Thin film electro-optical crystal-silicon nitride heterogeneous integrated N*N high-speed optical switch array and preparation method thereof

Abstract

The invention relates to a thin film electro-optical crystal-silicon nitride heterogeneous integrated N * N high-speed optical switch array and a preparation method thereof. The thin film electro-optical crystal-silicon nitride heterogeneous integrated N * N high-speed optical switch array is formed by connecting 1 * 2 or 2 * 2 Mach-Zehnder high-speed optical switch units according to a certain topological structure. Wherein the high-speed optical switch unit is composed of a silicon nitride passive waveguide and a thin-film electro-optical crystal / silicon nitride heterogeneous integrated phase shift region. Comprising but not limited to thin-film lithium niobate, thin-film lithium tantalate, barium titanate and lead zirconate titanate is integrated above the silicon nitride phase shifting arm through a micro transfer printing method to form a heterogeneous integrated phase shifting region. The large-scale high-speed optical switch array with excellent performance is realized by combining the characteristics of ultra-low loss, CMOS compatibility, thermal stability and the like of a silicon nitride material, the characteristics of low power consumption, low loss, high-speed modulation and the like of a thin film electro-optical crystal and the advantages of high flexibility, parallel integration, low cost, low loss and the like of a micro transfer printing method.

Description

Technical Field

[0001] This invention relates to the field of optical communication device technology, specifically to a thin-film electro-optic crystal-silicon nitride heterogeneous integrated N×N high-speed optical switch array and its fabrication method. Background Technology

[0002] With the rapid development of technologies such as artificial intelligence, machine learning, and big data analytics, modern computing systems are increasingly demanding high-throughput, low-latency data communication. To cope with the increasing system scale and complexity, traditional electrical interconnects face fundamental limitations in bandwidth, speed, and energy efficiency, becoming a major obstacle to further improving computing performance. Optical switching, due to its inherent advantages in bandwidth, latency, and power consumption, is widely considered a promising solution to these problems. High-speed optical switch chips, supporting nanometer-scale data switching, play a crucial role in applications such as optical packet switching, optical burst switching, and high-speed optical interconnects for cluster computing. Photonic integrated circuits (PICs) provide a scalable and compact platform for realizing high-performance, high-speed optical switches, enabling low-loss, high-speed, and low-power signal routing, suitable for advanced optical switching networks.

[0003] Currently, several mainstream PIC platforms have been used in the research and development of high-speed optical switches. The silicon nitride (SiN) platform possesses ultra-low transmission loss, a wide transparency window, full compatibility with Complementary Metal-Oxide-Semiconductor Transistor (CMOS) processes, and excellent thermal stability, enabling the realization of ultra-large-scale, wide optical bandwidth, low-cost, and thermally insensitive optical switch chips. A research team from Shanghai Jiao Tong University reported an 8×8 micro-ring optical switch chip based on a silicon-silicon nitride-silicon nitride three-dimensional integrated platform [X.Li,W.Gao,L.Lu,J.Chen,L.Zhou,Photon.Res.2023,11,712.], which is composed of cascaded dual micro-ring units in a crossbar structure. However, this platform mainly relies on thermo-optical effects for optical modulation, resulting in slow switching speeds (typically on the order of tens to hundreds of microseconds) and high power consumption, making it more suitable for manufacturing passive optoelectronic devices. Silicon-on-Insulator (SOI) platforms are widely used due to their mature CMOS processes and compact size. A research team from Shanghai Jiao Tong University reported a 16×16 silicon-based electro-optic switch chip [L.Lu,S.Zhao,L.Zhou,D.Li,Z.Li,M.Wang,X.Li,J.Chen,Opt.Express 2016,24(9),9295.], which is constructed by cascading seven 2×2 MZI stages using a Benes structure, achieving a switching time of approximately 3 nanoseconds. Although it can achieve high-speed response through carrier dispersion, the absorption of free carriers leads to significant insertion loss and crosstalk, making it difficult to realize large-scale high-speed optical switch arrays. Thin-film lithium niobate (TFLN) electro-optic crystal platforms utilize the Pockels effect to achieve low-loss, high-speed, and static-power-free modulation over an extremely wide transparency range. A research team from Zhejiang University reported a polarization-independent lithium niobate electro-optic switch chip [X. Liu, N. Chen, T. Chu, ACS Photon. 2024, 11(7), 2556.], which utilizes the high-speed modulation characteristics of lithium niobate to achieve picosecond switching times. However, the manufacturing process of TFLN waveguides is demanding, with few large-scale fabrication plants, resulting in high manufacturing costs. Furthermore, the strong birefringence of the material can degrade the performance of passive devices.

