Iterative Hartmann–Hahn Protocol

• The Iterative Hartmann–Hahn Protocol is a repeated double-spin-lock method that uses optically addressable NV centers to polarize and probe dark P1 spin ensembles in diamond.
• It employs simultaneous microwave driving to create matched dressed-state splittings, facilitating resonant exchange despite significant energy mismatches between NV and P1 spins.
• The protocol's iterative sequence builds up polarization cycle by cycle, offering insights into coherent driving, relaxation, spin diffusion, and the impact of local disorder.

The iterative Hartmann–Hahn protocol is a repeated double-spin-lock polarization-transfer method in which an optically addressable spin ensemble is used as a polarization source and probe for a nominally dark spin ensemble. In the reported diamond implementation, optical polarization is transferred from dense nitrogen-vacancy (NV) centers to substitutional nitrogen (P1) centers by repeatedly reinitializing the NVs with laser light and reapplying a Hartmann–Hahn contact, so that polarization accumulates cycle by cycle in the P1 bath until saturation is reached. In this setting, the protocol is both a control primitive and a diagnostic: it produces a coherent mesoscopic P1 spin state, yields a 740-fold enhancement over room-temperature thermal equilibrium as revealed by differential readout, and exposes the interplay of coherent driving, dipolar exchange, relaxation, spin diffusion, and local disorder in a disordered electron-spin network (Yoon et al., 19 Feb 2026).

1. Physical system and the need for Hartmann–Hahn matching

The experimental platform is a chemical-vapor-deposition diamond with a $40~\mu\mathrm m$ thick $99.9\%$ $^{12}\mathrm C$-enriched layer, so $^{13}\mathrm C$ nuclear-spin noise is strongly suppressed and the relevant many-body dynamics are dominated by electron-spin dipolar interactions between NV centers and P1 centers. The NV centers are the bright spins: they can be optically initialized and read out through spin-dependent fluorescence, and the NV is treated as an effective two-level system in the $\{|0\rangle, |-1\rangle\}$ manifold of the spin-1 ground state. The P1 centers are dark electron spins: they are ESR-addressable but not optically read out.

At the main field $B=446$ G aligned along the $[111]$ axis, the NV and P1 spin transitions are far apart in the laboratory frame. The supplement states that the NV–P1 frequency mismatch is about

$$|\omega_{\mathrm{NV}}-\omega_{\mathrm{P1}}| \sim 370~\mathrm{MHz},$$

which suppresses direct flip-flops. In the secular rotating-frame description, same-species spins with matching resonance frequencies retain exchange terms, but heterospecies NV–P1 pairs reduce to an Ising coupling,

$$H_{\mathrm{inter}}^{ij} = -\frac{J_0}{r_{ij}^3}(3\cos^2\theta_{ij}-1) S_i^z S_j^z,$$

with

$$J_0=\frac{\mu_0 \gamma_e^2\hbar}{4\pi}\approx 52~\mathrm{MHz\cdot nm^3}.$$

Without additional driving, the P1 bath therefore shifts NV frequencies and causes dephasing, but does not exchange polarization efficiently with the NVs. The protocol is designed precisely to overcome this large bare energy mismatch and convert an otherwise static heterospecies interaction into a usable transfer channel (Yoon et al., 19 Feb 2026).

2. Dressed-state mechanism

The key theoretical ingredient is the dressed-frame Hartmann–Hahn mechanism. Under simultaneous resonant microwave driving of NV and P1 spins,

$99.9\%$0

the drive defines the quantization axis along $99.9\%$1 provided $99.9\%$2. In this spin-lock frame, the NV–P1 Ising interaction acquires oscillating terms at the difference of the Rabi frequencies. Under the Hartmann–Hahn condition,

$99.9\%$3

those terms become stationary, and the inter-species coupling becomes

$99.9\%$4

where

$99.9\%$5

This is the operative meaning of Hartmann–Hahn matching in the heterospecies NV–P1 system. A recurring misconception is to treat matching as equality of bare resonance frequencies. In the reported implementation, the two species remain on distinct ESR transitions at different carrier frequencies; what must be matched are the dressed splittings, implemented experimentally as matched Rabi frequencies. Continuous driving thus turns a static heterospecies $99.9\%$6 coupling into resonant exchange in the rotating frame, enabling coherent or semiclassical polarization transfer between bright NV spins and dark P1 spins despite the large laboratory-frame mismatch (Yoon et al., 19 Feb 2026).

