Benes-architecture-based multi-port optical switch device using chalcogenide phase-change material, and implementation method therefor
Patent Information
- Application Number
- PCT/CN2025/122087
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2025-03-28
- Filing Date
- 2025-09-18
- Publication Date
- 2026-10-01
Smart Images
Figure CN2025122087_01102026_PF_FP_ABST
Abstract
Description
Benes architecture-based multi-channel chalcogenide phase change material optical switching device and its implementation method Technical Field
[0001] This invention relates to the field of optical communication technology, and in particular to a multi-channel chalcogenide phase change material optical switch device based on the Benes architecture and its implementation method. Background Technology
[0002] In recent years, the continuous growth of data traffic has posed severe challenges to communication networks. To further increase communication capacity, silicon-based photonic integration platforms have stood out among many integration platforms due to their strong potential to integrate electrical and photonic chips on a single chip, and their ability to be manufactured using mature, low-cost, large-scale CMOS integrated circuit manufacturing processes. This has led to widespread attention and investment. Among these, silicon-based on-chip optical switches are the core devices for achieving high-efficiency optical switching.
[0003] Existing on-chip waveguide photonic devices primarily utilize modulation methods based on thermo-optical effects and the dispersion effect of free carriers, directly altering the effective refractive index of the silicon waveguide to achieve optical switching. However, thermo-optical phase shifters suffer from slow modulation rates and high power consumption, typically on the order of milliseconds. Phase shifters based on carrier dispersion effects exhibit significant optical losses due to phosphorus or boron doping, and their refractive index change is less than 10⁻³, requiring dimensions larger than 100 μm to achieve a 2π phase shift.
[0004] Based on the research of 2×2 optical switches, in order to realize the expansion of multi-channel optical switches, various network architectures have been proposed, including binary tree architecture, crossbar architecture, PILOSS architecture, S&S architecture, Benes architecture, etc.
[0005] The binary tree architecture primarily uses a 2×2 optical switch as the central switching layer, expanding outwards in a tree structure from the input and output ports. Each port then connects to another optical switch, controlling the state of each level of switches to achieve path selection between the input and output ports. This architecture offers advantages such as low fabrication complexity and ease of integration. However, because the central switching layer is controlled by only a single 2×2 optical switch, transmission congestion can occur, meaning synchronous parallel data transmission is impossible, severely limiting the transmission efficiency of multiple optical switches. Furthermore, the other optical switches extending from the central switching layer only transmit data in a 1×2 manner, failing to fully utilize the port advantage of the 2×2 optical switches and significantly restricting the flexibility of data transmission.
[0006] The crossbar architecture primarily consists of a switch matrix formed by arranging switch units in a horizontal and vertical pattern. It is simple in structure, easy to design and expand, and features strict non-blocking characteristics, allowing any input to be independently connected to any unused output. During data transmission, since each input / output port connects to only a single switch unit, there is no waveguide crossing, resulting in low insertion loss and low routing power consumption. However, this architecture requires a large number of 2×2 optical switch units and suffers from severe loss unevenness, making it susceptible to crosstalk cascading issues from multiple paths.
[0007] The PILOSS architecture is a strictly non-blocking topology, consisting of multiple levels of 2×2 optical switch ports cross-connected. It features uniform link loss and simple link switching control, offering advantages in integrated chip design. However, when constructing an N×N optical switch, the PILOSS architecture requires the use of N... 2 Each optical switch unit would occupy a large amount of space, making large-scale integration difficult, and it would be susceptible to multipath coherent crosstalk.
[0008] The S&S architecture uses a 1×2 beam splitter structure to provide transmission paths for the input and output ports, and then utilizes a large number of cross-connect structures to complete the channel interconnection. Since optical path switching is completed by the 1×2 beam splitter structure, it has the characteristic of no first-order crosstalk. However, because a large number of cross waveguides need to be constructed in the intermediate layer, the fabrication process is relatively difficult.
[0009] Therefore, current chip-level optical switches are mainly based on silicon photonics platforms, relying on volatile modulation of thermo-optics or electro-optics to achieve switching functions. Such modulation causes a slight change in the effective refractive index within the waveguide (typically less than 0.01). Consequently, the system requires continuous power to maintain the switching state, which not only leads to high power consumption but also results in a large area occupied by the switching device, limiting its power efficiency and hindering the large-scale and compact integration of photonic integrated circuits.
[0010] In terms of multi-path network architecture, the binary tree topology suffers from transmission congestion when transmitting data in parallel at the ports because the intermediate stage only uses a single 2×2 optical switch, which is not conducive to improving data transmission efficiency. The crossbar architecture has a large loss difference due to the different number of switching units traversed by different paths, resulting in significant path loss non-uniformity. In the PILOSS architecture, the number of switching units is quadratically related to the number of ports, and there are many cross waveguides, making process implementation and optoelectronic hybrid packaging of devices challenging. The S&S architecture, with its numerous cross waveguide structures, increases the design difficulty of multi-path integrated networks. Summary of the Invention
[0011] The purpose of this invention is to provide a multi-channel chalcogenide phase change material optical switch device based on the Benes architecture and its implementation method, which can improve performance and transmission efficiency.
