TE mode and TM mode parallel transmission low crosstalk cross waveguide and design method thereof

Abstract

This invention relates to the field of photonic integration technology, and discloses a low-crosstalk cross-waveguide for parallel transmission of TE and TM modes and its design method. The cross-waveguide includes a first waveguide, a second waveguide, a longitudinal multimode interference region, and a transverse multimode interference region. It is composed only of a standard straight waveguide and a gradually varying tapered plate, requiring no complex micro / nano structures or special materials. It is fully compatible with mainstream silicon photonics / silicon nitride processes, has high manufacturing tolerance, and high yield. The design method is based on mature multimode interference theory, and the physical picture of the design process is intuitive. It does not rely on computationally intensive inverse design or topology optimization, resulting in a short development cycle. This invention can simultaneously achieve low-loss, low-crosstalk parallel pass-through of TE and TM modes, effectively supporting polarization multiplexing technology and improving the channel capacity of on-chip optical interconnects. It provides a reliable and easily implemented basic cross-connection solution for high-density, high-performance polarization multiplexing photonic integrated chips.

Description

Technical Field

[0001] This invention relates to the field of photonic integration technology, and in particular to a low crosstalk cross waveguide for parallel transmission of TE and TM modes and its design method. Background Technology

[0002] In photonic integrated circuits, optical crossbars, which enable cross-connection of optical signal paths, are fundamental and critical passive devices. With the continuous improvement of chip integration and the widespread application of polarization multiplexing technology, cross-waveguides capable of simultaneously and efficiently handling two orthogonal polarization states (such as TE and TM modes) have become an important research direction for improving chip functional density and transmission capacity. Currently, existing technologies aiming to achieve low-crosstalk, cross-connected, parallel, and direct transmission mainly follow two different technical paths.

[0003] One mainstream technological approach is based on standard process materials, achieving functionality through complex local optimization and micro / nano structure design of the waveguide structure in the crossover region. The most representative approach involves introducing artificial microstructures such as subwavelength gratings, metasurfaces, or photonic crystals into the crossover region. The core principle is to artificially create an equivalent anisotropic optical environment by carefully designing the morphology, size, and arrangement of these subwavelength units, causing them to produce different effective refractive indices or phase responses for TE and TM modes, thereby guiding the light fields of the two polarization states to propagate with low crosstalk along a predetermined crossover path. While this method is theoretically feasible, its limitations are significant. First, the design process is extremely complex, relying on computationally intensive algorithms such as topology optimization and inverse design to search for the optimal microstructure, often resulting in an irregular and difficult-to-understand complex pattern. Second, the performance of such structures is strictly dependent on the fineness of the features, typically requiring the fabrication of complex geometries at deep subwavelength scales (e.g., tens of nanometers), which places extremely stringent demands on the precision and uniformity of photolithography and etching processes, resulting in very small manufacturing tolerances. This results in poor compatibility with mainstream silicon photonics or silicon nitride micro / nano fabrication processes, high manufacturing costs, and difficulty in guaranteeing yield, greatly hindering its large-scale application in actual chips. For the technical path based on local micro / nano structure optimization, its core drawback lies in its extremely high design complexity and manufacturing difficulty. The root cause of these drawbacks lies in its implementation principle itself. This method relies on introducing subwavelength-scale artificial microstructures in the intersection region to modulate the light field, which is essentially a complex electromagnetic inverse problem. To achieve precise and differentiated wavefront manipulation of both TE and TM modes simultaneously, the required structures typically lack an intuitive physical form and must rely on computationally intensive algorithms such as topology optimization and adjoint variable methods for iterative searches, making the design process complex, time-consuming, and heavily dependent on computing power. On the manufacturing side, these optimized structures often contain a large number of irregular, deep subwavelength features, and their performance is extremely sensitive to deviations in size and shape. This requires photolithography and etching processes to achieve near-limit nanometer-level precision and uniformity; any tiny process fluctuation will lead to a sharp deterioration in performance. Therefore, its high manufacturing difficulty and low process tolerance stem from its inherent reliance on extremely fine structures, making it incompatible with mainstream silicon photonics process platforms that emphasize standardization, reliability, and cost control. For technology paths based on special anisotropic materials, the main drawbacks are poor process compatibility and significant integration challenges. This drawback arises directly from the heterogeneity of the material system. Silicon-based photonics has become mainstream because it can fully utilize the mature complementary metal-oxide-semiconductor (CMOS) manufacturing ecosystem. However, special materials such as lithium niobate and indium phosphide possess physicochemical properties drastically different from silicon. Integrating them with silicon-based chips cannot directly utilize standard processes; complex back-end integration processes such as wafer bonding and heteroepitaxial growth must be introduced.This not only introduces additional interface losses and scattering centers, but also brings a series of engineering challenges such as thermal expansion coefficient mismatch, stress management, and reliability verification. Therefore, its high cost and high integration difficulty do not stem from the design itself, but are rooted in the fundamental mismatch between the two major systems of "special materials" and "standard processes" in terms of physical basis and manufacturing chain.

[0004] Another technological approach involves using specialized optical materials with strong anisotropy. This approach directly utilizes birefringent crystals such as lithium niobate and indium phosphide, or two-dimensional materials, leveraging the significant difference in the inherent refractive index of the material for TE and TM polarized light to naturally achieve the separation and directional guidance of the two modes. For example, waveguide cross-sections can be designed so that one polarization is cut off while the other passes through smoothly, or special mode evolution structures can be designed using the material's anisotropy. Compared to the first approach, this method is more direct in terms of physical concepts. However, its limitations are equally apparent. The most critical issues lie in process compatibility and integration difficulty. These specialized materials are often incompatible with the mainstream silicon-based complementary metal-oxide-semiconductor (CMOS) process platform, requiring complex heterogeneous integration processes, such as wafer bonding or epitaxial growth, which introduces additional interface losses, thermal stress matching, and reliability challenges. Simultaneously, the specialized materials themselves may be expensive, and the related microfabrication process chain is not as mature as silicon photonics, further increasing manufacturing costs. Furthermore, efficiently integrating these specialized material components with other standard functional modules on the chip based on silicon or silicon nitride is also a significant engineering challenge.

