An InP-based optical communication laser high-speed modulation stability improvement process and laser

Abstract

This invention provides a process for improving the high-speed modulation stability of InP-based optical communication lasers and the laser itself. Relating to the field of laser fabrication technology, the process involves the following steps: epitaxial wafer fabrication, formation of a distributed feedback grating and a quarter-wavelength phase shift region, etching of a ridged main active strip, opening of narrow injection windows on both sides, dual-energy proton injection, rapid thermal annealing, formation of an Fe-doped InP semi-insulating buried current-limiting structure, and fabrication of coplanar waveguide electrodes. This constructs carrier relief corridors with a longitudinally distributed central region, transition region, and end region on both sides of the main active strip. This structure allows excess carriers in the main active region to enter the deep-level defect region through the uninjected connection region during high-speed direct modulation and undergo non-radiative recombination, reducing the spatial peak value of carriers, decreasing refractive index disturbances and output optical power fluctuations, and improving high-speed modulation stability. It is suitable for high-speed optical communication transmitting devices while also ensuring single-mode output and wide-temperature operating reliability.

Description

Technical Field

[0001] This invention relates to the field of laser technology, specifically to a process for improving the high-speed modulation stability of InP-based optical communication lasers and the laser itself. Background Technology

[0002] InP-based optical communication lasers, due to their material system advantages suitable for operation in the 1.31-micron and 1.55-micron bands, have been widely used in data center interconnects, access network transmissions, mobile communication fronthauls, and short-to-medium distance single-mode fiber optic links. In existing applications, to meet the requirements of high speed, low cost, and miniaturization, directly modulated distributed feedback lasers have become one of the important solutions. These devices typically achieve single-longitudinal-mode output through distributed feedback gratings, current constraint through ridge waveguides and embedded current-limiting structures, and high-bandwidth radio frequency drive through high-speed electrodes.

[0003] However, existing InP-based direct-modulation lasers still suffer from insufficient stability under high-speed modulation conditions. This is because the main active region is prone to excessive carrier accumulation under high-speed rise times, large modulation swings, and high-temperature bias conditions, leading to significant peaks in carrier spatial distribution. This, in turn, causes dynamic spatial hole burning, rapid local refractive index swings, decreased mode stability, and increased output power fluctuations. Particularly near the quarter-wavelength phase shift region, due to the high photon density, the local carrier consumption and replenishment processes are more uneven, easily causing eye diagram deterioration, decreased side-mode suppression ratio, increased chirp, and degradation of modulation performance under high-temperature conditions. Existing technologies typically improve high-speed performance by optimizing the grating coupling coefficient, reducing parasitic capacitance, improving heat dissipation structures, or enhancing embedded current limiting capabilities. However, these methods mainly optimize the static structural parameters or external parasitic parameters of the device, making it difficult to directly and actively manage the excess carriers within the main active region during high-speed modulation. Therefore, under high bit rates, large swings, and wide operating temperatures, the device still easily exhibits insufficient dynamic stability. Based on this, it is necessary to propose a high-speed modulation stability enhancement process for InP-based optical communication lasers that can selectively laterally extract and non-radiative recombine excess carriers during high-speed modulation without significantly damaging the gain of the main active region, so as to improve the high-speed modulation stability, single-mode output capability and wide-temperature operating reliability of the device in practical communication systems. Summary of the Invention

[0004] Technical problems to be solved

[0005] To address the shortcomings of existing technologies, this invention provides a process for improving the high-speed modulation stability of InP-based optical communication lasers and a laser itself, thus solving the problems of existing technologies.

[0006] Technical solution

[0007] To achieve the above objectives, the present invention provides a high-speed modulation stability improvement process for InP-based optical communication lasers, the process comprising the following steps:

[0008] Sp1. Provides an epitaxial wafer for an InP-based laser, wherein the epitaxial wafer is provided from bottom to top as follows: an n-type InP substrate, an n-type InP lower cladding layer, a lower waveguide confinement layer, a multi-quantum-well active layer, an upper waveguide confinement layer, a p-type InP upper cladding layer, and a p-type contact layer;

[0009] Sp2, A distributed feedback grating is formed on the epitaxial wafer, and a quarter-wavelength phase shift region is formed in the distributed feedback grating;

[0010] Sp3, etched along the length of the laser cavity to form a ridge-shaped main active strip;

[0011] Sp4, forming a narrow injection window extending along the length of the cavity in the shoulder region on both sides of the ridge-shaped main active strip;

[0012] Sp5. Proton injection is performed on both sides of the ridge-shaped main active strip through the narrow injection window to form carrier pressure relief corridors on both sides of the ridge-shaped main active strip. Deep energy level defect regions are provided in the carrier pressure relief corridors, and uninjected connection regions are provided between the deep energy level defect regions and the main active regions.

[0013] Sp6. Annealing is performed on the structure after proton injection to fix the position and defect density of the deep level defect region;

[0014] Sp7, An embedded current limiting structure is formed on both sides of the ridge-shaped main active strip;

[0015] Sp8, A coplanar waveguide electrode structure is formed on the p-type contact layer;

[0016] The deep-level defect region is provided with a central region, a transition region, and an end region along the cavity length of the laser. The central region is located in the quarter-wavelength phase shift region and the cavity length segments on both sides. The defect density of the central region is greater than the defect density of the transition region, and the defect density of the transition region is greater than the defect density of the end region. When the laser is in a high-speed direct modulation state, the excess carriers in the main active region enter the carrier depressurization corridor through lateral diffusion and local electric field drive and undergo non-radiative recombination, thereby reducing the carrier swing in the main active region and improving the high-speed modulation stability.

[0017] Preferably, the lower waveguide confinement layer and the upper waveguide confinement layer are both separate confinement heterostructure waveguide confinement layers, the multi-quantum well active layer is located between the lower waveguide confinement layer and the upper waveguide confinement layer, the deep energy level defect region is located within the upper waveguide confinement layer and the p-type InP upper cladding, and the uninjected connection region is formed by the continuous portion of the upper waveguide confinement layer on both sides of the ridge-shaped main active stripe.

