Optical frequency mixer, fabrication control method, and silicon optical integrated circuit

Abstract

This application relates to an optical mixer, a fabrication and control method, and a silicon photonics integrated circuit, comprising: a silicon device layer having a modulation region and a first input port, a second input port, and multiple output ports respectively connected to the modulation region; the modulation region includes multiple discretely distributed dielectric pillars, which, together with the remaining silicon region of the modulation region, form the equivalent refractive index distribution of the modulation region, wherein the refractive index of the dielectric pillars differs from the refractive index of the remaining silicon region of the modulation region; wherein the modulation region is configured such that when a first optical signal is input individually to the first input port, the multiple output ports output multiple optical signals with a first phase relationship, and that when a second optical signal is input individually to the second input port, the multiple output ports output multiple optical signals with a second phase relationship. The above technical solution can meet the requirements of phase accuracy and device size for optical mixers.

Description

Technical Field

[0001] This application relates to the field of semiconductor technology, specifically to an optical mixer, a fabrication and control method, and a silicon photonic integrated circuit. Background Technology

[0002] Optical mixers are core passive devices in coherent optical communication, coherent photoradar, and quantum communication systems. They enable the mixing of in-phase and quadrature components of the signal light and the local oscillator light, and their performance directly determines the sensitivity, bit error rate, and operating bandwidth of the coherent receiver link. As optical communication evolves towards 400G / 800G high-speed and high-density on-chip integration, increasingly stringent requirements are being placed on the size, bandwidth, and process compatibility of optical mixers.

[0003] Existing optical mixers are mainly based on multimode interference couplers, star couplers, or cascaded directional couplers. All of these suffer from common drawbacks such as large size (typically tens to hundreds of micrometers), limited operating bandwidth (difficult to simultaneously cover the entire C+L band), and sensitivity to process variations in phase accuracy. Recent reverse engineering solutions often employ reverse excitation or single-source forward optimization, indirectly constraining phase relationships through transmittance. This makes it difficult to directly and accurately control the phase combination of multiple ports under independent dual-source excitation, resulting in a tradeoff between convergence and phase accuracy, and further compromising device size control. Summary of the Invention

[0004] This application provides an optical mixer, a fabrication and control method, and a silicon photonics integrated circuit to at least partially solve the above-mentioned technical problems.

[0005] To achieve the above objectives, according to a first aspect of this application, an optical mixer is provided, comprising: A silicon device layer is formed with a modulation region and a first input port, a second input port, and multiple output ports respectively connected to the modulation region; the modulation region includes multiple discretely distributed dielectric pillars, which together with the remaining silicon region of the modulation region form the equivalent refractive index distribution of the modulation region, and the refractive index of the dielectric pillars is different from the refractive index of the remaining silicon region of the modulation region; The modulation region is configured such that when a first optical signal is input individually to the first input port, the plurality of output ports output multiple optical signals with a first phase relationship, and when a second optical signal is input individually to the second input port, the plurality of output ports output multiple optical signals with a second phase relationship.

[0006] Furthermore, the first optical signal is a local oscillator light, and the second optical signal is a signal light; the modulation region is specifically configured such that the plurality of output ports output multiple optical signals with orthogonal phase relationship when the local oscillator light is input alone; and that the plurality of output ports output multiple optical signals with consistent phase when the signal light is input alone.

[0007] Furthermore, the optical mixer includes four output ports, and the orthogonal phase relationship includes: the relative phases between the optical signals output from the other three output ports and the optical signal output from the reference output port are 180 degrees, -90 degrees, and +90 degrees, respectively.

[0008] Furthermore, the dielectric pillar is a silicon dioxide dielectric pillar, which penetrates the silicon device layer along the thickness direction, and the two ends of the dielectric pillar are used to contact the buried oxide layer and the cladding layer on both sides of the silicon device layer, respectively.

[0009] Furthermore, the modulation region has a planar dimension of 4.8 micrometers × 4.2 micrometers, and the modulation region is virtually divided into 40 × 35 pixel units on the plane, wherein each pixel unit includes at most one of the dielectric pillars; and / or The average angular distance across the multiple output ports is less than or equal to 10 degrees.

[0010] According to a second aspect of this application, a method for controlling the fabrication of an optical mixer, used to fabricate the optical mixer as described above, includes: The modulation area to be designed is virtually divided into M×N pixel units, and each pixel unit corresponds to two states: retaining silicon material or replacing it with a dielectric pillar; The output status of multiple output ports of multiple pixel units under different states is obtained, so as to determine the distribution of the dielectric pillars of the modulation region to be designed based on the output status.

[0011] Further, acquiring the output status of multiple output ports of multiple pixel units under different states, so as to determine the distribution of dielectric pillars of the modulation region to be designed based on the output status, includes: The equivalent refractive index distribution of the modulation region to be designed is evaluated based on the output. The pixel units of the modulation region to be designed are traversed, and the state of each individual pixel unit is changed in turn and the equivalent refractive index distribution is re-evaluated. When the equivalent refractive index distribution after the state change is better than the equivalent refractive index distribution before the change, the state of the pixel unit is retained; otherwise, the state of the pixel unit is reverted. Repeat the above steps until the termination condition is met to determine the distribution of the dielectric pillars in the modulation region to be designed.

[0012] Further, the output conditions include the complex values ​​of the electric field at multiple output ports when independent positive excitations are applied to the first input port and the second input port; the evaluation of the equivalent refractive index distribution of the modulation region to be designed based on the output conditions includes: The relative phase and transmittance of each output port are determined based on the complex value of the electric field; The equivalent refractive index distribution is evaluated by constructing an evaluation function based on the relative phase and the transmittance.

[0013] Furthermore, the formula for determining the evaluation function includes: ; Wherein, the superscript s represents the corresponding quantity under signal light excitation, and the superscript l represents the corresponding quantity under local oscillator light excitation; This represents the sum of the transmittance of multiple output ports; The standard deviation between ports represents the transmittance of multiple output ports; This represents the sum of phase errors between the relative phases of multiple output ports and the target phase; This represents the wavelength spread of the relative phase of multiple output ports; This represents the span of transmittance across wavelength at multiple output ports, where α, β, γ, δ, and ε are preset weighting coefficients.

[0014] According to a third aspect of this application, a silicon photonic integrated circuit includes an optical mixer as described above, or includes an optical mixer fabricated based on a fabrication control method for the optical mixer as described above.

[0015] The optical mixer of this application embodiment forms an equivalent refractive index distribution by combining the discrete dielectric pillars in the modulation region with the remaining silicon region. It utilizes the refractive index difference between the dielectric pillars and the silicon material to scatter and interfere with the transmitted light field, thereby guiding the optical signals from the two input ports to multiple output ports and forming preset phase relationships. It achieves phase distribution and equalization of multi-port output optical signals with a compact device size. Furthermore, the refractive index distribution of the modulation region can be flexibly configured through the spatial distribution morphology of the dielectric pillars to adapt to different mixing phase requirements, and has good design flexibility and process compatibility.

