Vertical cavity surface emitting semiconductor laser and manufacturing method thereof

Abstract

The invention relates to the technical field of semiconductor lasers, in particular to a vertical-cavity surface-emitting semiconductor laser and a manufacturing method thereof.The vertical-cavity surface-emitting semiconductor laser comprises a substrate, an N-type DBR layer, an active region, an oxide layer, a P-type DBR layer, a P contact layer and a P metal layer which are sequentially arranged from bottom to top, oxidation holes are formed in the oxide layer, a carbon displacement modulation structure layer is arranged in the P contact layer, and the P metal layer is arranged in the N-type DBR layer. The lower end of the carbon displacement modulation structure layer is located in the P-type DBR layer, the carbon displacement modulation structure layer is right opposite to the oxidation hole, and the material of the carbon displacement modulation structure layer is GaAs doped with carbon atoms. The high-frequency characteristic, single-mode performance, ultra-long service life, wide-temperature work and extreme environment resistance which are difficult to reach by a traditional pure oxidation limiting scheme and a traditional diffusion technology and oxidation limiting VCSEL scheme are achieved through cooperation of precise control of the SEG technology and the ultra-low diffusion characteristic of carbon.

Description

Technical Field

[0001] This application relates to the technical field of semiconductor lasers, and in particular to a vertical cavity surface-emitting semiconductor laser and a method for manufacturing the same. Background Technology

[0002] With the increasing demands on optoelectronic device performance from scenarios such as high-speed optical interconnects in data centers, wide-temperature control in industrial applications, and resistance to extreme environments in aerospace, the market urgently needs single-mode vertical-cavity surface-emitting lasers (VCSELs) that meet the requirements of "high-frequency modulation above 100Gbps, stable operation over a wide temperature range of -55℃ to 125℃, long lifespan of 100,000 hours, and resistance to radiation and aging." While current mainstream VCSEL fabrication methods (traditional oxide-confined and Zn / Be / Mg-diffused types) can achieve basic single-mode and high-frequency characteristics, significant technical bottlenecks exist in the synergistic optimization of "long-term reliability, high-frequency stability, and mass production safety." Specific problems and solutions are compared below: Traditional oxidation-limited VCSELs suffer from both insufficient reliability and performance. Core technological bottlenecks: 1. Low current control accuracy and limited high-frequency performance: The oxidation process is affected by water vapor partial pressure and crystal orientation differences. The oxidation pore size edge is irregular (deviation ±0.5um) and the longitudinal oxidation depth is uneven (deviation ±10nm), resulting in a current lateral diffusion ratio of 3% to 5%. Carriers accumulate at the edge of the active region, the SHB effect is significant, and the 3dB bandwidth is mostly ≤20GHz. 2. Poor long-term reliability and severe aging degradation: The Al2O3-GaAs interface contains a large number of dangling bonds (interface state density ≥ 1×10⁻⁶). 11 cm -2 Under prolonged high temperature (125℃), it is prone to adsorbing water vapor and forming oxidation defects. At the same time, Al atoms diffuse into the active region. After 1000 hours of aging, the power decay is ≥10% and the threshold current increase is ≥8%. 3. Weak resistance to extreme environments: The oxide layer has rough edges (roughness ≥ 5 nm), making it prone to mode competition under radiation environments, 1×10 6 After rad-γ radiation, the RIN deteriorates to -125 dB / Hz, which cannot meet the radiation resistance requirements of aerospace.

[0003] Zn / Be / Mg diffusion-type VCSELs: Uncontrollable diffusion leads to insufficient stability. Core technology bottleneck: 1. Uncontrollable lateral diffusion and low precision of the current channel. Zn atoms have a high diffusion coefficient in GaAs, and during diffusion, they expand laterally by 1-2 μm, resulting in an actual current channel (4-6 μm) larger than the design value. This leads to a decrease in the matching degree with the high-intensity region of the optical field, incomplete suppression of the SHB effect, and bandwidth attenuation of ≥5% after long-term high-speed modulation operation. 2. Uneven concentration gradient, low carrier transport efficiency: Zn in high Al composition AlGaAs (such as Al 0.93 Ga 0.07 The diffusion coefficient of As is only 1 / 100 of that of the low Al component, resulting in a Zn concentration gradient of 10 within the DBR layer. 4 cm -3 / nm, the current transmission path has a "high resistance region", and the series resistance is 20% higher than the target value; 3. High-temperature diffusion damage and lattice quality deterioration: The diffusion temperature of 620℃ is close to the thermal decomposition temperature of GaAs, which easily leads to interlayer mixing in DBR (interdiffusion of Al components between high Al layer and low Al layer), the reflectivity drops from 99.5% to 98.5%, and the optical gain efficiency decreases by 8%-10%. Summary of the Invention

[0004] To address the aforementioned problems, this application provides a vertical cavity surface-emitting semiconductor laser and a method for manufacturing the same.

[0005] One objective of this application is to provide a vertical-cavity surface-emitting semiconductor laser, employing the following technical solution: A vertical cavity surface-emitting semiconductor laser includes, from bottom to top, a substrate, an N-type DBR layer, an active region, an oxide layer, a P-type DBR layer, a P-contact layer, and a P-metal layer. Oxide holes are formed in the oxide layer. A carbon substitution modulation structure layer is disposed within the P-contact layer. The lower end of the carbon substitution modulation structure layer is located within the P-type DBR layer. The carbon substitution modulation structure layer is directly opposite the oxide holes. The material of the carbon substitution modulation structure layer is GaAs doped with carbon atoms.

