Coherent spin-control of S = 1 vanadium and molybdenum complexes

The burgeoning field of quantum sensing hinges on the creation and control of quantum bits. To date, the most well-studied quantum sensors are optically active, paramagnetic defects residing in crystalline hosts. We previously developed analogous optically addressable molecules featuring a ground-state spin-triplet centered on a Cr4+ ion with an optical-spin interface. In this work, we evaluate isovalent V3+ and Mo4+ congeners, which offer unique advantages, such as an intrinsic nuclear spin for V3+ or larger spin–orbit coupling for Mo4+, as optically addressable spin systems. We assess the ground-state spin structure and dynamics for each complex, illustrating that all of these spin-triplet species can be coherently controlled. However, unlike the Cr4+ derivatives, these pseudo-tetrahedral V3+ and Mo4+ complexes exhibit no measurable emission. Coupling absorption spectroscopy with computational predictions, we investigate why these complexes exhibit no detectable photoluminescence. These cumulative results suggest that design of future V3+ complexes should target pseudo-tetrahedral symmetries using bidentate or tridentate ligand scaffolds, ideally with deuterated or fluorinated ligand environments. We also suggest that spin-triplet Mo4+, and by extension W4+, complexes may not be suitable candidate optically addressable qubit systems due to their low energy spin-singlet states. By understanding the failures and successes of these systems, we outline additional design features for optically addressable V- or Mo-based molecules to expand the library of tailor-made quantum sensors.


Introduction
9][20][21][22] The electronic structure of these color centers provides a valuable platform for quantum control.First, the ground state spin sublevels of these defects provide a two-level quantum system to act as the qubit.Crucially, the spin sublevel populations may be optically prepared into a non-thermal equilibrium state, i.e. initialized, while the spin state information may be read out using spindependent optical emission. 18,23These combined features enable remote optical control and single qubit addressability, both of which are valuable features for quantum sensing technologies. 24,25lthough solid-state color centers offer incredible coherence properties and quantum control, deterministic spin localization and defect tunability remain major challenges. 77][28][29][30][31] To create bespoke quantum sensors, we sought to develop molecular analogues of solid-state color centers that host a similar optical-spin interface, but with their molecular nature facilitating bottom-up design and solution processibility towards specic sensor-analyte interactions. 32,33n recapitulating this optical-spin interface, we needed (i) a ground state spin that can be coherently controlled, (ii) an electronic excited state that is 'connected' to this ground-state spin through a spin-selective optical process (e.g., excitation or relaxation, see ESI † for further discussion), (iii) radiative decay from the excited state to the ground state for optical readout, and (iv) a ground-state spin lifetime, or spin-lattice relaxation time (T 1 ), that is longer than the optical lifetime (s opt ). 34][37][38] While these generic features provide a framework to build molecular color centers, S = 1 molecular systems pose two key challenges for spin-dependent optical readout.First, to achieve a spin-selective optical process and emission in S = 1 complexes, a ligand eld that ensures that the rst excited state is a spin-singlet excited state ( 1 ES) is desirable. 39For resonant optical control (off-resonant optical control protocol addressed in the ESI †), the spin-ip transition from the 1 ES to the spintriplet ground state ( 3 GS) should exhibit limited vibronic coupling, allowing for both narrow optical linewidths for spinselective optical excitation and radiative decay to the ground state. 34,40This desired electronic structure requires a sufficiently strong ligand eld around the spin-bearing metal center such that the rst excited state is a 1 ES, eliminating lower energy 3 ES's that provide non-radiative decay pathways. 39Second, the ground-state spin must be capable of coherent control.However, the spin of S = 1 transition metal systems oen cannot be coherently controlled with readily accessible microwave frequencies (i.e., 1-40 GHz).The inability to coherent drive the ground-state spin of these systems arises from both the integer spin and large zero-eld splitting (ZFS) values, requiring high microwave frequencies (>95 GHz) to probe ground-state spin transitions. 41,42This challenge may be overcome by designing orbitally non-degenerate, S = 1 ground states, wherein the spin-bearing ion resides in a (near) cubic ligand eld.
