Analysis, recent challenges and capabilities of spin-photon interfaces in Silicon carbide-on-insulator

Abstract

Silicon-carbide (SiC) is a promising platform for long-distance quantum information transmission via single photons, offering long spin coherence qubits, excellent electronic and optical characteristics and CMOS-compatibility. We review key properties of spin-photon interface components for future deployment on the SiC-on-insulator platform with detailed insights provided for available color centers as well as integrated photonic circuits. The associated challenges to achieve high-fidelity multi-qubit control and photon-mediated entanglement on-chip are elaborated, perspectively.

Introduction

Photons can be utilized as flying qubits to permit quantum communication between nodes^1,2 or to operate quantum gates^3,4. The major advantage of flying qubits is the ability to achieve simultaneous super-position of quantum states and show quantum interference, which grants photons the ability to combine multiple quantum states, depending on their relative phase. This provides classical and quantum optical interference⁵ utilized for quantum operations. In general, photons are mostly designated for long distance quantum communication through optical fibers^6,7. Integrated quantum photonic circuits utilize photons in the same manner with increased scalability and miniaturization for their deployment in quantum processors or quantum networks as well as quantum information distribution^8,9. Silicon-carbide (SiC) can provide a platform for on chip integrated photonics which enhances favorable material nonlinear optical properties (second and third order optical nonlinearity) and thus enhance single photon emission of a color center¹⁰. In addition, SiC integrated photonics, especially nanopillars, can also enhance the characteristics of a vast range of radiative paramagnetic color centers (see Fig. 1a), demonstrated with silicon vacancies in a single photon configuration¹¹. Combining this with its wafer-scale nanofabrication capabilities, SiC could evolve into a superior host for quantum phenomena. The composition of these unrivaled advantages could lead to enhanced spin-photon interface (SPI), known as linking a photon to the electron spin associated to a single-photon emitting paramagnetic defect (see Fig. 1b). This provides the opportunity to have different types of qubits within the same system. Recent efforts to enable this ability in other integrated photonic host-platforms have been undertaken^12,13, with first insights given in bulk SiC¹⁴. However, a successful demonstration of this combination in Silicon-carbide on Insulator (SiCOI) is still lacking. Furthermore, not every available color center might be useful for integrated quantum applications due to the occurrence of spectral-broadening effects. The silicon vacancy in SiC does not suffer from excessive spectral broadening^15,16,17.

Fig. 1: Overview of SiCOI-platform illustrating the SiCOI with different implanted color centers combined with integrated photonics.

a Exemplary ensemble PL-measurements of the silicon vacancy, nitrogen vacancy (${{{{\rm{N}}}}}_{{{{\rm{C}}}}}{{{{\rm{V}}}}}_{{{{\rm{Si}}}}}^{-}$), divacancy and vanadium centers and for quantum information distribution in 4H-SiC with data from ref. ¹⁶⁸ (original data from ref. ¹⁶⁹). Silicon vacancy and divacancy are excited at 671 nm, ${{{{\rm{N}}}}}_{{{{\rm{C}}}}}{{{{\rm{V}}}}}_{{{{\rm{Si}}}}}^{-}$ is excited at 940 nm and vanadium at 441 nm, following ref. ¹¹²; silicon vacancy, divacancy and ${{{{\rm{N}}}}}_{{{{\rm{C}}}}}{{{{\rm{V}}}}}_{{{{\rm{Si}}}}}^{-}$ are measured at 12 K, while vanadium PL is at room temperature. A second-order correlation measurement is shown in b, from a visible single photon emitter in 4H-SiC¹⁷⁰; a color center electron spin with adjacent nuclear spins hosted in SiC is shown in c; the energy levels scheme of a silicon vacancy defect hosted in 4H-SiC is illustrated in d. A waveguide transmission is shown in e and a schematic spin-photon interface in f, This last consisting of a 2D-PhC and a chosen color center with SLT: slap layer thickness, λ: light wavelength, BA: base angle of waveguide, H: waveguide heigth, n_4H−SiC: refractive index of 4H-SiC at desired wavelength, W: waveguide width, D: diameter of holes from PhC.

Additionally, there are mainly two types of spins associated with color centers, which are relevant for quantum technologies applications due to the electron shell model of the considered atom or defect. The electrons residing in the color centers and surrounded by nuclei can posses a specific spin, called electron spin as well as the nuclei itself could possess a spin, called nuclear spin (see Fig. 1c). Depending on the considered host-material or atoms inside the lattice, these spins have different intrinsic values. When electron and nuclear spins are polarized or prepared in a given quantum state or energy level (see Fig. 1d), they naturally evolve into a mixed or classical state, identifying the relaxation time or spin-spin relaxation (spin coherence). This is generally much longer for nuclear spins, which can be utilized as quantum memories^18,19. Highly sought after are single-photon emitting (SPE) color centers, which are either fabricated utilizing various ion -implantation or -irradiation techniques. The SPE can be assessed utilizing different experimental setups such as second-order correlation measurements (see Fig. 1b). This particular measurement utilizes a photon arrival coincidence measurement based on either a Mach-Zehnder- or an Hanbury Brown and Twiss-Interferometer, consisting out of a 50:50 beam-splitter and a time-delay circuit for one channel of incoming photons. Time-tagging electronics are recording the arrival times of incoming photons from both channels to construct an histogram.

An integrated SPI in photonic nanostructures could enable the quantum memory capability of nuclear spins coupled with an electron spin^18,19, and operate quantum gates¹⁴, which then enables the transfer of information via entanglement distribution on-chip or free space utilizing photons²⁰.

Furthermore, SPIs can be applied for quantum repeater-based communication²¹ and allow communication between quantum computing units²². SiC is capable of hosting photonic devices such as waveguides (see Fig. 1e), photonic crystal cavities (PhCs) (see Fig. 1f) and on-chip photon detectors²³. A quantum SPI is therefore depending on the material-capabilities that permits to connect solid spin states, that are spatially far apart, by the efficient transfer of quantum information by photons. SPI’s can enable the realization of quantum nodes²⁴ and quantum memories in a quantum network^25,26, based on the electron/nuclear spins to interconnect spins and single photons.

In quantum networks, some qubits interact with each other directly to complete small-scale quantum operations, while other qubits can operate as repeaters to distribute quantum information and entanglement to other nodes utilizing photons. A quantum network is supposed to distribute quantum entanglement over long distance across many quantum memories, such as nuclear spins in atom-like solid-state qubits. Qubit entanglement can be generated by spin-photon entanglement combined with single photon-interference in specific photonic channels.

