Solid state device employing optical waveguide as floating gate electrode

Abstract

A semiconductor device includes a floating gate that can be charged in a non-volatile manner. The floating gate is also configured as an optical waveguide and can be optically coupled to a photonic circuit, such as an interferometer.

Description

[0001] Cross-reference to related applications

[0002] This Patent Cooperation Treaty patent application claims priority to U.S. Provisional Patent Application No. 63 / 014,806, filed April 24, 2020, entitled “Floating Gate Waveguide Nonvolatile Optical Switching,” the contents of which are incorporated herein by reference in their entirety. Technical Field

[0003] The embodiments described herein relate to solid-state devices comprising one or more floating gates configured for non-volatile storage of charge in optical waveguides. Background Technology

[0004] Semiconductor devices may include floating gates that can be charged or discharged via quantum mechanical tunneling through an insulating layer. More specifically, control gate electrodes disposed above the insulating layer can be driven to a specified voltage to sense or remove charge carriers from the floating gate. In these configurations, the charging state of the floating gate can be used as a non-volatile memory for digital circuits.

[0005] In some applications, photonic circuits can be chosen to replace semiconductor circuits to reduce power consumption and / or improve performance. Photonic circuits can include multiple optical waveguides configured to guide light to and between one or more passive or active optical circuits, photonic circuits, delay loops, input / output surfaces, etc. However, conventional photonic circuits cannot be implemented as non-volatile memories. Summary of the Invention

[0006] The embodiments described herein can take the form of a semiconductor device, which includes at least a source electrode, a control electrode, an insulating layer, and a floating gate separated from the source electrode and the control electrode by the insulating layer. In the configuration described herein, the floating gate is an optical waveguide optically coupled to a photonic circuit.

[0007] In these configurations, the floating gate can be charged or discharged through an insulating layer via quantum tunneling (e.g., Fowler-Nordheim tunneling and / or hot carrier injection). Within the insulating layer, the floating gate maintains its charged or uncharged state in a non-volatile manner.

[0008] As described above, in the configuration described herein, the floating grating also serves as an optical waveguide. As those skilled in the art will know, the presence or absence of charge in an optical waveguide affects its refractive index. Due to this photoelectric effect, the refractive index of the optical waveguide / floating grating differs between the charged and uncharged states of the floating grating. Furthermore, because the charged or uncharged state of the floating grating is non-volatile, the refractive index of the optical waveguide can also be modified / varied in a non-volatile manner. Thus, the optical waveguide can be used as part of a non-volatile optical switch.

[0009] More generally and more broadly, the constructions described herein can be used to form non-volatile memory cells for photonic circuits.

[0010] Related and additional embodiments may include a configuration in which the insulating layer comprises a first portion separating the source electrode from the floating gate and a second portion separating the control electrode from the floating gate. The first and second portions may be formed with different thicknesses to encourage charge retention in the floating gate / optical waveguide. In other cases, the second portion of the insulating layer may be an oxide-nitride-oxide dielectric layer.

[0011] In some configurations, the photonic circuit including the optical waveguide incorporates an interferometer, such as a Mach-Zehnder interferometer. In these examples, the charge state of the optical waveguide / floating grating introduces or does not introduce a phase shift in the light (e.g., infrared light) passing through it.

[0012] The light input to the photonic circuit is branched through two or more paths, one of which passes through an optical waveguide / floating grating, and these paths are recombined at one or more outputs. Any phase shift introduced by the optical waveguide / floating grating causes constructive or destructive interference at the recombined outputs, thus determining the amount of light at each output. In a simpler formulation, the charge state of the floating grating non-volatilely controls how (or whether) light passes through the photonic circuit.

[0013] The embodiments described herein can also take the form of a semiconductor device, which includes at least a conductive element, a layered dielectric disposed beneath the conductive element, and a silicon waveguide formed to bond with a silicon oxide layer of the layered dielectric. As with other examples, the silicon waveguide can be optically coupled to photonic circuits (e.g., Mach-Zehnder interferometers, ring resonators, etc.).

[0014] Furthermore, as with other embodiments described herein, the silicon waveguide can be electrically decoupled from the conductive element. As a result of this construction, in response to a voltage applied to the conductive element, the silicon waveguide accumulates charge through a quantum tunneling effect that affects the refractive index change of the silicon waveguide. This refractive index change can be used as a non-volatile optical memory and / or a non-volatile optical switch. In higher-order constructions, semiconductor devices as described herein can be used in silicon and / or photonic circuits, such as memory cells, field-programmable photonic gate arrays, etc. Any suitable photonic circuit of any appropriate complexity or scale can be implemented to utilize the systems and methods described herein.

[0015] Brief description of the attached diagram

[0016] Reference will now be made to the representative embodiments illustrated in the accompanying drawings. It should be understood that the following description is not intended to limit this disclosure to a single, encompassing embodiment. Rather, the disclosure provided herein is intended to cover alternatives, modifications, and equivalents that may be included within the spirit and scope of the described embodiments and as defined by the appended claims.

[0017] Figure 1 A simplified cross-section of the non-volatile programmable passive optical structure for photonic circuits, as described herein, is depicted.

[0018] Figure 2A-2B Each depicts a simplified schematic diagram of an optical waveguide that can be optically coupled to a non-volatile programmable passive optical structure for photonic circuits as described herein.

[0019] Figures 3A-3E Each is depicted with a simplified system diagram of the non-volatile programmable passive optical structure for photonic circuits as described herein.

[0020] Figure 4 Field-programmable gate arrays are described, which can utilize one or more non-volatile programmable passive optical structures for photonic circuits as described herein.

[0021] Figure 5 This is a flowchart illustrating an example operation of a method for operating a non-volatile programmable passive optical structure as described herein.

[0022] Figure 6 This is a flowchart illustrating an example operation of a method for operating a non-volatile programmable passive optical structure with photonic circuitry as described herein.

[0023] Using the same or similar reference numerals in different figures to indicate similar, related or identical items.

[0024] The use of crosshairs or shading in the accompanying drawings is generally to clarify the boundaries between adjacent elements and to improve readability. Therefore, the presence or absence of crosshairs or shading does not convey or indicate any preference or requirement for a particular material, material properties, element proportions, element dimensions, commonalities of similar illustrated elements, or any other characteristic, property, or property of any element shown in the accompanying drawings.

