Quantum well tunable laser and method of fabrication

Abstract

The application discloses a kind of quantum well tunable laser and preparation method, belong to semiconductor laser technical field, laser includes the fixed gain cavity and channel selection cavity of V-shaped position distribution, and the half-wave coupler of coupling phase is 180 °;Its quantum well active layer adopts sub-band micro-well structure, by a layer of In 0.15 Ga 0.85 As center well and the GaAs barrier layer of two sides symmetric arrangement, In 0.15 Ga 0.85 As thin well and GaAs waveguide layer constitute.The application is based on GaAs material system, by optimizing In component and symmetric sub-band micro-well design, reduce lattice mismatch strain, enhance carrier confinement and radiation recombination efficiency.The laser preparation method includes waveguide production, front electrode sputtering, deep etching surface production and back electrode sputtering step, using dry-wet method combination etching and planarization process.The application realizes that central wavelength 940nm, tuning range reaches 18.22nm, average side mode suppression ratio 34.5dB High-performance tunable output, simple and compact structure, without grating production and secondary epitaxial growth.

Description

Technical Field

[0001] This invention belongs to the field of semiconductor laser technology, specifically relating to a quantum well tunable laser and its fabrication method. Background Technology

[0002] Gallium arsenide (GaAs)-based semiconductor lasers are among the earliest types of semiconductor lasers researched. Their emission wavelength range covers 760 nm to 1060 nm, and they have been widely used in numerous fields such as biosensing, high-speed communication, optical measurement, fiber optic sensing, and weaponry. As a direct bandgap semiconductor material, gallium arsenide has a refractive index of 3.3 at room temperature, and its lattice constant is highly matched with that of AlAs. The ternary compound AlGaAs formed by GaAs can achieve lattice self-matching under any composition, greatly facilitating the design and growth of layered epitaxial structures. In particular, the emission wavelength of this material system falls precisely within the biosafety detection range. This wavelength has good penetration into human tissue, and the absorption peak of water is low, effectively ignoring the interference of water in biological tissues during detection. Furthermore, the small size and ease of integration of semiconductor lasers help reduce the size of detection equipment, showing promising application prospects in the field of portable medical testing.

[0003] Tunable semiconductor lasers based on GaAs materials have attracted much attention since the 1980s due to their advantages in coherent detection. Their output wavelength can be continuously varied within a certain range, and they typically operate in a single longitudinal mode. They offer advantages such as small size, low cost, and integration with other devices, making them promising candidates for applications in the sensing field. Current applications include TDLAS (Transient Digital Laser Assay) detection of gases such as O2, CO2, NH3, CH4, and H2S; solution composition and concentration detection when combined with microfluidic technology; liquid flow velocity detection using the laser Doppler effect; and distance and velocity measurement using frequency-modulated continuous wave technology.

[0004] Among existing tunable semiconductor laser technologies, external cavity lasers (ECLs) use gratings or etalons as external mode selection devices, offering advantages such as wide tuning range and narrow linewidth, and are widely used in coherent communication and high-precision detection. However, their tuning speed is slow, system size is large, accuracy requirements are high, and stability is poor, making large-scale adoption difficult in applications with strict cost and size constraints, such as access networks and sensor networks. Vertical cavity surface-emitting lasers (VCSELs) employ a vertical output structure with cavity lengths on the order of wavelength, achieving good single-mode characteristics, but their output power is low (maximum approximately 6mW). Tunable VCSELs with cantilever structures typically have output power below 1mW, making it difficult to meet the needs of sensing applications. Distributed feedback (DFB) lasers achieve wavelength variation through temperature tuning, with a tuning range of only about 5nm, and performance degrades significantly with increasing temperature. While DFB arrays can extend the tuning range, the integration of multiple grating periods increases fabrication complexity and manufacturing costs. While distributed Bragg reflector (DBR) lasers and sampled grating DBR (SGDBR) lasers can achieve a large tuning range (approximately 10 nm), they require complex grating fabrication processes and multiple tuning electrodes, and the current control algorithms are complex, making fabrication and packaging difficult.

[0005] The V-shaped coupled-cavity (VCL) semiconductor laser, first proposed by Zhejiang University in 2008, is primarily based on InP material and is mainly used in optical communication. This structure consists of two Fabry-Perot cavities with different optical lengths forming a V-shaped composite coupled cavity and a half-wave coupler, achieving dynamic wavelength tuning based on the vernier effect. Compared to the aforementioned lasers, the VCL structure is simpler and more compact, and its fabrication process is essentially the same as that of FP lasers, eliminating the need for complex processes such as grating fabrication and secondary epitaxial growth. It holds great promise for sensor and communication network applications requiring low-cost, high-performance tunable lasers. However, previous InP-based solutions have mainly been optimized for the communication band, and their application potential in the near-infrared band has not yet been fully explored. Summary of the Invention

[0006] To address the aforementioned technical problems, this invention proposes a quantum well tunable laser, comprising: The two resonant cavities are arranged in a V-shape, namely a fixed-gain cavity and a channel selection cavity; A half-wave coupler is used to select the output mode of the two resonant cavities to achieve single longitudinal mode output; And an epitaxial wafer, used to carry two resonant cavities and a half-wave coupler; The epitaxial wafer includes a quantum well active layer, which adopts a sub-band microwell structure. The fixed gain cavity is used to provide a fixed gain and determine the channel spacing. The channel selection cavity changes the refractive index by injecting current or temperature changes to achieve wavelength tuning. There is a length difference between the two resonant cavities to form a vernier effect to amplify the tuning range. The subband microwell structure consists of a layer of In 0.15 Ga 0.85 As serves as the central well, with GaAs barrier layers and In layers arranged symmetrically on both sides of the central well. 0.15 Ga 0.85 As thin well and the outermost GaAs waveguide layer.

