Gallium nitride epitaxy stability enhancement method incorporating surface mobility calibration

Abstract

This invention discloses a method for enhancing the stability of gallium nitride epitaxial growth by combining surface mobility calibration. The method involves neutralizing polarization residues on the growth layer surface through inert gas impingement and plasma activation to generate a regrowth substrate interface. Then, energy pulse-assisted control is used to alternately heat and cool the growth layer surface to obtain a regrowth layer of a predetermined thickness. The regrowth layer is then heated, and the ratio of ammonia to metal source flow rate is adjusted to generate a nitrogen-rich surface in the initial heating stage. Stress redistribution is performed in stress concentration areas to obtain a thick-layer sample. The mobility of the thick-layer sample is calibrated using a closed-loop method by comparing the mobility of a preset model with the measured surface diffusion response time to correct the timing of pulsed thermal flow period and nitrogen-rich surface formation. This invention solves the technical problems of delayed surface activity recovery, interlayer stress inhomogeneity, and rapid local dislocation propagation during the low-temperature regrowth stage, significantly improving the crystal quality and process stability of gallium nitride epitaxial layers.

Description

Technical Field

[0001] This application relates to the field of process calibration technology, and in particular to a method for enhancing the stability of gallium nitride epitaxy by combining surface mobility calibration. Background Technology

[0002] Gallium nitride (GaN), as a typical representative of third-generation semiconductor materials, has shown broad application prospects in power electronics, radio frequency communication, and optoelectronic conversion due to its excellent properties such as wide bandgap, high breakdown field strength, and high electron saturation drift velocity. With the development of power devices towards high voltage, high frequency, and high reliability, higher requirements are placed on the thickness, crystal quality, and stress control of GaN epitaxial layers. With the development of low-temperature thick-layer regeneration processes (i.e., multi-cycle high-temperature-low-temperature alternation) to improve stress relief capabilities, the diffusion distance of surface adsorbed atoms is extremely short at low temperatures, and the mobility is almost zero, causing local surface instability. Although existing mobility calibration systems issue temperature adjustment signals, the surface of the regenerated layer is covered with residual polarization charge and roughness from the previous layer, making it difficult to immediately restore surface diffusion activity upon temperature increase. This leads to a "freeze-thaw delay" phenomenon, resulting in uneven release of interlayer stress and rapid diffusion of local dislocations during the thawing stage. Summary of the Invention

[0003] This application provides a gallium nitride epitaxial stability enhancement method that combines surface mobility calibration, which is used to solve the technical problems in the prior art where interlayer stress is difficult to release uniformly and local dislocations spread rapidly during the thawing stage.

[0004] The present invention adopts the following technical solution.

[0005] The first aspect of this invention discloses a method for enhancing the stability of gallium nitride epitaxial growth by incorporating surface mobility calibration, the method comprising:

[0006] The surface of the growth layer is neutralized by inert gas impact and plasma activation to generate a regenerated substrate interface.

[0007] Based on the regenerated substrate interface, the surface of the growth layer is controlled by energy pulses to alternate between heating and cooling in order to obtain a regenerated layer of a set thickness.

[0008] The regenerated layer is heated at a set rate, and the ratio of ammonia to metal source flow rate is adjusted through a surface diffusion activation mechanism to generate a nitrogen-rich surface in the early stage of heating.

[0009] Real-time stress field monitoring was performed on the regenerated layer after heating treatment, and stress redistribution was carried out in the stress concentration area by fine-tuning the cavity gas pressure and carrier gas flow direction to obtain a thick layer sample.

[0010] The migration rate of the thick sample is calibrated by comparing the migration rate of the preset model with the measured surface diffusion response time to correct the pulse heat flow cycle and the generation sequence of the nitrogen-rich surface.

[0011] Furthermore, the polarization residue neutralization treatment of the growth layer surface through inert gas impact and plasma activation to generate a regrowth substrate interface includes:

[0012] After the high-temperature stage ends and before low-temperature regrowth, the epitaxial wafer is controlled at a preset transition temperature, and the total polarization charge on the surface of the growth layer is measured by an in-situ reflective ellipsoid or contact electrostatic sampling device.

[0013] The initial polarization surface charge density of the growth layer is calculated, and the average roughness of the growth layer surface is monitored using in-situ surface scattering or optical roughness monitoring technology. The initial polarization surface charge density is the ratio between the total polarization charge and the effective epitaxial area of ​​the sample.

[0014] Furthermore, the process of neutralizing polarization residues on the surface of the growth layer through inert gas impact and plasma activation to generate a regrowth substrate interface also includes:

[0015] Based on the initial polarization surface charge density and average roughness, the inert gas is controlled to impact the cavity according to the set instantaneous gas flow rate and gas impact time, so as to peel off the adsorbed residues and loaded clusters on the surface of the growth layer.

[0016] The first surface charge density of the grown layer surface after the inert gas impact is calculated based on the charge reduction coefficient of the inert gas impact and the initial polarization surface charge density, and the improved average roughness after the inert gas impact is calculated based on the roughness improvement coefficient of the inert gas impact and the average roughness of the grown layer surface.

[0017] Furthermore, the process of neutralizing polarization residues on the surface of the growth layer through inert gas impact and plasma activation to generate a regrowth substrate interface also includes:

[0018] Under a uniform cavity vacuum, a low-power plasma source is used to excite the surface of the growth layer with the first surface charge density and improved average roughness according to the set excitation power and plasma action time, so as to control the migration of the accumulated surface polarization charge to the surrounding area, and obtain the second surface charge density and activation average roughness of the growth layer surface after plasma activation.

[0019] The epitaxial surface under the second surface charge density and activation average roughness is cooled to the set regeneration preparation temperature at a set cooling rate, and the surface of the growth layer is controlled to be in a thermodynamically stable state at the regeneration preparation temperature to generate the regeneration substrate interface.

[0020] The thermodynamically stable state is a thermodynamic state in which both the second surface charge density and the average activation roughness are lower than the corresponding process set threshold.

[0021] Furthermore, the step of using energy pulses to assist in controlling the surface of the growth layer to alternate between heating and cooling, based on the regenerated substrate interface, to obtain a regenerated layer of a set thickness, includes:

[0022] Based on the surface charge density and average roughness of the regenerated substrate interface, the mobility target in the low-temperature section and the pulsed heat flow reference parameters corresponding to the mobility target are determined, and the regenerated substrate interface is periodically heated according to the pulsed heat flow reference parameters to determine the instantaneous surface mobility under pulse action.

[0023] The nascent low-temperature epitaxial layer that has been pulse-excited is determined based on the instantaneous surface mobility, and the ammonia or nitrogen carrier gas flow rate of the nascent low-temperature epitaxial layer is increased within a set time window and then restored to determine the effective surface mobility.