[0025] Some of the accompanying figures include or suggest vector, ray, trajectory, and / or other visual representations of one or more exemplary paths—which may include reflection, refraction, diffraction, etc. through one or more media—that may be taken or may be presented to represent one or more photons, wavelets, or other propagating electromagnetic energy that is derived from or generated from one or more light sources shown in the figures or, in some cases, omitted from the figures.

[0026] It is understood that, regardless of the spectrum (e.g., ultraviolet, visible, infrared, etc.), these simplified visual representations of the light or optical path, or more generally, their electromagnetic energy or waveguides, are provided only for the purpose of understanding the various embodiments described herein, and are therefore not necessarily presented or illustrated to scale or with angular precision or accuracy. Consequently, they are not intended to represent any preference or requirement for the illustrated embodiments to emit, reflect, refract, focus, and / or diffract light at any particular illustrated angle, orientation, polarization, color, or direction, except for other embodiments described or referenced herein.

[0027] Furthermore, it should be understood that the proportions and dimensions (relative or absolute) of various features and elements (as well as their sets and groups) and the boundaries, separations and positional relationships between them presented in the accompanying drawings are provided merely to facilitate understanding of the various embodiments described herein, and are therefore not necessarily presented or illustrated to scale, and are not intended to indicate any preference or requirement for the illustrated embodiments to exclude reference to the embodiments described herein. Detailed Implementation

[0028] The embodiments described herein relate to photonic circuits and structures having stable and non-volatile optical properties in two or more states. Such structures can be used in photonic circuits with non-volatile field configurations or non-volatile storage, for example, in one or more optical memory cells, an array of one or more memory cells, one or more configurable logic blocks of a field-programmable gate array, one or more bistable switching elements, and so on.

[0029] More generally, those skilled in the art will readily understand that optical elements exhibiting electrically programmable, non-volatile, and optical properties can be used in countless ways in photonic and electronic circuits.

[0030] Furthermore, because these characteristics are non-volatile, photonic circuits incorporating elements as described herein can operate at significantly reduced power compared to electrically switched photonic circuits known to those skilled in the art that exhibit significantly reduced power compared to semiconductor circuits. More specifically, photonic circuits as described herein can be operated and can perform one or more functions while consuming only the amount of electrical power required to generate light passing through the photonic circuit. In other words, the embodiments described herein can be used to create passive optical circuits that are electrically programmable and, once programmed, require no power to maintain their state or function.

[0031] For example, a single bistable optical element, as described herein, can be programmed to allow light to pass through or block light. More specifically, once programmed and regardless of the electrical power state, the bistable optical element allows light to pass through or blocks the provided light as an optical input. This controllable optical characteristic can be used as bits of digital memory, control gates for enabling or disabling other photonic circuit elements optically coupled to the bistable optical element, pixels of a powerless digital image, or for any other suitable purpose.

[0032] In other examples, arrays of bistable optical elements can be used to create non-volatile digital memory cells or arrays of non-volatile memory cells to persistently store any suitable amount of digital data.

[0033] In other examples, bistable operation may be unnecessary or preferred. For instance, a programmable optical element as described herein can be programmed to attenuate light by a specific programmed amount, or to switch entirely between outputs. More specifically, once programmed, the optical element can be configured to attenuate the light supplied to it by a specified amount regardless of the electrical power state. This controllable optical property can be used as part of a trained neural network or other machine learning data structure (e.g., a node or layer). More generally and more extensively, arrays of optical elements as described herein can be programmed to perform any machine learning or training-proficient computational function as a passive optical structure.

[0034] These structures, which may be referred to in this paper as “non-volatile programmable passive optical structures,” are particularly useful in low-power applications such as edge computing devices and small electronic devices such as personal electronic devices, robots, drones, IoT devices, satellites, etc.

[0035] In other implementations, non-volatile programmable passive optical structures, as described herein, can be used for large-scale, energy-efficient data storage.

[0036] In other examples, non-volatile programmable passive optical structures or networks thereof can be used or used as trained neural networks or other trained, proficient machine learning systems configured for one or more of the following: signal analysis (e.g., Fourier transform or other transforms, encoding, decoding, modulation, demodulation, encryption, decryption, etc.); cryptographic computation (e.g., cryptocurrency mining, data encryption or decryption); autonomous driving; image recognition; speech recognition; command and / or control of drones or robots; GPS computation; hardware virtualization; network switching; or for any other suitable computational or control task.

[0037] Therefore, it can be understood, generally and broadly, that non-volatile programmable passive optical structures or networks of such structures as described herein can be programmed to optically / physically implement neural networks or other trained machine learning structures, or more generally, any digital or analog logic structure or network of digital or analog logic structures.

[0038] Furthermore, non-volatile programmable passive optical structures can be used in or with any suitable computing resource configured to perform any computational operation or function. As used herein, the term "computing resource" (along with other similar terms and phrases, including but not limited to "computing device" and "computing network") may refer to any physical electronic device or machine component, or a set or group of interconnected and / or communicatively coupled physical electronic devices or machine components suitable for performing or causing one or more arithmetic or logical operations on digital data or analog signals.

[0039] Example computing resources incorporating one or more non-volatile programmable passive optical structures as envisioned herein include, but are not limited to: single-core or multi-core processors; single-threaded or multi-threaded processors; purpose-configurable coprocessors (e.g., graphics processing units, motion processing units, sensor processing units, etc.); configurable logic units (e.g., field-programmable gate arrays); memory units; volatile or non-volatile memories or memory units or arrays; application-specific integrated circuits (ASICs); field-programmable gate arrays (FPGAs); input / output devices and systems and their components (e.g., keyboards, mice, touchpads, general-purpose human-machine interface devices, cameras, microphones, speakers, etc.); network devices and systems and their components (e.g., routers, switches, firewalls, packet shapers, content filters, network interface controllers or cards, access points, etc.). Modems, etc.; embedded devices and systems and their components (e.g., system-on-a-chip, IoT devices, etc.); industrial control or automation devices and systems and their components (e.g., programmable logic controllers, programmable relays, supervisory control and data acquisition controllers, discrete controllers, etc.); vehicle or aviation control devices and systems and their components (e.g., navigation devices, safety devices or controllers, security devices, etc.); corporate or enterprise infrastructure equipment or appliances (e.g., dedicated small switch equipment, Internet Protocol voice host and controller, end-user terminals, etc.); personal electronic devices and systems and their components (e.g., mobile phones, tablets, desktop computers, laptops, wearable devices); personal electronic devices and their accessories (e.g., peripheral input devices, wearable devices, implantable devices, medical devices, etc.); and so on. It is understood that the foregoing examples are not exhaustive.

