In the search for a system which can provide technologically implementable, cost-effective and easy-to-deploy platform for various quantum technology applications, some paramagnetic point defects (Weber et al., 2010) in semiconductor materials seem to stand heads and shoulders above other candidates such as cold atoms (Beterov, 2020), trapped ions (Bruzewicz et al., 2019), and even technologically implemented superconductor-based systems (Arute et al., 2019); it seems that they yet have not been given the attention they deserve in the field of information technology.

These solid-state qubits, based on point defects with levels in the band gap of their host, are characterized by degenerate electronic ground states, the degeneracy of which can be removed either by the application of a magnetic field or via electric fields and spin-orbit interaction giving rise to the zero-field splitting (ZFS) (Lommer et al., 1988). These defects have to be in a specific charge state obtained by Fermi level engineering in order to have the required spin state S<=1; then the manipulation of their spin states via the application of appropriate microwave and optical pulses is considered as a practical route for information registration and processing. A key property is the presence of excited states within the band gap of the material, which allow the optical manipulation of the spin states and an optical read-out via the photoluminescence intensity (Weber et al., 2010). The spin-selective recombination from the excited state to the ground state spin state is a mean to readout the registered information. Combined with the possibility of optical re-initialization of the qubit after each readout through recombination through intermediate singlet states (Chakraborty et al., 2017), such solid-state qubits satisfy the well-known criteria proposed by Divincenzo for the realization of a quantum information system (DiVincenzo, 2000). More interestingly, all these processes, in the case of defect-based spin-qubits, are achievable through relatively simple experimental setups such as optically detected magnetic resonance and as shown recently also by photoelectric magnetic resonance (ODMR and PDMR respectively) measurements (Falk et al., 2013; Bourgeois et al., 2020).

Solid-state qubits in semiconductor materials are by far more advantageous over other solid-state qubits, due to their compatibility with the present microelectronic technology (de Leon et al., 2021). Large band gap semiconductors such as SiC and diamond are known as optimal matrices to host qubits in an electronically isolated environments (Balasubramanian et al., 2009), compared to vacuum chambers required to for implementation of trapped ions (Kaufmann et al., 2017). Also, compared to quantum dots where the physical properties considerably vary from one center to the other, isolated point defects in a specific host have close to identical characteristics (Kloeffel and Loss, 2013). This makes them reliable candidates for qubit applications which require defects with identical properties when addressing them optically or with microwaves.

While several color centers in both diamond and SiC have been identified as suitable, the focus of the research, in the past few years has mainly been upon the negatively charged NV center in diamond (Pezzagna and Meijer, 2021) and the neutral di-vacancy VV in SiC (Anderson et al., 2022). These defects have intrinsic properties based on their electronic structure, groundstate/first excited state, and interesting spin coherence times, Debye factors and second order correlation factors. Meanwhile the analogue of the NV center in diamond, the NV center in SiC, have been discovered and investigated in detail in different SiC polytypes (3C, 4H, 6H) (von Bardeleben et al., 2016). We will show that for practical applications, apparently NV centers in SiC are more advantageous (Figure 1) over the two former centers. In this manuscript, we revisit the previous studies on each of these systems, and show why the NV centers in SiC are favored.

The NV center in diamond is a point defect with a 0/- level deep in the band gap of diamond (5.5eV). It is formed by a carbon substituted nitrogen atom in the nearest neighbor position of a carbon vacancy. Nitrogen doping is generally performed by ion implantation which at the same time introduces carbon monovacancies, the second ingredient to form NV centers, with energies of keV to MeV giving a yield of NV formation <50% (Pezzagna et al., 2010). Post-implantation thermal treatment at temperatures around 600°C–800°C is then required to induce diffusion of the vacancies and the formation of nitrogen vacancy pairs. As in the annealing considerable number of vacancies can be captured by other type defects (other than nitrogen ions), electron irradiation at kinetic energies of 1 MeV can enhance their concentration and improve the yield of NV formation (Newton et al., 2002). It has trigonal C_3v symmetry oriented along the [111] direction of the diamond structure (Walker, 1979). Among the various stable charge states, namely, NV⁰, NV⁺¹ and NV⁻¹, it is only the latter which is of interest for qubit applications. The NV- center is characterized by a triplet ground state, ³A, with total spin S = 1 and a charge transition level at E = 2.545 eV from the top of the valance band (Subedi et al., 2019). The Hamiltonian H of a spin S = 1 center is given by: H = D S z + γ S S →   B → + S →   A ═   B → + γ I I →   B → (1)