[0024] Determine whether the effective surface mobility falls within the set range. If not, increase the ammonia or nitrogen carrier gas flow rate of the nascent low-temperature epitaxial layer again within the set time window and then restore it.

[0025] Furthermore, the step of heating the regenerated layer at a set rate and adjusting the ratio of ammonia to metal source flow rate through a surface diffusion activation mechanism to generate a nitrogen-rich surface in the initial stage of heating includes:

[0026] The temperature of the regenerated layer is linearly increased to the target regeneration temperature according to the set heating rate, and the brightness and temperature of the surface of the regenerated layer are monitored by in-situ optical reflection, infrared thermometry or electron diffraction technology to output the temperature distribution uniformity and reflection intensity change.

[0027] Based on the temperature distribution uniformity and reflection intensity changes, the medium-low temperature zone of the regenerated layer surface is determined. When the surface of the regenerated layer is heated to the medium-low temperature zone, the ammonia flow rate is increased while the carrier gas flow rate is reduced to generate a nitrogen-rich surface state.

[0028] During the heating process in the nitrogen-rich surface state, the heating is controlled by adjusting the power distribution in different areas of the cavity to obtain a temperature-uniformed epitaxial layer with a surface temperature difference lower than the set temperature difference threshold.

[0029] When the temperature of the temperature homogenization epitaxial layer reaches the target regrowth temperature, the stress change rate in different regions within the cavity is monitored, and when the stress change rate exceeds the average value, the stress in the local region is diffused.

[0030] Furthermore, the real-time stress field monitoring of the regenerated layer after the heating treatment, and the stress redistribution in the stress concentration area by fine-tuning the cavity gas pressure and carrier gas flow direction, to obtain a thick-layer sample, includes:

[0031] Read real-time stress data of different radial and angular directions obtained by online optical interference or wafer bending monitoring of epitaxial layer under isothermal growth conditions, and divide the real-time stress data into regions and stress levels to generate a stress distribution table;

[0032] Adjust the carrier gas or nitrogen guiding components in the cavity marked as high stress areas in the stress distribution table to control the surface temperature of the high stress areas to be lower than the surface temperature of other areas, reduce the effective incident flux of the high stress areas, and increase the effective incident flux of the low stress areas.

[0033] Furthermore, the step of performing closed-loop mobility calibration on the thick sample by comparing the preset model mobility with the measured surface diffusion response time to correct the pulsed thermal flux cycle and the generation sequence of the nitrogen-rich surface includes:

[0034] Based on the migration rate data of the thick sample at each stage, the migration rate deviation between the measured migration rate value and the target migration rate value at the corresponding stage is calculated to determine whether the migration rate deviation exceeds the allowable deviation threshold.

[0035] When the migration rate deviation exceeds the allowable deviation threshold, a correction amount is generated based on the migration rate deviation at the corresponding stage, and the heating rate, pulse heat flow, and ammonia or nitrogen carrier gas flow rate are adjusted according to the correction amount to control the migration rate deviation at the corresponding stage to meet the allowable deviation threshold.

[0036] A second aspect of the present invention discloses a gallium nitride epitaxial stability enhancement device incorporating surface mobility calibration, for implementing the gallium nitride epitaxial stability enhancement method incorporating surface mobility calibration as described in any one of the first aspects, the device comprising:

[0037] The substrate interface generation module is used to neutralize polarization residues on the surface of the growth layer through inert gas impact and plasma activation in order to generate a regenerated substrate interface.

[0038] The regenerated layer acquisition module is used to obtain a regenerated layer of a set thickness by using energy pulse-assisted control to keep the surface of the growth layer in an alternating state of heating and cooling based on the interface of the regenerated substrate.

[0039] The nitrogen-rich surface generation module is used to heat the regenerated layer at a set rate and adjust the ratio of ammonia to metal source flow through a surface diffusion activation mechanism to generate a nitrogen-rich surface in the initial stage of heating.

[0040] The stress redistribution module is used to monitor the stress field of the regenerated layer after heating treatment in real time, and redistribute the stress in the stress concentration area by finely adjusting the cavity air pressure and the direction of the carrier gas flow field to obtain a thick layer sample.

[0041] The mobility calibration module is used to perform closed-loop mobility calibration on the thick sample by comparing the migration rate of the preset model with the measured surface diffusion response time, so as to correct the pulse heat flow cycle and the generation sequence of the nitrogen-rich surface.

[0042] A third aspect of the present invention discloses a terminal, including a processor and a storage medium;

[0043] The storage medium is used to store instructions;

[0044] The processor is configured to operate according to the instructions to perform the steps of the method described in the first aspect.

[0045] A fourth aspect of the present invention discloses a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method described in the first aspect.

[0046] The beneficial effects of this invention are:

[0047] (1) This invention redistributes the surface polarization charge formed in the previous cycle by introducing inert gas impact and low-power plasma activation, while removing surface adsorption residues and rough micro-island structures to form an electrically neutral and smooth regeneration substrate interface. At the same time, a low-energy pulsed heat flow auxiliary system is set up to keep the surface in an alternating state of "micro-heating-micro-cooling". The heat flow pulse period is synchronized with the epitaxial deposition rate, so that the adsorbed atoms maintain a limited diffusion distance in the local energy fluctuation region, effectively avoiding the formation of a completely frozen layer, and enabling the mobility to maintain a stable lower limit value in the low temperature region.

[0048] (2) This invention introduces a surface diffusion activation mechanism, namely, synchronously adjusting the flow ratio of ammonia to metal source, to form a temporary nitrogen-rich surface in the early stage of heating, thereby enhancing the adsorption stability of atomic bonds and preventing surface roughening caused by sudden heating. At the same time, a thermal inertia compensation module is set up to keep the temperature rise rate consistent in different regions, effectively eliminating the spatial difference of "thawing delay". In addition, when a high stress area is detected in a local area, a slight cold end of airflow is formed in the stress concentration area by finely adjusting the cavity air pressure and the direction of the carrier gas flow field, which causes the thermal expansion rate of the area to decrease, causing the interlayer stress to redistribute. After balancing treatment, the dislocation paths that may have diffused can also be effectively suppressed under the weakening stress gradient. Attached Figure Description

[0049] Figure 1This is a schematic flowchart of the gallium nitride epitaxial stability enhancement method combined with surface mobility calibration provided by the present invention.

[0050] Figure 2 This is a schematic diagram of the structure of the gallium nitride epitaxial stability enhancement device combined with surface mobility calibration provided by the present invention. Detailed Implementation

[0051] The present invention will be further described below with reference to the accompanying drawings and specific embodiments. The illustrative embodiments and descriptions herein are used to explain the present invention, but are not intended to limit the present invention.

