A correction unit and an optical computing chip based on on-chip correction

Abstract

The present invention relates to the field of semiconductor technology, and more specifically to a correction unit and an optical computing chip based on on-chip correction. The correction unit comprises: a first waveguide; a cladding surrounding the outer surface of the first waveguide; a doped region disposed on the first waveguide, doped with a rare earth element; a third waveguide connected to the first surface of the cladding, and a first reflective layer disposed within the third waveguide at a first height Y1 from the first surface, at a second height Y2; and an amplifying portion disposed on the second surface of the cladding, the amplifying portion having a convex surface. The present invention proposes a local correction unit with a local focusing effect. This local correction unit is particularly suitable for signal correction scenarios in computing chips, reducing the difficulty of preparing correction-related components during production and processing while also lowering production costs.

Description

Technical Field

[0001] The present invention relates to the field of semiconductor technology, and in particular to a correction unit and an optical computing chip based on on-chip correction. Background Art

[0002] Rare-Earth Doped Waveguide Amplifiers (REDWA) is a new type of optical amplifier.

[0003] For example, patent application CN118259486A discloses a rare earth ion-doped optical waveguide amplifier with nanowires or thin films on its surface. The amplifier comprises a first lens, a first anti-reflection film, an optical waveguide, a second anti-reflection film, and a second lens, arranged in sequence. The optical waveguide is made of a rare earth ion-doped material, and its surface is coated with a metal film or metal nanowires. Signal light and pump light are sequentially transmitted through the first lens and the first anti-reflection film to the optical waveguide. The optical waveguide absorbs the energy of the pump light and transfers it to the signal light, amplifying the signal light. The amplified signal light is then transmitted through the second anti-reflection film and the second lens to the output optical fiber.

[0004] For example, patent application CN118264324A discloses a silicon-based erbium-doped waveguide amplifier with integrated optical modulation functionality. The amplifier comprises four main modules: an input module, an I / O coupling module, an optical modulation and amplification module, and an output module. The input module includes a signal light input module and a pump light input module; the I / O coupling module includes an input coupling module and an output coupling module; and the optical modulation and amplification module comprises four submodules: a modulation module, a multiplexing module, a gain module, and a filtering module.

[0005] For example, patent application CN115755279A discloses an on-chip optical amplifier coupler and a method for forming the same. The method includes depositing a predetermined thickness of a rare-earth ion-doped dielectric material on a multimode coupler to form a gain cladding layer of the multimode coupler; depositing a bonding dielectric layer on the gain cladding layer; and flip-chip bonding a pump laser on the bonding dielectric layer to form the on-chip optical amplifier coupler.

[0006] For example, patent application CN118169938A also discloses a rare-earth-doped polymer three-dimensional optical waveguide on-chip amplifier and its preparation method. The amplifier comprises, from bottom to top, a substrate, a pump light transmission waveguide, a signal light transmission waveguide, a wavelength division multiplexer, and an upper cladding. The pump light transmission waveguide and the signal light transmission waveguide have a three-dimensional spatial structure to support the transmission of high-order optical modes. The upper surface of the substrate has several parallel signal light gain channels, and each of the signal light gain channels has a pump light transmission channel to provide a pump source for it.

[0007] However, traditional scale-up solutions face huge challenges in terms of preparation process and cost reduction. Summary of the Invention

[0008] The object of the present invention is to provide a correction unit and an optical computing chip based on on-chip correction, which partially solve or alleviate the above-mentioned deficiencies in the prior art and can improve the correction capability of limited doping areas.

[0009] In order to solve the above-mentioned technical problems, the present invention specifically adopts the following technical solutions:

[0010] A first aspect of the present invention is to provide a correction unit, comprising:

[0011] First Waveguide;

[0012] The outer side of the first waveguide is wrapped with at least one cladding layer;

[0013] A doped region is provided on the first waveguide, the doped region penetrates the first waveguide and intersects at least a portion of the cladding, and the doped region is doped with at least one excitation element;

[0014] A third waveguide is connected to the first surface of the cladding, and a first reflective layer at a second height Y2 is provided in the third waveguide at a position at a first height Y1 from the first surface;

[0015] An enlarged portion is provided on the second surface of the cladding, wherein the enlarged portion has a convex surface, wherein when at least one beam of the modulated light source is incident on the enlarged portion along a first direction, the at least one beam of the modulated light source is focused in the enlarged portion;

[0016] When at least one beam of the adjustment light source is guided by the amplifying portion and enters the doped region, and penetrates the doped region and refracts into the third waveguide, and when the adjustment light source reaches the first reflective surface of the first reflective layer, the adjustment light source will be reflected on the first reflective surface, and at least one beam of the reflected light can be re-injected into the doped region;

[0017] In which, the excitation element can absorb the energy of the corresponding regulating light source so that the electrons inside it transition from a low energy level to a high energy level, thereby completing energy storage; when the first light signal is transmitted to the doped region through the first waveguide, the electrons that transition to the high energy level undergo stimulated radiation under the stimulation of the first light signal, releasing photons with the same frequency, phase and polarization state as the first light signal, and the photons are superimposed on the first light signal to form a second light signal, and the power of the second light signal is greater than the power of the first light signal.

