Multi-channel integrated photonic wavelength demultiplexer

Abstract

A multi-channel photonic demultiplexer includes: an input region receiving a multi-channel optical signal including four different wavelength channels; four output regions, each adapted to receive a corresponding one of the four different wavelength channels demultiplexed from the multi-channel optical signal; and a dispersion region optically disposed between the input region and the four output regions. The dispersion region includes a first material and a second material that are non-uniformly dispersed to form a plurality of interfaces, wherein each of the plurality of interfaces corresponds to a change in the refractive index of the dispersion region, and the dispersion region is collectively configured to optically separate each of the four different wavelength channels from the multi-channel optical signal and guide each of the four different wavelength channels to a corresponding one of the four output regions.

Description

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS

[0002] This application is based on U.S. application No. 16 / 679,579, filed on November 11, 2019, the contents of which are incorporated herein by reference in their entirety. Technical Field

[0003] The present disclosure relates generally to photonic devices and, in particular, but not exclusively, to optical multiplexers and demultiplexers. Background Art

[0004] Fiber-optic communications are commonly used to transmit information from one location to another via light that has been modulated to carry the information. For example, many telecommunications companies use optical fiber to transmit telephone signals, internet communications, and cable television signals. However, the cost of deploying optical fiber for fiber-optic communications can be prohibitive. Therefore, technologies have been developed to more efficiently use the available bandwidth within a single optical fiber. Wavelength division multiplexing is one such technology, which bundles multiple optical carrier signals onto a single optical fiber using different wavelengths. BRIEF DESCRIPTION OF THE DRAWINGS

[0005] Non-limiting and non-exhaustive embodiments of the present invention are described with reference to the following figures, wherein like reference numerals refer to like parts throughout the various views unless otherwise specified. Not all instances of an element must be labeled to avoid cluttering the drawings where appropriate. The drawings are not necessarily drawn to scale, emphasis instead being placed upon illustrating the principles being described.

[0006] Figure 1 is a functional block diagram illustrating a system for performing optical communication between two optical communication devices via optical signals according to an embodiment of the present disclosure.

[0007] Figure 2A and Figure 2B An example demultiplexer and multiplexer according to embodiments of the present disclosure are shown, respectively.

[0008] Figure 2C Example different wavelength channels of a multi-channel optical signal according to an embodiment of the present disclosure are shown.

[0009] Figures 3A-3D Different views of an example photonic demultiplexer according to embodiments of the present disclosure are shown.

[0010] Figure 4A-4B A more detailed cross-sectional view of the dispersive region of an example photonic demultiplexer according to an embodiment of the present disclosure is shown.

[0011] Figure 5is a functional block diagram illustrating a system for generating a design for a photonic integrated circuit according to an embodiment of the present disclosure.

[0012] Figure 6A An illustrative simulation environment describing a photonic integrated circuit according to an embodiment of the present disclosure is shown.

[0013] Figure 6B An example operational simulation of a photonic integrated circuit according to an embodiment of the present disclosure is shown.

[0014] Figure 6C An example adjoint simulation within a simulation environment by back-propagating loss values according to an embodiment of the present disclosure is shown.

[0015] Figure 7A is a flow chart illustrating example time steps for operating and accompanying simulations according to an embodiment of the present disclosure.

[0016] Figure 7B is a diagram illustrating the relationship between gradients determined from operational simulations and concomitant simulations according to an embodiment of the present disclosure.

[0017] Figure 8 An example method for generating a design for a photonic integrated circuit according to an embodiment of the present disclosure is shown. DETAILED DESCRIPTION

[0018] Embodiments of photonic integrated circuits including multi-channel photonic demultiplexers, as well as methods for generating designs for photonic integrated circuits, are described herein. In the following description, numerous specific details are set forth to provide a thorough understanding of the embodiments. However, one skilled in the relevant art will recognize that the techniques described herein can be practiced without one or more of these specific details, or with other methods, components, materials, and the like. In other instances, well-known structures, materials, or operations are not shown or described in detail to avoid obscuring particular aspects.

[0019] Reference throughout this specification to "one embodiment" or "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment of the present invention. Thus, the appearances of the phrases "in one embodiment" or "in an embodiment" throughout this specification are not necessarily all referring to the same embodiment. Furthermore, the particular features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.

[0020] Wavelength division multiplexing and its variants (e.g., dense wavelength division multiplexing, coarse wavelength division multiplexing, etc.) utilize the bandwidth of optical fiber by bundling multiple optical carrier signals onto a single optical fiber. Once the multiple carrier signals are bundled together, they are transmitted from one location to another through a single optical fiber, where they can be demultiplexed to be read by optical communication equipment. However, the equipment to decouple the carrier signals from each other remains prohibitive due to cost and size.

[0021] Furthermore, the design of photonic devices, such as those used for optical communications, has traditionally been performed via conventional techniques, sometimes determined by simple guess-and-check methods or manually guided grid searches, where a small number of design parameters from predetermined designs or building blocks are adjusted for suitability for a particular application. However, in reality, the design parameters of these devices can range from hundreds to billions or more, depending on the size and functionality of the device. Therefore, as the functionality of photonic devices increases and manufacturing tolerances improve to allow for smaller device feature sizes, it becomes increasingly important to fully exploit these improvements through optimized device designs.

[0022] Embodiments of photonic integrated circuits (e.g., multi-channel photonic demultiplexers and / or multiplexers) having designs that can be obtained through an inverse design process are described herein. More specifically, the techniques described in the embodiments herein utilize gradient-based optimization combined with first-principles simulations to generate designs based on an understanding of the underlying physics that is expected to govern the operation of the photonic integrated circuit. It should be understood that in other embodiments, design optimization of photonic integrated circuits without gradient-based techniques may also be used. Advantageously, the embodiments and techniques described herein are not limited to traditional techniques for photonic device design (wherein a small number of design parameters of a predetermined building block are adjusted based on suitability for a particular application). In contrast, the first-principles-based designs described herein do not necessarily rely on human intuition and can often result in designs that exceed current state-of-the-art designs in terms of performance, size, robustness, or a combination thereof. Furthermore, the embodiments and techniques described herein can provide scalable optimization of an almost unlimited number of design parameters, rather than being limited to a small number of design parameters due to traditional techniques.

[0023] Figure 1is a functional block diagram illustrating a system 100 for optical communication (e.g., via wavelength division multiplexing or other techniques) between optical communication devices 101-A and 101-B via an optical signal 110, according to an embodiment of the present disclosure. More generally, the optical communication device 101-A is configured to transmit information by modulating light from one or more light sources into a multi-channel optical signal 110 (e.g., a single optical signal including multiple different wavelength channels) that is then transmitted from the optical communication device 101-A to the optical communication device 101-B via an optical fiber, light guide, waveguide, or other photonic device. The optical communication device 101-B receives the multi-channel optical signal 110 and demultiplexes each of the multiple different wavelength channels from the multi-channel optical signal 110 to extract the transmitted information. It should be understood that in some embodiments, the optical communication devices 101-A and 101-B can be different and independent devices (e.g., optical transceivers or transmitters communicatively coupled to independent optical transceivers or receivers via one or more optical fibers). However, in other embodiments, the optical communication devices 101-A and 101-B may be part of a single component or device (e.g., a smartphone, a tablet, a computer, an optical device, etc.). For example, the optical communication devices 101-A and 101-B may both be components on a monolithic integrated circuit, coupled to each other via a waveguide embedded within the monolithic integrated circuit and adapted to carry the optical signal 110 between the optical communication devices 101-A and 101-B or otherwise transmit an optical signal between one location and another.

[0024] In the illustrated embodiment, the optical communication device 101-A includes a controller 105, one or more interface devices 107 (e.g., fiber couplers, optical guides, waveguides, etc.), a multiplexer (mux), a demultiplexer (demux), or a combination thereof 109, one or more light sources 111 (e.g., light emitting diodes, lasers, etc.), and one or more optical sensors 113 coupled to each other (e.g., photodiodes, phototransistors, photoresistors, etc.). The controller includes one or more processors 115 (e.g., one or more central processing units, dedicated circuits, field programmable gate arrays, or other) and a memory 117 (e.g., volatile memory such as DRAM and SAM, non-volatile memory such as ROM, flash memory, etc.). It should be understood that the optical communication device 101-B may include the same or similar elements as the optical communication device 101-A, which have been omitted for clarity.

[0025] The controller 105 coordinates the operation of the optical communication device 101-A for transmitting and / or receiving an optical signal 110 (e.g., a multi-channel optical signal having a plurality of different wavelength channels or other). The controller 105 includes software (e.g., instructions included in a memory 117 coupled to a processor 115) and / or hardware logic (e.g., an application specific integrated circuit, a field programmable gate array, etc.) that, when executed by the controller 105, causes the controller 105 and / or the optical communication device 101-A to perform operations.