For the NV⁻ center in diamond the zero-field splitting D of the ground-state levels m_s = 0, and m_s = ±1 is D = 2.87 GHz (Wrachtrup and Jelezko, 2006). The first excited state (³E triplet) is at 1.94 eV, which situates the zero phonon lines (ZPL) at λ = 637 nm (Chen et al., 2011). Of crucial importance is the existence of intermediate singlet states, ¹A₁,¹E₁ below the ³E state (Goldman et al., 2015), which allow to optically polarize the spins of the ground-state (Tetienne et al., 2012). To obtain the NV centers in the negative charge state the control of the Fermi level is required (Deák et al., 2014). Donor doping of the samples (Lühmann et al., 2019), hydrogen passivation (Mittiga et al., 2018), and electric field gating of the samples (Schreyvogel et al., 2016) have been shown to solve this problem. While the deep electronic level of the NV center in insulating matrix of diamond is a determining parameter to isolate them from environmental disturbance, still the readout of the qubits at room temperature is a limiting problem. In this context, one of the proposed strategies is the information transfer from electron spins to nuclear spins trough electron-nuclear spins interaction, using SWAP gates (Dolde et al., 2014). In this single shot process, using the relevant microwave pulse sequences, the spin state of the electrons are transferred to doublet nuclear spins prior to readout and then retransferred to them in order to prevent state modification upon quantum measurement. Electron-nuclear spins interactions are achieved, due to their short-range effectiveness, wherever the defect is in the vicinity of a non-zero nuclear spin. This is achieved either by the coupling to naturally present ¹³C isotopes (Smeltzer et al., 2011) or the implanted ones (Haruyama et al., 2019); their location is obviously uncontrollable (even though the interaction is very strong), or via ¹⁴N (I = 1) and ¹⁵N (I = 1/2) implantation (Rabeau et al., 2006) and doping (Felton et al., 2009). The dipolar hyperfine interaction has a 1/r³ distant dependence and its strength is only a few MHz (between 2 and 5 MHz) for nearest neighbor nitrogen atoms (Gali et al., 2008).

It was pointed out that several color centers in SiC present similar properties as NV center in diamond. In fact, the most studied polytypes of SiC, namely 3C and 4H have large indirect band gaps (E_G = 2.2 eV and 3.2 eV) respectively (Zhao and Bagayoko, 2000; Gali, 2009) in which the defect states can be well isolated. Color centers, such as the silicon/carbon divacancy pairs, V_SiV_C, in hexagonal 4H-SiC, have particularly been in the center of the research (Christle et al., 2017). This defect has been observed as a native defect at low 10¹⁵cm⁻³ concentrations in bulk 4H-SiC substrates (Koehl et al., 2011); but by high energy electron irradiation a large concentrations (>10¹⁸cm⁻³) can be generated in a controlled way (Hazdra and Vobecký, 2019; Crook et al., 2020). Here, again, post-irradiation annealing, at typical temperatures T = 850°C–1,050°C, is required for silicon monovacancy diffusion and V_SiV_C pair formation (Son and Ivanov, 2021). However, contrary to diamond, different type of NV centers can be formed depending on the lattice site of the carbon vacancy - 4H-SiC has two non-equivalent carbon sites with hexagonal (h) or quasi-cubic (k) neighboring symmetries and the location of the N atom, on-axis or in the (0001) plane (Davidsson et al., 2018). Thus we have two axial NV centers (kk,hh) and two basal NV centers (hk, kh). In addition, divacancy centers in the vicinity of stacking faults generate an additional variety, which results in the possible formation of seven distinct VV centers with slightly different magneto-optical properties (Ivády et al., 2019). For instance, the photoluminescence peaks (PL 1–7) vary from 1,037 to 1,132 nm in the near-infrared range, increasing their selectivity and addressability as distinguishable color centers. Their Debye-Waller factor of 5.3% (Christle et al., 2017) for isolated defects in epitaxial non-enriched layers, can be significantly increased to 75% using nanocavities (Gordon et al., 2015). The relatively small electron-phonon interaction is also reflected in the negligible Stokes shift of the optical absorption and photoluminescence spectra (60 nm) (Gordon et al., 2015). The VV centers has various stable charge states in 4H-SiC, among which the neutral state with S = 1 is of interest. Its groundstate zero-field splitting varies between 1.14 GHz and 1.365 GHz for the four distinct divacancies in 4H-SiC (Falk et al., 2013). The temperature dependence of the ZFS parameter D has been interpreted in the Debye model (Lin et al., 2021). Further, each divacancy defect, has its specific hyperfine structure related to ¹³C (I =1/2) and ²⁹Si (I = 1/2) neigbours. This increases the scalability of the defect and the number of addressable qubit states. In the case of proximity to ²⁹Si isotopes the so-called PL1-PL4 defects (kk,hh,hk,kh) have hyperfine splittings ranging from 9 to 13 MHz (Falk et al., 2015). The swap and CNOT gates between the nuclear and electronic spins for transferring the polarization states (information) have been realized via various MW pulse sequences (Bourassa et al., 2020). The time required to perform the gate is of the order of µs whereas the coherence time of the defect in the presence of electron-nuclear spin interaction is in the order of ms (1,000 operations); sufficient for error correction deployment.

