Temperature control method, device, apparatus, medium and program product for semiconductor thin film growth

Abstract

This invention relates to the field of semiconductor manufacturing equipment control technology, and provides a temperature control method, apparatus, device, medium, and program product for semiconductor thin film growth. The method includes: acquiring a temperature timing signal and rotation parameters of a wafer in a semiconductor thin film growth cavity; converting the temperature timing signal into temperature distribution data corresponding to the rotation parameters; extracting the temperature trajectory of the target wafer from the temperature distribution data; using the temperature trajectory of the target wafer as a reference signal for temperature control; and adjusting the output parameters of heaters distributed in the cavity according to the reference signal to regulate the temperature of the target wafer. This invention is used to achieve closed-loop temperature control during thin film growth based on the actual temperature of the target wafer.

Description

Technical Field

[0001] This invention relates to the field of semiconductor manufacturing equipment control technology, and in particular to a temperature control method, apparatus, equipment, medium, and program product for semiconductor thin film growth. Background Technology

[0002] In semiconductor thin film growth processes, such as Metal-Organic Chemical Vapor Deposition (MOCVD), wafers are typically placed on a rotating tray for film growth, with multiple temperature sensors and heaters distributed within the cavity. Current temperature control technologies generally employ fixed-position temperature sensors to collect time-series temperature signals from the cavity, and then mix and average these multi-channel temperature signals to create a global temperature control reference, thereby uniformly adjusting the heater output parameters. However, this method has significant drawbacks: firstly, due to the continuous rotation of the wafer, the time-series temperature signals cannot establish a stable correspondence with the wafer's physical location, leading to a disconnect between temperature data and the specific wafer; secondly, the tray often contains both process wafers and non-process objects such as dummy wafers, meaning the averaged temperature reference reflects the overall thermal state of the cavity, rather than the actual thermal behavior of the target wafer, resulting in a separation between the temperature control feedback source and the controlled object. This leads to a mismatch between temperature control commands and the actual thermal requirements of the target wafer, manifested as large temperature control deviations, lag in dynamic response, and insufficient uniformity of the cavity thermal field. Ultimately, this affects the thickness consistency, interface quality, and process repeatability of thin film growth, making it difficult to meet the stringent requirements of high-precision semiconductor manufacturing for thermal environment stability.

[0003] Therefore, there is an urgent need for a temperature control method, apparatus, equipment, medium, and process product for semiconductor thin film growth to improve the above-mentioned problems. Summary of the Invention

[0004] This invention provides a temperature control method, apparatus, equipment, medium, and program product for semiconductor thin film growth. This invention is used to achieve closed-loop temperature control during thin film growth based on the actual temperature of the target wafer.

[0005] According to a first aspect of the present invention, a temperature control method for semiconductor thin film growth is provided, comprising: acquiring a temperature timing signal and rotation parameters of a wafer in a semiconductor thin film growth cavity; converting the temperature timing signal into temperature distribution data corresponding to the rotation parameters; extracting a temperature trajectory of a target wafer from the temperature distribution data; using the temperature trajectory of the target wafer as a reference signal for temperature control; and adjusting the output parameters of heaters distributed in the cavity according to the reference signal to regulate the temperature of the target wafer.

[0006] In one embodiment, extracting the temperature trajectory of a target wafer from temperature distribution data includes: identifying the position range of the target wafer in the rotation angle domain based on at least one of temperature stability characteristics, temperature response rate characteristics, reflectivity signal-to-noise ratio characteristics, and reflectivity intensity characteristics, and extracting the temperature data corresponding to the position range as the temperature trajectory.

[0007] In one embodiment, converting a temperature time series signal into temperature distribution data corresponding to rotation parameters includes: mapping the temperature time series signal to a rotation angle domain according to the rotation parameters, and averaging the temperature data of multiple consecutive rotation cycles at the same rotation angle position to obtain the average temperature distribution data of the rotation parameters in the angle domain.

[0008] In one embodiment, adjusting the output parameters of heaters distributed in the cavity according to a reference signal to control the temperature of the target wafer includes: dividing the heating region into multiple process zones according to the actual distribution of the target wafer in the heating region of the cavity, each process zone including a positive integer number of heaters, and independently adjusting the output parameters of the heaters in each process zone based on the reference signal; the output parameters of the heaters in each process zone are positively correlated with the temperature of the wafer located in that process zone.

[0009] In one embodiment, the method further includes: real-time detection of the signal quality index of the temperature trajectory; when the signal quality index is lower than a preset threshold, re-execute the step of extracting the temperature trajectory of the target wafer from the temperature distribution data.

[0010] In one embodiment, using the temperature trajectory of the target wafer as a reference signal for temperature control includes: generating a virtual temperature control point from the temperature trajectory of the target wafer, and using the virtual temperature control point as a reference signal.

[0011] In one implementation, the temperature trajectory of the target wafer is used as a reference signal for temperature control, including: acquiring preset process partitioning rules; executing an optimal wafer strategy for each process partition based on the wafer identification result and the process partitioning rules, wherein the optimal wafer strategy includes selecting the tray position corresponding to the valid wafer with the highest confidence level within the process partition as the target tray position; generating a mapping relationship based on the target tray position, wherein the mapping relationship associates the virtual temperature control point of the process partition with the physical temperature measurement channel corresponding to the target tray position; and sending the mapping relationship to the temperature control logic execution module so that the temperature control logic execution module switches the temperature measurement channel according to the mapping relationship, and uses the temperature signal collected by the physical temperature measurement channel as the temperature control reference signal for the process partition.

