A photonic computing chip
By using fiber optic connections between VCSEL arrays and photodetector arrays in a photonic computing chip, combined with optoelectronic devices and multimode fiber, the problems of optical signal crosstalk and high bit error rate are solved, achieving efficient optical signal propagation and computation, and improving data processing capabilities.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Utility models(China)
- Current Assignee / Owner
- SHENZHEN ZHONGKE OPTICAL SEMICON TECH CO LTD
- Filing Date
- 2025-08-04
- Publication Date
- 2026-07-14
AI Technical Summary
Existing photonic computing chips suffer from problems such as complex structure, low modulation rate, severe crosstalk in optical signal channels, and high bit error rate, making it difficult to meet the needs of artificial intelligence computing.
A VCSEL array and a photodetector array are connected by optical fiber. Couplers at both ends of the optical fiber are used to connect the VCSEL unit and the photodetector unit to ensure that the optical signal propagates without crosstalk. Auxiliary calculations are performed by optoelectronic devices, and the beam quality is improved by combining multimode fiber and surface lens.
It effectively solves the problem of high bit error rate caused by crosstalk of optical signals, realizes accurate propagation and calculation of optical signals, increases data processing capability, reduces beam divergence angle, and improves calculation accuracy and data processing capability.
Smart Images
Figure CN224501223U_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of optical computing chip technology, and more specifically, to a photonic computing chip. Background Technology
[0002] With the rapid development of artificial intelligence (AI), the computing speed of electronic chips based on the von Neumann architecture is gradually becoming insufficient to meet the demands of AI computing, and electronic chips also consume a lot of energy. In order to improve computing power and save energy, people have begun to develop optical computing chips that use light as the information carrier.
[0003] In the process of developing this invention, the inventors discovered that with the development of the field of photonic computing, many different basic architectures have been proposed. For example, ordinary photonic computing devices use light-emitting diodes, which have relatively low speeds. This type of ordinary photonic computing architecture suffers from severe crosstalk and a high bit error rate. Such photonic computing can only achieve photocurrent superposition and cannot perform logical operations. Other architectures, such as the Stanford structure and the Reck architecture, involve the interaction of light and light, as well as the design of relatively complex circuits and optical paths. They also suffer from structural complexity, low modulation rates, and severe crosstalk between different optical signal channels. Summary of the Invention
[0004] In view of the above problems, the present invention is proposed to provide a photonic computing chip that overcomes or at least partially solves the above problems.
[0005] In a first aspect, embodiments of the present invention provide a photonic computing chip, comprising: a VCSEL array, wherein the VCSEL array includes a plurality of VCSEL units for converting electrical signals into optical signals and outputting them;
[0006] A photodetector array, comprising multiple photodetector units, for receiving optical signals;
[0007] An optical fiber is provided, with the VCSEL unit and the photodetector unit connected to its two ends respectively, and coupling mirrors are provided at both ends of the optical fiber.
[0008] Furthermore, the photonic computing chip also includes optoelectronic devices, which are connected in series or parallel on the optical fiber for processing optical signals to achieve auxiliary calculations.
[0009] Furthermore, the optoelectronic device includes an optoelectronic modulator, an optoelectronic switch, or an integrated chip.
[0010] Furthermore, the output terminal of the photodetector array is connected to a logic gate circuit to receive the electrical signal output by the photodetector array, calculate it, and output the logic calculation result.
[0011] Furthermore, the optical fiber is a multimode optical fiber.
[0012] Furthermore, the photodetector unit includes one of gallium arsenide photodetectors, photodiodes, or photoresistors.
[0013] Furthermore, the connection between the optical fiber and the VCSEL unit and the photodetector unit is such that: the VCSEL unit is connected to one of the photodetector units via an optical fiber; or,
[0014] Multiple VCSEL units are connected to one photodetector unit via optical fibers.
[0015] Furthermore, the photodetector array is designed to be combined for superposition of electrical signals.
[0016] The second aspect discloses an optical intelligent computing module, including the use of any photonic computing chip.
