Antimony volatilization control method based on high temperature time regulation

Abstract

The present application relates to the technical field of single crystal silicon pulling, and particularly relates to a method for controlling antimony evaporation amount based on high-temperature time regulation, comprising: obtaining antimony evaporation amount thresholds and stage temperature characteristics of each preset temperature control stage; inputting the antimony evaporation amount thresholds and the stage temperature characteristics into a preset temperature control duration evaluation model for each temperature control stage to obtain a temperature control duration of the temperature control stage; obtaining real-time working conditions of a pulling operation of each temperature control stage, and performing evaporation influence analysis on the real-time working conditions to obtain a real-time evaporation rate; based on the real-time evaporation rate and the temperature control duration, a real-time antimony evaporation amount corresponding to the temperature control stage is calculated; in response to the real-time antimony evaporation amount being greater than the antimony evaporation amount threshold, an evaporation amount overflow value between the two is calculated to determine an evaporation amount control strategy of the current temperature control stage and / or a subsequent temperature control stage, and the evaporation amount control strategy is executed.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of single crystal silicon pulling, and in particular to a high-temperature time-regulated antimony volatilization control method. BACKGROUND

[0002] In the process of single crystal silicon pulling, antimony, as an important N-type dopant, its volatilization behavior has a key influence on the resistivity uniformity and crystal quality of single crystal silicon. Due to the extremely low segregation coefficient and high volatilization constant of antimony, the volatilization rate of antimony will fluctuate significantly with changes in parameters such as temperature, furnace pressure, and gas flow during different pulling stages such as high-temperature melt, seeding, and constant diameter. If the volatilization amount is not properly controlled, it will easily lead to problems such as excessive difference in resistivity between the head and tail of the single crystal silicon rod, and reduction of minority carrier lifetime, which seriously affects product performance.

[0003] The existing control method mainly relies on fixed process parameter settings, such as high-temperature duration range or adjustment of single variables such as furnace pressure and argon flow. However, due to the lack of a quantitative correlation model between temperature characteristics and antimony volatilization amount, it is not possible to accurately assess the reasonable temperature control time under different temperature characteristics. At the same time, there is a lack of dynamic monitoring and analysis of real-time working conditions during pulling, making it difficult to adjust the control strategy according to the real-time volatilization rate. When the actual volatilization amount exceeds the preset threshold due to factors such as equipment differences and thermal field fluctuations, the existing method cannot timely identify volatilization amount overflows and implement targeted regulation, resulting in insufficient control accuracy of single crystal silicon resistivity uniformity and limited product yield improvement. SUMMARY

[0004] The present application provides a high-temperature time-regulated antimony volatilization control method that realizes precise identification and targeted regulation of volatilization abnormalities, effectively solving the problems in the background art.

[0005] To achieve the above purpose, the present application provides a high-temperature time-regulated antimony volatilization control method, comprising:

[0006] Obtaining the antimony volatilization threshold and the stage temperature characteristics of each preset temperature control stage;

[0007] For each of the temperature control stages, input the antimony volatilization threshold and the stage temperature characteristics into a preset temperature control time evaluation model to obtain the temperature control time of that temperature control stage;

[0008] Obtaining the real-time working conditions of the pulling operation of each temperature control stage and performing volatilization influence analysis to obtain the real-time volatilization rate;

[0009] Based on the real-time volatilization rate and the temperature control time, the corresponding real-time antimony volatilization amount of the temperature control stage is calculated;

[0010] In response to the real-time antimony evaporation amount being greater than the antimony evaporation amount threshold, an evaporation amount overflow value between the two is calculated to determine an evaporation amount control strategy for the current temperature control stage and / or a subsequent temperature control stage, and the determination is performed.

[0011] In one possible design, the antimony evaporation amount threshold and the stage temperature characteristic of each preset temperature control stage are obtained, including:

[0012] The single crystal silicon pulling process and the temperature characteristic of the antimony-doped raw material are obtained, and the resistivity design requirement is obtained by extracting the resistivity characteristic of the single crystal silicon pulling process.

[0013] According to the single crystal silicon pulling process and the temperature characteristic of the antimony-doped raw material, the pulling process is divided by temperature to obtain a plurality of temperature control stages.

[0014] According to the resistivity design requirement, the antimony evaporation amount threshold and the stage temperature characteristic of each temperature control stage are determined.

[0015] In one possible design, the calculation formula of the temperature control duration evaluation model is:

[0016] ;

[0017] wherein, represents the temperature control duration of the i th temperature control stage; represents the antimony evaporation amount threshold of the i th temperature control stage; represents the process constant of the i th temperature control stage; represents the stage temperature characteristic of the i th temperature control stage; represents the antimony evaporation activation energy; represents the gas constant; represents the natural constant.

