Epitaxial structure and method for manufacturing the same, optical communication chip

Abstract

The application discloses an epitaxial structure and a preparation method thereof and an optical communication chip, and relates to the technical field of semiconductors.The epitaxial structure comprises a substrate, a buffer layer, an N-type semiconductor layer, a multi-quantum well active layer and a P-type semiconductor layer which are sequentially stacked on the substrate; the multi-quantum well active layer comprises periodically and alternately stacked quantum well layers and quantum barrier layers; the quantum well layers comprise a first InGaN quantum well layer, a second InGaN quantum well layer and a third InGaN quantum well layer; along the stacking direction, the In component proportion of the first InGaN quantum well layer gradually changes from x1 to x2, the In component proportion of the second InGaN quantum well layer is x2, and the In component proportion of the third InGaN quantum well layer gradually changes from x2 to x3, wherein x2>x1, x2>x3, and x1>x3. By implementing the application, the overlap rate of electron-hole wave functions can be improved under optical communication working conditions, so that the radiation recombination rate and the modulation bandwidth are improved.

Description

Technical Field

[0001] This invention relates to the field of semiconductor technology, and in particular to an epitaxial structure and its fabrication method, and an optical communication chip. Background Technology

[0002] Visible Light Communication (VLC) and underwater optical communication, as emerging high-speed wireless communication technologies, have broad application prospects in fields such as smart lighting, high-speed indoor access, and underwater sensor networks, thanks to their advantages such as unlicensed spectrum, resistance to electromagnetic interference, and the ability to handle both lighting and communication. LEDs, as the core light source in VLC systems, have modulation bandwidth that directly determines the transmission rate of the communication system. However, the modulation bandwidth of traditional GaN-based InGaN / GaN LEDs is typically only in the tens of MHz range, far from meeting the demands of Gbps-level high-speed VLC transmission.

[0003] The 3dB modulation bandwidth of an LED is mainly determined by the differential carrier lifetime, and the two satisfy an inverse relationship: f -3dB =3 / (2πτ), where τ is the differential carrier lifetime, which is affected by Shockley-Read-Hall recombination, radiative recombination, and Auger recombination. Shortening the differential carrier lifetime is the fundamental way to improve the modulation bandwidth. However, the active regions of InGaN / GaN multi-quantum wells grown in the traditional c-plane polar direction generate a strong built-in polarization electric field due to the non-centrosymmetry of the crystal and lattice mismatch, which induces the quantum confinement Stark effect (QCSE). This causes the electron and hole wave functions in the quantum well to be offset in opposite directions and spatially separated, with the electron-hole wave function overlap rate being only 17%~25%. The reduction in wave function overlap rate directly leads to a decrease in the radiative recombination rate. Under low injection conditions, SRH recombination dominates, and the differential carrier lifetime is as long as 5ns~20ns, corresponding to a modulation bandwidth limited to tens of MHz, which becomes the main bottleneck for optical communication applications.

[0004] To improve modulation bandwidth, existing technologies mainly adopt the following solutions: (1) Semi-polar or non-polar surface growth, which reduces the polarization field by changing the crystal orientation, but has problems such as poor crystal quality, poor surface morphology, and complex epitaxial process; (2) Micro-LED technology, which increases the current density by reducing the device size to shield QCSE, but requires kA / cm 2 The extremely high current density leads to efficiency drops and reliability issues; (3) reducing the thickness of the quantum well / quantum barrier, but the reduction in the thickness of the active layer leads to a decrease in optical output power; (4) resonant cavity LEDs or plasmonic LEDs utilize the Purcell effect to improve the spontaneous emission rate, but there are problems such as difficulty in DBR growth, metal Joule loss, and complex processes. All of the above solutions start from the perspective of changing the device geometry or crystal orientation, but do not fundamentally solve the problem of wave function separation caused by the polarization electric field in the quantum well.

[0005] Furthermore, the low wavefunction overlap rate also leads to a decrease in spontaneous emission rate, requiring higher drive current to obtain sufficient optical output power, further exacerbating the efficiency droop effect, limiting the dynamic range and linearity of the device, and adversely affecting higher-order modulation formats such as OFDM and QAM in optical communication. Summary of the Invention

[0006] The technical problem to be solved by the present invention is to provide an epitaxial structure and its preparation method, and an optical communication chip, which can improve the overlap rate of electron-hole wave functions under optical communication working conditions, thereby improving the radiative recombination rate and modulation bandwidth.

[0007] To address the aforementioned problems, this invention discloses an epitaxial structure comprising a substrate, and a buffer layer, an N-type semiconductor layer, a multiple quantum well active layer, and a P-type semiconductor layer sequentially stacked on the substrate. The multi-quantum-well active layer comprises periodically alternating quantum well layers and quantum barrier layers. The quantum well layers include a first InGaN quantum well layer, a second InGaN quantum well layer, and a third InGaN quantum well layer. Along the stacking direction, the In composition ratio of the first InGaN quantum well layer gradually changes from x1 to x2, the In composition ratio of the second InGaN quantum well layer is x2, and the In composition ratio of the third InGaN quantum well layer gradually changes from x2 to x3, wherein x2 > x1, x2 > x3, and x1 > x3.

[0008] As an improvement to the above technical solution, the value of x1 ranges from 0.03 to 0.15; The value of x2 ranges from 0.1 to 0.35; The value of x3 ranges from 0.02 to 0.12; x2-x1>0.03, x2-x3>0.05.