[0016] Other features and advantages of this application will be described in detail in the following detailed description section. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] To gain a more complete understanding of this application and its beneficial effects, the following description will be provided in conjunction with the accompanying drawings, wherein the same reference numerals in the following description denote the same parts.

[0019] Figure 1 This is a schematic diagram of the silicon device layer of the optical mixer provided in one embodiment of this application; Figure 2 This is a schematic diagram of the structure of an optical mixer provided in one embodiment of this application; Figure 3 This is a schematic diagram of the target relative phase of the four output ports when the first input port and the second input port are independently input with positive excitation in the embodiments of this application; Figure 4 This is a schematic flowchart of the fabrication and control method of an optical mixer provided in one embodiment of this application; Figure 5 This is a flowchart illustrating the fabrication and control method of an optical mixer provided in another embodiment of this application; Figure 6 This is a flowchart illustrating the fabrication and control method of an optical mixer provided in another embodiment of this application; Figure 7 This is a schematic diagram of the actual relative phase curves of the four output ports under the condition of a single local oscillator input in one embodiment of this application; Figure 8 This is a schematic diagram of the phase error curves of the four output ports under the condition of a single local oscillator input in one embodiment of this application; Figure 9 This is a schematic diagram of the relative phase curves of the four output ports under the condition of a single signal light input in one embodiment of this application; Figure 10 This is a flowchart illustrating the fabrication and control method of an optical mixer provided in another embodiment of this application.

[0020] Explanation of reference numerals in the attached figures: 100, silicon device layer; 101, modulation region; 111, dielectric pillar; 102, waveguide region; 103, first input port; 104, second input port; 105, output port; 115, first output port; 125, second output port; 135, third output port; 145, fourth output port; 200, cladding; 300, buried oxide layer; 400, substrate. Detailed Implementation

[0021] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the protection scope of this application.

[0022] In the description of this application, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more features. In the description of this application, "multiple" means two or more, unless otherwise explicitly specified.

[0023] Reference Figure 1 and Figure 2 As shown, according to a first aspect of this application, this application provides an optical mixer, comprising: a silicon device layer 100, having a modulation region 101 and a first input port 103, a second input port 104, and a plurality of output ports 105 respectively connected to the modulation region 101; the modulation region 101 includes a plurality of discretely distributed dielectric pillars 111, the dielectric pillars 111 and the remaining silicon region of the modulation region 101 together forming the equivalent refractive index distribution of the modulation region 101, the refractive index of the dielectric pillars 111 being different from the refractive index of the remaining silicon region of the modulation region 101; wherein, the modulation region 101 is configured such that when the plurality of output ports 105 are individually input with a first optical signal at the first input port 103, they output multiple optical signals having a first phase relationship, and when the plurality of output ports 105 are individually input with a second optical signal at the second input port 104, they output multiple optical signals having a second phase relationship.

[0024] Specifically, refer to Figure 2 As shown, the optical mixer in this embodiment is constructed based on an SOI (Silicon-On-Insulator) platform. The SOI platform, from bottom to top, includes a substrate 400, a buried oxide layer 300 on the substrate 400, and a silicon device layer 100 on the buried oxide layer 300; a cladding layer 200 is also covered on the silicon device layer 100.

[0025] In this embodiment, the thickness of the silicon device layer 100 can be 220nm, that is, a standard SOI wafer using a 220nm process node, which is compatible with the current mainstream silicon-based photonics process platform.

[0026] The modulation region 101 is formed within the silicon device layer 100. The modulation region 101 is used to perform equivalent refractive index modulation on the optical signal from the input port, thereby realizing phase control and distribution of the optical field. Specifically, the modulation region 101 has multiple discretely distributed dielectric pillars 111. Each dielectric pillar 111 has a discrete spatial distribution within the modulation region 101. The region of the dielectric pillars 111 and the remaining silicon region not occupied by the dielectric pillars 111 together constitute the equivalent refractive index distribution of the modulation region 101.

[0027] Since the refractive index of the dielectric pillar 111 differs from that of the remaining silicon region in the modulation region 101, the effective refractive index difference at different locations causes scattering and interference modulation of the light field transmitted therein, thereby guiding light signals of different phases to converge to the corresponding output ports 105, thus realizing the optical mixing function.

[0028] In this embodiment, the first input port 103 and the second input port 104 are respectively located on the input side of the modulation region 101, referring to... Figure 2 The first input port 103 and the second input port 104 are also the waveguide region 102 located on the left side of the silicon device layer 100. The first optical signal and the second optical signal are independently coupled into the modulation region 101 through the waveguide region 102. The first input port 103 and the second input port 104 are independent of each other and do not interfere with each other.

[0029] The plurality of output ports 105 are located on the output side of the modulation region 101, as shown in the reference. Figure 2 Multiple output ports 105, also known as waveguide regions 102 located on the right side of silicon device layer 100, lead out the optical signals modulated by modulation region 101 to downstream devices through waveguide regions 102.

[0030] The modulation region 101 is configured by the discrete distribution morphology of the dielectric pillars 111 such that when the first optical signal is input alone, the plurality of output ports 105 output multiple optical signals with a first phase relationship; and when the second optical signal is input alone, the plurality of output ports 105 output multiple optical signals with a second phase relationship. This configuration enables the optical mixer to meet the operating conditions of an optical mixer and directly interface with downstream device groups in a coherent receiving link.

[0031] The optical mixer in this embodiment forms an equivalent refractive index distribution by combining the discrete dielectric pillars 111 in the modulation region 101 with the remaining silicon region. It uses the refractive index difference between the dielectric pillars 111 and the silicon material to scatter and interfere with the transmitted light field, thereby guiding the optical signals from the two input ports (103, 104) to multiple output ports 105 and forming preset phase relationships. It achieves phase distribution and equalization of multi-port output optical signals with a compact device size. Moreover, the refractive index distribution of the modulation region 101 can be flexibly configured through the spatial distribution morphology of the dielectric pillars 111 to adapt to different mixing phase requirements, and has good design flexibility and process compatibility.

[0032] Reference Figures 1-3 As shown, in an optional embodiment of this application, the first optical signal is a local oscillator light, and the second optical signal is a signal light; the modulation region 101 is specifically configured to: cause the plurality of output ports 105 to output multiple optical signals with orthogonal phase relationship when the local oscillator light is input alone; and cause the plurality of output ports 105 to output multiple optical signals with consistent phase when the signal light is input alone.