3. Iterative sequence and experimental realization

The protocol is iterative because a single Hartmann–Hahn contact does not fully polarize the P1 bath. Instead, the experiment repeatedly restores the NV polarization reservoir and re-establishes the dressed-frame exchange contact. In the implementation reconstructed from the experimental and simulation sections, the sequence is: optical initialization of the NV ensemble with a green laser pulse; simultaneous continuous-wave driving of the NV and a selected P1 subgroup with calibrated amplitudes satisfying $99.9\%$7; a spin-lock interaction interval during which polarization leaves the optically pumped NVs and is transferred into nearby P1 spins; and optical NV reinitialization while the P1 ensemble partially retains the acquired polarization. The protocol thereby differs from a single-shot transfer, for which the NV reservoir is depleted after one contact and transfer stalls.

The apparatus uses two independent microwave channels, one for NV and one for P1, generated by separate vector signal generators and shaped by an AWG. The microwaves are combined, amplified, and delivered through a coplanar waveguide. The laser is $99.9\%$8 nm. The initialization uses $99.9\%$9 for initial NV polarization and fluorescence readout, and $^{12}\mathrm C$0 during polarization-transfer cycles for faster NV re-initialization. In the iterative-transfer simulation, each interaction phase lasts $^{12}\mathrm C$1, the repumping phase also lasts $^{12}\mathrm C$2, and the total cycle time is $^{12}\mathrm C$3; the process is repeated up to $^{12}\mathrm C$4 cycles. The addressed P1 subgroup is one of the resolved hyperfine/Jahn–Teller lines in the P1 ESR spectrum; the supplement notes five resolvable dips with population fractions $^{12}\mathrm C$5, and the simulation uses an addressed subgroup concentration of about $^{12}\mathrm C$6 ppm, corresponding to $^{12}\mathrm C$7. Absolute concentrations extracted from DEER and simulation are

$^{12}\mathrm C$8

with $^{12}\mathrm C$9 (Yoon et al., 19 Feb 2026).

4. Polarization buildup, saturation, and readout

The buildup dynamics are governed by the competition between transfer and loss. During the Hartmann–Hahn dark phase, P1 relaxation is slow,

$^{13}\mathrm C$0

so polarization can accumulate. During the laser-on reinitialization phase, however, the effective P1 relaxation becomes much faster,

$^{13}\mathrm C$1

which is attributed to laser-induced P1 depolarization, plausibly through photo-ionization and charge-state dynamics. The initial NV polarization used in the simulation is

$^{13}\mathrm C$2

The buildup is fit phenomenologically by an exponential approach to saturation,

$^{13}\mathrm C$3

where $^{13}\mathrm C$4 is the number of transfer cycles; in the simulation reproducing the experiment, $^{13}\mathrm C$5.

The saturation polarization depends strongly on the spin-lock drive amplitude $^{13}\mathrm C$6, with the supplement fitting

$^{13}\mathrm C$7

This captures the crossover from a disorder-limited low-drive regime to a coherent-transfer regime at larger $^{13}\mathrm C$8. The simulation yields $^{13}\mathrm C$9, whereas the experimentally estimated P1 polarization is $\{|0\rangle, |-1\rangle\}$0. The authors explicitly note that the model overestimates the experiment by a factor of $\{|0\rangle, |-1\rangle\}$1–$\{|0\rangle, |-1\rangle\}$2, and discuss possible reasons: leakage into unaddressed spin subgroups, incomplete quasi-equilibrium assumptions used in the polarization estimate, neglected quantum correlations, and oversimplified modeling of laser-induced depolarization. Experimentally, the P1 polarization is extracted from differential Hartmann–Hahn readout using parallel and antiparallel initial configurations of NV and P1 spin-lock polarization, leading to

$\{|0\rangle, |-1\rangle\}$3

This estimate assumes linear contrast mapping, conservation of rotating-frame polarization, and full quasi-equilibration between the addressed NV and P1 groups. The resulting mesoscopic P1 spin ensemble exhibits collective Rabi oscillations and long-lived spin-lock and Hahn-echo coherences (Yoon et al., 19 Feb 2026).

