Abstract
There is a growing interest in hybrid solidstate quantum systems where nuclear spins, interfaced to the electron spin qubit, are used as quantum memory or qubit register. These approaches require long nuclear spin coherence, which until now seemed impossible owing to the disruptive effect of the electron spin. Here we study InGaAs semiconductor quantum dots, demonstrating millisecondlong collective nuclear spin coherence even under inhomogeneous coupling to the electron central spin. We show that the underlying decoherence mechanism is spectral diffusion induced by a fluctuating electron spin. These results provide new understanding of the manybody coherence in central spin systems, required for development of electronnuclear spin qubits. As a demonstration, we implement a conditional gate that encodes electron spin state onto collective nuclear spin coherence, and use it for a singleshot readout of the electron spin qubit with >99% fidelity.
Similar content being viewed by others
Introduction
The original approach to solidstate spin qubits considered the manybody environments, such as nuclear spins, as a decoherence source^{1,2,3}. On the other hand, nuclear spins offer uniquely long coherence storage, making them attractive as buffer memories in photonic quantum information processing systems^{4}. Thus, the attention shifted to hybrid systems, where nuclear spins are interfaced with the electron spin qubit^{5,6} and used as quantum information processing resource^{7,8,9,10}. Several material systems are of interest. In groupIV semiconductors, such as diamond and silicon, isotopic purification offers electron and nuclear spin coherence on the order of seconds^{11} and hours^{12}, respectively. The states of nuclear spins, adjacent to the electron spin qubit, are individually addressable and preserve coherence for tens of milliseconds^{5,13}, but the readout speed and fidelity are restricted by the tiny nuclear magnetic moments. In groupIII–V semiconductor quantum dots (QDs) all nuclei have spins that can be probed using efficient optical techniques^{14}. The electron spin can be controlled on a picosecond timescale^{15} and coherently interfaced to nuclear spin singlequantum excitations^{6}. On the other hand, preserving nuclear spin ensemble coherence in presence of the electron central spin proved challenging^{16}, and, due to the manybody nature, the underlying physics remain an open problem.
The central spin model is described by the hyperfine Hamiltonian \({\hat{{{{{{{{\mathcal{H}}}}}}}}}}_{{{{{{{{\rm{hf}}}}}}}}}={\sum }_{j}\;{A}_{j}\hat{{{{{{{{\bf{s}}}}}}}}}\cdot \hat{{{{{{{{\bf{I}}}}}}}}}\), where constants A_{j} (1 ≤ j ≤ N) characterize the coupling of the singleelectron spin s to the spins I_{j} of N lattice nuclei (Fig. 1a). Magnetic interactions between the nuclei with pairwise coupling constants b_{j,k}, together with inhomogeneity in A_{j}, result in complex dynamics^{2,3,17,18,19}, characteristic of manybody quantum phenomena^{20,21}. In a typical epitaxial device (Fig. 1a) a Fermi reservoir of electrons is introduced, and electric field is applied through the gate voltage V_{G}, to control the charge state of the QD. In an empty (0e) InGaAs QD the direct nuclearnuclear interactions limit the nuclear spin coherence time to a few millisecond range^{16,22} \({T}_{2,{{{{{{{\rm{N}}}}}}}}}^{(0e)}\propto h/\max ( {b}_{j,k} )\), where h is Planck’s constant. The spin of a single electron (1e) induces hyperfine shifts in precession frequencies, ranging from 0 for distant nuclei to \(\max ( {A}_{j} )/(2h)\approx 100\) kHz for nuclei at the center of the QD. These shifts, known as Knight shifts, result in short nuclear spin dephasing time \({T}_{2,{{{{{{{\rm{N}}}}}}}}}^{* ,(1e)}\propto 2\hslash /\sqrt{\langle {A}_{j}^{2}\rangle }\approx 3\) μs, determined simply by inhomogeneity of A_{j} (ℏ = h/(2π) is the reduced Planck’s constant). Dephasing is reversible and can be refocused with spin echo^{23} to reveal the timescale T_{2,N} of irreversible decoherence. Unlike with \({T}_{2,{{{{{{{\rm{N}}}}}}}}}^{* }\), predicting T_{2,N} is a difficult problem. Previous experiments^{16} gave rather short coherence \({T}_{2,{{{{{{{\rm{N}}}}}}}}}^{(1e)}\approx 20\) μs, that was ascribed to indirect nuclearnuclear interactions mediated by the electron spin. However, to fully understand the QD coherence, one must take into account the effects of the electron Fermi reservoir.
By optimizing the tunnel coupling between QD and the Fermi reservoir, we achieve here an isolatedqubit regime where the lifetime of a QD electron central spin is long and therefore spectral diffusion^{24,25} is reduced. We find that indirect nuclearnuclear interactions are negligible, and inhomogeneous interaction with the central spin is harmless under these conditions, allowing us to achieve long coherence \({T}_{2,{{{{{{{\rm{N}}}}}}}}}^{(1e)} \; > \; 1\) ms in a manybody nuclear spin ensemble. While manybody spin environment is often seen as a hindrance for qubits, here we design a protocol where nuclear spin coherence is harnessed for a singleshot readout of the electron spin qubit, with fidelity matching the state of the art^{26,27,28,29,30,31,32}.
Results
Coherence of a spin ensemble coupled to the central spin
We study diode structures (Fig. 1a) with tunnel barrier t_{B} = 37 or 52 nm, thick enough to ensure long singleelectron spin lifetimes T_{1,e} up to ≈1 ms or ≈1 s, respectively^{33}. We investigate spin3/2 nuclei of ^{75}As, ^{69}Ga and ^{71}Ga. With magnetic field B_{z} applied along the sample growth axis, nuclear states with spin projections I_{z} = ±1/2 form effective spin1/2 ensembles that we focus on, whereas I_{z} = ±3/2 states are spectrally detuned and can be ignored (see Methods). Hyperfine interaction between the electron spin and nuclear spins^{14} (see Supplementary Note 1) provides a tool both for hyperpolarisation of the nuclei via optical pumping, and for detection of the nuclear ensemble polarization via hyperfine shifts \({E}_{{{{{{{{\rm{hf}}}}}}}}}={\sum }_{j}\;{A}_{j}{\hat{s}}_{{{{{{{{\rm{z}}}}}}}}}\langle {\hat{I}}_{{{{{{{{\rm{z}}}}}}}},j}\rangle\) in the optical emission spectra of the QD electrons (here \(\langle {\hat{I}}_{{{{{{{{\rm{z}}}}}}}},j}\rangle\) is the expectation value for the z projection spin operator of the jth nuclear spin). The reciprocal hyperfine effect of the electron spin is characterized by the Knight shifts in the nuclear magnetic resonance (NMR) frequencies.
