Compact hybrid optical path high-density integration method of multi-level waveguide composite structure

Abstract

The invention relates to the technical field of optoelectronics, and discloses a compact mixed optical path high-density integration method of a multi-level waveguide composite structure, which comprises the following steps: S1, preparing a multi-material layer: sequentially depositing a silicon layer, a silicon nitride layer and a lithium niobate layer, preparing the silicon layer and the silicon nitride layer through a chemical vapor deposition process, preparing the lithium niobate layer through direct current magnetron sputtering, and preparing the multi-material layer; the refractive indexes of the silicon layer, the silicon nitride layer and the lithium niobate layer meet gradient distribution; s2, designing an interlayer coupling structure: forming a gradient waveguide coupler at an interface of adjacent material layers to realize interlayer light field transmission; and S3, stress regulation and control layer integration: arranging a buffer layer with a matched thermal expansion coefficient between the silicon layer and the silicon nitride layer. According to the invention, the silicon layer, the silicon nitride layer and the lithium niobate layer are sequentially deposited, and the refractive index of each layer satisfies gradient distribution, so that the problems of large light field reflection and scattering loss and low coupling efficiency caused by the fact that most of the traditional multilayer waveguide structures adopt single material layers are solved.

Description

Technical Field

[0001] The present invention relates to the technical field of optoelectronics, and specifically to a compact hybrid optical path high-density integration method for a multi-level waveguide composite structure. Background Art

[0002] A multi-level waveguide composite structure is an optical waveguide device formed by stacking two or more dielectric materials with different refractive indices in layers according to specific design rules. Its core advantage lies in regulating the optical transmission path, mode characteristics, and energy distribution through the multi-level structure, enabling functions such as efficient confinement, beam splitting, coupling, or filtering of optical signals. For example, a periodic structure formed by alternating high- and low-refractive-index layers can manipulate the propagation of light with specific wavelengths using the photonic bandgap effect; introducing a graded refractive index layer can optimize the mode field matching and reduce the transmission loss. This structure is widely used in the fields of integrated optics, photonic chips, optical sensors, etc., providing key technical support for the miniaturization and multi-functionality of optical communication and photonic integration systems.

[0003] Most traditional multi-layer waveguide structures use a single material layer or simple stacking without designing a refractive index gradient. Due to the abrupt change in refractive index between layers and the single material preparation process, problems such as large optical field reflection and scattering losses and low coupling efficiency are caused. Summary of the Invention

[0004] In view of the deficiencies of the prior art, the present invention provides a compact hybrid optical path high-density integration method for a multi-level waveguide composite structure, solving the problems that most traditional multi-layer waveguide structures use a single material layer or simple stacking without designing a refractive index gradient, resulting in large optical field reflection and scattering losses and low coupling efficiency due to the abrupt change in refractive index between layers and the single material preparation process.

[0005] To achieve the above objectives, the present invention is realized through the following technical solutions: A compact hybrid optical path high-density integration method for a multi-level waveguide composite structure, comprising the following steps:

[0006] S1. Preparation of multi-material layers: Deposit a silicon layer, a silicon nitride layer, and a lithium niobate layer in sequence. The silicon layer and the silicon nitride layer are prepared by chemical vapor deposition, and the lithium niobate layer is prepared by DC magnetron sputtering. The refractive indices of the silicon layer, the silicon nitride layer, and the lithium niobate layer satisfy a gradient distribution;

[0007] S2. Design of interlayer coupling structure: Form a graded waveguide coupler at the interface of adjacent material layers to achieve interlayer optical field transmission;

[0008] S3. Integration of stress regulation layer: Set a buffer layer with a matching coefficient of thermal expansion between the silicon layer and the silicon nitride layer;

[0009] S4. Multi-layer Stacking and Alignment: The multi-layer structure is stacked through bonding technology, and a laser interference alignment system is used to achieve inter-layer alignment;

[0010] S5. Waveguide Array Processing: A ridge waveguide array is formed on the top lithium niobate layer using nanoimprint lithography technology;

[0011] S6. Thermal Management Structure Integration: A composite heat dissipation layer is prepared on the back of the bottom silicon layer;

[0012] S7. Optical Loss Optimization: The surface of the coupler region is coated;

[0013] S8. Multi-physical Field Verification: The inter-layer optical field distribution and thermal stress state are analyzed by simulation means, and the structural parameters are adjusted until the design requirements are met.

[0014] By adopting the above technical solutions, a silicon layer, a silicon nitride layer, and a lithium niobate layer are sequentially deposited, and the refractive indices of each layer satisfy a gradient distribution. The silicon layer and the silicon nitride layer are prepared by chemical vapor deposition process, and the lithium niobate layer is prepared by DC magnetron sputtering. Furthermore, a multi-level waveguide composite structure is constructed and efficient coupling transmission of light between layers is achieved, thereby improving the problem that most traditional multi-layer waveguide structures use a single material layer or simple stacking and do not design a refractive index gradient. Due to the sudden change of refractive index between layers and the single material preparation process, large optical field reflection and scattering losses and low coupling efficiency are caused.

[0015] Preferably, in S1, the thickness of the silicon layer is 0.2 - 2 μm, the thickness of the silicon nitride layer is 0.1 - 1.5 μm and the deposition rate is controlled at 0.5 - 2 nm / s, the thickness of the lithium niobate layer is 0.3 - 3 μm, and it is prepared by DC magnetron sputtering with a sputtering power of 80 - 150 W and an argon pressure of 0.5 - 1.2 Pa.

[0016] Preferably, in S2, the length of the tapered waveguide coupler is 5 - 50 μm, the width gradient rate is 0.1 - 1 μm / μm, and an exponential tapered profile is adopted; when a grating coupler is used, the grating period is 200 - 500 nm, the duty cycle is 0.3 - 0.7, and the etching depth is 60% - 80% of the thickness of the adjacent silicon nitride layer.

[0017] Preferably, in S3, the buffer layer is a titanium dioxide - silicon dioxide composite layer with a thickness of 20 - 50 nm, the doping ratio of silicon dioxide is 5% - 15% atomic ratio, and it is deposited by ion beam sputtering with a sputtering current of 50 - 100 mA, a deposition temperature of 100 - 200 °C, and a thermal expansion coefficient of 8 - 12×10 -6 / °C, and the difference in thermal expansion coefficient from the silicon layer is ≤1×10 -6 / °C.

[0018] Preferably, in S4, the laser interference alignment system uses a helium-neon laser with a wavelength of 632.8 nm, the alignment accuracy is ≤20 nm, the alignment mark is a stacked grating of a silicon layer and a silicon nitride layer, the period is 2 μm, and the height is 200 - 500 nm; the bonding process uses plasma-activated bonding, the oxygen plasma power is 100 - 300 W, the bonding pressure is 0.1 - 0.5 MPa, the bonding time is 5 - 15 min, and the interfacial void ratio is ≤0.5%.

[0019] Preferably, in S5, the template in the nanoimprint lithography technology is a nickel-cobalt alloy, the hardness is 500 - 800 HV, the imprinting pressure is 5 - 15 MPa, the imprinting temperature is 80 - 120 °C, the template release uses critical CO2 drying, the pressure is 7.38 MPa, the temperature is 31.1 °C, the ridge waveguide width is 0.5 - 2 μm, the ratio of the height to the thickness of the lithium niobate layer is 0.6 - 0.9, and the sidewall roughness is ≤5 nm.