[0006] By adopting the above technical solution, this application achieves dual current confinement through the nested synergy of "SEG carbon-doped pore size plus oxide pore size": First-level current limiting: SEG carbon substitution modulation structure layer: only the carbon-doped GaAs carbon substitution modulation structure layer is grown within the window (low-resistance region, 1.2 × 10⁻⁶). -3 -1.5×10 -3 Ω·cm), outside the window is a covered high-resistivity region (resistivity ≥ 1×10 Ω·cm). 3 (Ω·cm), the current is initially confined within the carbon substitution modulation structure layer; Secondary current limiting: Oxidation pore size: After the carbon substitution modulation structure layer is grown, the underlying AlGaAs layer is subjected to wet oxygen oxidation to form oxidation pore size, which further limits the injection current.

[0007] Advantages in current limiting accuracy: The diameter of the carbon substitution modulation structure layer is determined by the photolithography accuracy (±0.1um), and the oxide aperture is achieved with an accuracy of ±0.2um by controlling the oxidation time (e.g., 30min). The nesting deviation between the two is ≤0.3um, resulting in more precise current limiting.

[0008] Suppression of spatial hole burning (SHB) effect: Precise matching of current injection and optical field: The optical field of the active region of VCSEL exhibits a Gaussian distribution, with 80% of the light intensity concentrated in the central region. The photon density is high in the center (cathode consumption is fast) and low at the edges (cathode consumption is slow). The SEG carbon substitution modulation structure layer is directly aligned with the central region of the optical field. Current is injected only into the central region where carrier consumption is fastest, and the uniformity of carrier concentration in the carbon substitution modulation structure layer is ≤7%. The current is evenly distributed in the central region, avoiding the problem of "large current injection area leading to carrier accumulation at the edges" in traditional schemes. The carrier injection rate is perfectly matched with the photon consumption rate in the central region. The low-frequency roll-off (≤1GHz) caused by SHB originates from the "supplementary delay" after the central carriers are depleted. In this application, due to the current-optical field matching degree ≥95%, the low-frequency modulation response flatness is improved by 15%, and the RIN is stabilized at -149dB / Hz.

[0009] Flattening of the light field distribution: The synergistic effect of the oxide apertures and the carbon substitution modulation structure layer makes the light field distribution flatter. The oxide apertures, through the difference in refractive index (Al2O3 refractive index 1.7, GaAs refractive index 3.6), initially confine the light field within the oxide apertures, preventing the light field from leaking to the edges; the carbon substitution modulation structure layer (high carrier concentration) has a slight absorption on the light field, which can weaken the local high intensity at the center of the light field, reducing the gradient of light field intensity from the center to the edge from 20% to 10%, reducing the difference in photon density—the carrier consumption is more uniform, there is no local "hole burning", and it provides support for a wide bandwidth greater than 30 GHz.

[0010] The dual-aperture optical field is synergistically reduced, resulting in more stable mode selection and enabling single-mode output. This application achieves more stable single-mode output by "preliminary constraint of the oxide aperture plus further reduction of the optical field by a carbon substitution modulation structure layer": precise control of the lateral dimension of the optical field; first-level optical field constraint: oxide aperture: the oxide aperture restricts the optical field within the oxide aperture range through total internal reflection, initially suppressing large-size high-order modes; second-level optical field reduction: carbon substitution modulation structure layer: the high carrier concentration of the carbon substitution modulation structure layer has a "mode filtering" effect on the optical field - only the fundamental mode can be stably transmitted in the narrow channel of the carbon substitution modulation structure layer, and high-order modes will be absorbed or leaked by the carbon substitution modulation structure layer because the width of the optical field distribution exceeds the diameter of the carbon substitution modulation structure layer (high-order mode loss ≥20dB / cm, fundamental mode loss ≤5dB / cm); when the lateral dimension of the optical field is 3.5-4.7 times the working wavelength, the mode gain of the high-order mode is lower than the threshold gain (fundamental mode gain ≥15dB, high-order mode gain ≤8dB), and only the fundamental mode achieves resonant output.

[0011] Single-mode stability advantages: No mode shift: The low diffusion characteristics of carbon ensure long-term stability of the diameter of the carbon substitution modulation structure layer (deviation ≤0.1um after 1000 hours of operation), and there is no shift in optical field constraint; while the Zn diffusion scheme has a diffusion aperture shift ≥0.3um due to Zn atom migration, and the single-mode performance deteriorates. High SMSR output: Through dual-aperture synergy, the solution stabilizes the fundamental mode suppression ratio (SMSR) at 38-40dB and the vertical divergence angle at ≤14°, resulting in more stable single-mode output.

[0012] Preferably, the carbon atom concentration in the carbon substitution modulation structure layer is 1×10⁻⁶. 19 -1×10 20 cm -3 .

[0013] Preferably, the diameter of the carbon substitution modulation structure layer is smaller than the pore size of the oxide pores.

[0014] By adopting the above technical solution, the oxide pore size covers the diameter of the carbon substitution modulation structure layer, further blocking the leakage current outside the carbon substitution modulation structure layer.

[0015] Another objective of this application is to provide a method for manufacturing a vertical-cavity surface-emitting semiconductor laser, employing the following technical solution: A method for manufacturing a vertical-cavity surface-emitting semiconductor laser (VCSEL), comprising the following steps: S1. Epitaxial growth: An epitaxial layer is grown on the substrate. The epitaxial layer includes an N-type DBR layer, an active region, an oxide layer, a P-type DBR layer, a P-contact layer, and a P-metal layer. S2, Preparation of the carbon substitution modulation structure layer; S2 includes the following steps. S221, Deposit a SiN layer on the P contact layer. X , S222, Photolithography plus dry etching in SiN X A circular window matching the high-intensity region of the Gaussian distribution of the light field is etched out in the center, and the circular window penetrates the P contact layer; S223. Carbon atom doping is performed using MOCVD, with CCl4 as the doping source to provide carbon atoms to be doped on the P contact layer through a circular window, forming a carbon substitution modulation structure layer.

