Molecular beam epitaxy system gallium source furnace temperature optimization method

Abstract

The invention provides a temperature optimization method for a gallium source furnace of a molecular beam epitaxy system, and relates to the technical field of semiconductor manufacturing. The method comprises the following steps: under a plurality of different filament temperature difference conditions, epitaxially growing epitaxial layers with preset structures to obtain a plurality of epitaxial wafers corresponding to the temperature difference conditions; measuring and acquiring surface defect data of the epitaxial wafer; establishing a model for representing the relationship between the defect data and the filament temperature difference; calculating a constant value in the model through numerical fitting; and substituting the constant value into the model, calculating and solving the minimum value of y in the model, and taking the filament temperature difference corresponding to the minimum value of y as the filament temperature difference of the gallium source furnace during next epitaxial growth. According to the method, the model is established, the model constant value is solved, and the filament temperature difference corresponding to the minimum surface defect is obtained by using the model, so that the filament temperature difference of the gallium source furnace is optimized and the surface defect is reduced. In mass production, the measurement results are continuously added to fitting data of the model, so that the process conditions can be continuously improved.

Description

Technical Field

[0001] The invention relates to the technical field of semiconductor manufacturing, and in particular to a method for optimizing the temperature of a gallium source furnace in a molecular beam epitaxy system. Background Art

[0002] In the large-scale production of molecular beam epitaxy (MBE), the number of defects per unit area on the surface of the grown epitaxial wafer is a key indicator of epitaxial wafer quality. Epitaxial wafer surface defects will deteriorate the performance of semiconductor devices prepared using epitaxial wafers and reduce the yield of semiconductor devices. Therefore, in the molecular beam epitaxy growth stage, it is expected to obtain epitaxial wafers with low surface defects.

[0003] Whether growing materials such as GaAs, AlGaAs or InGaP on a gallium arsenide substrate, or growing materials such as InGaAs on an indium phosphide substrate, the gallium molecular beam provided by the gallium source furnace is an indispensable molecular beam source material. The temperature control process conditions of the gallium source furnace seriously affect the number of surface defects of gallium-containing epitaxial materials. For conventional mass production molecular beam epitaxy systems, the gallium source furnace usually includes two upper and lower heating filaments for heating the crucible. The heating temperatures of the upper and lower heating filaments are independently controlled, and the upper heating filament directly affects the temperature at the crucible opening. Since the gallium molecular beam is easy to condense at the opening of the source furnace crucible and form gallium droplets, the gallium droplets accumulated near the opening edge of the source furnace crucible will affect the rate stability of the gallium molecular beam provided by the gallium source furnace. In addition, if the gallium droplets are deposited on the epitaxial wafer, defects will be formed on the surface of the epitaxial wafer, deteriorating the quality and yield of the epitaxial wafer. Therefore, it is usually necessary to ensure that the crucible opening has a higher temperature to limit condensation and gallium droplet formation. However, if the temperature at the crucible opening is too high, the excessive heat radiation at the crucible opening will cause impurities to be generated in the surrounding environment of the growth chamber. If these impurities reach the surface of the epitaxial wafer, the surface defects of the epitaxial wafer will increase. Therefore, it is necessary to provide a method to optimize the temperature difference between the upper and lower filaments of the gallium source furnace to reduce the number of surface defects of the epitaxial wafer containing gallium epitaxial material. Summary of the invention

[0004] The object of the present invention is to provide a method for optimizing the temperature of a gallium source furnace in a molecular beam epitaxy system in view of the above-mentioned deficiencies in the prior art, so as to solve the problem of reducing the number of surface defects of epitaxial wafers.

[0005] To achieve the above purpose, the technical solution adopted by the present invention is as follows: The present invention provides a method for optimizing the temperature of a gallium source furnace of a molecular beam epitaxy system. The gallium source furnace comprises a lower heating filament located at the bottom of a crucible of the source furnace and an upper heating filament located at the top of the crucible of the source furnace. The beam rate of gallium used in the molecular beam epitaxy is constant and equal to a preset beam rate. The method is used to optimize the filament temperature difference of the gallium source furnace. The filament temperature difference is equal to the temperature of the upper heating filament minus the temperature of the lower heating filament. The method comprises the following steps: S1. Under a plurality of different filament temperature difference conditions, epitaxially grow epitaxial layers of a preset structure on a substrate of the same specification, respectively, to obtain a plurality of epitaxial wafers corresponding to the filament temperature difference conditions, wherein the number of the plurality of different filament temperature differences is a preset number, and the preset number is greater than or equal to 6, and the value range of the filament temperature difference is greater than or equal to 0° C. and less than or equal to 200° C.; S2. Measure and obtain surface defect data of the plurality of epitaxial wafers, wherein the surface defect data includes the number of defects per unit area on the surface of the epitaxial wafers; S3. Establish a defect model for characterizing the relationship between surface defect data and filament temperature difference. The defect model is as follows: , Wherein, y represents the number of defects per unit area of ​​the epitaxial wafer surface, x represents the filament temperature difference, coefficient k1 and coefficient k2 are both constants greater than 0, base a1 is a constant greater than 0 and less than 1, base a2 is a constant greater than 1, and c is a constant; S4. Using the surface defect data, by numerical fitting, calculate the values ​​of all constants in the defect model; S5. Substitute the values ​​of all the constants obtained into the defect model, calculate and solve the minimum value of y in the defect model within the range of the filament temperature difference, and use the filament temperature difference corresponding to the minimum value of y as the filament temperature difference of the gallium source furnace during the next epitaxial growth.