[0005] In summary, the two main existing technical solutions face irreconcilable contradictions in achieving the goal of "low crosstalk parallel pass-through between TE and TM modes." Methods based on local microstructure optimization are limited by the extreme complexity of design and manufacturing; while methods based on special materials are hampered by process compatibility and integration challenges. Neither can guarantee high performance while simultaneously meeting the core requirements of large-scale photonic integrated circuits for ease of design, process standardization, and controllable manufacturing costs. Therefore, there is an urgent need in this field for a new solution that can circumvent these limitations and achieve high-performance dual-polarization cross-waveguides within standard process and material systems through a simpler and more robust design approach. Summary of the Invention

[0006] The purpose of this invention is to provide a low-crosstalk cross-waveguide for parallel transmission of TE and TM modes that is clear in principle, simple in design, and easy to manufacture, and its design method.

[0007] To achieve the above objectives, the present invention provides a low crosstalk cross waveguide for parallel transmission of TE mode and TM mode, including a first waveguide, a second waveguide, a longitudinal multimode interference region and a transverse multimode interference region, wherein the first waveguide extends along a first direction and has a first port and a second port, the second waveguide extends along a second direction and has a third port and a fourth port, and the first direction is orthogonal to the second direction. The longitudinal multimode interference region is disposed along the path of the first waveguide in the first direction, and the first port and the second port are respectively connected to the two ends of the longitudinal multimode interference region; the transverse multimode interference region is disposed along the path of the second waveguide in the second direction, and the third port and the fourth port are respectively connected to the two ends of the transverse multimode interference region; the longitudinal multimode interference region and the transverse multimode interference region orthogonally intersect in the plane to form an intersection center; The longitudinal multimode interference region has a first self-image length, which is such that: for the TE mode, the first self-image point generated by the optical signal injected from the first port in the longitudinal multimode interference region is exactly located at the intersection center; the transverse multimode interference region has a second self-image length, which is such that: for the TM mode, the first self-image point generated by the optical signal injected from the third port in the transverse multimode interference region is exactly located at the intersection center.

[0008] As a preferred embodiment, the longitudinal multimode interference region and the transverse multimode interference region are each rectangular.

[0009] As a preferred embodiment, both the first waveguide and the second waveguide are straight waveguides. The two ends of the first waveguide are connected to the first port and the second port respectively through a taper structure, and the two ends of the second waveguide are connected to the third port and the fourth port respectively through a taper structure.

[0010] As a preferred embodiment, the taper structure is a linear cone.

[0011] As a preferred embodiment, the length of the longitudinal multimode interference region along the first direction is not equal to the width of the transverse multimode interference region along the first direction.

[0012] This invention also provides a design method for low crosstalk cross-waveguides for parallel transmission of TE and TM modes, comprising: S1. Establish a longitudinal multimode interference waveguide model, which includes a first waveguide, a first port, a second port, and a longitudinal multimode interference region. Apply a TE polarization state light source to the first port. Optimize the length and width of the longitudinal multimode interference region through simulation scanning. Determine the first structural parameters that make the first self-image point of the TE mode meet the preset low insertion loss transmission conditions. Obtain the first waveguide structure with the first structural parameters. S2. Establish a transverse multimode interference waveguide model, which includes a second waveguide, a third port, a fourth port, and a transverse multimode interference region. Apply a TM polarization state light source to the third port, optimize the length and width of the transverse multimode interference region through simulation scanning, determine the second structural parameters that make the first self-image point of the TM mode meet the preset low insertion loss transmission conditions, and obtain a second waveguide structure with the second structural parameters. S3. Arrange the first waveguide structure obtained in step S1 and the second waveguide structure obtained in step S2 orthogonally on the planar layout so that their physical center lines intersect. Force the first self-image point of the two to be aligned in space to an alignment point, which is defined as the core intersection point of the entire device, thereby forming a preliminary cross waveguide structure. S4. The preliminary cross-waveguide structure formed in step S3 is modeled and simulated as a whole to obtain the transmission characteristic parameters of TE mode and TM mode. It is determined whether the transmission characteristic parameters meet the preset requirements of low insertion loss and low crosstalk. If not, the structural parameters of the longitudinal multimode interference region and / or the transverse multimode interference region are adjusted, and the overall modeling and simulation are repeated until the preset requirements of low insertion loss and low crosstalk are met, so as to obtain the final device parameters that meet the performance indicators.

[0013] As a preferred embodiment, in step S1, the longitudinal multimode interference waveguide model further includes a first taper structure connecting the first waveguide and the first port. When a TE polarization state light source is applied to the first port, the tapered waveguide is excited through the first taper structure so that the optical signal smoothly enters the longitudinal multimode interference region in the form of the fundamental mode. In step S2, the transverse multimode interference waveguide model further includes a second taper structure connected to the first port of the first waveguide. When a TE polarization state light source is applied to the third port, the second taper structure is used to excite the tapered waveguide so that the optical signal enters the transverse multimode interference region smoothly in the form of the fundamental mode.

[0014] As a preferred option, the simulation scans in steps S1 and S2 employ the finite-difference time-domain method.

[0015] As a preferred embodiment, the preset low insertion loss transmission conditions in steps S1 and S2 include at least one of the following: self-image condition, mode matching condition, destructive interference condition, multimode support condition, and polarization adaptation condition.

