Arxiv:2007.11529V1 [Cond-Mat.Mes-Hall] 22 Jul 2020 Aainhsfcsdo H Ieieo the of Lifetime Re- the Phonon-Induced on on Focused Work Has Prior Laxation [12–15]

Home , ArXiv

arXiv:2007.11529v1 [cond-mat.mes-hall] 22 Jul 2020 aainhsfcsdo h ieieo the of lifetime re- the phonon-induced on on focused work has Prior laxation [12–15]. perature fthe of coherence unclear. spin remained electronic has temperature NV result, room the at a to time As limit ultimate dynamics. relaxation the of spin-lattice picture NV’s incomplete the an leaving investigated, been ously h aeo eaainotof out relaxation for of of rate account dependence the to temperature observed understood experimentally and the are acoustic Raman phonons from di- two-phonon contributions quasilocalized bulk major Specifically, re- with high-purity spin transitions [12]. in for samples temperature mechanism amond room primary at the laxation are spin-lattice that interactions indicated have concentrations defect ing ape,teevle eanfrsoto h measured the of short the far of dy- remain lifetime sophisticated diamond values high-purity these of and use samples, sequences the decoupling despite namical However, temperature 11]. previously Co- room [10, at have diamond ms bulk temperature. in 2–3 demonstrated room been approximately elec- of at long times its the time herence is of coherence thermal applications performance spin these and of the tronic 7], many to local in [6, Critical center the strain NV environment. of 5], 9] [4, sensing [8, quantum electric and as [3], 2] such magnetic [1, applications, potential processing interest of information attracted range has a that for system quantum solid-state Va w-ee ytm[1.Frhroe h phonon- the the the of Furthermore, lifetimes treating [11]. limited system of two-level validity a as the NV of consideration prompting ueoseprmna esrmnso h lifetime the of measurements experimental Numerous h irgnvcny(V etri imn sa is diamond in center (NV) nitrogen-vacancy The 2 etod netgcńDiALb autdd suisIn Estudios de Facultad Lab, Investigación DAiTA de Centro | tt-eedn hnnlmtdsi eaaino nitrog of relaxation spin phonon-limited State-dependent m s .C Cambria, C. M. 0 = | civbechrnetm o nN t25Ki iie o6.8( to limited is K W 295 at interactions. NV concentrat an spin-phonon for NV by time of coherence limited achievable independent are are they observe we indicating rates The amond. uin u ohge-re em ntesi-hnnHamilt spin-phonon the in terms relaxati higher-order Orbach-like behind to suggests due that butions results our of analysis odi ia oraiigtefl oeta fti quantum this of potential full the realizing the to on vital is mond eedneo hsrlxto rate. relaxation this of dependence m nesadn h iist h pnchrneo h nitro the of spin-coherence the to limits the Understanding | i s m 3 nttt eFıia otfii nvria a´lc eC Católica de Universidad F´ısica, Pontificia de Instituto 0 = tt narneo ape ihvary- with samples of range a in state s INTRODUCTION 1 0 = | eateto hsc,Uiest fWsosn aio,W Madison, Wisconsin, of University Physics, of Department | | ↔ i m m s s i = | = m tt otncle the called (often state m − 4 − s 1, etod netgcńe aoenlgı aeilsA Materiales Nanotecnolog´ıa y Investigación en de Centro s 1 1 = ∗ = | ↔ i | ↔ i .Gardill, A. ± ± | otfii nvria a´lc eCie atao Chile Santiago, Chile, Católica de Universidad Pontificia m 1 1 i i s m m rniin ne min odtosi aieNsi high- in NVs native in conditions ambient under transitions ttshv o previ- not have states 0 = s s +1 = +1 = i erro tem- room near 1, i i ∗ rniinocr prxmtl wc sfs srelaxatio as fast as twice approximately occurs transition eaain oiaigftr esrmnso h tempera the of measurements future motivating relaxation, .Li, Y. T | Dtd uy2,2020) 23, July (Dated: m 1 s time), 1 0 = .Norambuena, A. i tt,hwvr elcigpsil eaaino the on relaxation possible neglecting | however, state, ersraeNshsdmntae atrlxto on relaxation fast demonstrated the has NVs near-surface edniecnb xetd evn pnteqeto of the electric question of the less rate open far relaxation leaving phonon-induced diamond, expected, the dia- be bulk can the in noise from Deep field emanating at- noise is surface. field relaxation mond this electric nanodiamond contexts to in both tributed NVs In and respectively. [16] [17] NVs shallow in observed n htpoo-nue eaaino the on relaxation phonon-induced that ing lattice. diamond the in ties | ae fu oapoiaey2 approximately to up of rates | frlxto nthe on relaxation of mmaheal Veetoi pnchrnetm of of time superposition coherence a for spin ms electronic 6.8(2) NV max- the achievable on limit imum ultimate an setting conditions, ambient huh osgicnl otiuet eaaino the on relaxation is to which | contribute relaxation, significantly theo- spin-lattice to standard NV thought processes the explain Raman that to two-acoustic-phonon invoked show of we model addition, qubit retical simplified In the than full rather dia- model. qutrit, the bulk NV considering the high-purity of of importance picture in the K) discuss We (295 mond. temperature room at o the for rniina omtmeaue uuesuiso re- of studies Future temperature. room at transition e nteaoi ipaeet r oiatsuc of the source on dominant a relaxation are spin-phonon displacements atomic or- the second in to der Hamiltonian that spin-phonon a suggests involving processes or This processes Orbach-like symmetry. two-quasilocalized-phonon time-reversal of quence m m m m edsilnro,UiesddMyr atao Chile Santiago, Mayor, Universidad terdisciplinarios, nti etrw rsn xeietlrslsindicat- results experimental present we Letter this In s s s s | = +1 = +1 = 0 = m nfo uslclzdpooso contri- or phonons quasilocalized from on − s n httemxmmtheoretically maximum the that find e | ↔ i | = )m.Fnly epeetatheoretical a present we Finally, ms. 2) 1 m 2 ie ail 0,Snig,Chile Santiago, 306, Casilla hile, | ↔ i i i na r h oiatmechanism dominant the are onian .R Maze, R. J. s − rniina ela h oeo te impuri- other of role the as well as transition ytm eso htrelaxation that show We system. rniinocr truhytieterate the twice roughly at occurs transition o vrfu reso magnitude, of orders four over ion = e-aac N)cne ndia- in center (NV) gen-vacancy 1 m | ↔ i − s m 1 = s | ↔ i sosn576 USA 53706, isconsin ± +1 = m | 1 m s vanzados, i ,4 3, +1 = s m rniin rdcsart fzero of rate a predicts transition, nvcnycenters en-vacancy i 0 = s n .Kolkowitz S. and rniin eetwr with work Recent transition. +1 = i | ↔ i rniin ihrelaxation with transition, uiybl di- bulk purity × | m | i m 10 m s rniina conse- a as transition s 3 s = nthe on n 0 = s = − − 1 ± i 1 ture n 2 and | ↔ i 1 and i | | 1, m m rniinin transition † s s | m m × = = s s 10 − − = +1 = 5 1 1 ↔ i ↔ i ± s − 1 1 i i 2 laxation on the ms = 1 ms = +1 transition as a curves isolates the dynamics of the single NV orientation function of temperature| − couldi ↔ resolve| thei relative contri- resonantly addressed by the microwaves from among the butions of these different mechanisms. four orientations allowed by the diamond’s tetrahedral As a matter of terminology, we note that the exist- lattice. ing literature refers to the same transitions with various Relaxation measurements were performed on native names. The ms = 1 ms = +1 and ms =0 NVs in three chemical-vapor-deposition (CVD) grown | − i ↔ | i | i ↔ ms = 1 transitions have been referred to as double- bulk diamond samples from different growers with dif- | ± i quantum and single-quantum transitions respectively, ferent native NV concentrations. Confocal microscope but for mixed eigenstates these terms may be mislead- images of the three samples can be seen in Fig. 2(a-c). ing. The names of the transitions were generalized to The NV concentrations of samples A and B were esti- account for this in Ref. [17], and we will adopt that nam- mated by NV counting. The NV concentration of sample ing scheme here. Specifically, with H; i denoting the C was estimated by comparing the typical fluorescence | i energy eigenstate with majority component ms = i , we of the sample to the fluorescence of the isolated single | i call the H; 1 H;+1 and H;0 H; 1 tran- NVs visible in sample A. Sample A, from Element Six, | − i ↔ | i | i ↔ | ± i sitions the qutrit and qubit transitions respectively. In has the lowest NV concentration at approximately 10−5 referring to the eigenstates going forward, we will omit ppb. Sample B, from Diamond Elements, has an NV the H identifier from kets, referring to H; i simply as i . | i | i γ Δ± METHODS