[0006] Optionally, the filament temperature difference has a value range of greater than or equal to 30° C. and less than or equal to 150° C.

[0007] Optionally, the steps for selecting the multiple different filament temperature difference values ​​are as follows: Step S21, assuming that B0 represents the minimum value in the value range of the filament temperature difference, and B1 represents the maximum value in the value range of the filament temperature difference, first determine that the first three values ​​of the multiple different filament temperature differences are x1, x2 and x3, respectively, x1=B0+(B1-B0) / 4, x2=B0+(B1-B0) / 2, x3=B0+3×(B1-B0) / 4; Step S22, respectively measuring and obtaining surface defect data y1, y2 and y3 of the epitaxial wafer grown under the corresponding filament temperature difference conditions, wherein y1 corresponds to x1, y2 corresponds to x2, and y3 corresponds to x3; Step S23, use m i Represents surface defect data, using n i Indicates that m i Corresponding filament temperature difference, i=1,2,3, initialize m i The values ​​are m1=y1, m2=y2, m3=y3, initialize n i The values ​​are n1=x1, n2=x2, n3=x3, and x 3+j represents other values ​​of the selected multiple different filament temperature differences except the first three values, y 3+j Represents x 3+j The corresponding surface defect data of the epitaxial wafer, the value of j is initialized to 1; Step S24: if the values ​​of m1, m2, and m3 decrease in sequence, the average value of n3 and b1 is selected as the value x of the plurality of different filament temperature differences. 3+j , b1 represents the selected values ​​of the filament temperature difference and the value in B1 that is larger than n3 and closest to n3, and then the surface defect data y of the epitaxial wafer grown under this condition is measured and obtained 3+j , n2, n3, x 3+j Arrange them from small to large to form new sequences P1, P2 and P3, and then update m i and n i The values ​​of are m1=p1, m2=p2, m3=p3, n1=P1, n2=P2, n3=P3, and then the value of j is increased by 1, p1 represents the surface defect data of the epitaxial wafer corresponding to P1, p2 represents the surface defect data of the epitaxial wafer corresponding to P2, and p3 represents the surface defect data of the epitaxial wafer corresponding to P3; Step S25: If the values ​​of m1, m2, and m3 increase in sequence, the average value of n1 and b0 is selected as the value x of the plurality of different filament temperature differences. 3+j , b0 represents the selected values ​​of the filament temperature difference and the value in B0 that is smaller than n1 and closest to n1, and then the surface defect data y of the epitaxial wafer grown under this condition is measured and obtained 3+j , n1, n2, x 3+j Arrange them from small to large to form new sequences F1, F2 and F3, and then update m i and n i The values ​​of are m1=f1, m2=f2, m3=f3, n1=F1, n2=F2, n3=F3, and then the value of j is increased by 1, f1 represents the surface defect data of the epitaxial wafer corresponding to F1, f2 represents the surface defect data of the epitaxial wafer corresponding to F2, and f3 represents the surface defect data of the epitaxial wafer corresponding to F3; Step S26: If m2 is the minimum value among m1, m2 and m3, the average value of n1 and n2 is selected as the value x of the plurality of different filament temperature differences. 3+j , and the average value of n2 and n3 is selected as the value x of the multiple different filament temperature differences 3+j+1 , measure and obtain the surface defect data y of the epitaxial wafer grown under the corresponding conditions 3+j and 3+j+1 , n2, x 3+j 、x 3+j+1 Arrange them from small to large to form new sequences D1, D2 and D3, and then update m i and n i The values ​​of are m1=d1, m2=d2, m3=d3, n1=D1, n2=D2, n3=D3, and then the value of j is increased by 2, d1 represents the surface defect data of the epitaxial wafer corresponding to D1, d2 represents the surface defect data of the epitaxial wafer corresponding to D2, and d3 represents the surface defect data of the epitaxial wafer corresponding to D3; Step S27, iteratively executing steps S24 to S26 until the preset number of filament temperature differences is selected.

[0008] Optionally, in step S5, the structure of the epitaxial layer grown during the next epitaxial growth is a preset structure.

[0009] Optionally, after step S5, the method further includes: adding the surface defect data obtained during the next epitaxial growth and the corresponding filament temperature difference to the surface defect data in step S4, re-performing numerical fitting to obtain a new constant value, and using the new constant value, performing step S5 again to obtain a re-optimized filament temperature difference.