[0020] Preferably, in S6, the composite heat dissipation layer includes a copper-nickel alloy layer with a thickness of 50 - 200 nm, the copper-nickel atomic ratio is 7:3, the interfacial roughness is ≤2 nm, the backside is etched with serpentine microchannels, the width is 5 - 10 μm, the depth is 2 - 5 μm, the spacing is 20 - 50 μm, the channels are filled with graphene-water nanofluid, the graphene concentration is 0.01 - 0.1 wt%, the thermal conductivity is 2.5 - 5.0 W / (m 2 ·K).

[0021] Preferably, in S7, the surface coating treatment uses PECVD technology to deposit a silica layer, the deposition temperature is 200 - 350 °C, the radio frequency power is 50 - 150 W, the coating thickness is 50 - 200 nm, it is controlled by measuring with an ellipsometer, the difference in refractive index from the waveguide layer is ≤0.03, and the surface flatness is ≤1 nm.

[0022] Preferably, in S8, the simulation method is optical simulation technology, and the optical simulation technology uses the beam propagation method. The calculation formula is:

[0023]

[0024] where λ is the working wavelength, taking values of 1.3 - 1.55 μm, k0 = 2π / λ is the vacuum wave number, n eff is the effective refractive index of the waveguide, n is the refractive index of the waveguide material, and the simulation grid accuracy is ≤λ / 20; the thermal stress simulation uses an elastoplastic mechanics model, the calculation step size is 0.1 - 1 μm, and the von Mises stress is ≤50 MPa.

[0025] Preferably, in S5, the waveguide array pitch is modulated by a superlattice structure, the superlattice period a = (1.5 - 2.5)w (w is the waveguide width), the slot line depth is 1 / 3 - 1 / 2 of the ridge waveguide height, and the etching rate of 20 - 50 nm / min is achieved by radio frequency plasma etching; the thermal cycle test conditions are -40°C to 85°C, 100 cycles, the waveguide loss change ≤ 0.1 dB, and the displacement ≤ 5 nm.

[0026] The present invention provides a compact hybrid optical path high-density integration method for a multi-level waveguide composite structure. It has the following beneficial effects:

[0027] 1. In the present invention, by sequentially depositing a silicon layer, a silicon nitride layer, and a lithium niobate layer and making the refractive indices of each layer satisfy a gradient distribution, and the silicon layer and the silicon nitride layer are prepared by chemical vapor deposition process, and the lithium niobate layer is prepared by DC magnetron sputtering, a multi-level waveguide composite structure is constructed and efficient coupling transmission of light between layers is realized, thereby improving the problem that most traditional multi-layer waveguide structures use a single material layer or simple stacking and do not design a refractive index gradient. Due to the sudden change of refractive index between layers and the single material preparation process, the light field reflection and scattering loss is large and the coupling efficiency is low.

[0028] 2. In the present invention, by forming a gradient waveguide coupler (using an exponential gradient profile) at the interface of adjacent material layers and using a laser interference alignment system to achieve nano-level alignment between layers, the inter-layer light field transmission efficiency is improved and the precise integration of the multi-layer structure is ensured, thereby improving the problem that most traditional inter-layer light transmission structures use simple linear coupling or no gradient profile design and the alignment process is rough. Due to the coupling mode mismatch and large alignment error, the optical signal transmission loss is high and the optical path integration density is limited.

[0029] 3. In the present invention, by preparing a composite heat dissipation layer containing a copper-nickel alloy layer, a serpentine microchannel, and graphene-water nanofluid on the back of the bottom silicon layer, and analyzing the inter-layer light field and thermal stress state by simulation means, the structural heat accumulation is efficiently derived and the multi-physical field performance synergy is optimized, thereby improving the problem that most traditional photonic integrated devices use a single metal heat dissipation layer or no active heat dissipation structure and lack multi-physical field simulation optimization. Due to low heat dissipation efficiency and thermal stress concentration, the device's optical performance deteriorates or even the structure fails due to temperature drift. Brief Description of the Drawings

[0030] Figure 1 It is a schematic diagram of the method steps of the present invention. Detailed Embodiments

[0031] The technical solution of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative work belong to the scope of protection of the present invention.

[0032] Please refer to the attached Figure 1 , the embodiment of the present invention provides a compact hybrid optical path high-density integration method for a multi-level waveguide composite structure, including the following steps:

[0033] S1. Preparation of multi-material layers: Deposit a silicon layer, a silicon nitride layer, and a lithium niobate layer in sequence. The silicon layer and the silicon nitride layer are prepared by chemical vapor deposition process, and the lithium niobate layer is prepared by DC magnetron sputtering. The refractive indices of the silicon layer, the silicon nitride layer, and the lithium niobate layer satisfy a gradient distribution;

[0034] S2. Design of interlayer coupling structure: Form a graded waveguide coupler at the interface of adjacent material layers to realize interlayer optical field transmission;

[0035] S3. Integration of stress control layer: Set a buffer layer with a matching coefficient of thermal expansion between the silicon layer and the silicon nitride layer;

[0036] S4. Multi-layer stacking and alignment: Complete the multi-layer structure stacking through bonding technology, and use a laser interference alignment system to achieve interlayer alignment;

[0037] S5. Processing of waveguide array: Use nanoimprint lithography technology to form a ridge waveguide array on the top lithium niobate layer;

[0038] S6. Integration of thermal management structure: Prepare a composite heat dissipation layer on the back of the bottom silicon layer;

[0039] S7. Optimization of optical loss: Perform surface coating treatment on the coupler area;

[0040] S8. Verification of multi-physical fields: Analyze the interlayer optical field distribution and thermal stress state through simulation means, and adjust the structural parameters until the design requirements are met.

