Compact cross-band mode division multiplexer based on sub-wavelength structure and optimization method thereof

Abstract

The invention discloses a compact cross-band mode division multiplexer based on a sub-wavelength structure and an optimization method of the compact cross-band mode division multiplexer, and belongs to the technical field of photonic devices and integration. The waveguide layer comprises a mode multiplexer composed of a plurality of cascaded directional coupling waveguides; each directional coupling waveguide comprises a narrow access waveguide positioned on the side and a multimode bus waveguide positioned in the center; the interval region of the narrow access waveguide and the multi-mode bus waveguide is of a sub-wavelength grating structure; the sub-wavelength grating structure is composed of a plurality of waveguides which are equal in period, equal in duty ratio and different in width, and each waveguide is optimally distributed through a reverse optimization method. According to the invention, the width and position distribution of the sub-wavelength grating array can be accurately controlled, the high-order mode conversion efficiency and the cross-band working capability of the mode division multiplexer are improved, meanwhile, the requirements of compact size and process compatibility of the mode division multiplexer are met, and at least the cross-band working bandwidth of 1500nm-2000nm is covered.

Description

Technical Field

[0001] The present invention relates to the field of photonic devices and integration technologies, in particular to a compact cross-band mode division multiplexer based on subwavelength structures and an optimization method therefor. Background Art

[0002] With the rapid development of optical communication technologies, the demand for transmission capacity and integration density in optical communication systems has been continuously increasing. Mode division multiplexing technology, by using multiple orthogonal guided modes as independent channels to transmit data, has become one of the key solutions to enhance data transmission capabilities. Currently, traditional mode division multiplexing devices still face many significant challenges in practical applications. Especially in achieving efficient conversion of multiple modes and cross-band operation, the performance of traditional devices is not satisfactory. Therefore, it is necessary to design a compact cross-band mode division multiplexer with a small overall structural size, easy mode channel expansion, and a cross-band operating bandwidth that can at least cover 1500 nm - 2000 nm. This is of great significance for designing a highly integrated on-chip mode division multiplexing integrated system. Currently, the design schemes of mode division multiplexers mainly include asymmetric directional coupler structures, Y-junction and multimode interference structures, subwavelength grating structures, and inverse-designed metawaveguide structures. The asymmetric directional coupling structure has attracted much attention in the design of on-chip mode division multiplexers due to its simple design and good scalability. However, this structure has extremely strict requirements for phase matching conditions, and the coupling length is generally in the range of dozens of micrometers to hundreds of micrometers, which greatly limits its application in highly compact on-chip systems. The Y-junction and multimode interference structure has a relatively wide bandwidth and also shows good tolerance in the manufacturing process. However, the design conditions are relatively strict, resulting in a relatively large device size and difficulty in meeting the development trend of high integration. Traditional subwavelength grating structures have certain advantages in achieving specific functions but are difficult to meet the requirements of cross-band operating devices in the field of optical communication. Although the inverse-designed metawaveguide structure is expected to achieve a small-size and high-performance mode division multiplexer, on the one hand, it highly depends on a large number of numerical simulations, which not only takes a long time to optimize but also requires a large amount of computing resources. On the other hand, this structure designed for specific modes is difficult to flexibly adapt to multi-channel multiplexing scenarios, and the mode scalability is severely limited. Generally speaking, the above technical solutions have their own advantages and disadvantages. The most crucial point is that the existing mode division multiplexing technology solutions can only achieve the working bandwidth of shorter wavelengths and lack a design solution for cross-band operation. Summary of the Invention

[0003] The technical problem to be solved by the present invention is to provide a compact cross-band mode division multiplexer based on subwavelength structure and its optimization method. By introducing an inverse optimization method, the width and position distribution of the subwavelength grating array are precisely controlled to improve the high-order mode conversion efficiency and cross-band working ability of the mode division multiplexer, while meeting the requirements of the compact size and process compatibility of the mode division multiplexer, providing key device support for on-chip densely integrated optical systems.