[0010] The beneficial effects of the present invention include: The method for optimizing the temperature of a gallium source furnace of a molecular beam epitaxy system provided by the present invention comprises the following steps: S1, epitaxially growing epitaxial layers of a preset structure on a substrate of the same specification under a plurality of different filament temperature difference conditions respectively, so as to obtain a plurality of epitaxial wafers corresponding to the filament temperature difference conditions, wherein the number of the plurality of different filament temperature differences is a preset number, and the preset number is greater than or equal to 6, and the value range of the filament temperature difference is greater than or equal to 0° C. and less than or equal to 200° C.; S2, measuring and obtaining surface defect data of the plurality of epitaxial wafers, wherein the surface defect data includes the number of defects per unit area of ​​the surface of the epitaxial wafer; S3, establishing a defect model for characterizing the relationship between the surface defect data and the filament temperature difference, wherein the defect model is as follows: , where y represents the number of defects per unit area on the surface of the epitaxial wafer, x represents the filament temperature difference, coefficients k1 and k2 are both constants greater than 0, base a1 is a constant greater than 0 and less than 1, base a2 is a constant greater than 1, and c is a constant; S4, using the surface defect data, by numerical fitting, calculate the values ​​of all constants in the defect model; S5, substitute the values ​​of all constants obtained into the defect model, calculate the minimum value of y in the defect model within the range of the filament temperature difference, and use the filament temperature difference corresponding to the minimum value of y as the filament temperature difference of the gallium source furnace during the next epitaxial growth. This application establishes a defect model for characterizing the relationship between surface defect data and filament temperature difference, and fits the constants in the model through multiple measurement data results, and then uses the model to obtain the filament temperature difference corresponding to the minimum surface defect. This method achieves the optimization of the filament temperature difference of the gallium source furnace and the reduction of the number of surface defects. In addition, in large-scale mass production, the filament temperature difference process conditions can be continuously improved and product quality can be improved by continuously adding the measurement results to the fitting data of the model. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0012] Figure 1 A schematic structural diagram of a gallium source furnace provided in an embodiment of the present invention is shown; Figure 2 A schematic flow chart of a method for optimizing the temperature of a gallium source furnace in a molecular beam epitaxy system provided by an embodiment of the present invention is shown; Figure 3 A curve schematic diagram showing the relationship between the number of surface defects per unit area of ​​the epitaxial wafer provided by an embodiment of the present invention and the temperature difference between the upper and lower filaments of the gallium source furnace. DETAILED DESCRIPTION

[0013] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0014] Figure 1 A schematic structural diagram of a gallium source furnace provided in an embodiment of the present invention is shown.

[0015] like Figure 1As shown, the gallium source furnace includes a source furnace crucible 10, the interior of the source furnace crucible 10 is used to accommodate gallium raw materials, the source furnace crucible 10 is a cone with a small bottom and a large top, and the top of the source furnace crucible 10 is open. The gallium source furnace also includes a lower heating filament 12 located at the lower part of the source furnace crucible 10 and an upper heating filament 11 located at the upper part of the source furnace crucible. The lower heating filament 12 and the upper heating filament 11 are independently controlled for heating power. The temperature of the lower heating filament 12 is sensed by a lower thermocouple 14 located at the bottom of the source furnace crucible 10, and the temperature of the upper heating filament 11 is sensed by an upper thermocouple 13 located at the upper part of the source furnace crucible 10. In the control interface of the molecular beam epitaxy system, the operator can set the temperature value of the corresponding filament, and then the system will use PID control to control the heating power of the corresponding filament so that the temperature of the corresponding filament reaches and stabilizes at the set temperature.

[0016] The source furnace crucible 10 is heated by the upper heating filament 11 and the lower heating filament 12. When the heating temperature reaches a certain level, the gallium raw material in the source furnace crucible 10 will be ejected from the top opening of the source furnace crucible 10 in the form of a molecular beam, and then deposited on the surface of the epitaxial wafer. The higher the heating temperature, the larger the beam, and the greater the deposition rate. It should be noted that this application aims to optimize the filament temperature difference and does not consider the effect of the beam rate on the number of surface defects. Therefore, in this application, the beam rate of gallium under all process conditions is equal to the preset beam rate, that is, when the filament temperature difference is changed, the beam rate of gallium remains unchanged.

[0017] Since the gallium molecular beam easily condenses at the opening of the source furnace crucible 10 and forms gallium droplets, the gallium droplets accumulated near the opening edge of the source furnace crucible 10 will affect the rate stability of the gallium molecular beam provided by the gallium source furnace. In addition, if the gallium droplets are deposited on the epitaxial wafer, defects will be formed on the surface of the epitaxial wafer, deteriorating the quality and yield of the epitaxial wafer. Therefore, it is usually necessary to ensure that the crucible opening has a higher temperature to limit condensation and gallium droplet formation. However, if the temperature at the crucible opening is too high, the excessive heat radiation at the crucible opening will cause impurities to be generated in the surrounding environment of the growth chamber. If these impurities reach the surface of the epitaxial wafer, it will cause an increase in surface defects of the epitaxial wafer. Therefore, it is necessary to provide a method to optimize the temperature difference between the upper and lower filaments of the gallium source furnace.

[0018] Figure 2 A schematic flow chart of a method for optimizing the temperature of a gallium source furnace in a molecular beam epitaxy system provided by an embodiment of the present invention is shown.

[0019] The beam rate of gallium used in molecular beam epitaxy is constant and equal to the preset beam rate. The method is used to optimize the filament temperature difference of the gallium source furnace. The filament temperature difference is equal to the temperature of the upper heating filament minus the temperature of the lower heating filament, such as Figure 2 As shown, the method for optimizing the temperature of a gallium source furnace in a molecular beam epitaxy system provided by the present invention comprises the following steps: S1. Under a plurality of different filament temperature difference conditions, epitaxial layers of a preset structure are grown on substrates of the same specification to obtain a plurality of epitaxial wafers corresponding to the filament temperature difference conditions.

[0020] The number of the multiple different filament temperature differences is a preset number, and the preset number is greater than or equal to 6, and the value range of the filament temperature difference is greater than or equal to 0° C. and less than or equal to 200° C.