[0018] In one possible design, the real-time working condition of the pulling operation includes furnace pressure, protective gas flow, and crystal pulling speed.

[0019] In one possible design, the real-time antimony evaporation amount corresponding to the temperature control stage is calculated based on the real-time evaporation rate and the temperature control duration, including:

[0020] Based on the real-time time, the temperature control stage is divided into an occurred period and an unoccurred period;

[0021] The antimony evaporation amount of the occurred period is calculated by piecewise integration on the occurred period;

[0022] The antimony evaporation amount of the unoccurred period is calculated by rolling prediction on the unoccurred period;

[0023] Sum the antimony volatilization amount of the occurred time period and the antimony volatilization amount of the non-occurred time period to obtain the real-time antimony volatilization amount corresponding to the temperature control stage.

[0024] In a possible design, the occurred time period is segmented and integrated to calculate the antimony volatilization amount of the occurred time period, including:

[0025] The real-time working condition of the occurred time period is subjected to mutation point detection to obtain at least one working condition mutation point;

[0026] The occurred time period is divided into multiple intervals with the working condition mutation points as boundaries;

[0027] In each interval, a preset volatilization influence analysis model is used to calculate the volatilization rate of a preset step length, to obtain a volatilization rate sequence of the interval;

[0028] Based on the volatilization rate sequence, the integral antimony volatilization amount of each interval is calculated, and the antimony volatilization amounts of all intervals are summed to obtain the antimony volatilization amount of the occurred time period.

[0029] In a possible design, the integral volatilization amount of each interval is calculated, and the calculation formula is as follows:

[0030]

[0031] wherein, represents the integral volatilization amount of the i th interval in the occurred time period; represents the preset step length; represents the number of the volatilization rate sequence in the i th interval; represents the k th volatilization rate in the volatilization rate sequence in the i th interval.

[0032] In a possible design, in the process of dividing the occurred time period into multiple sub-intervals with the mutation points as boundaries, if the interval between adjacent mutation points is less than a preset interval threshold, the adjacent mutation points are combined into the same interval.

[0033] In a possible design, the non-occurred time period is subjected to rolling prediction to calculate the antimony volatilization amount of the non-occurred time period, including:

[0034] A time series prediction model is used to predict the trend of the working condition parameter, the input data of the time series prediction model being the working condition parameter in a preset time window; and the output data being the working condition parameter in a future preset time length;

[0035] Every time the preset time length is advanced, the input data is updated with the latest data, and the working condition parameter in the future preset time length is re-predicted;

[0036] ​A time series prediction model is used to stepwise predict the volatile amount in the non-occurrence period, and the prediction results are sorted to obtain a working condition parameter prediction sequence;

[0037] According to the working condition parameter prediction sequence, a future volatile rate sequence is generated;

[0038] According to the future volatile rate sequence, the antimony volatile amount in the non-occurrence period is calculated.

[0039] In a possible design, the formula for calculating the antimony volatile amount in the non-occurrence period is:

[0040] ;

[0041] wherein, represents the antimony volatile amount in the non-occurrence period; j represents the data amount in the future volatile rate sequence; represents the kth volatile rate in the future volatile rate sequence; represents the preset time length.

[0042] The technical scheme of the present application can achieve the following technical effects:

[0043] By combining the antimony volatile amount threshold value and the stage temperature characteristics, the preset temperature control time length evaluation model is used to output the accurate temperature control time length, the dynamic quantitative correlation between the temperature characteristics and the antimony volatile amount is established, and the temperature control time length is changed from the experience setting to the model driving; the real-time production working condition is collected and the volatile influence is analyzed, the static preset parameter is combined with the dynamic real-time data; the real-time volatile rate and the temperature control time length are linked to calculate, forming a dynamic control chain of monitoring, analysis, prediction and adjustment; compared with the existing lag mode of after-detection and manual adjustment, the present method can give an early warning before the volatile amount deviates from the threshold value, reducing the crystal quality defects caused by response delay; through the comparison and overflow calculation of the real-time volatile amount and the threshold value, the accurate identification and targeted regulation of the volatile amount anomaly are realized. BRIEF DESCRIPTION OF DRAWINGS

[0044] Figure 1 The present application is described below in conjunction with the drawings in the present application. DETAILED DESCRIPTION

[0045] The present application is described below in conjunction with the drawings in the present application.

