include a second temperature detection value of a second temperature measurement point of a crystal growth device in multiple historical furnaces. The second temperature measurement point may refer to a position where the crystal growth device detects the temperature in the furnace in multiple historical furnaces, and the second temperature detection value may refer to a result obtained by the crystal growth device detecting the temperature at the second temperature measurement point in multiple historical furnaces. The second temperature measurement point may be the same as or different from the position of the aforementioned first temperature measurement point.

[0112] In some embodiments, the historical operating condition parameters may also include the historical growth time of the crystal growing equipment in the historical furnaces for growing crystals. The historical growth time refers to the time taken by the crystal growing equipment to grow crystals in multiple historical furnaces.

[0113] In some embodiments, the historical operating condition parameters may also include other parameters. For example, the historical operating condition parameters may also include but are not limited to the furnace power, furnace pressure, exhaust rate, etc. of historical furnaces.

[0114] The parameter determination module 710 can obtain the historical first parameters of the crystal growth equipment in multiple historical furnaces and the historical operating parameters corresponding to the historical first parameters through the historical data of the crystal growth equipment. Exemplarily, the crystal growth equipment can also include a storage module, and the crystal growth equipment can detect relevant data each time it performs crystal growth, for example, detect the historical first parameters and the historical operating parameters, and store them in the aforementioned storage module as historical data. The parameter determination module 710 can obtain the historical first parameters of the crystal growth equipment in multiple historical furnaces and the historical operating parameters corresponding to the historical first parameters from the aforementioned storage module when necessary.

[0115] Step 520, determining a target first parameter based on historical first parameters of multiple historical furnaces and historical operating condition parameters.

[0116] In some embodiments, when the historical operating condition parameter includes the second temperature detection value of the second temperature measuring point of the crystal growth equipment at multiple historical furnaces, the parameter determination module 710 can analyze and process the historical first parameters and the second temperature detection values ​​of the multiple historical furnaces to determine the second corresponding relationship between the temperature detection value of the second temperature measuring point and the first parameter. The aforementioned second corresponding relationship can reflect the corresponding relationship between the temperature detection value of the second temperature measuring point of the crystal growth equipment and the first parameter. The second corresponding relationship can be characterized as a fitting function, a machine learning model, etc. It can be understood that when the first parameter includes multiple types, the second corresponding relationship can include the corresponding relationship between each first parameter and the temperature detection value of the second temperature measuring point. For example, when the first parameter includes the thermal conductivity and electrical conductivity of the crucible, the aforementioned second corresponding relationship can include the corresponding relationship between the temperature detection value of the second temperature measuring point and the thermal conductivity and electrical conductivity of the crucible.

[0117] In some embodiments, the parameter determination module 710 can obtain a third temperature detection value of the second temperature measurement point of the crystal growth device in the target furnace, and determine the first parameter based on the third temperature detection value and the second corresponding relationship. The third temperature detection value refers to the result obtained by the crystal growth device performing temperature detection on the second temperature measurement point in the target furnace. It can be understood that when the first temperature measurement point is the same as the second temperature measurement point, the third temperature detection value is the first temperature detection value. The parameter determination module 710 can determine the first parameter corresponding to the third temperature detection value based on the second corresponding relationship, and determine it as the target first parameter.

[0118] In some embodiments, when the historical operating condition parameter includes the historical growth time of the crystal growing equipment in the historical furnace growth of the crystal, the parameter determination module 710 can determine the third corresponding relationship based on the historical first parameters of multiple historical furnaces and based on multiple historical related parameters and multiple historical growth times. Among them, the aforementioned third corresponding relationship can reflect the third corresponding relationship between the growth time of the crystal and the first parameter. The third corresponding relationship can be characterized as a fitting function, a machine learning model, etc. Similar to the second corresponding relationship, when the first parameter includes multiple types, the third corresponding relationship can include the corresponding relationship between each first parameter and the growth time of the crystal.

[0119] In some embodiments, the parameter determination module 710 can obtain the target growth time of the crystal growth device in the target furnace to grow the crystal; based on the target crystal growth time and the third corresponding relationship, determine the first parameter. The parameter determination module 710 can determine the first parameter corresponding to the target growth time based on the third corresponding relationship, and determine it as the target first parameter.