[0018] Preferably, the proton injection in Sp5 consists of a first proton injection and a second proton injection. The first proton injection forms a first defect layer in the p-type InP cladding, and the second proton injection forms a second defect layer in the upper waveguide confinement layer. The first defect layer and the second defect layer are connected in the vertical direction to form the deep energy level defect region, and no defect peak region is formed in the multi-quantum well active layer.

[0019] Preferably, the length of the central region along the cavity length direction accounts for 20% to 50% of the total cavity length of the laser, the total length of the transition region along the cavity length direction accounts for 20% to 40% of the total cavity length of the laser, and the total length of the end region along the cavity length direction accounts for 20% to 40% of the total cavity length of the laser.

[0020] Preferably, the annealing process in Sp6 is rapid thermal annealing, with an annealing temperature of 350 to 500 degrees Celsius and an annealing time of 10 to 120 seconds.

[0021] Preferably, the embedded current limiting structure in Sp7 is formed using Fe-doped InP semi-insulating material, and the embedded current limiting structure is located on both sides of the ridge-shaped main active strip.

[0022] Preferably, the lateral width of the uninjected connection region is 0.1 micrometer to 0.8 micrometers, the lateral width of the carrier pressure relief corridor is 0.5 micrometers to 3 micrometers, and the defect density of the deep level defect region and the lateral width of the uninjected connection region together define the lateral carrier pressure relief path from the main active region to the carrier pressure relief corridor.

[0023] Preferably, the laser fabricated using the high-speed modulation stability improvement process for an InP-based optical communication laser comprises an n-type InP substrate, an n-type InP lower cladding, a lower waveguide confinement layer, a multi-quantum-well active layer, an upper waveguide confinement layer, a p-type InP upper cladding, a p-type contact layer, a distributed feedback grating, a quarter-wavelength phase shift region, a ridge-shaped main active stripe, a carrier pressure relief corridor, a buried current confinement structure, and a coplanar waveguide electrode structure.

[0024] The distributed feedback grating and the quarter-wavelength phase shift region are located in the waveguide region corresponding to the multi-quantum-well active layer;

[0025] The ridge-shaped main active strip extends along the length of the laser cavity;

[0026] The charge carrier pressure relief corridor is located on both sides of the ridge-shaped main active strip;

[0027] The carrier pressure relief corridor is equipped with a deep energy level defect region;

[0028] An uninjected connection region is provided between the deep-level defect region and the main active region;

[0029] The deep energy level defect region is provided with a central region, a transition region, and an end region along the length of the laser cavity;

[0030] The central region is located in the quarter-wavelength phase shift region and the cavity length segments on both sides;

[0031] The defect density in the central region is greater than the defect density in the transition region;

[0032] The defect density in the transition region is greater than the defect density in the end region.

[0033] Preferably, the deep-level defect region is located within the upper waveguide confinement layer and the p-type InP upper cladding, the uninjected connection region is formed by the continuous portion of the upper waveguide confinement layer on both sides of the ridge-shaped main active strip, and the embedded current confinement structure is formed on the outside of the carrier relief corridor using Fe-doped InP semi-insulating material.

[0034] Preferably, the coplanar waveguide electrode structure is disposed above the p-type contact layer. The carrier depressurization corridor performs lateral extraction and non-radiative recombination of excess carriers in the main active region under high-speed direct modulation, thereby reducing the peak value of the spatial distribution of carriers in the main active region and reducing the fluctuation of output optical power during high-speed modulation.

[0035] Beneficial effects

[0036] This invention provides a process for improving the high-speed modulation stability of InP-based optical communication lasers and the laser itself. It offers the following advantages:

[0037] 1. This invention sets up carrier pressure relief corridors on both sides of the ridge-shaped main active strip and divides the deep level defect region into a central region, a transition region, and an end region along the cavity length. This allows the excess carriers generated in the main active region during high-speed direct modulation to diffuse laterally in a controlled manner along the uninjected connection region and undergo non-radiative recombination in the deep level defect region. This preferentially reduces the carrier peak near the phase shift region, effectively suppressing dynamic spatial hole burning, reducing transient refractive index swing, reducing output optical power overshoot and fluctuation, and improving the stability of the high-speed modulation eye diagram.

[0038] 2. This invention employs a combination of dual-energy proton injection and rapid thermal annealing to construct a vertically connected deep-level defect region within the upper waveguide confinement layer and the p-type InP cladding. Simultaneously, it prevents the formation of defect peak regions in the main region of the multi-quantum-well active layer. Therefore, it can form an effective pressure relief channel for excess carriers while maintaining the stimulated emission capability and high side-mode suppression ratio of the main active region, thus balancing high-speed modulation performance and laser output performance. It has strong fabrication feasibility.

[0039] 3. This invention integrates the carrier relief corridor, the Fe-doped InP semi-insulating embedded current limiting structure, and the coplanar waveguide electrode structure in a coordinated design, which spatially separates the main injection current, the lateral carrier relief path, and the parasitic current blocking path. This not only reduces lateral leakage and parasitic effects, but also helps maintain a high modulation bandwidth, low power drift, and good wide-temperature operating stability in the range of 10 Gbps to 25.8 Gbps. It is suitable for practical applications in data center optical modules, mobile communication fronthaul modules, and industrial communication optical transmitter components. Attached Figure Description

[0040] Figure 1 This is the InP-based epitaxial wafer structure of the present invention;

[0041] Figure 2 This is a schematic diagram of the carrier pressure relief corridor and gradient defect region of the present invention;

[0042] Figure 3 This is a schematic diagram of the laser structure and carrier decompression principle of the present invention;

[0043] Figure 4 This is a process flow diagram of the present invention;

[0044] Figure 5 This is a schematic diagram of the overall architecture of the present invention;

[0045] Figure 6 This is a schematic diagram of the device layer architecture of the present invention;

[0046] Figure 7 This is a schematic diagram of the electric drive and control architecture of the present invention;

[0047] Figure 8 This is a schematic diagram of the optical packaging architecture of the present invention. Detailed Implementation

[0048] 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. Specific Implementation Example 1:

[0050] like Figures 1 to 8As shown, a high-speed modulation stability improvement process for InP-based optical communication lasers is presented. The process consists of the following steps: The process is designed for direct modulation lasers in the 1.31-micron or 1.55-micron bands for optical communication. The device structure adopts a single-section distributed feedback laser structure combined with an embedded heterogeneous current-limiting structure to ensure single-mode emission, low parasitic capacitance, and high modulation bandwidth. The entire process is completed using a standardized workflow including metal-organic chemical vapor deposition epitaxy, grating definition, ridge etching, selective proton injection, rapid thermal annealing, selective regrowth of semi-insulating materials, electrode metallization, and high-speed packaging and testing. Sp1 provides an InP-based laser epitaxial wafer, which, from bottom to top, comprises an n-type InP substrate, an n-type InP lower cladding, a lower waveguide confinement layer, a multi-quantum-well active layer, an upper waveguide confinement layer, a p-type InP upper cladding, and a p-type contact layer. The n-type InP substrate has a thickness of 100 to 150 micrometers and a doping concentration of 5 x 10^17 to 2 x 10^18 per cubic centimeter. The n-type InP lower cladding has a thickness of 1.2 to 2.5 micrometers and a doping concentration of 1 x 10^17 to 1 x 10^18 per cubic centimeter, used to provide longitudinal optical confinement and longitudinal current conduction. The lower waveguide confinement layer has a thickness of 100 to 200 nanometers and a refractive index higher than the upper and lower claddings to form the main optical mode confinement region. The multi-quantum-well active layer... It consists of five to eight layers of compressive strain quantum wells and six to nine layers of barrier layers. The quantum well material is InGaAsP or InGaAlAs, with a single well thickness of six to eight nanometers and a barrier layer thickness of eight to twelve nanometers. The light emission center wavelength of the quantum well is matched with the target communication wavelength. The upper waveguide confinement layer has a thickness of one hundred to two hundred and fifty nanometers. The p-type InP cladding layer has a thickness of one to two micrometers and a doping concentration of five to ten-seventeen per cubic centimeter to three to ten-eighteen per cubic centimeter. The p-type contact layer uses a p-type InGaAs contact layer with a thickness of one hundred to two hundred and fifty nanometers and a doping concentration of one to ten-nineteen per cubic centimeter to five to ten-nineteen per cubic centimeter to reduce ohmic contact resistance and meet the requirements of high-speed injection. Sp2. A distributed feedback grating is formed on the epitaxial wafer, and a quarter-wavelength phase shift region is formed in the distributed feedback grating. The distributed feedback grating adopts a single-stage grating structure. The grating period is determined to be 190 nm to 250 nm according to the target emission wavelength and effective refractive index. The grating etching depth is set to 40 nm to 80 nm, and the coupling coefficient is set to 40 cm to 120 cm to maintain stable single longitudinal mode output. The quarter-wavelength phase shift region is preferably set at the center of the cavity length, and the position deviation is controlled within 5% of the total cavity length. The effective length of the phase shift region corresponds to the quarter-wavelength optical path, forming a local phase change point, which is used to select a single dominant mode in the longitudinal mode spectrum. The total cavity length of the laser is set to 200 μm to 600 μm, preferably 300 μm to 500 μm, to take into account the threshold current, output power and high-frequency modulation bandwidth.Sp3. Etching along the laser cavity length direction to form a ridge-shaped main active stripe; the width of the ridge-shaped main active stripe is set to 1.2 micrometers to 2.5 micrometers, preferably 1.5 micrometers to 2 micrometers; the etching method adopts dry etching combined with subsequent wet finishing, the etching depth extends from the top of the p-type contact layer downwards to the interior of the p-type InP cladding, or extends to the upper part of the upper waveguide confinement layer but does not enter the main region of the multi-quantum well active layer, so as to ensure that the main active region is not damaged by etching; the roughness of the sidewalls on both sides of the ridge-shaped main active stripe is controlled below 10 nanometers, and the sidewall tilt angle is controlled within the range of 10 degrees to 20 degrees, so as to ensure the smoothness of the subsequent regeneration interface and the stability of the transverse optical field distribution; the two ends of the ridge-shaped main active stripe are consistent with the subsequent end-face cleavage direction to ensure accurate definition of the effective cavity length. Sp4. A narrow injection window extending along the cavity length is formed in the shoulder region on both sides of the ridge-shaped main active strip. The narrow injection window is formed by photolithography and dielectric mask windowing process. The windowing dielectric layer is made of silicon nitride or silicon dioxide with a thickness of 100 nanometers to 300 nanometers. The lateral width of the narrow injection window on one side is set to 0.5 micrometers to 3 micrometers, and a 0.1 micrometer to 0.8 micrometer unwindowed area is reserved between the window and the sidewall of the ridge-shaped main active strip to form an unimplanted connection region after subsequent proton implantation. The narrow injection window runs through the entire effective cavity length along the cavity length direction, but different injection doses are designed at the central region, transition region and end region to form a vertically distributed pressure relief capability. Sp5. Proton injection is performed on both sides of the ridge-shaped main active strip through a narrow injection window, forming carrier decompression corridors on both sides of the ridge-shaped main active strip. Deep-level defect regions are located within these corridors, and uninjected connection regions are located between the deep-level defect regions and the main active region. A dual-energy proton injection process is used. The first proton injection energy is set to 80 keV to 160 keV, and the dose is set to 1 x 10^13 per square centimeter to 8 x 10^13 per square centimeter, forming a shallow defect distribution within the p-type InP cladding. The second proton injection energy is set to 180 keV to 320 keV, and the dose is set to 5 x 10^13 per square centimeter to 3 x 10^14 per square centimeter. A deep defect distribution is formed within the upper waveguide confinement layer; the damage distributions after the two injections are interconnected in the vertical direction, thus forming a deep-level defect region; the deep-level defect region undertakes the functions of trapping, recombination, and depressurization of excess carriers, and its lateral position is located outside the full width at half maximum (FWHM) of the main optical mode, avoiding becoming the main optical absorption center; the uninjected connection region is retained as a bridging region of continuous semiconductor material, and its function is to provide a confined lateral diffusion channel for excess carriers in the main active region during high-speed modulation, so that carriers enter the carrier depressurization corridor during high-level injection or rising edge overshoot, while maintaining normal stimulated emission of the main active region during low-level and steady-state bias; no defect peak region is formed in the main region of the multi-quantum-well active layer, thus avoiding a significant reduction in quantum well radiative recombination efficiency.Sp6. After proton implantation, the structure is annealed to fix the location and density of deep-level defect regions. Rapid thermal annealing in a nitrogen atmosphere is used. Before annealing, a silicon nitride protective layer with a thickness of 100 to 300 nanometers is deposited on the wafer surface to prevent phosphorus volatilization and surface roughening under high-temperature conditions. The annealing temperature is set to 350 to 500 degrees Celsius, the annealing time is set to 10 to 120 seconds, and the heating rate is set to 20 to 80 degrees Celsius per second. After annealing, sheet resistance measurement, photoluminescence mapping, and dark current testing confirm that the defect regions have been stabilized in the designed locations, and the threshold current increment of the main active region is controlled within 15% of that of the unimplanted control device. Annealing also repairs lattice damage in non-target regions caused by proton implantation, concentrating deep-level defects in the carrier relief corridor region. Sp7. Embedded current-limiting structures are formed on both sides of the ridge-shaped main active strip. The embedded current-limiting structures are formed by selective regrowth of Fe-doped InP semi-insulating material. The Fe doping concentration is set to 5 x 10^16 per cubic centimeter to 5 x 10^17 per cubic centimeter, and the regrowth thickness is set to 1 micrometer to 4 micrometers. Before regrowth, the ridge sidewalls are subjected to in-situ thermal deoxidation and low-damage surface cleaning to ensure a reduction in interface trap density. After regrowth, a conductive backfill layer is formed in the ridge top region, so that the longitudinally injected current is mainly concentrated in the ridge-shaped main active strip and diffuses to the carrier relief corridor through the uninjected connection region. The embedded current-limiting structure and the carrier relief corridor together constitute a transverse functional gradient structure, in which the ridge-shaped main active strip is used for main stimulated radiation, the uninjected connection region is used for restricted transverse carrier transport, the carrier relief corridor is used for excess carrier extraction and non-radiative recombination, and the semi-insulating embedded structure is used for current blocking and transverse parasitic suppression. Sp8. A coplanar waveguide electrode structure is formed on the p-type contact layer. The coplanar waveguide electrode structure adopts a titanium-platinum composite metal system. The signal electrode width is set to 8 to 15 micrometers, the gap between the signal electrode and the ground electrode is set to 6 to 12 micrometers, the total metal thickness is set to 0.8 to 2.5 micrometers, and the layout characteristic impedance is controlled to 45 to 55 ohms. The alignment deviation between the electrode and the center of the ridge-shaped main active strip is controlled within 0.5 micrometers to reduce additional parasitic inductance and parasitic resistance. The length of the feed line from the pad to the active area is controlled to 30 to 150 micrometers to meet the direct modulation bandwidth requirements of 10 Gb / s to 25.8 Gb / s. After the device is metallized, wafer-level RF probe testing is performed, and after cleavage coating, it enters the eutectic bonding and high-speed packaging process.The deep-level defect region is divided into a central region, a transition region, and an end region along the length of the laser cavity. The central region is located in the quarter-wavelength phase shift region and the cavity length segments on both sides. The defect density in the central region is greater than that in the transition region, and the defect density in the transition region is greater than that in the end region. When the laser is in a high-speed direct modulation state, the excess carriers in the main active region enter the carrier depressurization corridor through lateral diffusion and local electric field drive and undergo non-radiative recombination, thereby reducing the carrier swing in the main active region and improving the high-speed modulation stability. The central region, along the cavity length, corresponds to the photon density peak region and the dynamic spatial hole-burning sensitive region, with a length set to 20% to 50% of the total cavity length. The transition regions are arranged on both sides of the central region, with a total length set to 20% to 40% of the total cavity length, used to create a gradual change in pressure relief capability. The end regions are located at both ends, with a total length set to 20% to 40% of the total cavity length, used to suppress unnecessary non-radiative losses near the end faces. The three regions achieve defect density differences in the longitudinal direction through different proton injection doses, with the central region having the largest total dose, the transition region having a medium total dose, and the end region having the smallest total dose. In high-speed direct modulation, the central region preferentially extracts local excess carriers, the transition region acts as a buffer for pressure relief, and the end regions maintain low damage boundary conditions, thereby reducing the peak value of carrier spatial distribution, reducing the transient swing of refractive index, and suppressing mode drift and power overshoot.