[0016] As a preferred embodiment, in step S3, the preset requirements for low insertion loss and low crosstalk include both high transmittance and low crosstalk operating bandwidth.

[0017] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention addresses the inherent problems of high polarization crosstalk and high insertion loss from a physical perspective. By employing a cross-waveguide core design based on multimode interference principles, combined with a straight waveguide and taper structure, a waveguide configuration with a simple structure and clear physical picture is constructed. This method aims to precisely control the evolution paths and phase relationships of the TE and TM modes in the cross-region, thereby achieving low-crosstalk, low-loss parallel directional pass-through of the two polarization states, thus improving the signal transmission quality at the cross-node.

[0018] Secondly, this invention aims to provide a design method with a simple design process and rapid optimization. Addressing existing solutions that rely on complex inverse design or special materials, the method proposed in this invention is based on the principle of multimode interference, with intuitive design parameters and clear physical meaning. This aims to enable designers to efficiently complete designs using conventional simulation optimization methods, avoiding reliance on computationally intensive algorithms such as topology optimization, significantly reducing design complexity and development cycle.

[0019] Finally, this invention strives to achieve high compatibility with standard photonic integration processes and improve manufacturing robustness. The waveguide structure designed in this invention consists entirely of standard straight waveguides and gradually tapered waveguide structures, without containing any deep subwavelength complex micro / nano structures or anisotropic materials. This aims to ensure that the structure can be manufactured entirely using mainstream silicon-based or silicon nitride process platforms, has higher tolerance to process variations, thereby effectively reducing manufacturing costs, improving product yield, and meeting the reliability and consistency requirements of large-scale photonic integrated circuits.

[0020] Therefore, this invention provides a waveguide structure design method that is clear in principle, simple in design, and process-friendly, so as to simultaneously achieve excellent dual polarization transmission performance, efficient design process and perfect compatibility with standard processes, thereby providing a key basic device solution for high-density, high-performance polarization multiplexing photonic integrated chips. Attached Figure Description

[0021] Figure 1 This is a schematic diagram of the structure of the parallel transmission low crosstalk cross waveguide of TE mode and TM mode according to an embodiment of the present invention; Figure 2This is a detailed flowchart of the design method for low crosstalk cross waveguides for parallel transmission of TE mode and TM mode according to an embodiment of the present invention; Figure 3 This is a summary flowchart of the design method for low crosstalk cross-waveguides for parallel transmission of TE mode and TM mode according to an embodiment of the present invention. Figure 4 This is a transmission spectrum diagram of the cross waveguide according to an embodiment of the present invention; Figure 5 This is the transmission and crosstalk curve diagram of the longitudinal multimode interference region in step S1 of the present invention, which only optimizes the TE mode transmission. Figure 6 In this embodiment of the invention, step S2 only optimizes the transmittance and crosstalk curves of the transverse multimode interference region during TM mode transmission. Figure 7 This is a mode field distribution diagram of the overall cross-waveguide device according to an embodiment of the present invention at a specific wavelength; Figure 8 This is a mode field distribution diagram of the first port of an embodiment of the present invention at a specific wavelength; Figure 9 This is a mode field distribution diagram of the second port of an embodiment of the present invention at a specific wavelength; Figure 10 This is a mode field distribution diagram of the third port of an embodiment of the present invention at a specific wavelength; Figure 11 This is a mode field distribution diagram of the fourth port of an embodiment of the present invention at a specific wavelength.

[0022] In the figure, 1-first waveguide; 101-longitudinal multimode interference region; 2-second waveguide; 201-lateral multimode interference region; 3-first port; 4-second port; 5-third port; 6-fourth port; 7-taper structure. Detailed Implementation

[0023] 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.

[0024] 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.

[0025] 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.

[0026] Furthermore, in the description of this invention, unless otherwise stated, "a plurality of" means two or more.

[0027] Example 1 like Figure 1 As shown, a preferred embodiment of the present invention provides a low crosstalk cross-waveguide for parallel transmission of TE and TM modes, comprising a first waveguide 1, a second waveguide 2, a longitudinal multimode interference region 101, and a transverse multimode interference region 201. The first waveguide 1 extends along a first direction and has a first port 3 and a second port 4. The second waveguide 2 extends along a second direction and has a third port 5 and a fourth port 6. The first direction is orthogonal to the second direction. The longitudinal multimode interference region 101 is disposed along the path of the first waveguide 1 along the first direction, and the first port 3 and the second port 4 are respectively connected to the two ends of the longitudinal multimode interference region 101. The transverse multimode interference region 201 is disposed along the path of the second waveguide 2 along the second direction, and the first port 3 and the second port 4 are respectively connected to the two ends of the longitudinal multimode interference region 101. Port 5 and the fourth port 6 are respectively connected to the two ends of the transverse multimode interference region 201; the longitudinal multimode interference region 101 and the transverse multimode interference region 201 intersect orthogonally in the plane, forming a cross center; the longitudinal multimode interference region 101 has a first self-image length, which is such that: for the TE mode, the first self-image point generated by the optical signal injected from the first port 3 in the longitudinal multimode interference region 101 is exactly located at the cross center; the transverse multimode interference region 201 has a second self-image length, which is such that: for the TM mode, the first self-image point generated by the optical signal injected from the third port 5 in the transverse multimode interference region 201 is exactly located at the cross center.

[0028] The cross waveguide designed in this embodiment consists of two spatially orthogonal multimode interference regions with overlapping central regions. The longitudinal multimode interference region 101 dominates the transmission and routing of the TE mode, while the transverse multimode interference region 201 dominates the transmission and routing of the TM mode. Ports 3, 4, 5, and 6 are used for connections to other optical paths on the chip.