[0040] In view of the foregoing, it can be understood that, generally and broadly, the non-volatile programmable passive optical structures described herein can be used as basic components of any number of photonic circuits.

[0041] Therefore, for the sake of simplicity in description and illustration, the following embodiments refer to a non-volatile programmable passive optical structure implemented as part of a Mach-Zehnder interferometer used as an optical switch, configured to guide light to one of two output paths, depending on the programming state. In these examples, the non-volatile programmable passive optical structure is stable relative to each corresponding output (e.g., configured to allow light to pass through or block light from that output, depending on the programming state) and can be configured as a digital non-volatile optical switch. However, it should be understood that this is merely an exemplary construction. In other cases, a ring resonator or a chain thereof may be used instead of the interferometer structure or as a supplement to the interferometer structure.

[0042] In this exemplary configuration, the non-volatile programmable passive optical structure has at least one optical input and at least one optical output. At least one of the optical inputs and at least one of the optical outputs are optically coupled to one or more other components of a photonic circuit, which may vary depending on the embodiment.

[0043] The optical input of a non-volatile programmable passive optical structure can be coupled into a single path and then branch into two arms (of equal length) to achieve an interferometer structure. Subsequently, the first and second arms are recombined via a second coupler. Due to this configuration, any phase difference between the first and second arms will cause destructive or constructive interference at the optical output, thereby altering the relative power output between the optical outputs.

[0044] Those skilled in the art will readily understand that a phase shift can be introduced in the first or second arm using several optical, mechanical, or electrical properties, thereby introducing a phase difference between the arms of the interferometer structure, as described above. Examples include thermo-optical effects, thermal expansion / contraction, photoelectric effects, optional delay loops, etc.

[0045] In other cases, the non-volatile programmable passive optical structures described herein may include one or more ring resonators. In these configurations, one or more closed-loop optical waveguides may be positioned / placed near one or more other (typically linear, although this is not required) waveguides. Due to the proximity of the closed-loop waveguides to these one or more other waveguides, light couples between them. As those skilled in the art will know, light of a specific frequency (which varies from ring resonator to ring resonator) will constructively interfere in each successive round trip through each closed-loop waveguide, thereby achieving resonance therein. Due to this configuration, the geometry or optical characteristics of each ring resonator will affect the frequency of the light resonating in each respective ring resonator.

[0046] Similar to the examples described above, those skilled in the art will readily understand that variations in the resonant frequency of a ring resonator can be introduced using several optical, mechanical (thermal), or electrical properties, as described herein. Examples include thermo-optical effects, thermal expansion / contraction, photoelectric effects, and optional delay rings.

[0047] The embodiments described herein refer to methods and structures for utilizing the photoelectric effect (where the presence of an electric field causes a change in the effective refractive index of an optical waveguide, thereby imparting a phase shift to light passing through it) to introduce a controllable phase difference between the two arms of an optical interferometer structure within a ring resonator or, for example, any other suitable optical structure described above.

[0048] Specifically, in these embodiments, one arm of the interferometer structure—or more specifically, at least a portion of one arm of the interferometer structure—is configured to operate as a floating gate formed of a conductive material encapsulated in an insulator (e.g., silicon dioxide). Due to this encapsulation, the floating gate can be used to store charge in a non-volatile manner because there is no electrical path for dissipating any accumulated charge in the floating gate.

[0049] Charge can accumulate in the floating gate through quantum tunneling (e.g., via Fowler-Nordheim tunneling) or hot carrier injection through the potential barrier (i.e., insulator) surrounding the floating gate, even though the floating gate is encapsulated by an insulating layer.

[0050] More specifically, in many configurations, the source electrode and control electrode can be disposed above the insulating layer encapsulating the floating gate. Due to this configuration, the implementation and / or design of a voltage of a specific magnitude applied to the control electrode can induce one or more charge carriers to tunnel through the insulating layer between the source electrode and the floating gate, and accumulate or reduce the number of charge carriers in the floating gate.

[0051] Similarly, applying an opposite voltage to the control electrode can induce one or more charge carriers to tunnel in opposite directions between the floating gate and the source electrode, thereby changing the charge state of the floating gate again.

[0052] As is the convention in this paper, applying a signal to the control electrode to cause a change in the charge state of the floating gate to a specified / desired value is referred to as a "write" operation or a "programming" operation. Similarly, applying a signal to the control electrode that causes a change in the charge state of the floating gate to a default value is referred to as an "erase" operation or a "reset" operation.

[0053] As described above, in many embodiments, a floating grating can be optically coupled to one arm of the interferometer structure. In another formulation, the floating grating can also serve as an optical waveguide in an optical path defined by one arm of the interferometer structure.

[0054] Due to this architecture—where the optical waveguide is also a floating grating, configured to accumulate charge and retain that charge in a non-volatile manner—a phase shift is applied in the first arm proportional to the accumulated charge of the floating grating. The accumulated charge, in turn, causes a phase difference between the first and second arms of the interferometer structure, resulting in destructive interference at the optical output of the non-volatile programmable passive optical structure. By selecting the size of the floating grating and / or the charge state to impart a 180° phase difference, the optical interferometer structure can be used as a non-volatile bistable switching element; in the first charge state, light passes through the optical interferometer structure, while in the second charge state, the light destructively interferes and does not pass through the optical interferometer structure. In other structures, the optical interferometer structure may include multiple optical outputs; in such an example, in the first charge state, light passes to the first output, while in the second charge state, light passes to the second output. In still other examples, intermediate charge states may each be associated with a different optical power output from each output segment.

[0055] Therefore, more generally and more broadly, the embodiments described herein can take the form of a semiconductor device, which includes at least a source electrode, a control electrode, an insulating layer, and a floating gate spaced apart from the source electrode and the control electrode by the insulating layer. In the configuration described herein, the floating gate is an optical waveguide optically coupled to or into a photonic circuit.

[0056] In these configurations, the floating gate can be charged or discharged (programmed or erased) through the insulating layer via quantum tunneling (e.g., Fowler-Nordheim tunneling and / or hot carrier injection). Because the floating gate is electrically isolated by the insulating layer, it retains its charged or uncharged state in a non-volatile manner.

[0057] As described above, for these configurations, the floating grating also functions as an optical waveguide. Due to the photoelectric effect, the refractive index of the optical waveguide / floating grating differs between the charged and uncharged states of the floating grating. Furthermore, since the charged or uncharged state of the floating grating is non-volatile, the effective refractive index of the optical waveguide is also non-volatile and modifiable. In this way, the optical waveguide can be used as part of a non-volatile optical switch, as described above.