[0007] Furthermore, three electrodes are respectively covered above the two resonant cavities and the half-wave coupler, and the electrodes are separated by isolation grooves; wherein, in addition to covering the half-wave coupler, the coupler electrodes also cover a portion of the two resonant cavities, and the proportion of the covered portions is the same.

[0008] Furthermore, the front and rear ends of the two resonant cavities and the end of the half-wave coupler are respectively provided with three deep etching regions to form etched mirror surfaces as the reflecting surfaces of the laser; wherein, the deep etching surface at the end of the channel selection cavity serves as the light-emitting surface of the laser, and the other two reflecting surfaces are coated with high-reflectivity films.

[0009] Furthermore, the epitaxial wafer of the laser has a vertical structure comprising, from top to bottom, a contact layer, an upper cladding layer, a confinement layer, a quantum well active layer, a lower cladding layer, and a substrate; the upper and lower cladding layers are composed of AlGaAs material; the AlGaAs in the upper cladding layer has a layer of highly p-type doped GaAs for forming an ohmic contact with the electrode metal; the AlGaAs in the lower cladding layer has a layer of highly n-type doped GaAs below it; and an InGaP wet resist layer is disposed below the confinement layer.

[0010] Furthermore, the length difference between the two resonant cavities is 5%-20%, which allows the channel spacing determined by the fixed gain cavity to work together with the vernier effect to achieve a wavelength tuning range of not less than 15nm.

[0011] This invention also proposes a method for fabricating a quantum well tunable laser, comprising the following steps: S1. Using photoresist as a mask, the epitaxial wafer containing the sub-band microwell structure is etched to fabricate the waveguide and coupler parts; the etching adopts a combination of dry etching and wet etching to control the etching depth of the waveguide and ensure the flatness of the etched surface. S2. Planarize the epitaxial wafer on the fabricated waveguide to ensure the continuity of the subsequent sputtering of the front electrode, and then sputter the front electrode. S3. Use photoresist as a mask to etch a deep etched area that serves as the reflective surface; S4. Etch out the electrical isolation area between each electrode, and then thin and polish the back of the laser before sputtering the back electrode.

[0012] Furthermore, the epitaxial wafer includes an InGaP layer, which serves as an etch stop layer for the wet etching process.

[0013] Furthermore, in step S2, planarization is performed by spin-coating a dielectric layer; the sputtered front electrode is a multilayer metal structure containing an adhesion layer, a barrier layer and a conductive layer, wherein the adhesion layer is used to form an ohmic contact with the semiconductor and the barrier layer is used to prevent atomic diffusion of the conductive layer.

[0014] Furthermore, in step S3, a deep etched surface at the channel selection cavity end is used as the light-emitting surface, and a high-reflectivity film is deposited on other reflective surfaces.

[0015] Furthermore, in step S4, the photolithography process employs a multi-segment spin coating and a stepped post-baking process to form a photoresist mask of uniform thickness on a surface with a deep etched morphology.

[0016] Compared with the prior art, the present invention has the following beneficial technical effects: 1. A quantum well epitaxial structure with a center wavelength of 940 nm based on GaAs material is proposed. A sub-band micro-well design is used to replace the traditional structure. Through In composition optimization and symmetric layout design, the introduction of strain is effectively reduced, the carrier transition probability is increased, and local strain is decreased, achieving fine-grained control of electronic energy levels. Simulation results show that three symmetrical potential wells can efficiently confine electrons and holes within In... 0.15 Ga 0.85 Within the As well layer, the simulated bandgap E_PL=1.32eV corresponds to a PL wavelength of 939.90nm, which is highly consistent with the design target of 940nm.

[0017] 2. A 940nm tunable semiconductor laser based on a V-cavity structure and GaAs material system was successfully fabricated, and key fabrication parameters were optimized. A single laser can achieve 28 tunable channels, with a tuning range from 930.53nm to 948.75nm, achieving a wide tuning range of 18.22nm, far exceeding the approximately 5nm tuning range of DFB lasers, and also compensating for the limitation of conventional VCSELs with a tuning range of only 4.8nm. The average side-mode suppression ratio is 34.5dB, with performance comparable to the high-stability output of ECLs and SGDBRs, without relying on complex external mode selection or multi-electrode control.