We examine the collective coherence of the QD nuclear spins using optically detected NMR protocol (Fig. 1b). Figure 1c shows the result of an NMR experiment, where a single resonant radiofrequency (rf) burst of duration t_{rf} induces coherent Rabi oscillations between the I_{z} = −1/2 and +1/2 nuclear spin states. The inhomogeneous Knight shifts ν_{e,j} ∝ A_{j} induced by the electron (1e) result in faster damping of coherent nuclear spin rotations compared to an empty QD (0e). To characterize this inhomogeneity we use a spin echo sequence (Fig. 1b), where the initial π/2 pulse transforms nuclear spin polarization into collective (multiquantum) coherence, followed by free evolution over time τ_{0}, a refocusing π pulse, and a further free evolution time τ, before the final π/2 pulse converts the remaining coherence back into optically detectable nuclear spin polarization. The width of the spin echo peak observed around τ = τ_{0} (Fig. 1d) is proportional to dephasing time, which reduces from \({T}_{2,{{{{{{{\rm{N}}}}}}}}}^{* ,(0e)}\approx 35\) μs in a neutral QD to \({T}_{2,{{{{{{{\rm{N}}}}}}}}}^{* ,(1e)}\approx 4.3\) μs when QD is charged with a single electron. Using Fourier transform, the dephasing dynamics reveal the inhomogeneous broadening of the nuclear spin resonance (Fig. 1e). In an empty QD the broadening of ≈13 kHz is due to the second order quadrupolar shifts, whereas in presence of an electron the ≈ 126 kHz linewidth is dominated by the inhomogeneous Knight shifts.
By fixing τ = τ_{0} we remove the dephasing and examine pure nuclear spin decoherence by measuring spin echo amplitude decay with increasing total free evolution time τ_{evol} = τ_{0} + τ (Fig. 1f). From exponential fitting we find the coherence time \({T}_{2,{{{{{{{\rm{N}}}}}}}}}^{(0e)}\approx 1.8\) ms in a bare ^{69}Ga nuclear spin ensemble, in good agreement with previous studies^{16,22}. By contrast, the coherence time \({T}_{2,{{{{{{{\rm{N}}}}}}}}}^{(1e)}\approx 0.7\) ms in presence of a single QD electron is a factor of ≈ 30 longer than reported earlier^{16}. By varying the gate bias V_{G} (Fig. 2a), we observe a clear plateau in T_{2,N} around V_{G} ≈ 0.42 V, confirming that QD is in singleelectron (1e) Coulomb blockade. Nuclear coherence time \({T}_{2,{{{{{{{\rm{N}}}}}}}}}^{(1e)}\) in this regime is measured as a function of magnetic field B_{z} (solid symbols in Fig. 2b). At low fields, \({T}_{2,{{{{{{{\rm{N}}}}}}}}}^{(1e)}\) is nearly independent of B_{z}, but changes over to \({T}_{2,{{{{{{{\rm{N}}}}}}}}}^{(1e)}\propto {B}_{{{{{{{{\rm{z}}}}}}}}}^{4}\) scaling at B_{z} ≳ 4 T, in stark contrast to \({T}_{2,{{{{{{{\rm{N}}}}}}}}}^{(1e)}\propto {B}_{{{{{{{{\rm{z}}}}}}}}}^{2}{A}^{3}\) law predicted previously for hyperfinemediated nuclear spin decoherence mechanism^{16}. Furthermore, at high field B_{z} ≈ 7.8 T we find similar \({T}_{2,{{{{{{{\rm{N}}}}}}}}}^{(1e)}\approx 130\) μs for ^{75}As and ^{71}Ga nuclei, despite a factor of ≈ 1.8 difference in their hyperfine constants A.
Decoherence driven by spectral diffusion
To explain the observed \({T}_{2,{{{{{{{\rm{N}}}}}}}}}^{(1e)}\) values we invoke a model, historically known as “spectral diffusion”^{24,34} in solidstate magnetic resonance (not to be confused with physical diffusion of spins in gases and liquids). In this model, one type of spins is observed, while the second type is a source of fluctuating local fields. In our case, the nuclear spins are observed, while the singleelectron spin is the fluctuating environment. The evolution of the electron spin is intertwined with the nuclear spin dynamics^{35}, but for the present case, an adequate description is achieved by assuming that the electron spin makes uncorrelated random jumps between the s_{z} = ±1/2 states as sketched in Fig. 1b. The validity of this telegraphprocess approximation comes from the fact that the nuclear spin phase is a timeintegral of the electron spin evolution s_{z}(t) (see Supplementary Note 2). As a result, only the lowfrequency components play a role, and all the relevant information about s_{z}(t) is captured by the rate of the electron spin flips. In the studied range of magnetic fields B_{z} ≤ 8 T this rate 1/(2T_{1,e}) ≲ 6000 s^{−1} (ref. ^{33}) is small compared to root mean square Knight shift δν_{e} ≈ 30 kHz, derived from NMR spectra of Fig. 1e. In this regime of slow fluctuations, the nuclear spin decoherence time due to spectral diffusion approximately equals the electron spin lifetime T_{2,N,SD} ≈ 1.38T_{1,e} (see derivation in Supplementary Note 2). Spectral diffusion dominates nuclear spin decoherence at high magnetic fields where electron spin lifetimes shorten as \({T}_{1,{{{{{{{\rm{e}}}}}}}}}\propto {B}_{{{{{{{{\rm{z}}}}}}}}}^{4}\). In an empty QD nuclear dipoledipole interaction causes nuclear spin ensemble decoherence on a timescale \({T}_{2,{{{{{{{\rm{N}}}}}}}}}^{(0e)}\)—including this mechanism, we find the following approximation for nuclear spin coherence time in presence of the electron central spin:
This closed form model fully describes the experimental dependence of \({T}_{2,{{{{{{{\rm{N}}}}}}}}}^{(1e)}\) on B_{z} (lines in Fig. 2b), as well as isotopeindependent \({T}_{2,{{{{{{{\rm{N}}}}}}}}}^{(1e)}\) at high B_{z}. In the t_{B} = 52 nm sample, T_{1,e} is very long at B_{z} ≲ 3 T, making the dipoledipole mechanism dominant – and indeed we find \({T}_{2,{{{{{{{\rm{N}}}}}}}}}^{(1e)}\approx {T}_{2,{{{{{{{\rm{N}}}}}}}}}^{(0e)}\) in that case.