[0012] In one embodiment, the output parameters include output power. Adjusting the output parameters of the heaters distributed in the cavity according to the reference signal includes: acquiring the target temperature curve of each process zone in the current process recipe; for each process zone, generating a power adjustment command through proportional-integral-derivative (PID) control calculation based on the reference signal and target temperature curve corresponding to the process zone; and sending the power adjustment command to the lower-level machine to adjust the output power of the heaters in the process zone.

[0013] In one embodiment, identifying the position range of a target wafer in the rotation angle domain based on at least one of temperature stability characteristics, temperature response rate characteristics, and reflectivity intensity characteristics includes: inputting the temperature stability characteristics, temperature response rate characteristics, and reflectivity intensity characteristics into a trained classification model; outputting the category determination result and corresponding confidence level of the target wafer at each tray position from the classification model, wherein the category determination result is used to indicate whether the target wafer is a valid wafer; determining the tray position corresponding to the valid wafer with the highest confidence level based on the category determination result and confidence level; and using a preset angle range of the tray position in the rotation angle domain as the position range.

[0014] In one embodiment, after extracting the temperature trajectory of the target wafer from the temperature distribution data, the method further includes: generating a map characterizing the current physical state of the wafer disk based on the wafer identification result, the map being used to indicate the category of objects at each tray position; executing a dynamic mapping strategy based on the map and preset process partitioning rules to generate a mapping relationship between virtual temperature control points and physical temperature measurement channels, the mapping relationship being used to map virtual temperature control points to physical temperature measurement channels so that the temperature signal collected by the physical temperature measurement channels can be used as a reference signal.

[0015] According to a second aspect of the present invention, a temperature control device for semiconductor thin film growth is provided, used in any of the methods of the first aspect. The device includes: a lower-level machine and a higher-level machine; the lower-level machine includes a rotation speed calculation module, a data acquisition module, and a temperature control logic execution module; the higher-level machine includes a signal synchronization and preprocessing module, an object recognition module, a dynamic temperature control mapping module, and a temperature control decision module; the data acquisition module is used to acquire temperature timing signals of a wafer in a semiconductor thin film growth cavity; the rotation speed calculation module is used to acquire rotation parameters of the wafer in the semiconductor thin film growth cavity; the signal synchronization and preprocessing module is used to convert the temperature timing signals into temperature distribution data corresponding to the rotation parameters; the object recognition module is used to extract the temperature trajectory of a target wafer from the temperature distribution data; the dynamic temperature control mapping module is used to use the temperature trajectory of the target wafer as a reference signal for temperature control; the temperature control decision module is used to generate a temperature control command based on the reference signal; and the temperature control logic execution module is used to adjust the output parameters of heaters distributed in the cavity according to the temperature control command to regulate the temperature of the target wafer.

[0016] According to a third aspect of the present invention, an electronic device is provided, including a memory and a processor, wherein the memory is used to store a computer program executable by the processor; and the processor is used to execute the computer program in the memory to implement the method described above.

[0017] According to a fourth aspect of the present invention, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, enables the implementation of the above-described method.

[0018] According to a fifth aspect of the present invention, a computer program product is provided, comprising a computer program that, when executed by a processor, implements the method described above.

[0019] Compared with existing technologies, the advantages of this invention are as follows: This invention provides fundamental data for the correlation between temperature and rotational motion by acquiring the temperature timing signal and rotational parameters of the wafer in the semiconductor thin film growth cavity; it converts the temperature timing signal into temperature distribution data corresponding to the rotational parameters, enabling precise positioning of temperature information in the rotational angle domain and establishing a correspondence between temperature and the physical position of the wafer; it extracts the temperature trajectory of the target wafer from the temperature distribution data, obtaining a complete thermal behavior characterization of the target wafer; it uses the temperature trajectory of the target wafer as a reference signal for temperature control, directly anchoring the temperature control feedback source to the target wafer itself; and it adjusts the output parameters of the heaters distributed in the cavity according to the reference signal to regulate the temperature of the target wafer, constructing a closed-loop control circuit with the actual thermal state of the target wafer as input. This technical approach enables a direct and strongly correlated feedback mechanism between temperature control commands and the actual thermal requirements of the target wafer, fundamentally eliminating temperature control deviations and response lags caused by the separation of the measurement reference and the controlled object, significantly improving the accuracy of temperature control, dynamic response capability, and process stability, and effectively ensuring the uniformity of the thermal environment and process repeatability during semiconductor thin film growth. Attached Figure Description

[0020] Figure 1 This is a flowchart illustrating a temperature control method for semiconductor thin film growth according to an exemplary embodiment.

[0021] Figure 2 This is a schematic diagram of the wireframe structure of a temperature control device for semiconductor thin film growth according to an exemplary embodiment.

[0022] Figure 3 This is a block diagram illustrating an electronic device according to an exemplary embodiment.

[0023] Explanation of the reference numerals in the figure:

[0024] 1. Lower-level computer; 2. Upper-level computer; 3. Temperature control device for semiconductor thin film growth;

[0025] 11. Rotation speed calculation module; 12. Data acquisition module; 13. Temperature control logic execution module; 14. Communication interface;

[0026] 21. Signal synchronization and preprocessing module; 22. Object recognition module; 23. Dynamic temperature control mapping module; 24. Temperature control decision module; 25. Human-computer interaction module; 26. Database module;

[0027] 900. Electronic device; 922. Processing component; 926. Power supply component; 932. Memory; 950. Network interface; 958. Input / output interface. Detailed Implementation

[0028] Unless otherwise defined, the technical or scientific terms used in this specification should have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. Specific embodiments of the invention will be described below with reference to the accompanying drawings. It should be noted that, in order to provide a concise description, this specification cannot provide a detailed description of all features of the actual embodiments. Without departing from the spirit and scope of the invention, those skilled in the art can make modifications and substitutions to the embodiments of the invention, and the resulting embodiments are also within the protection scope of the invention.