[0004] Integrating thin-film electro-optic crystals, such as TFLNs, onto commercially fabricated silicon nitride platforms can combine the advantages of both, achieving high-speed optical switches with low loss, low power consumption, and low cost. In terms of technology routes for integrating multiple material devices onto the same chip platform, they can generally be divided into two main categories: hybrid integration and heterogeneous integration. Hybrid integration focuses on packaging, that is, opto-packaging a pre-fabricated chip and a target chip onto the same substrate. However, due to its relatively large integration spatial scale, alignment accuracy is limited. Furthermore, this method is serial integration, resulting in lower throughput and higher manufacturing costs, thus limiting its application in large-scale production scenarios. Heterogeneous integration focuses on manufacturing, bonding different materials or semi-finished devices onto the target chip; a common technique is wafer bonding. Wafer bonding involves growing on a native substrate and bonding it to the target wafer, making it the method closest to monolithic integration. A team at Ghent University in Belgium has achieved a silicon-on-a-chip light source in the 850nm band by bonding gallium arsenide wafers to silicon chips [E. Haglund, S. Kumari, P. Westbergh, J. Gustavsson, G. Roelkens, R. Baets, A. Larsson, Opt. Express 2015, 23(26), 33634.]. Wafer bonding has high alignment accuracy and parallel batch integration capabilities, making it very suitable for large-scale production. However, its process is relatively complex, has higher requirements for manufacturing conditions, and makes it difficult to integrate devices from different material platforms onto the same platform multiple times after bonding one material.

[0005] In summary, there is an urgent need to design and develop high-speed optical switching devices and integration methods for heterogeneous integrated thin-film electro-optic crystals on the silicon nitride platform. Summary of the Invention

[0006] To overcome the shortcomings of the prior art, this invention proposes a thin-film electro-optic crystal / silicon nitride heterogeneous integrated N×N high-speed optical switch array and its fabrication method. Leveraging the high flexibility of the micro-transfer method, it can support the implementation of any silicon / silicon nitride photonics platform, including a 400nm thick silicon nitride platform, and also supports the implementation of thin-film electro-optic crystal platforms, including but not limited to thin-film lithium niobate, thin-film lithium tantalate, barium titanate, and lead zirconate titanate. This effectively reduces the power consumption, loss, and response time of the electro-optic switch, supporting large-scale, low-cost, and ultra-high-speed optical switching.

[0007] The technical solution of the present invention is as follows:

[0008] On one hand, this invention proposes a thin-film electro-optic crystal / silicon nitride heterogeneous integrated N×N high-speed optical switch array, characterized in that it is formed by connecting several 1×2 or 2×2 thin-film electro-optic crystal / silicon nitride heterogeneous integrated high-speed optical switch units according to a specific topology. The N×N optical switch array includes N optical input ports and N optical output ports, with a routing link formed between any one optical input port and any one optical output port. This routing link passes through several 1×2 or 2×2 heterogeneous integrated high-speed optical switch units and several silicon nitride waveguide cross-junctions.

[0009] Furthermore, the 1×2 or 2×2 thin-film electro-optic crystal / silicon nitride heterostructure integrated high-speed optical switch unit has a topology of Mach-Zehnder Interferometer (MZI).

[0010] Furthermore, the 1×2 or 2×2 thin-film electro-optic crystal / silicon nitride hetero-integrated MZI high-speed optical switch unit is composed of a silicon nitride waveguide, a silicon nitride 3dB coupler, a thin-film electro-optic crystal / silicon nitride hetero-integrated evanescent wave coupler, and a thin-film electro-optic crystal / silicon nitride hetero-integrated phase-shifting region.

[0011] Furthermore, the waveguide structure of the thin-film electro-optic crystal / silicon nitride heterogeneous integrated phase-shifting region is a thin-film electro-optic crystal / silicon nitride composite waveguide, with metal electrodes on both sides of the composite waveguide.