Qualitative understanding of spectral diffusion is found by considering two scenarios: In the absence of electron spin flips, the static inhomogeneous Knight shifts are fully refocused to form nuclear spin echo. By contrast, even a singleelectron spinflip unbalances the phases acquired by the nuclei before and after the refocusing π pulse (except for those rare flips occurring within a short interval \({\approx}{T}_{2,{{{{{{{\rm{N}}}}}}}}}^{* ,(1e)}\) at the start or the end or the spin echo sequence). The probability to have zero electron flips in an increasing time interval decreases exponentially, and with it decays exponentially the average nuclear spin coherence on a timescale T_{2,N,SD} ≈ 1.38T_{1,e}. A direct observation of this process is shown in Fig. 3a, where we plot histograms of the spin echo amplitude ΔE_{hf} detected with singleshot probe pulses, as opposed to averaging over multiple pumprfprobe cycles used for the data of Figs. 1 and 2. At long evolution times τ_{evol} ≈ 1300 μs the echo is destroyed by decoherence, resulting in a single peak at ΔE_{hf} ≈ 0. At short τ_{evol} ≈ 0.4 μs the nuclear coherence is preserved in most cases, leading to a peak at ΔE_{hf} ≈ 1.9 μeV. At intermediate τ_{evol} a bimodal distribution is observed, demonstrating the two discrete possibilities of nuclear spin echo preservation or destruction, if electron spin does not or does flip, respectively.
Singleshot readout of the central spin qubit
Figures 3b, c show singleshot NMR measurement with a sequence (π/2)_{x} − τ_{evol} − (π/2)_{y}, which generates collective nuclear spin coherence and probes its quadrature component ΔE_{hf} following the free evolution time τ_{evol}. In an empty QD (0e, Fig. 3b) the distribution of detected ΔE_{hf} is unimodal with a width given by the optical readout noise. By contrast, a clear bimodal distribution is observed with a single electron in a QD (1e, Fig. 3c) at τ_{evol} ≈ 0.3 μs. The two modes correspond to the two discrete electron spin qubit states s_{z} = ±1/2, that add positive or negative phase ∝ τ_{evol} to the nuclear spin coherence. The free evolution can be seen as a quantum logic gate with unitary propagator \(=\left\uparrow \right\rangle \left\langle \uparrow \right{\otimes \bigotimes }_{j = 1}^{N}{e}^{+\;\frac{i{A}_{j}{\tau }_{{{{{{{{\rm{evol}}}}}}}}}}{\hslash }{\hat{I}}_{{{{{{{{\rm{z}}}}}}}},j}}+\left\downarrow \right\rangle \left\langle \downarrow \right{\otimes \bigotimes }_{j = 1}^{N}{e}^{\;\frac{i{A}_{j}{\tau }_{{{{{{{{\rm{evol}}}}}}}}}}{\hslash }{\hat{I}}_{{{{{{{{\rm{z}}}}}}}},j}}\), where the electron spin state ↑ or ↓ controls the phase of the nuclei ( ⊗ is the Kronecker product operator). Owing to the inhomogeneity of the QD electronnuclear spin coupling, the phase acquired by the jth individual nuclear spin (1 ≤ j ≤ N) is proportional to its hyperfine constant ∝ A_{j}. The final (π/2)_{y} pulse transforms this phase into the optically detected hyperfine shift, thus implementing a singleshot readout of the electron spin qubit, which is a key ingredient in quantum information processing.
In order to characterize the noise in the singleshot readout, we use the double Gaussian model^{36,37} for the distribution of the optically measured hyperfine shifts ΔE_{hf} (lines in Fig. 3c). Setting the threshold level in the middle between the two Gaussian modes (ΔE_{hf} = 0 in Fig. 3c), we interpret the abovethreshold readouts ΔE_{hf} > 0 as electron spin up s_{z} = +1/2, and ΔE_{hf} < 0 as electron spin down s_{z} = −1/2. Using the fitted parameters, we calculate ≳99.8% for the probability that the singleshot measurement of ΔE_{hf}, including Gaussian noise, gives the correct electron spin state s_{z} (see details in Methods). Previously, singleshot readout of electron spin state s_{z} in an optically active QD was achieved by measuring fluorescence under resonant optical excitation, which acts back on the electron spin, limiting the readout fidelity. Our approach is different, in that it first converts the fragile electron spin degree of freedom into a more robust nuclear spin ensemble polarization, which we then readout optically. The conversion approach is known to provide improved readout fidelity, for example by transforming spin into charge, as demonstrated in gatedefined quantum dots^{26,31}, and more recently in diamond spins^{32}.
There are two factors limiting the overall readout speed and fidelity. One is the noise in the optical signal, which requires detection time on the order of 1–10 ms to achieve the estimated above singleshot accuracy of ≳0.998. These parameters can be enhanced further by optimizing the spectral resolution and collection efficiency of the optical detection setup. The other factor is related to the finite lifetime of the electron spin qubit. The flips of the electron spin during its conversion into the quadrature nuclear spin polarization impose a fidelity limit on the order of \({\approx}1({T}_{2,{{{{{{{\rm{N}}}}}}}}}^{* ,(1e)}/{T}_{1,{{{{{{{\rm{e}}}}}}}}})\), where the microsecondrange nuclear spin dephasing time \({T}_{2,{{{{{{{\rm{N}}}}}}}}}^{* ,(1e)}\) is also the optimal conversion time. In the particular experiment of Fig. 3c, conducted at B_{z} = 7.8 T, we estimate the spinflip fidelity limit to be ≈0.99, although this bound can be increased to >0.9999 by simply reducing the field to B_{z} ≤ 2 T, where the electron spin lifetimes exceed T_{1,e} > 10 ms (see Methods). Taking the worst of the two limiting factors, we arrive to an overall fidelity estimate >0.99, comparable to the state of the art in spin^{26,27,30,31,32} and superconducting qubits^{28,29}. Notably, the singleshot readout is achieved using a small fraction (<13%, see Methods) of the QD nuclei, so that the remaining nuclear spins can be used as a quantum resource. For example, it is possible to store multiple snapshots of the electron spin qubit state, by converting it into different nuclear isotope states (^{69}Ga, ^{71}Ga, ^{75}As), followed by the optical retrieval at the end of operations on the qubit.