Considering these design parameters, we previously demonstrated that pseudo-tetrahedral (T d ) Cr 4+ molecules in a strong ligand eld exhibit small ground-state ZFS values of 1.8-4.1 GHz for microwave control at X-band frequency (9-10 GHz), and an optical-spin interface. 34,40Yet, the air-sensitivity, small spin-orbit coupling, and low percentage of nuclear spin-bearing isotopes of Cr 4+ -based systems may not be optimal for every sensing task.Herein, we aim to translate this combination of features from T d Cr 4+ systems into molecular hosts that offer (i) intrinsic nuclear spins, I, to serve as quantum memories for prolonged information storage, 7,43,44 and (ii) larger spin-orbit coupling for improved sensitivity to electric elds. 35,45To that end, we turned to T d trivalent vanadium (V 3+ ) and tetravalent molybdenum (Mo 4+ ) derivatives in the same ligand elds as their Cr 4+ congeners.The vanadium-51 (i.e., 51 V, I = 7/2, 99.8% abundance) center provides a potential nuclear spin memory that is intrinsically coupled to the electronic spin through hyperne interactions, 42,46 while the Mo 4+ center increases the spin-orbit coupling experienced by the electron spin, offering enhanced sensitivity of the spin Hamiltonian parameters to certain external elds through the Zeeman splitting term (e.g., strain or electric elds). 47,48However, the metal ion substitution should substantially change both the electronic structure and spin dynamics that dictate the quantum sensor performance.To initiate our studies with these ions, we synthesized and evaluated three spin-triplet systems with V 3+ and Mo 4+ to compare with previously studied Cr 4+ congeners.Coupling spectroscopic analysis with computational predictions, we evaluate how to design these systems, addressing their advantages and limitations.From these results, we suggest that reducing nearby high energy oscillators and using ligand scaffolds could enable both emission and low frequency microwave control for V 3+ systems while Mo 4+ systems may be fundamentally limited by their intrinsically low lying 1 ES.

Results and discussion
We synthesized [Li(THF)   5) 51 (Fig. 1).][51] We also note that previous efforts to synthesize the molybdenum analogue of Cr(trimethylsilylmethyl) 4 resulted in the hexakis(trimethylsilylmethyl)dimolybdenum species, so we could not investigate the Mo 4+ analogue of 5. 52,53 We rst characterized the solid-state structure of 1 and 3 through single crystal X-ray diffraction, nding that their inner MC 4 coordination sphere remains close to ideal tetrahedral symmetry based on both the s 4 and s 4 0 metrics (0.95-1.01). 54,552 had been previously characterized at room temperature and shows only a slight contraction in Mo-C bond lengths upon cooling to 100 K. 49 As such, these systems should exhibit relatively small, but non-zero, ZFS values.The ground-state structure may be approximated as a descent in symmetry from ideal T d to C 2v , as discussed for [Li(THF) 4 ][V(mesityl) 4 ] and [Li(THF) 4 ] [V(pentachlorophenyl) 4 ]. 56icrowave addressability remains a challenging parameter to design and tune in S = 1 compounds. 41,42Thus, we initiated our investigation by probing the ground-state spin structure of these systems to determine if they are capable of ground-state spin control.We characterized 1-3 with both continuous-wave (cw) and pulsed electron paramagnetic resonance (EPR) to understand their ground-state spin structure and dynamics.To reduce decoherence from electron spin-electron spin interactions, we cocrystallized 1-3 in their corresponding isostructural, diamagnetic analogues, [Li(THF 57 and [Li(12-crown-4) 2 ][Al(trimethylsilylmethyl) 4 ] (3-Al), respectively (see ESI † for details).We denote the cocrystallized samples as 1 0 -3 0 for all subsequent experiments with relative electron spin concentrations of 0.36-3% (Table S8 †).From the X-band cw-EPR spectra in Fig. 2a-c, we nd that the axial ZFS values, D, are larger for the V 3+ and Mo 4+ derivatives than the corresponding Cr 4+ congeners. 401 0 and 2 0 exhibit jDj values of 5.62 GHz and 7.3 GHz, respectively, while jDj for 4 0 at 5 K is 3.63 GHz. 34The largest contribution of the two-fold increase in jDj for 2 0 likely arises from the increased spin-orbit coupling of the heavier Mo 4+ ion, where the free ion spin-orbit coupling parameter, l, is ∼425 and 167.5 cm −1 for Mo 4+ and Cr 4+ . 58Given the similarity in l for V 3+ and Cr 4+ , the increase in jDj for 1 0 more likely results from symmetry lowering around the V 3+ center due to crystal packing with the [Li(THF) 4 ] + cation.For example, jDj = 5.62 GHz for 1 0 is similar to the jDj value of 5.55 GHz for Cr(otolyl) 4 (4) diluted in a lower symmetry (orthorhombic) host matrix. 