External cavity laser adopting multi-material micro-ring reflector

Abstract

There is provided an external cavity laser comprising: a photonic integration platform comprising a gain chip providing a gain medium and at least one reflector chip, the photonic integration platform comprising a first functional layer and a second functional layer made of different materials; a resonant cavity optically coupled to the gain medium and including first and second microring resonators (MRRs) in a vernier configuration, the first and second MRRs extending within respective ones of the first and second functional layers made of different materials and having respective group refractive indices, the first MRR and the second MRR have different radii selected according to the respective group refractive indexes; and a tuning mechanism for tuning a spectral response of at least one of the first MRR and the second MRR.

Description

Technical Field

[0001] This technical field is largely related to external cavity lasers. Background Technology

[0002] Widely tunable lasers with narrow linewidths are in high demand across a wide range of applications, including optical sensing, fiber optic communications, GPS clocks in space applications, and basic metrology. Semiconductor diode lasers, such as distributed feedback (DFB) lasers and distributed Bragg reflector (DBR) lasers, are attractive due to their small size, ease of mass production, and compatibility with integrated circuits. However, they are approaching their limits in terms of tuning range and spectral linewidth, with some offering only a few nanometers of tuning range and megahertz-level linewidths. A new type of laser, called an external cavity laser (ECL), can be created by placing the laser diode (or gain dielectric, essentially a laser diode coated with an anti-reflective coating on one or both sides) within an external cavity containing frequency-selective elements, further improving the tuning range and / or linewidth of the laser diode. With proper design, ECLs can provide a wider wavelength tuning range and narrower linewidth than a single laser diode.

[0003] ECLs typically combine a laser diode or gain medium with passive and active external feedback circuitry to form a Fabry-Perot (FP) cavity or ring cavity. The term "external cavity" refers to a configuration in which at least a portion of the feedback circuitry or cavity—including frequency selection elements, mirrors, or reflectors—is located outside the gain medium or a single laser diode; it can be single-sided or double-sided. This differs from a single-diode laser, where the cavity or feedback circuitry elements are integral components of the gain medium or the main cavity of the laser diode.

[0004] Currently, commercially available ECLs employ a variety of feedback circuit topologies, potentially incorporating free-space and fiber-based components such as blazed gratings, high-Q whispering resonators, and fiber Bragg gratings (FBGs). Recent advancements in photonic integration platforms have generated significant interest in ECLs that integrate frequency-selective elements within an external cavity and incorporate III-V gain media. Compared to traditional peer-to-peer solutions, this approach offers several advantages, including reduced size, weight, and cost. Furthermore, it provides an efficient pathway for the co-integration of ECLs with other highly functional component modules.

[0005] Various integrated ECL configurations have been demonstrated, differing in various aspects of the feedback circuitry, such as frequency selection elements, mirrors or reflectors, optical coupling systems, gain media, and the type of integration method. For example, ECLs on native III-V substrates (e.g., InP) have been shown, where the feedback circuitry and gain media are fabricated integrally on the same platform using a single process. Furthermore, with the rapid development of silicon photonics, recent integrated ECL demonstrations have shown the integration of silicon-based feedback circuitry with III-V gain media. These typically rely on two main integration methods: 1) heterogeneous integration, where the III-V gain material is chip-bonded to a passive waveguide for further front-end line processing at the wafer level; and 2) hybrid integration, where readily fabricated III-V gain material is attached to the passive feedback circuitry in the back end of an in-line assembly process. On the one hand, hybrid integration does not interfere with the front-end fabrication process flow of the passive external cavity circuitry and can therefore be applied to a variety of integration platforms. On the other hand, heterogeneous integration may be a clear path towards achieving scalability and mass production.

[0006] Regardless of the integration method employed, a key aspect of ECL is the design of the external cavity feedback circuit. This circuit is typically implemented using the vernier effect, either through a sampled Bragg grating reflector or, more commonly, a microring resonator (MRR), combined with other passive and active components. The design of the frequency selection element in the feedback circuit affects key laser parameters, including wavelength tuning range, linewidth, side-mode rejection ratio (SMSR), wavelength and single-mode laser stability, wavelength tuning power consumption, and to some extent, maximum output power. It is important to note that wavelength tuning power consumption refers to the energy consumed in tuning the wavelength through methods such as thermal tuning via metal or doped heaters, and / or through the plasmonic dispersion effect of a doped PN junction; while maximum output power represents the limitation that the mirror circuit may impose on the maximum output power due to non-ideal factors such as nonlinearity.

[0007] Figure 1A and 1B (Prior art) shows a typical structure of silicon-based ECL 100, which forms a Fabry-Perot cavity, for example, Tran, MA et al “ Ring-resonator based widely-tunable narrow- linewidth Si / InP integrated lasersAs shown in IEEE J. Sel. Top. Quantum Electron. 26, 1–14 (2020), both ECL 100s shown include a gain section 102 located between the passive cavity reflector 104 and the tunable cavity reflector 106, forming a frequency-selective feedback circuit. The tunable cavity reflector 106 includes a pair of microring resonators (MRRs) 108 and 110 cascaded in an additive-subtractive configuration. The MRRs can be placed in a ring configuration, such as... Figure 1A As shown, or a separate annular reflector 112 can be added at the end of the MRR cascade, as shown. Figure 1B As shown in the diagram. Broadband and / or tunable couplers are typically used to form the loop of the structure. The tunable cavity reflector 106 serves as a single-wavelength filter for the ECL 100, and the phase shifter 114 located between the gain segment 102 and the tunable cavity reflector 106 is used to tune the laser cavity mode. Figure 1C Another example of an ECL configuration is shown, where the vernier-based frequency feedback circuit consists of two MRRs 108 and 110 in a ring cavity. MRRs 108 and 110 are coupled via two bus waveguides 116a and 116b, with a gain region 102 situated between them. In all topologies, the cascaded MRRs 108 and 110 are designed in a vernier configuration, thus having slightly different radii, resulting in slightly different free spectral ranges (FSRs) to achieve the vernier effect and thereby shape the vernier spectrum required for wavelength tuning. The laser's optical output, whether as the primary output or the monitoring output, can be extracted from multiple points within the laser structure. These points include the high reflectivity (HR) mirror 104, the ports of the MRRs, and the couplers. For the latter, this is applicable if the coupler is asymmetric (not 50% / 50%) or if a tunable coupler is used.

[0008] As those skilled in the art will understand, the so-called vernier effect utilizes two (or more) resonators with two (or more) different resonant frequencies and slightly different FSRs, so that they only perfectly overlap at multiples of their respective FSRs. The FSR of light with wavelength l for a given MRR is given by the following equation. ,in FSR i It is the first i FSR of MRR, R i It is the radius of the MRR, and n g It is the group refractive index of the MRR waveguide. By convention, such as... Figure 1A and 1B The vernier FSR of the pair of MRRs shown is determined by the following formula: (1) (2) For any of the dual MRR vernier filters m , m and m- 1 is a coprime integer. m and m- 1 is called the resonance number of the corresponding MRR. m The improved tuning efficiency of the vernier MRR relative to the single MRR was also quantified, i.e. m ≈ FSR V / FSR 1 .

[0009] Figure 1D and 1E (Prior art) shows examples of the spectral response of the dual MRR mirror 106, such as Figure 1A and 1B As shown in any of the figures, its cross-coupling coefficient is = = 0.23, and MRR radius R 1 =32.5μm, R 2 =25μm. Figure 1C The individual spectra of MRR with a few nanometers of FSR are shown. Figure 1E The synthesized spectral response of the dual MRR mirrors shows a wider FSR ( FSR V ≈( m- 1) FSR 2 ≈ mFSR 1 The FSR cursor sets the wavelength tuning range of the ECL.

[0010] Various implementations of MRR-based reflectors using a variety of materials exist, including silicon, indium phosphide (InP), and other compound semiconductors. In particular, silicon-based reflectors are typically implemented on standard silicon-on-insulator (SOI) or silicon-nitride (SiN). The well-established SOI platform (often simply referred to as the silicon (Si) platform—these terms are used interchangeably in the art) offers particular advantages in ECL feedback circuits due to its compatibility with a wide range of active and passive silicon photonic devices, lower manufacturing costs, and smaller size.

[0011] Despite the numerous advantages of the SOI platform, it also faces some challenges in realizing ECLs. Standard Si photonic devices typically employ high-contrast waveguides, consisting of a silicon core surrounded by an oxide cladding that tightly confines the light within the core. This results in a compact ECL feedback circuit with small wiring bends, a low MRR radius, and moderate waveguide losses of approximately 1.5 to 3 dB / cm. ECLs with these waveguide losses exhibit linewidths between approximately 37 kHz and 27 MHz, and output power ranging from approximately 0 to 13 dBm; higher MRR quality factors result in lower linewidths and output power. To further reduce waveguide losses in the SOI platform, thereby reducing linewidth, J. Bowers' team employed a low-loss ridge waveguide (~0.2 dB / cm) in a heterogeneous integrated ECL, achieving a linewidth record of 220 Hz on the Si platform. The trade-off is the need for a larger MRR and more complex triple MRR mirrors to maintain broadband tunability [Tran, MA]. et al [As mentioned above]. In other existing technologies, a smaller 40Hz linewidth is achieved by using an ultra-low loss SiN waveguide (~0.1 dB / cm) in the feedback circuit. [Fan, Y. et al “ Hybrid integrated InP-Si3N4 diode laser with a 40-Hz intrinsic linewidth ", Opt. Express 28, 21713–21728 (2020); Y. Fan et al .,“ 290 Hz intrinsic linewidth from an integrated optical chip-based widely tunable InP-Si3N4 hybrid laser ”, in Proc. Conf. Lasers Electro-Opt., SanJose, CA, USA, Jun. 2017, p. 1].

[0012] Another challenge with Si-based feedback circuits is the nonlinear loss caused by two-photon absorption and the resulting free carrier absorption. Two-photon absorption ultimately limits the total output power and intracavity power of the ECL. This problem can also be overcome by employing SiN-based feedback circuits, such as, for example, Boller, K.-J. et al “ Hybrid integrated semiconductor lasers with silicon nitride feedback circuits As demonstrated in Photonics 7, 4 (2019), the authors developed a hybrid ECL that achieved a record-breaking high output power exceeding 20 dBm. Alternatively, high output power in a Si-based ECL can be achieved by utilizing an external SOA to amplify the laser emission (see, Y. Gao et al ., “ High-Power, narrow-linewidth, miniaturized silicon photonic tunable laser with accurate frequency control ”, J. Lightw. Technol., vol. 38, no. 2, pp. 265–271, 2020).

[0013] Another challenge in integrating gain and feedback circuitry is the waveguide mode mismatch between the gain and feedback chips. Si waveguides typically exhibit large mode mismatches, leading to high inter-chip coupling losses. To reduce losses below 2 dB, polymer spot size converters can be carefully designed on the Si chip, which introduces increased process complexity. Alternatively, inverse tapered structures can be designed on both the active and passive chips to minimize losses. Furthermore, photonic wire bonding techniques can be employed to further reduce coupling losses. [Xu, Yilin, et al . " Hybrid external-cavity lasers (ECL) using photonic wire bonds as coupling elements ", Scientific reports 11.1 (2021):16426].