[0012] Furthermore, the thin-film electro-optic crystal / silicon nitride composite waveguide has a silicon nitride waveguide below and a thin-film electro-optic crystal planar waveguide above.

[0013] Furthermore, the thin-film electro-optic crystal / silicon nitride heterogeneous integrated phase-shifting region is designed with a thin-film electro-optic crystal / silicon nitride heterogeneous integrated evanescent wave coupler composed of a tapered waveguide at its input and output ends, which is used to realize low-loss coupling between the silicon nitride waveguide and the thin-film electro-optic crystal / silicon nitride composite waveguide.

[0014] Furthermore, the optical switch array topology includes, but is not limited to, PILOSS, Butterfly, Benes and improved Benes, DLN or improved DLN, S&S, Crossbar, etc.

[0015] Furthermore, any given routing link is formed by connecting several 2×2 heterogeneous integrated high-speed optical switch units and silicon nitride waveguide cross junctions in series via waveguides.

[0016] Furthermore, the 1×2 or 2×2 thin-film electro-optic crystal / silicon nitride heterogeneous integrated MZI high-speed optical switch unit is characterized by comprising an input waveguide I1, an input waveguide I2 (the 1×2 optical switch unit only has I1), an output waveguide O1, an output waveguide O2, a first silicon nitride 3dB coupler, an evanescent wave coupler, a heterogeneous integrated phase shifting region, and a second silicon nitride 3dB coupler. When light is input from the input waveguide I1 (or input waveguide I2), it is uniformly split into two beams by the first silicon nitride 3dB coupler and then coupled to the thin-film electro-optic crystal / silicon nitride composite waveguide through the thin-film electro-optic crystal / silicon nitride heterogeneous integrated evanescent wave coupler. After entering the thin-film electro-optic crystal / silicon nitride heterogeneous integrated electro-optic phase shifter, it is coupled back to the silicon nitride waveguide through the thin-film electro-optic crystal / silicon nitride heterogeneous integrated evanescent wave coupler, and then combined by the second silicon nitride 3dB coupler. By applying voltage to the two phase shifters, a phase difference of 0 or π is generated, allowing the switching unit to switch to different operating states, namely cross or bar states.

[0017] Furthermore, the thin-film electro-optic crystal / silicon nitride heterogeneous integrated high-speed optical switch unit is characterized in that the structure of the 3dB coupler includes, but is not limited to, an adiabatic coupler, a multimode interferometer, or a cascaded directional coupler.

[0018] Furthermore, the phase-shifting function of the aforementioned thin-film electro-optic crystal / silicon nitride heterogeneous integrated high-speed optical switch unit is achieved through the Pockels effect of the thin-film electro-optic crystal.

[0019] Furthermore, the heterogeneous integrated evanescent wave coupler is characterized in that its structure includes, but is not limited to, a linearly broadened tapered waveguide, a nonlinearly broadened tapered waveguide, and a piecewise linearly broadened tapered waveguide.

[0020] On the other hand, the thin-film electro-optic crystal-silicon nitride heterogeneous integrated N×N high-speed optical switch array proposed in this invention is characterized in that the thickness of silicon nitride includes, but is not limited to, 300nm, 400nm, and 800nm, and the thin-film electro-optic crystal platform includes, but is not limited to, thin-film lithium niobate, thin-film lithium tantalate, barium titanate, and lead zirconate titanate.

[0021] This invention also provides a method for fabricating a thin-film electro-optic crystal-silicon nitride heterostructure integrated N×N high-speed optical switch array, comprising the following steps:

[0022] S1. A heterogeneous integrated evanescent wave coupler and a tether structure are fabricated on a thin-film electro-optic crystal wafer on an insulator by photolithography and dry etching. The tether structure is used to support the electro-optic crystal phase shifter.

[0023] S2. Wet etching of the lower cladding layer of the electro-optic crystal;

[0024] S3. Fabrication of silicon nitride chips;

[0025] S4. Clean silicon nitride chips using a microwave plasma descaling machine;

[0026] S5. Use micro-transfer technology to transfer the electro-optic crystal phase shifter to the silicon nitride waveguide;

[0027] S6. Deposit metal electrodes on both sides of the phase shifter waveguide;

[0028] S7. Deposit a silicon oxide cladding layer;

[0029] S8. Electrode windowing.