[0052] like Figure 1 As shown, in one embodiment, a gallium nitride epitaxial stability enhancement method incorporating surface mobility calibration includes the following steps:

[0053] Step S110 involves neutralizing the polarization residue on the surface of the growth layer by inert gas impact and plasma activation to generate a regrowth substrate interface.

[0054] In some embodiments, the gallium nitride epitaxial stability enhancement method combined with surface mobility calibration provided by the present invention specifically includes the following steps in step S110:

[0055] Step S111: After the high-temperature stage ends and before low-temperature regrowth, the epitaxial wafer is controlled at a preset transition temperature, and the total polarization charge on the surface of the growth layer is measured by an in-situ reflective ellipsoid or contact electrostatic sampling device.

[0056] Step S112: Calculate the initial polarization surface charge density of the growth layer surface, and monitor the average roughness of the growth layer surface using in-situ surface scattering or optical roughness monitoring technology. The initial polarization surface charge density is the ratio between the total polarization charge and the effective epitaxial area of ​​the sample.

[0057] In some embodiments, the gallium nitride epitaxial stability enhancement method combined with surface mobility calibration provided by the present invention further includes the following steps in step S110:

[0058] Step S113: Based on the initial polarization surface charge density and average roughness, the inert gas is controlled to impact the cavity according to the set instantaneous gas flow rate and gas impact time, so as to peel off the adsorbed residues and loaded clusters on the surface of the growth layer.

[0059] Step S114: Calculate the first surface charge density of the grown layer surface after inert gas impact based on the charge reduction coefficient of inert gas impact and the initial polarization surface charge density, and calculate the improved average roughness after inert gas impact based on the roughness improvement coefficient of inert gas impact and the average roughness of the grown layer surface.

[0060] In some embodiments, the gallium nitride epitaxial stability enhancement method combined with surface mobility calibration provided by the present invention further includes the following steps in step S110:

[0061] Step S115: Under a uniform cavity vacuum, a low-power plasma source is used to excite the surface of the growth layer with the first surface charge density and improved average roughness according to the set excitation power and plasma action time, so as to control the migration of the accumulated surface polarization charge to the surrounding area, and obtain the second surface charge density and activation average roughness of the growth layer surface after plasma activation.

[0062] Step S116: The epitaxial surface under the second surface charge density and activation average roughness is cooled to the set regeneration preparation temperature at a set cooling rate, and the surface of the growth layer is controlled to be in a thermodynamically stable state at the regeneration preparation temperature to generate the regeneration substrate interface.

[0063] The thermodynamically stable state is the thermodynamic state in which both the second surface charge density and the average activation roughness are lower than the corresponding process set thresholds.

[0064] In a specific embodiment, the gallium nitride epitaxial stability enhancement method combined with surface mobility calibration provided by the present invention includes steps 1 to 5:

[0065] Step 1: Interface preprocessing and polarization residue neutralization.

[0066] After the high-temperature stage and before entering the low-temperature regeneration stage, the surface of the grown layer undergoes polarization residue neutralization treatment. This is achieved through a short-term introduction of inert gas impact and low-power plasma activation, which redistributes the surface polarization charges formed in the previous cycle, while simultaneously removing surface adsorption residues and rough micro-island structures, forming an electrically neutral and smooth regeneration substrate interface. This includes the following sub-steps:

[0067] Sub-step 1.1: Initial confirmation of surface polarization and roughness.

[0068] Specifically, the epitaxial wafer, having just completed its high-temperature phase but not yet transitioned to low-temperature regrowth, is kept at a stable transition temperature, and the total surface polarization charge is measured using in-situ reflection / ellipsoidal or contact electrostatic sampling devices. The expression for calculating the surface charge density is:

[0069]

[0070] In the formula, The surface charge density is the initial polarization of the surface, expressed in coulombs per square meter. The measured total surface polarization charge, in coulombs, is obtained from an in-situ electrostatic sampler. The effective area of ​​the epitaxial growth layer prepared for low-temperature regrowth is given in square meters and is directly specified by the process formula.

[0071] Subsequently, the average surface roughness is obtained by in-situ surface scattering or optical roughness monitoring, and the average surface roughness is generally in the range of less than ten nanometers.

[0072] Sub-step 1.2: Short-term impact desorption of inert gas.

[0073] Specifically, without altering the overall cavity structure, argon gas is injected into the cavity at an instantaneous flow rate 1.5 to 3 times higher than the conventional growth flow rate to create an impact, with the impact duration maintained on the order of several seconds to tens of seconds. This causes the weakly adsorbed residues and loaded surface microclusters to be stripped away. The surface charge density can then be estimated using the following formula:

[0074]

[0075] In the formula, The surface charge density after an inert gas impact; is the charge reduction coefficient for inert gas impact, dimensionless, with an empirical range of 0.20 to 0.50, obtained through process calibration based on actual impact flow rate and time. The average surface roughness decreases slightly after weak deposits are washed away, expressed as:

[0076]

[0077] In the formula, The average roughness after inert gas impact; The roughness improvement factor ranges from 0.05 to 0.20 and is determined by the airflow alignment and the rotation state of the epitaxial wafer. The average surface roughness obtained in sub-step 1.1.

[0078] Sub-step 1.3, low-power plasma activation and polarization redistribution.

[0079] Specifically, while maintaining the same vacuum level in the process chamber, a low-power plasma source is activated, with the power controlled within a range sufficient to excite the surface without causing re-etching. For GaN epitaxy, the power is typically controlled in the tens to hundreds of watts range, and the time is controlled in the range of several to tens of seconds. The slight bombardment and chemical activation formed by the plasma on the surface will cause the still accumulated surface polarization charges to migrate to the surrounding low-potential region, the result of which can be expressed by the following formula:

[0080]

[0081] In the formula, The surface charge density after plasma activation; The coefficient for secondary reduction caused by plasma, ranging from 0.10 to 0.40, is calibrated by a combination of plasma power and treatment time. Plasma further weakens surface protrusions and fills in some low-potential points; the roughness can be expressed as:

[0082]

[0083] In the formula, The average roughness after plasma activation; The coefficient for optimizing the roughness of the plasma is 0.05 to 0.15. This process also generates a highly active epitaxial initiation layer with a relatively uniform chemical potential on the surface.

[0084] Sub-step 1.4: Maintaining the leveled steady state and confirming the regrowth interface.