[0058] It is understood that non-volatile programmable passive optical structures can be incorporated into multiple photonic and / or semiconductor circuits for any number of optical or electrical purposes or functions. For example, as described above, the interferometer structure is merely one exemplary optical structure that can utilize the methods and techniques described herein. In a ring resonator construction, one or more closed-loop waveguides (rings) can be implemented as floating gratings as described herein. In other constructions, a portion of the closed-loop waveguide can be passed through and / or optically coupled to a floating grating / optical waveguide as described herein.

[0059] Furthermore, in some exemplary configurations, semiconductor devices including non-volatile programmable passive optical structures (including at least one floating gate that also functions as an optical waveguide) as described herein may also include a drain electrode. In these exemplary configurations, the electric field generated by the charge accumulated in the floating gate can be measured electrically or optically by determining the electrical characteristics of the voltage difference between the drain and source. In other words, in these configurations, the floating gate can operate as an optical memory or an electrical memory (e.g., flash memory).

[0060] In some implementations, manipulating the control electrodes to accumulate charge in the floating gate / waveguide during a write operation can be a time-consuming process; in such examples, a non-volatile programmable passive optical structure can be electrically readable via the drain electrode before the photonic circuitry optically coupled to it becomes optically readable. In these constructions, different read operations (i.e., electrical and optical) can be utilized to improve overall performance. For example, an electrical read can be performed until the floating gate is fully charged to enable an optical read operation. Those skilled in the art will understand that many different read / write techniques are possible, both electrical and optical.

[0061] See below for reference. Figure 1-6 The discussion includes these and other constructions of non-volatile programmable passive optical structures. However, those skilled in the art will readily understand that the detailed descriptions of these figures given herein are for illustrative purposes only and should not be construed as limiting.

[0062] Generally and widely, Figure 1-3E The depiction and / or reference to exemplary semiconductor structures may include non-volatile programmable passive optical structures as described herein. For the sake of simplicity of description and illustration, the depicted and described non-volatile programmable passive optical structures are implemented as non-volatile memory cells or non-volatile optical switches; however, as stated above, this is merely an exemplary construction and other implementations are possible.

[0063] Figure 1 A simplified cross-section of the non-volatile programmable passive optical structure 100 for photonic circuits as described herein is depicted. The non-volatile programmable passive optical structure 100 can be included as part of any suitable electronic or photonic circuit.

[0064] The non-volatile programmable passive optical structure 100 includes an optical waveguide 102. The optical waveguide 102 can be a ribbed waveguide, or it can take any suitable optical waveguide shape. The optical waveguide 102 can be formed of any suitable material, although crystalline silicon may be used in many embodiments. Therefore, in many examples, the optical waveguide 102 is a silicon optical waveguide.

[0065] The optical waveguide 102 is embedded in an insulating layer 104, which may be silicon oxide, such as silicon dioxide. The thickness of the cladding / encapsulation of the optical waveguide 102 may vary depending on the location and embodiment.

[0066] The optical waveguide 102 may include ribs and two wings extending from the ribs. The ribs are depicted at the center of the optical waveguide 102, but this may not be necessary in all embodiments. In some examples, the optical waveguide 102 may be offset relative to the depicted position.

[0067] Any suitable technique can be used to form the optical waveguide 102. In some examples, the optical waveguide 102 is formed on a silicon oxide region of the initial substrate. In other cases, the optical waveguide 102 can be formed in a first process, followed by the formation of an insulating layer 104 around the optical waveguide 102. Those skilled in the art will readily understand that any suitable silicon fabrication or micromachining technique can be used to form one or more features of the optical waveguide 102 and to position the optical waveguide 102 within the insulating layer 104.

[0068] The non-volatile programmable passive optical structure 100 also includes a control electrode 106 and a source electrode 108, each formed of a conductor or a semiconductor material such as polysilicon. The control electrode 106 may be isolated from the optical waveguide 102 via a dielectric layer 110 (e.g., electrically decoupled), and the source electrode 108 may be isolated from the optical waveguide 102 via a dielectric layer 112 (e.g., electrically decoupled).

[0069] In some examples, dielectric layer 110 and dielectric layer 112 may be formed of silicon oxide. In a further embodiment, dielectric layer 110 and dielectric layer 112 may be formed of the same material as insulating layer 104.

[0070] In some embodiments, dielectric layer 110 and dielectric layer 112 may be formed with different thicknesses; in a typical embodiment, dielectric layer 110 may be formed with a greater thickness than dielectric layer 112, but this may not be necessary in all embodiments.

[0071] In other cases, dielectric layer 110 and dielectric layer 112 may be formed of different materials or combinations of different materials. In one embodiment, dielectric layer 110 is formed as a layered substrate comprising a first portion / layer of silicon oxide, a second portion / layer of silicon nitride, and a third portion / layer of silicon oxide.

[0072] In more specific terms, a first layer of silicon oxide may be formed to bond with the wing region of the optical waveguide 102, a layer of silicon nitride may be formed to bond with the first layer of silicon nitride, and a second layer of silicon oxide may be formed to bond with the silicon nitride layer and the control electrode 106.

[0073] Similar to the optical waveguide 102, the dielectric layer 110, dielectric layer 112, control electrode 106, and source electrode 108 can all be formed in a variety of suitable ways with any number of suitable thicknesses, made of any suitable amount of material, etc. It is understood that the cross-section shown is merely a simplified example.

[0074] In many cases, the control electrode 106 and the source electrode 108, as well as each corresponding dielectric layer separating these electrodes from the optical waveguide 102, are also encapsulated within the insulating layer 104. This architecture can be chosen to passivate the control electrode 106 and the source electrode 108 and prevent them from being corroded or damaged.

[0075] The control electrode 106 and the source electrode 108 may each be electrically coupled to one or more metallized portions (e.g., metallized regions 114 and 116). In some examples, the control electrode 106 may include an extension 106a to which the metallized region 116 extends via a via for electrical coupling to the control electrode 106. Similarly, the source electrode 108 may include an extension 108a to which the metallized region 114 extends via a separate via for electrical coupling to the source electrode 108. The metallized portions may be formed of any suitable metal (e.g., gold or aluminum).

[0076] The non-volatile programmable passive optical structure 100 can be coupled to circuitry configured to perform one or more write or erase operations. In particular, the circuitry—which may be referred to as a state controller—can be configured to apply a design-specific or implementation-specific voltage signal relative to system ground to the control electrode 106 via the metallized region 116.

