Ultra-compact reconfigurable light beam splitter based on reverse design

Abstract

The invention discloses an ultra-compact reconfigurable light beam splitter based on reverse design, and the device comprises a multimode interference coupler which is realized based on a ridge waveguide and comprises a phase change material region, an N-type doped silicon waveguide and a P-type doped silicon waveguide, and the N-type doped silicon waveguide and the P-type doped silicon waveguide are arranged at the two sides of the phase change material region. The phase change material region is arranged on the ridge of the ridge waveguide and is obtained through reverse design; the metal electrode is used for applying voltage to the N-type doped silicon waveguide and the P-type doped silicon waveguide on the two sides of the phase change material area, so that the phase state of the phase change material in the phase change material area is changed; and the conical input waveguide and the conical output waveguide are arranged at the input end and the output end of the multimode interference coupler. The fixed light beam splitting mode of a traditional device is broken through, the light beam splitter integrates the optical regulation and control characteristics of the phase change material and the degree of freedom of a reversely designed structure, and an innovative solution with high dynamic range, low insertion loss and chip-level integration is provided for an optical interconnection system.

Description

Technical Field

[0001] This invention relates to the field of integrated optical interconnects, and in particular to an ultra-compact reconfigurable optical beam splitter based on reverse design. Background Technology

[0002] In high-speed optical interconnect systems, optical beamsplitters primarily function to rationally allocate optical power, serving as core components for intelligent power management, link quality monitoring, and fault diagnosis. While traditional optical beamsplitters based on fixed beam splitting ratios (such as Y-branch and MMI couplers) can achieve power distribution, they suffer from inherent limitations: the fixed beam splitting ratio cannot adapt to the dynamic power balancing requirements of multiple nodes, leading to wasted signal margins or receiver overload. Especially in data center optical interconnect scenarios, existing devices struggle to meet stringent requirements such as millisecond-level power reconfiguration and nanosecond-level fault response, severely hindering improvements in system energy efficiency and reliability.

[0003] To overcome the limitations of fixed beam splitting ratios, current optical beam splitters are designed by deducing their structural composition based on optical theories and extensive practical experience. However, relying on traditional analytical design methods suffers from high computational complexity, limited analytical solutions, and long processing times, resulting in large device sizes that hinder further integration and expansion, failing to meet the industry's current demands for integration and miniaturization. Furthermore, while traditional reverse engineering methods can optimize device structures, they are often limited to single material systems and lack the ability to synergistically optimize materials with dynamically adjustable refractive indices, leading to insufficient flexibility in functional reconfiguration. Summary of the Invention

[0004] To address the problems existing in the background art, the purpose of this invention is to disclose an ultra-compact reconfigurable optical beamsplitter based on reverse design, which replaces the traditional method of manually adjusting structural parameters. It uses reverse design to optimize structural parameters and combines the modulation of phase change materials to achieve dynamic optical beam splitting power detection of the entire optical signal path in a reconfigurable optical network / system.

[0005] According to a first aspect of the embodiments of this application, an ultracompact reconfigurable optical beam splitter based on reverse design is provided, comprising:

[0006] A multimode interference coupler, which is based on a ridge waveguide, includes a phase change material region and N-type doped silicon waveguides and P-type doped silicon waveguides on both sides of the phase change material region. The phase change material region is disposed on the ridge of the ridge waveguide and is obtained through reverse design.

[0007] Metal electrodes are used to apply voltage to N-type doped silicon waveguides and P-type doped silicon waveguides on both sides of the phase change material region, thereby changing the phase state of the phase change material in the phase change material region.

[0008] a tapered input waveguide and a tapered output waveguide arranged at an input end and an output end of the multimode interference coupler.

[0009] Further, a voltage driving circuit is further included, which is connected with the metal electrode and used to apply a voltage to the N-doped silicon waveguide and the P-doped silicon waveguide on both sides of the phase change material region.

[0010] Further, a plurality of grating couplers are further included, each of which is connected with the tapered input waveguide and the tapered output waveguide and used to connect the multimode interference coupler to an optical fiber.