[0018] In some embodiments, further comprising:

[0019] A second reflective layer is provided on the cladding and has a third height Y3, and the second reflective layer has an opening area formed at a position corresponding to the amplifying portion, so as to reserve a light source path between the amplifying portion and the doped region, so that the regulating light source concentrated in the amplifying portion can enter the doped region through the light source path; wherein, when at least one beam of reflected light is deflected and reaches the second reflective surface of the second reflective layer, it is reflected at the second reflective surface and re-enters the doped region under the effect of reflection.

[0020] In some embodiments, the cladding is provided with a first isolation groove and a second isolation groove at a first length X1 between the amplifying portion, wherein; a third isolation groove is provided on the third waveguide corresponding to the second isolation groove; wherein the first isolation groove, the second isolation groove, and the third isolation groove arranged along the same height direction form an isolation layer, and the isolation layer is filled with a medium, and the refractive index of the medium satisfies the following rule: 1<refractive index of medium<refractive index of cladding.

[0021] In some embodiments, the thickness of the insulating layer is 40um-100um, and the depth of the insulating layer is 80um-100um.

[0022] In some embodiments, the medium is air.

[0023] In some embodiments, the first length X1 is greater than 20 μm.

[0024] In some embodiments, the first height Y1 is 20-50 μm.

[0025] In some embodiments, the doped region includes: a central doped region, and an edge doped region surrounding the central doped region, and the element doping concentration in the edge doped region is lower than the doping concentration in the central doped region.

[0026] The present invention also provides an optical computing chip based on on-chip correction, comprising:

[0027] N rows of first waveguides;

[0028] N rows of second waveguides, the N rows of second waveguides being intersected with the first waveguide to form a plurality of calculation units, wherein the first waveguide in at least one of the calculation units includes: a first waveguide segment and a third waveguide segment, and a correction unit as described in any one of the embodiments of the present invention is connected between the first waveguide segment and the third waveguide segment;

[0029] Correspondingly, the optical computing chip further includes: a light source device, which is used to provide an adjustment light source to the doping region in the correction unit, wherein the adjustment light source is incident on the doping region in the form of spatial light.

[0030] Beneficial Technical Effect: This invention proposes a local correction unit capable of efficiently amplifying spatial light within a limited region. Specifically, through a segmented design, a local correction unit with a local focusing effect is introduced within the middle of a first waveguide, enabling signal correction requirements for large-scale optical computing chips through limited local doping. Specifically, the spatial synergy between the localized doping region and the first reflective layer enables the introduction of a modulated light source in the form of spatial light. This, while creating a local focusing effect, reduces the impact on the optical path design structure of the existing optical computing chip in batch correction scenarios.

[0031] Furthermore, the present invention utilizes multiple reflection paths and isolation spaces within the waveguide to achieve sufficient signal amplification even in limited doped regions. In other words, the signal amplification achieved in limited doped regions through the interaction of reflection paths and isolation spaces further facilitates the implementation of the technology for modulating the light source using spatial light.

[0032] Specifically, the present invention preferably introduces a correction signal (such as pump light) via spatial light, which avoids complicating the circuit structure of the optical computing chip. In one exemplary embodiment, a laser can be used to emit pump light into the doped region in the form of spatial light, guiding the pump light to be efficiently focused within a limited doped region. This improves signal amplification while reducing the risk of crosstalk between the pump light and the signal light. Specifically, introducing pump light into the doped region in the form of spatial light avoids adding new optical transmission paths (i.e., adding new waveguide structures) to the optical waveguide, thereby simplifying the array structure of the optical computing chip.

[0033] From another perspective, this local correction unit with a local focusing effect can limit the interaction between pump light and signal light to a limited space, thereby reducing the possibility or degree of crosstalk between the pump light and signal light, thereby improving the accuracy and reliability of the correction process. In other words, this local correction unit with a local focusing effect is particularly suitable for large-scale computing chips used in small model reasoning processes. This local correction unit can reduce interference with existing devices during the large-scale introduction of optical computing chips, and the local focusing effect helps to achieve effective amplification of optical signals within a limited doping area (thereby reducing both the difficulty and cost of the doping process). BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following is a brief introduction to the drawings required for the embodiments or the description of the prior art. In all drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the various elements or parts are not necessarily drawn according to the actual scale. Obviously, the drawings described below are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can also be obtained based on these drawings without inventive work.

