Composite strain buffer layer low temperature gallium oxide epitaxy method

Abstract

The application discloses a composite strain buffer layer low-temperature gallium oxide epitaxial method and belongs to the technical field of semiconductors. x Ga 1‑x N dislocation filtering layer; in-situ transfer to cavity B through inert gas blowing and positive pressure protection; growing a low-temperature amorphous Ga2O3 buffer layer in the cavity B; growing a beta-Ga2O3 main epitaxial layer by adopting a two-stage process of a low-temperature low-speed nucleation section and a higher-temperature main growth section; and finally performing annealing post-processing. The application eliminates NH3 / O2 cross contamination through double-cavity in-situ atmosphere isolation, effectively filters penetrating dislocations through a composite strain buffer system composed of a gradient AlGaN layer and an amorphous buffer layer, inhibits thermal stress and improves surface morphology through low-temperature two-stage growth, and realizes controllable growth of a high-quality beta-Ga2O3 epitaxial film at low temperature.

Description

Technical Field

[0001] This invention belongs to the field of semiconductor technology, and more specifically, relates to a low-temperature gallium oxide epitaxy method with a composite strain buffer layer. Background Technology

[0002] Gallium oxide (Ga2O3) is a novel wide-bandgap semiconductor material. Its stable phase, β-Ga2O3, possesses an ultra-wide bandgap of 4.8 eV and a theoretical breakdown electric field strength of 8 MV / cm. In the field of power devices, β-Ga2O3 exhibits a significantly higher Baliga quality factor than the currently mainstream GaN and SiC, a key performance indicator for semiconductor materials. Based on these superior properties, power switching devices made from β-Ga2O3 can theoretically withstand higher voltages and have lower conduction losses, making it one of the current research hotspots in the field of semiconductor materials.

[0003] However, the preparation of high-quality β-Ga2O3 thin films faces a core challenge: a significant performance mismatch between β-Ga2O3 and the substrate, making it difficult to grow high-quality crystalline thin films at low cost and high efficiency, thus hindering their industrial application. This mismatch is mainly manifested in two aspects: firstly, lattice mismatch, where β-Ga2O3 (monoclinic crystal system) and commonly used c-plane sapphire substrates exhibit lattice mismatch, resulting in a high through-dislocation density of up to 10 in the epitaxial layer. 9 ~10 10 cm -2 First, it degrades device performance; second, it causes thermal mismatch, with the thermal expansion coefficients of the two differing by about 40%. The thermal stress generated during the cooling of the existing 650~900℃ high-temperature epitaxial process can easily induce dislocations or even thin film cracking.

[0004] In addition to the aforementioned material physics-level "mismatch," existing fabrication processes also suffer from the following technical bottlenecks: 1) High-temperature metal-organic chemical vapor deposition (MOCVD) processes exacerbate thermal mismatch, leading to gallium volatilization and a growth rate of only 0.1–0.3 μm / h, increasing costs; 2) Buffer layer solutions have defects, with single low-temperature Ga2O3 buffer layers offering limited improvement, and other oxide buffer layers being incompatible with existing MOCVD processes; 3) Direct switching between NH3 / H2 and O2 atmospheres poses risks of NOx poisoning and explosion, and the incorporation of N elements into the epitaxial layer disrupts carrier concentration control.

[0005] Therefore, it is of great significance to provide a safe process compatible with existing MOCVD equipment to obtain high-quality β-Ga2O3 epitaxial films on sapphire or SiC substrates at temperatures below 600°C. Summary of the Invention

[0006] To address the aforementioned problems in existing technologies, the technical problem to be solved by this invention is to provide a low-temperature gallium oxide epitaxy method with a composite strain buffer layer, which eliminates NH3 / O2 cross-contamination by introducing a dual-cavity series in-situ atmosphere isolation system, and constructs a gradient Al x Ga 1-x A composite strain buffer system, which combines an N-dislocation filter layer with a low-temperature amorphous oxide buffer layer, reduces the main epitaxial temperature of β-Ga2O3 to a low-temperature range, while simultaneously achieving the growth of high-quality β-Ga2O3 epitaxial films with low dislocation density, low surface roughness, and controllable doping.

[0007] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows:

[0008] A low-temperature gallium oxide epitaxy method with a composite strain buffer layer employs an MOCVD device equipped with a dual-cavity tandem in-situ transfer system. The dual-cavity system includes a nitride growth cavity and an oxide growth cavity connected by an in-situ transfer channel sealed with an inert gas. The method includes the following steps:

[0009] 1) Substrate pretreatment: The semiconductor substrate is sequentially cleaned and baked at high temperature;

[0010] 2) Gradient Al x Ga 1-x N-dislocation filter layer growth: In a nitride growth chamber, a low-temperature AlN nucleation layer is grown at a first temperature, followed by the sequential growth of multiple Al layers with an Al composition decreasing in a gradient from bottom to top at a second temperature. x Ga 1-x N sublayer;

[0011] 3) In-situ atmosphere isolation transfer: Turn off the NH3 and metal-organic sources in the nitride growth chamber, purge with inert gas until the NH3 concentration is below 1 ppm, and transfer the sample to the oxide growth chamber through the in-situ transfer channel under the protection of inert gas positive pressure, without contact with the atmosphere throughout the process;

[0012] 4) Low-temperature amorphous Ga2O3 buffer layer growth: Low-temperature amorphous Ga2O3 buffer layer is grown in the oxide growth chamber under low pressure conditions at 430~470℃.

[0013] 5) β-Ga2O3 main epitaxial layer growth: Two-stage growth is adopted in the oxide growth cavity, including an initial nucleation stage at 490~510℃ and a subsequent main growth stage at 520~565℃, wherein the main growth stage is isothermal growth or contains multiple sub-segments with increasing temperature.

