Direct Observation of Dipolar-Driven Anisotropic Quantum Projection Noise in a Solid-State Spin Ensemble

AI summary

Overview

Research area: Quantum sensing with solid-state spin defects, specifically nitrogen-vacancy (NV) centers in diamond, quantum projection noise, dipolar spin interactions, and the precursor dynamics to spin squeezing.
Technical level: Intermediate. The paper assumes familiarity with NV spin physics, quantum projection noise, dynamical decoupling, and collective spin measurements, but its central question is accessible.
Scope: The paper demonstrates quantum-projection-noise-resolved readout of a dense, strongly interacting two-dimensional NV ensemble and directly observes how intrinsic dipolar interactions shear the projection noise into an anisotropic distribution.

What This Paper Is About

Advanced quantum sensing with NV centers aims to combine readout at the quantum projection noise limit with large numbers of sensing spins. In dense NV ensembles, however, dipolar interactions normally broaden the linewidth and obscure signals, while conventional fluorescence readout remains dominated by photon shot noise. This paper asks whether a dense, strongly interacting NV ensemble can be read out near the projection-noise limit, and whether its own dipolar interactions can then be seen directly in the noise distribution.

Key Contributions

1. QPN-resolved readout in a dense solid-state ensemble. The authors use repetitive readout via the NV’s intrinsic (^{15}\mathrm{N}) nuclear spin at a moderate magnetic field of approximately (0.3\ \mathrm{T}) ((2700\ \mathrm{G})) to resolve the quantum projection noise of a coherent spin state in a two-dimensional NV ensemble.
2. Characterization across four NV densities. They study four regions labeled A, B, C, and D with different NV densities and extract the number of participating spins both from the projection-noise variance and independently from (T_{2}) under XY-8 dynamical decoupling.
3. First direct observation of dipolar-driven anisotropic QPN. The paper reports, for the first time, the interaction-driven evolution of quantum projection noise in a dense NV ensemble, showing that dipolar interactions shear an initially isotropic noise distribution into an anisotropic one.
4. Control experiment isolating the dipolar origin. Repeating the measurement under the DROID-60 decoupling sequence, which SU(2)-symmetrizes the NV–NV interactions, eliminates the anisotropy and confirms that it arises from dipolar twisting dynamics rather than readout artifacts.

Main Findings

• Projection noise is directly resolved. The residual variance (\sigma_{\mathrm{res}}^{2}), obtained after subtracting photon shot noise from the total measured variance, oscillates twice per Rabi cycle as the preparation angle (\theta) is swept from (-\pi) to (+\pi). This twice-per-cycle dependence is characteristic of quantum projection noise and cannot arise from the photon-shot-noise subtraction, which oscillates only once per cycle.
• Readout approaches the QPN limit. Conventional fluorescence readout gives (\sigma_{R} \gg 1) in all four regions. Repetitive readout yields (\sigma_{R}=6.8) for spot A, (2.7) for spot B, (3.1) for spot C, and (38) for spot D, as listed in Table 1. The text describes a near-order-of-magnitude improvement in readout noise for region B and a (\lesssim 3\times) improvement for regions A, C, and D.
• Spin numbers are extracted consistently. Table 1 reports (N=41\pm11) (repetitive-readout fit) and (75\pm23) (XY-8) for A; (48\pm6) and (36\pm12) for B; (83\pm20) and (86\pm27) for C; and (229\pm74) from XY-8 for D, where the QPN fit was not constraining enough. The two estimates agree for A, B, and C.
• Dipolar interactions create anisotropic projection noise. After preparing the ensemble in a coherent spin state along (\hat{x}) and evolving under native dipolar interactions with XY-8 decoupling, the residual noise is initially independent of the measurement angle (\phi). At intermediate times it becomes anisotropic: the noise is suppressed below the coherent-state level along one quadrature and amplified along the orthogonal quadrature. The anisotropy grows monotonically with dipolar evolution time.
• The anisotropy is not a readout artifact. The repetitive-readout noise depends only on spin populations and hence on the measured collective spin (S_{\phi}^{\mathrm{tot}}), which is independent of (\phi). The observed (\phi)-dependence therefore reflects shearing of the underlying quantum projection noise.
• A control sequence removes the effect. Under DROID-60, which decouples the spin bath and SU(2)-symmetrizes the NV–NV interactions, the initial state becomes an eigenstate of the evolution and the dipolar dynamics are arrested. The anisotropy in (\sigma_{\mathrm{res}}^{2}) completely disappears and the noise remains isotropic at all measured times.
• Higher density produces faster twisting. Spot C has a stronger average NV–NV interaction than spot B, and the separation between the residual noise at (\phi\approx\pi/4) and (3\pi/4) grows substantially faster for spot C than for spot B.
• The interaction model is two-dimensional. The NVs sit in a delta-doped layer of thickness (\lesssim 7\ \mathrm{nm}), and the selected ([111])-oriented NV subgroup has its quantization axis perpendicular to the plane. The dipolar Hamiltonian is (H=-\sum_{i,j}\frac{J_{0}}{r_{ij}^{3}}\left(S_{x}^{i}S_{x}^{j}+S_{y}^{i}S_{y}^{j}-S_{z}^{i}S_{z}^{j}\right)), with (J_{0}=52\ \mathrm{MHz}\cdot\mathrm{nm}^{3}).