[0041] Specifically, through S1: depositing a silicon layer, a silicon nitride layer, and a lithium niobate layer in sequence, the basic architecture of a multi-level waveguide composite structure is formed, providing a physical carrier for subsequent optical transmission and integration; each layer of material is processed through specific preparation processes (chemical vapor deposition for the silicon layer and the silicon nitride layer, and direct current magnetron sputtering for the lithium niobate layer), ensuring the quality and performance of the materials; at the same time, the refractive indices of the silicon layer, the silicon nitride layer, and the lithium niobate layer satisfy a gradient distribution, which helps to achieve efficient coupling and transmission of light between different layers, reduce reflection and scattering losses, and improve the performance of the optical system; through S2: a tapered waveguide coupler is formed at the interface between adjacent material layers, and through its geometric shape (such as the width, height, or the gradient characteristics of the refractive index), the optical field is made to achieve mode matching between the waveguides of different material layers, reducing the reflection loss caused by the difference in material refractive indices and ensuring the efficient transmission of optical signals between layers; the design of the tapered coupler can shorten the optical transmission path between layers and reduce the space occupied by the coupling structure, thereby allowing more waveguide layers and functional devices to be integrated in a limited area and enhancing the integration density of the hybrid optical path; at the same time, by adjusting the tapered parameters of the coupler (such as length, taper rate), the optical characteristics of different material layers can be flexibly adapted, enhancing the compatibility of the optical path design; the tapered structure avoids the leakage of optical energy into non-target waveguide layers by gradually adjusting the optical field distribution, reducing interlayer crosstalk; at the same time, the optimized coupling interface can reduce scattering and mode mismatch losses, ensuring the transmission stability of optical signals in the multi-layer waveguide composite structure; through S3: a buffer layer with a matching coefficient of thermal expansion is set between the silicon layer and the silicon nitride layer, which can reduce the thermal stress generated by the difference in the coefficient of thermal expansion between the two layers of materials, and avoid cracking, deformation, or interface separation of the multi-layer structure caused by temperature changes during preparation, testing, and operation; the buffer layer evenly disperses the thermal stress to the entire composite structure through physical isolation and stress conduction path adjustment, suppressing local stress concentration, thereby enhancing the mechanical stability and environmental adaptability of the multi-layer waveguide composite structure (such as reliability in high and low temperature cycling scenarios); a stable mechanical structure can prevent changes in the geometric dimensions of the waveguide layer (such as width, spacing) due to stress, ensuring the consistency of the interlayer optical field transmission characteristics (such as coupling efficiency, transmission loss), and maintaining the long-term stability of the optical path performance; through S4: the multi-layer materials such as the silicon layer, the silicon nitride layer, and the lithium niobate layer are firmly stacked through a bonding technique to form a complete composite waveguide structure, providing a physical carrier for three-dimensional optical signal transmission; a laser interference alignment system is used to achieve nanoscale alignment between layers (accuracy ≤ ±20 nm), ensuring the position matching of the waveguide structures of each layer (such as couplers, ridge waveguides) in the vertical direction, and avoiding the deviation of the optical field transmission path or the increase in loss caused by alignment deviation; high-precision alignment can ensure the accurate docking of the geometric features (such as length, width taper path) of the interlayer coupling structure (such as the tapered waveguide coupler), ensuring the efficient coupling of the optical field between layers; the bonding technique controls the interfacial void fraction (≤0.5%) and bonding strength, avoid introducing additional optical losses due to interlayer gaps or bubbles, and improve the overall reliability of the structure; precise multi-layer stacking and alignment techniques allow for the integration of more functional layers (such as stress regulation layers, thermal management layers) within a limited space, and enable the design of complex three-dimensional optical paths (such as crossing waveguides, vertical couplings), providing a process guarantee for the high-density integration of compact hybrid optical paths; by S5: Using nanoimprint lithography technology to form a ridge waveguide array on the top lithium niobate layer, providing a regular transmission path for optical signals, restricting the optical field to transmit within the ridge waveguide region, reducing the leakage of optical energy to the surrounding medium, and ensuring the efficiency and stability of signal transmission; nanoimprint lithography technology can batch fabricate ridge waveguides with micron to nanometer sizes (width 0.5 - 2μm, spacing ≥ 3 times the waveguide width) on the surface of the lithium niobate layer through high-precision template replication (template hardness 500 - 800HV, imprinting pressure 5 - 15MPa), significantly increasing the waveguide integration density per unit area and meeting the requirements of compact optical paths for space utilization; by controlling the imprinting temperature (80 - 120°C), template release process (critical CO2 drying), and etching parameters (sidewall roughness ≤ 5nm), ensure the consistency of the geometric dimensions (such as the ratio of height to lithium niobate layer thickness 0.6 - 0.9) and surface quality of the ridge waveguide array, reduce the impact of process fluctuations on optical performance, and achieve the stability of mass production; the regular ridge waveguide array can provide a standardized interface for subsequent integration of functional devices such as modulators and detectors, facilitating the construction of complex hybrid photon integration systems, such as optical communication modules or optical signal processing units; by S6: Prepare a composite heat dissipation layer (such as a 50 - 200nm thick copper-nickel alloy layer) on the back of the bottom silicon layer, and utilize the high thermal conductivity characteristics of the metal material (copper-nickel atomic ratio 7:3, thermal conductivity 2.5 - 5.0W / (m·K)) to quickly conduct the heat generated during the operation of the multi-layer waveguide composite structure to the external environment, avoiding material property degradation or optical parameter drift caused by heat accumulation; the composite heat dissipation layer combines serpentine microchannels (width 5 - 10μm, depth 2 - 5μm) and graphene-water nanofluid filling, and through increasing the heat dissipation surface area and strengthening convective heat transfer, achieve a uniform distribution of the temperature field within the structure, suppress the formation of local hot spots, and improve the operating stability of the device under high power density; the heat dissipation layer and the silicon layer are bonded by magnetron sputtering technology to achieve a low interface roughness (≤ 2nm), reducing the thermal contact resistance while reducing the mechanical stress caused by the difference in thermal expansion coefficients, avoiding delamination or cracking of the multi-layer structure due to thermal cycling (-40°C to 85°C, 100 times), and extending the service life of the device; the efficient thermal management structure provides a thermal environment guarantee for the high-density integration of the multi-layer waveguide composite structure, allowing for the integration of more heat-generating devices (such as lasers, modulators) within a limited space, breaking through the integration density limitation of traditional optical paths due to heat dissipation bottlenecks; by S7: Perform surface coating treatment on the coupler region (such as depositing a silicon dioxide layer by PECVD), and by controlling the refractive index difference between the coating and the adjacent waveguide layer ≤ 0.03. Make the refractive index transition of the optical field at the interlayer interface smoother, reduce the Fresnel reflection loss caused by the sudden change in refractive index, and improve the transmission efficiency of the optical signal through the coupling interface; the cladding treatment can fill the microscopic defects on the surface of the coupler (such as grooves or particles generated by etching), control the surface flatness within ≤1nm, and reduce the scattering loss of the optical field caused by surface roughness; at the same time, the uniform cladding thickness (50 - 200nm) can eliminate the interference of local protrusions or depressions on the optical transmission path and ensure the stability of the optical field distribution; the cladding, as a physical barrier, can prevent the coupler area from being directly exposed to the external environment (such as moisture, dust), prevent the deterioration of optical performance caused by material oxidation or pollution, and extend the long-term working reliability of the device; in addition, the chemical stability of the cladding material (such as silica) can enhance the mechanical strength of the coupler structure and reduce the risk of damage under external forces; through S8: Analyze the optical field distribution between layers by optical simulation (such as the beam propagation method), quantify parameters such as transmission loss and coupling efficiency, and verify the rationality of the optical design of structures such as tapered waveguide couplers and ridge waveguide arrays. By adjusting parameters such as the coupler length (5 - 50μm) and waveguide width (0.5 - 2μm), make the optical field distribution meet the low-loss transmission requirements (such as transmission loss ≤0.5dB / coupling interface); use thermal stress simulation (elastoplastic mechanics model) to analyze the stress distribution of the multi-layer structure under temperature changes, calculate the von Mises stress (≤50MPa) and waveguide displacement (≤5nm), and optimize parameters such as the thickness of the stress control layer (20 - 50nm) and the structure of the heat dissipation layer to ensure that the structure has no cracking or performance drift under thermal cycling (-40°C to 85°C, 100 times); through the combined simulation of the coupling effect of the optical field and thermal stress, identify the conflict points between optical performance and thermal stability (such as insufficient heat dissipation caused by high integration density), and achieve the balance of optical efficiency and thermal management through parameter iteration (such as adjusting the waveguide spacing, heat dissipation layer coverage area ≥80%), improving the comprehensive performance of the multi-layer waveguide composite structure; before the process implementation, pre-verify the design scheme through simulation, discover potential defects (such as mode mismatch, thermal stress concentration) in advance, reduce the dependence on expensive manufacturing processes (such as nanoimprint lithography, ion beam sputtering), shorten the R & D cycle and reduce the mass production cost; by sequentially depositing silicon layer, silicon nitride layer, and lithium niobate layer and making the refractive index of each layer satisfy the gradient distribution, and the silicon layer and silicon nitride layer are prepared by chemical vapor deposition process, and the lithium niobate layer is prepared by DC magnetron sputtering, thus constructing a multi-level waveguide composite structure and realizing the efficient coupling transmission of light between layers, thereby improving the problem that most traditional multi-layer waveguide structures use single material layers or simple stacks without designing refractive index gradients, resulting in large light field reflection and scattering losses and low coupling efficiency due to sudden changes in refractive index between layers and single material preparation processes.