[0026] In one embodiment, the controller 105 can orchestrate the operation of the optical communication device 101-A so that the light source 111 generates a plurality of different wavelength channels, wherein these channels are multiplexed into a multi-channel optical signal 110 via the mux / demux 109, and the multi-channel optical signal 110 is then transmitted to the optical communication device 101-B via the interface device 107. In other words, the light source 111 can output light having different wavelengths (e.g., 1271 nm, 1291 nm, 1311 nm, 1331 nm, 1511 nm, 1531 nm, 1551 nm, 1571 nm, or other wavelengths), which can be modulated or pulsed via the controller 105 to generate a plurality of different wavelength channels representing information. The plurality of different wavelength channels are then combined or otherwise multiplexed into the multi-channel optical signal 110 via the mux / demux 109, and the multi-channel optical signal 110 is transmitted to the optical communication device 101-B via the interface device 107. In the same or another embodiment, controller 105 may orchestrate the operation of optical communication device 101 -A to demultiplex multiple different wavelength channels from a multi-channel optical signal 110 received from optical communication device 101 -B via interface device 107 via mux / demux 109 .

[0027] It should be understood that in some embodiments, certain elements of the optical communication devices 101-A and / or 101-B may have been omitted to avoid obscuring certain aspects of the present disclosure. For example, the optical communication devices 101-A and 101-B may include amplification circuits, lenses, or components to facilitate sending and receiving the optical signal 110. It should also be understood that in some embodiments, the optical communication devices 101-A and / or 101-B may not necessarily include Figure 1 For example, in one embodiment, the optical communication devices 101-A and / or 101-B are passive devices that operate as intermediate devices, wherein the intermediate devices can passively multiplex a plurality of different wavelength channels into the multi-channel optical signal 110 and / or demultiplex a plurality of different wavelength channels from the multi-channel optical signal 110.

[0028] Figure 2A and Figure 2B2 and 3 show an example demultiplexer 220 and multiplexer 250 according to an embodiment of the present disclosure. Figure 1 Possible embodiments of mux / demux 109 are shown, and may be part of an integrated photonic circuit, a silicon photonic device, or other.

[0029] like Figure 2A As shown, the demultiplexer 220 includes an input region 202 and a plurality of output regions 204. The demultiplexer 220 is configured to receive the data from the input region 202 (e.g., a plurality of output regions 204). Figure 1 The waveguide of the interface device 107 shown in FIG1 receives a multi-channel optical signal 110, wherein the signal includes a plurality of different wavelength channels (e.g., Ch.1, Ch.2, Ch.3, ..., Ch.N, each having a wavelength corresponding to λ1, λ2, λ3, ..., λ N to optically separate each of the plurality of different wavelength channels from the multi-channel optical signal 110 and guide each of the plurality of different wavelength channels to a plurality of output regions 204 (e.g., which may correspond to Figure 1 More specifically, in the illustrated embodiment, each output region receives a plurality of waveguides of the interface device 107, corresponding to or otherwise representing a plurality of optical signals (e.g., λ1, λ2, λ3, ..., λ N ) outputs a portion of one of a plurality of different wavelength channels. The plurality of output regions may each be coupled to a corresponding light sensor (e.g., corresponding to Figure 1 The optical sensor 113 is shown, which can be used to convert the optical signal demultiplexed from the multi-channel optical signal 110 into an electrical signal for further processing.

[0030] exist Figure 2B In the illustrated embodiment, the multiplexer 250 includes a plurality of input regions 254 and an output region 252. The multiplexer is configured to receive a plurality of different optical signals (eg, λ1, λ2, λ3, ..., λ N ), each optical signal is located at a plurality of input regions 254 (eg, may correspond to Figure 1 The multiplexer 250 is constructed or otherwise configured to optically combine (i.e., multiplex) each of the plurality of different wavelength channels into a multi-channel optical signal 110 that is directed to an output region 252 (e.g., which may correspond to Figure 1 It should be understood that in some embodiments, Figure 2A The demultiplexer 220 and Figure 2BThe illustrated multiplexer 250 may be bi-directional such that each device may function as both a demultiplexer and a multiplexer.

[0031] Figure 2C shows a multi-channel optical signal (eg, Ch.N is Figure 1 、 Figure 2A and Figure 2B The example channels may represent individual channels included in a plurality of different wavelength channels of the multi-channel optical signal. Figure 2A The demultiplexer 220 demultiplexes and / or Figure 2B Each different wavelength channel may have a different central wavelength (λ N ), including at least one of 1271nm, 1291nm, 1311nm, 1331nm, 1511nm, 1531nm, 1551nm or 1571nm, or others. Figure 2C In the embodiment shown, the different wavelength channels have a channel bandwidth 212 that is approximately 13 nm wide. However, in other embodiments, the channel bandwidth may be different from 13 nm wide. Instead, the channel bandwidth may be considered a configurable parameter that depends on the channel bandwidth. Figure 1 mux / demux 109, Figure 2A The demultiplexer 220 and / or Figure 2B For example, in some embodiments, each of the multiplexer 250 may share a common bandwidth corresponding to 13 nm or other. Figure 2C , the channel bandwidth 212 can be defined as the width of the passband region 213 (i.e., defined as the region between PB1 and PB2). The passband region 213 can represent the approximate power transfer of the demultiplexer or multiplexer. It should be understood that in some embodiments, the passband region 213 can include, for example, Figure 2C The ripple shown corresponds to the fluctuation within the passband region 213. In one or more embodiments, the ripple within the passband region can be + / - 2 dB or less, + / - 1 dB or less, + / - 0.5 dB or less, or other. In some embodiments, the channel bandwidth 212 can be defined by the passband region 213. In other embodiments, the channel bandwidth 212 can be defined as a threshold value (e.g., dB th ) or above. For example, Figure 2A The demultiplexer 220 shown in FIG. 1 can optically separate channel N from the multi-channel optical signal 110 and has a corresponding channel bandwidth for channel N, which is equal to the channel bandwidth transmitted to the channel mapped to channel N (ie, λ N) above the threshold of the output region 204 of the photonic demultiplexer. In the same or other embodiments, the isolation of the channels (i.e., defined by the channel bandwidth 212) may also be considered when optimizing the design. Isolation may be defined as the ratio between the passband region 213 and the stopband region (e.g., a region less than SB1 and greater than SB2). It should also be understood that the transition band regions (e.g., a first transition band region between SB1 and PB1 and a second transition band region between PB2 and SB2) are exemplary and may be exaggerated for illustrative purposes. In some embodiments, optimization of the photonic demultiplexer design may also include target metrics for the slope, width, etc. of the transition band regions.

[0032] Figures 3A-3D 1 shows different views of an example photon demultiplexer 320 according to an embodiment of the present disclosure. The photon demultiplexer 320 is Figure 1 The mux / demux 109 and Figure 2A . It should also be understood that although the discussion hereafter may be directed to a photonic integrated circuit capable of demultiplexing a plurality of different wavelength channels from a multi-channel optical signal, in other embodiments, according to embodiments of the present disclosure, a demultiplexer (e.g., demultiplexer 320) may also or alternatively be capable of multiplexing a plurality of different wavelength channels into a multi-channel optical signal.

[0033] Figure 3A A cross-sectional view of a demultiplexer 320 is shown along a transverse plane within an active layer defined by a width 321 and a length 323 of the demultiplexer 320. As shown, the demultiplexer 320 includes an input region 302 (e.g., which may be equivalent to Figure 2A The input area 202 shown), a plurality of output areas 304 (e.g., which may be equivalent to Figure 2A The plurality of output regions 204 shown in FIG. 2 ) and a dispersion region optically disposed between the input region 302 and the plurality of output regions 304. The input region 302 and the plurality of output regions 304 (e.g., 304-A, 304-B, 304-C, and 304-D) can each be a waveguide (e.g., a slab waveguide, a strip waveguide, a slot waveguide, etc.) capable of propagating light along a waveguide path. The dispersion region 330 includes a first material and a second material that are non-uniformly dispersed to form a plurality of interfaces (see, e.g., FIG. 2 ). Figure 3D ), each interface corresponds to a change in the refractive index of the dispersion region 330 and together constructs the dispersion region 330 to receive a multi-channel optical signal from the multi-channel optical signal (e.g., Figure 2A The optical signal 110 shown) is optically separated into multiple different wavelength channels (e.g., Figure 2A, Ch. 1, Ch. 2, Ch. 3, ..., Ch. N) shown in the figure, and guides each of the multiple different wavelength channels to a corresponding one of the multiple output regions 304. In other words, the input region 302 is adapted to receive a multi-channel optical signal including a plurality of different wavelength channels, and the multiple output regions 304 are adapted to each receive a corresponding one of the multiple different wavelength channels demultiplexed from the multi-channel optical signal via the dispersion region 330.

[0034] like Figure 3A shown, and in Figure 3D and Figures 4A-4B As more clearly shown in FIG, the shape and arrangement of the non-uniformly dispersed first and second materials create a plurality of interfaces that collectively form a material interface pattern along the cross-sectional area of the dispersion region 330, which is at least partially surrounded by a peripheral boundary region 322 comprising the second material. In some embodiments, the peripheral region 322 has a substantially uniform composition comprising the second material. In the illustrated embodiment, the dispersion region 330 includes a first side 331 and a second side 333, each of which is contiguous with an inner boundary (i.e., an unlabeled dashed line of the peripheral region 322 disposed between the dispersion region 330 and a dashed-dotted line corresponding to the outer boundary of the peripheral region 322). The first side 331 and the second side 333 are disposed corresponding to opposite sides of the dispersion region 330. The input region 302 is disposed proximate the first side 331 (e.g., one side of the input region 302 abuts the first side 331 of the dispersion region 330), while each of the plurality of output regions 304 is disposed proximate the second side 333 (e.g., one side of each of the plurality of output regions 304 abuts the second side 333 of the dispersion region 330).