[0029] In related technologies, existing semiconductor thin film growth equipment commonly employs a fixed temperature measurement point-average temperature control mode for temperature control. Infrared thermometers measure the radiation temperature of a specific annular area on a graphite tray. The collected signal is a mixture of the thermal contributions from the process wafer and non-process objects, such as dummy components. Because dummy components differ significantly from process wafers in surface condition, reflectivity, and thermal capacity, their temperature response often deviates from the actual wafer behavior. The system uses the average value of this mixed signal as a single temperature control reference input to the PID loop. This results in the feedback source essentially reflecting the "contaminated" temperature of a localized area within the cavity, rather than the true thermal state of the target wafer. This introduces a systematic temperature control deviation, directly affecting the radial uniformity of thin film growth and batch process repeatability.

[0030] Some high-end systems attempt to adopt a manually specified temperature measurement point - single-point temperature control mode. This involves operators pre-setting a specific tray position (assuming a process wafer is placed there) based on the loading drawings and statically binding the corresponding temperature measurement channel as the feedback source. While this approach avoids signal mixing issues, it heavily relies on manual prediction and static configuration: each time the product formula or loading scheme is changed, the drawings must be manually consulted and reset, making the process cumbersome and prone to configuration errors. If wafer breakage or displacement occurs during process operation, the system cannot autonomously identify signal degradation and switch to the temperature measurement channel of other healthy wafers, posing a risk of temperature instability. Furthermore, this mode can only utilize temperature information from a single location and cannot integrate the multi-wafer distribution characteristics of the tray to achieve fine-grained zone control, thus limiting the flexibility and adaptability of thermal field control.

[0031] At a deeper level, the existing temperature control architecture suffers from a fragmented system functionality. Data flow between the lower-level computer (responsible for data acquisition and execution) and the upper-level computer (responsible for monitoring and display) is unidirectional. The upper-level computer only has data recording and visualization functions, lacking real-time intelligent analysis capabilities for raw temperature measurement data (such as automatic identification of wafers and dies) and the ability to dynamically reconstruct control logic. This break in the "sensing-decision" link prevents the equipment from autonomously optimizing temperature control strategies based on actual operating conditions, limiting its level of intelligence and making it difficult to meet the stringent requirements of high-precision semiconductor manufacturing for dynamic adaptability to thermal environments and process robustness.

[0032] like Figure 1 As shown, the first embodiment of the present invention provides a temperature control method for semiconductor thin film growth, comprising the following steps:

[0033] S1, acquire the temperature timing signal and rotation parameters of the wafer in the semiconductor thin film growth cavity.

[0034] S2 converts the temperature time series signal into temperature distribution data corresponding to the rotation parameters.

[0035] S3 extracts the temperature trajectory of the target wafer from the temperature distribution data.

[0036] S4 uses the temperature trajectory of the target wafer as a reference signal for temperature control.

[0037] S5 adjusts the output parameters of the heaters distributed in the cavity according to the reference signal to regulate the temperature of the target wafer.

[0038] In one embodiment, extracting the temperature trajectory of a target wafer from temperature distribution data includes: identifying the position range of the target wafer in the rotation angle domain based on at least one of temperature stability characteristics, temperature response rate characteristics, reflectivity signal-to-noise ratio characteristics, and reflectivity intensity characteristics, and extracting the temperature data corresponding to the position range as the temperature trajectory.

[0039] In some specific embodiments, the temperature trajectory of the target wafer is extracted from the temperature distribution data. Specifically, this includes an intelligent identification process between the wafer and its counterpart: First, the rotational angle domain (0°~360°) is uniformly divided into multiple angle intervals corresponding to the number of tray positions based on the physical structure of the graphite tray, with each interval corresponding to one tray position. Then, for each angle interval, feature vectors are extracted from the pre-processed multi-channel temperature and reflectivity signals. These feature vectors include temperature stability features (standard deviation of the temperature signal within the interval), temperature response rate features (mean value of the slope of the temperature curve during the process heating stage), and reflectivity intensity features. The system identifies at least one of the following features: the average amplitude of the reflectivity signal and the reflectivity signal-to-noise ratio feature (the ratio of reflectivity amplitude to background noise). The feature vector is then input into a trained classification model, which outputs the category determination result and corresponding confidence level for each object at each tray position. The category determination result indicates whether the object is a process wafer. Finally, based on the category determination result and confidence level, the tray position corresponding to the process wafer with the highest confidence level is determined, and a preset angle range within the rotation angle domain is used as the position interval of the target wafer. The temperature data corresponding to this position interval is then extracted as the temperature trajectory. This identification process achieves accurate differentiation between process wafers and accompanying wafers, providing reliable data support for subsequent dynamic temperature control based on the actual process wafer temperature.

[0040] In one embodiment, converting a temperature time series signal into temperature distribution data corresponding to rotation parameters includes: mapping the temperature time series signal to a rotation angle domain according to the rotation parameters, and averaging the temperature data of multiple consecutive rotation cycles at the same rotation angle position to obtain the average temperature distribution data of the rotation parameters in the angle domain.