[0085] Specifically, the epitaxial surface treated in sub-step 1.3 is slowly cooled to the low-temperature regrowth preparation temperature and held at this temperature for a short period to allow the plasma-activated surface to regain thermodynamic stability and confirm the surface charge density. The surface roughness has fallen below the process-defined threshold, and this has been confirmed. The value is already below the process-defined threshold. These two thresholds are fixed process constants, predetermined by the requirements for electrical neutrality and step height during thick-layer regrowth of the product. Generally, the surface charge density process-defined threshold is set to [value missing]. For less than 30% of the roughness, the roughness process setting threshold is set to 0. Less than half of the actual value. If measured... Still above the surface charge density process threshold, short-duration plasma can be repeated; if It is still higher than the roughness process setting threshold, which can extend the steady-state holding time to complete the surface rearrangement.

[0086] Step S120: Based on the regenerated substrate interface, the surface of the growth layer is controlled by energy pulses to alternate between heating and cooling in order to obtain a regenerated layer of a set thickness.

[0087] In some embodiments, the gallium nitride epitaxial stability enhancement method combined with surface mobility calibration provided by the present invention specifically includes the following steps in step S120:

[0088] Step S121: Based on the surface charge density and average roughness of the regenerated substrate interface, determine the mobility target of the low-temperature section and the pulsed heat flow reference parameters corresponding to the mobility target, and periodically heat the regenerated substrate interface according to the pulsed heat flow reference parameters to determine the instantaneous surface mobility under pulse action.

[0089] Step S122: Determine the nascent low-temperature epitaxial layer that has been pulse-excited based on the instantaneous surface mobility, and increase the ammonia or nitrogen carrier gas flow rate of the nascent low-temperature epitaxial layer within a set time window before restoring it, so as to determine the effective surface mobility.

[0090] Step S123: Determine whether the effective surface mobility falls within the set range. If not, increase the ammonia or nitrogen carrier gas flow rate of the nascent low-temperature epitaxial layer again within the set time window and then restore it.

[0091] In a specific embodiment, the gallium nitride epitaxial stability enhancement method combined with surface mobility calibration provided by the present invention includes step 2, dynamic maintenance of surface mobility in the low-temperature stage. When entering the low-temperature thick-layer regrowth stage at the neutralization interface output in step 1, a low-energy pulsed heat flow auxiliary system is set up to keep the surface in an alternating state of "micro-heating - micro-cooling". The period of this heat flow pulse is synchronously matched with the epitaxial deposition rate, allowing adsorbed atoms to maintain a limited diffusion distance in the local energy fluctuation region without forming a completely frozen layer. This method enables the mobility to maintain a stable lower limit value in the low-temperature region, including the following sub-steps:

[0092] Sub-step 2.1: Setting the target range for low-temperature mobility.

[0093] Specifically, after obtaining the interface data output in step 1, first according to Determine the surface charge neutrality level, and then based on Determine whether the step height amplifies the diffusion barrier at low temperatures. The surface charge density is below the set threshold, which is the process set threshold. If the roughness of the process is below a set threshold, the target mobility for the low-temperature range can be set at the lower limit of normal; if it is close to the threshold, the target should be set slightly higher to avoid freezing. The target mobility is calculated using the following formula:

[0094]

[0095] In the formula, The target value for surface mobility to be maintained in the low-temperature range; The empirical low-temperature mobility benchmark value for this equipment in this material system was obtained from long-term process statistics. To increase the coefficient, the range is 0 to 0.30, when , The closer to the threshold, The closer the value is to the upper limit.

[0096] Then The pulsed heat flow is then mapped in reverse to the reference power and pulse period, forming a set of pulse reference parameters.

[0097] Sub-step 2.2, pulsed heat flow injection and surface energy maintenance.

[0098] Specifically, after lowering the epitaxial cavity to the set low-temperature regrowth temperature, the surface is not kept at a constant low temperature for an extended period. Instead, it is periodically micro-heated according to the pulse reference given in step 2.1. The peak power of the pulsed heat flux is controlled between 30% and 60% of the device's allowable upper limit, the pulse duration is controlled between several hundred milliseconds and several seconds, and the pulse interval is controlled between several seconds and tens of seconds, causing the surface to oscillate within a relatively narrow temperature fluctuation band. In this way, surface adsorbed atoms do not remain at the lowest diffusion energy point for an extended period, thus ensuring that the mobility remains around [a certain value]. It fluctuates slightly up and down. This process can be characterized by the following formula:

[0099]

[0100] In the formula, This represents the actual instantaneous surface mobility formed under pulse action; This is the mobility fluctuation coefficient caused by the pulse, ranging from 0 to 0.10. The smaller the value, the more stable the surface diffusion. If a lag in temperature rise is detected in some areas, the pulse interval can be temporarily shortened in those areas to compensate.

[0101] Sub-step 2.3, atmosphere and pulse-coordinated diffusion lower limit locking.

[0102] Specifically, a heat pulse alone can sometimes cause the surface energy to drop too quickly. Therefore, it is necessary to simultaneously fine-tune the atmosphere in the low-temperature zone to extend the lifetime of the adsorbed state on the surface. Within a short time window after the pulsed heat flow is applied, the ammonia or nitrogen carrier gas flow rate is increased to 1.05 to 1.20 times the upper limit of the low-temperature zone, maintained for several seconds to tens of seconds, and then returned to the baseline flow rate of the low-temperature zone. The purpose of this is to utilize the brief enrichment of the atmosphere to slow down the rate of surface energy loss, preventing the diffusion capacity from immediately dropping after the pulse ends. This process can be described by the following equation:

[0103]

[0104] In the formula, The effective surface mobility under the combined effects of heat flow and atmosphere; This is the atmosphere synergistic enhancement coefficient, ranging from 0 to 0.08, primarily used to compensate for areas where the local cooling rate is too rapid or where slight polarization remains at the interface. If Still below If the value is 90%, it indicates that the atmosphere holding time is insufficient or the pulse interval is too long, and the pulse parameters of the previous step need to be adjusted back.

[0105] Sub-step 2.4: Confirm the effect of maintaining the low temperature range.

[0106] Specifically, after completing one or two pulse-atmosphere co-processing cycles, a rapid scan of the surface is performed to confirm that the surface temperature fluctuation is within the allowable range, and to confirm that no new rough islands have appeared due to excessive energy, nor has localized deactivation occurred due to insufficient energy. If the mobility points obtained from the scan are all within the acceptable range... If the concentration is between 90% and 110%, it indicates that the low temperature was successfully maintained in this cycle. If there is a local area below 90%, sub-steps 2.2 and 2.3 need to be reverted to shorten the pulse interval or extend the atmosphere enrichment time for that area until the entire film enters the target range.

[0107] In step S130, the regenerated layer is heated at a set rate, and the ratio of ammonia to metal source flow rate is adjusted through a surface diffusion activation mechanism to generate a nitrogen-rich surface in the initial stage of heating.