[0004] To solve the above technical problems, the technical solution adopted by the present invention is:

[0005] A compact cross-band mode division multiplexer based on subwavelength structure is constructed on a silicon platform on an insulating substrate, and includes a waveguide layer, a cladding layer, and a base layer from top to bottom; the waveguide layer includes a mode multiplexer composed of a plurality of cascaded directional coupling waveguides; each directional coupling waveguide includes a narrow access waveguide on the side and a multimode bus waveguide in the center; the interval region between the narrow access waveguide and the multimode bus waveguide is a subwavelength grating structure; the subwavelength grating structure is composed of a plurality of waveguides with equal periods, equal duty ratios, and different widths, and each waveguide obtains an optimal distribution using an inverse optimization method to couple the fundamental mode to the multimode bus waveguide more efficiently; the multimode bus waveguide mode channels include a TE0 mode multiplexing channel, a TE1 mode multiplexing channel, a TE2 mode multiplexing channel, and a TE3 mode multiplexing channel.

[0006] A further improvement of the technical solution of the present invention is that the materials of the waveguide layer and the base layer are both silicon, and the material of the cladding layer is silicon dioxide.

[0007] An optimization method for a compact cross-band mode division multiplexer based on subwavelength structure includes the following steps:

[0008] S1. Determine the design objectives of the compact cross-band mode division multiplexer and design a dual-taper coupling structure;

[0009] S2. Based on the dual-taper coupling structure, design a subwavelength grating structure and use it as the initial structure for inverse design;

[0010] S3. Use an inverse optimization method combining a particle swarm algorithm and a direct binary search method to optimize the subwavelength grating structure.

[0011] A further improvement of the technical solution of the present invention is that in S1, the design objective of the compact cross-band mode division multiplexer is four-mode multiplexing in the wavelength range of 1500 nm - 2000 nm, and the loss is less than or equal to 0.5 dB.

[0012] A further improvement of the technical solution of the present invention is that in S1, a dual-taper coupling structure is designed, where the width of the fundamental mode input waveguide is 0.55 μm and the coupling interval is 0.3 μm.

[0013] A further improvement of the technical solution of the present invention lies in that: in S2, the coupling region length of the TE0-TE1 mode conversion structure is 8 μm, and the width of the multimode bus waveguide is 1.2 μm; the coupling region length of the TE0-TE2 mode conversion structure is 12 μm, and the width of the multimode bus waveguide is 1.7 μm; the coupling region length of the TE0-TE3 mode conversion structure is 15 μm, and the width of the multimode bus waveguide is 2.2 μm; it is determined that the subwavelength grating structure is distributed as 60 waveguides at a period of P = 0.2 μm and an interval of α = 0.1 μm, and the y-direction center value at the starting position of each waveguide is defined as l1, l2, … l n …l 60 , and the y-direction width at the starting position of each waveguide is defined as w1, w2, … w n …w 60 .

[0014] A further improvement of the technical solution of the present invention lies in that: S3 specifically includes the following steps:

[0015] S31 uses the particle swarm optimization algorithm to optimize the subwavelength grating structure model, and sequentially adjusts the positions l n and widths w n within a preset range to search for the global optimal solution;

[0016] The number of particles in the particle swarm is set to Num_particles = 20, the weight parameter w = 0.5, the learning factor c1 = 1 controls the individual to approach its historical optimal solution, and the learning factor c2 = 1 controls the individual to approach the global optimal solution; the mode conversion efficiency of each particle is calculated through simulation and compared with the current highest efficiency; when the efficiency is improved, the individual extreme value L_bestfitness of this particle is updated and saved; further compare the individual extreme values of all particles to determine the global extreme value G_bestfitness;

[0017] The position update of the particle follows the following formula:

[0018] x i t+1 = x i t + v i t+1

[0019] Among them, the velocity update formula is:

[0020] v i t+1 = w·v i t + c1·rand1·(L_bestfitness - x i t) + c2·rand2·(G_bestfitness - x i t )

[0021] Among them, rand1 and rand2 are random numbers between 0 and 1; after updating the particle positions through multiple iterations, the algorithm will approach the global optimal solution;

[0022] S32 Optimize the sub-wavelength grating structure using the direct binary search method;

[0023] Based on the optimization of the sub-wavelength grating structure model by the particle swarm algorithm, the device is further refined by the binary search method with dynamically adjustable step size. The step size ΔL is dynamically adjustable according to the transmittance T, and the formula is as follows:

[0024]

[0025] Among them, ΔL n+1 is the step size to be updated, ΔL n is the existing step size, T n is the transmittance at this time, T n-1 is the transmittance of the previous generation.