[0041] In some specific embodiments, the temperature timing signal is converted into temperature distribution data corresponding to the rotation parameters, including signal synchronization and preprocessing: First, the raw data packet from the lower-level machine is received. Based on the rotation speed and phase reference information in the rotation parameters, the temperature timing signal and reflectivity timing signal are synchronously mapped to the rotation angle domain to establish the spatiotemporal correspondence between the multi-channel signal and the physical location of the wafer; the rotation angle of the wafer at sampling time t. satisfy:

[0042] ;

[0043] Where RPM is the wafer rotation speed and mod is the modulo operator.

[0044] Secondly, for the same rotational angle position, the temperature data within multiple consecutive rotation cycles are arithmetically averaged to effectively suppress random noise interference and significantly improve the signal-to-noise ratio and stability of the temperature signal in the angle domain. Simultaneously, the reflectivity signal undergoes the same multi-cycle alignment and averaging process to enhance the reliability of subsequent feature extraction. Finally, the temperature and reflectivity data from all sensors are normalized to a standard angular coordinate system of 0°~360°, forming the average temperature distribution data of the wafer in the rotational angle domain. This process not only achieves accurate reconstruction of the time-domain signal into the angular domain but also provides a highly consistent and reliable data foundation for subsequent target wafer temperature trajectory extraction and intelligent wafer / wafer identification through multi-cycle averaging and coordinate unification, fundamentally solving the technical challenge of temperature signal decoupling from wafer position in dynamic rotating scenarios.

[0045] In other specific embodiments, temperature data from multiple consecutive rotation cycles at the same rotation angle position are averaged to satisfy:

[0046] ;

[0047] in, Indicates angle The average temperature signal at that location, This indicates that the k-th rotation period is in the angle The temperature signal at that location, where N represents the number of rotation cycles involved in the averaging.

[0048] In one embodiment, adjusting the output parameters of heaters distributed in the cavity according to a reference signal to control the temperature of the target wafer includes: dividing the heating region into multiple process zones according to the actual distribution of the target wafer in the heating region of the cavity, each process zone including a positive integer number of heaters, and independently adjusting the output parameters of the heaters in each process zone based on the reference signal; the output parameters of the heaters in each process zone are positively correlated with the temperature of the wafer located in that process zone.

[0049] In some specific embodiments, the output parameters of the heater are adjusted based on a reference signal to control the temperature of the target wafer. This includes a dynamic temperature control mapping process: First, the wafer identification result and preset process partitioning rules are obtained. The wafer identification result represents the category determination result and confidence level of the object at each tray position. The process partitioning rules divide the heating area into multiple process partitions based on the actual distribution of the target wafer in the cavity heating area. Then, for each process partition, a predefined mapping strategy is executed (e.g., the "optimal wafer" strategy: among all tray positions identified as valid wafers in the process partition, the tray position with the highest confidence level is selected as the target tray). Next, a mapping relationship is established between the virtual temperature control point of the process partition and the physical temperature measurement channel corresponding to the target tray. Finally, the mapping relationship is encoded into a configuration command and sent to the lower-level machine, so that the lower-level machine uses the temperature signal collected by the physical temperature measurement channel as the temperature control reference signal of the process partition, thereby independently adjusting the output parameters of the heater in the process partition based on the reference signal. This mapping process enables dynamic and precise binding between the temperature control reference signal and the actual target wafer within the process zone. This allows the heating control logic of each process zone to directly respond to the actual thermal state of the wafer within the zone, effectively eliminating temperature control deviations caused by temperature signal source distortion, rigid manual configuration, or abnormal operating conditions. It significantly improves the accuracy, adaptability, and process robustness of zone temperature control, providing core support for the refined and intelligent control of the thermal field during semiconductor thin film growth.

[0050] In other specific embodiments, the output parameters of the heaters in each process zone are independently adjusted based on a reference signal. This includes a zone-specific PID temperature control decision process: for each process zone, the target temperature curve preset in the current process recipe is obtained; the temperature signal collected by the physical temperature measurement channel bound to the process zone after dynamic mapping is used as real-time feedback input; based on the dynamic deviation between the feedback input and the target temperature curve, a power adjustment command corresponding to the process zone is generated through PID control calculation; the power adjustment command is sent to the lower-level machine, which independently adjusts the output power of the heater group within the process zone. This process constructs a dedicated temperature closed-loop control loop for each process zone, enabling the heating control logic to directly respond to the actual thermal state of the target wafer within the zone, effectively eliminating cross-zone interference and signal distortion. The collaborative work of each zone's temperature control loop significantly improves the radial uniformity and dynamic response accuracy of the cavity thermal field, while enhancing the system's adaptive adjustment capability to process disturbances (such as sudden wafer state changes or environmental fluctuations), providing a highly stable and consistent thermal environment guarantee for semiconductor thin film growth processes.

[0051] In one embodiment, the method further includes: real-time detection of the signal quality index of the temperature trajectory; when the signal quality index is lower than a preset threshold, re-execute the step of extracting the temperature trajectory of the target wafer from the temperature distribution data.

