Multi-stage antimony volatility coefficient optimization method applied to single crystal silicon antimony doping process

Abstract

The present application relates to the technical field of single crystal silicon doping, and particularly relates to a multi-stage antimony evaporation coefficient optimization method applied to a single crystal silicon antimony doping process, comprising: obtaining process operation parameters of each process stage and inputting the process operation parameters into an antimony evaporation coefficient simulation quantization space to obtain a first antimony evaporation coefficient corresponding to the process stage; obtaining a single crystal furnace structure feature and correcting the first antimony evaporation coefficient based on the single crystal furnace structure feature to obtain a second antimony evaporation coefficient; and taking an antimony content design requirement of a single crystal silicon product as a target, determining a doping ratio of each process stage according to pre-obtained antimony dopant purity information and the second antimony evaporation coefficient of each process stage. The method can comprehensively consider various influencing factors and accurately optimize, and can effectively reduce product quality fluctuations caused by inaccurate antimony evaporation coefficient control and unreasonable doping ratio.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of single crystal silicon doping, and in particular to a multi-stage antimony volatility coefficient optimization method applied to a single crystal silicon antimony doping process. BACKGROUND

[0002] In the field of single crystal silicon production, the antimony doping process is a key technology for preparing N-type single crystal silicon, which realizes semiconductor performance regulation by adding antimony elements to the silicon melt; however, antimony elements have high volatility at high temperatures, and the accurate control of their volatility coefficient directly determines the doping uniformity and product quality.

[0003] Due to the structural differences of single crystal furnaces, key parameters such as temperature gradient distribution and heat flux density will change significantly; however, existing antimony volatility coefficient optimization methods do not consider the structural characteristics of single crystal furnaces in the volatility coefficient calculation system, ignoring the implicit influence of the inherent physical properties of single crystal furnace equipment on volatility behavior; when structural parameters change due to the aging of the thermal field components or equipment maintenance, the existing method cannot correct the volatility coefficient in real time, thereby causing doping ratio deviation. SUMMARY

[0004] The present application provides a multi-stage antimony volatility coefficient optimization method applied to a single crystal silicon antimony doping process, which can consider various influencing factors and perform accurate optimization, effectively reducing product quality fluctuations caused by inaccurate antimony volatility coefficient control and unreasonable doping ratio, and can effectively solve the problems in the background art.

[0005] To achieve the above purpose, the present application provides a multi-stage antimony volatility coefficient optimization method applied to a single crystal silicon antimony doping process, comprising:

[0006] Obtain the process operation parameters of each process stage and input them into the antimony volatility coefficient simulation quantization space to obtain the first antimony volatility coefficient corresponding to the process stage;

[0007] Obtain the structural characteristics of the single crystal furnace and correct the first antimony volatility coefficient based on the structural characteristics of the single crystal furnace to obtain the second antimony volatility coefficient;

[0008] With the antimony content design requirement of single crystal silicon products as the target, determine the doping ratio of each process stage according to the pre-obtained antimony dopant purity information and the second antimony volatility coefficient of each process stage.

[0009] In one possible design, the structural characteristics of the single crystal furnace include at least heating element influence characteristics, draft tube influence characteristics and insulation layer influence characteristics.

[0010] In one possible design, the heating element influence characteristics include the number of turns of the heating coil, the heating power density distribution and the material of the heating element.

[0011] In a possible design, the shroud influence features include shroud height, inner diameter size, shroud material, and relative position of the shroud and the silicon melt.

[0012] In a possible design, the thermal insulation layer influence features include thermal conductivity of the thermal insulation material, thickness, and laying method.

[0013] In a possible design, the thermal field influence correction is performed by combining finite element fluid mechanics simulation and orthogonal experimental design to determine an antimony volatilization coefficient correction coefficient matrix under different single crystal furnace structure feature combinations, and the correction coefficient matrix is used to quantify the influence of different single crystal furnace structure features on the antimony volatilization coefficient in each process stage.