Methodology in Plain English

The sample is a ((111))-oriented, isotopically purified (^{12}\mathrm{C}) diamond epilayer with a (^{15}\mathrm{N}) delta-doped layer confining the NVs to a sheet of thickness (\lesssim 7\ \mathrm{nm}). Electron-beam irradiation and annealing create 18 circular high-NV-density regions with diameters of (2)–(12\ \mu\mathrm{m}); the paper uses four of them, A–D. Measurements are performed in a room-temperature scanning confocal microscope with (532\ \mathrm{nm}) laser excitation, microwave control of the electronic spin, and radio-frequency control of the (^{15}\mathrm{N}) nuclear spin.

The key readout trick is repetitive readout. The electronic spin state is first swapped onto the (^{15}\mathrm{N}) nuclear spin, which acts as a quantum memory. A hyperfine-selective (\pi)-pulse implements a (\mathrm{C}{n}\mathrm{NOT}{e}) gate that maps the nuclear population back onto the electron for fluorescence readout. Because the nuclear spin is left nearly unperturbed, the cycle can be repeated (k) times; the paper states (k\gtrsim100) repetitions in spots B and C and (k\gtrsim50) in spots A and D, while Table 1 lists (k_{\mathrm{max}}) values of (17) (A), (1.8\times10^{2}) (B), (45) (C), and (10) (D). A permanent magnet supplies (2700\ \mathrm{G}), extending the nuclear-spin lifetime, and the temperature is stabilized to within (\pm0.01\ \mathrm{K}) to keep the magnetic field and gate frequencies stable.

To resolve projection noise, the authors prepare each NV in ((\cos(\theta/2)|0\rangle_{e}+\sin(\theta/2)|-1\rangle_{e})|-1/2\rangle_{n}), apply repetitive readout, and record a weighted photon-count signal. They measure the total variance, calculate the Poisson photon-shot-noise contribution from the measured mean photon count, and subtract it to obtain the residual variance. A classical Monte Carlo simulation of the full readout sequence incorporates the measured hyperfine linewidth, electronic-spin repolarization, and nuclear-spin relaxation; the number of spins (N) is the only free parameter.

For the dipolar dynamics, the ensemble is prepared in (|+\rangle_{e}^{\otimes N}) along (\hat{x}) and allowed to evolve under the native dipolar interaction while XY-8 decoupling suppresses extrinsic noise. The authors measure the variance of the collective spin (S_{\phi}^{\mathrm{tot}}=S_{z}^{\mathrm{tot}}\cos\phi+S_{y}^{\mathrm{tot}}\sin\phi) as a function of (\phi), normalize by the average at (\phi=0) and (\pi), and compare with the same experiment under DROID-60.

Why This Matters

This work reaches a regime that is central to quantum-enhanced sensing in the solid state: dense, strongly interacting spins with readout close to the projection-noise limit. The paper states that its results open the door to direct measurements of spin squeezing and entanglement-enhanced sensing in the solid state. More broadly, QPN-resolved readout gives access to fluctuations and correlations that conventional fluorescence measurements obscure.

Potential applications, as framed by the paper’s introduction and results:

• Nanoscale magnetic and electric field imaging in condensed-matter systems, where NV centers are already used as probes.
• Intracellular thermometry in living cells, an existing NV sensing application.
• High-sensitivity magnetometry and quantum sensing with dense NV ensembles, where sensitivity improves with the number of spins but is limited by readout noise and dipolar broadening.
• Studies of strongly correlated and non-Gaussian quantum states, including higher-order correlations and nonequilibrium many-body dynamics in solid-state spin ensembles.

For industry, the work is relevant to diamond-based quantum sensors, magnetometry instrumentation, bio-sensing, and materials characterization. The use of a compact permanent magnet at (2700\ \mathrm{G}), rather than a tesla-scale superconducting magnet, is presented as keeping the protocol compatible with practical sensing platforms. The paper does not report a direct end-to-end sensitivity measurement; its reported metrics are readout noise (\sigma_{R}), extracted spin number (N), and the anisotropic projection-noise dynamics.

Future Directions

• Observe metrologically useful spin squeezing. The anisotropic reshaping of the QPN is described as the immediate precursor to spin squeezing. Combining repetitive readout with techniques that suppress positional disorder could enable direct observation of squeezing in dense NV ensembles.
• Study higher-order correlations and non-Gaussian states. QPN-resolved readout could provide access to fluctuations and correlations hidden by conventional fluorescence, enabling studies of higher-order correlations, non-Gaussian states, and nonequilibrium many-body dynamics.
• Push readout closer to (\sigma_{R}\to1). The current results reach (\sigma_{R}=2.7)

Authors’ abstract

The nitrogen-vacancy (NV) center in diamond is a prominent quantum-sensing platform. Combining readout at the quantum projection noise limit with strong dipolar interactions promises substantial gains in sensitivity. However, experimentally accessing this regime has remained a longstanding challenge. In this work, we demonstrate quantum-projection-noise-resolved readout of a strongly-interacting, two-dimensional ensemble of NV centers. Our approach leverages repetitive readout via the NV's intrinsic $^{15}$N nuclear memory at a moderate magnetic field ($\sim 0.3$ T), improving the readout fidelities by nearly an order of magnitude. This enables us to directly resolve the quantum projection noise of a coherent spin state and to watch the ensemble's intrinsic dipolar interactions shear this noise into an anisotropic profile. Our results open the door to direct measurements of spin squeezing and entanglement-enhanced sensing in the solid state.