[0035] In the illustrated embodiment, each of the plurality of output regions 304 is parallel to every other of the plurality of output regions 304. However, in other embodiments, the plurality of output regions 304 may not be parallel to each other or may even be disposed on the same side (e.g., one or more of the plurality of output regions 304 and / or input region 302 may be disposed proximate to a side of the dispersion region 330 that is adjacent to the first side 331 and / or the second side 333). In some embodiments, when the plurality of output regions includes at least three output regions, adjacent output regions in the plurality of output regions are separated from each other by a common separation distance. For example, as shown, adjacent output regions 304-A and 304-B are separated from each other by a distance 316, where the distance 316 may be the same as the separation distance between other pairs of adjacent output regions.

[0036] like Figure 3AAs shown in the embodiment of FIG. 3 , the demultiplexer 320 includes four output regions 304 (e.g., 304-A, 304-B, 304-C, and 304-D), each of which is mapped (i.e., by virtue of the structure of the dispersion region 330) to a corresponding one of four channels included in a plurality of different wavelength channels. More specifically, a plurality of interfaces of the dispersion region 330 defined by the non-uniform dispersion of the first material and the second material are arranged along a cross-sectional area of the dispersion region 330 (e.g., as shown in FIG. 3 ). Figure 3A 、 Figure 4A or Figure 4B As shown in FIG, a material interface pattern is formed so that the dispersion region 330 optically separates each of the four channels from the multi-channel optical signal and routes each of the four channels to a corresponding one of the four output regions 304 when the multi-channel optical signal is input to the input region 302.

[0037] Note that the first material and the second material of the dispersion region 330 are arranged and shaped within the dispersion region so that the material interface pattern is substantially proportional to a design obtainable using an inverse design process, which will be discussed in more detail later in this disclosure. More specifically, in some embodiments, the inverse design process can include iterative gradient-based optimization of the design based at least in part on a loss function, wherein the loss function comprises performance losses (e.g., enforcing functionality) and manufacturing losses (e.g., enforcing manufacturability and binarization of the first and second materials), wherein the manufacturing losses are reduced or otherwise adjusted via the iterative gradient-based optimization to generate the design. In the same or other embodiments, other optimization techniques can be used in place of or in combination with gradient-based optimization. Advantageously, this allows for the optimization of a virtually unlimited number of design parameters to achieve functionality and performance within a predetermined region that would not be possible using traditional design techniques.

[0038] For example, in one embodiment, dispersion region 330 is configured to optically separate each of the four channels from the multi-channel optical signal within a predetermined 35 μm x 35 μm region (e.g., as defined by width 325 and length 327 of dispersion region 330) when the multi-channel optical signal is received by input region 302. In the same or another embodiment, the dispersion region is configured to accommodate a common bandwidth for each of the four channels, each of the four channels having a different central wavelength. In one embodiment, the common bandwidth is approximately 13 nm wide, and the different central wavelengths are selected from the group consisting of 1271 nm, 1291 nm, 1311 nm, 1331 nm, 1511 nm, 1531 nm, 1551 nm, and 1571 nm. In some embodiments, the entire structure of demultiplexer 320 (e.g., including input region 321, peripheral region 322, dispersion region 330, and multiple output regions 304) fits within the predetermined region (e.g., as defined by width 321 and length 323). In one embodiment, the predetermined area is 35 μm×35 μm. It should be understood that in other embodiments, the dispersion region 330 and / or the demultiplexer 320 can be adapted to fit within other areas that are larger or smaller than 35 μm×35 μm, which can result in changes to the structure of the dispersion region 330 (e.g., the arrangement and shape of the first and second materials) and / or other components of the demultiplexer 320.

[0039] In the same or other embodiments, for a given wavelength within one of the plurality of different wavelength channels, the dispersion region is configured to have a power transmission of -2 dB or greater from input region 302 through dispersion region 330 to a corresponding one of the plurality of output regions 304. For example, if channel 1 of a multi-channel optical signal is mapped to output region 304-A, when demultiplexer 320 receives the multi-channel optical signal at input region 302, dispersion region 330 will optically separate channel 1 from the multi-channel optical signal and direct a portion of the multi-channel optical signal corresponding to channel 1 to output region 304-A at a power transmission of -2 dB or greater. In the same or other embodiments, dispersion region 330 is configured such that for a given wavelength, the adverse power transmission (i.e., isolation) from the input region to any one of the plurality of output regions other than the corresponding one of the plurality of output regions is -30 dB or less, -22 dB or less, or other. For example, if channel 1 of a multi-channel optical signal is mapped to output region 304-A, the unfavorable power transfer from input region 302 to any of the output regions (e.g., 304-B, 304-C, 304-D) other than the corresponding one of the plurality of output regions (e.g., 304-A) is -30 dB or less, -22 dB or less, or other. In some embodiments, the maximum power reflection of an input signal (e.g., a multi-channel optical signal) received at an input region (e.g., input region 302) from demultiplexer 320 and reflected back to the input region by dispersion region 330 is either -40 dB or less, -20 dB or less, -8 dB or less, or other. It should be understood that in other embodiments, the power transfer, unfavorable power transfer, maximum power, or other performance characteristics may vary from the values discussed herein, but the structure of dispersion region 330 may vary due to the inherent relationship between the structure, function, and performance of demultiplexer 320.

[0040] Figure 3B A vertical schematic diagram or stacking of the various layers included in the illustrated embodiment of the demultiplexer 320 is shown. However, it should be understood that the illustrated embodiment is not exhaustive and that certain features or elements may be omitted to avoid obscuring certain aspects of the invention. In the illustrated embodiment, the demultiplexer 320 includes 312, a dielectric layer 314, an active layer 306 (e.g., Figure 3A ) and cladding layer 308. In some embodiments, demultiplexer 320 may be partially or otherwise a photonic integrated circuit or silicon photonic device compatible with conventional manufacturing techniques (e.g., lithography such as photolithography, electron beam lithography, etc., sputtering, thermal evaporation, physical and chemical vapor deposition, etc.).

[0041] In one embodiment, a silicon on insulator (SOI) wafer may initially be provided, comprising a support substrate (e.g., a silicon substrate) corresponding to substrate 312, a silicon dioxide dielectric layer corresponding to dielectric layer 314, a silicon layer (e.g., intrinsic, doped, or other), and an oxide layer (e.g., intrinsic, grown, or other). In one embodiment, the silicon in active layer 306 may be selectively etched by photolithographically creating a pattern on the SOI wafer, which pattern is transferred to the SOI wafer via a dry etching process (e.g., through a photoresist mask or other hard mask) to remove a portion of the silicon. The silicon may be etched all the way to dielectric layer 314 to form a void, which may then be backfilled with silicon dioxide and subsequently encapsulated with silicon dioxide to form cladding layer 308. In one embodiment, several etch depths may be achieved, including a full etch depth of silicon, to obtain the target structure. In one embodiment, the silicon may be 220 nm thick, so the full etch depth may be 220 nm. In some embodiments, this can be a two-step packaging process where two silicon dioxide depositions are performed with an intermediate chemical mechanical planarization to produce a planar surface.

[0042] Figure 3C Shown along with Figure 3A The active layer 306 (relative to the active layer 306) is cut out of a portion of the peripheral region 322 of the input region 302 Figure 3B ) is a more detailed view of FIG. In the embodiment shown, the active layer 306 includes a first material 332 having a refractive index of ε1 and a second material 334 having a refractive index of ε2 that is different from ε1. Figure 3A and Figure 3C As shown, homogenous regions of first material 332 and second material 334 may form waveguides or waveguide portions corresponding to input region 302 and plurality of output regions 304 .

[0043] Figure 3D The active layer 306 is shown cut along the dispersion region 330 (relative to Figure 3B ) is a more detailed view of the active layer 306. As previously described, the active layer 306 includes a first material 332 (e.g., silicon) and a second material 334 (e.g., silicon dioxide) that are non-uniformly dispersed to form a plurality of interfaces 336 that collectively form a material interface pattern. Each of the plurality of interfaces 336 forming the interface pattern corresponds to a change in the refractive index of the dispersion region 330, so as to configure the dispersion region (i.e., the shape and arrangement of the first material 332 and the second material 334) to at least partially provide the functionality of the demultiplexer 320 (i.e., when the input region 302 receives the multi-channel optical signal, the optical separation of the plurality of different wavelength channels from the multi-channel optical signal and the corresponding guidance of each of the plurality of different wavelength channels to a corresponding one of the plurality of output regions 304).

[0044] It should be understood that in Figures 3A-3D In the illustrated embodiment of the demultiplexer 320, the change in refractive index is shown as being vertically uniform (i.e., the first material 332 and the second material 334 form an interface that is substantially perpendicular or perpendicular to a transverse plane or cross-section of the demultiplexer 320). However, in the same or other embodiments, multiple interfaces (e.g., Figure 3D The interface 336 shown in FIG. 3 may not be substantially perpendicular to a lateral plane or cross-section of the demultiplexer 320 .

