On-chip multimode end-face coupler based on silicon nitride structure

Abstract

An on-chip multimode end-face coupler based on a silicon nitride structure includes: a first silicon oxide cladding layer, a first tapered silicon nitride waveguide, a second silicon oxide cladding layer, a second tapered silicon nitride waveguide, and a third silicon oxide cladding layer, sequentially disposed on a tapered silicon waveguide from bottom to top. The tapered silicon waveguide is used for mode input and coupling within the waveguide. The three silicon oxide cladding layers have different thicknesses. The wider port of the tapered silicon waveguide receives the mode input, and as the waveguide gradually narrows, the mode is gradually coupled into the silicon nitride waveguide. The tapered silicon nitride waveguide and the silicon oxide cladding layers together amplify the mode beam. This invention applies a multilayer silicon nitride structure to an end-face waveguide structure, achieving efficient coupling of TE0, TE1, TM0, and TM1 modes.

Description

Technical Field

[0001] This invention relates to a technology in the field of silicon-based integrated photonic chips, specifically an on-chip multimode end-face coupler based on a silicon nitride structure. Background Technology

[0002] In existing on-chip optical interconnect systems, mode multiplexing technology can simultaneously utilize multiple mutually orthogonal modes as independent channels to transmit information, thereby significantly improving system capacity and spectral efficiency. However, due to the large differences in mode field volume between different modes, it is difficult to couple multiple modes from the chip to the optical fiber simultaneously. To achieve high-efficiency, low-loss coupling between multimode fiber and chip, multimode on-chip couplers have become an important research direction in current optical communication systems. Summary of the Invention

[0003] This invention addresses the shortcomings of existing multimode coupling technologies, such as large device size, low coupling efficiency, excessive loss, inability to achieve high-order mode multiplexing and coupling, and high requirements for manufacturing processes. It proposes an on-chip multimode end-face coupler based on a silicon nitride structure, which applies a multilayer silicon nitride structure to the end-face waveguide structure to achieve efficient coupling of TE0, TE1, TM0, and TM1 modes.

[0004] This invention is achieved through the following technical solution:

[0005] This invention relates to an on-chip multimode end-face coupler based on a silicon nitride structure, comprising: a first silicon oxide cladding layer, a first tapered silicon nitride waveguide, a second silicon oxide cladding layer, a second tapered silicon nitride waveguide, and a third silicon oxide cladding layer, sequentially disposed on a tapered silicon waveguide from bottom to top, wherein: the tapered silicon waveguide is used to realize mode input and coupling in the waveguide; the thicknesses of the three silicon oxide cladding layers are different; the wider port of the tapered silicon waveguide inputs the mode in the waveguide; as the waveguide gradually narrows, the mode in the waveguide is gradually coupled into the silicon nitride waveguide; the tapered silicon nitride waveguide and the silicon oxide cladding layer together realize mode amplification.

[0006] The widths of the tapered silicon waveguide and the tapered silicon nitride waveguide are both tapered and gradually narrow, so that the mode achieves refractive index matching in the two waveguides during transmission and gradually enters the upper silicon nitride waveguide for transmission.

[0007] The aforementioned mode amplification refers to the following: when a mode enters the silicon oxide capping layer above the tapered silicon waveguide, the refractive index of silicon nitride is higher than that of silicon oxide, creating a refractive index difference between the two materials. Therefore, the light in the silicon oxide gradually focuses onto the silicon nitride layer and propagates within the silicon nitride waveguide. Since the refractive index difference between silicon nitride and silicon oxide is much smaller than that between the silicon waveguide and silicon oxide, the mode field propagating in the silicon nitride waveguide is more divergent, and the mode size is much larger than the mode size propagating within the waveguide itself, thus achieving mode amplification.

[0008] The silicon waveguide has a width and a thickness of 220 nm, and the waveguide width is reduced from 1.2 μm to 0.01 μm.