[0012] To achieve the above objectives, the present invention provides a multi-channel chalcogenide phase change material optical switch device based on the Benes architecture, comprising a plurality of 2×2 optical switches, all of which form a Benes architecture. Each optical switch includes an incident waveguide, a directional coupler, and an exit waveguide. The incident waveguide is connected to the input end of the directional coupler, and the exit waveguide is connected to the output end of the directional coupler. The directional coupler includes a first waveguide and a second waveguide. A chalcogenide phase change material layer is disposed on the second waveguide. A metal electrode is connected to the first waveguide and the second waveguide, respectively.
[0013] As a preferred embodiment, the sulfur-based phase change material layer is an Sb2Se3 layer.
[0014] As a preferred embodiment, the Sb:Se composition ratio of the sulfur-based phase change material layer (5) is (20-50):(50-80).
[0015] As a preferred embodiment, both the first waveguide and the second waveguide are ridge waveguides.
[0016] As a preferred embodiment, the optical switch further includes a substrate, an insulating layer, and a dielectric layer. The insulating layer is disposed above the substrate. The incident waveguide, the first waveguide, the second waveguide, and the outgoing waveguide are all disposed above the insulating layer. The dielectric layer is disposed on the incident waveguide, the first waveguide, the second waveguide, the outgoing waveguide, and the portion of the insulating layer not covered by the waveguide.
[0017] As a preferred embodiment, the substrate is a silicon wafer, the insulating layer is a silicon dioxide buried oxide layer, and the dielectric layer is aluminum oxide.
[0018] As a preferred embodiment, the Benes architecture includes an input switching layer and an output switching layer, each of which includes multiple 2×2 optical switches, and the number of optical switches used in the input switching layer and the output switching layer is equal.
[0019] The input switching layer uses recursive grouped input, and the ports of all optical switches in the input switching layer are divided into upper and lower parts, with each part recursively constructing a Benes sub-network; the output switching layer uses grouped recursive output, which is the opposite of the setting logic of the input switching layer.
[0020] As a preferred option, the device is an N×N optical switch network, with 2×2 optical switches having the following number: The number of optical switching stages to be constructed can be expressed by the following formula: S = 2log2N-1
[0021] Where S represents the number of network levels to be constructed, and N represents the number of optical switch ports to be implemented.
[0022] This invention also provides a method for implementing a multi-channel chalcogenide phase change material optical switch device based on the Benes architecture. Using the aforementioned optical switch, the method includes the following steps:
[0023] S1: Select the appropriate number of 2×2 optical switches according to the N×N optical switch network to be constructed;
[0024] S2: Divide the optical switch into two equal parts, and the two optical switches constitute the input switching layer and output switching layer of the Benes architecture, respectively;
[0025] S3: In constructing the input switching layer, the number of optical switches in each level is equal. Each optical switch is numbered, and the port of each optical switch is numbered.
[0026] S4: According to the number, connect the two ports of the previous stage optical switch to the ports of the next stage optical switch to realize the connection between adjacent stages; each stage within the layer is connected to the two adjacent stages to complete the connection of the input switching layer;
[0027] S5: Using the backtracking method, starting from the output port of the output switching layer, each level in the layer is connected to the two adjacent levels respectively, so that the output switching layer is symmetrical with the input switching layer, and the connection of the output switching layer is completed.
[0028] As a preferred embodiment, in step S3, each optical switch port is assigned a number (X, Y, Z), where X represents the group number of each level, Y represents the number of each group of optical switches, and Z represents the number of each group of optical switch ports; the optical switches of the nth level can be divided into 2... n-1 The Y and Z numbers for each group are obtained using the following formula: Y n,i =i Z n,j =j
[0029] Where N is the number of ports, n is the number of stages, i represents the i-th optical switch in a group, and j represents the j-th port in a group;
[0030] In step S4, in each group, the two ports of the Y-numbered optical switch of the previous stage are connected in ascending order to the ports of the Z-numbered optical switch of the next stage, which has a value equal to Y.