[0077] When a voltage signal is applied, one or more charge carriers can be induced to tunnel through the dielectric layer 112 between the optical waveguide 102 and the source electrode 108. In these examples, the voltage can be selected such that the charge carriers induced to move into the optical waveguide 102 may not tunnel further through the dielectric layer 110. As a result, the tunneled charge carriers are effectively trapped within the optical waveguide 102 of the non-volatile programmable passive optical structure 100. In this way, the optical waveguide 102 serves as a floating gate as described above.

[0078] When the voltage is removed from the control electrode 106, charge remains in the optical waveguide 102, affecting its refractive index. In some cases, the optical waveguide 102 may include a doped pattern that induces any charge carriers in the optical waveguide 102 to concentrate in specific regions of the optical waveguide 102 (e.g., rib regions of the optical waveguide 102); however, this is not required in all embodiments.

[0079] Based on the described structure, it can be understood that the charge state of the optical waveguide 102 is a variable non-volatile characteristic that does not require electricity to maintain. Conversely, since the charge state affects the refractive index of the optical waveguide 102 through the photoelectric effect, the refractive index of the optical waveguide is also a variable non-volatile characteristic that does not require electricity to maintain.

[0080] In this way, the structure depicted is a non-volatile passive optical structure with at least one electrically programmable optical property.

[0081] Generally, for the purposes of explanation, and to facilitate understanding of the various configurations and constructions of the non-volatile programmable passive optical structures described herein, are presented as follows: Figure 1 The foregoing embodiments and their various alternatives and variations are described. However, it will be apparent to those skilled in the art that some of the specific details presented herein may not be necessary to practice the particular described embodiments or their equivalents.

[0082] For example, in other embodiments, the proportions and relative dimensions of the layers of the non-volatile programmable passive optical structure 100 can vary. In some cases, the wing region of the optical waveguide can extend further than illustrated; in others, the ribs of the optical waveguide can take on different shapes.

[0083] Furthermore, it is understandable that Figure 1 The structure shown can be implemented at any suitable scale. In many embodiments, the structure can be a micrometer-scale structure with a width of approximately 5 μm and a height of less than 1 μm. In other cases, different proportions of different elements, cross-sectional profiles, relative dimensions, and layer thicknesses can be used.

[0084] In some cases, optical waveguides can extend linearly (inside or outside the page), while in others they can be at least partially bent. In some cases, optical waveguides can be optically coupled to one or more photonic circuits.

[0085] Furthermore, as described above, one or more dimensions or dimensional characteristics of the optical waveguides described herein can be selected to impart—when exhibiting a particular charge—a 180° phase shift relative to another parallel waveguide (e.g., which may be included in an interferometer structure). In these examples, the charged state of the optical waveguide / floating grating corresponds to completely destructive interference (i.e., no light passes through the interferometer structure), while the uncharged state of the optical waveguide / floating grating corresponds to completely in-phase constructive interference (i.e., allowing light to pass through the interferometer structure). In other structures, light may selectively pass between different outputs based on different charge states.

[0086] In other examples, the silicon nitride layer disposed between the optical waveguide and the control electrode can be used as a second waveguide coupled to different (or the same) photonic circuits. In other cases, the silicon nitride layer can be set to a thickness unsuitable for waveguide operation.

[0087] In such an embodiment, the floating gate / optical waveguide should be electrically decoupled from any circuit or path to circuit or system ground so that the floating gate / optical waveguide can maintain its charged state in a non-volatile manner.

[0088] In one exemplary configuration, the waveguide optically coupled to the waveguide of the non-volatile programmable passive optical structure as described herein can be segmented, for example... Figure 2A As shown in the diagram. In this embodiment, waveguide 200a may include a first segment 202, which is electrically decoupled from a second segment 204, which in turn is electrically decoupled from a third segment 206. In this exemplary configuration, each segment may be optically coupled due to their relative proximity while maintaining electrical decoupling. More specifically, each illustrated segment may be separated by an insulator (e.g., silicon oxide). In this example, the second segment 204 may be a floating grating / optical waveguide of, for example, a non-volatile programmable passive optical structure described herein.

[0089] In another example, optical coupling can be formed by an out-of-plane adiabatic cone. Figure 2B Waveguide 200b is depicted. Waveguide 200b may include a first segment 202, which is electrically decoupled from a second segment 204, and the second segment 204 is electrically decoupled from a third segment 206. Figure 2A The embodiment shown is the same. However, in this exemplary configuration, each segment includes at least one adiabatic cone that overlaps with an optical jumper such as optical jumpers 208 and 210. More specifically, optical jumpers 208 and 210 may also include corresponding adiabatic cones and may be formed and / or arranged out of plane together with the respective segments of waveguide 200b. In this way, light can pass through waveguide 200b, through the overlapping adiabatic cones, and across different material layers. In this example, with Figure 2A As in the illustrated embodiment, the segments of waveguide 200b are optically coupled but electrically decoupled. More specifically, each illustrated segment may be separated by an insulator (e.g., silicon oxide). Figure 2A As in the illustrated embodiment, in this example, the second segment 204 can be a floating grating / waveguide of a non-volatile programmable passive optical structure as described herein.

[0090] Generally, for the purposes of explanation, and to facilitate understanding of the various configurations and constructions of non-volatile programmable passive optical structures including gratings / waveguides as described herein, are presented in Figure 1-2BThe foregoing embodiments and their various alternatives and variations are described herein. However, it will be apparent to those skilled in the art that some of the specific details presented herein may not be necessary to practice the particular described embodiments or their equivalents.

[0091] Therefore, it should be understood that the foregoing and following description of the specific embodiments are presented for limited purposes of illustration and description. These descriptions are not intended to be exhaustive or to limit this disclosure to the precise forms described herein. Rather, it will be apparent to those skilled in the art that many modifications and variations are possible in light of the foregoing teachings.

[0092] For example, generally and broadly, it can be understood, for example, by referring to Figure 1-2B The described non-volatile programmable passive optical structure can be implemented in any number of suitable ways. In some examples, the source electrode can be positioned along or overlapped with the edge of the optical waveguide / floating grating. In some examples, the control electrode can be positioned along or overlapped with the edge of the optical waveguide / floating grating. In some cases, the control electrode can be positioned on one surface (e.g., the bottom surface) of the optical waveguide, while the source electrode can be positioned on the opposite surface of the same waveguide. In some cases, more than one optical waveguide / floating grating can be associated with the same control electrode / source electrode pair. In some cases, as described above, a drain electrode may also be included to impart both electrical and optical readout functionality to the non-volatile programmable passive optical structure.