[0017] The beneficial effects of the above-described technical solutions provided in the embodiments of the present invention include at least the following:
[0018] This invention provides a photonic computing chip. In this embodiment, a VCSEL array is used as the transmitter, a photodetector array as the receiver, and optical signals are propagated between the transmitter and receiver via optical fiber. This effectively solves the problem of high bit error rate caused by optical signal crosstalk.
[0019] In this embodiment, the VCSEL unit and the photodetector unit are connected at both ends of an optical fiber, ensuring that the optical signal emitted by each VCSEL unit can be accurately propagated to the photodetector unit without crosstalk. At the same time, due to the optical fiber connection, the VCSEL array and the photodetector array can be expanded horizontally or vertically according to the actual situation. That is, when the horizontal space of the chip architecture is limited, it can be expanded vertically; when the vertical space of the chip architecture is limited, it can be expanded horizontally, so as to support a larger amount of data and increase the computing power.
[0020] Other features and advantages of the invention will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the written description and the accompanying drawings.
[0021] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description
[0022] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the accompanying drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn to scale.
[0023] Figure 1 This is a schematic diagram of the photonic computing chip architecture according to the first embodiment of the present invention;
[0024] Figure 2 This is a schematic diagram of the photonic computing chip architecture according to the second embodiment of the present invention;
[0025] Figure 3 This is a schematic diagram of the VCSEL unit according to the first embodiment of the present invention;
[0026] Figure 4 This is a schematic diagram of the surface lens according to the first embodiment of the present invention;
[0027] Figure 5 This is the present invention. Figure 3 A diagram from another perspective;
[0028] Figure 6 This is the present invention. Figure 5 AA section view;
[0029] Figure 7 This is a schematic diagram of the surface lens according to the second embodiment of the present invention;
[0030] Figure 8 This is the present invention. Figure 7 BB section view;
[0031] Figure 9 This is a diagram showing the output light field distribution of the fundamental mode and higher-order modes of a traditional VCSEL.
[0032] Figure 10 These are the light field distribution diagrams of the fundamental mode and higher-order modes of this invention;
[0033] Figure 11 This is a magnified view of the output fundamental mode and higher-order mode optical field distribution of a traditional VCSEL.
[0034] Figure 12 This is an enlarged view of the light field distribution of the fundamental mode and higher-order modes of the present invention.
[0035] The image shows:
[0036] 1. VCSEL array; 2. Photodetector array;
[0037] 10. VCSEL unit; 20. Optical fiber; 30. Coupler mirror; 40. Photodetector unit;
[0038] 101. Substrate; 102. N-type DBR section; 103. N-electrode; 104. Active region layer; 105. P-type DBR section; 106. P-electrode; 107. Surface lens; 108. h-BN thin film;
[0039] 1070, First lens layer; 1071, Second lens layer; 1072, Third lens layer; 1073, Fourth lens layer; 1074, Fifth lens layer; 1075, Sixth lens layer; 1076, Oxide layer. Detailed Implementation
[0040] The embodiments of the technical solution of the present invention will now be described in detail with reference to the accompanying drawings. These embodiments are merely illustrative of the technical solution of the present invention and are therefore intended to limit the scope of protection of the present invention.
[0041] It should be noted that, unless otherwise stated, the technical or scientific terms used in this application should have the ordinary meaning as understood by those skilled in the art to which this invention pertains.
[0042] In the description of this application, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", "clockwise", "counterclockwise", "axial", "radial", "circumferential", etc., indicating the orientation or positional relationship are based on the orientation or positional relationship shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, and are not intended to indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the present invention.
[0043] Furthermore, the terms "first," "second," etc., 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. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly defined.
[0044] In this application, unless otherwise expressly specified and limited, the terms "installation," "connection," "linking," and "fixing," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.
[0045] In this application, unless otherwise expressly specified and limited, "above" or "below" the second feature can mean that the first feature is in direct contact with the second feature, or that the first feature is in indirect contact with the second feature through an intermediate medium. Furthermore, "above," "on top of," and "over" the second feature can mean that the first feature is directly above or diagonally above the second feature, or simply that the first feature is at a higher horizontal level than the second feature. "Below," "below," and "under" the second feature can mean that the first feature is directly below or diagonally below the second feature, or simply that the first feature is at a lower horizontal level than the second feature.
