Demolding method in molecular beam epitaxy process

Abstract

The invention provides a demolding method in a molecular beam epitaxy process, and relates to the technical field of semiconductor manufacturing. The method comprises the following steps: heating an InP substrate to a first demolding temperature under a preset As pressure condition; keeping at the first demolding temperature for a first time period; raising the temperature from the first demolding temperature to a second demolding temperature; and keeping for a third time period at the second demolding temperature, and then ending the demolding process. According to the method, the first time period is kept at the first demolding temperature slightly lower than the reference temperature, the substrate is fully heated, meanwhile, desorption of phosphorus is reduced, due to sufficient heating, the time needed for complete desorption of surface oxide at the second demolding temperature is shortened, therefore, surface defects of the epitaxial wafer can be remarkably reduced, and the yield of the epitaxial wafer is improved. And epitaxial wafer surface quality is improved.

Description

Technical Field

[0001] The invention relates to the technical field of semiconductor manufacturing, and in particular to a film stripping method in a molecular beam epitaxy process. Background Art

[0002] For the molecular beam epitaxy (MBE) growth of InP-based epitaxial materials, the InP substrate must first be subjected to high-temperature stripping treatment in the molecular beam epitaxy equipment to remove the natural oxide on the surface of the InP substrate, and then the corresponding epitaxial layer is deposited and grown according to the device structure design. In the large-scale batch production of InP-based semiconductor devices, the quality of the molecular beam epitaxy high-temperature stripping process directly affects the surface defect density of the prepared InP-based semiconductor device epitaxial wafer, and thus affects the yield and reliability of the final device. If the stripping time is too short, the surface oxide cannot be completely desorbed, which leads to defects in the subsequent epitaxial growth process; if the stripping time is too long, phosphorus will be desorbed to form a large number of InAs microcrystals, which will also cause defects in the subsequent epitaxial growth process. Therefore, if the stripping time is too long or too short, the defect density of the substrate surface and the epitaxial wafer will increase. The more surface defects there are on the epitaxial wafer, the lower the yield of the device, and the worse the reliability of the device.

[0003] Therefore, for the mass production of InP-based semiconductor device epitaxial wafers, it is necessary to propose a molecular beam epitaxy stripping process to reduce the surface defect density of InP-based semiconductor device epitaxial wafers and improve the quality of InP-based semiconductor device epitaxial wafers. Summary of the invention

[0004] The object of the present invention is to provide a film stripping method in a molecular beam epitaxy process in view of the above-mentioned deficiencies in the prior art, so as to solve the problem of reducing the surface defects caused by film stripping of an InP substrate.

[0005] To achieve the above purpose, the technical solution adopted by the present invention is as follows: The present invention provides a film stripping method in a molecular beam epitaxy process, the method is used to strip an InP substrate, the method comprising: Step a, under a preset As pressure condition, heating the InP substrate to a first stripping temperature, the first stripping temperature being 4°C to 6°C lower than a pre-obtained reference temperature, the reference temperature being the substrate temperature when the InP substrate is heated under a preset As pressure condition to a RHEED pattern showing ×4 reconstruction fringes; Step b, maintaining the first time period at the first demolding temperature, wherein the first time period t1 is in the range of 6 min ≤ t1 ≤ 14 min; Step c, heating from the first demolding temperature to the second demolding temperature, the time required for the heating process is the second time period, the range of the second time period t2 is 0.5min ≤ t2 ≤ 2 min, and the range of the difference between the second demolding temperature and the reference temperature is greater than or equal to zero and less than or equal to 15°C; Step d, maintaining the second demolding temperature for a third period of time, and then the demolding process ends.

[0006] Optionally, step a comprises: under a preset As pressure condition, the temperature of the InP substrate is linearly increased to a preset temperature at a first preset rate, and after the substrate temperature reaches the preset temperature, the temperature is immediately linearly increased to a first stripping temperature at a second preset rate, the first preset rate is greater than the second preset rate, and the preset temperature is greater than the reference temperature T c -15℃ to 20℃ low.

