Single-Shot Cross-Spectroscopy (SSCS)

• SSCS is a noise-spectroscopy protocol that extracts auto- and cross-power spectral densities from dephasing noise in qubit pairs using synchronized single-shot Ramsey experiments.
• It overcomes the intermediate-frequency gap in conventional methods by forming logarithmic correlators that isolate noise correlations while ensuring SPAM resilience.
• Experimental implementations on silicon spin-qubits have demonstrated SSCS’s capability to reproduce benchmark spectra and resolve detailed noise features such as 1/f decay and phase switches.

Searching arXiv for the cited SSCS paper and closely related single-shot spectroscopic methods. arXiv search: (Rojas-Arias et al., 26 Sep 2025) Single-Shot Cross-Spectroscopy (SSCS) is a noise-spectroscopy protocol for extracting both the auto-power spectral density (auto-PSD) and cross-power spectral density (cross-PSD) of dephasing noise acting on a pair of qubits, using only synchronized single-qubit Ramsey experiments with single-shot readout (Rojas-Arias et al., 26 Sep 2025). Its central innovation is to bypass the conventional step of estimating a qubit’s energy from many Ramsey experiments and instead work directly with raw single-shot outcomes, from which suitable correlators between qubits and between Ramsey subsequences are formed to isolate the underlying noise correlations. In the literature assembled here, the acronym SSCS is explicitly introduced in this qubit-noise context, while several earlier optical and computational-imaging methods are conceptually related because they also seek single-shot recovery of spectrally differentiated information without temporal scanning (Mathur, 2020).

1. Terminological scope and problem setting

The term Single-Shot Cross-Spectroscopy is introduced as a method for measuring correlated dephasing noise in qubit pairs over a frequency range that had been difficult to access with prior techniques (Rojas-Arias et al., 26 Sep 2025). The motivating problem is a frequency-gap structure in conventional noise spectroscopy. Low-frequency noise is usually accessed by repeatedly estimating qubit energies from Ramsey experiments and Fourier transforming the resulting time trace, whereas high-frequency noise is usually accessed with dynamical-decoupling sequences such as CPMG. For cross correlations, low frequencies can be accessed by tracking multiple qubits’ energies, but high-frequency cross-spectroscopy with multi-qubit dynamical decoupling is much harder and not experimentally mature. SSCS is introduced specifically to bridge this intermediate-frequency gap.

The protocol is formulated for a pair of qubits, denoted $Q_1$ and $Q_2$, and requires only synchronized single-qubit operations. No two-qubit gates are required. The physical quantity of interest is the correlation function

$$\langle \delta\omega_\alpha(t')\,\delta\omega_\beta(t'+t)\rangle,$$

where $\delta\omega_\alpha(t)$ is the fluctuating frequency of qubit $\alpha$. By reconstructing this time-domain correlator and Fourier transforming it, SSCS obtains the corresponding spectral density.

A common misconception is to treat SSCS as a generic synonym for any single-shot spectroscopic architecture. The available record does not support that usage. The explicit acronym refers to the qubit-noise protocol of (Rojas-Arias et al., 26 Sep 2025). Earlier solar, astronomical, ptychographic, and ultrafast-laser diagnostics are better regarded as conceptually adjacent single-shot cross-spectral or cross-wavelength methods rather than nominal instances of the same technique.

2. Ramsey subsequences, single-shot observables, and correlators

SSCS operates by running, on each qubit, two alternating Ramsey-type subsequences (Rojas-Arias et al., 26 Sep 2025). These are

• $R_{XX}$: $\pi/2$ about $X$, free evolution for $\tau_\alpha$, then $\pi/2$ about $Q_2$0, then single-shot readout.
• $Q_2$1: the same sequence, but with the second $Q_2$2 pulse about $Q_2$3.

For qubit $Q_2$4, the single-shot result is encoded as $Q_2$5, where $Q_2$6 and $Q_2$7. Each Ramsey subsequence consists of the following steps: initialize to $Q_2$8, apply $Q_2$9 pulse about $$\langle \delta\omega_\alpha(t')\,\delta\omega_\beta(t'+t)\rangle,$$0, free evolution for time $$\langle \delta\omega_\alpha(t')\,\delta\omega_\beta(t'+t)\rangle,$$1, apply final $$\langle \delta\omega_\alpha(t')\,\delta\omega_\beta(t'+t)\rangle,$$2 pulse about $$\langle \delta\omega_\alpha(t')\,\delta\omega_\beta(t'+t)\rangle,$$3 or $$\langle \delta\omega_\alpha(t')\,\delta\omega_\beta(t'+t)\rangle,$$4, and perform projective single-shot readout. The readout outcomes are collected over many repetitions with repetition period $$\langle \delta\omega_\alpha(t')\,\delta\omega_\beta(t'+t)\rangle,$$5. The qubits need only be aligned on the scale of $$\langle \delta\omega_\alpha(t')\,\delta\omega_\beta(t'+t)\rangle,$$6, not necessarily exactly simultaneously at the level of the Ramsey pulses.