[0014] 6) Post-annealing treatment: Anneal in an oxygen-containing atmosphere and control the cooling rate.

[0015] Preferably, in step 2), the thickness of the low-temperature AlN nucleation layer is 15~40 nm, and the growth temperature is 535~565℃; multiple Al x Ga 1-x The N-sublayer comprises high-Al, medium-Al, and low-Al molecular layers, with a growth temperature of 600–640℃; gradient Al x Ga 1-x The total thickness of the N-dislocation filter layer is 100~200 nm.

[0016] Preferably, the x-value of the high-Al group molecular layer is 0.70~0.80, the x-value of the medium-Al group molecular layer is 0.40~0.50, and the x-value of the low-Al group molecular layer is 0.10~0.20.

[0017] Preferably, in step 3), the inert gas is N2, and the NH3 concentration at the outlet of the nitride growth chamber is <1ppm; the N2 is under positive pressure, and its gauge pressure is 0.03~0.08 bar; the in-situ transfer channel is purged with N2 again for no less than 5 minutes after the sample transfer is completed.

[0018] Preferably, in step 4), the pressure is 3-8 Torr, the gallium source is trimethylgallium (TMGa), the oxygen source is O2, the O2 / TMGa molar ratio is 50-80, and the thickness of the low-temperature amorphous Ga2O3 buffer layer is 15-25 nm. The buffer layer is amorphous or contains metastable nanocrystalline regions, and its grazing incidence X-ray diffraction (XRD) pattern shows a diffuse scattering broadening peak in the range of 2θ = 30°-33°.

[0019] Preferably, in step 5), the initial nucleation segment conditions are: temperature 490~510℃, pressure 3~8 Torr, O2 / TMGa molar ratio 100~150, growth rate 0.05~0.15 μm / h, and thickness 50~100 nm.

[0020] Preferably, in step 5), the conditions for the subsequent main growth stage are: temperature 520~565℃, pressure 8~15 Torr, O2 / TMGa molar ratio 100~150, growth rate 0.2~0.35 μm / h, and thickness 0.5~3 μm.

[0021] Preferably, in step 5), n-type doping is performed by introducing SiH4 into the oxide growth cavity, with a SiH4 / TMGa molar ratio of 0.003~0.008, and the carrier concentration is controlled at 5×10⁻⁶. 16 ~5×10 18 cm -3 .

[0022] Preferably, in step 1), the semiconductor substrate is c-plane sapphire or 4H-SiC.

[0023] Preferably, when the substrate is 4H-SiC, the high-temperature baking temperature in step 1) is 1100℃ and the time is 15 min, and the AlN nucleation layer thickness in step 2) is increased to 30~40 nm, with a gradient Al... x Ga 1-x The thickness of each N sublayer is increased by 5~15 nm.

[0024] Beneficial effects: Compared with the prior art, the present invention has the following advantages:

[0025] 1) This invention employs a dual-cavity series in-situ transfer system, isolating the nitride growth chamber and the oxide growth chamber through an in-situ transfer channel sealed with inert gas. After the gradient Al growth is completed... x Ga 1-x After layer N, the sample is purged with inert gas until the NH3 concentration is below the safe threshold, and then transferred to chamber B under positive pressure protection of inert gas, without contact with the atmosphere throughout the process; this fundamentally eliminates direct contact between NH3 and O2, and avoids NO. x The generation of toxic gases and the risk of H2 / O2 mixed explosion are eliminated, while the possibility of trace N elements being incorporated into the β-Ga2O3 epitaxial layer is eliminated, thereby ensuring the stoichiometric purity and carrier concentration control accuracy of the epitaxial layer.

[0026] 2) This invention controls the main epitaxial temperature of β-Ga2O3 in the low-temperature range of 500~580℃, which is significantly lower than that of traditional MOCVD process; low-temperature growth effectively suppresses the generation of gallium volatile by-products, which is conducive to improving growth rate and raw material utilization; at the same time, it significantly reduces the thermal mismatch effect between epitaxial layer and substrate (sapphire or SiC), reduces the dislocation density and wafer warpage induced by thermal stress during cooling, and provides a process basis for realizing epitaxial growth of large-size and thin substrates;

[0027] 3) This invention employs a two-stage process in the growth of the β-Ga2O3 main epitaxial layer: first, nucleation stage growth is performed at a lower temperature and lower growth rate to promote two-dimensional layered nucleation; then, the temperature is increased to the main growth stage for thick film growth at a higher growth rate. The low-speed nucleation stage provides a flat, continuous, and oriented nucleation template for subsequent main growth, suppressing three-dimensional island growth. The increased temperature in the main growth stage enhances the surface atomic mobility, which is beneficial for lateral grain growth and the formation of surface step flow. The synergistic effect of the two stages enables the surface roughness of the epitaxial layer to be controlled at the nanometer level, meeting the stringent requirements of heterojunction devices for interface flatness. Attached Figure Description

[0028] Figure 1This is a process flow diagram of the low-temperature gallium oxide epitaxial method for the composite strain buffer layer of the present invention;

[0029] Figure 2 This is a schematic diagram of the cross-sectional structure of the epitaxial layer in Example 1;

[0030] Figure 3 This is a schematic diagram of the cross-sectional structure of the epitaxial layer in Example 2;

[0031] Figure 4 This is a schematic diagram of the cross-sectional structure of the epitaxial layer in Example 3;

[0032] Figure 5 This is a schematic diagram of the cross-sectional structure of the epitaxial layer in Example 4;

[0033] Figure 6 This is a schematic diagram of the cross-sectional structure of the epitaxial layer in Example 5. Detailed Implementation

[0034] To make the objectives, technical solutions, and advantages of this invention clearer, the invention is further described below with reference to specific embodiments. Unless otherwise specified, the technical means used in the following embodiments are all conventional means well known to those skilled in the art. Where specific conditions are not specified in the embodiments, conventional conditions or conditions recommended by the manufacturer are followed. Where the manufacturers of reagents or instruments are not specified, they are all conventional products that can be purchased commercially.