[0018] 3. Unlike ECL lasers, which rely on complex structures involving discrete device assembly and μm-level optical path alignment, and which require the nanoscale grating etching and secondary epitaxial growth processes necessary for DFB, DBR, and SGDBR, the laser structure of this invention is simpler and more compact. The manufacturing process is essentially the same as that of FP lasers, without involving complex processes such as grating fabrication and secondary epitaxial growth, making it easy to integrate with other devices. Compared to existing laser designs with similar structures, this invention expands the design scheme to a 940nm center wavelength, and is expected to be applied in the sensing field as a low-cost light source for biological detection systems. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of a V-cavity semiconductor laser. Figure 2 Diagram of the conduction band structure; Figure 3 This is a diagram of the price band structure. Figure 4 This is a full-band structure diagram; Figure 5 This is a diagram of the conduction band structure of a traditional single quantum well; Figure 6 This is a diagram of the valence band structure of a traditional single quantum well. Figure 7 This is a full-band structure diagram of a traditional single quantum well; Figure 8 It is a layered structure of waveguide in the vertical direction; Figure 9 The wavelength of the laser varies with the channel-selective cavity current under different conditions; Figure 10 In the diagram, (a) is the full-channel spectrum of the laser, and (b) shows the variation of the side-mode suppression ratio for each channel. Detailed Implementation

[0020] Example 1 The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] The laser of this invention comprises two resonant cavities arranged in a V-shape. The fixed-gain cavity has a length of 217 μm, and the channel selection cavity has a length of 238.7 μm, with a 10% length difference between the two cavities. Coupling is achieved through a half-wave coupler with a 180° coupling phase. This coupler is used to select the laser's output mode, achieving high side-mode suppression ratio (SMSR) while maintaining single-longitudinal-mode output. The shorter of the two resonant cavities primarily provides the fixed gain of the laser and determines the channel spacing to be 0.6 nm when tuning the laser wavelength. The other resonant cavity, called the channel selector cavity, has its refractive index altered by changes in injected current or temperature, and is used to switch between different output wavelengths during wavelength tuning. The wavelength tuning range is amplified by a vernier effect created by the length difference between the two resonant cavities. The optimal structural parameters for the half-wave coupler are a waveguide spacing gap of 1.7 μm and a half-wave coupler length Lc of 45 μm.

[0022] The main structure of this type of laser is as follows: Figure 1 As shown, three electrodes, separated by isolation slots, cover the two resonant cavities and the half-wave coupler, respectively, to provide the current required for laser gain. The coupler electrode provides the main gain required by the laser; it covers a portion of both resonant cavities in addition to the half-wave coupler, with the covered portions being proportionally equal. Typically, the injection current of the channel selection cavity electrode is changed to achieve laser output wavelength selection, while the injection current of the fixed-gain cavity remains constant. However, theoretically, both electrodes can achieve wavelength selection. The ground electrode shared by the entire laser is located on the back side. Three deep etching zones are located at the front and rear ends of the two resonant cavities and at the half-wave coupler end, forming etched mirror surfaces as the laser's reflecting surfaces. The deep etched surface at the channel selection cavity end serves as the laser's output surface, while the other two reflecting surfaces are coated with high-reflectivity films made of titanium and gold. The direction of the deep etching surface at the end of the half-wave coupler needs to be parallel to the wafer's main plane, and the curvature of the two resonant cavities also needs to be consistent during design. This ensures that the losses introduced by process errors during laser fabrication remain consistent, positively impacting laser performance. The area of ​​the deep etched surface at the end of the half-wave coupler is larger than that of the other two deep etched surfaces. Applying a high-reflectivity film here will be more effective in reducing laser loss than applying a high-reflectivity film to the other two deep etched surfaces. Therefore, choosing the channel selection cavity as the output surface can improve the performance of the laser.

[0023] The epitaxial layer structure design of a laser significantly impacts its optical and electrical properties in the vertical direction, such as lasing wavelength, refractive index distribution, and carrier concentration distribution. From an energy engineering perspective, two semiconductor materials with different bandgap widths should be selected as the active layer of the laser. Epitaxial growth must be carried out in a cavity pressure of 8.86E-8. The GaAs substrate is pretreated with a deoxidation temperature controlled at 730℃, a reconstruction temperature of 560℃, and a subsequent growth temperature stabilized at 690℃. During epitaxial growth, the material should be grown sequentially in the order of wide bandgap, narrow bandgap, and wide bandgap again. The deposition rate of GaAs is 1.1 A / s, and the deposition rate of AlGaAs is 1.5 A / s. The wide bandgap material serves as the barrier layer, and the narrow bandgap material serves as the quantum well layer. In this quantum well structure, the narrow bandgap material is at a lower energy level, and the wide bandgap material is at a higher energy level, thereby confining carriers and generating quantum size effects. This means that electrons in the well wall direction lose their degree of freedom in the direction perpendicular to the heterojunction plane. At this point, the charge carriers will exhibit a quantized distribution, and only a small current injection is needed to achieve population inversion.