The relative weights of the two modes in Fig. 3c reveal the electron spin polarization degree ρ_{e} ≈ 0.35, corresponding to equilibrium of an electron with g_{e} ≈ −0.63, in good agreement with electron gfactors found in similar QDs^{38}. At long evolution times \({\tau }_{{{{{{{{\rm{evol}}}}}}}}}\gg {T}_{2,{{{{{{{\rm{N}}}}}}}}}^{* ,(1e)}\) the coherence imprinted by the electron spin is dephased, resulting in a unimodal distribution around ΔE_{hf} ≈ 0 (e.g., at τ_{evol} ≈ 500 μs).
Dynamical decoupling of the spin ensemble
Spin echo sequence with a single refocusing π pulse is the simplest case of a more general concept of dynamical decoupling^{39}, where fast (bangbang) rf control pulses are applied to filter out the unwanted interactions and increase the coherence time. For a single spin, decoupling from external environment can be achieved with a sequence of π pulses^{40,41}, whereas for nuclear spin ensembles, sequences of π/2 and π pulses, such as CHASE^{22}, have been designed to suppress both the ensemble inhomogeneity and the internal dipolar interactions. We examine ^{69}Ga nuclear spin coherence under dynamical decoupling in two regimes of magnetic field.
At low magnetic fields (B_{z} = 2.5 T), where electroninduced spectral diffusion is weak, we find that a 5pulse sequence CHASE5 provides an improved coherence \({T}_{2,{{{{{{{\rm{N}}}}}}}}}^{(1e)}\approx 2.2\) ms, compared to a singlepulse spin echo \({T}_{2,{{{{{{{\rm{N}}}}}}}}}^{(1e)}\approx 1.5\) ms. This confirms the validity of the dynamical decoupling approach for a spin ensemble inhomogeneously coupled to the central spin. Dynamical decoupling sequences are designed such that higher control rate (larger number of control pulses) converges the residual Hamiltonian to zero, progressively increasing the coherence time T_{2,N}. Experiments with longer decoupling sequences CHASE10 and CHASE20 show reduction of the echo amplitude to a level where T_{2,N} becomes unmeasurable (see Supplementary Note 5), which we ascribe to strong inhomogeneous Knight shift broadening. However, this limitation is technical, and can be addressed by increasing the rf control field amplitude and bandwidth. Since nuclear spin decoherence in both neutral (0e) and charged (1e) QDs is governed by the same mechanisms of inhomogeneous broadening and nuclear dipolar interactions, we expect that \({T}_{2,{{{{{{{\rm{N}}}}}}}}}^{(1e)}\) can be at least as large as \({T}_{2,{{{{{{{\rm{N}}}}}}}}}^{(0e)} \; > \; 10\) ms, as found previously in neutral QDs^{9,22}. Thus, dynamical decoupling can be a route to long coherence storage in manybody ensembles, with coherence times comparable to dilute individual nuclear spins^{5}.
Experiments at high magnetic field (B_{z} = 7.8 T), where nuclear spin coherence is dominated by the random electron spin flips, show that CHASE dynamical decoupling no longer provides any improvement over spin echo (see data in Supplementary Note 5). The decoherence mechanism is similar to the case of a singlepulse spin echo. A singleelectron spinflip is sufficient to completely destroy the balance of phases accumulated by spins in the free evolution intervals of the decoupling sequence. As a result, nuclear spin coherence is essentially limited by the electron spin lifetime \({T}_{2,{{{{{{{\rm{N}}}}}}}}}^{(1e)}\;\lesssim \;1.38{T}_{1,{{{{{{{\rm{e}}}}}}}}}\), and dynamical decoupling at high fields would require the control rate to be high enough to perform multiple decoupling cycles on a timescale shorter than T_{1,e}. A more feasible approach would be to operate in the lowfield regime, where electron spin is essentially static. In this case, optimal coherence of the nuclear spin environment would be achieved by synchronizing its dynamical decoupling with the coherent control of the electron central spin qubit.
Discussion
Spectral diffusion was previously observed for individual electron spins subject to fluctuations of the dilute nuclear spin ensembles^{11}. Our results show that the same concept works in reverse, and describes the decoherence of 100% abundant nuclear spin ensembles subject to fluctuations of a single electron in a III–V semiconductor quantum dot. We show experimentally that manybody interactions and inhomogeneous coupling to the common central spin are not obstacles to long coherence storage in a nuclear spin ensemble, which is required to implement recent proposals for QD spin–photon networks with nuclear spin quantum memories^{7} and registers^{9}. In contrast to previous studies^{16}, we find no signature of decoherence arising from indirect hyperfinemediated nuclearnuclear interactions, which are expected to be prominent only below a certain magnetic field. Earlier works, not related to nuclear spin coherence, reported varying upperboundary values 0.02 − 0.75 T, for the range of magnetic fields where hyperfinemediated interactions play an important role^{42,43,44}. Our results show that nuclear spin coherence is unaffected by hyperfinemediated nuclearnuclear couplings at least down to B_{z} ≈ 0.25 T, offering a wide range of magnetic fields suitable for long coherence storage.
While our experiments are conducted on collective multiquantum coherent excitations of the nuclear spin ensemble, we expect the same mechanisms and similarly long coherence times for the singlequantum spinwave excitations. Generation of single nuclear spin waves (single magnons) in quantum dots has been reported recently^{6}, but with short dephasing time (few microseconds). Our results suggest that once inhomogeneity is refocused, the coherence of the spinwave quantum memories in quantum dots may be extended to milliseconds. The findings of this work on InGaAs/GaAs QDs are expected to apply to highquality optically active GaAs/AlGaAs QDs, where small intrinsic strain^{9} holds a promise for even longer coherence and complete control of the hybrid electronnuclear spin quantum system.
We have demonstrated generation of the collective nuclear spin phase, controlled by the electron spin qubit state. The fidelity of this operation is limited by inhomogeneity of the electronnuclear interactions. This problem can be addressed by operating on small subensembles of nuclear spins with similar hyperfine coupling to the electron spin. Such manipulation can be achieved with spectrally selective control pulses^{45}—this approach is feasible with the highsensitivity optical techniques, capable of detecting a fewpercent fraction of the nuclear spin ensemble, as demonstrated in our experiments. Even without any additional development, we demonstrate that controlledphase operation on a small fraction (<13%) of nuclear spins can be used for a singleshot highfidelity readout of the electron spin qubit. The remaining nuclei can be used as a spinwave quantum memory. Alternatively, the QD nuclear spins can be divided into small subensembles, e.g., by the isotope type (^{69}Ga, ^{71}Ga or ^{75}As), and used to record the electron spin state at multiple timepoints. This for example, can alleviate the need to initialize the electron spin qubit prior to its coherent manipulation—both the initial and final electron spin states can be stored in the nuclei for subsequent optical retrieval. Future work will include exploration of the new opportunities offered by a highly coherent electronnuclear spin system in individual quantum dots.