[0035] The MOCVD equipment used in the following embodiments is a customized device configured with a dual-chamber in-situ transfer system: Chamber A is a Group III nitride MOCVD chamber (rotating base, 2-inch substrate × 6 wafers), and Chamber B is a dedicated oxide growth chamber resistant to O2 corrosion (quartz inner wall, rotating base). The two chambers are connected by a magnetic wafer transfer mechanism and an N2 sealed gate valve. The outlet of Chamber B is equipped with an alkaline absorption + high-temperature oxidation exhaust gas treatment device. Unless otherwise specified, the sapphire substrates are all c-plane (0001), with a 0.2° off-cut angle, a diameter of 2 inches, and a thickness of 430 μm.

[0036] This application provides a low-temperature gallium oxide epitaxy method using a composite strain buffer layer, such as... Figure 1 As shown, the method includes the following steps:

[0037] 1) Substrate pretreatment: The semiconductor substrate is sequentially cleaned and baked at high temperature;

[0038] 2) Gradient Al x Ga 1-x N-dislocation filter layer growth: In a nitride growth chamber, a low-temperature AlN nucleation layer is grown at a first temperature, followed by the sequential growth of multiple Al layers with an Al composition decreasing in a gradient from bottom to top at a second temperature. x Ga 1-x N sub-layers, multiple Al x Ga1-x The N sublayer includes high-Al molecular layers, medium-Al molecular layers, and low-Al molecular layers;

[0039] 3) In-situ atmosphere isolation transfer: Turn off the NH3 and metal-organic sources in the nitride growth chamber, purge with inert gas until the NH3 concentration is below 1 ppm, and transfer the sample to the oxide growth chamber through the in-situ transfer channel under the protection of inert gas positive pressure, without contact with the atmosphere throughout the process;

[0040] 4) Low-temperature amorphous Ga2O3 buffer layer growth: Low-temperature amorphous Ga2O3 buffer layer is grown in the oxide growth chamber under low pressure conditions at 430~470℃.

[0041] 5) β-Ga2O3 main epitaxial layer growth: Two-stage growth is adopted in the oxide growth cavity, including an initial nucleation stage at 490~510℃ and a subsequent main growth stage at 520~565℃, wherein the main growth stage is isothermal growth or contains multiple sub-segments with increasing temperature.

[0042] 6) Post-annealing treatment: Anneal in an oxygen-containing atmosphere and control the cooling rate to obtain β-Ga2O3 epitaxial thin film.

[0043] Example 1

[0044] This embodiment describes a standard process using undoped β-Ga₂O₃ epitaxy on a sapphire substrate to obtain a β-Ga₂O₃ epitaxial film with the structure shown below. Figure 2 As shown, the structure includes: a β-Ga2O3 main epitaxial layer 1, a β-Ga2O3 low-speed nucleation segment 2, an amorphous Ga2O3 buffer layer 3, a low Al sublayer 4, a medium Al sublayer 5, a high Al sublayer 6, a low-temperature AlN nucleation layer 7, and a sapphire substrate 8.

[0045] The method includes the following steps:

[0046] 1. Substrate pretreatment

[0047] The sapphire substrate was ultrasonically cleaned with acetone / ethanol for 5 min each, treated with piranha solution (H2SO4:H2O2=3:1, 70℃, 10 min), rinsed with deionized water, dried with N2, placed in cavity A, H2 (10 slm), and baked at 1050℃ for 10 min. In-situ laser reflectivity monitoring showed that the surface tended to be atomically flat and had the rudimentary characteristics of step flow morphology.

[0048] 2. Gradient Al x Ga 1-x N-dislocation filter layer (cavity A)

[0049] a) Low-temperature AlN nucleation layer: 550℃, 60 Torr, trimethylaluminum (TMAl) 25 μmol / min, NH3 2.0 slm, H2:N2=1:1, growth for 90 s, thickness 20 nm, in amorphous / polycrystalline transition state;

[0050] b) High Al sublayer (x≈0.75): 620℃, 110 Torr, TMAl 75 / TMGa 25 μmol / min, NH3 1800sccm, growth for 6 min, thickness 40 nm;

[0051] c) Al sublayer (x≈0.45): TMAl 40 / TMGa 50 μmol / min, other parameters unchanged, growth for 7 min, thickness 50 nm;

[0052] d) Low Al sublayer (x≈0.15): TMAl 13 / TMGa 77 μmol / min, growth for 5 min, thickness 40 nm.

[0053] The total thickness of the gradient layer is 130 nm. EDS line scan confirmed the Al composition gradient. High-resolution X-ray diffraction (HRXRD) showed broadened sub-peaks corresponding to different sub-layers near the GaN (0002) peak.

[0054] 3. In-situ atmosphere isolation and transfer

[0055] Turn off the NH3 and organometallic sources, purge with N2 (5 slm) for 10 min, and confirm the NH3 concentration at the outlet of chamber A with an online mass spectrometer (<1 ppm). Under N2 positive pressure (0.05 bar above atmospheric pressure), transfer the sample to chamber B through the in-situ transfer channel without contact with the atmosphere for 3 min. Then purge the transfer channel with N2 again for 5 min.

[0056] 4. Low-temperature amorphous Ga2O3 buffer layer (cavity B)

[0057] The sample was stabilized at 450℃, pressure 5 Torr, O2 (30 sccm) + TMGa (15 μmol / min), O2 / TMGa molar ratio 60, for 12 min, with a thickness of 20 nm. Grazing incidence XRD (incident angle 0.5°) confirmed that it was mainly amorphous, exhibiting diffuse broadening peaks in the range of 2θ=30°~33°, without sharp β-Ga2O3 diffraction peaks.