[0033] Specifically, in this embodiment, the first input port 103 serves as the local oscillator end, used for coupling input local oscillator light; the second input port 104 serves as the signal end, used for coupling input signal light. In actual operation, the signal light and the local oscillator light are coupled into the modulation region 101 from the signal end and the local oscillator end, respectively. After being processed by the equivalent refractive index distribution within the modulation region 101, a light field distribution with a specific phase relationship is formed at the plurality of output ports 105.

[0034] Under the condition of single input of the local oscillator light, the refractive index distribution of the modulation region 101 causes the optical signals output by the multiple output ports 105 to have an orthogonal phase relationship, that is, the relative phase difference between adjacent output ports is orthogonal, so as to provide the I (In-phase) component and Q (Quadrature) component required for coherent reception, and meet the requirements of coherent optical communication system for I / Q demodulation.

[0035] Under the condition of a single signal light input, the refractive index distribution of the modulation region 101 ensures that the optical signals output from the multiple output ports 105 are in phase, thereby ensuring a balanced distribution of the signal light at each port and preventing a decrease in the demodulation sensitivity of the downstream balanced detector group due to phase deviation of the signal light at each port. The above phase configuration constitutes the basic operating conditions of a standard 90-degree optical mixer.

[0036] The optical mixer in this embodiment performs differentiated phase modulation on the local oscillator light and the signal light through the equivalent refractive index distribution of the modulation region 101. It outputs an orthogonal phase relationship for the local oscillator light to provide I / Q components and a consistent phase relationship for the signal light to ensure balanced distribution. This allows a single device to simultaneously meet the orthogonal mixing and signal equalization requirements of the coherent receiving link. There is no need to set up different optical path structures for the local oscillator light and the signal light, which simplifies the device architecture and reduces the complexity of on-chip integration.

[0037] Reference Figure 1 and Figure 3 As shown, in an optional embodiment of this application, the plurality of output ports 105 includes four output ports, namely a first output port 115, a second output port 125, a third output port 135, and a fourth output port 145; with the first output port 115 as the reference output port, the orthogonal phase relationship includes: the relative phases between the optical signals output by the other three of the four output ports and the optical signal output by the first output port 115 are 180 degrees, -90 degrees, and +90 degrees, respectively.

[0038] Specifically, refer to Figure 3 As shown, the four output ports are arranged from top to bottom along the output side of the modulation region 101, and are defined as the first output port 115, the second output port 125, the third output port 135 and the fourth output port 145.

[0039] During operation, the signal light and the local oscillator light are coupled into the modulation region 101 from their respective input ports. After being affected by the equivalent refractive index distribution within the modulation region 101, they form a light field distribution with a specific phase relationship at the four output ports.

[0040] Reference Figure 1 As shown, the waveguide width of the multiple output ports 105 can be 0.5 micrometers.

[0041] In this embodiment, using the first output port 115 as a reference port (denoted as 0 degrees phase), under the condition of a single local oscillator input: the relative phase of the optical signal output from the second output port 125 is 180 degrees, the relative phase of the optical signal output from the third output port 135 is -90 degrees, and the relative phase of the optical signal output from the fourth output port 145 is +90 degrees. The above four-port phase combinations [0°, 180°, ...] [90°, +90°] achieves the quadrature mixing conditions of a standard 90° optical mixer, with the four output ports corresponding to in-phase I+, in-phase I+, and in-phase I+, respectively. Orthogonal Q+, Orthogonal Q The four quadrant components can be directly connected to the downstream balanced photodetector array to achieve complete I / Q four-quadrant demodulation through differential detection.

[0042] Under the condition of individual signal light input, the relative phase of the four output ports with respect to the first output port 115 is all 0 degrees, meaning the signal light is evenly distributed to the four output channels. This configuration enables the optical mixer provided in this application to meet all the operating conditions of a standard 90-degree optical mixer.

[0043] It is understood that the selection of the reference output port is arbitrary. The above description uses the first output port 115 as a reference only for ease of description. In other embodiments, the second output port 125, the third output port 135, or the fourth output port 145 can also be used as references. As long as the relative phase difference between the four output ports satisfies the corresponding orthogonal relationship, it falls within the protection scope of this application.

[0044] The optical mixer in this embodiment can directly output four-quadrant I / Q components without the need for a separate phase separation or phase shifting structure at the output end, which simplifies the overall architecture of the coherent receiving link and reduces the complexity of subsequent signal processing. At the same time, the signal light is in phase at each port, ensuring balanced signal light power at the four ports, which is beneficial for common-mode noise suppression of the downstream balanced detector group.

[0045] Reference Figure 1 and Figure 3 As shown, in an optional embodiment of this application, the dielectric pillar 111 is a silicon dioxide dielectric pillar, which penetrates the silicon device layer 100 along the thickness direction. The two ends of the dielectric pillar 111 are used to contact the buried oxide layer 300 and the cladding layer 200 on both sides of the silicon device layer 100, respectively.

[0046] Specifically, refer to Figure 1 As shown in the embodiment of this application, the dielectric pillar 111 is made of silicon dioxide, and the other areas of the modulation region 101 are made of silicon material. The refractive index of silicon dioxide is less than that of silicon. The silicon dioxide dielectric pillar penetrates the entire silicon device layer 100 along the thickness direction of the silicon device layer 100 (i.e., the direction perpendicular to the SOI platform surface).

[0047] In one specific embodiment, the height of the silicon dioxide dielectric pillar is equal to the thickness of the silicon device layer 100, for example, both being 220 nm; in other embodiments, the thickness of the silicon device layer 100 may be in the range of 200 nm to 340 nm to accommodate wafer specifications of different standard SOI process nodes.

[0048] The lower end face of the silicon dioxide dielectric pillar contacts the buried oxide layer 300 located below the silicon device layer 100, and the upper end face of the silicon dioxide dielectric pillar contacts the cladding layer 200 located above the silicon device layer 100. Thus, in the thickness direction of the silicon device layer 100, the silicon dioxide dielectric pillar region constitutes a symmetrically clad vertical waveguide structure of silicon dioxide / silicon / silicon dioxide. This vertically symmetrical cladding structure ensures that the optical field in the region of the dielectric pillar 111 is uniformly constrained in the vertical direction, avoiding additional phase errors introduced by refractive index asymmetry in the vertical direction, and simultaneously preventing optical field leakage or higher-order mode excitation caused by incomplete penetration.

[0049] In some optional embodiments of this application, the cross-section of the silicon dioxide dielectric pillar is circular; optionally, the cross-sectional radius of the silicon dioxide dielectric pillar can be adjusted within the range of 30nm to 55nm to adapt to different process capabilities and design requirements. In a specific embodiment of this application, the cross-section of the silicon dioxide dielectric pillar is a circle with a radius of 45nm, corresponding to a minimum feature size of 90nm in diameter. Within the lithography and etching accuracy range of the standard SOI process platform, no additional process steps or high-precision lithography equipment are required. During the fabrication process, a circular via is etched at the position of the dielectric pillar 111 in the silicon device layer 100 and filled with silicon dioxide to form the silicon dioxide dielectric pillar; the remaining areas remaining as silicon material do not require processing. The entire patterning and etching process can be completed in one step on the standard SOI process platform, without the need for multi-layer stacking or metal reflective layers. The process flow is simple and highly compatible with mainstream silicon-based photonics process platforms.

