Multi-mode interference coupler based on thin-film lithium niobate

Abstract

The utility model discloses a multimode interference coupler based on thin film lithium niobate. The multimode interference coupler comprises a silicon-based substrate, a protective layer and an optical waveguide. The optical waveguide is composed of an input waveguide, a multimode interference area and two identical output waveguides. The multi-mode interference area is composed of a multi-mode interference area A and a multi-mode interference area B; the output waveguide is composed of an output gradual change area C, an output gradual change area D and an output straight waveguide. The input gradual change waveguide, the multimode interference area A, the multimode interference area B, the output gradual change area C and the output gradual change area D are all of isosceles trapezoid structures. According to the utility model, the additional loss is 0.009 dB under the wavelength of 1.55 [mu] m. Under the additional loss of 1dB, the bandwidth of the device is greater than 0.8 [mu] m. Compared with other MMI structures, the MMI structure provided by the utility model has the characteristics of large bandwidth, low loss and high process tolerance.

Description

Technical Field

[0001] The utility model relates to the technical field of optical communications, in particular to a multimode interference coupler based on thin-film lithium niobate. Background Art

[0002] Multimode interference (MMI) couplers have the advantages of excellent device performance, simpler manufacturing process, and more compact device structure. A multimode interference coupler with high bandwidth, low loss, and high process tolerance occupies a dominant position in photonic circuits such as electro-optic modulators, optical switches, and optical multiplexers. Utility Model Content

[0003] The purpose of this utility model is to provide a high process tolerance, low loss MMI structure design based on thin-film lithium niobate. While ensuring a compact structure and good integration, it solves the problems of high insertion loss, small working bandwidth and low process tolerance of the splitter. It can be applied to related fields such as optical transceiver modules and optical quantum computing.

[0004] The multimode interference coupler provided by the utility model comprises a silicon-based substrate, a protective layer and an optical waveguide. The optical waveguide is made of a thin-film lithium niobate material with a thickness of 0.3 μm. The protective layer is located on the silicon-based substrate and the optical waveguide, completely wrapping the optical waveguide.

[0005] The optical waveguide exhibits a tapered structure that gradually expands in the transmission direction. The first section is where the optical waveguide transmission transitions to a larger waveguide width; the second section is the multimode interference region, which is split into two sections; and the third section converts the waveguide with a wider multimode interference region into two single-mode transmission waveguides.

[0006] The optical waveguide consists of an input waveguide, a multimode interference region, and two identical output waveguides. The input waveguide consists of an input straight waveguide and an input gradient waveguide. The multimode interference region consists of a multimode interference region A and a multimode interference region B. The output waveguide consists of an output gradient region C, an output gradient region D, and an output straight waveguide. The input gradient waveguide, multimode interference region A, multimode interference region B, output gradient region C, and output gradient region D all have isosceles trapezoidal structures.

[0007] The optical waveguide is made of thin-film lithium niobate material with a thickness of 0.3um. The input straight waveguide is connected to one end of the input gradient waveguide, and the width of the connection is W1 = 1.5um; the other end of the input gradient waveguide is connected to one end of the multimode interference area A, and the width of the connection is W2 = 2.5um; the other end of the multimode interference area A is connected to one end of the multimode interference area B, and the width of the connection is W3 = 5um; the other end of the multimode interference area B is connected to one end of the two output gradient areas C, and the width of the connection is W4 = 6um; the other ends of the two output gradient areas C are connected to the two output gradient areas D respectively, and the connection is The width at the junction is W5 = 2.5um; the other ends of the two output gradient areas D are respectively connected to the two output straight waveguides, and the width of the connection is W6 = 1.5um; the length of the input straight waveguide is L1 = 10um, the length of the input gradient waveguide is L2 = 10um, the length of the multimode interference area A is L3 = 5um, the length of the multimode interference area B is L4 = 22um, the length of the output gradient area C is L5 = 0.5um, the length of the output gradient area D is L6 = 5um, and the length of the output straight waveguide is L7 = 10um; the error of the above dimensions is ±10%.