[0014] In a possible design, each matrix element in the correction coefficient matrix is calculated by the following formula:

[0015] ;

[0016] wherein, represents a correction coefficient of the i th single crystal furnace structure feature on the antimony volatilization coefficient in the j th process stage; represents a volatilization coefficient change amount measured by experiment; represents a relative deviation of an actual value of the i th single crystal furnace structure feature parameter from a design value; represents a correction function based on average temperature.

[0017] In a possible design, the calculation formula of the second antimony volatilization coefficient is as follows:

[0018] ;

[0019] wherein, represents the second antimony volatilization coefficient in the j th process stage; represents the first antimony volatilization coefficient in the j th process stage; and n represents the number of single crystal furnace structure features.

[0020] In a possible design, the process stages include at least a melt stage, a doping stage, and a crystal growth stage.

[0021] In a possible design, the process operation parameters include at least any one or a combination of two or more of temperature, pressure, gas flow rate, crystal pulling speed, and crystal rotation speed in the single crystal furnace.

[0022] The technical scheme of the present application can achieve the following technical effects:

[0023] When the structure parameters change due to the aging of the hot field components of the single crystal furnace, equipment maintenance, etc., the application can obtain the new structure characteristics of the single crystal furnace in a timely manner, and re-correct and optimize the antimony volatilization coefficient, improve the dynamic adaptability of the antimony doping process, quickly respond to equipment changes, ensure the stability of product quality, and avoid production interruption or product quality decline caused by equipment factors; by combining process operation parameter acquisition, hot field influence correction and doping ratio determination, precise control and collaborative optimization of the single crystal silicon antimony doping process are realized, and the optimization of the antimony volatilization coefficient is more accurate; in summary, the application can consider various influencing factors and perform precise optimization, effectively reduce product quality fluctuations caused by inaccurate antimony volatilization coefficient control and unreasonable doping ratio, and help improve the quality consistency of single crystal silicon products. BRIEF DESCRIPTION OF DRAWINGS

[0024] Figure 1 The logic flow chart of the multi-stage antimony volatilization coefficient optimization method applied to the single crystal silicon antimony doping process in the embodiments of the application. DETAILED DESCRIPTION

[0025] The application will be described below in conjunction with the drawings in the application.

[0026] As shown in Figure 1 The multi-stage antimony volatilization coefficient optimization method applied to the single crystal silicon antimony doping process of the application specifically includes the following steps:

[0027] Step S100, acquire the process operation parameters of each process stage and input them into the antimony volatilization coefficient simulation quantization space to obtain the first antimony volatilization coefficient corresponding to the process stage;

[0028] Step S200, acquire the structure characteristics of the single crystal furnace and perform hot field influence correction on the first antimony volatilization coefficient based on the structure characteristics of the single crystal furnace to obtain the second antimony volatilization coefficient;

[0029] Step S300, take the antimony content design requirement of the single crystal silicon product as the target, determine the doping ratio of each process stage according to the pre-acquired antimony dopant purity information and the second antimony volatilization coefficient of each process stage.

[0030] In the embodiment, the process running parameters of each process stage are obtained, and the first antimony volatilization coefficient is calculated by inputting the process running parameters into the antimony volatilization coefficient simulation quantization space, so as to realize dynamic pre-evaluation of the antimony volatilization behavior. The existing method relies on fixed parameters or empirical values, while the present application utilizes the simulation quantization space to quickly generate the initial volatilization coefficient based on real-time process data, which can effectively reflect the volatilization differences in different process stages. The first antimony volatilization coefficient is corrected by using the structural characteristics of the single crystal furnace, and the device structure parameters are converted into correction coefficients, so that the volatilization coefficient calculation is more in line with the actual working condition. Especially when the heat field assembly is aging or the device maintenance causes structural changes, the volatilization coefficient can be corrected in real time to avoid doping ratio deviation. The doping ratio is determined by taking the single crystal silicon product antimony content design requirement as the target, combining the antimony dopant purity information and the second antimony volatilization coefficient, so as to realize precise control of the doping process. The existing method is relatively extensive in doping ratio design, which is easy to cause large differences in rod head and tail resistivity, uneven doping and other problems. By quantifying the relationship between the volatilization coefficient and the product requirement, the antimony dopant addition amount of each stage can be accurately planned to ensure the performance consistency of the single crystal silicon product and reduce the waste rate caused by improper doping.