[0016] By adopting the above technical solution, SiN XThe mask window size is precisely controlled by photolithography, and the carbon substitution modulation structure layer grows only within the window, resulting in absolutely clear lateral boundaries (no edge blurring issues caused by oxidation), with a lateral current diffusion ratio ≤0.5%. C atoms form acceptor levels (ionization energy 26 meV) through "Ga substitution sites," providing hole conductivity. The synergistic effect of a high V / III ratio and high temperature—As atoms occupy surface adsorption sites, reducing C atom aggregation, while high temperature promotes C atom migration in the lattice, dispersing locally high-concentration areas; the solid solubility limitation of carbon leads to a doping concentration (8 × 10⁻⁶). 18 ) lower than Zn (1×10 19 The resistivity is slightly high. By increasing the thickness by 50nm (resistance is proportional to thickness), the total resistance can be kept consistent without affecting the bandwidth.

[0017] As a preferred embodiment, in S223, the growth temperature is 600-610℃; the growth pressure is 200 mTorr; and the As / Ga ratio is 60.

[0018] By adopting the above technical solution, the carbon atom substitution rate is ≥90% at around 600℃; however, it should not exceed 610℃ to avoid GaAs decomposition.

[0019] Preferably, S2 also includes a step preceding S221. S211. A two-step method of diamond grinding followed by colloidal silica polishing is used to control the surface roughness of the p-contact layer to Ra≤0.3nm, and AFM full-wafer scanning is used to ensure 5×5um. 2 The protrusions within the region are ≤1nm; S212. Clean with acetone and isopropanol for 10 minutes each. S213. Soak in diluted HF for 5 minutes, then rinse with DIW. S214. Ashing was performed using O2 / Ar plasma, and the surface carbon content was determined by XPS to be ≤0.05%.

[0020] By adopting the above technical solution, the subsequent carbon-doped epitaxial temperature reaches 600℃, thus thickening SiN. X The mask can enhance high temperature resistance and prevent the mask from shrinking at high temperatures, which would cause window deformation (diameter deviation ≤0.1um); residue at the window edge can block the diffusion of C atoms, resulting in a 20% decrease in local doping concentration. BOE treatment can ensure uniform doping at the edge.

[0021] As a preferred method, during the S223 process, the intensity of the C-C bond characteristic peaks is monitored in real time using in-situ infrared spectroscopy.

[0022] As a preferred embodiment, a step following S223 is also included. S231. Rapid thermal annealing in a N2 atmosphere at 400°C for 30 seconds; S232. Dry etching is performed using a Cl2 / BCl3 mixed gas to remove the 5-10nm high protrusions at the edge of the SEG epitaxial layer window, making the epitaxial layer surface flush with the SiN. X The mask surface is flush.

[0023] By adopting the above technical solution, the carbon substitution modulation structure layer may have a small number of "C interstitial atoms" (shallow energy level defects that lead to carrier traps). Annealing at 400℃ can promote the migration of interstitial C to Ga vacancies (substitution sites), increasing the activation rate from 90% to 95%.

[0024] Preferably, in step S222, after etching out the circular window, it is immersed in BOE to remove the SiN at the edge of the circular window. X Residue.

[0025] By adopting the above technical solution, the GaAs surface is fully exposed, avoiding the carbon doping shadowing effect.

[0026] In summary, this application includes at least one of the following beneficial technical effects: 1. Higher current limiting accuracy and a more than 20% improvement in high-frequency performance (bandwidth).

[0027] 2. SiN X The window size of the mask is precisely controlled by photolithography, and the carbon substitution modulation structure layer grows only within the window, with absolutely clear lateral boundaries (no edge blurring problem due to oxidation), and the lateral current diffusion ratio is ≤0.5%.

[0028] 3. The resistivity uniformity of the carbon substitution modulation structure layer is ≤7% (the resistivity uniformity of the oxide layer in the traditional oxide scheme is ≥15%), the current density distribution is more concentrated (overlap rate with the high intensity region of the light field is ≥95%), the SHB effect is extremely suppressed (RIN≤-149dB / Hz, while the traditional scheme is -135dB / Hz), the carrier consumption and replenishment are more balanced during high-frequency modulation, and the bandwidth is improved.

[0029] 4. Long-term reliability is improved by more than 30%, and lifespan is extended to 100,000 hours.

[0030] 5. The current limiting region is composed of "SiN" X The structure consists of a mask and carbon-doped GaAs, and SiN. X It has extremely strong chemical inertness (does not react with water vapor) and low interfacial state density with GaAs (≤5×10⁻⁶). 10 cm -2 There is no interface aging issue.

[0031] 6. Carbon has an extremely low diffusion coefficient, only 1 / 100 that of Al. During high-temperature aging, carbon atoms migrate almost no further (diffusion distance ≤ 0.1 nm / 1000 hours), the doping distribution is stable, and the resistance change of the current channel is ≤ 1%, avoiding the increase in threshold current caused by increased resistance.

[0032] 7. Optimal long-term doping stability, suitable for extreme environments (-55℃ to 125℃). After wide temperature cycling (-55℃ to 125℃, 1000 cycles), the carrier concentration change in the carbon substitution modulation structure layer is ≤2%. The diffusion coefficient of carbon is much lower than that of Zn, Be, and Mg, meaning that carbon atoms will hardly leave their lattice positions (substitution Ga sites) when the temperature fluctuates, and the carrier concentration remains stable.