[0009] The coupling length, width, and thickness of the tapered silicon nitride waveguide are as follows: thickness 50 nm, and waveguide width increased from 0.01 μm to 16 μm.

[0010] The first layer of the silicon oxide capping layer has a thickness of 2 μm, and the second and third layers of the silicon oxide capping layer have thicknesses of 1.5 μm each.

[0011] The aforementioned end-face coupler uses the central axis of symmetry of the tapered silicon waveguide as the axis of symmetry of the entire device.

[0012] An air groove with a width of 5μm and a height of 5μm is provided on each side of the 8μm axis of symmetry to reduce the lateral divergence of the mode during transmission, thereby reducing transmission loss and improving coupling efficiency.

[0013] The width and thickness of the tapered silicon waveguide and the tapered silicon nitride waveguide, along with the silicon oxide cladding layer, are optimized by calculating the electric field distribution function in the waveguide using the finite difference eigenmode expansion (FDE) method. Specifically, the parameters of the waveguide width and thickness are optimized as follows:

[0014] The electric field of a light wave (electromagnetic wave) in a uniform waveguide can be known from Maxwell's equations: For the optical field (electromagnetic field) in a waveguide, suppose that the electromagnetic field varies with a certain angular frequency in a time-harmonic manner, and its electric field is expressed as an electric field that varies with time: E ~ e [-iωt) Assuming the electromagnetic wave propagates along the z-axis, the electric field along that direction can be expressed as: E ~ e (-iβz) Substituting into Maxwell's electric field divergence equation, we get: The subscript t denotes the vector within the waveguide cross section. The electric field E of the waveguide cross section is then calculated. t From (x, y), the electric field distribution E in the direction of propagation can be calculated. z (x, y).

[0015] Using the half-vector approximation, Et Let the dominant component in (x, y) be E x Substituting (x, y) into Maxwell's equations, we can obtain the equation regarding E. x The eigenvalue equations of (x, y): Therefore, the process of solving the electric field within the waveguide cross section is to solve the above eigenmode equations using the finite difference eigenmode expansion method.

[0016] For a rectangular waveguide with length *a* and width *b*, the continuous waveguide plane is first divided into finite discrete units using a fixed dimensional accuracy (~50 nm). The electric field *E* at (i, j) of each discrete unit is then calculated. x (i, j) satisfy the following relationship: Substituting the two equations into the above eigenvalue equation and setting the boundary conditions to 0, the eigenvalues ​​E of the electric field can be obtained. x (i, j) are obtained, and the corresponding propagation constant β is also obtained. By taking the working wavelength as 1550 nm, the electric field distribution of the target mode can be calculated.

[0017] The coupling lengths of the tapered silicon waveguide and the tapered silicon nitride waveguide to the silicon oxide cladding layer are calculated using the eigenmode expansion (EME) method, specifically:

[0018] a) Using the finite difference eigenmode expansion method described above, the frequency domain eigenvalues ​​of Maxwell's equations can be solved in the direction of electromagnetic wave propagation, which are the electric field functions at each discrete element.

[0019] b) Based on the electric field function mentioned above, and taking advantage of the reversibility of the optical path, each electric field function can propagate bidirectionally, thereby constructing a transmission matrix that shows the electromagnetic wave changing with the propagation distance. Attached Figure Description

[0020] Figure 1 This is a schematic diagram of the device structure of the present invention;

[0021] In the figure: 1) is the front view of the device, 2) is the side view of the device, and 3) is the top view of the device;

[0022] Figure 2 This is a simulation field distribution diagram of the TE0, TE1, TM0, and TM1 modes of the present invention;

[0023] Figure 3 This is a diagram showing the coupling efficiency of the TE0, TE1, TM0, and TM1 modes of the present invention.