[0021] In practical applications, in order to prevent the gallium beam from condensing at the crucible opening, the minimum value of the filament temperature difference needs to be greater than or equal to 0°C. In other words, the temperature of the upper heating filament 11 is greater than or equal to the temperature of the lower heating filament 12. If the temperature of the upper heating filament 11 is too high, a large amount of impurities in the surrounding environment of the filament will be released. In addition, the heat radiation between the upper and lower heating filaments will affect each other, so that the filament temperature difference cannot be too large. Therefore, according to experience, the maximum value of the filament temperature difference can be set to be less than or equal to 200°C. Preferably, the range of the filament temperature difference is greater than or equal to 30°C and less than or equal to 150°C.

[0022] It should be understood that the larger the preset number is, the more data is obtained and the more accurate the final result is, but at the same time, the more material cost and time are required. Therefore, considering that there are 5 constants that need to be fitted and calculated below, the preset number is at least 6. Preferably, the preset number is greater than or equal to 10, for example, the preset number is 10.

[0023] When epitaxial growth is performed, the substrate is first transferred into the growth chamber, and then during the epitaxial growth process, in order to obtain a stable beam current, the filament temperature difference is stable and unchanged. After the epitaxial growth is completed, a new filament temperature difference can be set and stabilized for a period of time, and then the new filament temperature difference conditions can be used to perform epitaxial growth on the new substrate, and a new epitaxial wafer can be obtained. In other words, in step S1, different filament temperature differences correspond to different epitaxial wafers. In order to reduce the impact of substrates of different specifications on the number of surface defects of epitaxial wafers, it is defined in step S1 that substrates of the same specifications are used for different filament temperature differences, that is, the conventional parameters of the multiple substrates used in step S1 (for example, the material of the substrate, conductivity, single-sided polishing or double-sided polishing, substrate size and deflection angle, etc.) are the same.

[0024] S2. Measure and obtain surface defect data of the plurality of epitaxial wafers.

[0025] The surface defect data includes the number of defects per unit area on the surface of the epitaxial wafer. The surface defect data in step S2 of the present application is automatically measured by a surface defect optical scanning measurement device commonly used in the field of semiconductor epitaxial wafer manufacturing.

[0026] S3. Establish a defect model for characterizing the relationship between surface defect data and filament temperature difference.

[0027] The defect model is as follows: , Among them, y represents the number of defects per unit area of ​​the epitaxial wafer surface, x represents the filament temperature difference, coefficient k1 and coefficient k2 are both constants greater than 0, base a1 is a constant greater than 0 and less than 1, base a2 is a constant greater than 1, and c is a constant.

[0028] As mentioned above, the higher the heating temperature of the gallium source furnace, the larger the molecular beam of gallium and the greater the deposition rate. Since the molecular beam of gallium is easy to condense at the opening of the source furnace crucible 10 and form gallium droplets, if the gallium droplets are deposited on the epitaxial wafer, defects will be formed on the surface of the epitaxial wafer, deteriorating the quality and yield of the epitaxial wafer. The larger the filament temperature difference, that is, the higher the temperature near the opening of the source furnace crucible 10 relative to the lower part of the source furnace crucible 10, the smaller the possibility of condensation and formation of gallium droplets, and thus the smaller the defect density on the surface of the epitaxial wafer caused by gallium droplets (that is, the number of surface defects per unit area). It should be understood that when the filament temperature difference is small, the defect density on the surface of the epitaxial wafer caused by gallium droplets is large. At this time, increasing the filament temperature difference can significantly reduce the defect density on the surface of the epitaxial wafer caused by gallium droplets; as the filament temperature difference further increases, the defect density on the surface of the epitaxial wafer caused by gallium droplets itself is decreasing, so the change in the filament temperature difference on the defect density on the surface of the epitaxial wafer caused by gallium droplets is also decreasing. Therefore, within a certain range, it is possible to adopt the following Figure 3 Curve 1 in the figure schematically shows the relationship between the defect density on the surface of the epitaxial wafer caused by the gallium droplets and the change in the filament temperature difference. Therefore, the following functional relationship can be used to approximately characterize the relationship between the defect density on the surface of the epitaxial wafer caused by the gallium droplets and the change in the filament temperature difference: , Wherein, y1 represents the number of defects per unit area of ​​the epitaxial wafer surface caused by gallium droplets, x represents the filament temperature difference, coefficient k1 is a constant greater than 0, base a1 is a constant greater than 0 and less than 1, and c1 is a constant.