[0033] 8. It exhibits minimal electrical performance degradation at high temperatures, making it suitable for wide-temperature industrial applications. At high temperatures, the ionization energy of the doped atoms determines the stability of carrier concentration: the ionization energy of carbon in GaAs is only 26 meV, and its ionization rate is even higher at high temperatures; 9. Carbon has a higher lattice matching degree with GaAs (carbon atom radius is 0.077nm, Ga is 0.126nm, and lattice distortion after substitution is ≤0.5%), making it less prone to lattice relaxation at high temperatures and resulting in less carrier mobility decay.

[0034] 10. Excellent radiation resistance, suitable for aerospace and nuclear industry applications. (In 1×10...) 6 After rad-γ-ray radiation, the power decay of carbon-doped VCSELs is ≤8%. Radiation generates "vacancy-interstitial atom pairs" (defects) in semiconductors, which can scatter charge carriers or become non-radiative recombination centers. The high binding energy of carbon atoms with the GaAs lattice can "pin" radiation-induced defects (reducing defect migration) and lessening the impact on charge carrier transport, thus making the performance more robust in radiation environments. Attached Figure Description

[0035] Figure 1 This is a schematic diagram of the overall structure of Embodiment 1.

[0036] Explanation of reference numerals in the attached figures: 1. Substrate; 2. N-type DBR layer; 3. Active region; 4. Oxide layer; 5. P-type DBR layer; 6. P-contact layer; 7. P-metal layer; 8. Oxide hole; 9. Carbon substitution modulation structure layer. Detailed Implementation

[0037] The present application will be further described in detail below with reference to all the accompanying drawings.

[0038] Example 1

[0039] This application discloses a vertical-cavity surface-emitting semiconductor laser, with reference to... Figure 1The structure includes, from bottom to top, a substrate 1, an N-type DBR layer 2, an active region 3, an oxide layer 4, a P-type DBR layer 5, a P-contact layer 6, and a P-metal layer 7. Oxide holes 8 are formed on the oxide layer 4. A carbon substitution modulation structure layer 9 is formed in the P-contact layer 6. The lower end of the carbon substitution modulation structure layer 9 is located in the P-type DBR layer 5. The carbon substitution modulation structure layer 9 is directly opposite the oxide holes 8. The material of the carbon substitution modulation structure layer 9 is GaAs doped with carbon atoms.

[0040] The parameters of carbon substitution modulation structure layer 9 are: substitution pore size Wz = 3-4 μm (±0.1 μm), substitution growth depth d = 0.5 μm, and carbon atom concentration = 1 × 10⁻⁶. 19 -1×10 20 cm -3 The pore size of oxide pore 8 is Wo = 6-7 μm (±0.2 μm).

[0041] Example 2

[0042] This application discloses a method for manufacturing a vertical-cavity surface-emitting diode (VCSEL) laser, used to manufacture a VCSEL laser as described in the above embodiments, comprising the following steps: S1, epitaxial growth; The specific technical solution is as follows: S11, N-type GaAs substrate crystal orientation ( <100> Offset angle 2°, Si doped, concentration 1×10 18 cm -3 Pretreatment is performed to ensure that epitaxial growth is carried out on a "clean substrate".

[0043] Process steps: The substrate is ultrasonically cleaned sequentially with acetone and isopropanol for 10 minutes (to remove organic contaminants), rinsed with deionized water, and then immersed in 10% HF solution for 30 seconds (to remove the natural oxide layer on the surface). Finally, it is dried with nitrogen. The substrate is then placed in the MOCVD reaction chamber and annealed at 800°C for 10 minutes in an H2 atmosphere to further remove residual oxides and adsorbed impurities from the surface.

[0044] The purpose of substrate pretreatment is to reduce the dislocation density of the epitaxial layer (target <10). 4 cm -2 This ensures lattice matching of subsequent semiconductor layers (lattice mismatch rate between GaAs substrate and AlGaAs <0.1%).

[0045] S12, N-type DBR layer growth; Process steps: Metal-organic chemical vapor deposition (MOCVD) is used. The growth temperature is 700℃ (high temperature is beneficial to reduce interface roughness), and the pressure is maintained at 200 Torr. The carrier gas is H2. Group III reactive gas sources (TMGa, TMAl) and Group V sources (AsH3) are used, with SiH4 as the doping source. Trimethylaluminum (TMAl, Al source) and TMGa are introduced. The Al composition is controlled to 0.9 by adjusting the TMAl / TMGa flow ratio (set to 9:1) to form Al. 0.93 Ga 0.07 As (high Al layer), with a refractive index of approximately 3.0; TMAl is shut off, and only TMGa is introduced, forming Al. 0.15 Ga 0.85 As (low-Al layer), refractive index approximately 3.6; alternating growth in 39±1 cycles, with each layer thickness strictly controlled to 1 / 4 of the laser wavelength (when λ=850nm, Al...). 0.93 Ga 0.07 The As layer is approximately 70 nm thick, and the Al layer... 0.15 Ga 0.85 The As layer thickness is approximately 58 nm, with an error of <1 nm; N-type doping is formed (Si concentration is approximately 5 × 10⁻⁶). 17 ±10%cm -3 ), to ensure good conductivity.

[0046] The function of N-type DBR is to maximize reflectivity by maximizing the refractive index difference between the high-Al and low-aluminum components; the thickness of each layer, λ / 4n (where n is the refractive index of the material), ensures that the reflected light from each interface is superimposed in phase, forming a highly reflective "optical mirror".