[0052] In some specific embodiments, the method further includes a monitoring and adaptive control process: during the semiconductor thin film growth process, the system continuously monitors the signal quality indicators of the reference signal used for temperature control. The signal quality indicators include at least one of the target wafer identification confidence level and the temperature signal noise level. When any signal quality indicator is detected to be lower than a preset threshold, it is determined that the current temperature control reference is not reliable enough. The adaptive control process for the affected process partition is automatically triggered, the step of extracting the temperature trajectory of the target wafer from the temperature distribution data is re-executed, and the temperature control reference signal of the process partition is updated based on the newly extracted temperature trajectory. The temperature control signal source is dynamically switched to the physical temperature measurement channel corresponding to other process wafers in normal condition within the same partition. This process is completed in real time during the process operation without interrupting the thin film growth process. It effectively avoids temperature control reference distortion and temperature fluctuation caused by abnormalities of a single process wafer (such as cracking, contamination, or signal degradation), significantly improves the fault tolerance, dynamic adaptability, and process robustness of the temperature control system, and provides a continuous, stable, and highly reliable thermal environment guarantee for the semiconductor thin film growth process.

[0053] In one embodiment, using the temperature trajectory of the target wafer as a reference signal for temperature control includes: generating a virtual temperature control point from the temperature trajectory of the target wafer, and using the virtual temperature control point as a reference signal.

[0054] In some specific embodiments, the temperature trajectory of the target wafer is used as the reference signal for temperature control. This includes: abstracting and encapsulating the temperature data sequence corresponding to the position interval of the target wafer in the rotational angle domain into a logical entity—a virtual temperature control point—at the control logic layer. The virtual temperature control point does not rely on a fixed physical sensor but is a logical feedback node that dynamically characterizes the complete thermal behavior of the target wafer. Through a dynamic mapping relationship, the virtual temperature control point is linked in real time to the physical temperature measurement channel that collects the temperature of the target wafer, so that the temperature signal stream output by the physical temperature measurement channel continuously updates the temperature value of the virtual temperature control point. In the temperature control loop, the real-time temperature value of the virtual temperature control point is directly used as the feedback reference signal, and the deviation is calculated with the target temperature curve at the corresponding position in the process recipe to generate a temperature control command. This mechanism transforms the temperature sensing source from a fixed hardware configuration into a software-defined logical temperature control unit, ensuring that the temperature control logic is always anchored to the thermal state of the actual process wafer. This provides a unified and flexible logical interface for subsequent advanced functions such as independent partition control and adaptive switching of abnormal signals, significantly improving the accuracy, dynamic adaptability, and architectural scalability of the temperature control system.

[0055] In one implementation, the temperature trajectory of the target wafer is used as a reference signal for temperature control, including: acquiring preset process partitioning rules; executing an optimal wafer strategy for each process partition based on the wafer identification result and the process partitioning rules, wherein the optimal wafer strategy includes selecting the tray position corresponding to the valid wafer with the highest confidence level within the process partition as the target tray position; generating a mapping relationship based on the target tray position, wherein the mapping relationship associates the virtual temperature control point of the process partition with the physical temperature measurement channel corresponding to the target tray position; and sending the mapping relationship to the temperature control logic execution module so that the temperature control logic execution module switches the temperature measurement channel according to the mapping relationship, and uses the temperature signal collected by the physical temperature measurement channel as the temperature control reference signal for the process partition.

[0056] In some specific embodiments, the temperature trajectory of the target wafer is used as a reference signal for temperature control, which is specifically achieved through the collaboration of a dynamic temperature control mapping engine and a reconfigurable execution unit in the lower-level machine.

[0057] The dynamic temperature control mapping engine constructs a software-defined virtual temperature control point system in the host computer. The virtual temperature control point is a logical temperature control unit, independent of the physical sensor layout. Its data structure includes process partition identifier, bound tray position, associated temperature measurement channel index and confidence weight. Based on the wafer identification results (including object category determination and confidence of each tray position) and preset process partition rules, the engine executes a dynamic mapping strategy for each process partition (such as the "optimal wafer strategy": selecting the tray position with the highest confidence among all valid wafers in the partition; or an extended strategy: weighted fusion of multi-wafer temperature signals based on confidence), generates the mapping relationship between virtual temperature control points and physical temperature measurement channels, and encodes it into a structured mapping configuration instruction to be sent to the lower computer in real time.

[0058] The lower-level reconfigurable temperature control execution unit has a built-in non-volatile feedback channel configuration table that stores the binding relationship between virtual temperature control points and physical temperature measurement channels in each process zone. After receiving the mapping configuration instruction, the temperature control logic execution module dynamically updates the configuration table and redirects the feedback signal source of the corresponding PID controller within milliseconds through hardware register reconfiguration or firmware logic switching, so that the temperature signal collected by the physical temperature measurement channel can be seamlessly integrated into the real-time reference signal of the virtual temperature control point.

[0059] This design transforms the temperature sensing source from a fixed hardware configuration into a closed-loop reconstruction mechanism of "software-defined - instruction-driven - hardware response": the virtual temperature control point serves as a unified logical interface, supporting flexible expansion of partitioning strategies; the online dynamic update capability of the feedback channel configuration table enables the system to autonomously respond to scenarios such as wafer anomalies and loading changes during process operation, achieving zero-interruption switching of the temperature control reference, and significantly improving the architectural flexibility, operational robustness and process adaptability of the temperature control system.

[0060] In one embodiment, the output parameters include output power. Adjusting the output parameters of the heaters distributed in the cavity according to a reference signal includes: acquiring the target temperature curve of each process zone in the current process recipe; generating a power adjustment command for each process zone by PID control calculation based on the reference signal and target temperature curve corresponding to the process zone; and sending the power adjustment command to the lower-level machine to adjust the output power of the heaters in the process zone.