Experiments were conducted under ambient conditions ΩΩ using a homebuilt confocal microscope with a temperature stable to 295 K within 1 K. Applied magnetic fields were typically < 100 G,± with few measurements conducted at higher fields. (Table S1 of the Supplemen- tal Material [18] contains a complete summary of the data collected for this paper, including applied magnetic fields.) The components of the applied magnetic field are defined as on- and off-axis with respect to the NV’s spatial axis of symmetry. Spin polarization and readout were accomplished with approximately 1 mW of 532-nm light. To extract the qubit and qutrit relaxation rates, defined as Ω and γ respectively as in Fig. 1(a), we use state- selective π-pulses to measure the population decay into and out of each spin state individually, as described in [16, 17]. Under a classical three-level population model, subtraction of these curves yields single-exponential de- cays from which we isolate the relaxation rates for each FIG. 1. State-dependent spin relaxation rates of nitrogen- transition. In particular, with P (τ) denoting the pop- i,j vacancy centers. (a) Ground state level-structure of the NV− ulation in j at time τ after initialization in i , we de- | i | i electronic spin. Qubit transitions between |0i and |±1i occur termine Ω and γ using the differences defined at rate Ω, while qutrit transitions between |+1i and |−1i occur at rate γ. An applied magnetic field lifts the |±1i degeneracy by the splitting ∆±. (b) Measurement of state- −3Ωτ FΩ(τ)= P0,0(τ) P0,+1(τ)= ae , (1) dependent decay curves. Relaxation out of |0i (cyan squares) − is described by a single exponential with decay rate 3Ω (cyan −(2γ+Ω)τ curve). Relaxation out of |+1i (yellow diamonds) is described Fγ (τ)= P , (τ) P , (τ)= ae , (2) +1 +1 − +1 −1 by a biexponential decay that depends on both Ω and γ (yellow curve). Solid lines are fits to population dynamics equa- where a is the fluorescence contrast between the two sub- tions of the ground state [17]. (The deviation of the yellow tracted data sets. Fig. 1(c) shows a representative result curve from 1 at τ = 0 is due to π-pulse infidelity.) For data −1 −1 of the single-exponential decay curves FΩ and Fγ . In shown, Ω = 56(3) s , γ = 132(11) s . (c) Semilog plot of ensemble measurements, where multiple NVs with differ- subtracted population curves, FΩ and Fγ in Eqs. 1 and 2, ent orientations contribute to the measured signal, the used to extract Ω and γ respectively. Displayed data is from applied magnetic field and the subtraction of population the same measurement series as that shown in (b). 3 concentration of approximately 10−3 ppb. The NV con- the result of interactions between defects. In the small centrations of samples A and B are low enough to allow off-axis magnetic field regime, we find that relaxation on for measurements to be performed on single NVs using the qutrit transition is approximately twice as fast as re- confocal microscopy. The NV concentration of Sample C, laxation on the qubit transition. Specifically, we measure from Chenguang Machinery & Electric Equipment Com- the weighted averages γ = 117(5) s−1 andΩ = 59(2) s−1. pany, is high enough to prohibit measurement on single Importantly, this indicates that significant dynamics are NVs in our confocal microscope, at approximately 0.1 missed if the NV is treated as a qubit by considering only ppb. For this sample we performed measurements on one of the 0 and 1 subspaces; population leakage to the ensemble of NVs within the confocal volume of the the third state| i actually|± i occurs faster than relaxation be- microscope. In the data, NVs are indexed first by sam- tween the states intended to comprise the qubit. ple and then numerically to distinguish single NVs where The standard T1 relaxation experiment that is most appropriate. Using the ratio of approximately 300 sub- commonly used with NVs consists of measuring the life- − stitutional nitrogen defects per NV determined in [19] time of the 0 state. The T1 times quoted in the lit- for native defects in CVD-grown (100)-oriented samples, erature based| i on these measurements can be related to (0) we estimate the substitutional nitrogen concentrations to our values of Ω by defining T1 = 1/3Ω [16]. As two be approximately 3 10−3 ppb, 3 10−1 ppb, and 3 10 (0) × × × examples, the T1 times reported from both Naydenov ppb in samples A, B, and C respectively. et al. [10] and Bar-Gill et al. [11] in isotopically pure diamond samples at room temperature corresponds to Ω = 56(4) s−1, in agreement with our measurements. To our RESULTS knowledge, the qutrit relaxation rate γ has not been sys- tematically studied for NVs deep within bulk diamond The measured relaxation rates γ and Ω under small samples. off-axis magnetic fields (quantitatively, B⊥ < 1 G) for We next examine the limit to coherence time T2 im- native NVs deep in the bulk in the three samples are posed by the qubit and qutrit relaxation rates. We intro- shown in Fig. 2(d). We find that neither γ nor Ω de- duce, T2,max, the maximum theoretically achievable co- pends on the NV concentration over a range covering herence time due to incoherent relaxation between states. −5 −1 four orders of magnitude from approximately 10 –10 For a qubit basis formed of 0 and either 1 [16, 17], ppb. This indicates that the relaxation we observe is not | i |± i 2 T2,max = . (3) (a) (b) (c) 3Ω+ γ 2 µm 2 µm 2 µm For the values of Ω and γ measured at small off-axis magnetic fields, we find T2,max = 6.8(2) ms. This value rep- resents the maximum achievable coherence time at room temperature for NVs deep in high-purity bulk diamond Sample A Sample B Sample C samples if all pure dephasing sources are eliminated. As (d) a caveat, we note that T2,max could in principle be ex- γ tended by engineering the vibrational modes of the lattice ) 120 1 Ω

− to suppress relaxation due to spin-lattice interactions. As an example of the implications of this limit, T2,max 100 2.2 determines the minimum theoretically achievable sensi- 2.0 tivity of a room temperature single NV electronic spin to 80 Ω γ/ 1.8 magnetic ﬁelds. If we assume 100% charge state initial-

Relaxation rate (s ization and spin state polarization as well as zero dead 60 NVA3 NVB1 NVC time between measurements, the ultimate limit is given NVA3 NVB1 NVC by [20, 21]:

FIG. 2. Rates γ and Ω for samples with differing defect con- ~ δB = (4) centration. (a-c) Confocal microscope images of samples A, min gµBCpT2,maxT B, and C, displaying NV concentrations ranging from approx- −5 −1 imately 10 ppb to 10 ppb. (d) Average measured rates where T is the total averaging time, ~ is the reduced γ and Ω at small off-axis magnetic fields in samples A, B, Planck constant, g 2 is the electronic Landé g-factor, and C. On-axis magnetic fields are in the range 20 − 60 G. µ is the Bohr magneton,≈ and C 1 describes the mea- Measurements of both γ and Ω are equivalent to within error B surement efficiency. In our experiment≤ C 0.02. Using in all three samples, indicating no dependence on defect con- ≈ centration in the measured regime. Inset: The ratio γ/Ω ≈ 2 the value of the qubit T2,max determined by our measure- for each pair of points shown in the main plot. Error bars are ment of γ and Ω, we calculate δBmin = 69(1) pT/√Hz for one standard error. the idealized case C = 1. As a practical example, this 4 corresponds to the ability to detect the magnetic field NVA3 NVB1 NVC from a single nuclear spin 30 nm away from the NV after (a) 1 s of measurement time. 70 )

There has been considerable interest in using coherent 1 − 60 superpositions of the +1 and 1 states for magnetic (s

| i |− i Ω field sensing. This “double-quantum” basis eliminates 50 noise from temperature drift and fluctuations of the elec- (b) tric and scaled-strain field along the NV axis while also 250 doubling the frequency shift due to the magnetic field )

1 200 along the NV axis [22–25]. However, a sensor in this ba- − (s sis is also more susceptible to the faster qutrit decay γ. γ 150 We next investigate the sensitivity of an ideal NV sensor making use of this basis, where the maximum theoreti- 100 0 10 20 30 40 50 60 cally achievable coherence time is [16] B k (G) (c) 1 70 T = . (5) 2,max ) Ω+ γ 1 − 60 (s Ω Using the same values of Ω and γ as before, we ﬁnd 50 the maximum theoretically achievable coherence time for (d) 250 the double-quantum basis is T2,max = 5.7(2) ms, which is approximately 1 ms shorter than that in the “single- ) 1 200 −