[0042] In S1, the thickness of the silicon layer is 0.2 - 2 μm, the thickness of the silicon nitride layer is 0.1 - 1.5 μm and the deposition rate is controlled at 0.5 - 2 nm / s, the thickness of the lithium niobate layer is 0.3 - 3 μm, and it is prepared by DC magnetron sputtering with a sputtering power of 80 - 150 W and an argon pressure of 0.5 - 1.2 Pa.

[0043] Specifically, controlling the thickness of the silicon layer at 0.2 - 2 μm can provide sufficient mechanical support strength. At the same time, as the bottom waveguide layer, its thickness directly affects the optical mode confinement ability of the silicon-based waveguide, ensuring that the optical field is effectively confined within the waveguide for transmission and reducing leakage loss. By controlling the thickness of the silicon nitride layer (0.1 - 1.5 μm) and the deposition rate (0.5 - 2 nm / s): optimizing the thickness realizes the refractive index gradient matching with the silicon layer and the lithium niobate layer (silicon layer > silicon nitride layer > lithium niobate layer), enhancing the interlayer optical coupling efficiency; the deposition rate affects the film stress and crystallization quality, and a lower rate (0.5 - 1 nm / s) can reduce the accumulation of internal stress (≤ ±50 MPa), avoid film cracking, and ensure the structural stability of the waveguide. By controlling the thickness of the lithium niobate layer (0.3 - 3 μm) and the DC magnetron sputtering parameters (power 80 - 150 W, argon pressure 0.5 - 1.2 Pa): the thickness determines the optical mode size of the lithium niobate ridge waveguide, matching the mode field of the single-mode fiber (such as the mode field diameter is about 9 μm at a wavelength of 1.55 μm), reducing the coupling loss; the sputtering power and gas pressure affect the film density and crystal orientation, and a high power (100 - 150 W) combined with a low gas pressure (0.5 - 0.8 Pa) can prepare a high-purity single-crystal lithium niobate film, improving the nonlinear optical performance (such as the electro-optic effect coefficient) to meet the integration requirements of functional devices such as modulators and detectors. By matching the thicknesses of the silicon layer, the silicon nitride layer, and the lithium niobate layer, an optical confinement system for the multi-layer waveguide composite structure is constructed to ensure the efficient transmission of the optical field in the interlayer tapered coupler (such as the coupling loss ≤ 0.5 dB / interface), while meeting the strict space size limitations of compact integration (total thickness ≤ 5 μm).

[0044] In S2, the length of the tapered waveguide coupler is 5 - 50 μm, the width gradient rate is 0.1 - 1 μm / μm, and an exponential tapered profile is adopted; when a grating coupler is used, the grating period is 200 - 500 nm, the duty cycle is 0.3 - 0.7, and the etching depth is 60% - 80% of the thickness of the adjacent silicon nitride layer.

[0045] Specifically, the length determines the action distance of the optical field in the interlayer coupling. A shorter length (5 - 20 μm) is suitable for compact integration scenarios, while a longer length (20 - 50 μm) can improve the coupling efficiency through progressive mode conversion; the width gradient rate controls the rate of change of the waveguide cross-section. A lower gradient rate (0.1 - 0.5 μm / μm) reduces the mode mismatch loss, and a higher gradient rate (0.5 - 1 μm / μm) adapts to the interlayer coupling with a high refractive index difference to ensure the effective transfer of optical field energy (such as the coupling efficiency ≥ 95%); the waveguide width change is regulated by an exponential function (such as ), enabling the Gaussian distribution matching of the optical field mode at the interlayer interface, more effectively suppressing the excitation of high-order modes compared with linear gradient, and reducing the scattering loss (such as the transmission loss ≤ 0.3 dB / coupler); the period matches the optical wavelength (such as a period of 200 - 400 nm corresponding to a wavelength of 1.3 - 1.55 μm), satisfying the phase matching condition (Δβ ≤ 10 -3 rad / m) to achieve vertical optical coupling (such as alignment with the fiber array); the duty cycle controls the ratio of the grating tooth width. A range of 0.3 - 0.7 balances the reflection and transmission efficiencies, avoiding total reflection or energy leakage; the etching depth reaches 60% - 80% of the thickness of the silicon nitride layer, ensuring the effective modulation of the optical field phase by the grating structure while retaining sufficient material strength to prevent structural failure caused by etching through; the combination of the tapered and grating couplers provides the dual capabilities of planar interlayer coupling and vertical optical input / output, supporting three-dimensional hybrid optical path integration to meet the high-density wiring requirements of complex optical interconnection scenarios (such as chip-level optical networks).

[0046] In S3, the buffer layer is a titanium dioxide - silicon dioxide composite layer with a thickness of 20 - 50 nm, and the doping ratio of silicon dioxide is 5% - 15% atomic ratio. It is deposited by ion beam sputtering, with a sputtering current of 50 - 100 mA, a deposition temperature of 100 - 200 °C, and a thermal expansion coefficient of 8 - 12×10 -6 / °C, and the difference in thermal expansion coefficient from the silicon layer is ≤ 1×10 -6 / °C.

[0047] Specifically, the thermal expansion coefficient of the buffer layer (8 - 12×10 -6 / °C) and the silicon layer (the thermal expansion coefficient is about 2.6×10 -6 / °C) have a difference ≤ 1×10 -6 / °C. Through the design of gradient thermal expansion coefficient, the interfacial thermal stress (von Mises stress ≤ 30 MPa) generated by the temperature change (-40 °C to 85 °C) between the silicon layer and the upper layer (such as silicon nitride / niobate lithium) is reduced, avoiding interlayer cracking or waveguide structure deformation; the titanium dioxide - silicon dioxide composite layer (refractive index 1.6 - 2.0, adjusted with the silicon dioxide doping ratio of 5% - 15% atomic ratio) serves as an intermediate transition layer to bridge the silicon (n si ≈ 3.4) and silicon nitride (n siN ≈ 2.0) / niobate lithium( ) refractive index difference, reducing the reflection loss of the optical field at the interlayer interface (reflectivity ≤ 1%), and improving the optical transmission efficiency; The ion beam sputtering process (sputtering current 50 - 100 mA, deposition temperature 100 - 200 °C) promotes atomic-level diffusion through high-energy ion bombardment to form a dense amorphous composite layer (density ≥ 2.8 g / cm 3 ), reducing the pinhole defect rate (≤ 0.1 per mm 2 ), and enhancing the interlayer adhesion (binding energy ≥ 2 J / m 2 ); The thickness is controlled within 20 - 50 nm to avoid optical mode leakage caused by excessive thickness or stress regulation failure caused by excessive thinness, while meeting the requirements of nanoscale lithography process for the flatness of the bottom layer (surface roughness ≤ 1 nm); The composite layer has both thermal stress buffering and optical matching functions, and dynamically adjusts the thermal-optical performance balance through the doping ratio of silicon oxide. For example, when highly doped (10% - 15%), it focuses on reducing the difference in thermal expansion coefficients, and when lowly doped (5% - 10%), it focuses on optimizing the refractive index gradient to ensure that the silicon-based photon integrated device maintains stable optical performance (such as insertion loss fluctuation ≤ 0.2 dB) and structural reliability in a wide temperature range (operating temperature range - 20 °C to 60 °C).