[0031] When the single crystal furnace appears heat field assembly aging, device maintenance and other situations that cause structural parameter changes, the method can timely obtain new single crystal furnace structure characteristics and re-correct and optimize the antimony volatilization coefficient. The whole antimony doping process has strong dynamic adaptability, can quickly respond to device changes, ensure the stability of product quality, and avoid production interruption or product quality decline caused by device factors. By combining process running parameter acquisition, heat field influence correction and doping ratio determination, precise control and collaborative optimization of the single crystal silicon antimony doping process are realized, so that the optimization of the antimony volatilization coefficient is more accurate and efficient. For example, when the process running parameters change or the single crystal furnace structure characteristics change, the influence of these changes on the antimony content can be compensated by timely adjusting the doping ratio, so as to realize accurate regulation and control of the performance of the single crystal silicon product. Since the method can consider various factors and optimize accurately, it can effectively reduce the product quality fluctuation caused by inaccurate antimony volatilization coefficient control and unreasonable doping ratio, and help to improve the quality consistency of the single crystal silicon product.

[0032] Specifically, in the multi-stage process of single crystal silicon antimony doping, such as melt, doping and crystal growth, the parameters such as temperature, pressure, gas flow, crystal pulling speed and crystal rotation speed show significant differences. For example, the temperature in the melt stage can reach 1420-1450℃, and the temperature in the crystal growth stage decreases to 1380-1400℃. The crystal pulling speed is dynamically adjusted according to the crystal diameter at 5-15mm / h, and the volatility of antimony is strongly sensitive to the above parameters. The existing method uses fixed coefficient or unified model across stages, which cannot capture the differential influence of parameter fluctuations in each stage on the volatilization behavior. Therefore, it is necessary to independently build the dynamic mapping relationship between process parameters and volatilization coefficient for each stage.

[0033] More specifically, the division of the process stages and the process operation parameters corresponding to each process stage are as follows:

[0034] The melt stage: the process operation parameters of this stage at least include the heating rate, the holding time, the furnace pressure, etc., and the initial influence of the temperature overshoot on the volatilization of the melt surface during the melting process of the silicon raw material is focused on;

[0035] The doping stage: the process operation parameters of this stage at least include the antimony doping agent feeding time, the inert gas flow, the stirring speed, etc., and the promoting effect of the melt surface renewal frequency caused by stirring on the volatilization rate is focused on;

[0036] The crystal growth stage: the process operation parameters of this stage at least include the monitoring of the crystal pulling speed and the crystal rotation speed, etc., and the influence of the melt flow state change caused by the solid-liquid interface movement on the volatilization path is focused on;

[0037] The process operation parameters of the above process stages are collected in real time by the built-in sensors of the single crystal furnace, including thermocouples, pressure transmitters, mass flow meters, etc.

[0038] As a preferred embodiment of the present application, the antimony volatilization coefficient simulation quantization space includes simulation quantization models corresponding to different process stages, the simulation quantization model adopts an LSTM neural network enhanced by an attention mechanism, the input is a process operation parameter sequence within a preset time of each stage, and the output is the first antimony volatilization coefficient of the current stage, for example: when the crystal growth stage pulling speed is accelerated, the model learns from historical data that the first antimony volatilization coefficient is improved by 1.2% for every 1mm / h increase in speed;

[0039] When constructing the antimony volatilization coefficient simulation quantization space, the measured antimony volatilization amount and process parameter sequence in different batches of production are collected to construct a sample training set, and the antimony volatilization rate can also be tested in a controllable environment through physical experiments, such as different temperature gradients, gas flow rates, etc., to supplement small sample scene data; the mean square error is used as the loss function to optimize the fitting ability of the model to the nonlinear volatilization behavior, such as the exponential relationship between the pressure fluctuation in the high temperature zone and the volatilization coefficient; for the unique physical mechanisms of the melt, doping and crystal growth stages, such as the temperature-driven volatilization in the melt stage and the interface flow-driven volatilization in the crystal growth stage, the corresponding models are trained respectively to avoid cross-stage parameter interference, such as the meaningless of the stirring speed in the doping stage to the melt stage model.

