Stroboscopic Squeezing in Quantum Systems

• Stroboscopic squeezing is a protocol that synchronizes squeezing operations to discrete time windows, enabling controlled manipulation of quantum states.
• It employs pulsed interactions and measurement-based schemes in systems such as spinor Bose–Einstein condensates, optomechanics, and trapped ions to mitigate decoherence.
• By isolating optimal squeezing phases through periodic sampling, the technique enhances metrological precision and facilitates robust quantum sensing.

Stroboscopic squeezing denotes a family of protocols in which squeezing, or the dynamics that produce it, is synchronized to a periodic reference and is therefore effective only at discrete times or within narrow time windows. In some systems the squeezed state itself appears only at special phases of a fast oscillation, as in the large-$|q|$ regime of a spin-1 spinor Bose–Einstein condensate; in others, the control or measurement is pulsed so that the same quadrature is repeatedly addressed, as in back-action-evading measurement of oscillators and stroboscopic quantum optomechanics. The term is therefore not attached to a single Hamiltonian or platform, but to a timing structure: periodic sampling, pulsed interaction, or discrete-time capture of an otherwise nonstationary squeezed configuration (Giovanoudis et al., 7 Jul 2026, Vasilakis et al., 2014, Brunelli et al., 2020).

1. Conceptual structure

Across the literature, stroboscopic squeezing is tied to synchronization between the control sequence and an intrinsic oscillator, collective spin, or driven Floquet orbit. In measurement-based settings, the synchronization is chosen so that the same effective quadrature is probed repeatedly. In interaction-engineering settings, the system is acted on only during short intervals, so useful nonlinear evolution accumulates while unwanted rotation, decoherence, or loss is suppressed. In collective-spin dynamics, the squeezing may be present only at discrete observation times because fast oscillatory motion periodically restores a favorable squeezing axis (Vasilakis et al., 2014, Ge et al., 2019, Giovanoudis et al., 7 Jul 2026).

A common timing condition is half-period sampling. In stroboscopic back-action-evading measurement of an oscillator, the probe is modulated at $2\Omega$, so that the same rotating-frame quadrature reappears at successive pulses. In pulsed optomechanics, backaction-evading squeezing uses pulse spacing

$T=\frac{\pi}{\omega_m},$

while mode-selective squeezing of condensate excitations uses a stroboscopic frequency

$\varpi = 2\omega_j$

for a target Bogoliubov mode $j$ (Vasilakis et al., 2014, Brunelli et al., 2020, Wade et al., 2015).

Taken together, these works suggest that “stroboscopic” has two technically distinct but related meanings. First, it can describe discrete-time appearance of squeezing in an otherwise rapidly oscillating evolution. Second, it can describe discrete-time application of measurement or control that engineers squeezing by repeatedly aligning with the same quadrature or by interleaving squeezing operations with free evolution.

2. Collective-spin realizations

A recent collective-spin realization appears in spin-1 spinor Bose–Einstein condensates in the single-mode regime, where all atoms share the same spatial wave function and the relevant dynamics is entirely in the internal spin degrees of freedom. The model Hamiltonian is

$$\frac{H}{c} = \frac{\mathbf{J}^2}{2N} + q \sum_m m^2 a_m^\dagger a_m ,$$

with $m=-1,0,1$, $N=\sum_m a_m^\dagger a_m$, and

$$J_\mu = \sum_{m,m'} a_m^\dagger (S_\mu)_{m,m'} a_{m'} .$$

For spin-1, the quadratic Zeeman term can be rewritten as

$$q \sum_m m^2 a_m^\dagger a_m = -q N_0 + \text{const.}, \quad N_0=a_0^\dagger a_0.$$

The relevant initial state is a coherent spin state aligned along $2\Omega$0,

$2\Omega$1

which lies in the maximal-spin Dicke manifold $2\Omega$2 (Giovanoudis et al., 7 Jul 2026).