[0045] Figure 4A A more detailed cross-sectional view of the dispersion region of an example photonic demultiplexer 420 is shown in accordance with an embodiment of the present disclosure. Figure 4B Shown by the Figure 4A A more detailed view of the interface pattern formed by the shape and arrangement of the first material 432 and the second material 434 of the dispersion region of the photon demultiplexer 420. The demultiplexer 420 is Figure 1 The mux / demux109 shown, Figure 2A The demultiplexer 220 and Figures 3A-3D One possible implementation of the demultiplexer 320 is shown.

[0046] like Figure 4A-4B As shown, the demultiplexer 420 includes an input region 402, a plurality of output regions 404, and a dispersion region 430 optically disposed between the input region 402 and the plurality of output regions 404. The dispersion region 430 is at least partially surrounded by a peripheral region 422, wherein the peripheral region 422 includes an inner boundary 436 and an outer boundary 438. It should be understood that similarly named or labeled elements of the demultiplexer 420 may similarly correspond to similarly named or labeled elements of other demultiplexers described in embodiments of the present disclosure.

[0047] The first material 432 (i.e., the black region within the dispersion region 430) and the second material 434 (i.e., the white region within the dispersion region 430) of the photon demultiplexer 420 are non-uniformly dispersed to produce a plurality of interfaces that collectively form a material interface pattern 431, as shown in FIG. Figure 4BMore specifically, the inverse design process utilizes iterative gradient-based optimization, Markov chain Monte Carlo optimization, or other optimization techniques combined with first-principles simulations to generate a design that is essentially replicated by dispersion region 430 in a proportional or scaled manner, such that photonic demultiplexer 420 provides the desired functionality. In the illustrated embodiment, dispersion region 430 is configured to optically separate each of a plurality of different wavelength channels from a multi-channel optical signal and, upon receipt of the multi-channel optical signal by input region 402, direct each of the plurality of different wavelength channels to a corresponding one of a plurality of output regions 404. More specifically, a plurality of output regions 404-A, 404-B, 404-C, and 404-D are mapped to wavelength channels having center wavelengths corresponding to 1271 nm, 1291 nm, 1311 nm, and 1331 nm, respectively. In another embodiment, output regions 404-A, 404-B, 404-C, and 404-D are mapped to wavelength channels having center wavelengths corresponding to 1511 nm, 1531 nm, 1551 nm, and 1571 nm, respectively.

[0048] like Figure 4B As shown, the material interface pattern 431 defined by the black lines within the dispersion region 430 and corresponding to the change in refractive index within the dispersion region 430 includes a plurality of protrusions. The first protrusion 442-A is formed of the first material 432 and extends from the peripheral region 422 into the dispersion region 430. Similarly, the second protrusion 442-B is formed of the second material 434 and extends from the peripheral region 422 into the dispersion region 430. Figure 4B As further shown in FIG, the dispersion region 430 includes a plurality of islands formed of the first material 432 or the second material 434. The plurality of islands includes a first island 444-A formed of the first material 432 and surrounded by the second material 434. The plurality of islands 442 also includes a second island 444-B formed of the second material 434 and surrounded by the first material 434.

[0049] In some embodiments, the material interface pattern 431 includes one or more dendritic shapes, wherein each of the one or more dendritic shapes is defined as a branched structure formed from the first material 432 or the second material 434 and having a width that alternates between increasing and decreasing in size along a corresponding direction. Figure 4AFor clarity, dendritic structures 446 are marked with white arrows with black borders. It can be seen that the width of dendritic structures 446 alternately increases and decreases in size along corresponding directions (i.e., the white-marked arrows cover the length of dendritic structures 446) to produce a branched structure. It should be understood that in other embodiments, there may be no protrusions, no islands, no dendritic structures, or any number (including zero) of protrusions, islands, dendritic structures, or combinations thereof of any material included in dispersion region 430.

[0050] In some embodiments, the reverse design process includes implementing a minimum feature size manufacturing penalty, for example, to ensure the manufacturability of the design. Figure 4A and Figure 4B In the illustrated embodiment of the photon demultiplexer 420, the interface pattern 431 is shaped to implement a minimum feature size within the dispersion region 430 such that the plurality of interfaces within the cross-sectional area formed by the first material 432 and the second material 434 do not have a radius of curvature less than a threshold size. For example, if the minimum feature size is 150 nm, then the radius of curvature of any of the plurality of interfaces has a size less than the threshold size, which corresponds to the inverse of half the minimum feature size (i.e., 1 / 75 nm). -1 ). By taking into account manufacturing constraints, limitations, and / or yields, the implementation of such a minimum feature size prevents the reverse design process from producing a non-manufacturable design. In the same or other embodiments, a minimum width or spacing can be implemented as the minimum feature size using different or additional checks on manufacturability-related metrics.

[0051] Figure 5 is a functional block diagram illustrating a system 500 for generating a design for a photonic integrated circuit (i.e., a photonic device) according to an embodiment of the present disclosure. The system 500 can be used to perform an inverse design process that generates a design with iterative gradient-based optimization that takes into account the underlying physics that governs the operation of the photonic integrated circuit. More specifically, the system 500 is a design tool that can be used to optimize structural parameters of the photonic integrated circuit (e.g., the shape and arrangement of the first material and the second material within the dispersion region of the embodiments described in the present disclosure) based on first principles simulations (e.g., electromagnetic simulations to determine the field response of the photonic device to an excitation source) and iterative gradient-based optimization. In other words, the system 500 can provide a design obtained via the inverse design process that is respectively Figure 3A and 4A The dispersion regions 330, 430, and 472 of the demultiplexers 320 and 420 shown in FIG are substantially replicated (ie, scaled).

[0052] As shown, system 500 includes a controller 505, a display 507, input device(s) 509, communication device(s) 511, a network 513, a remote resource 515, a bus 521, and a bus 523. Controller 505 includes a processor 531, a memory 533, a local storage device 535, and a photonic device simulator 539. Photonic device simulator 539 includes an operation simulation engine 541, manufacturing loss calculation logic 543, calculation logic 545, a companion simulation engine 547, and an optimization engine 549. It should be understood that in some embodiments, controller 505 may be a distributed system.

[0053] Controller 505 is coupled to display 507 (e.g., a light-emitting diode display, a liquid crystal display, etc.), which is coupled to bus 521 via bus 523 for displaying information to a user using system 500 to optimize the structural parameters of the photonic device (i.e., the demultiplexer). Input device 509 is coupled to bus 521 via bus 523 for communicating information and command selections to processor 531. Input device 509 may include a mouse, trackball, keyboard, stylus, or other computer peripheral device to facilitate interaction between the user and controller 505. In response, controller 505 may provide verification of the interaction via display 507.

[0054] Another device optionally coupled to the controller 505 is a communication device 511 for accessing remote resources 515 of the distributed system via a network 513. The communication device 511 may include any number of network peripheral devices, such as devices for coupling to an Ethernet network, the Internet, or a wide area network. The communication device 511 may further include a mechanism for providing a connection between the controller 505 and the outside world. Note that Figure 5 Any or all components of the illustrated system 500 and associated hardware may be used in various embodiments of the present disclosure. Remote resources 515 may be part of a distributed system and include any number of processors, memory, and other resources for optimizing structural parameters of a photonic device.

[0055] Controller 505 coordinates the operation of system 500 to optimize the structural parameters of photonic devices. Processor 531 (e.g., one or more central processing units, graphics processing units, and / or tensor processing units, etc.), memory 533 (e.g., volatile memory such as DRAM and SRAM, non-volatile memory such as ROM, flash memory, etc.), local storage 535 (e.g., magnetic storage such as a computer disk drive), and photonic device simulator 539 are coupled to each other via bus 523. Controller 505 includes software (e.g., instructions included in memory 533 coupled to processor 531) and / or hardware logic (e.g., an application-specific integrated circuit, a field-programmable gate array, etc.) that, when executed by controller 505, causes controller 505 or system 500 to perform operations. Operations can be based on instructions stored in any one or a combination of memory 533, local storage 535, photonic device simulator 539, and remote resources 515 accessed via network 513.