[0011] Further, in the inverse design process, a density-based topology algorithm, a gradient-based optimization algorithm, a non-gradient-based optimization algorithm or a deep learning-based data-driven algorithm is used, the gradient-based optimization algorithm includes a concomitant method and a level set method, and the non-gradient-based optimization algorithm includes a genetic algorithm, a particle swarm algorithm, a binary search algorithm and a simulated annealing algorithm.

[0012] Further, the density-based topology algorithm specifically includes:

[0013] S1: selecting a multimode interference coupler as an initial structure;

[0014] S2: determining a design region based on a coupling region in the initial structure, covering the phase change material on the design region, and randomly generating an N×M two-dimensional pixel structure of the design region;

[0015] S3: optimizing the material properties of each unit in the two-dimensional pixel structure by parameterizing the dielectric constant, and assigning the refractive index of the phase change material in different phase states to different units;

[0016] S4: taking the design parameters of the design region as the independent variable, constructing a target function based on the refractive index in different phase states, respectively performing forward simulation calculation and concomitant simulation calculation, and solving the gradient of the target function;

[0017] S5: based on the gradient of the target function, approaching the required target function value according to gradient descent optimization, and obtaining the corresponding design parameters of the design region;

[0018] S6: repeating steps S3-S5 to iteratively optimize the two-dimensional pixel structure.

[0019] Further, after S6, further comprising:

[0020] S7: performing binary discrete processing on the optimized two-dimensional pixel structure.

[0021] Further, in step S2, the N×M two-dimensional pixel structure of the design region is randomly generated, specifically:

[0022] According to the size of the design region and the process level, the design region is divided into an N*M grid;

[0023] Each grid unit is randomly assigned a value of 0 / 1, where 0 represents that the unit material attribute is air or silicon dioxide, and 1 represents that the unit material attribute is a phase change material.

[0024] Further, in step S3, the permittivity is parameterized using a density interpolation method, and the refractive index of the phase change material in different phase states is assigned, specifically:

[0025] The two-dimensional pixel structure is taken as the original design parameter to obtain a filtering field;

[0026] A projection field is obtained by using a differentiable nonlinear function through a heaviside projection function;

[0027] The permittivity distribution of the phase change material in different phase states is calculated through the projection field, so as to optimize the material attribute of each unit in the two-dimensional pixel structure and assign the refractive index of the phase change material in different phase states.

[0028] Further, in step S4, forward simulation calculation and adjoint simulation calculation are performed to obtain the gradient of the objective function, specifically:

[0029] The electromagnetic field distribution corresponding to the design parameter is calculated through forward simulation;

[0030] The sensitivity of the objective function to the electromagnetic field is taken as an excitation source, and the adjoint field distribution is calculated through adjoint simulation;

[0031] The gradient of the objective function is calculated by point product real part integration of the electromagnetic field distribution and the adjoint field distribution in the target region.

[0032] According to a second aspect of the embodiments of the present application, a reverse design method of the super-compact reconfigurable optical splitter is provided, and the method comprises:

[0033] The super-compact reconfigurable optical splitter of the first aspect is connected in series to a link to be detected;

[0034] The phase state of the phase change material in the phase change material region is changed, so that more than a predetermined proportion of optical power is output from the first output end of the super-compact reconfigurable optical splitter, and the optical signal output from the second output end of the super-compact reconfigurable optical splitter is detected;

[0035] If the optical signal is detected to be abnormal, the optical power more than the predetermined proportion is switched to the second output end for deep diagnosis.