[0108] In some embodiments, the gallium nitride epitaxial stability enhancement method combined with surface mobility calibration provided by the present invention specifically includes the following steps in step S130:

[0109] Step S131: The temperature of the regenerated layer is linearly increased to the target regeneration temperature according to the set heating rate, and the brightness and temperature of the surface of the regenerated layer are monitored by in-situ optical reflection, infrared thermometry or electron diffraction technology to output the temperature distribution uniformity and reflection intensity change.

[0110] Step S132: Based on the temperature distribution uniformity and reflection intensity changes, determine the medium-low temperature zone on the surface of the regenerated layer, and when the surface of the regenerated layer is heated to the medium-low temperature zone, increase the ammonia flow rate while reducing the carrier gas flow rate to generate a nitrogen-rich surface state.

[0111] In step S133, during the heating process in the nitrogen-rich surface state, the heating is controlled by adjusting the power distribution in different areas of the cavity to obtain a temperature-uniformed epitaxial layer with a surface temperature difference lower than the set temperature difference threshold.

[0112] Step S134: When the temperature of the temperature homogenization epitaxial layer reaches the target regrowth temperature, the stress change rate in different regions of the cavity is monitored, so that when the stress change rate exceeds the average value, the stress in the local area is diffused.

[0113] In a specific embodiment, the gallium nitride epitaxial stability enhancement method combined with surface mobility calibration provided by the present invention includes step 3, surface diffusion activation and delay compensation during the heating stage. After the low-temperature regenerated layer reaches a predetermined thickness, the system is gradually heated. At this time, a surface diffusion activation mechanism is introduced, that is, the ratio of ammonia to metal source flow rate is simultaneously adjusted to form a transient nitrogen-rich surface in the early stage of heating, enhancing the atomic bond adsorption stability and preventing surface roughening caused by sudden heating. Simultaneously, a thermal inertia compensation mechanism is set to maintain a consistent temperature rise rate in different regions, eliminating the spatial difference of "thawing delay," including the following sub-steps:

[0114] Sub-step 3.1, controlled temperature rise triggering and surface state monitoring.

[0115] Specifically, after the low-temperature phase, the system activates the temperature control module to gradually raise the temperature from the set value of the low-temperature phase to the target regrowth temperature. The heating rate is primarily linearly controlled and limited to a few degrees Celsius per second to ensure that thermal inertia does not cause abrupt heating on the surface. At this time, the atmosphere parameters are kept constant, and surface brightness changes and temperature uniformity are monitored through in-situ optical reflection, infrared thermography, or electron diffraction. The core parameters output by surface condition monitoring include temperature distribution uniformity (in degrees Celsius) and changes in reflectance intensity (reflecting changes in surface smoothness).

[0116] Sub-step 3.2, nitrogen-rich surface formation and initial diffusion activation.

[0117] Specifically, as the temperature gradually rises to the mid-to-low temperature range (still tens of degrees Celsius away from the target temperature), the ammonia flow rate is briefly increased to 1.2 to 1.5 times the baseline, while the carrier gas flow rate is slightly reduced, creating a nitrogen-rich environment on the surface. After adsorption, nitrogen atoms form a temporarily stable saturated adsorption layer. This nitrogen layer reduces the surface binding potential of the metal atoms, promoting lateral diffusion of surface atoms. The surface mobility expression at this point is:

[0118]

[0119] In the formula, The migration rate after diffusion activation in a nitrogen-rich environment; The nitrogen enrichment promotion coefficient ranges from 0.05 to 0.15, depending on the nitrogen flow rate increase ratio. If a stable upward trend in the reflection intensity is observed (the surface becomes smoother), it indicates successful diffusion activation.

[0120] Sub-step 3.3, Heating homogenization and thermal inertia compensation.

[0121] Specifically, during the continued heating process, the power distribution in different areas of the cavity is automatically adjusted by the zoned heating control system to ensure surface temperature difference. The temperature difference remains consistently below the set threshold (typically 3°C to 5°C). If a region is detected to have a delayed or excessively rapid temperature rise, the system compensates by altering the airflow direction and local heating power in that region. At this point, the surface mobility distribution tends to become uniform, and the expression for the mobility after temperature homogenization is:

[0122]

[0123] In the formula, The migration rate after temperature homogenization; The coefficient representing the inhibition of migration rate by temperature difference ranges from 0.01 to 0.03 empirically; for every 1°C increase in temperature difference, the migration rate decreases. Times. Adjustments are made based on real-time feedback, so that... near This ensures the continuity of the overall diffusion process.

[0124] Sub-step 3.4, diffusion delay compensation and stress redistribution.

[0125] Specifically, after the target regrowth temperature is reached, the stress evolution in different regions is assessed. If the stress change rate in certain regions is found to be significantly higher than the average, it indicates a lag in migration recovery during the thawing phase, leading to local atomic re-aggregation. In these regions, the system temporarily reduces heating power and enhances airflow guidance to slow diffusion; simultaneously, it slightly increases power in lower stress regions to rebalance the stress. The correction period is typically tens of seconds to several minutes, until the stress change rate returns to stability, i.e., the spatial fluctuation of surface migration is less than 5%.

[0126] Sub-step 3.5: Confirmation and output of the activity of the regenerated layer.

[0127] Specifically, maintain steady-state observation for several minutes at the target temperature. Once the mobility fluctuation is confirmed to be within 5% and the stress fluctuation is below the set threshold, terminate the heating control and maintain constant-temperature growth. If any area shows a new decrease in reflectivity or a sudden change in surface brightness, it indicates that the local area has not been completely "thawed" and needs to return to sub-step 3.4 for rebalancing. After confirming that the entire film meets the standard, use this state as the input for step 4.

[0128] Step S140: Real-time stress field monitoring is performed on the regenerated layer after the heating treatment, and stress redistribution is performed in the stress concentration area by finely adjusting the cavity air pressure and the direction of the carrier gas flow field to obtain a thick layer sample.

[0129] In some embodiments, the gallium nitride epitaxial stability enhancement method combined with surface mobility calibration provided by the present invention specifically includes the following steps in step S140:

[0130] Step S141: Read the real-time stress data of different radial and angular directions of the epitaxial layer under the isothermal growth state through online optical interference or wafer bending monitoring, and divide the real-time stress data into regions and stress levels to generate a stress distribution table.

[0131] Step S142: Adjust the carrier gas or nitrogen guiding component in the cavity marked as a high-stress area in the stress distribution table to control the surface temperature of the high-stress area to be lower than the surface temperature of other areas, reduce the effective incident flux of the high-stress area, and increase the effective incident flux of the low-stress area.