[0048] In S4, the laser interference alignment system uses a helium-neon laser with a wavelength of 632.8 nm, the alignment accuracy ≤ 20 nm, and the alignment marks are stacked gratings of silicon layer and silicon nitride layer with a period of 2 μm and a height of 200 - 500 nm; The bonding process uses plasma-activated bonding, the oxygen plasma power is 100 - 300 W, the bonding pressure is 0.1 - 0.5 MPa, the bonding time is 5 - 15 min, and the interface void ratio ≤ 0.5%.

[0049] Specifically, the helium-neon laser (wavelength 632.8 nm) combines laser interference technology, and uses the diffraction signal of the stacked grating marks (silicon layer / silicon nitride layer period 2 μm, height 200 - 500 nm) to achieve an alignment accuracy ≤ 20 nm, ensuring that the lateral position deviation of the waveguide layer (such as silicon-based optical waveguide and silicon nitride modulation layer) is less than the full width at half maximum of the optical field mode (such as when the single-mode waveguide mode width ≤ 400 nm, the deviation ≤ 10%), avoiding mode mismatch loss caused by alignment error (such as the coupling loss increase ≥ 1 dB); The stacked grating structure enhances the optical signal intensity through the contrast of the double-layer material (diffraction efficiency ≥ 15%), improves the signal-to-noise ratio of the alignment system, and supports sub-wavelength-level alignment correction (such as the step adjustment accuracy ≤ 5 nm); The oxygen plasma power of 100 - 300 W regulates the degree of surface hydroxylation, generating a density ≥ 1×10 15 OH / cm 2The active groups enhance the initial bonding strength (interface shear strength ≥ 10 MPa) through hydrogen bonding, and at the same time remove surface organic contamination (residual carbon content ≤ 0.1 at%). The combined action of a bonding pressure of 0.1 - 0.5 MPa and a time of 5 - 15 min promotes the diffusion of interface atoms (diffusion depth ≥ 5 nm) to form a covalent bonding layer (such as the Si - O - Si bond density ≥ 5×10 14 / cm 2 ), making the interface void ratio ≤ 0.5% (void size ≤ 5 μm) to avoid interlayer delamination or optical scattering defects caused by local stress concentration (such as scattering loss ≤ 0.05 dB / mm). The alignment and bonding processes are jointly controlled to ensure that the silicon - based platform and functional layers such as silicon nitride / niobate maintain strict coplanarity (height deviation ≤ 50 nm) after bonding, avoiding optical field leakage (such as an increase in the bending waveguide loss ≥ 0.1 dB / cm) or thermal stress concentration (bonding interface thermal resistance ≤ 0.5 K·cm 2 / W) caused by interlayer misalignment, while meeting the structural reliability requirements for long - term device operation (lifetime ≥ 10 years) (such as the bonding strength attenuation ≤ 5% after 1000 thermal cycles). The non - contact measurement characteristics of laser interference alignment (alignment time ≤ 30 s / chip) and the large - area uniformity of plasma bonding (thickness uniformity ≤ 2% when the bonding area ≥ 100 cm 2 ) support wafer - level batch bonding processes, meet the large - scale manufacturing requirements of silicon - based photonic chips (such as production capacity ≥ 1000 wafers / month), and at the same time keep the process cost controllable (bonding defect rate ≤ 1%).

[0050] In S5, the template in nanoimprint lithography is a nickel - cobalt alloy with a hardness of 500 - 800 HV, an imprinting pressure of 5 - 15 MPa, an imprinting temperature of 80 - 120 °C, and the template release uses critical CO2 drying with a pressure of 7.38 MPa and a temperature of 31.1 °C. The ridge waveguide width is 0.5 - 2 μm, the ratio of the height to the thickness of the niobate layer is 0.6 - 0.9, and the side - wall roughness ≤ 5 nm.

[0051] Specifically, a nickel-cobalt alloy template (hardness 500 - 800 HV) is formed into a nanoscale pattern (such as a ridge waveguide width of 0.5 - 2 μm) through electron beam direct writing or focused ion beam machining. Using an imprinting pressure of 5 - 15 MPa and a temperature of 80 - 120 °C, a photoresist (such as a thermoplastic polymer) undergoes plastic deformation to achieve a high-fidelity transfer of the pattern from the template to the lithium niobate layer (line width error ≤ 5%). Ensure that the ratio of the ridge waveguide height to the lithium niobate layer thickness is controlled within 0.6 - 0.9 to avoid a decrease in the mode confinement efficiency caused by over-etching (ratio < 0.6) or an overly thick residual layer (ratio > 0.9) (such as an effective refractive index deviation ≤ 0.05); The critical CO2 drying process (pressure 7.38 MPa, temperature 31.1 °C) eliminates capillary forces through the characteristics of supercritical fluids, preventing the collapse of the photoresist structure (such as height loss ≤ 2%), while ensuring that the sidewall roughness of the ridge waveguide ≤ 5 nm, reducing the optical field scattering loss (such as scattering loss ≤ 0.1 dB / cm at a wavelength of 1550 nm); The hardness of the template matches that of the lithium niobate material (hardness ≈ 600 HV) to avoid template wear during the imprinting process (pattern distortion rate ≤ 1% when the cumulative number of uses ≥ 1000 times) or material damage (such as the microcrack density in the imprinted area ≤ 0.01 pieces / μm); Temperature-pressure co-regulation optimizes the molecular chain orientation, so that the internal stress of the waveguide structure ≤ 50 MPa after cooling and solidification, reducing structural cracking caused by stress concentration (such as crack propagation rate ≤ 10 -10 m / s), while enhancing the interfacial bonding strength between the waveguide and the underlying buffer layer (adhesion ≥ 3 N / cm); The nanoimprinting technology has a single imprinting area ≥ 100 cm 2 , and the production efficiency is 5 - 10 times higher than that of the traditional lithography process. Combined with the template reuse characteristics (pattern residual pollutants ≤ 0.1 μg / cm 2 ), it significantly reduces the manufacturing cost of high-resolution patterns (such as submicron waveguides) (single chip cost ≤ $10); The fully automated process of critical CO2 drying (processing time ≤ 30 min) is adapted to roll-to-roll or wafer-level continuous production, supporting the large-scale preparation of photonic integrated devices (such as the batch consistency error of arrayed waveguide grating devices ≤ 2%); Precise control of the ridge waveguide geometric parameters (width, height, sidewall roughness) and the internal stress state of the material ensure that the optical field is confined in the waveguide core layer (mode field confinement factor ≥ 90%), while reducing the refractive index drift caused by the thermo-optic effect (such as refractive index change ≤ 1×10 -5 ) when the temperature changes by 1 °C, meeting the requirements of low loss (transmission loss ≤ 0.5 dB / cm) and high stability (wavelength drift ≤ 0.01 nm / °C) of the device in the communication band (1260 - 1625 nm).