[0030] Furthermore, step S3 includes:

[0031] S31. Silicon nitride passive waveguides and devices are fabricated on silicon-silicon dioxide-silicon nitride wafers by photolithography and dry etching;

[0032] S32. Deposit a silica overlay;

[0033] S33. Use CMP to polish the wafer to achieve wafer surface planarization;

[0034] Furthermore, step S5 includes:

[0035] S51. Use a PDMS stamp to quickly pick up the electro-optical phase shifter and transfer the electro-optical phase shifter over the silicon nitride chip via PDMS.

[0036] S52. Adjust the silicon nitride chip and align the electro-optic crystal phase shifter to the target position;

[0037] S53. Move the PDMS stamp to bring the electro-optic crystal phase shifter into contact with the silicon nitride chip and apply a certain pressure to press it, so that the electro-optic crystal phase shifter and the surface of the silicon nitride chip produce a sufficiently strong van der Waals contact.

[0038] S54. Apply a specific shear force in the horizontal direction to the PDMS stamp, and slowly lift the PDMS stamp in the vertical direction;

[0039] S55. Separate the PDMS stamp and the electro-optical phase shifter, and complete the transfer of the electro-optical phase shifter to the silicon nitride chip.

[0040] Furthermore, the PDMS stamp is made of polydimethylsiloxane.

[0041] Furthermore, the PDMS stamp consists of a flat PDMS and PDMS pillars on the flat panel, with the pillars in contact with the device to be transferred.

[0042] Furthermore, by increasing the number of pillars in the PDMS stamp, batch transfer integration of electro-optic crystal phase shifter arrays can be achieved.

[0043] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0044] This invention proposes a heterogeneous integrated N×N high-speed optical switch array and fabrication method using thin-film electro-optic crystals and silicon nitride, which fully combines the respective technological advantages of silicon nitride photonics platforms and thin-film electro-optic crystal platforms. The silicon nitride platform features low cost, ultra-low transmission loss, and excellent thermal stability, enabling long-distance, low-loss signal transmission; while the thin-film electro-optic crystal possesses high-speed, low-power electro-optic modulation characteristics, achieving rapid switching times from nanoseconds to picoseconds. Through heterogeneous integration of these two technologies, this invention enables low-loss, high-speed switching of arbitrary routing links in an N×N optical switch array on a single chip, meeting the stringent requirements of high throughput and low latency for large-scale optical interconnect networks. Furthermore, the micro-transfer heterogeneous integration method employed in this invention offers high flexibility. Compared to traditional wafer bonding or hybrid integration processes, micro-transfer not only achieves low-coupling-loss, low-cost, and high-yield heterogeneous integration but also supports parallel transfer over large areas or multiple chips, thereby increasing production throughput, reducing manufacturing costs, and enhancing process scalability, thus enabling highly integrated, modular, and scalable optical switch array designs. This invention provides strong technical support for the manufacturing of large-scale, high-speed, low-power optical switch arrays, and has significant engineering application value and industrialization prospects. Attached Figure Description

[0045] Figure 1 This is a schematic diagram of the structure of the thin-film electro-optic crystal-silicon nitride hetero-integrated N×N optical switch high-speed array of the present invention;

[0046] Figure 2 This is a schematic diagram of the structure of the 1×2 heterogeneous integrated MZI optical switch unit in the thin-film electro-optic crystal-silicon nitride heterogeneous integrated N×N high-speed optical switch array of the present invention;

[0047] Figure 3 This is a schematic diagram of the structure of the 2×2 heterogeneous integrated MZI optical switch unit in the thin-film electro-optic crystal-silicon nitride heterogeneous integrated N×N high-speed optical switch array of the present invention;

[0048] Figure 4 This is a schematic diagram of the waveguide cross-section in the thin-film electro-optic crystal-silicon nitride hetero-integrated N×N high-speed optical switch array of the present invention, wherein (a) is Figure 2 The waveguide cross-section at point 2.1, (b) is Figure 2 The waveguide cross-section at 2.6, (c) is Figure 2 Waveguide cross-section at 2.7;

[0049] Figure 5This is a schematic diagram of the structure of the thin-film electro-optic crystal / silicon nitride hetero-integrated N×N high-speed optical switch array of the present invention;