[0031] Compared with the prior art, the beneficial effects of the present invention are as follows:
[0032] This invention incorporates a chalcogenide phase change material layer on the waveguide of the directional coupler of an optical switch. The phase change material itself has a high refractive index and low loss, and the large difference in refractive index before and after the phase change results in a small extinction coefficient. This reduces the voltage required to switch between crystalline and amorphous states, enabling faster state switching and thus lowering the overall power consumption of the device. It significantly alters the supermode structure in the waveguide system, affecting the light coupling path and output position, achieving a high-performance optical switch. Furthermore, this invention constructs a multi-channel optical switch device based on the Benes architecture. Utilizing the reconfigurable non-blocking transmission characteristics of the Benes architecture, it greatly improves data transmission flexibility and efficiency. It also enables the same-scale multi-channel optical switch design with fewer switch units. When the scale requirement increases, compared to other solutions, it significantly reduces the number of switch units used, saving chip space and reducing overall device loss, further improving the chip's overall performance and integrability. Attached Figure Description
[0033] Figure 1 shows the phase change film of the Sb2Se3 system after EDS analysis with optimized composition;
[0034] Figure 2 shows the refractive index and extinction coefficient of Sb2Se3 thin films in amorphous and crystalline states;
[0035] Figure 3 shows the pulse width for photo-controlled crystallization of Sb2Se3 thin films;
[0036] Figure 4 shows the pulse width of the light-controlled amorphization of Sb2Se3 thin film;
[0037] Figure 5 shows a silicon-based chalcogenide optical waveguide based on Sb₂Se₃;
[0038] Figure 6 is a schematic diagram of an optical switch for a silicon-based chalcogenide phase change material two-waveguide directional coupler.
[0039] Figure 7 is a side view of the optical switch of a silicon-based chalcogenide phase change material two-waveguide directional coupler.
[0040] Figure 8 shows the electric field distribution of the amorphous symmetric mode of the optical switch in a silicon-based chalcogenide phase change material two-waveguide directional coupler.
[0041] Figure 9 shows the electric field distribution of the amorphous antisymmetric mode in the optical switch of a silicon-based chalcogenide phase change material two-waveguide directional coupler.
[0042] Figure 10 shows the electric field distribution of the first crystalline mode of the optical switch in the silicon-based chalcogenide phase change material two-waveguide directional coupler.
[0043] Figure 11 shows the electric field distribution of the second crystalline mode of the optical switch in the silicon-based chalcogenide phase change material waveguide directional coupler.
[0044] Figure 12 is a network block diagram of multi-channel optical switch routing based on the Benes architecture;
[0045] Figure 13 is a simplified model of a single optical switch in the Benes architecture;
[0046] Figure 14 shows the input layer of the Benes architecture for the 8×8 optical switch after numbering processing;
[0047] Figure 15 shows the input layer of the 8×8 optical switch Benes architecture;
[0048] Figure 16 is a schematic diagram of an 8×8 optical switch based on the Benes architecture. In the figure, 1-substrate; 2-insulating layer; 3-first waveguide; 4-second waveguide; 5-chalcogenide phase change material layer; 6-dielectric layer; 7-incident waveguide; 8-outgoing waveguide. Detailed Implementation
[0049] The specific embodiments of the present invention will be described in further detail below with reference to the accompanying drawings and examples. The following examples are for illustrative purposes only and are not intended to limit the scope of the invention.
[0050] In the description of this invention, it should be noted that the terms "center," "longitudinal," "lateral," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing the invention and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.
[0051] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.
[0052] Furthermore, in the description of this invention, unless otherwise stated, "a plurality of" means two or more.
[0053] Example 1
[0054] As shown in Figures 1 to 16, a preferred embodiment of the present invention provides a multi-channel chalcogenide phase change material optical switch device based on a Benes architecture, comprising several 2×2 optical switches. All optical switches form a Benes architecture. Each optical switch includes an incident waveguide 7, a directional coupler, and an exit waveguide 8. The incident waveguide 7 is connected to the input end of the directional coupler, and the exit waveguide 8 is connected to the output end of the directional coupler. The directional coupler includes a first waveguide 3 and a second waveguide 4. A chalcogenide phase change material layer 5 is disposed on the second waveguide 4. A metal electrode is connected to each of the first waveguide 3 and the second waveguide 4. This embodiment utilizes the high refractive index and low loss of the chalcogenide phase change material, along with the large difference in refractive index and small extinction coefficient before and after the phase change. This results in a lower voltage required for switching between crystalline and amorphous states, enabling faster state switching and reducing the overall power consumption of the device. This significantly alters the supermode structure in the waveguide system, affecting the light coupling path and output position, thus achieving a high-performance optical switch.
[0055] Benes architecture-based multi-path chalcogenide phase change material optical switching devices are formed by routing several 2×2 optical switching devices. Each 2×2 switching module can control the switching path of the input signal, allowing the signal to switch between different ports.
[0056] The crystalline and amorphous states of chalcogenide phase change materials can be rapidly and reversibly switched on the μs-ns scale via thermal pulses. Furthermore, once a chalcogenide phase change material switches to a certain state, no additional power is required to maintain that state, indicating non-volatility. Chalcogenide phase change materials exhibit a large refractive index, a near-zero extinction coefficient over a wide spectral range, a significant refractive index difference between the crystalline and amorphous states, and good reversibility between the amorphous and crystalline states.
