High-fidelity single-shot readout of single electron spin in diamond with spin-to-charge conversion

The study addresses the challenge of high-fidelity single-shot readout of solid-state spins, specifically NV center electron spins in diamond, which is vital for fault-tolerant quantum computing and scalable quantum networks. Conventional resonance fluorescence readout at cryogenic temperatures suffers from spin non-conserving processes that limit the optical readout window and induce spin-flip errors, making it difficult to surpass fault-tolerant thresholds. Prior approaches use optical microstructures and nanocavities to enhance photon collection and emission rates, but these are technically challenging and can introduce strain and defects that degrade spin and optical properties. The authors propose a method that converts the spin state into a charge state using spin-selective photoionization aided by NIR light before significant spin-flip relaxation occurs, enabling robust, high-fidelity readout via stable charge detection. The key objective is to increase the ratio of ionization rate to spin-flip rate to improve single-shot readout fidelity without complex nanophotonic engineering.

The paper situates the work within efforts to achieve single-shot spin readout in solid-state systems via resonance fluorescence, applied to platforms such as quantum dots, rare-earth ions, silicon-vacancy centers, and NV centers. Enhancements using solid immersion lenses and nanocavities can boost collection efficiency and emission rates, with notable demonstrations in nanophotonic cavities. However, implementing high-quality cavities involves complex engineering and can introduce strain and surface defects that compromise performance. Previous NV center spin-to-charge conversion schemes typically used non-resonant excitation to improve readout efficiency. Additionally, literature establishes fault-tolerant thresholds for quantum computing and networking, highlights successes in quantum teleportation and entanglement distribution using NVs, and explores photoelectric detection approaches. The present work builds on these by introducing a cryogenic, resonantly excited, NIR-assisted SCC that targets suppression of spin-flip-induced errors while avoiding stringent optical microcavity requirements.

Experiments are conducted on a single NV center in bulk diamond at 8 K within a solid immersion lens. The NV exhibits non-axial strain δ = 5.9 GHz, leading to a measured excited-state spin-flip rate Γ_flip = 0.75 ± 0.02 MHz. The scheme exploits spin-selective optical transitions: E_γ drives the m_s = 0 cycling transition and E_1,2 addresses m_s = ±1. Under selective E_γ excitation, the |0⟩ ground state is promoted to the excited state and then photoionized to NV^0 by a 1064 nm NIR laser before spin-flip relaxation can occur. In contrast, |±1⟩ remains unexcited and stays NV^−. The spin state is thus mapped to a charge state (SCC), which is read out with near-unity fidelity because NV^− and NV^0 exhibit distinct fluorescence under simultaneous E_γ and E_1,2 excitation. Charge-state readout fidelity is characterized via two consecutive 500 µs charge readouts following a 3 µs 532 nm reset pulse and post-selection. Photon-count histograms and correlations distinguish NV^− from NV^0 using a threshold (≤11 photons indicates NV^0). The charge lifetime under continuous E_γ + E_1,2 excitation is measured to quantify readout-induced charge conversion. Ionization dynamics are probed by initializing NV to NV^− with a 532 nm pulse and charge post-selection, preparing the spin to |0⟩ with a 20 µs E_1,2 pulse, applying a simultaneous E_γ + NIR pulse for a variable duration to induce SCC, and then reading out the charge. NV^− population decay as a function of NIR power is recorded. A rate-equation model incorporating the relevant energy levels and processes (detailed in Supplementary Information) is used to fit the decay curves with ionization rate Γ_ion as the free parameter, revealing Γ_ion's dependence on NIR power. To mitigate leakage due to spin-flip during SCC, an auxiliary correction uses the m_s = −1 (AUX) level: after each short SCC pulse, a microwave pulse (MW_AUX) transfers population from AUX back to |0⟩, and the SCC-AUX cycle is repeated n times to increase the |0⟩→NV^0 conversion without affecting |1⟩. The SCC duration equals 2 µs × n and is optimized for maximal average readout fidelity. The SCC-based single-shot readout is compared to resonance fluorescence readout under the same conditions.