[0029] However, if the temperature at the opening of the source furnace crucible 10 is too high, the excessive heat radiation at the opening of the source furnace crucible 10 will cause impurities to be generated in the surrounding environment of the growth chamber. For example, in the historical growth process, various source materials that have not been deposited on the epitaxial wafer will be deposited and adsorbed on the cold bottle wall in the growth chamber (the temperature of the cold bottle is usually liquid nitrogen temperature, which is used to adsorb impurities at low temperature). Under normal circumstances, these adsorbed materials will not affect the subsequent epitaxial wafer growth. However, when the ambient temperature rises, some of the adsorbed materials will be desorbed due to thermal radiation. Once these desorbed materials are deposited on the surface of the epitaxial wafer, they will act as impurities, causing an increase in surface defects of the epitaxial wafer and deteriorating the quality of the epitaxial wafer. In addition to the cold bottle wall, impurities deposited in other places may also form a beam due to thermal radiation, affecting the surface quality of the epitaxial wafer. As mentioned above, the beam rate of gallium used in molecular beam epitaxy is constant and equal to the preset beam rate. In order to ensure that the beam rate of gallium remains constant, when adjusting the filament temperature difference, if the filament temperature difference decreases, the lower heating filament needs to be heated up, while the upper heating filament needs to be cooled down; if the filament temperature difference increases, the lower heating filament needs to be cooled down, while the upper heating filament needs to be heated up. In other words, when the beam rate of gallium remains constant, the trend of decreasing or increasing the temperature of the upper heating filament is consistent with the trend of decreasing or increasing the filament temperature difference. Figure 1 As shown, the lower heating filament 12 is located at the lower part of the source furnace crucible 10, and the outer side of the lower heating filament 12 is the water cooling jacket of the source furnace. The temperature of the lower heating filament 12 only affects the beam rate of the gallium source furnace, and has almost no effect on the surrounding environment in the growth chamber. The upper heating filament 11 is located at the upper part of the source furnace crucible 10, and the temperature at the opening of the source furnace crucible 10 is mainly determined by the temperature of the upper heating filament 11. Therefore, when the filament temperature difference is small, the temperature of the upper heating filament 11 is relatively low, the temperature at the opening of the source furnace crucible 10 is low, and the thermal radiation at the opening has little effect on the environment in the chamber. At this time, the defect density on the surface of the epitaxial wafer caused by thermal radiation is small; as the filament temperature difference increases, the temperature of the upper heating filament 11 increases accordingly, the temperature at the opening of the source furnace crucible 10 increases, and the thermal radiation at the opening has a greater impact on the environment in the chamber. At this time, the defect density on the surface of the epitaxial wafer caused by thermal radiation also increases. As the filament temperature difference increases, the change in the filament temperature difference also increases the effect on the defect density on the surface of the epitaxial wafer caused by thermal radiation. Therefore, within a certain range, it is possible to use Figure 3 Curve 2 in the figure schematically shows the relationship between the defect density on the surface of the epitaxial wafer caused by the heat radiation at the opening of the source furnace crucible and the change of the filament temperature difference. Therefore, the following functional relationship can be used to approximately characterize the relationship between the defect density on the surface of the epitaxial wafer caused by the heat radiation at the opening of the source furnace crucible and the change of the filament temperature difference: , Among them, y2 represents the number of defects per unit area on the surface of the epitaxial wafer caused by thermal radiation at the opening of the source furnace crucible, x represents the filament temperature difference, the coefficient k2 is a constant greater than 0, the base a2 is a constant greater than 1, and c2 is a constant.

[0030] In summary, a defect model that characterizes the relationship between surface defect data and filament temperature difference can be established. Specifically, the total number of defects per unit area on the epitaxial wafer surface caused by the heat radiation from the gallium droplets and the opening of the source furnace crucible can be expressed as follows: , Where y represents the number of defects per unit area on the surface of the epitaxial wafer, x represents the temperature difference of the filament, coefficient k1 and coefficient k2 are both constants greater than 0, base a1 is a constant greater than 0 and less than 1, base a2 is a constant greater than 1, and c is a constant. Figure 3 Curve 3 in FIG. 1 schematically shows the relationship between the total defect density on the surface of the epitaxial wafer and the change in the filament temperature difference.

[0031] It should be understood that the purpose of this application is to optimize the filament temperature difference of the gallium source furnace. The surface defects in the above model only consider the factors related to the filament temperature difference (that is, the two aspects y1 and y2 mentioned above). Therefore, the surface defects in the above model do not include surface defects caused by other reasons unrelated to the filament temperature difference (such as substrate quality, demolding process, etc.).

[0032] S4. Using the surface defect data, the values ​​of all constants in the defect model are calculated through numerical fitting.

[0033] All the constants here include k1, k2, a1, a2 and c. There are 5 unknown constants in the defect model. In step S1 and step S2, for each selected value of the independent variable x, the corresponding value of the variable y can be obtained by measurement. Therefore, using the surface defect data that has been obtained, the values ​​of the 5 constants in the defect model can be calculated by numerical fitting (for example, least squares fitting).

[0034] S5. Substitute the values ​​of all the constants obtained into the defect model, calculate and solve the minimum value of y in the defect model within the range of the filament temperature difference, and use the filament temperature difference corresponding to the minimum value of y as the filament temperature difference of the gallium source furnace during the next epitaxial growth.

[0035] After substituting the values ​​of all constants obtained in step S4 into the defect model, the functional relationship of y changing with x is obtained. Therefore, the minimum value of y when x changes within the range of the filament temperature difference can be analytically calculated by solving the derivative of the function y. Figure 3As shown in the figure, when the temperature difference between the upper and lower filaments is ΔT0, the value of the number of surface defects per unit area is the smallest. If ΔT0 is within the value range of the filament temperature difference, ΔT0 can be used as the filament temperature difference of the gallium source furnace for the next epitaxial growth. If ΔT0 is not within the value range of the filament temperature difference, due to the properties of the function y, the y value corresponding to the value closest to ΔT0 within the value range of the filament temperature difference is the smallest. At this time, the value closest to ΔT0 within the value range of the filament temperature difference can be used as the filament temperature difference of the gallium source furnace for the next epitaxial growth.