[0046] With the rapid development of artificial intelligence (AI), the computing speed of current electronic chips based on the von Neumann architecture is gradually becoming insufficient to meet the demands of AI computing, and the excessive energy consumption of electronic chips may lead to a serious energy crisis. Considering both computing power and energy efficiency, optical computing chips, which use light as the information carrier, represent one of the development directions for next-generation chips.
[0047] See attached document Figure 3 As shown, a VCSEL is a semiconductor laser structure in which an optical resonant cavity is formed perpendicular to the semiconductor epitaxial wafer, and the emitted laser beam is perpendicular to the substrate surface. It is widely used in consumer electronics, military, and medical fields, such as 3D sensing, AR / VR, robotics, LiDAR, consumer electronics, smart manufacturing, IoT, and data centers. Specific applications include autonomous driving, video surveillance, LiDAR, machine vision, gesture recognition, and facial recognition.
[0048] VCSEL is one of the important active devices used in optical computing. It has the advantages of small size, high speed, addressability and arrayability, and is used in optical intelligent computing.
[0049] The inventors of this application continued their research on how to effectively solve the problems of complex structure, low modulation rate, severe crosstalk between different optical signal channels, and high bit error rate in photonic computing chips. The inventors' research included at least: how to connect the optical signal transmitter and receiver via optical fiber 20, and how to connect VCSEL unit 10 and photodetector unit 40 at both ends of optical fiber 20 respectively, ensuring that the optical signal emitted by each VCSEL unit 10 can accurately propagate to the photodetector unit 40 without crosstalk. After extensive and repeated research, the inventors proposed the photonic computing chip and system of this application.
[0050] Before describing the embodiments of the present invention in detail, the design concept of the present invention will be summarized below. (Refer to the appendix.) Figure 1As shown, the present invention designs a photonic computing chip. The photonic computing chip connects the optical signal transmitter and the optical signal receiver through optical fiber 20, ensuring that the optical signal emitted by each VCSEL unit 10 can be accurately propagated to the photodetector unit 40 without crosstalk, effectively improving the quality of the beam during propagation, improving the anti-crosstalk capability of the optical signal, and ensuring the calculation accuracy.
[0051] See attached document Figure 1 As shown, this embodiment of the invention provides a photonic computing chip, including: a VCSEL array 1, wherein the VCSEL array 1 includes a plurality of VCSEL units 10, used to convert electrical signals into optical signals and output them;
[0052] A photodetector array 2, comprising a plurality of photodetector units 40, for receiving optical signals;
[0053] Optical fiber 20, with VCSEL unit 10 and photodetector unit 40 connected to its two ends respectively, and coupling mirrors 30 are provided at both ends of optical fiber 20.
[0054] Understandably, traditional photonic computing chip architectures suffer from severe crosstalk and high bit error rates, or architectures using beam-to-beam coherence suffer from complex circuitry and optical paths.
[0055] In this embodiment, VCSEL array 1 is used as the transmitter and photodetector array 2 is used as the receiver. The transmitter and receiver transmit optical signals through optical fiber 20, which can effectively solve the problem of high bit error rate caused by optical signal crosstalk.
[0056] In this embodiment, the VCSEL unit 10 and the photodetector unit 40 are connected at both ends of the optical fiber 20, respectively, to ensure that the optical signal emitted by each VCSEL unit 10 can be accurately propagated to the photodetector unit 40 without crosstalk. At the same time, since they are connected by the optical fiber 20, the VCSEL array 1 and the photodetector array 2 can be expanded horizontally or vertically according to the actual situation. That is, when the horizontal space of the chip architecture is limited, it can be expanded vertically; when the vertical space of the chip architecture is limited, it can be expanded horizontally, so as to support a larger amount of data and increase the computing power.
[0057] In this embodiment, the two ends of the optical fiber 20 are coupled to the VCSEL unit 10 and the photodetector unit 40 respectively through the coupling mirror 30, so as to ensure that the light beam can accurately reach the corresponding photodetector.