[0029] The core working principle of this embodiment is to utilize the self-image effect of multimode interference and combine it with the inherent mode field distribution differences between the TE mode and the TM mode in the waveguide to achieve low crosstalk transmission at the spatial intersection point. Based on the path separation principle, for a specific polarization state (such as the TE mode), when it is injected from the port into the corresponding multimode interference region (such as the longitudinal multimode interference region 101), the input mode field will be reproduced at a specific location (called the "self-image point"). By precisely designing the length of this multimode interference region, its first self-image point is located exactly at the intersection center with another multimode interference region (the transverse multimode interference region 201). Similarly, the length of the transverse multimode interference region 201 is independently designed for the TM mode, so that its first self-image point is also located at the same intersection center. In this way, the predetermined optical paths of the TE mode and the TM mode intersect orthogonally at the intersection point in space. Based on the principle of low crosstalk, although the optical paths of the TE mode and the TM mode overlap at the intersection point, the mode field distributions of the two modes are spatially offset because the field energy of the TE mode is mainly concentrated in the center of the waveguide core layer, while the field energy of the TM mode is closer to the interface between the waveguide core layer and the upper and lower cladding layers. When their first self-image points are aligned in this overlapping region, the energy centers of their mode fields can still be well aligned in space to complete their respective transmissions. However, the interaction between the mode fields in the overlapping region is very weak, which greatly suppresses the coupling between modes physically and achieves low crosstalk. Based on the principle of direct transmission, the optical signal enters from the input port, is extended through the tapered waveguide, and enters the multimode interference region. It completes the "crossing" at the pre-designed first self-image point (i.e., the intersection point), and then continues to propagate forward, finally converging again at the output port to form the fundamental mode output, realizing a low-loss 1x1 direct transmission function.

[0030] In this embodiment, the longitudinal multimode interference region 101 and the transverse multimode interference region 201 are each rectangular. Furthermore, both the first waveguide 1 and the second waveguide 2 are straight waveguides. The two ends of the first waveguide 1 are connected to the first port 3 and the second port 4 respectively via a taper structure 7, and the two ends of the second waveguide 2 are connected to the third port 5 and the fourth port 6 respectively via a taper structure 7. Optionally, the taper structure 7 is a linear cone.

[0031] This embodiment employs four taper structures 7 to connect four ports to their corresponding multimode interference regions, achieving efficient mode matching with standard single-mode straight waveguides and reducing coupling loss. The four input / output ports—first port 3, second port 4, third port 5, and fourth port 6—are located outside their respective taper structures 7. The taper structure 7 refers to a waveguide that changes slowly and continuously along the direction of light propagation, used to achieve smooth mode transitions between waveguides of different sizes or types. The taper structure 7 can be linear tapered, parabolic, or exponential; in this embodiment, the taper structure 7 is linear tapered. The first waveguide 1 and second waveguide 2 are straight waveguide segments, and the taper structure 7 is a gradually tapered waveguide. Therefore, the waveguide structure in this embodiment is entirely composed of straight waveguide segments and gradually tapered waveguide structures, without containing any deep subwavelength complex micro / nano structures or anisotropic materials. This ensures that the structure can be manufactured using mainstream silicon-based or silicon nitride process platforms, exhibiting higher tolerance to process variations, thereby effectively reducing manufacturing costs, improving product yield, and meeting the reliability and consistency requirements of large-scale photonic integrated circuits.

[0032] Furthermore, in this embodiment, the length L of the longitudinal multimode interference region 101 along the first direction is not equal to the width W of the transverse multimode interference region 201 along the first direction. This results in the two multimode interference regions having different dimensional parameters in the orthogonal direction. This structural feature provides geometric freedom for independently optimizing the self-image lengths of the TE mode and TM mode, avoiding the defect that the two polarization states cannot simultaneously satisfy the self-image condition due to the use of a symmetrical square cross structure, thereby achieving low crosstalk transmission in all polarization states.

[0033] Example 2 like Figures 2 to 11 As shown, a preferred embodiment of the present invention provides a design method for a low-crosstalk cross-waveguide for parallel transmission of TE and TM modes, comprising: S1. Establish a longitudinal multimode interference waveguide model. The longitudinal multimode interference waveguide model includes a first waveguide 1, a first port 3, a second port 4, and a longitudinal multimode interference region 101. Apply a TE polarization state light source to the first port 3. Optimize the length and width of the longitudinal multimode interference region 101 through simulation scanning. Determine the first structural parameters that make the first self-image point of the TE mode meet the preset low insertion loss transmission conditions. Obtain the first waveguide 1 structure with the first structural parameters. S2. Establish a transverse multimode interference waveguide model. The transverse multimode interference waveguide model includes a second waveguide 2, a third port 5, a fourth port 6, and a transverse multimode interference region 201. Apply a TM polarization state light source to the third port 5. Optimize the length and width of the transverse multimode interference region 201 through simulation scanning. Determine the second structural parameters that make the first self-image point of the TM mode meet the preset low insertion loss transmission condition. Obtain the second waveguide 2 structure with the second structural parameters. S3. Arrange the first waveguide 1 structure obtained in step S1 and the second waveguide 2 structure obtained in step S2 orthogonally on the planar layout so that their physical center lines intersect. Force the first self-image points of the two to be aligned in space to an alignment point, which is defined as the core intersection point of the entire device, thereby forming a preliminary cross waveguide structure. Make the longitudinal multimode interference region 101 and the transverse multimode interference region 201 orthogonally intersect in space, and make the end point of the longitudinal multimode interference region 101 segment coincide with the end point of the transverse multimode interference region 201 segment at the intersection center of the orthogonal intersection. S4. Model and simulate the preliminary cross-waveguide structure formed in step S3 as a whole to obtain the transmission characteristic parameters of TE mode and TM mode; determine whether the transmission characteristic parameters meet the preset requirements of low insertion loss and low crosstalk. If not, adjust the structural parameters of the longitudinal multimode interference region 101 and / or the transverse multimode interference region 201, and repeat the overall modeling and simulation until the preset requirements of low insertion loss and low crosstalk are met, and obtain the final device parameters that meet the performance indicators.