- Non-demolition charge-state readout fidelity: 99.96 ± 0.02% overall; specifically, unity for NV^0 and 99.92 ± 0.03% for NV^−. The NV^− charge lifetime under E_γ + E_1,2 illumination is 400.7 ± 9.7 ms, implying a ~0.12% charge-conversion error during readout, consistent with observed imperfection. - NIR-induced ionization from the E_γ-excited state is approximately single-photon at 1064 nm with Γ_ion proportional to NIR power: coefficient 67.0 ± 6.7 kHz/mW at 8 K. Maximum extracted Γ_ion = 2.79 ± 0.08 MHz. This is lower than earlier room-temperature estimates (~1.2 ± 0.33 MHz/mW), suggesting temperature or setup dependence. - The NV used has high strain (δ = 5.9 GHz), yielding a fast spin-flip rate Γ_flip = 0.75 ± 0.02 MHz. - Baseline SCC (without AUX correction) yields average single-shot spin readout fidelity of 89.1 ± 0.2% under current Γ_ion and Γ_flip. - Implementing the AUX correction improves SCC conversion of |0⟩ to NV^0 and achieves an average single-shot spin readout fidelity of 95.4 ± 0.2% at ~10 µs SCC duration. - Resonance fluorescence readout under the same high-spin-flip conditions achieves an optimal average fidelity of 79.6 ± 0.8%, substantially lower than SCC with AUX correction. - Modeling shows readout fidelity depends critically on the ratio Γ_ion/Γ_flip. For Γ_flip ≈ 0.2 MHz (low-strain NV), reaching Γ_ion with modest >1 W NIR power on diamond could enable overall fidelities >99.9%, surpassing fault-tolerant thresholds.

By converting the spin state to a stable charge state before significant excited-state spin-flip occurs, the NIR-assisted SCC directly addresses the core limitation of resonance fluorescence readout—the spin-flip-induced interruption of optical cycling. The high-fidelity, non-demolition charge readout enables reliable discrimination between NV^− and NV^0, ensuring that once SCC occurs, the spin information is preserved. The observed 95.4% single-shot spin readout fidelity in a high-strain NV (Γ_flip ≈ 0.75 MHz) demonstrates the robustness of SCC even under unfavorable conditions where resonance fluorescence performance degrades (79.6%). The dependence of fidelity on Γ_ion/Γ_flip highlights clear engineering targets: increasing NIR power and improving optical delivery to boost Γ_ion, while selecting low-strain NVs or optimizing environment to minimize Γ_flip. The approach reduces reliance on challenging nanocavity fabrication and alignment, offering a more accessible route to high-fidelity readout. Projections indicate that with higher NIR power and/or lower spin-flip rates, fidelities beyond 99.9% are achievable, meeting fault-tolerance requirements for quantum computation and networking.

The work introduces and experimentally validates an NIR-assisted spin-to-charge conversion scheme for single-shot readout of NV center electron spins at cryogenic temperatures. Leveraging spin-selective resonant excitation and NIR photoionization, the method suppresses spin-flip errors and achieves 95.4% average single-shot fidelity in a high-strain NV. The charge-state readout itself reaches 99.96% fidelity, and the ionization rate scales linearly with NIR power. Compared to resonance fluorescence, SCC provides superior performance without demanding nanophotonic cavities. Modeling suggests that with increased NIR power and/or lower spin-flip rates, the method can exceed the 99.9% fault-tolerant threshold. Future directions include improving NIR power delivery and optics to raise Γ_ion, applying SCC to integrated optoelectronic devices (e.g., coupling to single-electron transistors for electrical readout), leveraging SCC for high-efficiency quantum sensing, and exploiting compatibility of NIR with biological samples. Although SCC is demolition for the electron spin, projective readout of weakly coupled nuclear spins remains feasible.

Key limitations include: (1) Current continuous-wave NIR laser output power and optical delivery (transmission losses and chromatic aberration) limit the achievable ionization rate Γ_ion, which in this work is only ~3.7× the fast spin-flip rate, constraining fidelity. (2) The studied NV exhibits high strain, yielding a large Γ_flip (0.75 MHz), which is an unfavorable regime for resonance fluorescence and stresses SCC performance; lower-strain NVs would perform better. (3) The measured ionization coefficient at 8 K (67.0 ± 6.7 kHz/mW) is much smaller than room-temperature estimates, and the underlying mechanisms require further study. (4) SCC is a demolition readout for the electron spin, precluding reuse of the same electron spin state post-measurement. (5) Modeling and optimization rely on assumptions about level structure and rates; deviations or sample-to-sample variability may impact generalizability.