In this setting, stroboscopic squeezing refers specifically to the regime of large quadratic Zeeman shift, $2\Omega$3, where squeezing is not continuously maintained but instead appears in narrow time windows separated by rapid oscillations in the collective-spin observables. The quadratic Zeeman term induces a fast oscillation at frequency $2\Omega$4; squeezing survives only at special times when the oscillatory contribution effectively aligns favorably; between those times the state can become anti-squeezed and the collective spin contrast $2\Omega$5 can collapse. Nevertheless, the optimal squeezing continues to follow the universal one-axis-twisting scaling

$2\Omega$6

for all explored $2\Omega$7, including the stroboscopic regime. The squeezing measure is the Wineland parameter

$2\Omega$8

with $2\Omega$9 signaling squeezing and $T=\frac{\pi}{\omega_m},$0 witnessing $T=\frac{\pi}{\omega_m},$1-partite entanglement (Giovanoudis et al., 7 Jul 2026).

The same paper distinguishes two mechanisms. For $T=\frac{\pi}{\omega_m},$2, projection onto the maximal-spin Dicke subspace gives an effective one-axis-twisting Hamiltonian,

$T=\frac{\pi}{\omega_m},$3

For $T=\frac{\pi}{\omega_m},$4, the interaction picture with respect to $T=\frac{\pi}{\omega_m},$5 yields a rapidly oscillating spin-changing term,

$T=\frac{\pi}{\omega_m},$6

and a rotating-wave approximation gives

$T=\frac{\pi}{\omega_m},$7

This effective model contains an OAT-like nonlinear term, but now accompanied by nematic and detuning corrections, so the observed squeezing is stroboscopic rather than smooth (Giovanoudis et al., 7 Jul 2026).

A closely related discrete-time theme appears in effective two-axis-countertwisting constructions built from one-axis twisting. In repeated $T=\frac{\pi}{\omega_m},$8-pulse switching and periodically modulated drive schemes, the exact dynamics only approximates effective TACT at specific instants, so the squeezing parameter oscillates in time. The freezing protocol of persistent atomic spin squeezing addresses precisely this nonstationarity: after the squeezing reaches its optimal value, a rapid rotation aligns the direction of minimal quantum noise with the $T=\frac{\pi}{\omega_m},$9-axis, after which evolution under $\varpi = 2\omega_j$0 produces only a nearly global phase and the squeezing becomes effectively persistent. In the repeated-pulse realization, the number of pulse pairs for $\varpi = 2\omega_j$1 can be reduced to about

$\varpi = 2\omega_j$2

and in the modulated-drive realization the needed modulation and Rabi frequencies can be reduced to roughly

$\varpi = 2\omega_j$3

(Wu et al., 2016).

3. Measurement-induced stroboscopic squeezing

A major class of stroboscopic squeezing protocols is measurement-based. In a collective-spin oscillator formed by the transverse components of an atomic ensemble precessing in a magnetic field, canonical variables are defined as

$\varpi = 2\omega_j$4

with $\varpi = 2\omega_j$5. The probe intensity is modulated as short pulses synchronized to the oscillator motion, with the pulse train repeated at twice the Larmor frequency $\varpi = 2\omega_j$6. In the ideal $\varpi = 2\omega_j$7 limit, the measured quadrature

$\varpi = 2\omega_j$8

approaches $\varpi = 2\omega_j$9. The back-action parameter becomes

$j$0

so $j$1 for ideal delta-function strobes. Experimentally, this protocol produced a squeezed state conditioned on the light-polarization measurement with $j$2 dB noise reduction below the zero-point fluctuations for the measured quadrature, starting from an average thermal occupancy number $j$3 (Vasilakis et al., 2014).

The same stroboscopic QND logic appears in multimode condensates. For a quasi-1D trapped Bose gas, density fluctuations are expanded in Bogoliubov modes,

$j$4

and dispersive imaging measures linear combinations of mode quadratures $j$5. Continuous probing squeezes, but the oscillator rotation at rate $j$6 continuously mixes the squeezed and anti-squeezed quadratures. Properly timed stroboscopic imaging instead applies short pulses at

$j$7

with $j$8, so one effective quadrature is repeatedly measured. For small phase window $j$9, the variance follows the QND-like result

$$\frac{H}{c} = \frac{\mathbf{J}^2}{2N} + q \sum_m m^2 a_m^\dagger a_m ,$$0

and the protocol can selectively squeeze one mode while largely avoiding crosstalk with others. The same framework also generates bipartite mode entanglement when the probe is modulated at $$\frac{H}{c} = \frac{\mathbf{J}^2}{2N} + q \sum_m m^2 a_m^\dagger a_m ,$$1 (Wade et al., 2015).