[0036] The technical scheme of the present application can be summarized as: an integrated optical splitter based on dynamic regulation of phase change materials and gradient-driven reverse design, by introducing a chalcogenide phase change material (such as Sb2S3, Sb2Se3) as a reconfigurable medium, using its nanosecond amorphous-crystalline phase change characteristics (refractive index change Δn>0.5) to realize dynamic switching of optical splitting ratio; combined with the gradient optimization technology of the accompanying method, an N*M pixelized phase change unit topology is constructed, breaking through the fixed optical splitting mode of traditional devices. The optical regulation characteristics of the phase change material and the structural freedom of reverse design are fused, providing an innovative solution for optical interconnection systems with high dynamic range (optical splitting ratio 1:99~99:1), low insertion loss (<1dB) and chip-level integration, meeting the core needs of real-time reconfigurability and accurate detection for the next generation of intelligent optical networks.

[0037] The technical scheme provided by the embodiment of the present application can include the following beneficial effects:

[0038] ①The present application uses the phase change characteristics of phase change materials to realize the function of dynamic optical splitter, laying a solid scientific foundation for on-chip optical interconnection, programmable photonics and other fields.

[0039] ②The present application based on density topology optimization is a device design that can break through the traditional device based on physical theory limitations, without excessive concern about the complex physical processes occurring in the ultra-compact region. The optimized structure can not be limited to symmetrical and regular shapes, and has more freedom to find the optimal solution, making it low loss, high extinction ratio, and compact and miniaturized, paving the way for improving the integration density of integrated photonics chips.

[0040] ③Compared with active devices that require driving voltage to maintain configuration, the method based on non-volatile phase change optical devices has energy consumption advantages, thereby providing opportunities for integrated photonics platforms to reduce power consumption.

[0041] ④The present application has the synergistic optimization capability of refractive index dynamically adjustable material and high flexibility of function reconfiguration.

[0042] ⑤The present application breaks through the limitation of fixed optical splitting ratio, realizes nanosecond phase change state switching, can realize intelligent power management and fault detection, and has reconfigurability.

[0043] ⑥The present application is compatible with mainstream CMOS process, facilitating device processing and manufacturing.

[0044] It should be understood that the above general description and the following detailed description are only exemplary and explanatory, and cannot limit the present application. BRIEF DESCRIPTION OF DRAWINGS

[0045] The drawings herein are incorporated into the specification and form part of the specification, showing embodiments consistent with the present application, and together with the specification, serve to explain the principles of the present application.

[0046] Figure 1 is a structural schematic diagram of an ultra-compact reconfigurable optical splitter based on reverse design of the application

[0047] Figure 2 is a flow chart of a reverse design method based on topological optimization of the application.

[0048] Figure 3 is a simulation result schematic diagram of the ultra-compact reconfigurable optical splitter structure of the application when the phase change material is in amorphous state, wherein (a) is an electric field distribution diagram, and (b) is a corresponding spectral transmission diagram.

[0049] Figure 4 is a simulation result schematic diagram of the ultra-compact reconfigurable optical splitter structure of the application when the phase change material is in crystalline state. Wherein (a) is an electric field distribution diagram, and (b) is a corresponding spectral transmission diagram.

[0050] The reference signs: 1, input end; 2, first output end; 3, second output end; 4, N-type doped silicon waveguide; 5, phase change material region; 6, P-type doped silicon waveguide; 7, metal electrode; 8, grating coupler.

[0051] Figure 5 is a schematic diagram of a programmable optical splitting detection method of the application. DETAILED DESCRIPTION

[0052] The exemplary embodiments will be described in detail herein with reference to the attached drawings. In the following description, the same numbers are used to indicate the same or similar components. The embodiments described in the following exemplary embodiments are not intended to represent all embodiments consistent with the present application.

[0053] The terminology used in the present application is for the purpose of describing particular embodiments only and is not intended to be limiting of the present application. As used in the present application and the appended claims, the singular forms "a," "an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms "comprises" and / or "comprising," when used in this specification, specify the presence of stated features, integers, steps, operations, elements, and / or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof.

[0054] It should be understood that although the terms first, second, third, etc. can be used herein to describe various information, these terms are not intended to denote a particular order or hierarchy. These terms are used only to distinguish one from another. For example, a first information can be termed a second information, and similarly, a second information can also be termed a first information, without departing from the scope of the present application. The word "if" as used herein means "when" or "upon" or "in response to the determination" depending on the context.