While the divacancy color centers in 4H-SiC has gained a lot of attention for quantum technology applications, the NV center, i.e., the V_SiN_C pair defect, was also identified in the three main polytypes 3C,4H, 6H (Zargaleh et al., 2018a; von Bardeleben et al., 2015; Khazen et al., 2019; Csóré et al., 2017) and shown to combine the advantages of both systems. Similar to diamond, the center is formed by a nitrogen substitutional (N_C) defect paired to a vacancy (V_Si), and similar to divacancies in SiC, depending on the polytype in which it is formed, it exists in various configurations with non-equivalent magneto-optical properties. For instance, in 4H we have four and in 6H in six distinct NV center configurations. The nitrogen impurity in SiC is a low concentration native defect (Nashiyamal et al., 1992) and higher concentrations can be obtained by doping during the growth (Slack and Scace, 1965; Ivanov et al., 1996) or by ion implantation (Hirano and Inada, 1995; Gimbert et al., 1999). Silicon and carbon monovacancies can be generated by electron irradiation at energies E > 100 keV and it is again the post irradiation annealing at T = 700°C, which finalize the pair formation. For NV centers in SiC it is the negative charge state–as in diamond-which produces the interesting qubit. The spin-photon interface has already been demonstrated in the case of negatively charged NV centers with a S = 1 ground state and a ZFS parameter D of 1.3GHz; is numerical value is polytype-, symmetry- and temperature dependent (see Table 1) (Csóré et al., 2017). The first excited state E₁ is located within the band gap for all NV center and polytypes. The lowest energy optical transition is in the infrared range (1100–1300 nm) (see Table 1 for details) (Zargaleh et al., 2018b; Khazen et al., 2019; von Bardeleben et al., 2021) and interestingly due to this energy range, contrary to NV center in diamond, the photoexcitation, does not lead to a simultaneous photoionization (Fu et al., 2010; Löfgren et al., 2020). This eliminates blinking effects, making the NV centers much more practical for quantum information applications. As for NV centers in diamond and VV in SiC the presence of an intermediate singlet state allows optically induced ground state spin polarization. The NV center in SiC has been shown both theoretically and experimentally to be a single photon emitter with a small second order correlation factor g² = 0.03 (Wang et al., 2020a). The NV centers are also expected to present high quantum efficiency due to their small Huang-Rhys factor of 2.64 and 2.85 for 4H and 3C-SiC respectively (Hashemi et al., 2021; von Bardeleben et al., 2021), and the larger Debye-Waller factor at both low and high temperatures (two-fold enhancement) (von Bardeleben et al., 2021). Their weak electron-phonon interactions with the lattice shows up in the low temperature photoluminescence spectra. They present high intensity ZPL photoluminescence lines observable up to T = 100K (Zargaleh et al., 2018a). At higher temperatures (T = 300K) the phonon sidebands are clearly dominating and can be used for optical readout of the spin state (Wang et al., 2020b), again in analogy to the NV center in diamond and the divacancy centers in SiC. A further important parameter is the count rate of the collected photons; it does not only depend on the physical properties (quantum efficiency) of the defect itself, but also on the total reflection of the surface (sample dependence), interference of other defects and the collection efficiency (measurement setup dependence). In the case of NV center in diamond it has been shown that the photoluminescence emission reaches a saturated value of the order of 10⁸ counts/s as a function of laser power. Due to experimental and sample dependent limitations, the experimental count rates are often only a few percent of this value decreasing the sensitivity of this system. However, using photocurrents as a probe of the spin state, one can increase the sensitivity by two orders of magnitude (Bourgeois et al., 2020). The reported count rates for single NV center in 4H-SiC are in the same range as for NV centers in diamond (30kc/s) (Zargaleh et al., 2016); up to now, no PDMR measurements have yet been reported for NV centers in SiC. From EPR measurements the hyperfine interaction between the spin of electronic states and the neighboring ¹³C and ¹⁴N nuclear spins has been determined (Csóré et al., 2017). It was pointed out that this property is essential for storage of the information, prior to its readout process, which eventually would destroy the information written in the electronic states. Hence, even though the application of any intra-atomic quantum gate, such as SWAP, has not yet been demonstrated for NV centers in SiC, they have all the potential to be efficient. Moreover, a major advantage of NV center compared to divacancies is that they benefit from the presence an on-site nuclear spin with 14N atoms. This, indeed largely facilitates the control over the location and magnitude of the interactions, particularly in isotopically purified samples, where one eliminate the randomly distributed nuclear spins.