[0014] For ECL, a reflector configuration based on MRR is still needed to mitigate some of the drawbacks of existing technologies. Summary of the Invention

[0015] According to one aspect, an external cavity laser is provided, comprising: Gain chips that provide gain media; A reflector chip comprising functional layers made of different materials; and The resonant cavity includes: A fixed-cavity reflector, optically coupled to one end of a gain chip; A tunable cavity reflector, disposed on a reflector chip and optically coupled to the end of a gain chip opposite to a fixed cavity reflector, comprises a first microring resonator (MRR) and a second MRR in a vernier configuration. The first and second MRRs extend in corresponding functional layers made of different materials and have corresponding group refractive indices. The first and second MRRs have different radii selected according to the corresponding group refractive indices. A tuning mechanism for tuning the spectral response of at least one of the first MRR and the second MRR.

[0016] According to one aspect, an external cavity laser is provided, comprising: a photonic integrated platform including a gain chip providing a gain medium and at least one reflector chip, the photonic integrated platform including a first functional layer and a second functional layer made of different materials; a resonant cavity optically coupled to the gain medium and including a first microring resonator (MRR) and a second MRR in a vernier configuration, the first MRR and the second MRR extending within a respective functional layer of the first functional layer and the second functional layer made of different materials and having a corresponding group refractive index, the first MRR and the second MRR having different radii selected according to the corresponding group refractive index; and a tuning mechanism for tuning the spectral response of at least one of the first MRR and the second MRR.

[0017] In some embodiments, the gain chip includes a semiconductor optical amplifier.

[0018] In some embodiments, the gain medium is based on a quantum well, a quantum wire, a quantum dot, or a quantum dot array.

[0019] In some embodiments, each of at least one reflector chip includes a substrate and a cladding extending on the substrate, with each of a first functional layer and a second functional layer embedded in the cladding of one of the at least one reflector chip.

[0020] In some embodiments, the substrate of the reflector chip is made of Si, and the cladding of the reflector chip is made of SiO2.

[0021] In some embodiments, the material of the first functional layer is Si, and the material of the second functional layer is SiN.

[0022] In some embodiments, the materials of the first functional layer and the second functional layer are each independently selected from a list including the following materials: silicon (Si), silicon nitride (SiN), indium phosphide (InP), gallium arsenide (GaAs), gallium nitride (GaN), aluminum nitride (AlN), indium gallium arsenide phosphide (InGaAsP), indium gallium arsenide antimony (InGaAsSb), mercury cadmium telluride (HgCdTe), gallium nitride (GaN), aluminum nitride (AlN), gallium arsenide (GaAs), aluminum gallium arsenide (AlGaAs), aluminum nitride (AlN), QXP, lithium niobate (LiNbO3), barium titanate (BaTiO3, BTO) or their binary, ternary and quaternary compounds and alloys.

[0023] In some embodiments, the ratio of the radii of the first MRR and the second MRR is proportional to the ratio of their group refractive indices.

[0024] In some embodiments, the ratio of the radii of the first MRR and the second MRR approximately corresponds to: , in R 1 It is the radius of the first MRR. R 2 It is the radius of the second MRR. n g1 It is the group refractive index of the material in the first MRR. n g2 It is the group refractive index of the material of the second MRR, and m It is a tuning enhancement factor.

[0025] In some embodiments, the ratio of the radii of the first MRR and the second MRR approximately corresponds to: , in R 1 It is the radius of the first MRR. R 2 It is the radius of the second MRR. n g1 It is the group refractive index of the material in the first MRR. n g2 It is the group refractive index of the material in the second MRR. m It is a tuning enhancement factor, and N It is a reduction factor related to the material of the first MRR.

[0026] In some embodiments, the resonant cavity has a Fabry-Perot configuration and includes a fixed cavity reflector optically coupled to one end of the gain medium; a first MRR and a second MRR form a tunable cavity reflector optically coupled to the end of the gain medium opposite to the fixed cavity reflector.

[0027] In some embodiments, the fixed cavity reflector includes a layer of reflective material extending along the surface of the gain chip.

[0028] In some embodiments, the resonant cavity has a ring configuration.

[0029] In some embodiments, at least one reflector chip is composed of a monolithic reflector chip integrating a first functional layer and a second functional layer.

[0030] In some embodiments, the external cavity laser includes a waveguide structure defining the resonant cavity, the waveguide structure including: a first MRR and a second MRR; one or more waveguide branches extending in each functional layer; and one or more optical vias that allow light to travel between functional layers of a monolithic reflector chip.

[0031] In some embodiments, at least one reflector chip includes a first reflector chip and a second reflector chip in a hybrid configuration, wherein the first reflector chip integrates a first functional layer and the second reflector chip integrates a second functional layer.

[0032] In some embodiments, the external cavity laser includes a waveguide structure defining the resonant cavity, the waveguide structure including: a first MRR and a second MRR; one or more waveguide branches extending in each functional layer; and one or more optical edge couplers or grating couplers that allow light to travel between the functional layers of the first reflector chip and the second reflector chip.

[0033] In some embodiments, the first functional layer is located in the gain chip, and the second functional layer is integrated in one of the at least one reflector chips.

[0034] In some embodiments, the external cavity laser includes a waveguide structure defining the resonant cavity, the waveguide structure including: a first MRR and a second MRR; one or more waveguide branches extending in each functional layer; and one or more optical vias that allow light to travel between one of the at least one reflector chip and a functional layer of the gain chip.

[0035] In some embodiments, the waveguide structure includes a coupler that splits light into clockwise and counterclockwise portions that travel in opposite directions through the waveguide structure.

[0036] In some embodiments, the waveguide structure includes a ring-shaped reflector that closes the feedback circuit.

[0037] In some embodiments, one of the first MRR and the second MRR is located inside the annular reflector.

[0038] In some embodiments, the first MRR and the second MRR are optically connected through at least two of the waveguide branches.

[0039] In some embodiments, at least one of the first MRR and the second MRR includes a plurality of MRR components.

[0040] In some embodiments, the waveguide structure includes a Mach-Zehnder interferometer integrated with at least one of the first MRR and the second MRR.

[0041] In some embodiments, at least one of the first MRR and the second MRR has a variable width portion.

[0042] In some embodiments, the external cavity laser further includes a frequency stabilization mechanism comprising: a monitoring photodetector configured to measure an estimated optical power output from the external cavity laser; a first heater thermally connected to a second MRR to tune the resonance of the second MRR based on the estimated optical power; and a second heater thermally connected to a first MRR to tune the resonance of the first MRR based on the estimated optical power.

[0043] In some embodiments, the distance between the first heater and the second MRR is less than the distance between the second heater and the first MRR.

[0044] In some embodiments, the external cavity laser further includes a 90-degree coupler that is optically coupled to a first MRR and a second MRR to measure the phase difference between the external laser cavity outputs.

[0045] In some embodiments, the first heater and the second heater are integrated with the second MRR and the first MRR monolith, respectively.

[0046] In some embodiments, the external cavity laser further includes a plurality of temperature sensors located between the gain chip and the resonant cavity, the plurality of temperature sensors being configured to monitor heat wave propagation, wherein a first heater and a second heater are adjusted based on the monitored heat transfer.

[0047] In some embodiments, the external cavity laser includes a plurality of lugs, each lug being connected to a corresponding one of a plurality of temperature sensors and extending through a first functional layer and a second functional layer.

[0048] In some embodiments, the external cavity laser further includes an auxiliary temperature sensor configured to measure external thermal disturbances.

[0049] Other features and advantages will be better understood after reading the detailed embodiments with reference to the accompanying drawings. Attached Figure Description

[0050] Figure 1A , 1B 1C (Prior Art) is a schematic diagram of an ECL configuration based on the prior art; Figure 1D and 1E These are the various spectra of MRR ( Figure 1D ) and vernier reflectance spectrum ( Figure 1E ).

[0051] Figure 2A This is a schematic diagram of an ECL configuration according to one embodiment; Figure 2B It includes Figure 2A A partially transparent side view of the configured photonic integration platform; Figure 2C It is along Figure 2B Cross-sectional view of the CC line; Figure 2D It includes Figure 2A A partially transparent side view of another photonic integration platform configuration.

[0052] Figure 3A This is a front view of the MRR in the SiN functional layer and the associated waveguide branches in the Si functional layer; Figure 3B This is a side view of a heater containing doped elements.

[0053] Figures 4A to 4C The changes in group refractive index and wavelength for Si and SiN waveguides are shown respectively. Figure 4A ), lies in R SiN When the value is fixed at 115 μm, the change in vernier FSR and SMSR varies with R Si Changes ( Figure 4B ), and simulated transmission spectra of hybrid Si-SiN mirrors.

[0054] Figure 5A and 5B These are a schematic diagram and a partially transparent side view of an ECL construction based on one embodiment.

[0055] Figures 6A to 6E They respectively show: for fixed R SiN ≈115μm, and N =3( Figure 6A middle =0.23, Figure 6C middle =0.2), vernier FSR and SMSR with R Si The relationship; for Figure 6B middle =0.23 and Figure 6D middle =0.2, simulated transmission spectrum of hybrid Si-SiN mirror; targeting N, and Different combinations of ECL linewidths serve as a function of detuning.

[0056] Figure 7A This is a schematic diagram of a Si-SiN MRR inside a ring-shaped mirror and interconnected by two waveguide branches; Figure 7B This is a schematic diagram of a Si-SiN MRR within a ring mirror, where the two output waveguide branches from the Si MRR are combined before being connected to the SiN MRR; and Figure 7C This is a schematic diagram of a Si-SiN MRR connected to each other through two waveguide branches and separated from the ring mirror.

[0057] Figure 8A and 8B This is a simulated transmission spectrum of a hybrid Si-SiN mirror with p output waveguide branches, where p=1 ( Figure 8A ) and p=2 ( Figure 8B ).

[0058] Figure 9 This is a schematic diagram of an ECL configuration using interconnected Si MRR ring components.

[0059] Figure 10A 10b and 10b are Si MRRs with integrated MZI ( Figure 10A ) and Si and SiN MRR integrating two MZIs ( Figure 10B ) is a schematic diagram of ECL configuration.

[0060] Figure 11A and 11B This is a schematic diagram of an ECL configuration including an MRR with a variable width portion.

[0061] Figure 12 It is a combination Figure 2A and Figure 5A A schematic diagram illustrating the advantages of ECL configuration.

[0062] Figure 13A This is a schematic diagram of an ECL annular cavity configuration according to one embodiment; Figure 13B Is to implement Figure 13A A partially transparent side view of the configured photonic integration platform.

[0063] Figure 14 This is a schematic diagram of a dual-wavelength laser source.

[0064] Figure 15 (a) to Figure 15 (b) shows a wavelength control method utilizing the multi-material properties of the proposed ECL.

[0065] Figure 16 An example of a distributed network of compact temperature sensors is shown for predicting and correcting frequency shifts caused by thermal disturbances, in this case, from RSOA.

[0066] Figure 17 This demonstrates how to use metal vias or lugs to increase the thermal resistance of the sensing path leading to the sensor.

[0067] Figure 18A and Figure 18B An example of a temperature sensor from a PIC that uses MOSFET transistors is shown, along with the sensor's layout within the PIC.

[0068] Figure 19This demonstrates how a temperature circuit sensor can be used to resolve thermal crosstalk between MRR heaters.

[0069] Figure 20 This is a block diagram of the ECL and related control components. Detailed Implementation

[0070] To provide a more concise description, some quantitative representations given herein may be limited by the word “approximately.” It is to be understood that, whether or not the word “approximately” is explicitly used, each quantity given herein is intended to refer to an actual given value, and also to a value that is reasonably inferred from ordinary techniques in the art, including approximations due to experimental and / or measurement conditions of the given value.