[0024] First, materials need to be selected based on the design target of 940nm lasing wavelength. Based on the quantum well material properties and subband control requirements, In... x Ga 1-x InGaAs, a ternary compound, is used as the quantum well material for the active layer. The emission wavelength of the quantum well is determined by the transition energy between the first conduction band level and the first valence band level. This transition energy depends on the band gap structure of InGaAs, and the material properties of InGaAs directly determine the band gap energy of the quantum well. The thickness of the quantum well affects the position of the energy levels within the well. At the same time, the compressive strain caused by the lattice mismatch between InGaAs and GaAs further modifies the band gap and valence band structure of the quantum well. Therefore, the In composition and the quantum well thickness are key parameters for controlling the lasing wavelength, optical gain, and threshold current density.

[0025] This invention employs a subband microwell structure as the active layer, which consists of a 4.4 nm thick In layer. 0.15 Ga 0.85 As serves as the central well, with 1.3 nm thick GaAs barrier layers and 1.3 nm thick In layers arranged symmetrically on both sides. 0.15 Ga 0.85 A thin GaAs well as an outermost 4nm thick GaAs waveguide layer, through symmetry optimization in bandgap engineering, introduces an ultrathin barrier layer to finely modulate the potential energy profile, thereby forming a multi-subband energy level layout within a single recombination potential well. This achieves spatial expansion and overlap enhancement of the carrier wavefunction, significantly improving radiative recombination efficiency. Compared to ordinary In... 0.17 Ga 0.83Based on the As single quantum well structure and this multi-subband quantum well structure, while maintaining the target 940nm lasing wavelength, this invention optimizes the In composition from 0.17 to 0.15. This adjustment brings key performance improvements in reducing lattice mismatch strain, decreasing nonradiative recombination, and increasing carrier transition probability. First, the introduction of strain is reduced. The lattice mismatch between the InGaAs quantum well and the GaAs substrate generates strain, as shown in the lattice mismatch formula (1-1): (1-1) Where f is the lattice mismatch degree. x is the In component. , When the In composition x=0.17, the lattice mismatch f=1.218%; when x=0.15, f decreases to 1.075%, and the absolute value of compressive strain decreases by about 11.8%. This directly reduces the defect density introduced by lattice mismatch and significantly suppresses nonradiative recombination.

[0026] Secondly, it increases the carrier transition probability. Compressive strain causes the valence band to split, leading to the separation of the heavy hole band and the light hole band, and this valence band subband splitting energy increases with increasing strain. The valence band subband splitting energy is negatively correlated with the carrier transition probability; a decrease in splitting energy increases the wavefunction overlap between heavy holes and electrons. When the In composition changes from 0.17 to 0.15, the decrease in strain reduces the splitting energy by approximately 12%, and the wavefunction overlap between heavy holes and electrons increases accordingly. Simultaneously, the reduced In composition increases the band gap, moving the carriers further away from the nonradiative recombination centers of deep-level defects, collectively significantly increasing the radiative transition probability of carriers.

[0027] Unlike the single-order valence band subbands of traditional single potential wells, a third-order valence band subband layout was constructed through band engineering design of symmetrical subband micro-wells. This was achieved using a GaAs barrier layer and In... 0.15 The bandgap difference in the GaAs potential well layer forms three quantum potential wells of uniform depth in the valence band, corresponding to three quantized heavy hole subbands. This three-order subband layout efficiently confines heavy holes within the potential wells, providing ample carrier reserves for high-gain lasing.

[0028] Finally, the local strain was reduced. Due to the constraint of the GaAs substrate, the quantum well experiences compressive strain ε along the growth direction Z-axis. zz : (1-2) Where ε zz To induce compressive strain along the Z-axis of the quantum well growth direction, the elastic constant C of InGaAs... 11 =122GPa, C 12 =57 GPa. When x=0.17, εzz =-0.777%; when x=0.15, ε zz =-0.685%, the decrease in In composition reduces the absolute value of local elastic strain by about 11.8%. The strain magnitude of the quantum well directly affects the integrity of the subband structure. Reducing the strain can reduce nonradiative recombination centers caused by lattice defects, allowing injected charge carriers to be converted into laser output more efficiently.

[0029] To verify the band structure characteristics of the optimized design, the present invention performed simulation calculations on the conduction band, valence band, and full band structure of the above-mentioned subband micro-well structure.

[0030] like Figure 2 As shown, the conduction band exhibits a typical quantum well "plateau-barrier" structure. Figure 2 The black lines represent the conduction band barrier structure, i.e., how the conduction band energy levels change with spatial position. The concave region represents the area below the quantum well, the blue curve represents the ground state electronic energy level, and the upper blue curve represents the first excited state. The high-energy regions at both ends correspond to the GaAs barrier layer, and the flat region in the middle corresponds to In. 0.15 The GaAs potential well layer embodies the quantized confinement of electrons in the Z-direction, forming discrete sub-band energy levels. The flatness of the energy plateau indicates a quasi-continuous energy state distribution of electrons within the potential well, while the steep energy rise at the potential barrier effectively prevents lateral leakage of charge carriers. This band structure efficiently confines electrons to specific sub-bands within the potential well, avoiding energy loss due to electron overflow, enhancing the radiative recombination efficiency of charge carriers, and providing a band structure foundation for achieving low threshold current and high output power.