[0058] See attached document Figure 1As shown, in some embodiments, the optical fiber 20 is connected to the VCSEL unit 10 and the photodetector unit 40 in such a way that the VCSEL unit 10 is connected to one of the photodetector units 40 via the optical fiber 20.
[0059] Understandably, by connecting a single VCSEL unit 10 and a single photodetector unit 40 at each end of the optical fiber 20, it is ensured that the optical signal emitted by each VCSEL unit 10 can accurately propagate to the photodetector unit 40 without crosstalk, thus improving crosstalk immunity. In other words, each VCSEL unit 10 on the VCSEL array 1 is connected one-to-one with the photodetector unit 40 on the photodetector array 2.
[0060] See attached document Figure 2 As shown, in some other embodiments, the optical fiber 20 is connected to the VCSEL unit 10 and the photodetector unit 40 such that multiple VCSEL units 10 are connected to one photodetector unit 40 via the optical fiber 20.
[0061] Understandably, by connecting a single VCSEL unit 10 to multiple photodetector units 40 at each end of the optical fiber 20, it is ensured that the optical signal emitted by each VCSEL unit 10 can accurately propagate to the photodetector unit 40 without crosstalk, thus improving crosstalk immunity. That is, the VCSEL units 10 on the VCSEL array 1 are respectively connected to the multiple photodetector units 40 on the photodetector array 2.
[0062] In a further embodiment, the photodetector unit 40 includes one of a gallium arsenide photodetector, a photodiode, or a photoresistor.
[0063] In another embodiment, the photonic computing chip further includes a surface lens 107 disposed on the light-emitting surface of the VCSEL unit 10. The refractive index of the surface lens 107 gradually increases from bottom to top, and the refractive index of the surface lens 107 gradually increases or decreases from bottom to top, for focusing the beam spot.
[0064] See attached document Figure 3 and 4 As shown in the embodiment, to ensure the quality of the light beam, a surface lens 107 is provided on the light-emitting surface of the VCSEL unit 10. This surface lens 107 has a focusing function, which can converge the light beam into a smaller spot and significantly reduce the beam divergence angle. This metasurface effectively improves the quality of the light beam during propagation, enhances the anti-crosstalk capability of the optical signal, and ensures the accuracy of calculations.
[0065] In this embodiment, each VCSEL unit 10, photodetector unit 40, and optical fiber 20 connecting them, along with the corresponding optical path and optical components, constitute a basic optical computing unit. When the VCSEL unit 10 receives an electrical signal and emits an optical signal, it propagates the optical signal to the photodetector unit 40 through the optical fiber 20. The photodetector unit 40 senses the optical signal and generates a photocurrent, which is a binary '1'. When the VCSEL unit 10 is not working, the photodetector unit 40 does not generate a photocurrent, which is a binary '0'. For example, 4 × 4 = 16 optical computing units can carry 2^16 = 65536 bits of information.
[0066] In a further embodiment, the photodetector array 2 is designed to be combined for superposition of electrical signals.
[0067] The merging design here is to combine the circuits output by the photodetector array 2 together to detect the total current, and to obtain the superposition operation based on the current value.
[0068] The superposition operation of this photonic computing chip is described below in an example:
[0069] By merging the independent lines at the receiving end, the corresponding information can be superimposed. According to Kirchhoff's circuit laws, the photocurrent of the photodetector array 2 is linearly superimposed, and the photocurrent is directly proportional to the number of detected '1's, such as... Figure 1 A single channel has a current of 1 mA, and all channels lit up equals 16 mA. Sequence calculation can also be achieved by detecting the current in the photodetector array 2. For example, if the transmitter is 011100, the receiver can detect the current to achieve three "1"s.
[0070] For example:
[0071] A VCSEL array 1 is composed of 16 VCSEL units 10 in a 4×4 configuration. The VCSEL array 1 serves as a transmitter. The VCSEL array 1 can receive electrical signals and convert them into optical signals.
[0072] A photodetector array 2 is composed of 16 photodetector units 40 in a 4×4 configuration, and the photodetector array 2 serves as the receiving end.