[0058] 5. β-Ga2O3 master epitaxial layer (cavity B, two-segment)

[0059] Low-speed nucleation stage: 500℃, 5 Torr, TMGa 20 μmol / min, O2 60 sccm, growth for 60 min, thickness 65 nm, selected area electron diffraction (SAED) confirmed (-201) oriented β-Ga2O3 nucleation; Main body stage: heating to 550℃ at 5℃ / min, TMGa 50 / O2 150 sccm, 10 Torr, growth for 200 min, in-situ reflectance calculation rate ≈ 0.28 μm / h, main body thickness 0.93 μm, total thickness 1.0 μm.

[0060] 6. Annealing treatment

[0061] An O2 / N2 mixed atmosphere (O2 volume fraction 10%, total flow rate 1 slm), annealed at 500℃ for 30 min, then cooled to below 100℃ at a rate of ≤2℃ / min.

[0062] The performance test results are as follows:

[0063] XRD results showed that the full width at half maximum (FWHM) of the X-ray rocking curve in the (-201) plane of β-Ga2O3 was 248 arcsec, and the FWHM in the (020) plane was 265 arcsec. Atomic force microscopy (AFM) (5×5 μm2) results showed that Ra=0.48 nm, Rq=0.61 nm, and the rudimentary step flow was visible, with some areas containing nanohillocks (a characteristic morphology of low-temperature epitaxy). Transmission electron microscopy (TEM) results showed that the penetration dislocation density was 6×10⁻⁶. 8 cm -2 The interfaces between layers are clear, and a large number of dislocations terminate at gradient Al. x Ga1₋ x N layer. Secondary ion mass spectrometry (SIMS) results showed that the N concentration in the epitaxial layer was 3 × 10⁻⁶. 16 cm -3 The O vacancy concentration is 7 × 10⁻⁶. 16 cm -3 Hall effect measurements showed a background electron concentration of 8 × 10⁻⁶. 16 cm -3 μ≈38 cm 2 / (V·s), resistivity 2.1 Ω·cm. PL (room temperature) results show: band-edge emission 4.78 eV, FWHM≈72 meV; 490 nm O vacancy peak intensity decreased to 40% after annealing. Etching pit density (EPD, hot phosphoric acid corrosion) results show: 5.8 × 10⁻⁶ / (V·s). 8 cm -2 This is consistent with the TEM results.

[0064] Example 2

[0065] This embodiment uses a Si-doped β-Ga2O3 epitaxial growth method, with the process steps being exactly the same as in Example 1. The only difference is the additional introduction of SiH4 (N2 diluted to 100 ppm, flow rate 50 sccm) in the main growth stage of step 5. The experiment was conducted in three subgroups:

[0066] Subgroup 2-A: SiH4 / TMGa molar ratio ≈ 0.003, target carrier concentration is 10. 17 cm -3 ;

[0067] Subgroup 2-B: SiH4 / TMGa molar ratio ≈ 0.005, target carrier concentration is 10. 18 cm -3 ;

[0068] Subgroup 2-C: SiH4 / TMGa molar ratio ≈ 0.008, target carrier concentration 5 × 10⁻⁶ 18 cm -3 .

[0069] The structure of the β-Ga2O3 epitaxial thin film is as follows: Figure 3 As shown, the structure includes: a Si-doped β-Ga2O3 main epitaxial layer 1, a β-Ga2O3 low-speed nucleation section 2, an amorphous Ga2O3 buffer layer 3, a low Al sublayer 4, a medium Al sublayer 5, a high Al sublayer 6, a low-temperature AlN nucleation layer 7, a sapphire substrate 8, and a SiH4 introduction section 9.

[0070] The performance test results are as follows:

[0071] Subgroup 2-A: (-201) plane XRC-FWHM = 252 arcsec; AFM Ra = 0.51 nm; SIMS mean Si concentration = 8.3 × 10⁻⁶ 16 cm -3 Longitudinal fluctuation <15%; Hall effect test n=7.6×10 16 cm -3 μ=43cm 2 / (V·s), ρ≈1.9 Ω·cm.

[0072] Subgroup 2-B: (-201) plane XRC-FWHM = 261 arcsec; AFM Ra = 0.53 nm; SIMS mean Si concentration = 1.1 × 10⁻⁶ 18 cm -3 Fluctuation <18%; Hall effect test n=9.4×10 17 cm -3 μ=41 cm 2 / (V·s), ρ≈0.16 Ω·cm.

[0073] Subgroup 2-C: (-201) plane XRC-FWHM = 269 arcsec; AFM Ra = 0.55 nm; SIMS mean Si concentration = 4.8 × 10⁻⁶ 18 cm -3 Fluctuation <22%; Hall effect test n=4.1×10 18 cm -3 μ=35 cm 2 / (V·s), ρ≈0.043 Ω·cm.

[0074] The above results show that as the Si doping concentration increases, XRC-FWHM and AFM Ra only increase slightly, the decrease in mobility is consistent with the scattering law of ionized impurities, and the SIMS has good longitudinal concentration uniformity, proving that the scheme of the present invention has excellent doping engineering compatibility.

[0075] Example 3

[0076] This embodiment uses a 4H-SiC (0001) substrate instead of sapphire and employs a β-Ga2O3 epitaxial growth method on a 4H-SiC substrate, including the following steps:

[0077] Step 1: 4H-SiC (0001) substrate, off-cut angle 4°, high temperature baking at 1100℃ in H2 atmosphere for 15 min to remove the oxide layer and carbon contamination on the SiC surface, and in-situ reflectivity confirmation that the Si surface reconstruction is complete.