Among the material platforms that could enable monolithic SPI with potentials of being on-chip and complementary metal-oxide-semiconductor (CMOS)-compatible at large wafer scale with the ability of hosting nanophotonic devices, SiC has emerged as a prominent candidate. This is due to some key characteristics: the long spin coherence of some available color centers electron spin qubits^27,28; the low density of the nuclear spins bath, that deems SiC as the best non-chalcolgenide material platform with the least nuclear spin noise²⁹; the entanglement of electron and nuclear spins with single photon emission (SPE) read-out¹⁹ within the infrared wavelength-region^30,31. In particular, compared to the visible or near infrared spin-photon interfaces found in other material platforms, such as diamond, with long coherence time in the mK regime, SPI’s based on infrared photon emitters in SiC promise to extend ten times the distances obtained in diamond due to 10 times lower losses inside the fiber or within the photonic circuits³².

In this review, we summarize all necessary elements required to establish SPIs based on color centers as spin qubits in SiCOI. So far, only inversion-symmetry defects as well as the silicon vacancy have shown promise to be utilized in these applications. We include their fabrication methods via multiple approaches, such as ions implantation, electrons irradiation and direct laser writing, as discussed in Section “Color centers fabrication methods”. We elaborate the control, manipulation and the optical read-out of single spins, and the verification of indistinguishable single photons associated with spin qubits. Considering the importance of nuclear spins in SPIs, available either in the lattice of the host-platform or within the implanted ions, we outline their role embedded in SPIs. We describe relevant available spins in SiC known from color centers and the material challenges in fabricating them in the SiCOI-platform. Additionally, we point out the most recent material challenges in terms of fabrication and wafer purity of SiCOI. Furthermore, we give an overview of the most recent photonic structures in SiC and associated nonlinear phenomena necessary to control photons and improve the optical read-out of the spin qubits. Finally, we will highlight the benefits of SPI’s in integrated circuits for possible future applications in quantum technologies.

Spin-photon interface requirements

A SPI in solid state is a physical system which hosts an interaction of spin-, optical- and charge-states properties that can be utilized to realize complex experiments and technologies. While several physical systems are currently being explored, implementations based on interfacing photons to individual solid state spins³³ have been particularly explored due to long spin coherence times that can be achieved in host materials such as SiC, with low nuclear spins concentration, which can be further reduced via nuclear spin material purification.

In these systems, the electron spin is primarily initialized and read-out optically via photons, while the electron spin phase accumulation/shift in different nano-environments is utilized for quantum sensing^34,35. The electron spin can be coupled with large clusters of nuclear spin registers for quantum computation³⁶, with the inbuilt quantum gate control and read-out mediated by the SPI. Quantum communication between quantum computing nodes requires spin-dependent optical transitions³⁷ and quantum memories based on spin-capabilties. The key components of SPI are therefore grouped in electron and nuclear spin states associated to the emission of indistinguishable single photons, corresponding to spin-selective transitions (see Fig. 1b), which can be detected with high fidelity. This requires high spin (spin-lattice, spin-spin coherence) and optical coherence properties, that can be controlled utilizing Rabi-³⁸ or Spin Echo-methodologies³⁹, while photonic local density of states control via photonic cavities is another optical approach to control the emission. While these properties of SPI’s can be controlled by the previously mentioned methods, they are also dictated by the material crystallographic structure and purity as well as by the device fabrication necessary to enhance them. For spin read-out purposes, optically detected magnetic resonance (ODMR)-capability of the host-material is highly desirable. The required properties for SPI’s can be summarized:

The electron spin should possess ODMR-capabilities (see Fig. 2a, b), ideally in a single-shot state. Additionally, it should be efficiently initialized and have a long spin lattice relaxation and spin coherence time, to allow quantum operations. ODMR is widely understood as a spectra-measurement technique, where a magnetic field or a microwave/radio-frequency excitation is applied to an optically active transition of a semiconductor material. The external magnetic field can cause Zeeman-splitting⁴⁰, with the microwave/radio-frequency excitation impacting the population states of the ground state spin levels, providing a modulation of the optical transition. Single-shot ODMR read-out is highly desirable as it allows to read-out single quantum states with the highest fidelity, which is crucial for quantum technologies.

Fig. 2: Spin-photon interface key measurements.

Illustration of major SPI requirements like ODMR-measurement data shown in a from ref. ¹⁶⁴ with the subsequent experimental setup in b, Hong-Ou-Mandel experimental setup in c and experimental HOM-demonstration from ref. ¹⁶ in d. Panels a reprinted with permission from ref. ¹⁶⁴ and d from ref. ¹⁶ permitted under the Creative Commons license CC BY 4.0.

The initialization and read-out of the spin should have high fidelity to permit addressing and manipulating additional qubits, such as nearby nuclear spins to enable quantum-error correction and quantum-computational tasks as well as entanglement purification. Electrical spin read-out has been utilized in silicon (Si) due to the availability of integrated single electron transistors on chip⁴¹, but due to single electron transistors thermal limitations, generally optical read-out is preferred as the read-out enhancement can be achieved via photonics and enable long distance communication from quantum nodes. Spin to charge conversion properties of the SPI could be relevant to achieve single-shot read-out. In fact, single shot read-out can be achieved utilizing photonic cavities to control the spontaneous emission rate of the spin selective transitions and limiting spin-flips caused by transitions to metastable states. As reported by Calusine et al.⁴², utilizing photonic cavities increases the ability to control the spontaneous emission rate of the color center and the optical spin read-out, by increasing the photon collection efficiency of the the photoluminescence, thus improving the ODMR signal strength and the ground state spin initialization. As high Purcell enhancement of color centers is generally challenging in hybrid photonic cavities, other methods can be utilized i.e. spin to charge conversion^28,43 or nuclear spin assisted single shot read-out⁴⁴. Furthermore, these alternative methods do not provide sufficient high fidelity due to the low count rate of the photon-emission and require additional nuclear spins in the proximity of color centers.

Single photon emission is essential to determine that only one SPI is addressed and utilized to transfer quantum information. To be able to determine if the photon source is a single photon emitter, a second order correlation-measurement can be performed (Fig. 1b). Equation (1) will be applied onto the examined photon source-data during this measurement⁴⁵:

$${g}^{(2)}(\tau )=\frac{\langle {n}_{1}(t){n}_{2}(t+\tau )\rangle }{\langle {n}_{1}(t)\rangle \langle {n}_{2}(t+\tau )\rangle },$$

(1)

where n_i(t) represents the number of photons detected at timestamp t and n_i(t + τ) at a delay-time τ. We can assume that an implanted color center can be identified as SPE if the condition (2) is satisfied^46,47

$${g}^{(2)}(\tau =0)\;<\;0.5.$$

(2)

Regardless the host platform, highly desirable are indistinguishable photons with high emission rate emitted by a single source to host a spin-photon interface⁴⁸.