[0056] In the illustrated embodiment, modules 541-549 of the photonic device simulator 539 are used to optimize photonic devices (e.g., Figure 1 mux / demux 109, Figure 2A Demultiplexer 220, Figure 2B Demultiplexer 250, Figures 3A-3D The demultiplexer 320 and Figure 4A-4BIn some embodiments, the system 500 can optimize the structural parameters of the photonic device by simulations (e.g., operational simulations and adjoint simulations) that model field responses (e.g., electric and magnetic fields within the photonic device), particularly using finite-difference time-domain (FDTD) methods. The operational simulation engine 541 provides instructions for performing electromagnetic simulations of the photonic device operating in response to an excitation source within a simulation environment. Specifically, the operational simulation determines the field response of the simulation environment (and therefore the photonic device described by the simulation environment) in response to the excitation source to determine performance metrics of the physical device (e.g., an initial description or input design of the photonic device based on a description of the structural parameters of the photonic device within the simulation environment using a plurality of voxels). The structural parameters can correspond to, for example, a specific design, material composition, size, etc. of the physical device. The manufacturing loss calculation logic 543 provides instructions for determining manufacturing losses that are used to enforce minimum feature sizes to ensure manufacturability. In some embodiments, manufacturing losses are also used to implement binarization of the design (i.e., such that the photonic device includes a first material and a second material that are interspersed to form multiple interfaces). Calculation logic 545 calculates a loss metric determined by a loss function that includes a performance loss based on the performance metric and a manufacturing loss. An adjoint simulation engine 547 is used in conjunction with the operational simulation engine 541 to perform an adjoint simulation of the photonic device, thereby backpropagating the loss metric through the simulation environment via the loss function to determine how changes to the structural parameters of the photonic device affect the loss metric. An optimization engine 549 is used to update the structural parameters of the photonic device to reduce the loss metric and generate a revised description of the photonic device (i.e., a revised design).

[0057] Figures 6A-6C 6. The initial setup of a simulation environment 601-A describing a photonic device, the operational simulation of the photonic device in response to an excitation source performed within the simulation environment 601-B, and the adjoint simulation of the photonic device performed within the simulation environment 601-C are shown, respectively. The initial setup of the simulation environment 201, the 1-dimensional representation of the simulation environment 201, the operational simulation of the physical device, and the adjoint simulation of the physical device can be shown as follows: Figure 1 The system 100 shown in FIG. Figures 6A-6C As shown, the simulation environment 601 is represented in two dimensions. However, it should be understood that other dimensions (e.g., 3-dimensional space) can also be used to describe the simulation environment 601 and the photonic device. In some embodiments, Figures 6A-6C The optimization of the structural parameters of the photonic device shown in can be achieved through an inverse design process, wherein the inverse design process particularly includes simulations (e.g., operational simulations and adjoint simulations) of the field response (e.g., electric and magnetic fields) to the excitation source using a finite-difference time-domain (FDTD) method.

[0058] Figure 6A An illustrative simulation environment 601-A for describing a photonic integrated circuit (i.e., a photonic device such as a waveguide, a demultiplexer, etc.) according to an embodiment of the present disclosure is shown. More specifically, in response to receiving an initial description of a photonic device defined by one or more structural parameters (e.g., an input design), a system (e.g., Figure 5 The system 500 of FIG. 5 ) configures the simulation environment 601 to represent a photonic device. As shown, the simulation environment 601 (and subsequently the photonic device) is described by a plurality of voxels 610 representing individual elements (i.e., discretized) of a two-dimensional (or other dimensional) space. Each voxel is shown as a two-dimensional square; however, it should be understood that the voxel can be represented as a cube or other shape in a three-dimensional space. It should be understood that the specific shape and dimensions of the plurality of voxels 610 can be adjusted according to the simulation environment 601 and the photonic device being simulated. It should also be noted that only a portion of the plurality of voxels 610 is shown to avoid obscuring other aspects of the simulation environment 601.

[0059] Each of the plurality of voxels 610 may be associated with a structure value, a field value, and a source value. Collectively, the structure values of the simulation environment 601 describe structural parameters of the photonic device. In one embodiment, the structure values may correspond to structural (i.e., material) boundaries or interfaces (e.g., Figure 4B The relative permittivity, permeability, and / or refractive index of the interface pattern 431 of the simulation environment 601 are represented. For example, the interface 636 represents a location where the relative permittivity changes in the simulation environment 601 and can define a boundary of the photonic device where a first material meets or otherwise interfaces with a second material. The field value describes a field (or loss) response calculated (e.g., via Maxwell's equations) in response to an excitation source described by the source value. For example, the field response can correspond to a vector describing the electric and / or magnetic field (e.g., in one or more orthogonal directions) of each of the plurality of voxels 610 at a particular time step. Thus, the field response can be based at least in part on structural parameters of the photonic device and the excitation source.

[0060] In the illustrated embodiment, the photonic device corresponds to a device having a design region 630 (eg, corresponding to Figure 3A The dispersion region 330, and / or Figure 4A In the embodiment of the present invention, an optical demultiplexer (e.g., a dispersion region 430) is provided, wherein the structural parameters of the physical device can be updated or otherwise revised. More specifically, through an inverse design process, an iterative gradient-based optimization of the loss metric determined from the loss function is performed to generate a design of the photonic device, wherein the design functionally enables the multi-channel optical signal to be demultiplexed and directed from the input port 602 to a corresponding one of the output ports 604-A and 604-B. Thus, the input port 602 of the photonic device (e.g., corresponding to Figure 3AInput area 302, Figure 4A The input region 402 of the design environment 601, etc., corresponds to the location of an excitation source that provides an output (e.g., a Gaussian pulse, a wave, a waveguide mode response, etc.). The output of the excitation source interacts with the photonic device based on the structural parameters (e.g., as the wave propagates through the photonic device within the simulation environment 601, the electromagnetic wave corresponding to the excitation source may be perturbed, retransmitted, attenuated, refracted, reflected, diffracted, scattered, absorbed, dispersed, amplified, or otherwise). In other words, the excitation source can cause the field response of the photonic device to change, depending on the underlying physics governing the physical domain and the structural parameters of the photonic device. The excitation source originates from or is otherwise proximate to the input port 602 and is positioned to propagate through the design region 630 toward the output ports 604-A and 604-B of the photonic device (or otherwise affect the field values of multiple voxels). In the illustrated embodiment, the input port 602 and the output ports 604-A and 604-B are positioned outside the design region 630. In other words, in the illustrated embodiment, only a portion of the structural parameters of the photonic device are optimizable.

[0061] However, in other embodiments, the entirety of the photonic device can be placed within the design region 630 such that the structural parameters can represent any portion or the entirety of the photonic device design. The electric and magnetic fields within the simulation environment 601 (and subsequently the photonic device) can change in response to the excitation source (e.g., represented by field values at individual voxels that collectively correspond to the field response of the simulation environment). The output ports 604-A and 604-B of the optical demultiplexer can be used to determine performance metrics of the photonic device in response to the excitation source (e.g., power transfer from the input port 602 to a particular one of the output ports 604-A and 604-B). An initial description of the photonic device, including initial structural parameters, excitation source, performance parameters or metrics, and other parameters describing the photonic device, is provided by a system (e.g., Figure 5 The system 500) is received and used to configure the simulation environment 601 to perform first-principles simulations of photonic devices. These specific values and parameters can be determined by a user (e.g., Figure 5 directly defined by a user of the system 500 in FIG. 5 , indirectly defined (eg, via the controller 505 selecting a predetermined value stored in the memory 533, the local storage device 535, or the remote resource 515), or a combination thereof.

[0062] Figure 6BAn operational simulation of a photonic device in response to an excitation source within a simulation environment 601-B according to an embodiment of the present disclosure is shown. In the illustrated embodiment, the photonic device is an optical demultiplexer structure that optically separates each of a plurality of different wavelength channels included in a multi-channel optical signal received at an input port 602 and directs each of the plurality of different wavelength channels to a corresponding one of a plurality of output ports 604. The excitation source can be selected (randomly or otherwise) from the plurality of different wavelength channels and originates from the input port 602 having a specified spatial, phase, and / or temporal profile. The operational simulation occurs over a plurality of time steps, including the illustrated time step. While the operational simulation is executed, changes in the field response (e.g., field values) of each of a plurality of voxels 610 in response to the excitation source over the plurality of time steps are incrementally updated. The change in the field response at a particular time step is based at least in part on the structural parameters, the excitation source, and the field response of the simulation environment 601 at the immediately preceding time step included in the plurality of time steps. Similarly, in some embodiments, the source values of the plurality of voxels 610 are updated (e.g., based on spatial and / or temporal properties describing the excitation source). It should be understood that the operational simulation is incremental, and that during the operational simulation, for each of a plurality of time steps, as time moves forward, the field values (and source values) of the simulation environment 601 are incrementally updated at each time step. It should also be noted that in some embodiments, the updating is an iterative process, and that the update of each field and source value is based at least in part on the previous update of each field and source value.

[0063] Once the operational simulation reaches a steady state (e.g., substantially stabilizes or decreases to a negligible value in response to changes in the field value of the excitation source) or otherwise ends, one or more performance metrics can be determined. In one embodiment, the performance metric corresponds to the power transfer at a corresponding one of the output ports 604 mapped to the different wavelength channels simulated by the excitation source. In other words, in some embodiments, the performance metric represents the power in the target mode shape (at one or more frequencies of interest) at a particular location of the output port 604. A loss value or metric for an input design (e.g., an initial design and / or any refined designs in which structural parameters have been updated) based at least in part on the performance metric can be determined by a loss function. The loss metric, in conjunction with the accompanying simulation, can be used to determine structural gradients (e.g., the effect of structural parameters on the loss metric) to update or otherwise revise the structural parameters to reduce the loss metric (i.e., improve the performance metric). Note that the loss metric is also based on a manufacturing loss value for a minimum feature size used to implement the photonic device to improve the manufacturability of the device.