[0071] In this specification, the term "about" means within an acceptable range of error for a particular value as determined by a person skilled in the art, the range of error depending in part on the method of measurement or determination of that value, i.e., the limitations of the measurement system. Measurements with 10% accuracy are generally considered acceptable and are covered by the term "about".

[0072] In this specification, when a wide range of values ​​is provided, any possible narrower ranges within the boundaries of that wide range should also be considered. For example, if a wide range of values ​​from 0 to 1000 is provided, any narrower range between 0 and 1000 should also be considered. If a wide range of values ​​from 0 to 1 is mentioned, any narrower range between 0 and 1 (i.e., the decimal part) should also be considered.

[0073] According to one aspect, an external cavity laser, hereinafter referred to as ECL, is provided. ECLs according to embodiments described herein can be used in various application environments, such as optical sensing, fiber optic communication, Global Positioning System (GPS), space clock applications, basic metrology, etc. According to one aspect, as further described below, the ECLs described herein offer a new perspective by integrating multiple materials into an MRR-based vernier frequency selective reflector in the ECL design. In some embodiments, MRR-based ECLs are described herein that leverage the advantages of different material platforms in the feedback circuit to enhance performance and broaden application range.

[0074] In some implementations, the ECL includes a photonic integration platform, a resonant cavity, and a tuning mechanism. As understood by those skilled in the art, integrated photonics refers to an on-chip technology in which at least one component with optical functionality is provided. The photonic integration platform includes a gain chip and at least one reflector chip. Various non-limiting configurations of the photonic integration platform are described below. As will also be explained below, the photonic integration platform includes a first functional layer and a second functional layer made of different materials. The first and second functional layers may both extend on one of the chips of the photonic integration platform, typically on the reflector chip, or may be disposed on two separate chips.

[0075] ECL Configuration refer to Figure 2A , 2B Figures 2C and 2C schematically illustrate an ECL 20 according to one embodiment. As previously described, the ECL 20 includes a photonic platform 21, which in this example includes a gain chip 22 and one or more reflector chips 30. While these illustrative examples serve only as a basis for describing the components of an ECL, it will be readily understood that the same considerations and options are applicable to other variations, such as those described further below.

[0076] The gain chip 22 of the ECL 20 provides a gain medium configured to amplify light. The gain chip 22 can have any structure known in the art. For example, depending on the desired wavelength, the gain chip can be implemented by a semiconductor optical amplifier (SOA) using a direct bandgap III-V medium (e.g., InP, etc.). In some embodiments, the gain chip includes a gain waveguide 24 implementing the gain medium. The gain medium can be quantum well-based, quantum dot-based, or quantum dot-based. It will be readily understood that in other variations, the gain medium can be implemented by any material or structure that amplifies light by stimulated emission, the choice of which determines the operating wavelength of the laser. SOAs, such as those based on InP, are commonly used in lasers operating in the 1300-1600 nm wavelength range. Fabry-Perot diode lasers can also be used as the gain medium for the ECL. In other examples, gain media without semiconductors, such as EDFA, can be used. The gain medium can have a bandwidth covering the C-band, L-band, O-band, or any other wavelength range, thereby providing amplification at the desired wavelength, enabling the ECL to produce laser light at the desired output wavelength.

[0077] ECL 20 also includes a resonant cavity 26 optically coupled to the gain medium 22. In the illustrated variant, the resonant cavity has a Fabry-Perot configuration, also known as a linear cavity. It is well known that a Fabry-Perot cavity can be defined by a pair of optical reflectors or mirrors positioned at opposite ends of the gain waveguide 24. In some implementations, for example in… Figure 2A , 2B As shown in the variants of 2C, the pair of optical reflectors may include: a fixed cavity reflector 28, which is optically coupled to one end 23 of the gain waveguide 24; and a tunable cavity reflector 32, which is optically coupled to the end 25 of the gain waveguide 24 opposite to the fixed cavity reflector 28.

[0078] Those skilled in the art will understand the term "fixed cavity reflector" to mean a non-tunable or passive reflector. In some implementations, such as in... Figure 2A and Figure 2B As shown, the fixed-cavity reflector 28 can be implemented by a layer of reflective material deposited on the surface of the corresponding end 23 of the gain chip 22, thereby realizing a reflective semiconductor optical amplifier (RSOA) design. However, it will be readily understood that the fixed-cavity reflector can also alternatively be located outside the gain chip, for example, including a separate mirror, a Sagnac ring mirror (see...). Figure 5A )wait.

[0079] As described above, the illustrated variant of ECL 20 includes a reflector chip 30 on which a tunable cavity reflector 32 is disposed. The reflector chip 30 is located outside the gain chip 22, which is an inherent feature of ECL design. The reflector chip 30 can be implemented using a multi-layered SOI structure. In some implementations, such as... Figure 2C As best seen, the reflector chip 30 may include a substrate 60, for example made of Si, with a cladding 62, for example made of SiO2, on which a cladding layer 62 is disposed. In a typical embodiment, the reflector chip includes one or more functional layers 31a, 31b of different materials embedded in the cladding 62. These different materials may be silicon-based, such as Si, SiN, etc. Figure 2B and 2CThe embodiments shown, as well as several other figures herein, illustrate Si functional layers 31a and 31b. However, those skilled in the art will readily understand that any embodiments disclosed herein, and variations thereof, can be implemented using different combinations of materials as functional layers 31a and 31b. In some embodiments, non-silicon materials, such as indium phosphide (InP), gallium arsenide (GaAs), gallium nitride (GaN), aluminum nitride (AlN), and their respective binary, ternary, and quaternary compounds and alloys, can be used. For example, indium gallium phosphide (InGaAsP) is used in the CO band (1200 nm–1700 nm), indium gallium antimony arsenide (InGaAsSb) and mercury cadmium telluride (HgCdTe) are suitable for wavelengths above 1700 nm in the mid-infrared region, gallium nitride (GaN), aluminum nitride (AlN), and their alloys (e.g., InGaN, AlGaN) cover the ultraviolet to blue light range (100 nm–400 nm), and gallium arsenide (GaAs) and aluminum gallium arsenide (AlGaAs) are used in the visible to near-infrared range (400 nm–1200 nm). Furthermore, materials such as aluminum nitride (AlN), silicon nitride (SiN), QXP, lithium niobate (LiNbO3), and barium titanate (BaTiO3) (BTO) are attracting attention due to their wider coverage across multiple wavelength regions; for example, they also provide larger Pockels coefficients and, in the latter two cases, slow down wave propagation. Any combination of two or more of these materials (including Si) can be used to form MRR-based vernier reflectors, such as InP and SiN, AlN and SiN, etc. In another example, one of the functional layers 31a and 31b can be made of QXP, also known as high-refractive-index silica glass or Hydex (see example, Kues, M., Reimer, C., Roztocki, P. et al. On-chip generation of high-dimensional entangled quantum states and their coherent control. Nature 546,622–626 (2017)).

[0080] In some implementations, the first functional layer 31a and the second functional layer 31b can be located on the same reflector chip, or on two different reflector chips. Figure 2B and 2C In the illustrated variant, at least one reflector chip is composed of a monolithic reflector chip 30 integrating both the first functional layer 31a and the second functional layer 31b. In other variants, such as, for example Figure 2DIn the variant shown, the at least one reflector chip may include a first reflector chip 30a and a second reflector chip 30b in a hybrid configuration, wherein the first reflector chip 30a integrates a first functional layer 31a and the second reflector chip 30b integrates a second functional layer. Other configurations are also possible.

[0081] In some embodiments, a waveguide structure 33 may be provided in the reflector chip 30 for receiving, guiding, and reflecting light to form an ECL feedback circuit, thereby performing the reflection function of the tunable cavity reflector 32. The waveguide structure 33 may typically include one or more waveguide branches 43, 46 in functional layers 31a, 31b. These branches are designed to guide light through the tunable cavity reflector 32 and are coupled together in various possible configurations, and coupled to a first MRR and a second MRR (defined below), several of which are listed below. Light can be coupled between the gain chip 22 and the reflector chip 30 in any manner known in the art. Other optical elements may also be integrated into the tunable cavity reflector 32, such as one or more couplers 40, phase shifters 42, etc.

[0082] The tunable cavity reflector 32 includes a first microring resonator (MRR) 34 and a second MRR 36. As their names suggest, the first MRR and the second MRR each define a ring waveguide provided as part of the waveguide structure 33 in the reflector chip 30, each forming a ring waveguide with a corresponding radius. R i The closed loop is formed. MRR 34 and MRR 36 preferably extend substantially parallel to the substrate 60 and are respectively disposed in different functional layers 31a and 31b of the reflector chip 30. In some embodiments, the first MRR 34 and the second MRR 36 may extend from the substrate 60 at different heights h. In one variant, for example, as shown in the figure... Figure 2B As shown, the first MRR 34 and the second MRR 36 can be substantially or partially stacked on top of each other. In other implementations, such as... Figure 5B As shown, the first MRR 34 and the second MRR 36 can extend laterally away from each other.

[0083] The first and second MRRs are located within the reflector chip 30 to perform the reflection function of the tunable cavity reflector 32. For example, Figure 2A and Figure 2BThe following configuration is shown, in which the first MRR 34 and the second MRR 36 are positioned in a ring mirror configuration. It will be apparent that this configuration is only one way of implementing the tunable cavity reflector 32, and in other variations, the different components of the tunable cavity reflector may be arranged differently and may be located in different functional layers 31 than shown here, depending on design considerations and target specifications. In this configuration, the waveguide structure 33 of the tunable cavity reflector 32 includes an input-output waveguide 41 disposed in the Si functional layer 31a and optically coupled to the gain waveguide 24 via a tilted taper spot size converter 39. An optional phase shifter 42 may be provided to tune the cavity mode. The waveguide structure 33 then includes a coupler 40 that optically couples the input / output waveguide 41 to a first Si layer waveguide branch 43a and a second Si layer waveguide branch 43b, which are also disposed in the silicon functional layer 31a. The first optical via 44a optically couples the first Si waveguide branch 43a located in the Si functional layer 31a with the first SiN layer waveguide branch 46a located in the SiN functional layer 31b and optically coupled to the first MRR 34. The second optical via 44b allows light to travel between the second SiN layer waveguide branch 46b and the third Si layer waveguide branch 43c, wherein the second SiN layer waveguide branch 46b is located in the SiN layer 31b and is also optically coupled to the first MRR 34, and the third Si layer waveguide branch 43c is located in the Si functional layer 31a and is optically coupled to the second MRR 36. The second MRR 36 is finally optically coupled to the second Si layer waveguide branch 43b.

[0084] In this configuration, light from gain chip 22, entering tunable cavity reflector 32, travels from input / output waveguide 41 to coupler 40, where it is then split into two optical portions that travel clockwise and counterclockwise through waveguide structure 33, respectively. The clockwise portion travels sequentially through first Si-layer waveguide branch 43a, through first optical via 44a into first SiN-layer waveguide branch 46a, and then around first MRR 34. The clockwise portion then couples into second SiN-layer waveguide branch 46b, through second optical via 44b into third Si-layer waveguide branch 43c, around second MRR 36, and returns to coupler 40 through second Si-layer waveguide branch 23b. The counterclockwise portion travels from second Si-layer waveguide branch 43b through the same components in the reverse order. The clockwise and counterclockwise light components are combined by coupler 40 and finally output from tunable cavity reflector 32 through input / output waveguide 41.

