A dual-mode power beam splitter

Abstract

This invention provides a dual-mode power beamsplitter, comprising a substrate with a top silicon layer. The top silicon layer includes an optimization region. An input waveguide is disposed at one end of the optimization region, and a first output waveguide and a second output waveguide are disposed at the other end. The first and second output waveguides output different modes. The optimization region is divided into N×M equally sized square units. By adjusting the state of the center of the square units, a non-periodic perforated array is formed to meet a predetermined output target. The output target refers to a 50:50 ratio of optical power between the first and second output waveguides. The dual-mode power beamsplitter provided by this invention can achieve a 50:50 power ratio beam split while simultaneously enabling mode switching, and can be widely used in mode-division multiplexing systems.

Description

Technical Field

[0001] This invention belongs to the field of micro-nano optoelectronic components technology, specifically relating to a dual-mode power beam splitter optimized based on a multi-dimensional direct binary search algorithm. Background Technology

[0002] Photonic integrated circuits (PICs) offer a promising platform for realizing large-scale optical interconnects on chips. Mode division multiplexing (MDM) technology can transmit multi-channel data through orthogonal eigenmodes in multi-mode waveguides, making it an effective means of increasing communication capacity. Among these technologies, multi-mode power beamsplitters are an essential component of MDM systems, widely used in feedback circuits, power beam splitting, and power monitoring.

[0003] In fact, power beam splitters have long attracted the interest of researchers, such as power beam splitters designed using the "Y" branch model. However, devices designed using traditional methods that rely on prior physical models have a small parameter space for optimization, low degrees of freedom, and large device size, making them difficult to apply to future large-scale photonic integrated systems.

[0004] With the advancements in micro-nano fabrication technology and the development of computer technology in recent years, intelligent algorithms have emerged. Many devices designed based on these algorithms have broken through the limitations of traditional thinking, achieving superior performance compared to traditional devices. In 2016, Luluzi Lu et al. designed and fabricated a power beam splitter with a size of only 2.72μm × 2.72μm using the Direct Binary Search (DBS) algorithm; however, this device only supported TE0 mode operation. Subsequently, in 2018 and 2020, Weijie Chang and Hucheng Xie respectively used intelligent algorithms to implement small-sized power beam splitters supporting dual-mode and tri-mode operation. However, these devices used the same input and output modes, making it difficult to monitor and test different modes.

[0005] Therefore, designing a multi-mode power beam splitter using intelligent algorithms to monitor and test different modes is of great significance for modular division multiplexing systems. Summary of the Invention

[0006] The technical problem to be solved by this invention is to address the current lack of power beam splitters that can simultaneously output different modes. This invention provides a dual-mode power beam splitter based on the multi-dimensional direct binary search (MDBS) algorithm that can achieve a 50:50 power beam split output while simultaneously switching modes. This dual-mode power beam splitter can be widely used in modular division multiplexing systems.

[0007] To achieve the above objectives, the technical solution of the present invention is as follows: a dual-mode power beam splitter includes a substrate, on which a top silicon layer is provided. The top silicon layer includes an optimization region. An input waveguide is provided at one end of the optimization region, and a first output waveguide and a second output waveguide are provided at the other end. The first output waveguide and the second output waveguide output different modes. The optimization region is divided into N×M square units of equal size. By adjusting the state of the center of the square units, a non-periodic perforated array that satisfies a predetermined output target is formed. The output target refers to the ratio of the optical power of the first output waveguide to the second output waveguide being 50:50.

[0008] Preferably, the center of the square unit is either perforated or not perforated.

[0009] Preferably, the center hole of the square unit includes a circular hole or an oblong hole.

[0010] Preferably, the substrate has a thickness of 3 μm, the top silicon layer has a thickness of 220 nm, the arc diameter of the waist-shaped hole is 90 nm, the straight side length of the waist-shaped hole is 30 nm, the diameter of the circular hole is 90–120 nm, and the width of the input waveguide is 500 nm.

[0011] Preferably, the size of the optimization region is 2520nm×2520nm, the distances from the center line of the input waveguide to the upper and lower edges of the optimization region are 1060nm and 1460nm respectively, the width of the first output waveguide is 500nm, and the width of the second output waveguide is 900nm.

[0012] Preferably, the size of the optimization region is 3300nm×2400nm, the distances from the center line of the input waveguide to the upper and lower edges of the optimization region are 1450nm and 1850nm respectively, the width of the first output waveguide is 900nm, and the width of the second output waveguide is 1300nm.

[0013] Preferably, the size of the square unit in the optimization region is 150nm × 150nm.