[0120] In some embodiments, the parameter determination module 710 can also establish a parameter prediction model based on the historical first parameters and historical operating parameters of multiple historical furnaces; determine the preset operating parameters of the crystal growth equipment in the target furnace; and determine the target first parameter based on the preset operating parameters and the parameter prediction model.

[0121] The preset operating parameters may refer to the operating parameters in the predetermined target furnace. The preset operating parameters may include but are not limited to the predetermined furnace temperature, furnace pressure, exhaust rate, and furnace power of the crystal growth equipment in each time period of the target furnace. The parameter determination module 710 or the user may plan the growth of the target furnace based on the crystal growth information, thereby determining the preset operating parameters of the crystal growth equipment in the target furnace. For more information about crystal growth information, see Figure 2 and its related description.

[0122] In some embodiments, the parameter determination module 710 may input the preset operating condition parameters of the target furnace into the parameter prediction model, and the output of the parameter prediction model is the first parameter under the preset operating condition parameters. The parameter determination module 710 may determine the first parameter under the preset operating condition parameters as the target parameter. The parameter prediction model may be a machine learning model. For example, the parameter prediction model may include but is not limited to one or more combinations of a convolutional neural network model, a deep learning model, and a random forest model. Some embodiments of this specification can quickly and accurately determine the first parameter through the parameter prediction model to improve processing efficiency.

[0123] The parameter determination module 710 may train the initial parameter prediction model based on multiple groups of fourth training samples with fourth training labels to obtain a trained parameter prediction model. For each group of fourth training samples with fourth training labels, the fourth training samples may include historical operating parameters of a certain historical furnace, and the fourth training labels may include the historical first parameters corresponding to the historical furnace.

[0124] Some embodiments of the present specification determine the first parameter by analyzing and processing the historical first parameters and historical operating parameters of multiple historical furnaces of the crystal growth equipment, thereby more accurately determining the first parameter of the crystal growth equipment in the target furnace and avoiding making judgments directly based on the initial first parameter, which may lead to errors in determining the crystal surface temperature.

[0125] Figure 6 FIG. 6 is another exemplary flow chart of determining a target first parameter according to some embodiments of the present specification. In some embodiments, process 600 may be executed by parameter determination module 710. Figure 6 As shown, process 600 may include the following steps:

[0126] Step 610: Obtain a first corresponding relationship between a furnace of a crystal growth device and a first parameter.

[0127] In some embodiments, for each of the multiple preset furnaces of the crystal growth equipment, the parameter determination module 710 can process the initial first parameter based on the reinforcement learning model to determine the reference first parameter corresponding to each preset furnace. The multiple preset furnaces can refer to the multiple furnaces pre-set by the crystal growth equipment. For example, the multiple preset furnaces can be the first 20 furnaces of the crystal growth equipment. The reference first parameter is the first parameter of the crystal growth equipment in the preset furnace. For more information about the reinforcement learning model, see Figure 3 and its related description.

[0128] The parameter determination module 710 can determine the first corresponding relationship between the furnace of the crystal growth equipment and the first parameter based on the reference first parameter corresponding to the plurality of preset furnaces. Similar to the second corresponding relationship and the third corresponding relationship, the first corresponding relationship can also be characterized by a fitting function or a machine learning model. For example, the parameter determination module 710 can perform data fitting on the reference first parameter corresponding to the plurality of preset furnaces, determine the fitting function between the furnace of the crystal growth equipment and the first parameter, and determine the aforementioned fitting function as the first corresponding relationship. For another example, the parameter determination module 710 determines the preset furnace as the fifth training sample, determines the reference first parameter corresponding to the preset furnace as the fifth training label, performs training based on the aforementioned fifth training sample and the fifth training label, determines the machine learning model, and determines the aforementioned machine learning model as the first corresponding relationship.

[0129] Step 620, determining a target first parameter based on the target heat and the first corresponding relationship.

[0130] For example, when the first correspondence is a fitting function, the parameter determination module 710 can determine the target first parameter by calculation based on the target heat. For another example, when the first correspondence is a machine learning model, the parameter determination module 710 can input the target heat into the aforementioned machine learning model, and the output of the aforementioned machine learning model is the target first parameter.

[0131] Some embodiments of the present specification determine the target first parameter through the first corresponding relationship, thereby avoiding the amount of calculation caused by continuous processing based on the reinforcement learning model, and improving the calculation efficiency of the crystal growth system 700 while ensuring the accuracy of the target first parameter.