[0035] Figure 1 is a schematic structural diagram of an optical computing chip in an exemplary embodiment of the present invention;

[0036] Figure 2 is a schematic structural diagram of an amplifying unit in a first exemplary embodiment of the present invention;

[0037] Figure 3 is a schematic structural diagram of an amplifying unit in a second exemplary embodiment of the present invention;

[0038] Figure 4 is a schematic structural diagram of an amplifying unit in a third exemplary embodiment of the present invention;

[0039] Figure 5 is a schematic structural diagram of an optical computing chip in another exemplary embodiment of the present invention;

[0040] Figure 6 for Figure 5 Schematic diagram of the structure of the amplification unit in the optical computing chip shown;

[0041] Figure 7 Schematic diagram of the structure of a pump light adjustment unit in another exemplary embodiment of the present invention;

[0042] Figure 8 Schematic diagram of the structure of a pump light adjustment unit in another exemplary embodiment of the present invention.

[0043] Summary of reference numerals:

[0044] First waveguide 1, cladding 2, amplifying portion 21, first reflective layer 3, third waveguide 4, doped region 5, second reflective layer 6, isolation layer 7, first isolation trench 71, second isolation trench 72, third isolation trench 73, second waveguide 8, laser 9, fourth waveguide 10, phase-change material layer 101, first phase-change material layer 101a, second phase-change material layer 101b, third phase-change material layer 101c, first constriction 1011, second constriction 11, pump light 011, second optical signal 010, first waveguide segment I, second waveguide segment II, third waveguide segment III, computing unit L. DETAILED DESCRIPTION

[0045] To make the purpose, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0046] Herein, suffixes such as "module," "component," or "unit" used to represent elements are only used to facilitate description of the present invention and have no specific meaning. Therefore, "module," "component," or "unit" may be used interchangeably.

[0047] As used herein, terms such as "upper," "lower," "inner," "outer," "front," "back," "one end," and "the other end" indicate positions or locations based on those shown in the accompanying drawings. These terms are intended solely to facilitate and simplify the description of the present invention and are not intended to indicate or imply that the devices or components referred to must have, be constructed, or operate in a specific orientation. Therefore, they should not be construed as limitations on the present invention. Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0048] As used herein, unless otherwise expressly specified or limited, the terms "installed," "provided with," and "connected" should be understood broadly. For example, "connected" may refer to a fixed connection, a detachable connection, or an integral connection; it may refer to a mechanical connection, a direct connection, an indirect connection via an intermediate medium, or internal communication between two components. Those skilled in the art will understand the specific meanings of the above terms in the present invention on a case-by-case basis.

[0049] As used herein, "and / or" includes any and all combinations of one or more of the associated listed items.

[0050] Herein, "plurality" means two or more than two, ie, it includes two, three, four, five, etc.

[0051] As used in this specification, the term "about" typically means + / - 5% of the stated value, more typically + / - 4% of the stated value, more typically + / - 3% of the stated value, more typically + / - 2% of the stated value, even more typically + / - 1% of the stated value, and even more typically + / - 0.5% of the stated value.

[0052] In this specification, certain embodiments may be disclosed in a format that is within a range. It should be understood that this description of "within a range" is merely for convenience and brevity and should not be interpreted as a rigid limitation on the disclosed range. Therefore, the description of a range should be considered to have specifically disclosed all possible subranges and individual numerical values ​​within this range. For example, the description of a range of 1-6 should be considered to have specifically disclosed subranges such as from 1 to 3, from 1 to 4, from 1 to 5, from 2 to 4, from 2 to 6, from 3 to 6, etc., as well as individual numbers within this range, such as 1, 2, 3, 4, 5, and 6. Regardless of the breadth of the range, the above rules apply.

[0053] In this article, "steering" refers to the change in the direction of light transmission due to refraction or reflection.

[0054] See also Figure 1-Figure 5 As shown, the present invention proposes a local correction unit with a local focusing effect based on a segmented design. This local correction unit can efficiently correct signal light (e.g., amplify the signal) within a limited space. Therefore, when this local correction unit is applied to a large-scale optical computing chip (e.g., when the local correction unit is correspondingly arranged in n*n computing units), it can not only meet the correction requirements of n*n optical signals in the large-scale optical computing chip, but also reduce the impact or damage to the optical path of the original optical computing chip. Therefore, through limited local doping, the production cost required for the correction unit when applied to the optical computing chip can be reduced (e.g., the segmented arrangement reduces the doping length of the doped region, thereby reducing consumables and lowering production costs).

[0055] Example 1

[0056] The present invention provides a locally arranged low-power correction unit for an optical computing chip (see Figure 1-Figure 7 In order to facilitate the description of the positional relationship between the various components of the optical waveguide, this article will construct a three-dimensional coordinate system with the length of the optical waveguide as the X-axis, the height as the Y-axis, and the width as the Z-axis, where: Figure 2-Figure 4 The cross-section shown refers to a cross-section obtained by cutting the corresponding region of the optical waveguide along a vertical plane (i.e., the XY plane). Correspondingly, in the present invention, "length" or "thickness" is used to describe the length of a structure in the X-axis direction, and "depth" or "height" is used to describe the length of a structure in the Y-axis direction.