[0009] As an improvement to the above technical solution, the thickness of the first InGaN quantum well layer is L1, the thickness of the second InGaN quantum well layer is L2, and the thickness of the third InGaN quantum well layer is L3, wherein L2 < L1, L2 < L3, and L2 < 10 nm.

[0010] As an improvement to the above technical solution, the value of L1 ranges from 0.5nm to 3nm; The value of L2 ranges from 0.3nm to 1.5nm; The value of L3 ranges from 0.5nm to 2.5nm; 1.5nm≤L1+L2+L3≤5nm.

[0011] As an improvement to the above technical solution, the quantum barrier layer includes a first InGaN quantum barrier layer, a GaN quantum barrier layer, and a second InGaN quantum barrier layer; The thickness of the first InGaN quantum barrier layer is 0.3 nm to 1.5 nm; along the stacking direction, the In composition ratio of the first InGaN quantum barrier layer gradually changes from 0.02 to 0.08 to 0. The thickness of the GaN quantum barrier layer is 3nm~8nm; The thickness of the second InGaN quantum barrier layer is 0.3nm~1.5nm; along the stacking direction, the In composition ratio of the second InGaN quantum barrier layer gradually changes from 0 to 0.02~0.08.

[0012] As an improvement to the above technical solution, in the M quantum well layers close to the N-type semiconductor layer, along the stacking direction, the proportion of In composition in the first InGaN quantum well layer is increased by x. 11 Gradually becomes x 21 The In composition ratio of the second InGaN quantum well layer is x. 21 The In composition ratio of the third InGaN quantum well layer is determined by x. 21 Gradually becomes x 31 In the N quantum well layers close to the P-type semiconductor layer, along the stacking direction, the In composition ratio of the first InGaN quantum well layer is increased by x. 12 Gradually becomes x 22 The In composition ratio of the second InGaN quantum well layer is x. 22 The In composition ratio of the third InGaN quantum well layer is determined by x. 22 Gradually becomes x 32 , where x 21 >x 22 , and / or x 31 <x 32 ; The value of M is 1 or 2, and the value of N is 1 or 2.

[0013] As an improvement to the above technical solution, in the M quantum well layers close to the N-type semiconductor layer, the thickness of the second InGaN quantum well layer is L. 21 In the N quantum well layers close to the P-type semiconductor layer, the thickness of the second InGaN quantum well layer is L. 22 L 21 >L 22 .

[0014] Accordingly, the present invention also discloses a method for preparing an epitaxial structure, which includes the following steps: A substrate is provided on which a buffer layer, an N-type semiconductor layer, a multiple quantum well active layer and a P-type semiconductor layer are sequentially grown. The multi-quantum-well active layer comprises periodically alternating quantum well layers and quantum barrier layers. The quantum well layers include a first InGaN quantum well layer, a second InGaN quantum well layer, and a third InGaN quantum well layer. Along the stacking direction, the In composition ratio of the first InGaN quantum well layer gradually changes from x1 to x2, the In composition ratio of the second InGaN quantum well layer is x2, and the In composition ratio of the third InGaN quantum well layer gradually changes from x2 to x3, wherein x2 > x1, x2 > x3, and x1 > x3.

[0015] As an improvement to the above technical solution, the growth temperature of the quantum well layer is 700℃~780℃, and the growth temperature of the quantum barrier layer is 830℃~870℃.

[0016] Accordingly, the present invention also discloses an optical communication chip, including the above-described epitaxial structure.

[0017] Implementing this invention has the following beneficial effects: The epitaxial structure provided by this invention includes a first InGaN quantum well layer, a second InGaN quantum well layer, and a third InGaN quantum well layer. The In component ratio gradually changes in an inverted V-shape from both sides of the quantum well towards the center. Firstly, volume charge density is introduced into the first and third InGaN quantum well layers to generate a compensating electric field opposite to the polarization field. This compensating electric field forms electric field components opposite to the polarization field on the N-type side (first InGaN quantum well layer) and the P-type side (third InGaN quantum well layer) of the quantum well, respectively. These components work together to cancel the polarization field from both sides of the well towards the center, effectively offsetting the influence of the polarization field at the center and sides of the well. Secondly, the inverted V-shaped In component ratio distribution causes the conduction band bottom to form the lowest potential energy point at the center of the well, and the valence band top to form the highest potential energy point at the center of the well. After applying a forward bias, the bias electric field and the compensation electric field work together to form a potential energy gradient pointing towards the center of the well at the bottom of the conduction band. Corresponding electrons are attracted to the center, and simultaneously, a potential energy gradient also forms at the top of the valence band, attracting corresponding holes, thus increasing the overlap rate of the wave function distribution. Thirdly, by controlling x2 > x1, x2 > x3, and x1 > x3, an asymmetric inverted V-shaped gradient distribution is formed. By adjusting the barrier heights of the conduction and valence bands on both sides of the well, the injection efficiency of charge carriers from adjacent quantum barrier layers and the transport characteristics between quantum wells are optimized. Compared to conventional quantum well structures, the asymmetric inverted V-shaped quantum well provided by this invention can reduce the net electric field strength within the quantum well by 40% to 60%. Under the forward bias of the optical communication device, the overlap rate of holes and electrons can be increased to 45% to 60%, the radiative recombination rate can be increased by 2 to 4 times, the efficiency drop rate can be reduced to below 18%, and the modulation bandwidth can be increased to 2.3 to 3 times that of conventional quantum well structures. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of the epitaxial structure provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of the structure of a multi-quantum-well active layer provided in an embodiment of the present invention; Figure 3 This is a flowchart of a method for preparing an epitaxial structure according to an embodiment of the present invention. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be described in further detail below.

