Equipment self-adaptive control method for growth of silicon carbide epitaxial film

By using multi-source temperature monitoring and PID optimization, adaptive control of the silicon carbide epitaxial thin film growth equipment was achieved, solving the problems of defect positioning ambiguity and C/Si ratio imbalance, and improving growth quality and efficiency.

CN121344760APending Publication Date: 2026-01-16JIANGSU HI-PRINT TECH CO LTD

Patent Information

Application Number
CN202511280701.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-09-09
Publication Date
2026-01-16

AI Technical Summary

Technical Problem

Existing silicon carbide epitaxial thin film growth equipment lacks multi-source data fusion verification, resulting in ambiguity in the location of defect types and causes. PID parameters rely on empirical tuning, which can easily lead to temperature oscillations. Power regulation does not have over-limit protection. Asynchronous gas flow regulation leads to C/Si ratio imbalance, which may cause thermal deformation of the graphite substrate.

Method used

Temperature is monitored using a multi-wavelength infrared thermometer and a distributed fiber optic sensor. Combined with anomaly analysis signals, PID parameters are optimized using the critical proportional method to achieve adaptive adjustment of power and gas flow. Differentiated treatment is applied to center and edge defects, and a mapping between defect type, cause, and adjustment strategy is established.

Benefits of technology

It can accurately distinguish between central defects and edge defects, improve the accuracy of defect classification, increase the response speed of temperature regulation, avoid equipment damage, enhance the control accuracy of C/Si ratio, reduce the deviation of edge thickness uniformity, improve the batch yield of epitaxial wafers, and shorten the preparation cycle.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN121344760A_ABST
    Figure CN121344760A_ABST
Patent Text Reader

Abstract

The invention discloses an equipment self-adaptive control method for growth of a silicon carbide epitaxial thin film, relates to the technical field of growth of thin film materials, and solves the problems that thermal deformation of a graphite base is possibly caused due to the fact that over-limit protection is not set in power regulation; in order to solve the technical problems of C / Si ratio imbalance caused by temperature deviation due to the fact that the temperature deviation cannot be counteracted due to asynchronous power and gas flow regulation, space-parameter-time recognition logic is innovatively adopted, an abnormity-defect association knowledge base is matched, a center defect and an edge defect can be accurately distinguished, PID parameters are optimized by adopting a critical proportion method, and the accuracy and the reliability of the system are improved. The temperature regulation response speed is increased, a power upper and lower limit judgment mechanism is increased, equipment damage caused by sudden power change is avoided, synchronous linkage of power regulation and H2 / C / Si ratio is achieved, accurate mapping of defect type-cause-regulation strategy is established, differential schemes are designed for edge warping, polycrystal / particles, impurity aggregation and uneven thickness, and the defect type-cause-regulation strategy is optimized. And the edge thickness uniformity deviation and the edge defect rate are reduced.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of thin film material growth technology, specifically to an adaptive control method for equipment used in silicon carbide epitaxial thin film growth. Background Technology

[0002] Silicon carbide (SiC) epitaxial films are the core material for fabricating high-performance power semiconductor devices. Their quality (defect density, thickness uniformity, and crystal integrity) directly determines the breakdown voltage, on-resistance, and long-term reliability of the devices.

[0003] According to patent application number 202111287621.9, an apparatus and control method for silicon carbide epitaxial thin film growth are disclosed. The process includes: S1. Cleaning the reaction chamber, which involves using a gas supply system to reduce the concentration of impurity gases in the reaction chamber to below 1 ppb by coordinating the gas intake and vacuum pump operation; S2. Pre-growth preparation, including preparing the gas supply and heating the sidewall heaters and substrate heaters of the reaction chamber to a tray loading temperature of 920°C; S3. Tray loading, which involves a robotic arm transferring a tray with the substrate placed on it from the TM cavity to a rotating device within the reaction chamber; S4. The sidewall heaters and substrate heaters continue to heat up to a preset temperature of 1620-1680°C. Simultaneously, the tray is rotated by a drive unit, and the rotation speed is gradually increased at a certain slope to a preset speed of 600-1000°C. S5. While heating, adjust the flow rates of the reaction gas and purge gas to meet the growth requirements; S6. After the temperature reaches the set value, wait for the second preset time of 2-5 minutes to stabilize the temperature in the reaction chamber and on the tray; S7. Enter the growth mode, at which time the temperature fluctuation of the reaction chamber and on the tray is controlled within ±1℃ until a film of the preset thickness is grown; S8. After growth ends, wait for cooling, switch all H2-containing carrier gas and purge gas to Ar gas and reduce the flow rate to the set value, turn off the side wall heater and substrate heater to lower the temperature of the reaction chamber and on the tray until the temperature drops to the tray removal temperature, turn off the drive unit to slowly reduce the rotation speed of the tray to 0; S9. Remove the tray. When the actual temperature of the reaction chamber and on the tray drops to the 920℃ tray removal temperature, the robot arm removes the tray from the reaction chamber to the cooling chamber for cooling.