[0050] Figure 6 This is the simulation result of the thin-film electro-optic crystal / silicon nitride hetero-integrated evanescent wave coupler in the N×N high-speed optical switch array of the present invention. The thin-film electro-optic crystal is taken as thin-film lithium niobate as an example. The thin-film electro-optic crystal can also include, but is not limited to, thin-film lithium tantalate, barium titanate, and lead zirconate titanate. Among them, (a) is the transmission spectrum of the thin-film electro-optic crystal / silicon nitride hetero-integrated evanescent wave coupler, (b) is the alignment tolerance of the thin-film electro-optic crystal / silicon nitride hetero-integrated evanescent wave coupler perpendicular to the optical transmission direction, (c) is the alignment tolerance of the thin-film electro-optic crystal / silicon nitride hetero-integrated evanescent wave coupler parallel to the optical transmission direction, and (d) is the alignment tolerance of the thin-film electro-optic crystal / silicon nitride hetero-integrated evanescent wave coupler in the rotation direction.

[0051] Figure 7 This invention relates to a method for fabricating a thin-film electro-optic crystal-silicon nitride heterostructure integrated N×N high-speed optical switch array. Detailed Implementation

[0052] The present invention will be further described below with reference to the accompanying drawings and embodiments, but these should not be construed as limiting the scope of protection of the present invention. The accompanying drawings are for illustrative purposes only, representing schematic diagrams rather than actual physical objects, and should not be construed as limiting the scope of this patent. To better illustrate the embodiments of the present invention, some components in the drawings may be omitted, enlarged, or reduced, and do not represent the actual dimensions of the product. It is understandable to those skilled in the art that some well-known structures and their descriptions may be omitted in the drawings.

[0053] Please see Figure 1 , Figure 1 This is a schematic diagram of the structure of the thin-film electro-optic crystal-silicon nitride heterogeneous integrated N×N high-speed optical switch array of the present invention. In this example, the N×N heterogeneous integrated Mach-Zehnder optical switch array is constructed using a PILOSS structure, as shown in the figure, including N 2 One 2×2 heterogeneous integrated MZ optical switch unit 1.1, (N-1) 2 1.2 silicon nitride waveguide cross junctions.

[0054] In this embodiment, the N×N heterogeneous integrated high-speed optical switch array is formed by cascading 2×2 heterogeneous integrated MZI high-speed optical switch units 1.1 in a PILOSS structure, consisting of N 2The 2×2 heterogeneous integrated MZI high-speed optical switching units 1.1 are arranged into an N-row N-column switching matrix. The switching array can also be constructed from the switching units 1.1 using classic topologies such as Butterfly, Benes and improved Benes, DLN and improved DLN, S&S, Crossbar, and PILOSS.