[0040] In this embodiment, compared with the traditional fixed coefficient method, the single-stage volatility coefficient error is controlled within the allowable range by inputting dynamic parameters in real time; the core influence parameters of each stage are independently modeled, which can accurately reflect the dominant physical mechanism of different stages and avoid systematic deviation caused by one-size-fits-all modeling; at the same time, the analog quantitative space continuously accumulates new batches of data in operation, and through the online learning mechanism, it automatically adapts to process improvement, avoiding the problem that the traditional experience model is invalid due to long-term invariability.

[0041] As a preferred embodiment of the present application, in the single crystal silicon antimony doping process, the structural features of the single crystal furnace have a decisive influence on the thermal field distribution in the furnace; the existing method does not include the structural features of the single crystal furnace in the volatility coefficient calculation system, ignores the implicit influence of the inherent physical properties of the equipment on the volatility behavior, and cannot real-time correct the volatility coefficient when the thermal field components are aged and the structural parameters are changed due to equipment maintenance, thereby causing doping ratio deviation and affecting product quality;

[0042] Specifically, the structural features of the single crystal furnace, such as heating elements, flow guide tubes and thermal insulation layers, will significantly affect the temperature gradient distribution and heat flux density in the furnace, and the above factors directly determine the volatility coefficient of antimony; for example, uneven power density distribution of the heating element will cause local temperature to be too high or too low, thereby affecting the volatilization rate of antimony; the height and inner diameter size of the flow guide tube will affect the airflow state of the melt surface, thereby affecting the diffusion and convection of antimony vapor; the material and thickness of the thermal insulation layer will affect the heat transfer and distribution, and indirectly affect the volatilization behavior of antimony;

[0043] Therefore, obtaining the structural features of the single crystal furnace and correcting the thermal field influence on the first antimony volatility coefficient based on these features to obtain the second antimony volatility coefficient can more accurately reflect the actual production of antimony volatilization, improve the accuracy of doping ratio and the stability of product quality.

[0044] More specifically, the method for obtaining the structural features of the single crystal furnace is as follows:

[0045] Step S210, using high-precision laser scanners, thermal imagers and other equipment to detect the key structural components of the single crystal furnace; for example, using a laser scanner to measure the height, inner diameter size and other geometric parameters of the flow guide tube; using a thermal imager to detect the temperature distribution of the heating element, and then calculating the heating power density distribution; using a high-temperature-resistant high-precision range finder to measure the relative position of the flow guide tube and the silicon melt; at the same time, querying the historical data of the heating element material, the thermal conductivity, the thickness and the laying method of the thermal insulation material from the equipment management system of the single crystal furnace;

[0046] Step S220, the collected heating coil turns, heating power density distribution and heating element material information are quantitatively processed, for example, the heating power density distribution is converted into a specific numerical matrix, indicating the power density size at different positions; for the heating element material, it can be classified and coded according to its thermal conductivity, emissivity and other physical properties; the height of the flow guide cylinder, the inner diameter size, the material of the flow guide cylinder and the relative position of the flow guide cylinder and the silicon melt and other information are quantified, such as the relative position of the flow guide cylinder and the silicon melt is represented by a distance value, so as to facilitate subsequent calculation and analysis; the thermal conductivity, thickness and laying method of the heat insulation material are quantified, for example, the laying method of the heat insulation material is divided into different types such as multi-layer paving and wave-shaped laying, and corresponding numerical codes are assigned.