[0055] As Figure 1 shown, the application provides an ultra-compact reconfigurable optical splitter based on reverse design, including a multimode interference coupler, a metal electrode 7, a tapered input waveguide and a tapered output waveguide, the multimode interference coupler is realized based on a ridge waveguide, including a phase change material region 5 and N-type doped silicon waveguide 4 and P-type doped silicon waveguide 6 on both sides of the phase change material region 5, the phase change material region 5 is arranged on the ridge of the ridge waveguide and obtained by reverse design; the metal electrode 7 is connected with the N-type doped silicon waveguide 4 and the P-type doped silicon waveguide 6 on both sides of the phase change material region 5 to apply voltage to change the phase state of the phase change material in the phase change material region 5; the tapered input waveguide and the tapered output waveguide are arranged at the input end 1 and the output end of the multimode interference coupler respectively.

[0056] In specific implementation, a plurality of grating couplers 8 can be further included, each grating coupler 8 is connected with the tapered input waveguide and the tapered output waveguide respectively, for connecting the multimode interference coupler to an optical fiber.

[0057] In specific implementation, the metal electrode 7 can be connected with a voltage driving circuit to apply voltage to the N-type doped silicon waveguide 4 and the P-type doped silicon waveguide 6 on both sides of the phase change material region 5.

[0058] Preferably, a chalcogenide phase change material (such as Sb2S3, Sb2Se3) is selected as the phase change material in the phase change material region 5.

[0059] The overall silicon waveguide part of the device adopts a shallow etching design, the phase change material region 5 adopts a ridge waveguide design, and the remaining regions are arranged on a flat plate waveguide. In the optical splitter, the phase change material region 5 is uniformly divided into a plurality of unit structures and optimized by reverse design, covering the ridge waveguide; the optical wave signal is transmitted to the tapered input waveguide through the grating coupler 8, passes through the ridge waveguide covered by the reverse designed phase change material region 5, and is output from the first tapered output waveguide and the second tapered output waveguide through different voltage control phase change states, and finally transmitted to the optical fiber through the grating coupler 8 for signal output detection.

[0060] In specific implementation, the state of the phase change material is not limited to two phase states of crystalline and amorphous, but also includes a partial crystalline state (mixed amorphous / crystalline region), which can realize different optical splitting ratios of the device with different refractive index differences.

[0061] The overall device is manufactured by SOI process, and the manufacturing process includes photolithography, electron beam exposure, focused ion beam etching, nanoimprint, etc.

[0062] In a specific implementation, the reverse design can adopt a density-based topology algorithm, a gradient-based optimization algorithm, a non-gradient-based optimization algorithm, a deep learning-based data-driven algorithm, etc. The gradient-based optimization algorithm includes a companion method and a level set method. The non-gradient-based optimization algorithm includes a genetic algorithm, a particle swarm algorithm, a binary search algorithm, and a simulated annealing algorithm.

[0063] For example, as shown in FIG. 2, the density-based topology algorithm first randomly initializes two-dimensional data according to the size of the design area, secondly projects the permittivity by density filtering, then respectively performs forward simulation and companion simulation to solve the gradient under the structure of different phase change material states, then updates the permittivity parameter according to the gradient descent optimization, and finally performs pixel binarization to facilitate processing and manufacturing. The specific steps are as follows: Figure 2

[0064] S1: Select a multi-mode interference coupler as an initial structure;

[0065] Specifically, a 1x2, 2x2 or 1x4 multi-mode interference coupler (MMI) can be used. In this embodiment, a ridge silicon waveguide 1x2 MMI structure based on a silicon-on-insulator (SOI) process is used as the starting point of topology optimization. The input end 1, the first output end 2 and the second output end 3 all adopt a tapered waveguide structure to realize efficient optical coupling. The MMI structure is designed based on the self-mapping theory, covering a working waveband of 1530nm-1570nm, with an initial optical splitting ratio of about 50:50, wideband characteristics (bandwidth>40nm) and low coupling loss (<0.5dB). The MMI structure is used as the initial design for subsequent topology optimization to realize efficient optical field splitting and an optimal starting point. The wideband characteristics (bandwidth>40nm) of the MMI provide a robust starting point for optimization, avoiding local optimal solutions, and providing a good convergence starting point for dynamic optical splitting ratio optimization based on phase change material control. The tapered waveguide reduces end reflection, improves optimization convergence efficiency, and is more convenient for approaching the target value.