[0085] As those skilled in the art will understand, the optical output of the ECL, whether as the main output or the monitoring output, can be provided from different locations within the laser structure to extract the light therein. For example, the optical output can be provided at any port of the fixed reflectors 28, MRR 34, and MRR 36, or at coupler 40. For the latter, this approach is applicable when the coupler is an asymmetric coupler (not 50% / 50%) or, if a tunable coupler is used.

[0086] It will be readily understood that references to the “first” MRR and the “second” MRR in the context of this specification are for convenience only and are not intended to assign any particular order or importance to these MRRs. Furthermore, it will be readily understood that an ECL may contain more than one MRR, and that MRRs may be configured in a variety of other ways. More examples are provided below.

[0087] As described above, the vertical interlayer transition between MRRs for light transmission can be achieved using optical vias (e.g., thermally insulating tapered portions). The term "optical via" can be understood as any device or structure that allows light to travel between different functional layers, similar to the use of electrical vias in electronic chip design. In some embodiments, such as Figure 2B As shown, the optical via 45 can be implemented using a SiN-Si double-layer transition structure, such as that shown in Figure 6 (see Figure 6). Wang, Pengfei et al The entire contents of the paper “Design and fabrication of a SiN-Si opticalphased array chip”, Photonic Research, Vol. 8, Issue 6, pp 912-919 (2020) are hereby incorporated by reference. In other implementations, optical coupling between different functional layers can be achieved through direct vertical coupling of light without the need for typical optical vias. For example, see [link to relevant documentation]. Figure 3A An embodiment is shown in which a SiN functional layer 31b is stacked on top of a Si functional layer 31a, and a SiN ring defining an MRR 34 is formed in the SiN functional layer 31b. A first Si layer waveguide branch 43a and a second Si layer waveguide branch 43b are formed in the Si functional layer 31a. In this variant, the Si layer waveguide branches 43a and 43b are directly optically coupled to the SiN ring 34 located in the upper layer 31b. The coupling strength k is controlled by adjusting the gap 47 between the Si and SiN functional layers 31a and 31b through the Si waveguide branches 43a and 43b. Light from the first Si layer waveguide branch 43a is coupled into the SiN MRR 34, resonates in the ring, and then couples to the bottom second Si layer waveguide branch 43b. For example, Li, Xin, et al Another embodiment of optical coupling between a waveguide and an MRR in two different materials is shown in “Ultra-low-loss multi-layer 8x8 microring opticalswitch”, Photonics Research, Vol. 11, Issue 5, pp.712-723 (2023), the entire contents of which are incorporated herein by reference.

[0088] It will be easy to understand that optical coupling between different functional layers can also be performed in other ways. For example, in Figure 2D In the hybrid configuration shown, one or more optical edge couplers or grating couplers can be provided to allow light to travel between the functional layers of the first reflector chip 30a and the second reflector chip 30b.

[0089] According to one aspect, and as stated above, the first MRR 34 and the second MRR 36 may be made of different materials. Figures 2A to 2C In one embodiment, the first MRR 34 is made of Si, while the second MRR 36 is made of SiN. Therefore, the reflector chip includes a Si functional layer and a SiN functional layer carrying their respective MRRs. The first MRR 34 and the second MRR 36 are preferably formed with different radii. R i and different effective refractive indices n effi Group refractive index n gi The toroidal waveguide is determined by the material and waveguide structure of the corresponding functional layers 31a and 31b. Note that the MRR does not need to be a perfect circle; they can have any closed-loop shape, with a perimeter satisfying the target radius, i.e. L i =2πR i 。 Each MRR supports [the function] within the corresponding radius. R i Group refractive index n gi Light propagation at a determined resonant frequency (i.e., wavelength). It is well known that the intervals between the resonant frequencies of each MRR define the corresponding free spectral range of that MRR, given by the following formula. .

[0090] The first MRR 34 and the second MRR 36 are arranged in a vernier configuration. As described above, the vernier FSR of a pair of MRRs ( FSR V Based on the respective FSRs of the two MRRs (see Equation (1))， The wavelength tuning range of the ECL was also defined. In the case of an ECL based on an MRR provided by two different materials, their group refractive index difference will affect the above considerations. In some implementations, a first MRR and a second MRR with different radii are selected according to their respective group refractive indices. In one embodiment, the radius of the MRR can be proportional to the ratio of its group refractive indices. For example, for a SiN-on-Si platform, the radius of the MRR can be approximately expressed as: (3) Therefore, the vernier FSR is: (4) in, R si 、R SiN Let be the radius of MRR, and n g_Si 、n g_SiN , respectively, are the group refractive indices of Si and SiN waveguides.

[0091] Restating equation (3) and applying it to different materials not specified, the ratio of the radii of the first MRR and the second MRR may in some variants correspond to essentially the following: (3') in R 1 It is the radius of the first MRR. R 2 It is the radius of the second MRR. n g1 It is the group refractive index of the material in the first MRR. n g2 is the group refractive index of the material in the second MRR, and m is the tuning enhancement factor.

[0092] Restating equation (3) and applying it to different materials not specified, the ratio of the radius of the first MRR to the radii of the remaining MRRs—which may be greater than 1—may in some variations correspond to: ,in, (3”) in R 1 It is the radius of the first MRR. R i+1 It is the radius of the (i+1)th MRR. n g1 It is the group refractive index of the material in the first MRR. n g(i+1)The group refractive index of the material of the (i+1)th MRR, m1, and the other resonance numbers of the corresponding MRR (m i+1 () are coprime numbers.

[0093] See Figure 4A The figure shows the relationship between the group refractive index and wavelength for Si and SiN waveguides, which were calculated using the standard mode solver. The group refractive index of the strip Si waveguide is approximately twice that of SiN, which in turn determines the corresponding MRR radius, for example, using equation (3) above. Figure 4B Displayed in fixed R SiN With a value of 115μm, for R Si An example of vernier FSR variation (i.e., wavelength tuning range variation). The estimated Si radius is close to... R Si_est ≈ ≈57.5μm (assuming a large value) m , making ≈1). The cross-coupling coefficient of this simulation is =0.23, with waveguide losses of 2 dB / cm and 0.1 dB / cm for Si and SiN, respectively. The tuning range is within... R Si = R Si_est The point exhibits a vertical asymptote, that is, R Si Approaching from both sides R Si_est hour, FSR V It tends towards infinity. However, R Si_est It does not necessarily mean R Si The optimal point should be considered because the SMSR of the feedback circuit should also be taken into account. For example... Figure 4B As described in the text, when R Si near R Si_est At this point, the vernier SMSR approaches zero, and reaches zero when all transmission resonances of the two MRRs are aligned with each other. When calculating the vernier SMSR, we calculate the height of the first adjacent mode of the fundamental mode. Maximization or optimization can be chosen. FSR V Both the and the SMSR vernier R Si Value. Different design factors can be considered. For example, for stable laser operation, the ECL feedback circuit design may need to maintain the vernier SMSR above 3-5 dB. Therefore, it is best to choose... RSi This value allows the SMSR to exceed 3 dB. In another example, the radius of the MRR can be chosen to ensure a higher value. FSR V A proper balance is achieved between the value and the acceptable cursor SMSR value. Figure 4C The simulated reflection spectrum of the tunable cavity reflector is shown, where the radius of the Si MRR is based on... Figure 4B The selected curve in the graph ( R SiN =115μm) in order to achieve 10 dB SMSR, thus leading to selection R Si ≈55μm. As shown in the simulated spectrum, this selection achieves a tuning range of approximately 50nm. The peak height at the common resonance represents the loss that occurs during propagation through a pair of MRRs. In this example, we obtain... T IL A value of approximately 1.3 dB.

[0094] In some implementations, the radius of the MRR is chosen to be large enough to neglect bending loss, while being small enough to minimize waveguide loss. Once the radius of one of the MRRs is determined, the other can be estimated using equation (3). As mentioned above, in Figures 4A to 4C In the example, the radius of the SiN MRR is first set large enough to ignore bending losses. R SiN =115μm), and then calculate the radius of Si MRR according to equation (3). R Si (≈55μm). However, for MRR radii larger than approximately 10-30μm, the bending loss of Si waveguides is usually negligible. Therefore, choosing an excessively large... R Si Values ​​can lead to additional waveguide losses but do not provide any additional benefit regarding bending losses. The situation can be even worse if a larger SiN MRR radius is required, or if a material with a lower refractive index than SiN is used (resulting in a larger Si MRR radius). To address these issues, in some embodiments, an improved version of the multi-material vernier MRR design can be used. With careful design, this improvement can also provide additional advantages. The improved FSR of dual MRRs on the SiN-on-Si platform is given by the following equation: (5) Where N is the Si MRR reduction factor relative to Equation 3. The radius of the MRR can be approximately expressed in the corrected form as follows: (6) Therefore, in some embodiments, the ratio of the radii of the first MRR and the second MRR can substantially correspond to the following in some variations: (6') in, R 1 It is the radius of the first MRR. R 2 It is the radius of the second MRR. n g1 It is the group refractive index of the material in the first MRR. n g2 is the group refractive index of the material of the second MRR, m is the tuning enhancement factor, and N is the reduction factor related to the material of the first MRR.

[0095] Reference Figures 2A to 2C The ECL 20 also includes a tuning mechanism for tuning the spectral response of at least one or both of the first MRR 34 and the second MRR 36. In some implementations, the thermo-optical (TO) effect is utilized to tune and switch adjacent and non-adjacent wavelength channels. For example, the tuning mechanism may include a heater 50 located above each MRR 34, 36 to tune their respective resonances. Figure 2B and 2C As best seen in the illustrated embodiment, each heater 50 includes a metal resistance thermometer 52 through which current flows to generate heat, and the temperature of the resistor changes the refractive index of the semiconductor material. The metal used as the heater should have a higher resistivity than Cu and Al, which are typically used for wiring / pads and metallization. TiN or tungsten (W) wire is typically used above the waveguide. In the illustrated implementation, for example, the metal resistance thermometer 52 is implemented as a TiN arch structure extending into the cladding 62, covering a portion of the associated MRR, its arch shape matching a portion of the underlying MRR ring. A metal electrode 54 is disposed above the cladding 62 of the reflector chip and connected to the end of the TiN arch structure 52. A metal via 56 provides electrical connection between the electrode 54 and the TiN arch structure 52. The heaters can be controlled together for continuous wavelength tuning within the FSR of a single MRR, or individually for discrete wavelength tuning in steps of the FSR of a single MRR.

[0096] Reference Figure 3BIn some implementations, heater 50 may include N-type or P-type doped silicon elements 58 to alter the refractive index of the Si or SiN waveguide. A shorter distance between the heater and the waveguide can increase the temperature of the MRR waveguide, thereby changing its refractive index, while requiring lower electrical power than the metal electrodes. Similar to metal heaters, Si-based heaters can also generate heat through bias and cause a change in the refractive index of nearby waveguides via a thermo-optic effect. The Si-doped heater can be positioned next to the MRR waveguide, or in some implementations, the MRR waveguide itself can act as a resistor. Other methods utilizing thermo-optic effects to change the refractive index of the MRR waveguide can also be employed.