[0093] Figure 3A A system diagram 300a depicts a non-volatile programmable passive optical structure for photonic circuits as described herein. The non-volatile programmable passive optical structure 302 receives light from an optical input terminal 304 and provides the output to further photonic elements or photonic circuits via an optical output terminal 306.

[0094] The non-volatile programmable passive optical structure 302 is described as realizing an interferometer structure, although this is not required in all embodiments. In other cases, it may include a ring resonator, a delay ring, and other optical waveguides or optical / photonic structures.

[0095] The interferometer structure is defined by the optical input terminal 304 splitting or branching into a first arm 308 and a second arm 310. The first arm 308 can also be called the reference arm, and the second arm can also be called the variable arm.

[0096] In this example, the second arm 310 is configured for optical coupling via a floating gate / optical waveguide 312, which can be configured as shown above. Figure 1-2B As described, the floating gate / optical waveguide 312 can be completely encapsulated in an insulating layer (e.g., silicon oxide). However, the floating gate / optical waveguide 312 can be optically coupled to the second arm 310.

[0097] The floating gate / waveguide 312 can be formed adjacent to the source electrode 314 and the control electrode 316, and can be configured to operate as described above. More specifically, in this configuration, when the non-volatile programmable passive optical structure 302 receives a voltage signal—such as a programming signal or an erase signal—at the signal input 318, the control electrode 316 can be driven to a specified voltage. This specified voltage can then cause charge to accumulate in the floating gate / waveguide 312 due to tunneling through the insulating layer that separates the source electrode 314 from the floating gate / waveguide 312.

[0098] Once charge accumulates in the floating grating / optical waveguide 312, the effective group refractive index of the second arm 310 changes, which in turn changes the phase of the light passing through the second arm 310 relative to the reference arm.

[0099] As a result, when the light passing through the variable arm recombines with the light passing through the reference arm before the optical output terminal 306, the phase difference causes destructive interference, thereby attenuating the light output from the optical output terminal 306 relative to the amplitude of the light supplied as input to the optical input terminal 304. As described above, any suitable phase difference, including a 180° phase difference, can be achieved by controlling the dimensions (e.g., length, width, shape, etc.) of the floating grating / optical waveguide 312 and the charge applied thereto, which, as those skilled in the art, effectively prevents any light from passing through the optical output terminal 306.

[0100] In this way, the non-volatile programmable passive optical structure 302 can be used as a non-volatile, low-power optical storage unit or a non-volatile optical switch.

[0101] Generally, for the purposes of explanation, and to facilitate understanding of the various configurations and constructions of the non-volatile programmable passive optical structures described herein, are presented in Figure 1-3A The foregoing embodiments and their various alternatives and variations are described herein. However, it will be apparent to those skilled in the art that some of the specific details presented herein may not be necessary to practice the particular described embodiments or their equivalents.

[0102] Therefore, it should be understood that the foregoing and following description of the specific embodiments are presented for limited purposes of illustration and description. These descriptions are not intended to be exhaustive or to limit this disclosure to the precise forms described herein. Rather, it will be apparent to those skilled in the art that many modifications and variations are possible in light of the foregoing teachings.

[0103] For example, as described above, in some examples, a drain electrode 320 may also be included. Measuring the voltage between the drain electrode 320 and the source electrode 314 can probe the charge state of the floating gate / optical waveguide 312. In this way, the drain electrode 320 can be used to verify that charge has been correctly applied to the floating gate / optical waveguide 312 and / or as an electronic bit in digital memory. In another non-limiting statement, the drain electrode can be used as a redundant electrical method for determining the non-volatile charge state of the floating gate / optical waveguide 312. In some cases, this can improve the apparent speed of operation of the memory cells described herein; in a first mode, when the floating gate / optical waveguide is being charged and / or programmed, the drain electrode 320 can be used as a digital non-volatile bit in a semiconductor circuit. Once the floating gate is fully charged, the drain electrode 320 may no longer be needed and / or used; thereafter the circuit can only be read optically. It is understood that these structures are merely examples; in other cases, other methods and / or mutual electrical and optical operations can be performed using the structures described and depicted herein.

[0104] Furthermore, as mentioned above, in certain cases, an interferometer structure with multiple inputs and multiple outputs can be constructed. For example, as with the other embodiments presented herein, Figure 3B A system diagram 300b depicts a non-volatile programmable passive optical structure for photonic circuits as described herein. The non-volatile programmable passive optical structure 302 receives light from optical inputs 304a and 304b and provides output to other photonic elements or photonic circuits via optical outputs 306a and 306b. The optical inputs 304a and 304b can be coupled together at a junction (also called coupler 304c).

[0105] and Figure 3A The embodiments shown are the same. Figure 3B The interferometer structure depicted may include splitting or branching the output of coupler 304c into a first arm 308 and a second arm 310. The first arm 308 may also be referred to as a reference arm, and the second arm may also be referred to as a variable arm, a control arm, a floating gate arm, etc.

[0106] In this example, as in the other examples presented herein, the second arm 310 is configured for optical coupling via a floating gate / optical waveguide 312, which can be configured as described above. Figure 1-3A As described, the floating gate / optical waveguide 312 can be completely encapsulated in an insulating layer (e.g., silicon oxide), thereby electrically decoupling the floating gate / optical waveguide 312 from other electrical components. Despite being electrically decoupled from other circuit or ground paths, the floating gate / optical waveguide 312 can be optically coupled to (or into) the second arm 310.

[0107] The floating gate / waveguide 312 can be formed adjacent to the source electrode 314 and the control electrode 316, which can be configured to operate as described above. More specifically, in this configuration, when the non-volatile programmable passive optical structure 302 receives a voltage signal—such as a programming signal or an erase signal—at the signal input 318, the control electrode 316 can be driven to a specified voltage. This specified voltage can then cause charge to accumulate in the floating gate / waveguide 312 due to tunneling through the insulating layer separating the source electrode 314 from the floating gate / waveguide 312.

[0108] Once charge accumulates in the floating grating / optical waveguide 312, the effective group refractive index of the second arm 310 changes, which in turn changes the phase of the light passing through the second arm 310 relative to the reference arm.