Spin coherence times, T1 and T2 (or T2*), have been obtained via ODMR from Ramsey oscillation and Hahn echo measurements. They correspond to the spin–lattice relaxation times between two electron states and the time the superposed states persist before dephasing occurs, respectively. Hence, they determine the time during which various quantum gates and readout procedures can be applied in a reliable way. The determination of intrinsic T1 and T2 times is rather complicated, as they are highly sensitive to the presence of spins in the defect environment. The spin coherence can be disturbed via interactions with the neighboring defects via spin-spin interactions, via nuclear spin-electronic spin interactions, structural defects, electron-spin interactions with surrounding charges, environmental noises as inhomogeneous electric and magnetic fields, and also with spin-phonon interactions, which become particularly important at high temperatures. This, on one hand led to numerous reports with different experimentally determined coherence times for the same defect, and on the other hand, has led to new approaches to minimize these effects. The very first of these attempts concerned the isolation of the defects and the use of isotopically modified material to reduce the influence of ²⁹Si with a nuclear spin I = 1/2 (Christle et al., 2014). Other protocols include specific decoupling microwave pulse sequences and microwave dressings to extrinsically isolate the defects during the gate applications (Miao et al., 2020). To date, the highest reported values for NV centers in diamond, are T₁ > 1 h and T₂ = 1s at 4K (Abobeih et al., 2018). The dephasing time T₂ is the limiting parameter of the coherence as T₁ is orders of magnitude larger at low temperatures (where the contribution of the phonons is small enough). Even at high temperatures, as the phonon contribution to dephasing increases significantly, both the coherence times and their differences are reduced (they can become comparable). For instance, at room temperature the values for NV in diamond reach T1 = 6 ms and T2 = 3.3 ms, as determined by multiple decoupling pulse sequences (Bar-Gill et al., 2013). Knowing that each gate operation requires a few ns and for each operation several thousand error correction operations are necessary, the minimum coherence times for quantum computing applications are of the order of few ms. This has indeed been largely achieved for the NV center in diamond even at room temperature. Interestingly, the coherence times of the divacancies in SiC have shown to be more promising. Very recently, a value of T2 = 5s at T = 5K for the neutral divacancy in isotopically purified 4H-SiC, using microwave dressing protocols, has been reported (Anderson et al., 2022). This is the highest recorded value among for solid state qubits. Yet, it should be noted that for room-temperature applications this value is expected to reduce significantly to a few tens of milliseconds. Among the many studies, which have been dedicated to the optimization of the coherence times for the aforementioned defects, it is surprising that only very few are reported for the case of NV centers in SiC. The only reported values correspond to ¹²C-implanted isotopically natural 4H-SiC samples; there at low-temperature (T = 15K) dephasing time for basal and axial defects are in the range of 1 ms, in agreement with DFT calculations (Zhu et al., 2021). The measurement of the coherence time on a similar sample at room-temperature has presented T₂ ≈ 20µs using conventional Hahn-echo measurement without defect isolation (Li et al., 2021). It is very important to note that, when comparing results for these three defects, one should take into account the similar sample conditions. The initial reports concerning the dephasing times in NV centers in diamond and divacancies in SiC at room temperature have shown values ranging from 6 μs at room temperature to few hundreds of µs at T = 20K (Falk et al., 2013). These values are nearly one order of magnitude smaller than the reported value for NV center in 4H-SiC. Hence, although the coherence times for NV centers in SiC have not been measured in detail and using various defect isolation procedures, a simple comparison between the presented values and their extrapolation to NV centers under optimized conditions, lead us to expect coherence at least as large as the other two systems.