[0097] ECL configuration variants See Figure 5A and 5B The figure shows Figure 2A A variant of the ECL configuration is provided, in which a separate annular mirror 48 is used to close the loop formed by the tunable cavity resonator 32. In this configuration, the waveguide structure 33 of the tunable cavity reflector 32 includes a first SiN layer waveguide branch 46a optically coupled to a gain waveguide 24. No coupler is required in this variant, and no coupler is needed at the entrance of the tunable cavity reflector 32. A phase shifter 42 is used to tune the cavity mode. The waveguide structure 33 also includes a first MRR 34 disposed in the SiN functional layer, with a second SiN layer waveguide branch 46b opposite to the first SiN layer waveguide branch 46a coupled thereto. An optical via 44 optically couples the second SiN layer waveguide branch 46b in the Si functional layer to the first Si layer waveguide branch 43a in the Si functional layer. A second MRR 36 disposed in the Si functional layer is optically coupled to both the first Si layer waveguide branch 43a and the second Si layer waveguide branch 43b, and optically coupled to the annular mirror 48 via an optical coupler 49.

[0098] In this configuration, light entering the tunable cavity reflector 32 first travels into the first SiN waveguide branch 46a, from which it is directly coupled to the first MRR 34. The light then travels from the first MRR 34 to the second SiN waveguide branch 46b, passes through the optical via 44 to the first SiN waveguide branch 43a, and from there enters the second MRR 36. The light travels from the second MRR to the second SiN waveguide branch 43b and enters the annular mirror 48 through the coupler 49. The light then reflects back along its original path until it exits the tunable cavity reflector 32 through the first SiN waveguide branch 46a.

[0099] It will be noted that, Figure 5A In the example shown, the fixed cavity mirror 28 is also implemented using a Sagnac annular mirror, although the ROSA design can also be used in other variants.

[0100] Figure 6A The fixation at 115 μm was shown. R SiN In the case of values, for R Si Examples of cursor FSR and SMSR, such as in Figure 2A The ECL configuration shown simulates a scenario where Si and SiN are used as the functional layer materials carrying the first and second MRRs. The cross-coupling coefficient is... =0.23. Using the above equation (6) and N =3, the estimated Si radius was determined to be close to R Si ≈ The diameter of the Si waveguide is approximately 19.2 μm, which is about 1 / 6 of the radius of the SiN MRR waveguide. A smaller Si radius can improve the thermal insulation performance of SiN and Si MRR waveguides when they are stacked on top of each other, such as... Figure 2B As shown. The vernier FSR and vernier SMSR are... N =1 follows the same pattern ( Figure 4B For the vernier SMSR calculation, the height of the third adjacent mode is calculated, since the first two adjacent modes are completely distinguishable by Si MRR, given... FSR Si ≈ N × FSR SiN . Figure 6B The simulated transmission spectrum of the tunable reflector is shown, with the goal of achieving a ~10 dB SMSR. R Si ≈18.4μm, which yields a wavelength tuning range of approximately 60 nm. By maintaining =0.23 and decrease With a value of 0.2, the wavelength tuning range can be further optimized. Figure 6C The resulting change in vernier SMSR relative to the Si MRR radius is shown, while the vernier FSR remains constant. However, in this case, to achieve a 10 dB SMSR, the vernier FSR can be increased to as high as 85 nm, such as... Figure 6D As presented, it can be inferred that employing a variety of materials in the design of MRR-based ECLs would provide a wider range of parameters, thereby optimizing the performance of all lasers, as will be further elaborated below.

[0101] Based on The RSOA characteristics and equations used in Tran, MA et al., “Ring-resonator based widely-tunable narrow-linewidth Si / InP integrated lasers”, IEEE J. Sel. Top. Quantum Electron. 26, 1–14 (2020) (hereinafter referred to as Tran et al.), were used to estimate the ECL linewidth. For this purpose, the method described by Tran et al. was followed. Considering the inter-chip coupling of 2 dB and the aforementioned waveguide losses, we estimated the linewidth of the Si-SiN ECL feedback circuit. N =1 and the cross-coupling coefficient value With a value of 0.23, we observed a linewidth of approximately 3.3 kHz at the 1.4 GHz detuning. For N Decreasing the radius of the Si MRR by 2 results in a decrease in the Q factor of the Si MRR, with the linewidth only increasing to 5.5 kHz. In this case, increasing the cross-coupling coefficient to 2... Further reducing the Q factor of Si MRR to 0.35 results in a slight increase in linewidth to 7.8 kHz. However, if we also increase the cross-coupling coefficient of SiN to... When the value is 0.35, the linewidth increases to 22 kHz, indicating that the SiN MRR has a significant impact on the linewidth value. Further optimization can be performed to further reduce the linewidth.

[0102] To maximize output power, multi-material ECLs may still face some limitations due to their reliance on Si MRRs. However, in such ECLs, the ECL parameters can be partially decoupled from each other, allowing each MRR to contribute to optimizing specific ECL parameters. Therefore, a primary goal may be to leverage SiN MRRs to maximize ECL performance. Doing so reduces the requirements for Si MRRs, thereby enabling increased output power through various techniques described below.

[0103] Increasing optical power in a Si MRR waveguide leads to nonlinear losses due to two-photon absorption (TPA) and subsequent free carrier absorption (FCA). To quantify these effects and the impact (at least approximatedly) of techniques for increasing output power, we base our analysis on... Quantitative estimation was performed using the method reported in Xu, Yilin, et al., "Hybrid external-cavity lasers (ECL) using photonicwire bonds as coupling elements." Scientific reports 11.1 (2021): 16426 (hereinafter referred to as Xu et al.). 。 The maximum output power can be estimated based on two parameters: P max ( dBm )≈ NL ( dBm )- M ( dB ),in, NL It is the length and the length of the feedback circuit ( L equ The nonlinear threshold of an equivalent straight Si waveguide with the same waveguide cross-section (here, 500 nm × 220 nm) is given, and M It is the power enhancement factor in MRR relative to the power in the corresponding bus waveguide, which can be defined as: (7) in, t It is a self-coupling coefficient, and α It is the single-pass amplitude transmission rate. The first term in equation (7) is the enhancement of the optical portion entering the MRR in the counterclockwise direction. The second term in the above equation is the enhancement of the optical portion entering the target MRR in the clockwise direction after passing through another MRR. Here, we assume that the RSOA can still provide sufficient power after coupling and feedback circuit losses, so the output power is mainly limited by the nonlinear threshold. Xu et al. reported that the on-chip power is approximately NL At approximately 19 dBm (80 mW), signs of nonlinear losses such as TPA and TPA-FCA appeared. Therefore, when =0.23( t = ≈0.973), α ≈0.994 and M ≈14.4 dB, estimated maximum power is approximately P max ( dBm )≈4.6 dBm .

[0104] In some implementation schemes, to improve the output power of the ECL, one or both of the following strategies can be adopted: reducing the power enhancement factor of the Si MRR. M and / or increase the nonlinear threshold NLThis paper proposes several methods consistent with these two strategies. In the first, more intuitive method, we can reduce the... This can improve SiN MRR. M This allows us to maintain the low linewidth of ECL. Simultaneously, we can significantly improve... To reduce Si MRR M However, we cannot increase indefinitely. This is because it reduces the SMSR and thus the wavelength tuning range. To establish a baseline for observing the impact of the following techniques, we used an example. Figure 6A and Figure 6B Parameters in =0.23). Given the same L equ and NL Under these conditions, the maximum power was indeed achieved to be approximately [value missing]. P max ( dBm )≈4.6 dBm, Similar to the results obtained by Xu et al., this setting also yielded SMSR≈10. dB and a linewidth of ~5.5kHz. By modifying =0.4 and =0.15 will increase the maximum power by 3.9 dB, that is P max ( dBm )≈8.5 dBm Meanwhile, it provides an SMSR of approximately 3.3. dB And a linewidth of ~2kHz. Further optimization may be needed to find the optimal point.

[0105] In the second method, methods such as Figure 7A The feedback circuit design is shown. In this design, to close the feedback loop, the two MRR 34 and 36 are interconnected through more than one waveguide branch. Figure 7A In the variant, the waveguide structure 33 of the tunable cavity mirror 32 and Figure 2A The example is similar, and includes a first Si layer waveguide branch 43a, a second Si layer waveguide branch 43b, and a third Si layer waveguide branch 43c, a first SiN layer waveguide branch 46a and a second SiN layer waveguide branch 46b, and a first optical via 44a and a second optical via 44b for coupling light between the Si and SiN functional layers, all of which together constitute a... Figure 2AA similar feedback circuit. Furthermore, an auxiliary optical via 44c is disposed between the first MRR 34 and the second MRR 36, and coupled to it via a fourth Si layer waveguide branch 43d and a third SiN layer waveguide branch 46c. When each MRR 34 and 36 is considered as a temporary storage element for optical power, it is filled with the input waveguide branch that brings about RSOA (e.g., Figure 7A The first SiN layer waveguide branch 46a and the second Si layer waveguide branch 43b in the configuration. Adding more output waveguide branches at the MRR output port can be seen as providing additional paths for the optical power to leave the tunable cavity reflector 42, which may result in a decrease in the M value of Si.

[0106] The power enhancement factor in Equation 7 can be adjusted to account for the number of output waveguides from which optical power is extracted from the Si MRR. p ),like (8) in, t Si1 and t SiN1 These are the self-coupling coefficients between the input waveguide branch and the Si and SiN MRR, respectively. t Si2 and t SiN2 This represents the self-coupling coefficient of the output waveguide branches associated with Si and SiN MRR, specifically represented here by the third Si layer waveguide branch 43c and the fourth Si layer waveguide branch 43d. This equation can be further generalized for cases where each output waveguide branch has a different coupling coefficient. p In the case of two output waveguide branches with a value of 2, such as Figure 7A As shown, where =0.25, =0.3, and =0.15, =0.2, we can reduce M Si And increase the maximum output power to P max ≈0.95 dBm It also offers linewidths of approximately 3 dB and ~14 kHz with an SMSR of approximately 3 dB. Furthermore, in this example, by appropriately adjusting Δ... L = L 2- L 1, where L1 and L2 are the waveguide lengths of the two branches. A phase shifter exists in the L2 path, which allows us to generate interference between the two paths, producing twice the effect. FSR V The FSR allows us to double the tuning range, such as Figure 8Aand 8B As shown in the image.

[0107] We can also implement the same principle in different embodiments, such as Figure 7B and 7C The two examples shown. For example, Figure 7B Implementation examples and Figure 2A The difference in this embodiment is that it adds a fourth Si layer waveguide branch 43d, which is connected to the second MRR 36 from the third Si layer waveguide branch 43c using a coupler 40'. Figure 7C In the embodiments, with Figure 5A Similar to the embodiment, the first MRR 34 and the second MRR 36 are interconnected by two sets of waveguide branches 43a, 43b and 46b, 46c, which are linked by optical vias 44a, 44b.

[0108] Reference Figure 9 Figures A and 9B show another example of an ECL configuration, which in some implementations can increase the output power limit and extend the tuning range to over 100 nm, while achieving a linewidth close to 10 kHz. Figure 9 In the ECL configuration of A, the second MRR comprises two ring components 36a and 36b, a portion of which is used to separate light along two distinct paths using two Y-shaped branches 37a and 37b and a waveguide cross 45. This configuration reduces the light intensity within the ring, thereby reducing nonlinearity. Furthermore, the phase tuners placed on each branch are used not only to tune the ring resonant wavelength, but their combination also functions as a Mach-Zehnder interferometer within the ring to suppress the next resonant mode and increase the wavelength tunable range.