[0066] S2: Determine a design area based on a coupling region in the initial structure, cover the design area with a phase change material, and randomly generate an NXM two-dimensional pixel structure of the design area;

[0067] ​In a specific implementation, based on the coupling region of the initial structure MMI, the design region is determined in combination with design requirements and process level. In an embodiment, the size of the design region, i.e., the size of the phase change material region 5, is 8*4 μm, the minimum size in each optimization simulation is 50 nm, the design region is divided into an N*M grid, and each grid unit is randomly assigned a value of 0 / 1, where 0 represents the material property of the unit as air / silicon dioxide, and 1 represents the material property of the unit as phase change material (Sb2Se3). The performance of the target device is adjusted by switching the crystalline / amorphous state of the phase change material of each grid in the phase change material region 5.

[0068] S3: The material properties of each unit in the two-dimensional pixel structure are optimized, and the refractive index of the phase change material in different phases of different units is assigned by parameterizing the dielectric constant.

[0069] In a specific implementation, the refractive index of the phase change material in the crystalline state, amorphous state, and partial crystalline state can be assigned according to actual requirements. In this embodiment, the crystalline state and amorphous state are taken as examples for illustration. In order to parameterize the dielectric constant, a density interpolation method can be used, the two-dimensional pixel structure randomly generated in step S2 is taken as the original design parameter p, a Gaussian filter is used, and the filtered design parameter p is obtained. , w(x) is the kernel, and * is a two-dimensional convolution. Then, the filtered field is projected by a heaviside projection function using a differentiable nonlinear function, and the projected field is: and are threshold parameters, <0.5 produces expansion, >0.5 produces erosion, is a differentiable nonlinear function. Finally, the dielectric constant required for design is represented by the following formula: is the dielectric constant of the "empty" region, is the dielectric constant of the "solid" region, different dielectric constants in different phase change states are obtained, thereby optimizing the material properties of each unit in the two-dimensional pixel structure and assigning different refractive indices n1, n2 in different phase change states.

[0070] S4: Taking the design parameters of the design region as the independent variable, constructing a target function based on the refractive index in different phases, respectively performing forward simulation calculation and adjoint simulation calculation, and solving to obtain the gradient of the target function.

[0071] In a specific implementation, the two different states of the phase change material are expressed as a maximum value optimization problem on two different target functions: is the refractive index when the phase change material is in the amorphous state,​​​​ the refractive index of the phase change material in the crystalline state, the transmittance of the first tapered output waveguide, the transmittance of the second tapered output waveguide, the design parameters (material distribution and / or geometric parameters), the objective function (transmittance of different design parameter structures in different phase change states); satisfying Maxwell's equations at different frequencies.

[0072] Preferably, in order to meet the process manufacturing constraints, the objective function is reconstructed into a differentiable maximization problem with additional nonlinear constraints.

[0073] Sequentially perform forward electromagnetic field simulation and adjoint electromagnetic field simulation, and solve the gradient of the objective function based on the simulation results;

[0074] Specifically, the following steps are implemented using Lumerical FDTD solver:

[0075] (a) Forward simulation: calculate the electromagnetic field distribution E forward corresponding to the design parameters;

[0076] (b) Adjoint simulation: take the sensitivity of the objective function to the electromagnetic field as the excitation source, and calculate the adjoint field distribution E adjoint ;

[0077] (c) Gradient generation: calculate the gradient by pointwise integration of the real part of the dot product of E forward and E adjoint in the target region, and the mathematical expression is:

[0078]

[0079] where Ω represents the design region, and Re represents the real part of a complex number.