Methods
Sample structures
The samples are lowdensity InAs selfassembled QDs grown on a GaAs substrate using molecular beam epitaxy. The dots are placed between two distributed Bragg reflectors consisting of GaAs and AlAs layers and forming a λ/2 optical cavity. The low temperature groundstate optical emission of the studied QDs is at ≈ 950 nm. The Fermi reservoir is a Si doped GaAs layer (Si concentration ≈ 1.1 × 10^{18} cm^{−3}), separated from QDs by a GaAs tunnel barrier layer of thickness t_{B}. The samples are processed into planar Schottky diode structure, allowing for the charge state of the dots to be controlled by applying external bias V_{G} to the top metal gate. The studied semiconductor device have been examined previously using photoluminescence, resonance fluorescence and spin lifetime measurement techniques—the detailed results can be found in ref. ^{33} and the Supplementary Information therein.
Electron spin relaxation rates Γ_{e} = 1/T_{1,e} measured at low temperature T = 4.5 K are well described by the following model^{33}:
where the first term Γ_{e,cotun} accounts for the fieldindependent relaxation induced by electron cotunneling and the second term describes the fielddependent relaxation induced by acoustic phonons. Equation (2) is then substituted into Eq. (1) to find a closed form dependence of the nuclear spin coherence time on the external magnetic field B_{z}. The phonon mechanism parameters are Γ_{e,ph} ≈ 2.27 ± 0.48 \({{{{{{{{\rm{s}}}}}}}}}^{1}\times {{{{{{{{\rm{T}}}}}}}}}^{{k}_{{{{{{{{\rm{ph}}}}}}}}}}\) and k_{ph} ≈ 4.1 ± 0.13 in both samples^{33}. Cotunneling depends on the barrier thickness. For the t_{B} = 52 nm sample the range of values is 1/Γ_{e,cotun} ≈ 1.26 − 1.65 s for different individual QDs. For the t_{B} = 37nm sample the dottodot variation is more pronounced 1/Γ_{e,cotun} ≈ 1.2 − 5.8 ms. Electron spin lifetimes have not been measured for the particular QDs used here for nuclear spin coherence studies, and thus we treat Γ_{e,cotun} in the t_{B} = 37 nm sample as the only adjustable parameter. The best fit value plotted by the solid line in Fig. 2b is 1/Γ_{e,cotun} ≈ 0.96 ms, in good agreement with the range of rates found from direct measurements on other QDs in the same sample.
Experimental techniques
The sample is placed in a liquid helium bath cryostat equipped with a superconducting magnet, providing a field up to B_{z} = 8 T, parallel to sample growth direction and optical axis z (Faraday geometry). An aspheric lens is used for optical excitation of the QD and photoluminescence (PL) collection. Diode lasers emitting at 850 nm are used both for optical polarization of the nuclear spins (optical pump pulses) and PL excitation (optical probe pulses). The collected PL is dispersed with a 1 m double grating spectrometer and recorded with a chargecoupled device camera. The changes in the spectral splitting of a negatively charged trion X^{−}, derived from the PL spectra, are used to measure the hyperfine shifts E_{hf} proportional to the nuclear spin polarization degree (see Supplementary Fig. 3). The oscillating magnetic field B_{x}⊥z implementing the coherent control of the collective nuclear spin state is produced by a coil placed at a distance of ≈ 0.5 mm from the QD sample. The coil is made of 10 turns of a 0.1 mm diameter enameled copper wire wound on a ≈ 0.4 mm diameter spool in five layers, with two turns in each layer. The coil is fed by a 1 kW rf amplifier through a resonant impedance matching network made of 50 Ω coaxial cables.
Optically detected NMR
The details of NMR implementation are given in Supplementary Note 3. Overall the experimental cycle follows the timing diagram shown in Fig. 1b. As in previous work^{16,22,46}, the optically induced nuclear spin state is augmented with adiabatic rf frequency sweeps over the inhomogeneously broadened satellite NMR transitions −3/2 ↔ −1/2 and +1/2 ↔ +3/2. The sweeps exchange the populations of the I_{z} = −3/2 and −1/2 pair of states, as well as populations of the I_{z} = +1/2 and +3/2 pair. This way the population difference of the I_{z} = ±1/2 pair is maximized, prior to coherent manipulation of the I_{z} = ±1/2 subspace. The amplitude of the frequencyswept rf excitation is chosen to produce Rabi frequency between ≈ 1 − 4 kHz, and the typical sweep rates are between 5−10 MHz s^{−1}. In some experiments, such as singleshot measurements of Fig. 3, the same adiabatic sweeps are applied for the second time, after the coherent manipulation, transferring the final populations of the I_{z} = ±1/2 states back into I_{z} = ±3/2 states. This gives a factor of ≈3 increase in the variation of the optically detected hyperfine shifts ΔE_{hf} at the expense of a longer experimental time. In the case such second set of sweeps is used, all the measured ΔE_{hf} are divided by 3 to obtain the values directly comparable with the measurements where adiabatic sweeps are applied only after the optical pumping. In all experiments, except for singleshot measurements, the hyperfine shift ΔE_{hf} is acquired by averaging over 15 − 60 pumprfprobe cycles shown in Fig. 1b to obtain an approximation to a statistical average NMR signal. In echo decay experiments, such as shown in Fig. 1f, the dependence of the echo amplitude on the total free evolution time τ_{evol} is fitted with stretched or compressed exponentials \(\propto {{{{{{{{\rm{e}}}}}}}}}^{{({\tau }_{{{{{{{{\rm{evol}}}}}}}}}/{T}_{{{{{{{{\rm{2,N}}}}}}}}})}^{\eta }}\), where η is the parameter describing stretching (η < 1) or compression (η > 1).