[0055] Figure 2 Figures 1 and 2 are schematic diagrams of the structure of a 1×2 / 2×2 heterogeneous integrated MZI high-speed optical switch unit in an N×N high-speed optical switch array of thin-film electro-optic crystal / silicon nitride heterogeneous integration according to the present invention. As shown, the 1×2 or 2×2 heterogeneous integrated MZI high-speed optical switch unit 1.1 comprises one input waveguide I12.1. The 2×2 optical switch unit comprises two input waveguides I13.1 and I23.2, two output waveguides O12.2 / O13.3 and O22.3 / O23.4, a first silicon nitride 3dB coupler 2.4 / 3.5, a second silicon nitride 3dB coupler 2.5 / 3.6, a heterogeneous integrated evanescent wave coupler 2.6 / 3.7, and an electro-optic crystal / silicon nitride hybrid waveguide phase shifter 2.7 / 3.8. When light enters from the I1 (or I2) input waveguide 2.1 / 3.1 or 3.2, the waveguide cross-section is shown in the figure. Figure 4 (a) The light is uniformly split into two beams by the first silicon nitride 3dB coupler 2.4 / 3.5, and then coupled to the thin-film electro-optic crystal / silicon nitride heterogeneous integrated phase shifter 2.7 / 3.8 via heterogeneous integrated evanescent wave couplers 2.6 / 3.7. The waveguide cross-sections of the heterogeneous integrated evanescent wave couplers 2.6 / 3.7 and the thin-film electro-optic crystal / silicon nitride heterogeneous integrated phase shifter 2.7 / 3.8 are respectively referred to [reference needed]. Figure 4(b) and (c). After generating a phase difference of 0 or π, the two thin-film electro-optic crystal / silicon nitride heterogeneous integrated phase shifters 2.7 / 3.8 are coupled back to the silicon nitride waveguide via heterogeneous integrated evanescent wave couplers 2.6 / 3.7, and then combined by the second silicon nitride 3dB coupler 2.5 / 3.6. For a 1×2 heterogeneous integrated MZI high-speed optical switching unit, when the thin-film electro-optic crystal / silicon nitride heterogeneous integrated phase shifter 2.7 generates a phase difference of -π / 2, the two beams interfere constructively at the O2 output waveguide 2.4 of the switching unit 1.1, and the switching unit operates in the cross state; when the thin-film electro-optic crystal / silicon nitride heterogeneous integrated phase shifter 2.7 generates a phase difference of π / 2, the two beams interfere constructively at the O1 output waveguide 2.2 of the switching unit 1.1, and the switching unit operates in the bar state. For a 2×2 heterogeneous integrated MZI high-speed optical switching unit, when the thin-film electro-optic crystal / silicon nitride heterogeneous phase shifter 3.8 generates a 0-phase difference, the two beams interfere constructively at the output waveguide 3.4 or 3.3 of the O2 (or O1) of the switching unit 1.1, and the switching unit operates in a cross state. When the thin-film electro-optic crystal / silicon nitride heterogeneous phase shifter 3.8 generates a π-phase difference, the two beams interfere constructively at the output waveguide 3.3 or 3.4 of the O1 (or O2) of the switching unit 1.1, and the switching unit operates in a bar state. The first and second silicon nitride 3dB couplers 2.4 / 3.5 and 2.5 / 3.6 can be implemented using thermally adiabatic couplers, multimode interferometers, cascaded directional couplers, etc.

[0056] The thin-film electro-optic crystal / silicon nitride heterojunction coupler 2.6 / 3.7 in the embodiments is as follows: Figure 5 As shown, in this example, the thin-film electro-optic crystal is exemplified by thin-film lithium niobate. The thin-film electro-optic crystal may also include, but is not limited to, thin-film lithium tantalate, barium titanate, and lead zirconate titanate. In this example, the design band is C-band, which includes visible light, near-infrared, communication bands, and far-infrared bands. The heterogeneous integrated coupler is formed by overlapping a lower silicon nitride waveguide 5.1 and an upper lithium niobate waveguide 5.2 parallel to the transmission direction. To increase the alignment tolerance of the micro-transfer process, the silicon nitride waveguide 5.1 is widened to 2–3 μm. To reduce mode mismatch loss, the initial end 5.3 of the lithium niobate waveguide 5.2 is designed as a narrow waveguide on the order of hundreds of nanometers, and the final end 5.4 is designed as a wide waveguide that completely covers the silicon nitride waveguide 5.1. Finally, the silicon nitride waveguide 5.1 evolves to the required width 5.5. Both the lower silicon nitride waveguide 5.1 and the upper lithium niobate waveguide 5.2 are tapered waveguides with gradually changing widths. Its structure includes, but is not limited to, linearly stretched tapered waveguides, nonlinearly stretched tapered waveguides, and piecewise linearly stretched tapered waveguides. For the simulation structure of this heterogeneous integrated coupler, please refer to [link to simulation]. Figure 6 .in, Figure 6 (a) is the transmission spectrum of the evanescent wave coupler, with a coupling loss of approximately 0.1 dB at one end in the range of 1500-1600 nm. Figure 6 (b) is the alignment tolerance of the evanescent wave coupler perpendicular to the optical transmission direction, with the micro-transfer alignment tolerance being greater than ±1.5μm when the loss increases to 1dB at one end. Figure 6 (c) is the alignment tolerance of the evanescent wave coupler in the direction parallel to the optical transmission direction, where the coupler is almost insensitive to micro-transfer errors in this direction. Figure 6 (d) represents the alignment tolerance of the evanescent wave coupler in the rotation direction, where the coupler is almost insensitive to micro-transfer errors in this direction.