[0020] like Figure 1 and Figure 2 As shown, the present invention provides an epitaxial structure including a substrate 100, and a buffer layer 200, an N-type semiconductor layer 300, a multiple quantum well active layer 400 and a P-type semiconductor layer 500 sequentially stacked on the substrate 100. The multi-quantum-well active layer 400 includes periodically alternating quantum well layers 410 and quantum barrier layers 420. The quantum well layer 410 includes a first InGaN quantum well layer 411, a second InGaN quantum well layer 412, and a third InGaN quantum well layer 413. Along the stacking direction, the In composition ratio of the first InGaN quantum well layer 411 gradually changes from x1 to x2, the In composition ratio of the second InGaN quantum well layer 412 is x2, and the In composition ratio of the third InGaN quantum well layer 413 gradually changes from x2 to x3, where x2 > x1, x2 > x3, and x1 > x3.

[0021] In other words, the first InGaN quantum well layer 411 is a quantum well layer with a gradually increasing In content, the second InGaN quantum well layer 412 is a quantum well layer with a constant In content, and the third InGaN quantum well layer 413 is a quantum well layer with a gradually decreasing In content. The In content gradually changes in an asymmetric inverted V-shape (or Λ-shape) from the sides of the quantum well towards the center, meaning the In content reaches its peak at the center of the well and decreases towards the sides, with different magnitudes of change on the N-type and P-type sides. Firstly, by introducing volume charge density into the first InGaN quantum well layer 411 and the third InGaN quantum well layer 413, a compensating electric field opposite to the polarization field is generated. This compensating electric field acts bidirectionally within the well, effectively counteracting the influence of the polarization field at the center and sides, reducing the net electric field strength within the well by 40% to 60%. Secondly, the inverted V-shaped In content distribution causes the conduction band bottom to form the lowest potential energy point at the center of the well, and the valence band top to form the highest potential energy point at the center of the well. After applying a forward bias voltage, the bias electric field and the compensation electric field work together to form a potential energy gradient pointing towards the center of the well from the bottom of the conduction band. Corresponding electrons are attracted to the center, and simultaneously, a potential energy gradient also forms pointing towards the center from the top of the valence band, attracting corresponding holes to the center, thus maximizing the overlap of the wave function distribution. Compared to the 17%~25% overlap rate of rectangular wells, the quantum well of this invention can increase the overlap rate to 45%~60% under a forward bias voltage. Third, by adjusting the difference between x1 and x3, the initial In component ratio of the first InGaN quantum well layer 411 is higher than the termination In component ratio of the third InGaN quantum well layer 413, thereby adjusting the barrier heights of the conduction band and valence band on both sides of the well, optimizing the injection efficiency of charge carriers from adjacent quantum barrier layers 420 and the transport characteristics between quantum wells, further improving the optical response speed under high-speed modulation, reducing the chip efficiency drop rate to below 18%, increasing the modulation bandwidth to 2.3 to 3 times that of conventional quantum well structures, and significantly reducing the dependence of the emission wavelength on the current, which is beneficial to the wavelength stability of the optical communication channel and the maintenance of linearity under high-speed modulation.

[0022] It is understood that the gradual change in the In composition ratio of the first InGaN quantum well layer 411 and the gradual change in the In composition ratio of the third InGaN quantum well layer 413 can be one of linear gradual change, parabolic gradual change or multi-step step gradual change, with linear gradual change being preferred.

[0023] In one embodiment, the value of x1 ranges from 0.03 to 0.15, exemplarily 0.05, 0.08, 0.1, 0.12, or 0.14, but is not limited thereto. Preferably, the value of x1 ranges from 0.05 to 0.12. The value of x2 ranges from 0.1 to 0.35, exemplarily 0.12, 0.18, 0.22, 0.25, or 0.3, but is not limited thereto. x2 determines the emission wavelength, and preferably, the value of x2 ranges from 0.10 to 0.32. The value of x3 ranges from 0.02 to 0.12, exemplarily 0.03, 0.05, 0.08, 0.1, or 0.12, but is not limited thereto. Preferably, the value of x3 ranges from 0.03 to 0.1. In one embodiment, x2 - x1 > 0.03, and x2 - x3 > 0.05.

[0024] In one embodiment, the thickness of the first InGaN quantum well layer 411 is L1, the thickness of the second InGaN quantum well layer 412 is L2, and the thickness of the third InGaN quantum well layer 413 is L3, wherein L2 < L1, L2 < L3, and L2 < 10 nm. The central region has the highest In composition and the thinnest thickness. This combination of high In composition and thinness provides a deeper potential well, enhancing the confinement of charge carriers. Furthermore, the quantum confinement effect compresses the spatial distribution of charge carriers, making the electron and hole wave functions more concentrated in the central region. This combination of high composition and thinness reduces the radiative recombination lifetime from 5 ns to 20 ns in traditional rectangular wells to 1 ns to 3 ns, directly increasing the device's modulation bandwidth by 3 dB.