Coherent control of the nuclear spins is achieved using high power rf pulse bursts with rotatingframe amplitude of up to B_{1} ≈ 10 mT, which corresponds to laboratoryframe amplitude B_{x} ≈ 20 mT and ^{75}As Rabi frequency ν_{1} = 2γB_{1}/(2π) ≈ 140 kHz, where the additional factor of 2 is from the matrix element of the spin3/2 operator projected onto the I_{z} = ±1/2 subspace. Owing to the internal strain of the QDs, the quadrupolar shifts of the I_{z} = ±3/2 nuclear spin states are in MHz range, much larger than ν_{1}. As a result, the I_{z} = ±3/2 states are strongly detuned, making it possible to perform selective coherent rotations within the I_{z} = ±1/2 subspace of interest. In order to achieve broadband uniform rotation of the spin ensemble, ν_{1} must be larger than the resonance spectral broadening of the I_{z} = ±1/2 subspace. While larger B_{1} can be achieved by increasing the rf power, the practical limitations arise from the parasitic effects of the rf electric field. Above certain level, typically corresponding to B_{1} ≈ 5 mT, high power rf pulses are found to induce electron spin flips, which then disrupt the formation of the nuclear spin echo. In singleshot spin echo experiments with short τ_{evol} the rfinduced electron spin flips are observed as nearly equal weights of the two modes in the histogram. Consequently, experiments shown in Fig. 3a are conducted at a reduced B_{1} ≈ 5 mT, where the probability of parasitic electron spin flipping is estimated from fitting to be within <0.05. The downside of a low B_{1} is the reduced spectral bandwidth of the control pulses, which leads to incomplete rotation for some of the spins and a reduced spin echo amplitude. In those spin echo experiments where the readout is averaged over multiple pumpprobe cycles, the contributions of the cycles where electron spin is flipped by the rf field can be ignored, leading to correct spin echo decay time \({T}_{2,{{{{{{{\rm{N}}}}}}}}}^{(1e)}\) but with a reduced echo amplitude. Future work may include optimization of rf circuitry with the aim of maximizing B_{1} while reducing the parasitic rf electric field.
Most of the pulsed nuclear magnetic resonance experiments are conducted on the ^{69}Ga isotope due to its favorable balance between the hyperfine shift amplitude, Knight shift inhomogeneity, and the quadrupolar inhomogeneity. Additional results for ^{75}As isotope can be found in Supplementary Note 4.
Estimate of the number of nuclei used for singleshot electron spin detection
The singleshot NMR experiments presented in Fig. 3 are conducted on ^{75}As nuclei, which have 100% natural abundance and thus make up 50% of all the nuclei in a QD, the rest being the groupIII Ga, In and possibly some Al nuclei. The nuclear spin polarization degree produced by optical pumping is estimated to be ∣ρ_{N}∣ ≈ 0.65 (refs. ^{47,48}), which corresponds to a dimensionless inverse spin temperature ∣β∣ ≈ 0.98 in a Boltzmann distribution of the nuclear spin level population probabilities \(p({I}_{{{{{{{{\rm{z}}}}}}}}})\propto \exp (\beta {I}_{{{{{{{{\rm{z}}}}}}}}})\). The nuclei that are initially in the I_{z} = ±3/2 states are transferred into the I_{z} = ±1/2 states by adiabatic rf frequency sweeps—only these nuclei contribute to NMR signal, and from the p(I_{z}) distribution their fraction is ≈0.67. Finally, the reduced rf amplitude used in singleshot NMR (^{75}As CT Rabi frequency ν_{1} ≈ 70 kHz) results in selective rf pulses, which excite only the nuclei whose NMR frequencies are within the pulse bandwidth ∝ ν_{1}. In spin echo experiments, such lowering of the rf amplitude results in spin echo amplitude that is a factor of ≈0.38 smaller compared to the echo amplitude in experiments conducted with the highest possible rf amplitude. Combining all these factors we find an upper estimate <0.5 × 0.67 × 0.38 ≈ 0.13 of the fraction of the QD nuclei taking part in a singleshot detection of the electron spin state. With the total number of nuclei estimated to be N ≈ 4 × 10^{4} in the studied dots^{33}, this corresponds to ≈5000 nuclei participating in electron spin detection.
Estimate of the singleshot electron spin readout fidelity and its limiting factors
The histograms of the singleshot quadrature NMR signal ΔE_{hf} shown in Fig. 3c are fitted with a double Gaussian function \({A}_{()}{2}^{{\left(\frac{{{\Delta }}{E}_{{{{{{{{\rm{hf}}}}}}}}}+{{\Delta }}{E}_{{{{{{{{\rm{hf}}}}}}}},0}/2}{w/2}\right)}^{2}}+{A}_{(+)}{2}^{{\left(\frac{{{\Delta }}{E}_{{{{{{{{\rm{hf}}}}}}}}}{{\Delta }}{E}_{{{{{{{{\rm{hf}}}}}}}},0}/2}{w/2}\right)}^{2}}\), where ΔE_{hf,0} is the splitting of the two modes, w is the full width at half maximum of each mode peak and A_{(+)} (A_{(−)}) is the amplitude of the mode corresponding to positive (negative) average hyperfine shift variation ΔE_{hf}. The relative difference of the amplitudes reflects the electron spin polarization degree ρ_{e} = (A_{(+)} − A_{(−)})/(A_{(+)} + A_{(−)}). For the measurement at τ_{evol} = 0.3 μs in Fig. 3c we find ΔE_{hf,0} ≈ 1.75 μeV, and the width w ≈ 0.67 μeV determined by collection efficiency and spectral resolution of the instruments used to analyze QD PL. When the quadrature NMR signal ΔE_{hf} is measured, it is interpreted as electron spin state s_{z} = +1/2 (s_{z} = −1/2) for ΔE_{hf} > 0 (ΔE_{hf} < 0). Thus, the total probability of correct detection is
Taking the integrals we find for the readout fidelity:
where erf is the standard Gauss error function. Evaluating this for the measurement with τ_{evol} = 0.3 μs, we find F ≈ 0.9989. This is the value quoted in the main text above and it is determined by the accuracy with which the collective nuclear spin polarization of the quantum dot is readout optically. Increasing the optical probe duration T_{Probe} increases the number of the PL photons collected and reduces w, thus improving the measurement accuracy of the spectral splitting variation ΔE_{hf}. However, if T_{Probe} is too long, it depolarizes the nuclei and reduces the separation ΔE_{hf,0} of the histogram modes, thus reducing the readout fidelity. Therefore, there is an optimal probe duration T_{Probe}, which maximizes the fidelity F. In our experiments T_{Probe} is set to this optimal value. Further improvement of the electron spin readout fidelity can be achieved using, e.g., solid immersion lenses^{49} in order to increase probe PL photon collection without the need for a longer T_{Probe}.