[0034] Furthermore, in step S1, the longitudinal multimode interference waveguide model also includes a first taper structure connecting the first waveguide 1 and the first port 3. When a TE polarization state light source is applied to the first port 3, the tapered waveguide is excited through the first taper structure so that the optical signal smoothly enters the longitudinal multimode interference region 101 in the form of the fundamental mode. In step S2, the transverse multimode interference waveguide model also includes a second taper structure that connects the first waveguide 1 to the first port 3. When a TE polarization state light source is applied to the third port 5, the tapered waveguide is excited through the second taper structure so that the optical signal smoothly enters the transverse multimode interference region 201 in the form of the fundamental mode.

[0035] The device of this invention consists only of a straight waveguide and a gradually varying tapered plate, requiring no complex micro / nano structures or special materials. Its simple structure is fully compatible with mainstream silicon photonics / silicon nitride processes, offering high manufacturing tolerance and yield. Furthermore, the design method is clear and direct, based on mature multimode interference theory. The physical picture of the design process is intuitive, eliminating the need for computationally intensive inverse design or topology optimization, resulting in a short development cycle. This invention enables low-loss, low-crosstalk parallel pass-through of both TE and TM modes, effectively supporting polarization multiplexing technology, increasing the channel capacity of on-chip optical interconnects, and exhibiting excellent dual-polarization performance. This invention provides a reliable and easily implemented basic cross-connection solution for high-density, high-performance polarization-multiplexed photonic integrated chips.

[0036] Therefore, the present invention has the following advantages: (1) simple structure, high process compatibility, and large manufacturing tolerance; (2) clear and intuitive design principle, direct and efficient method, and short development cycle; (3) can realize high-performance parallel processing of TE and TM modes at the same time, effectively supporting polarization multiplexing; (4) realizes low-loss through transmission function with low insertion loss; (5) practical function, providing a reliable solution for high-density photonic integration.

[0037] Specifically, existing technologies often require the introduction of micro / nano structures such as subwavelength gratings, metasurfaces, or complex curved waveguides to achieve low crosstalk. These structures are extremely sensitive to fabrication precision, resulting in high manufacturing difficulty and low yield. This invention eliminates all complex micro / nano structures, with its core consisting solely of a standard straight waveguide segment and a gradually varying tapered waveguide orthogonally arranged. This structure is entirely based on standard waveguide processes from mainstream silicon photonics or silicon nitride process platforms, requiring no special materials or unconventional processing steps. This minimalist structure and perfect compatibility with standard processes allow for greater process tolerance, facilitating high-yield, low-cost mass production.

[0038] While some existing high-performance cross-waveguide designs (such as those based on reverse engineering or topology optimization) offer excellent performance, their design processes heavily rely on computationally intensive global optimization algorithms, resulting in unclear physical images and often complex, difficult-to-adjust structures. This invention, based on mature multimode interferometry (MMI) self-image theory, presents a design process with a clear physical picture: precisely controlling the spatial intersection of the first self-image points of the TE and TM modes by independently designing the lengths of the two MMI regions. The proposed design method features clear steps and a direct flow (S1 independently designing the TE path, S2 independently designing the TM path, S3 integration, S4 fine-tuning), eliminating the need for "black box" reverse engineering. This principle-driven, step-by-step approach significantly reduces design difficulty and substantially shortens the device design and development cycle.

[0039] Many existing cross-waveguides are typically optimized only for a single polarization (such as the TE mode), or while supporting dual polarization, their crosstalk performance is limited. This invention cleverly utilizes the inherently different mode field distribution characteristics of the TE and TM modes in the waveguide (TE mode energy is concentrated at the core center, while TM mode energy is closer to the interface). By precisely aligning the first self-image points of both modes at the cross-center, the two light paths physically converge, but due to the spatial offset of their mode field energy centers, inter-mode coupling is greatly suppressed. Therefore, this invention can provide a low insertion loss and low crosstalk through-path transmission for both TE and TM polarization states on a single device, providing a key foundational component for realizing on-chip polarization multiplexing systems and directly improving the channel capacity and signal transmission efficiency of photonic integrated circuits.

[0040] This invention not only focuses on crosstalk suppression but also strives to minimize transmission loss. By introducing carefully designed tapered, gradually varying waveguides at the four ports, efficient mode matching between the MMI region and the standard single-mode waveguide is achieved, significantly reducing end-face reflection and mode mismatch loss. Simultaneously, the optical path based on the MMI self-image effect inherently possesses high transmission efficiency. After overall parameter optimization in step S4, the device can simultaneously guarantee low insertion loss in both TE and TM modes over a wide operating bandwidth, meeting the stringent requirements for signal integrity (insertion loss, bandwidth) in practical applications.

[0041] In on-chip optical interconnect networks that pursue higher integration density and more complex functions, low crosstalk and low-loss optical path crossing are essential basic functions. The cross waveguide solution provided by this invention has a compact structure, excellent performance, and is easy to manufacture. It provides a reliable, practical, and easy-to-implement basic connection unit for building large-scale, high-performance polarization-multiplexed photonic integrated circuits, and has significant practical application value and market prospects.

[0042] In one specific embodiment, the simulation scans in steps S1 and S2 employ the finite-difference time-domain method.