[0109] Reference Figure 10A and 10B The figure shows another tunable cavity design aimed at increasing the tuning range, reducing the vernier SMSR, and improving the output power limit. In this configuration, the MZI 70 is integrated with one of the MRRs, such as a second MRR 36 extending in the Si functional layer, so that the waveguide of the second MRR 36 is used as one of the MZI arms. The MZI configuration is formed when light is split into two paths by a coupler and enters two waveguides with a length difference of DL = L2 - L1, and then they are recombined by a coupler. The MZI 70 produces a periodic spectral response, which can be used to control the overall spectral response of the vernier configuration. For example, in Figure 10AIn the process, the light in the Si waveguide branch 43b is partially split into two paths. One part enters the second (Si)MRR 36, and the other part enters the section of the Si waveguide branch 43b that passes through DL1. Then, it is recombined with the light from the second MRR 36 in the coupler region, thus forming the MZI. This is achieved by optimizing the parameters. , Δ L 1 and Δ L 2. We anticipate that the proposed integrated MZI MRR structure can potentially reduce M This achieves a lower SMSR value while maintaining a wider tuning range. Note that the same design can also be implemented in a single-silicon ECL. Figure 10B Another example is shown, in which MZI 70a and MZI 70b are formed in both the first MRR 34 and the second MRR 36.

[0110] The second method to increase the output power level involves increasing the nonlinear threshold. NL :according to P max ( dBm )≈ NL ( dBm )- M ( dB See also Figure 11A and Figure 11B We can achieve this by using a variable width ( W The second (Si)MRR (i.e., increasing the effective Si waveguide cross-sectional area) in segment 35 is used to improve the nonlinear threshold in the second (Si)MRR 36. NL In the illustrated embodiment, the variable-width portion overlaps with the heater, which may help increase the interaction volume between the heater and the Si MRR, thereby improving the efficiency of the thermo-optical phase transition. However, it will be understood that different arrangements can be considered. In this MRR structure, the waveguide cross-sectional area forming the second MRR ( A CS It is designed to reduce light intensity without interfering with single-mode coupling from the bus to the MRR coupling region. The width of the second MRR 36 is W = W 0 From the initial angle ( —Bus to MRR coupling region—up to the output angle (OA), at which light is coupled out from the waveguide forming the second MRR 36, for example in Figure 11A middle, Counterclockwise. W 0 The choice is made to ensure single-mode coupling between the bus and the MRR, and between the MRR and the bus, in this example. W0 = 500 nm. After OA, the waveguide width thermally extends to its maximum value. W max At this point, the rate of change of width gradually stabilizes. Subsequently, the width adiabatically decreases to... W 0 As it approaches the initial angle IA This variation in waveguide width can be represented by an appropriate function, such as a trigonometric function, a polynomial function, or an exponential function. Other adiabatic MRR structures have been applied to Si. [Yang, Shuyu, et al. "A single adiabatic microring-based laser in 220nm silicon-on-insulator." Optics express 22.1 (2014): 1172-1180] and SiN ECL [Corato-Zanarella, Mateus, et al. "Widely tunable and narrow-linewidth chip-scale lasers from near-ultraviolet to near-infrared wavelengths." Nature Photonics 17.2 (2023): 157-164], the entire contents of which are incorporated herein by reference, aims to reduce propagation loss by minimizing the overlap between the optical mode and the sidewalls. Nevertheless, the design disclosed herein can be specifically addressed for nonlinear problems and provides for... NL The possibility of improvement up to approximately 3 dB. Compared to conventional MRR, adiabatic MRR also exhibits improved robustness against manufacturing variations and requires less power to thermally tune the resonant wavelength. Adiabatic MRR can be used in different embodiments of ECL configurations, another example being... Figure 11B As shown.

[0111] Return to reference Figure 2A and Figure 5A The figure shows two typical basic configurations of multi-material ECL, in which: on the one hand, two MRRs are placed inside a ring mirror (verb configuration 1, see Figure 2A On the other hand, a separate annular reflector (vernier configuration 2, see...) is added to the end of the cascaded MRR. Figure 5A ). Figure 12 A third basic configuration is shown, combining the advantages of the first two configurations, especially in ECL designs using multi-material MRR. In this configuration, waveguide structure 33 and Figure 5A Similar to the structure in the previous configuration, light from gain chip 22 first encounters the first SiN layer waveguide branch 46a, couples into the first MRR 34, and then exits through the second SiN layer waveguide branch 46b. Optical via 44 provides a bridge to the first SiN layer waveguide branch 43a, leading to coupler 49. Here, the second MRR 36 is positioned after coupler 49 and forms part of the annular mirror 48. Therefore, light arriving at coupler 49 is split into clockwise and counterclockwise portions, which couple into the second SiN layer waveguide branch 43b and the third SiN layer waveguide branch 43c, respectively, both of which are optically coupled to the second MRR 36. In effect, one MRR is located outside the annular mirror, while the other is added inside. Therefore, while using the same multi-material and cross-coupling coefficient in the three basic configurations, this new configuration benefits from the lower linewidth, lower cavity loss, and higher stability of vernier configuration 1.

[0112] The example above shows the resonant cavity of an ECL in a Fabry-Perot configuration. Reference Figure 13A and 13B In other embodiments, the resonant cavity may have a ring configuration. In one example of a ring configuration, waveguide structure 33 may form a ring connected to the two ends 23 and 25 of gain waveguide 24. In the illustrated variant, a first functional layer 31a is located in gain chip 22, for example, made of III-V material; a second functional layer 31b is integrated in reflector chip 30, for example, made of SiN. The second (III-V) MRR 36 and gain dielectric 24 are monolithically fabricated on the same III-V platform. The III-V platform is then heterogeneously integrated onto the SiN platform housing the first (SiN) MRR 34. In this example, MRRs 34 and 36, along with the gain dielectric and phase shifter 42, constitute a ring cavity. For example, coupling between the III-V and SiN waveguides can be achieved using optical vias 44a and 44b.

[0113] Reference Figure 14The figure shows a dual-wavelength laser source 300. This laser source can be constructed from two independently tunable cavity reflectors 32a and 32b, defining an adjustable vernier filter that can be used to provide two independently tunable wavelengths. Dual-wavelength sources have been extensively studied using various different approaches. In the scheme proposed in this paper, both wavelengths use the same gain chip 22 as the gain medium. A tunable coupler 302 provides coupling between the gain chip 28 and the corresponding tunable cavity reflector 32. In some implementations, to counteract spectral compression to a single wavelength through gain competition, a thermally tunable Mach-Zehnder coupler (not shown) can be used to adjust the relative strength of the feedback from the vernier circuit. However, the tuning speed (i.e., the kilohertz modulation speed range) of the thermally tunable coupler can cause instability in the laser output of one or both wavelengths. In other implementations, for example... Figure 20 In the implementation shown, the high-speed tunable coupler 302 loaded with a doped PN junction for dual-wavelength multi-material ECL allows tuning speeds up to GHz levels, thereby achieving stable dual-wavelength laser output.

[0114] Frequency control In some embodiments, reference Figures 15 to 20 The external cavity laser also includes a frequency stabilization mechanism. This mechanism includes: a monitoring photodetector configured to measure an estimated optical power output of the external cavity laser; a first heater thermally connected to a second MRR for tuning the resonance of the second MRR based on the estimated optical power; and a second heater thermally connected to the first MRR for tuning the resonance of the first MRR based on the estimated optical power. In some embodiments, the distance between the first heater and the second MRR is less than the distance between the second heater and the first MRR. In some embodiments, the external cavity laser also includes a 90-degree coupler optically coupled to the first and second MRRs for measuring the phase difference between the outputs of the external laser cavity. In some embodiments, the first and second heaters are monolithically integrated with the second and first MRRs, respectively. In some embodiments, the external cavity laser also includes a plurality of temperature sensors located between the gain chip and the resonant cavity, configured to monitor heat wave propagation, wherein the first and second heaters are adjusted based on the monitored heat transfer. In some embodiments, the external cavity laser also includes a plurality of lugs (e.g., see...). Figure 17 (The element labeled 72 in the attached figure), each lug is connected to a corresponding temperature sensor among the plurality of temperature sensors and extends through the first and second functional layers. Note that... Figure 17 The MRR, temperature sensor, and resistors in the external cavity laser are similar to those described above. In some embodiments, the external cavity laser also includes an auxiliary temperature sensor configured to measure external thermal disturbances.

[0115] In some implementations, tunable ECLs may require high-precision control of their frequency and demand both short-term and long-term stability. For example, in coherent communication systems, both the amplitude and phase of the optical signal are used to encode information. Any instability in the laser frequency can lead to errors in the received signal, negatively impacting the overall performance of the coherent communication system. In this regard, the ECL embodiments disclosed herein can leverage the thermal stability advantages of low-refractive-index silicon materials, such as, for example, SiN or QXP, while maintaining the fast switching or tuning time of Si and reducing tuning power. This can result in more robust laser operation, exhibiting better short-term and long-term stability. However, it will be readily understood that the same advantages can be achieved using other multi-material combinations, such as III-V materials and their corresponding binary, ternary, and quaternary compounds or alloys, for example, InP and InGaAsP, to name just a few. In the examples below, SiN, with a refractive index lower than silicon, is used in combination with Si as an example. Note that while the following paragraphs only refer to SiN, other similar materials can also be used.

[0116] At room temperature (approximately 300 K), the thermal conductivity (TC) of Si and SiN is approximately 150 W / m / K and 55 W / m / K, respectively. Therefore, the thermal conductivity of Si is approximately three times that of SiN. At room temperature, the thermo-optical coefficient (TOC) of Si and SiN is approximately 1.86 × 10⁻⁶. - 4 K -1 and 2.45×10 -5 K -1 Therefore, SiN requires a much higher temperature than Si to produce the same phase shift. SiO2 has a TOC of approximately 1 × 10⁻⁶. -5 K -1 Combining the thermal advantages of both materials, the proposed hybrid Si-SiN ECL provides stable and fast switching / tuning laser operation suitable for a wide range of applications. For example, given its ability to provide high output optical power, the proposed hybrid MRR structure enables fast wavelength switching / tuning, for example, but not limited to, which could be of great significance in LiDAR applications. Faster wavelength tuning gain is achieved by coarse tuning via a Si MRR heater followed by fine tuning via a SiN MRR heater. Thus, due to TOC... Si Compared to TOC SiN About an order of magnitude higher, temperature change ΔTThis can cause a shift in the resonant wavelength of a Si ring filter that is several times greater than that of SiN. Silicon dioxide is the primary heat transfer medium used in heat transfer systems. Using the aforementioned heater, SiN can be placed closer to its associated heater, thus contributing to operation with lower power consumption. Thermal crosstalk can be ignored if sufficient distance is provided between the MRRs.