[0047] S13, Lower isolation layer growth; Process steps: Metal-organic chemical vapor deposition (MOCVD) is used, with the temperature reduced to 680℃ and the pressure to 200 Torr. TMAl, TMGa, and AsH3 are introduced, with the TMAl / TMGa flow ratio set to 9:1 (consistent with the high Al layer of N-type DBR), to form Al. 0.93 Ga 0.07 As, the doping source is diethylberyllium (DEBe, a p-type doping source), with a flux of 8 sccm, resulting in a Be doping concentration of approximately 1 × 10⁻⁶. 18 cm -3 P-type high-Al composition AlGaAs; thickness 80nm (error ±2nm).

[0048] Function of the lower isolation layer: The lower isolation layer is located between the N-type DBR and the active region. Its function is to isolate the active region from the N-type DBR, and at the same time, it can assist in optical field confinement. The wide bandgap characteristics of high Al composition materials can reduce the absorption of photons in the active region, and low P-type doping can reduce the nonradiative recombination of charge carriers in the spacer layer. S14, Active region (multiple quantum well) growth; Process steps: Using metal-organic chemical vapor deposition (MOCVD), the growth temperature is reduced to 600℃ (low temperature can suppress In atom diffusion and ensure the flatness of the quantum well interface), and the pressure is reduced to 100 Torr (reduce gas phase reaction and improve layer thickness uniformity); trimethylindium (TMIn, In source), TMGa and AsH3 are introduced, and the In composition is controlled to 0.15 (adjusted by the TMI / (TMIn plus TMGa) flow ratio) to form In. 0.07 Ga 0.93 As (well layer) / , the bandgap corresponds to a wavelength of 850nm; then TMIn is turned off, and only TMGa and AsH3 are introduced, Al 0.3 Ga 0.7 As (barrier layer); growth of 4±1 period quantum wells (MQWs); well layer thickness 8±0.5 nm, In composition 7±0.3%; barrier layer thickness 20±1 nm, Al composition 30±0.5%.

[0049] The active region (multiple quantum wells) plays a crucial role in laser generation. Composed of InGaAs quantum wells and GaAs barriers, it generates optical gain through carrier recombination. Strict control of the well width and In composition is necessary to match the laser wavelength. The quantum confinement effect of the quantum well quantizes carrier energy, allowing only photons with specific energies (corresponding to 850nm) to be amplified. A thinner well width (8nm) enhances quantum confinement and improves the gain coefficient.

[0050] S15, Upper isolation layer growth; Process steps: Metal-organic chemical vapor deposition (MOCVD) is used, with the growth temperature raised to 680℃ and the pressure to 200 Torr; TMAl, TMGa, and AsH3 are introduced, with the TMAl / TMGa flow ratio set to 3:7, to form Al... 0.3 Ga 0.7 As (low Al content, reducing photon absorption); DEBe is introduced (flow rate 8 sccm), Be doping concentration approximately 1 × 10⁻⁶. 18 cm -3 (P-type); grown at 80nm (error ±2nm), symmetrical with the lower isolation layer to ensure symmetrical distribution of the light field in the active region.

[0051] The function of the upper isolation layer: Located between the active region and the P-type DBR, the upper isolation layer isolates the active region from the P-type DBR, reducing non-radiative recombination of holes in the DBR. The refractive index of the low-Al composition material is closer to that of the active region InGaAs, which can reduce interface reflection loss and improve the overlap between the optical field and the active region.

[0052] S16, P-type DBR growth; Process steps: Metal-organic chemical vapor deposition (MOCVD) is used, with a growth temperature of 700℃ (consistent with N-type DBR growth) and a pressure of 200 Torr; DEBe is introduced (flow rate 10 sccm) to achieve a Be doping concentration of approximately 5 × 10⁻⁶. 17 ±10%cm -3 The p-type layer (high doping ensures efficient hole transport). Alternating growth of Al 0.93 Ga 0.07 As (higher Al layer) / Al 0.15 Ga 0.85 As (low Al layer), grown with 24±1 pairs of cycles, total thickness of each pair = λ / 4; Be doping concentration 5×10⁻⁶. 17 ±10%cm -3 P-type DBR layer.

[0053] Function: The P-type DBR and N-type DBR have symmetrical structures and together form a vertical resonant cavity. The high reflectivity of the P-type DBR and N-type DBR (total >99.9%) forms a closed resonant cavity, which allows photons to be reflected back and forth in the active region, continuously gaining gain until laser light is formed.

[0054] S17, P-type contact layer growth; Process steps: Metal-organic chemical vapor deposition (MOCVD) was used at a growth temperature of 650℃ and a pressure of 200 Torr, with only TMGa and AsH3 introduced to form GaAs; the Be doping concentration was approximately 1×10⁻⁶. 19 ±10%cm -3 Heavy doping of P + Layer; thickness approximately 200±5nm.

[0055] Function of P-type contact layer: Heavily doped P + The layer can eliminate the Schottky barrier at the metal-semiconductor contact, forming a low-resistance ohmic contact (contact resistance < 1 × 10⁻⁶). -4 Ω·cm 2 ).

[0056] S2, preparation of carbon substitution modulation structure layer; The specific technical solution is as follows: S21. Pre-treatment: Ultra-high cleanliness substrate pretreatment (inhibiting carbon cluster nucleation sites).