[0004] However, existing defect judgments rely on a single signal and lack multi-source data fusion verification, resulting in ambiguity in the location of defect types, morphologies, and causes, making targeted adjustments difficult. At the same time, PID parameters rely on empirical tuning, which is prone to temperature oscillations; power regulation does not have over-limit protection, which may cause thermal deformation of the graphite base; and power and gas flow regulation are not synchronized, which cannot work together to offset the C / Si ratio imbalance caused by temperature deviation. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides an adaptive control method for equipment used in silicon carbide epitaxial thin film growth. This method solves the problems of power regulation not being set with over-limit protection, which may cause thermal deformation of the graphite substrate; and power and gas flow regulation not being synchronized, which cannot work together to offset the C / Si ratio imbalance caused by temperature deviation.

[0006] To achieve the above objectives, the present invention provides the following technical solution: an adaptive control method for equipment used in silicon carbide epitaxial thin film growth, the method specifically comprising the following steps: Step 1: Compare the temperature of the silicon carbide epitaxial film with the standard temperature range. If they match, a normal monitoring signal is generated; if they do not match, an abnormal analysis signal is generated. Step 2: Based on the anomaly analysis signal, extract the anomaly location, parameter deviation and change trend, and combine real-time monitoring data and physical detection after extension to determine the central defect or edge defect; Step 3: Handle the central defect, calculate the central temperature deviation, and use the critical proportional method to perform PID control to calculate the adjustment power. Determine the adjustment standard by considering the upper and lower limits of the equipment power, and generate power adjustment information. Step 4: Based on the power adjustment information, perform carrier gas flow compensation and carbon-silicon ratio fine-tuning simultaneously. For carrier gas flow compensation, calculate the carrier gas flow compensation. For carbon-silicon ratio fine-tuning, control the carbon source flow and generate carbon-silicon ratio fine-tuning information. Step 5: Process edge defects. Classify edge defect types to obtain edge warping, edge polycrystalline / granular defects, edge impurity aggregation, and edge thickness unevenness. Adjust according to different defect types to generate edge defect processing information.

[0007] As a further aspect of the present invention, the method for comparing the temperature of the silicon carbide epitaxial thin film with the standard temperature range is as follows: One multi-wavelength infrared thermometer is set up to monitor the temperature at the center of the wafer, and three distributed fiber optic sensors are set up to monitor the temperature at the edge. A moving average filter is used to eliminate high-temperature thermal radiation noise, and data from different sensors are aligned by timestamps to ensure spatiotemporal consistency. The system synchronously acquires corresponding standard information, including the standard temperature ranges for the heating, constant temperature, and cooling stages. It compares the collected temperature data with the standard information, and if the data exceeds the allowable range for the corresponding stage, it generates an abnormal analysis signal; otherwise, it generates a normal monitoring signal.

[0008] As a further aspect of the present invention, the method for determining a central defect or an edge defect is as follows: The abnormal analysis signal is disassembled to extract the abnormal location, parameter deviation, and trend characteristics. The defect association area and type direction are initially located. Combined with real-time monitoring data and external physical detection, the existence, type, and distribution characteristics of the defect are verified. Based on the location, shape, and signal characteristics of the defect, the defect is classified as a central defect or an edge defect.