[0109] As a result, when the light passing through the variable arm is recombined by the second coupler 306c with the light passing through the reference arm before the optical output terminals 306a and 306b, the phase difference causes destructive or constructive interference, thereby dividing the power of the light input to the optical input terminals 304a and 304b between the optical output terminals 306a and 306b in a proportion related to the charge state of the floating grating / waveguide 312. As described above, any suitable phase difference between the first arm 308 and the second arm 310 can be achieved by controlling the dimensions (e.g., length, width, shape, etc.) and the charge applied to the floating grating / waveguide 312 via programming, thus providing different power outputs through each of the optical output terminals 306a and 306b.

[0110] In this way, the non-volatile programmable passive optical structure 302 can be used as a non-volatile, low-power optical storage unit, an optical hidden layer node of a trained neural network, a non-volatile optical switch, etc.

[0111] and Figure 3A The embodiments shown are the same. Figure 3B The optical structure shown may optionally include a drain electrode 320 configured as described above or in another suitable manner. This description will not be repeated.

[0112] Generally, for the purposes of explanation, and to facilitate understanding of the various configurations and constructions of the non-volatile programmable passive optical structures described herein, are presented in Figures 3A-3B The foregoing embodiments and their various alternatives and variations are described herein. However, it will be apparent to those skilled in the art that some of the specific details presented herein may not be necessary to practice the particular described embodiments or their equivalents.

[0113] Therefore, it should be understood that the foregoing and following description of the specific embodiments are presented for limited purposes of illustration and description. These descriptions are not intended to be exhaustive or to limit this disclosure to the precise forms described herein. Rather, it will be apparent to those skilled in the art that many modifications and variations are possible in light of the foregoing teachings.

[0114] For example, as mentioned above, in some cases, optical structures that are not interferometer structures can be used. As an example, a ring resonator can be used. Figure 3C-3E An exemplary ring resonator construction is depicted.

[0115] Figure 3C A ring resonator 322 is depicted. The ring resonator 322 is positioned to be optically coupled to a waveguide 324, which receives light from an input terminal 326 and provides light as an output terminal 328. As those skilled in the art will know, the optical and mechanical properties of the ring resonator 322 define the frequencies at which the ring resonator resonates through a continuous path around it. For example, a larger ring resonator (with a larger radius or defining a longer path) resonates at different low frequencies than a smaller ring resonator. Furthermore, ring resonators of different structures can resonate with different harmonics. As those skilled in the art will know, a heater can be used to change one or more dimensions of the ring resonator through thermal expansion. For the sake of simplicity of description and illustration, Figure 3C It is depicted as having no heating element that can be used in some embodiments.

[0116] In addition to altering one or more mechanical characteristics of the ring resonator (e.g., ring resonator 322), or instead of altering them, the ring resonator 322 can be a floating grating / optical waveguide as described herein. More specifically, the ring resonator 322 can be electrically decoupled from other circuitry (as well as system and circuit ground), but optically coupled to waveguide 324. In this way, as with other optical structures described herein, the charge state of the ring resonator 322 defines which frequencies resonate through the ring resonator 322, and thus defines which frequencies are filtered and / or reflected within waveguide 324 due to optical coupling with the ring resonator 322.

[0117] In some configurations, the ring resonator 322 can be optically coupled to multiple waveguides. For example, as... Figure 3DAs shown, the ring resonator 322—which may also be a floating grating / waveguide as described herein and / or optically coupled to a floating grating / waveguide as described herein—may be optically coupled to (e.g., positioned close to) a second waveguide 332, which receives optical input through a second optical input terminal 332 and provides output through a second optical output terminal 334. In some cases, optical input through the second optical input terminal 332 may not be required; light from the first waveguide—waveguide 324—may be coupled into the ring resonator 322 to resonate at one or more frequencies and coupled into the second waveguide 330. In these configurations, very specific and narrowband light (e.g., light of a specific frequency) can be delivered to the second output terminal 334. In these configurations, as referenced above… Figure 3C The ring resonator 322 can be a floating gate / waveguide as described herein or can be optically coupled to a floating gate / waveguide as described herein. In this way, the charging state of the ring resonator 322, programmed as described above, defines which frequencies of light are output from the second output terminal 334, which frequencies of light are filtered out from the output terminal 328, and / or which frequencies of light are reflected back in the opposite direction to the input terminal 326. By changing the charge state, the power output and frequency output from each optical output terminal can be precisely controlled and maintained in a non-volatile manner.

[0118] In other cases, multiple ring resonators can be chained together and / or optically coupled. For example, such as Figure 3E As shown, the second ring resonator 336 can be optically coupled to the ring resonator 322. In this configuration, higher frequency harmonics resonating within the first ring resonator 322 can be coupled into the second ring resonator 336, and further into the second waveguide 330.

[0119] In some cases, as described herein, the floating grating / waveguide can be used in conjunction with one or more other control elements (such as heating elements).

[0120] Furthermore, in some cases, multiple non-volatile programmable passive optical structures can be combined to create higher-order programmable optical structures. For example, as described above, in some embodiments, multiple non-volatile programmable passive optical structures can be arranged in an optical network, and each can be uniquely programmed to impart different attenuations (different phase differences resulting from different programming) to different hidden layer nodes corresponding to the trained neural network.

[0121] In other embodiments, the non-volatile programmable passive optical structure described herein can be used to store configuration variables or parameters for configurable logic blocks or input / output blocks for field-programmable gate arrays. Figure 4A simplified system diagram 400 depicts a field-programmable gate array 402 that can utilize one or more non-volatile programmable passive optical structures for photonic circuits as described herein.

[0122] Specifically, in the depicted configuration, an array of optically configurable logic blocks can be implemented, one of which is designated as configurable logic block 404, each having at least one non-volatile programmable passive optical structure as described herein. In some examples, optical signal routing between different optically configurable logic blocks can be controlled by, for example, a bistable programmable passive optical structure as described above. In other cases, one or more input blocks / output blocks, such as input block 406 / output block 408, may include at least one non-volatile programmable passive optical structure as described herein.

[0123] In other implementations, non-volatile programmable passive optical structures can be used in other photonic and / or electronic circuits. For example, in one configuration, an array of non-volatile programmable passive optical structures can be coupled together to form a non-volatile optical memory array. In particular, each set of non-volatile programmable passive optical structures defining discrete memory cells can be mutually controlled using word line and bit line control paradigms used to control memory cell blocks in conventional flash memory implementations.

[0124] For example, Figure 5 This is a flowchart illustrating example operations of a method for operating a non-volatile programmable passive optical structure as a memory cell as described herein. Method 500 includes operation 502, in which a specific memory cell is selected from a set of memory cells. In some examples, the memory cell can be selected by address.