[0051] Both the lower and upper waveguide confinement layers are separate confinement heterostructure waveguide confinement layers. The multi-quantum-well active layer is located between the lower and upper waveguide confinement layers. The deep-level defect region is located within the upper waveguide confinement layer and the p-type InP cladding. The uninjected connection region is formed by the continuous portion of the upper waveguide confinement layer on both sides of the ridge-shaped main active stripe. The lower waveguide confinement layer uses undoped or lightly doped InGaAsP material, with a bandgap wavelength set from 1 μm to 1.1 μm. The upper waveguide confinement layer also uses undoped or lightly doped InGaAsP material, with a bandgap wavelength also set from 1 μm to 1.1 μm, to provide stable longitudinal optical confinement at an operating wavelength of 1.31 μm or 1.55 μm. The deep-level defect region is located within the upper waveguide confinement layer and the p-type InP cladding, which helps to maintain physical separation between the defect region and the main quantum well emitting region in the vertical direction. The uninjected connection region is composed of the continuous portion of the upper waveguide confinement layer on both sides of the ridge-shaped main active strip. The lateral width, longitudinal continuity and material integrity of this continuous portion directly determine the response time of excess carrier diffusion into the pressure relief corridor. The response time is controlled within the range of 10 picoseconds to 200 picoseconds to match the edge of the high-speed direct modulation pulse.

[0052] In Sp5, proton injection consists of a first proton injection and a second proton injection. The first proton injection forms a first defect layer within the p-type InP cladding, and the second proton injection forms a second defect layer within the upper waveguide confinement layer. The first and second defect layers are vertically connected to form a deep-level defect region, while no defect peak region is formed within the multi-quantum-well active layer. The peak depth of the first defect layer formed by the first proton injection is 100 to 400 nanometers from the epitaxial surface, and the peak depth of the second defect layer formed by the second proton injection is 400 to 900 nanometers from the epitaxial surface. The two defects overlap through damage tails to form a continuous deep-level trap distribution. After proton injection, secondary ion mass spectrometry, cross-sectional transmission electron microscopy, and cathodoluminescence scanning are used to confirm the location of the defect layers and their relative relationship with the multi-quantum-well active layer, ensuring that the photoluminescence peak shift in the central region of the multi-quantum-well active layer does not exceed 5 nanometers and the full width at half maximum (FWHM) increment does not exceed 10%.