[0043] The low insertion loss transmission conditions preset in steps S1 and S2 include at least one of the following: self-image condition, mode matching condition, destructive interference condition, multimode support condition, and polarization adaptation condition.

[0044] The self-image condition refers to the following: the length of the longitudinal multimode interference region 101 is equal to the first self-image length of the TE mode, so that the TE mode optical signal injected from the first port 3 forms the first self-image point at the cross center; and the length of the transverse multimode interference region 201 is equal to the second self-image length of the TM mode, so that the TM mode optical signal injected from the third port 5 forms the first self-image point at the cross center. The mode matching condition refers to the following: the first waveguide 1 is connected to both ends of the longitudinal multimode interference region 101 through the taper structure 7, and the second waveguide 2 is connected to both ends of the transverse multimode interference region 201 through the taper structure 7, to suppress the excitation of higher-order modes when entering the multimode interference region from the single-mode waveguide. The destructive interference condition refers to the following: at the cross center, the TE mode optical field forms a node distribution with zero field strength along the axis along the second direction, and the propagation direction of the TE mode and the node distribution direction are orthogonal; the TM mode optical field forms a node distribution with zero field strength along the axis along the first direction, and the propagation direction of the TM mode and the node distribution direction are orthogonal; to suppress energy leakage to non-target ports. The multimode support condition refers to the fact that the width of the longitudinal multimode interference region 101 and the width of the transverse multimode interference region 201 are each greater than the cutoff width of the single-mode waveguide, so as to support at least three transverse modes. The polarization adaptation condition refers to the fact that the length of the longitudinal multimode interference region 101 is different from the length of the transverse multimode interference region 201, so as to adapt to the difference in propagation constants of the TE mode and the TM mode respectively, so that the two polarization states simultaneously satisfy their respective self-image conditions; therefore, the length of the longitudinal multimode interference region 101 in this invention is not equal to the length and width of the transverse multimode interference region 201, and the width of the longitudinal multimode interference region 101 is not equal to the length and width of the transverse multimode interference region 201.

[0045] Furthermore, in step S3, the preset requirements for low insertion loss and low crosstalk simultaneously include high transmittance, low crosstalk, and wide operating bandwidth.

[0046] The high transmittance index refers to the following: at the operating wavelength, the transmittance of the TE-mode optical signal input from the first port 3 to the second port 4 is greater than or equal to a first preset threshold, and the transmittance of the TM-mode optical signal input from the third port 5 to the fourth port 6 is greater than or equal to the first preset threshold. The low crosstalk index refers to the following: at the operating wavelength, the power of the TE-mode optical signal leaking from the first port 3 to the third port 5 and the fourth port 6 is less than or equal to a second preset threshold, and the power of the TM-mode optical signal leaking from the third port 5 to the first port 3 and the second port 4 is less than or equal to a second preset threshold. The wide operating bandwidth index refers to the following: within a preset wavelength range, the transmittance of both the TE and TM modes at their respective target ports remains greater than or equal to the first preset threshold, and the crosstalk at non-target ports remains less than or equal to the second preset threshold.

[0047] like Figure 2As shown, step S1 of this embodiment designs the multimode interference region for transmitting the TE mode. Using electromagnetic simulation tools such as the finite-difference time-domain (FDTD) method, a longitudinal multimode interference waveguide model is established. A TE-polarized light source is applied at the input end, excited through a tapered waveguide. The length and width of this multimode interference region are optimized by scanning to produce a clear self-image sequence at a specific propagation length. An optimal length is determined such that at this length, a suitable first self-image point appears as a crossover point, and the light field can continue to propagate forward and form a high-quality final output image at the output end. The structural parameters of the MMI region at this point are recorded. Step S2 designs the multimode interference region for transmitting the TM mode. The same method as in step S1 is used, but a transverse multimode interference waveguide model is established, and a TM-polarized light source is applied. Its length and width are independently optimized by scanning, with the goal of matching the position of its first self-image point with the first self-image point of the TE mode in step S1 in terms of performance, and forming a good final output image. The structural parameters of this MMI region are recorded. Step S3 constructs the cross waveguide and determines the crossover point. The transverse and longitudinal multimode interference waveguides designed in steps S1 and S2 are orthogonally arranged on the layout so that the physical center lines of the two waveguides intersect. The first self-image points of the two are aligned in space, and this alignment point is defined as the center intersection point of the entire cross waveguide. Thus, a preliminary cross waveguide structure is formed. Step S4 is to perform parameter optimization and performance verification. The preliminary structure obtained in step S3 is modeled and simulated as a whole. The precise length and width of the two MMI regions, as well as the size of the tapered waveguide, are scanned and optimized with small parameters. The optimization goal is to achieve the following simultaneously under TE and TM polarization inputs: (1) high transmittance (low insertion loss); (2) low crosstalk (extremely low power leakage from TE to TM port and from TM to TE port); (3) wide operating bandwidth. Through iterative optimization, the final device parameters that meet the performance indicators are obtained.