[0007] Optionally, the first preset rate ranges from 50°C / min to 100°C / min, and the second preset rate ranges from 10°C / min to 30°C / min.

[0008] Optionally, the first demolding temperature is higher than the reference temperature T c 5℃ lower.

[0009] Optionally, the difference between the second demolding temperature and the reference temperature is in the range of greater than or equal to 5 and less than or equal to 15°C.

[0010] Optionally, the method further comprises a step of optimizing the demolding conditions, wherein the step of optimizing the demolding conditions comprises: Step a1, the second time period is fixed, and for three different first time periods, the corresponding third time period t3 is calculated according to the following relationship between the third time period t3 and the first time period t1, , respectively stripping the InP substrates, and growing epitaxial layers of a preset structure on the stripped InP substrates to obtain three corresponding epitaxial wafers, wherein m represents the expected minimum value of t3 obtained in advance, n represents the expected maximum value of t3 obtained in advance, k is a preset constant, and 0.2≤k≤0.3; Step a2, performing surface defect characterization tests on the three epitaxial wafers respectively to obtain surface defect data corresponding to the three epitaxial wafers; Step a3, performing data fitting on the surface defect data and the three different first time periods to obtain a corresponding relationship between the surface defect data and the first time periods; Step a4, according to the corresponding relationship determined in step a3, within the value range of the first time period, calculate and determine the minimum value S of the surface defect data and the corresponding data tolerance range, the data tolerance range is S to S×(1+f%), and f is a real number from 5 to 15; Step a5: determining a value sub-range of the first time period corresponding to the data tolerance range within the value range of the first time period; Step a6, within the value sub-range, combined with the relationship between the third time period and the first time period in step a1, calculate the value of t1 and the value of t3 corresponding to the minimum sum of t1 and t3, and use the value as the optimized value of the first time period and the third time period.

[0011] Optionally, in step a1, the values ​​of the three different first time periods are 6 min, 10 min, and 14 min, respectively.

[0012] Optionally, m=3 min, and n=8 min.

[0013] Optionally, f=10.

[0014] The beneficial effects of the present invention include: The stripping method in the molecular beam epitaxy process provided by the present invention comprises: step a, under a preset As pressure condition, heating the InP substrate to a first stripping temperature, the first stripping temperature being 4°C to 6°C lower than a pre-obtained reference temperature, the reference temperature being the substrate temperature when the InP substrate is heated to a RHEED pattern showing ×4 reconstruction stripes under the preset As pressure condition; step b, maintaining a first time period at the first stripping temperature, the first time period t1 being in the range of 6 min ≤t1≤14 min; step c, heating from the first stripping temperature to a second stripping temperature, the time required for the heating process being a second time period, the second time period t2 being in the range of 0.5 min ≤t2≤2 min, the difference between the second stripping temperature and the reference temperature being in the range of greater than or equal to zero and less than or equal to 15°C; step d, maintaining a third time period at the second stripping temperature, and then the stripping process ends. The method reduces the desorption of phosphorus by maintaining a first time period at a first stripping temperature slightly lower than a reference temperature, while sufficiently heating the substrate. Due to sufficient heating, the time required for complete desorption of the surface oxide at a second stripping temperature is reduced, thereby significantly reducing surface defects of the epitaxial wafer and improving the surface quality of the epitaxial wafer. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] 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.

[0016] Figure 1A flow chart showing a film stripping method in a molecular beam epitaxy process provided by an embodiment of the present invention is shown; Figure 2 A schematic curve diagram showing the change of substrate temperature over time during the film stripping process provided by an embodiment of the present invention; Figure 3 A schematic diagram of determining a value sub-range of a first time period provided by an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0017] 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.