quantum” basis. This corresponds to a magnetic ﬁeld (s

γ 150 sensitivity of δBmin = 38(1) pT/√Hz, indicating that despite the shorter coherence time, an overall sensitivity 100 advantage is conferred due to the factor of two increase 0 10 20 30 40 50 60 B in signal. We note that if we set γ = 2Ω in accordance ⊥ (G) with our measurements for small off-axis magnetic fields, we find T2,max =2/(5Ω) in the single-quantum basis and FIG. 3. Measurements of Ω and γ as functions of Bk and T2,max =1/(3Ω) in the double-quantum basis. B⊥, the on-axis and off-axis components of the applied magnetic field respectively. To allow for visual separation between Figure 3 shows measurements of Ω and γ for vary- the points in the figure, only measurements made at applied ing applied magnetic field magnitudes and orientations magnetic field magnitudes under 65 G are shown. (See [18] plotted as functions of Bk and B⊥, the on-axis and off- for an extended version of this figure and a table of complete axis components of the magnetic field respectively. Lin- measurements.) Green lines are weighted linear regressions ear regressions are shown to demonstrate correlations be- intended to show correlations. The regressions account for all tween the rates and the field components. As functions data, including that not shown in the plots. Shaded regions −1 demonstrate one standard error. The regressions evidence of Bk, Ω and γ display slopes of 0.01(2) s /G and −1 − a small correlation of γ with B⊥; the other slopes are zero 0.02(5) s /G respectively. As functions of B⊥, Ω and −1 −1 within two standard errors. Note that the same values of Ω γ display slopes of 0.07(5) s /G and 0.81(14) s /G re- are plotted in (a) and (c). Likewise, the same values of γ are spectively. Interestingly, the slope of γ as a function of plotted in (b) and (d). B⊥ (Fig. 3(d)) indicates there is a significant but small correlation between the two variables, with γ accruing an approximate 1% increase per G of applied field. The DISCUSSION analysis of directional correlations shown here provides additional information which is not typically captured in We now turn to a discussion of the origins of relaxation studies of phonon-limited relaxation that focus on tem- on the qutrit transition. In high-purity bulk diamond at perature dependence. Consideration of such properties room temperature, NV spin relaxation is attributed to may help uncover and explain rich spin-lattice dynamics spin-lattice interactions [12–15]. The analysis by Jarmola in NVs. In the discussion of maximum coherence times et al. [12] found that a sum of the form above and for the remainder of this work we focus on measurements conducted under small off-axis magnetic 1 A2 5 (0) = A1(S)+ + A3T (6) fields because superior phonon-limited coherence times exp(∆/kBT ) 1 T1 are offered in this limit and because NV applications are − commonly conducted with magnetic fields applied along provides a good fit to experimental measurements of the the NV axis. rate of relaxation out of 0 as a function of tempera- | i 5 ture in a range of samples. The fit parameter A1(S) is a (a) (b) 3 −1 sample-dependent constant, while A2 =2.1(6) 10 s −11 −1 × and A3 =2.2(5) 10 s are coefficients weighting the relative contributions× of the exponential and T 5 terms. The parameter ∆ = 73(4) meV is an empirical activa- Energy tion energy. Only the latter two terms in Eq. 6 provide significant contributions at temperatures T & 50 K, in agreement with the earlier analysis of Ref. [14] and the Energy work of Ref. [13], which considers a wider variety of low- temperature relaxation phenomena. While past results uniformly emphasize the significance of the exponential FIG. 4. Two examples of two-phonon processes driving the 5 and T terms at high temperatures, claims regarding qubit transition. In both processes, a phonon of energy ~ωem which of these terms is dominant at room temperature is emitted and a phonon of energy ~ωabs is absorbed. Im- have been mixed. Using the numbers from Ref. [12], we portantly, the emission-first process (a) is associated with a calculate that the exponential term contributes approx- different suppressive factor than the absorption-first process imately 120 s−1 and the T 5 term contributes approxi- (b) in perturbation theory. mately 50 s−1 to 1/T (0) 170 s−1 at T = 295 K. 1 ≈ The mechanism behind the exponential term has been the observation that γ > Ω, this suggests that contribu- identified as a two-quasilocalized-phonon Orbach-like tions from quasilocalized phonons may also be dominant process. The empirical activation energy associated with for relaxation on the qubit transitions. While the sim- this term is roughly consistent with ab initio calculations plicity of the ratio γ/Ω 2 is suggestive, we are unable that have revealed a band of quasilocalized phonon modes to say conclusively whether≈ this is a coincidence arising centered at 65 meV [26–28]. The T 5 term is attributed to from various contributions to the two rates, or whether it a two-acoustic-phonon Raman process. The theoretical is because a single mechanism dominates both the qutrit basis for this identification comes from Ref. [29], in which and qubit relaxation rates and naturally gives rise to the Walker demonstrates a two-acoustic-phonon Raman pro- ratio between the rates. cess leading to transition rates with a thermal scaling of T 5. Other possible contributions from two-phonon pro- The Supplemental Material [18] contains derivations cesses arise due to the action of a spin-lattice Hamiltonian of the two-acoustic-phonon Raman process in Ref. [29] taken to second order in the atomic displacements and applied to first order in perturbation theory [30]. These and a phenomenological model of a two-quasilocalized- phonon Orbach-like process. Both derivations are based contributions are not typically considered for the NV; on a spin-lattice Hamiltonian taken to first order in where they have been considered, symmetry arguments such as the one discussed here have not been applied the atomic displacements and applied to second-order in time-dependent perturbation theory. The processes [13]. We include a brief discussion of these contributions may proceed by either absorption-followed-by-emission in the Supplemental Material [18]. We leave for future or emission-followed-by-absorption, as shown in Fig. 4. work a complete investigation of the effect of the second- Due to the time-reversal symmetry of the spin-lattice order terms in the spin-lattice Hamiltonian in light of symmetry considerations. Ultimately, the difference in Hamiltonian, a generic two-phonon Raman process to first-order in the atomic displacements and second order expected temperature scalings of the relaxation rates in- in perturbation theory only provides a nonzero contribu- duced by the various processes provides a way to poten- tially disentangle their contributions, motivating future tion to γ by the inequality of these two variants of the measurements of γ, and the ratio γ/Ω, as a function of process. Because this restriction is inconsistent with the Ab initio assumptions in Ref. [29] leading to a T 5 thermal scaling temperature. calculations of the coupling con- of a two-acoustic-phonon Raman process (which has pre- stants describing the spin-lattice Hamiltonian may also 5 be critical in explaining the ratio γ/Ω 2 [31]. viously been invoked to explain the T term in Eq. 6), we ≈ conclude that this theoretical model cannot explain our observation of significant relaxation on the qutrit transition. CONCLUSIONS In contrast, a phenomenological model of the effect of two quasilocalized phonons results in an Orbach-like We have presented observations of spin state- process (the exponential term in Eq. 6) that contributes dependent relaxation rates under ambient conditions in nonzero γ and Ω relaxation rates. These observations high-purity bulk diamond samples. We have measured indicate that two-quasilocalized-phonon Orbach-like pro- the rate of phonon-limited relaxation on the qutrit transi- cesses may be the dominant mechanism for relaxation on tion, γ, at room temperature. Our measurements provide the qutrit transition at room temperature. Along with a complete experimental picture of the three-level relax- 6 ation dynamics in high-purity diamond at room temper- D. Budker, N. Y. Yao, et al., Imaging the local charge en- ature, indicating T2,max = 6.8(2) ms as the limit on the vironment of nitrogen-vacancy centers in diamond, Phys. maximum theoretically achievable coherence time of an Rev. Lett. 121, 6 (2018). NV electronic spin at 295 K. Furthermore, we have shown [6] P. Ovartchaiyapong, K. W. Lee, B. A. Myers, and A. C. B. Jayich, Dynamic strain-mediated coupling of that the standard model of two-acoustic-phonon Raman a single diamond spin to a mechanical resonator, Nat. processes predicts no relaxation on the qutrit transition, Commun. 5, 4429 (2014). motivating measurement of the temperature dependence [7] J. Teissier, A. Barfuss, P. Appel, E. Neu, and of γ and providing evidence that quasilocalized phonons P. Maletinsky, Strain coupling of a nitrogen-vacancy or contributions from higher order terms in the spin- center spin to a diamond mechanical oscillator, phonon Hamiltonian provide dominant contributions to Phys. Rev. Lett. 113, 020503 (2014). spin-lattice relaxation on the qutrit transition at room [8] G. Kucsko, P. C. Maurer, N. Y. Yao, M. Kubo, H. J. Noh, P. K. Lo, H. Park, and M. D. Lukin, Nanometre- temperature. scale thermometry in a living cell, Nature (London) 500, 54 (2013). [9] D. M. Toyli, C. F. de las Casas, D. J. Christle, V. V. ACKNOWLEDGMENTS Dobrovitski, and D. D. Awschalom, Fluorescence thermometry enhanced by the quantum coherence of single spins in diamond, P. Natl. Acad. Sci. USA 110, 8417 The authors thank Jeff Thompson, Nathalie de Leon, (2013). Gregory Fuchs, Dolev Bluvstein, Nir Bar-Gill, and Nor- [10] B. Naydenov, F. Dolde, L. T. Hall, C. Shin, H. Fed- man Yao for enlightening discussions, helpful insights, der, L. C. L. Hollenberg, F. Jelezko, and J. Wrachtrup, and comments on the manuscript. Experimental work, Dynamical decoupling of a single-electron spin at room data analysis, and theoretical efforts conducted at UW– temperature, Phys. Rev. B 83, 081201 (2011). Madison were supported by the U.S. Department of En- [11] N. Bar-Gill, L. Pham, A. Jarmola, D. Budker, and ergy (DOE), Office of Science, Basic Energy Sciences R. Walsworth, Solid-state electronic spin coherence time approaching one second, Nat. Commun. 4, 1743 (2013). (BES) under Award #de-sc0020313. Theoretical contri- [12] A. Jarmola, V. M. Acosta, K. Jensen, S. Chemerisov, and butions conducted at Pontificia Universidad Católica de D. Budker, Temperature- and magnetic-field-dependent Chile by J. R. M. and A. N. were supported by ANID longitudinal spin relaxation in nitrogen-vacancy ensem- Fondecyt 1180673 and ANID PIA ACT192023. A. N. ac- bles in diamond, Phys. Rev. Lett. 108, 197601 (2012). knowledges financial support from Universidad Mayor [13] A. Norambuena, E. Muñoz, H. T. Dinani, A. Jar- through the Postdoctoral Fellowship. A. G. acknowl- mola, P. Maletinsky, D. Budker, and J. R. Maze, edges support from the Department of Defense through Spin-lattice relaxation of individual solid-state spins, Phys. Rev. B 97, 094304 (2018). the National Defense Science and Engineering Graduate [14] D. A. Redman, S. Brown, R. H. Sands, and S. C. Fellowship (NDSEG) program. Rand, Spin dynamics and electronic states of N-V centers in diamond by epr and four-wave-mixing spectroscopy, Phys. Rev. Lett. 67, 3420 (1991). [15] S. Takahashi, R. Hanson, J. van Tol, M. S. Sher- win, and D. D. Awschalom, Quenching spin deco- ∗ These authors contributed equally. herence in diamond through spin bath polarization, † [email protected] Phys. Rev. Lett. 101, 047601 (2008). [1] P. Neumann, R. Kolesov, B. Naydenov, J. Beck, [16] B. A. Myers, A. Ariyaratne, and A. C. B. Jayich, F. Rempp, M. Steiner, V. Jacques, G. Balasubramanian, Double-quantum spin-relaxation limits to coher- M. L. Markham, D. J. Twitchen, S. Pezzagna, J. Mei- ence of near-surface nitrogen-vacancy centers, jer, J. Twamley, F. Jelezko, J. Wrachtrup, et al., Quan- Phys. Rev. Lett. 118, 197201 (2017). tum register based on coupled electron spins in a room- [17] A. Gardill, M. Cambria, and S. Kolkowitz, Fast relax- temperature solid, Nat. Phys. 6, 249 (2010). ation on qutrit transitions of nitrogen-vacancy centers in [2] F. Dolde, I. Jakobi, B. Naydenov, N. Zhao, S. Pezzagna, nanodiamonds, Phys. Rev. Appl. 13, 034010 (2020). C. Trautmann, J. Meijer, P. Neumann, F. Jelezko, and [18] See Supplemental Material for experimental details and J. Wrachtrup, Room-temperature entanglement between additional theoretical details. single defect spins in diamond, Nat. Phys. 9 (2013). [19] A. M. Edmonds, U. F. S. D’Haenens-Johansson, R. J. [3] L. Rondin, J.-P. Tetienne, T. Hingant, J.-F. Roch, Cruddace, M. E. Newton, K.-M. C. Fu, C. Santori, R. G. P. Maletinsky, V. Jacques, et al., Magnetometry with Beausoleil, D. J. Twitchen, and M. L. Markham, Pro- nitrogen-vacancy defects in diamond, Rep. Prog. Phys. duction of oriented nitrogen-vacancy color centers in syn- 77, 056503 (2014). thetic diamond, Phys. Rev. B 86, 035201 (2012). [4] F. Dolde, H. Fedder, M. W. Doherty, T. Nöbauer, [20] J. R. Maze, P. L. Stanwix, J. S. Hodges, S. Hong, J. M. F. Rempp, G. Balasubramanian, T. Wolf, F. Reinhard, Taylor, P. Cappellaro, L. Jiang, M. V. G. Dutt, E. To- L. C. L. Hollenberg, F. Jelezko, J. Wrachtrup, et al., gan, A. S. Zibrov, A. Yacoby, R. L. Walsworth, M. D. Electric-field sensing using single diamond spins, Nat. Lukin, et al., Nanoscale magnetic sensing with an indi- Phys. 7, 459 (2011). vidual electronic spin in diamond, Nature (London) 455, [5] T. Mittiga, S. Hsieh, C. Zu, B. Kobrin, F. Machado, 644 (2008). P. Bhattacharyya, N. Z. Rui, A. Jarmola, S. Choi, 7