[0014] A method for manufacturing a dual-mode power beam splitter includes the following steps:

[0015] Step S1: Set the objective function for reflector performance in the multi-dimensional direct binary search algorithm;

[0016] Step S2: Arbitrarily select one of the square cells as the starting point for optimization;

[0017] Step S3: Set the selected square cell to the un-drilled state and calculate the objective function value;

[0018] Step S4: Set the center of the selected square unit as the waist-shaped hole, and change its rotation angle with the center of the waist-shaped hole as the origin until it rotates 180° back to the original position. Calculate the objective function value once for each change of rotation angle.

[0019] Step S5: Set the center of the selected square cell as a circular hole and change the diameter of the circular hole. Calculate the objective function value each time the size is changed.

[0020] Step S6: Compare the objective function values ​​calculated for all state parameters at the center of the selected square cell, select the optimal value, and set the parameters of each dimension corresponding to the optimal value as the final shape of the square cell.

[0021] Step S7: Select the next square cell and repeat steps S3 to S6. One iteration is defined as traversing all square cells once. After multiple iterations, compare the objective function value with the objective function value after the previous iteration. If the change in the objective function value between the two iterations is less than 0.1%, the objective function converges and the algorithm stops.

[0022] Preferably, the waist-shaped hole rotates at 10° increments until it rotates 180° back to its original position, and the diameter of the circular hole is between 90nm and 120nm, with the diameter changing at 10nm increments.

[0023] Preferably, the objective function value is the sum of the transmittance of the first output waveguide and the second output waveguide in the corresponding mode.

[0024] The beneficial effects of this invention are that the multi-dimensional direct binary search algorithm optimizes the multi-dimensional parameters of the square units in the initial structure. The parameters of the state, shape, rotation angle and size of each square unit are determined by the algorithm to satisfy the objective function. The computational freedom is higher, and it can achieve a near 50:50 power ratio beam output while realizing mode switching. It is of great significance for monitoring and measuring different modes in the mode division multiplexing system and can be widely used in mode division multiplexing systems. Attached Figure Description

[0025] Figure 1 This is a schematic diagram of one embodiment of the present invention.

[0026] Figure 2for Figure 1 The diagram shows the structure of the top silicon layer.

[0027] Figure 3 This is a schematic diagram of the structure of the central waist-shaped hole in a square unit.

[0028] Figure 4 This is a schematic diagram of the structure of a square unit with a circular hole at its center.

[0029] Figure 5 for Figure 1 The transmission spectrum of the embodiment shown.

[0030] Figure 6 This is a schematic diagram of another embodiment of the present invention.

[0031] Figure 7 for Figure 6 The diagram shows the structure of the top silicon layer.

[0032] Figure 8 for Figure 6 The transmission spectrum of the embodiment shown.

[0033] Figure 9 A schematic diagram of a beam splitter structure designed for the direct binary search algorithm; Figure 9 (a) is a device with a circular aperture radius of 45 nm. Figure 9 (b) is a device with a circular aperture radius of 52.5 nm. Figure 9 (c) is a device with a circular aperture radius of 60 nm. Figure 9 (d) is the transmission spectrum.

[0034] In the figure, 10 is the substrate; 20 is the top silicon layer; 1 is the input waveguide; 2 is the optimization region; 21 is the square cell; 3 is the first output waveguide; and 4 is the second output waveguide. Detailed Implementation

[0035] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments:

[0036] The dual-mode power beam splitter provided by the present invention includes a substrate 10, on which a top silicon layer 20 is provided. The top silicon layer 20 includes an optimization region 2. An input waveguide 1 is provided at one end of the optimization region 2, and a first output waveguide 3 and a second output waveguide 4 are provided at the other end. The first output waveguide 3 and the second output waveguide 4 output different modes. The optimization region 2 is divided into N×M square units 21 of equal size. By adjusting the state of the center of the square units 21, a non-periodic perforated array that meets a predetermined output target is formed. The output target refers to a ratio of 50:50 of the optical power of the first output waveguide 3 and the second output waveguide 4 in the wavelength range of 1540nm to 1560nm.

[0037] More specifically, the center of the square unit 21 is either perforated or not perforated.

[0038] More specifically, the center hole of the square unit 21 includes a circular hole or an oblong hole.

[0039] More specifically, the substrate 10 has a thickness of 3 μm, the top silicon 20 has a thickness of 220 nm, the arc diameter of the waist-shaped hole is 90 nm, the straight side length of the waist-shaped hole is 30 nm, the diameter of the circular hole is 90–120 nm, and the width of the input waveguide 1 is 500 nm.