Indistinguishable single photons are required for a variety of reasons for complex spin-photon interactions. In general, the indistinguishability is an assumption indicating that every emitted photon from the source is identical and indiscernible from another^16,49,50. To prove the indistinguishability of a certain single photon source, the most reported solution^16,51,52,53 is to conduct a Hong–Ou–Mandel two-photon interference experiment⁵⁴. This experiment is predominantly utilized to test two consecutive incoming photons interfering an unbalanced Mach–Zehnder interferometer and conduct a time-delay measurement of these incoming photons (Fig. 2c). In a time-delay plot, the Hong–Ou–Mandel experiment provides a very characteristic dip at zero time-delay (Fig. 2d), which is a clear indicator if the photon source can be deemed indistinguishable.

Nuclear spins coupling with electron spins via hyperfine interaction is needed. The first technologically useful quantum computation application of nuclear spins was proven in specific doped silicon-substrate ²⁹Si⁵⁵. Due to its unique material properties SiC combines both, ¹³C- and ²⁹Si- nuclear spins with their respective natural abundance of p_C = 1.1% and p_Si = 4.7%^56,57, which are both contributing to an increased decoherence time of specific color centers. SiC has the ability to couple one electronic spin to multiple nuclei, which then generates a quantum register^58,59. Further investigation is required to develop a quantum processor which scales the atomic-scale spin to a semiconductor chip⁵⁸. This field of research has experienced a significant impact from the report of the control of an isolated ²⁹Si-spin with an non-zero nuclear spin, which then gets utilized to entangle an optically excited divacancy-spin with a quantum register, based on the natural 1/2-spin of ²⁹Si¹⁹. However, very recently, it was reported that the utilization of an electronically spin-active silicon vacancy in combination with nuclear spin of either ¹³C and ²⁹Si forms a scalable quantum memory node¹⁸. Before that, a theory was developed where transition metal defects, especially the vanadium vacancy, were investigated in terms of their ability to couple the electron spin of the implanted atom to nuclear spins hosted in either 4H- or 6H-SiC⁶⁰.

Coherent spin-photon interfacing requires narrow spectrally stable, coherent spin-selective optical transitions, as close as possible to the lifetime limit. So far, only the silicon vacancy has shown the required capabilities. Additionally, noise-free quantum frequency conversion should be available to convert SPI in the desired spectral range or compensate for the optical emission frequency differences among different emitters. Frequency conversion can also aid in interfering different color centers by tuning the emission to the desired wavelength.

High-quality nanophotonic platforms are relevant to enhance spin-photon interaction on a single spin-/photon-level as well as to achieve single-shot read-out of the spin. Optical nanophotonic cavities are required to increase photon generation rates and to improve collection efficiency. Scalability considerations requires this platform to embed all standard integrated photonic circuitry, including frequency conversion in addition to the compatibility with industrial nanofabrication standards considering full-size wafer-scale. High cooperativity in light matter interaction could also be utilized to produce quantum optical entanglement states via nonlinear quantum optics or high-order spin-photon entangled states, i.e. Greenberger–Horne–Zeilinger (GHZ) states⁶¹, with specific color centers hosted in SiC being identified and deemed suitable for the generation of these quantum systems⁶².

Various necessary elements for quantum networks implementation such as, electron spin quantum control⁶³, electron-spin mediated nuclear spins control¹⁹, quantum operations¹⁴ as well as indistinguishable single photons¹⁶ have been identified in SiC so far, which could lead to further technological development in integrated quantum circuits as well as spin-photon entanglement¹⁷. Single shot electron spin readout via spin to charge conversion was achieved in SiC with a fidelity over 80%²⁸. SPI’s are not exclusively tied to SiC as host platform³³. However, SiC is well suited to couple the electronic/nuclear spin of color centers with photonic structures⁴⁸, ideal prerequisites have been shown by the silicon vacancy color center.

SiC platform

There are different types of color center available, mostly studied in bulk-SiC as well as SiC-epilayers. These reported findings can be utilized to generate these centers in a superior SiCOI-structure. Their properties depend heavily on the crystallographic sites, local material impurities and the utilized SiC-polytype, mostly recognized are hexagonal-(4H, 6H) and cubic (3C)-polytypes. These centers introduce excellent proven material quantum properties as SPE-sources in SPIs with long spin coherence times⁶⁴. Reported are a total of approximately 20 different color centers⁶⁵. Most studied, however, are two specific color centers known as silicon vacancy ${{{{\rm{V}}}}}_{{{{\rm{Si}}}}}^{-}$⁶⁶ and divacancy ${{{{\rm{V}}}}}_{{{{\rm{Si}}}}}{{{{\rm{V}}}}}_{{{{\rm{C}}}}}^{0}$⁶⁷.

Here, we will narrow down this list to approx. 5–6 specific centers (see Table 1) which have unique properties in terms of spin, charge state, zero phonon line (ZPL), spin coherence as well as spin-photon transitions and point out recent challenges. Interestingly, so far, the ${{{{\rm{V}}}}}_{{{{\rm{Si}}}}}^{-}$ is the only proven emitter which does not suffer from major spectral broadening. Before that, we will elaborate the state-of-the-art fabrication methods of color centers.

Table 1 Comparison of optically addressable color centers in SiC

Color centers fabrication methods

The ability to precisely fabricate the color center at given depths and lateral positions in the material is cruical in order to raise the abilities of this technology and can impact the quantum properties of the color center such as spin coherence time, emission linewidth, photostability, quantum efficiency and spectral diffusion. Electrons-^66,67,68, neutrons-⁴⁷ and protons-^69,70 irradiation as well as focused ions beams^70,71, with subsequent thermal annealing can be utilized to manufacture the desired color center, generated by vacancy creation and diffusion. The range of accessible color center with the irradiation technique is mostly limited to the intrinsic ones, such as the silicon vacancy (${{{{\rm{V}}}}}_{{{{\rm{Si}}}}}^{-}$), divacancy (${{{{\rm{V}}}}}_{{{{\rm{Si}}}}}{{{{\rm{V}}}}}_{{{{\rm{C}}}}}^{0}$), and carbon antisite vacancy pair (CAV). In addition, the temperature stable (TS) center, classified as another intrinsic defect, has been created via proton irradiation⁷² with a recent investigation indicating photo-stability after thermal annealing at temperature above 1200 °C⁷³. The removal-principle of the irradiation-technique remains the same regardless the particular utilized fabrication technique. The incoming particle is aimed to remove either a Si-, C-atom or both out of the SiC lattice and therefore generate a vacancy which is optically addressable after thermal homogeneous annealing across the sample. The electron and neutron irradiation methods are not able to provide the in plane- and in depth-location information due to the nature of the irradiation. The most commonly utilized technique since many years is the ion implantation (see Fig. 3a) to obtain extrinsic color center, providing an approximation of the implanted depth of the color center based on the implantation energy due to the implanted ions stopping range. Subsequently, the implantation dose impacts the classification of the generated center as SPE or ensemble.

Fig. 3: Color centers from fabrication to spin-coherence measurements.