[0050] The optical mixer in this embodiment uses a silicon dioxide dielectric pillar 111 that penetrates the silicon device layer 100 and is in contact with the buried oxide layer 300 and the cladding layer 200 at both ends as a refractive index modulation unit. This forms a symmetrically clad vertical waveguide structure of silicon dioxide / silicon / silicon dioxide in the vertical direction, ensuring the symmetrical constraint of the light field in the vertical direction and effectively avoiding additional phase errors caused by vertical asymmetry. At the same time, the minimum feature size (diameter 90nm) of the silicon dioxide dielectric pillar 111 is within the lithography and etching accuracy range of the standard SOI process. It can be patterned and filled in one step on the standard SOI process platform without the need for multi-layer stacking or metal reflective layers, and has good process compatibility and manufacturability.

[0051] In an optional embodiment of this application, the modulation region 101 has a size of 4.8 micrometers × 4.2 micrometers in the planar direction, and the modulation region 101 is virtually divided into 40 × 35 pixel units on the plane, wherein each pixel unit includes at most one of the dielectric pillars 111; and / or the average angular distance of the full band between the plurality of output ports 105 is less than or equal to 10 degrees.

[0052] Specifically, refer to Figure 2 As shown in this embodiment, the modulation region 101 has a planar dimension of 4.8 μm × 4.2 μm. To describe the distribution density and arrangement of the dielectric pillars 111 within the modulation region 101, the modulation region 101 can be virtually divided into a 40 × 35 pixel unit array on the plane. That is, it is divided into 40 pixel units along the direction perpendicular to the light propagation direction (the width direction of the modulation region) and 35 pixel units along the light propagation direction, for a total of 40 × 35 = 1400 pixel units. The side length of each pixel unit on the plane can be 120 nm, and each pixel unit includes at most one dielectric pillar 111, meaning that each pixel position can be reserved as silicon material or a dielectric pillar 111 can be set at the center.

[0053] It is understood that the above division of pixel units is only an auxiliary description at the logical level, used to explain the positional occupancy relationship and spatial arrangement of each dielectric pillar 111 in the modulation region 101, and is not a physical boundary existing in the actual fabricated device structure (for example, there are no visible pixel grid lines on the device silicon board), and therefore does not constitute a limitation on the scope of protection of this application.

[0054] In some alternative embodiments, the side length of the pixel unit can be adjusted in the range of 100nm to 150nm, and the number of pixel units can be adjusted accordingly to maintain the design area.

[0055] Based on the above pixelated layout, in a specific embodiment of this application, the number of dielectric pillars 111 is the number of pixel units in a low refractive index state determined by optimized design, while the remaining pixel units are retained as silicon material. The independent selection between two refractive index states for each pixel unit within the modulation region 101 allows the overall equivalent refractive index distribution of the modulation region 101 to be flexibly configured in discrete degrees of freedom space to meet preset multi-port phase relationships and bandwidth requirements.

[0056] In an optional embodiment of this application, the optical mixer operates in a band covering 1530nm to 1625nm, that is, simultaneously covering the full range of both the C-band (1530nm to 1565nm) and the L-band (1565nm to 1625nm), enabling a single device to perform mixing functions across the entire C+L band. FDTD (Finite-Difference Time-Domain) simulations have verified that, under local oscillator excitation alone, the average angular distance across the entire band between the actual relative phase of the four output ports 105 and the target phase (0 degrees, 180 degrees, -90 degrees, +90 degrees) in the 1530nm to 1625nm band is no greater than 10 degrees, with a typical value reaching 5 degrees (see [link to relevant documentation]). Figure 4 and Figure 5 Under signal light excitation alone, the relative phase of the four output ports 105 is close to 0 degrees (see...). Figure 6 The total insertion loss of the device under signal source excitation is approximately 2.25 dB, while the total insertion loss of the device under local oscillator excitation is approximately 2.35 dB.

[0057] Specifically, the angular distance is used to measure the degree of deviation between two phase values, for any two phases and The angular distance between the two is: .

[0058] The full-band average angular distance refers to the average wavelength of the angular distance between the actual relative phase and the target phase at each wavelength point within the operating band. This indicator comprehensively reflects the phase accuracy and phase flatness of the device within the operating band.

[0059] Compared to existing 90-degree optical mixers based on multimode interference (MMI) technology, whose imaging length is proportional to the square of the waveguide width and whose typical size ranges from tens to hundreds of micrometers; 90-degree optical mixers based on star couplers, whose Rowland circles and waveguide arrays also occupy a large area; and 90-degree optical mixers based on cascaded 2×2 directional couplers, which require multiple cascades leading to a significant increase in size, the modulation region 101 in this embodiment can be set much smaller than the aforementioned traditional schemes, exhibiting a significant advantage in compactness and facilitating high-density on-chip integration. Furthermore, the phase response of traditional devices based on interference principles naturally depends on the wavelength dispersion of the optical path difference, making it difficult for a single device to simultaneously cover the C-band and L-band, often requiring separate designs for C-band and L-band devices. The optical mixer in this embodiment achieves low phase error mixing across the entire C+L band in a single device through optimized design of the discrete refractive index modulation region. This eliminates the need to design or switch devices for different bands, reducing system integration complexity and maintenance costs.

[0060] In some optional embodiments of this application, the operating band of the 90-degree optical mixer can also be extended to other optical communication bands such as the O-band (Original Band), E-band (Extended Band), S-band (Short Band), and U-band (Ultra-long Band). When extending the band, the optimization design can be re-executed for the target band (e.g., the distribution optimization of the dielectric pillars 111 can be re-executed) to obtain the optimal distribution morphology of the dielectric pillars 111 under the corresponding band, while the basic architecture and principle of the device remain unchanged.

[0061] Reference Figure 4 As shown, a second aspect of this application provides a method for controlling the fabrication of an optical mixer, used to determine the distribution of dielectric pillars 111 in the modulation region 101 of the optical mixer, so as to achieve optimized design of the equivalent refractive index distribution of the modulation region 101. The fabrication control method includes the following steps: The modulation area to be designed (i.e.) Figure 2 The design area is divided into M×N pixel units, where M and N are both positive integers. Each pixel unit has two states: it is retained as silicon material or replaced with dielectric pillar 111.

[0062] The output status of multiple output ports 105 of multiple pixel units in different states is obtained, so as to determine the distribution of dielectric pillars 111 of the modulation region 101 to be designed based on the output status.

[0063] Specifically, the methods in the embodiments of this application can be implemented through simulation or actual preparation operations.