Even for a perfect (noisefree) optical probing, the readout fidelity is limited by the flips of the electron spin during its conversion into the nuclear spin coherence. The conversion takes place both during the free evolution time τ_{evol}, and partially during the π/2 control pulses of the (π/2)_{x} − τ_{evol} − (π/2)_{y} sequence. Simple calculations show that for any realistic distribution of the inhomogeneous Knight shifts, the peak amplitude of the quadrature signal generated by the (π/2)_{x} − τ_{evol} − (π/2)_{y} sequence is reached when the evolution time approximately equals the dephasing time \({\tau }_{{{{{{{{\rm{evol}}}}}}}}}\approx {T}_{2,{{{{{{{\rm{N}}}}}}}}}^{* ,(1e)}\). However, in the experiment (Fig. 3c), the quadruature signal (splitting ΔE_{hf,0} of the modes) is maximized at τ_{evol} = 0.3 μs, much shorter than \({T}_{2,{{{{{{{\rm{N}}}}}}}}}^{* ,(1e)}\approx 4.5\) μs. This suggests that some conversion takes place during the control pulses, whose duration is t_{π/2} = 3.75 μs. Owing to the spin locking^{50} arising when the nuclear spin control pulses are strong (i.e., Rabi frequency of the rf pulse ν_{1} ≈ 70 kHz is sufficiently large compared to Knight shift broadening δν_{e} ≈ 30 kHz), the effect of the electron spin flips during rf excitation is smaller than during the free evolution. Thus, we estimate that the effective conversion time τ_{conv} in our experiment is longer than τ_{evol} = 0.3 μs, but does not exceed \({T}_{2,{{{{{{{\rm{N}}}}}}}}}^{* ,(1e)}\approx 4.5\) μs. By increasing the rf field amplitude B_{1}, one can achieve τ_{conv} ≈ τ_{evol}. The probability that the electron spin is preserved during its conversion into nuclear spin is ≈ 1 − (τ_{conv}/T_{1,e}). Taking τ_{conv} ≈ 1 μs and T_{1,e} ≈ 90 μs, measured previously^{33} at high magnetic field B_{z} = 7.8 T, we find the fidelity upper bound of ≈ 0.99. Electron spin flips occurring at random times during the conversion result in quadrature nuclear spin polarization values spread over the entire range [−ΔE_{hf,0}/2, + ΔE_{hf,0}/2]. Such additional signal must appear as broad background in the histograms of Fig. 3c, but in practice is too small to be observed directly, explaining why the fidelity derived from Gaussian fitting exceeds the upper bound associated with electron spin flips. On the other hand, the role of the electron spin flips during the conversion stage of the readout can be suppressed if magnetic field is reduced to ≤2 T, where T_{1,e} ≥ 10 ms (ref. ^{33}), and the fidelity would be governed by the statistical noise in the optically detected hyperfine shift ΔE_{hf}, as derived in the previous paragraph.
In summary, the overall fidelity of the electron spin readout is limited by the strongest of the two competing effects. One mechanism is due to the electron spin flips taking place during its conversion into the nuclear spin coherence—this factor is largely related to the quality of the electron spin qubit, and would affect any readout method. The resulting fundamental limit on fidelity \({\approx}1({T}_{2,{{{{{{{\rm{N}}}}}}}}}^{* ,(1e)}/{T}_{1,{{{{{{{\rm{e}}}}}}}}})\) can exceed 0.9999 for QDs at reasonably low fewTesla magnetic fields. The other mechanism is related to the second stage of the readout, where the nuclear spin coherence is retrieved optically. Here, optimized collection of the light emitted by the QD, can improve fidelity beyond the ≳0.998 level demonstrated in this work.
Data availability
The data that support the findings of this study are available from the corresponding authors upon reasonable request.
References
Khaetskii, A. V., Loss, D. & Glazman, L. Electron spin decoherence in quantum dots due to interaction with nuclei. Phys. Rev. Lett. 88, 186802 (2002).
Yao, W., Liu, R.B. & Sham, L. J. Theory of electron spin decoherence by interacting nuclear spins in a quantum dot. Phys. Rev. B 74, 195301 (2006).
Fischer, J. & Loss, D. Dealing with decoherence. Science 324, 1277–1278 (2009).
Heshami, K. et al. Quantum memories: emerging applications and recent advances. J. Mod. Opt. 63, 2005–2028 (2016).
Hensen, B. et al. A silicon quantumdotcoupled nuclear spin qubit. Nat. Nanotechnol. 15, 13–17 (2020).
Gangloff, D. A. et al. Quantum interface of an electron and a nuclear ensemble. Science 364, 62–66 (2019).
Denning, E. V., Gangloff, D. A., Atatüre, M., Mørk, J. & Le Gall, C. Collective quantum memory activated by a driven central spin. Phys. Rev. Lett. 123, 140502 (2019).
Bradley, C. E. et al. A tenqubit solidstate spin register with quantum memory up to one minute. Phys. Rev. X 9, 031045 (2019).
Chekhovich, E. A., Covre da Silva, S. F. & Rastelli, A. Nuclear spin quantum register in an optically active semiconductor quantum dot. Nat. Nanotechnol. 15, 999–1004 (2020).
Ruskuc, A., Wu, C.J., Rochman, J., Choi, J. & Faraon, A. Nuclear spinwave quantum register for a solidstate qubit. Nature 602, 408–413 (2022).
Tyryshkin, A. M. et al. Electron spin coherence exceeding seconds in highpurity silicon. Nat. Mater. 11, 143 (2012).
Saeedi, K. et al. Roomtemperature quantum bit storage exceeding 39 minutes using ionized donors in silicon28. Science 342, 830–833 (2013).
Gurudev Dutt, M. V. et al. Quantum register based on individual electronic and nuclear spin qubits in diamond. Science 316, 1312–1316 (2007).
Urbaszek, B. et al. Nuclear spin physics in quantum dots: An optical investigation. Rev. Mod. Phys. 85, 79–133 (2013).
Press, D. et al. Ultrafast optical spin echo in a single quantum dot. Nat. Photon. 4, 367–370 (2010).
Wüst, G. et al. Role of the electron spin in determining the coherence of the nuclear spins in a quantum dot. Nat. Nanotechnol. 11, 885–889 (2016).
Witzel, W. M. & Das Sarma, S. Quantum theory for electron spin decoherence induced by nuclear spin dynamics in semiconductor quantum computer architectures: spectral diffusion of localized electron spins in the nuclear solidstate environment. Phys. Rev. B 74, 035322 (2006).