[21] J. Taylor, P. Cappellaro, L. Childress, L. Jiang, D. Bud- [26] J. Zhang, C.-Z. Wang, Z. Zhu, and V. Dobrovitski, Vi- ker, P. Hemmer, A. Yacoby, R. Walsworth, and M. Lukin, brational modes and lattice distortion of a nitrogen- High-sensitivity diamond magnetometer with nanoscale vacancy center in diamond from first-principles calcula- resolution, Nature Physics 4, 810 (2008). tions, Physical Review B 84, 035211 (2011). [22] H. J. Mamin, M. H. Sherwood, M. Kim, C. T. [27] A. Gali, T. Simon, and J. Lowther, An ab initio study of Rettner, K. Ohno, D. D. Awschalom, and D. Ru- local vibration modes of the nitrogen-vacancy center in gar, Multipulse double-quantum magnetome- diamond, New Journal of Physics 13, 025016 (2011). try with near-surface nitrogen-vacancy centers, [28] A. Alkauskas, B. B. Buckley, D. D. Awschalom, and C. G. Phys. Rev. Lett. 113, 030803 (2014). Van de Walle, First-principles theory of the luminescence [23] M. Kim, H. J. Mamin, M. H. Sherwood, K. Ohno, D. D. lineshape for the triplet transition in diamond NV cen- Awschalom, and D. Rugar, Decoherence of near-surface tres, New Journal of Physics 16, 073026 (2014). nitrogen-vacancy centers due to electric field noise, Phys. [29] M. Walker, A T 5 spin–lattice relaxation rate for non- Rev. Lett. 115, 087602 (2015). Kramers ions, Canadian Journal of Physics 46, 1347 [24] E. Bauch, C. A. Hart, J. M. Schloss, M. J. Turner, J. F. (1968). Barry, P. Kehayias, S. Singh, and R. L. Walsworth, Ul- [30] J. Van Kranendonk, Theory of quadrupolar nuclear spin- tralong dephasing times in solid-state spin ensembles via lattice relaxation, Physica 20, 781 (1954). quantum control, Phys. Rev. X 8, 031025 (2018). [31] J. Gugler, T. Astner, A. Angerer, J. Schmiedmayer, [25] D. Bluvstein, Z. Zhang, C. A. McLellan, N. R. J. Majer, and P. Mohn, Ab initio calculation of the spin Williams, and A. C. B. Jayich, Extending the quan- lattice relaxation time T 1 for nitrogen-vacancy centers in tum coherence of a near-surface qubit by coher- diamond, Physical Review B 98, 214442 (2018). ently driving the paramagnetic surface environment, Phys. Rev. Lett. 123, 146804 (2019). arXiv:2007.11529v1 [cond-mat.mes-hall] 22 Jul 2020 abnao,wihcmrssteltie sdenoted is lattice, the comprises which atom, carbon T edis field Φ( where and rpriso cutcpoosaddmntae h eea m general the demonstrates perf m and the vibrational phonons of bulk acoustic Examination of of perturbat section. band properties subsequent considered the a in be in fall may phenomenologically frequencies modes their modes quasilocalized since quasilocalized the crystal describe to lattice, own the its in on insufficient is Hamiltonian this steN lcrncsi yoantcrto h pn1oeaosa operators spin-1 The ratio. gyromagnetic spin electronic NV the is bevdtmeauedpnec ftelftm of lifetime provide the to of shown dependence been temperature together have observed which phonons, quasilocalized where hl w-uslclzdpoo rahlk rcse a ngener rates. in can pr processes Orbach-like Raman two-quasilocalized-phonon two-acoustic-phonon while that symmetry time-reversal to oenn h VadteHmloingvrigtelattice. eigenbasis the The governing Hamiltonian perturbation. the a and as NV lattice the diamond governing the and center NV 5 h atc aitna o efc atc ie atc ihu n d any without lattice a (i.e. lattice perfect a for Hamiltonian lattice The eew eieepesosfrterlxto ae u otwo-ph to due rates relaxation the for expressions derive we Here h aclto rcesb plcto fFrisgle uet secon to rule golden Fermi’s of application by proceeds calculation The ecnie nN rudsaetiltHmloino h form the of Hamiltonian triplet ground-state NV an consider We ∆ em nE.6o h antx.Tecluaincniesprocesses considers calculation The text. main the of 6 Eq. in terms upeetlMtrasfr“tt-eedn phonon-limi “State-dependent for Materials Supplemental u 2 ii etod netgcńDiALb autdd suisIn Estudios de Facultad Lab, Investigación DAiTA de Centro iα ′ h ,where ), i B h neato ewe tm sdsrbdt eododri the in order second to described is atoms between interaction The . 2 = nee h tm opiigteltie and lattice, the comprising atoms the indexes . .C Cambria, C. M. π ~ stePac constant, Planck the is 3 ∆ nttt eFıia otfii nvria a´lc eC Católica de Universidad F´ısica, Pontificia de Instituto ii 1 ′ eateto hsc,Uiest fWsosn aio,W Madison, Wisconsin, of University Physics, of Department = R EIAINO HNNIDCDRLXTO RATES RELAXATION PHONON-INDUCED OF DERIVATION i 4 1, − etod netgcńe aoenlgı aeilsA Materiales Nanotecnolog´ıa y Investigación en de Centro ∗ R .Gardill, A. i otfii nvria a´lc eCie atao Chile Santiago, Chile, Católica de Universidad Pontificia ′ stedsac ewe h qiiru oiin fthe of positions equilibrium the between distance the is H l D = gs 1, irgnvcnycenters” nitrogen-vacancy 2 ∗ 2 = 1 m H .Li, Y. NV X . iα 7Gzi h rudsaezr-edsltig and splitting, zero-field ground-state the is GHz 87 atc Hamiltonian Lattice /h VHamiltonian NV p 1 iα 2 = .Norambuena, A. m | + 0 D i h tmcmmnaadeulbimdslcmnsare displacements equilibrium and momenta atomic The . gs 2 1 α at S ii pn h atsa opnns( components Cartesian the spans X z 2 ′ αα T + ′ & edsilnro,UiesddMyr atao Chile Santiago, Mayor, Universidad terdisciplinarios, gµ Φ( cse ontcnrbt to contribute not do ocesses to ywihltievbain a equantized. be can vibrations lattice which by ethod 0K[,2.I diin edmntaeta due that demonstrate we addition, In 2]. [1, K 50 B ∆ lcnrbt obt h urtadqbtrelaxation qubit and qutrit the both to contribute al nnpoesscnitn ihteepnniland exponential the with consistent processes onon B ii ds[–] uslclzdmdswl etreated be will modes Quasilocalized [3–5]. odes 2 ′ c atc aitna eel oeimportant some reveals Hamiltonian lattice ect · infiatcnrbtost h experimentally the to contributions significant ie ail 0,Snig,Chile Santiago, 306, Casilla hile, ) edenoted re oso xsigmdswti h uko the of bulk the within modes existing of ions ssmlaeul ignli h Hamiltonian the in diagonal simultaneously is S .R Maze, R. J. u , re,tetn h neato ewe the between interaction the treating order, d iα novn cutcpoosi h ukand bulk the in phonons acoustic involving hc rs u otepeec fdefects of presence the to due arise which , u fcs is efects) i ′ α sosn576 USA 53706, isconsin ′ ipaeet ihiorpccoupling isotropic with displacements , S vanzados, ,4 3, e pnrlxto of relaxation spin ted ( = n .Kolkowitz S. and S i x hand th S , γ y ,y z y, x, S , ntesadr theory standard the in z i n h magnetic the and ) ′ gµ haos Though atoms. th .Tems fa of mass The ). B 2 = 1, † . MHz/G 8 p (2) (1) iα 2