[0064] Figure 6CAn example adjoint simulation within a simulation environment 601-C by backpropagating a loss metric according to an embodiment of the present disclosure is shown. More specifically, the adjoint simulation is a time-reverse simulation in which the loss metric is treated as an excitation source that interacts with the photonic device and induces a loss response. In other words, an adjoint (or virtual source) based on the loss metric is placed at an output region (e.g., output port 604) or other location corresponding to the location used in determining the performance metric. During the adjoint simulation, the adjoint source(s) are treated as physical stimuli or excitation sources. In response to the adjoint source(s), a loss response of the simulation environment 601 is calculated for each of a plurality of time steps (e.g., backward in time). The loss response collectively refers to the loss values of a plurality of voxels that are incrementally updated over the plurality of time steps in response to the adjoint source(s). Changes in the loss response based on the loss metric can correspond to loss gradients, where the loss gradients indicate how changes in the field response of the physical device affect the loss metric. The loss gradients and field gradients can be combined in an appropriate manner to determine the structural gradients of the photonic device / simulation environment (e.g., how changes in the structural parameters of the photonic device in the simulation environment affect the loss metric). Once the structural gradients for a particular cycle (e.g., operation and accompanying simulations) are known, the structural parameters can be updated to reduce the loss metric and generate a revised description or design of the photonic device.

[0065] In some embodiments, as part of an inverse design process utilizing iterative gradient-based optimization, an iterative loop of performing operational simulations and adjoint simulations, determining structural gradients, and updating structural parameters to reduce a loss metric is continuously performed. An optimization scheme such as gradient descent can be utilized to determine a specific amount or degree of change in the structural parameters of the photonic device to incrementally reduce the loss metric. More specifically, after each loop, the structural parameters are updated (e.g., optimized) to reduce the loss metric. The operational simulations, adjoint simulations, and updating of the structural parameters are iteratively repeated until the loss metric substantially converges or is otherwise below or within a threshold or range such that the photonic device provides desired performance while maintaining manufacturability.

[0066] Figure 7A 7 is a flowchart 700 illustrating example time steps of an operational simulation 710 and a companion simulation 750 according to an embodiment of the present disclosure. Flowchart 700 is a diagram of a system (e.g., Figure 5 The system 500 of FIG. 5 can be used to execute simulation environments (e.g., Figures 6A-6C The operational simulation 710 and the adjoint simulation 750 of the simulation environment 601 of FIG. 1 are one possible implementation of the operational simulation 710 and the adjoint simulation 750 of FIG. 1 . In the illustrated embodiment, the operational simulation utilizes a finite-difference time-domain (FDTD) method to model a plurality of voxels (e.g., Figures 6A-6CThe field response (both electric and magnetic) or loss response at each of the plurality of voxels 610 shown in FIG.

[0067] like Figure 7A As shown, flowchart 700 includes update operations for an operational simulation 710 and a portion of an accompanying simulation 750. The operational simulation 710 occurs over a plurality of time steps (e.g., from an initial time step through a predetermined or conditional number of time steps having a specified time step size to a final time step) and models changes in electric and magnetic fields (e.g., from initial field values 711) for a plurality of voxels describing a simulation environment and / or photonic device that collectively correspond to field responses. More specifically, update operations (e.g., 712, 714, and 716) are iterative and based on the field responses, structural parameters 704, and one or more excitation sources 708. Each update operation is followed by another update operation that represents a sequential step forward in time over the plurality of time steps. For example, update operation 714 updates field values 713 based on the field responses determined from the previous update operation 712, sources 708, and structural parameters 704 (see, e.g., Figure 7B Similarly, update operation 716 updates field value 715 based on the field response determined from update operation 714 (see, e.g., Figure 7B ). In other words, at each time step of the operational simulation, the field values (and therefore the field response) are updated based on the previous field response and structural parameters of the photonic device. Once the final time step of the operational simulation 710 is performed, the loss metric 718 can be determined (e.g., based on a predetermined loss function 720). The loss gradient determined from block 752 can be treated as an adjoint or virtual source (e.g., a physical stimulus or excitation source originating from an output region or port) that is inversely back-propagated (incrementally from the final time step through multiple time steps until the initial time step is reached) to determine the structural gradient 768.

[0068] In the illustrated embodiment, the FDTD solve (e.g., operation simulation 710) and inverse solve (e.g., adjoint simulation 750) problems are described at a high level using only "update" and "loss" operations and their corresponding gradient operations. A simulation is initially set up where the structural parameters of the simulation environment (and photonic device), the physical stimulus (i.e., the excitation source), and the initial field state are provided (e.g., via an initial description and / or input design). As previously described, the field values are updated in response to the excitation source based on the structural parameters. More specifically, the update operation is given by φ, where for Here, n corresponds to the total number of time steps (e.g., a number of time steps) of operation simulation, where x i corresponds to the field response of the simulated environment at time step i (the field values associated with the electric and magnetic fields for each of the plurality of voxels), corresponds to the (multiple) excitation sources (source values associated with the electric and magnetic fields for each of the multiple voxels) of the simulated environment at time step i, while z corresponds to structural parameters describing the topology and / or material properties of the physical device (e.g., relative permittivity, refractive index, etc.).

[0069] Note that using the FDTD method, the update operation can be specifically expressed as:

[0070]

[0071] That is, the FDTD update is linear with respect to the field and source terms. Specifically, and are linear operators that depend on the structural parameter z and act on the field x i Heyuan Here, assuming Where N is the number of FDTD field components in the operational simulation. In addition, the loss operation (e.g., the loss function) can be expressed as L = f(x i ,…,x n ) which takes the computed field as input and produces a single real-valued scalar (e.g., a loss metric) that can be reduced and / or minimized.

[0072] In terms of revising or otherwise optimizing the structural parameters of a physical device, the relevant quantity to be produced is It is used to describe the effect of changes in structural parameters on the loss value and is expressed as Figure 7A The structural gradient 768 is shown in .

[0073] Figure 7B is a diagram 780 illustrating the relationship between an operation simulation and an update operation accompanying the simulation (eg, backpropagation), according to an embodiment of the present disclosure. More specifically, Figure 7B Summarizes the calculation of structural gradient The operations and accompanying simulation relationships involved include and The update operation 714 of the operation simulation converts the field value 713 of the plurality of voxels at the i-th time step into i Update to the next time step (i.e., time step i+1), which corresponds to the field value 715x i+1 The gradient 755 is used to determine the gradient for backpropagation (e.g., update operation 356 backward in time). which is used at least in part in conjunction with gradient 769 to calculate the structural gradient is the contribution of each field to the loss metric L. Note that this is a partial derivative and therefore does not take into account x i →x i+1 Therefore, the use of It contains x i →x i+1 Relationship. Loss gradient Can also be used to calculate structural gradients and corresponds to the total derivative of the field with respect to the loss value L. The gradient of the loss at a specific time step i is equal Finally, using the field gradient corresponding to It is the pair from each time / update step contribution.

[0074] Specifically, directly calculate and The memory footprint of is very large, and it is difficult to store many state tensors. A state tensor corresponds to storing the values of all FDTD cells (e.g., multiple voxels) for a single simulation time step. It should be understood that the term "tensor" can refer to a tensor in the mathematical sense or a tensor as described by the TensorFlow framework developed by Alphabet. In some embodiments, the term "tensor" refers to a mathematical tensor corresponding to a multidimensional array that follows a specific transformation law. However, in most embodiments, the term "tensor" refers to a TensorFlow tensor, where tensors are described as generalizations of vectors and matrices to potentially higher dimensions (e.g., n-dimensional arrays of primitive data types) and are not necessarily limited to a specific transformation law. For example, for a general loss function f, it may be necessary to store the field x for all time steps i. i This is because, for most choices of f, the gradient will be a function of the parameters of f. Due to incremental updates of the field response and / or backpropagation through the loss metric, larger values of i are required before smaller values of i This difficulty is exacerbated by the fact that the value of Value scheme.

[0075] When calculating the structural gradient Another difficulty is further illustrated when the structural gradient It is given by:

[0076]

[0077] For completeness, the sum The first term in the full form is:

[0078]

[0079] Based on the definition of φ as described in equation (1), note that It can be substituted into Equation (3) to obtain the adjoint update for backpropagation (e.g., an update operation such as update operation 756), which can be expressed as:

[0080]

[0081] Or

[0082]

[0083] The adjoint update is the backpropagation of the loss gradient (e.g., from the loss metric) from a later time step to an earlier time step, and can be referred to as the backward solution for. More specifically, the loss gradient can initially be based on the backpropagation of the loss metric determined from the operation simulation using the loss function. The structural gradient and the second term in correspond to the field gradient, expressed as:

[0084]

[0085] For the specific form of φ described in Equation (1). Thus, each term in the relevant sum depends on both for i >= i0 and for i < i0. Since the direction of the dependency chain for these two terms is opposite, it can be concluded that calculating in this way requires storing the x i values for all i. In some embodiments, the need to store all field values can be alleviated by a simplified representation of the field.