[0048] This invention abandons the traditional approach of designing complex micro / nano structures or relying on special materials in the intersection region. It proposes and implements for the first time a cross-waveguide configuration based on the principle of "dual-path independent optimized multimode interference." This configuration cleverly utilizes the spatial difference in the eigenmode field distributions of the two polarization states by independently designing and orthogonally integrating two dedicated multimode interference waveguides for the TE and TM modes in space, and precisely controlling the overlap of their first self-image points as the intersection point. This achieves low crosstalk and low-loss parallel direct transmission of the two polarization states simultaneously, with an extremely simple structure and using only standard processes. This method achieves a unity of high performance, high process compatibility, and low design complexity from a physical principle perspective. This invention is based on the design method of "dual-path independent optimized multimode interference." The core of this design concept lies in designing a dedicated multimode interference waveguide path independently for the TE and TM modes, and making them orthogonally intersect in space. This design concept involves independently determining the structural parameters (especially the length) of the multimode interference region for different polarization states; and a construction method that precisely overlaps the first self-image point (i.e., the middle image point in the 1x1 straight path) of two independently designed interference structures in space, using this point as the physical crossover center of the entire device. This approach is the fundamental principle for achieving low crosstalk parallel transmission with dual polarization. Furthermore, a waveguide functional structure of "orthogonal integrated dual-path multimode interference" was designed. This structure is characterized by containing a horizontally arranged first multimode interference waveguide and a vertically arranged second multimode interference waveguide, both orthogonally arranged in the plane with their waveguide regions overlapping at the center, forming a crossover region. This structure explicitly excludes crossover region designs based on complex micro / nano structures such as subwavelength gratings, photonic crystals, or metasurfaces, and also excludes the use of anisotropic crystal materials as the core functional medium. The design method of this invention is a complete design and optimization process, including complete steps from independent design and cross-synthesis to overall optimization. Specifically, this includes: determining the lengths of the longitudinal and transverse multimode interference regions 201 that satisfy low-loss direct transmission of TE and TM modes respectively; arranging the two orthogonally and aligning the self-image points to determine the cross center; and performing parameter scanning and collaborative optimization on the overall structure, especially the cross-over overlapping area, to simultaneously achieve low crosstalk and low-loss performance of TE and TM modes.

[0049] This invention proposes and implements a cross-waveguide configuration based on the principle of "dual-path independent optimized multimode interference". This configuration uses two dedicated multimode interference waveguides, designed independently for the TE mode and the TM mode respectively and orthogonally integrated in space. It precisely controls the overlap of their first self-image points as the intersection point, and utilizes the spatial difference in the distribution of the eigenmode fields of the two polarization states across the waveguide cross-section. With an extremely simple structure and using standard manufacturing processes, it achieves low crosstalk and low-loss parallel direct transmission of dual polarization states. This method unifies high performance, high process compatibility, and low design complexity at the physical principle level.

[0050] Figure 1 A schematic diagram of the basic structure of the cross waveguide designed in this invention is shown. This structure comprises two spatially orthogonal multimode interference waveguides that overlap in the central region, specifically including a horizontal multimode interference region, a vertical multimode interference region, four tapered, gradually varying waveguides, and four input / output ports. Figure 1 As shown, a horizontally oriented multimode interferometric waveguide is designed to transmit the TE mode, and a vertically oriented multimode interferometric waveguide is designed to transmit the TM mode. The two waveguides are designed so that their first self-image points coincide, serving as the crossover center. Since the field energy of the TE mode is mainly concentrated in the central region of the waveguide core, while the field energy of the TM mode is more distributed near the interface between the upper and lower cladding layers, there is a spatial offset in the vertical direction of the mode field distributions. When the first self-image points of the two waveguides are aligned in this overlapping region, their energy centers can still be efficiently transmitted along predetermined paths, while the interaction of the mode fields in the overlapping region is significantly suppressed, thus achieving low crosstalk in a physical mechanism.

[0051] Figure 4 It is a transmission spectrum diagram, also called a wavelength scan transmittance curve, which shows the through transmission spectrum and crosstalk performance of the cross waveguide designed in this invention, that is, the law of device performance changing with wavelength. Figure 4 This indicates that the cross waveguide exhibits both low insertion loss and low polarization crosstalk characteristics within the target band.

[0052] Figure 5 and Figure 6 The performance simulation results of the independent multimode interferometry regions corresponding to steps S1 and S2 in the design process are presented respectively. Figure 5 The transmittance and crosstalk curves corresponding to the vertical multimode interference region when only TE mode transmission is optimized; Figure 6 Transmittance and crosstalk curves corresponding to the horizontal multimode interference region when only TM mode transmission is optimized.

[0053] Figure 7 The image illustrates the mode field distribution of the cross waveguide at a specific wavelength. It is clearly observed that the optical fields of the TM mode in the horizontal multimode interference waveguide and the TE mode in the vertical multimode interference waveguide coincide and cross at the first self-image point, achieving spatial convergence and separation of the optical paths. This mode field distribution visually verifies the core design concept of this invention.

[0054] Figure 8 , Figure 9 , Figure 10 , Figure 11The diagram illustrates the mode field distribution characteristics of the cross-waveguide at its input and output ports under the operating wavelength. The waveguide transmitting horizontally has its left port as input and its right port as output, transmitting the TM mode; the waveguide transmitting vertically has its upper port as input and its lower port as output, transmitting the TE mode. The figure clearly shows that the output mode field of the optical signal after passing through the waveguide cross-section maintains its original polarization mode, visually verifying the physical mechanism by which this structure achieves low crosstalk parallel pass-through with dual polarization states. The upper port is port 3 (first port), the lower port is port 4 (second port), the left port is port 5 (third port), and the right port is port 6 (fourth port).

[0055] In summary, this invention provides a low-crosstalk cross-waveguide for parallel transmission of TE and TM modes and its design method. The cross-waveguide consists only of a standard straight waveguide and a gradually varying tapered plate, requiring no complex micro / nano structures or special materials. It is fully compatible with mainstream silicon photonics / silicon nitride processes, offering high manufacturing tolerance and yield. The design principle is clear and the method is direct, based on mature multimode interference theory. The physical picture of the design process is intuitive, eliminating the need for computationally intensive inverse design or topology optimization, resulting in a short development cycle. It exhibits excellent dual-polarization performance, enabling low-loss, low-crosstalk parallel pass-through of TE and TM modes simultaneously, effectively supporting polarization multiplexing technology and increasing the channel capacity of on-chip optical interconnects. Functionally practical, it provides a reliable and easily implemented basic cross-connect solution for high-density, high-performance polarization-multiplexed photonic integrated chips.