Schematic illustration of color center hosted in SiC with a fabrication, b defect lattice, c illustrating spin read-out microwave/radiofrequency excitation sequences embedded between laser-excitation and read-out measurement with shared x-axis labels from ref. ¹⁷¹. The corresponding spin dephasing/control measurements: continuous ODMR measurements of the divacancy PL 1 reported in d-i from ref. ¹⁷²; a dynamic spin-decoupling measurement utilizing a neutral PL1-divacancy in combination with Carr-Purcell-Meiboom-Gill (CPMG)-N pulses from ref. ¹⁷³ is illustrated in d-ii; a Spin echo spectroscopy of an ${{{{\rm{N}}}}}_{{{{\rm{C}}}}}{{{{\rm{V}}}}}_{{{{\rm{Si}}}}}^{-}$ center is shown in d-iii) from ref. ³⁹ and Ramsey-excitation was utilized to manipulate an ${{{{\rm{N}}}}}_{{{{\rm{C}}}}}{{{{\rm{V}}}}}_{{{{\rm{Si}}}}}^{-}$ center (kk) in d-iv from ref. ¹⁰⁷. Panel d-i reprinted with permission from ref. ¹⁷² ⓒPhotonics Research with d-ii reprinted with permission¹⁷³. d-iii distributed from ref. ³⁹ under the CC BY-NC 4.0 ⓒThe Authors, some rights reserved, exclusive licensee AAAS and d-iv reprinted with permission¹⁰⁷. Copyright 2024 American Chemical Society.

In order to obtain intrinsic defects, the electron irradiation technique is most commonly utilized as less damaging to the material, which is also able to provide an approximate depth-information. Additionally, the laser-writing technique is also arising for various depth implantation and potentially less material damage.

Various ion beams have been utilized for implantation such as He⁺, ${{{{\rm{H}}}}}_{2}^{-}$, Si²⁺, C⁻, Er³⁺,N⁻ into the SiC lattice (see Fig. 3b), responsible for vacancy creation and replacement of either a Si- or C-Atom. Depending on the desired color center, the corresponding, to be implanted ions are determined together with the implantation energy and dose to minimize material damage.

In general, the ion-irradiation technique can have a severe impact on spin coherence times^74,75 compared to grown emitters or electron-irradiated color center, thus increasing the demand of less material-damaging fabrication methods. The spin coherence is also affected by the material surface species therefore thin films with larger surface area and color centers at nm-distance from the surface have unfavorable optical and spin quantum properties, including reduced photostability and intermittence in some cases^32,42. Especially, spectrally stable emitter are highly favorable for future integrated quantum photonics applications. So far, the silicon vacancy possesses the most favorable characteristics for these applications.

Silicon vacancy

One of the most studied centers is the negatively charged silicon vacancy ${{{{\rm{V}}}}}_{{{{\rm{Si}}}}}^{-}$. This intrinsic defect can be generated by removing a silicon-atom out of the SiC-lattice⁷⁶, which can be achieved utilizing various irradiation techniques⁷⁷, such as initially electron irradiation⁶⁶, later ion implantation^78,79 or most recently reported focused He⁺-Ion beam (FIB) implantation⁸⁰.

For fabrication purposes, many studies reported that the utilization of Helium ions (He⁺-Ions) in combination with a PMMA-mask for precise positioning, can be advantageous to preserve the outstanding spin and optical properties of ${{{{\rm{V}}}}}_{{{{\rm{Si}}}}}^{-}$¹⁴. Maskless methods are sought after as they can be even more precise and directly applied to photonic devices.

Focused ion beams (FIB) are under investigation to create color centers with a He⁺-FIB already shown the ability to provide ${{{{\rm{V}}}}}_{{{{\rm{Si}}}}}^{-}$ with overall high yield and simple localization⁸⁰. This technology can only be utilized for near-surface (20–100 nm depths) emitters due to the low energy of the FIB-beam. It is not always applicable for in depth fabrications of color centers. Laser writing based on femtosecond lasers was also demonstrated successfully to generate ${{{{\rm{V}}}}}_{{{{\rm{Si}}}}}^{-}$^81,82 at depths from 5 to 40 μm and most recently, a nanosecond laser was utilized to generate ${{{{\rm{V}}}}}_{{{{\rm{Si}}}}}^{-}$ directly into nanophotonic crystal cavities⁸³. Laser writing could provide a very beneficial impact on the spin coherence, however experimental evidence is lacking.

The unique optical-properties are based on the electronic arrangement of the ${{{{\rm{V}}}}}_{{{{\rm{Si}}}}}^{-}$, which consists out of five active and three vacant electrons. That points directly to a spin of 3/2^17,76, where the sub-states m_s = ± 1/2 and m_s = ± 3/2 have been identified as Kramer’s doublets⁸⁴. When occurring in 4H-SiC, ${{{{\rm{V}}}}}_{{{{\rm{Si}}}}}^{-}$ can be distinguished into two different crystallographic sites which are called hexagonal (h) and quasi-cubic (k), with each of them providing specific characteristics⁸⁵. For both crystallographic sites, additional main no-phonon photoluminescence (PL) lines V1 and V2 can be observed⁸⁵ which are related to the placement of the defect in the lattice^86,87. The occurrence of V1 and V2 ZPL-lines in 4H-SiC is caused by the existence of two excited states called ⁴E and ⁴A₂^17,88 with slightly different excitation energies. Depending on the selected excitation from the ground state, either a V1-ZPL at 861 nm¹⁷, which corresponds to a ⁴A₂ to ⁴A₂ ZPL-transition or a V1’-ZPL can be generated based on a ⁴E to ⁴A₂ transition⁸⁸. Both of them can be considered spin conserving⁸⁹. However, these excited states can also result from a mixture of both transitions which can occur due to polaronic mixing of the existing electronic states⁸⁷. Both of these states are associated with a h-defect configuration in the SiC lattice. Conversely, attributed with a k-defect configuration is a V2-ZPL at 916 nm⁹⁰.

Additionally, ${{{{\rm{V}}}}}_{{{{\rm{Si}}}}}^{-}$ exhibits a Debye-Waller-Factor (DWF) of 6% theoretically and experimentally determined to be 8–9%⁸⁷ with an optical lifetime of 5.03 ns and 6.26 ns considering specific spin-states m_s = ± 1/2 and m_s = ± 3/2, respectively for V1-configurations^21,76,91. For V2-configurations, a theoretical-DWF of approx. 6.2% and experimental-DWF of 6-9% has been reported⁹² as well as optical lifetimes of 6.1 ns (m_s = ± 1/2) and 11.3 ns (m_s = ± 3/2)⁹³. The spin coherence time T₂ is reported at 20 ms^17,94 measured by utilizing ODMR-setups in combination with echo-radio frequency excitation sequences, as illustrated in Fig. 3c), which are applied between the laser-excitation functioning as spin polarization and subsequent optical read-out measurement.