[0026] Due to the adoption of the above technical solutions, the technical progress achieved by the present invention is:

[0027] 1. The compact cross-band mode division multiplexer proposed by the present invention can increase the working bandwidth of the mode division multiplexer by optimizing the width and position distribution of the sub-wavelength structure through the inverse design algorithm. The mode division multiplexer covers at least a cross-band working bandwidth of 1500 nm - 2000 nm.

[0028] 2. The compact cross-band mode division multiplexer proposed by the present invention can accelerate the mode coupling process through the sub-wavelength grating structure, thereby shortening the coupling length, and the mode multiplexing channels at all levels can be directly connected end to end without cascading tapered transition waveguides, further shortening the device size. The overall size of the device is small, easy to realize mode expansion, and the device preparation is compatible with the standard wafer process.

[0029] 3. This patent optimizes the sub-wavelength grating structure through an inverse optimization method combining the particle swarm algorithm and the direct binary search method. The binary search method is slow in finding the optimal solution and may also fall into a local optimal solution. Therefore, in the inverse optimization method, the particle swarm algorithm is first used to quickly find the approximate position of the global optimal solution, and then the binary search method is used for detailed optimization Brief Description of the Drawings

[0030] Figure 1 It is a three-dimensional structure schematic diagram of a compact cross-band mode division multiplexer based on a sub-wavelength structure provided in an embodiment of the present invention;

[0031] Figure 2 Schematic diagram of the planar structure and the optimized structure for TE0-TE3 mode multiplexing in the embodiments of the present invention;

[0032] Figure 3 Schematic diagram of the optimization process of the subwavelength grating structure in the embodiments of the present invention;

[0033] Figure 4 Optimized structure and performance diagram of TE0-TE1 mode multiplexing in the embodiments of the present invention;

[0034] Figure 5 Optimized structure and performance diagram of TE0-TE2 mode multiplexing in the embodiments of the present invention;

[0035] Figure 6 Optimized structure and performance diagram of TE0-TE3 mode multiplexing in the embodiments of the present invention;

[0036] Wherein, 1. waveguide layer, 2. cladding layer, 3. substrate layer, 4. TE0 mode multiplexing channel, 5. TE1 mode multiplexing channel, 6. TE2 mode multiplexing channel, 7. TE3 mode multiplexing channel. Detailed implementation manners

[0037] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments:

[0038] As Figure 1 shown, this embodiment is a compact cross-band mode division multiplexer based on a subwavelength structure, constructed on a silicon-on-insulator (SOI) platform on an insulating substrate, and including a waveguide layer 1, a cladding layer 2, and a substrate layer 3 from top to bottom; specifically, the materials of the waveguide layer 1 and the substrate layer 3 are both silicon, and the material of the cladding layer 2 is silicon dioxide. The mode channels include a TE0 mode multiplexing channel 4, a TE1 mode multiplexing channel 5, a TE2 mode multiplexing channel 6, and a TE3 mode multiplexing channel 7.

[0039] As Figure 2 shown, the main structure of this embodiment is in the waveguide layer. The waveguide layer 1 includes a mode multiplexer composed of a plurality of cascaded directional coupling waveguides; each directional coupling waveguide includes a narrow access waveguide on the side and a multimode bus waveguide in the center; wherein, the spaced area between the narrow access waveguide and the multimode bus waveguide is a subwavelength grating structure; the subwavelength grating structure is composed of a plurality of waveguides with equal periods, equal duty cycles, and different widths, and each waveguide uses a reverse optimization method to obtain an optimal distribution, so that the fundamental mode can be more efficiently coupled to the multimode bus waveguide; the multimode bus waveguide supports at least four mode multiplexings of TE0-TE0, TE0-TE1, TE0-TE2, and TE0-TE3.