[0117] Reference Figure 20 The figure shows a block diagram of the components of an ECL according to a variant example and its associated control elements. A control unit is provided, and a multi-channel ADC and DAC in a general-purpose form are embedded within it. The minimum configuration for wavelength tuning may contain only a PD and a temperature sensor. The box labeled "Heating 1" is the first MRR 34 (made of SiN) heater 50, while the box labeled "Heating 2" is the second MRR 36 (made of Si) heater 50. Note that in this document, "Heating 1" and H... SiN They can be used interchangeably, such as "Heating 2" and H. Si They can also be used interchangeably. The box labeled "MPD 1" is the monitoring photodetector 202 used to read the estimated output optical power from the ECL. Note that in this document, PD and MPD are used interchangeably to refer to the monitoring photodetector. In the illustrated embodiment, there are five temperature sensors, labeled "TS 1" through "TS 5"; each sensor can refer to any passive or active device embedded inside or mounted on top of the micro MRR. Also note that in this document, TS 1 and T... SiN They can be used interchangeably, TS 2 and T SiThey can also be used interchangeably. In practice, to accurately measure the temperature of a silicon waveguide, ideally, a temperature sensor (e.g., a thermistor) needs to be directly integrated onto the silicon ring resonator. However, this direct integration significantly increases optical propagation loss due to the interaction between the optical mode within the silicon ring waveguide and the thermistor material. To mitigate this problem, the temperature sensor can be placed above the silicon ring resonator and isolated by an upper silicon dioxide cladding; or, using doped silicon, it can be placed parallel to the silicon ring resonator on the same silicon device layer. Known examples of the latter are I (or V)-PTAT circuits (where the current (or voltage) is proportional to the absolute temperature) and I (or V)-CTAT circuits (where the current (or voltage) is complementary to the absolute temperature) for temperature sensing. The techniques involved in these examples outline a general approach to accurately setting and controlling the frequency, stability, and accuracy of multi-material ECLs. Integrated ECLs are particularly sensitive to thermal disturbances after initial setup. These disturbances can arise from sudden or gradual changes in package temperature and / or variations in gain dielectric current, which can have a relatively significant impact on the frequency accuracy of the ECL due to the high thermal sensitivity of the MRR. Existing techniques can stabilize or adjust back to the initial spectral position, although the stabilization time is relatively long, which is not ideal for many applications. Note that an important step before switching stabilization algorithms is to collect contour plots of the proposed hybrid ECL, i.e., the current values ​​of the heaters of the first and second MRRs and the corresponding derived wavelengths.

[0118] Varying conditions (e.g., laser current variations of hundreds of milliamps and wide ambient temperature fluctuations throughout the frequency tuning range and the entire device lifetime) have traditionally hindered or limited the real-time, near-real-time, or faster recovery of the initial spectral characteristics of a signal while maintaining high frequency accuracy. The multi-material approach described in this paper addresses this challenge. One solution is to utilize the multi-material properties of the proposed ECL. In the case of ECLs based on single-material MRRs, when thermal perturbations occur, the responses of the two MRRs shift with almost the same magnitude and in the same direction, such as... Figure 15 As shown in (a)-(b). Therefore, without using a thermal sensor, MPD alone may not be able to detect, measure, and correct wavelength shifts. Figure 15 (c) The first row of the figure shows the transfer function of each MRR at the resonant frequency under a set temperature T0 and a temperature variation ±ΔT, while the second row shows the vernier transfer function under the same temperature variation. A significant advantage of ECLs based on multi-material MRRs (e.g., Si and SiN) is that, due to different TOCs, the filter response offsets of the two MRRs are not the same under a given temperature perturbation, such as... Figure 15As shown in (d). This effect can be utilized to detect the power drop in the laser output by monitoring the MPD and adjusting the wavelength accordingly. Furthermore, the direction of the offset can be detected by utilizing the difference in the TOC of the two MRRs, for example, by measuring the phase difference between the ECL outputs using a 90-degree hybrid coupler connected to the MRR ports. Figure 15 As shown in (d). Figure 15 As shown in (b), this detection cannot be achieved using a single-material ECL.

[0119] In some embodiments, a distributed network of compact temperature sensors can be used to refine frequency control in order to predict and correct frequency shifts caused by thermal disturbances from RSOA or other external sources. For example, such as Figure 16 As shown, temperature changes and heat wave propagation within the RSOA can be effectively monitored using temperature sensors located near the RSOA and close to the MRR, allowing for adjustment of the metal heaters associated with the first and second MRRs. The heat wave velocity in the SOI can be calculated from real-time or near-real-time temperature readings from the temperature sensors (compared to the dynamics of the heat wave and control loop). Knowing the distance between the temperature sensors, temperature changes can be accurately predicted and compensated for through robust real-time feedback or correction, thereby dynamically adjusting the heaters associated with the first and second MRRs. This approach can help maintain frequency accuracy and / or stability in integrated ECLs, which may be necessary or desirable in many applications, including, but not limited to, optical coherent communication systems.

[0120] In some embodiments, metal connected to the temperature sensor and passing through the lugs or via walls of the SiO2 layer can be used to improve the sensing speed of the temperature sensor. This configuration allows for a more efficient internal heat transfer process because, in the PIC, the thermal conductivity of the metal is much higher than that of SiO2 in both the vertical and horizontal directions, in order to compensate for disturbances from other external sources.

[0121] Several methods exist to utilize the current (or equivalent, power) of the MRR heaters to tune the laser. The control unit is configured to calculate the current pair that achieves minimum power consumption. One known solution to improve heater power efficiency is to use air trenching and / or air undercutting to isolate each MRR, while improving ECL controllability by further reducing thermal crosstalk between MRRs. Implementing trenching and undercutting incurs additional cost and may not be ideal because it can reduce the yield of the final chip. In this case, another sensor ( Figure 20A sensor, designated "TS 5", is used to sense thermal crosstalk between the two rings. Combined with two proximity sensors, TS 1 and TS 2, on the ring, the crosstalk effect can be mitigated. TS 4 is positioned between RSOA and the ring, but at a greater distance, to indicate the ambient temperature as a reference. Later, in the crosstalk cancellation algorithm, it can be used to counteract the influence of the reference on the calculation of the optimal heater current pair. Utilizing feedback circuits made of various materials with different TOCs can help achieve a temperature-insensitive ECL.

[0122] In some embodiments, auxiliary temperature sensors located near each sensor and / or via wall are used to detect external thermal disturbances (e.g., convection). The control unit can predict the next value of the heater current based on information collected from all thermal sensors. In some embodiments, integrated sensors can be used instead of thermal sensors with via walls. In some embodiments, the via wall is located between two MRRs, and auxiliary adjacent sensors without via walls are provided in the control unit. This configuration allows thermal crosstalk between the heaters of the MRRs to be calculated without interfering with the waveguide optical mode, which can then be used to adjust the wavelength in response to disturbances. This distributed approach can be used to quickly predict the wavelength shift direction by the direction of temperature change, and thus adjust the heater current accordingly.

[0123] To realize a distributed network of compact sensors, it is feasible to utilize the thermal properties of doped semiconductor electronics, such as transistors. Unlike previous approaches that involved building devices on dedicated electronics platforms and then connecting them to the ECLPIC in a hybrid configuration, recent advances have shown that bipolar junction transistors (BJTs) and metal-oxide-semiconductor field-effect transistors (MOSFETs) can be integrated into SOI technology. By designing these transistors to optimize their thermal properties, it is feasible to design and monolithically fabricate temperature sensor circuitry on the same SOI platform as ECL. This approach allows for the placement of sensors near the MRR (Medium Temperature Regulator), enabling precise measurement of localized temperatures within the ECL PIC. Figure 18A -B shows an example of a temperature sensor circuit schematic for laying out MOSFET transistors on an SOI platform, demonstrating the feasibility of the method.

[0124] Integrating these sensors directly onto the SOI platform and placing them near the MRR offers several advantages. First, it reduces the overall system size. Second, it enhances thermal management by providing precise local temperature monitoring. Third, it allows for compensation for thermal crosstalk between the heaters of each MRR, where a heater in one MRR might also affect the refractive index change of another. As mentioned above, previous studies have typically used deep trenches and undercuts for temperature isolation. This type of solution increases design complexity and manufacturing processes and cannot completely eliminate thermal crosstalk effects.

[0125] In some embodiments, for example Figure 19 The illustrated embodiment uses a temperature circuit sensor to solve this problem. In this method, the temperature from sensor T... SiN and T Si (like Figure 16 The two temperature-dependent currents (shown in the diagram) are subtracted from each other and used to eliminate crosstalk effects by controlling the heater. It is important to note that this cancellation (or subtraction) is achieved using a coefficient called α. This coefficient represents H... SiN For T Si The impact of H. Si When closed, it will be disconnected from the output stage of T1 via a switch. For simplicity, it is... Figure 19 As shown in the image. T SiN and T Si This can be an I (or V)-PTAT circuit or an I (or V)-CTAT circuit for sensing temperature. These current-dependent temperature sensors are connected together in the output stage, and their currents will be subtracted, such as... Figure 19 As shown in the figure, this diagram illustrates how a temperature circuit sensor can be used to address the problem of thermal crosstalk between MRR heaters. It goes without saying that if other deterministic heat sources exist around the MRR (e.g., heat sources from the RSOA), the same technique can be used to eliminate the crosstalk effect.

[0126] Following a similar approach, other integrated circuits can also be integrated into the SOI platform to build more robust ECLs. For example, a transimpedance amplifier (TIA) can be directly added to the SOI platform where the ECL is fabricated. By placing the TIA next to the monitoring MPD, the MPD current can be amplified directly at the source. Although current applications require low speeds of around kHz, integrating the TIA on the SOI platform can improve performance by reducing noise and parasitic effects. This approach ensures that initial amplification occurs near the MPD, minimizing signal attenuation, before further amplification is performed off-chip using a conventional amplifier. Compared to the conventional method of wire-bonding the MPD to the TIA within a CMOS chip, this integration on the SOI platform provides better noise performance and lower parasitic losses. Furthermore, the temperature effects of the TIA can be mitigated by slotting or integrating a temperature sensor, ensuring stable performance.

[0127] Furthermore, integrating the TIA circuitry into the PIC can resolve margin issues associated with the low-voltage power supply of advanced CMOS nodes. By level-shifting the TIA output for CMOS compatibility, the PIC's compatibility with advanced CMOS technologies is enhanced.

[0128] Monitoring circuitry and drivers employing metal-semiconductor field-effect transistors (MESFETs) and BJT components can be integrated into the SOI platform. These devices offer high current drive capability and are compatible with advanced CMOS processes that may not support BJTs or MESFETs. This integration not only improves our system's compatibility with advanced nodes but also allows for more precise temperature control and reduced thermal crosstalk.

[0129] The electronic circuitry integrated into the SOI PIC can be used to characterize process variations in active components, such as modulators, MPDs, and heaters with PN junctions, for example, to monitor doping concentration variations between chips and wafers. These circuits can detect and quantify changes in doping concentration, which is crucial for ensuring the long-term consistency and reliability of active components. This information is invaluable for calibrating thermal sensor and power readings from the MPD. By accurately understanding and compensating for process variations, the accuracy of the ECL is improved. For example, if a specific doping concentration variation is identified, the calibration parameters of the MPD and temperature sensors can be adjusted accordingly, resulting in more accurate measurements and improved overall system performance. This approach ultimately simplifies the calibration process and ensures reliable operation of our ECL across different production batches.

[0130] Furthermore, for long-term stability and performance, aging effects can be compensated for using integrated circuits that facilitate the recalibration of the PIC and coefficients (e.g., regarding temperature sensors, MPDs, etc.). As components age, the properties of the PN junction and metals change due to heat and current. These aging effects can be monitored using circuits designed to detect performance changes. Implementing such aging monitors in the PIC can provide real-time feedback on component health, enabling timely recalibration and maintaining optimal performance. One concern regarding the application of these aging monitoring circuits is their requirement for a large number of pads on the chip, which may make them seem impractical. However, by designing electronic multiplexers or switches in the PIC, multiple circuits can be connected to fewer pads, simplifying the integration process and reducing the complexity of interfacing with external circuitry.