59hile 1 0 and 2 0 show modest increases in jDj relative to their Cr 4+ analogue (4 0 ), 3 0 exhibits a dramatic increase to 16.6 GHz, which is ∼13 times greater than its Cr 4+ analogue (5 0 ) where jDj = 1.23 GHz. 40To understand this order-of-magnitude increase in jDj for 3 0 , we quantied the deviation from ideal T d symmetry using the s 4 and s 4 0 parameters 54,55 for 3 0 and 5 0 .Between 3 0 and 5 0 , the s 4 and s 4 0 values vary by only 0.004 and 0.015, respectively.While these geometric deviations are small, the absolute spin splitting energies of 3 0 and 5 0 vary by only 63 meV, suggesting that subtle structural changes may result in signicant variations in jDj.However, variations in jDj may also arise from a convolution of competing effects, such as changes in SOC, ligand eld strength, symmetry-driven coupling to excited states, and electron delocalization. 60,61Thus, to maintain microwave addressability across diverse host media for sensing, it will be key to accurately predict and measure how growth on different surfaces or substrates affects local structure around the spin center and, consequently, jDj. 62he two V 3+ complexes, 1 0 and 3 0 , also exhibit two advantageous features.First, the crystal symmetries of 1 0 and 3 0 result in non-zero transverse ZFS, jEj, values of 0.8 and 2.2 GHz, respectively.Non-zero jEj should result in clock-like transitions around zero magnetic eld at frequencies of jD − Ej, jD + Ej and 2jEj, resulting from mixing of the M S = +1 and −1 states (Fig. S1 †). 63Clock-like transitions show little variation in their transition frequency with small changes in applied magnetic eld. 64,655][66][67] Second, the cw-EPR spectra in Fig. 2a and c exhibit hyperne coupling to the 51 V nucleus of 155 and 165 MHz for 1 0 and 3 0 , respectively.Thus, the built-in 51 V nuclear spin and relatively large hyperne coupling to the electron spin provides a potential nuclear spin memory to prolong the storage time of quantum information. 44,68,69For example, previous work with S = 1/2 V 4+ as well as 173 Yb 3+ molecular spins suggested that the coherence of strongly coupled nuclear spin may be prolonged beyond electronic spin coherence time by increasing the magnetic eld strength. 46,70As a result, both the clock-like transitions around zero eld and the intrinsic nuclear spin of the V 3+ -based systems provide potential routes to lengthen coherence lifetimes.
To then determine the spin dynamics of these Mo 4+ and V 3+ systems in their isostructural matrices, we examined the temperature and magnetic eld dependence on spin-lattice relaxation (T 1 ) and coherence (T m ) times for 1 0 -3 0 at X-band frequency (9-10 GHz).In each case, these complexes exhibit shorter T 1 times (10-15 ms) at low temperature than 4 0 or 5 0 , where we previously measured T 1 times of 2-3 ms at 5 K. 40 However, the concentration of both 1 0 and 3 0 were 1.5 and 3% in the spin-diluted lattice, which we have previously shown leads to a reduction in T 1 at low temperatures for 4 0 . 40Thus, T 1 for 1 0 and 3 0 may likely be improved with further dilution.Notably, 1 0 and 3 0 exhibit T 1 times of >0.5 ms up to 30 K while 2 0 shows a steep decline in T 1 < 0.5 ms by 18 K (Fig. 2g-i).These contrasting temperature dependencies likely arise from the larger spin-orbit coupling of the Mo 4+ -based spin centers, as well as lower energy phonon modes arising from the heavier metal center. 71,72urning to the T m times, a key metric to evaluate the performance of qubit candidates, we nd relatively similar temperature dependencies for 1 0 -3 0 .In general, each compound exhibits T m times of approximately 1 ms at 6 K, likely limited by the high density of 1 H nuclei on the ligands. 23,73,74To verify that the nuclear spin environment inhibits T m , we also performed power-dependent Hahn-echo experiments to mitigate the inuence of instantaneous diffusion from nearby paramagnetic spin centers in relatively spin concentrated samples. 64,75We nd that the T m times of 1 0 and 3 0 show no signicant change with decreasing power (Fig. S2 †), suggesting that the nuclear spin environment of the matrix limits T m in these samples.In contrast, the T m time of 2 0 approaches ∼2 ms at decreasing small microwave powers, which is similar to the T m time previously measured for 4 0 at similar concentrations 40 (see ESI † for further details).This analysis does not result in large increases in T m times, indicating that coherence times reported here are likely limited by the 1 H nuclear spin environment.The T m times then decrease with increasing temperature until the T m times are limited by T 1 .The larger spin-orbit coupling in 2 0 likely leads to T 1 -limited T m times by 18 K while the V 3+ -based systems are measurable up to at least 30 K.However, both the T 1 and T m times in 1 0 and 3 0 are shorter than their Cr 4+ congeners, 4 0 and 5 0 . 