[0009] As a further aspect of the present invention, the power regulation information is generated in the following manner: According to the formula =T0-T s Calculate the center temperature deviation T0 represents the standard temperature, T s This represents the measured temperature. The critical proportional method is used to optimize the PID parameters. First, K is set... i =0、K d =0, increase K p Once the temperature exhibits constant amplitude oscillations, press K. p =0.6 × critical value, K i =2×K p / Critical period, K d =K p Set the initial value to ×critical period / 8, then fine-tune it through on-site debugging, while also following the formula. =K p × +K i × +K d ×d / dt, calculate the regulating power ; Obtain the upper and lower limits of the device's power and determine the current power adjustment. Check if there are any over-limit situations during adjustment. If so, adjust according to the corresponding upper or lower limit. If not, adjust the power accordingly. The current power of the equipment is adjusted according to the standard, and power adjustment information is generated.

[0010] As a further aspect of the present invention, the method for calculating H2 carrier gas flow compensation is as follows: Obtain regulated power Synchronously according to the formula =k1× The increased flow rate of H2 was calculated. , where k1 is the proportionality coefficient.

[0011] As a further aspect of the present invention, the method for generating C / Si ratio fine-tuning information is as follows: For center temperature deviation The value is used to determine if... If the temperature is less than 0, the flow rate of carbon source C3H8 will be reduced by 0.5%-1% for every 1°C increase in temperature; if the C3H8 flow rate reaches the lower limit, the flow rate of silicon source SiH4 will be increased by 0.3%-0.5% for every 1°C increase in temperature, and the flow rate of H2 will be increased simultaneously.

[0012] As a further aspect of the present invention, regarding the center temperature deviation... The value is used to determine if... >0, based on the temperature being 1°C lower, increase the C3H8 flow rate by 0.5%-1%, with a single adjustment not exceeding 4%; if the C3H8 flow rate reaches the upper limit, then reduce the SiH4 flow rate by 0.2%-0.4% for every 1°C lower temperature, and simultaneously reduce the H2 flow rate.

[0013] As a further aspect of the present invention, the edge defect is processed in the following way: For edge warping: reduce the temperature difference between the edge and the center by optimizing the power of the edge heating module. If the edge temperature is >10℃ lower than the center, increase the power of the edge heating coil by 5%-8% to within ±5℃ of the temperature difference. If the edge temperature is too high, reduce the power and check the distance between the heating coil and the substrate. For edge polycrystalline / particle defects: Install a high-purity graphite annular porous edge airflow guide ring to increase the edge H2 flow rate by 10%-15%, and tilt the SiH4 / C3H8 nozzles near the edge towards the center by 5°-10°.

[0014] As a further aspect of the present invention, the edge defect is processed in the following way: To address the accumulation of edge impurities: H2 is purified in three stages: deoxygenation, dehydration, and metal removal. Before each growth, the reactor wall is cleaned in situ for 30 minutes at a flow rate of 800-1000 sccm and a temperature of 1600℃. For uneven edge thickness: If the thickness is too thin, increase the H2 flow rate at the edge to 575±25sccm. If the thickness difference is still >3%, increase the flow rates of SiH4 and C3H8 at the edge by 4%±1%. If the thickness is too thick, reduce the flow rate of SiH4 at the edge by 4%±1% and increase the air extraction rate at the edge. If necessary, add 1-2 small holes at the edge of the distribution plate.

[0015] This invention provides an adaptive control method for equipment used in silicon carbide epitaxial thin film growth. Compared with the prior art, it has the following advantages: This invention innovatively adopts a space-parameter-time identification logic and, by matching an anomaly-defect association knowledge base, can accurately distinguish between central and peripheral defects, improving defect classification accuracy. It optimizes PID parameters using the critical proportional method to improve temperature regulation response speed, adds a power upper and lower limit judgment mechanism to avoid equipment damage caused by sudden power changes, achieves synchronous linkage between power regulation and H2 / C / Si ratio, and improves C / Si ratio control accuracy through a graded fine-tuning strategy.