We diagonalize the lattice Hamiltonian with discrete Fourier transforms over the N atoms that comprise the lattice: 1 piα = pkekα exp(ik Ri), (3) √ · N k X 1 uiα = ukαekα exp(ik Ri), (4) √ · N k X where k indexes the vibrational mode with wavevector k and polarization unit vector ek. The αth component of ek is denoted ekα. We may then abbreviate the wavevector space momentum and position as

pkα = pkekα, (5)

ukα = ukekα. (6)

With k indicating a vibrational mode with the same polarization but opposite wavevector as k, the transformation of Eqs. 3− and 4 results in the Hamiltonian

1 mω2 H = p p + k u u , (7) l 2m kα −kα 2 kα −kα kα X where ~ωk describes the energy associated with one quantum of the vibrational mode k. We can rewrite the Hamil- tonian using the creation and annihilation operators b† and b, which obey

~ mωk † pkα = i b bk ekα, (8) 2 −k − r ~ † ukα = b−k + bk ekα. (9) 2mωk r Using Eqs. 8 and 9, we obtain

~ † 1 Hl = ωk bkbk + 2 , (10) k X which gives the energy of the lattice in terms of the occupation numbers of the vibrational modes. The eigenstates of the Hamiltonian are Fock, or number, states. A single quantum of a mode can be treated as a bosonic quasiparticle, the phonon. In general, the quantization of lattice vibrations described by Eq. 10 can be used as the Hamiltonian for a more complicated lattice by extending the range of k to cover vibrational modes beyond those of a perfect lattice.

Interaction Hamiltonian

A phenomenological model of the spin-lattice interaction to second order in the atomic displacements uiα is described by:

′ ′ H = λmm′iα m m uiα + λmm′ii′αα′ m m uiαui′α′ (11) sl | i h | | i h | mm′iα ! mm′ii′ αα′ ! X X ′ where m and m are eigenstates of the spin Hamiltonian. The λmm′iα are coupling constants to ﬁrst order in the | i | i atomic displacements and the λmm′ii′αα′ are coupling constants to second order in the atomic displacements. To avoid an unnecessary proliferation of terms, we will consider only the ﬁrst-order contributions as we demonstrate how this interaction Hamiltonian may be expressed in terms of the phonon creation and annihilation operators. Excluding the second-order terms, Eq. 11 can also be written

Hsl = λβiαFβuiα, (12) βiα X 3

′ where the Fβ are linear combinations of the operators m m , β indexes the combinations, and the λβiα are coupling constants. For the NV ground-state triplet, consideration| i h | of time-reversal and spatial symmetries imposes the restrictions

β = (z, x′,y′,x,y) , (13)

2 2 2 (Fz, Fx′ , Fy′ , Fx, Fy) = (S , Sx,Sz , Sy,Sz , S S , Sx,Sy ), (14) z { } { } y − x { } which may be demonstrated by either linear algebra or group theory [6]. Working from Eq. 12, the Fourier transform given by Eq. 4 yields 1 H = λβiα exp(ik Ri)Fβukα. (15) sl √ · βikα N X Using Eq. 9 to expand in terms of the creation and annihilation operators, ~ k R † Hsl = λβiα exp(i i)Fβ b−k + bk ekα. (16) 2Nmωk · βikα r X We deﬁne ~ λβk = λβiαekα exp(ik Ri) (17) 2Nmω · iα r k X and obtain the spin-phonon Hamiltonian

† Hsl = λβkFβ b−k + bk . (18) βk X It can easily be shown that the spin-phonon Hamiltonian including second-order terms is analogously

† † † H = λβkFβ b + bk + λβkk′ Fβ b + bk b ′ + bk′ , (19) sl  −k   −k −k  βk βkk′ X X     where ~ ′ λβkk′ = λβii′ αα′ ekα exp(ik Ri)ek′α′ exp(ik Ri′ ) (20) 2Nm ωkωk′ · · ii′ αα′ √ X We note that if we assume the most signiﬁcant contributions to λβk come from atoms close to the defect, then for long wavelength phonons in the context of the spin-phonon Hamiltonian to ﬁrst order in the atomic displacements we can write ~ λβk λβiαekα (1 + ik Ri) . (21) ≈ 2Nmω · iα r k X Because the system is invariant under translation,

λβiα =0, (22) i X and Eq. 21 simpliﬁes to

~ωk λβk = λβiαekα ikˆ Ri , (23) 2Nmv2 · iα s s X where kˆ is the unit vector along k, vs is the speed of sound in diamond, and we have used the dispersion relation for acoustic phonons ωk = vs k . Eq. 23 displays the proportionality λβk √ωk, a property which will be invoked to obtain the T 5 scaling of two-acoustic-phonon| | Raman processes. ∝ 4

Evaluation of Fermi’s golden rule for Raman processes

From time-dependent perturbation theory, Fermi’s golden rule allows us to calculate Γi→f , the transition rate from an initial state i to a ﬁnal continuum state f . To second order, | i | i 2 2π VfmVmi Γi→f = ~ Vfi + δ(Ei Ef ), (24) Ei Em − m − X where Vαβ = α V β are the matrix elements of the perturbation, E α is the energy of α , and m is an intermediate state. h | | i | i | i Treating the spin-phonon Hamiltonian as a perturbation, the eigenstates of the system are described by the tensor product of the NV Hamiltonian eigenstates and the lattice Fock states. We assume the initial Fock state occupation numbers nk are described by the Bose-Einstein distribution

1 nk = (25) exp(~ωk/k T ) 1 B − for temperature T , where kB is the Boltzmann constant. We consider all ﬁnal states with the NV component of interest, regardless of the vibrational mode occupation numbers. That is, the phonon-induced NV transition rate is given by