40We posit that in these low eld measurements, we may be either simultaneously driving multiple overlapping spin transitions (Fig. S1 †) or driving a highly mixed spin transition, which has previously resulted in decreased spin-lattice relaxation times for an S = 5/2 Fe 3+ complex. 76However, this assertion requires a multifrequency EPR study for validation, which will be the subject of future work.Additionally, we note that the T 1 and T m times reported here for 1 0 -3 0 will vary substantially when measured in other matrices as the matrix tends to most strongly inuence the spin dynamics of molecular qubits. 77e further evaluate the magnetic eld dependence of these T m times.Similar to the behavior of 4 0 , 40 we nd local maxima in T m for 2 0 at 170, 460, and 620 mT, where the applied eld is parallel or perpendicular to the principal axis of the ZFS tensor (Fig. 2e).Across this eld range, T m changes by a factor of ∼4.For 1 0 and 3 0 where E > 0 and I = 7/2, the eld dependence of T m is less exaggerated since these systems exhibit are more transitions over a similar eld range (Fig. S1 †).As a result, T m changes by less than a factor of 2 for 1 0 over a similar eld range to 2 0 .While the variation is less signicant for 1 0 and 3 0 , these data still highlight that no single value of T m provides a complete picture of the spin relaxation, especially for anisotropic metal complexes.We similarly expect T 1 to show a strong magnetic eld dependence for 1 0 -3 0 , as demonstrated with Cr(otolyl) 4 . 78ost importantly, even the maxima in T m times measured here still fall well short of coherence times for state-of-the-art quantum sensors, such as anionic nitrogen vacancy centers in diamond (e.g., from 10 s to 100 s of microseconds). 79In fact, low temperature (#10 K) T m times across most transition metalbased systems to date are #15 ms, regardless of spin state, 30,73,77,80,81 ligand nuclear spin environment, 77,82 or magnetic eld, 40,83 illustrating how the surrounding matrix promotes decoherence.Thus, introducing clock-like transitions (e.g., Fig. S1 †) into molecular sensors while be critical to improve their coherence properties, and consequently their sensitivity, 84 in magnetically noisy matrices. 23,63urning to the excited state structure, we initially measured optical absorption spectra of these systems in solution at room temperature.The solution-phase UV-visible near-infrared spectra in Fig. 3a illustrate that the Mo 4+ ion in 2 leads to both higher energy and more intense transitions than either of the rst row congeners, 1 and 4. The UV-vis-NIR spectrum of 2 also appears qualitatively similar to 4, but the transitions in 2 are shied to higher energy (Fig. S6 †).This spectrum suggests that, similar to 4, the 1 ES of 2 is the lowest energy excited state, which should yield the correct energy level structure for opticalspin control. 34Conversely, the V 3+ ion in the pseudo-T d o-tolyl ligand eld of 1 leads to lower energy transitions than both 2 or 4. In fact, 1 exhibits absorption extending well into the NIR, leading to substantial spectral overlap with aromatic C-H overtones (see green and blue shaded regions in Fig. 3a), which may provide a multi-phonon mediated non-radiative decay pathway. 85hen replacing the aryl ligands in 1 with (trimethylsilyl) methyl ligands in 3, we nd that the absorption features from 1100-1400 nm are suppressed while the shorter wavelength transitions from 400-1000 nm appear similar between 1 and 3.The reduction in absorption between 1100-1400 nm for 3 suggests that either the stronger alkyl ligand eld increases the transition energy of the 3 ES manifold or the s-only alkyl ligands do not mix with the lowest energy 3 ES metal-centered transitions, resulting in very low oscillator strengths of the d-d transitions.We observed similar behavior in tetraaryl-and tetralkyl-Cr 4+ systems.For example, 4 exhibits intense absorption between 600 and 800 nm while those transitions are present, but much weaker, for 5. 40 To determine if 1-3 exhibited emission from a 1 ES, we then performed steady-state photoluminescence measurements.When exciting pure, single-crystalline samples of 1-3 at 4 K with 660 or 785 nm excitation, we observe no emission in the range of 900-1700 nm.Even performing photoluminescence experiments on dilute single-crystalline samples of 1 0 -3 0 or lms of 5-10% (w/w) of 1-3 in polystyrene at 4 K yielded no emission.The lack of emission prevents optical readout of the ground-state spin, the essential component for molecular color centers.We sought to understand why some systems emit and other similar ones did not, to see if failure might serve as a guide for future systems.