[0025] Specifically, in one embodiment, the value of L1 ranges from 0.5nm to 3nm, exemplarily 0.8nm, 1nm, 1.5nm, 2nm, or 2.5nm, but is not limited thereto. Preferably, the value of L1 ranges from 1nm to 2.5nm. The value of L2 ranges from 0.3nm to 1.5nm, exemplarily 0.5nm, 0.8nm, 1nm, 1.2nm, or 1.4nm, but is not limited thereto. Preferably, the value of L2 ranges from 0.5nm to 1.2nm. The value of L3 ranges from 0.5nm to 2.5nm, exemplarily 0.8nm, 1nm, 1.5nm, 2nm, or 2.2nm, but is not limited thereto. Preferably, the value of L3 ranges from 0.8nm to 2nm. In one embodiment, 1.5nm≤L1+L2+L3≤5nm, that is, the thickness of the quantum well layer 410 is 1.5nm~5nm, preferably, the thickness of the quantum well layer 410 is 2nm~4nm.

[0026] In one embodiment, among the M quantum well layers 410 near the N-type semiconductor layer 300, along the stacking direction, the In composition ratio of the first InGaN quantum well layer 411 is x 11 Gradually becomes x 21 The In composition ratio of the second InGaN quantum well layer 412 is x. 21 The In composition ratio of the third InGaN quantum well layer 413 is determined by x. 21 Gradually becomes x 31 In the N quantum well layers 410 near the p-type semiconductor layer 500, along the stacking direction, the In composition ratio of the first InGaN quantum well layer 411 is increased by x. 12 Gradually becomes x 22 The In composition ratio of the second InGaN quantum well layer 412 is x. 22 The In composition ratio of the third InGaN quantum well layer 413 is determined by x. 22 Gradually becomes x 32 , where x 21 >x 22 , and / or x 31 <x 32 ; In the M quantum well layers 410 adjacent to the N-type semiconductor layer 300, the thickness of the second InGaN quantum well layer 412 is L. 21 In the N quantum well layers 410 close to the p-type semiconductor layer 500, the thickness of the second InGaN quantum well layer 412 is L. 22 L 21 >L 22 ; The value of M is 1 or 2, and the value of N is 1 or 2.

[0027] By appropriately increasing x2 and L2 in one or two quantum well layers 410 near the N-type semiconductor layer 300 to capture holes, and by appropriately increasing x3 in one or two quantum well layers 410 near the P-type semiconductor layer 500 to enhance electron reflection, the uniformity of carrier distribution in the multiple quantum wells is improved, making the contribution of each well to the total light emission more balanced. This further suppresses efficiency drops and improves the linearity of optical response under high-speed modulation, thereby reducing inter-symbol interference and nonlinear distortion in the optical communication system.

[0028] It is understood that the quantum barrier layer 420 can be a GaN quantum barrier layer, an AlGaN quantum barrier layer, or an InGaN quantum barrier layer, but is not limited thereto. In one embodiment, the quantum barrier layer 420 includes a first InGaN quantum barrier layer, a GaN quantum barrier layer, and a second InGaN quantum barrier layer. The thickness of the first InGaN quantum barrier layer is 0.3 nm to 1.5 nm; along the stacking direction, the In composition percentage of the first InGaN quantum barrier layer gradually changes from 0.02 to 0.08 to 0; the thickness of the GaN quantum barrier layer is 3 nm to 8 nm; the thickness of the second InGaN quantum barrier layer is 0.3 nm to 1.5 nm; along the stacking direction, the In composition percentage of the second InGaN quantum barrier layer gradually changes from 0 to 0.02 to 0.08.

[0029] In one embodiment, the number of alternating stacking periods of the multi-quantum well active layer 400 is 3 to 10, exemplarily 4, 5, 6, 7 or 8, but not limited thereto. Preferably, the number of alternating stacking periods of the multi-quantum well active layer 400 is 4 to 7.

[0030] In addition to the multi-quantum-well active layer 400, the other layered structures of the present invention have the following characteristics: The substrate 100 may be a sapphire substrate or a silicon substrate, but is not limited thereto. In one embodiment, the substrate 100 is selected as a sapphire substrate.

[0031] The buffer layer 200 can be an AlN buffer layer, a GaN buffer layer, or an AlGaN buffer layer, but is not limited thereto. In one embodiment, the buffer layer 200 is a GaN buffer layer with a thickness of 15nm~35nm. The GaN buffer layer serves as a nucleation layer, forming high-density three-dimensional island-like nuclei on the sapphire substrate, providing nucleation centers for the subsequent two-dimensional epitaxial growth of the high-temperature GaN layer. Its lattice mismatch is partially relaxed through the Volmer-Weber growth mode, reducing the dislocation density from approximately 102. 10 cm -2 Dropped to about 10 8 cm -2This ensures high-quality crystal growth in the active region of subsequent optical communication transmitter devices.

[0032] The N-type semiconductor layer 300 can be an N-type GaN layer with a thickness of 1μm~5μm and a Si doping concentration of 1×10⁻⁶. 18 atoms / cm 3 ~5×10 18 atoms / cm 3 The N-type GaN layer provides electron injection and current spreading capabilities. A sufficiently thick N-type GaN layer ensures a uniform lateral current distribution, reducing current congestion. For high-speed optical communication devices, a uniform current distribution helps reduce the RC time constant and modulation distortion.

[0033] In one embodiment, an electron blocking layer is further included stacked on the multi-quantum-well active layer 400. The electron blocking layer may be a p-type Al. 0.1 Ga 0.9 An N electron blocking layer, 10 nm to 20 nm thick, with an Mg doping concentration of 1 × 10⁻⁶. 19 atoms / cm 3 ~5×10 19 atoms / cm 3 .