[0008] The technical features and effects of this utility model are as follows:

[0009] The input and output waveguides of this utility model are both constructed from tapered waveguides, and the multimode interference region utilizes a tapered structure (or isosceles trapezoidal structure), which gradually increases in size. The imaging principle of MMI is a process in which a single mode excites multiple modes, and the imaging position also continuously widens. The use of a tapered structure increases the effective width of the multimode interference region, thereby reducing the influence of wavelength on the imaging position, increasing the bandwidth of the device, and simultaneously improving process tolerance and bandwidth while reducing excess loss. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] Figure 1 It is a schematic diagram of the overall structure of the utility model;

[0011] Figure 2 This is a top view of the optical waveguide portion of the utility model;

[0012] Figure 3 Comparison of process tolerances for a conventional structure, a rectangular multi-mode interference region, and the structure of the present invention;

[0013] Figure 4 The bandwidth comparison is for a conventional structure, a multi-mode interference region with a rectangle, and a structure of the present invention.

[0014] Figure 1 Middle: 1. Input straight waveguide; 2. Input gradient waveguide; 3. Multimode interference region A; 4. Multimode interference region B; 5. Output gradient region C; 6. Output gradient region D; 7. Output straight waveguide. DETAILED DESCRIPTION

[0015] In this embodiment, the waveguide is made of thin-film lithium niobate, with an anisotropic refractive index of {2.21, 2.13, 2.21}. The protective layer surrounding the waveguide is made of silicon dioxide, with a refractive index of 1.44. The substrate is made of silicon, with a refractive index of 3.9.

[0016] In this embodiment, the thickness of the silicon substrate is 300 μm, the height of the silicon dioxide protection layer is 4.70 μm, and the height of the lithium niobate film is 0.6 μm.

[0017] The straight waveguide is composed of a ridge waveguide. In order to meet the conditions of single-mode transmission, the ridge width W1 of the ridge waveguide is designed to be 1.5μm, the ridge height is 0.3μm, and according to the process requirements, the etching angle is 70°, and the straight waveguide length L1 is 10μm.

[0018] The input and output waveguides are both constructed from tapered waveguides. Each tapered waveguide is continuous, with optimized width and length to achieve a low-loss MMI result. The wide waveguide input effectively reduces the number of input modes, thereby minimizing phase loss.

[0019] Due to existing process accuracy issues and etching precision errors, even small dimensional changes in MMI can have a significant impact on imaging, making process tolerance analysis essential. Compared to the Y-beam splitter, the multimode interferometer has improved process tolerance, and a gradient MMI has been proposed to improve process tolerance.

[0020] The following is a simulation of the entire device model using the lumenical eigenmode expansion method (EME). The model parameters are optimized. The specific steps for modeling are as follows:

[0021] Modeling: A three-dimensional parametric model of the MMI was constructed. The width of the multimode interference was preset to 6μm. The interference length for a one-to-two splitting beam was theoretically calculated to be 22μm. The height of one multimode interference beam was 0.3μm, and the height of the lithium niobate at the bottom was also 0.3μm. Thin-film lithium niobate was coated with silicon dioxide, with the material priority of the lithium niobate set to 2 and the silicon dioxide set to 3.

[0022] Material settings: The material used for thin film lithium niobate is anisotropic dispersion data. Since there is no lithium niobate data in Lumerical, it needs to be imported. For silicon dioxide, you can directly use the material provided in the software.

[0023] Boundary condition setting: Set the four sides of the model as PML boundary conditions.

[0024] Set the excitation port: There are three ports to set, and each port is set to the mode field. For each port mode analysis, set it to TE mode transmission.

[0025] Solution setup: In EME, the entire MMI structure is divided into six sections, each with a different segmentation unit and excitation mode. The first section is the input straight waveguide, divided into 5 cells; the second section is the input gradient waveguide, divided into 10 cells; the third section is multimode interference region 1, divided into 1 cell, with a mode setting of 50; the fourth section is multimode interference region 2, with a mode setting of 50 and a single cell; the fifth section is the output gradient region, divided into 10 cells; and the sixth section is the output straight waveguide region, divided into 10 cells.