[0046] As Figure 1 shown, the antimony volatile amount control method based on high-temperature time regulation of the present application specifically includes the following steps:

[0047] Step S100, obtaining the antimony volatile amount threshold value and the stage temperature characteristics of each preset temperature control stage;

[0048] In step S200, the antimony volatilization threshold and the stage temperature feature are input into a preset temperature control duration evaluation model for each temperature control stage to obtain a temperature control duration of the temperature control stage, wherein the temperature control duration represents a duration for maintaining the stage temperature feature in the temperature control stage.

[0049] In step S300, a real-time working condition of the drawing operation in each temperature control stage is obtained, and a volatilization influence analysis is performed to obtain a real-time volatilization rate.

[0050] In step S400, a real-time antimony volatilization amount corresponding to the temperature control stage is calculated based on the real-time volatilization rate and the temperature control duration.

[0051] In step S500, in response to the real-time antimony volatilization amount being greater than the antimony volatilization threshold, a volatilization overflow value between the two is calculated to determine a volatilization control strategy for the current temperature control stage and / or a subsequent temperature control stage, and the volatilization control strategy is executed.

[0052] In the embodiment, by combining the antimony volatilization threshold with the stage temperature feature, a precise temperature control duration is output by using a preset temperature control duration evaluation model, a rough mode of combining existing fixed process parameters with single variable adjustment is changed, a dynamic quantitative correlation between the temperature feature and the antimony volatilization amount is established, and the temperature control duration is changed from an experience setting to a model driving. For example, in the high-temperature molten material stage, the model can automatically calculate an optimal holding time that can ensure uniformity of doping and avoid excessive volatilization according to a real-time temperature curve, and the problem of temperature control blindness caused by lack of quantitative correlation is solved.

[0053] By real-time acquisition of the drawing operation working condition and analysis of the volatilization influence, static preset parameters are combined with dynamic real-time data. The existing method only relies on preset thresholds and cannot cope with sudden situations such as device differences or thermal field fluctuations. For example, when the argon flow fluctuates due to device aging in a batch of drawing, the temperature control duration can be adjusted in real time to avoid volatilization overflow caused by uncaught device differences. The linkage calculation of the real-time volatilization rate and the temperature control duration forms a dynamic control chain of monitoring, analysis, prediction, and adjustment. Compared with the existing lag mode of post-detection and manual adjustment, the method can provide early warning before the volatilization amount deviates from the threshold, and reduce crystal quality defects caused by response delay.

[0054] By comparing the real-time evaporation amount with the threshold value and calculating the overflow value, the precise identification and targeted regulation of the evaporation amount anomaly are realized. The existing method lacks quantitative evaluation of the overflow value, and often uses a one-size-fits-all adjustment, which may exacerbate the evaporation imbalance in other stages. According to the size of the overflow value and the characteristics of the stage, the method dynamically selects the regulation strategy to avoid the limitations of single variable adjustment. For example, if the evaporation amount exceeds the threshold value due to excessively high temperature in the seed crystal stage, the overflow value can be compensated by reducing the temperature rise rate in the constant diameter stage instead of blindly extending the high temperature time, thereby balancing the evaporation amount distribution in each stage. The quantitative overflow value calculation and multi-stage strategy cooperate to form an adaptive control network. Each stage is no longer an independent control unit, but forms a global optimization mechanism through overflow value transmission.

[0055] In some embodiments of the present application, for step S100, the following implementation is specific:

[0056] Step S110, obtain the temperature characteristics of the single crystal silicon pulling process and the antimony-doped raw material, and extract the resistivity characteristics of the single crystal silicon pulling process to obtain the resistivity design requirements, specifically;

[0057] Step S111, obtain the temperature characteristics of the single crystal silicon pulling process. The single crystal silicon pulling process involves multiple stages, including high-temperature melt, seed crystal, constant diameter, and end, etc. Each stage has specific temperature requirements. For example, in the high-temperature melt stage, the silicon material needs to be heated to about 1420℃ to completely melt it. In the seed crystal stage, the temperature is slightly reduced to ensure stable growth of the crystal. In the constant diameter stage, the temperature needs to be accurately controlled to maintain uniform growth of the crystal. The above temperature characteristics are the basic parameters of the single crystal silicon pulling process, which can be obtained through process files or actual measurement data.

[0058] Step S112, obtain the temperature characteristics of the antimony-doped raw material. Antimony, as an N-type dopant, will volatilize at high temperatures. Its volatilization rate is closely related to temperature. Generally, the higher the temperature, the faster the volatilization rate of antimony. Through experiments or literature, the volatilization characteristic curve of antimony at different temperatures is obtained.