[0016] This invention establishes a precise mapping between defect type, cause, and adjustment strategy. Differentiated solutions are designed for edge warping, polycrystalline / granular structures, impurity aggregation, and thickness inhomogeneity. Compared with the existing uniform adjustment mode, this reduces edge thickness uniformity deviation and edge defect rate. Through standardized control of the entire process, it improves the batch yield of epitaxial wafers and shortens the single-wafer preparation cycle by 10%-15%. At the same time, PID parameters, gas regulation coefficients, etc., can be iteratively reused through a knowledge base, reducing reliance on process debugging and lowering the experience requirements of operators. This provides stable and controllable technical support for the large-scale mass production of SiC epitaxy. Attached Figure Description

[0017] Figure 1 This is a diagram illustrating the steps and methods of the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Please see Figure 1 This application provides an adaptive control method for equipment used in silicon carbide epitaxial thin film growth, which specifically includes the following steps: Step 1: Acquire growth information of the silicon carbide epitaxial thin film, specifically temperature data. A multi-wavelength infrared thermometer is set up with parameters including a focusing diameter of 20mm, a sampling frequency of 10Hz, and an accuracy of ±0.5℃ to monitor the wafer center temperature. Three distributed fiber optic sensors are set up with parameters including uniform distribution along the circumference at 120°, a distance of 5mm from the edge, and a sampling frequency of 8Hz to monitor the edge temperature. Simultaneously, a moving average filter is used to eliminate high-temperature thermal radiation noise. Data from different sensors is aligned using timestamps to ensure spatiotemporal consistency. Corresponding standard information is acquired synchronously, including the heating, isothermal, and cooling stages. Each stage corresponds to a different standard temperature range, and the two are compared. By analyzing the difference between the two and the allowable range for the corresponding stage (5℃ for the heating stage, 2℃ for the isothermal stage, and 5℃ for the cooling stage), if the growth information does not match the standard information, it indicates an anomaly, and an anomaly analysis signal is generated. Conversely, if the growth information matches the standard information, it indicates no anomaly, and a normal monitoring signal is generated. Step 2: Deconstruct the abnormal analysis signals generated in the early stage, extract key features, including abnormal location, parameter deviation, and change trend, and preliminarily locate possible defect-related areas and types. Combine real-time monitoring data and physical detection after extension to verify the actual existence, type, and distribution characteristics of defects. Based on the location, shape, and signal characteristics of defects, accurately classify them to obtain central defects or edge defects, and then analyze and process them separately. Specifically, the abnormal analysis signal is decomposed into three dimensions: space, parameter, and time. Quantitative features are extracted, including abnormal location, parameter deviations (such as the absolute value and percentage of temperature / flow rate deviation from the standard), and trend of change (such as deviation rate, duration, and periodicity). Feature vectors are formed, and based on these feature vectors, a pre-defined abnormal-defect association knowledge base is matched. For example, a 3°C overshoot at the center temperature lasting 10 seconds indicates a tendency for microtube defects, and an 8°C temperature difference at the edge indicates a tendency for annular defects. This process initially identifies the defect association area and its possible types. Surface defects are detected by laser scattering imaging, and the dynamic changes of defects are verified simultaneously by RHEED (crystal form analysis). The morphology is observed by OM, the crystal integrity is identified by PL imaging, and the three-dimensional morphology is measured by AFM to confirm the existence and precise morphology of defects, such as microtube diameter and stacking fault length; spatial distribution, such as 5mm ring distribution at the edge and random distribution at the center. Defects are classified by two dimensions: region and type: central defects (such as microtubes and stacking faults) and edge defects (such as ring defects, triangular defects, and carbon inclusions).