[0018] InP-based semiconductor devices can be prepared using molecular beam epitaxy technology. During the preparation process, the InP substrate needs to be subjected to high-temperature stripping treatment in a molecular beam epitaxy device to remove the natural oxide on the surface of the InP substrate, and then the corresponding epitaxial layer is deposited and grown according to the device structure design. In the large-scale batch production of InP-based semiconductor devices, the quality of the molecular beam epitaxy high-temperature stripping process directly affects the surface defect density of the prepared InP-based semiconductor device epitaxial wafer, and then affects the yield and reliability of the final device. If the stripping time is too long or too short, the defect density of the substrate surface and the epitaxial wafer will increase. The more surface defects there are on the epitaxial wafer, the lower the yield of the device, and the worse the reliability of the device. Therefore, for the batch production of InP-based semiconductor devices, it is necessary to propose an optimization technology for the molecular beam epitaxy stripping process to reduce the surface defect density of the InP-based semiconductor device epitaxial wafer and improve the quality of the InP-based semiconductor device epitaxial wafer.

[0019] Figure 1 The present invention provides a method for stripping a film in a molecular beam epitaxy process, which is used to strip the film of an InP substrate when growing an InP-based semiconductor device epitaxial wafer by molecular beam epitaxy. Figure 1 As shown, the method includes: Step a: Under a preset As pressure condition, the temperature of the InP substrate is raised to a first stripping temperature T1, which is higher than the pre-obtained reference temperature T c 4℃ to 6℃ lower, base temperature T c The first stripping temperature T1 is higher than the reference temperature T2. c 5℃ lower.

[0020] Generally, different As pressure conditions correspond to different optimal demolding conditions. All demolding processes involved in the embodiments of the present invention are performed under the same preset As pressure conditions. Optionally, the preset As pressure range is: greater than or equal to 1×10 -6 Torr, and less than or equal to 6×10 -6 Torr. For example, the preset As pressure can be 1×10 -6 Torr, 2×10 -6 Torr, or 6×10 -6 Torr, or other values ​​within the above range. It should be noted that the temperatures described in this application refer to the thermocouple temperatures obtained by the thermocouples used to sense the temperature of the corresponding components in the molecular beam epitaxy equipment. For example, in general, for the same molecular beam epitaxy equipment, the reference temperature T of the InP substrate of the same specification is c is stable and constant, therefore, for any specification of InP substrate, the reference temperature T c , which can be obtained in advance.

[0021] Step b: maintaining the first time period t1 at the first demolding temperature T1, wherein the first time period t1 is in the range of 6 min ≤ t1 ≤ 14 min.

[0022] Step c: heating from the first demolding temperature T1 to the second demolding temperature T2. The time required for the heating process is the second time period t2. The range of the second time period t2 is 0.5min ≤ t2 ≤ 2 min. The second demolding temperature T2 minus the reference temperature T c The difference is greater than or equal to zero and less than or equal to 15°C, that is, T c ≤T2≤T c +15°C. Optionally, the second demolding temperature T2 minus the reference temperature T c The difference is greater than or equal to 5 and less than or equal to 15°C, that is, T c +5℃≤T2≤T c +15℃.

[0023] Step d: maintaining the second demolding temperature T2 for a third period of time t3, and then the demolding process ends.

[0024] In conventional stripping technology, the InP substrate is directly heated from the pre-stripping temperature (e.g., linearly heated) to a predetermined stripping temperature (equivalent to the second stripping temperature T2 in this application) and maintained for a predetermined stripping time. When the substrate temperature reaches the predetermined stripping temperature, the oxide film begins to desorb. If the stripping time is too short, the surface oxide cannot be completely desorbed, which leads to defects in the subsequent epitaxial growth process; if the stripping time is too long, phosphorus will be desorbed to form a large number of InAs microcrystals, which will also cause defects in the subsequent epitaxial growth process. Therefore, in conventional technology, surface defects are reduced by optimizing the stripping time (equivalent to the third time period in this application). Since it takes a certain amount of time for the oxide to completely desorb, phosphorus in the area where the oxide desorbs earlier will desorb during the entire stripping process. Although the optimization of the stripping time in conventional technology can improve surface defects, this improvement is limited and it is difficult to further reduce surface defects.