[0057] Figure 7 This invention relates to a method for fabricating a thin-film electro-optic crystal-silicon nitride heterogeneously integrated N×N high-speed optical switch array. In this example, the thin-film electro-optic crystal is a thin-film lithium niobate crystal. The thin-film electro-optic crystal may also include, but is not limited to, thin-film lithium tantalate, barium titanate, and lead zirconate titanate. The method includes the following steps:

[0058] S11. First, the lithium niobate chip on the insulator is organically cleaned using acetone / isopropanol / deionized water and dried with a nitrogen gun. Then, it is placed in a plasma desmearing machine to remove any residual colloidal particles and organic matter on the surface.

[0059] S12. Amorphous silicon (α-Si) is deposited on a lithium niobate substrate by plasma-enhanced chemical vapor deposition and used as a hard mask for subsequent waveguide etching;

[0060] S13. The heterogeneous integrated evanescent wave coupler and the tethered structure are exposed by photolithography.

[0061] S14. The structure is etched using argon (Ar) reactive ion etching (RIE);

[0062] S15. Use SC1 solution (H2O2:NH4OH:H2O) to clean the residual organic matter, then use potassium hydroxide aqueous solution (KOH:H2O=1:1) to remove α-Si, and finally use SC1 solution to clean again to complete the entire process of lithium niobate chip processing.

[0063] S2. Use hydrofluoric acid solution to wet etch the silicon oxide lower cladding of the lithium niobate chip, at which point the lithium niobate phase shifter is suspended;

[0064] S31. The silicon nitride chip on the insulator is organically cleaned, coated with photoresist and exposed through a photolithography machine to reveal the silicon nitride chip structure. It is then etched using a RIE etching machine, and finally the organic photoresist is removed to complete the entire process of silicon nitride chip processing.

[0065] S32. Deposit a silicon dioxide cladding layer to fill the etched area of ​​the waveguide;

[0066] S33. Use CMP to polish the wafer to achieve wafer surface planarization;

[0067] S4. Use a microwave plasma descaling machine to treat the silicon nitride chip to activate the surface;

[0068] S51. Use a PDMS stamp to quickly pick up the suspended lithium niobate phase shifter and transfer the lithium niobate phase shifter over the silicon nitride chip via PDMS.

[0069] S52. Adjust the silicon nitride chip and align the lithium niobate phase shifter to the target position;

[0070] S53. Move the PDMS stamp to bring the lithium niobate phase shifter into contact with the silicon nitride chip and apply a certain pressure to press it, so that the lithium niobate phase shifter and the silicon nitride chip surface produce a sufficiently strong van der Waals contact.

[0071] S54. Apply a specific shear force in the horizontal direction to the PDMS stamp, and slowly lift the PDMS stamp in the vertical direction;

[0072] S55. Separate the PDMS stamp and the lithium niobate phase shifter, and complete the transfer of the lithium niobate phase shifter to the silicon nitride chip.

[0073] S6. Spin-coating is performed on the silicon nitride chip and the electrode structure is exposed by a photolithography machine. Metal electrodes are deposited on both sides of the phase shifter waveguide by electron beam evaporation and lift-off processes.

[0074] S7. Deposit a silicon oxide overlayer on a silicon nitride chip by plasma-enhanced chemical vapor deposition;

[0075] S8. Spin-coating is performed on the silicon nitride chip and the electrode structure is exposed using a photolithography machine. The electrode window structure is then etched using a RIE etching machine to expose the metal electrodes.

[0076] At this point, the process flow for the N×N high-speed optical switch array based on thin-film electro-optic crystal / silicon nitride heterostructure integration is complete. Subsequent electrical testing can be performed by directly contacting the electrodes with probes or by using wire bonding to bring out the package.

Claims

1. A thin-film electro-optic crystal-silicon nitride heterogeneous integrated N×N high-speed optical switch array, characterized in that, include: N optical input ports and N optical output ports; A switching network consisting of several M×2 Mach-Zehnder (MZI) high-speed optical switching units connected according to a preset topology, where M = 1 or 2; Multiple silicon nitride waveguide cross structures; Each M×2MZI high-speed optical switch unit includes: Silicon nitride input waveguide at the input end; The first silicon nitride 3dB coupler is directly connected to the input waveguide; A pair of thin-film electro-optic crystal / silicon nitride heterogeneous evanescent wave couplers are respectively connected to the two output terminals of the first 3dB coupler; A pair of thin-film electro-optic crystal / silicon nitride heterogeneous phase-shifting regions are respectively connected to the evanescent wave coupler, and metal electrodes are provided on both sides of each phase-shifting region; A pair of thin-film electro-optic crystal / silicon nitride heterogeneous evanescent wave couplers are connected to the output terminals of the phase-shifting region, respectively; A second silicon nitride 3dB coupler is connected to the output terminal of the evanescent wave coupler; Silicon nitride output waveguide at the output end; A reconfigurable optical routing link is formed between any input port and the output port through the switching network. By controlling the phase shifting voltage of each M×2MZI high-speed optical switching unit, the cross or bar state switching of the optical path can be realized.