[0031] like Figure 3 As shown, the black lines represent the variation of valence band energy levels with spatial location, the bulges correspond to hole potential wells, the red curves represent heavy hole wave functions, the blue curves represent light hole wave functions, and the high-energy region in the middle of the valence band structure diagram corresponds to In. 0.15 The GaAs potential well layer, with its low-energy regions at both ends corresponding to the GaAs potential barrier layer, exhibits a complementary energy distribution of "high in the middle and low at both ends" to the conduction band structure. The significant rise in valence band energy within the potential well region indicates that holes are strongly quantized, forming discrete valence band subbands. The relatively gentle band curvature in the potential well region and the steep energy drop in the potential barrier region demonstrate effective hole confinement, corresponding to the electron distribution region in the conduction band, further enhancing the electron-hole recombination efficiency.

[0032] like Figure 4 As shown, the full-band structure diagram comprehensively illustrates the overall band profile of the entire subband micro-trap structure.

[0033] Figure 4 The meanings of the curves from top to bottom: The thick red solid line represents the distribution of conduction band energy levels. The dipped portion corresponds to the electron quantum well, where electrons are confined to these low-energy regions.

[0034] Red dashed line: indicates the spatial distribution of electrons in the quantum well.

[0035] Blue dashed lines: indicate the distribution of holes in the quantum well.

[0036] The thin red solid line represents the valence band potential well structure, and the raised region corresponds to the heavy hole quantum well.

[0037] The thin blue solid line represents the valence band potential well structure, and the raised region corresponds to the light hole quantum well.

[0038] Blue step lines: Hole discrete energy levels, representing valence band bound state energy levels formed by quantum confinement.

[0039] The figure shows symmetrical and deep potential wells in the conduction and valence bands, visually demonstrating that electrons and holes are efficiently confined within a specified In band. 0.15 Within the GaAs well layer, discrete quantized energy levels are formed. This strong confinement effect is the basis for achieving high population inversion efficiency and low threshold current.

[0040] To further illustrate the advantages of the sub-band microwell structure of this invention, it is now compared and analyzed with the traditional single quantum well structure.

[0041] like Figure 5 As shown, the conduction band structure diagram of a conventional single quantum well illustrates the conventional In... 0.17 The normalized potential profile of GaAs single quantum wells shows a single, shallow potential well structure with limited carrier confinement capability. Figure 5 The black lines in the diagram represent the potential barrier structure of the conduction band, i.e., how the conduction band energy levels change with spatial position, and the concave regions represent quantum wells; the blue curve below represents the ground state electronic energy level, and the blue curve above represents the first excited state.

[0042] like Figure 6 As shown, the valence band of a traditional single quantum well exhibits a single potential energy profile of "high in the middle and low at both ends," which can only form a single subband layout. This results in insufficient confinement of heavy holes, making it easy for holes to diffuse into the barrier region, and limiting the carrier reserve. Figure 6 The black lines represent the changes in valence band energy levels with spatial position, the raised parts correspond to hole potential wells, and the blue curves represent hole wave functions.

[0043] like Figure 7 As shown, the full-band structure diagram of a traditional single quantum well exhibits a single narrow peak with low peak intensity and a wide half-width at half-maximum, indicating that its radiative recombination efficiency is limited and its energy level uniformity is poor.

[0044] Figure 7The meanings of the curves from top to bottom: The thick red solid line represents the distribution of conduction band energy levels. The dipped portion corresponds to the electron quantum well, where electrons are confined to these low-energy regions.

[0045] Red dashed line: indicates the spatial distribution of electrons in the quantum well.

[0046] Blue dashed lines: Reflect the distribution of holes in the quantum well.

[0047] The thin blue solid line represents the valence band potential well structure, and the raised region corresponds to the hole quantum well.

[0048] Blue step lines: Hole discrete energy levels, representing valence band bound state energy levels formed by quantum confinement.

[0049] In contrast, the sub-band micro-well structure used in this invention significantly enhances the spatial confinement of charge carriers and the overlap of wave functions through a symmetrical third-order potential well layout, achieving more efficient radiative recombination and a lower threshold current.

[0050] Based on this band structure, the band gap E in the simulation results is calculated. PL The PL wavelength of 1.32 eV is 939.90 nm, which is highly consistent with the design target of 940 nm. This confirms the superiority of the subband micro-well structure in band engineering and spectral performance, and provides a key theoretical basis for subsequent epitaxial growth and device fabrication.

[0051] Based on the above analysis, the final designed structure of the laser epitaxial wafer in the vertical direction is as follows: Figure 8 As shown. From top to bottom, it includes a contact layer, an upper cladding layer, an active layer, a lower cladding layer, and a substrate.

[0052] In 0.15 GaAs has the highest refractive index and the smallest band gap in the material system. At room temperature, near a wavelength of 940 nm, the refractive index of GaAs is 3.5518, while that of In... 0.15 The refractive index of GaAs is 3.8126. Furthermore, since the laser designed in this paper requires TE-mode lasing, compressive strain was introduced into the quantum well. This strain was used to control the splitting of the valence band and subband, optimizing the gain characteristics. The upper cladding (also called the waveguide layer) is composed of AlGaAs material. This is because AlAs and GaAs have the same lattice constant, and the ternary compound AlGaAs formed by them matches the substrate lattice, making material growth very convenient. The band gap width and refractive index respectively confine the charge carriers and improve the optical confinement factor. At room temperature, Al... x Ga 1-x The relationship between the refractive index of As material and the component ratio x can be calculated using equation (1-3): (1-3) Among them, when hour, ;when hour, As the Al content increases, Al... x Ga 1-x The refractive index of As gradually decreases.