[0036] Since the number of filament temperature difference values ​​is limited, in order to obtain data that is more conducive to accurate model fitting under limited value conditions, the present application designs the following filament value selection steps.

[0037] The steps for selecting the values ​​of the multiple different filament temperature differences are as follows: Step S21, assuming that B0 represents the minimum value in the value range of the filament temperature difference, and B1 represents the maximum value in the value range of the filament temperature difference, first determine that the first three values ​​of the multiple different filament temperature differences are x1, x2 and x3, respectively, x1=B0+(B1-B0) / 4, x2=B0+(B1-B0) / 2, x3=B0+3×(B1-B0) / 4.

[0038] Step S22, respectively measure and obtain surface defect data y1, y2 and y3 of the epitaxial wafer grown under the corresponding filament temperature difference conditions, wherein y1 corresponds to x1, y2 corresponds to x2, and y3 corresponds to x3.

[0039] Step S23, use m i Represents surface defect data, using n i Indicates that m i Corresponding filament temperature difference, i=1,2,3, initialize m i The values ​​are m1=y1, m2=y2, m3=y3, initialize n i The values ​​are n1=x1, n2=x2, n3=x3, and x 3+j represents other values ​​of the selected multiple different filament temperature differences except the first three values, y 3+j Represents x 3+j The corresponding surface defect data of the epitaxial wafer, the value of j is initialized to 1.

[0040] Step S24: if the values ​​of m1, m2, and m3 decrease in sequence, the average value of n3 and b1 is selected as the value x of the plurality of different filament temperature differences. 3+j , b1 represents the selected values ​​of the filament temperature difference and the value in B1 that is larger than n3 and closest to n3, and then the surface defect data y of the epitaxial wafer grown under this condition is measured and obtained 3+j, n2, n3, x 3+j Arrange them from small to large to form new sequences P1, P2 and P3, and then update m i and n i The values ​​of are m1=p1, m2=p2, m3=p3, n1=P1, n2=P2, n3=P3, and then the value of j is increased by 1, p1 represents the surface defect data of the epitaxial wafer corresponding to P1, p2 represents the surface defect data of the epitaxial wafer corresponding to P2, and p3 represents the surface defect data of the epitaxial wafer corresponding to P3. In this step, since n2, n3, x 3+j It increases in sequence, so at this time P1=n2, P2=n3, P3=x 3+j .

[0041] Step S25: If the values ​​of m1, m2, and m3 increase in sequence, the average value of n1 and b0 is selected as the value x of the plurality of different filament temperature differences. 3+j , b0 represents the selected values ​​of the filament temperature difference and the value in B0 that is smaller than n1 and closest to n1, and then the surface defect data y of the epitaxial wafer grown under this condition is measured and obtained 3+j , n1, n2, x 3+j Arrange them from small to large to form new sequences F1, F2 and F3, and then update m i and n i The values ​​of are m1=f1, m2=f2, m3=f3, n1=F1, n2=F2, n3=F3, and then the value of j is increased by 1, f1 represents the surface defect data of the epitaxial wafer corresponding to F1, f2 represents the surface defect data of the epitaxial wafer corresponding to F2, and f3 represents the surface defect data of the epitaxial wafer corresponding to F3. In this step, since n1, n2, x 3+j Medium 3+j minimum, n2 maximum, therefore, F1=x 3+j , F2=n1, F3=n2.

[0042] Step S26: If m2 is the minimum value among m1, m2 and m3, the average value of n1 and n2 is selected as the value x of the plurality of different filament temperature differences. 3+j , and the average value of n2 and n3 is selected as the value x of the multiple different filament temperature differences 3+j+1 , measure and obtain the surface defect data y of the epitaxial wafer grown under the corresponding conditions 3+j and 3+j+1 , n2, x 3+j 、x 3+j+1 Arrange them from small to large to form new sequences D1, D2 and D3, and then update m i and n iThe values ​​of are m1=d1, m2=d2, m3=d3, n1=D1, n2=D2, n3=D3, and then the value of j is increased by 2, d1 represents the surface defect data of the epitaxial wafer corresponding to D1, d2 represents the surface defect data of the epitaxial wafer corresponding to D2, and d3 represents the surface defect data of the epitaxial wafer corresponding to D3. In this step, since n2, x 3+j 、x 3+j+1 Medium 3+j Minimum, x 3+j+1 Maximum, therefore, at this time D1=x 3+j , D2=n2, D3=x 3+j+1 .

[0043] Step S27, iteratively executing steps S24 to S26 until the preset number of filament temperature differences is selected.

[0044] The introduction of P1, P2 and P3; F1, F2 and F3; D1, D2 and D3 in steps S24 to S26 is to facilitate the assignment of values ​​to n in the next iteration. i Sort the three numbers from smallest to largest.

[0045] b1 represents the selected values ​​of the filament temperature difference and the specific meaning of the value in B1 that is larger than n3 and closest to n3 is as follows: For example, assuming that t values ​​of the filament temperature difference have been selected, these t values ​​are x1, x2, ..., x t , then at x1, x2, …, x t Among the t+1 values ​​of B1 and B2, select the value that is larger than n3 and closest to n3 as b1.

[0046] b0 represents the selected values ​​of the filament temperature difference and the specific meaning of the value in B0 that is smaller than n1 and closest to n1 is as follows: For example, assuming that t values ​​of the filament temperature difference have been selected, these t values ​​are x1, x2, ..., x t , then in x1, x2, …, x t Among the t+1 values ​​of B0 and B1, select the value that is smaller than n1 and closest to n1 as b0.

