Skip to content

Observation of period doubling and higher multiplicities in a driven single-spin system

FieldValue
JournalPhysical Review Research
Published2026-06-04
AuthorsDhruv Deshmukh, Raúl González, Roberto Sailer, Fedor Jelezko, Ressa S. Said
DOI10.1103/wj1t-vk39
Citations6

A single nitrogen-vacancy center in diamond exhibited period doubling and higher multiplicities through (k = 5) under strong RF driving. The finest scan located period doubling at a generator amplitude of (35.844 \pm 0.018) mV with a mean transition spacing of 9.21 MHz, while a second scan resolved all four multiplicities at 7.88 MHz.

The authors modeled the driven spin as a two-level system with level spacing (\Delta_0) and drive frequency (\omega_d). At resonance, (\hbar\omega_d = \Delta_0), period (k)-tupling occurs when (\Delta\epsilon/(\hbar\omega_d) = j/k) for integers (j < k). A general state then returns after (k) drive periods, giving stroboscopic observables a period (kT_d).

For (k = 2), the state rotates by (\pi) about the Floquet mode axis after each (T_d). The minimal model predicts (A_{Pk} \approx \Delta_0/k) and gives (A_{P2} \approx 0.5042\Delta_0) at resonance. A small amplitude mismatch produces a slow modulation with period (\tau \propto 1/|\delta A|). This modulation period diverges at the exact response manifold, so its growth provides a sensitive indicator of period doubling.

The sample contained one optically resolved (^{15}\mathrm{N}) NV center. It used a 2 mm diameter type IIa (100) single-crystal substrate polished into a hemisphere. A plasma-enhanced CVD layer added about 100 nm of isotopically purified 99.999% (^{12}\mathrm{C}). The hemispherical solid immersion geometry increased collection efficiency by a factor of 4.4.

Shallow defects were created by (^{15}\mathrm{N}^+) implantation at (10^9\ \mathrm{cm}^{-2}) and 5 keV. The sample was then annealed at 1000 °C for 3 h. A static field near 1020 G aligned with the NV axis placed the experiment near the ground-state level anticrossing.

The main period-doubling dataset used a mean driven spacing of 9.21 MHz. The higher-multiplicity dataset used 7.88 MHz. The measured (T_2^) exceeded (100T_d), while (T_1) was about 5 ms and greatly exceeded (T_2^).

Each trace began with 3 μs of optical pumping at 532 nm. The spin was then driven by rectangular RF pulses of length (nT_d), with photoluminescence recorded after each pulse up to (n = 200). Drive frequency was matched to the measured transition before each amplitude scan.

The arbitrary waveform generator provided 0.02 ns timing resolution and 10-bit voltage digitization. Its stated voltage resolution was about 0.5 mV at 500 mV peak-to-peak.

The authors applied a discrete Fourier transform to locate each dominant response. They then split the trace into (k) phase-shifted subsequences. Each subsequence received a damped sinusoidal fit, and the modulation periods were averaged with inverse-variance weighting. A weighted hyperbolic fit, (\tau = C_k/|A-A_{Pk}|), extracted each response amplitude and its uncertainty.

Dataset-2 used a coarse amplitude scan and a mean transition spacing of 7.88 MHz. Its averaged fits gave the following generator amplitudes.

(k)Response(A_{Pk}) (mV)(C_k/T_d) (mV)
2Period doubling(33.294 \pm 0.010)(82.429 \pm 3.104)
3Period tripling(20.870 \pm 0.036)(67.167 \pm 1.622)
4Period quadrupling(15.408 \pm 0.035)(61.740 \pm 2.117)
5Period quintupling(12.249 \pm 0.043)(59.371 \pm 4.137)

The fitted (C_k/T_d) coefficients decreased as (k) increased, matching theory. The generator voltages also fell with increasing (k). Initial states sampled at different offsets within each (k)-cycle gave fitted amplitudes differing by roughly 0.1%. This supported the predicted independence from initial state.

The extended NV model reproduced the weak extra fringes seen in photoluminescence. It included hyperfine hybridization with the nearby (^{15}\mathrm{N}) nuclear spin and a longitudinal RF component caused by field misalignment. The best simulation used (|\mu|^2 = 0.9145) and (\theta = 9^\circ), with (B_z = 1020.678) G and (D_{\mathrm{ZFS}} = 2.87) GHz. The simulation used an 8.00 MHz transition with a correction for drift, while the experimental mean was 9.21 MHz.