[0061] In some specific embodiments, after the dynamic mapping relationship takes effect, for each process partition, the target temperature curve preset in the current process recipe for that process partition is obtained; the temperature signal originating from the target wafer and bound by the mapping is used as the real-time feedback benchmark for that process partition, and dynamic deviation calculation is performed between it and the target temperature curve; corresponding power adjustment instructions are generated through PID control calculations and sent to the lower-level machine; the lower-level machine independently adjusts the output power of the heater group in that process partition according to the instructions. This process constructs a dedicated closed-loop temperature control loop for each process partition with the actual thermal state of the process wafer as the feedback source, so that the power adjustment instructions are strongly correlated with the actual thermal demand of the target wafer in the partition, effectively avoiding the influence of cross-regional thermal interference and temperature measurement signal distortion; the temperature control loops of each partition operate in concert, significantly improving the radial uniformity, dynamic response accuracy and process stability of the cavity thermal field, realizing a complete thermal control closed loop of "sensing the real object - making precise decisions - executing partition control", providing a highly consistent and reliable thermal environment guarantee for semiconductor thin film growth processes.

[0062] In one embodiment, identifying the position range of a target wafer in the rotation angle domain based on at least one of temperature stability characteristics, temperature response rate characteristics, and reflectivity intensity characteristics includes: inputting the temperature stability characteristics, temperature response rate characteristics, and reflectivity intensity characteristics into a trained classification model; outputting the category determination result and corresponding confidence level of the target wafer at each tray position from the classification model, wherein the category determination result is used to indicate whether the target wafer is a valid wafer; determining the tray position corresponding to the valid wafer with the highest confidence level based on the category determination result and confidence level; and using a preset angle range of the tray position in the rotation angle domain as the position range.

[0063] In some specific embodiments, the host computer first receives the raw temperature distribution data and uses the wafer rotation speed information to convert the time-domain signal and accurately align it to the angular domain. To improve the signal-to-noise ratio, the system averages the data from multiple rotations. Subsequently, within the angular domain, for each preset tray interval, multiple dimensional features are extracted, including but not limited to temperature stability features, temperature response rate features, and reflectivity intensity features. These features are fed as input into a pre-trained classification model to complete the online identification task between the target wafer and non-target objects (such as dummy). Based on this identification result, a map characterizing the current physical state of the wafer tray surface is generated, which indicates the object category at each tray location.

[0064] In one embodiment, after extracting the temperature trajectory of the target wafer from the temperature distribution data, the method further includes: generating a map characterizing the current physical state of the wafer disk based on the wafer identification result, the map being used to indicate the category of objects at each tray position; executing a dynamic mapping strategy based on the map and preset process partitioning rules to generate a mapping relationship between virtual temperature control points and physical temperature measurement channels, the mapping relationship being used to map virtual temperature control points to physical temperature measurement channels so that the temperature signal collected by the physical temperature measurement channels can be used as a reference signal.

[0065] In some specific embodiments, based on the aforementioned "physical state map" and preset process partitioning rules, the system further executes a dynamic mapping strategy to select the most suitable wafer (e.g., the effective wafer with the highest confidence level) for each process partition, thereby generating a mapping relationship between virtual temperature control points and actual physical temperature measurement channels. Through this mapping, the system can establish the temperature signal acquired by the selected physical temperature measurement channel as a reference signal for subsequent temperature control processes. This step not only enhances the accuracy of temperature control but also provides a foundation for more refined thermal management in semiconductor manufacturing. The entire process embodies a complete intelligent sensing flow from data acquisition, feature extraction, intelligent recognition to dynamic mapping, providing a reliable basis for optimizing the temperature control system.

[0066] like Figure 2As shown in the second embodiment of the present invention, a temperature control device 3 for semiconductor thin film growth is provided, used in any of the methods described in the above embodiments. The device includes: a lower-level computer 1 and a higher-level computer 2; the lower-level computer 1 includes a rotation speed calculation module 11, a data acquisition module 12, and a temperature control logic execution module 13; the higher-level computer 2 includes a signal synchronization and preprocessing module 21, an object recognition module 22, a dynamic temperature control mapping module 23, and a temperature control decision module 24; the data acquisition module 12 is used to acquire the temperature timing signal of the wafer in the semiconductor thin film growth cavity; the rotation speed calculation module 11 is used to acquire the rotation parameters of the wafer in the semiconductor thin film growth cavity; the signal synchronization and preprocessing module 21 is used to convert the temperature timing signal into temperature distribution data corresponding to the rotation parameters; the object recognition module 22 is used to extract the temperature trajectory of the target wafer from the temperature distribution data; the dynamic temperature control mapping module 23 is used to use the temperature trajectory of the target wafer as a reference signal for temperature control; the temperature control decision module 24 is used to generate a temperature control command based on the reference signal; the temperature control logic execution module 13 is used to adjust the output parameters of the heaters distributed in the cavity according to the temperature control command to regulate the temperature of the target wafer.