[0132] Figure 7 is an exemplary module diagram of a crystal growth system according to some embodiments of the present specification.

[0133] like Figure 7 As shown, the crystal growth system 700 may include a parameter determination module 710 , a temperature detection module 720 , and a temperature determination module 730 .

[0134] The parameter determination module 710 can be used to determine a parameter set of a crystal growth device in a target furnace, wherein the parameter set includes at least one target parameter, which is a parameter in the crystal growth device that affects crystal growth. In some embodiments, the target parameter includes a target first parameter and a target second parameter, wherein the target first parameter includes a first parameter of a preset object in the crystal growth device in the target furnace, wherein the first parameter reflects a material physical property in the preset object related to crystal growth, and the target second parameter includes a second parameter of the crystal growth device in the target furnace, wherein the second parameter includes a parameter related to the crystal growth device.

[0135] In some embodiments, the parameter determination module 710 can be further used to obtain an initial first parameter, wherein the initial first parameter reflects the first parameter of a preset object in the crystal growth equipment in an initial environment; the initial first parameter is processed based on a reinforcement learning model to determine a target first parameter, and the reward value of the reinforcement learning model is related to the furnace information of the crystal growth equipment.

[0136] In some embodiments, the parameter determination module 710 may be further used to obtain a first corresponding relationship between a furnace of a crystal growth device and a first parameter; and determine a target first parameter based on the target furnace and the first corresponding relationship.

[0137] In some embodiments, the parameter determination module 710 can also be used to process the initial first parameter for each of multiple preset furnaces of the crystal growth equipment based on the reinforcement learning model, and determine the reference first parameter corresponding to the preset furnace, the initial first parameter reflects the first parameter of the preset object in the crystal growth equipment in the initial environment, the reference first parameter is the first parameter of the crystal growth equipment in the preset furnace, and the reward value of the reinforcement learning model is related to the furnace information of the crystal growth equipment; based on the reference first parameters corresponding to multiple preset furnaces, the first corresponding relationship between the furnace of the crystal growth equipment and the first parameter is determined.

[0138] In some embodiments, the parameter determination module 710 can be further used to obtain the historical first parameters of the crystal growth equipment in multiple historical furnaces and the historical operating parameters corresponding to the historical first parameters, wherein the historical first parameters include the first parameters reflecting the preset objects in the crystal growth equipment in the historical furnaces; based on the historical first parameters of the multiple historical furnaces and the historical operating parameters, the target first parameters are determined. In some embodiments, the historical operating parameters include the second temperature detection value of the second temperature measurement point of the crystal growth equipment in multiple historical furnaces, and the parameter determination module 710 can be further used to determine the second corresponding relationship based on the historical first parameters of the multiple historical furnaces and the second temperature detection value, wherein the second corresponding relationship reflects the corresponding relationship between the temperature detection value of the second temperature measurement point of the crystal growth equipment and the first parameter; obtain the third temperature detection value of the second temperature measurement point of the crystal growth equipment in the target furnace; and determine the target first parameter based on the third temperature detection value and the second corresponding relationship. In some embodiments, the historical operating condition parameters include the historical growth time of the crystal growing equipment in the historical furnaces, and the parameter determination module 710 can be further used to determine the third corresponding relationship based on the historical first parameters and historical growth time of multiple historical furnaces, and the third corresponding relationship reflects the corresponding relationship between the growth time of the crystal and the first parameter; obtain the target growth time of the crystal growing equipment in the target furnace; determine the target first parameter based on the target growth time and the third corresponding relationship. In some embodiments, the parameter determination module 710 can be further used to establish a parameter prediction model based on the historical first parameters and historical operating condition parameters of multiple historical furnaces; determine the preset operating condition parameters of the crystal growing equipment in the target furnace; determine the target first parameter based on the preset operating condition parameters and the parameter prediction model.

[0139] In some embodiments, the parameter determination module 710 can also be used to obtain multiple sets of reference data, each set of reference data includes multiple candidate first parameters and verification information corresponding to the multiple candidate first parameters, each candidate first parameter reflects a material physical property of a preset object in the crystal growth equipment; the multiple sets of reference data are processed by an importance analysis model to determine the type of the first parameter, and the importance analysis model is a random forest model.

[0140] The temperature detection module 720 may be configured to obtain a first temperature detection value at a first temperature measurement point of the crystal growth device during a target time period of a target furnace.