[0064] The modulation region 101 to be designed corresponds to the two-dimensional planar region to be optimized in the silicon device layer 100. The virtual partitioning discretizes the modulation region 101 into an M-row N-column pixel array in the planar direction. Each pixel unit has the same planar size, which transforms the optimization problem of the equivalent refractive index distribution in continuous space into a binary selection problem between two refractive index states for each pixel unit in discrete space, thereby restricting the design freedom to a finite searchable range.

[0065] In this embodiment, M can be 40, N can be 35, and the side length of each pixel unit can be 120nm, corresponding to a planar size of 4.8μm × 4.2μm for the modulation region 101. Optionally, the side length of the pixel unit can be adjusted within the range of 100nm to 150nm to adapt to process nodes with different feature sizes. During initialization, an all-silicon initial layout (i.e., all pixel units are in a state where they are retained as silicon material) can be selected, or the initial layout can be selected from the results of previous optimization, random distribution, or symmetric prior structure to accelerate convergence; the initialization method does not affect the physical realizability of the final optimization result.

[0066] The two states corresponding to each pixel unit are as follows: State 1 (first refractive index state), the pixel unit is retained as silicon material, that is, the position is a continuous silicon layer after device fabrication; State 0 (second refractive index state), the pixel unit is replaced at the center position by a dielectric pillar 111 that penetrates the silicon device layer 100 along the thickness direction.

[0067] In this embodiment, the dielectric pillar 111 is a silicon dioxide dielectric pillar with a circular cross-section and a radius of 45 nm. The dielectric materials in the two states are differentiated in refractive index, thereby forming a configurable binarized equivalent refractive index distribution within the modulation region 101.

[0068] The output information is obtained in the following way: For example, in a simulation environment, independent forward excitations are applied to the first input port 103 (signal end) and the second input port 104 (local oscillator end), that is, the signal source and the local oscillator source are enabled separately in sequence, and the optical field response at the four output ports 105 is obtained in two independent forward simulations. The output information can be directly obtained from the measured data of the electric field monitor at the output port, rather than indirectly constructed through reverse propagation or adjoint field inference, which makes the constraint on the phase relationship of the four ports more direct and accurate in the optimization process.

[0069] The method in this embodiment of the application transforms the optimization problem of the equivalent refractive index distribution in continuous space into a discrete binary combination optimization problem by virtually dividing the modulation region 101 into a discrete M×N pixel array. Each pixel unit is independently configured with either silicon material or dielectric pillar 111 refractive index states. This significantly reduces the design degree of freedom. At the same time, by directly obtaining the optical field response of the output port by applying a positive excitation to the input port to determine the distribution of dielectric pillar 111, the method avoids the indirect constraint of phase relationship in the optimization target of existing reverse design methods. This results in a direct correspondence between the optimization target and the actual performance index of the device, which is beneficial for obtaining high-precision optimization results under multi-port composite constraints.

[0070] Reference Figure 5 As shown, in an optional embodiment of this application, the process of acquiring the output status and determining the distribution of the modulation region 101 dielectric pillars 111 based on the output status specifically includes the following sub-steps.

[0071] The equivalent refractive index distribution of the modulation region 101 to be designed is evaluated based on the output.

[0072] The pixel units of the modulation region 101 to be designed are traversed, and the state of each individual pixel unit is changed in turn and the equivalent refractive index distribution is re-evaluated. Wherein, when the equivalent refractive index distribution after the state change is better than the equivalent refractive index distribution before the change, the changed state of the pixel unit is retained; otherwise, the pixel unit is reverted to the state before the change.

[0073] Repeat step 202 until the termination condition is met to determine the distribution of the dielectric pillars 111 of the modulation region 101 to be designed.

[0074] The system iterates through each of the M×N pixel units in a random order. For the currently visited pixel unit, its state 1 is flipped to state 0 (i.e., a dielectric pillar 111 is added at that position, changing the refractive index state), or its state 0 is flipped to state 1 (i.e., the dielectric pillar 111 at that position is removed, reverting the refractive index state), generating a candidate pixel state distribution. For this candidate state distribution, two forward simulations are performed again (one for signal source excitation and one for local oscillator excitation), and the equivalent refractive index distribution after the flip is calculated. If the equivalent refractive index distribution after the flip is better than before the flip, the flip is accepted, and the modulation region 101 is kept in the candidate state distribution; if the equivalent refractive index distribution after the flip is worse than or equal to before the flip, the state change is rejected, and the pixel unit is reverted to its original state. After each pixel state flip, the system directly calls full-wave simulation to complete the structural evaluation, without introducing any approximate fast evaluation model, nor relying on the adjoint field back-inference or gradient approximation based on the continuous dielectric constant.

[0075] It can be understood that a complete traversal of the M×N pixel units is defined as one iteration. When no accepted flip occurs in any pixel unit within one iteration, it means that no single-pixel flip can be found that can further improve the equivalent refractive index distribution under the current pixel dispersion and search scale, the current structure has reached a local optimum, and the optimization process terminates; or it terminates after reaching the preset maximum number of iterations, and outputs the current best pixel state distribution as the final medium pillar 111 distribution scheme.

[0076] Optionally, a maximum number of iterations (e.g., 500 or 1000) can be set as a termination condition to avoid the process not stopping due to non-convergence of the optimization.

[0077] After optimization is terminated, the final pixel state distribution output corresponds to the complete scheme of whether dielectric pillars 111 are set at each position in the modulation area 101, which is used to guide the fabrication of actual devices. Circular through holes are etched and filled with silicon dioxide at the pixel positions corresponding to dielectric pillars 111 in the silicon device layer 100, while the silicon material remains unchanged at the other pixel positions.

[0078] In an optional embodiment of this application, the output includes: the complex values ​​of the electric field at multiple output ports 105 when independent positive excitations are applied to the first input port 103 and the second input port 104. The equivalent refractive index distribution of the modulation region 101 to be designed is evaluated based on the output, specifically including: The relative phase and transmittance of each of the output ports 105 are determined based on the complex value of the electric field; The equivalent refractive index distribution is evaluated by constructing an evaluation function based on the relative phase and the transmittance.

[0079] Specifically, in the two rounds of forward simulation where the positive excitation is input at the first input port 103 and the second input port 104, the complex electric field value of the output port 105 is obtained in each round in the following manner. In the first round, a signal source (local oscillator light) is preset at the first input port 103 (local oscillator end), the second input port 104 (signal end) is turned off, and an FDTD simulation is run once; Ey (y-direction electric field) component monitors are set at the four output ports 105 on the output side of the modulation region 101, and the complex values ​​of each monitor after spatial averaging at each frequency point (corresponding to each wavelength in the working band) are recorded; with the first output port 115 as the reference port, the electric field phase of the other three output ports is converted into the relative phase with respect to the first output port 115; the power spectrum of the monitors at the four output ports 105 in the working band is read synchronously and the wavelength average is taken to obtain the transmittance value of the corresponding port. In the second round, the signal source (signal light) of the second input port 104 is enabled, and the local oscillator source (local oscillator light) of the first input port 103 is turned off. Another FDTD simulation is run in the same way to obtain the four-port relative phase and transmittance under the excitation of the signal source.