′ ′ ′ ms,nk,nk′ ,... ′ Γms↔ms = Γms,nk,nk′ ,... (26) n′ ,n′ ,... k Xk′

′ ′ ′ ms,nk,nk′ ,... ′ ′ where Γms,nk,nk′ ,... is the transition between an initial state ms ..., nk, ..., nk , ... to a ﬁnal state ms ′ ′ | i ⊗ | i | i ⊗ ..., n , ..., n ′ , ... . | k k i Because the majority of phonons at room temperature have energies much greater than the splitting between the levels of the NV ground-state triplet and the NV has no low-lying excited states, relatively few phonons satisfy the energy requirements for direct two-phonon processes. Therefore, we consider only Raman processes, in which one phonon is absorbed and another is emitted. The NV transition rate due solely to Raman processes is

′ ms,nk−1,nk′ +1 ′ Γms↔ms = Γms,nk,nk′ . (27) kk′ X Evaluating Eq. 27 using the spin-phonon Hamiltonian of Eq. 19, we obtain a general expression for the transition rate associated with two-phonon Raman processes with intermediate states restricted to the multiplet containing the initial and ﬁnal states:

2 ′ ′′ ′ ′′ ′ ′′ ′′ ′ 2π Hmsms k Hms msk Hmsms kHms msk ′ ′ ′ ′ ′ ~ ~ ′ Γms↔ms = nk (nk + 1) Hmsmskk + + δ(∆msms + ωk ωk ) ~ ′′ ~ ′′ ~ ′ ′ ′′ ∆msms + ωk ∆msms ωk − kk ms X X − (28)

′ ′ ′ where the matrix element Hmsmskk and Hmsmsk are deﬁned

′ ′ ′ ′ Hmsm kk = ms λβkk Fβ m , (29) s h | | si β X

′ ′ Hmsm k = ms λβkFβ m , (30) s h | | si β X and

′ ′ ′ ∆msm = ms H ms m H m (31) s h | NV | i − h s| NV | si 5

′ is the energy diﬀerence between the NV states ms and m . Evaluating the squared magnitude in Eq. 28, | i | si 2π ′ ′ ′ ~ ~ ′ Γms m = nk (nk + 1) δ(∆msm + ωk ωk ) ↔ s ~ s − kk′ X

′ ′′ ′ ′′ ′ ′′ ′′ ′ 2 ∗ Hmsms k Hms msk Hmsms kHms msk ′ ′ ′ ′ Hmsmskk +2Re Hmsm kk +  s ′′ ~ ′′ ~ ′ ×  ′′ ∆msms + ωk ∆msms ωk   ms −   X   2  ′ ′′ ′ ′′ ′ ′′ ′′ ′ Hm m k Hm msk Hm m kHm msk + s s s + s s s . (32) ′′ ~ ′′ ~ ′  ′′ ∆msms + ωk ∆msms ωk ms X −   We see that two-phonon Raman processes arise in three ways according to the three major terms in the large paren- thesis in Eq. 28. The ﬁrst term,

2 ′ ′ Hmsmskk , corresponds to the second-order terms in the spin-phonon Hamilt onian taken to ﬁrst order in perturbation theory. The mixed second term,

′ ′′ ′ ′′ ′ ′′ ′′ ′ ∗ Hmsms k Hms msk Hmsms kHms msk ′ ′ 2 Re Hmsm kk + , s ′′ ~ ′′ ~ ′  ∆msm + ωk ∆msm ωk  m′′ s s  Xs −  results from the interference of the ﬁrst- and second-order terms in the spin-phonon Hamiltonian. The third term,

2 ′ ′′ ′ ′′ ′ ′′ ′′ ′ Hm m k Hm msk Hm m kHm msk s s s + s s s ′′ ~ ′′ ~ ′ ′′ ∆msms + ωk ∆msms ωk ms X −

corresponds to the ﬁrst-order terms in the spin-phonon Hamiltonian taken to second order in perturbation theory. For the remainder of this supplement, we explicitly consider only the third term, since this is the premise of both the exponential and T 5 terms in Eq. 6 of the main text. We will comment on the possible contributions of the other terms in a subsequent section. With this simpliﬁcation, we can write

2 ′ ′′ ′ ′′ ′ ′′ ′′ ′ 2π Hmsms k Hms msk Hmsms kHms msk ′ ′ ′ ~ ~ ′ Γms↔ms = nk (nk + 1) + δ(∆msms + ωk ωk ). (33) ~ ′′ ~ ′′ ~ ′ ′ ′′ ∆msms + ωk ∆msms ωk − kk ms X X −

This generic second-order two-phonon Raman process may be interpreted as follows. For each intermediate NV state ′′ ms , the terms in the sum in Eq. 33 represent an absorption-followed-by-emission variant (ﬁrst term in the sum, depicted in Fig. 4(a) of the main text) and an emission-followed-by-absorption variant (second term in the sum, depicted in Fig. 4(b) of the main text) of the two-phonon process. As we will see, it is important that these processes are not identical, but are associated with diﬀerent denominators and may interfere with one another. Eq. 33 will be the starting point for our analyses of transitions involving acoustic phonons and quasilocalized phonons, which will be treated separately.

Properties of the generic second-order two-phonon Raman process under time-reversal

Before considering the eﬀects of diﬀerent types of phonons, it is helpful to examine the properties of the generic two-phonon process under time-reversal. The restriction that the spin-lattice Hamiltonian is symmetric under time- reversal excludes odd powers of the spin operators from appearing in the Hamiltonian. The operator β λβkFβ is thus also symmetric under time reversal. With Θ denoting the time-reversal operator, the Sz eigenstates obey P ms Θ ms = ( 1) ms . (34) | i − |− i 6

′ Thus the matrix elements Hmsmsk follow

′ ′ ms+ms ′ ∗ ms λβkFβ m = ( 1) ms λβkFβ m (35) h | | si − h− | |− si β β X X ′ ms+ms ′ = ( 1) m λβkFβ ms , (36) − h− s| |− i β X where we have used the antiunitarity of time-reversal and taken the Hermitian conjugate. More concisely,

′ s ′ m +ms ′ Hmsm k = ( 1) H m msk. (37) s − − s− Inspecting Eq. 33, it may seem appropriate to make the approximation

2 2 ′ ′′ ′ ′′ ′ ′′ ′′ ′ Hmsms k Hms msk Hmsms kHms msk 1 ′ ′′ ′ ′′ ′ ′′ ′′ ′ + Hmsms k Hms msk Hmsms kHms msk (38) ′′ ~ ′′ ~ ′ ~ 2 ′′ ∆msms + ωk ∆msms ωk ≈ ( ωk) ′′ − ms − ms X X

where we have assumed the delta function and used the fact that the energy scale of the NV is small in comparison to ′ the energies of typical phonons at room temperature. In this approximation, Γms↔ms is only nonzero if the quantity

′ ′ ′ ′′ ′ ′′ ′ ′′ ′′ ′ Amsmskk = Hmsms k Hms msk Hmsms kHms msk (39) ′′ − ms X

′ ′ is nonzero in general. We will now show that Amsm kk is exactly zero for the transition ms ms by using s | i ↔ |− i the time-reversal symmetry described by Eq. 37. The demonstration will in fact apply broadly to ms ms ′′ ′ | i ↔ |− i transitions within a multiplet of an integer-spin system. The terms of the sum over ms in Ams−mskk may be divided into two simple cases. Case 1, m′′ =0: If m′′ = 0, then s 6 s 6 ′′ ′′ s s ′′ ′ ′′ −m +ms +ms +m ′′ ′ ′′ H msm k Hm msk = ( 1) H m msk H ms m k (40) − s s − − s − − s

′′ ′′ ′ = H−ms−ms kH−ms msk . (41)

′ ′′ The opposite of this term enters Ams−mskk via the intermediate ms . Therefore, for each term associated with a ′′ − ′′ particular ms , there is a cancellation that occurs with a term from ms and the total contribution from intermediates ′′ − ms = 0 is 0. 6 ′′ Case 2, ms = 0: This is just a special case of Eq. 41. We have

′ ′ H−ms0kH0msk = H−ms0k H0msk (42)

′′ for which the opposite term comes from the remaining part of the ms = 0 contribution. ′ Because Ams−mskk = 0, it is clear that the approximation of Eq. 38 is insufficient to explain relaxation on ms ms transitions. More specifically, nonzero relaxation rates for a two-phonon process on ms ms |transitionsi ↔ |− arei contingent on the difference between the denominators for the absorption-followed-by-emission| i ↔ |− andi ′ emission-followed-by-absorption variants of the process. The finding Ams−mskk = 0 can also be easily verified for the specific case of the NV’s qutrit transition by evaluating the matrix elements explicitly. It should be noted that the ′ same symmetry argument clearly does not hold if ms = ms, as in the case of the NV’s qubit transition. A similar symmetry argument to the one presented here has historically6 − been applied to half-integer-spin systems, where it is referred to as Van Vleck cancellation [7].