[0040] As Figure 3As shown, an optimization method for a compact cross-band mode division multiplexer with subwavelength structure includes the following steps:

[0041] S1. Determine the design objectives of the compact cross-band mode division multiplexer and design a dual-taper coupling structure;

[0042] Determine that the design objectives of the compact cross-band mode division multiplexer are four-mode multiplexing in the wavelength range of 1500 nm - 2000 nm with a loss less than or equal to 0.5 dB; and design a dual-taper coupling structure, where the width of the fundamental mode input waveguide is 0.55 μm and the coupling interval is 0.3 μm;

[0043] S2. Based on the dual-taper coupling structure, design a subwavelength grating structure as the initial structure for inverse design;

[0044] Among them, the coupling region length of the TE0-TE1 mode conversion structure is 8 μm, and the width of the multimode bus waveguide is 1.2 μm; the coupling region length of the TE0-TE2 mode conversion structure is 12 μm, and the width of the multimode bus waveguide is 1.7 μm; the coupling region length of the TE0-TE3 mode conversion structure is 15 μm, and the width of the multimode bus waveguide is 2.2 μm; determine that the subwavelength grating structure is distributed with 60 waveguides at a period of P = 0.2 μm and an equal interval of α = 0.1 μm, and the y-direction center value at the starting position of each waveguide is defined as l1, l2,...l n …l 60 , and the y-direction width at the starting position of each waveguide is defined as w1, w2,...w n …w 60 ;

[0045] S3. Use an inverse optimization method combining the particle swarm algorithm and the direct binary search method to optimize the subwavelength grating structure;

[0046] S3 specifically includes the following steps:

[0047] S31. Use the particle swarm algorithm to optimize the subwavelength grating structure model, and sequentially adjust the positions l n and widths w n within the preset range to search for the global optimal solution;

[0048] Set the number of particles in the particle swarm to Num_particles = 20, the weight parameter w = 0.5, the learning factor c1 = 1 to control the individual to approach its historical optimal solution, and the learning factor c2 = 1 to control the individual to approach the global optimal solution; calculate the mode conversion efficiency of each particle through simulation and compare it with the current highest efficiency; when the efficiency improves, update and save the individual extreme value L_bestfitness of the particle; further compare the individual extreme values of all particles to determine the global extreme value G_bestfitness;

[0049] The position update of the particle follows the following formula:

[0050] x i t+1 = x i t + v i t+1

[0051] where the velocity update formula is:

[0052] v i t+1 = w·v i t + c1·rand1·(L_bestfitness - x i t ) + c2·rand2·(G_bestfitness - x i t )

[0053] where rand1 and rand2 are random numbers between 0 and 1; after updating the particle position through multiple iterations, the algorithm will approach the global optimal solution;

[0054] S32 uses the direct binary search method to optimize the sub-wavelength grating structure;

[0055] On the basis of optimizing the sub-wavelength grating structure model by the particle swarm algorithm, the device is further refined and optimized by the binary search method with dynamically adjustable step size. The step size ΔL is dynamically adjustable according to the transmittance T, and the formula is as follows:

[0056]

[0057] where ΔL n+1 is the step size to be updated, ΔL n is the existing step size, T n is the transmittance at this time, and T n-1 is the transmittance of the previous generation;

[0058] Through the preliminary optimization of the particle swarm algorithm, the result can approach the global optimal solution, effectively preventing the risk of falling into the local optimal solution when directly using the binary search method for optimization; by further refining the step size of the optimization, the global optimal solution can be found more efficiently. Thus, the optimal design of the sub-wavelength grating structure is realized.

[0059] With this technical solution, the width distribution of the waveguide array is changed one by one through reverse design to obtain an optimized sub-wavelength grating structure. Under the control of the sub-wavelength grating structure, the coupling length can be shortened to achieve adiabatic transmission of each order mode, replacing the cascaded tapered waveguide, and further shortening the device size; the structure can cover a cross-band working bandwidth of 1500nm - 2000nm. Compared with the existing technology, the overall structure size of this on-chip mode division multiplexing device is small, easy to achieve mode expansion, the device fabrication is compatible with the standard chip manufacturing process, and it can be widely applied to on-chip ultra-wideband and cross-band mode division multiplexing transmission systems, providing key device support for on-chip dense integrated optical systems.