From the measured outcomes $$\langle \delta\omega_\alpha(t')\,\delta\omega_\beta(t'+t)\rangle,$$7, the empirical mean and two-time correlators are estimated as

$$\langle \delta\omega_\alpha(t')\,\delta\omega_\beta(t'+t)\rangle,$$8

and

$$\langle \delta\omega_\alpha(t')\,\delta\omega_\beta(t'+t)\rangle,$$9

The subtraction of the product of means is essential. It removes state-preparation-and-measurement bias at the correlator level and is therefore part of the protocol’s built-in SPAM resilience. This measurement architecture is deliberately minimal: synchronized single-qubit Ramsey cycles, single-shot readout, and correlator estimation from the raw binary outcomes.

3. Statistical model and reconstruction of the cross-spectrum

The average Ramsey readout for qubit $\delta\omega_\alpha(t)$0 is modeled as

$\delta\omega_\alpha(t)$1

with $\delta\omega_\alpha(t)$2 the readout bias or offset from SPAM, $\delta\omega_\alpha(t)$3 the readout visibility or contrast from SPAM, $\delta\omega_\alpha(t)$4 the average qubit frequency in the chosen rotating frame, and $\delta\omega_\alpha(t)$5, so that the $\delta\omega_\alpha(t)$6 versus $\delta\omega_\alpha(t)$7 final pulse introduces a phase shift (Rojas-Arias et al., 26 Sep 2025).

Assuming quasi-static, wide-sense stationary, Gaussian frequency noise, the averaged correlator is

$\delta\omega_\alpha(t)$8

where $\delta\omega_\alpha(t)$9, and

$\alpha$0

The extraction step is based on two logarithmic combinations of correlators,

$\alpha$1

and

$\alpha$2

Under the stated assumptions, both isolate the same time-domain frequency correlator up to an additive constant: $\alpha$3

The cross-PSD then follows by Fourier transform,

$\alpha$4

The auto-PSD appears as a special case. For $\alpha$5, the protocol recommends

$\alpha$6

which yields

$\alpha$7

This algebraic isolation of the correlator is the defining analytical step of SSCS. It converts raw single-shot statistics into a spectral estimate without first constructing an intermediate qubit-energy time trace.

4. SPAM resilience, aliasing, and range-setting mechanisms

A defining practical property of SSCS is its resilience to SPAM errors (Rojas-Arias et al., 26 Sep 2025). In the raw single-shot means, SPAM enters through $\alpha$8 and $\alpha$9. The correlator definition

$R_{XX}$0

removes the bias $R_{XX}$1. The remaining visibility factors $R_{XX}$2 appear only as multiplicative prefactors in $R_{XX}$3, and after taking logarithms they become additive constants. After Fourier transform, those constants contribute only $R_{XX}$4 terms at zero frequency and therefore do not contaminate the finite-frequency spectrum. The method thus does not require prior calibration of SPAM parameters.

The accessible frequency range is set primarily by the experiment repetition rate rather than by the time required to estimate a qubit energy from multiple Ramsey shots. This is the basic reason SSCS extends the upper frequency bound relative to conventional energy-tracking methods. The demonstrated range is approximately

$R_{XX}$5

and the paper states that the accessible range is limited only by the experiment repetition rate and can scale accordingly on other platforms.

Aliasing is an explicit concern because the protocol alternates $R_{XX}$6 and $R_{XX}$7, so the two estimators $R_{XX}$8 and $R_{XX}$9 are sampled on offset time grids. Individually, each estimator suffers from aliasing, but averaging the Fourier transforms of the two substantially suppresses folded high-frequency contributions. This provides improved reliability near the Nyquist limit.

The paper also notes a generalized non-quasi-static expression in which the prefactor $\pi/2$0 is replaced by

$\pi/2$1

indicating a more general filter-function treatment beyond the quasi-static approximation. This suggests that SSCS is not restricted to a single asymptotic noise model, although the principal derivation is presented in the quasi-static regime.

5. Experimental implementation and observed spectra

SSCS was tested on a silicon spin-qubit device consisting of a linear five-quantum-dot array in a $\pi/2$2Si/SiGe heterostructure, using qubits 2 and 3 (Rojas-Arias et al., 26 Sep 2025). Two measurement modes were used.

Mode	Parameters	Acquisition
Cross-PSD mode	$\pi/2$3, $\pi/2$4	$\pi/2$5 repetitions per batch, 8 batches total
Auto-PSD mode	$\pi/2$6, $\pi/2$7	$\pi/2$8 repetitions per batch, 10 batches total

The conventional benchmark spectra were obtained by estimating qubit energies from interleaved Ramsey experiments over hours-long time traces. SSCS reproduced the benchmark spectra where they overlapped while extending the range to much higher frequencies.