[0020] Step 3: Analyze and process the obtained central defects. Use a multi-wavelength infrared thermometer to focus on the central region of the wafer to obtain the center temperature of the silicon carbide epitaxial film, and calculate the center temperature deviation. Specifically... =T0-Ts Where T0 represents the standard temperature, T s This represents the measured temperature. Next, the critical proportional method is used to optimize the PID parameters. First, K is set... i =0、K d =0, increase K p Once the temperature exhibits constant amplitude oscillations, press K. p =0.6 × critical value, K i =2×K p / Critical period, K d =K p Set the initial value to ×critical period / 8, then fine-tune it through on-site debugging, while also following the formula. =K p × +K i × +K d ×d / dt, calculate the regulating power ; Obtain the upper and lower limits of the device's power and determine the current power adjustment. Check if there are any over-limit situations during adjustment. If so, adjust according to the corresponding upper or lower limit. If not, adjust the power accordingly. To standardize the current power of the equipment, power adjustment information is generated; Step 4: Based on the obtained adjustment power Simultaneously, gas flow rate adjustment is performed, specifically including H2 carrier gas flow rate compensation and C / Si ratio fine-tuning, with different adjustment processes applied to each. The processing method for H2 carrier gas flow rate compensation is as follows: Obtain regulated power Synchronously according to the formula =k1× The increased flow rate of H2 was calculated. Where k1 is a proportionality coefficient, for example, the center temperature is 3℃ higher ( =+3℃), then =0.02x3=0.06slm (that is, the total H2 flow rate increases from 50slm to 50.06slm), and by enhancing the convective heat dissipation in the central area, the temperature will be brought back to the target range within 5-10 seconds; The method for fine-tuning the C / Si ratio is as follows: [This refers to addressing the center temperature deviation.] The value is used to determine if... A value less than 0 indicates that the temperature is too high, and the Si atom volatilization rate exceeds the deposition rate, resulting in a decrease in the actual amount of Si participating in the reaction. Specifically, for every 1°C increase in temperature, the C3H8 carbon source flow rate is reduced by 0.5%-1% of the original flow rate, with each adjustment controlled to not exceed 3%. The C3H8 carbon source flow rate is assessed; if it has reached the lower limit, such as 90% of the original flow rate, the SiH4 silicon source flow rate is increased, specifically by 0.3%-0.5% for every 1°C increase in temperature. Conversely, the C3H8 carbon source flow rate is further reduced. Based on the above adjustment method, the H2 carrier gas flow rate is increased simultaneously, and the calculated increase in flow rate is used as the reference. Adjust the settings to the standard until everything is working properly; For example, during the isothermal stage of a 6-inch 4H-SiC epitaxial growth circuit, the target central temperature was 1600℃, but the actual monitored temperature was 1604℃, which was 4℃ higher than expected. The initial carbon source C3H8 flow rate was 2.0 slm, the SiH4 flow rate was 1.5 slm, and the initial C / Si ratio was approximately 1.33. The adjustment method was as follows: the carbon source C3H8 flow rate was reduced by 1.5% / ℃ × 4℃ = 6%, specifically from 2.0 slm to 1.88 slm, and the H2 flow rate was increased by 0.025 slm / ℃ × 4℃ = 0.1 slm, specifically from 50 slm to 50.1 slm.

[0021] Conversely if A value greater than 0 indicates that the temperature is too low, weakening the diffusion ability of Si and reducing the deposition rate. Specifically, for every 1°C decrease in temperature, the C3H8 carbon source flow rate is increased by 0.5%-1% of the original flow rate, with a single adjustment not exceeding 4%. Simultaneously, it is determined whether the C3H8 carbon source flow rate has reached its upper limit. If it has, the Si input is reduced; specifically, for every 1°C decrease in temperature, the SiH4 flow rate is reduced by 0.2%-0.4%. Conversely, if the upper limit has not been reached, the C3H8 carbon source flow rate is further increased, simultaneously adjusting the calculated flow rate. To standardize the reduction of H2 carrier gas flow rate until the overall system returns to normal, C / Si ratio fine-tuning information is generated.

[0022] For example, during the isothermal stage of a 6-inch 4H-SiC epitaxial growth circuit, the target center temperature was 1600℃, but the actual monitored temperature was 1595℃, which was 5℃ lower than expected. The initial carbon source C3H8 flow rate was 2.0 slm, the SiH4 flow rate was 1.5 slm, and the initial C / Si ratio was approximately 1.33. The adjustment method was to increase the C3H8 flow rate by 1% / ℃ × 5℃ = 5%, from 2.0 slm to 2.1 slm; increase the center heating power by 2.5%, from 12kW to 12.3kW; and decrease the H2 flow rate by 0.02 slm / ℃ × 5℃ = 0.1 slm, from 50 slm to 49.9 slm.