Two-acoustic-phonon Raman processes

The two-acoustic-phonon Raman process with T 5 thermal scaling presented by Walker in [8] depends on the approximation of Eq. 38, which we have just shown to be insuﬃcient to describe relaxation on the qutrit transition. For completeness, we will demonstrate the rest of the derivation here, showing how the temperature dependence of 7 the process arises. We restrict consideration to acoustic vibrational modes (denoted aco) and make the approximation of Eq. 38. This yields

2 (aco) 2π nk (nk + 1) ′ ′′ ′ ′′ ′ ′′ ′′ ′ ~ ~ ′ Γ ′ = Hm m k Hm msk Hm m kHm msk δ( ωk ωk ). (43) ms↔ms ~3 2 s s s s s s ′ ωk ′′ − − kk ∈ aco ms X X

From here, our derivation will diverge slightly from the original derivation in [8]. In particular, while the original derivation was made with arguments speciﬁc to a simple cubic lattice, we will generalize those arguments for an arbitrary lattice. Transforming from wavevector space back to position space and taking the long wavelength ′ approximation of Eq. 23 allows the matrix elements Hmsmsk to be expressed

~ωk ′ kˆ R ′ Hmsm k ekα i i Hmsm iα (44) s ≈ 2Nmv2 · s iα s s X

′ ′ where Hmsmsiα is the position space analog of Hmsmsk deﬁned

′ ′ Hmsm iα = ms λβiαFβ m . (45) s h | | si β X Substitution of Eq. 44 into Eq. 43 yields

(aco) π ′ ′ ′ Γ ′ = nk (nk + 1) Bmsm kk δ(ωk ωk ) (46) ms↔ms 2~2N 2m2v4 s − s kk′ X∈ aco with

′ ′ ′ ′ kˆ R kˆ′ R ′ ′ ′′ ′′ ′ ′ ′ ′′ ′ ′ ′′ Bmsmskk = ekαek α i i Hmsms iαHms msi α Hmsms i α Hms msiα . (47) ′ ′ ′′ · · − ii αα ms X

′ ′ ′ We observe that Bmsm skk depends only on the spatial characteristics of the vibrational modes k and k . In the ′ ′ ′ continuum limit, we may therefore reasonably replace Bmsmskk by a spatial average Bmsms deﬁned

1 ′ P n e P ′ n′ e n R n′ R ′ Bmsms = 2 ( γ( ) α) ( γ ( ) α) ( i) ( i ) 144π ′ ¨ ′ ′ ′′ · · · · γγ ii αα ms X X

2 ′ ′′ ′′ ′ ′ ′ ′′ ′ ′ ′′ ′ Hm m iαHm msi α Hm m i α Hm msiα dS dS (48) s s s − s s s

′ where the integrals each cover the unit sphere and γ and γ each index the one longitudinal and two transverse polarizations. The unit vector n is the normal associated with the surface element dS, and the function Pγ(n) returns the γth polarization unit vector associated with n. We take the continuum limit with the transformations

ωk ω, (49) → 1 nk n(ω)= , (50) → exp(~ω/k T ) 1 B −

′ ′ ′ Bmsm kk Bmsm . (51) s → s The sums over vibrational modes become integrals over frequencies with measure provided by the density of states of acoustic phonons. We assume the Debye model density of states given by

3V ω D(ω)= ω2 Π , (52) 2π2v3 ω s D 8 where ωD is the Debye frequency and V is the volume of the crystal. We have deﬁned the rectangle function

1, 0

2 (aco) 9V ω ′ 4 ′ ′ Γ ′ = Bmsm n(ω) (n(ω)+1) ω δ(ω ω ) Π dω dω . (54) ms↔ms 8π3~2N 2m2v10 s ¨ − ω s D Evaluating the integral over ω′,

ωD (aco) 9 ′ 4 Γ ′ = Bmsm n(ω) (n(ω)+1) ω dω (55) ms↔ms 3~2 2 2 10 s 8π d m vs ˆ0 where d = N/V is the atomic number density. With the change of variable ~ω x = , (56) kBT

~ωD xD = , (57) kBT 1 η(x)= , (58) exp(x) 1 − Eq. 55 can be rewritten

5 5 xD (aco) 9k T B ′ 4 Γ ′ = Bmsm η(x) (η(x)+1) x dx. (59) ms↔ms 3~7 2 2 10 s 8π n m vs ˆ0 Because the integral depends only weakly on temperature for T . 300 K, Eq. 59 exhibits a thermal scaling of 5 5 approximately T at room temperature. In the limit ωD , Eq. 59 exhibits an exact T scaling. We now enumerate three diﬀerent schemes by which acoustic→ ∞ phonons may contribute to two-phonon relaxation as alternatives to the derivation shown in this section. Alternative 1, restrict intermediate NV states to higher-lying states beyond the ground-state triplet: Because the splittings associated with these states are much greater than typical phonon energies at room temperature, we ′′ ~ ′′ 7 approximate ∆msm ωk for any ms, m , and ωk. This results in a scaling of T . s ≫ s Alternative 2, neglect the spatial characteristics of the acoustic phonon coupling constants λβk: If we do not make the approximation of Eq. 38 and instead neglect the spatial characteristics of the coupling constants λβk (which is ′ ′′ ′ ′′ ′ ′′ ′′ ′ necessary to make the calculation tractable in this case), then Hmsms k Hms msk = Hmsms kHms msk for phonons of ′ ′ the same energy. Now for any ms and ms the rate Γms↔ms is only nonzero by the inequality of the denominators for the absorption-followed-by-emission and emission-followed-by-absorption variants in the generic two-phonon process of Eq. 33. This assumption results in a scaling of T 3. Alternative 3, consider the spin-lattice Hamiltonian up to second order in the atomic displacements (see Eq. 32), which was originally proposed to explain quadrupolar nuclear spin-lattice relaxation [9]. The second-order contributions to the spin-phonon Hamiltonian taken to ﬁrst order in perturbation theory exhibit a T 7 thermal scaling for acoustic phonons, which has been shown to contribute negligibly to Ω at room temperature [2]. The time-reversal symmetry argument developed in the previous section does not apply to this mechanism since there is no interference of terms by which cancellation can occur. We leave exploration of the mixed term in Eq. 32 for future work.

Two-quasilocalized-phonon Orbach-like processes

Quasilocalized vibrational modes are expected to couple more strongly to the NV spin than delocalized bulk modes because the atoms with the most significant displacements for a quasilocalized mode are in close proximity to the defect. In Ref. [5], Alkausas et al. demonstrated from first principles the existence of a vibrational resonance induced by the NV at ~ωloc = 65 meV with a full width at half maximum (FWHM) of ~∆loc = 32 meV, consistent 9 with experimental measurements of NV photoluminescence and prior calculations of quasilocalized vibrational modes [3, 4]. The effect of the quasilocalized modes comprising the resonance was phenomenologically modeled in Ref. [2] by including a delta function in the density of states. Because the resonance is a relatively wide feature, a less simplified model should account for its width. We demonstrate a model following this consideration here, and show that its temperature dependence is consistent with the exponential term in the empirical model of the NV qubit relaxation rate described by Eq. 6 of the main text. Because the quasilocalized modes comprising the resonance are associated with a single spatial vibrational pattern, we model the coupling of the NV spin to the quasilocalized modes by a Lorentzian which depends only on the mode frequency:

λβ,loc(ω)= λβ,loc S(ω), (60) where the λβ,loc are constants describing the spatial character of the quasilocalized vibrational pattern and S(ω) is the Lorentzian 1 1 2 ∆loc S(ω)= 2 . (61) π (ω ω )2 + 1 ∆ − loc 2 loc The coupling constants are associated with the matrix elements

′ ′ Hmsm , = ms λβ, Fβ m . (62) s loc h | loc | si β X Evaluating Eq. 33 in the continuum limit, the transition rate for Raman processes with quasilocalized phonons is therefore 2 (loc) 2π ′ ′ 1 1 ′ ′ ′′ ′′ Γ = n(ω) (n(ω )+1) S(ω)S(ω ) + Hmsms ,locHms ms,loc ms↔ms ~ ′′ ~ ′′ ~ ′ ¨ ′′ ∆msms + ω ∆msms ω ms X −

′ ~ ~ ′ ′ ′ δ(∆msms + ω ω )Dloc(ω)Dloc (ω ) dω dω (63) × − where Dloc(ω) is the density of states for quasilocalized modes, which can be assumed to take the form Dloc(ω) = 2 Dloc ω as the quasilocalized mode frequencies are within the band of bulk modes. The delta function in Eq. 63 ′ ′ ′ ~ ﬁxes the value of ω to ω = ω + ∆msms / . Because the transition frequencies associated with the NV are small in comparison to both the vibrational resonance FWHM and center frequency, we can take n(ω′) n(ω), S(ω′) S(ω), and D (ω′) D (ω). Performing the integral over ω′ then yields ≈ ≈ loc ≈ loc 2 2 ωD (loc) 2πD 2 loc ′′ ′ ′′ ′ ′′ ′′ Γ ′ ∆msm + ∆m m Hm m ,locHm ms,loc n(ω) (n(ω)+1)(S(ω)) dω. (64) ms↔ms ~6 s s s s s s ≈ ′′ ˆ0 ms X