[0057] Specifically, in this embodiment, the chalcogenide phase change material layer 5 is an Sb₂Se₃ layer, and both the first waveguide 3 and the second waveguide 4 are ridge waveguides. Therefore, this embodiment achieves a 2×2 optical switch using a two-waveguide directional coupler composed of a ridge silicon waveguide and a ridge silicon-on-Sb₂Se₃ hybrid waveguide. By switching the crystalline and amorphous states of Sb₂Se₃ on the hybrid waveguide, the supermode structure in the waveguide system is significantly altered, affecting the coupling path and output position of the light, thus realizing the optical switch function. When Sb₂Se₃ is in the amorphous state, this device satisfies the phase matching condition, that is, when the effective refractive indices of the modes in different waveguides are equal, efficient coupling occurs between modes, and the optical power is output from the output port on the opposite side. When Sb₂Se₃ is in the crystalline state, the refractive index of Sb₂Se₃ increases, and the input light no longer undergoes cross-coupling, but is directly output from the port of the same-side strip silicon waveguide. Therefore, this device structure can realize the optical switch function through the conversion between the amorphous and crystalline states of the chalcogenide material Sb₂Se₃.
[0058] This embodiment uses the low-loss chalcogenide phase change material Sb₂Se₃ as one of the waveguide materials. High-quality Sb₂Se₃ thin films are fabricated to reduce device power consumption. Utilizing the high refractive index of the phase change material itself, the large refractive index difference before and after the phase transition of the low-loss chalcogenide phase change material Sb₂Se₃, and its small extinction coefficient, the optical switching function can be achieved with a relatively thin film. This embodiment uses only a 20-30 nm thick phase change material film to fabricate a hybrid waveguide, resulting in a lower voltage required to switch between crystalline and amorphous states, enabling faster state switching and thus reducing the overall power consumption of the device. This material also exhibits non-volatility, significantly reducing the device's power supply requirements to achieve the goal of power reduction. The material has a significant refractive index difference between its crystalline and amorphous states, allowing for thermally driven phase transitions between crystalline and amorphous states via electrical or optical methods. Furthermore, by controlling the crystallinity of the phase change material, its refractive index can be continuously adjusted.
[0059] Using chalcogenide phase change material Sb₂Se₃ as one of the waveguide materials for directional couplers, this study addresses the issue of inconsistent evaporation rates of Sb and Se elements during fabrication. By optimizing the amount of material used in thin film deposition, compositional fluctuations are reduced, improving thickness uniformity, refractive index uniformity, and defect control. This leads to the fabrication of Sb₂Se₃ thin film materials with ultra-low extinction coefficients (k < 0.01).
[0060] The evaporation source material for the sulfur-based phase change material Sb₂Se₃ is mainly prepared from high-purity elemental Sb and Se through a series of preparation methods including raw material and quartz tube pretreatment, weighing, batching, packaging, melting, quenching, and annealing. Based on the low-loss Sb₂Se₃ thin film material with smooth surface, uniform refractive index and thickness distribution, and good defect control, the optimized preparation process closest to its stoichiometry was selected. The component ratio of the prepared Sb₂Se₃ thin film material was characterized using EDS, and the results, shown in Figure 1, indicate an Sb:Se composition of 37.87:62.13. In this embodiment, the Sb:Se ratio of the sulfur-based phase change material layer 5 is (20-50):(50-80). For example, the Sb:Se ratio can be 20:80, 35:65, 45:55, 40:60, or 50:50. More specifically, the Sb:Se ratio is (35-45):(55-65). In this embodiment, the Sb:Se ratio is 37.87:62.13. The sulfur-based phase change material layer 5 is specifically a thin film structure. In this embodiment, the thickness of the sulfur-based phase change material layer 5 is 20-30 nm, specifically 20 nm, 25 nm, or 30 nm.
[0061] The prepared thin film material was annealed on a hot plate under different temperature conditions to induce crystallization. The refractive index and extinction ratio in the crystalline and amorphous states were measured using an ellipsometry, as shown in Figure 2. Before and after the crystallization phase transition, the Sb₂Se₃ phase change thin film material with optimized composition exhibited an extinction coefficient k ≤ 0.01 and a refractive index change Δn ≥ 0.7 in the near-infrared band, which is beneficial for improving device integration and reducing power consumption. Furthermore, the phase transition rate was tested using a 532nm nanosecond-level pulsed optically controlled reconfigurable phase transition transient response system. The pulse width for optically controlled crystallization is shown in Figure 3, indicating that the optically controlled crystallization rate of the Sb₂Se₃ thin film with optimized composition and process is approximately 436.07 ns. The pulse width for optically controlled amorphization is shown in Figure 4, indicating that the optically controlled amorphization rate of the Sb₂Se₃ thin film with optimized composition and process is approximately 2.01 ns, enabling ultra-high-speed phase transition control. Low-loss thin film materials were prepared by thermal evaporation, and dry etching was performed by electron beam exposure and plasma etching. A dielectric protective layer was prepared by inductively coupled plasma chemical vapor deposition. The waveguide morphology is shown in Figure 5. The middle part is a ridge-shaped silicon waveguide, and a chalcogenide phase change material is grown in the box to prepare a silicon-based chalcogenide optical waveguide.