[0078] Step 2: The thickness of the low-temperature AlN nucleation layer is increased to 35 nm (550℃) to accommodate the larger SiC / AlN lattice mismatch. SiC a = 3.073 Å, AlN a = 3.112 Å, mismatch approximately 1.3%, lower than that of sapphire; gradient Al x Ga 1-x The thickness of each of the N three sublayers is increased by 10 nm, for a total thickness of 160 nm, to provide more adequate strain release. The remaining steps are exactly the same as in Example 1.

[0079] The structure of the β-Ga2O3 epitaxial thin film is as follows: Figure 4 As shown, the structure includes: a β-Ga2O3 main epitaxial layer 1, a β-Ga2O3 low-speed nucleation segment 2, an amorphous Ga2O3 buffer layer 3, a low Al sublayer 4, a medium Al sublayer 5, a high Al sublayer 6, a low-temperature AlN nucleation layer 7, and a 4H-SiC (0001) substrate 8.

[0080] The performance test results are as follows:

[0081] XRD results show that the XRC-FWHM of the (-201) plane is 231 arcsec, consistent with the physical expectation of smaller thermal mismatch in SiC / β-Ga2O3. AFM (5×5 μm) 2 The results showed that Ra = 0.43 nm, the surface morphology was flatter than that of sapphire-based Example 1, and the hillock density was lower. TEM results showed that the through-dislocation density was 4.2 × 10⁻⁶. 8 cm -2 Lower than Example 1; cross-sectional HRTEM confirmed the β-Ga2O3 epitaxial layer and the amorphous buffer layer, gradient Al x Ga 1-x The interfaces between the N layers are clear. SIMS results show that the N concentration is 3.5 × 10⁻⁶. 16 cm -3 Comparable to the sapphire-based embodiment, this demonstrates that the dual-cavity atmosphere isolation process is equally effective on different substrates. Hall effect measurements show a background electron concentration of 6.3 × 10⁻⁶. 16 cm -3 μ≈42 cm 2 / (V·s), ρ≈2.4 Ω·cm. The EPD result is 3.9×10 8 cm -2 .

[0082] Example 4

[0083] This embodiment systematically examines gradient Al. x Ga 1-x The influence of the total thickness of the N-dislocation filter layer on the epitaxial quality of β-Ga2O3 was investigated to determine the optimal buffer layer thickness range. The substrate was c-plane sapphire. Three experimental groups were conducted, with the Al composition gradient (x=0.75, 0.45, 0.15) kept constant, only varying the thickness of each sublayer.

[0084] Subgroup 4-A (thin buffer layer): The thicknesses of the high Al / medium Al / low Al sublayers are 25 / 30 / 25 nm, respectively, with a total thickness of 80 nm.

[0085] Subgroup 4-B (standard buffer layer, same as in Example 1): The thicknesses of the high Al / medium Al / low Al sublayers are 40 / 50 / 40 nm, respectively, with a total thickness of 130 nm.

[0086] Subgroup 4-C (thick buffer layer): The thicknesses of the high Al / medium Al / low Al sublayers are 60 / 80 / 60 nm, respectively, with a total thickness of 200 nm.

[0087] Except for the gradient layer thickness, the low-temperature AlN nucleation layer (20 nm), the amorphous Ga2O3 buffer layer, the two-stage growth conditions of β-Ga2O3 and the annealing conditions are the same as in Example 1. The total thickness of the β-Ga2O3 epitaxial layer is 1.0 μm.

[0088] The structure of the β-Ga2O3 epitaxial thin film is as follows: Figure 5 As shown, the structure of subgroup 4-A includes: β-Ga2O3 main epitaxial layer A-1, β-Ga2O3 low-speed nucleation segment A-2, amorphous Ga2O3 buffer layer A-3, low Al sublayer A-4, medium Al sublayer A-5, high Al sublayer A-6, low-temperature AlN nucleation layer A-7, and sapphire substrate A-8.

[0089] The structure of subgroup 4-B includes: β-Ga2O3 main epitaxial layer B-1, β-Ga2O3 low-speed nucleation segment B-2, amorphous Ga2O3 buffer layer B-3, low Al sublayer B-4, medium Al sublayer B-5, high Al sublayer B-6, low-temperature AlN nucleation layer B-7, and sapphire substrate B-8.

[0090] The structure of subgroup 4-C includes: β-Ga2O3 main epitaxial layer C-1, β-Ga2O3 low-speed nucleation segment C-2, amorphous Ga2O3 buffer layer C-3, low Al sublayer C-4, medium Al sublayer C-5, high Al sublayer C-6, low-temperature AlN nucleation layer C-7, and sapphire substrate C-8.

[0091] The performance test results are as follows:

[0092] Subgroup 4-A (total thickness 80 nm): (-201) plane XRC-FWHM = 362 arcsec; AFM Ra = 0.72 nm; EPD ≈ 1.5 × 10⁻⁶ 9 cm -2 Hall effect μ = 29cm 2 / (V·s).

[0093] Subgroup 4-B (total thickness 130 nm, reference): (-201) plane XRC-FWHM = 248 arcsec; AFM Ra = 0.48 nm; EPD ≈ 5.8 × 10⁻⁶ 8 cm -2 Hall effect μ=38cm 2 / (V·s).

[0094] Subgroup 4-C (total thickness 200 nm): (-201) plane XRC-FWHM = 239 arcsec; AFM Ra = 0.44 nm; EPD ≈ 4.7 × 10⁻⁶ 8 cm -2 Hall effect μ = 41cm 2 / (V·s).

[0095] The results show that as the total thickness of the buffer layer increases from 80 nm to 200 nm, both XRC-FWHM and EPD decrease. However, the improvement from 130 nm to 200 nm is much smaller than that from 80 nm to 130 nm, and the 200 nm thick buffer layer increases wafer warpage by 18 μm compared to 4-B. Taking all factors into consideration, this invention uses gradient Al... x Ga 1-x The preferred total thickness of the N-dislocation filter layer is determined to be 115–145 nm.