[0086] Figure 8 FIG. shows an example method 800 for generating a design of a photonic integrated circuit according to an embodiment of the present disclosure. It should be understood that method 800 is an inverse design process, which can be implemented by performing iterative gradient-based optimization of the loss metric determined from a loss function by using a system (e.g., Figure 5 system 500). In the same or other embodiments, method 800 can be included as instructions provided by at least one machine-accessible storage medium (e.g., non-transitory memory), which, when executed by a machine, will cause the machine to perform operations for generating a design of a photonic integrated circuit. It should also be understood that the order in which some or all of the process blocks appear in method 800 should not be considered restrictive. Instead, those of ordinary skill in the art benefiting from the present disclosure will understand that some process blocks can be executed in various orders not shown, or even in parallel.

[0087] Block 810 illustrates configuring a simulation environment to represent an initial description of a photonic integrated circuit (e.g., a photonic device) that has been received or otherwise obtained. In some embodiments, the photonic integrated circuit may be expected to have a specific functionality after optimization (e.g., to perform as an optical demultiplexer). The initial description may describe structural parameters of the photonic integrated circuit within the simulation environment. The simulation environment may include a plurality of voxels that collectively describe the structural parameters of the photonic device. Each of the plurality of voxels is associated with a structural value describing the structural parameter, a field value describing a field response to a physical stimulus (e.g., one or more excitation sources) (e.g., electric and magnetic fields in one or more orthogonal directions), and a source value describing the physical stimulus. Once the initial description is received or otherwise obtained, the simulation environment is configured (e.g., the number of voxels, the shape / arrangement of the voxels, and specific values for the structural, field, and / or source values of the voxels are set based on the initial description). In some embodiments, the initial description may be a first description of the physical device, wherein the values of the structural parameters may be random or null values outside of the input and output regions so as to provide no deviation from the initial (e.g., first) design. It should be understood that the initial description or input design may be relative terms. Thus, in some embodiments, the initial description may be a first description of a physical device described in the context of a simulation environment (eg, a first input design for performing a first operational simulation).

[0088] However, in other embodiments, the term initial description may refer to an initial description of a particular cycle (e.g., performing an operational simulation, an operational concomitant simulation, and updating structural parameters). In such embodiments, the initial description or design of the particular cycle may correspond to a revised description or refined design (e.g., generated from a previous cycle). In one embodiment, the simulation environment includes a design region, wherein the design region includes a portion of a plurality of voxels having structural parameters, wherein the structural parameters may be updated, revised, or otherwise changed to optimize the structural parameters of the photonic device. In the same or other embodiments, the structural parameters are associated with geometric boundaries and / or material composition of the physical device based on material properties of the simulation environment (e.g., relative permittivity, refractive index, etc.).

[0089] In one embodiment, the simulation environment includes a design area optically coupled between a first communication area and a plurality of second communication areas. In some embodiments, the first communication area may correspond to an input area or port (e.g., where an excitation source originates), and the second communication area may correspond to a plurality of output areas or ports (e.g., when designing an optical demultiplexer that optically separates a plurality of different wavelength channels included in a multi-channel optical signal received at an input port and guides each different wavelength channel to a corresponding one of a plurality of output ports). However, in other embodiments, the first communication area may correspond to an output area or port, and the plurality of second communication areas may correspond to a plurality of input ports or areas (e.g., when designing an optical demultiplexer that optically combines a plurality of different wavelength signals received at corresponding input ports among a plurality of input ports to form a multi-channel optical signal guided to an output port).

[0090] Block 815 shows mapping each of the plurality of different wavelength channels to a corresponding one of the plurality of second communication zones. The different wavelength channels can be mapped to the second communication zones by virtue of an initial description of the photonic device. For example, a loss function can be selected that associates a performance metric of the photonic device with power transfer from an input port to each output port of the mapped channel. In one embodiment, a first channel included in the plurality of different wavelength channels is mapped to a first output port, meaning that the performance metric of the photonic device of the first channel is associated with the first output port. Similarly, other output ports can be mapped to the same or different channels included in the plurality of different wavelength channels, such that each different wavelength channel is mapped to a corresponding one of the plurality of output ports (i.e., the second communication zones) within the simulation environment. In one embodiment, the plurality of second communication zones include four zones, and the plurality of different wavelength channels include four channels, each channel being mapped to a corresponding one of the four zones. In other embodiments, there can be a different number of second communication zones (e.g., eight zones) and a different number of channels (e.g., eight channels), each channel being mapped to a corresponding one of the second communication zones.

[0091] Block 820 illustrates performing an operational simulation of a photonic integrated circuit operating in response to one or more excitation sources in a simulation environment to determine a performance metric. More specifically, an electromagnetic simulation is performed in which a field response of the photonic integrated circuit is incrementally updated over a plurality of time steps to determine how the field response of the physical device changes due to the excitation sources. Field values for a plurality of voxels are updated in response to the excitation sources and based at least in part on structural parameters of the integrated photonic circuit. Furthermore, each update operation at a particular time step may also be based at least in part on a previous (e.g., immediately preceding) time step.

[0092] Thus, an operational simulation simulates the interaction between a photonic device (i.e., a photonic integrated circuit) and a physical stimulus (i.e., one or more excitation sources) to determine a simulated output of the photonic device in response to the physical stimulus (e.g., at one or more output ports or regions). The interaction can correspond to any one or combination of perturbations, retransmissions, attenuations, dispersions, refractions, reflections, diffractions, absorptions, scatterings, amplifications, or the like of the physical stimulus within the electromagnetic domain due, at least in part, to the structural parameters of the photonic device and the underlying physics governing the operation of the photonic device. Thus, an operational simulation simulates how the field response of the simulated environment changes due to the excitation source over multiple time steps (e.g., from an initial time step with a predetermined step size to a final time step).

[0093] In some embodiments, the simulated output can be used to determine one or more performance metrics of the photonic integrated circuit. For example, an excitation source can correspond to a selected one of a plurality of different wavelength channels, each of which is mapped to one of a plurality of output ports. When performing an operational simulation, the excitation source can originate from a first communication region (i.e., an input port) or be disposed proximate to the first communication region. During the operational simulation, the field response at the output port mapped to the selected one of the plurality of different wavelength channels can then be used to determine a simulated power transfer of the photonic integrated circuit for the selected different wavelength channel. In other words, the operational simulation can be used to determine a performance metric by determining a simulated power transfer of the excitation source from the first communication region through the design region to a corresponding one of a plurality of second communication regions mapped to the selected one of the plurality of different wavelength channels. In some embodiments, the excitation source can cover the spectrum of all of the plurality of output ports (e.g., the excitation source spans at least a target frequency range of a passband region, a corresponding transition band region, and at least a portion of a corresponding stopband region of each of the plurality of different wavelength channels) to determine a performance metric (i.e., a simulated power transfer) associated with each different wavelength channel of the photonic integrated circuit. In some embodiments, one or more frequencies across a passband of a given one of the plurality of different wavelength channels are randomly selected to optimize the design (e.g., batch gradient descent while ensuring that the full width of each passband includes ripple in the passband that meets target specifications). In the same or other embodiments, each of the plurality of different wavelength channels has a common bandwidth with a different center wavelength.

[0094] Block 825 shows determining a loss metric based on a performance loss associated with the performance metric and a manufacturing loss associated with the minimum feature size. In some embodiments, the loss metric is determined by a loss function, wherein the loss function includes both the performance loss and the manufacturing loss as input values. The performance loss can correspond to the difference between the performance metric of the photonic integrated circuit and the target performance metric. In some embodiments, a minimum feature size of a design area of a simulation environment can be provided to improve the manufacturability of the design generated by the inverse design process. The manufacturing loss is based at least in part on the minimum feature size and structural parameters of the design area. More specifically, the manufacturing loss implements the minimum feature size of the design so that the design area does not have a structural element with a diameter less than the minimum feature size. This helps the system provide a design that meets specific manufacturability and / or yield requirements. In some embodiments, the manufacturing loss also helps implement binarization of the design (i.e., the design includes areas of a non-uniformly dispersed first material and a second material, rather than mixing the first material and the second material together to form a third material).

[0095] In some embodiments, manufacturing losses are determined by generating a convolution kernel (e.g., circular, square, octagonal, or other) with a width equal to the minimum feature size. The convolution kernel is then moved through the design area of the simulation environment to determine voxel positions (i.e., individual voxels) within the design area, where the voxel positions match the convolution kernel within the design area without extending outside the design area. The convolution kernel is then convolved with the structural parameters associated with the voxel position at each voxel position to determine a first manufacturing value. The structural parameters are then inverted, and the convolution kernel is convolved again with the inverted structural parameters at each voxel position to determine a second manufacturing value. The first and second manufacturing values are then combined to determine the manufacturing losses of the design area. This process of determining manufacturing losses can promote structural elements of the design area to have a radius of curvature that is less than a threshold size (i.e., the inverse of half the minimum feature size).

[0096] Block 830 shows backpropagating the loss metric through the simulation environment via the loss function to determine the effect of changes in the structural parameters on the loss metric (i.e., the structural gradient). The loss metric is treated as an adjoint source or virtual source and is incrementally backpropagated from the final time step to earlier time steps in the reverse simulation to determine the structural gradient of the photonic integrated circuit.