Greilich, A. et al. Nucleiinduced frequency focusing of electron spin coherence. Science 317, 1896–1899 (2007).
Barnes, E., Cywiński, L. & Das Sarma, S. Nonperturbative master equation solution of central spin dephasing dynamics. Phys. Rev. Lett. 109, 140403 (2012).
Cai, Z. & Barthel, T. Algebraic versus exponential decoherence in dissipative manyparticle systems. Phys. Rev. Lett. 111, 150403 (2013).
Zeiher, J. et al. Coherent manybody spin dynamics in a longrange interacting Ising chain. Phys. Rev. X 7, 041063 (2017).
Waeber, A. M. et al. Pulse control protocols for preserving coherence in dipolarcoupled nuclear spin baths. Nat. Commun. 10, 3157 (2019).
Hahn, E. L. Spin echoes. Phys. Rev. 80, 580–594 (1950).
Klauder, J. R. & Anderson, P. W. Spectral diffusion decay in spin resonance experiments. Phys. Rev. 125, 912–932 (1962).
Hu, P. & Hartmann, S. R. Theory of spectral diffusion decay using an uncorrelatedsuddenjump model. Phys. Rev. B 9, 1–13 (1974).
Elzerman, J. M. et al. Singleshot readout of an individual electron spin in a quantum dot. Nature 430, 431–435 (2004).
Morello, A. et al. Singleshot readout of an electron spin in silicon. Nature 467, 687–691 (2010).
Arute, F. et al. Quantum supremacy using a programmable superconducting processor. Nature 574, 505–510 (2019).
Dassonneville, R. et al. Fast highfidelity quantum nondemolition qubit readout via a nonperturbative crossKerr coupling. Phys. Rev. X 10, 011045 (2020).
Neumann, P. et al. Singleshot readout of a single nuclear spin. Science 329, 542–544 (2010).
HarveyCollard, P. et al. Highfidelity singleshot readout for a spin qubit via an enhanced latching mechanism. Phys. Rev. X 8, 021046 (2018).
Zhang, Q. et al. Highfidelity singleshot readout of single electron spin in diamond with spintocharge conversion. Nat. Commun. 12, 1529 (2021).
Gillard, G. et al. Fundamental limits of electron and nuclear spin qubit lifetimes in an isolated selfassembled quantum dot. NPJ Quantum Inf. 7, 43 (2021).
Herzog, B. & Hahn, E. L. Transient nuclear induction and double nuclear resonance in solids. Phys. Rev. 103, 148–166 (1956).
Kleinjohann, I. et al. Magnetic field dependence of the electron spin revival amplitude in periodically pulsed quantum dots. Phys. Rev. B 98, 155318 (2018).
Barthel, C., Reilly, D. J., Marcus, C. M., Hanson, M. P. & Gossard, A. C. Rapid singleshot measurement of a singlettriplet qubit. Phys. Rev. Lett. 103, 160503 (2009).
Urdampilleta, M. et al. Gatebased high fidelity spin readout in a CMOS device. Nat. Nanotechnol. 14, 737–741 (2019).
Prechtel, J. H. et al. Electrically tunable hole g factor of an optically active quantum dot for fast spin rotations. Phys. Rev. B 91, 165304 (2015).
Viola, L., Knill, E. & Lloyd, S. Dynamical decoupling of open quantum systems. Phys. Rev. Lett. 82, 2417–2421 (1999).
Carr, H. Y. & Purcell, E. M. Effects of diffusion on free precession in nuclear magnetic resonance experiments. Phys. Rev. 94, 630–638 (1954).
Uhrig, G. S. Keeping a quantum bit alive by optimized πpulse sequences. Phys. Rev. Lett. 98, 100504 (2007).
Reilly, D. J. et al. Exchange control of nuclear spin diffusion in a double quantum dot. Phys. Rev. Lett. 104, 236802 (2010).
Gong, Z.X., Yin, Z.q. & Duan, L.M. Dynamics of the Overhauser field under nuclear spin diffusion in a quantum dot. New J. Phys. 13, 033036 (2011).
Malinowski, F. K. et al. Spectrum of the nuclear environment for GaAs spin qubits. Phys. Rev. Lett. 118, 177702 (2017).
Freeman, R. Selective excitation in highresolution NMR. Chem. Rev. 91, 1397–1412 (1991).
Chekhovich, E. A., Hopkinson, M., Skolnick, M. S. & Tartakovskii, A. I. Suppression of nuclear spin bath fluctuations in selfassembled quantum dots induced by inhomogeneous strain. Nat. Commun. 6, 6348 (2015).
Chekhovich, E. A. et al. Pumping of nuclear spins by optical excitation of spinforbidden transitions in a quantum dot. Phys. Rev. Lett. 104, 066804 (2010).
Puebla, J. et al. Dynamic nuclear polarization in InGaAs/GaAs and GaAs/AlGaAs quantum dots under nonresonant ultralowpower optical excitation. Phys. Rev. B 88, 045306 (2013).
Serrels, K. A. et al. Solid immersion lens applications for nanophotonic devices. J. Nanophotonics 2, 1–29 (2008).
Ostroff, E. D. & Waugh, J. S. Multiple spin echoes and spin locking in solids. Phys. Rev. Lett. 16, 1097–1098 (1966).
Acknowledgements
Royal Society provided funding support to E.A.C. through University Research Fellowship and grant RGF\EA\180117, and to G.G. and E.A.C through grant RG150465. The work was supported by EPSRC grant EP/V048333/1.
Author information
Authors and Affiliations
Contributions
G.G. and E.A.C developed experimental techniques, analyzed the data, and wrote the manuscript. G.G. conducted experiments. E.C. grew the samples. E.A.C. coordinated the project.
Corresponding authors
Ethics declarations
Competing interests
The authors declare no competing interests.
Peer review
Peer review information
Nature Communications thanks the anonymous reviewer(s) for their contribution to the peer review of this work. Peer reviewer reports are available.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary information
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Gillard, G., Clarke, E. & Chekhovich, E.A. Harnessing manybody spin environment for long coherence storage and highfidelity singleshot qubit readout. Nat Commun 13, 4048 (2022). https://doi.org/10.1038/s41467022316184
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41467022316184
This article is cited by

Cavityenhanced singleshot readout of a quantum dot spin within 3 nanoseconds
Nature Communications (2023)

Nuclear spin diffusion in the central spin system of a GaAs/AlGaAs quantum dot
Nature Communications (2023)
Comments
By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.