[0052] In S6, the composite heat dissipation layer includes a copper-nickel alloy layer with a thickness of 50 - 200 nm, a copper-nickel atomic ratio of 7:3, an interface roughness ≤ 2 nm, and serpentine microchannels etched on the back with a width of 5 - 10 μm, a depth of 2 - 5 μm, and a pitch of 20 - 50 μm. The channels are filled with graphene-water nanofluid with a graphene concentration of 0.01 - 0.1 wt%, a thermal conductivity of 2.5 - 5.0 W / (m·K), and a heat dissipation coefficient ≥ 1000 W / (m 2 ·K).

[0053] Specifically, the 50 - 200 nm thick copper-nickel alloy layer (copper-nickel atomic ratio 7:3) utilizes the high thermal conductivity of copper (thermal conductivity ≈ 400 W / (m·K)) and the oxidation resistance of nickel to form a low contact thermal resistance interface (interface thermal resistance ≤ 0.5×10 -6 m 2 ·K / W), combined with an interface roughness ≤ 2 nm, to ensure that the heat generated (power density ≤ 10 6 W / m 2 ) in the active area of the chip (such as modulators, detectors) is quickly laterally diffused through the metal layer (thermal diffusion rate ≥ 10 3 m 2 / s); the difference in the thermal expansion coefficient of the copper-nickel alloy (≈ 13×10 -6 / ℃) and the silicon-based substrate (≈ 2.6×10 -6 / ℃) is further reduced through gradient composition design (such as introducing titanium or chromium in the transition layer) to avoid interface delamination (fatigue life ≥ 10 4 cycles) under thermal cycling (-40℃ to 85℃); the serpentine microchannels etched on the back (width 5 - 10 μm, depth 2 - 5 μm, pitch 20 - 50 μm) increase the effective heat dissipation area to 3 - 5 times that of the traditional planar structure, and the channel density ≥ 2×10 4 channels / cm 2 , combined with graphene-water nanofluid (thermal conductivity 2.5 - 5.0 W / (m·K), 20 - 50% higher than pure water), to control the chip junction temperature (Tj) ≤ 85℃ (when the ambient temperature is 25℃) by enhancing convective heat transfer (Nusselt number increased by 40 - 60%); the optimization of the microchannel pitch and width (aspect ratio 1:0.4 - 1:0.5) balances the flow resistance (pressure drop ≤ 10 kPa) and heat dissipation efficiency, avoiding local hot spots (temperature uniformity ≤ ±2℃) caused by too low a flow rate (< 0.1 m / s), and at the same time preventing hydrodynamic noise (sound pressure level ≤ 40 dB) caused by too high a flow rate (> 1 m / s); the dispersion system with a graphene concentration of 0.01 - 0.1 wt% is stably suspended through π-π bond interactions (sedimentation rate ≤ 0.01% / h), and its nanoscale sheet structure (diameter 1 - 10 μm, thickness 1 - 10 nm) forms a "thermal network" during flow, strengthening the heat exchange between the fluid and the channel wall (convective heat transfer coefficient ≥ 1000 W / (m2 ·K), which is 2 - 3 times that of pure water); the high specific heat capacity (≈4.2 kJ / (kg·K)) and low viscosity (≈1 mPa·s) characteristics of nanofluids enable them to achieve high heat transport capacity while maintaining a laminar flow state (Reynolds number ≤ 2000) in microchannels, effectively dissipating the local peak heat flux density of the chip (≥5×10 6 W / m 2 ), avoiding material thermal failure (such as the thermal-induced phase change error of the lithium niobate layer ≤ 0.1 rad) caused by an excessive temperature gradient (>10 4 K / m); the electrochemical corrosion potential of the copper-nickel alloy layer (≈ -0.25 V vs SHE) and the passivation treatment of the inner wall of the microchannel (such as electroless nickel-phosphorus alloy plating with a thickness of 1 - 3 μm) enable the corrosion rate of the heat dissipation structure to be ≤ 0.1 μm / year in an acidic (pH = 4) or alkaline (pH = 10) nanofluid environment; the critical CO2 cleaning process (pressure 7.38 MPa, temperature 31.1 °C) can periodically remove the graphene aggregates deposited in the channel (residual amount ≤ 10 -6 g / cm 2 ), maintaining long-term stable heat dissipation performance (the thermal conductivity decay ≤ 5% after 500 hours of continuous operation).

[0054] In S7, for the surface coating treatment, a silica layer is deposited by PECVD technology. The deposition temperature is 200 - 350 °C, the radio frequency power is 50 - 150 W, the coating thickness is 50 - 200 nm, and it is controlled by measuring with an ellipsometer. The difference in refractive index from the waveguide layer is ≤ 0.03, and the surface flatness is ≤ 1 nm.

[0055] Specifically, the difference in refractive index between the silica coating (refractive index ≈ 1.44) deposited by PECVD technology and the waveguide layer (such as the refractive index of silicon nitride ≈ 2.0) is controlled to be ≤ 0.03, which can form a cladding structure with a low refractive index contrast, reducing the waveguide mode confinement loss (≤ 0.01 dB / cm), and at the same time avoiding mode leakage caused by a high refractive index difference (leakage loss ≤ 1×10 - 4 dB / mm); the coating thickness of 50 - 200 nm ensures that the waveguide mode is effectively confined within the core layer (mode field confinement factor ≥ 95%). Combining with the control of surface flatness ≤ 1 nm, it reduces the scattering loss during light transmission (scattering loss caused by roughness ≤ 0.005 dB / cm), maintaining the stability of the transmitted light field (phase error ≤ 2° / cm); the process window of deposition temperature of 200 - 350 °C and radio frequency power of 50 - 150 W forms a dense silica layer (density ≥ 2.2 g / cm 3 , hydrogen content ≤ 1 at%), effectively blocking environmental water vapor (permeability ≤ 10 - 12 g / (cm·s·Pa)) and ion contamination (Na + concentration ≤ 1×1012 cm -3 ), protecting the underlying waveguide layer from chemical corrosion (corrosion rate ≤ 0.01 nm / year); the thermal expansion coefficients of the cladding and the waveguide layer are matched (difference ≤ 1×10 -6 / °C), suppressing the interfacial stress (thermal stress ≤ 50 MPa) during temperature cycling (-40°C to 85°C), avoiding crack generation (crack density ≤ 0.1 cm -1 ), or delamination defects (interfacial binding energy ≥ 2 J / m 2 ), ensuring long-term environmental reliability (lifetime ≥ 10 years); the ellipsometer monitors the cladding thickness (accuracy ≤ 1 nm) and refractive index (accuracy ≤ 0.001) in real time, providing a uniform substrate (thickness uniformity ≤ ±1%) for the subsequent lithography process, ensuring the transfer accuracy (line width error ≤ 5 nm) of nanoscale patterns (such as grating period ≤ 200 nm); the cladding with a surface flatness ≤ 1 nm can be directly used as a bonding interface (such as plasma-activated bonding with a silicon substrate), the void ratio of the bonding interface ≤ 0.5%, the bonding strength ≥ 100 MPa, meeting the requirements of the three-dimensional integration process for surface quality (such as the alignment error of multi-layer waveguide stacking ≤ 20 nm); the anti-laser damage threshold of the dense silica cladding (≥ 100 MW / cm 2 ) is higher than the optical power density requirement in the communication band (1550 nm), avoiding material damage (damage probability ≤ 10 -6 ) caused by high-power light transmission (such as ≥ 100 mW); the ultraviolet light absorption coefficient of the cladding (≤ 0.1 cm - -1 at a wavelength of 300 nm) and the visible light transmittance (≥ 95%) meet the requirements of wide-spectrum applications, and the refractive index drift ≤ 0.0005 under long-term light illumination (1000 hours, 1 sun intensity), ensuring the stability of optical performance (insertion loss fluctuation ≤ 0.05 dB).