[0040] Example 1

[0041] Please refer to the following: Figures 1-5 In this embodiment, the dual-mode power beamsplitter is a TE0 and TE1 dual-mode power beamsplitter. The optical device is designed on a silicon-on-insulator (SOI) platform with a top silicon layer 20 thickness of 220 nm, a substrate 10 silicon dioxide thickness of 3 μm, and an air cladding. The width W1 of the input waveguide 1 and the width W6 of the first output waveguide 3 are both 500 nm, ensuring that the input waveguide 1 and the first output waveguide 3 can support the TE0 mode without loss. The second output waveguide 4 is a TE1 mode output waveguide with a width W7 of 900 nm, ensuring that the second output waveguide 4 can support the TE1 mode without loss. The distance W2 from the centerline of the input waveguide 1 to the top edge of the optimization region 2 is 1060 nm, and the distance W3 from the centerline of the input waveguide 1 to the bottom edge of the optimization region 2 is 1460 nm. The optimization region 2 is square with a side length W4 of 2520 nm. Compared with traditional methods for designing optical devices, the intelligent algorithm can improve the optimization freedom and is more conducive to designing small-sized devices. Before optimization using the multi-dimensional direct binary search algorithm, optimization region 2 was divided into 16×16 square units 21 with a side length W5 of 150nm. Each square unit has two states: the center of the square unit 21 is punched or not punched. Each punched air hole has a corresponding shape, rotation angle and size. Considering the machinability, the edges of the air holes should be as smooth as possible and without sharp corners. Therefore, circular holes and waist-shaped holes are selected here.

[0042] A multi-dimensional direct binary search algorithm optimizes the parameters of the air hole in multiple dimensions, including state, shape, rotation angle, and size. The state parameter refers to the filling state, i.e., whether a hole is drilled or not at the center of the square unit 21; the shape parameter refers to whether the hole is oblong or circular, with the oblong hole having an arc diameter d1 of 90 nm and a straight side length W. 12The diameter d2 of the circular hole is 90-120 nm, and the depth of the circular hole is 0-220 nm. The rotation angle parameter refers to the rotation angle of the oblong hole with its center point as the axis. The size parameter refers to the diameter of the circular hole. In this embodiment, the depth of the oblong hole is 220 nm, and the depth of the circular hole is 220 nm.

[0043] The parameters of the state, shape, rotation angle, and size of each air hole are determined by a multi-dimensional direct binary search algorithm to satisfy the objective function. The objective function reflecting the device performance is set in the algorithm.

[0044] First, arbitrarily select a square cell 21, set its state to no hole, and calculate the objective function value. In order to balance time cost and device performance, the rotation angle step and diameter step can be flexibly set. Here, the rotation angle step is set to 10° and the diameter step is set to 10nm.

[0045] Then, the center position of the square unit 21 is set as an oblong hole, rotated 180° with a step size of 10°, and the objective function value is calculated every time the rotation angle is changed;

[0046] Finally, the center of the square unit 21 is set as a circular hole, and the diameter is changed from 90nm to 120nm in increments of 10nm. The objective function value is calculated for each change in size. All calculated objective function values ​​are compared, the optimal value is selected, and the parameters of each dimension corresponding to the optimal value are set to the final shape of the square unit.

[0047] One iteration is defined as traversing all square cells once. After multiple iterations, the objective function value is compared with the value after the previous iteration. If the change in the objective function value between the two iterations is less than 0.1%, the objective function converges, and the algorithm stops. The algorithm optimizes and determines parameters of the square cells in multiple dimensions, including state, shape, rotation angle, and size, increasing the degree of computational freedom. This allows for adjustment of the refractive index distribution in the optimization region, thereby realizing the device's function. Specifically, a TE0 mode light source is injected into input waveguide 1, passes through optimization region 2, and is converted into TE0 and TE1 modes. A primary mode and a converted mode are output from the first output waveguide 3 and the second output waveguide 4 respectively, with a power ratio of approximately 50:50. This not only allows for the simultaneous and uniform output of two different modes but also achieves mode conversion, which is of great significance for monitoring and measuring different modes in a mode-division multiplexing system.