[0034] The P-type semiconductor layer 500 may include a P-type GaN layer and a P-type GaN ohmic contact layer. The thickness of the P-type GaN layer is 100 nm to 150 nm, and the Mg doping concentration is 3 × 10⁻⁶. 19 atoms / cm 3 ~8×10 19 atoms / cm 3 The thickness of the p-type GaN ohmic contact layer is 15nm~25nm, and the Mg doping concentration is 5×10⁻⁶. 19 atoms / cm 3 ~2×10 20 atoms / cm 3 .

[0035] Correspondingly, such as Figure 3 As shown, the present invention also discloses a method for preparing an epitaxial structure, which includes the following steps: S1. Provide a substrate 100.

[0036] In one embodiment, the substrate 100 is selected as a sapphire substrate. Specifically, the substrate 100 is placed on a graphite base in the MOCVD reaction chamber, the rotation speed of the graphite base is set to 500 r / min to 800 r / min, H2 is introduced as a carrier gas, the temperature of the reaction chamber is controlled at 1050℃ to 1100℃, the pressure is 100 mbar to 200 mbar, and in-situ annealing is performed for 5 min to 10 min to remove organic contaminants and the natural oxide layer on the surface of the substrate 100.

[0037] S2. A buffer layer 200, an N-type semiconductor layer 300, a multi-quantum-well active layer 400, and a P-type semiconductor layer 500 are sequentially grown on the substrate 100.

[0038] It is understandable that layered structures can be grown by MOCVD, MBE, PVD, or VPE, but are not limited to these methods.

[0039] In one implementation, S2 includes the following steps: S21, growth buffer layer 200.

[0040] In one embodiment, a GaN buffer layer is grown using MOCVD, with the reaction chamber temperature controlled at 500℃~550℃ and the pressure at 100mbar~200mbar. NH3 is introduced as the N source at a flow rate of 2slm~4slm, and TMGa is introduced as the Ga source at a flow rate of 30sccm~60sccm.

[0041] S22, grow an N-type semiconductor layer 300.

[0042] In one embodiment, an N-type GaN layer is grown using MOCVD. The reaction chamber temperature is controlled at 1050℃~1100℃, the pressure at 200mbar~300mbar, NH3 is introduced as the N source at a flow rate of 30slm~80slm, TMGa is introduced as the Ga source at a flow rate of 200sccm~500sccm, and SiH4 is introduced as the N-type doping source at a flow rate of 50sccm~200sccm to grow the N-type GaN layer.

[0043] S23, grow a multi-quantum-well active layer 400.

[0044] The fabrication of the multi-quantum-well active layer 400 includes: growing a quantum well layer 410 and a quantum barrier layer 420 by MOCVD, and repeatedly stacking and periodically growing the quantum well layer 410 and the quantum barrier layer 420.

[0045] Specifically, it includes the following steps: S231, grow quantum well layer 410.

[0046] The first InGaN quantum well layer 411 was grown using MOCVD. The reaction chamber temperature was controlled at 720℃~780℃, and the pressure at 100mbar~200mbar. NH3 was introduced as the N source, TMGa as the Ga source, and TMIn as the In source. During the growth process, the TMIn / TMGa flow rate was gradually increased, causing the In component percentage to gradually change from x1 to x2. The growth rate was 0.01nm / s~0.03nm / s. The second InGaN quantum well layer 412 was grown using MOCVD, maintaining a constant TMIn / TMGa flow rate during the growth process. The third InGaN quantum well layer 413 was grown using MOCVD, gradually decreasing the TMIn / TMGa flow rate during the growth process, causing the In component percentage to gradually change from x2 to x3.

[0047] S232, growth of quantum barrier layer 420.

[0048] GaN quantum barrier layer 420 was grown by MOCVD, with the reaction chamber temperature controlled at 830℃~870℃ and the pressure at 100mbar~400mbar. NH3 was introduced as the N source, TMGa as the Ga source, and SiH4 as the N-type doping source.

[0049] In one embodiment, a first InGaN quantum barrier layer is grown using MOCVD, with the reaction chamber temperature controlled at 720°C~780°C and the pressure at 100mbar~200mbar. NH3 is introduced as the N source, TMGa as the Ga source, and TMIn as the In source. A GaN quantum barrier layer is grown using MOCVD, with the reaction chamber temperature controlled at 830°C~870°C and the pressure at 100mbar~400mbar. NH3 is introduced as the N source, TMGa as the Ga source, and SiH4 as the N-type doping source. A second InGaN quantum barrier layer is grown using MOCVD, with the reaction chamber temperature controlled at 720°C~780°C and the pressure at 100mbar~200mbar. NH3 is introduced as the N source, TMGa as the Ga source, and TMIn as the In source.

[0050] S233, a quantum well layer 410 and a quantum barrier layer 420 are periodically grown in repeated stacks.

[0051] In one implementation, it further includes: S24, grow an electron blocking layer.

[0052] In one implementation, MOCVD is used to grow p-type Al. 0.1 Ga 0.9An N-type electron blocking layer was used as the electron blocking layer. The reaction chamber temperature was controlled at 860℃~920℃, and the pressure at 200mbar~300mbar. NH3 was introduced as the N source at a flow rate of 40slm~80slm, TMGa was introduced as the Ga source at a flow rate of 60sccm~200sccm, TMAl was introduced as the Al source at a flow rate of 30sccm~200sccm, and Cp2Mg was introduced as the P-type doping source at a flow rate of 50sccm~300sccm to grow P-type Al. 0.1 Ga 0.9 N electron blocking layer.