Major advantages of ${{{{\rm{V}}}}}_{{{{\rm{Si}}}}}^{-}$ are the near-infrared photon emission and the already demonstrated ability of its implementation within existing opto-electronic devices operating at room temperature. A bright and highly photo-stable photon emission even at room temperature as well as a fairly long spin coherence time T₂ is also provided by ${{{{\rm{V}}}}}_{{{{\rm{Si}}}}}^{-}$ in combination with low Rayleigh scattering losses which are related to the emission wavelength⁸⁵. Additionally, the narrow spectral-linewidths can be preserved even when coupled to a photonic device, as reported by¹⁰. Linewidths of 57 ± 6 MHz and 48 ± 6 MHz were measured for the V1-${{{{\rm{V}}}}}_{{{{\rm{Si}}}}}^{-}$, which permitted to verify single photon indistinguishability¹⁶. For the V2-${{{{\rm{V}}}}}_{{{{\rm{Si}}}}}^{-}$, linewidths of less than 200 MHz (116–187 MHz) have been observed in SiC membranes less than 250-nm thick, while in the thicker membranes, the linewidths were 26 MHz and 14 MHz for the two spin-selective transitions¹⁵. This indicates that sub-micron membranes containing the ${{{{\rm{V}}}}}_{{{{\rm{Si}}}}}^{-}$ can be utilized for SPI protocols.

To the best of our knowledge, the ${{{{\rm{V}}}}}_{{{{\rm{Si}}}}}^{-}$ is currently the only color center successfully integrated in a SiCOI-chip^76,95. A first insight of combining a light confinement structure with a ${{{{\rm{V}}}}}_{{{{\rm{Si}}}}}^{-}$ hosted in bulk SiC was given by Babin et al.¹⁴. The utilization of the long spin coherence time of ${{{{\rm{V}}}}}_{{{{\rm{Si}}}}}^{-}$ in combination with the characteristics of bulk SiC could enable spin-photon manipulation which could lead to a remote generation of quantum spin-photon entanglement. Local spin-photon entanglement has been demonstrated by Fang et al.⁹⁶ utilizing 4H-SiC epilayers with a V1-${{{{\rm{V}}}}}_{{{{\rm{Si}}}}}^{-}$ within a solid immersion lens, suggesting that utilizing an integrated photonic cavity higher visibility and scalability could be achieved for entangling distant qubits.

Interestingly, a quantum efficiency of 30%⁹¹ has been reported when ${{{{\rm{V}}}}}_{{{{\rm{Si}}}}}^{-}$ are tuned into a SPE at room temperature with a lower limit estimated at ~23% by Morioka et al.⁹¹. This can be related to the dipole polarization along the vertical axis and the complex transition to the inter-system level-crossing⁸⁶. The reported properties of the implanted ${{{{\rm{V}}}}}_{{{{\rm{Si}}}}}^{-}$ differ slightly from the bulk demonstration of ${{{{\rm{V}}}}}_{{{{\rm{Si}}}}}^{-}$. The V2-ZPL in 4H-SiCOI is reported at 916.5 nm at a temperature of 4.3 K⁹⁵.

Divacancy

${{{{\rm{V}}}}}_{{{{\rm{Si}}}}}{{{{\rm{V}}}}}_{{{{\rm{C}}}}}^{0}$ is another example for an well studied photon emitter in SiC. The neutrally charged ${{{{\rm{V}}}}}_{{{{\rm{Si}}}}}{{{{\rm{V}}}}}_{{{{\rm{C}}}}}^{0}$ is one of the more common defects occurring in semiconductors like SiC⁹⁷. This specific color center consists out of one missing Si-atom with one missing C-atom adjacent inside the lattice. The most recognized option to fabricate ${{{{\rm{V}}}}}_{{{{\rm{Si}}}}}{{{{\rm{V}}}}}_{{{{\rm{C}}}}}^{0}$ is the application of thermal annealing onto previously created ${{{{\rm{V}}}}}_{{{{\rm{Si}}}}}^{-}$-defects, as demonstrated in^98,99. However, for optimal divacancy creation, electrons- or ion-beam fluences should also be optimized. Most recently, ${{{{\rm{V}}}}}_{{{{\rm{Si}}}}}{{{{\rm{V}}}}}_{{{{\rm{C}}}}}^{0}$ were generated close to the surface at 200-nm depth in bulk SiC, utilizing the laser writing-technique with femtosecond pulses and subsequent thermal annealing⁹⁸. Its unique orientation in the lattice results in four reported configurations of ${{{{\rm{V}}}}}_{{{{\rm{Si}}}}}{{{{\rm{V}}}}}_{{{{\rm{C}}}}}^{0}$, which are either orientated along the c-axis (called hh and kk) or along the basal bond direction (called hk and kh)¹⁰⁰, causing four different ZPLs reported between 1078 nm (kh), 1108 nm (hk), 1131 nm (kk) and 1132 nm (hh)^65,78,101. An extension to these reported ZPL’s was introduced recently with three additional configurations inside stacking faults of the lattice, which can act as quantum wells¹⁰². Most recognizable from that extension is the PL6-subtype, which can be observed at room temperature with an ZPL reported at 1038 nm¹⁰³. This configuration is particularly known for its very high single-photon emission count-rate at approximately 150 kcts/s and ODMR contrast above 25%¹⁰³. Conversely, the PL1-configuration (kh) has a low ODMR-contrast even based on dynamic spin-decoupling measurement with Carr-Purcell-Meiboom-Gill (CPMG)-N pulses, as shown in Fig. 3d-i and d-ii, respectively.

Although the emissions of various subtypes of ${{{{\rm{V}}}}}_{{{{\rm{Si}}}}}{{{{\rm{V}}}}}_{{{{\rm{C}}}}}^{0}$ aren’t relying directly in the telecommunication range, it is still considered as advantage in terms of fiber transmission⁶⁵. Additionally, ${{{{\rm{V}}}}}_{{{{\rm{Si}}}}}{{{{\rm{V}}}}}_{{{{\rm{C}}}}}^{0}$ provides good control over the nuclear spin¹⁴. ${{{{\rm{V}}}}}_{{{{\rm{Si}}}}}{{{{\rm{V}}}}}_{{{{\rm{C}}}}}^{0}$ can be applied in solid-state quantum sensing, communication and information processing¹⁰ and can be utilized for biomedical imaging in deep tissue applications¹⁰³.