[0096] Example 5

[0097] This embodiment verifies the effect of the three-step temperature increment sub-segment (500℃, 520℃, 535℃, 550℃) strategy in step 5 on improving surface morphology and crystal quality.

[0098] In step 5, the low-speed nucleation stage was as follows: 500℃, TMGa 20 μmol / min, O2 60 sccm, 5 Torr, 60 min, thickness 65 nm. The first sub-segment of the main growth stage was: 520℃, TMGa 35 μmol / min, O2 105 sccm, 8 Torr, 60 min, thickness 0.20 μm (rate ≈ 0.20 μm / h). The second sub-segment of the main growth stage was: 535℃, TMGa 45 μmol / min, O2 135 sccm, 9 Torr, 60 min, thickness 0.22 μm (rate ≈ 0.22 μm / h). The third sub-segment of the main growth stage was: 550℃, TMGa 50 μmol / min, O2 150 sccm, 10 Torr, 120 min, thickness 0.48 μm (rate ≈ 0.24 μm / h). The remaining steps were identical to those in Example 1. The total thickness of β-Ga2O3 is 0.97 μm (including the nucleation segment).

[0099] The structure of the β-Ga2O3 epitaxial thin film is as follows: Figure 6 As shown, the structure includes: a β-Ga2O3 main epitaxial layer, a β-Ga2O3 low-speed nucleation segment 2, an amorphous Ga2O3 buffer layer 3, a low Al sublayer 4, a medium Al sublayer 5, a high Al sublayer 6, a low-temperature AlN nucleation layer 7, and a sapphire substrate 8, wherein the β-Ga2O3 main epitaxial layer includes a first sub-segment 1-1, a second sub-segment 1-2, and a third sub-segment 1-3.

[0100] The performance test results are as follows:

[0101] XRD results show that the XRC-FWHM of the (-201) plane is 232 arcsec. AFM (5×5 μm) 2The results showed that Ra = 0.36 nm, Rq = 0.45 nm, the surface hillock density was significantly reduced, and the clarity of the step flow morphology was improved. TEM results showed that the penetrating dislocation density was 5.1 × 10⁻⁶. 8 cm -2 Cross-sectional HRTEM showed no obvious structural defects at the sublayer interfaces of the three main growth segments. Hall effect measurements showed a background electron concentration of 7.2 × 10⁻⁶. 16 cm -3 μ≈42 cm 2 / (V·s), resistivity 2.1 Ω·cm. EPD results show: 4.8×10 8 cm -2 This is the lowest value found in the sapphire-based embodiments. The multi-step temperature-incremental growth strategy effectively improves the surface morphology and crystal quality without significantly increasing the process time, verifying the effectiveness of the preferred embodiment of the present invention.

[0102] Comparative Example 1

[0103] This comparative example replicates the epitaxial process of Example 1 by removing the dual-cavity in-situ atmosphere isolation system and using a single-cavity direct atmosphere switching method to quantitatively evaluate the impact of NH3 / O2 cross-contamination on the quality of the epitaxial layer.

[0104] The only difference from Example 1 is step 3: the in-situ N2 positive pressure wafer transfer is cancelled, and instead, NH3 is directly shut off in cavity A, followed by purging with O2 / N2 (1:9 mixture) for 20 min. Then, in the same reaction cavity, the temperature is raised to 450°C and O2 and TMGa are introduced to grow an amorphous buffer layer, and then the two-stage growth of β-Ga2O3 continues. That is, all epitaxial steps are completed in cavity A (cavity A is modified to be compatible with both NH3 and O2 systems). The remaining process parameters are completely consistent with those of Example 1.

[0105] The performance test results are as follows:

[0106] XRD results show that the XRC-FWHM of the (-201) plane is 379 arcsec. AFM (5×5 μm) 2 The results showed that Ra = 1.08 nm, with obvious island-like aggregates on the surface and a significantly increased hillock density. TEM results showed a penetrating dislocation density of 2.3 × 10⁻⁸ nm. 9 cm -2 Cross-sectional TEM revealed a 3–5 nm wide blurred transition region near the interface between the β-Ga₂O₃ layer and the amorphous buffer layer. EDS surface analysis showed that the N signal in this region was higher than that on the background, confirming the presence of a GaON miscible region. SIMS results showed that the N concentration in the β-Ga₂O₃ layer was 8.7 × 10⁻⁶. 17 cm -3O vacancy concentration 2.1 × 10⁻⁶ 17 cm -3 Hall effect measurements showed a background electron concentration of 6.2 × 10⁻⁶. 17 cm -3 μ≈18 cm 2 / (V·s). X-ray photoelectron spectroscopy (XPS) results show that the N / Ga atomic ratio near the surface of the β-Ga2O3 epitaxial layer (sputtered to a depth of 50 nm) is 0.032.

[0107] Comparative Example 2

[0108] This comparative example completely removes gradient Al. x Ga 1-x The N-dislocation filter layer retains the dual-cavity atmosphere isolation and amorphous buffer layer. The difference from Example 1 is that step 2 (gradient Al) is skipped. x Ga 1-x (N-layer growth), Step 3, in-situ transfer, is performed immediately after the substrate has only undergone high-temperature baking (skipping AlN and AlGaN growth in cavity A), directly transferring the bare sapphire substrate in-situ to cavity B. All other process parameters are identical to those in Example 1. Dual-cavity atmosphere isolation conditions, N2 purging, and N2 positive pressure transfer are also performed to ensure atmosphere purity.