[0057] The present invention provides a local correction unit (or amplification unit), which is arranged on the first waveguide 1, preferably, see Figure 1As shown, the waveguide in the computing unit L (specifically, the first waveguide 1) is configured as a first waveguide segment I, a second waveguide segment II, and a third waveguide segment III. The first waveguide segment I is used to receive the input optical signal (or referred to as the first optical signal) input by the second waveguide 8 or the previous first waveguide segment, and the second waveguide segment II is used to correct the input first optical signal to output a second optical signal. The second optical signal is then transmitted again to the next first waveguide segment 1 via the third waveguide segment III. It is worth noting that the localized doping configuration of the correction unit in the present invention can effectively reduce the length of the second waveguide segment II. This, on the one hand, allows for longer first waveguide segments I and third waveguide segments III, thereby facilitating the easy placement of corresponding optical path components (such as a mode division multiplexer, a wavelength division multiplexer, or a PD detector, etc.) thereon. Furthermore, the concentrated doping can reduce the difficulty of the doping process, such as avoiding doping of waveguide bend structures or cross-connection structures, while also reducing the cost of the doping process.

[0058] See also Figure 2 As shown, the correction unit includes:

[0059] A first waveguide 1 (preferably, it refers to the second waveguide section II);

[0060] The outer side of the first waveguide 1 is wrapped with at least one cladding layer 2;

[0061] A doped region 5 is provided on the first waveguide, the doped region 5 passes through the first waveguide and intersects at least a portion of the cladding, and the doped region is doped with at least one excitation element;

[0062] Preferably, the first surface of the cladding 2 is further connected to a third waveguide 4, and a first reflective layer 3 with a second height Y2 is provided in the third waveguide at a position at a first height Y1 from the first surface;

[0063] An amplifying portion 21 is provided on the second surface of the cladding 2. The amplifying portion 21 has a surface that bulges outward along the second surface of the cladding (therefore, the amplifying portion 21 in this embodiment is also referred to as a convex lens). Correspondingly, the first reflective layer and the amplifying portion are arranged on both sides of the first waveguide 1. When at least one beam of adjustment light (or correction signal) is incident on the amplifying portion 21 along a first direction, at least one beam of the adjustment light is focused in the amplifying portion 21. Focusing refers to the ability to converge two or more beams of light at the same point or on the same optical path, or in other words, to reduce the degree of diffusion of the light source.

[0064] Preferably, the modulation light source is injected into the doped region in the form of spatial light. Specifically, when at least one beam of modulation light is injected into the doped region 5 under the guidance of the amplifying unit, penetrates the doped region 5 and refracts into the third waveguide, and when the modulation light source reaches the first reflective surface of the first reflective layer 3, the modulation light source is reflected on the first reflective surface to form at least one beam of reflected light (or so-called reflected light), and at least one beam of reflected light can be re-injected into the doped region. In other words, the first reflective layer located in the third waveguide 4 cooperates with the amplifying unit to concentrate the external modulation light source (such as pump light) to reduce the loss of the modulation light source and enhance the signal correction capability (such as the signal amplification capability) of the limited doping region. In other words, the amplifying unit in this embodiment has the function of locally concentrating the modulation light source.

[0065] The excitation element can absorb the energy of the corresponding modulating light source, causing electrons within it to transition from a low energy level to a high energy level to complete energy storage. When the first optical signal is transmitted through the first waveguide (such as the first waveguide segment I) to the doped region, the electrons that transition to the high energy level undergo stimulated radiation under the stimulation of the first optical signal, releasing photons with the same frequency, phase, and polarization state as the first optical signal. These photons are superimposed on the first optical signal to form a second optical signal, and the power of the second optical signal is greater than that of the first optical signal. Therefore, in this embodiment, correction of the optical signal by the local correction unit can be achieved by introducing spatial light.

[0066] For example, see Figure 2 As shown, at least one beam of pump light entering the doped region 5 through the amplifying portion 21 can move through the first optical path L1 and enter the third waveguide 4 through refraction, can move along the second optical path L2 in the third waveguide 4, and be reflected along the third optical path L3 through the first reflection plane, and return to the doped region through refraction again, such as returning to the doped region through the fourth optical path L4.

[0067] It is understood that the direction of light movement may vary depending on the refractive index of the material. The arrows in the accompanying drawings are used to exemplify the direction of light movement. Their designation is only to facilitate the extended description of the process of light focusing by the doped region in the present invention and does not represent a limitation on the direction of light movement.

[0068] In some embodiments, the doped region at least covers the second waveguide segment II. Preferably, the doped region 5 further covers at least a portion of the cladding. Preferably, the third length X3 of the doped region 5 is approximately 0.5 mm to 1 cm, and the fifth height Y5 is approximately 3 μm to 20 μm.