[0023] Step 5: Analyze and process the obtained edge defects, determine the types of edge defects, and establish a mapping table between edge defect types and their causes. At the same time, adjust the processing based on different edge defects. The types of edge defects include edge warping, edge polycrystalline / particle defects, edge impurity aggregation, and edge thickness unevenness. To address edge warping, which is caused by thermal stress differences, the key is to reduce the temperature difference between the substrate edge and center, while optimizing the heat conduction path. This is achieved by optimizing the power of the edge heating module. If the edge temperature is >10°C lower than the center, the power of the edge heating coil is gradually increased, with a single increase of 5%-8%, until the temperature difference between the edge and center is controlled within ±5°C. If the edge temperature is too high, the power is reduced and the distance between the heating coil and the substrate is checked. To address edge polycrystalline / particle defects, an edge airflow guide ring with a porous annular structure made of high-purity graphite is installed to guide the carrier gas H2 towards the edge, increasing the flow rate by 10%-15%. For example, if the central airflow velocity is 500 sccm, the edge velocity can be adjusted to 550-575 sccm, removing excess Si source. At the same time, the SiH4 / C3H8 nozzles near the edge are tilted 5°-10° towards the center to reduce direct impact of reactants in the edge region. To regulate the accumulation of edge impurities, the carrier gas H2 needs to undergo three-stage purification: deoxygenation, dehydration, and metal removal. Before each growth, the H2 flow rate is controlled at 800-1000 sccm, and the reactor wall is cleaned in situ at 1600℃ for 30 minutes to remove residual carbon deposits or metal impurities. To address uneven edge thickness, the edge thickness was analyzed to determine whether it was too thin or too thick, and then dealt with accordingly. When the thickness was too thin, the H2 flow rate at the edge was increased to 575±25 sccm to enhance the convective transport of reactants to the edge. After 2 minutes, the thickness change was monitored. If the thickness difference was still >3% after the first adjustment, the SiH4 flow rate was increased by 4%±1% through the independent gas path at the edge, and the C3H8 flow rate was increased by 4% simultaneously, while maintaining the C / Si ratio unchanged. When the thickness is too thick, reduce the SiH4 flow rate at the edge by 4%±1%, and at the same time increase the opening of the vacuum pump valve in the edge area from 50% to 60% to accelerate the discharge of unreacted gas. If the first-stage adjustment is not effective, add 1-2 smaller diameter openings at the edge of the distribution plate. Verify the reduction of the edge reactant injection amount by 8%-10% through CFD simulation.

[0024] Some of the data in the above formulas are numerical calculations with dimensions removed, and the contents not described in detail in this specification are all prior art known to those skilled in the art.

[0025] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. A method for adaptive control of equipment for epitaxial thin film growth of silicon carbide, characterized in that, The method specifically comprises the following steps: Step one, compare the temperature of the silicon carbide epitaxial film with the standard temperature range, if matched, generate a normal monitoring signal, if not matched, generate an abnormal analysis signal; Step two, based on the abnormal analysis signal, extract the abnormal position, parameter deviation and change trend, combine the real-time monitoring data and the physical detection after epitaxy to determine the center defect or edge defect; Step three, handle the center defect, calculate the center temperature deviation, and use the critical proportion method for PID control calculation to adjust the power, determine the adjustment standard by comprehensively considering the upper and lower limits of the equipment power, and generate power adjustment information; Step four, according to the power adjustment information, simultaneously carry out carrier gas flow compensation and carbon-silicon ratio fine tuning, for carrier gas flow compensation, calculate the carrier gas flow compensation, for carbon-silicon ratio fine tuning, control the carbon source flow, and generate carbon-silicon ratio fine tuning information; Step five, handle the edge defect, classify the edge defect types to obtain edge warping, edge polycrystalline / grain defect, edge impurity aggregation and edge thickness unevenness, and adjust according to different defect types to generate edge defect handling information.

2. The apparatus adaptive control method for silicon carbide epitaxial thin film growth according to claim 1, wherein, The way to compare the temperature of the silicon carbide epitaxial film with the standard temperature range is: Set 1 multi-wavelength infrared temperature measuring instrument to monitor the wafer center temperature, set 3 distributed optical fiber sensors to monitor the edge temperature, use sliding average filtering to eliminate high-temperature thermal radiation noise, and align different sensor data through time stamp to ensure spatiotemporal consistency; Synchronously acquire corresponding standard information, the standard information includes the standard temperature range of the heating stage, the constant temperature stage and the cooling stage, compare the collected temperature data with the standard information, if it exceeds the allowed range of the corresponding stage, generate an abnormal analysis signal, otherwise generate a normal monitoring signal.