[0056] 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 low-crosstalk cross-waveguide for parallel transmission of TE mode and TM mode, characterized in that, It includes a first waveguide (1), a second waveguide (2), a longitudinal multimode interference region (101), and a transverse multimode interference region (201). The first waveguide (1) extends along a first direction and has a first port (3) and a second port (4). The second waveguide (2) extends along a second direction and has a third port (5) and a fourth port (6). The first direction is orthogonal to the second direction. The longitudinal multimode interference region (101) is disposed along the first direction on the path of the first waveguide (1), and the first port (3) and the second port (4) are respectively connected to the two ends of the longitudinal multimode interference region (101); the transverse multimode interference region (201) is disposed along the second direction on the path of the second waveguide (2), and the third port (5) and the fourth port (6) are respectively connected to the two ends of the transverse multimode interference region (201); the longitudinal multimode interference region (101) and the transverse multimode interference region (201) intersect orthogonally in the plane to form an intersection center; The longitudinal multimode interference region (101) has a first self-image length, which is such that: for the TE mode, the first self-image point generated by the optical signal injected from the first port (3) in the longitudinal multimode interference region (101) is exactly located at the intersection center; the transverse multimode interference region (201) has a second self-image length, which is such that: for the TM mode, the first self-image point generated by the optical signal injected from the third port (5) in the transverse multimode interference region (201) is exactly located at the intersection center.

2. The low crosstalk cross-waveguide for parallel transmission of TE mode and TM mode according to claim 1, characterized in that, The longitudinal multimode interference region (101) and the transverse multimode interference region (201) are each rectangular.

3. The low crosstalk cross-waveguide for parallel transmission of TE mode and TM mode according to claim 2, characterized in that, Both the first waveguide (1) and the second waveguide (2) are straight waveguides. The two ends of the first waveguide (1) are connected to the first port (3) and the second port (4) respectively through the taper structure (7). The two ends of the second waveguide (2) are connected to the third port (5) and the fourth port (6) respectively through the taper structure (7).

4. The low crosstalk cross-waveguide for parallel transmission of TE mode and TM mode according to claim 3, characterized in that, The taper structure (7) is a linear cone.

5. The low crosstalk cross-waveguide for parallel transmission of TE mode and TM mode according to claim 1, characterized in that, The length of the longitudinal multimode interference region (101) along the first direction is not equal to the width of the transverse multimode interference region (201) along the first direction.

6. A design method for a low-crosstalk cross-waveguide for parallel transmission of TE mode and TM mode, characterized in that, include: S1. Establish a longitudinal multimode interference waveguide model, which includes a first waveguide (1), a first port (3), a second port (4) and a longitudinal multimode interference region (101). Apply a TE polarization state light source to the first port (3), optimize the length and width of the longitudinal multimode interference region (101) through simulation scanning, determine the first structural parameter that makes the first self-image point of the TE mode meet the preset low insertion loss transmission condition, and obtain the first waveguide (1) structure with the first structural parameter. S2. Establish a transverse multimode interference waveguide model. The transverse multimode interference waveguide model includes a second waveguide (2), a third port (5), a fourth port (6), and a transverse multimode interference region (201). Apply a TM polarization state light source to the third port (5). Optimize the length and width of the transverse multimode interference region (201) through simulation scanning. Determine the second structural parameters that make the first self-image point of the TM mode meet the preset low insertion loss transmission conditions. Obtain the second waveguide (2) structure with the second structural parameters. S3. Arrange the first waveguide (1) structure obtained in step S1 and the second waveguide (2) structure obtained in step S2 orthogonally on the planar layout so that their physical center lines intersect. Force the first self-image point of the two to be aligned to an alignment point in space. This alignment point is defined as the core intersection point of the entire device, thereby forming a preliminary cross waveguide structure. S4. The preliminary cross-waveguide structure formed in step S3 is modeled and simulated as a whole to obtain the transmission characteristic parameters of TE mode and TM mode. It is determined whether the transmission characteristic parameters meet the preset requirements of low insertion loss and low crosstalk. If not, the structural parameters of the longitudinal multimode interference region (101) and / or the transverse multimode interference region (201) are adjusted, and the overall modeling and simulation are repeated until the preset requirements of low insertion loss and low crosstalk are met, so as to obtain the final device parameters that meet the performance indicators.

7. The design method according to claim 6, characterized in that, In step S1, the longitudinal multimode interference waveguide model further includes a first taper structure connecting the first waveguide (1) and the first port (3). When a TE polarization state light source is applied to the first port (3), the tapered waveguide is excited through the first taper structure so that the optical signal smoothly enters the longitudinal multimode interference region (101) in the form of the fundamental mode. In step S2, the transverse multimode interference waveguide model also includes a second taper structure that connects the first waveguide (1) to the first port (3). When a TE polarization state light source is applied to the third port (5), the second taper structure is used to excite the tapered waveguide so that the optical signal enters the transverse multimode interference region (201) smoothly in the form of the fundamental mode.

8. The design method according to claim 6, characterized in that, The simulation scans in steps S1 and S2 employ the finite-difference time-domain method.

9. The design method according to claim 6, characterized in that, The preset low insertion loss transmission conditions mentioned in steps S1 and S2 include at least one of the following: self-image condition, mode matching condition, destructive interference condition, multimode support condition, and polarization adaptation condition.

10. The design method according to claim 6, characterized in that, In step S3, the preset requirements for low insertion loss and low crosstalk include high transmittance, low crosstalk, and operating bandwidth.

Images