Nitrogen vacancy

The negatively charged nitrogen vacancy center ${{{{\rm{N}}}}}_{{{{\rm{C}}}}}{{{{\rm{V}}}}}_{{{{\rm{Si}}}}}^{-}$ in SiC emerged recently and consists out of an nitrogen-atom, replacing a carbon-atom, which is implanted adjacent to a ${{{{\rm{V}}}}}_{{{{\rm{Si}}}}}^{-}$^30,103,104. The implantation can be achieved by either electron irradiation, ions bombardment¹⁰⁵ or precise utilization of a PMMA-mask with subsequent ions-flooding, which has to be followed by thermal annealing of the sample at approximately 1050 °C (1h)¹⁰⁴. However, depending on the concentration of N-dopant during growth in the SiC host material, the ${{{{\rm{N}}}}}_{{{{\rm{C}}}}}{{{{\rm{V}}}}}_{{{{\rm{Si}}}}}^{-}$ centers can be formed from the contained N rather than the implanted N¹⁰⁶. Similar to ${{{{\rm{V}}}}}_{{{{\rm{Si}}}}}{{{{\rm{V}}}}}_{{{{\rm{C}}}}}^{0}$, the ${{{{\rm{N}}}}}_{{{{\rm{C}}}}}{{{{\rm{V}}}}}_{{{{\rm{Si}}}}}^{-}$ center has four available configurations, which can be detected by ODMR spectroscopy and also be distinguished by its orientation in c-axis (kk and hh) or basal bond direction (kh and hk)¹⁰⁷. This leads to related but different optical and magnetic characteristics for each configuration. Reported photon-emission ranges from 1180 nm until 1240 nm^21,108 with kh emitting at 1176 nm, hh at 1223 nm, kk at 1241 nm and hk at 1242 nm³⁰. The lifetime of an ${{{{\rm{N}}}}}_{{{{\rm{C}}}}}{{{{\rm{V}}}}}_{{{{\rm{Si}}}}}^{-}$ center is reported at 2.7 ns^30,107 with an intrinsic spin equal to 1^108,109. However, hosted in 4H-SiC epilayers, the reported lifetime is stated at (2.2 ± 0.4) ns at cryogenic temperatures¹¹⁰, observed with an excitation-wavelength of 785 nm. The zero-field splitting (ZFS) is determined at 1.27–1.313 GHz^108,109.

Especially, the emission near the telecom-band as well as the proven occurrence as single photon emitter can be identified as advantageous. However, the reported spin coherence time T₂ of 17.2 μs (for hh configuration)⁶⁴ and low ODMR contrast¹⁰⁴ as well as spectral diffusion are not optimal (i.e. Fig. 3d-iii, iv). By applying spin-echo and ramsey sequences, environmental factors as well as inhomogenous magnetic-fields will be minimized during the ODMR-measurements, due to spin-refocusing owing to pulse-application, as illustrated in Fig. 3d-iii, iv.

Most recently, the extension of the spin coherence time T₂ of an ${{{{\rm{N}}}}}_{{{{\rm{C}}}}}{{{{\rm{V}}}}}_{{{{\rm{Si}}}}}^{-}$ center in bulk 4H-SiC utilizing a spin echo with dynamic decoupling sequence has been demonstrated. This led to an measured increase of T₂ by up to 10 times³⁹. Additionally, Rabi oscillations has been achieved with a corresponding Rabi frequency of 3.73 Mhz^39,107.

Vanadium and erbium defect

There are many more color centers available in SiC such as Mo⁵⁺, V⁴⁺, Er³⁺ or the carbon antisite-vacancy pair CAV, which have been less studied.

The vanadium defect V⁴⁺ was one of the first emitters discovered in SiC¹¹¹. It is known for an emission in the telecom range³¹ combined with amphoteric characteristics. Reported is that the neutrally charged V⁴⁺ impurity has a ZPL in the telecom range at around 1300 nm with an excited optical lifetime of less than 50 ns¹¹². The spin relaxation time has been reported in μs with no specific numerical value at cryogenic temperatures³¹. Although the electronic spin is recorded as 1/2^31,113, vanadium impurities are still considered prospects for electrical excitation¹¹² with a nuclear spin at 7/2. A very recent demonstration¹¹⁴ of the vanadium color center in bulk 4H-SiC points out a charge state lifetime of 135 ms of the neutral V⁴⁺ site. Additionally, a spectral distribution study of the PL emission range, indicated a small emission-band distribution of solid-state single-photon emitter^113,114, with values as low as 100 MHz. Further spin selective optical transitions have been identified. That can be highly beneficial for applications in quantum networks as well as communication.

Another color center which is attracting lots of interest, at least in other materials, is the ion implanted Er³⁺-defect¹¹⁵. It is known for its photon emission at 1540 nm, which is extremely robust due to inner-shell electron screening of electrons from the crystal field^116,117. The ZPL is reported to be temperature-independent, however other works have indicated ZPL’s between 1490 nm and 1640 nm¹¹⁵. The optical lifetime for Er³⁺ has been reported for patterned at T₂ = 1.16 ± 0.04 ms as well as unpatterned at T₂ = 1.56 ± 0.005 ms by Parker et al.¹¹⁶. In general, the Er³⁺ is the least characterized defect in SiC with most recent demonstrations reported in other materials^118,119. Its integration in photonic cavities is essential to increase the photon emission rate which is limited by the long optical lifetime in bulk material. However, Er³⁺-center hosted in CaWO₄ has shown indistinguishable single photon emission at telecommunication wavelengths utilizing a Si nanophotonic beam with optical linewidths of 150 kHz, spectral diffusion of 63 kHz¹²⁰ and a very long spin coherence time (23 ms)¹²¹. Er³⁺-ions implanted in Si have a reported spin coherence time of 0.8–1.2 ms¹²² with highly inhomogeneous optical linewidths (100 MHz) and homogeneous linewidths below 70 kHz at ≈ 11 mT and 20 mK. It is expected that in SiC, the spin coherence time T₂ could reach values in the 1.1 ms as predicted in ref. ²⁹, slightly higher than in diamond and silicon, but at higher temperature than in Si, due to higher Debye temperature of SiC compared to Si.

