States with long coherence are a crucial requirement for qubits and quantum memories. Nuclear spins in epitaxial GaAs/AlGaAs quantum dots are a great candidate, offering excellent isolation from external environments and on-demand coupling to optical flying qubits. However, coherence times are limited to ≲1 ms by the dipole-dipole interactions between the nuclei and by the nuclear quadrupolar coupling to inhomogeneous crystal strain. Here, we combine strain engineering of the nuclear spin ensemble and tailored dynamical decoupling sequences to achieve nuclear spin coherence times exceeding 100 ms. Recently, a reversible transfer of quantum information into nuclear spin ensembles has been demonstrated in quantum dots: our results provide a path to develop this concept into a functioning solid-state quantum memory suitable for quantum repeaters in optical quantum communication networks.
Figures
13 panels with data across 5 figures. Each panel page shows the plot, its columns and its files; each data.csv begins with a header naming the paper, the panel, the source, the license and the provenance route.
Fig. 1
panel (a): NMR spectrum of the spin-3/2 $^{75}$As nuclei in an uncharged (0e) quantum dot: NMR signal $\Delta E_{\mathrm{hf}}$ ($\mu$eV) against radio frequency offset (kHz), satellite transitions in black and central transition in blue. The dashed Rf pulse spectral profiles on the right-hand log scale and the insets are not shown. data.csv
Fig. 2
panel (a): Photoluminescence spectra of a neutral exciton in one quantum dot, PL intensity (arb. units) against PL photon energy (eV), after optical pumping with $\sigma^{+}$ (red) and $\sigma^{-}$ (blue) polarized light. The hosted range, 1.5995 to 1.6060 eV, opens on the printed window; photon energies are given to $10^{-5}$ eV. data.csv
panel (b): Rabi oscillations of the nuclear spins in a neutral (0e) quantum dot: NMR signal $\Delta E_{\mathrm{hf}}$ ($\mu$eV) against the duration $T_{\mathrm{Rf}}$ ($\mu$s) of a constant-amplitude Rf pulse. data.csv
panel (c): Measured nuclear spin decoherence in a neutral (0e) quantum dot: NMR signal $\Delta E_{\mathrm{hf}}$ ($\mu$eV) against free evolution time $T_{\mathrm{FreeEvol}}$ (ms) under free induction decay (FID), Hahn echo, and $n_{\mathrm{Cycles}} = 1$, 12, 32 and 48 cycles of CHASE-40 dynamical decoupling, as the legend names them. data.csv
panel (d): Numerical model of nuclear spin decoherence for a homogeneous ensemble of $N = 12$ spins: normalized echo amplitude against free evolution time $T_{\mathrm{FreeEvol}}$ under a Hahn echo and $n_{\mathrm{Cycles}} = 1$, 12, 32 and 48 cycles of CHASE-40 decoupling, $T_{\mathrm{Rf}} = 20\,\mu$s. Time is in seconds, where the print uses ms. data.csv
Fig. 3
panel (a): Measured decoherence of the $I_z = \pm 1/2$ subspace of the $^{75}$As nuclei in a neutral (0e) quantum dot under dynamical decoupling with $T_{\mathrm{Rf}} = 20\,\mu$s: echo amplitude ($\mu$eV; the print normalizes it) as colour against total evolution time $T_{\mathrm{EvolTot}}$ (s) and CHASE-40 cycle time $T_{\mathrm{Cycle}}$ (s). Points are squares; the print fills the plane between them. data.csv
panel (b): Measured decoherence of the $I_z = (-3/2, -1/2)$ subspace of the $^{75}$As nuclei under dynamical decoupling with $T_{\mathrm{Rf}} = 20\,\mu$s: echo amplitude ($\mu$eV; the print normalizes it) as colour against total evolution time $T_{\mathrm{EvolTot}}$ (s) and CHASE-40 cycle time $T_{\mathrm{Cycle}}$ (s). Points are squares; the print fills the plane between them. data.csv
panel (c): Numerical model of decoherence under dynamical decoupling for a homogeneous ensemble of $N = 12$ nuclei: normalized echo amplitude as colour against total evolution time $T_{\mathrm{EvolTot}}$ (s) and CHASE-40 cycle time $T_{\mathrm{Cycle}}$ (s). Points are squares; the print fills the plane between them. data.csv
panel (d): Numerical model of decoherence for $N = 12$ nuclei with inhomogeneous spectral broadening: normalized echo amplitude as colour against total evolution time $T_{\mathrm{EvolTot}}$ (s) and CHASE-40 cycle time $T_{\mathrm{Cycle}}$ (s). Points are squares; the print fills the plane between them. data.csv
panel (e): Nuclear spin memory time $T_{\mathrm{M}}$ (s) against CHASE-40 cycle time $T_{\mathrm{Cycle}}$ (s) for the central transition (CT, blue) and satellite transition (ST, red): fits to the experiment (solid), fits to the numerical model (dotted) and the first-principle $\sqrt{2/M_2}$ estimate (dashed). The print draws the ST lines double. data.csv
Fig. 4
panel (a): Nuclear spin decoherence under 4 cycles of CHASE-40 ($T_{\mathrm{Rf}} = 20\,\mu$s): QD NMR signal $\Delta E_{\mathrm{hf}}$ ($\mu$eV) against free evolution time $T_{\mathrm{FreeEvol}}$ (ms) for initialization Rf pulse phases $\phi = -\pi/4$, 0, $\pi/4$ and $\pi/2$, and without the initialization pulse ($\theta = 0$, open diamonds). data.csv
panel (b): Nuclear spin decay times (ms) fitted to the decays in (a), with 95% confidence intervals: $T_2$ for $\theta = \pi/2$ with $\phi = -\pi/4$, 0, $\pi/4$, $\pi/2$ (positions 0 to 3) and $T_1$ for $\theta = 0$ (position 4). The Bloch-sphere sketches are not shown. data.csv
Fig. 5
panel (1): Numerical model, homogeneous ensemble of $N = 12$ nuclei: overlap fidelity $|\langle \psi_{\mathrm{Init}} | \psi_{\mathrm{Fin}} \rangle|^2$ against free evolution time $T_{\mathrm{FreeEvol}}$ under free induction decay (squares) and 4 cycles of CHASE-40 (triangles), for full transverse polarization (filled) and a spin wave superposition (open). Time is in seconds, where the print uses ms. data.csv
Cite
Harry E. Dyte, Santanu Manna, Saimon F. Covre da Silva, Armando Rastelli, Evgeny A. Chekhovich. Storing quantum coherence in a quantum dot nuclear spin ensemble for over 100 milliseconds. Nat. Commun. 17, 239 (2025). https://doi.org/10.1038/s41467-025-66948-6
When you use hosted data, cite the original paper and give the panel's URL so a reader can find the exact values you used.