[0053] The central region accounts for 20% to 50% of the total cavity length of the laser, the transition region accounts for 20% to 40% of the total cavity length, and the end region accounts for 20% to 40% of the total cavity length. The central region is preferentially aligned with the quarter-wavelength phase-shift region and the high photon density regions on both sides. When the cavity length is 400 micrometers, the length of the central region is set to 100 to 180 micrometers. The transition region is located on both sides of the central region, with a length of 40 to 100 micrometers on each side. The end regions are located at both ends of the cavity, with a length of 40 to 80 micrometers on each side. Precise control of the longitudinal defect distribution in the central, transition, and end regions is achieved through independent masking or multiple exposure dose adjustments, with the longitudinal position error controlled within 5 micrometers to ensure that the pressure relief intensity corresponds to the photon density distribution within the cavity.

[0054] The annealing process in Sp6 employs rapid thermal annealing at temperatures ranging from 350°C to 500°C for 10 to 120 seconds. Rapid thermal annealing is performed under a nitrogen protective atmosphere. A protective layer is deposited before annealing, and removed after annealing, followed by re-passivation of the surface. After annealing, threshold current, side-mode suppression ratio (SMRR), dark current, differential resistance, and carrier lifetime are used as screening criteria. The threshold current is controlled within the range of 5 mA to 25 mA, the SMRR is maintained above 35 dB, the dark current is not higher than 1.5 times that of the unimplanted control device, and the differential resistance is maintained within the range of 2 ohms to 10 ohms. This ensures that the annealing process effectively fixes defects without compromising the main lasing performance.

[0055] The embedded current limiting structure in Sp7 is formed using Fe-doped InP semi-insulating material and is located on both sides of the ridge-shaped main active strip. The Fe-doped InP semi-insulating material is formed by selective region epitaxy, with a regrowth temperature of 580°C to 650°C, a reaction pressure of 50 mbar to 150 mbar, and an Fe doping concentration of 1 x 10^17 per cubic centimeter to 5 x 10^17 per cubic centimeter. After regrowth, a conductive p-type InP backfill layer and a p-type InGaAs contact cap layer are formed at the ridge top. The semi-insulating embedded layer performs lateral current limiting and parasitic suppression functions, while the conductive backfill layer functions as the main injection channel. With the embedded current limiting structure located on both sides of the ridge-shaped main active strip, the lateral leakage current of the device decreases and the current mainly concentrates near the main lasing region and the uninjected connection region, thereby enhancing the directionality of the pressure relief path.

[0056] The lateral width of the uninjected connection region ranges from 0.1 to 0.8 micrometers, while the lateral width of the carrier relief corridor ranges from 0.5 to 3 micrometers. The defect density of the deep-level defect region and the lateral width of the uninjected connection region together define the lateral carrier relief path from the main active region to the carrier relief corridor. When the lateral width of the uninjected connection region is 0.1 to 0.3 micrometers, the time constant for excess carriers in the main active region to enter the relief corridor is shorter, making it suitable for high-speed direct modulation above 20 gigabits per second. When the lateral width of the uninjected connection region is 0.3 to 0.8 micrometers, the threshold current rise in the main active region is smaller, making it suitable for direct modulation from 10 to 20 gigabits per second. When the lateral width of the carrier relief corridor is less than 0.5 micrometers, the deep-level defect volume is insufficient; when the lateral width is greater than 3 micrometers, optical perturbations and non-target absorption increase. Therefore, this width range is used to balance relief efficiency, mode stability, and output power.

[0057] A laser fabricated using a high-speed modulation stability enhancement process for InP-based optical communication lasers is disclosed. The InP-based optical communication laser comprises an n-type InP substrate, an n-type InP lower cladding, a lower waveguide confinement layer, a multi-quantum-well active layer, an upper waveguide confinement layer, a p-type InP upper cladding, a p-type contact layer, a distributed feedback grating, a quarter-wavelength phase shift region, a ridged main active stripe, a carrier discharge corridor, a buried current confinement structure, and a coplanar waveguide electrode structure. The laser's emission wavelength is set to 1290 nm to 1330 nm or 1530 nm to 1565 nm, the static side-mode suppression ratio is set to above 35 dB, the threshold current is set to 5 mA to 25 mA, the operating bias current is set to 1.2 to 2.5 times the threshold current, the junction capacitance is controlled to 0.1 picofarad to 0.6 picofarads, and the total series resistance is controlled to 2 ohms to 10 ohms. The feedback grating and quarter-wavelength phase shift region are located in the waveguide region corresponding to the multi-quantum-well active layer. The grating is preferably formed in the region adjacent to the waveguide confinement layer to balance coupling strength and fabrication stability. The ridge-shaped main active strip extends along the laser cavity length. The two ends of the ridge-shaped main active strip correspond to the cleaved end faces, which are composed of a combination of high-reflectivity and low-reflectivity films or a combination of two low-reflectivity films, with the end face reflectivity set to 1% to 30%. Carrier pressure relief corridors are located on both sides of the ridge-shaped main active strip. Deep-level defect regions are provided within the carrier pressure relief corridors. Uninjected connection regions are provided between the deep-level defect regions and the main active region. The deep-level defect regions are provided with a central region, a transition region, and an end region along the laser cavity length. The central region is located in the quarter-wavelength phase shift region and the cavity length segments on both sides. The defect density in the central region is greater than the defect density in the transition region. The defect density in the transition region is greater than the defect density in the end region. After completion, the device is mounted on a eutectic bonding subcarrier, a butterfly package base, or a high-speed emission component substrate. The heat sink material is made of aluminum nitride, copper tungsten, or diamond composite material. The ridge-shaped main active strip is fixed towards the heat sink along the shortest thermal path. The bonding wire length is controlled between 200 and 500 micrometers, and the package parasitic inductance is controlled between 0.2 and 1 nanohenry to maintain modulation response flatness and high-temperature stability.