[0053] Unlike other epitaxial structures, this invention incorporates a layer of highly p-doped GaAs in the upper AlGaAs cladding to form an ohmic contact with the electrode metal. For the same reason, a layer of highly n-doped GaAs is added below the lower AlGaAs cladding. Below the confinement layer, a 10nm thick InGaP wet resist layer is added to precisely control the waveguide etching depth during the etching process. To prevent light leakage to the contact layer, the upper cladding needs a certain thickness to separate the contact layer from the active layer and minimize losses. However, an excessively thick upper cladding complicates the manufacturing process and increases the total resistance of the laser. Therefore, considering various factors, the layered laser structure designed in this invention uses an upper cladding thickness of 1.55μm, a lower cladding thickness of 2.85μm, and a contact layer thickness of 0.15μm.

[0054] Example 2 The fabrication process of V-cavity semiconductor lasers made of GaAs / AlGaAs materials consists of four stages: waveguide fabrication, front electrode sputtering, deep etching fabrication, and back electrode sputtering.

[0055] Waveguide fabrication: First, photoresist is used as a mask to etch the waveguide and coupler parts of the laser. The contact layer and upper cladding need to be etched away to a depth of approximately 1.7 μm. The waveguide fabrication stage includes three steps: surface pretreatment, photolithography, and etching. For surface pretreatment, the epitaxial wafer is ultrasonically cleaned sequentially in acetone, methanol, and isopropanol for 6 minutes each, rinsed with deionized water, dried with nitrogen, immersed in HF buffer for 30 seconds to remove surface oxides, and finally rinsed with deionized water and dried. For photolithography, an ABM contact UV lithography machine is used. First, positive photoresist RZJ304 is spin-coated, with the thickness controlled to approximately 1.1–1.3 μm using an optimized three-stage spin-coating program. The epitaxial wafer is then baked at 105°C for 40 seconds to enhance photoresist adhesion and remove solvent. Next, exposure is performed at 320W for 4.2 seconds, followed by development with DPD-200 developer for 28 seconds to dissolve the photoresist in unexposed areas. To further increase the photoresist's density and make it more robust, thus better protecting the underlying structure, post-baking involves placing the epitaxial wafer on a heated plate and slowly raising the temperature from 60°C to 90°C. After post-baking, the epitaxial wafer is cleaned using the low-power oxygen cleaning function of a dry etching machine to remove residual photoresist. The next step is to etch the areas not covered by photoresist to obtain a waveguide shape on the epitaxial wafer that is exactly the same as that on the photomask. This invention adopts a combined dry and wet etching scheme. A 0.01 μm thick InGaP layer is designed as a resist layer in the layered structure of the epitaxial wafer. In the experiment, dry etching was first used to etch the waveguide depth to about 1.6 μm, and then wet etching was used to make the etching depth as close as possible to 1.7 μm. At the same time, the solution ratio of the wet etching was carefully selected. The wet etching formula is shown in Table 1. It can etch the AlGaAs layer while having almost no reaction with GaAs and InGaP, and the etching rate is relatively slow, which makes it easier to control the etching depth.

[0056] Table 1

[0057] Front Electrode Sputtering: Compared to most semiconductor processing methods, this invention adds a planarization step before electrode sputtering. Benzocyclobutene (BCB) is spin-coated as a filler, and the shape of the front electrode is etched using a negative photoresist KMPE3130A as a mask. The front electrode of the laser is fabricated using magnetron sputtering, with Ti, Pt, and Au selected as the constituent electrode metal materials. During fabrication, Ti is sputtered first to form an ohmic contact with P+-doped GaAs, and it also exhibits good adhesion to GaAs. Next, Pt is sputtered, which prevents Au from diffusing into the epitaxial wafer. Finally, Au is sputtered on the top layer, as its good conductivity allows it to directly contact the probe during testing. After sputtering the electrode metal, the unused portions of the electrode are stripped away.

[0058] Deep etching fabrication: Using positive photoresist RZJ304 as a mask, the deep etching area serving as the reflective surface is etched. In the previous steps, this area is protected by photoresist, and the etching depth is between 2.6-3μm. Back electrode sputtering: The electrical isolation region between the three electrodes is etched, thus completing the fabrication of the front part of the laser. Before sputtering the back electrodes, the back side of the epitaxial wafer needs to be thinned and polished. Then, metal is sputtered and annealed. Finally, the laser is cleaved according to requirements.

[0059] Example 3 Taking the design and fabrication process of a laser with a center wavelength of 940nm as an example 1. Quantum well epitaxial wafer design: Select the materials for each layer according to the target center wavelength and process conditions, and optimize the composition, band structure and strain of the quantum well layer.