The possibility to select a particular defect for information registration, processing and readout relies on two parameters. The first one is the electronic structure (electron and nuclear), which is specific for each defect and each addressed state. For all of the aforementioned defects, its value is comparable and of the order of MHz to GHz, hence addressable via microwave pulses. The second one is the spectral range for optical excitation/recombination. Due to the weak electron-phonon interactions in the host materials, the Stokes shift is small and the spectral range has been measured precisely from the photoluminescence experiments at various temperatures. At low temperatures, where the zero phonon lines (ZPL) dominate the optical absorption and emission spectra, high resolution measurements of the relevant excitation wavelengths are possible. In the case of the NV center in diamond the absorption ZPL is at λ = 637 nm and can still be observed at temperatures above room temperature (Zhao et al., 2012); this is one of the parameters why this defect is applicable for practical quantum technology applications. However, due to Franck-Condon shift in optical absorption/emission transitions of NV center, the excitation should be performed at higher energies (Gali et al., 2011). This results in the photoionization of the center due to simultaneous defect to band transitions (Razinkovas et al., 2021). Usually, photoexcitation with λ = 583 nm is used for initializing the spin state of the NV center in diamond (Aslam et al., 2013). Several defects in SiC can also be spin polarized by optical excitation. However, the spectral range (near infrared) for VV and NV centers in 3C,4H,6H-SiC polytypes is advantageous over NV in diamond. It is situated in the range (1000–1300 nm) more compatible for information transfer through optical fibers, generally used in the present state of the art of telecommunication (Wang et al., 2018; Cao et al., 2019). Further, each defect (hh,kk,hk,kh) (VV, NV) in the different polytypes of SiC (3C, 4H,6H) has a distinct ZPL, depending on their local symmetries, which increases the selectivity and flexibility of the qubits depending on the application and technical resources. For the neutral divacancy centers VV with ZPL named PL1-PL4 and VV-X complexes (PL5 and 6) in 4H-SiC, six different ZPL have been observed, with wavelength from 1,000 nm to 1,260 nm (Table 1) (Koehl et al., 2011). The ZPL of the NV center in SiC (Csóré et al., 2017) and particularly in 3C-SiC (1,289 nm) (Felton et al., 2009) is well suited for transition in the telecom O-band. By this technology the information, transferred to flying qubits from the localized electronic qubits, can be transported easily over long distances using quantum repeaters embedded inside the fibers (Buterakos et al., 2017) and quantum entanglement (Whiteley et al., 2019; Bourassa et al., 2020) at large distances are feasible.

Apart from the intrinsic properties of the solid-state spin-qubits, their efficient and cost-effective scalability and addressability are determining factors for their implementation as practical technological information systems. These factors are mainly governed by the host matrix and defect generation process. It is worth mentioning that these points are different to the intrinsic capability of the qubits to be scaled or addressed which were discussed in previous sections.

In Table 1, we have summarized the main conclusions of our discussions. It is surprising that despite the clear advantages of NV centers in SiC over divacancies in SiC and NV center in diamond, much less attention and work has been dedicated to development of this high potential qubit. There are several other potential qubit are presently under examination and investigation and there is an ever-growing number of reports regarding the achievements in elucidation of, both, their fundamental properties and devices and technologies, which are designed based on them. This includes all systems from trapped ions ¹ , superconductors ² , quantum dots ³ , and neutral atoms ⁴ to the very recently developed 2D systems such as h-BN (Li and Gali, 2022). They have even been successful to receive significant financial supports to build technology startup units. In this context, it is a surprise that the case of NV center in SiC, with its high potential suffers from the lack of attention. NV centers in 4H-SiC not only dispose of properties adapted to quantum computing, but also for other applications functioning at room-temperature such as quantum sensing, metrology, and communication applications; further they are biocompatible, have a large natural abundance of its constituent elements, are suitable for industrial fabrication and processing, and cost efficient.