[0069] In some embodiments, see Figure 3As shown, it also includes: a second reflective layer 6 arranged on the cladding 2 and spaced apart from the first waveguide by a fourth height Y4 and having a third height Y3, so as to reserve a light source path between the amplifying part and the doped region, so that the regulating light source gathered in the amplifying part 21 can enter the doped region through the light source path.

[0070] Furthermore, in some embodiments, a certain interval, such as about 20-50 μm, is formed between the second reflective layer 6 and the doped region in the height direction (ie, the Y-axis direction).

[0071] For example, in some embodiments, a second reflective layer may be provided on the second surface of the cladding 2, and the second reflective layer 6 may be formed with an opening area at a position corresponding to the amplifying portion 21 to expose the amplifying portion 21; wherein, when at least one beam of reflected light undergoes at least one deflection, such as refraction and / or reflection, and reaches the second reflective surface of the second reflective layer, it is reflected at the second reflective surface and re-enters the doped region under the action of reflection.

[0072] That is to say, in this embodiment, the signal light in a limited space can be repeatedly collected through the third waveguide and the double reflection layer arrangement on the cladding, so as to further enhance the correction capability of the local correction unit.

[0073] In other embodiments, the second reflective layer can also be arranged on the inner side of the cladding 2, and the first waveguides 1 are spaced a certain distance apart. Correspondingly, the second reflective layer is also provided with an opening area corresponding to the amplifying part so that the light source output by the amplifying part can smoothly enter the doping region 5.

[0074] Preferably, in some embodiments, a fourth height Y4 between the second reflective layer 6 and the first waveguide 1 (i.e., the distance therebetween) is greater than 2 μm. The second reflective layer 6 may be a metal layer, preferably disposed on the surface of the cladding. This simplifies the fabrication of the correction unit and prevents absorption loss of the optical signal within the waveguide caused by the metal layer.

[0075] Preferably, in some embodiments, a fourth height Y4 between the metal layer on the surface of the cladding 2 and the first waveguide 1 is approximately 2 μm to 8 μm. The thickness of the metal layer may be between 100 nm and 300 nm, and the material of the metal layer may be one or more materials such as chromium, aluminum, and gold.

[0076] In some embodiments, see Figure 4As shown, the cladding 2 is provided with a first isolation groove 71 and a second isolation groove 72 at a first length X1 spaced apart from the amplifying portion 21, wherein a third isolation groove 73 connected to the third waveguide 3 is provided corresponding to the second isolation groove 72; wherein the isolation groove is filled with a medium, and the refractive index of the medium satisfies the following rule: 1<refractive index of medium<refractive index of cladding.

[0077] In some embodiments, the first isolation trench 71 has an opening exposed through the second reflective layer.

[0078] In some embodiments, the first length X1 is greater than about 20 μm, which can reduce or prevent the enlarged portion 21 from being affected by the etching of the isolation trench.

[0079] In this embodiment, Figure 4 As shown, the first, second, and third isolation trenches, arranged on the same side (or at the same height), form a vertical isolation layer 7, and the two isolation layers 7 cooperate to form an isolation space. Preferably, the first or second reflective layer at least covers the space between two adjacent isolation layers (that is, two isolation layers are arranged adjacent to the same doped region). The isolation layers on either side of the doped region 5 cooperate to form an isolation space, allowing the first and second reflective layers to guide and modulate the light source to converge within the isolation space.

[0080] In this embodiment, two adjacent isolation layers are separated by a second length X2. Correspondingly, the length of the first reflective layer is preferably greater than the second length X2, and the first reflective layer can respectively cover the openings of the two isolation layers at the bottom (i.e., inside the third waveguide).

[0081] Of course, in other embodiments, the first reflective layer may also be continuously disposed in the third waveguide.

[0082] In some embodiments, the thickness X4 of the isolation layer (i.e., its length in the X-axis direction) is 40um-100um, and the height of the isolation layer (multiple vertical isolation grooves form an isolation layer) (i.e., its width in the Y-axis direction) is 80um-100um.

[0083] Preferably, in this embodiment, the aspect ratio of the isolation trench (ie, the ratio of the lengths along the Y-axis to the lengths along the X-axis) is approximately 2:1.

[0084] Alternatively, in some embodiments, when the first reflective layer and the second reflective layer are provided simultaneously, the thickness X4 of the isolation layer (i.e., its length in the X-axis direction) can be set to approximately 60 μm to 150 μm, and the height of the isolation layer (one isolation layer is formed by a plurality of vertical isolation grooves) (i.e., its width in the Y-axis direction) is approximately 90 μm to 150 μm.

[0085] In some embodiments, the medium is air.

[0086] In some embodiments, the distance between the opening of the first isolation trench 71 and the enlarged portion is greater than 20 μm.

[0087] In some embodiments, the first height Y1 is 20-50 μm, and the first reflective layer 3 is a metal layer, preferably having a thickness between 100 nm and 300 nm. In this embodiment, by placing the metal layer within the third waveguide at a safe distance from the doped region 5, effective collection of the regulated light source is ensured while also avoiding the risk of mutual diffusion between the doped region and the metal layer in a confined space. However, if metal ions in the metal layer diffuse toward the doped region, this could affect the correction capability of the doped region.