[0132] In a specific embodiment, the gallium nitride epitaxial stability enhancement method combined with surface mobility calibration provided by the present invention includes step 4, interlayer stress balancing and diffusion path suppression. Real-time stress field monitoring is performed on the regenerated layer output from step 3. When a high-stress region is detected in a localized area, a slight cold end of airflow is formed in the stress concentration region by finely adjusting the cavity gas pressure and carrier gas flow direction, causing a decrease in the thermal expansion rate of that region and redistributing the interlayer stress. After balancing, the dislocation paths that might have previously diffused are suppressed under the weakened stress gradient, including the following sub-steps:

[0133] Sub-step 4.1, Initial identification and classification of stress in multiple zones.

[0134] Specifically, the stable epitaxial layer output from step 3 is grown under isothermal conditions for a short period, while simultaneously reading stress data in different radial and angular directions obtained from online optical interferometry or wafer bending monitoring, and dividing the area into several monitoring units. The interlayer equivalent stress of each unit is calculated and expressed as:

[0135]

[0136] In the formula, The equivalent stress of a certain unit is expressed in megapascals (MPa). The equivalent internal forces detected are derived from curvature conversion or stress sensing results; The growth area corresponding to this unit is obtained by dividing the grid.

[0137] Then, all the data collected at the same time They are divided into three levels: low, medium, and high, based on size, and the location, stress level, and gradient between adjacent elements are recorded in the data table.

[0138] Sub-step 4.2, stress reduction peak guided by local airflow.

[0139] Specifically, for the high-stress region marked in sub-step 4.1, the carrier gas or nitrogen guiding component within the cavity is adjusted to create a slight cold end above this region. This slight cold end causes the actual surface temperature of this region to be slightly lower than the surrounding area, thereby reducing the rate of thermal expansion of the material at that location. The equivalent stress after stress clipping is expressed as:

[0140]

[0141] In the formula, The equivalent stress after airflow clipping is expressed in units of 100%. ; This is the airflow clipping factor, ranging from 0.05 to 0.30, calibrated based on airflow rate and guide angle. If a certain area belongs to the medium stress level, it can be... Take the lower limit; if it belongs to a high stress level, then... Take the upper limit, but ensure that the surface temperature difference is still within the allowable range determined in step 3.

[0142] Sub-step 4.3, growth rate micro-allocation to eliminate stress gradient.

[0143] Specifically, even after airflow clipping, a situation may still exist where a low-stress region is adjacent to a high-stress region. If constant-rate growth continues, dislocations will still preferentially form in the high-stress region. Therefore, while maintaining the overall composition, a slight decrease is made to the effective incident flux above the high-stress region, and a slight increase is made to the incident flux above the low-stress region, resulting in a slight difference in the deposition rate per unit time, expressed as:

[0144]

[0145]

[0146] In the formula, This is the baseline growth rate for the current batch; The growth rate adjusted for high-stress regions; The growth rate adjusted for the low-stress region; and This is the growth rate adjustment coefficient, ranging from 0.02 to 0.10. The specific value is determined according to the stress gradient output from the previous sub-step.

[0147] Then, by slowing down the material growth in high-stress areas and accelerating it in low-stress areas, the stress difference from the previous moment is partially offset in the next moment. At this point, it is necessary to check whether the surface mobility is still within the stable range output in step 3. If the mobility shows a downward trend due to changes in the growth rate, a small amount of compensation can be made to the atmosphere in that area simultaneously.

[0148] Sub-step 4.4: Diffusion path determination and final stress balance output.

[0149] Specifically, after completing one round of "airflow peak reduction + growth rate fine-tuning", a full-wafer stress scan is performed on the epitaxial wafer again. If the stress values ​​of all monitoring units are... If the stress values ​​obtained from the iteration do not exceed the corresponding set stress threshold and the mobility fluctuation does not exceed the allowable fluctuation value, it indicates that the stress in the current layer has been balanced. At the same time, because the high-stress area is weakened and the deposition above it is suppressed, the path that could have become a dislocation diffusion path is "cut off" and does not have the conditions to continue to penetrate upwards. If any element is still higher than the set stress threshold, return to sub-step 4.2 to perform local airflow peak shaving and growth rate redistribution in that area again until the entire sheet meets the balance requirements.

[0150] Step S150: The migration rate of the thick sample is closed-loop calibrated by comparing the migration rate of the preset model with the measured surface diffusion response time, so as to correct the pulse heat flow cycle and the generation sequence of the nitrogen-rich surface.

[0151] In some embodiments, the gallium nitride epitaxial stability enhancement method combined with surface mobility calibration provided by the present invention specifically includes the following steps in step S150:

[0152] Step S151: Based on the migration rate data of each stage of the thick sample, calculate the migration rate deviation between the measured migration rate value and the target migration rate value of the corresponding stage, so as to determine whether the migration rate deviation exceeds the allowable deviation threshold.

[0153] Step S152: When the migration deviation exceeds the allowable deviation threshold, a correction amount is generated based on the migration deviation in the corresponding stage, and the heating rate, pulse heat flow and ammonia or nitrogen carrier gas flow rate are adjusted according to the correction amount to control the migration deviation in the corresponding stage to meet the allowable deviation threshold.

[0154] In a specific embodiment, the gallium nitride epitaxial stability enhancement method combined with surface mobility calibration provided by the present invention includes step 5, mobility closed-loop calibration and stability solidification. After completing a full high-temperature-low-temperature regrowth cycle, the thick-layer sample output in step 4 undergoes mobility closed-loop calibration. By comparing the preset model mobility with the measured surface diffusion response time, the pulsed thermal flow cycle and the nitrogen-rich surface formation sequence are corrected, so that the low-temperature segment of the next cycle can automatically avoid the mobility freezing interval. After several cycles, the system forms an adaptive correction mechanism, enabling the mobility curve to converge stably in the low-temperature segment, including the following sub-steps:

[0155] Sub-step 5.1: Growth cycle results collection and target comparison.

[0156] Specifically, the qualified epitaxial layer confirmed in step 4 is considered the end point of a complete high-temperature-low-temperature alternating growth cycle, and the mobility data recorded for each stage within this cycle are retrieved. The difference between the measured mobility values ​​collected on the same time axis and the target values ​​corresponding to that time period is calculated, expressed as:

[0157]

[0158] In the formula, This represents the migration rate deviation at a certain point in time, with units consistent with the migration rate. This represents the actual surface migration rate monitored at that time point; The target migration rate that should be achieved at this point in time (e.g., the corresponding low-temperature range). Nitrogen-rich activation segment corresponding to ).