[0109] The performance test results are as follows:

[0110] XRD results show that the XRC-FWHM of the (-201) plane is 447 arcsec. AFM (5×5 μm) 2 Ra=0.89 nm, poor surface uniformity, obvious island nucleation characteristics, and lack of clear step flow morphology. TEM: Penetrating dislocation density 3.6×10⁻⁶ 9 cm -2 TEM showed that dislocations originated extensively at the sapphire / amorphous Ga2O3 buffer layer interface and extended upwards. SIMS results showed an N concentration of 2.8 × 10⁻⁶. 16 cm -3 The dual-cavity isolation was effective, and nitrogen contamination was controlled, proving that nitrogen contamination and dislocation density are two independent variables. Hall effect measurements showed a background electron concentration of 1.5 × 10⁻⁶. 17 cm -3 μ≈25 cm 2 / (V·s). EPD results showed: 3.2×10 9 cm -2 .

[0111] Comparative Example 3

[0112] This comparative example increases the epitaxial temperature of β-Ga2O3 to 700℃ and uses a single-stage high-temperature continuous growth method instead of the two-stage low-temperature growth strategy of this invention, while retaining the dual-cavity atmosphere isolation and composite buffer layer system.

[0113] The only difference from Example 1 is step 5: the low-speed nucleation stage is omitted, and β-Ga₂O₃ is directly grown in a single-segment continuous manner under the conditions of 700℃, 20 Torr, TMGa 60 μmol / min, O₂ 200 sccm (O₂ / TMGa≈130), with a growth rate of 0.45 μm / h, a target thickness of 1.0 μm, and a growth time of 135 min; the annealing conditions are the same. All other process parameters are completely consistent with those of Example 1.

[0114] The performance test results are as follows:

[0115] XRD results show that the XRC-FWHM of the (-201) plane is 283 arcsec. AFM (5×5 μm) 2 The results showed that Ra = 0.75 nm, and the Ga volatilization side reaction was enhanced at high temperature, resulting in obvious pit morphology on the surface. TEM results showed that the through-dislocation density was approximately 1.1 × 10⁻⁶. 9 cm -2 Cross-sectional TEM revealed a high-defect-density transition region of approximately 5–8 nm thick near the interface between the amorphous Ga2O3 buffer layer and the β-Ga2O3 epitaxial layer. SIMS results showed an N concentration of 3.1 × 10⁻⁶. 16 cm -3 O vacancy concentration 1.8 × 10⁻⁶ 17 cm -3 Hall effect measurements showed a background electron concentration of approximately 3.4 × 10⁻⁶. 17 cm -3 μ≈30 cm 2 / (V·s). Wafer warpage results show that: 35 μm, thermal mismatch during high-temperature epitaxial cooling causes greater thermal stress warpage.

[0116] Comparative Example 4

[0117] This comparative example was used to verify the growth temperature of the low-temperature amorphous Ga2O3 buffer layer in step 4. Two groups of experiments were conducted, both based on Example 1, with only the growth temperature in step 4 changed; all other conditions remained identical to Example 1.

[0118] Subgroup 4-D1 (Temperature Too Low): The growth temperature in step 4 is set to 400℃.

[0119] Subgroup 4-D2 (Temperature Too High): The growth temperature in step 4 is set to 500℃.

[0120] The performance test results are as follows:

[0121] Subgroup 4-D1 (400℃): The grazing incidence XRD diffuse scattering peak intensity is weak, indicating poor compactness of the amorphous layer; polycrystalline diffraction peaks appear in the β-Ga2O3 epitaxial layer, XRC-FWHM on the (-201) plane = 412 arcsec; AFM Ra = 1.21 nm.

[0122] Subgroup 4-D2 (500℃): The presence of a weak β-Ga2O3 diffraction peak in grazing incidence XRD indicates partial crystallization of the amorphous layer, introducing additional dislocations; XRC-FWHM = 331 arcsec; EPD ≈ 1.9 × 10⁻⁶ 9 cm -2 .

[0123] When the temperature in step 4 is below 430℃, insufficient amorphous layer quality leads to polycrystalline formation of the epitaxial layer; when the temperature is above 470℃, partial crystallization of the amorphous layer results in a significant increase in dislocation density. High-quality β-Ga2O3 epitaxial layers can only be obtained in the temperature range of 430–470℃.

[0124] Comparative Example 5

[0125] This comparative example omits the low-speed nucleation stage in step 5 while maintaining a constant temperature of 550℃, and directly uses single-segment growth. The only difference from Example 1 is step 5: the low-speed nucleation stage is omitted, and β-Ga2O3 is directly grown continuously in a single stage at 550℃, 10 Torr, 50 μmol / min TMGa, and 150 sccm O2, with a growth rate of 0.28 μm / h, a target thickness of 1.0 μm, and a growth time of 215 min; the annealing conditions are the same, and all other process parameters are completely consistent with Example 1.

[0126] The performance test results are as follows:

[0127] XRD results show that the XRC-FWHM of the (-201) plane is 305 arcsec. AFM (5×5 μm) 2 The results showed that Ra = 0.68 nm, with obvious island-like nucleation features and unclear step flow morphology. TEM results showed that the penetration dislocation density was 1.3 × 10⁻⁶. 9 cm -2 Cross-sectional TEM revealed numerous mismatch dislocations near the β-Ga2O3 / amorphous buffer layer interface, and the dislocation density in the epitaxial layer did not decrease effectively with increasing thickness. SIMS results showed an N concentration of 3.2 × 10⁻⁶. 16 cm -3 O vacancy concentration 8.5 × 10 16 cm -3 Hall effect measurements showed a background electron concentration of 1.1 × 10⁻⁶. 17cm -3 μ≈33 cm 2 / (V·s). EPD: 1.2×10 9 cm -2 .

[0128] Comparative Example 6

[0129] This comparative example omits the AlN nucleation layer and directly grows Al with high Al content. x Ga 1-x Feasibility of the N sublayer. In step 2: the low-temperature AlN nucleation layer is omitted, and the growth temperature is reduced to 580±15℃. A high Al sublayer (x≈0.75) is grown directly on the sapphire (0001) substrate. To ensure the dislocation filtering effect, the thickness of each sublayer is uniformly increased by 15 nm (high Al / medium Al / low Al are 55 / 65 / 55 nm, and the total thickness is 175 nm). The remaining parameters and steps are exactly the same as in Example 1.