[0059] Step S113, resistivity characteristic extraction. According to the specifications of the single crystal silicon product, such as the target resistivity range and uniformity requirements of N-type silicon wafer, the key indicators of resistivity design are extracted, such as the head-to-tail resistivity difference threshold, the whole rod resistivity fluctuation range, etc. Combined with industry standards or customer requirements, the resistivity design requirements are converted into control targets for antimony evaporation amount, such as the total evaporation amount needs to be controlled within a certain proportion of the initial doping amount to ensure that the final resistivity meets the standards.

[0060] Step S120, according to the single crystal silicon drawing process and the temperature characteristics of the antimony-doped raw material, the drawing process is divided into temperature control stages, and a plurality of temperature control stages are obtained; combining the temperature characteristics of the single crystal silicon drawing process and the volatilization characteristics of antimony obtained in step S110, the drawing process is divided into a plurality of temperature control stages; for example, according to the temperature range, the drawing process can be divided into the following stages:

[0061] High-temperature melt stage: the temperature range is 1400-1450℃, and the main purpose of this stage is to completely melt the silicon material, and the volatilization rate of antimony in this stage is relatively high;

[0062] Crystal pulling stage: the temperature range is 1350-1400℃, and the temperature of this stage is slightly lower to ensure stable growth of the crystal, and the volatilization rate of antimony is relatively reduced;

[0063] Constant diameter stage: the temperature range is 1300-1350℃, and the temperature of this stage needs to be accurately controlled to maintain uniform growth of the crystal, and the volatilization rate of antimony is relatively stable;

[0064] Finishing stage: the temperature range is 1250-1300℃, and the temperature of this stage is further reduced to complete the finishing work of the crystal, and the volatilization rate of antimony is relatively low;

[0065] By dividing the drawing process into a plurality of temperature control stages, the volatilization amount of antimony can be more finely controlled, and in each stage, a corresponding temperature control strategy can be developed according to the temperature characteristics and the volatilization characteristics of antimony, thereby improving the uniformity of the resistivity of the single crystal silicon and the quality of the crystal.

[0066] Step S130, according to the resistivity design requirement, determine the antimony volatilization amount threshold and the stage temperature characteristics of each temperature control stage; according to the resistivity design requirement determined in step S110, the total allowable volatilization amount is allocated to each temperature control stage through reverse engineering; for example: assuming that the head-to-tail resistivity difference of the single crystal silicon rod needs to be ≤5%, the maximum volatilization amount allowed in each stage is calculated through the relationship between the doping concentration and the resistivity, such as the high-temperature melt stage allows volatilization amount to account for 60% of the total volatilization amount, the constant diameter stage accounts for 30%, etc.; combined with the volatilization rate characteristics of each stage, such as the high-temperature stage volatilizes fast, and a differentiated threshold is set: because the volatilization amount has a greater impact on the total doping concentration, the threshold of the high-temperature stage is more stringent, and the threshold of the low-temperature stage is relatively relaxed;

[0067] In step S120, the drawing process has been divided into multiple temperature control stages, each stage has its own specific temperature range and temperature change law; for example, in the high-temperature melt stage, the temperature range is 1400℃~1450℃, and the temperature change law may be first quickly heating up to 1450℃, and then slowly cooling down to 1400℃ and maintaining for a period of time; in the seeding stage, the temperature range is 1350℃~1400℃, and the temperature change law may be slowly cooling down from 1400℃ to 1350℃ and maintaining stability, etc.; the stage temperature characteristics are set for each temperature control stage, and any one or more of the target temperature value, temperature fluctuation range and heating / cooling rate are coupled.

[0068] In some embodiments of the present invention, the antimony volatilization threshold and stage temperature characteristics of each temperature control stage are obtained in step S100. Then, for each temperature control stage, the temperature control duration is calculated using a temperature control duration evaluation model. The temperature control duration evaluation model quantifies the relationship between temperature, volatilization, and time through a specific formula to achieve accurate temperature control duration calculation. Specifically, the calculation formula of the temperature control duration evaluation model is:

[0069] ;

[0070] in, represents the temperature control duration of the i-th temperature control stage, that is, the time during which the specific temperature characteristics are maintained in this temperature control stage. It is the control parameter output by the temperature control duration evaluation model; The antimony volatilization threshold value in the i-th temperature control stage is determined in step S130 according to the resistivity design requirements and represents the maximum antimony volatilization amount allowed in this temperature control stage; It represents the process constant of the i-th temperature control stage, reflecting the comprehensive influence of other process conditions on the volatilization rate except temperature in this stage. It is determined by the previous process calibration. Without considering the influence of temperature on the volatilization rate, Used to determine the basic volatilization rate, its physical meaning is the basic volatilization amount per unit time; represents the stage temperature characteristics of the i-th temperature control stage; It represents the activation energy of antimony volatilization, a physical quantity that characterizes the difficulty of antimony volatilization. It is determined by the material properties and reflects the energy barrier of the volatilization process. Represents the gas constant, a universal physical constant, used as a conversion factor in the formula relating temperature to energy; Represents a natural constant.