5. Disorder, diffusion, and network-level description

The transfer dynamics are further modeled on a disordered spin network using a rate-equation description derived from Fermi’s Golden Rule. Local disorder appears as a detuning $\{|0\rangle, |-1\rangle\}$4 for each spin from the drive frequency, so the local effective Rabi frequency is

$\{|0\rangle, |-1\rangle\}$5

and the local tilt angle satisfies

$\{|0\rangle, |-1\rangle\}$6

Only the transverse component in the dressed frame contributes to exchange, so the effective coupling is reduced by $\{|0\rangle, |-1\rangle\}$7:

$\{|0\rangle, |-1\rangle\}$8

The pairwise flip-flop rate is then

$\{|0\rangle, |-1\rangle\}$9

where

$B=446$0

and the Hartmann–Hahn resonance half-width is taken from experiment as

$B=446$1

The full polarization dynamics are evolved via

$B=446$2

This formalism makes the role of disorder explicit. Even if the nominal Rabi frequencies of the two microwave drives are matched globally, local detunings shift the effective dressed splittings and reduce spectral overlap, so not all NV–P1 pairs transfer polarization equally well. The disorder scale used in the simulation is

$B=446$3

taken as the standard deviation of a Gaussian distribution for the local detunings $B=446$4. The measured crossover of $B=446$5 therefore acts as a probe of intrinsic disorder in the dark-spin ensemble: by observing at what $B=446$6 the polarization begins to saturate, one infers the characteristic local energy mismatch inside the ensemble. Spin diffusion is an essential companion process. At the experimental Rabi frequency

$B=446$7

the extrapolated infinite-size spin diffusion coefficient is

$B=446$8

Polarization is thus injected locally into P1 spins near an NV by the Hartmann–Hahn step and then redistributed through the dark ensemble by intra-P1 flip-flops (Yoon et al., 19 Feb 2026).

6. Stability, limitations, and broader implications

A separate theoretical analysis of Hartmann–Hahn double resonance shows that matching alone does not guarantee dynamically benign operation. In a minimal problem with an undriven spin $B=446$9 coupled to a driven ancilla spin $[111]$0, the relevant matching condition is

$[111]$1

and the ancilla modifies the target-spin damping according to

$[111]$2

Red-detuned driving $[111]$3 gives a positive contribution to the relaxation rate, whereas blue-detuned driving $[111]$4 can give a negative contribution and produce instability or self-excited oscillation when the threshold condition $[111]$5 is reached. In the regime $[111]$6, $[111]$7, the threshold can be summarized by the cooperativity parameter

$[111]$8

The theory is mean-field, reduces the problem to two effective two-level systems, neglects spin-spin correlations and entanglement, and does not include many-body effects, disorder, spectral diffusion, or ensemble inhomogeneity. It therefore does not directly model the disordered mesoscopic NV–P1 network, but it establishes that ancilla-induced dynamical back-action is an intrinsic consideration near Hartmann–Hahn matching (Levi et al., 2020).

For iterative operation, this suggests that the sign of detuning can matter in addition to matching. A plausible implication is that repeated contacts on the red-detuned side of Hartmann–Hahn matching should be more stable and reproducible, whereas blue-detuned near-threshold operation may amplify transverse motion and distort polarization transfer. In the NV–P1 experiment, the iterative protocol is significant because it does more than demonstrate one-time polarization transfer. By repeatedly applying Hartmann–Hahn contacts with bright-spin reset, it actively polarizes a disordered dark ensemble and then uses the resulting response to characterize coherence, transfer efficiency, diffusion, and intrinsic disorder. This suggests utility in quantum sensing, where a polarized dark-spin environment may become a resource rather than a noise source, and in quantum many-body simulation, where the protocol provides controlled access to driven, disordered dipolar dynamics in a mesoscopic ensemble (Yoon et al., 19 Feb 2026).