[0060] Figure 4 In [figure], (a) is a schematic planar structure diagram of the TE0-TE1 mode conversion of the cross-band mode division multiplexer according to the embodiment of the present invention; the three-dimensional finite-difference time-domain method is used to simulate and analyze the device, (b) is the simulated transmission spectrum diagram of the TE0 mode input light in the 1500nm - 2000nm band under the TE0-TE1 mode conversion structure of the embodiment of the present invention; (c) is the optical field distribution of the designed multi-mode bus waveguide at 1550nm, 1750nm, and 1950nm under the TE0-TE1 mode conversion structure of the embodiment of the present invention when the TE0 mode light is input. It can be seen that when the TE0 fundamental mode is incident from the narrow access waveguide and transmitted to the coupling region, due to the design of the sub-wavelength grating structure through the optimization algorithm, the TE0 fundamental mode can maintain high-efficiency mode coupling and conversion to the TE1 mode in the multi-mode bus waveguide within the 1500nm - 2000nm band range.

[0061] Figure 5 In [figure], (a) is a schematic planar structure diagram of the TE0-TE2 mode conversion of the cross-band mode division multiplexer according to the embodiment of the present invention; the three-dimensional finite-difference time-domain method is used to simulate and analyze the device, (b) is the simulated transmission spectrum diagram of the TE0 mode input light in the 1500nm - 2000nm band under the TE0-TE2 mode conversion structure of the embodiment of the present invention; (c) is the optical field distribution of the designed multi-mode bus waveguide at 1550nm, 1750nm, and 1950nm under the TE0-TE2 mode conversion structure of the embodiment of the present invention when the TE0 mode light is input. It can be seen that when the TE0 fundamental mode is incident from the narrow access waveguide and transmitted to the coupling region, due to the design of the sub-wavelength grating structure through the optimization algorithm, the TE0 fundamental mode can maintain high-efficiency mode coupling and conversion to the TE2 mode in the multi-mode bus waveguide within the 1500nm - 2000nm band range.

[0062] Figure 6Among them, (a) is a schematic plan view of the TE0-TE3 mode conversion of the cross-band mode division multiplexer according to an embodiment of the present invention; the three-dimensional finite difference time domain method is used to simulate and analyze the device, and (b) is the TE0 mode input light in the 1500 nm - 2000 nm band under the TE0-TE3 mode conversion structure according to an embodiment of the present invention. The simulated transmission spectrum diagram; (c) is the optical field distribution of the TE0 mode light input at 1550 nm, 1750 nm, and 1950 nm of the multimode bus waveguide designed under the TE0-TE3 mode conversion structure according to an embodiment of the present invention. It can be seen that when the TE0 fundamental mode is incident from the narrow access waveguide and transmitted to the coupling region, due to the design of the sub-wavelength grating structure through the optimization algorithm, the TE0 fundamental mode can maintain high-efficiency mode coupling and conversion to the TE3 mode in the multimode bus waveguide within the 1500 nm - 2000 nm band range.

[0063] In summary, the design scheme of the present invention has mode scalability, and the cross-band mode division multiplexer supports at least four mode multiplexing of TE0-TE0, TE0-TE1, TE0-TE2, and TE0-TE3; and the working band range is the cross-band region of 1500 nm - 2000 nm, and it is expected to be widely used in ultra-wideband and cross-band on-chip mode division multiplexing systems.

Claims

1. A compact cross-band mode division multiplexer based on subwavelength structures, constructed on a silicon platform on an insulating substrate, comprising a waveguide layer (1), a cladding layer (2), and a base layer (3) from top to bottom; characterized in that: The waveguide layer (1) includes a mode multiplexer composed of a plurality of cascaded directional coupled waveguides; each directional coupled waveguide includes a narrow access waveguide located on the side and a multimode bus waveguide located in the center; the interval region between the narrow access waveguide and the multimode bus waveguide is a sub-wavelength grating structure; the sub-wavelength grating structure is composed of a plurality of waveguides with equal periods, equal duty cycles, and different widths, and each waveguide is obtained by an inverse optimization method to obtain an optimal distribution; the multimode bus waveguide mode channels include a TE0 mode multiplexing channel (4), a TE1 mode multiplexing channel (5), a TE2 mode multiplexing channel (6), and a TE3 mode multiplexing channel (7).