[0088] In some embodiments, the metal material of the metal layer may be one or more materials such as chromium, aluminum, and gold.

[0089] In some embodiments, a fourth height Y4 between the second reflective layer 6 and the first waveguide 1 is greater than 2 μm.

[0090] In some embodiments, the height (or thickness) of the cladding layer wrapped around the outside of the first waveguide 1 is approximately in the range of 6 um to 20 um.

[0091] Alternatively, in some embodiments, the interval between the first waveguide 1 and the third waveguide 4 is approximately 3 um to 10 um.

[0092] This invention proposes a localized correction unit capable of efficiently amplifying spatial light within a limited region. This unit utilizes multiple reflection paths and isolation spaces on a waveguide to achieve sufficient signal amplification even within a limited doped region. In other words, the signal amplification achieved within this limited doped region, achieved through the interaction of reflection paths and isolation spaces, makes it easier to implement correction signals in the form of spatial light.

[0093] The present invention introduces a correction signal (such as pump light) via spatial light, thus avoiding complicating the circuit structure of the optical computing chip. In one exemplary embodiment, a laser can be used to emit pump light into the doped region in the form of spatial light. This light is then efficiently focused within a limited doped region, thereby increasing signal amplification while reducing the risk of crosstalk between the pump light and the signal light. Specifically, introducing the pump light into the doped region in the form of spatial light avoids adding new optical transmission paths (i.e., adding new waveguide structures) to the optical waveguide, thereby simplifying the array structure of the optical computing chip.

[0094] From another perspective, this local correction unit with a local focusing effect can ensure that the pump light and signal light only affect each other within a limited space, thereby reducing the possibility or degree of crosstalk between the pump light and the signal light, thereby improving the accuracy and reliability of the correction process.

[0095] Preferably, each doping region uses an independent laser to perform signal correction.

[0096] Preferably, the maximum power of the optical signal after amplification is limited to approximately 2-3 mW.

[0097] It is worth noting that the local correction unit proposed in the present invention is particularly suitable for small model reasoning scenarios, such as being applied to optical computing chips designed for small model reasoning scenarios.

[0098] For example, an optical computing chip typically includes n*n computing units. For example, an optical computing chip includes: n row waveguides and n column waveguides arranged to intersect with the n row waveguides in sequence. The intersection of the row waveguides and the column waveguides forms n*n computing units.

[0099] During the computational process, optical computing chips need to perform optical splitting on the input signal light, dividing it into multiple sub-signals for transmission to different computing units. However, as the scale of optical computing chips continues to increase (e.g., as the chip size gradually expands or the scale of optical components in its optoelectronic circuits continues to increase), light attenuation may occur during transmission, and the overall on-chip loss seriously affects the transmission performance of the entire computing system. To address this issue, the present invention introduces a local correction unit with a local concentration effect into each computing unit. This local correction unit can limit the length range of the doped region through the local concentration effect and reduce the required doping concentration, thereby reducing the impact of the large-scale introduction of correction units on the existing circuitry of the optical computing chip.

[0100] At the same time, the local focusing effect is also beneficial to reducing the interference problem caused by diffusion and other reasons between different computing units of pump light.

[0101] Specifically, the local correction unit with a local focusing effect in the present invention can be introduced in large-scale batches in large-scale optical computing chips, and the local focusing effect of the local correction unit reduces the set concentration of the doped region. From another perspective, the local focusing effect generated by the cooperation of the isolation layer and the reflective layer in the present invention is conducive to large-scale small-scale (i.e., relatively small amplification) signal amplification of signals in batches of computing units. Therefore, the present invention is particularly suitable for small-model reasoning processes in AI applications (in small-model reasoning processes, the intensity of the optical signal used in the computing unit is relatively small, such as the maximum intensity of about 2-3mW).

[0102] In some embodiments, the doping region includes: a central doping region, and an edge doping region surrounding the central doping region, and the element doping concentration in the edge doping region is lower than the doping concentration in the central doping region.

[0103] Preferably, in some embodiments, the doping concentration of the central doping region is in the range of 5.5 to 6.5×10 20 / cm 3 The doping concentration in the edge doping region ranges from 0.5 to 1×10 20 / cm 3 .

[0104] Preferably, the excitation element is a rare earth element, and the rare earth element may preferably be one or more of erbium, thulium, praseodymium, and rubidium.

[0105] Correspondingly, see Figure 1 As shown, the present invention further provides a computing chip comprising: N rows of first waveguides 1; N columns of second waveguides 8, wherein the N columns of second waveguides 8 intersect with the first waveguides 1 to form a plurality of computing units, wherein the first waveguide 1 in at least one computing unit comprises: a first waveguide segment I and a third waveguide segment III, and a correction unit as described in any embodiment of the present invention is connected between the first waveguide segment I and the third waveguide segment III; the optical computing chip further comprises: a light source device, wherein the light source device is configured to provide a regulated light source to the doped region in the correction unit, wherein the regulated light source is incident on the doped region in the form of spatial light. In some embodiments, the light source device is a laser.