Stroboscopic quantum optomechanics extends the measurement-based picture to cavity optomechanical systems driven by a train of short pulses. The ideal pulse implements

$$\frac{H}{c} = \frac{\mathbf{J}^2}{2N} + q \sum_m m^2 a_m^\dagger a_m ,$$2

and homodyne detection of the output gives the Kraus operator

$$\frac{H}{c} = \frac{\mathbf{J}^2}{2N} + q \sum_m m^2 a_m^\dagger a_m ,$$3

For backaction-evading squeezing, the pulse spacing is again $$\frac{H}{c} = \frac{\mathbf{J}^2}{2N} + q \sum_m m^2 a_m^\dagger a_m ,$$4, so the same quadrature is sampled repeatedly. The steady-state variances satisfy

$$\frac{H}{c} = \frac{\mathbf{J}^2}{2N} + q \sum_m m^2 a_m^\dagger a_m ,$$5

with $$\frac{H}{c} = \frac{\mathbf{J}^2}{2N} + q \sum_m m^2 a_m^\dagger a_m ,$$6, and in the large-$$\frac{H}{c} = \frac{\mathbf{J}^2}{2N} + q \sum_m m^2 a_m^\dagger a_m ,$$7 limit

$$\frac{H}{c} = \frac{\mathbf{J}^2}{2N} + q \sum_m m^2 a_m^\dagger a_m ,$$8

Mechanical squeezing below the zero-point level requires

$$\frac{H}{c} = \frac{\mathbf{J}^2}{2N} + q \sum_m m^2 a_m^\dagger a_m ,$$9

A central implication is that many weak stroboscopic measurements can accumulate enough information to generate squeezing even with moderate radiation-pressure interaction (Brunelli et al., 2020).

4. Pulsed interaction engineering

In trapped-ion quantum information processing, stroboscopic squeezing refers to a protocol that alternates short bursts of parametric amplification with spin-dependent force pulses. Parametric amplification is generated by modulating the trap at roughly twice the motional frequency, with squeezing operator

$m=-1,0,1$0

Because

$m=-1,0,1$1

a small spin-dependent displacement is amplified by a factor $m=-1,0,1$2. In the square protocol, the full loop encloses a phase-space area enlarged by $m=-1,0,1$3, yielding the geometric phase

$m=-1,0,1$4

For total time

$m=-1,0,1$5

the optimized phase is

$m=-1,0,1$6

and the enhancement factor relative to the no-PA square gate is

$m=-1,0,1$7

The gain can therefore become exponential in gate time, but the same squeezing also amplifies sensitivity to mode-frequency fluctuations and relative phase errors (Ge et al., 2019).

A different interaction-pulsing realization is stroboscopic Rydberg dressing in arrays of optical microtraps containing cesium atoms. Here the atoms are not dressed continuously; instead, the dressing light is applied in a sequence of short pulses with a dark time between pulses. The purpose is to preserve the coherent one-axis-twisting interaction while suppressing facilitated excitation and super-Poissonian loss. In the coherence study, the sequence used $m=-1,0,1$8 dressing pulses and total dressing time $m=-1,0,1$9. At zero delay, broadening and loss are large; at delay $N=\sum_m a_m^\dagger a_m$0, broadening is strongly suppressed, with the observed noise decaying toward the coherent-spin-state level with a characteristic timescale

$N=\sum_m a_m^\dagger a_m$1

For the squeezing experiment, each pulse was limited to about $N=\sum_m a_m^\dagger a_m$2, and the best measured result was

$N=\sum_m a_m^\dagger a_m$3

corresponding to about $N=\sum_m a_m^\dagger a_m$4 dB of squeezing. The experiment realized metrological gain across three spatially separated ensembles in parallel (Hines et al., 2023).