3. The apparatus adaptive control method for silicon carbide epitaxial thin film growth according to claim 1, wherein, The way to determine the center defect or edge defect is: Disassemble the abnormal analysis signal, extract the abnormal position, parameter deviation, change trend characteristics, preliminarily locate the defect related area and type direction, combine the real-time monitoring data and the physical detection after epitaxy to verify the existence, type and distribution characteristics of the defect, based on the position, shape and signal characteristics of the defect, classify the defect as a center defect or an edge defect.

4. The apparatus adaptive control method for silicon carbide epitaxial thin film growth of claim 1, wherein, The way to generate power adjustment information is: According to the formula = T0- T s The central temperature deviation is calculated , T0 represents the standard temperature, T s represents the measured temperature, the PID parameters are optimized by using the critical ratio method, first set K i = 0, K d = 0, increase K p to the temperature appears equal amplitude oscillation, then according to K p = 0.6 × critical value, K i = 2 × K p / critical period, K d = K p × critical period / 8 set initial value, and through the field debugging fine tuning, according to the formula = K p × + K i × + K d × d / dt, the adjusting power is calculated; acquire the upper and lower limits of the power of the device, judge whether the current adjustment power exceeds the limits if the adjustment exceeds the limits, adjust the power according to the corresponding upper or lower limit, if not, adjust the current power of the device according to the adjustment power and generate power adjustment information.

5. The apparatus adaptive control method for silicon carbide epitaxial thin film growth according to claim 1, wherein, The way to calculate H2 carrier gas flow compensation is: Acquiring adjustment power , the synchronization is according to formula =k1× The increased flow of H2 is calculated , wherein k1 is a proportional coefficient.

6. The apparatus adaptive control method for silicon carbide epitaxial thin film growth of claim 1, wherein, The way to generate C / Si ratio fine tuning information is: The numerical value of the center temperature deviation is judged, if <0, based on the temperature, the carbon source C3H8 flow is reduced by 0.5%-1% for every 1℃ higher; if the C3H8 flow reaches the lower limit, the silicon source SiH4 flow is increased by 0.3%-0.5% for every 1℃ higher, and the H2 flow is simultaneously increased.

7. The apparatus adaptive control method for silicon carbide epitaxial thin film growth according to claim 6, wherein, The numerical value of the center temperature deviation is judged, if > 0, increase the C3H8 flow by 0.5%-1% for every 1°C lower temperature, and the single adjustment does not exceed 4%; if the C3H8 flow reaches the upper limit, decrease the SiH4 flow by 0.2%-0.4% and the H2 flow simultaneously for every 1°C lower temperature.

8. The apparatus adaptive control method for silicon carbide epitaxial thin film growth of claim 1, wherein, The way to handle the edge defect is: For edge warping: reduce the temperature difference between the edge and the center by optimizing the power of the edge heating module, if the edge temperature is lower than the center temperature by >10℃, increase the power of the edge heating ring by 5%-8% at a time until the temperature difference is within ±5℃, if the edge temperature is too high, reduce the power and check the distance between the heating ring and the substrate; For edge polycrystalline / grain defect: install a ring-shaped porous edge airflow guide ring made of high-purity graphite material, increase the edge H2 flow rate by 10%-15%, and tilt the SiH4 / C3H8 nozzle near the edge by 5°-10° towards the center.

9. The apparatus adaptive control method for silicon carbide epitaxial thin film growth of claim 1, wherein, The way to handle the edge defect is: For edge impurity aggregation: H2 is purified by three stages of deoxidation, dehydration and metal removal, before each growth, clean the reactor wall in situ at 1600℃ for 30 minutes at a flow rate of 800-1000sccm; For uneven edge thickness: if the thickness is too thin, increase the edge H2 flow rate to 575±25sccm, if the thickness difference is still >3%, increase the edge SiH4 and C3H8 flow rate by 4%±1%; if the thickness is too thick, decrease the edge SiH4 flow rate by 4%±1% and increase the edge pumping rate, if necessary, add 1-2 small holes at the edge of the distribution plate.

Citation Information

Patent Citations

  • An apparatus and control method for silicon carbide epitaxial thin film growth

    CN113718331B

Cited By

  • Abnormity detection and in-situ repair method and system for self-healing epitaxial growth

    CN122304022A