[0053] S25, grow a P-type semiconductor layer 500.

[0054] In one embodiment, a P-type GaN layer and a P-type GaN ohmic contact layer are grown using MOCVD as the P-type semiconductor layer 500. The reaction chamber temperature is controlled at 980°C~1030°C, the pressure at 200mbar~300mbar, NH3 is introduced as the N source at a flow rate of 40slm~90slm, TMGa is introduced as the Ga source at a flow rate of 300sccm~800sccm, and Cp2Mg is introduced as the P-type doping source at a flow rate of 200sccm~500sccm to grow the P-type GaN layer. The reaction chamber temperature is then lowered to 800°C~900°C, NH3 is introduced as the N source at a flow rate of 10slm~60slm, TMGa is introduced as the Ga source at a flow rate of 100sccm~300sccm, and Cp2Mg is introduced as the P-type doping source at a flow rate of 200sccm~600sccm to grow the P-type GaN ohmic contact layer. After growth, the reaction chamber temperature was lowered to 700℃~750℃, and in-situ annealing was performed for 10min~20min under a N2 atmosphere. Annealing decomposed the Mg-H complex, allowing H atoms to diffuse out of the crystal, activating the Mg acceptors into effective acceptor centers, and reducing the hole concentration from approximately 10% before annealing. 17 cm -3 Increased to approximately 10 19 cm -3 The resistivity of the P-type semiconductor layer 500 is reduced by 2 to 3 orders of magnitude, which helps to reduce the series resistance and RC time constant of the device and plays an important role in improving the modulation bandwidth of optical communication.

[0055] The present invention will be further described below with reference to specific embodiments: Example 1 This embodiment provides an epitaxial structure, including a substrate, and a buffer layer, an N-type semiconductor layer, a multiple quantum well active layer, an electron blocking layer and a P-type semiconductor layer sequentially stacked on the substrate.

[0056] The substrate is a c-plane sapphire substrate with a thickness of 660 μm.

[0057] The buffer layer consists of an LT-GaN layer and a U-type GaN layer. The thickness of the LT-GaN layer is 25 nm, and the thickness of the U-type GaN layer is 3 μm.

[0058] The N-type semiconductor layer is an N-type GaN layer with a thickness of 2.5 μm and a Si doping concentration of 3 × 10⁻⁶. 18 atoms / cm 3 .

[0059] The multi-quantum-well active layer comprises periodically alternating quantum well layers and quantum barrier layers, with a period number of 5. The quantum well layers include a first InGaN quantum well layer, a second InGaN quantum well layer, and a third InGaN quantum well layer.

[0060] In each cycle, the In composition ratio of the first InGaN quantum well layer gradually changes from 0.1 to 0.2, and the thickness is 1.5 nm; the In composition ratio of the second InGaN quantum well layer is 0.2, and the thickness is 0.8 nm; the In composition ratio of the third InGaN quantum well layer gradually changes from 0.2 to 0.08, and the thickness is 1.2 nm.

[0061] The quantum barrier layer is a GaN quantum barrier layer with a thickness of 6 nm.

[0062] The electron blocking layer is a p-type Al 0.1 Ga 0.9 An N electron blocking layer, 15 nm thick, with a Mg doping concentration of 2 × 10⁻⁶. 19 atoms / cm 3 .

[0063] The p-type semiconductor layer consists of a p-type GaN layer and a p-type GaN ohmic contact layer. The thickness of the p-type GaN layer is 120 nm, and the Mg doping concentration is 5 × 10⁻⁶. 19 atoms / cm 3 The thickness of the p-type GaN ohmic contact layer is 20 nm, and the Mg doping concentration is 1 × 10⁻⁶. 20 atoms / cm 3 .

[0064] The above epitaxial structure was fabricated into a 350μm×350μm blue light chip using standard chip manufacturing processes, and its performance was tested. The results are as follows: Peak wavelength: 455nm, Full width at half maximum (FWHM): 20nm, Forward voltage (20mA): 3.1V, Optical output power (20mA): 28mW, Optical output power (350mA): 340mW, Peak external quantum efficiency (EQE): 55%, 3dB modulation bandwidth: 95MHz (350mA drive), OOK modulation rate: 80Mb / s (BER < 1×10-9), 16-QAM-OFDM modulation: supports 256 subcarriers, Data rate: 400Mb / s, Efficiency dropout rate (100A / cm²) 2 The peak wavelength shift (vs. peak) was 13%, and the peak wavelength shift within the injection current range of 10mA to 350mA was 2.8nm. For comparison, a conventional quantum well structure with the same structural parameters (i.e., a constant In content of 0.2%) was prepared. This comparative sample was identical to Example 1 in all aspects except for the constant In content distribution of the quantum well layer (0.2, not a gradual change). The performance test results of the comparative sample are as follows: EQE: 45%, modulation bandwidth: 40MHz. The blue light chip provided by this embodiment of the invention shows an approximately 22% improvement in EQE and an approximately 138% improvement in modulation bandwidth.