[0057] Process steps: S211. Using a two-step method of "diamond grinding (0.1um) followed by colloidal silica polishing", P + The surface roughness of the contact layer (GaAs, 200nm) is controlled to Ra≤0.3nm, and AFM full-sheet scanning (5×5um² area) is used to ensure that there are no protrusions ≥1nm (to avoid becoming the nucleation center of carbon clusters). S212, Deep Cleaning: Organic Cleaning: Acetone (99.9%) ultrasonic (300W, 10 minutes), followed by isopropanol ultrasonic cleaning (same power, 10 minutes) (extend the time to remove micron-sized organic residues). S213, Oxide layer removal: Soak in diluted HF (1:100, concentration 0.4%) for 5 minutes (low concentration, slow corrosion, avoid surface damage), then rinse 5 times with DIW. S214. Ultimate removal of carbon pollution: O2 / Ar plasma ashing (power 200W, O2:Ar=1:1, flow rate 100sccm, 5 minutes), surface carbon content detected by XPS is ≤0.05% (conventional solution is 0.1%).

[0058] Function and Principle: The nucleation of carbon clusters depends on "surface defects or impurity sites." An ultra-smooth surface with Ra ≤ 0.3 nm can reduce nucleation centers by 90%. Residual carbon impurities (such as surface-adsorbed CO2) combine with C atoms from CCl4 decomposition to form clusters. Deep cleaning reduces the cluster density from 1 × 10⁻⁶. 6 cm -2 Reduced to 1×10 4 cm -2 .

[0059] S22, preparation of carbon substitution modulation structure layer.

[0060] Dielectric mask fabrication process steps: S221, PECVD deposition of 200nm SiN X (50nm thicker than conventional film; too thin and it is easily corroded by epitaxial gas, too thick and it causes stress concentration at the window edge), Parameters: SiH4 / NH3=1:6, temperature 350℃ (higher temperature increases film density), stress control ≤8MPa (avoid cracking during high-temperature epitaxy); CF4 / O2 (5:1).

[0061] S222, Dry etching (power 120W, 10mTorr): Photolithography plus dry etching defines a circular window with a diameter of 3-4µm (perfectly matching the high-intensity region of the Gaussian distribution of the light field, with a lateral accuracy of ±0.1µm and no lateral expansion problem caused by zinc diffusion); after etching, immerse in BOE (buffered oxide etchant) for 10 seconds to remove SiN from the window edges. X Residue (to ensure the GaAs surface is fully exposed and avoid carbon doping shadowing effect).

[0062] Function and principle: Subsequent carbon doping epitaxial temperature reaches 600℃, thickening SiN XThe mask can enhance high temperature resistance and prevent the mask from shrinking at high temperatures, which would cause window deformation (diameter deviation ≤0.1um); residue at the window edge can block the diffusion of C atoms, resulting in a 20% decrease in local doping concentration. BOE treatment can ensure uniform doping at the edge.

[0063] Carbon-doped selective epitaxial growth (MOCVD parameter optimization).

[0064] Process steps: S223. Using Metal-Organic Chemical Vapor Deposition (MOCVD), growth temperature: 600℃ (±1℃) (high temperature promotes the diffusion of C atoms in the GaAs lattice, reducing local enrichment (cluster density reduced by 50%), but not exceeding 610℃ (to avoid GaAs decomposition)). Growth pressure: 200 mTorr (±10 mTorr) (high pressure extends the residence time of the vapor precursor in the reaction chamber (from 0.5s to 1s), making the C atom distribution more uniform (in-wafer doping uniformity ≤7%)). V / III ratio (As / Ga): 60 (±2) (a high As atmosphere can suppress Ga atom surface migration, reducing "Ga vacancies" (C atoms easily occupy Ga vacancies to form clusters), reducing cluster size from 5nm to below 2nm). Doping source flux: CCl4 (CCl4 decomposes into CCl2 at 600℃). + and Cl - CCl2 + Provides C atoms). Corresponding doping concentration 5 × 10⁻⁶ 18 -8×10 18 cm -3 To avoid exceeding the solid solubility. Growth thickness: 350-550nm (±5nm) (add 50nm redundancy: the resistivity of the carbon substitution modulation structure layer is slightly higher (1.5×10⁻⁶)). -3 Ω·cm), thickening to compensate for resistance loss); Real-time monitoring: 1. In-situ infrared spectroscopy (IR) monitoring: 1500cm -1 CCl4 characteristic peak intensity (threshold: ≤0.1au, if exceeded, immediately reduce CCl4 flow rate); 2. Thickness measurement with laser interferometer: Ensure thickness deviation ≤ 5nm (to avoid resistance differences caused by uneven thickness).

[0065] Function and principle: C atoms form acceptor levels (ionization energy 26 meV) by "substituting Ga sites", providing hole conduction. The substitution rate is ≥90% at 600℃ (higher than 70% at low temperatures). The synergistic effect of high V / III ratio and high temperature - As atoms occupy surface adsorption sites, reducing C atom aggregation. At the same time, high temperature promotes the migration of C atoms in the lattice, dispersing local high concentration areas. The solid solubility limitation of carbon leads to a lower doping concentration than Zn, and a slightly higher resistivity. By increasing the thickness by 50nm (resistance is proportional to thickness), the total resistance is kept consistent without affecting the bandwidth.

[0066] S23. Post-processing optimization (defect control for carbon substitution modulation structure layer). Low-temperature annealing eliminates shallow-level defects. Process steps: Rapid thermal annealing (RTA) in S231 and N2 atmosphere, temperature 400℃, time 30 seconds (heating rate 50℃ / s).

[0067] Function and principle: The carbon substitution modulation structure layer may contain a small number of "C interstitial atoms" (shallow energy level defects, leading to carrier traps). Annealing at 400℃ can promote the migration of interstitial C to Ga vacancies (substitution sites), increasing the activation rate from 90% to 95%.