The simple TLS model predicted (A_{P2}) near (0.5042\Delta_0), and the observed hyperbolic modulation agreed with its inverse-distance scaling. The fine scan produced slightly asymmetric fits on either side of (A_{P2}). The authors attribute this asymmetry to longitudinal driving and note that hybridization can also contribute.

The measured period-(k) structure was invariant across the tested initial states, as Floquet theory requires for a generic superposition. Earlier work had reported indications of subharmonic behavior, but the authors state that systematic exploration of a driven single TLS had not been performed. Their result covers (k = 2, 3, 4,) and 5 for one NV center.

Exact placement on each response manifold is difficult because the transition spacing drifts between amplitude scans. The analysis recorded the spacing once per amplitude scan and ignored drift within that scan. Dataset-2 also allowed up to 2% variation in drive frequency because it contained fewer iterations near each target.

Visual identification became harder for larger (k) because the pattern contains more phases, the scan was coarse, and relaxation limited repeated confirmations. Optical initialization produced only 95% to 98% fidelity, which slightly reduced fluorescence contrast.

The extended model matched the dominant structure but did not include weak relaxation, which the authors identify as a route to higher accuracy. Targeted experiments are still needed to determine the initial population distribution, branching ratio (|\mu|^2), and hyperfine constants without relying on ODMR or EPR spectra. The supporting data are not public, but the authors state that they are available on reasonable request.

Coherence sets whether the interference can be observed at all. The paper attributes the clear response to an NV sample whose (T_2^*) exceeded (100T_d). It also notes that relaxation erases the repeating pattern during longer measurements.

The results identify two selection criteria for this device class. The host needs long spin coherence and a controlled nuclear-spin environment near the ground-state level anticrossing. The observed (^{15}\mathrm{N}) hyperfine structure affected both the response and the model. The paper used an isotopically purified (^{12}\mathrm{C}) layer before introducing (^{15}\mathrm{N}).

The response is sensitive to small amplitude mismatches because the modulation period diverges near (A_{Pk}). The authors suggest that strongly driven NV spins may probe weak electromagnetic fields or residual spins. They also propose gate operations based on (k)-tupling, with timing tied more closely to the bare transition than conventional Rabi gates.

The material requirement is single-crystal diamond for a shallow NV spin device that must support implantation, annealing, and optical readout. The 6CCVD diamond materials catalog lists single-crystal diamond grown by MPCVD, with polishing to Ra below 1 nm for single-crystal material. Engineering support can review the paper’s substrate geometry and defect-preparation requirements before selecting a material configuration.

View Original Abstract

One of the prime features of quantum systems strongly driven by external time-periodic fields is the subharmonic response with integer multiples of the drive period <a:math xmlns:a=“http://www.w3.org/1998/Math/MathML”> <a:mrow> <a:mi>k</a:mi> <a:mspace width=“0.16em”/> <a:msub> <a:mi>T</a:mi> <a:mi>d</a:mi> </a:msub> </a:mrow> </a:math> due to long-lived interference. Here, we demonstrate experimentally, based on a careful theoretical analysis, period doubling and higher multiplicities ( <c:math xmlns:c=“http://www.w3.org/1998/Math/MathML”> <c:mrow> <c:mi>k</c:mi> <c:mo>=</c:mo> <c:mn>2</c:mn> <c:mo>,</c:mo> <c:mo>…</c:mo> <c:mo>,</c:mo> <c:mn>5</c:mn> </c:mrow> </c:math> ) for one of the most fundamental systems, namely, an individual spin <d:math xmlns:d=“http://www.w3.org/1998/Math/MathML”> <d:mrow> <d:mn>1</d:mn> <d:mo>/</d:mo> <d:mn>2</d:mn> </d:mrow> </d:math> . Nitrogen-vacancy centers in diamond support sufficiently stable coherent dynamics owing to long coherence times and allow for optical addressability of their spin states. This allows to monitor coherent period <e:math xmlns:e=“http://www.w3.org/1998/Math/MathML”> <e:mi>k</e:mi> </e:math> -tupling oscillations over a broad set of driving parameters in the vicinity of the ideal manifolds. In this domain, superimposed low-frequency modulations serve as unique proxy for the approach toward period <f:math xmlns:f=“http://www.w3.org/1998/Math/MathML”> <f:mi>k</f:mi> </f:math> -tupling.