[0058] The deep-level defect region is located within the upper waveguide confinement layer and the p-type InP cladding. The uninjected connection region is formed by the continuous portion of the upper waveguide confinement layer on both sides of the ridge-shaped main active strip. The buried current confinement structure is formed outside the carrier relief corridor using Fe-doped InP semi-insulating material. The deep-level defect region is located beyond the full width at half maximum (FWHM) of the main optical mode in the lateral direction. The optical overlap integral between the main optical mode and the deep-level defect region is controlled to be less than 10%, so that the deep-level defect region mainly undertakes the carrier dynamics function rather than the main absorption function. After the buried current confinement structure is formed outside the carrier relief corridor, it sequentially forms a four-level functional structure in the lateral direction: main active region, uninjected connection region, carrier relief corridor, and semi-insulating buried layer. This spatially separates the main injection current, the lateral relief of excess carriers, and the blocking of parasitic currents.

[0059] The coplanar waveguide electrode structure is located above the p-type contact layer. The carrier depressurization corridor performs lateral extraction and non-radiative recombination of excess carriers in the main active region under high-speed direct modulation, thereby reducing the peak value of the spatial distribution of carriers in the main active region and reducing the fluctuation of output optical power during high-speed modulation. High-speed direct modulation is defined as non-return-to-zero direct modulation of 10 gigabits per second to 25.8 gigabits per second, or four-level pulse amplitude modulation of 20 gigabits to 28 gigabits per second. Within an ambient temperature range of 25 degrees Celsius to 85 degrees Celsius, the device maintains a small-signal bandwidth of 3 dB at 8 GHz to 20 GHz, a large-signal extinction ratio above 3 dB, and an eye diagram crossover point stable within 40% to 60%. The peak-to-peak value of the output optical power fluctuation is lower than that of the control device without a carrier relief corridor. Through joint testing of near-field spot size, far-field distribution, side-mode suppression ratio, relative intensity noise, and transient chirp, it is confirmed that the carrier relief corridor selectively extracts excess carriers laterally, while the stimulated emission gain of the main mode remains within the range required for communication applications. Specific Implementation Example 2:

[0061] Based on the technical solution of Specific Embodiment 1, further practical application examples are provided:

[0062] Application case of a direct-modulation laser for the transmitter of a 25 Gigabit / s, 10 km / h single-mode data center optical module. The process is used to fabricate a 1.31-micron band direct-modulation distributed feedback laser and integrate it into the transmitter of an SFP28 or similar high-speed optical module for 25 Gigabit / s, 10 km / h single-mode fiber transmission. First, an InP-based laser chip is fabricated using Sp1 to Sp8 processes. The epitaxial wafer consists of an n-type InP substrate, an n-type InP lower cladding, a lower waveguide confinement layer, a multi-quantum-well active layer, an upper waveguide confinement layer, a p-type InP upper cladding, and a p-type contact layer. The target emission wavelength is designed to be 1310 nm. The total cavity length of the laser is set to 400 μm, the width of the ridge-shaped main active strip is set to 1.8 μm, the distributed feedback grating coupling coefficient is set to 70 μm, the quarter-wavelength phase shift region is located at the center of the cavity length, the length of the central region along the cavity length is set to 140 μm, the total length of the transition region is set to 140 μm, and the total length of the end region is set to 120 μm. A narrow injection window with a width of 1.2 μm is formed in the shoulder region on both sides, leaving a 0.25 μm unwindowed region between the window and the ridge-shaped main active strip to form an uninjected connection region. Subsequently, dual-energy proton injection is performed, with the first proton injection energy set to... The first proton injection was performed at 120 kiloelectron volts with a dose of 3 x 10^13 per square centimeter. The second proton injection was performed at 250 kiloelectron volts with a dose of 1 x 10^14 per square centimeter. The highest dose weight was applied to the central region, the intermediate dose weight to the transition region, and the lowest dose weight to the end region, thus forming a longitudinally distributed carrier depressurization corridor on both sides of the main active region. After proton injection, the deep-level defect region was fixed by rapid thermal annealing at 430 degrees Celsius for 45 seconds. The embedded current-limiting structure was then formed by selective regrowth of Fe-doped InP semi-insulating material. A p-type conductive backfill layer and a p-type InGaAs contact cap layer were then formed in the ridge region. Finally, a 50-ohm coplanar waveguide electrode structure was fabricated. The signal electrode width was set to 10 micrometers, and the gap between the signal electrode and the ground electrode was set to 8 micrometers. The metal system used a titanium-platinum composite stack and the total thickness was controlled to be around 1.5 micrometers. After device completion, wafer-level screening is performed, with screening criteria set as follows: threshold current not exceeding 15 mA, side-mode rejection ratio exceeding 40 dB, series resistance less than 6 ohms, and small-signal bandwidth of 3 dB at 25 degrees Celsius not less than 15 GHz. Chips that pass screening are mounted on a high-speed ceramic subcarrier using AuSn eutectic bonding, with aluminum nitride ceramic heat sinks and bonding wire lengths controlled within 300 micrometers. They are then packaged together with a driver chip, monitoring photodiode, isolator, lens assembly, and single-mode pigtail to form a 25 Gbps, 10 km / h optical module transmitter assembly. During module operation, the laser bias current is set to 1.6 to 2 times the threshold current, the modulation current swing is set to 12 to 20 mA, and the modulation format is a 25.78125 Gbps non-return-to-zero (NRZ) code. The target application scenarios are 10 km single-mode interconnects between data center switches or campus aggregation layer links.In this scenario, traditional InP-based DMLs are prone to carrier overshoot and dynamic spatial hole burning near the phase shift region under conditions of large modulation swing and increased temperature. This leads to a decrease in extinction ratio, collapse of the upper edge of the eye diagram, spectral broadening, and increased power jitter. However, the device fabricated using this technical solution has carrier relief corridors with longitudinal partitions on both sides. The central region preferentially extracts local excess carriers during high-speed rising edges and high-level periods. The transition region provides a buffer for the lateral diffusion process, and the end region maintains low-damage boundary conditions. Therefore, the spatial peak of carriers in the main active region is weakened, the transient swing of the refractive index is reduced, and the main longitudinal mode retention capability is enhanced. Specifically, at the module level, this manifests as a more stable eye diagram opening, a more concentrated crossover point, and smaller output optical power fluctuations under temperature drift conditions. In actual verification, the emitted optical power, extinction ratio, eye diagram template margin, side-mode rejection ratio, relative intensity noise, and transient chirp were measured at three ambient temperatures: 25°C, 55°C, and 85°C. The performance differences between the device of this invention and a control device without a carrier pressure relief corridor under high-temperature and large-amplitude conditions were compared. During the test, it was observed that the optical power overshoot amplitude of the device of this invention was significantly reduced under the drive of an eight-code-long pseudo-random sequence, and the receiving eye diagram decision window was more stable after 25 gigabits per second and 10 kilometers of transmission. The module can still maintain a high side-mode rejection ratio and low power drift at 85°C. Therefore, it is particularly suitable for medium- and short-distance single-mode high-speed interconnect scenarios in data centers with high-density ports and weak or even no cooling. Specific Implementation Example 3:

[0064] Based on the technical solution of Specific Embodiment 1, further practical application examples are provided:

[0065] This case study presents a direct-modulation laser application for a 25 Gbps industrial-grade high-temperature optical module used in fifth-generation mobile communication fronthaul. The technology is applied to a 1.31-micron industrial-grade fronthaul optical module laser transmitter chip, serving outdoor base station fronthaul equipment, high-speed optical links between AAU and DU, and telecom-grade transmitter modules requiring high temperature, vibration, and wide-temperature operation. Compared to data center scenarios, this scenario is characterized by greater environmental temperature fluctuations, more significant power supply disturbances, and longer continuous link operation times. Therefore, it places more stringent requirements on the laser's high-temperature modulation stability, long-term bias reliability, and modulation eye diagram stability. The process technology follows Sp1 to Sp8, but the parameters are further enhanced to meet high-temperature stability requirements: the target epitaxial wavelength is selected as 1310 nm, the quantum well active layer adopts a six-layer compressive strain quantum well structure, both the lower and upper waveguide confinement layers adopt separated confinement heterostructure waveguide layers, the width of the ridge-shaped main active strip is set to 1.6 μm, and the total cavity length is set to 350 μm to achieve a balance between a lower threshold current and a higher resonant frequency; the quarter-wavelength phase shift region is still located at the center of the cavity length, but the length of the central region is increased to 45% of the total cavity length to enhance the carrier depressurization control in the high photon density region near the phase shift; the width of the narrow injection windows on both sides is set to 1.5 μm each, and the width of the non-injected connection region is set to 0.2 μm, further shortening the lateral diffusion time constant between the main active region and the carrier depressurization corridor to adapt to more easily occurring under outdoor high-temperature conditions. The carrier accumulation is achieved through a process where, during proton injection, the first proton injection energy is set to 140 kiloelectron volts and the dose to be 5 x 10^13 per square centimeter, while the second proton injection energy is set to 280 kiloelectron volts and the dose to be 1.5 x 10^14 per square centimeter. Furthermore, the total injection dose is increased in the central region to enhance the excess carrier extraction capability of the deep-level defect region. The annealing conditions are selected as 450 degrees Celsius for 60 seconds to balance defect fixation and long-term high-temperature stability. The embedded current-limiting structure continues to use Fe-doped InP semi-insulating material, but the regeneration thickness is increased to over three micrometers to enhance lateral leakage resistance and thermal diffusion path stability. The coplanar waveguide electrodes employ a thicker metal stack and a shorter feed line layout, with the characteristic impedance controlled around 50 ohms, and are coupled with a low-inductance packaging base to meet the higher RF integrity requirements of the fronthaul module. After the device is fabricated, it is packaged into an industrial-grade SFP28 fronthaul module or an equivalent telecom-grade small transmitter assembly. The heat sink uses a copper-tungsten or diamond composite heat sink substrate. The chip eutectic mounting direction is along the shortest thermal path toward the heat sink. The module introduces a thermal monitoring and automatic power control loop, but does not rely on complex external pre-distortion compensation. It mainly relies on the carrier depressurization structure of the device body to improve dynamic stability.When the system is in use, the module operates on a single-mode fronthaul link at a distance of 10 to 20 kilometers, with a data rate of 25 gigabits per second non-return-to-zero code pattern. The ambient temperature range is from -40 degrees Celsius to 85 degrees Celsius. The key evaluation items are average transmitted optical power drift at high temperatures, extinction ratio maintenance, eye diagram template throughput, wavelength stability after long-term continuous operation, and dynamic code pattern dependent jitter. In actual link operation, traditional direct-modulated lasers typically experience increased threshold drift, enhanced local heat accumulation, and deteriorated spatial distribution of carrier density in the main active region under high temperature and large bias conditions. This leads to eye diagram edge broadening, increased pattern-dependent noise, and degraded bit error rate. However, with the device using this technical solution, the deep-level defects in the central region preferentially depressurize excess carriers, preventing excessive carrier accumulation near the phase shift region due to continuous high-level patterns. The transition region provides a smooth buffer for the lateral descent process, and the end region maintains a low level of non-radiative loss near the end face. Therefore, the device can still maintain small transient chirp and low output optical power fluctuations even at a high temperature of 85 degrees Celsius. Module-level verification can be carried out through long-term lamp-on tests under wide-temperature chamber conditions, 25 gigabits per second pseudo-random sequence transmission tests, 10-kilometer and 20-kilometer standard single-mode link bit error rate tests, and joint analysis of spectrum and eye diagram. Under high temperature and high bias stress conditions, the power drift, extinction ratio attenuation, side-mode suppression ratio changes and bit error rate performance of the device of this invention are compared with those of conventional DML devices. The device of this solution still maintains relatively stable transmit optical power and clear eye diagram boundaries after more than 200 hours of continuous operation. It is especially suitable for real engineering scenarios with high requirements for wide-temperature stability and long-term reliability, such as 5G mobile communication fronthaul, industrial switching networks, rail transit communication links and outdoor high-temperature telecommunications transmission modules.