[0067] In some specific embodiments, the data acquisition module 12 of the lower computer 1 acquires the raw temperature signals of multiple temperature sensors in the cavity in real time through a multi-channel synchronous acquisition circuit, and embeds a hardware-generated timestamp and phase reference mark for each frame of data, forming a data packet with spatiotemporal identification and then uploading it to the upper computer 2; the rotation speed calculation module 11 calculates the instantaneous rotation speed of the wafer in real time and extracts the phase zero position information based on the pulse sequence output by the rotary encoder through a hardware counting unit or embedded algorithm, and generates a structured parameter packet that is synchronously transmitted with the temperature data. The signal synchronization and preprocessing module 21 of the host computer 2 converts the time-domain temperature signal to a standard angular coordinate system using an angle mapping algorithm based on rotation parameters. It then performs alignment and averaging processing on continuous multi-cycle temperature data at the same angular position to generate high signal-to-noise ratio angular domain temperature distribution data. The object recognition module 22 divides the angular domain data into corresponding intervals according to the physical layout of the trays, extracts multi-dimensional features such as temperature stability, temperature response rate, and reflectivity intensity for each interval, inputs them into a trained classification model for online recognition, outputs the object category determination results and confidence levels for each tray position, generates a structured physical state map, and determines the tray position of the target wafer based on the map to extract its corresponding angle. Temperature data within a certain temperature range is used as the temperature trajectory. The dynamic temperature control mapping module 23 loads preset process partitioning rules and executes dynamic mapping strategies for each process partition in conjunction with the physical state map. For example, it selects valid wafers within the partition and chooses the tray position with the highest confidence as the target, generating a mapping relationship between the virtual temperature control point and the physical temperature measurement channel associated with the target tray. The mapping relationship is encoded into a structured configuration instruction and sent to the lower-level machine 1 in real time. The temperature control decision module 24 retrieves the target temperature curve of each process partition from the process recipe library, uses the real-time temperature of the virtual temperature control point as feedback input, calculates the power adjustment amount through the PID control algorithm, and generates a temperature control instruction package containing the partition identifier and power instruction, which is then sent to the lower-level machine 1. The temperature control logic execution module 13 of the lower-level machine 1 has a built-in dynamically updatable feedback channel configuration table. After receiving the mapping configuration instruction, it updates the feedback signal source binding relationship of the PID controller in milliseconds through hardware logic reconfiguration or firmware switching mechanism, and adjusts the drive signal of the corresponding heater group according to the temperature control instruction to achieve partitioned closed-loop temperature control based on the actual wafer temperature. This device, through the collaborative work of upper and lower computers, deeply integrates hardware perception, software intelligent recognition and dynamic mapping, and precise execution, constructing a complete thermal control closed loop of "perceiving real objects - making precise decisions - executing zoned regulation", which significantly improves the accuracy, adaptability and process robustness of temperature control.

[0068] In other specific embodiments, the device also includes a human-machine interaction module 25, which provides a graphical user interface, renders a top-down layout of the wafer tray in real time and distinguishes effective wafers from non-process objects with differentiated visual identifiers, dynamically presents the temperature tracking curves of each process zone, and visualizes the mapping relationship and switching status between the current virtual temperature control point and the physical temperature measurement channel, providing operators with an intuitive monitoring view of the process and the basis for necessary intervention.

[0069] In some specific embodiments, the device also includes a database module 26, which is used to centrally store and manage the process recipe library, the parameter set of the wafer identification model, and historical process data. The historical process data covers complete wafer identification logs, dynamic mapping records and corresponding temperature control trajectories, supporting process traceability, anomaly analysis, model parameter optimization and process knowledge accumulation, providing a data foundation for continuous system iteration and process quality improvement.

[0070] In some specific embodiments, the device also includes a communication interface 14, which is used to encapsulate the rotation speed parameters, phase reference and raw temperature measurement data generated by the lower computer 1 into a structured data packet and upload it to the upper computer 2 in real time through a high-speed industrial communication bus; at the same time, it receives the temperature control command and mapping configuration command issued by the upper computer 2 and accurately distributes them to the corresponding execution unit of the lower computer 1, so as to realize low-latency and high-reliability data interaction and command synchronization between the upper and lower computers, and provide stable communication support for temperature sensing, intelligent decision-making and precise execution.

[0071] A third embodiment of the present invention provides an electronic device, including a memory and a processor, wherein the memory is used to store a computer program executable by the processor; and the processor is used to execute the computer program in the memory to implement the method of any of the above embodiments.

[0072] Figure 3 This is a block diagram illustrating an electronic device according to an exemplary embodiment. For example, electronic device 900 may be provided as a server. (Refer to...) Figure 3 The electronic device 900 includes a processing component 922, which further includes one or more processors, and memory resources represented by memory 932 for storing instructions, such as application programs, that can be executed by the processing component 922. The application programs stored in memory 932 may include one or more modules, each corresponding to a set of instructions. Furthermore, the processing component 922 is configured to execute instructions to perform the methods described above.

[0073] Electronic device 900 may also include a power supply component 926 configured to perform power management of electronic device 900, a wired or wireless network interface 950 configured to connect electronic device 900 to a network, and an input / output interface 958. Electronic device 900 may operate on an operating system stored in memory 932, such as Windows Server™, MacOS X™, Unix™, Linux™, FreeBSD™, or similar.

[0074] In an exemplary embodiment, a non-transitory computer-readable storage medium including instructions is also provided, such as a memory 932 including instructions, which can be executed by a processing component 922 of an electronic device 900 to perform the above-described method. For example, the non-transitory computer-readable storage medium may be a ROM, random access memory (RAM), CD-ROM, magnetic tape, floppy disk, and optical data storage device.

[0075] A fourth embodiment of the present invention provides a readable storage medium storing a program, which, when executed, implements the method of any of the above embodiments.

[0076] The fifth embodiment of the present invention provides a computer program product, including a computer program, which, when executed, implements the method of any of the above embodiments.

[0077] In this invention, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance. The term "multiple" refers to two or more unless otherwise expressly defined.