[0060] (1) Identify the In component: ternary compound In x Ga 1-x As the performance parameters of semiconductor materials can be calculated by equivalently representing two binary compounds: (1-4) P represents the parameter to be calculated, In x Ga 1-x The x-value in As material is related to the lasing wavelength.

[0061] The relationship between the In component and the gain wavelength was calculated using simulation software, and a linear formula was obtained through fitting. (1-5) Substituting the target wavelength of 940nm, the In composition is determined to be 0.15.

[0062] (2) Band structure and energy level calculation Using the effective mass approximation model, the quantized energy levels of the conduction band and valence band are calculated: (1-6) The effective mass associated with carriers in the barrier and well layers is m b * and m W * E indicates i v0 corresponds to the eigenvalues ​​related to carrier energy and potential well depth. z The thickness of the potential well layer, To reduce Planck's constant, substituting it into the calculation yields the energy level E of the first subband of the conduction band. c1 ≈0.035 eV, first subband energy level of the valence band E v1 ≈ 0.12 eV, energy level difference E g,QW =E c1 -E v1 The value is approximately 1.32 eV, corresponding to a PL wavelength of λ = 1240 / 1.32 ≈ 939.9 nm, which is highly consistent with the 940 nm target.

[0063] (3) Lattice mismatch and strain calculation The lattice mismatch between the InGaAs quantum well and the GaAs substrate will produce strain, according to the lattice mismatch formula: (1-7) Where f is the lattice mismatch degree and the lattice constant is... x is the In component. , When the In composition x = 0.17, the lattice mismatch f = 1.218%; when x = 0.15, f decreases to 1.075%.

[0064] Due to the constraint of the GaAs substrate, the quantum well experiences compressive strain along the Z-axis of the growth direction. According to the elastic strain formula: (1-8) Where ε zz To induce compressive strain along the Z-axis of the quantum well growth direction, the elastic constant C of InGaAs... 11 =122GPa, C 12 =57 GPa. When x=0.17, ε zz =-0.777%; when x=0.15, ε zz = -0.685%, the absolute value of local elastic strain decreased by about 11.8%.

[0065] 2. Waveguide Design: (1) Determine the waveguide height and width: Based on the designed epitaxial layer structure, determine the etching depth of the waveguide, input the material parameters of each layer in the simulation software, calculate the effective refractive index value and mode distribution at the center wavelength, and determine the etching depth to be 1.7 μm and the waveguide width to be 3 μm.

[0066] (2) Determine the parameters of the half-wave coupler: The threshold gain difference and normalized intensity cross-coupling coefficient between the master mode and the side modes are calculated using the transfer matrix method (TMM) through Matlab software programming. The number of tunable channels and the side mode suppression ratio are calculated under different cavity length differences to obtain the cavity length and coupler coupling coefficient when the performance is optimal. A complete half-wave coupler model is established in the simulation software, the coupling coefficient is set, and the relationship between the half-wave coupler length Lc, the gap between the two resonant cavities Gap, the normalized intensity cross-coupling coefficient X, and the coupling phase is obtained through simulation calculation to determine the final parameters of the half-wave coupler.

[0067] (3) Determine the polynomial parameters of the bent waveguide: In this section, the two cavities adopt a symmetrical structure. The cavity length difference is adjusted by adding a straight waveguide with an oblique direction after the curved waveguide, while the curved sections are identical. The polynomial curved waveguide can be defined by four conditions: the starting direction, the ending direction, and the projected distances in the X and Y directions between the starting and ending points. Considering both the smoothness of the curved waveguide and the computational complexity, this invention uses a fifth-order polynomial to realize the curve. The commercial software Rsoft is used with the Beam Propagation Method (BPM) to perform a two-dimensional scan and obtain the optimal parameters. This invention sets the projected distances in the X and Y directions between the starting and ending points to 10 μm and 160 μm, respectively, with the ending axis direction deflected by 10° relative to the starting point. Then, the fitted univariate fifth-order polynomial can accurately represent the fifth-order normalized curve of the curved waveguide, and the expression is: (1-9) Next, the beam propagation method is used in Rsoft software to simulate the structure of the curved waveguide. By optimizing the parameter values ​​of the polynomial, these two variables are scanned. and By monitoring the transmitted optical power, the transmission loss of the curved waveguide can be minimized.

[0068] 3. Preparation process: This invention employs a combined dry-wet etching approach. A 0.01 μm thick InGaP layer is designed as a resist layer within the layered structure of the epitaxial wafer. In the experiment, dry etching was first used to etch the waveguide depth to approximately 1.6 μm, followed by wet etching to bring the etching depth as close to 1.7 μm as possible. The solution ratio for the wet etching was carefully selected; the wet etching formulation is shown in Table 2. This ensures that the solution can etch the AlGaAs layer while exhibiting minimal reaction with GaAs and InGaP, and that the etching rate is relatively slow, allowing for better control of the etching depth.

[0069] Table 2

[0070] 4. Test results of the prepared laser: The optimized design of the subband microwell structure enables the laser to achieve wide-range wavelength tuning through a combination of carrier injection and thermal effects. A single laser can achieve 28 channels, with wavelengths tuned from 930.53 nm to 948.75 nm, a total width of 18 nm, and an average side-mode rejection ratio of over 34.45 dB across all channels.