[0141] The temperature determination module 730 may be configured to determine the simulated crystal surface temperature corresponding to the first temperature detection value through a crystal surface temperature determination model based on the target parameter and the first temperature detection value.

[0142] In some embodiments, Figure 7 The crystal growth system 700 shown may also include a temperature control module 740. The aforementioned temperature control module 740 may be used to determine the temperature control parameters of the crystal growth device in the target time period of the target furnace based on the simulated crystal surface temperature. In some embodiments, the temperature control module 740 may be further used to obtain multiple reference crystal surface temperatures of each reference device in the reference furnace and multiple reference control parameters corresponding to the multiple reference crystal surface temperatures, wherein the reference device is a device of the same type as the crystal growth device; based on the reference crystal surface temperatures, reference control parameters and simulated crystal surface temperatures of the multiple reference devices in the reference furnace, the temperature control parameters are determined. In some embodiments, the temperature control module 740 may be further used to obtain crystal detection data of crystals generated by the multiple reference devices in the reference furnace; based on the reference crystal surface temperatures, reference control parameters, crystal detection data and simulated crystal surface temperatures of the multiple reference devices in the reference furnace, the temperature control parameters are determined.

[0143] In some embodiments, Figure 7 The crystal growth system 700 shown may further include a temperature adjustment module 750. The temperature adjustment module 750 may be used to automatically send a temperature adjustment instruction based on the temperature control parameter to adjust the crystal growth temperature of the crystal growth equipment in a target time period of a target furnace.

[0144] It should be understood that Figure 7 The crystal growth system 700 and its modules shown can be implemented in various ways. Exemplarily, the crystal growth system 700 can be applied to a processor of a crystal growth device, which can also include a crystal growth device, and the processor can control the crystal growth device through the crystal growth system 700 to implement the crystal growth method described in any embodiment below in this specification.

[0145] It should be noted that the above description of the crystal growth system 700 and its modules is only for convenience of description and does not limit the present specification to the scope of the embodiments. It is understandable that, after understanding the principle of the system, those skilled in the art may arbitrarily combine the modules or form a subsystem to connect with other modules without deviating from the principle. In some embodiments, Figure 7 The parameter determination module 710, temperature detection module 720 and temperature determination module 730 disclosed in the specification can be different modules in a system, or a module can realize the functions of two or more modules. For example, each module can share a storage module, or each module can have its own storage module. Such variations are all within the protection scope of this specification.

[0146] The basic concepts have been described above. Obviously, for those skilled in the art, the above detailed disclosure is only for example and does not constitute a limitation of this specification. Although not explicitly stated here, those skilled in the art may make various modifications, improvements and corrections to this specification. Such modifications, improvements and corrections are suggested in this specification, so such modifications, improvements and corrections still belong to the spirit and scope of the exemplary embodiments of this specification.

[0147] At the same time, this specification uses specific words to describe the embodiments of this specification. For example, "one embodiment", "an embodiment", and / or "some embodiments" refer to a certain feature, structure or characteristic related to at least one embodiment of this specification. Therefore, it should be emphasized and noted that "one embodiment" or "an embodiment" or "an alternative embodiment" mentioned twice or more in different positions in this specification does not necessarily refer to the same embodiment. In addition, certain features, structures or characteristics in one or more embodiments of this specification can be appropriately combined.

[0148] In addition, unless explicitly stated in the claims, the order of the processing elements and sequences described in this specification, the use of alphanumeric characters, or the use of other names are not intended to limit the order of the processes and methods of this specification. Although the above disclosure discusses some invention embodiments that are currently considered useful through various examples, it should be understood that such details are only for illustrative purposes, and the attached claims are not limited to the disclosed embodiments. On the contrary, the claims are intended to cover all modifications and equivalent combinations that are consistent with the essence and scope of the embodiments of this specification. For example, although the system components described above can be implemented by hardware devices, they can also be implemented only by software solutions, such as installing the described system on an existing server or mobile device.

[0149] Similarly, it should be noted that in order to simplify the description disclosed in this specification and thus help understand one or more embodiments of the invention, in the above description of the embodiments of this specification, multiple features are sometimes combined into one embodiment, figure or description thereof. However, this disclosure method does not mean that the features required by the subject matter of this specification are more than the features mentioned in the claims. In fact, the features of the embodiments are less than all the features of the single embodiment disclosed above.