[0068] Epitaxial layer edge planarization etching process steps: S232. Dry etching is performed using a Cl2 / BCl3 mixed gas (volume ratio 3:1) at an etching power of 50W, a pressure of 10mTorr, a flow rate of 30sccm, and an etching depth of 5-10nm (monitored by a profilometer). This removes the tiny protrusions (height ≤10nm) at the edge of the SEG epitaxial layer window, ensuring the epitaxial layer surface is flush with the SiN. x The mask surface is flush (step height ≤ 5nm).

[0069] Function and Principle: During SEG growth, tiny protrusions (5-10 nm in height) may form at the window edge due to "gas phase supersaturation." If left untreated, this can lead to "excessively thick edge electrodes" (current concentrated at the edge) during subsequent electrode deposition, causing localized carrier accumulation. By using a light dry etching process (removing only 5-10 nm), a "step-free" surface can be achieved, ensuring uniform electrode contact (current density deviation ≤5%).

[0070] Explanation of the design principle of the carbon substitution modulation structure layer: Selective epitaxial growth combined with carbon substitution doping, through "physical mask defining the epitaxial region and precise carbon atom lattice occupancy," achieves three core functions: current limiting, SHB suppression, and single-mode output. The following is a detailed explanation of the design principle: This patented solution achieves dual current confinement through a nested synergy of "SEG carbon-doped pore size plus oxide pore size": First-level current limiting: SEG carbon-doped pore size (Wc=3-4um): through SiN X A 3-4µm circular window is etched using photolithography, and a carbon-doped GaAs epitaxial layer (low-resistivity region, 1.2 × 10⁻⁶) is grown only within the window. -3 -1.5×10 -3 Ω·cm), outside the window is SiN X High-resistivity region covered (resistivity ≥ 1×10⁻⁶) 3 (Ω·cm), the current is initially confined within the 3-4 μm carbon substitution modulation structure layer; Secondary current limiting: Oxide pore size (Wo = 6-7 μm): After the growth of the carbon substitution modulation structure layer, the current limiting is applied to the underlying AlGaAs layer (e.g., Al...). 0.98 Ga 0.02 As undergoes wet oxygen oxidation to form an oxide pore size of 6-7 μm (Al2O3 insulating region) – the oxide pore size covers the carbon doped pore size (3-4 μm), further limiting the injected current; Advantages in current limiting accuracy: The carbon doped pore size is determined by the photolithography accuracy (±0.1um), while the oxide pore size achieves an accuracy of ±0.2um by controlling the oxidation time (e.g., 30min). The nesting deviation between the two is ≤0.3um, resulting in more precise current limiting.

[0071] Suppression of spatial hole burning (SHB) effect: Precise matching of current injection and optical field: The optical field of the active region of VCSEL exhibits a Gaussian distribution, with 80% of the light intensity concentrated in the central region of 3-4µm (within the 6-7µm oxide aperture). The photon density is high at the center (fast carrier consumption) and low at the edges (slow carrier consumption). The carbon-doped aperture (3-4µm) of the SEG is directly aligned with the central region of the optical field, and the current is injected only into the central region where carrier consumption is fastest. Furthermore, the uniformity of carrier concentration in the carbon substitution modulation structure layer is ≤7%, and the current is uniformly distributed in the central region—avoiding the problem of "large current injection area leading to carrier accumulation at the edges" in traditional schemes. The carrier injection rate is perfectly matched with the photon consumption rate in the central region. The low-frequency roll-off (≤1GHz) caused by SHB originates from the "supplementary delay" after the central carriers are depleted. Due to the current-optical field matching degree of ≥95%, the low-frequency modulation response flatness is improved by 15%, and the RIN is stabilized at -149dB / Hz.

[0072] Flattening of the optical field distribution: The synergistic effect of the oxide aperture (6-7µm) and the carbon-doped aperture (3-4µm) makes the optical field distribution flatter. The 6-7µm oxide aperture, through the difference in refractive index (Al2O3 refractive index 1.7, GaAs refractive index 3.6), initially confines the optical field within the oxide aperture, preventing the optical field from leaking to the edges; the 3-4µm carbon substitution modulation structure layer (high carrier concentration) has a slight absorption on the optical field, which can weaken the local high intensity at the center of the optical field, reducing the gradient of the optical field intensity from the center to the edge from 20% to 10%, reducing the difference in photon density—the carrier consumption is more uniform, there is no local "hole burning", and it provides support for a wide bandwidth greater than 30GHz.

[0073] The dual-aperture optical field is synergistically reduced, resulting in more stable mode selection and enabling single-mode output. The solution achieves more stable single-mode output by "initial constraint of the oxide aperture and further reduction of the optical field by carbon doping": precise control of the lateral size of the optical field, first-order optical field constraint: oxide aperture (6-7µm): the oxide aperture confines the optical field to the range of 6-7µm through total internal reflection, initially suppressing large-size high-order modes (such as TEM). 20 TEM 11 Second-order light field reduction: carbon-doped aperture (3-4µm): the high carrier concentration of the carbon substitution modulation structure layer has a "mode filtering" effect on the light field—only the fundamental mode (TEM) 00 It can transmit stably in narrow channels of 3-4µm, and higher-order modes (such as TEM) can also transmit stably. 01 TEM 10 Because the width of the optical field distribution exceeds the carbon-doped aperture, it will be absorbed or leaked by the carbon substitution modulation structure layer (higher-order mode loss ≥20dB / cm, fundamental mode loss ≤5dB / cm); when the lateral dimension of the optical field (3-4um) is 3.5-4.7 times the working wavelength (850nm), the mode gain of the higher-order mode is lower than the threshold gain (fundamental mode gain ≥15dB, higher-order mode gain ≤8dB), and only the fundamental mode achieves resonant output.