[0106] See also Figure 1 As shown, pump light 011 is emitted into the doped region 5 by the laser 9 , whereby the pump light 011 excites the doped region, and the doped region can amplify the intensity of the first optical signal passing therethrough to output a second optical signal 010 .

[0107] Example 2

[0108] See also Figure 6-Figure 8 As shown, the present invention provides another exemplary correction system, which includes:

[0109] A correction unit, wherein the correction unit is provided with a first waveguide 1; the outer side of the first waveguide 1 is wrapped with at least one cladding layer 2;

[0110] Correspondingly, the correction unit includes: a doping region 5 provided on the first waveguide 1, the doping region passing through the first waveguide and intersecting at least a portion of the cladding, and the doping region is doped with at least one excitation element;

[0111] The correction system further includes a fourth waveguide 10, having at least one phase-change material layer 101 disposed thereon. The fourth waveguide 10 is connected to the doped region 5 via the first waveguide. A first constricted portion 1011 is disposed at the output end of the fourth waveguide 10, and a second constricted portion 11 corresponding to the first constricted portion 1011 is disposed on the first waveguide (for example, a waveguide with gradually varying dimensions may be extended from the first waveguide to mate with the fourth waveguide 10). Specifically, the first constricted portion and the second constricted portion at least partially overlap.

[0112] In which, the excitation element can absorb the energy of the corresponding regulating light source so that the electrons inside it transition from a low energy level to a high energy level to complete energy storage; when the first light signal is transmitted to the doped region through the first waveguide, the electrons that transition to the high energy level undergo stimulated radiation under the stimulation of the first light signal, releasing photons with the same frequency, phase and polarization state as the first light signal, and the photons are superimposed on the first light signal to form a second light signal, and the power of the second light signal is greater than the power of the first light signal.

[0113] The radius of the first contraction portion 1011 gradually decreases along the transmission direction of the signal (or light), while the radius of the second contraction portion 11 gradually increases along the transmission direction of the signal. In this embodiment, the radius refers to its width in the Z-axis direction.

[0114] For example, in some embodiments, multiple independent phase change material layers 101 may be provided, and the independent phase change material layers 101 are interconnected via a fourth waveguide, such as a first phase change material layer 101a, a second phase change material layer 101b, and a third phase change material layer 101c. By modulating different numbers of phase change material layers, the intensity (or power) of the light source can be adjusted.

[0115] Specifically, the phase change material layer is deposited on the optical waveguide (i.e., the fourth waveguide 10). The phase change material of the phase change material layer uses an electrical signal as an external stimulus to change its own state. The own state is manifested as a modification of the absorption coefficient α of the optical waveguide containing the phase change material to light. The phase change material has at least two stable states at normal pressure within the CMOS (Complementary Metal Oxide Semiconductor) operating temperature range, including a crystalline state and an amorphous state. The phase change material has a difference in absorption coefficient between its different stable states (including the crystalline state and the amorphous state).

[0116] Specifically, at least one phase change material layer can be modulated to be in a specific state, where the ratio of the phase change material in the crystalline state to the phase change material in the amorphous state in the phase change material layer is a fixed value. Thus, the intensity of the signal light (i.e., the pump light) can be precisely modulated according to the state of the phase change material layer.

[0117] In this embodiment, by providing multiple phase-change material layers, the number of phase-change material layers and their absorption coefficients can be flexibly adjusted based on the amplification requirements for the first optical signal (i.e., the required pump light intensity). For example, when the absorption coefficient α is set to a larger value, multiple phase-change material layers can be modulated simultaneously. Alternatively, when the absorption coefficient is set to a smaller value, only one or a limited number of phase-change material layers can be modulated (i.e., limited modulation). The multi-layer configuration employed in this embodiment reduces the number of necessary phase-change material layer modulations while maintaining signal modulation flexibility, further enhancing the stability and reliability of the optical computing chip's modulation function during long-term inference applications.

[0118] Preferably, the phase change material layer (or phase change material) can be formed of a chalcogenide containing antimony or selenium or contain the chalcogenide, such as antimony selenide (Sb2Se3 or SbSe), antimony sulfide (Sb2S3 or SbS), Ge2Sb2Se4Te (GSST); the phase change material may include or be composed of a compound or alloy of a combination of elements selected from GeSbTe, VOx; NbOx, GeTe, GeSb, GaSb, AgInSbTe, InSb, InSbTe, InSe, SbTe, TeGeSbS, AgSbSe, SbSe, GeSbMnSn, AgSbTe, AuSbTe and AlSb, and the phase change material may include a mixture of the above compounds; the phase change material may be a superlattice material.