[0025] In the film stripping method provided in the present application, at a temperature higher than the reference temperature T c The first stripping temperature T1, which is 4°C to 6°C lower, needs to be maintained for a first period of time (6min to 14min). By adding this maintenance process, compared with the effect in conventional technology, the actual test results show that the method of the present application can significantly reduce surface defects. The reasons may be as follows: Since the first stripping temperature T1 is lower than the second stripping temperature T2 actually used for oxide film desorption, during the maintenance of the first stripping temperature T1, while the substrate is fully heated, the desorption of phosphorus can be significantly reduced. Since the first stripping temperature T1 has been maintained for 6min to 14min, the substrate has been fully heated. Therefore, when the substrate temperature rises to the second stripping temperature T2, the oxide on the surface of the substrate will immediately begin to desorb and the desorption is relatively uniform. At the same time, the total time required for the oxide to be completely desorbed can be reduced, thereby significantly reducing the surface defect density. It has been experimentally verified that the first stripping temperature T1 is lower than the reference temperature T2. c A temperature of 4 to 6°C lower than the reference temperature T1 can significantly reduce the surface defect density. c If the temperature is too low, the substrate cannot be heated sufficiently, thus affecting the subsequent oxide desorption time. If the first stripping temperature T1 is too close to the reference temperature T c Or greater than the reference temperature T c , which will lead to increased desorption of phosphorus in the first time period. Similarly, the range of the first time period is also a consideration. If the first time period is less than 6 minutes, the substrate cannot be fully heated; if the first time period is too long (for example, greater than 14 minutes), it will lead to unnecessary waste of growth time.

[0026] Optionally, step a includes: under a preset As pressure condition, the temperature of the InP substrate is linearly increased to a preset temperature at a first preset rate, and after the substrate temperature reaches the preset temperature, the temperature is immediately linearly increased to a first stripping temperature T1 at a second preset rate, the first preset rate is greater than the second preset rate, and the preset temperature is higher than the reference temperature T c 15°C to 20°C lower. Optionally, the first preset rate ranges from 50°C / min to 100°C / min, and the second preset rate ranges from 10°C / min to 30°C / min. By introducing the preset temperature, it is possible to quickly heat up to the preset temperature and reduce the heating time, and then reduce the heating rate from the preset temperature to the first demolding temperature T1 to avoid temperature overshoot.

[0027] The specific demolding heating process is as follows: Figure 2 As shown, from the start of stripping to time f1, the substrate temperature is increased to the preset temperature T0 at a first preset rate. After the substrate temperature reaches the preset temperature T0, it is immediately increased to the first stripping temperature T1 at a second preset rate. The process of increasing from the preset temperature T0 to the first stripping temperature T1 corresponds to from time f1 to time f2. Then, the first time period is maintained at the first stripping temperature T1, which corresponds to from time f2 to time f3. The process of increasing from the first stripping temperature T1 to the second stripping temperature T2 corresponds to from time f3 to time f4. The process of maintaining the third time period at the second stripping temperature T2 corresponds to from time f4 to time f5. At time f5, the stripping process ends.

[0028] In summary, the method reduces the desorption of phosphorus by maintaining the first stripping temperature slightly lower than the reference temperature for a first period of time, while sufficiently heating the substrate. Due to sufficient heating, the desorption of oxides on the substrate surface is relatively uniform, which reduces the time required for complete desorption of surface oxides at the second stripping temperature, thereby significantly reducing surface defects of epitaxial wafers and improving the surface quality of epitaxial wafers.

[0029] Optionally, the method further comprises a demolding condition optimization step, wherein the demolding condition optimization step is used to optimize the duration of the first time period and the third time period.