[0062] According to coupled-mode theory, in a directionally coupled structure, after each transmission length Lc, the optical power is completely transferred from one waveguide to another. This is called complete coupling. The Lc at which complete coupling occurs is the coupling length, which is typically expressed by the following formula:
[0063] Where m is a natural number, n eff1 and n eff2 These are the effective refractive indices of the modes in the two waveguides, respectively.
[0064] The directional coupler in this embodiment is a two-waveguide directional coupler, consisting of a ridged silicon waveguide and a ridged silicon-on-Sb2Se3 hybrid waveguide. A pair of metal electrodes are located on both sides of the waveguides. Joule heating can be generated by applying an electrical pulse to the conductive electrodes, causing the material temperature to rise above the glass transition temperature and begin crystallization, forming a crystalline structure. Alternatively, a narrow pulse with high energy density but short duration can be applied to rapidly heat the material above its melting temperature, preventing the atoms in the molten state from rearranging and forming bonds, resulting in an amorphous structure. A schematic diagram is shown in Figure 6. When Sb2Se3 is in the amorphous state, the device satisfies the phase-matching condition, i.e., when the effective refractive indices of the modes in different waveguides are equal, efficient coupling between modes occurs, and optical power is output from the Cross port. When Sb2Se3 is in the crystalline state, the refractive index of Sb2Se3 increases, and the input light no longer undergoes cross-coupling but is directly output from the Bar port of the same-side strip silicon waveguide.
[0065] Specifically, the optical switch also includes a substrate 1, an insulating layer 2, and a dielectric layer 6. The insulating layer 2 is disposed above the substrate 1. The incident waveguide 7, the first waveguide 3, the second waveguide 4, and the outgoing waveguide 8 are all disposed above the insulating layer 2. The dielectric layer 6 is disposed on the incident waveguide 7, the first waveguide 3, the second waveguide 4, the outgoing waveguide 8, and the portion of the insulating layer 2 not covered by the waveguides. The substrate 1 is a silicon wafer, the insulating layer 2 is a buried oxide layer of silicon dioxide, and the dielectric layer 6 is aluminum oxide.
[0066] Through simulation optimization and calculation using the commercial software Ansys Lumerical Mode, a set of parameters for a silicon-based chalcogenide phase change material two-waveguide directional coupler structured optical switch can be obtained. The side view of the model is shown in Figure 7. In the model, 1 is the substrate with a thickness of 2 μm; 2 is the insulating layer, made of silicon dioxide with a thickness of 3 μm; 3 and 4 are ridge silicon waveguides, with 3 being the first waveguide and 4 being the second waveguide. The bottom silicon layer has a thickness of 90 nm, and the ridge layer has a thickness of 130 nm. The width of the ridge silicon waveguide without phase change material (first waveguide) is 500 nm, and the width of the ridge silicon waveguide with phase change material (second waveguide) is 434 nm; 5 is the Sb₂Se₃ material on the hybrid ridge silicon waveguide, with a thickness of 30 nm and a width of 394 nm; 6 is the dielectric layer, made of alumina, encapsulating the entire optical switch device, with a thickness of 40 nm. The gap between the two waveguides is 250 nm, and the waveguide coupling length is 19.2165 μm.
[0067] Mode simulations were performed on a 2×2 two-waveguide directional coupler grown from the phase change material Sb₂Se₃ at 1550 nm. In the amorphous state, the simulated refractive index was approximately 3.476, and in the crystalline state, it was approximately 4.367. In the amorphous state, due to the symmetry principle, the eigenmodes of the directional coupler system can be decomposed into two modes satisfying symmetry: a symmetric mode and an antisymmetric mode, as shown in Figures 8 and 9. The effective refractive indices of the two modes are different. The symmetric mode indicates that the electric field distributions within the two waveguides are in phase, the field shape is symmetric about the center, the energy is more concentrated between the two waveguides, and the effective refractive index is higher, with a simulated value of approximately 2.59. The antisymmetric mode indicates that the electric field distributions within the two waveguides are in opposite phase, the field shape is antisymmetric about the center, the energy distribution is more dispersed, and the effective refractive index is lower, with a simulated value of approximately 2.55. When light enters from one input port, as it propagates through the directional coupler, the phase difference accumulates due to the different propagation speeds of the two modes in the coupling region. By setting an appropriate coupling length, directional distribution can be achieved at the output port; that is, the signal at the coupled port is a coherent superposition result, while the signal at the isolated port is a cancellation result, ultimately allowing light to be output from the opposite output port. When the phase change material is in a crystalline state, the simulation results are shown in Figures 10 and 11. The crystallization of the material breaks the coupling condition, and the system's eigenmodes cannot be decomposed into symmetric and antisymmetric modes. This means the device loses its directional coupling function, and light cannot be coupled to another waveguide in the directional coupler, supporting only single-mode transmission. Ultimately, light is output from the same output port. The simulation results of the two-waveguide directional coupler in crystalline and amorphous states verify that the device possesses the switching function of a 2×2 optical switch, providing fundamental device support for realizing high-performance multi-channel optical switch networks.