[0080] The core feature of the above method is the use of dual-source forward direct excitation: forward excitation sources are applied separately to the signal terminal and the local oscillator terminal, and complex values ​​are directly read from the electric field monitor at the output port 105. The relative phase and transmittance of each port are directly obtained from these complex values ​​through mathematical conversion, without the need to pre-arrange reverse excitation sources with target phase differences at the four output ports 105 and obtain the optimization target value through backpropagation, nor is it necessary to construct gradient information by back-calculation using the adjoint field. This allows the phase and transmittance data used in the optimization process to come directly from the measured response of the output port, which is completely consistent with the output performance of the device in actual operation, avoiding the disconnect between the objective function and the actual phase performance in the indirect phase constraint method.

[0081] Based on the two sets of relative phase and transmittance data obtained from the above two rounds of simulation, the relative phase of each port under signal source excitation is calculated (refer to...). Figure 9 , which is the phase error between the actual relative phase curves of the four output ports under the condition of single signal light input in one embodiment and their target phase [0°, 0°, 0°, 0°], and the relative phase of each port under local oscillator excitation (refer to...). Figure 7 (This is the actual relative phase curve of the four output ports under the condition of a single local oscillator input in one embodiment) and its target phase [0°, 180°, ... Phase error between 90° and +90° (refer to) Figure 8The phase error curves of the four output ports are calculated under the condition of a single local oscillator input. Simultaneously, the sum of the transmittance of each port and the standard deviation of the transmittance between ports (measuring the degree of signal distribution balance across ports) are calculated, as well as the wavelength spread of the above quantities (measuring broadband flatness). The above calculation results are input into the evaluation function FOM to obtain a comprehensive quantitative evaluation score of the equivalent refractive index distribution under the current structure.

[0082] This application embodiment directly reads the complex electric field values ​​at 105 points on the four output ports in each forward simulation and directly calculates the relative phase and transmittance of each port. The optimization target is established on the electric field data of the output ports, so that the performance reflected by the evaluation function is completely consistent with the output performance of the device under actual working conditions, eliminating the phase constraint deviation in the indirect construction method of the objective function. At the same time, the independent forward excitation method of dual sources activated in sequence avoids the complexity of field decomposition under multi-source superposition, so that each simulation only needs to process a single excitation source, simplifying the simulation settings and reducing the computational cost.

[0083] In an optional embodiment of this application, the formula for determining the evaluation function FOM includes: ; Wherein, the superscript 's' represents the corresponding quantity under signal light excitation, and the superscript 'l' represents the corresponding quantity under local oscillator light excitation; T total This represents the sum of the transmittance of multiple output ports 105; T uniform φ represents the standard deviation of transmittance among multiple output ports 105, used to measure the balance of output power across ports; err This represents the sum of phase errors between the relative phases of multiple output ports 105 and their respective target phases, i.e., the sum of the angular distances between the actual relative phases of each port and the target phases; φ flat T represents the wavelength-wise breadth of the relative phase of multiple output ports 105, used to measure the flatness of the phase change with wavelength; flat This represents the transmittance breadth across wavelength of multiple output ports 105, used to measure the flatness of transmittance variation with wavelength; α, β, γ, δ, and ε are preset weighting coefficients.

[0084] Specifically, the evaluation function adopts a five-term weighted combination, simultaneously constraining the comprehensive performance under both excitation sources. The first term is the transmittance gain term, encouraging the FOM (Forward Oscillator Metric) to increase with the sum of transmittances, meaning the optimization direction tends to maximize the total output power of each port, thereby reducing the device's insertion loss. The second term is the transmittance equalization penalty term; if the transmittance distribution of each port is uneven (large standard deviation between ports), this term takes a larger value, reducing the FOM and thus pushing the optimization direction towards a more balanced distribution of port power. The third term is the phase error penalty term; if the relative phase of each port deviates from the target value (target phase under signal light excitation [0°, 0°, 0°, 0°] and target phase under local oscillator excitation [0°, 180°, ...), the error is penalized accordingly. [90°, +90°]), increasing this value decreases the FOM, thereby driving the structure to evolve towards phase accuracy. The fourth and fifth terms are the phase broadband flatness penalty term and the transmittance broadband flatness penalty term, respectively, used to suppress drastic fluctuations in phase and transmittance in the wavelength dimension, ensuring consistent device performance at all wavelengths within the operating band.

[0085] In the embodiments of this application, the weighting coefficient α can be 0.75, β can be 1, γ can be 3, δ can be 2, and ε can be 1. The above values ​​can be adjusted under specific design objectives. For example, if the requirement for phase accuracy is higher, the value of γ can be appropriately increased; if the requirement for broadband flatness is higher, the values ​​of δ and ε can be appropriately increased.

[0086] The target phase under signal light excitation is [0°, 0°, 0°, 0°], meaning the relative phase of all four output ports 105 with respect to the reference output port is 0 degrees. The target phase under local oscillator light excitation is [0°, 180°, ...]. [90°, +90°], meaning the relative phases of the four output ports 105 with respect to the reference output port are 0 degrees, 180 degrees, -90 degrees, and +90 degrees, respectively. Phase error is calculated based on the principle of minimizing angular distance for any two phases. and angular distance between the two for: ; For example, the angular distance between the measured phase of 355 degrees and the target phase of 0 degrees is 5 degrees, not 355 degrees. The phase error at each port is the sum or average of the angular distances between the measured relative phase and the target phase at that port. Phase bandwidth expansion φ flat This is used to measure the deviation between the relative phase of each output port at each wavelength point within the operating band and the average relative phase of that port within the operating band. It can be characterized using statistical measures such as standard deviation or maximum deviation; similarly, T... flatIt is used to measure the degree of deviation between the transmittance of each output port at each wavelength point in the operating band and the average transmittance of the port in the operating band.

[0087] This application embodiment constructs an evaluation function covering five indicators: transmittance gain, transmittance equalization, phase error, phase broadband flatness, and transmittance broadband flatness. This unifies and quantifies the multi-dimensional performance constraints of the multi-port mixer under two independent excitation sources into a single objective function, achieving a weighted balance between different performance indicators. This eliminates the need to handle multiple potentially contradictory optimization objectives separately during the optimization process, thus improving optimization efficiency. At the same time, the phase error is calculated using the principle of shortest angular distance, avoiding misjudgments caused by phase periodicity.