[0047] In some embodiments of the present application, the thermal field influence correction aims to consider the influence of the structure characteristics of the single crystal furnace on the antimony volatility coefficient, by combining finite element fluid mechanics simulation with orthogonal experimental design, the correction coefficient matrix under different single crystal furnace structure characteristic combinations is determined, and then the first antimony volatility coefficient is corrected to obtain a second antimony volatility coefficient that is more in line with the actual situation; Specifically:

[0048] A thermal field model of the single crystal furnace is constructed using finite element analysis software to simulate the temperature field and flow field distribution under different structure characteristic parameter combinations, and to obtain the theoretical trend of the antimony volatility coefficient, providing theoretical guidance and expected result reference for experimental design, and reducing the blindness of experiments;

[0049] Due to the large number of structure characteristic parameters of the single crystal furnace, comprehensive experiments are time-consuming and laborious, and orthogonal experimental design can screen out key parameter combinations and obtain sufficient data information through a small number of experiments; such as the single crystal furnace has heating elements, flow guide cylinders, heat insulation layers and other structure characteristics, each characteristic has multiple variable parameters, and orthogonal experiments can select the most representative parameter combinations for experiments, reducing the amount of experiments and improving the efficiency of experiments.

[0050] Further, the correction coefficient matrix is used to quantify the influence of different structure characteristics of the single crystal furnace on the antimony volatility coefficient of each process stage, wherein the calculation formula of each matrix element is as follows:

[0051] ;

[0052] Wherein, represents the correction coefficient of the i th single crystal furnace structure characteristic to the j th process stage antimony volatility coefficient, reflecting the influence degree of the change of a specific structure characteristic on the volatility coefficient of a specific process stage;

[0053] represents the experimentally measured volatile coefficient variation; through experimental means, the i th single crystal furnace structure characteristic parameter is changed alone under the condition of controlling other variables, and the change value of the antimony volatile coefficient caused thereby is measured; for example, the number of turns of the heating coil is changed, and the change of the antimony volatile coefficient at this process stage is measured;

[0054] represents the relative deviation of the actual value of the i th single crystal furnace structure characteristic parameter from the design value; for example, if the design value of the draft tube height is h0, and the actual value is h1, then ; Z is a parameter for measuring the degree of deviation of the structure characteristic parameter from the design;

[0055] represents the correction function based on the average temperature; because the temperature has a significant effect on the volatilization of antimony, the average temperature at different process stages is different, so the correction function of the average temperature is needed to correct the correction coefficient according to the average temperature at this process stage; for example, in the high-temperature melt stage and the relatively low-temperature crystal growth stage, the same structure characteristic change may have different effects on the volatile coefficient due to the different temperatures, which plays a role in adjusting the temperature difference at different process stages;

[0056] Through the calculation of each matrix element in the above-mentioned correction coefficient matrix, the contribution degree of the deviation of each structure characteristic parameter to the change of the volatile coefficient can be quantified, while considering the influence of the temperature, that is, the change of the volatile coefficient caused by each unit of the structure characteristic parameter deviation, and then adjusting according to the temperature.

[0057] In some embodiments of the present application, the second antimony volatile coefficient is obtained by considering the comprehensive influence of multiple single crystal furnace structure characteristics on the first antimony volatile coefficient; the calculation formula of the second antimony volatile coefficient is:

[0058] ;

[0059] wherein, represents the second antimony volatile coefficient at the j th process stage; it is determined after considering the comprehensive influence of multiple structure characteristics of the single crystal furnace;

[0060] represents the first antimony volatile coefficient at the j th process stage, which is calculated by inputting the process operation parameters of each process stage into the antimony volatile coefficient simulation quantization space in step S100, and reflects the volatilization of antimony based only on the process operation parameters, but does not consider the influence of the structure characteristics of the single crystal furnace on the thermal field; for example, in the melt stage, the first antimony volatile coefficient is calculated according to the temperature, the heating rate and other operation parameters at this stage, without considering the additional effect of the heat field change caused by the structure characteristics such as heating elements and draft tubes on the volatilization of antimony;

[0061] n represents the number of single crystal furnace structure characteristics;

[0062] The calculation formula of the second antimony volatility coefficient realizes the coupling effect of multiple structure parameters in a product form. The reason for using the form of continuous multiplication is that the effects of multiple structure characteristics on the antimony volatility coefficient are mutually superimposed and jointly acted. For example, the change of the power density distribution of the heating element will affect the temperature of the thermal field, and the change of the inner diameter size of the flow guide tube will affect the gas flow field, which will all affect the antimony volatility coefficient. Through continuous multiplication, the correction effect of different structure characteristics on the first antimony volatility coefficient under the joint action can be comprehensively reflected, and the volatilization of antimony under the actual structure condition of the single crystal furnace at a specific process stage can be more accurately reflected.