[0068] Example 2
[0069] The difference between this embodiment and Embodiment 1 is that, based on Embodiment 1, this embodiment provides a further explanation of the multi-channel optical switching device.
[0070] Benes architecture is a non-blocking network in a multi-level interconnected network that can perform all permutations from input to output. A Benes network consists of a series of basic units (such as 2×2 switches) that are cascaded and cross-connected to form a multi-level non-blocking network capable of performing any type of permutation, that is, transforming any arrangement of inputs into any arrangement of outputs.
[0071] Specifically, the Benes architecture in this embodiment includes an input switching layer and an output switching layer. The input switching layer and the output switching layer each include multiple 2×2 optical switches, and the number of optical switches used in the input switching layer and the output switching layer is equal. The input switching layer adopts recursive grouped input, and the ports of all optical switches in the input switching layer are divided into upper and lower parts, and each part recursively constructs a Benes sub-network. The output switching layer adopts grouped recursive output, which is the opposite of the setting logic of the input switching layer.
[0072] When the device is an N×N optical switch network, the number of 2×2 optical switches is: The number of optical switching stages to be constructed can be expressed by the following formula: S = 2log2N-1
[0073] Where S represents the number of network levels to be constructed, and N represents the number of optical switch ports to be implemented.
[0074] The Benes architecture is used to implement multi-channel optical switches by constructing a multi-layer 2×2 switch network, which exhibits a certain degree of symmetry. As shown in Figure 12, block 1 is the input switching layer, which uses recursive grouping of inputs, dividing the N ports into upper and lower parts, with each part recursively constructing a Benes sub-network; block 2 is the output switching layer, which uses grouped recursive outputs, with the logic reversed from that of the input switching layer. This can be achieved using a backtracking method, that is, deriving the switching unit state from the output end. If an N×N optical switch network is constructed, the number of 2×2 optical switches required is: The number of optical switching stages to be constructed can be expressed by the following formula: S = 2log2N-1
[0075] Where S represents the number of network levels to be constructed, and N represents the number of optical switch ports to be implemented.
[0076] In this embodiment, the optical switch is combined with the Benes network architecture. By constructing a multi-layer 2×2 optical switch module, multiple optical switches are realized, ensuring efficient signal routing and flexible connection control, and realizing multi-port switching in optical communication.
[0077] The other structures in this embodiment are the same as in Embodiment 1, and will not be described again here.
[0078] Example 3
[0079] This embodiment provides a method for implementing a multi-channel chalcogenide phase change material optical switch device based on the Benes architecture, using the optical switch of Embodiment 1 or Embodiment 2. The method includes the following steps:
[0080] S1: Select the appropriate number of 2×2 optical switches according to the required N×N optical switch network; if constructing an N×N optical switch network, the number of 2×2 optical switches needed is: The number of optical switching stages to be constructed can be expressed by the following formula: S = 2log2N-1
[0081] Where S represents the number of network levels to be constructed, and N represents the number of optical switch ports to be implemented;
[0082] S2: Divide the optical switch into two equal parts, and the two optical switches constitute the input switching layer and output switching layer of the Benes architecture, respectively;
[0083] S3: In constructing the input switching layer, the number of optical switches in each level is equal. Each optical switch is numbered, and the port of each optical switch is numbered.
[0084] S4: According to the number, connect the two ports of the previous stage optical switch to the ports of the next stage optical switch to realize the connection between adjacent stages; each stage within the layer is connected to the two adjacent stages to complete the connection of the input switching layer;
[0085] S5: Using the backtracking method, starting from the output port of the output switching layer, each level in the layer is connected to the two adjacent levels respectively, so that the output switching layer is symmetrical with the input switching layer, and the connection of the output switching layer is completed.
[0086] Specifically, in step S3, each optical switch port is assigned a number (X, Y, Z), where X represents the group number of each level, Y represents the number of each group of optical switches, and Z represents the number of each group of optical switch ports; the optical switches of the nth level can be divided into 2... n-1 The Y and Z numbers for each group are obtained using the following formula: Y n,i =i Z n ,j=j
[0087] Where N is the number of ports, n is the number of stages, i represents the i-th optical switch in a group, and j represents the j-th port in a group;
[0088] In step S4, in each group, the two ports of the Y-numbered optical switch of the previous stage are connected in ascending order to the ports of the Z-numbered optical switch of the next stage, which has a value equal to Y.