[0030] The steps to optimize the stripping conditions include: Step a1, the second time period is fixed, and for three different first time periods, the corresponding third time period t3 is calculated according to the following relationship between the third time period t3 and the first time period t1, , respectively, the InP substrates are stripped, and an epitaxial layer of a preset structure is grown on the stripped InP substrates to obtain three corresponding epitaxial wafers, wherein m represents the expected minimum value of t3 obtained in advance, n represents the expected maximum value of t3 obtained in advance, k is a preset constant, and 0.2≤k≤0.3. Optionally, in step a1, the values ​​of the three different first time periods are 6 min, 10 min, and 14 min, respectively. The present application does not limit the specific structure of the preset structure. For example, an epitaxial structure including materials such as InP, InGaAs, and InAlAs can be grown, as long as the structures used for the three epitaxial growths in the three different first time periods are consistent.

[0031] Step a2: Perform surface defect characterization tests on the three epitaxial wafers respectively to obtain surface defect data corresponding to the three epitaxial wafers. Step a3: Perform data fitting on the surface defect data and the three different first time periods to obtain the corresponding relationship between the surface defect data and the first time period. Step a4: According to the corresponding relationship determined in step a3, within the value range of the first time period, calculate and determine the minimum value S of the surface defect data and the corresponding data tolerance range, the data tolerance range is S to S×(1+f%), and f is a real number from 5 to 15. Optionally, f=10. Step a5: Within the value range of the first time period, determine the value sub-range of the first time period corresponding to the data tolerance range. Step a6: Within the value sub-range, in combination with the relationship between the third time period and the first time period in step a1, calculate the value of t1 and the value of t3 corresponding to the minimum sum of t1 and t3, and use the value as the optimized value of the first time period and the third time period.

[0032] Since the temperature rise in the second time period is not large (about 5°C to 20°C) and the duration is short, the second time period remains unchanged during the optimization process. For example, the second time period can be fixed to 1 minute. Based on experience, the minimum and maximum values ​​of the third time period can be expected. The minimum value here means that if it is less than the minimum value, defects caused by the oxide film not being desorbed cleanly will be significantly found; the maximum value here means that if it exceeds the maximum value, defects caused by InAs microcrystals caused by phosphorus desorption will be significantly found. For example, based on experience, the expected minimum value m=3 min can be set, and the expected maximum value n=8 min can be set.

[0033] After setting the values ​​of m and n, for any first time period, there is theoretically an optimal third time period corresponding thereto. The relationship between the corresponding optimal third time period and the first time period is discussed below. As the first time period increases, the corresponding optimal third time period decreases accordingly, but the minimum value is m; as the first time period decreases, the corresponding optimal third time period increases accordingly, but the maximum value is n, so that a relationship can be established to approximately characterize the value of the optimal third time period t3 corresponding to the first time period t1. The optimal third time period here means: for any first time period, within the value range of the third time period, the third time period corresponding to the minimum surface defects. By establishing the above relationship, when optimizing the time period in the stripping process, the variables can be reduced to only one (that is, the first time period). In this case, only three experiments can be used to obtain the data for optimizing the first time period (that is, the data obtained in step a2), and then the corresponding relationship is established by step a3, and the minimum value S is determined in step a4, so that a large number of experiments when optimizing two or more variables are avoided, and the optimization cost is greatly reduced. Considering the errors caused by various reasons such as random factors, the expected optimal surface defect data is considered acceptable in the range of S to S×(1+f%). Therefore, for S to S×(1+f%), that is, for the data tolerance range, a corresponding value sub-range of the first time period can be determined.

[0034] like Figure 3 As shown, the surface defect data is the number of surface defects per unit area, the experimental test data is the data corresponding to the black dots, and the data fitting is performed to obtain the fitting corresponding relationship curve (the dotted line in the figure), determine the minimum value S on the corresponding relationship curve, and then determine the data tolerance range. Finally, the value sub-range of the first time period corresponding to the data tolerance range can be determined to be st1 to st2. It can be considered that any value of the first time period within the value sub-range is optimal for the corresponding surface defect data. In order to save the total demolding time and machine time, the relationship between the third time period and the first time period in step a1 can be combined to calculate the value of t1 and the value of t3 corresponding to the minimum sum of t1 and t3, and use the value as the optimized value of the first time period and the third time period. In the subsequent demolding process, the optimized values ​​of the first time period and the third time period are used, which can not only obtain smaller surface defect data, but also reduce the total demolding time and avoid unnecessary time waste.