[0063] In the embodiment, the correction coefficient is determined by combining experiments and simulations, which can more accurately capture the actual influence of the structure characteristics of the single crystal furnace on the antimony volatility coefficient, thereby improving the accuracy of the calculation result. By considering the comprehensive influence of multiple structure characteristics of the single crystal furnace on the antimony volatility coefficient, systematic deviation caused by ignoring the structure characteristics of the equipment in the traditional method is avoided. When the structure characteristics of the single crystal furnace change, such as thermal field component aging, equipment maintenance, etc., the second antimony volatility coefficient can be updated by recalculating the correction coefficient, so as to ensure that the corrected volatility coefficient always conforms to the actual working condition.

[0064] As a preferred embodiment of the above embodiment, in the antimony-doped single crystal silicon process, precise control of the doping ratio of antimony can ensure the performance stability and consistency of the single crystal silicon product. Antimony has high volatility at high temperatures, and its volatility coefficient is affected by multiple factors, including process operating parameters and structure characteristics of the single crystal furnace. Step S100 obtains the process operating parameters and calculates the first antimony volatility coefficient, and step S200 corrects the first antimony volatility coefficient by the thermal field influence of the structure characteristics of the single crystal furnace to obtain the second antimony volatility coefficient. However, the above steps only solve the problems of calculation and correction of the volatility coefficient, and have not yet solved how to determine a reasonable doping ratio according to these coefficients.

[0065] The existing method is relatively extensive in doping ratio design, usually relying on empirical values or fixed formulas, and cannot accurately reflect the antimony volatilization at different process stages, resulting in large differences in head and tail resistivity of the crystal bar, uneven doping, etc., which not only affects the performance consistency of the single crystal silicon product, but also increases the scrap rate. Therefore, in step S300, based on the second antimony volatility coefficient and the antimony content design requirement of the single crystal silicon product, the doping ratio of each process stage is accurately calculated to realize precise control of the doping process.

[0066] Step S310, according to the specific application scenario and performance requirement of the single crystal silicon product, the required antimony content range is determined; for example, for N-type single crystal silicon which requires high resistivity, the antimony content needs to be accurately controlled at a low level; and according to different stages of single crystal silicon production, the antimony content targets of each stage are set respectively; the process parameters and thermal field conditions are different in different stages, and the control requirements for antimony volatilization are also different.

[0067] Step S320, select appropriate antimony dopant such as antimony trioxide or antimony metal; the purity and chemical properties of different dopants will affect their dissolution and volatilization behavior in silicon melt; the purity of antimony dopant is accurately determined by chemical analysis or spectral analysis method; the purity information is the basis for calculating the doping amount, and directly affects the accuracy of doping ratio.

[0068] Step S330, based on the second antimony volatilization coefficient of each process stage, the doping ratio of each process stage is determined by taking the antimony content design requirement of the single crystal silicon product as the target through the following specific method:

[0069] Step S331, in the melting stage, assuming that the initial silicon raw material does not contain antimony, a certain amount of antimony dopant is added during the melting process, at this time the antimony content in the melt mainly depends on the addition amount of the dopant and the volatilization loss of antimony, which is expressed as:

[0070] ;

[0071] Wherein, represents the antimony content in the melt after the melting stage, usually expressed in mass concentration, reflecting the mass of antimony contained in unit mass of silicon melt during the melting process;

[0072] represents the mass of antimony dopant added in the melting stage, that is, the total amount of antimony-containing substances added to the silicon raw material;

[0073] represents the purity of antimony dopant, which refers to the mass proportion of antimony element in the antimony dopant;