[0047] It should be noted that in step S26, two values ​​are added at a time. If the number of values ​​exceeds the preset number after the two values ​​are added, one value is randomly selected from the two values ​​determined in step S26.

[0048] In order to achieve as wide a coverage of the entire value range as possible and avoid the model fitting deviation caused by taking values ​​only in a small area within the value range, in step S21, the initial three values ​​are set as the four equal points of the entire value range, and in the subsequent steps, the added value points are also taken based on the adjacent known average values. In order to make the subsequent values ​​as close to the optimal point as possible and avoid blind global average values, in steps S24 to S26, each subsequent value of the filament temperature difference is based on the surface defect test data corresponding to the previous value, and the new value position is selected in the direction close to the theoretical optimal value. Therefore, through the preset number of values ​​selected in the above steps S21 to S27, data that is more conducive to accurate model fitting can be obtained under limited value conditions, thereby facilitating the acquisition of a more optimized filament temperature difference with as little experimental cost as possible.

[0049] It should be understood that the filament temperature difference optimized through steps S1 to S5 is an optimization of the gallium source furnace and has nothing to do with the specific structure of the epitaxial layer. Therefore, the optimized filament temperature difference can be used for the epitaxial growth of the preset structure in step S1, and can also be used for the growth of other epitaxial structures that require a gallium source furnace.

[0050] Optionally, in step S5, the structure of the epitaxial layer grown during the next epitaxial growth is a preset structure. After step S5, the method further includes: adding the surface defect data obtained during the next epitaxial growth and the corresponding filament temperature difference to the surface defect data in step S4, re-performing numerical fitting to obtain a new constant value, and using the new constant value, performing step S5 again to obtain a re-optimized filament temperature difference. It should be understood that in order to eliminate the influence of the structural factors themselves on the surface defects, when the subsequently grown structure is the same as the structure in step S1, the subsequent measurement result data can be added to the fitting data of the model to make the model fitting result more accurate. If the subsequently grown epitaxial structure is different from the structure in step S1, the subsequent measurement results cannot be directly added to step S4.

[0051] It should be noted that, in a short period of time, when the filament temperature remains unchanged, the beam rate of gallium presents a relatively stable state. As the gallium source is used, the remaining gallium source in the crucible gradually decreases. After a certain period of use, in order to maintain the preset beam rate, the heating temperature of the filament needs to be increased. If the filament temperature difference remains unchanged, then the temperature of the upper and lower heating filaments must be increased. When the increase reaches a certain level, the effect of the filament temperature difference on surface defects may be different from before. Therefore, it is necessary to re-execute steps S1 to S5, re-fit the constant value of the solution model according to the latest measurement data, and determine the new optimized filament temperature difference.

[0052] In summary, this application establishes a defect model for characterizing the relationship between surface defect data and filament temperature difference, and fits the constant value of the model through multiple measurement data results, and then uses the model to obtain the filament temperature difference corresponding to the minimum surface defect. This method achieves the optimization of the gallium source furnace filament temperature difference and the reduction of the number of surface defects. In addition, in large-scale mass production, by continuously adding measurement results to the model fitting data, the filament temperature difference process conditions can be continuously improved to improve product quality.

[0053] The above embodiments are only for illustrating the technical concept and features of the present invention, and their purpose is to enable ordinary technicians in the field to understand the content of the present invention and implement it. They cannot be used to limit the protection scope of the present invention. Any equivalent changes or modifications made according to the spirit of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for optimizing the temperature of a gallium source furnace in a molecular beam epitaxy system. It is characterized in that The gallium source furnace comprises a lower heating filament located at the bottom of the source furnace crucible and an upper heating filament located at the top of the source furnace crucible. The beam rate of gallium used in molecular beam epitaxy is constant and equal to a preset beam rate. The method is used to optimize the filament temperature difference of the gallium source furnace, and the filament temperature difference is equal to the temperature of the upper heating filament minus the temperature of the lower heating filament. The method comprises the following steps: S1. Under a plurality of different filament temperature difference conditions, epitaxially grow epitaxial layers of a preset structure on a substrate of the same specification, respectively, to obtain a plurality of epitaxial wafers corresponding to the filament temperature difference conditions, wherein the number of the plurality of different filament temperature differences is a preset number, and the preset number is greater than or equal to 6, and the value range of the filament temperature difference is greater than or equal to 0° C. and less than or equal to 200° C.; S2. Measure and obtain surface defect data of the plurality of epitaxial wafers, wherein the surface defect data includes the number of defects per unit area on the surface of the epitaxial wafers; S3. Establish a defect model for characterizing the relationship between surface defect data and filament temperature difference. The defect model is as follows: ， Among them, y represents the number of defects per unit area on the surface of the epitaxial wafer, x represents the filament temperature difference, and the coefficient k 1 and coefficient k 2 are all constants greater than 0, base a 1 is a constant greater than 0 and less than 1, base a 2 is a constant greater than 1, c is a constant; S4, using the surface defect data to calculate and obtain the values ​​of all constants in the defect model by numerical fitting; S5. Substitute the values ​​of all the constants obtained into the defect model, calculate and solve the minimum value of y in the defect model within the range of the filament temperature difference, and use the filament temperature difference corresponding to the minimum value of y as the filament temperature difference of the gallium source furnace during the next epitaxial growth.