We performed spin polarized density functional theory calculations using VASP (Vienna Ab initio Simulation Package) 6.3.2 (ref.86-89) with projector augmented wave pseudopotentials 90,91 and the PBE exchange correlation functional 92,93 (see ESI † for further details).For calculations of the charged molecules 1 and 3, we explicitly included counterions but, for clarity, we only highlight spin-bearing orbitals here.Fig. 4 shows the electronic structures of 1-3, highlighting the d character of the molecular orbitals.We visualized the highest occupied molecular orbital (HOMO) and spin up and down lowest unoccupied molecular orbital (LUMO) orbitals.Similar to 4 and 5, 40 all molecules have a HOMO and LUMO with signicant d character.Thus, the qualitative picture of all ve compounds is similar, yet only 4 and 5 exhibit measurable emission.
To evaluate where the 1 ES lies relative to the 3 GS manifold, we performed excited state calculations using the DSCF (delta self consistent eld) method. 94,95We calculated the 1 ES-3 GS gap by subtracting the total energy of a constrained occupancy calculation, where the electron in the spin-up HOMO is promoted to the spin-down LUMO, from a standard groundstate DFT calculation (see ESI † for further details).We calculate 1 ES-3 GS gaps of 0.355 eV, 0.319 eV, and 0.407 eV, eV for 1, 2, and 3, respectively.Previously, we estimated the 1 ES-3 GS gap is 0.537 eV for 4 using the same methodology. 62Thus, the calculated 1 ES energies for 1, 2, and 3 are only 66, 59, and 76% of 4.Although these values have limited quantitative accuracy, the trend suggests that the 1 ES energy of 1 and 2 is signicantly lower than 4.These results align with the expectation that the spin-pairing energetic penalty decreases with increasing ionic radii, such that the 1 ES energies should be Mo 4+ < V 3+ < Cr 4+ in the same ligand eld.We further nd that 3 also exhibits a lower calculated 1 ES value than 4, despite the stronger alkyl ligand eld of 3.
If we estimate hypothetical emission wavelengths by assuming the ratio between the calculated 1 ES and experimental emission energies is similar to 4, where calculated 1 ES/ emission energy = 0.537/1.209eV, 34,62 we nd that 1, 2, and 3 would emit at roughly 1550, 1800, and 1400 nm.Considering this wavelength range, the lack of measurable emission may therefore be rationalized by four possible effects.First, any photons emitted at longer wavelengths than 1700 nm are outside of the range of commercially available InGaAs NIR detectors that are optimal for typical NIR emission measurements.Second, if any emission occurs in the measurement window, the higher strain sensitivity of 1-3 with lower symmetry or larger spin orbit couplings could result in broader emission than Cr 4 derivatives, resulting in fewer photons counted per pixel in the CCD measurements, rendering the measurement less sensitive.Third, the posited lower energy 1 ES ([1200 nm) may exhibit signicant spectral overlap with high energy C-H stretching overtones (e.g., Fig. 3) that provide highly favorable non-radiative decay through multi-phonon mediated relaxation pathways.This potential explanation for the lack of emission could be ameliorated through selective deuteration or uorination of the ligands. 96Further experimental and theoretical studies are required to investigate this pathway.Fourth, the energy-gap law suggests that as the emissive state becomes lower in energy, non-radiative decay rates will increase. 97,98In each compound, a combination of these effects may be operative, leading to no detectable emission.0][101] These results suggest a pathway forward for the design of molecular color centers featuring depleted C-H modes for compounds that emit in the near-IR or telecom region.