[0048] Example 2

[0049] Please see Figures 6-9The technical solution provided in this embodiment is basically the same as that in Embodiment 1, except that: the dual-mode power beam splitter is a TE1 and TE2 dual-mode power beam splitter, the width W1 of the input waveguide 1 is 500nm, ensuring that the input waveguide 1 can support the TE0 mode without loss; the first output waveguide 3 is a TE1 mode output waveguide, and its width W7 is 900nm, ensuring that the first output waveguide 3 can support the TE1 mode without loss; the second output waveguide 4 is a TE2 mode output waveguide, and its width W7 is 900nm. 13 The wavelength is 1300nm, ensuring that the second output waveguide 4 can support TE2 mode losslessly; the distance W8 from the centerline of input waveguide 1 to the top edge of optimization region 2 is 1450nm, and the distance W9 from the centerline of input waveguide 1 to the bottom edge of optimization region 2 is 1850nm. Optimization region 2 is rectangular with a side length W 10 The wavelength is 3300nm, and the side length is W. 11 The wavelength is 2400nm; the optimization region 2 is divided into 16×22 square units 21 with a side length W5 of 150nm; the TE0 mode light source is injected into the input waveguide 1, and after passing through the optimization region 2, it is converted into TE1 and TE2 modes, and outputs two different modes from the first output waveguide 3 and the second output waveguide 4 with a power ratio of approximately 50:50, which can be widely used in mode division multiplexing systems.

[0050] This invention also provides a method for manufacturing a dual-mode power beam splitter, the specific optimization steps of which are as follows:

[0051] Step S1: Set the objective function for reflector performance in the multi-dimensional direct binary search algorithm;

[0052] Optimize the objective function of the TE0 and TE1 dual-mode power beam splitter: In the formula, Transmittance of TE0 mode in the 1540nm to 1560nm band; Transmittance of TE1 mode in the 1540nm to 1560nm band.

[0053] Optimize the objective function of the TE1 and TE2 dual-mode power splitter: In the formula, Transmittance of TE1 mode in the 1540nm to 1560nm band; Transmittance of TE2 mode in the 1540nm to 1560nm band.

[0054] Step S2: Arbitrarily select one of the square cells as the starting point for optimization. Generally, the first square cell in the first row is chosen as the starting point, which facilitates sequential searching.

[0055] Step S3: Set the selected square cell to the un-drilled state and calculate the objective function value.

[0056] Step S4: Set the center of the selected square cell as the waist-shaped hole, and change its rotation angle in 10° increments until it rotates 180° back to the original position. Calculate the objective function value every time the rotation angle is changed.

[0057] Step S5: Set the center of the selected square cell to a circular hole, and change its diameter in 10nm increments between 90nm and 120nm. Calculate the objective function value each time the size is changed.

[0058] Step S6: Compare all calculated objective function values, select the optimal value, and set the parameters of each dimension corresponding to the optimal value to the final shape of a square cell.

[0059] Step S7: Select the next square unit in sequence, either horizontally or vertically, and repeat steps S3 to S6.

[0060] One iteration is defined as traversing all cells once. After multiple iterations, the objective function value is compared with the value after the previous iteration. If the change in the objective function value between the two iterations is less than 0.1%, the objective function converges, and the algorithm stops. The algorithm optimizes and determines the parameters of the square cells in multiple dimensions, including state, shape, rotation angle, and size, to adjust the refractive index distribution in the optimization region, thereby realizing the function of the device.

[0061] After algorithm optimization, the device exhibits better performance. Figure 5 Simulation results for the TE0 and TE1 dual-mode power beam splitter show that, within a bandwidth range of 1540nm to 1560nm, the insertion loss (IL) of the TE0 mode in the first output waveguide 3 is less than 3.2dB; the insertion loss (IL) of the TE1 mode in the second output waveguide 4 is less than 3.2dB; and the crosstalk (CT) of the TE0 mode in the second output waveguide 4 is less than -25.7dB. Figure 8 Simulation results for the TE1 and TE2 dual-mode power beam splitter show that, within a bandwidth range of 1540nm to 1560nm, the insertion loss of the TE1 mode in the first output waveguide 3 is less than 3.3dB, and the crosstalk of the TE0 mode in the first output waveguide 3 is less than -37.4dB. Similarly, the insertion loss of the TE2 mode in the second output waveguide 4 is less than 3.4dB, and the crosstalk between the TE0 and TE1 modes in the second output waveguide 4 is less than -21.9dB. Furthermore, the loss curves of the two devices in this invention almost overlap in their desired output modes, and the transmittance is close to 50:50, achieving even power distribution.

[0062] The existing direct binary search algorithm only optimizes the search based on the state of the square unit, that is, it determines whether to punch a hole or not at the center of the square unit. The multi-dimensional direct binary search algorithm used in this invention, in addition to searching for the state, adds the search for shape, size and rotation angle, making the algorithm more flexible in calculation.