[0073] Each VCSEL unit 10 of the VCSEL array 1 corresponds one-to-one with the photodetector unit 40 of the photodetector array 2, and the VCSEL unit 10 and the photodetector unit 40 are connected in a single correspondence via optical fiber 20, realizing one input end corresponding to one output end. The embodiment improves anti-interference capability through the connection via optical fiber 20.
[0074] In another embodiment, multiple VCSEL units 10 can be connected to a single photodetector unit 40 via optical fiber 20, meaning multiple input terminals correspond to one output terminal. For example:
[0075] A VCSEL array 1 is composed of 64 VCSEL units 10 in an 8×8 configuration. The VCSEL array 1 serves as a transmitter, and the electrical signals received by the VCSEL array 1 are converted into optical signals.
[0076] A photodetector array 2 is composed of 16 photodetector units 40 in a 4×4 configuration, and the photodetector array 2 serves as the receiving end.
[0077] Among them, four VCSEL units 10 serve as transmitters and are connected to one photodetector unit 40 via four optical fibers 20. That is, multiple inputs correspond to one output, which improves the computing power of complex data.
[0078] In a further embodiment, the photonic computing chip also includes optoelectronic devices, which are connected in series or in parallel on the optical fiber 20 for processing optical signals to achieve auxiliary calculations.
[0079] Understandably, in order to improve computing power, in this embodiment, an optoelectronic device is installed on the optical fiber 20 connecting the transmitter and receiver. The optoelectronic device is connected in series or parallel on the optical fiber 20. By adding external optoelectronic signals through the optoelectronic device, certain characteristics of the electrical signal emitted by the transmitter are changed, such as changing the polarity, phase, amplitude, and frequency. Then, the changed optical signal participates in the calculation. In this way, by adding external electrical signals, programmability can be achieved, increasing the ability to process complex data.
[0080] For example,
[0081] In one embodiment, a single or multiple optical switch is added between the VCSEL array 1 and the detector array to control the 'on' or 'off' of the optical signal emitted by the VCSEL array 1, that is, to realize the 'OR gate' or 'NOT gate' in the optical path.
[0082] In another embodiment, a specific silicon photonic integrated chip is added between a pair (two-way) VCSEL unit 10 and a detector unit. By adjusting the phase change of one of the channels, the optical path of the VCSEL unit 10 and the detector unit can be made to interfere with the optical path of the other VCSEL unit 10 and the detector unit, thereby realizing the logical addition and logical multiplication calculation of the two VCSEL units 10 and the detector unit.
[0083] In a further embodiment, the optoelectronic device includes an optoelectronic modulator, an optoelectronic switch, or an integrated chip.
[0084] In order to change certain characteristics of the electrical signal emitted by VCSEL array 1, such as changing polarity, phase, amplitude and frequency, in this embodiment, the optoelectronic device is an optoelectronic modulator, optoelectronic switch or integrated chip, etc., and the optoelectronic device is connected in parallel or in series to optical fiber 20.
[0085] In a further embodiment, the output terminal of the photodetector array 2 is connected to a logic gate circuit, which is used to receive the electrical signal output by the photodetector array 2, calculate and output the logic calculation result.
[0086] In this embodiment, a logic gate circuit can be provided at the electrical signal output terminal of the photodetector array 2 of the photonic computing chip. The electrical signal is then calculated using the logic gate circuit to obtain the final calculation result.
[0087] The application of logic computing in chips (integrated circuits) is the core of modern digital electronics technology. Almost all digital chips (such as CPUs, GPUs, memory, FPGAs, and application-specific integrated circuits) rely on logic gate circuits to implement their functions. Among them, the basic computing unit in a chip is a logic gate, which is composed of transistors (such as MOSFETs in CMOS technology) and implements Boolean operations. A logic gate is implemented through several transistors.
[0088] In a further embodiment, the optical fiber 20 is a multimode optical fiber.
[0089] In this embodiment, multimode fiber is used to connect the VCSEL unit 10 and the detector unit, which can reduce or eliminate crosstalk between different optical paths, accurately transmit optical signals, and reduce the bit error rate.