[0089] When the number of groups equals N / 2, it represents an intermediate level, which is the last level of the input layer and also an intermediate level. Step S4 establishes connections between levels. Connections are made within the same group based on the grouping of the previous level. In each group, the two ports of the Y-numbered optical switch from the previous level are connected sequentially, from smallest to largest, to the Z-numbered optical switch port of the next level, whose value is equal to Y. Step S5 uses a backtracking method, starting from the output end, to construct the output switching layer in reverse order of the steps used to construct the input switching layer.
[0090] In step S3, if an N×N optical switch network needs to be constructed, then there are N input terminals and N output terminals, requiring... There are N / 2 optical switches in each stage, starting from the input end, with stages 1, 2, 3, ..., n. Each stage can be divided into 2 optical switches. n-1 Optical switches are grouped into groups. For example, all optical switches in level 1 are grouped together, optical switches in level 2 are divided into two groups, optical switches in level 3 are divided into four groups, and so on. The groups are arranged sequentially from top to bottom within each level. Each group of optical switches is assigned an X number. For instance, if an optical switch belongs to a level divided into 3 groups and is located in group 2, its X number is 2. Optical switches within the same group at the same level are also arranged sequentially from top to bottom within their respective groups. Each group of optical switches is assigned a Y number. For instance, if an optical switch belongs to a level divided into 3 groups and is located in group 2, and is the first switch in that group, its X number is 2 and its Y number is 1. The ports of optical switches within the same group at the same level are numbered sequentially from top to bottom. For instance, if an optical switch belongs to a level divided into 3 groups and is located in group 2, and is the third switch in that group, its X number is 2 and its Y number is 3. The Z numbers for the ports on both sides are 5 and 6.
[0091] Taking the Benes architecture of an 8×8 optical switch as an example, the input switching layer is constructed. First, numbering is assigned: Level 1 is divided into 1 group, Level 2 into 2 groups, and Level 3 into 4 groups. Since the number of groups is equal to 4, Level 3 is the intermediate level, i.e., the last level of the input switching layer. Numbering is performed according to the formula mentioned above, as shown in Figure 14. When connecting between two levels, according to the grouping of the previous level, within the same group, the two ports of the Y-numbered optical switch of the previous level are connected sequentially from smallest to largest to the Z-numbered port of the next level with a value equal to Y. To reduce the number of cross waveguides, the two ports of the previous level switch unit are connected sequentially from top to bottom, completing the connection of the input switching layer, as shown in Figure 15. Finally, the previous steps are repeated starting from the output port using a backtracking method to complete the connection of the output switching layer. The complete 8×8 reconfigurable optical interconnect topology Benes network architecture is shown in Figure 16.
[0092] The Benes architecture's routing protocol is based on a multi-level 2×2 switch configuration, enabling non-blocking connections from any input port to any output port. Its core routing protocol utilizes recursive decomposition, breaking down the entire path selection problem into the configuration choices of multiple 2×2 switches. Specifically, the routing protocol progressively selects between through or cross states at each level of the switches based on the required input-output pairs. In this way, signals can be transmitted smoothly to the destination port without collisions. Furthermore, the recursive symmetry of the Benes architecture allows the routing protocol to be effectively reused in higher-order networks, achieving scalability while maintaining network flexibility and efficiency. This routing strategy ensures connectivity between each pair of ports and can dynamically adjust to adapt to real-time changing routing requirements.
[0093] In summary, this invention provides a multi-channel chalcogenide phase change material optical switch device based on the Benes architecture. By placing a chalcogenide phase change material layer on the waveguide of the directional coupler of the optical switch, the high refractive index and low loss of the phase change material, along with the large difference in refractive index and small extinction coefficient before and after the phase change, reduces the voltage required to switch between crystalline and amorphous states, enabling faster state switching and thus reducing the overall power consumption of the device. This significantly alters the supermode structure in the waveguide system, affecting the light coupling path and output position, achieving a high-performance optical switch. Furthermore, this invention constructs a multi-channel optical switch device based on the Benes architecture, leveraging its reconfigurable non-blocking transmission characteristics to significantly improve data transmission flexibility and efficiency. It also enables the same-scale multi-channel optical switch design with fewer switch units. When the scale requirement increases, compared to other solutions, it greatly reduces the number of switch units used, saving chip space and reducing overall device loss, further improving the chip's overall performance and integrability. This invention also provides a method for implementing the aforementioned multi-channel chalcogenide phase change material optical switch device based on the Benes architecture.