[0119] The correction unit proposed in this embodiment can flexibly adjust the intensity of the light source to be suitable for different application scenarios, such as inference scenarios of different models.

[0120] It is understandable that, in the absence of conflict, the correction unit in this embodiment may include a structure that is identical or similar to that in the first embodiment, and details thereof will not be repeated here.

[0121] It is worth noting that the amplifying unit proposed in the present invention can also be applied to other semiconductor devices, such as other types of chips that need to transmit and process light.

[0122] It should be noted that, in this document, the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, article, or apparatus comprising a series of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article, or apparatus comprising the element.

[0123] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of the present invention, ordinary technicians in this field can also make many forms without departing from the scope of protection of the present invention and the claims, all of which are protected by the present invention.

Claims

1. A correction unit, characterized in that: include: First waveguide (1); The outer side of the first waveguide (1) is wrapped with at least one cladding layer (2); A doping region (5) is provided on the first waveguide, the doping region penetrates the first waveguide and intersects with at least a portion of the cladding, and the doping region is doped with at least one excitation element; The first surface of the cladding is connected to a third waveguide (4), and a first reflection layer (3) having a second height Y2 is provided in the third waveguide at a position at a first height Y1 from the first surface; An enlarged portion (21) is provided on a second surface of the cladding layer opposite to the first surface, the enlarged portion (21) having a convex surface, wherein when at least one beam of the modulated light source is incident on the enlarged portion (21) along a first direction, the at least one beam of the modulated light source is focused in the enlarged portion (21); Wherein, when at least one beam of the adjustment light source is injected into the doped region (5) under the guidance of the amplifying portion, and penetrates the doped region (5) and refracts into the third waveguide, and when the adjustment light source reaches the first reflection surface of the first reflection layer (3), the adjustment light source will be reflected on the first reflection surface, and the generated at least one beam of reflected light source can be injected into the doped region again; In which, the excitation element can absorb the energy of the corresponding light source so that the electrons inside it transition from a low energy level to a high energy level, thereby completing energy storage; when the first light signal is transmitted to the doped region through the first waveguide, the electrons that transition to the high energy level undergo stimulated radiation under the stimulation of the first light signal, releasing photons with the same frequency, phase and polarization state as the first light signal, and the photons are superimposed on the first light signal to form a second light signal, and the power of the second light signal is greater than the power of the first light signal.

2. A correction unit according to claim 1, characterized in that: Also includes: A second reflective layer (6) is provided on the second surface of the cladding (2) and is spaced apart from the first waveguide by a fourth height Y4, and the second reflective layer is provided with an opening area at a position corresponding to the amplifying portion (21) so as to reserve a light source path between the amplifying portion and the doped region, so that the regulating light source gathered in the amplifying portion (21) can enter the doped region through the light source path; wherein, when at least one beam of the reflected light source reaches the second reflective surface of the second reflective layer after being deflected, reflection occurs at the second reflective surface, and the corresponding at least one beam of the reflected light source can be re-emitted into the doped region.

3. A correction unit according to claim 2, characterized in that: The fourth height Y4 is greater than 2 μm.

4. The correction unit according to claim 1, characterized in that: The cladding is provided with a first isolation groove (71) and a second isolation groove (72) at a first length X1 spaced apart from the amplifying portion (21), wherein; a third isolation groove (73) is provided on the third waveguide (4) corresponding to the second isolation groove (72); wherein the first isolation groove (71), the second isolation groove (72), and the third isolation groove (73) provided along the same height direction form an isolation layer (7), and the isolation layer (7) is filled with a medium, and the refractive index of the medium satisfies the following rule: 1<refractive index of the medium<refractive index of the cladding.

5. A correction unit according to claim 4, characterized in that: The thickness of the insulating layer is 40um~100um, and the depth of the insulating layer is 80um~100um.

6. The correction unit according to claim 4, characterized in that: The medium is air.

7. The correction unit according to claim 4, characterized in that: The first length X1 is greater than 20 μm.

8. The correction unit according to claim 1, characterized in that: The first height Y1 is 20-50 μm.

9. The correction unit according to claim 1, characterized in that: The doping region includes: a central doping region, and an edge doping region wrapped outside the central doping region, and the element doping concentration in the edge doping region is lower than the doping concentration in the central doping region.

10. An optical computing chip based on on-chip correction, characterized in that: include: N rows of first waveguides (1); N rows of second waveguides (8), the N rows of second waveguides (8) and the first waveguide (1) are cross-arranged to form a plurality of calculation units, and the first waveguide (1) in at least one of the calculation units includes: a first waveguide segment (I) and a third waveguide segment (III), and a correction unit according to any one of claims 1 to 9 is connected between the first waveguide segment (I) and the third waveguide segment (III); Correspondingly, the optical computing chip further includes: a light source device, which is used to provide an adjustment light source to the doping region in the correction unit, wherein the adjustment light source is incident on the doping region in the form of spatial light.

Images