[0074] represents the mass of melt silicon, that is, the total mass of silicon melt formed after the complete melting of silicon raw material at the end of the melting stage;

[0075] represents the second antimony volatilization coefficient of the melting stage;

[0076] The formula represents the actual retention of antimony content in the melt silicon after adding antimony dopant to the silicon raw material in the melting stage; when calculating, the total amount of added antimony dopant, the purity of the dopant and the volatilization loss of antimony in the melting process need to be considered; wherein the volatilization loss of antimony is represented by the second antimony volatilization coefficient; through the formula, the antimony content in the silicon melt after the end of melting can be accurately predicted or calculated, providing basic data for the subsequent doping process;

[0077] Step S332, in the doping stage, continue to add antimony dopant and stir the melt and other operations, at this time the antimony content in the melt is not only related to the newly added dopant, but also affected by the fluctuation of the original antimony content and the volatilization of antimony, which is represented as:

[0078]

[0079] Among them, represents the antimony content in the melt after the end of the doping stage;

[0080] represents the antimony content brought from the melting stage, that is, the antimony content in the melt at the end of the melting stage, which is the basis for the initial antimony content in the doping stage;

[0081] represents the second antimony volatilization coefficient in the doping stage, which reflects the volatilization loss of antimony in the silicon melt during the doping process, and is also corrected by the influence of the hot field, considering the influence of the structure characteristics and process parameters of the single crystal furnace on the volatilization of antimony;

[0082] represents the mass of the newly added antimony dopant in the doping stage, that is, the amount of antimony dopant additionally added on the basis of the melting stage;

[0083] In the doping stage, continue to add antimony dopant to the silicon melt, and stir the melt and other operations; the antimony content in this stage is composed of two parts: one is the remaining amount of the antimony content in the melting stage after volatilization loss, and the other is the contribution of the newly added antimony dopant to the antimony content in the melt after considering the volatilization loss; through the formula, the antimony content in the melt silicon at the end of the doping stage can be accurately calculated, so as to accurately control the amount of antimony doping and meet the design requirements of the single crystal silicon product on the antimony content;

[0084] Step S333, in the crystal growth stage, as the crystal grows, the antimony in the melt will gradually be brought into the crystal, and the volatilization of antimony also continues; the change of the antimony content in this stage can be described as:

[0085]

[0086] Among them, represents the antimony content in the remaining melt at the end of the crystal growth stage;

[0087] It represents the antimony content in the melt after the doping stage, which serves as the antimony content at the beginning of the crystal growth stage;

[0088] Indicates the mass of the remaining silicon melt after the crystal growth stage. As the silicon crystal grows, part of the silicon melt is consumed and crystallized into silicon crystals. This parameter is the mass of the remaining uncrystallized silicon melt;

[0089] It represents the initial mass of the silicon melt at the beginning of the crystal growth stage, that is, the total mass of the silicon melt after the doping stage is completed;

[0090] The second antimony volatility coefficient represents the second antimony volatility coefficient during the crystal growth stage. It comprehensively considers the influence of factors such as the thermal field change in the single crystal furnace and the crystal growth rate on antimony volatility during the crystal growth process, and is used to calculate the degree of antimony volatility loss during the crystal growth stage.

[0091] During the crystal growth stage, silicon from the silicon melt is gradually pulled into a single silicon crystal. As the crystal grows, the mass of silicon in the melt decreases, and the antimony content also changes accordingly. At the same time, antimony volatilization continues. This formula calculates the antimony content in the remaining melt after the crystal growth stage by taking into account the changes in silicon melt mass caused by crystal growth and the volatilization losses of antimony. This formula can accurately assess the changes in antimony content throughout the single silicon crystal growth process, helping to precisely control the antimony content of the final single silicon crystal product, ensuring that product quality meets design requirements.

[0092] Step S334: Substitute the antimony content design requirements of the single crystal silicon product into the above mathematical model, combine the second antimony volatility coefficient and the antimony dopant purity information of each process stage, and solve to obtain the antimony dopant addition amount required for each process stage, that is, determine the doping ratio.