[0080] S5: Based on the gradient of the objective function, the gradient descent optimization is performed to approach the required objective function value, and the corresponding design parameters of the design region are obtained;

[0081] In an embodiment, the first-order optimization algorithm ADAM optimizer is used, which can replace the traditional stochastic gradient descent process. By calculating the first-order estimate and second-moment estimate of the gradient, independent adaptive learning is designed by different parameters, so as to realize the gradient descent optimization.

[0082] S6: Repeat steps S3-S5 to iteratively optimize the two-dimensional pixel structure;

[0083] Returning to step S3, adjust the material properties (such as dielectric constant, phase transition state) or topological shape of the pixel unit to make the optical response approximate the target beam splitting ratio. Optimization terminates when the objective function converges to a threshold, the number of optimization iterations reaches the maximum, or the beam splitting ratio error is less than a pre-set threshold (e.g., <2%). The optimized N×M pixel structure is fabricated using photolithography or electron beam etching, where each pixel corresponds to a phase transition material state (amorphous / crystalline) or dimensional parameter.

[0084] It should be noted that steps S3-S6 above use density-based topology algorithms. However, in specific implementations, gradient-based optimization algorithms such as the adjoint method and level set method can also be used. This is also applicable to non-gradient-based optimization algorithms such as genetic algorithms, particle swarm optimization, binary search algorithms, and simulated annealing algorithms, as well as data-driven algorithms based on deep learning.

[0085] S7: Perform binarization discretization on the optimized two-dimensional pixel structure to facilitate processing;

[0086] In practice, by adjusting the state of the phase change material (amorphous / crystalline), the dynamic beam splitting ratio of the input light between the first output terminal 2 and the second output terminal 3 is adjusted. The specific implementation is as follows:

[0087] 1. Amorphous beam splitting mode (high transmission to the second output terminal 3)

[0088] Applying a voltage causes the phase change material to switch to an amorphous state (refractive index n1). The input light is mainly directed to the second output terminal 3 (beam splitting ratio P3:P2 ≈ 99:1), and the electric field Ey distribution is as follows: Figure 3 As shown in (a) in the figure.

[0089] Optical performance ( Figure 3 (b) in the figure: Bandwidth: 1537nm–1560nm (bandwidth 23nm), Insertion loss: <1 dB (0.34 dB at 1550nm), Extinction ratio: >15 dB (20 dB at 1550nm)

[0090] 2. Crystalline beam splitting mode (high transmission to first output terminal 2)

[0091] Applying voltage causes the phase change material to switch to a crystalline state (refractive index n2). The input light is mainly directed to the first output terminal 2 (beam splitting ratio P2:P3≈99:1). The electric field Ey distribution is as follows: Figure 4 As shown in (a) in the figure.

[0092] Optical performance ( Figure 4(b) in the above table): Wavelength range: 1540 nm - 1560 nm (bandwidth 20 nm), insertion loss: <1 dB (0.28 dB at 1550 nm), extinction ratio: >28 dB (42 dB at 1550 nm).

[0093] 3. Dynamic light splitting ratio control: Through partial phase change (mixed amorphous / crystalline region) or multi-level voltage control, any light splitting ratio (such as P2:P3=50:50, 30:70, etc.) of the first output end 2 and the second output end 3 is realized.

[0094] In the above embodiments, although 1550nm light waves are shown, the device provided by the present application is not limited to this one wavelength, and is also applicable to other wavelengths, including typical optical communication wavelength bands such as O-band and C-band.