[0026] Mesh refinement: To increase simulation accuracy and keep the simulation speed low, the simulation mesh accuracy needs to be properly set. Introduce mesh analysis to refine the mesh in the simulation. The refinement accuracy is set to one-tenth of the height of the ridge waveguide, and the refinement area is selected as the input and output parts.

[0027] An additional loss for an MMI device is calculated as the sum of each output divided by the input:

[0028] EL(dB)=-10lg(∑P OUT / P IN );

[0029] According to the formula, the law of imaging single shot length:

[0030] Table 1 shows the output positions of single and multiple images under different input stimuli.

[0031] This design uses symmetrical interference, aiming for a bisection image output. Substituting the multimode interference width into the table equation determines the interference length. In other words, each multimode interference width corresponds to an optimal interference length. In the next section, for multimode interference width optimization, both the width and length are varied simultaneously to determine the optimal interference length.

[0032] In order to satisfy the symmetrical interference condition, the input waveguide is selected at the center of the device, while the output position is symmetrical to the center.

[0033] Table 1

[0034]

[0035] In order to better reflect the advantages of the structure of this utility model, the bandwidth and process tolerance of the designed structure will be analyzed. The process tolerance is mainly reflected in the error when etching the waveguide. During the simulation, the error caused by the length of the entire device is mainly selected. When the deviation is 10%, the performance change of the device is reflected as follows Figure 3 As shown in the figure, the structure of the present invention has a significant improvement in resistance to process tolerance. As for bandwidth analysis, the wavelength is analyzed from 1.2μm to 2μm. Under the condition of 1dB loss, the bandwidth of structure II and structure III is greater than 0.8μm, which is significantly better than structure I. Figure 4 The structure of the utility model has excellent advantages in bandwidth, additional loss and process tolerance. At a light wavelength of 1.55, the additional loss of the utility model is 0.009dB.

Claims

1. A multimode interference coupler comprising a silicon-based substrate, a protective layer, and an optical waveguide, characterized in that: The optical waveguide is composed of an input waveguide, a multimode interference region and two identical output waveguides; the input waveguide is composed of an input straight waveguide and an input gradient waveguide; the multimode interference region is composed of a multimode interference region A and a multimode interference region B; the output waveguide is composed of an output gradient region C, an output gradient region D and an output straight waveguide; the input straight waveguide is butted against one end of the input gradient waveguide, and the width of the butt joint is W1=1.5um; the other end of the input gradient waveguide is butted against one end of the multimode interference region A, and the width of the butt joint is W2=2.5um; the other end of the multimode interference region A is butted against one end of the multimode interference region B, and the width of the butt joint is W3=5um; the other end of the multimode interference region B is butted against one end of the two output gradient regions C, and the width of the butt joint is W4=6um; the two output gradient regions The other end of domain C is respectively connected to the two output gradient areas D, and the width of the connection is W5 = 2.5um; the other ends of the two output gradient areas D are respectively connected to the two output straight waveguides, and the width of the connection is W6 = 1.5um; the input straight waveguide length L1 = 10um, the input gradient waveguide length L2 = 10um, the multimode interference area A length L3 = 5um, the multimode interference area B length L4 = 22um, the output gradient area C length L5 = 0.5um, the output gradient area D length L6 = 5um, and the output straight waveguide length L7 = 10um; the input gradient waveguide, multimode interference area A, multimode interference area B, output gradient area C and output gradient area D are all isosceles trapezoidal structures, and the optical waveguide is made of thin film lithium niobate material with a film thickness of 0.3um.

2. The multimode interference coupler according to claim 1, wherein: The dimensional error of W1, W2, W3, W4, W5, W6, L1, L2, L3, L4, L5, L6, and L7 is ±10%.

3. The multimode interference coupler according to claim 1 or 2, wherein: The protective layer is located on the silicon-based substrate and the optical waveguide, and completely wraps the optical waveguide.

Images