Designing emissive V 3+ systems with a microwave addressable ground-state spin The aggregate of these results provides us with a series of new S = 1 molecules featuring coherent control over their groundstate spin and a set of design principles for next generation quantum sensors.Previous data illustrate that emissive V 3+ systems may be achieved using trigonal bipyramidal or octahedral geometries. 102,103However, trigonal bipyramidal or octahedral geometries with d 2 metal ions result in either a noncubic (e.g., C 3v , D 3h ) symmetry or an orbitally degenerate ( 3 T 2 ) ground state, respectively.These features result in jDj values that well exceed frequencies of conventional microwave sources, making their ground-state spin control more challenging.Thus, we attempted to unify the emissive nature of ve and six coordinate V 3+ with the low jDj values of pseudo-T d by introducing the V 3+ ion into sufficiently strong ligand elds.However, we never observed emission for V 3+ systems, even with the strong-eld alkyl donors of 3.These results suggest that a stronger ligand eld should be enforced through a rigid ligand coordination with, for example, C 3 symmetric ligand scaffolds (e.g.tris(pyridyl)methane, tris(N-heterocyclic carbene)borate) or C 2 -symmetric bidentate ligand scaffolds (e.g.1,1 0 -binaphthyl-2,2 0 -diamine). 104Notably, increased rigidity is oen ascribed to improved radiative efficiency of Cr 3+ systems. 96However, the hard nature and high charge of early transition metals likely necessitates using suitably hard, anionic ligand donors, precluding the use of soer strong eld ligands such as phosphines.Additionally, the lower energy transitions in V 3+ systems likely undergo rapid non-radiative decay mediated by nearby C-H groups.Thus, the spectral overlap of these modes with excited states may be mitigated through deuteration or uorination, also demonstrated with Cr 3+ systems to greatly enhance emission. 96,99,100A similar approach may be employed with future systems to realize emission from these V 3+ systems that feature generally small (e.g., <30 GHz) jDj values.

Designing emissive Mo/W systems with microwave addressable ground-state spin
In contrast to the V 3+ analogues, Mo 4+ -based systems pose a unique challenge.The energy gap between the 1 ES and 3 GS is generally quite small given the large size of the Mo 4+ center.Additionally, to unlock the full potential of large spin-orbit coupling from the Mo 4+ center to achieve enhanced sensitivities of the spin to electric eld perturbations, the ion should sit in a polar molecular symmetry, such as C 3v . 45For Mo 4+ systems, these symmetries may simply yield an S = 0 complexes or prohibitively large jDj values.Thus, for Mo-and even W-based quantum sensors, an alternative spin-state may be better suited to enable strong spin-electric eld coupling. 47,105For example, Mo 5+ S = 1/2 defects in silicon carbide have demonstrated NIR emission 106 and may be capable of coherent optical control, similar to anionic tin vacancies in diamond. 107Thus, future investigations with S = 1/2 Mo 5+ /W 5+ systems may provide a more suitable framework to achieve a ground-state spin capable of coherent control connected to an emissive optical state than the S = 1 derivatives with second or third row metal ions.

Conclusions
In this work, we examined the electronic structure and spin dynamics of three novel V 3+ and Mo 4+ molecular color center candidates that could feature valuable spectroscopic handles for sensing.Notably, each system showed appropriately small ground-state anisotropy for coherent spin control.We demonstrated this control over a series of molecules, including a rare example of spin control in a second row transition metal with a spin-triplet ground state.We also suggest that the transverse anisotropy (i.e., jEj > 0) and nuclear spin of the V 3+ analogues may offer two avenues to extend quantum coherence.Despite these potential advantageous features, none of the systems investigated here exhibited measurable emission, likely resulting from a combination of a low energy 1 ES and rapid nonradiative decay mediated by the C-H-rich ligand environment.Coupling experimental and computational insight, we suggest that the use of ligand scaffolds and deuteration/uorination may result in emissive V 3+ while maintaining ground-state spin control, while larger transition metal ions, such as Mo and W, in S = 1/2 states may be better suited for optical-spin control.Thus, we offer design considerations that will hopefully lead to the realization of V-or Mo-based molecular quantum sensors in future studies.

Fig. 2
Fig. 2 (top) cw-EPR spectra of (a) 1 0 , (b) 2 0 , and (c) 3 0 at 5 K and 9.37-9.39GHz with resulting D and E values.* denotes V 4+ impurity in 3 0 , representing 1% of spin density in the sample.(middle) Magnetic field dependence of T m (hollow symbols) at 6 K for (d) 1 0 , (e) 2 0 , and (f) 3 overlayed on the echo-detected field swept spectrum.Error bars are within the data points.(bottom) Temperature dependence of T 1 (filled circles) and T m (hollow circles) for (g) 1 0 , (h) 2 0 , and (i) 3 0 at X-band frequency and measured at the magnetic field indicated by the arrow in the (d-f).