[0095] The ultra-compact reconfigurable optical splitter based on reverse design proposed by the present application breaks through the design limitations of traditional "trial and error method" by fusing intelligent optimization algorithm and refractive index dynamically adjustable material characteristics, realizes intelligent optical power management and precise matching of phase change response. This method uses gradient descent topology optimization algorithm technology to automatically search for the optimal device structure that meets the target optical function in the parameter space, significantly improves the extinction ratio, response speed and energy efficiency of the device, and further reduces the size and facilitates device integration. In addition, by introducing ultra-low loss chalcogenide phase change material (such as Sb2Se3), the high absorption loss and thermal crosstalk problems of traditional phase change devices can be solved at the same time, realizing non-volatile optical regulation. This technology not only provides scalable unit devices for high-density optical interconnection chips, but also can be expanded to emerging fields such as optical storage and photonic neural networks, helping to build low-power and high-robust integrated photonics systems, providing bottom-layer hardware support for the next generation of data centers, quantum computing and 6G communication. Its core value lies in breaking through the whole chain innovation of "material characteristics-structure design-function implementation", accelerating the transformation of optical interconnection technology from laboratory prototypes to industrial applications.

[0096] Based on the foregoing ultra-compact reconfigurable optical splitter based on reverse design, the present application further provides a programmable optical splitting and detection method, comprising:

[0097] The ultra-compact reconfigurable optical splitter is connected in series to the link to be detected;

[0098] The phase state of the phase change material in the phase change material region 5 is changed, so that more than a predetermined proportion of optical power is output from the first output end 2 of the ultra-compact reconfigurable optical splitter, and the optical signal output from the second output end 3 of the ultra-compact reconfigurable optical splitter is detected;

[0099] If an abnormal optical signal is detected, the more than a predetermined proportion of optical power is switched to the second output end 3 for deep diagnosis.

[0100] Specifically, in the optical interconnection system, the intelligent detection of link quality is realized by the programmable characteristics of the phase change material, such as Figure 5 As shown, the optical signal is transmitted from the input end to the optical splitter and then to the output end to be detected by the optical splitting ratio, the phase change of the phase change material is controlled by the voltage driving circuit on the optical splitter, and finally the closed-loop regulation is realized by the field programmable gate array (FPGA), and the required pulse voltage is regulated in real time according to the optical power distribution. Under normal circumstances, the phase change material is set to be amorphous, so that more than 90% of the optical power is transmitted from the main output end (the second output end 3), and at the same time, 1-10% of the optical power is continuously collected through the monitoring port (port 2) for basic quality analysis; when abnormal signals are detected (such as power fluctuation > ± 0.5 dB or error rate rises sharply), it is switched to crystalline state within 5 ms, and 90% of the optical power is switched to the first output end 2 for deep diagnosis, and any optical splitting ratio can also be controlled to detect the link transmission quality, meet the requirements of millisecond power redistribution, nanosecond fault response, etc., and improve the energy efficiency and reliability of the system.

[0101] Other embodiments of the application will be apparent to those skilled in the art from consideration of the specification and practice of the application disclosed herein. It is intended that the application cover any and all variations of the application that come within the scope of the general inventive concepts and the following claims and that the application is not limited to any one specific embodiment.

[0102] It should be understood that the application is not limited to the precise construction that has been described above and illustrated in the accompanying drawings, and that various modifications and changes can be made by those skilled in the art without departing from the scope of the application. The scope of the application is limited only by the appended claims.

Claims

1. An ultra-compact reconfigurable optical beam splitter based on reverse design, characterized in that, include: A multimode interference coupler, which is based on a ridge waveguide, includes a phase change material region and N-type doped silicon waveguides and P-type doped silicon waveguides on both sides of the phase change material region. The phase change material region is disposed on the ridge of the ridge waveguide and is obtained through reverse design. Metal electrodes are used to apply voltage to N-type doped silicon waveguides and P-type doped silicon waveguides on both sides of the phase change material region, thereby changing the phase state of the phase change material in the phase change material region. A tapered input waveguide and a tapered output waveguide are disposed at the input and output ends of the multimode interference coupler.

2. The ultra-compact reconfigurable optical beam splitter according to claim 1, characterized in that, It also includes a voltage driving circuit for connecting to the metal electrode and applying voltage to the N-type doped silicon waveguide and P-type doped silicon waveguide on both sides of the phase change material region.