[0071] The above formula establishes a strong correlation between temperature and volatilization rate through the exponential term, that is, the higher the temperature, the higher the exponential term. The greater the value of the numerator, the greater the denominator, resulting in a decrease in the temperature control duration, indicating that the antimony volatilization rate is fast at the high temperature stage, and the time required to reach the volatilization threshold is shorter; otherwise, the low temperature stage is increased, which conforms to the physical law of antimony volatilization.

[0072] In this embodiment, the calculated temperature control duration will be used to guide the actual drawing process, ensuring that the volatilization amount of antimony is accurately controlled at each temperature control stage; the temperature control duration is calculated based on the preset antimony volatilization threshold and temperature characteristics, which can help operators more accurately control the temperature during the drawing process to avoid excessive volatilization or insufficient volatilization of antimony.

[0073] In some embodiments of the present application, during the single crystal silicon drawing process, real-time working condition data related to the antimony volatilization rate covers multiple aspects, which needs to be accurately collected by various sensors. It should be noted that the temperature working condition is not included in the drawing operation real-time working condition, because the temperature is controlled according to the stage temperature characteristics determined in step S130. Therefore, the drawing operation real-time working condition includes:

[0074] Furnace pressure: the pressure value in the furnace is obtained in real time through a pressure sensor; the change of the furnace pressure will affect the movement state of gas molecules and their interaction with antimony, thereby affecting the volatilization of antimony; for example, excessively high furnace pressure may inhibit the volatilization of antimony, while excessively low furnace pressure may accelerate the volatilization of antimony;

[0075] Gas flow: a flow meter is used to accurately measure the flow of protective gas and other gases that may participate in the reaction into the furnace; changes in gas flow will change the gas environment in the furnace, affecting the diffusion and volatilization process of antimony atoms; for example, appropriately increasing the argon flow may dilute other components in the furnace that may inhibit the volatilization of antimony to some extent, thereby affecting the volatilization rate of antimony;

[0076] Crystal pulling speed: the control system of the crystal pulling equipment can feedback the speed information of the crystal pulling in real time; different crystal pulling speeds will cause changes in the state of the silicon melt surface, thereby affecting the concentration distribution and volatilization of antimony on the melt surface; faster crystal pulling speed may cause faster update of the melt surface, affecting the accumulation and volatilization of antimony atoms.

[0077] Further, the collected original real-time working condition data may have noise interference, data anomalies and other problems, which need to be preprocessed to improve the accuracy and reliability of the data; specifically, for continuous change analog signal data such as furnace pressure and gas flow, digital filtering algorithm is used to remove high-frequency noise and smooth the data curve, so that the data can better reflect the real working condition changes; through statistical analysis method or threshold judgment based on historical data, the abnormal values in the data are identified; for abnormal values that deviate obviously from the normal range, methods such as rejection, interpolation, etc. can be used for processing according to the specific circumstances to ensure the accuracy of subsequent analysis.

[0078] Furthermore, based on the collected and preprocessed real-time operating data, a volatilization impact analysis model was established to evaluate the impact of various factors on the antimony volatilization rate. Taking into account the complex nonlinear relationship between the various influencing factors, a neural network model was constructed with furnace pressure, gas flow rate, and crystal pulling speed as input layer nodes and the antimony volatilization rate as the output layer node. The network was trained using a small amount of training sample data, and the weights and thresholds were adjusted to achieve an approximate assessment of the antimony volatilization rate.

[0079] Specifically, a large amount of historical data under different working conditions is collected, including furnace pressure, gas flow, crystal pulling speed and the corresponding antimony volatilization rate measurement values; the volatilization impact analysis model automatically extracts the complex nonlinear relationship between input parameters (furnace pressure, gas flow, crystal pulling speed) and output parameters (antimony volatilization rate) by learning from a large amount of historical data; for example, a certain combination of furnace pressure and gas flow has a specific influence pattern on the antimony volatilization rate, which is difficult to accurately describe with a simple mathematical formula, but the neural network can capture this relationship by learning the patterns in the data; the activation function is used to realize the nonlinear mapping from input to output, so that the model can handle the complex interactions between the input parameters and more accurately reflect the changing law of the antimony volatilization rate in actual conditions.