[0150] In some embodiments, numbers describing the number of components and attributes are used. It should be understood that such numbers used in the description of the embodiments are modified by the modifiers "about", "approximately" or "substantially" in some examples. Unless otherwise specified, "about", "approximately" or "substantially" indicate that the numbers are allowed to vary by ±20%. Accordingly, in some embodiments, the numerical parameters used in the specification and claims are approximate values, which may change according to the required features of individual embodiments. In some embodiments, the numerical parameters should take into account the specified significant digits and adopt the general method of retaining digits. Although the numerical domains and parameters used to confirm the breadth of their range in some embodiments of this specification are approximate values, in specific embodiments, the setting of such numerical values ​​is as accurate as possible within the feasible range.

[0151] Each patent, patent application, patent application publication, and other materials, such as articles, books, specifications, publications, documents, etc., cited in this specification are hereby incorporated by reference in their entirety. Except for application history documents that are inconsistent with or conflicting with the contents of this specification, documents that limit the broadest scope of the claims of this specification (currently or later attached to this specification) are also excluded. It should be noted that if the descriptions, definitions, and / or use of terms in the materials attached to this specification are inconsistent or conflicting with the contents described in this specification, the descriptions, definitions, and / or use of terms in this specification shall prevail.

[0152] Finally, it should be understood that the embodiments described in this specification are only used to illustrate the principles of the embodiments of this specification. Other variations may also fall within the scope of this specification. Therefore, as an example and not a limitation, alternative configurations of the embodiments of this specification may be considered consistent with the teachings of this specification. Accordingly, the embodiments of this specification are not limited to those explicitly described in this specification.

Claims

1. A crystal growth method, characterized in that: The method comprises: Determining a parameter set of a crystal growth device in a target furnace, wherein the parameter set includes at least one target parameter, and the target parameter is a parameter in the crystal growth device that affects crystal growth; In a target time period of the target furnace, obtaining a first temperature detection value of a first temperature measuring point of the crystal growth equipment; Based on the target parameter and the first temperature detection value, determining a simulated crystal surface temperature corresponding to the first temperature detection value by a crystal surface temperature determination model, wherein the crystal surface temperature determination model is a numerical simulation model or a machine learning model; Determining temperature control parameters of the crystal growth equipment in a target time period of a target furnace based on the simulated crystal surface temperature; Based on the temperature control parameter, automatically sending a temperature adjustment instruction to adjust the crystal growth temperature of the crystal growth device in the target time period; When the crystal plane temperature determination model is the numerical simulation model, determining the simulated crystal plane temperature corresponding to the first temperature detection value through the crystal plane temperature determination model based on the target parameter and the first temperature detection value includes: Inputting the target parameter and the first temperature detection value into the crystal surface temperature determination model, wherein the crystal surface temperature determination model performs simulation based on the target parameter and the first temperature detection value to determine the simulated crystal surface temperature; When the crystal plane temperature determination model is the machine learning model, determining the simulated crystal plane temperature corresponding to the first temperature detection value through the crystal plane temperature determination model based on the target parameter and the first temperature detection value includes: The target parameter and the first temperature detection value are input into the crystal surface temperature determination model, and the output of the crystal surface temperature determination model is the simulated crystal surface temperature.

2. The method according to claim 1, characterized in that The target parameters include a first target parameter and a second target parameter. The target first parameter includes a first parameter of a preset object in the crystal growth device at a target furnace, wherein the first parameter reflects a physical property of a material in the preset object related to crystal growth. The target second parameter includes a second parameter of the crystal growth equipment in the target furnace, and the second parameter includes a parameter related to the setting of the crystal growth equipment.

3. The method according to claim 2, characterized in that The parameter set for determining the target heat of the crystal growth equipment includes: Acquiring an initial first parameter, wherein the initial first parameter reflects the first parameter of the preset object in the crystal growth device in an initial environment; The initial first parameter is processed based on a reinforcement learning model to determine the target first parameter, and a reward value of the reinforcement learning model is related to the furnace information of the crystal growth equipment.