Integrated silicon carbide photonics

With a wide bandgap, high optical nonlinearities, high refractive index, controllable artificial spin defects, and a fabrication process compatible with CMOS technology, SiC presents a substantial prospect for the progressive evolution of photonic integrated circuits (PICs). These exceptional properties of SiC enable the realization of advanced SiC PICs for both classical and quantum applications^64,65,76. Planar optical waveguides play a crucial role in PICs by providing a means to guide and manipulate light within a planar structure. These waveguides are essential components that enable the integration of various optical functions on a single chip. Enabling the fabrication of an entire optical system on a chip provides unrivaled scalability, weight reduction, cost efficiency, and power efficiency. In the initial stages of investigating SiC waveguides, suspended SiC in the air was utilized to create photonic structures on a bulk SiC substrate, taking advantage of the availability of commercial SiC epilayers on silicon substrates. Given that Si has a higher refractive index than SiC, it necessitates removal in the proximity of SiC waveguides^14,123. However, waveguides fabricated on this platform typically exhibit heightened intrinsic loss, primarily due to the substantial lattice mismatch between SiC and Si. Moreover, any photonic components fabricated on this platform require suspension in air because of the higher refractive index of the Si layer. This necessity results in complex fabrication procedures, including patterning and top-cladding deposition, leading to reduced reliability^124,125. As an alternative, the current literature predominantly features the utilization of SiCOI-platform, a development that addresses the challenges associated with the earlier structure^{126,127,128,129}. Three main approaches to SiCOI formation, outlined in Fig. 4a, include the following: Approach in Fig. 4a-i adopts the smart/ion cut technique from silicon-on-insulator (SOI) stack formation, substituting the Si wafer with a SiC wafer, mainly yielding 4H-SiCOI stacks^130,131. However, due to distinct physical properties, SiCOI experiences considerable material loss compared to SOI after ion implantation. Approach in Fig. 4a-ii involves directly thinning bonded SiC using grinding and subsequent chemical mechanical polishing (CMP) to achieve a smooth surface, thereby bypassing the ion implantation step in previous method and achieving minimal loss in SiCOI stacks¹³². This approach yields stacks of SiCOI in both 4H and 3C configurations. In approach in Fig. 4a-iii, amorphous SiC (a-SiC) is deposited directly onto the Si substrate with an SiO2 layer in between, achievable through plasma-enhanced chemical vapor deposition (PECVD) and sputtering¹²⁹.

Fig. 4: Silicon carbide on insulator fabrication with derived integrated photonics.

a Three approaches for creating SiCOI stacks. DL: damaged layer, induced by H⁺ implantation. a-SiC: amorphous SiC. b Designing directional couplers using two parallel 4H-SiC rib/ridge waveguides. (i) and (ii) illustrate simulated mode profiles of the even and odd modes of the directional coupler, respectively, while (iii) shows the simulated optical field distribution when the directional coupler operates as a 3-dB coupler. n_eff,even: effective index of even mode. n_eff,odd: effective index of odd mode. λ: light wavelength. κ: field coupling coefficient. G: gap width between two waveguides. L_c: straight coupling length. H: height of waveguides. BW: base width of waveguides. SLT: slap layer thickness. BA: base angle. c Micrograph of high-confinement SiC waveguide fabricated on 4H-SiCOI¹²⁶. d SEM image of RRs fabricated on 3C-SiCOI platform¹²⁷. e SEM image of a RR fabricated on a-SiCOI platform¹²⁹. f Ultrahigh-Q PhC nanocavity structure fabricated on 4H-SiCOI¹⁷⁴. g SEM and optical images of a fabricated RR on 4H-SiCOI platform designed for second-harmonic generation¹⁰. h SEM images of a fabricated RR on 3C-SiCOI platform characterized for SHG, difference-frequency generation, and spontaneous parametric down conversion processes¹⁵⁸. i SEM image of fabricated RRs (false-colored) on 4H-SiCOI platform designed for optical parametric oscillation and optical microcombs¹³⁵. j SEM image of a fabricated RR on 4H-SiCOI platform designed for for broadband comb generations¹⁷⁵. k Overview of an integrated Pockels modulator fabricated on 3C-SiCOI platform¹⁶¹. l SEM image of a RR on 4H-SiCOI platform employed for photon pair generation¹⁵⁹. Panel c reprinted with permission from ref. ¹²⁶ ⓒOptical Society of America with d, f, i and j reutilized under the Creative Commons license CC BY 4.0. Panel e reprinted with permission from ref. ¹²⁹ Copyright 2024 American Chemical Society. g Reprinted with permission from ref. ¹⁰ with h, k and l reprinted from refs. ^158,161 as well as ref. ¹⁵⁹ permitted under the Creative Commons license CC BY 4.0.

Directional couplers serve as fundamental components integral to Mach-Zehnder interferometers (MZIs), ring resonators (RRs), and Sagnac interferometers (SIs), all of which constitute essential building blocks in PICs. These couplers are created by closely positioning two waveguides with mutual energy coupling, allowing them to divide a guided optical wave into two physically separated coherent components and vice versa. Precise control over the coupling strength between optical waveguides is essential for the design and implementation of the MZIs, RRs, and SI-based devices. In a directional coupler, the coupling strength can be adjusted by varying either the interaction length or the separation gap between them. According to the coupled mode theory^133,134, the operational principle of a directional coupler can be simplified and elucidated based on the phase matching condition between the two fundamental eigenmodes of the coupled waveguides, commonly referred to as even and odd modes, or symmetric and anti-symmetric modes. Figure 4b-i, ii illustrates the mode profiles of the even and odd modes of a directional coupler created by two parallel 4H-SiC rib/ridge waveguides, respectively. Figure 4b-iii shows the simulated optical field distribution when the directional coupler functions as a 3-dB coupler on the 4H-SiCOI platform.

As micro/nano fabrication techniques progress, a variety of applications for SiC photonic integrated devices has been consistently demonstrated. Investigating the utilization of SiC PICs necessitates the creation of reliable photonic elements, including low-loss waveguides for efficient optical signal transmission and high quality-factor (Q) integrated optical cavities to amplify optical field strength^126,127,128. To date, various chip-scale SiCOI schemes with diverse applications have been proposed, including high-confinement waveguides for self-phase modulation (Fig. 4c), high-Q and high-confinement RRs (Fig. 4d, e), ultrahigh-Q PhC nanocavities (Fig. 4f), RRs for second-harmonic generation (Fig. 4g, h), RR for difference-frequency generation, and spontaneous parametric down conversion processes (Fig. 4h, high-Q RRs Kerr frequency comb generation (Fig. 4i, j), a Pockels modulator formed by a RR and microwave stripline electrodes (Fig. 4k), and a RR employed for photon pair generation (Fig. 4l). In contrast to discrete off-chip components encountering limitations in system complexity and production scale, the implementation of SiC components in integrated form, fabricated through well-established CMOS technologies, offers significant advantages, including a substantially reduced footprint and enhanced scalability.

Conclusions and outlook on SPI in SiCOI

The ‘on-Insulator’-structure implies highly favorable capabilities to host PICs in SiC and has been object of many recent efforts. It is the most scalable platform, the best in terms of low-loss/noise provided by the host and promises an exceptional Q-factor reported with 9.7 × 10⁵ in a 4H-SiCOI fabricated with wafer-bonding technique^135,136. A new fabrication-process for 3C-SiCOI chips is proposed by Li et al.¹³⁶, who utilizes an anodic bonding process in combination with borosilicate glass which provides an Q-factor of 1.4 × 10⁵. However, the nonlinear frequency conversion efficiency still possesses room for improvements compared to other competitive platforms, which indicates necessary further developments of SiCOI.