[0080] In some embodiments of the present invention, since the volatilization rate may change dynamically over time in actual production, such as due to furnace pressure fluctuations, gas flow adjustment, etc., it is necessary to adopt a segmented integration combined with a rolling prediction method to divide the temperature control stage into a period in which the volatilization has occurred and a period in which the volatilization has not occurred, and calculate the volatilization amount separately and then sum them up.

[0081] Perform segmented integration for the period of occurrence and calculate the antimony volatilization amount during the period of occurrence. The specific implementation is as follows:

[0082] The sliding window variance analysis method is used to detect mutation points in real-time operating condition data such as furnace pressure, gas flow, and crystal pulling speed. If the furnace pressure drops suddenly and exceeds the threshold, it is determined to be a condition mutation point.

[0083] The mutation point is used as the boundary to divide the occurred period into multiple sub-intervals. If the interval between adjacent change points is less than 5 seconds, they are merged into the same interval.

[0084] In each interval, the volatility impact analysis model of step S300 is called to calculate the volatility rate of a preset step length to obtain the volatility rate sequence of the interval;

[0085] Calculate the integrated volatility for each interval using the following formula:

[0086] ;

[0087] in, represents the integrated volatilization amount of the i-th interval in the occurred period; represents the preset step length; represents the number of volatilization rate sequences in the i-th interval; represents the k-th volatilization rate in the volatilization rate sequence in the i-th interval;

[0088] The antimony volatilization amount calculation formula of the occurred period is:

[0089]

[0090] wherein, represents the antimony volatilization amount of the occurred period, and n represents the number of intervals.

[0091] The volatilization amount of the non-occurred period is predicted, the time series prediction model is used to predict the working condition parameter trend, the input data of the time series prediction model is the working condition parameter in the last 60 seconds; the output data is the working condition parameter in the future preset time length, that is, the furnace pressure, the gas flow and the pulling speed; every time the preset time length is advanced, the input window is updated with the latest data, and the working condition parameter in the future preset time length is predicted again;

[0092] The time series prediction model is used to stepwise predict the volatilization amount of the non-occurred period, and the prediction results are sorted to obtain a working condition parameter prediction sequence;

[0093] The working condition parameter prediction sequence is input into the neural network model of step S300, and a future volatilization rate sequence is output;

[0094] The antimony volatilization amount of the non-occurred period is calculated by the formula:

[0095] ;

[0096] wherein, represents the antimony volatilization amount of the non-occurred period; and j represents the data amount in the future volatilization rate sequence. represents the k-th volatilization rate in the future volatilization rate sequence; represents the preset time length.

[0097] The real-time antimony volatilization amount calculation formula of the temperature control stage is:

[0098] ;

[0099] wherein, represents the real-time antimony volatilization amount of the temperature control stage.

[0100] In this embodiment, by using the method of piecewise integration combined with rolling prediction, the temperature control stage is divided into the occurred period and the unoccurred period, and the volatilization amount of antimony is calculated and predicted respectively, so that the actual volatilization amount of antimony in each temperature control stage can be more accurately evaluated; the sliding window variance analysis method is used to detect the working condition mutation point in the occurred period, and the volatilization rate sequence is calculated in each interval by using the volatilization influence analysis model, so that the volatilization amount of antimony in the occurred period can be more accurately calculated; and the time series prediction model is used to predict the trend of the working condition parameters in the unoccurred period, and the future volatilization rate sequence is output by using the neural network model, so that the volatilization amount of antimony in the unoccurred period can be predicted, and the prediction accuracy of the volatilization amount of antimony is improved; since the rolling prediction method is used, the input window is updated with the latest data every time the preset time length is advanced, so that the change of the working condition parameters in the production process can be responded in real time, the prediction of the volatilization amount and the future working condition can be adjusted in time, the control strategy of the production process can be more closely matched with the actual working condition, and the adaptability of the production process to various changes is enhanced.

[0101] In some embodiments of the present application, when the real-time volatilization amount of antimony is greater than the volatilization amount threshold of antimony, the difference between the two, i.e. the volatilization amount overflow value, is calculated; according to the size of the volatilization amount overflow value and the situation of the current drawing stage, a corresponding control strategy is determined; the control strategy includes adjusting the parameters of the current temperature control stage, and if the current temperature control stage cannot completely eliminate the influence of the volatilization amount overflow value, or in order to avoid the volatilization amount exceeding the standard further in the subsequent stage, the process parameters of the subsequent temperature control stage need to be adjusted.