[0035] 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 film stripping method in a molecular beam epitaxy process, characterized in that: The method is used to strip an InP substrate, and the method comprises: Step a, under a preset As pressure condition, heating the InP substrate to a first stripping temperature, the first stripping temperature being 4°C to 6°C lower than a pre-obtained reference temperature, the reference temperature being the substrate temperature when the InP substrate is heated under a preset As pressure condition to a RHEED pattern showing ×4 reconstruction fringes; Step b, maintaining the first time period at the first demolding temperature, wherein the first time period t1 is in the range of 6 min ≤ t1 ≤ 14 min; Step c, heating from the first demolding temperature to the second demolding temperature, the time required for the heating process is the second time period, the range of the second time period t2 is 0.5min ≤ t2 ≤ 2 min, and the range of the difference between the second demolding temperature and the reference temperature is greater than or equal to zero and less than or equal to 15°C; Step d, maintaining the second demolding temperature for a third period of time, and then the demolding process ends.

2. The film stripping method in the molecular beam epitaxy process according to claim 1, characterized in that: Step a includes: under a preset As pressure condition, linearly increasing the temperature of the InP substrate at a first preset rate to a preset temperature, and immediately linearly increasing the temperature at a second preset rate to a first stripping temperature after the substrate temperature reaches the preset temperature, the first preset rate is greater than the second preset rate, and the preset temperature is 15°C to 20°C lower than the reference temperature.

3. The film stripping method in the molecular beam epitaxy process according to claim 2, characterized in that: The first preset rate ranges from 50° C. / min to 100° C. / min, and the second preset rate ranges from 10° C. / min to 30° C. / min.

4. The film stripping method in the molecular beam epitaxy process according to claim 1, characterized in that: The first demolding temperature is 5°C lower than the reference temperature.

5. The film stripping method in the molecular beam epitaxy process according to claim 1, characterized in that: The difference between the second demolding temperature and the reference temperature is in the range of 5 or more and 15° C. or less.

6. The film stripping method in the molecular beam epitaxy process according to claim 1 or 5, characterized in that: The method further comprises a step of optimizing the stripping conditions, wherein the step of optimizing the stripping conditions comprises: Step a1, the second time period is fixed unchanged, for three different first time periods, the corresponding third time period t3 is calculated according to the following relationship between the third time period t3 and the first time period t1, , respectively stripping the InP substrates, and growing epitaxial layers of a preset structure on the stripped InP substrates to obtain three corresponding epitaxial wafers, wherein m represents the expected minimum value of t3 obtained in advance, n represents the expected maximum value of t3 obtained in advance, k is a preset constant, and 0.2≤k≤0.3; Step a2, performing surface defect characterization tests on the three epitaxial wafers respectively to obtain surface defect data corresponding to the three epitaxial wafers; Step a3, performing data fitting on the surface defect data and the three different first time periods to obtain a corresponding relationship between the surface defect data and the first time periods; Step a4, according to the corresponding relationship determined in step a3, within the value range of the first time period, calculate and determine the minimum value S of the surface defect data and the corresponding data tolerance range, the data tolerance range is S to S×(1+f%), and f is a real number from 5 to 15; Step a5: determining a value sub-range of the first time period corresponding to the data tolerance range within the value range of the first time period; Step a6, within the value sub-range, combined with the relationship between the third time period and the first time period in step a1, calculate the value of t1 and the value of t3 corresponding to the minimum sum of t1 and t3, and use the value as the optimized value of the first time period and the third time period.

7. The film stripping method in the molecular beam epitaxy process according to claim 6, characterized in that: In the step a1, the values ​​of the three different first time periods are 6 min, 10 min, and 14 min respectively.

8. The film stripping method in the molecular beam epitaxy process according to claim 6, characterized in that: m=3 min, and n=8 min.

9. The film stripping method in the molecular beam epitaxy process according to claim 6, characterized in that: f=10。

Images