[0125] The method also includes operation 504, in which a control gate or control electrode associated with the memory cell can be driven to a specific voltage to change the charging state of the floating gate / waveguide associated with the memory cell.

[0126] Optionally, method 500 may also include operation 506, wherein the drain electrode / gate can be used together with the source electrode / gate to determine the charge state of the floating gate / waveguide. This may not be required in all embodiments.

[0127] Figure 6This is a flowchart illustrating an example operation of a method for operating a non-volatile programmable passive optical structure with photonic circuitry as described herein. Method 600 includes operation 602, in which a specific memory cell is selected from a set of memory cells, for example, by addressing. Furthermore, the selected memory cell is programmed to a specific charge state, or more generally, set to a specific digital value. Next, in operation 604, optical or photonic circuitry optically coupled through a floating gate / waveguide of the memory cell can be operated. More simply, light can pass through the floating gate, which may be affected by the charge state set in operation 602.

[0128] It will be understood that although many embodiments have been disclosed above, the operations and steps presented with respect to the methods and techniques described herein are exemplary and therefore not exhaustive. It will also be understood that, for a particular embodiment, an alternating sequence of steps or fewer or additional operations may be required or desired.

[0129] Although the foregoing disclosure has been described with reference to various exemplary embodiments and implementations, it should be understood that the various features, aspects, and functions described in one or more individual embodiments are not limited to their applicability to the particular embodiments in which they are described, but may be applied individually or in various combinations to one or more embodiments of the invention, whether or not those embodiments are described and whether these features are presented as part of the described embodiments. Therefore, the breadth and scope of the invention should not be limited by any of the exemplary embodiments described above, but is defined by the claims set forth herein.

[0130] As used herein, the phrase "at least one" preceding a series of items separated by the terms "and" or "or" modifies the entire list, not each member of the list. The phrase "at least one" does not require selection of at least one of each listed item; rather, the phrase allows for the meaning of at least one of any item, and / or at least one of any combination of items, and / or at least one of each item. For example, the phrases "at least one of A, B, and C" or "at least one of A, B, or C" each refer to only A, only B, or only C; any combination of A, B, and C; and / or one or more of each of A, B, and C. Similarly, it is to be understood that the order of elements presented in the conjunction or disjunction lists provided herein should not be construed as limiting this disclosure to only that order.

Claims

1. A semiconductor device comprising: a source electrode; a control electrode; an insulating layer; and a floating gate separated from the source electrode and the control electrode by the insulating layer, the floating gate optically coupled to a photonic circuit; wherein the source electrode and the control electrode are disposed above the insulating layer that encapsulates the floating gate, wherein the floating gate is a first waveguide; wherein the insulating layer includes a first portion that separates the source electrode from the floating gate and a second portion that separates the control electrode from the floating gate, and the second portion of the insulating layer is an oxide-nitride-oxide dielectric layer, wherein the nitride includes silicon nitride as a second waveguide. The floating gate includes a silicon optical waveguide.

2. The semiconductor device of claim 1, wherein, The floating gate is electrically isolated from the photonic circuit.

3. The semiconductor device of claim 1, wherein, 4. The semiconductor device of claim 1, further comprising a drain electrode separated from the floating gate by the insulating layer. The photonic circuit includes an interferometer or a ring resonator.

5. The semiconductor device of claim 1, wherein, The interferometer is a Mach-Zehnder interferometer.

6. The semiconductor device of claim 5, wherein, The interferometer includes:

7. The semiconductor device of claim 5, wherein, a first arm; and a second arm that includes the floating gate.

8. A semiconductor device comprising: a conductive element; a layered dielectric disposed below the conductive element, the layered dielectric including: a first silicon oxide layer formed in junction with the conductive element; a silicon nitride layer formed in junction with the first silicon oxide layer; and a second silicon oxide layer formed in junction with the silicon nitride layer; a silicon waveguide formed in junction with the second silicon oxide layer of the layered dielectric, the silicon waveguide: optically coupled to a photonic circuit; and electrically decoupled from the conductive element; wherein: in response to a voltage applied to the conductive element, the silicon waveguide accumulates charge by quantum tunneling, the accumulated charge causing a change in refractive index of the silicon waveguide, wherein the silicon waveguide is a first waveguide and the silicon nitride layer is a second waveguide. The silicon waveguide is a floating gate of a memory cell, the conductive element is a control electrode of the memory cell.

9. The semiconductor device of claim 8, wherein, The memory cell includes a source electrode separated from the silicon waveguide by an insulating layer, the source electrode configured to provide charge accumulated by the silicon waveguide in response to a voltage applied to the control electrode.

10. The semiconductor device of claim 9, wherein, 11. The semiconductor device of claim 8, wherein: the photonic circuit includes an interferometer that defines a first optical path and a second optical path; and the silicon waveguide is optically coupled into the first optical path. The silicon waveguide includes a first end that defines a first adiabatic taper and a second end that defines a second adiabatic taper.

12. The semiconductor device of claim 8, wherein, 13. A non-volatile semiconductor memory comprising: an array of memory cells, each memory cell including: a source electrode; a control electrode; an optical input; an optical output; a first waveguide that defines a first optical path that optically couples the optical input to the optical output; and a second waveguide that defines a second optical path that optically couples the optical input to the optical output, the second waveguide functioning as a floating gate, the second waveguide electrically decoupled from the source electrode and the control electrode, and configured to accumulate and hold charge provided by the source electrode in response to a voltage applied to the control electrode, such that the second optical path exhibits a different refractive index than the first optical path; ​ wherein the source electrode and the control electrode are disposed above an insulating layer encapsulating the floating gate; wherein the insulating layer includes a first portion separating the source electrode from the second waveguide and a second portion separating the control electrode from the second waveguide, and the second portion of the insulating layer is an oxide-nitride-oxide dielectric layer.

14. The non-volatile semiconductor memory of claim 13, wherein: a subset of the array of memory cells is conductively coupled by a word line; and each memory cell in the subset of the array is conductively coupled to a respective one bit line.

15. The non-volatile semiconductor memory of claim 13, further comprising a light source optically coupled to a respective light input of at least one memory cell in the array of memory cells.

16. The nonvolatile semiconductor memory according to claim 13, wherein, the first waveguide and the second waveguide define a Mach-Zehnder interferometer.

17. The nonvolatile semiconductor memory according to claim 13, wherein, at least one memory cell in the array of memory cells is part of a configurable logic block of a field programmable gate array.

Images