[0102] The strategy of adjusting the parameters of the current temperature control stage includes:

[0103] Temperature adjustment: if the temperature has a greater influence on the volatilization rate of antimony, and the temperature in the current stage is within the adjustable range, the temperature can be appropriately reduced to slow down the volatilization rate of antimony; for example, in the high-temperature melt stage, if the temperature is slightly higher than the set value and the volatilization amount overflow value is large, the temperature can be reduced by 5-10°C, but attention should be paid to that the temperature adjustment should not affect the melting effect of the silicon material;

[0104] Furnace pressure adjustment: according to the relationship between the furnace pressure and the volatilization rate of antimony, the furnace pressure is appropriately increased to inhibit the volatilization of antimony; for example, if the furnace pressure is lower than the normal range, which leads to excessive volatilization, the furnace pressure can be slowly increased to an appropriate level, but attention should be paid to that the furnace pressure should not be too high to adversely affect the equipment and crystal growth;

[0105] Gas flow adjustment: the flow of protective gas is increased or decreased to change the gas environment in the furnace, so as to affect the volatilization of antimony; for example, when it is found that the volatilization amount overflow value is caused by insufficient gas flow, the argon flow can be appropriately increased, but it should be ensured that the change of the gas flow will not cause instability of the thermal field.

[0106] The strategy of adjusting the parameters of the subsequent temperature control stage includes:

[0107] Temperature control duration: According to the current stage of the volatile value overflow, predict the possible volatile amount of the subsequent stage, and adjust the temperature control duration of the subsequent stage accordingly; for example, if the current stage volatile value overflow is large, the temperature control duration of the subsequent stage may need to be shortened to ensure that the total volatile amount of antimony in the entire drawing process is within a reasonable range;

[0108] Temperature setting: optimize the temperature setting of the subsequent stage to better control the volatilization of antimony; for example, in the crystal pulling stage, if the volatile amount of the previous stage is too much, the temperature setting value of the crystal pulling stage can be appropriately reduced, but the crystal should be able to grow stably;

[0109] Other parameters: according to the actual situation, adjust the furnace pressure, gas flow and other parameters of the subsequent stage to realize precise control of the volatile amount of antimony.

[0110] After determining the control strategy, send the adjusted parameters to the control system of the crystal pulling equipment in a timely manner to ensure that the equipment operates according to the new parameters; during the execution of the control strategy, the real-time working condition data and the change of the antimony volatilization rate should be closely monitored, and the control strategy should be further adjusted according to the actual situation to ensure that the volatile amount of antimony is always controllable.

[0111] In this embodiment, the current and subsequent stage parameters are flexibly adjusted according to the volatile value overflow, which is targeted. The current temperature control stage is adjusted by temperature, furnace pressure and gas flow to timely suppress excessive volatilization of antimony; at the same time, in order to avoid further exceeding the volatile amount in the subsequent stage, the temperature control duration, temperature and other parameters of the subsequent temperature control stage are optimized to control the volatile amount from the whole, avoid the defects of traditional method of lagging behind in overflow value control, accurately control the volatile amount of antimony, effectively reduce the difference of single crystal silicon head and tail resistivity, and improve the crystal quality and product yield.

[0112] Some steps in the above method embodiments can be equivalently replaced by other possible steps. Alternatively, some steps in the method embodiments can be optional and can be deleted in some use scenarios. Alternatively, other possible steps can be added to the method embodiments. In addition, each method embodiment can be implemented separately or in combination.

[0113] The basic principles, main features and advantages of the present application are shown and described above. Those skilled in the art should understand that the present application is not limited to the above embodiments, and the above embodiments and descriptions in the specification are only to illustrate the principles of the present application. Without departing from the spirit and scope of the present application, various changes and improvements can be made to the present application, and these changes and improvements all fall within the scope of the claimed present application. The scope of protection of the present application is defined by the appended claims and their equivalents.