2. The method for optimizing the temperature of a gallium source furnace in a molecular beam epitaxy system according to claim 1, It is characterized in that The filament temperature difference has a value range of greater than or equal to 30° C. and less than or equal to 150° C.

3. The method for optimizing the temperature of a gallium source furnace in a molecular beam epitaxy system according to claim 1 or 2, It is characterized in that The steps for selecting the values ​​of the multiple different filament temperature differences are as follows: Step S21: Assumption B 0 represents the minimum value in the range of the filament temperature difference, B 1 represents the maximum value in the value range of the filament temperature difference, then first determine that the first three values ​​of the multiple different filament temperature differences are x 1 、x 2 and x 3 , x 1 =B 0 +(B 1 -B 0 ) / 4,x 2 =B 0 +(B 1 -B 0 ) / 2,x 3 =B 0 +3×(B 1 -B 0 ) / 4; Step S22: Measure and obtain surface defect data y of the epitaxial wafer grown under the corresponding filament temperature difference conditions. 1 ,y 2 and 3 , where y 1 With x 1 Corresponding, y 2 With x 2 Corresponding, y 3 With x 3 correspond; Step S23, use m i Represents surface defect data, using n i Indicates that m i Corresponding filament temperature difference, i=1,2,3, initialize m i The value of m 1 =y 1 , m 2 =y 2 , m 3 =y 3 , initialize n i The value of 1 =x 1 , n 2 =x 2 , n 3 =x 3 , use x 3+j represents other values ​​of the selected multiple different filament temperature differences except the first three values, y 3+j Represents x 3+j The corresponding surface defect data of the epitaxial wafer, the value of j is initialized to 1; Step S24: If m 1 、m 2 、m 3 The values ​​of n decrease in sequence. 3 With b 1 The average value of the multiple different filament temperature differences is selected as x 3+j , b 1 Indicates the values ​​of the selected filament temperature differences and B 1 Medium ratio 3 Large and closest to n 3 Then, the surface defect data y of the epitaxial wafer grown under this condition is measured. 3+j , n 2 、n 3 、x 3+j Arrange them from small to large to form a new sequence P 1 , P 2 and P 3 , then update m i and n i The value of m 1 =p 1 , m 2 =p 2 , m 3 =p 3 , n 1 =P 1 , n 2 =P 2 , n 3 =P 3 , then the value of j is increased by 1, p 1 Indicates that P 1 The corresponding surface defect data of the epitaxial wafer, p 2 Indicates that P 2 The corresponding surface defect data of the epitaxial wafer, p 3 Indicates that P 3 Surface defect data of the corresponding epitaxial wafer; Step S25: If m 1 、m 2 、m 3 The values ​​of n increase successively. 1 With b 0 The average value of the multiple different filament temperature differences is selected as x 3+j , b 0 Indicates the values ​​of the selected filament temperature differences and B 0 Medium ratio 1 Small and closest to n 1 Then, the surface defect data y of the epitaxial wafer grown under this condition is measured. 3+j , n 1 、n 2 、x 3+j Arrange them from small to large to form a new sequence F 1 、F 2 and F 3 , then update m i and n i The value of m 1 =f 1 , m 2 =f 2 , m 3 =f 3 , n 1 =F 1 , n 2 =F 2 , n 3 =F 3 , then the value of j is increased by 1, f 1 Indicates that F 1 Corresponding surface defect data of epitaxial wafer, f 2 Indicates that F 2 Corresponding surface defect data of epitaxial wafer, f 3 Indicates that F 3 Surface defect data of the corresponding epitaxial wafer; Step S26: If m 2 is m 1 、m 2 、m 3 The minimum value among the three values, then n 1 With n 2 The average value of the multiple different filament temperature differences is selected as x 3+j , and n 2 With n 3 The average value of the multiple different filament temperature differences is selected as x 3+j+1 , measure and obtain the surface defect data y of the epitaxial wafer grown under the corresponding conditions 3+j and 3+j+1 , n 2 、x 3+j 、x 3+j+1 Arrange them from small to large to form a new sequence D 1 , D 2 and D 3 , then update m i and n i The value of m 1 =d 1 , m 2 =d 2 , m 3 =d 3 , n 1 =D 1 , n 2 =D 2 , n 3 =D 3 , then the value of j is increased by 2, d 1 Indicates that D 1 Corresponding surface defect data of epitaxial wafer, d 2 Indicates that D 2 Corresponding surface defect data of epitaxial wafer, d 3 Indicates that D 3 Surface defect data of the corresponding epitaxial wafer; Step S27, iteratively executing steps S24 to S26 until the preset number of filament temperature differences is selected.

4. The method for optimizing the temperature of a gallium source furnace in a molecular beam epitaxy system according to claim 1, It is characterized in that In step S5, the structure of the epitaxial layer grown during the next epitaxial growth is the preset structure.

5. The method for optimizing the temperature of a gallium source furnace in a molecular beam epitaxy system according to claim 4, It is characterized in that After step S5, the method further includes: adding the surface defect data obtained during the next epitaxial growth and the corresponding filament temperature difference to the surface defect data in step S4, re-performing numerical fitting to obtain a new constant value, and using the new constant value, performing step S5 again to obtain a re-optimized filament temperature difference.

Images