[0056] In S8, the simulation method is the optical simulation technology, and the optical simulation technology uses the beam propagation method. The calculation formula is:

[0057]

[0058] where λ is the working wavelength, with a value of 1.3 - 1.55 μm, k0 = 2π / λ is the vacuum wave number, n eff is the effective refractive index of the waveguide, n is the refractive index of the waveguide material, and the simulation grid accuracy ≤ λ / 20; the thermal stress simulation uses an elastoplastic mechanics model, the calculation step size is 0.1 - 1 μm, and the von Mises stress ≤ 50 MPa.

[0059] Specifically, the beam propagation method (BPM) is used to solve the wave equation, through the formula It can simulate the optical field distribution at the working wavelength of 1.3 - 1.55 μm (such as single-mode / multimode transmission modes). Through spatial discretization with a grid accuracy ≤ λ / 20 (i.e., ≤ 77.5 nm), it ensures that the mode field calculation error ≤ 2%, accurately predicts waveguide bending loss (≤ 0.1 dB / cm @ R = 5 mm), coupling efficiency (≥ 95% @ end-face spacing ≤ 100 nm), and mode crosstalk (≤ -30 dB @ adjacent waveguide spacing ≥ 2 μm); the analysis of the difference between the effective refractive index n eff and the refractive index N of the material can optimize the waveguide cross-sectional dimensions (such as the ridge waveguide width of 0.5 - 2 μm and the ratio of height to lithium niobate layer thickness of 0.6 - 0.9), ensure the single-mode transmission condition (normalized frequency V ≤ 2.405), and suppress the excitation of higher-order modes (excitation probability ≤ 10 -4 ); Based on the elastoplastic mechanics model for thermal stress simulation, it analyzes the stress concentration regions of micro-nano structures (such as serpentine microchannels, stacked gratings) with a calculation step size of 0.1 - 1 μm, controls the von Mises stress ≤ 50 MPa, which is lower than the yield strength of the material (such as the yield strength of silicon ≈ 700 MPa), and avoids plastic deformation (strain ≤ 0.01%) and crack initiation (crack initiation stress ≥ 60 MPa); combined with the design of thermal expansion coefficient matching (such as the difference in thermal expansion coefficient between the composite heat dissipation layer and the silicon substrate ≤ 1×10 -6 / ℃), the simulation can predict the evolution of interface stress under temperature cycling (-40℃ to 85℃), ensure that the bonding strength of the bonding interface (such as the plasma-activated bonding interface) ≥ 100 MPa, and suppress delamination defects (interface energy release rate ≤ 0.5 J / m 2 ); The optical simulation results can inversely guide the setting of process parameters such as waveguide layer thickness (such as lithium niobate layer thickness of 500 - 1000 nm) and cladding refractive index (the difference from the waveguide layer ≤ 0.03). For example, by adjusting the PECVD radio frequency power (50 - 150 W), the refractive index of the silica cladding is controlled to make the mode field confinement factor ≥ 95%, while controlling the transmission loss (including scattering / absorption) ≤ 0.5 dB / cm; The thermal stress simulation can optimize the heat dissipation structure parameters (such as the serpentine microchannel width of 5 - 10 μm and the spacing of 20 - 50 μm). Combined with graphene-water nanofluid (thermal conductivity of 2.5 - 5.0 W / (m·K)), it predicts the effect of reducing the junction temperature when the heat dissipation coefficient ≥ 1000 W / (m 2 ·K) (such as the chip hot spot temperature ≤ 60℃ @ 10 W / cm 2 thermal power consumption), and avoids the influence of thermo-optic refractive index drift (Δn ≤ 1×10 -4 ) on optical performance; Combining the optical and thermal stress simulation results, the thermo-optic effect (such as the thermo-optic coefficient of silicon of 1.86×10 -5 / ℃) on the effective refractive index of the waveguide can be quantified (Δn eff ≤ 5×10 -4(@ΔT = 20 °C), evaluate the wavelength shift caused by temperature drift (such as the wavelength drift of the ring resonator ≤ 0.1 nm / °C), and provide parameter basis for the design of the temperature control system (such as the temperature control accuracy ≤ 0.1 °C); the simulation results can verify the deformation of the nanoimprint template (the hardness of nickel-cobalt alloy is 500 - 800 HV) under the imprinting pressure of 5 - 15 MPa (elastic deformation ≤ 5 nm), ensure that the sidewall roughness of the ridge waveguide ≤ 5 nm, and avoid pattern distortion caused by plastic deformation of the template (line width error ≤ 10 nm).

[0060] In S5, the waveguide array spacing is modulated by the superlattice structure, the superlattice period a = (1.5 - 2.5)w (w is the waveguide width), the slot line depth is 1 / 3 - 1 / 2 of the ridge waveguide height, and the etching rate of 20 - 50 nm / min is achieved by radio frequency plasma etching; the thermal cycle test conditions are -40 °C to 85 °C, 100 cycles, the waveguide loss change ≤ 0.1 dB, and the displacement ≤ 5 nm.

[0061] Specifically, through the design of the superlattice structure with the period a = (1.5 - 2.5)w (w is the waveguide width), the coupling strength and mode distribution between waveguides are regulated, adjacent waveguide crosstalk is suppressed (crosstalk suppression ≥ 25 dB), and signal interference is avoided; the slot line depth is 1 / 3 - 1 / 2 of the ridge waveguide height, the optical isolation between waveguides is precisely controlled, the light field confinement effect is optimized, the transmission loss is reduced (transmission loss ≤ 0.5 dB / cm), and stable transmission of optical signals is ensured; the radio frequency plasma etching rate is controlled at 20 - 50 nm / min, combined with the nanoimprint lithography technology, precise control of the sidewall perpendicularity (≥ 85°) and roughness (≤ 5 nm) of the ridge waveguide is achieved, and the light field scattering loss is reduced; the stable etching rate ensures the dimensional consistency of the waveguide array structure (dimensional error ≤ ±5%), improves the yield of mass production (yield ≥ 90%), and meets the requirements of industrial manufacturing; the thermal cycle test from -40 °C to 85 °C for 100 times simulates extreme environments and examines the reliability of the waveguide array under temperature changes; the indexes of waveguide loss change ≤ 0.1 dB and displacement ≤ 5 nm ensure good matching of the thermal expansion coefficients of the structure, avoid failure problems such as waveguide deformation and interlayer separation caused by thermal stress, and extend the service life of the device (working life ≥ 10 years); the optimized waveguide array spacing and superlattice structure can meet the requirements of multiple scenarios such as optical communication and optical computing, support high-density integration and complex optical path design, and at the same time, through the thermal cycle reliability verification, ensure the stable operation of the device in a wide temperature environment and improve the environmental adaptability of the product.