Claims

1. A method for controlling antimony volatilization based on high temperature time regulation, characterized in that: include: Obtaining the antimony volatilization threshold and stage temperature characteristics of each preset temperature control stage; For each temperature control stage, the antimony volatilization threshold and the stage temperature characteristics are input into a preset temperature control duration evaluation model to obtain the temperature control duration of the temperature control stage; Obtaining the real-time working conditions of the drawing operation in each temperature control stage, and performing volatilization impact analysis thereon to obtain the real-time volatilization rate; Based on the real-time volatilization rate and the temperature control time, the real-time antimony volatilization amount corresponding to the temperature control stage is calculated; In response to the real-time antimony volatilization amount being greater than the antimony volatilization amount threshold, calculating a volatilization amount overflow between the two, thereby determining and executing a volatilization amount control strategy for a current temperature control stage and / or a subsequent temperature control stage; Obtain the antimony volatilization threshold and stage temperature characteristics of each preset temperature control stage, including: Obtaining the temperature characteristics of a single crystal silicon pulling process and an antimony-doped raw material, and extracting resistivity characteristics of the single crystal silicon pulling process to obtain resistivity design requirements; According to the single crystal silicon pulling process and the temperature characteristics of the antimony-doped raw material, the pulling process is divided by temperature to obtain multiple temperature control stages; Determining the antimony volatilization threshold and stage temperature characteristics of each temperature control stage according to the resistivity design requirements; The calculation formula of the temperature control time evaluation model is: ； in, Indicates the temperature control duration of the i-th temperature control stage; represents the antimony volatilization threshold value in the i-th temperature control stage; It represents the process constant of the i-th temperature control stage, reflecting the comprehensive influence of other process conditions on the volatilization rate except temperature in this stage; represents the stage temperature characteristics of the i-th temperature control stage; represents the activation energy of antimony volatilization; represents the gas constant; Represents a natural constant.

2. The method for controlling antimony volatilization based on high temperature time regulation according to claim 1, characterized in that: The real-time working conditions of the pulling operation include furnace pressure, protective gas flow rate and crystal pulling speed.

3. The method for controlling the volatilization amount of antimony based on high temperature time regulation according to claim 2, characterized in that: Based on the real-time volatilization rate and the temperature control time, the real-time antimony volatilization amount corresponding to the temperature control stage is calculated, including: Based on real-time time, the temperature control stage is divided into the period in which the temperature control has occurred and the period in which the temperature control has not occurred; Performing segmented integration on the period of occurrence to calculate the antimony volatilization amount in the period of occurrence; Performing a rolling forecast for the non-occurrence period to calculate the antimony volatilization amount during the non-occurrence period; The antimony volatilization amount during the period in which antimony volatilization has occurred and the antimony volatilization amount during the period in which antimony volatilization has not occurred are summed to obtain the real-time antimony volatilization amount corresponding to the temperature control stage.

4. The method for controlling the volatilization amount of antimony based on high temperature time regulation according to claim 3, characterized in that: Performing segmented integration on the period of occurrence to calculate the antimony volatilization amount of the period of occurrence, including: Performing mutation point detection on the real-time operating condition during the period of occurrence to obtain at least one operating condition mutation point; Divide the occurred period into multiple intervals based on the boundary of the operating condition mutation point; In each interval, the volatilization rate of the preset step length is calculated using the preset volatilization impact analysis model to obtain the volatilization rate sequence of the interval; Based on the volatilization rate sequence, the integrated antimony volatilization amount of each interval is calculated, and the antimony volatilization amounts of all intervals are summed up to obtain the antimony volatilization amount of the occurred period.

5. The method for controlling the volatilization amount of antimony based on high temperature time regulation according to claim 4, characterized in that: Calculate the integrated volatility for each interval using the following formula: ； in, represents the integrated volatility of the ith interval within the period of occurrence; Indicates the preset step size; represents the number of volatilization rate sequences in the i-th interval; represents the kth volatility rate in the volatility rate sequence in the i-th interval.

6. The method for controlling the volatilization amount of antimony based on high temperature time regulation according to claim 4, characterized in that: In the process of dividing the occurred period into multiple sub-intervals with the mutation point as the boundary, if the interval between adjacent change points is less than the preset interval threshold, they are merged into the same interval.

7. The method for controlling antimony volatilization based on high temperature time regulation according to claim 5, characterized in that: Perform rolling forecasts for the non-occurrence period and calculate the antimony volatilization amount during the non-occurrence period, including: The time series prediction model is used to predict the trend of operating parameters. The input data of the time series prediction model are the operating parameters within the most recent preset time window; the output data are the operating parameters within the future preset time period; Every time the preset time is advanced, the input data is updated with the latest data, and the operating parameters within the future preset time are re-predicted; The above-mentioned time series prediction model is used to make a step-by-step prediction of the volatility during the period when no volatility occurs, and the prediction results are sorted to obtain the working condition parameter prediction sequence; Generate future volatilization rate sequence based on the prediction sequence of operating condition parameters; Based on the future volatilization rate series, calculate the antimony volatilization amount in the period when no volatilization occurs.

8. The method for controlling the volatilization amount of antimony based on high temperature time regulation according to claim 7, characterized in that: The formula for calculating the antimony volatilization amount during the period without antimony volatilization is: ； in, represents the amount of antimony volatilization during the period when no volatilization occurs; j represents the amount of data in the future volatilization rate series; represents the kth volatility rate in the future volatility rate sequence; Indicates the preset duration.

Images