The experimentally observed spectra exhibited approximately $\pi/2$9-type decay, clear 50 Hz and 100 Hz peaks from electrical mains and its harmonic, a correlation-phase switch near $X$0 Hz, another $X$1-phase switch around $X$2 Hz in the cross-PSD, and a Lorentzian-like negative-correlation region consistent with a two-level fluctuator between the qubits. These observations establish that the method does not merely detect aggregate noise strength; it also resolves structure in the sign and phase of inter-qubit correlations.

The paper further states that SSCS enables access to noise correlations in the previously inaccessible intermediate-frequency range $X$3–$X$4 Hz for spin qubits and could potentially extend above $X$5 Hz with faster readout. This suggests a direct role for SSCS in multi-qubit calibration and noise-characterization workflows on platforms where Ramsey-type single-shot readout is available.

6. Relation to earlier single-shot cross-spectral and multiplexed methods

The broader literature represented here contains several single-shot measurement architectures that are conceptually related to SSCS but not identical to it. Their common feature is the replacement of temporal scanning by spatial, angular, wavelength, or channel multiplexing in a single exposure or shot.

Work	Domain	Relation to SSCS
“Single Shot Spectroscopic Design Aspects” (Mathur, 2020)	Solar spectroscopy	Snapshot $X$6 acquisition by pupil-plane multiplexing and FP angle encoding
“Optical design of a multi-resolution, single shot spectrograph” (Henault et al., 2016)	Astronomical spectroscopy	Parallel spectral channels at different resolving powers via pupil slicing
“Wavelength-multiplexed single-shot ptychography” (Barolak et al., 2020)	Plasma imaging	Single-shot multiwavelength reconstruction of phase and amplitude
“Resolving ultrahigh-contrast ultrashort pulses with single-shot cross-correlator at the photon noise limit” (Ma et al., 2021)	Ultrafast laser diagnostics	Single-shot nonlinear correlation with time-to-space mapping

In solar spectroscopy, a pupil-sampling snapshot spectrograph samples the pupil plane with a lenslet array, sends the re-collimated beams through a Fabry–Pérot etalon in collimated mode, and uses a prefilter with full width half maximum smaller than the free spectral range to isolate a single order (Mathur, 2020). The result is multiple subimages of the same field, each corresponding to a slightly different wavelength because the transmitted wavelength shifts toward the blue as the distance from the FP axis increases. This architecture is closely aligned with the idea of recovering spectrally resolved information without scanning either the field of view or the wavelength axis, but it is an optical snapshot spectrograph rather than a qubit-noise cross-spectroscopy protocol.

In astronomical instrumentation, a multi-resolution single-shot spectrograph uses a Type 2 pupil slicer to divide the beam into sub-pupils and direct them through different dispersive elements, producing simultaneous spectra at $X$7 and $X$8 over specified sub-bands (Henault et al., 2016). This is again a single-shot, parallel-channel realization of multiplexed spectroscopy, but its purpose is simultaneous multi-resolution readout rather than extraction of a cross-PSD from stochastic qubit fluctuations.

In plasma diagnostics, wavelength-multiplexed single-shot ptychography performs SSP simultaneously with probes of multiple wavelengths and reconstructs wavelength-dependent phase and amplitude using ptychographic information multiplexing (Barolak et al., 2020). In that context, simultaneous reconstructions at different wavelengths permit separation of electron and neutral density contributions from a single-shot dataset. The paper explicitly frames this as a closely related analogue to single-shot cross-spectral imaging rather than generic spectroscopy.

In ultrafast-laser metrology, a single-shot cross-correlator based on noncollinear third-harmonic generation maps temporal contrast into a spatial coordinate and then into delayed temporal channels using a 100-pixel fiber array (Ma et al., 2021). The demonstrated detection limit reaches a contrast of $X$9, with the lowest pedestal set by single-photon detection and the highest signal set by the SSCC damage threshold. This shares with SSCS the structural principle of a single-event nonlinear or correlational measurement followed by a mapping stage, but it measures temporal pulse contrast rather than dephasing-noise spectra.

Taken together, these works define a broader methodological family of single-shot, multiplexed, or cross-domain measurement strategies. Within that family, SSCS in the strict sense denotes the qubit-noise protocol of (Rojas-Arias et al., 26 Sep 2025): synchronized single-qubit Ramsey experiments, correlator algebra that isolates $\tau_\alpha$0, Fourier reconstruction of auto- and cross-PSD, SPAM resilience, and frequency coverage governed primarily by repetition rate rather than by slow energy-tracking workflows.