[0024] In the figure: 1. Substrate 2. Silicon waveguide 3. Tapered silicon nitride waveguide 4. Silicon oxide capping layer 5. Detailed Implementation

[0025] like Figure 1 As shown, this embodiment relates to an on-chip multimode end-face coupler based on a silicon nitride structure, comprising: a tapered silicon waveguide 3, two tapered silicon nitride waveguides 4 of different lengths for mode conversion of TE0, TE1, TM0, and TM1 modes, and three silicon oxide capping layers 5 of different thicknesses, wherein: the silicon waveguide 3 is located above the substrate 1 and the substrate 2; the first silicon oxide capping layer 5, the first tapered silicon nitride waveguide 4, the second silicon oxide capping layer 5, the second tapered silicon nitride waveguide 4, and the third silicon oxide capping layer 5 are located above the silicon waveguide 3 from bottom to top.

[0026] The three silicon oxide capping layers 5 of different thicknesses are as follows: the thickness of the first silicon oxide capping layer 5 is 2 μm, the thickness of the second silicon oxide capping layer 5 is 1.5 μm, and the thickness of the third silicon oxide capping layer 5 is 1.5 μm.

[0027] The coupling length, width, and thickness of the silicon waveguide were calculated using the finite difference eigenmode expansion method to obtain the optimal solutions, which are: a thickness of 220 nm and a waveguide width reduced from 1.2 μm to 0.01 μm.

[0028] The coupling length, width, and thickness of the tapered silicon nitride waveguide were calculated using the finite difference eigenmode expansion method to obtain the optimal solutions, which are: a thickness of 50 nm and a waveguide width that increases from 0.01 μm to 16 μm.

[0029] The tapered silicon nitride waveguide 4 achieves mode conversion of the target mode by coupling and amplifying the modes in the silicon waveguide.

[0030] like Figure 2 The figure shows the electric field distributions of the TE0, TE1, TM0, and TM1 modes calculated using the finite difference eigenmode expansion method.

[0031] In this embodiment, the effective refractive indices of the TE0 mode, TE1 mode, TM0 mode, and TM1 mode at the end face of the silicon nitride waveguide are calculated to be 1.444, 1.455, 1.444, and 1.455, respectively, by means of the finite difference intrinsic mode expansion method.

[0032] The silicon waveguide 3 is realized by electron beam exposure and inductively coupled plasma etching on a silicon substrate on an insulating sheet, forming a 220nm silicon layer.

[0033] The tapered silicon nitride waveguide 4 achieves a 50nm thick silicon nitride film deposition through plasma-enhanced chemical vapor deposition, followed by electron beam exposure and inductively coupled plasma etching.

[0034] The silicon oxide capping layer 5 is achieved by depositing three SiO2 thin films with thicknesses of 2μm, 1.5μm, and 1.5μm on the chip using plasma-enhanced chemical vapor deposition.

[0035] The aforementioned device fabrication process is compatible with mature complementary metal-oxide-semiconductor (CMOS) fabrication technology, enabling large-scale industrial production.

[0036] Calculations using the intrinsic mode expansion method show that, with a waveguide length of 4000 μm, the device can simultaneously achieve high-efficiency, low-loss coupling of TE0, TE1, TM0, and TM1 modes. The coupling losses are 0.07 dB, 8 dB, 0.28 dB, and 4.45 dB, respectively. Figure 3 As shown.

[0037] Compared to existing technologies, this invention achieves efficient coupling of multiple modes. Except for the TE1 mode, the coupling efficiency of the TE0, TM0, and TM1 modes is superior to that of other reported on-chip multimode couplers. The fabrication methods involved rely on mature complementary metal-oxide-semiconductor (CMOS) processing technology, enabling large-scale production. By introducing a double-layer silicon nitride waveguide, an optimizable dimension is added to the device design process. Future work can further expand the number of coupling modes through the design of the silicon nitride waveguide, thus the device structure has good scalability.

[0038] The above-described specific implementations can be partially adjusted by those skilled in the art in different ways without departing from the principles and purpose of the present invention. The scope of protection of the present invention is defined by the claims and is not limited to the above-described specific implementations. All implementation schemes within the scope of the claims are bound by the present invention.