[0131] Now that the general concept of frequency control has been presented, it will be described in more detail. Figure 20 The embodiments presented herein. The boxes marked "TS" represent monitoring the temperature near the first MRR 34. In some embodiments, coarse adjustment can be based on readings from this temperature sensor, while for fine wavelength adjustment, a more precise and faster sensing device (e.g., MPD1) is used to readjust the required current to force the laser back to its locked mode. Furthermore, by Figure 20The “TS” box in the figure indicates the temperature near the second thermal sensor monitoring gain chip 22. In some embodiments, a thermoelectric cooler can be used to remove heat from the combined chip, thereby maintaining the desired temperature. The PID controller can be virtually embedded in the control unit, or a more efficient algorithm (e.g., a finite state machine (FSM), fuzzy controller, or model predictive control) can be considered to implement the tuning process and lock the feedback system to the correct wavelength in a short time.

[0132] In some implementations, to further overcome the challenges of low thermal tuning speeds of approximately tens of kHz and increased power consumption of approximately tens of mW for wavelength tuning, one strategy is to embed a PN junction in the Si MRR core waveguide using p-type and n-type doped Si, and to achieve electro-optic tuning using plasmonic dispersion effects in addition to the thermo-optic mechanism. Using an additional electro-optic tuning mechanism that relies solely on Si reflection can significantly improve the tuning speed and enable more degrees of freedom to control the system. However, doping Si in the core waveguide increases losses and reduces the requirement for narrow linewidths. To address this issue, it is feasible to redesign the SiN or QXP ring to achieve an ultra-high quality factor (Q) and significantly reduce the linewidth.

[0133] Advantages and Applications The ECL described herein, implemented using Si and SiN as the first and second functional layers, advantageously combines the compactness and compatibility of the Si platform with the low propagation loss and high nonlinear threshold of the low-refractive-index silicon platform. By combining these properties, the resulting ECL offers enhanced static performance compared to Si-based ECLs, such as narrower linewidth, wider tunable range, and higher output power (e.g., using a relatively low Q-value Si MRR and a higher Q-value SiN MRR). Furthermore, embodiments of the ECL benefit from the thermal stability advantages of low-refractive-index silicon while maintaining the fast switching / tuning time and low tuning power of Si. Advantageously, in some implementations, the laser output power can be directly coupled to passive / active devices (e.g., modulators or photodiodes) via Si waveguides. SiN waveguides can be used to minimize mode mismatch and optimize optical coupling between the gain chip, fiber, and silicon chip, thereby improving overall performance and integration efficiency.

[0134] Advantageously, the ECL embodiments described herein, employing a combination of multiple materials, offer broad applications across various fields based on their combined performance characteristics. In communications, the potentially lower linewidth, higher output power, and short- and long-term stability make it ideal for integrated intensity-modulated direct detection (IM-DD) and coherent silicon photonic transceivers, potentially operating across a wide spectral range including the C, L, and O bands. In sensing, for example, ECLs can utilize SiN MRRs as interfaces for biomolecular detection, providing value in environmental monitoring, medical diagnostics, and industrial process control. For LiDAR, the potential high power and fast tuning of ECLs can facilitate accurate, high-resolution distance measurements, showing great promise in autonomous vehicles and 3D mapping. In quantum optics, SiN ECLs have recently been demonstrated to generate tunable quantum light sources, enhancing quantum information processing systems. In some implementations, ECLs can meet the unique requirements of SOI and SiN quantum circuits, potentially enabling precise manipulation of the quantum states of light.

[0135] Finally, to seamlessly realize this multi-material, multi-layer ECL in practice, existing manufacturing processes from various foundries can be utilized, integrating multiple silicon materials such as SiN / Si-on-InP, SiN-on-Si, and QXP-on-Si. Therefore, the proposed method can provide a practical and easily implemented ECL, potentially advancing quantum computing, communication, and sensing technologies.

[0136] Of course, many other modifications can be made to the above embodiments without departing from the scope of protection defined by the appended claims.

Claims

1. An external cavity laser, comprising: A photonic integrated platform comprising a gain chip providing a gain medium and at least one reflector chip, the photonic integrated platform comprising a first functional layer and a second functional layer made of different materials; The resonant cavity is optically coupled to the gain medium and includes a first microring resonator MRR and a second MRR in a vernier configuration. The first MRR and the second MRR extend within corresponding functional layers of the first and second functional layers made of different materials and have corresponding group refractive indices. The first MRR and the second MRR have different radii selected according to the corresponding group refractive indices. A tuning mechanism for tuning the spectral response of at least one of a first MRR and a second MRR.

2. The external cavity laser according to claim 1, wherein, The gain chip includes a semiconductor optical amplifier.

3. The external cavity laser according to claim 1 or 2, wherein, The gain medium is based on quantum wells, quantum wires, quantum dots, or quantum dot arrays.

4. The external cavity laser according to any one of claims 1 to 3, wherein, Each of the at least one reflector chip includes a substrate and a cladding extending on the substrate, with each of the first functional layer and the second functional layer embedded in the cladding of one of the at least one reflector chips.

5. The external cavity laser according to claim 4, wherein, The substrate of the reflector chip is made of Si, and the cladding of the reflector chip is made of SiO2.

6. The external cavity laser according to any one of claims 1 to 5, wherein, The first functional layer is made of Si, and the second functional layer is made of SiN.

7. The external cavity laser according to any one of claims 1 to 5, wherein, The materials for the first and second functional layers are each independently selected from the list of materials including: silicon (Si), silicon nitride (SiN), indium phosphide (InP), gallium arsenide (GaAs), gallium nitride (GaN), aluminum nitride (AlN), indium gallium arsenide phosphide (InGaAsP), indium gallium arsenide antimony (InGaAsSb), mercury cadmium telluride (HgCdTe), gallium nitride (GaN), aluminum nitride (AlN), gallium arsenide (GaAs), aluminum gallium arsenide (AlGaAs), aluminum nitride (AlN), QXP, lithium niobate (LiNbO3), barium titanate (BaTiO3) (BTO), or their binary, ternary and quaternary compounds and alloys.

8. The external cavity laser according to any one of claims 1 to 7, wherein the ratio of the radii of the first MRR and the second MRR is proportional to the ratio of their group refractive indices.

9. The external cavity laser according to claim 8, wherein, The ratio of the radii of the first MRR and the second MRR roughly corresponds to: ， in, R 1 It is the radius of the first MRR. R 2 It is the radius of the second MRR. n g1 It is the group refractive index of the first MRR material. n g2 It is the group refractive index of the second MRR material, and m It is a tuning enhancement factor.

10. The external cavity laser according to claim 8, wherein, The ratio of the radii of the first MRR and the second MRR roughly corresponds to: ， in, R 1 It is the radius of the first MRR. R 2 It is the radius of the second MRR. n g1 It is the group refractive index of the first MRR material. n g2 It is the group refractive index of the second MRR material. m It is a tuning enhancement factor, and N It is a reduction factor related to the material of the first MRR.

11. The external cavity laser according to any one of claims 1 to 10, wherein, The resonant cavity has a Fabry-Perot configuration and includes a fixed cavity reflector optically coupled to one end of the gain medium, and a first MRR and a second MRR forming a tunable cavity reflector optically coupled to the end of the gain medium opposite to the fixed cavity reflector.

12. The external cavity laser according to claim 11, wherein, The fixed cavity reflector includes a reflective material layer extending along the surface of the gain chip.

13. The external cavity laser according to any one of claims 1 to 10, wherein, The resonant cavity has a ring configuration.

14. The external cavity laser according to any one of claims 1 to 13, wherein, The at least one reflector chip is composed of a monolithic reflector chip integrating a first functional layer and a second functional layer.

15. The external cavity laser of claim 14, further comprising a waveguide structure defining the resonant cavity, the waveguide structure comprising: First MRR and Second MRR; One or more waveguide branches extending in each functional layer; as well as One or more optical vias that allow light to travel between the functional layers of the monolithic reflector chip.

16. The external cavity laser according to any one of claims 1 to 13, wherein, The at least one reflector chip includes a first reflector chip and a second reflector chip in a hybrid configuration, wherein the first reflector chip integrates a first functional layer and the second reflector chip integrates a second functional layer.

17. The external cavity laser of claim 16, further comprising a waveguide structure defining the resonant cavity, the waveguide structure comprising: First MRR and Second MRR; One or more waveguide branches extending in each functional layer; as well as One or more optical edge couplers or grating couplers that allow light to travel between functional layers of a first reflector chip and a second reflector chip.

18. The external cavity laser according to claims 1 to 13, wherein, The first functional layer is located in the gain chip, and the second functional layer is integrated in one of the at least one reflector chips.

19. The external cavity laser of claim 18, further comprising a waveguide structure defining the resonant cavity, the waveguide structure comprising: First MRR and Second MRR; One or more waveguide branches extending in each functional layer; as well as One or more optical vias that allow light to travel between one of the at least one reflector chip and a functional layer of the gain chip.

20. The external cavity laser according to any one of claims 15, 17 and 19, wherein, The waveguide structure includes a coupler that splits light into a clockwise portion and a counterclockwise portion that travel in opposite directions through the waveguide structure.

21. The external cavity laser according to claim 20, wherein, The waveguide structure includes a ring-shaped reflector that closes the feedback circuit.

22. The external cavity laser according to claim 21, wherein, One of the first MRR and the second MRR is located inside the annular reflector.

23. The external cavity laser according to any one of claims 20 to 22, wherein, The first MRR and the second MRR are optically connected through at least two of the waveguide branches.

24. The external cavity laser according to any one of claims 20 to 23, wherein, At least one of the first MRR and the second MRR includes multiple MRR components.

25. The external cavity laser according to any one of claims 20 to 24, wherein, The waveguide structure includes a Mach-Zehnder interferometer integrated with at least one of the first MRR and the second MRR.

26. The external cavity laser according to any one of claims 20 to 25, wherein, At least one of the first MRR and the second MRR has a variable width portion.

27. The external cavity laser according to any one of claims 1 to 26, further comprising a frequency stabilization mechanism, the frequency stabilization mechanism comprising: A monitoring photodetector configured to measure the estimated optical power output of the external cavity laser; A first heater, which is thermally connected to a second MRR, is used to tune the resonance of the second MRR based on the estimated optical power. as well as A second heater, which is thermally connected to the first MRR, is used to tune the resonance of the first MRR based on the estimated optical power.

28. The external cavity laser according to claim 27, wherein, The distance between the first heater and the second MRR is less than the distance between the second heater and the first MRR.

29. The external cavity laser of claim 27 or 28 further includes a 90-degree coupler optically coupled to the first MRR and the second MRR to measure the phase difference between the external laser cavity outputs.

30. The external cavity laser according to any one of claims 27 to 29, wherein, The first heater and the second heater are respectively integrated with the second MRR and the first MRR monolithically.

31. The external cavity laser according to claims 27 to 30, further comprising a plurality of temperature sensors located between the gain chip and the resonant cavity, the plurality of temperature sensors being configured to monitor thermal wave propagation, wherein, The first and second heaters are adjusted based on the monitored heat transfer.

32. The external cavity laser of claim 31 further includes a plurality of lugs, each lug being connected to a corresponding one of the plurality of temperature sensors and extending through the first functional layer and the second functional layer.

33. The external cavity laser according to any one of claims 1 to 32, further comprising an auxiliary temperature sensor configured for measuring external thermal disturbances.

Images