[0071] like Figure 9 As shown, the laser has 8 switchable channels at 15℃, 16 switchable channels at 10℃, and 7 switchable channels at 20℃.

[0072] By superimposing the emission spectra of the laser under different conditions, the complete wavelength tuning range of the GaAs V-cavity semiconductor laser designed in this invention can be obtained, as shown below. Figure 10 As shown, coverage of 28 channels was ultimately achieved, and the average side-mode rejection ratio of all channels remained above 34.45dB.

[0073] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

[0074] The above embodiments are merely illustrative examples and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.

Claims

1. A quantum well tunable laser, characterized in that, include: The two resonant cavities are arranged in a V-shape, namely a fixed-gain cavity and a channel selection cavity; A half-wave coupler is used to select the output mode of the two resonant cavities to achieve single longitudinal mode output; And an epitaxial wafer, used to carry two resonant cavities and a half-wave coupler; The epitaxial wafer includes a quantum well active layer, which adopts a sub-band microwell structure. The fixed gain cavity is used to provide a fixed gain and determine the channel spacing. The channel selection cavity changes the refractive index by injecting current or temperature changes to achieve wavelength tuning. There is a length difference between the two resonant cavities to form a vernier effect to amplify the tuning range. The subband microwell structure consists of a 4.4 nm thick In layer. 0.15 Ga 0.85 As serves as the central well, with 1.3 nm thick GaAs barrier layers and 1.3 nm thick In layers symmetrically arranged on both sides of the central well. 0.15 Ga 0.85 A thin As well and an outermost 4nm thick GaAs waveguide layer were constructed; a third-order valence band subband layout was built, utilizing the GaAs barrier layer and In 0.15 The band gap difference in the GaAs potential well layer forms three quantum potential wells of uniform depth in the valence band, corresponding to three quantized heavy hole subbands.

2. The quantum well tunable laser according to claim 1, characterized in that, Three electrodes are respectively covering the top of the two resonant cavities and the half-wave coupler, and the electrodes are separated by isolation grooves; wherein, in addition to covering the half-wave coupler, the coupler electrodes also cover a portion of the two resonant cavities, and the proportion of the covered portions is the same.

3. The quantum well tunable laser according to claim 1, characterized in that, The front and rear ends of the two resonant cavities and the end of the half-wave coupler are respectively provided with three deep etching regions to form etched mirror surfaces as the reflecting surfaces of the laser; wherein, the deep etching surface at the end of the channel selection cavity serves as the light-emitting surface of the laser, and the other two reflecting surfaces are coated with high-reflectivity films.

4. The quantum well tunable laser according to claim 1, characterized in that, The epitaxial wafer of the laser has a vertical structure comprising, from top to bottom, a contact layer, an upper cladding layer, a confinement layer, a quantum well active layer, a lower cladding layer, and a substrate. The upper and lower cladding layers are composed of AlGaAs material. The upper cladding layer contains a layer of highly p-doped GaAs for forming an ohmic contact with the electrode metal. Below the AlGaAs in the lower cladding layer is a layer of highly n-doped GaAs. Below the confinement layer is an InGaP wet resist layer.

5. The quantum well tunable laser according to claim 1, characterized in that, The length difference between the two resonant cavities is 5%-20%, which allows the channel spacing determined by the fixed gain cavity to work together with the vernier effect to achieve a wavelength tuning range of not less than 15nm.

6. A method for fabricating a quantum well tunable laser as described in any one of claims 1-5, characterized in that, Includes the following steps: S1. Using photoresist as a mask, the epitaxial wafer containing the sub-band microwell structure is etched to fabricate the waveguide and coupler portions; The etching process employs a combination of dry and wet etching methods to control the etching depth of the waveguide and ensure the flatness of the etched surface. S2. Planarize the epitaxial wafer on the fabricated waveguide to ensure the continuity of the subsequent sputtering of the front electrode, and then sputter the front electrode. S3. Use photoresist as a mask to etch a deep etched area that serves as the reflective surface; S4. Etch out the electrical isolation area between each electrode, and then thin and polish the back of the laser before sputtering the back electrode.

7. The preparation method according to claim 6, characterized in that, The epitaxial wafer includes an InGaP layer, which serves as the etching stop layer for the wet etching process.

8. The preparation method according to claim 6, characterized in that, In step S2, planarization is performed by spin-coating a dielectric layer; the sputtered front electrode is a multilayer metal structure containing an adhesion layer, a barrier layer and a conductive layer, wherein the adhesion layer is used to form an ohmic contact with the semiconductor and the barrier layer is used to prevent atomic diffusion of the conductive layer.

9. The preparation method according to claim 6, characterized in that, In step S3, a deep etched surface at the channel selection cavity end is used as the light-emitting surface, and a high-reflectivity film is deposited on other reflective surfaces.

10. The preparation method according to claim 6, characterized in that, In step S4, the photolithography process employs a multi-segment spin coating and a stepped post-baking process to form a photoresist mask of uniform thickness on a surface with a deep etched morphology.

Images