Claims

1. An on-chip multimode end-face coupler based on silicon nitride structure, characterized in that, include: A first silicon oxide cladding layer, a first tapered silicon nitride waveguide, a second silicon oxide cladding layer, a second tapered silicon nitride waveguide, and a third silicon oxide cladding layer are sequentially disposed on a tapered silicon waveguide from bottom to top. The tapered silicon waveguide is used for mode input and coupling within the waveguide. The three silicon oxide cladding layers have different thicknesses. The wider port of the tapered silicon waveguide receives the mode input. As the waveguide gradually narrows, the mode is gradually coupled into the silicon nitride waveguide. The two tapered silicon nitride waveguides and the three silicon oxide cladding layers together amplify the mode's pattern. The widths of both the tapered silicon waveguide and the two tapered silicon nitride waveguides gradually taper from wide to narrow, ensuring that the mode achieves refractive index matching in both waveguides during transmission, gradually transferring from the silicon waveguide into the upper silicon nitride waveguide. The coupling lengths, widths, and thicknesses of the tapered silicon waveguide, the two tapered silicon nitride waveguides, and the three silicon oxide capping layers are calculated using the finite difference eigenmode expansion method to obtain the optimal solution, thereby achieving efficient mode coupling.

2. The on-chip multi-mode end-face coupler based on silicon nitride structure according to claim 1, characterized in that, At the end face of the tapered silicon nitride waveguide, the mode pattern amplified by two layers of silicon nitride waveguide is matched with the mode pattern of the multimode fiber to achieve high-efficiency end face coupling.

3. The on-chip multi-mode end-face coupler based on silicon nitride structure of claim 1, characterized in that, The width and thickness of the tapered silicon waveguide, the two-layer tapered silicon nitride waveguide, and the three-layer silicon oxide capping layer are calculated using the finite difference eigenmode expansion method to optimize the waveguide width and thickness parameters. Specifically: i) The electric field of a light wave in a uniform waveguide can be known from Maxwell's equations: For the optical field in the waveguide, it is a time-harmonic varying electromagnetic field E ~ e (-iωt) When the propagation direction of the electromagnetic wave is along the z-axis, the electric field E ~ e in the propagation direction is... (-iβz) Substituting this into Maxwell's electric field divergence equation, we get: Where: the subscript t denotes the vector within the waveguide cross section, and the electric field of the waveguide cross section is calculated. Find the electric field distribution in the direction of propagation. ; ii) using the semi-vector approximation, the dominant component in is set as , and the eigen equation about is obtained by substituting it into Maxwell equation: The process of solving the electric field in the waveguide cross section is to solve the above eigen equation by using finite difference eigen mode expansion method; iii) for a rectangular waveguide with length a and width b in a cross section, first divide the continuous waveguide plane into finite discrete units with a fixed size precision, and then calculate the electric field at each discrete unit (i, j) satisfy the following relationship: , Substitute the above two equations into the eigen equation, take the boundary condition as 0, and solve the eigenvalue of the electric field and get the corresponding propagation constant β, when the working wavelength is 1550 nm, the electric field distribution of the target mode is obtained.

4. The on-chip multimode end-face coupler based on silicon nitride structure according to claim 1, characterized in that, The coupling lengths of the tapered silicon waveguide, the two tapered silicon nitride waveguides, and the three silicon oxide capping layer were calculated using the finite difference eigenmode expansion method, specifically: a) Using the finite difference eigenmode expansion method described above, the frequency domain eigenvalues ​​of Maxwell's equations can be solved in the direction of electromagnetic wave propagation, which is the electric field function at each discrete unit. b) Based on the electric field function mentioned above, and taking advantage of the reversibility of the optical path, each electric field function can propagate bidirectionally, thereby constructing a transmission matrix that shows the electromagnetic wave changing with the propagation distance.

Images