[0078] The above description of the embodiments is intended to enable those skilled in the art to understand and apply the present invention. It will be apparent to those skilled in the art that various modifications can be made to these embodiments, and the general principles described herein can be applied to other embodiments without creative effort. Therefore, the present invention is not limited to the embodiments described herein, and any improvements and modifications made by those skilled in the art based on the disclosure of the present invention without departing from the scope and spirit of the invention are within the scope of the present invention.

Claims

1. A temperature control method for semiconductor thin film growth, characterized by, include: Acquire the temperature timing signal and rotation parameters of the wafer in the semiconductor thin film growth chamber; The temperature time-series signal is converted into temperature distribution data corresponding to the rotation parameters; Extract the temperature trajectory of the target wafer from the temperature distribution data; Using the temperature trajectory of the target wafer as a reference signal for temperature control includes: acquiring preset process partitioning rules; based on wafer identification results and the process partitioning rules, where the wafer identification results characterize the category determination result and confidence level of objects at each tray position, and the process partitioning rules divide the heating area into multiple process partitions according to the actual distribution of the target wafer in the cavity heating area; executing an optimal wafer strategy for each process partition, where the optimal wafer strategy includes selecting the tray position corresponding to the valid wafer with the highest confidence level within the process partition as the target tray position; generating a mapping relationship based on the target tray position, where the mapping relationship associates the virtual temperature control point of the process partition with the physical temperature measurement channel corresponding to the target tray position; and sending the mapping relationship to the temperature control logic execution module, so that the temperature control logic execution module switches the temperature measurement channel according to the mapping relationship, and uses the temperature signal collected by the physical temperature measurement channel as the temperature control reference signal for the process partition. The output parameters of the heaters distributed in the cavity are adjusted according to the reference signal to regulate the temperature of the target wafer.

2. The method according to claim 1, characterized in that, Extracting the temperature trajectory of the target wafer from the temperature distribution data includes: identifying the position range of the target wafer in the rotation angle domain based on at least one of temperature stability characteristics, temperature response rate characteristics, reflectivity signal-to-noise ratio characteristics, and reflectivity intensity characteristics, and extracting the temperature data corresponding to the position range as the temperature trajectory.

3. The method according to claim 1, characterized in that, Converting the temperature time series signal into temperature distribution data corresponding to the rotation parameter includes: mapping the temperature time series signal to the rotation angle domain according to the rotation parameter, and averaging the temperature data of multiple consecutive rotation cycles at the same rotation angle position to obtain the average temperature distribution data of the rotation parameter in the angle domain.

4. The method according to claim 1, characterized in that, Adjusting the output parameters of the heaters distributed in the cavity according to the reference signal to regulate the temperature of the target wafer includes: dividing the heating region into multiple process zones according to the actual distribution of the target wafer in the heating region of the cavity, each process zone including a positive integer number of heaters, and independently adjusting the output parameters of the heaters in each process zone based on the reference signal; the output parameters of the heaters in each process zone are positively correlated with the temperature of the wafer located in that process zone.

5. The method according to claim 1, characterized in that, The method further includes: real-time detection of the signal quality index of the temperature trajectory; when the signal quality index is lower than a preset threshold, re-execute the step of extracting the temperature trajectory of the target wafer from the temperature distribution data.

6. The method according to claim 4, characterized in that, The output parameters include output power. Adjusting the output parameters of the heaters distributed in the cavity according to the reference signal includes: Obtain the target temperature profile for each process zone in the current process formulation; For each process zone, a power adjustment command is generated through PID control calculation based on the reference signal corresponding to the process zone and the target temperature curve. The power adjustment command is sent to the lower-level machine to adjust the output power of the heaters in the process zone.

7. The method according to claim 2, characterized in that, The method of identifying the position range of a target wafer in the rotational angle domain based on at least one of temperature stability features, temperature response rate features, and reflectivity intensity features includes: inputting the temperature stability features, temperature response rate features, and reflectivity intensity features into a trained classification model; outputting the category determination result and corresponding confidence level of the target wafer at each tray position by the classification model, wherein the category determination result is used to indicate whether the target wafer is a valid wafer; determining the tray position corresponding to the valid wafer with the highest confidence level based on the category determination result and confidence level; and using a preset angle range of the tray position in the rotational angle domain as the position range.

8. A temperature control device for semiconductor thin film growth, used in the method according to any one of claims 1 to 7, characterized in that, The device includes: a lower-level machine and a higher-level machine; The lower-level computer includes a rotation speed calculation module, a data acquisition module, and a temperature control logic execution module; the upper-level computer includes a signal synchronization and preprocessing module, an object recognition module, a dynamic temperature control mapping module, and a temperature control decision module. The data acquisition module is used to acquire the temperature timing signal of the wafer in the semiconductor thin film growth cavity; The rotation speed calculation module is used to obtain the rotation parameters of the wafer in the semiconductor thin film growth cavity; The signal synchronization and preprocessing module is used to convert the temperature time series signal into temperature distribution data corresponding to the rotation parameters; The object recognition module is used to extract the temperature trajectory of the target wafer from the temperature distribution data; The dynamic temperature control mapping module is used to use the temperature trajectory of the target wafer as a reference signal for temperature control; The temperature control decision module is used to generate temperature control commands based on the reference signal; The temperature control logic execution module is used to adjust the output parameters of the heaters distributed in the cavity according to the temperature control command, so as to regulate the temperature of the target wafer.

9. An electronic device, characterized in that, The device includes a memory and a processor, wherein the memory stores a computer program executable by the processor; and the processor executes the computer program in the memory to implement the method as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the executable computer program in the storage medium is executed by a processor, it can implement the method as described in any one of claims 1 to 7.

11. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1 to 7.

Images