[0159] Then, the mobility deviations at all time points are packaged together with the temperature, ammonia flow rate, carrier gas ratio, and local airflow guidance status at the same time to form a closed-loop comparison dataset for this cycle. If the mobility deviation is less than the set allowable deviation threshold at most time points, it indicates that the control for this cycle is effective; if there are consecutive periods of large mobility deviations, parameter adjustments are required in subsequent cycles.

[0160] Sub-step 5.2: Abnormal segment identification and parameter backtracking.

[0161] Specifically, the closed-loop comparison dataset was segmented according to process stages, separating the low-temperature holding stage, reheating stage, and stress equilibrium stage, and examining the mobility deviation for each time period. If a persistent negative mobility deviation occurred in the low-temperature holding stage, it indicated insufficient pulsed heat flow power or atmosphere coordination time; if a positive mobility deviation occurred in the reheating stage accompanied by surface brightness fluctuations, it indicated an excessively rapid heating rate or insufficient nitrogen enrichment formation time; if a decrease in mobility occurred in the stress equilibrium stage, it indicated a slight deficiency in surface energy due to local airflow clipping. For each type of problem, the parameters used at the time were reviewed, resulting in the following correction values:

[0162]

[0163]

[0164] In the formula, This refers to the pulsed heat flow power originally used in this abnormal section. The corrected power; The original pulse or atmosphere duration in this abnormal section. This is the corrected time; , This is a correction factor, typically ranging from 0.05 to 0.20. The deviation range is determined.

[0165] If the deviation occurs during the heating stage, the heating gradient is used as the backtracking object, following the same principles as above, and a list of "abnormal section - correction object - correction coefficient" is finally output.

[0166] Sub-step 5.3: Calibration model update and next cycle process writing.

[0167] Specifically, the correction values ​​obtained in sub-step 5.2 are matched one by one to the original control model timing sequence. For example, the pulse power segment for cryogenic holding is replaced with... The duration of atmosphere coordination is replaced with Then, the starting point for further heating is shifted slightly forward or backward by a few seconds, reducing the threshold for local airflow clipping to a slightly lower stress level. The updated model retains the original five-segment structure (low temperature target, pulse hold, heating activation, thermal field homogenization, stress balance), but the amplitude and duration of each segment have been refined based on the actual performance of the previous cycle. If any item in the corrected model exceeds the maximum energy or flow rate allowed by the equipment, the correction amount must be brought back to the equipment's allowable range before being written.

[0168] Sub-step 5.4, stability curing and process archiving.

[0169] Specifically, after updating the model, a final inspection is performed on the epitaxial layers of this batch, including surface roughness, dislocation density distribution, and residual stress distribution. If these indicators all fall within the acceptable range defined by the product or R&D, the new model generated in sub-step 5.3 is fixed as the default model for the current equipment and archived together with the interface state parameters obtained in step 1, the cryogenic holding parameters in step 2, the temperature rise compensation parameters in step 3, and the stress balance parameters in step 4 to form a reproducible process package. If areas of "dislocation upward movement after thawing" are still found, the time periods of these areas are added to the list of abnormal sections, and a small-scale backtracking is performed in sub-step 5.2 until the final inspection is passed.

[0170] The following describes the gallium nitride epitaxial stability enhancement device combined with surface mobility calibration provided by the present invention. The gallium nitride epitaxial stability enhancement device combined with surface mobility calibration described below can be referred to in correspondence with the gallium nitride epitaxial stability enhancement method combined with surface mobility calibration described above.

[0171] like Figure 2 As shown, in one embodiment, a gallium nitride epitaxial stability enhancement device incorporating surface mobility calibration includes a substrate interface generation module, a regenerated layer acquisition module, a nitrogen-rich surface generation module, a stress redistribution module, and a mobility calibration module.

[0172] The substrate interface generation module is used to neutralize polarization residues on the surface of the growth layer through inert gas impact and plasma activation in order to generate a regrowth substrate interface.

[0173] The regenerated layer acquisition module is used to obtain a regenerated layer of a set thickness by using energy pulses to control the surface of the growth layer to alternate between heating and cooling based on the regenerated substrate interface.

[0174] The nitrogen-rich surface generation module is used to heat the regenerated layer at a set rate and adjust the ratio of ammonia to metal source flow through a surface diffusion activation mechanism to generate a nitrogen-rich surface in the early stage of heating.

[0175] The stress redistribution module is used to monitor the stress field of the regenerated layer after heating treatment in real time, and redistribute the stress in the stress concentration area by finely adjusting the cavity air pressure and the direction of the carrier gas flow field to obtain a thick layer sample.

[0176] The mobility calibration module is used to perform closed-loop mobility calibration on thick samples by comparing the mobility of a preset model with the measured surface diffusion response time, so as to correct the pulsed heat flow cycle and the generation sequence of nitrogen-rich surfaces.

[0177] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are all optional embodiments, and the actions and modules involved are not necessarily essential to this application.

[0178] It should be further noted that although the steps in the flowchart are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowchart may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the sub-steps or stages of other steps.

[0179] It should be understood that the above-described device embodiments are merely illustrative, and the device of this application can also be implemented in other ways. For example, the division of units / modules in the above embodiments is only a logical functional division, and there may be other division methods in actual implementation. For example, multiple units, modules, or components may be combined, or integrated into another system, or some features may be ignored or not executed.

[0180] Furthermore, unless otherwise specified, the functional units / modules in the various embodiments of this application can be integrated into one unit / module, or each unit / module can exist physically separately, or two or more units / modules can be integrated together. The integrated units / modules described above can be implemented in hardware or in the form of software program modules.

[0181] The technical solutions of the present invention are not limited to the specific embodiments described above. Any technical modifications made in accordance with the technical solutions of the present invention fall within the protection scope of the present invention.