[0090] See attached document Figure 5 and 6 As shown in the embodiment, the outer surface of the surface lens 107 is provided with an oxide layer 1076.
[0091] Understandably, by using a wet oxidation process to form an oxide layer 1076 on the outer surface of the surface lens 107, and the refractive index of the oxide layer 1076 in the surface lens 107 is lower than the refractive index of the unoxidized layer, the oxidation reduces the refractive index of the oxidized portion of the surface lens 107, and the refractive index of the lens layer in the surface lens 107 is gradually changed. This structural arrangement makes the light beam more focused when it is emitted from the surface lens 107.
[0092] See attached document Figure 8 As shown in the embodiment, the surface lens 107 is composed of multiple lens layers stacked together, and the lens layers are configured with a gradient refractive index, wherein the refractive index of the lens layer gradually decreases from the center to the edge.
[0093] Understandably, in order to focus the diverging beam emitted by the VCSEL, the embodiment is that the surface lens 107 is composed of multiple lens layers stacked together, the refractive index of the surface lens 107 gradually increases from bottom to top, and the refractive index at the edge position and the center position of the lens layer are different, thus increasing the focusing effect when the beam is emitted from the surface lens 107.
[0094] In some embodiments, the surface lens 107 is composed of 32 lens layers stacked together, and the refractive index of the 32 lens layers gradually increases from bottom to top, and the refractive index of each lens layer gradually decreases from the center to the edge; see attached figure. Figure 7 and 8 As shown, in other embodiments, the surface lens 107 is composed of 30 lens layers stacked together, and the refractive index of the 30 lens layers gradually increases from bottom to top. Furthermore, the refractive index of the center position and the edge position of the lens layer are different. That is, the refractive index of the oxidized part at the edge of the lens layer is lower than that of the unoxidized part due to changes in internal composition. Thus, at least two parts of a single lens layer have different refractive indices.
[0095] In a further embodiment, the refractive index of the lens layer near the edge is controlled to be 1.5-1.65; the refractive index of the lens layer near the center is controlled to be 2.8-3.4.
[0096] Understandably, in order to make the beam more focused, the embodiment stacks multiple lens layers to form the surface lens 107, and the refractive index of the lens layers increases from bottom to top, and the refractive index is controlled between 2.8 and 3.4; then, the outer surface of the surface lens 107 is oxidized by a wet oxidation process to obtain an oxide layer 1076, so that the refractive index of the edge of the lens layer is controlled between 1.5 and 1.65. In this embodiment, not only does the overall refractive index gradually change from bottom to top, but the refractive index of the edge of the lens layer is also different from that of the center. This embodiment can efficiently focus the diverging beam emitted by the VCSEL, converge the beam into a smaller spot, and significantly reduce the beam divergence angle.
[0097] In some embodiments, the surface lens 107 is composed of multiple lens layers, and the refractive index of the lens layers is set from bottom to top to 2.8-3.4, wherein the refractive index of the bottommost lens layer is 3, the refractive index of the topmost lens layer is 3.4, and the refractive index of the outer surface of the surface lens 107 after oxidation is 1.5; in other embodiments, the surface lens 107 is composed of multiple lens layers, and the refractive index of the lens layers is set from bottom to top to 2.8-3.4, wherein the refractive index of the bottommost lens layer is 2.8, the refractive index of the topmost lens layer is 3.1, and the refractive index of the outer surface of the surface lens 107 after oxidation is 1.65; in still other embodiments, the surface lens 107 is composed of multiple lens layers, and the refractive index of the lens layers is set from bottom to top to 2.8-3.4, wherein the refractive index of the bottommost lens layer is 2.8, the refractive index of the topmost lens layer is 3.2, and the refractive index of the outer surface of the surface lens 107 after oxidation is 1.6.
[0098] See attached document Figure 5 and 6 As shown in the embodiment, the stepped structure is configured such that the side of the surface lens 107 has multiple steps, and each step corresponds to each of the lens layers.