[0066] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising a reference structure" does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes the element.

[0067] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for improving high-speed modulation stability of an InP-based optical communication laser, characterized in that, The method consists of the following steps: Sp1. Provides an epitaxial wafer for an InP-based laser, wherein the epitaxial wafer is provided from bottom to top as follows: an n-type InP substrate, an n-type InP lower cladding layer, a lower waveguide confinement layer, a multi-quantum-well active layer, an upper waveguide confinement layer, a p-type InP upper cladding layer, and a p-type contact layer; Sp2, A distributed feedback grating is formed on the epitaxial wafer, and a quarter-wavelength phase shift region is formed in the distributed feedback grating; Sp3, etched along the length of the laser cavity to form a ridge-shaped main active strip; Sp4, forming a narrow injection window extending along the length of the cavity in the shoulder region on both sides of the ridge-shaped main active strip; Sp5. Proton injection is performed on both sides of the ridge-shaped main active strip through the narrow injection window to form carrier pressure relief corridors on both sides of the ridge-shaped main active strip. Deep energy level defect regions are provided in the carrier pressure relief corridors, and uninjected connection regions are provided between the deep energy level defect regions and the main active regions. Sp6. Annealing is performed on the structure after proton injection to fix the position and defect density of the deep level defect region; Sp7, An embedded current limiting structure is formed on both sides of the ridge-shaped main active strip; Sp8, A coplanar waveguide electrode structure is formed on the p-type contact layer; The deep-level defect region is provided with a central region, a transition region, and an end region along the cavity length of the laser. The central region is located in the quarter-wavelength phase shift region and the cavity length segments on both sides. The defect density of the central region is greater than that of the transition region, which is greater than that of the end region. When the laser is in a high-speed direct modulation state, the excess carriers in the main active region enter the carrier depressurization corridor through lateral diffusion and local electric field drive and undergo non-radiative recombination, thereby reducing the carrier swing in the main active region and improving the high-speed modulation stability. Both the lower waveguide confinement layer and the upper waveguide confinement layer are separate confinement heterostructure waveguide confinement layers. The multi-quantum well active layer is located between the lower waveguide confinement layer and the upper waveguide confinement layer. The deep energy level defect region is located within the upper waveguide confinement layer and the p-type InP upper cladding. The uninjected connection region is formed by the continuous portion of the upper waveguide confinement layer on both sides of the ridge-shaped main active strip. The proton injection in Sp5 consists of a first proton injection and a second proton injection. The first proton injection forms a first defect layer in the p-type InP cladding, and the second proton injection forms a second defect layer in the upper waveguide confinement layer. The first defect layer and the second defect layer are connected in the vertical direction to form the deep energy level defect region. No defect peak region is formed in the multi-quantum well active layer. The length of the central region along the cavity length direction accounts for 20% to 50% of the total cavity length of the laser, the total length of the transition region along the cavity length direction accounts for 20% to 40% of the total cavity length of the laser, and the total length of the end region along the cavity length direction accounts for 20% to 40% of the total cavity length of the laser.

2. The method for improving high-speed modulation stability of InP-based optical communication lasers according to claim 1, characterized in that, The annealing process in Sp6 is rapid thermal annealing, with an annealing temperature of 350 to 500 degrees Celsius and an annealing time of 10 to 120 seconds.

3. The method for improving high-speed modulation stability of InP-based optical communication lasers according to claim 1, characterized in that, The embedded current limiting structure in Sp7 is formed using Fe-doped InP semi-insulating material, and the embedded current limiting structure is located on both sides of the ridge-shaped main active strip.

4. The method for improving high-speed modulation stability of InP-based optical communication lasers according to claim 1, characterized in that, The lateral width of the uninjected connection region is 0.1 micrometer to 0.8 micrometers, and the lateral width of the carrier pressure relief corridor is 0.5 micrometers to 3 micrometers. The defect density of the deep level defect region and the lateral width of the uninjected connection region together define the lateral carrier pressure relief path from the main active region to the carrier pressure relief corridor.

5. A laser fabricated based on the method for improving high-speed modulation stability of an InP-based optical communication laser according to any one of claims 1-4, characterized in that, The InP-based optical communication laser consists of an n-type InP substrate, an n-type InP lower cladding, a lower waveguide confinement layer, a multi-quantum-well active layer, an upper waveguide confinement layer, a p-type InP upper cladding, a p-type contact layer, a distributed feedback grating, a quarter-wavelength phase shift region, a ridge-shaped main active strip, a carrier pressure relief corridor, a buried current confinement structure, and a coplanar waveguide electrode structure. The distributed feedback grating and the quarter-wavelength phase shift region are located in the waveguide region corresponding to the multi-quantum-well active layer; The ridge-shaped main active strip extends along the length of the laser cavity; The charge carrier pressure relief corridor is located on both sides of the ridge-shaped main active strip; The carrier pressure relief corridor is equipped with a deep energy level defect region; An uninjected connection region is provided between the deep-level defect region and the main active region; The deep energy level defect region is provided with a central region, a transition region, and an end region along the length of the laser cavity; The central region is located in the quarter-wavelength phase shift region and the cavity length segments on both sides; The defect density in the central region is greater than the defect density in the transition region; The defect density in the transition region is greater than the defect density in the end region.

6. The laser fabricated according to the method for improving the high-speed modulation stability of an InP-based optical communication laser as described in claim 5, is characterized in that... The deep-level defect region is located within the upper waveguide confinement layer and the p-type InP upper cladding. The uninjected connection region is formed by the continuous portion of the upper waveguide confinement layer on both sides of the ridge-shaped main active strip. The embedded current confinement structure is formed on the outside of the carrier relief corridor using Fe-doped InP semi-insulating material.

7. The laser fabricated according to the method for improving high-speed modulation stability of an InP-based optical communication laser as described in claim 5, characterized in that, The coplanar waveguide electrode structure is disposed above the p-type contact layer. The carrier depressurization corridor performs lateral extraction and non-radiative recombination of excess carriers in the main active region under high-speed direct modulation, thereby reducing the peak value of the spatial distribution of carriers in the main active region and reducing the output optical power fluctuation during high-speed modulation.

Images