[0065] Example 2 This embodiment provides an epitaxial structure that differs from Embodiment 1 in that, in the first period, the In composition ratio of the first InGaN quantum well layer gradually changes from 0.1 to 0.2, and the thickness is 1.5 nm; the In composition ratio of the second InGaN quantum well layer is 0.2, and the thickness is 1.2 nm; the In composition ratio of the third InGaN quantum well layer gradually changes from 0.2 to 0.08, and the thickness is 1.3 nm. In the second period, the In composition ratio of the first InGaN quantum well layer gradually changes from 0.08 to 0.18, and the thickness is 1.5 nm; the In composition ratio of the second InGaN quantum well layer is 0.18, and the thickness is 1 nm; the In composition ratio of the third InGaN quantum well layer gradually changes from 0.18 to 0.05, and the thickness is 1.3 nm. In the third cycle, the In composition of the first InGaN quantum well layer gradually changes from 0.08 to 0.18, with a thickness of 1.3 nm; the In composition of the second InGaN quantum well layer is 0.18, with a thickness of 0.8 nm; and the In composition of the third InGaN quantum well layer gradually changes from 0.18 to 0.05, with a thickness of 1.2 nm. In the fourth cycle, the In composition of the first InGaN quantum well layer gradually changes from 0.06 to 0.16, with a thickness of 1.2 nm; the In composition of the second InGaN quantum well layer is 0.16, with a thickness of 0.6 nm; and the In composition of the third InGaN quantum well layer gradually changes from 0.16 to 0.06, with a thickness of 1 nm. In the fifth cycle, the In composition ratio of the first InGaN quantum well layer gradually changes from 0.06 to 0.15, and the thickness is 1.2 nm; the In composition ratio of the second InGaN quantum well layer is 0.15, and the thickness is 0.5 nm; the In composition ratio of the third InGaN quantum well layer gradually changes from 0.15 to 0.08, and the thickness is 1 nm.

[0066] Everything else is the same as in Example 1.

[0067] The above epitaxial structure was fabricated into a 350μm×350μm blue light chip using standard chip manufacturing processes, and its performance was tested. The results are as follows: Peak wavelength: 455nm, Full width at half maximum (FWHM): 18nm, Forward voltage (20mA): 3.05V, Optical output power (20mA): 32mW, Optical output power (350mA): 380mW, Peak external quantum efficiency (EQE): 62%, 3dB modulation bandwidth: 120MHz (350mA drive), OOK modulation rate: 100Mb / s (BER < 1×10⁻⁶). -9 ), 16-QAM-OFDM modulation: supports 256 subcarriers, data rate 500Mb / s, efficiency dropout rate (100A / cm). 2 (vs peak): 10%, peak wavelength shift: 2.1nm in the injection current range of 10mA~350mA.

[0068] The performance test results of the conventional quantum well structure with the same structural parameters (i.e., the In component ratio is constant at 0.18%) are as follows: EQE: 48%, modulation bandwidth: 45MHz. The blue light chip provided in this embodiment of the invention improves EQE by 29% and modulation bandwidth by 167%.

[0069] Example 3 This embodiment provides an epitaxial structure that differs from Embodiment 1 in that the number of alternating stacks of the multiple quantum well active layers is 6.

[0070] The quantum barrier layer consists of a first InGaN quantum barrier layer, a GaN quantum barrier layer, and a second InGaN quantum barrier layer. The first InGaN quantum barrier layer has a thickness of 0.8 nm, and the In composition gradually changes from 0.05 to 0. The GaN quantum barrier layer has a thickness of 5 nm. The second InGaN quantum barrier layer has a thickness of 0.8 nm, and the In composition gradually changes from 0 to 0.05.

[0071] In the first cycle, the In composition of the first InGaN quantum well layer gradually changes from 0.22 to 0.32, with a thickness of 1.2 nm; the In composition of the second InGaN quantum well layer is 0.32, with a thickness of 0.8 nm; and the In composition of the third InGaN quantum well layer gradually changes from 0.32 to 0.18, with a thickness of 1 nm. In the second cycle, the In composition of the first InGaN quantum well layer gradually changes from 0.2 to 0.31, with a thickness of 1 nm; the In composition of the second InGaN quantum well layer is 0.31, with a thickness of 0.8 nm; and the In composition of the third InGaN quantum well layer gradually changes from 0.31 to 0.16, with a thickness of 0.9 nm. In the third cycle, the In composition of the first InGaN quantum well layer gradually changes from 0.2 to 0.3, with a thickness of 1.0 nm; the In composition of the second InGaN quantum well layer is 0.3, with a thickness of 0.6 nm; and the In composition of the third InGaN quantum well layer gradually changes from 0.3 to 0.16, with a thickness of 0.9 nm. In the fourth cycle, the In composition of the first InGaN quantum well layer gradually changes from 0.18 to 0.29, with a thickness of 0.9 nm; the In composition of the second InGaN quantum well layer is 0.29, with a thickness of 0.5 nm; and the In composition of the third InGaN quantum well layer gradually changes from 0.29 to 0.14, with a thickness of 0.8 nm. In the fifth cycle, the In composition of the first InGaN quantum well layer gradually changes from 0.18 to 0.28, with a thickness of 0.9 nm; the In composition of the second InGaN quantum well layer is 0.28, with a thickness of 0.5 nm; and the In composition of the third InGaN quantum well layer gradually changes from 0.28 to 0.16, with a thickness of 0.8 nm. In the sixth cycle, the In composition of the first InGaN quantum well layer gradually changes from 0.18 to 0.28, with a thickness of 0.9 nm; the In composition of the second InGaN quantum well layer is 0.28, with a thickness of 0.5 nm; and the In composition of the third InGaN quantum well layer gradually changes from 0.28 to 0.18, with a thickness of 0.7 nm.

[0072] Everything else is the same as in Example 1.