4. The method according to claim 2, characterized in that The parameter set for determining the target heat of the crystal growth equipment includes: Acquire a first corresponding relationship between the furnace of the crystal growth equipment and the first parameter; Based on the target heat and the first corresponding relationship, the target first parameter is determined; wherein the first corresponding relationship is determined by the following steps: For each of a plurality of preset furnaces of the crystal growth device, an initial first parameter is processed based on a reinforcement learning model to determine a reference first parameter corresponding to the preset furnace, wherein the initial first parameter reflects the first parameter of the preset object in the crystal growth device in an initial environment, the reference first parameter is the first parameter of the crystal growth device in the preset furnace, and the reward value of the reinforcement learning model is related to the furnace information of the crystal growth device; Based on the reference first parameters corresponding to the plurality of preset furnaces, a first corresponding relationship between the furnaces of the crystal growth equipment and the first parameters is determined.

5. The method according to claim 2, characterized in that The parameter set for determining the target heat of the crystal growth equipment includes: Acquire historical first parameters of the crystal growth equipment in multiple historical furnaces and historical operating condition parameters corresponding to the historical first parameters, wherein the historical first parameters include the first parameters reflecting the preset object in the crystal growth equipment in the historical furnaces; The target first parameter is determined based on the historical first parameters of the plurality of historical furnaces and the historical operating condition parameters.

6. The method according to claim 5, characterized in that The historical operating condition parameters include the second temperature detection values ​​of the crystal growth equipment at the second temperature measurement points of the plurality of historical furnaces; The determining the first parameter based on the historical first parameters of the plurality of historical heats and the historical operating condition parameters comprises: Determine a second corresponding relationship based on the historical first parameters and the second temperature detection values ​​of the plurality of historical furnaces, wherein the second corresponding relationship reflects the corresponding relationship between the temperature detection value of the second temperature measurement point of the crystal growth equipment and the first parameter; Acquire a third temperature detection value of the crystal growth equipment at the second temperature measurement point of the target furnace; The target first parameter is determined based on the third temperature detection value and the second corresponding relationship.

7. The method according to claim 5, characterized in that The historical operating condition parameters include the historical growth time of the crystal growth equipment in growing crystals in historical furnaces; The determining the target first parameter based on the historical first parameters of the plurality of historical heats and the historical operating condition parameters comprises: Based on the historical first parameters of the plurality of historical furnaces and the historical growth durations, determining a third corresponding relationship, wherein the third corresponding relationship reflects a corresponding relationship between the growth duration of the crystal and the first parameter; Obtaining a target growth time for the crystal growing device to grow the crystal in the target batch; Based on the target growth duration and the third corresponding relationship, the target first parameter is determined.

8. The method according to claim 2, characterized in that The type of the first parameter is determined by the following steps: Acquire multiple groups of reference data, each group of the reference data includes multiple candidate first parameters and verification information corresponding to the multiple candidate first parameters, each of the candidate first parameters reflects a material physical property of the preset object in the crystal growth device; The multiple groups of reference data are processed by an importance analysis model to determine the type of the first parameter, and the importance analysis model is a random forest model.

9. A crystal growth system, characterized in that: The system comprises: A parameter determination module, used to determine a parameter set of a crystal growth device in a target furnace, wherein the parameter set includes at least one target parameter, and the target parameter is a parameter in the crystal growth device that affects crystal growth; A temperature detection module, used for obtaining a first temperature detection value of a first temperature measuring point of the crystal growth device in a target time period of the target furnace; a temperature determination module, configured to determine, based on the target parameter and the first temperature detection value, a simulated crystal surface temperature corresponding to the first temperature detection value through a crystal surface temperature determination model, wherein the crystal surface temperature determination model is a numerical simulation model or a machine learning model; A temperature control module, configured to determine temperature control parameters of the crystal growth device in a target time period of a target furnace based on the simulated crystal surface temperature; A temperature adjustment module, configured to automatically send a temperature adjustment instruction based on the temperature control parameter to adjust the crystal growth temperature of the crystal growth device in the target time period; When the crystal surface temperature determination model is the numerical simulation model, the temperature determination module is further used to: Inputting the target parameter and the first temperature detection value into the crystal surface temperature determination model, wherein the crystal surface temperature determination model performs simulation based on the target parameter and the first temperature detection value to determine the simulated crystal surface temperature; When the crystal surface temperature determination model is the machine learning model, the temperature determination module is further used to: The target parameter and the first temperature detection value are input into the crystal surface temperature determination model, and the output of the crystal surface temperature determination model is the simulated crystal surface temperature.

Images