Claims

1. A method for enhancing the stability of gallium nitride epitaxial growth by combining surface mobility calibration, characterized in that, The method includes: The surface of the growth layer is neutralized by inert gas impact and plasma activation to generate a regenerated substrate interface. Based on the regenerated substrate interface, the surface of the growth layer is controlled by energy pulses to alternate between heating and cooling in order to obtain a regenerated layer of a set thickness. The regenerated layer is heated at a set rate, and the ratio of ammonia to metal source flow rate is adjusted through a surface diffusion activation mechanism to generate a nitrogen-rich surface in the early stage of heating. Real-time stress field monitoring was performed on the regenerated layer after heating treatment, and stress redistribution was carried out in the stress concentration area by fine-tuning the cavity gas pressure and carrier gas flow direction to obtain a thick layer sample. The migration rate of the thick sample was closed-loop calibrated by comparing the migration rate of the preset model with the measured surface diffusion response time, so as to correct the pulse heat flow cycle and the generation sequence of the nitrogen-rich surface. The step of heating the regenerated layer at a set rate and adjusting the ratio of ammonia to metal source flow rate through a surface diffusion activation mechanism to generate a nitrogen-rich surface in the initial stage of heating includes: The temperature of the regenerated layer is linearly increased to the target regeneration temperature according to the set heating rate, and the brightness and temperature of the surface of the regenerated layer are monitored by in-situ optical reflection, infrared thermometry or electron diffraction technology to output the temperature distribution uniformity and reflection intensity change. Based on the temperature distribution uniformity and reflection intensity changes, the medium-low temperature zone of the regenerated layer surface is determined. When the surface of the regenerated layer is heated to the medium-low temperature zone, the ammonia flow rate is increased while the carrier gas flow rate is reduced to generate a nitrogen-rich surface state. During the heating process in the nitrogen-rich surface state, the heating is controlled by adjusting the power distribution in different areas of the cavity to obtain a temperature-uniformed epitaxial layer with a surface temperature difference lower than the set temperature difference threshold. When the temperature of the temperature homogenization epitaxial layer reaches the target regrowth temperature, the stress change rate in different regions within the cavity is monitored, and when the stress change rate exceeds the average value, the stress in the local region is diffused.

2. The gallium nitride epitaxial stability enhancement method combined with surface mobility calibration according to claim 1, characterized in that, The process of neutralizing polarization residues on the surface of the growth layer through inert gas impact and plasma activation to generate a regrowth substrate interface includes: After the high-temperature stage ends and before low-temperature regrowth, the epitaxial wafer is controlled at a preset transition temperature, and the total polarization charge on the surface of the growth layer is measured by an in-situ reflective ellipsoid or contact electrostatic sampling device. The initial polarization surface charge density of the growth layer is calculated, and the average roughness of the growth layer surface is monitored using in-situ surface scattering or optical roughness monitoring technology. The initial polarization surface charge density is the ratio between the total polarization charge and the effective epitaxial area of ​​the sample.

3. The gallium nitride epitaxial stability enhancement method combined with surface mobility calibration according to claim 2, characterized in that, The method of neutralizing polarization residues on the surface of the growth layer through inert gas impact and plasma activation to generate a regrowth substrate interface further includes: Based on the initial polarization surface charge density and average roughness, the inert gas is controlled to impact the cavity according to the set instantaneous gas flow rate and gas impact time, so as to peel off the adsorbed residues and loaded clusters on the surface of the growth layer. The first surface charge density of the grown layer surface after the inert gas impact is calculated based on the charge reduction coefficient of the inert gas impact and the initial polarization surface charge density, and the improved average roughness after the inert gas impact is calculated based on the roughness improvement coefficient of the inert gas impact and the average roughness of the grown layer surface.

4. The gallium nitride epitaxial stability enhancement method combined with surface mobility calibration according to claim 3, characterized in that, The method of neutralizing polarization residues on the surface of the growth layer through inert gas impact and plasma activation to generate a regrowth substrate interface further includes: Under a uniform cavity vacuum, a low-power plasma source is used to excite the surface of the growth layer with the first surface charge density and improved average roughness according to the set excitation power and plasma action time, so as to control the migration of the accumulated surface polarization charge to the surrounding area, and obtain the second surface charge density and activation average roughness of the growth layer surface after plasma activation. The epitaxial surface under the second surface charge density and activation average roughness is cooled to the set regeneration preparation temperature at a set cooling rate, and the surface of the growth layer is controlled to be in a thermodynamically stable state at the regeneration preparation temperature to generate the regeneration substrate interface. The thermodynamically stable state is a thermodynamic state in which both the second surface charge density and the average activation roughness are lower than the corresponding process set threshold.

5. The gallium nitride epitaxial stability enhancement method combined with surface mobility calibration according to claim 1, characterized in that, The method of obtaining a regenerated layer of a set thickness by using energy pulse-assisted control to keep the surface of the growth layer in an alternating state of heating and cooling based on the regenerated substrate interface includes: Based on the surface charge density and average roughness of the regenerated substrate interface, the mobility target in the low-temperature section and the pulsed heat flow reference parameters corresponding to the mobility target are determined, and the regenerated substrate interface is periodically heated according to the pulsed heat flow reference parameters to determine the instantaneous surface mobility under pulse action. The nascent low-temperature epitaxial layer that has been pulse-excited is determined based on the instantaneous surface mobility, and the ammonia or nitrogen carrier gas flow rate of the nascent low-temperature epitaxial layer is increased within a set time window and then restored to determine the effective surface mobility. Determine whether the effective surface mobility falls within the set range. If not, increase the ammonia or nitrogen carrier gas flow rate of the nascent low-temperature epitaxial layer again within the set time window and then restore it.

6. The gallium nitride epitaxial stability enhancement method combined with surface mobility calibration according to claim 1, characterized in that, The process involves real-time stress field monitoring of the regenerated layer after heating treatment, and stress redistribution in stress concentration areas by fine-tuning the cavity gas pressure and carrier gas flow direction to obtain a thick-layer sample, including: Read real-time stress data of different radial and angular directions obtained by online optical interference or wafer bending monitoring of epitaxial layer under isothermal growth conditions, and divide the real-time stress data into regions and stress levels to generate a stress distribution table; Adjust the carrier gas or nitrogen guiding components in the cavity marked as high stress areas in the stress distribution table to control the surface temperature of the high stress areas to be lower than the surface temperature of other areas, reduce the effective incident flux of the high stress areas, and increase the effective incident flux of the low stress areas.

7. The gallium nitride epitaxial stability enhancement method combined with surface mobility calibration according to claim 1, characterized in that, The step of performing closed-loop calibration of the migration rate of the thick sample by comparing the migration rate of the preset model with the measured surface diffusion response time, in order to correct the pulsed heat flow cycle and the generation sequence of the nitrogen-rich surface, includes: Based on the migration rate data of the thick sample at each stage, the migration rate deviation between the measured migration rate value and the target migration rate value at the corresponding stage is calculated to determine whether the migration rate deviation exceeds the allowable deviation threshold. When the migration rate deviation exceeds the allowable deviation threshold, a correction amount is generated based on the migration rate deviation at the corresponding stage, and the heating rate, pulse heat flow, and ammonia or nitrogen carrier gas flow rate are adjusted according to the correction amount to control the migration rate deviation at the corresponding stage to meet the allowable deviation threshold.

8. A terminal, comprising a processor and a storage medium; characterized in that: The storage medium is used to store instructions; The processor is configured to operate according to the instructions to perform the steps of the method according to any one of claims 1-7.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the program implements the steps of the method according to any one of claims 1-7.

Images