[0063] Specifically, the direct binary search algorithm and the multi-dimensional direct binary search algorithm can be used to optimize and compare the same device respectively. Under the same conditions as in Example 1, such as... Figure 9 (a)~ Figure 9 (c) Devices with circular aperture radii of 45nm, 52.5nm, and 60nm, designed using the direct binary search algorithm. Figure 9 (d) shows the simulation results. It can be seen that the device with a radius of 52.5nm has better performance. The insertion loss of TE0 mode in the first output waveguide 3 is less than 3.7dB, and the insertion loss of TE1 mode in the second output waveguide 4 is less than 3.6dB.

[0064] We can use the multi-dimensional direct binary search algorithm employed in this invention to design devices and... Figure 9 (d) Comparison. The insertion loss of the TE0 mode in the first output waveguide 3 of this invention is less than 3.2 dB, and the insertion loss of the TE1 mode in the second output waveguide 4 is less than 3.2 dB. The overall performance is nearly 10% higher than that of devices designed using the direct binary search algorithm. Furthermore, the crosstalk of the TE0 mode in the second output waveguide 4 is -25.7 dB, which is significantly lower than the crosstalk of the TE0 mode in the second output waveguide 4 of devices designed using the direct binary search algorithm (-18.2 dB). Therefore, the multi-dimensional direct binary search algorithm used in this invention offers greater computational freedom and can be used to design micro / nano photonic devices with superior performance.

[0065] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A dual-mode power beam splitter, characterized in that: The device includes a substrate, on which a top silicon layer is provided. The top silicon layer includes an optimization region. An input waveguide is provided at one end of the optimization region, and a first output waveguide and a second output waveguide are provided at the other end. The first output waveguide and the second output waveguide output different modes. The optimization region is divided into N×M square units of equal size. By adjusting the state of the center of the square units, a non-periodic perforated array that meets a predetermined output target is formed. The output target refers to the ratio of the optical power of the first output waveguide to the second output waveguide being 50:

50. The state of the center of the square unit is either punched or not punched; The center hole of the square unit includes both circular and oblong holes.

2. The dual-mode power beam splitter as described in claim 1, characterized in that: The substrate has a thickness of 3 μm, the top silicon layer has a thickness of 220 nm, the arc diameter of the waist-shaped hole is 90 nm, the straight side length of the waist-shaped hole is 30 nm, the diameter of the circular hole is 90–120 nm, and the width of the input waveguide is 500 nm.

3. The dual-mode power beam splitter as described in claim 2, characterized in that: The size of the optimization region is 2520nm×2520nm. The distances from the center line of the input waveguide to the top and bottom edges of the optimization region are 1060nm and 1460nm, respectively. The width of the first output waveguide is 500nm, and the width of the second output waveguide is 900nm.

4. The dual-mode power beam splitter as described in claim 2, characterized in that: The size of the optimization region is 3300nm×2400nm. The distances from the center line of the input waveguide to the top and bottom edges of the optimization region are 1450nm and 1850nm, respectively. The width of the first output waveguide is 900nm and the width of the second output waveguide is 1300nm.

5. The dual-mode power beam splitter as described in claim 3 or 4, characterized in that: The size of the square unit in the optimization region is 150nm × 150nm.

6. The method for manufacturing a dual-mode power beam splitter as described in claim 1, characterized in that, Includes the following steps: Step S1: Set the objective function for reflector performance in the multi-dimensional direct binary algorithm; Step S2: Arbitrarily select one of the square cells as the starting point for optimization; Step S3: Set the selected square cell to the un-drilled state and calculate the objective function value; Step S4: Set the center of the selected square unit as the waist-shaped hole, and change its rotation angle with the center of the waist-shaped hole as the origin until it rotates 180° back to the original position. Calculate the objective function value once for each change of rotation angle. Step S5: Set the center of the selected square cell as a circular hole and change the diameter of the circular hole. Calculate the objective function value each time the size is changed. Step S6: Compare the objective function values ​​calculated for all state parameters at the center of the selected square cell, select the optimal value, and set the parameters of each dimension corresponding to the optimal value as the final shape of the square cell. Step S7: Select the next square cell and repeat steps S3 to S6. One iteration is defined as traversing all square cells once. After multiple iterations, compare the objective function value with the objective function value after the previous iteration. If the change in the objective function value between the two iterations is less than 0.1%, the objective function converges and the algorithm stops.

7. The method for manufacturing a dual-mode power beam splitter as described in claim 6, characterized in that: The waist-shaped hole rotates in 10° increments until it rotates 180° back to its original position. The diameter of the circular hole is between 90nm and 120nm, and its diameter is changed in 10nm increments.

8. The method for manufacturing a dual-mode power beam splitter as described in claim 6, characterized in that: The objective function value is the sum of the transmittance of the first output waveguide and the second output waveguide in the corresponding mode.

Images