[0073] The above epitaxial structure was fabricated into a 350μm×350μm green light chip using standard chip manufacturing processes, and its performance was tested. The results are as follows: Peak wavelength: 525nm, Full width at half maximum (FWHM): 28nm, Forward voltage (20mA): 3.15V, Optical output power (20mA): 18mW, Optical output power (200mA): 180mW, Peak external quantum efficiency (EQE): 38%, 3dB modulation bandwidth: 85MHz (200mA drive), Efficiency drop (100A / cm²) 2(vs peak): 18%, underwater transmission rate in 3m clear water: 10Mb / s (BER<1×10) -6 ).

[0074] The performance test results of the conventional quantum well structure with the same structural parameters (i.e., the In component ratio is constant at 0.3%) are as follows: EQE: 22%, modulation bandwidth: 30MHz. The green light chip provided in this embodiment of the invention has an EQE improvement of approximately 73% and a modulation bandwidth improvement of approximately 183%.

[0075] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. An epitaxial structure, characterized in that, It includes a substrate, and a buffer layer, an N-type semiconductor layer, a multi-quantum-well active layer and a P-type semiconductor layer sequentially stacked on the substrate; The multi-quantum-well active layer comprises periodically alternating quantum well layers and quantum barrier layers. The quantum well layers include a first InGaN quantum well layer, a second InGaN quantum well layer, and a third InGaN quantum well layer. Along the stacking direction, the In composition ratio of the first InGaN quantum well layer gradually changes from x1 to x2, the In composition ratio of the second InGaN quantum well layer is x2, and the In composition ratio of the third InGaN quantum well layer gradually changes from x2 to x3, wherein x2 > x1, x2 > x3, and x1 > x3.

2. The epitaxial structure as described in claim 1, characterized in that, The value of x1 ranges from 0.03 to 0.15; The value of x2 ranges from 0.1 to 0.35; The value of x3 ranges from 0.02 to 0.12; x2-x1>0.03, x2-x3>0.

05.

3. The epitaxial structure as described in claim 1, characterized in that, The thickness of the first InGaN quantum well layer is L1, the thickness of the second InGaN quantum well layer is L2, and the thickness of the third InGaN quantum well layer is L3, wherein L2 < L1, L2 < L3, and L2 < 10 nm.

4. The epitaxial structure as described in claim 3, characterized in that, The value of L1 ranges from 0.5nm to 3nm; The value of L2 ranges from 0.3nm to 1.5nm; The value of L3 ranges from 0.5nm to 2.5nm; 1.5nm≤L1+L2+L3≤5nm.

5. The epitaxial structure as described in claim 1, characterized in that, In the M quantum well layers close to the N-type semiconductor layer, along the stacking direction, the In composition ratio of the first InGaN quantum well layer is x 11 Gradually becomes x 21 The In composition ratio of the second InGaN quantum well layer is x. 21 The In composition ratio of the third InGaN quantum well layer is determined by x. 21 Gradually becomes x 31 In the N quantum well layers close to the P-type semiconductor layer, along the stacking direction, the In composition ratio of the first InGaN quantum well layer is increased by x. 12 Gradually becomes x 22 The In composition ratio of the second InGaN quantum well layer is x. 22 The In composition ratio of the third InGaN quantum well layer is determined by x. 22 Gradually becomes x 32 , where x 21 >x 22 , and / or x 31 <x 32 ; The value of M is 1 or 2, and the value of N is 1 or 2.

6. The epitaxial structure as described in claim 5, characterized in that, In the M quantum well layers located close to the N-type semiconductor layer, the thickness of the second InGaN quantum well layer is L. 21 In the N quantum well layers close to the p-type semiconductor layer, the thickness of the second InGaN quantum well layer is L. 22 L 21 >L 22 .

7. The epitaxial structure as described in claim 1, characterized in that, The quantum barrier layer includes a first InGaN quantum barrier layer, a GaN quantum barrier layer, and a second InGaN quantum barrier layer; The thickness of the first InGaN quantum barrier layer is 0.3 nm to 1.5 nm; along the stacking direction, the In composition ratio of the first InGaN quantum barrier layer gradually changes from 0.02 to 0.08 to 0. The thickness of the GaN quantum barrier layer is 3nm~8nm; The thickness of the second InGaN quantum barrier layer is 0.3nm~1.5nm; along the stacking direction, the In composition ratio of the second InGaN quantum barrier layer gradually changes from 0 to 0.02~0.

08.

8. A method for preparing an epitaxial structure, used to prepare the epitaxial structure as described in any one of claims 1 to 7, characterized in that, Includes the following steps: A substrate is provided on which a buffer layer, an N-type semiconductor layer, a multiple quantum well active layer and a P-type semiconductor layer are sequentially grown. The multi-quantum-well active layer comprises periodically alternating quantum well layers and quantum barrier layers. The quantum well layers include a first InGaN quantum well layer, a second InGaN quantum well layer, and a third InGaN quantum well layer. Along the stacking direction, the In composition ratio of the first InGaN quantum well layer gradually changes from x1 to x2, the In composition ratio of the second InGaN quantum well layer is x2, and the In composition ratio of the third InGaN quantum well layer gradually changes from x2 to x3, wherein x2 > x1, x2 > x3, and x1 > x3.

9. The method for preparing the epitaxial structure as described in claim 8, characterized in that, The growth temperature of the quantum well layer is 700℃~780℃, and the growth temperature of the quantum barrier layer is 830℃~870℃.

10. An optical communication chip, characterized in that, Includes the epitaxial structure as described in any one of claims 1 to 7.

Images