2. The compact cross-band mode division multiplexer based on subwavelength structure according to claim 1, characterized in that: The materials of the waveguide layer (1) and the base layer (3) are both silicon, and the material of the cladding layer (2) is silica.

3. An optimization method for a compact cross-band mode division multiplexer based on subwavelength structures, characterized in that: Applied to the compact cross-band mode division multiplexer based on the sub-wavelength structure as described in any one of claims 1 to 2, it includes the following steps: S1. Determine the design objectives of the compact cross-band mode division multiplexer and design a dual-taper coupling structure; S2. Based on the dual-taper coupling structure, design a sub-wavelength grating structure and use it as the initial structure for inverse design; S3. Use an inverse optimization method combining a particle swarm algorithm and a direct binary search method to optimize the sub-wavelength grating structure.

4. The optimization method of the compact cross-band mode division multiplexer based on sub-wavelength structure according to claim 3, characterized in that: In S1, the design objective of the compact cross-band mode division multiplexer is four-mode multiplexing in the wavelength range of 1500 nm - 2000 nm, and the loss is less than or equal to 0.5 dB.

5. The optimization method of the compact cross-band mode division multiplexer based on sub-wavelength structure according to claim 3, characterized in that: In S1, design a dual-taper coupling structure, where the width of the fundamental mode input waveguide is 0.55 μm and the coupling interval is 0.3 μm.

6. The optimization method of the compact cross-band mode division multiplexer based on sub-wavelength structure according to claim 3, characterized in that: In S2, the coupling region length of the TE0-TE1 mode conversion structure is 8 μm, and the width of the multimode bus waveguide is 1.2 μm; The coupling region length of the TE0-TE2 mode conversion structure is 12 μm, and the width of the multimode bus waveguide is 1.7 μm; The length of the coupling region for TE0 - TE3 mode conversion structure is 15 μm, and the width of the multimode bus waveguide is 2.2 μm; the sub - wavelength grating structure is determined to be distributed with a period of P = 0.2 μm and an interval of α = 0.1 μm into 60 waveguides. The y - direction center value of the starting position of each waveguide is defined as l1, l2, … l n …l 60 , and the y - direction width of the starting position of each waveguide is defined as w1, w2, … w n …w 60 .

7. The optimization method of the compact cross-band mode division multiplexer based on subwavelength structure according to claim 6, characterized in that: S3 specifically includes the following steps: S31 uses the particle swarm optimization algorithm to optimize the sub-wavelength grating structure model, and sequentially adjusts the values of the position l n and the width w n within the preset range to search for the global optimal solution; The number of particles in the particle swarm is set to Num_particles = 20, the weight parameter w = 0.5, the learning factor c1 = 1 controls the individual to approach its historical optimal solution, and the learning factor c2 = 1 controls the individual to approach the global optimal solution; calculate the mode conversion efficiency of each particle through simulation and compare it with the current highest efficiency; when the efficiency is improved, update and save the individual extreme value L_bestfitness of the particle; further compare the individual extreme values of all particles to determine the global extreme value G_bestfitness; The position update of the particle follows the following formula: x i t+1 = x i t + v i t+1 Among them, the velocity update formula is: v i t+1 = w·v i t + c1·rand1·(L_bestfitness - x i t ) + c2·rand2·(G_bestfitness - x i t ) Among them, rand1 and rand2 are random numbers between 0 and 1; after updating the particle position through multiple iterations, the algorithm will approach the global optimal solution; S32 uses a direct binary search method to optimize the sub-wavelength grating structure; On the basis of the particle swarm algorithm optimizing the sub-wavelength grating structure model, a more refined optimization of the device is carried out through a binary search method with a dynamically adjustable step size. The step size ΔL is dynamically adjustable according to the transmittance T, and the formula is as follows: Among them, ΔL n+1 is the step length to be updated, and ΔL n is the existing step length, T n is the transmittance at this time, and T n-1 is the transmittance of the previous generation.

Images