[0130] The performance test results are as follows:

[0131] XRD results show that the XRC-FWHM of the (-201) plane is 294 arcsec. AFM (5×5 μm) 2 The results showed that Ra = 0.59 nm. TEM results showed that the penetration dislocation density was 8.4 × 10⁻⁶. 8 cm -2 SIMS results showed that the nitrogen concentration was 4.2 × 10⁻⁶. 16 cm -3 It is still below 5×10 16 cm -3 The performance indicators are as follows. Hall effect test results show a background electron concentration of 1.3 × 10⁻⁶. 17 cm -3 μ≈34 cm 2 / (V·s), ρ≈1.4 Ω·cm. Wafer warpage: 22μm.

[0132] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A low-temperature gallium oxide epitaxial method with a composite strain buffer layer, characterized in that, The method employs an MOCVD device equipped with a dual-cavity tandem in-situ transfer system, wherein the dual-cavity system comprises a nitride growth chamber and an oxide growth chamber connected by an in-situ transfer channel sealed with an inert gas; the method includes the following steps: 1) Substrate pretreatment: The semiconductor substrate is sequentially cleaned and baked at high temperature; 2) Gradient Al x Ga 1-x N-dislocation filter layer growth: In a nitride growth chamber, a low-temperature AlN nucleation layer is grown at a first temperature, followed by the sequential growth of multiple Al layers with an Al composition decreasing in a gradient from bottom to top at a second temperature. x Ga 1-x N sublayer; 3) In-situ atmosphere isolation transfer: Turn off the NH3 and metal-organic sources in the nitride growth chamber, purge with inert gas until the NH3 concentration is below 1 ppm, and transfer the sample to the oxide growth chamber through the in-situ transfer channel under the protection of inert gas positive pressure, without contact with the atmosphere throughout the process; 4) Low-temperature amorphous Ga2O3 buffer layer growth: Low-temperature amorphous Ga2O3 buffer layer is grown in the oxide growth chamber under low pressure conditions at 430~470℃. 5) β-Ga2O3 main epitaxial layer growth: Two-stage growth is adopted in the oxide growth cavity, including an initial nucleation stage at 490~510℃ and a subsequent main growth stage at 520~565℃, wherein the main growth stage is isothermal growth or contains multiple sub-segments with increasing temperature. 6) Post-annealing treatment: Anneal in an oxygen-containing atmosphere and control the cooling rate.

2. The low-temperature gallium oxide epitaxy method with a composite strain buffer layer according to claim 1, characterized in that, In step 2), the thickness of the low-temperature AlN nucleation layer is 15~40 nm, and the growth temperature is 535~565℃; multiple Al x Ga 1-x The N-sublayer comprises high-Al, medium-Al, and low-Al molecular layers, with a growth temperature of 600–640℃; gradient Al x Ga 1-x The total thickness of the N-dislocation filter layer is 100~200 nm.

3. The low-temperature gallium oxide epitaxy method with a composite strain buffer layer according to claim 2, characterized in that, The x value of the high Al group molecular layer is 0.70~0.80, the x value of the medium Al group molecular layer is 0.40~0.50, and the x value of the low Al group molecular layer is 0.10~0.

20.

4. The low-temperature gallium oxide epitaxial method with a composite strain buffer layer according to claim 1, characterized in that, In step 3), the inert gas is N2, and the NH3 concentration at the outlet of the nitride growth chamber is <1 ppm; the N2 is under positive pressure, and its gauge pressure is 0.03~0.08 bar; the in-situ transfer channel is purged with N2 again for no less than 5 min after the sample transfer is completed.

5. The low-temperature gallium oxide epitaxy method with a composite strain buffer layer according to claim 1, characterized in that, In step 4), the pressure is 3~8 Torr, the gallium source is TMGa, the oxygen source is O2, the O2 / TMGa molar ratio is 50~80, and the thickness of the low-temperature amorphous Ga2O3 buffer layer is 15~25 nm.

6. The low-temperature gallium oxide epitaxy method with a composite strain buffer layer according to claim 1, characterized in that, In step 5), the initial nucleation segment conditions are: temperature 490~510℃, pressure 3~8 Torr, O2 / TMGa molar ratio 100~150, growth rate 0.05~0.15 μm / h, and thickness 50~100 nm.

7. The low-temperature gallium oxide epitaxy method with a composite strain buffer layer according to claim 1, characterized in that, In step 5), the conditions for the subsequent main growth section are: temperature 520~565℃, pressure 8~15 Torr, O2 / TMGa molar ratio 100~150, growth rate 0.2~0.35 μm / h, and thickness 0.5~3 μm.

8. The low-temperature gallium oxide epitaxy method with a composite strain buffer layer according to claim 1, characterized in that, In step 5), n-type doping is performed by introducing SiH4 into the oxide growth cavity, with a SiH4 / TMGa molar ratio of 0.003~0.008, and the carrier concentration is controlled at 5×10⁻⁶. 16 ~5×10 18 cm -3 .

9. The low-temperature gallium oxide epitaxy method with a composite strain buffer layer according to claim 1, characterized in that, In step 1), the semiconductor substrate is c-plane sapphire or 4H-SiC.

10. The low-temperature gallium oxide epitaxy method with a composite strain buffer layer according to claim 9, characterized in that, When the substrate is 4H-SiC, the high-temperature baking temperature in step 1) is 1100℃ and the time is 15 min. In step 2), the AlN nucleation layer thickness is increased to 30~40 nm, and the gradient Al... x Ga 1-x The thickness of each N sublayer is increased by 5~15 nm.

Images