3. The ultra-compact reconfigurable optical beam splitter according to claim 1, characterized in that, It also includes several grating couplers, each of which is connected to the tapered input waveguide and the tapered output waveguide, respectively, for connecting the multimode interference coupler to the optical fiber.

4. The ultra-compact reconfigurable optical beam splitter according to claim 1, characterized in that, In the reverse engineering process, density-based topology algorithms, gradient-based optimization algorithms, non-gradient-based optimization algorithms, or data-driven algorithms based on deep learning are employed. The gradient-based optimization algorithms include the adjoint method and the level set method; the non-gradient-based optimization algorithms include genetic algorithms, particle swarm optimization, binary search algorithms, and simulated annealing algorithms.

5. The ultra-compact reconfigurable optical beam splitter according to claim 4, characterized in that, The density-based topology algorithm is specifically as follows: S1: Select a multimode interference coupler as the initial structure; S2: Determine the design region based on the coupling region in the initial structure, cover the design region with phase change material, and randomly generate an N×M two-dimensional pixel structure of the design region; S3: By parameterizing the dielectric constant, the material properties of each unit in the two-dimensional pixel structure are optimized, and the refractive index of the phase change material of different units in different phase states is assigned. S4: Using the design parameters of the design region as independent variables, construct an objective function based on the refractive index under different phase states, perform forward simulation calculations and adjoint simulation calculations respectively, and solve for the gradient of the objective function; S5: Based on the gradient of the objective function, optimize according to gradient descent to approach the desired objective function value, and obtain the corresponding design parameters of the design region; S6: Repeat steps S3-S5 to iteratively optimize the two-dimensional pixel structure.

6. The ultracompact reconfigurable optical beamsplitter according to claim 5, characterized in that, Following S6, it also includes: S7: Perform binarization discretization on the optimized two-dimensional pixel structure.

7. The ultracompact reconfigurable optical beamsplitter according to claim 5, characterized in that, In step S2, an N×M two-dimensional pixel structure of the design area is randomly generated, specifically as follows: Based on the size and manufacturing process of the design area, the design area is divided into an N×M grid; Each grid cell is randomly assigned a value of 0 or 1, where 0 represents that the cell material is air or silicon dioxide, and 1 represents that the cell material is a phase change material.

8. The ultracompact reconfigurable optical beamsplitter according to claim 5, characterized in that, In step S3, the dielectric constant is parameterized using a density interpolation method, and the refractive index of the phase change material in different phase states is assigned accordingly. Specifically: Using the two-dimensional pixel structure as the original design parameters, the filter field is obtained; The projection field is obtained by using the heaviside projection function and a differentiable nonlinear function; The dielectric constant distribution of the phase change material in different phase states is obtained by calculating the projection field, thereby optimizing the material properties of each unit in the two-dimensional pixel structure and assigning the refractive index of the phase change material in different phase states.

9. The ultracompact reconfigurable optical beamsplitter according to claim 5, characterized in that, In step S4, forward simulation and adjoint simulation are performed to obtain the gradient of the objective function, specifically: The electromagnetic field distribution corresponding to the design parameters is calculated through forward simulation. Using the sensitivity of the objective function to the electromagnetic field as the excitation source, the adjoint field distribution is calculated through adjoint simulation; The gradient of the objective function is calculated by integrating the real part of the dot product of the electromagnetic field distribution and the associated field distribution within the target region.

10. A programmable optical beam splitting detection method, characterized in that, include: The ultra-compact reconfigurable optical beam splitter according to any one of claims 1-9 is connected in series to the link to be tested; The phase state of the phase change material in the phase change material region is changed so that a predetermined proportion of optical power is output from the first output end of the ultra-compact reconfigurable optical beam splitter, and the optical signal output from the second output end of the ultra-compact reconfigurable optical beam splitter is detected. If an abnormal optical signal is detected, the optical power of a predetermined proportion or higher is switched to the second output terminal for in-depth diagnosis.

Images