[0094] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and substitutions can be made without departing from the technical principles of the present invention, and these improvements and substitutions should also be considered within the scope of protection of the present invention.
Claims
1. A multi-channel chalcogenide phase change material optical switch device based on the Benes architecture, characterized in that, The device includes several 2×2 optical switches, all of which form a Benes architecture. Each optical switch includes an incident waveguide (7), a directional coupler, and an outgoing waveguide (8). The incident waveguide (7) is connected to the input end of the directional coupler, and the outgoing waveguide (8) is connected to the output end of the directional coupler. The directional coupler includes a first waveguide (3) and a second waveguide (4). A chalcogenide phase change material layer (5) is provided on the second waveguide (4). The first waveguide (3) and the second waveguide (4) are each connected to a metal electrode.
2. The multi-channel chalcogenide phase change material optical switch device based on the Benes architecture according to claim 1, characterized in that, The sulfur-based phase change material layer (5) is an Sb2Se3 layer.
3. The multi-channel chalcogenide phase change material optical switch device based on the Benes architecture according to claim 2, characterized in that, The Sb:Se composition ratio of the sulfur-based phase change material layer (5) is (20-50):(50-80).
4. The multi-channel chalcogenide phase change material optical switch device based on the Benes architecture according to claim 1, characterized in that, Both the first waveguide (3) and the second waveguide (4) are ridge waveguides.
5. The multi-channel chalcogenide phase change material optical switch device based on the Benes architecture according to claim 1, characterized in that, The optical switch further includes a substrate (1), an insulating layer (2), and a dielectric layer (6). The insulating layer (2) is disposed above the substrate (1). The incident waveguide (7), the first waveguide (3), the second waveguide (4), and the outgoing waveguide (8) are all disposed above the insulating layer (2). The dielectric layer (6) is disposed on the portion of the incident waveguide (7), the first waveguide (3), the second waveguide (4), the outgoing waveguide (8), and the portion of the insulating layer (2) not covered by the waveguide.
6. The multi-channel chalcogenide phase change material optical switch device based on the Benes architecture according to claim 1, characterized in that, The substrate (1) is a silicon wafer, the insulating layer (2) is a silicon dioxide buried oxide layer, and the dielectric layer (6) is aluminum oxide.
7. The multi-channel chalcogenide phase change material optical switch device based on the Benes architecture according to claim 1, characterized in that, The Benes architecture includes an input switching layer and an output switching layer, each of which includes multiple 2×2 optical switches, and the number of optical switches used in the input switching layer and the output switching layer is equal. The input switching layer uses recursive grouped input, and the ports of all optical switches in the input switching layer are divided into upper and lower parts, with each part recursively constructing a Benes sub-network; the output switching layer uses grouped recursive output, which is the opposite of the setting logic of the input switching layer.
8. The multi-channel chalcogenide phase change material optical switch device based on the Benes architecture according to claim 1, characterized in that, The device is an N×N optical switch network, with 2×2 optical switches having the following number of switches: The required number of optical switching stages can be expressed by the following formula: S = 2 log₂N⁻¹ Where S represents the number of network levels to be constructed, and N represents the number of optical switch ports to be implemented.
9. A method for implementing a multi-channel chalcogenide phase change material optical switch device based on the Benes architecture, characterized in that, The method using the optical switch according to any one of claims 1-6 includes the following steps: S1: Select the appropriate number of 2×2 optical switches according to the N×N optical switch network to be constructed; S2: Divide the optical switch into two equal parts, and the two optical switches constitute the input switching layer and output switching layer of the Benes architecture, respectively; S3: In constructing the input switching layer, the number of optical switches in each level is equal. Each optical switch is numbered, and the port of each optical switch is numbered. S4: According to the number, connect the two ports of the previous stage optical switch to the ports of the next stage optical switch to realize the connection between adjacent stages; each stage within the layer is connected to the two adjacent stages to complete the connection of the input switching layer; S5: Using the backtracking method, starting from the output port of the output switching layer, each level in the layer is connected to the two adjacent levels respectively, so that the output switching layer is symmetrical with the input switching layer, and the connection of the output switching layer is completed.
10. The method for implementing a multi-channel chalcogenide phase change material optical switch device based on the Benes architecture according to claim 9, characterized in that, In step S3, each optical switch port is assigned a number (X, Y, Z), where X represents the group number of each level, Y represents the number of each group of optical switches, and Z represents the number of each group of optical switch ports; the optical switches of the nth level can be divided into 2 equal groups. n-1 The Y and Z numbers for each group are obtained using the following formula: Y n,i =i Z n,j =j Where N is the number of ports, n is the number of stages, i represents the i-th optical switch in a group, and j represents the j-th port in a group; In step S4, in each group, the two ports of the Y-numbered optical switch of the previous stage are connected in ascending order to the ports of the Z-numbered optical switch of the next stage, which has a value equal to Y.