[0088] Reference Figure 10 As shown, in a specific embodiment of this application, the fabrication and control method of the optical mixer includes: The pixel distribution of the modulation region 101 is initialized. During the initialization phase, the modulation region 101 is virtually divided into M×N pixel units on the plane, and an initial state is assigned to each pixel unit. The initial state can be an all-silicon initial layout, or an initial layout derived from the results of previous optimization, random distribution, or symmetric prior structure.

[0089] After initialization, the forward simulation phase begins. First, the signal source at the first input port 103 is enabled and the local oscillator at the second input port 104 is disabled. An electromagnetic simulation (e.g., FDTD simulation) is run. Electric field component monitors are set at the four output ports 105 on the output side of the modulation region 101. The complex electric field values ​​at each frequency point of each output port 105 are read, thereby obtaining the transmittance and relative phase of the four output ports 105 under the condition of the signal light being excited alone. Subsequently, the signal source is disabled and the local oscillator is enabled. Another electromagnetic simulation is run in the same manner to obtain the transmittance and relative phase of the four output ports 105 under the condition of the local oscillator light being excited alone.

[0090] After acquiring the output data from two rounds of simulation, the evaluation function calculation stage begins. Based on the relative phase and transmittance of the four output ports 105 under signal source excitation and the relative phase and transmittance of the four output ports 105 under local oscillator excitation, the equivalent refractive index distribution of the modulation region 101 under the current pixel state distribution is comprehensively and quantitatively evaluated according to the preset evaluation function FOM. The evaluation function consists of a weighted gain term of the sum of transmittance, minus a port transmittance uniformity penalty term, a relative phase error penalty term, a phase wavelength flatness penalty term, and a transmittance wavelength flatness penalty term, which comprehensively reflect the overall design quality of the modulation region 101 in multiple dimensions such as transmission efficiency, port equalization, phase accuracy, and broadband flatness with a single scalar value.

[0091] After evaluating the current structure, the pixel-by-pixel flip search phase begins. Each pixel in the M×N pixel units is visited sequentially in a random order. For the currently visited pixel unit, its binary state is flipped—either from a state where it is retained as silicon material to a state where it is set with dielectric pillar 111, or from a state where it is set with dielectric pillar 111 to a state where it is retained as silicon material. After the state flip, the structure is reconstructed, and the dual-source forward simulation phase and the evaluation function calculation phase are re-executed to obtain a new FOM value after the flip. The new FOM value is compared with the current FOM value before the flip: if the new FOM value is higher than the current FOM value, it indicates that this flip has improved the overall performance of the equivalent refractive index distribution of the modulation region 101, so the flip is accepted, the changed state of the pixel unit is retained, and the current FOM value is updated to the new FOM value; if the new FOM value is not higher than the current FOM value, it indicates that this flip has not brought about a performance improvement, so the flip is rejected, the pixel unit is reverted to its original state before the flip, and the current FOM value remains unchanged.

[0092] After evaluating the flip of a pixel unit, check if there are any unvisited pixel units in the current round. If there are unvisited pixel units in the current round, return to the previous round and continue visiting the next pixel unit in random order and perform the above flip evaluation process. When all M×N pixel units in the current round have been visited, count whether there is at least one accepted flip in the current iteration. If there is at least one accepted flip in the current round, it indicates that the equivalent refractive index distribution of the modulation region 101 still has room for further optimization at the current search scale, so return to the dual-source forward simulation stage and start a new round of pixel-by-pixel flip search. If no accepted flip occurs in any pixel unit in the current round, it indicates that under the current pixel dispersion and search scale, it is impossible to further improve the FOM value through any single pixel flip, the current structure has reached a local optimum, the termination condition is met, the entire optimization process is terminated, and the final pixel state distribution is output as the design scheme for the distribution of dielectric pillars 111 in the modulation region 101.

[0093] The final design scheme output is directly used to guide the fabrication of actual devices: through holes are etched at the pixel positions corresponding to the dielectric pillars 111 in the silicon device layer 100 and filled with a low refractive index dielectric material (e.g., silicon dioxide), while the silicon material remains unchanged at the other pixel positions, thereby physically realizing the optimized pixel state distribution as the modulation region 101 structure of the optical mixer.

[0094] A third aspect of this application provides a silicon photonic integrated circuit, including the optical mixer as described above, or including an optical mixer prepared based on the optical mixer preparation and control method described above.

[0095] The multiple output ports 105 of the optical mixer are used to connect to the downstream balanced detector and TIA (Trans-Impedance Amplifier) ​​to form a complete coherent optical receiving link.

[0096] Specifically, in this embodiment, the optical mixer, as a core passive mixing unit, is integrated into the silicon photonics integrated circuit. The silicon photonics integrated circuit can be built on the same SOI platform. The silicon device layer 100 of the optical mixer shares the same device layer with other active and passive devices of the silicon photonics integrated circuit (e.g., modulators, detectors, wavelength division multiplexing / demultiplexers, etc.), achieving monolithic integration. The four output ports 105 of the optical mixer, namely the first output port 115, the second output port 125, the third output port 135, and the fourth output port 145, are respectively connected to two sets of downstream balanced photodetectors through waveguide regions 102. Specifically, the first output port 115 and the second output port 125 have a relative phase difference of 180 degrees, corresponding to in-phase I+ and I+. The components are connected to the first set of balanced detectors; the third output port 135 and the fourth output port 145 (with a relative phase difference of 180 degrees, corresponding to orthogonal Q+ and Q) are connected to the first set of balanced detectors; The first component (I-Q component) is connected to the second set of balanced detectors. Each set of balanced detectors performs differential photocurrent detection on the two input optical signals with a 180-degree phase difference to suppress common-mode noise, obtaining the I-component and Q-component electrical signals. The output current signal of the balanced detector is further input to the TIA, where the TIA converts the weak photocurrent signal into a voltage signal that can be processed by the subsequent digital signal processing unit, thus completing a complete coherent receiving link from optical signal input to I / Q electrical signal output.

[0097] In some embodiments of this application, the silicon photonic integrated circuit can be a coherent optical receiver front-end chip in a coherent optical communication system, an optical receiving unit in a coherent laser radar system, or a quantum state measurement front-end in a quantum communication system. The compact size of the optical mixer (modulation region 101 is 4.8 μm × 4.2 μm) allows the silicon photonic integrated circuit to integrate more functional units within a smaller chip area, which is beneficial to improving the functional density and integration level of the chip.

[0098] This technical solution integrates an optical mixer into a silicon photonics integrated circuit and forms a coherent optical receiving link with a balanced detector and a TIA. This enables the optical mixer to form a complete signal processing closed loop, from receiving optical signals and achieving I / Q phase separation to outputting electrical signals that can be used for digital signal processing. The chip only needs to provide optical signal input and electrical signal output interfaces, resulting in a simple external interface and high internal integration. At the same time, the compact size of the optical mixer significantly reduces the area occupied by the entire coherent receiving link on the chip, which is beneficial for integrating multiple coherent receiving channels on a single optical chip, thereby improving the channel density and system bandwidth density of the silicon photonics integrated circuit.