[0093] In this embodiment, the dynamic relationship between antimony volatilization loss and doping ratio is accurately quantified through a staged mathematical model, and the changes in antimony content in the melt, doping, and crystal growth stages are coupled with the volatility coefficient and purity parameters for calculation, thereby solving the problem of uneven doping caused by the traditional empirical method that ignores the differences in volatilization in multiple stages. An iterative equation based on the second antimony volatility coefficient is used to achieve accurate inversion of the addition amount in each stage, combined with purity correction and compensation for melt mass changes to reduce the resistivity fluctuation of the crystal rod. Through closed-loop feedback of the residual melt antimony content in the crystal growth stage, the doping ratio is dynamically corrected, and the antimony concentration stability can still be maintained when the thermal field component ages.

[0094] The above shows and describes the basic principles, main features and advantages of the present application. Those skilled in the art should understand that the present application is not limited to the above-mentioned embodiments, and the above-mentioned 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 multi-stage antimony volatility coefficient optimization method for single crystal silicon antimony doping process, characterized in that: include: Obtaining process operating parameters of each process stage and inputting them into the antimony volatility coefficient simulation quantification space to obtain the first antimony volatility coefficient corresponding to the process stage; Acquiring structural characteristics of a single crystal furnace, and performing thermal field effect correction on the first antimony volatility coefficient based on the structural characteristics of the single crystal furnace to obtain a second antimony volatility coefficient; Taking the antimony content design requirement of the single crystal silicon product as the target, the doping ratio of each process stage is determined according to the pre-acquired antimony dopant purity information and the second antimony volatility coefficient of each process stage; The thermal field effect correction is carried out by combining finite element fluid dynamics simulation with orthogonal experimental design to determine the antimony volatility coefficient correction coefficient matrix under different combinations of single crystal furnace structural features. The correction coefficient matrix is ​​used to quantify the influence of different single crystal furnace structural features on the antimony volatility coefficient at each process stage. Each matrix element in the correction coefficient matrix is ​​calculated by the following formula: ； in, It represents the correction coefficient of the antimony volatility coefficient of the jth process stage due to the structural characteristics of the i-th single crystal furnace; It represents the variation of volatility coefficient measured experimentally; Indicates the relative deviation between the actual value and the design value of the structural characteristic parameter of the i-th single crystal furnace; represents the correction function based on the average temperature; The calculation formula of the second antimony volatility coefficient is: ； in, represents the second antimony volatility coefficient of the jth process stage; represents the first antimony volatility coefficient of the jth process stage; n represents the number of structural features of the single crystal furnace.

2. The multi-stage antimony volatility coefficient optimization method for single crystal silicon antimony doping process according to claim 1, characterized in that: The structural features of the single crystal furnace at least include the features affecting the heating element, the features affecting the guide tube and the features affecting the thermal insulation layer.

3. The multi-stage antimony volatility coefficient optimization method for single crystal silicon antimony doping process according to claim 2, characterized in that: The influencing characteristics of the heating element include the number of turns of the heating coil, the heating power density distribution and the material of the heating element.

4. The multi-stage antimony volatility coefficient optimization method for single crystal silicon antimony doping process according to claim 3, characterized in that: The influencing features of the guide tube include the height of the guide tube, the inner diameter size, the material of the guide tube, and the relative position of the guide tube and the silicon melt.

5. The multi-stage antimony volatility coefficient optimization method for single crystal silicon antimony doping process according to claim 4, characterized in that: The influencing characteristics of the thermal insulation layer include thermal conductivity, thickness and laying method of the thermal insulation material.

6. The multi-stage antimony volatility coefficient optimization method for single crystal silicon antimony doping process according to any one of claims 1 to 5, characterized in that: The process stages include at least a melting stage, a doping stage and a crystal growth stage.

7. The multi-stage antimony volatility coefficient optimization method for single crystal silicon antimony doping process according to any one of claims 1 to 5, characterized in that: The process operating parameters include at least one or a combination of two or more of the temperature, pressure, gas flow, crystal pulling speed and crystal rotation speed in the single crystal furnace.

Images