[0097] Block 835 illustrates revising the design of the photonic integrated circuit (e.g., generating a revised description) by adjusting the loss metric by updating structural parameters. In some embodiments, adjusting the loss metric can reduce the loss metric. However, in other embodiments, the loss metric can be adjusted or otherwise compensated in a manner that does not necessarily reduce the loss metric. In one embodiment, adjusting the loss metric can maintain manufacturability while providing a general direction within the parameterization space to achieve a design that will ultimately lead to improved performance while also maintaining device manufacturability and a target performance metric. In some embodiments, the revised description is generated by utilizing an optimization scheme after a round of operations and accompanying simulations, using a gradient descent algorithm, a Markov Chain Monte Carlo algorithm, or other optimization techniques. In other words, an iterative cycle of simulating the photonic integrated circuit, determining the loss metric, backpropagating the loss metric, and updating the structural parameters to adjust the loss metric can be continuously performed until the loss metric substantially converges, such that the difference between the performance metric and the target performance metric is within a threshold range, while also accounting for manufacturability and binarization due to manufacturing losses. In some embodiments, the term "converged" can simply mean that the difference is within a threshold range and / or below a certain threshold.

[0098] Block 840 illustrates determining whether the loss metric has substantially converged such that the difference between the performance metric and the target performance metric is within a threshold range. An iterative loop is performed to simulate a photonic integrated circuit having an excitation source selected from a plurality of different wavelength channels, backpropagate the loss metric, and revise the design by updating structural parameters to reduce the loss metric until the loss metric substantially converges such that the difference between the performance metric and the target performance metric is within the threshold range. In some embodiments, when the loop is executed, the structural parameters of the design region of the integrated photonic circuit are revised such that the design region of the photonic integrated circuit optically separates each of the plurality of different wavelength channels from a multi-channel optical signal received via a first communication region and directs each of the plurality of different wavelength channels to a corresponding one of a plurality of second communication regions based on the mapping of block 815.

[0099] Block 845 illustrates outputting an optimized design of the photonic integrated circuit, wherein the structural parameters have been updated to bring the difference between the performance metric and the target performance metric within a threshold range, while also enforcing minimum feature size and binarization.

[0100] The processes explained above are described in terms of computer software and hardware. The described techniques may constitute machine-executable instructions embodied in a tangible or non-transitory machine (e.g., computer) readable storage medium, which, when executed by a machine, causes the machine to perform the described operations. Furthermore, these processes may be embodied in hardware, such as an application-specific integrated circuit ("ASIC") or other.

[0101] Tangible machine-readable storage media include any mechanism that provides (i.e., stores) information in a non-transitory form accessible to a machine (e.g., a computer, a network device, a personal digital assistant, a manufacturing tool, any device having a collection of one or more processors, etc.). For example, machine-readable storage media include recordable / non-recordable media (e.g., read-only memory (ROM), random access memory (RAM), magnetic disk storage media, optical storage media, flash memory devices, etc.).

[0102] The above description of the illustrated embodiments of the present invention, including the description in the Abstract, is not intended to be exhaustive or to limit the invention to the precise forms disclosed. Although specific embodiments and examples of the invention are described herein for illustrative purposes, those skilled in the relevant art will recognize that various modifications are possible within the scope of the invention.

[0103] These modifications may be made to the invention in light of the above detailed description. The terms used in the appended claims should not be construed to limit the invention to the specific embodiments disclosed in the specification. Rather, the scope of the invention is to be determined entirely by the appended claims, which are to be construed in accordance with established doctrines of claim interpretation.

Claims

1. A multi-channel photon demultiplexer comprising: An input region receiving a multi-channel optical signal including four channels of different wavelengths; four output areas, each output area receiving a corresponding one of the four different wavelength channels demultiplexed from the multi-channel optical signal; and a dispersion region optically disposed between the input region and the four output regions, wherein the dispersion region comprises a first material and a second material non-uniformly dispersed to form a plurality of interfaces, wherein each of the plurality of interfaces corresponds to a change in a refractive index of the dispersion region, and the dispersion region is collectively configured to optically separate each of the four different wavelength channels from the multi-channel optical signal when the input region receives the multi-channel optical signal, and to guide each of the four different wavelength channels to a corresponding one of the four output regions; wherein the plurality of interfaces form a material interface pattern along a cross-sectional area of the dispersion region, the dispersion region being at least partially surrounded by a peripheral region comprising the second material; and wherein the material interface pattern comprises a plurality of islands, wherein a first island included in the plurality of islands is formed of the first material and surrounded by the second material, and wherein a second island included in the plurality of islands is formed of the second material and surrounded by the first material.

2. The multi-channel photon demultiplexer according to claim 1, wherein: The material interface pattern includes protrusions formed of the second material, the protrusions extending from the peripheral region into the dispersion region.

3. The multi-channel photon demultiplexer according to claim 1, wherein: The material interface pattern includes one or more dendritic shapes, wherein each of the one or more dendritic shapes is defined as a branched structure formed from the first material or the second material and has a width that alternates between increasing and decreasing in size along a corresponding direction.

4. The multi-channel photon demultiplexer according to claim 1, wherein: The first material and the second material are arranged and shaped within the dispersion region such that the material interface pattern is substantially proportional to a design obtainable using an inverse design process.

5. The multi-channel photon demultiplexer according to claim 4, wherein: The inverse design process includes iteratively optimizing a design based at least in part on a loss function comprising a performance loss and a manufacturing loss, the manufacturing loss being adjusted via the iterative optimization to generate the design.

6. The multi-channel photon demultiplexer according to claim 1, wherein: The dispersion region is further configured to optically separate each of the four different wavelength channels from the multi-channel optical signal within a predetermined area of 100 μm×100 μm or less when the input region receives the multi-channel optical signal.

7. The multi-channel photon demultiplexer according to claim 1, wherein: The dispersion region is configured to accommodate a common bandwidth for each of the four different wavelength channels, wherein each different wavelength channel has a different center wavelength, wherein the common bandwidth is 13 nm wide, and wherein the different center wavelengths include at least one of 1271 nm, 1291 nm, 1311 nm, 1331 nm, 1511 nm, 1531 nm, 1551 nm, or 1571 nm.

8. The multi-channel photon demultiplexer according to claim 1, wherein: The material interface pattern is shaped to implement a minimum feature size within the dispersion region such that the multiple interfaces within a cross-sectional area formed by the first material and the second material do not have at least one of: a radius of curvature less than a first threshold size, a minimum width less than a second threshold size, or a minimum spacing less than a third threshold size.

9. The multi-channel photon demultiplexer according to claim 1, wherein: The dispersion region is configured to have a power transfer of -2 dB or greater from the input region through the dispersion region to a corresponding one of the four output regions for a given wavelength within one of the four different wavelength channels.

10. The multi-channel photon demultiplexer according to claim 9, wherein: The dispersion region is further configured to have at least one of the following: for a given wavelength, an unfavorable power transmission from the input region to any one of the four output regions except the corresponding one of the four output regions is -30 dB or less, a ripple within a bandpass region of each of the four different wavelength channels is 1 dB or less, or a maximum power reflection of the multi-channel optical signal is -40 dB or less.

11. The multi-channel photon demultiplexer according to claim 1, wherein: The dispersion region includes a first side and a second side opposite the first side, wherein the input region is disposed proximate the first side, wherein the four output regions are disposed proximate the second side, and wherein each of the four output regions is positioned parallel to every other of the four output regions.

12. The multi-channel photon demultiplexer according to claim 11, wherein: Adjacent pairs of the four output regions are separated from each other by a common separation distance.

13. A photonic integrated circuit, comprising: first communication area; A second communication area, including four communication areas; and A dispersion region optically disposed between the first communication region and the second communication region, wherein the dispersion region includes at least a first material and a second material non-uniformly dispersed to form a plurality of interfaces, wherein each of the plurality of interfaces corresponds to a change in the refractive index of the dispersion region, wherein the plurality of interfaces form a material interface pattern along a cross-sectional area of the dispersion region, the dispersion region is at least partially surrounded by a peripheral region including the second material, wherein the material interface pattern is shaped to be substantially proportional to a design obtainable using an inverse design process, and wherein the material interface pattern includes a plurality of islands, wherein a first island included in the plurality of islands is formed by the first material and surrounded by the second material, and wherein a second island included in the plurality of islands is formed by the second material and surrounded by the first material.

14. The photonic integrated circuit according to claim 13, wherein: The first communication area is configured to receive a multi-channel optical signal including four different wavelength channels, wherein the four communication areas are configured to each receive a corresponding one of the four different wavelength channels demultiplexed from the multi-channel optical signal, and wherein the dispersion region is constructed based on the design to, when the first communication area receives the multi-channel optical signal, optically separate each of the four different wavelength channels from the multi-channel optical signal and respectively guide each of the four different wavelength channels to a corresponding one of the four communication areas.

15. The photonic integrated circuit according to claim 13, wherein: The four communication areas are configured to each receive a corresponding one of four different wavelength channels, wherein the dispersion area is constructed based on the design to optically combine the four different wavelength channels into a multi-channel optical signal when the four communication areas receive a corresponding one of the four different wavelength channels, and guide the multi-channel optical signal to the first communication area.

Images