2. The thin-film electro-optic crystal-silicon nitride heterogeneous integrated N×N high-speed optical switch array according to claim 1, characterized in that, The thin-film electro-optic crystal / silicon nitride heterostructure integrated phase-shifting region includes: The underlying silicon nitride waveguide; The upper thin-film electro-optic crystal planar waveguide is heterogeneously integrated with the silicon nitride waveguide through micro-transfer printing technology; Metal electrodes disposed on both sides of the thin-film electro-optic crystal waveguide; In this process, the silicon nitride waveguide and the thin-film electro-optic crystal waveguide achieve optical field coupling through evanescent wave coupling.

3. The thin-film electro-optic crystal-silicon nitride heterogeneous integrated N×N high-speed optical switch array according to claim 1, characterized in that, The first silicon nitride 3dB coupler and the second silicon nitride 3dB coupler are thermally adiabatic couplers, multimode interferometers, or cascaded directional couplers.

4. The thin-film electro-optic crystal-silicon nitride heterogeneous integrated N×N high-speed optical switch array according to claim 1, characterized in that, The topology is any one of PILOSS, Butterfly, Benes and improved Benes, DLN or improved DLN, S&S or Crossbar, wherein: The PILOSS structure consists of N×N 2×2 MZI units arranged in an N-row N-column switch matrix; The Benes structure is composed of log2N levels of 2×2 MZI units cascaded together; Each MZI unit is connected by a silicon nitride waveguide, and a silicon nitride waveguide cross structure is used at the intersection.

5. The thin-film electro-optic crystal-silicon nitride heterogeneous integrated N×N high-speed optical switch array according to claim 1, characterized in that, The thin-film electro-optic crystal is a thin-film lithium niobate, a thin-film lithium tantalate, barium titanate, or lead zirconate titanate.

6. A method for fabricating a thin-film electro-optic crystal-silicon nitride heterogeneous integrated N×N high-speed optical switch array, characterized in that, Includes the following steps: S1. On a thin-film electro-optic crystal wafer on an insulator: - Evanescent wave couplers and tethered structures were fabricated using photolithography and dry etching; -Wet etching removes the lower cladding of the electro-optic crystal, forming a suspended phase shifter structure; S2. On a silicon-silicon dioxide-silicon nitride wafer: - Silicon nitride passive waveguides and devices are fabricated by photolithography and dry etching; -Deposit a silica cladding layer; - Use chemical mechanical polishing to polish the wafer to achieve surface planarization; S3. Use micro-transfer technology to transfer the electro-optic crystal phase shifter to the silicon nitride waveguide; S4. Deposit metal electrodes on both sides of the phase-shifting region; S5. Deposit the silicon oxide cladding and complete the electrode windowing.

7. The method for fabricating a thin-film electro-optic crystal-silicon nitride heterogeneous integrated N×N high-speed optical switch array according to claim 6, characterized in that, The micro-transfer technology in step S3 specifically includes: S51. Use a polydimethylsiloxane (PDMS) stamp to pick up the electro-optic crystal phase shifter and transfer the electro-optic crystal phase shifter over the silicon nitride chip via PDMS; S52. Adjust the silicon nitride chip and align the electro-optic crystal phase shifter to the target position; S53. Move the PDMS stamp to bring the electro-optic crystal phase shifter into contact with the silicon nitride chip and apply a certain pressure to press it, so that the electro-optic crystal phase shifter and the surface of the silicon nitride chip produce a sufficiently strong van der Waals contact. S54. Apply a specific shear force in the horizontal direction to the PDMS stamp, and slowly lift the PDMS stamp in the vertical direction; S55. Separate the PDMS stamp and the electro-optical phase shifter, and complete the transfer of the electro-optical phase shifter to the silicon nitride chip.

Images