[0099] It is understood that the surface lens 107 is composed of multiple lens layers, and the side surface of the surface lens 107 has multiple steps. Specifically, each lens layer in the surface lens corresponds to a step. It can also be considered that the cross-section of each lens layer forming the surface lens 107 is different, and the cross-section of the lens layer in the surface lens 107 gradually decreases from bottom to top, and the side surface of the surface lens 107 forms a stepped structure.
[0100] For further explanation, please refer to the appendix. Figure 5 As shown, in order to enhance the beam focusing effect, this embodiment sets the outer surface of the surface lens 107 as a stepped structure surface; however, it is not excluded that, referring to the appendix... Figure 7 and 8 As shown, the outer surface of the surface lens 107 is set as a smooth surface.
[0101] In a further embodiment, the VCSEL unit 10 also includes an h-BN film 108, which is disposed on the side of the surface lens 107 away from the VCSEL unit 10.
[0102] Understandably, by depositing a certain thickness of h-BN film 108 on the VCSEL unit 10, the h-BN film 108 can quickly dissipate the heat generated by the VCSEL unit 10 and the heat conducted from the active region; at the same time, it has good optical transparency in the visible and near-infrared bands, has minimal light absorption loss on the VCSEL unit 10, does not affect the beam quality, and can also protect the underlying oxide layer 1076.
[0103] In some embodiments, a surface lens 107 is disposed on the beam emitting surface of the VCSEL unit 10. The surface lens 107 has a stepped structure and is a multi-layered surface lens 107. An h-BN film 108 with a thickness of 50 nm is disposed on the side of the surface lens 107 away from the VCSEL unit 10. In other embodiments, a surface lens 107 is disposed on the beam emitting surface of the VCSEL unit 10. The surface lens 107 has a stepped structure and is a multi-layered surface lens 107. An h-BN film 108 with a thickness of 70 nm is disposed on the side of the surface lens 107 away from the VCSEL unit 10. In still other embodiments, a surface lens 107 is disposed on the beam emitting surface of the VCSEL unit 10. The surface lens 107 has a stepped structure and is a multi-layered surface lens 107. An h-BN film 108 with a thickness of 60 nm is disposed on the side of the surface lens 107 away from the VCSEL unit 10.
[0104] For example:
[0105] A surface lens 107 is obtained on the upper surface of the VCSEL unit 10. The surface lens 107 consists of six layers with an overall height of 60 nm. The refractive index of the outer surface of the surface lens 107 is 1.58 through oxidation. The specific data of each layer are shown in the table below:
[0106]
[0107] In this experiment, the surface lens 107 includes a first lens layer 1070 to a sixth lens layer 1075, forming a circular boss structure with a stepped outer surface. The sixth lens layer 1075, with the largest cross-sectional area, is in contact with the VCSEL unit 10, while the first lens layer 1070, with the smallest cross-sectional area, is furthest from the VCSEL unit 10. The refractive index gradually increases from the sixth lens layer 1075 to the first lens layer 1070. Simultaneously, through an oxidation process, the refractive index near the edge of the sixth lens layer 1075 to the first lens layer 1070 is 1.58, and the oxidation distance from the edge to the center is approximately 5-6 μm.
[0108] Then, the VCSEL chip with the above-mentioned graded refractive index surface lens 107 was compared with a traditional VCSEL chip. By observing the output light field distribution diagrams of the two, it can be seen that the VCSEL chip prepared by this method has a significantly better light spot focusing effect, as shown below:
[0109] See attached document Figure 9 , 10 11 and 12, Figure 9 The output light field distribution diagrams for the fundamental mode and higher-order modes of a traditional VCSEL; Figure 10 Light field distribution diagrams of the fundamental mode and higher-order modes of the refractive index graded surface lens 107 on the VCSEL are set; Figure 11 A magnified view of the output fundamental mode and higher-order mode optical field distribution of a traditional VCSEL; Figure 12 Enlarged view of the fundamental and higher-order mode light field distributions of the VCSEL with a refractive index graded surface lens 107.