Quantum photonics relies up-to-date quite heavily on fiber-coupling of light in and out of the chip, providing remarkable losses into the system which can lead to a complete loss of photons. For instance, a more integrated two-qubit network node with silicon vacancy-center and proximal silicon-29 nuclear spin in diamond nanophotonics has been demonstrated¹³⁷, and similarly, a nanophotonic diamond waveguide hosting a tin-vacancy center coupled to a spin-1/2 Sn nucleus was achieved¹³⁸. In both cases, the coupling of the emitted photons into optical fibers (regardless the high extraction efficiency from the photonic waveguide to the fiber), resulted in considerable photon loss, degrading system robustness and performance, thus limiting nuclear spins initialization, readout, control and remote entanglement of electron-nuclear spins. Therefore, the utilization of waveguides in SiCOI to route photons to on-chip interferometer is a major advantage in integrated SiC quantum photonics with increased modal coupling and collection efficiency of photons into the waveguide, absent in diamond. This could improve a possible two-qubit network in SiCOI based on ${{{{\rm{V}}}}}_{{{{\rm{Si}}}}}^{-}$ with small spectral diffusion. A quantum photonic link, a key element for quantum networks, can be demonstrated, aiming to couple two spin qubits in separated waveguides on a SiCOI-chip via interference into a on-chip beam-splitter. This can also allow to achieve more than two-qubit networks and photonic multidimensional clusters state based on a single SPI’s coupled strongly to a nuclear register¹³⁹. In fact, the rate R at which a photonic cluster states containing M × N photons can be produced and utilized is stated in Eq. (3):

$$R\;\approx\; {(qe\times DWF\times ce\times de)}^{M\times N}$$

(3)

exponentially dependent on the generation efficiency (with quantum efficiency (qe), collection efficiency (ce) and detection efficiency (de)). On-chip SPI integrated circuits can remarkably improve the ce- and de-values, and thus the overall efficiency which is currently limited by the waveguide to fiber coupling of 57(6)% in diamond nanocavities¹³⁸.

P-i-n junctions are designed to control the charge state or tune the PL emission of an designated color center based on the Stark shift effect, which has been demonstrated in SiC^140,141,142. In combination with on-chip single photon detectors, that could enable a fully integrated on-chip SPI with the elimination of fiber coupling, providing a much higher overall efficiency of the quantum system (see Fig. 5).

Fig. 5

Schematic illustration of an on-chip vision of future applications of integrated quantum photonics hosted in SiCOI.

On chip single photon detectors (see Fig. 5) as reported in^{23,143,144,145} and recently on SiCOI¹⁴⁶ based on superconducting nanowire single-photon detectors (SNSPDs) could be a major step towards a high system-efficiency of a PIC in combination with small spectral diffusion ZPL-lines from color center’s. In particular, future utilization of high-Tc superconducting nanowire detectors that can detect single photons operating at 25 K would be ideal for SiCOI with applications in quantum sensing and quantum information processing¹⁴⁷.

The most advanced emitters reported in SiC are discussed previously. However in recent years, a couple of novel emitters emerged due to efforts in adjusting the implantation process. Interesting characteristics have been shown by chlorine center’s in 4H-SiC reported by Bulancea-Lindvall et al.¹⁴⁸, the carbon vacancy hosted in SiC¹⁴⁹ as well as the G-center in Si^150,151. However, from an integrated quantum photonics point of view the only promising emitters, so far, are ${{{{\rm{V}}}}}_{{{{\rm{Si}}}}}^{-}$’s as well as the inversion-symmetry silicon vacancy-center in diamond with a reported spectral linewidth of 94 Mhz¹⁵². Especially, inversion-symmetry defects have drawn some attention recently due to the fact that they are unable to host a permanent electric dipole, which deems them unresponsive to external electrical fields, minimizing spectral diffusion of specific transitions¹⁵³.

In a long-term perspective, color centers based on a vacancy (either C- or Si-vacancy) could be generated by the focused electron beam technique¹⁵⁴, which can be mask-less. For that, an irradiation energy of at least 200 keV needs to be provided to the material for Si-displacement¹⁵⁵. The major benefit is the highly precise irradiation via electron beam with a resolution in the nm-range. Ion-irradiation typically implements damage to the chip-surface, a not precise color center lateral placement and mask requirement for photonics component alignment. This is particularly of interest if a waveguide is fabricated first on top of an SiC-chip with subsequent color center fabrication.

Babin et al. demonstrated the ability to fabricate triangular-shaped photonic devices embedding an implanted ${{{{\rm{V}}}}}_{{{{\rm{Si}}}}}^{-}$-center¹⁴ obtained with angle etching of bulk SiC (see Fig. 5). That enabled nuclear spin control by the ${{{{\rm{V}}}}}_{{{{\rm{Si}}}}}^{-}$-center over adjacent spin qubits which are progressing in the photonic structure.

Photonic inverse design is an emerging photonic-device development tool which utilizes a unique approach to provide focused optimization of the photonic device utilizing multiple unknown structure-parameter. This technique has been proposed in combination with intrinsic color centers to establish a solid-state quantum simulator for implanted atoms¹⁵⁶.

Phase matching is the most important factor to achieve efficient nonlinear frequency conversion utilizing the nonlinear abilities of the chosen platform. It is therefore crucial to achieve it in SiCOI. So far, latest efforts were reported by Yang et al.¹⁵⁷, who introduced an inverse design process which enables researchers to controlling dispersion. This leads to a 2nd-order phase matching condition which enabled second- and third-order nonlinear light generation¹⁵⁷. Using type-I (signal and idler are linearly polarized orthogonally to the pump) modal phase-matching spontaneous parametric down-conversion was shown in 3C-SiCOI RRs¹⁵⁸.

In conclusion, we can envision in Fig. 5 that future SiCOI integrated quantum photonics will incorporate single SPIs containing electron- and nuclear-spins as well as coupled and entangled photons. SPI’s are connected by photons propagating in waveguides, collectively utilized to form photonic multi-dimensional clusters states. Spin-entangled photons from different SPI’s interfere on chip via 50:50 beam splitter and are then detected utilizing waveguide-integrated SNSPD’s. Each SPI will contain photonic cavities to enhance the spin-photon coupling. The microwave antenna and electrodes could be fabricated on chip for spin- or charge-control as well as Stark-shift tuning of the SPI’s. Nuclear spins can be utilized as memories for quantum information storage and processing. In addition, quantum memories based on resonators can be realized (see Fig. 5). Finally, parametric nonlinear, classical as well as quantum, photon sources^{157,158,159,160} and electro-optic modulators¹⁶¹ can be also implemented in SiCOI chips.

Compared to other platforms, integrated quantum photonic links in SiC stand out from a sustainability point of view due to the fact that ${{{{\rm{V}}}}}_{{{{\rm{Si}}}}}^{-}$ center’s have the potential to be utilized at liquid hydrogen temperature T = 20 K⁸⁷.