[0099] In some optional embodiments of this application, the optical mixer and its preparation and control method described above also have the following alternative implementations, all of which fall within the protection scope of this application.

[0100] In an alternative embodiment, the pixel layout of the modulation region 101 can be extended to a universal grid of M×N (where M and N are both positive integers). While maintaining a pixel unit side length of 120nm, the planar dimensions of the modulation region 101 are 120nm×M×120nm×N. Optionally, the side length of the pixel unit can be adjusted within the range of 100nm to 150nm, and the values ​​of M and N can be adjusted accordingly to maintain the overall planar area of ​​the modulation region 101 or to adapt to process nodes with different feature sizes.

[0101] In an alternative embodiment, the cross-sectional shape of the dielectric pillar 111 is not limited to a circle, but may also be a square, rectangular, elliptical, or other polygonal cross-section, as long as it can be reliably patterned and filled with a low-refractive-index dielectric in the fabrication process. The material of the dielectric pillar 111 is not limited to silicon dioxide, but may also be other low-refractive-index dielectric materials compatible with SOI processes, such as silicon nitride (SiN) or polymer dielectrics.

[0102] In an alternative implementation, the number of output ports of the optical mixer is four in a standard 90-degree mixing application scenario; in other multi-port mixing application scenarios, the number of output ports can also be extended to more than four (e.g., six, eight or more). Different numbers of output ports and target phase relationships can be adapted by adjusting the equivalent refractive index distribution of the modulation region 101 to optimize the target.

[0103] In an alternative embodiment, the thickness of the silicon device layer 100 is not limited to 220 nm, and can be in the range of 200 nm to 340 nm in a standard SOI process platform to accommodate SOI wafer specifications at different process nodes. With different silicon device layer thicknesses, the height of the dielectric pillar 111 remains consistent with the thickness of the silicon device layer 100, ensuring that the upper and lower end faces of the dielectric pillar 111 contact the buried oxide layer 300 and the cladding layer 200 respectively, maintaining symmetrical coverage in the vertical direction.

[0104] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0105] The embodiments, implementation methods, and related technical features of this application can be combined and substituted for each other without conflict.

[0106] The above are merely preferred embodiments of this application and are not intended to limit this application in any way. Any simple modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of this application without departing from the scope of the technical solution of this application shall still fall within the scope of the technical solution of this application.

Claims

1. An optical mixer, characterized in that, include: A silicon device layer is formed with a modulation region and a first input port, a second input port, and multiple output ports respectively connected to the modulation region; The modulation region includes multiple discretely distributed dielectric pillars, which together with the remaining silicon region of the modulation region form the equivalent refractive index distribution of the modulation region. The refractive index of the dielectric pillars is different from the refractive index of the remaining silicon region of the modulation region. The modulation region is configured such that when a first optical signal is input individually to the first input port, the plurality of output ports output multiple optical signals with a first phase relationship, and when a second optical signal is input individually to the second input port, the plurality of output ports output multiple optical signals with a second phase relationship.

2. The optical mixer according to claim 1, characterized in that, The first optical signal is a local oscillator light, and the second optical signal is a signal light; the modulation region is specifically configured such that the plurality of output ports output multiple optical signals with orthogonal phase relationship when the local oscillator light is input alone; and that the plurality of output ports output multiple optical signals with consistent phase when the signal light is input alone.

3. The optical mixer according to claim 2, characterized in that, It includes four output ports, and the orthogonal phase relationship includes: the relative phases between the optical signals output by the other three output ports and the optical signal output by the reference output port are 180 degrees, -90 degrees and +90 degrees, respectively.

4. The optical mixer according to any one of claims 1-3, characterized in that, The dielectric pillar is a silicon dioxide dielectric pillar that penetrates the silicon device layer along its thickness direction. The two ends of the dielectric pillar are used to contact the buried oxide layer and the cladding layer on both sides of the silicon device layer, respectively.

5. The optical mixer according to any one of claims 1-3, characterized in that, The modulation region has a planar dimension of 4.8 micrometers × 4.2 micrometers, and is virtually divided into 40 × 35 pixel units on the plane, wherein each pixel unit includes at most one of the dielectric pillars; and / or The average angular distance across the multiple output ports is less than or equal to 10 degrees.

6. A method for controlling the fabrication of an optical mixer, used to fabricate the optical mixer as described in any one of claims 1-5, characterized in that, include: The modulation area to be designed is virtually divided into M×N pixel units, and each pixel unit corresponds to two states: retaining silicon material or replacing it with a dielectric pillar; The output status of multiple output ports of multiple pixel units under different states is obtained, so as to determine the distribution of the dielectric pillars of the modulation region to be designed based on the output status.

7. The method according to claim 6, characterized in that, The step of acquiring the output status of multiple output ports of multiple pixel units under different states, and determining the distribution of dielectric pillars of the modulation region to be designed based on the output status, includes: The equivalent refractive index distribution of the modulation region to be designed is evaluated based on the output. The pixel units of the modulation region to be designed are traversed, and the state of each individual pixel unit is changed in turn and the equivalent refractive index distribution is re-evaluated. When the equivalent refractive index distribution after the state change is better than the equivalent refractive index distribution before the change, the state of the pixel unit is retained; otherwise, the state of the pixel unit is reverted. Repeat the above steps until the termination condition is met to determine the distribution of the dielectric pillars in the modulation region to be designed.

8. The method according to claim 7, characterized in that, The output conditions include the complex electric field values ​​of multiple output ports when independent positive excitations are applied to the first input port and the second input port; the evaluation of the equivalent refractive index distribution of the modulation region to be designed based on the output conditions includes: The relative phase and transmittance of each output port are determined based on the complex value of the electric field; The equivalent refractive index distribution is evaluated by constructing an evaluation function based on the relative phase and the transmittance.

9. The method according to claim 8, characterized in that, The formula for determining the evaluation function includes: ； Wherein, the superscript s represents the corresponding quantity under signal light excitation, and the superscript l represents the corresponding quantity under local oscillator light excitation; This represents the sum of the transmittance of multiple output ports; The standard deviation between ports represents the transmittance of multiple output ports; This represents the sum of phase errors between the relative phases of multiple output ports and the target phase; This represents the wavelength spread of the relative phase of multiple output ports; This represents the span of transmittance across wavelength at multiple output ports, where α, β, γ, δ, and ε are preset weighting coefficients.

10. A silicon photonics integrated circuit, characterized in that, It includes the optical mixer as described in any one of claims 1 to 5, or the optical mixer prepared based on the preparation control method of the optical mixer as described in any one of claims 6 to 9.

Images