[0110] According to the appendix Figure 9 , 10 As can be seen from 11 and 12, the output beam diameter of the VCSEL with the surface lens 107 having a focusing function is reduced from 1.2 μm to 0.6 μm compared to the traditional VCSEL, and the beam focusing effect is obvious. That is, compared with the traditional VCSEL, the VCSEL of this embodiment can reduce the beam divergence angle by more than 70% and reduce the beam size to 1 / 4 of the original, effectively improving the concentration of light energy.
[0111] Based on the same inventive concept, an optical intelligent computing system is also disclosed, which includes the aforementioned photonic computing chip.
[0112] Based on the same inventive concept, a method for applying a photonic computing chip is also disclosed, which includes the use of the aforementioned photonic computing chip.
[0113] For example, this photonic computing chip can be used in fields such as holographic optics, autonomous driving, remote sensing, and ultra-high resolution imaging.
[0114] For example, this photonic computing chip can be applied to optical neural network computing:
[0115] Neural network: A computational model that simulates biological neural networks, applied in fields such as pattern recognition, classification, regression, speech recognition, image processing, and natural language processing.
[0116] Optical neural network: A computational model that uses optical elements (such as lasers, waveguide devices, etc.) to simulate biological neural networks.
[0117] Neuron: The basic unit of a neural network, also known as a node or processing unit, used to accept input, perform calculations, and produce output. Each neuron can have multiple inputs and one output.
[0118] Weights: Used to adjust the importance of input signals, thereby affecting the output of neurons.
[0119] On-chip integration: Integrating multiple functional modules and components onto a single chip to perform complex computing and control tasks.
[0120] This embodiment proposes a vertically integrated chip architecture based on a photonic computing chip, enabling chip-level VCSEL-based optical neural networks and expanding the potential application areas of VCSEL optical neural networks. Utilizing the high modulation rate (GHz) of VCSELs, the computing power of this photonic computing chip will far exceed that of existing electronic chips, and will increase with the scaling up of VCSELs. Based on the passive propagation characteristics of light, the energy consumption of this optical computing chip will also be much lower than that of existing electronic chips, solving the energy problem faced by AI computing. This optical computing chip can be used in various application scenarios such as facial recognition, optical computing, image classification, 6G communication, optical encryption, and autonomous driving.
[0121] For specific examples and explanations of the beneficial effects of the method described in this embodiment, please refer to the above description of the photonic computing chip, which will not be repeated here.
[0122] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. This disclosure is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this disclosure is limited only by the appended claims. Thus, if these modifications and variations of the invention fall within the scope of the claims of the invention and their equivalents, the invention is also intended to include these modifications and variations.
Claims
1. A photonic computing chip, characterized in that, include: VCSEL array, which includes multiple VCSEL units for converting electrical signals into optical signals and outputting them; A photodetector array, comprising multiple photodetector units, for receiving optical signals; An optical fiber is provided, with the VCSEL unit and the photodetector unit connected to its two ends respectively, and coupling mirrors are provided at both ends of the optical fiber.
2. The photonic computing chip according to claim 1, characterized in that, The photonic computing chip also includes optoelectronic devices, which are connected in series or parallel on an optical fiber to process optical signals for auxiliary computation.
3. The photonic computing chip according to claim 2, characterized in that, The optoelectronic devices include optoelectronic modulators, optoelectronic switches, or integrated chips.
4. The photonic computing chip according to claim 1, characterized in that, The output terminal of the photodetector array is connected to a logic gate circuit, which is used to receive the electrical signal output by the photodetector array, calculate it, and output the logic calculation result.
5. The photonic computing chip according to claim 1, characterized in that, The optical fiber is a multimode optical fiber.
6. The photonic computing chip according to claim 1, characterized in that, The photodetector unit includes one of gallium arsenide photodetectors, photodiodes, or photoresistors.
7. The photonic computing chip according to claim 1, characterized in that, The optical fiber is connected to the VCSEL unit and the photodetector unit in the following manner: the VCSEL unit is connected to one of the photodetector units via an optical fiber; or... Multiple VCSEL units are connected to one photodetector unit via optical fibers.
8. The photonic computing chip according to claim 1, characterized in that, The photodetector array is designed to be combined for superposition of electrical signals.