[0062] Although the embodiments of the present invention have been shown and described, those of ordinary skill in the art can understand that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A compact hybrid optical path high-density integration method for a multi-level waveguide composite structure, characterized in that It includes the following steps: S1. Preparation of multi-material layers: Deposit a silicon layer, a silicon nitride layer, and a lithium niobate layer in sequence. The silicon layer and the silicon nitride layer are prepared by chemical vapor deposition process, and the lithium niobate layer is prepared by DC magnetron sputtering. The refractive indices of the silicon layer, the silicon nitride layer, and the lithium niobate layer satisfy a gradient distribution; S2. Design of interlayer coupling structure: Form a tapered waveguide coupler at the interface of adjacent material layers to achieve interlayer optical field transmission; S3. Integration of stress control layer: Set a buffer layer with a matching coefficient of thermal expansion between the silicon layer and the silicon nitride layer; S4. Multi-layer stacking and alignment: Complete the stacking of the multi-layer structure through bonding technology, and use a laser interference alignment system to achieve interlayer alignment; S5. Processing of waveguide array: Use nanoimprint lithography technology to form a ridge waveguide array on the top lithium niobate layer; S6. Integration of thermal management structure: Prepare a composite heat dissipation layer on the back of the bottom silicon layer; S7. Optimization of optical loss: Perform surface coating treatment on the coupler area; S8. Verification of multi-physical fields: Analyze the interlayer optical field distribution and thermal stress state through simulation means, and adjust the structural parameters until the design requirements are met.

2. The compact hybrid optical path high-density integration method for the multi-level waveguide composite structure according to claim 1, characterized in that: In S1, the thickness of the silicon layer is 0.2 - 2μm, the thickness of the silicon nitride layer is 0.1 - 1.5μm and the deposition rate is controlled at 0.5 - 2nm / s, the thickness of the lithium niobate layer is 0.3 - 3μm, which is prepared by DC magnetron sputtering, the sputtering power is 80 - 150W, and the argon pressure is 0.5 - 1.2Pa.

3. The compact hybrid optical path high-density integration method for the multi-level waveguide composite structure according to claim 1, characterized in that: In S2, the length of the tapered waveguide coupler is 5 - 50μm, the width gradient rate is 0.1 - 1μm / μm, and an exponential tapered profile is adopted; when using a grating coupler, the grating period is 200 - 500nm, the duty cycle is 0.3 - 0.7, and the etching depth is 60% - 80% of the thickness of the adjacent silicon nitride layer.

4. The compact hybrid optical path high-density integration method for the multi-level waveguide composite structure according to claim 3, wherein In S3, the buffer layer is a titanium dioxide-silicon dioxide composite layer with a thickness of 20 - 50 nm, a silicon dioxide doping ratio of 5% - 15% atomic ratio, deposited by ion beam sputtering, with a sputtering current of 50 - 100 mA, a deposition temperature of 100 - 200 °C, and a thermal expansion coefficient of 8 - 12×10 -6 / °C, and the difference in thermal expansion coefficient from the silicon layer is ≤1×10 -6 / °C.

5. The compact hybrid optical path high-density integration method for the multi-level waveguide composite structure according to claim 1, characterized in that In S4, the laser interference alignment system uses a helium-neon laser with a wavelength of 632.8nm, the alignment accuracy ≤20nm, the alignment mark is a stacked grating of the silicon layer and the silicon nitride layer, the period is 2μm, and the height is 200 - 500nm; the bonding process uses plasma-activated bonding, the oxygen plasma power is 100 - 300W, the bonding pressure is 0.1 - 0.5MPa, the bonding time is 5 - 15min, and the interfacial void ratio ≤0.5%.

6. The compact hybrid optical path high-density integration method for a multi-level waveguide composite structure according to claim 1, characterized in that: In S5, the template in the nanoimprint lithography technology is a nickel-cobalt alloy, the hardness is 500 - 800HV, the imprinting pressure is 5 - 15MPa, the imprinting temperature is 80 - 120℃, the template release uses critical CO2 drying, the pressure is 7.38MPa, the temperature is 31.1℃, the width of the ridge waveguide is 0.5 - 2μm, the ratio of the height to the thickness of the lithium niobate layer is 0.6 - 0.9, and the sidewall roughness ≤5nm.

7. The compact hybrid optical path high-density integration method for a multi-level waveguide composite structure according to claim 1, characterized in that: In the S6, the composite heat dissipation layer includes a copper-nickel alloy layer with a thickness of 50 - 200 nm, a copper-nickel atomic ratio of 7:3, an interface roughness ≤ 2 nm, and a serpentine microchannel etched on the back, with a width of 5 - 10 μm, a depth of 2 - 5 μm, and a spacing of 20 - 50 μm. The channel is filled with graphene-water nanofluid, with a graphene concentration of 0.01 - 0.1 wt%, a thermal conductivity of 2.5 - 5.0 W / (m 2 ·K), and a heat dissipation coefficient ≥ 1000 W / (m 8. The compact hybrid optical path high-density integration method for the multi-level waveguide composite structure according to claim 1, wherein: In S7, the surface coating treatment uses PECVD technology to deposit a silica layer, the deposition temperature is 200 - 350℃, the RF power is 50 - 150W, the coating thickness is 50 - 200nm, which is measured and controlled by an ellipsometer, the difference in refractive index from the waveguide layer ≤0.03, and the surface flatness ≤1nm.

9. The compact hybrid optical path high-density integration method for a multi-level waveguide composite structure according to claim 1, characterized in that: In S8, the simulation means is optical simulation technology, and the optical simulation technology uses the beam propagation method. The calculation formula is: where λ is the working wavelength, with a value range of 1.3 - 1.55 μm, k0 = 2π / λ is the vacuum wave number, n eff is the effective refractive index of the waveguide, n is the refractive index of the waveguide material, and the simulation grid accuracy ≤ λ / 20; the thermo - stress simulation uses an elastoplastic mechanics model, the calculation step size is 0.1 - 1 μm, and the von Mises stress ≤ 50 MPa.

10. The compact hybrid optical path high-density integration method for a multi-level waveguide composite structure according to claim 1, wherein: In the above S5, the waveguide array pitch is modulated by a superlattice structure, and the superlattice period a = (1.5 - 2.5)w (where w is the waveguide width, the slot line depth is 1 / 3 - 1 / 2 of the ridge waveguide height, and the etching rate is 20 - 50 nm / min achieved by radio frequency plasma etching); the thermal cycle test conditions are -40°C to 85°C, with 100 cycles, the waveguide loss change ≤ 0.1 dB, and the displacement ≤ 5 nm.

Images