[0074] Single-mode stability advantages: No mode shift: The low diffusion characteristics of carbon ensure long-term stability of carbon-doped pore size (deviation ≤0.1um after 1000 hours of operation), and there is no shift due to optical field confinement; while the Zn diffusion scheme has a diffusion pore size shift ≥0.3um due to Zn atom migration, and the single-mode performance deteriorates. High SMSR output: The solution achieves a stable fundamental mode suppression ratio of 38-40dB and a vertical divergence angle of ≤14° through dual aperture synergy, resulting in more stable single-mode output.

[0075] The remaining fabrication steps are the same as those of traditional laser fabrication processes, so a simplified description is provided.

[0076] S3. A layer of SiN is grown using plasma-enhanced chemical vapor deposition (PECVD). X Low-stress SiN X A passivation protective film is applied to the device surface for passivation protection and optical modulation; S4. Remove the dielectric film above the current injection area in step 4 by dry etching process, and prepare a P-type metal contact layer (seed gold) by magnetron sputtering. S5. The etching of the oxide trench is completed by dry etching, followed by wet oxidation, and the oxide pore size is controlled at 6-7um. S6. Use benzocyclobutene as a dielectric material to fill the positive and negative electrode areas of the chip to reduce the parasitic parameters of the chip itself. S7. Use ALD process to passivate the chip device area, protect the internal structure of the chip, and improve chip reliability. S8. Using vapor deposition process, the front-side N-type electrode will be grown. S9, grinding and thinning; S10, RTA treatment (annealing).

[0077] The above are all preferred embodiments of this application, and are not intended to limit the scope of protection of this application. Therefore, all equivalent changes made in accordance with the structure, shape and principle of this application should be covered within the scope of protection of this application.

Claims

1. A vertical cavity surface-emitting semiconductor laser, comprising, from bottom to top, a substrate (1), an N-type DBR layer (2), an active region (3), an oxide layer (4), a P-type DBR layer (5), a P-contact layer (6), and a P-metal layer (7), wherein oxide holes (8) are formed on the oxide layer (4), characterized in that: The P contact layer (6) is provided with a carbon substitution modulation structure layer (9). The lower end of the carbon substitution modulation structure layer (9) is located in the P-type DBR layer (5). The carbon substitution modulation structure layer (9) is directly opposite the oxide hole (8). The material of the carbon substitution modulation structure layer (9) is GaAs doped with carbon atoms. The diameter of the carbon substitution modulation structure layer (9) is smaller than the pore size of the oxide hole (8).

2. A vertical-cavity surface-emitting semiconductor laser according to claim 1, characterized in that: The carbon atom concentration in the carbon substitution modulation structure layer (9) is 1×10 19 -1×10 20 cm -3 .

3. A method for manufacturing a vertical-cavity surface-emitting semiconductor laser, characterized in that: The method for fabricating a vertical-cavity surface-emitting semiconductor laser as described in any one of claims 1-2 includes the following steps: S1. Epitaxial growth: An epitaxial layer is grown on the substrate. The epitaxial layer includes an N-type DBR layer, an active region, an oxide layer, a P-type DBR layer, a P-contact layer, and a P-metal layer. S2, Preparation of carbon substitution modulation structure layer; S2 Includes the following steps, S221, Deposit a SiN layer on the P contact layer. X , S222, Photolithography plus dry etching in SiN X A circular window matching the high-intensity region of the Gaussian distribution of the light field is etched out at the center; S223. Carbon atom doping is performed using MOCVD, with CCl4 as the doping source to provide carbon atoms to be doped on the P contact layer through a circular window, forming a carbon substitution modulation structure layer.

4. The method for manufacturing a vertical-cavity surface-emitting semiconductor laser according to claim 3, characterized in that: In S223, the growth temperature is 600-610℃; the growth pressure is 200 mTorr; and the As / Ga ratio is 60.

5. The method for manufacturing a vertical-cavity surface-emitting semiconductor laser according to claim 3, characterized in that: S2 also includes the steps preceding S221. S211. A two-step method of diamond grinding and colloidal silica polishing is adopted to control the surface roughness of the p contact layer to Ra≤0.3nm. AFM full-wafer scanning is used to ensure that the protrusions in the 5×5um² area are ≤1nm. S212. Clean with acetone and isopropanol for 10 minutes each. S213. Soak in diluted HF for 5 minutes, then rinse with DIW. S214. Ashing was performed using O2 / Ar plasma, and the surface carbon content was determined by XPS to be ≤0.05%.

6. The method for manufacturing a vertical-cavity surface-emitting semiconductor laser according to claim 3, characterized in that: During the S223 process, the intensity of the C-C bond characteristic peaks was monitored in real time using in-situ infrared spectroscopy.

7. The method for manufacturing a vertical-cavity surface-emitting semiconductor laser according to claim 3, characterized in that: It also includes the steps following S223. S231. Rapid thermal annealing in a N2 atmosphere at 400°C for 30 seconds; S232. Dry etching is performed using a Cl2 / BCl3 mixed gas to remove the 5-10nm high protrusions at the edge of the SEG epitaxial layer window, making the epitaxial layer surface flush with the SiN. X The mask surface is flush.

8. A method for manufacturing a vertical-cavity surface-emitting semiconductor laser according to claim 3, characterized in that: In step S222, after etching out the circular window, it is immersed in BOE to remove the SiN at the edge of the circular window. X Residue.

Images