To demonstrate a simple temperature dependence for Eq. 64, we approximate n(ω) n(ωloc) for phonon frequencies with major contributions to the integral and obtain ≈ 2 2 ωD (loc) 2πD 2 loc ′′ ′ ′′ ′ ′′ ′′ Γ ′ n(ωloc) (n(ωloc)+1) ∆msm + ∆m m Hm m ,locHm ms,loc (S(ω)) dω. (65) ms↔ms ~6 s s s s s s ≈ ′′ ˆ0 ms X (loc) The temperature dependence of Γ ′ can therefore be described by the Bose-Einstein distribution term ms↔ms n(ωloc) (n(ωloc)+1). For ~ωloc kBT , we have n(ωloc) (n(ωloc)+1) n(ωloc), which is the temperature dependence typically invoked in studies≫ of the lifetime of 0 as a function of≈ temperature [1, 2, 10]. Treating ω as a | i loc fit parameter (i.e. equating ωloc with ∆ in Eq. 6 of the main text), Ref. [1] finds ωloc = 73(4) meV, well within the line of the vibrational resonance at 65 meV. Because the Orbach process exhibits the same temperature dependence as the process shown in this section, the exponential term in Eq. 6 of the main text has previously been identified explicitly as an Orbach process [10] or referred to as the result of an Orbach-type process [1, 2]. It is important to recognize, however, that while the Raman process for quasilocalized phonons and the Orbach process have the same temperature scaling, the two represent physically different mechanisms. Specifically, the Orbach process is a two-phonon process in which the intermediate state of the defect is a low-lying excited state and the intermediate state of the combined defect-lattice system is real (i.e., its energy is the same as that of the initial and final states) [11, 12]. In contrast the intermediate system state in the Raman process is virtual (i.e., its energy differs from that of the initial and final states). For these reasons we refer to the two-quasilocalized-phonon process discussed in this section as Orbach-like. 10

NVA3 NVB1 NVC (a) 70 ) 1

− 60 (s Ω 50 (b) 250 )

1 200 − (s

γ 150

100 0 50 100 150 200 250 300 B k (G) (c) 70 ) 1

− 60 (s Ω 50 (d) 250 )

1 200 − (s

γ 150

100 0 20 40 60 80 100 B⊥ (G)

FIG. S1. Extended version of Fig. 3 of the main text, showing all data taken in this work for which the magnitude and orientation of the applied magnetic ﬁeld is known. The vertical gray line at 65 G indicates the range of the data displayed in the main text version of this ﬁgure. 11

TABLE S1. Complete set of data collected for this work. The magnitude and orientation of the applied magnetic ﬁeld were determined with spin echo measurements. Information on the applied magnetic ﬁeld is not included for measurements where spin echo was not performed. 3 −1 3 −1 NV ∆± (MHz) B (G) Bk (G) B⊥ (G) γ (×10 s ) Ω(×10 s ) γ/Ω NVA1 23.9(6) n/a n/a n/a 0.13(2) 0.063(9) 2.1(4) NVA1 125.9(6) n/a n/a n/a 0.111(9) 0.053(3) 2.1(2) NVA1 233.2(6) n/a n/a n/a 0.132(17) 0.061(6) 2.2(4) NVA2 129.7(6) n/a n/a n/a 0.114(12) 0.060(4) 1.9(2) NVA3 121.2(6) 32.1 21.6 23.8 0.133(10) 0.050(4) 2.7(3) NVA3 120.8(6) 21.6 21.6 0.6 0.114(10) 0.059(4) 1.9(2) NVB1 167.1(6) n/a n/a n/a 0.132(11) 0.056(3) 2.4(2) NVB1 831.6(6) n/a n/a n/a 0.17(4) 0.08(2) 2.0(7) NVB1 1207.1(6) n/a n/a n/a 0.10(3) 0.048(13) 2.1(8) NVB1 40.6(6) n/a n/a n/a 0.12(1) 0.049(3) 2.3(2) NVB1 412.7(6) n/a n/a n/a 0.135(13) 0.065(5) 2.1(3) NVB1 623.8(6) n/a n/a n/a 0.138(14) 0.061(5) 2.3(3) NVB1 187.2(6) 33.4 33.4 0.0 0.121(10) 0.062(5) 2.0(2) NVB1 80.4(6) 32.5 14.4 29.2 0.145(11) 0.06(5) 2.4(3) NVB1 177.2(6) 33.0 31.6 9.3 0.130(10) 0.058(5) 2.2(3) NVB1 121.4(6) 32.1 21.7 23.7 0.158(12) 0.052(4) 3.0(3) NVB1 33.7(6) 32.9 6.0 32.4 0.139(13) 0.059(5) 2.4(3) NVB1 154.5(6) 32.6 27.6 17.3 0.135(11) 0.056(5) 2.4(3) NVB1 122.1(6) 32.7 21.8 24.4 0.150(12) 0.055(5) 2.7(3) NVB1 122.7(6) 21.9 21.9 0.0 0.113(10) 0.055(4) 2.1(2) NVB1 16.6(6) 32.2 2.9 32.1 0.128(19) 0.054(5) 2.4(4) NVB1 555.7(6) 146.5 98.8 108.1 0.16(2) 0.064(9) 2.5(5) NVB1 1662.4(6) 300.0 297.5 38.7 0.149(15) 0.053(7) 2.8(5) NVE 303.4(6) 55.1 54.0 10.6 0.140(12) 0.047(3) 3.0(3) NVE 191.1(6) 54.9 34.1 43.1 0.147(14) 0.065(5) 2.3(3) NVE 261.3(6) 55.0 46.6 29.3 0.149(13) 0.067(6) 2.2(3) NVE 130.4(6) 56.1 23.7 50.8 0.185(17) 0.057(5) 3.2(4) NVE 25.3(6) 53.4 4.3 53.2 0.24(3) 0.064(7) 3.7(6) NVE 324.3(6) 57.8 57.8 0.0 0.120(11) 0.061(4) 2.0(2) NVE 294.0(6) 58.6 52.5 26.1 0.151(15) 0.065(5) 2.3(3) NVE 29.4(6) 54.5 4.8 54.3 0.20(3) 0.070(6) 2.8(4) NVE 251.9(6) 60.9 45.1 41.0 0.140(14) 0.052(4) 2.7(3) NVE 217.4(6) 60.0 38.9 45.8 0.154(15) 0.064(5) 2.4(3) NVE 40.3(6) 53.7 7.1 53.3 0.181(19) 0.055(4) 3.3(4) 12

∗ These authors contributed equally. † [email protected] [1] A. Jarmola, V. M. Acosta, K. Jensen, S. Chemerisov, and D. Budker, Temperature- and magnetic-field-dependent longitudinal spin relaxation in nitrogen-vacancy ensembles in diamond, Phys. Rev. Lett. 108, 197601 (2012). [2] A. Norambuena, E. Muñoz, H. T. Dinani, A. Jarmola, P. Maletinsky, D. Budker, and J. R. Maze, Spin-lattice relaxation of individual solid-state spins, Phys. Rev. B 97, 094304 (2018). [3] A. Gali, T. Simon, and J. Lowther, An ab initio study of local vibration modes of the nitrogen-vacancy center in diamond, New Journal of Physics 13, 025016 (2011). [4] J. Zhang, C.-Z. Wang, Z. Zhu, and V. Dobrovitski, Vibrational modes and lattice distortion of a nitrogen-vacancy center in diamond from first-principles calculations, Physical Review B 84, 035211 (2011). [5] A. Alkauskas, B. B. Buckley, D. D. Awschalom, and C. G. Van de Walle, First-principles theory of the luminescence lineshape for the triplet transition in diamond nv centres, New Journal of Physics 16, 073026 (2014). [6] P. Udvarhelyi, V. O. Shkolnikov, A. Gali, G. Burkard, and A. Pályi, Spin-strain interaction in nitrogen-vacancy centers in diamond, Phys. Rev. B 98, 075201 (2018). [7] J. Van Vleck, Paramagnetic relaxation times for titanium and chrome alum, Physical Review 57, 426 (1940). [8] M. Walker, A T 5 spin–lattice relaxation rate for non-Kramers ions, Canadian Journal of Physics 46, 1347 (1968). [9] J. Van Kranendonk, Theory of quadrupolar nuclear spin-lattice relaxation, Physica 20, 781 (1954). [10] D. A. Redman, S. Brown, R. H. Sands, and S. C. Rand, Spin dynamics and electronic states of N-V centers in diamond by epr and four-wave-mixing spectroscopy, Phys. Rev. Lett. 67, 3420 (1991). [11] R. Orbach, Spin-lattice relaxation in rare-earth salts, Proceedings of the Royal Society of London. Series A. Mathematical and Physical Sciences 264, 458 (1961). [12] C. Finn, R. Orbach, and W. Wolf, Spin-lattice relaxation in cerium magnesium nitrate at liquid helium temperature: a new process, Proceedings of the Physical Society 77, 261 (1961).