These pulsed-interaction schemes differ from measurement-based stroboscopic squeezing, but they share the same operational logic: the desired nonlinear action is concentrated into short intervals, while periods of free evolution or dark time are used to suppress deleterious dynamics.

5. Freezing, selectivity, and sensing relevance

A recurring advantage of stroboscopic protocols is that they separate state preparation from state use. In the spin-1 spinor condensate, once the quadratic Zeeman shift is switched off, the remaining Hamiltonian is

$N=\sum_m a_m^\dagger a_m$5

which is SU(2)-symmetric. Then

$N=\sum_m a_m^\dagger a_m$6

leaves collective-spin observables unchanged. The squeezed state is therefore effectively frozen, after which an external field $N=\sum_m a_m^\dagger a_m$7 can be sensed through

$N=\sum_m a_m^\dagger a_m$8

Because the interaction no longer limits the interrogation time, the sensing time $N=\sum_m a_m^\dagger a_m$9 can in principle be made arbitrarily long (Giovanoudis et al., 7 Jul 2026).

In persistent atomic spin squeezing, the freezing step is geometric rather than Hamiltonian. After the squeezing optimum is reached, a rapid rotation aligns the narrow noise axis with $$J_\mu = \sum_{m,m'} a_m^\dagger (S_\mu)_{m,m'} a_{m'} .$$0, converting a transient stroboscopic optimum into a stationary one under $$J_\mu = \sum_{m,m'} a_m^\dagger (S_\mu)_{m,m'} a_{m'} .$$1. In the repeated-pulse effective-TACT scheme, the squeezing axis at maximum squeezing lies at angle $$J_\mu = \sum_{m,m'} a_m^\dagger (S_\mu)_{m,m'} a_{m'} .$$2 to the $$J_\mu = \sum_{m,m'} a_m^\dagger (S_\mu)_{m,m'} a_{m'} .$$3-axis in the $$J_\mu = \sum_{m,m'} a_m^\dagger (S_\mu)_{m,m'} a_{m'} .$$4-$$J_\mu = \sum_{m,m'} a_m^\dagger (S_\mu)_{m,m'} a_{m'} .$$5 plane, so a $$J_\mu = \sum_{m,m'} a_m^\dagger (S_\mu)_{m,m'} a_{m'} .$$6 rotation about the $$J_\mu = \sum_{m,m'} a_m^\dagger (S_\mu)_{m,m'} a_{m'} .$$7-axis stores the state (Wu et al., 2016).

Mode selectivity is another important consequence of stroboscopic timing. In condensate imaging, the nonuniform Bogoliubov spectrum allows a pulse train at $$J_\mu = \sum_{m,m'} a_m^\dagger (S_\mu)_{m,m'} a_{m'} .$$8 to squeeze one selected mode while leaving others essentially untouched, as diagnosed by $$J_\mu = \sum_{m,m'} a_m^\dagger (S_\mu)_{m,m'} a_{m'} .$$9 and $$q \sum_m m^2 a_m^\dagger a_m = -q N_0 + \text{const.}, \quad N_0=a_0^\dagger a_0.$$0 for small $$q \sum_m m^2 a_m^\dagger a_m = -q N_0 + \text{const.}, \quad N_0=a_0^\dagger a_0.$$1. Feedback via a weak time-dependent trapping potential damps stochastic mean displacements without substantially altering the covariance squeezing (Wade et al., 2015).

The sensing significance is explicit across several platforms. Spinor condensates are proposed for entanglement-assisted quantum sensing (Giovanoudis et al., 7 Jul 2026). Rydberg-dressed arrays are presented as a route to enhance the precision of tests of fundamental physics based on arrays of atomic clocks and to enable quantum-enhanced imaging of electromagnetic fields (Hines et al., 2023). Measurement-based atomic squeezing is directly connected to metrological advantage in quantum sensing through reduced uncertainty in a selected quadrature below the standard quantum limit (Vasilakis et al., 2014).

6. Terminological boundaries and related usages

The literature does not use “stroboscopic squeezing” in a single universal sense. Several distinctions are technically important.

First, some works use the term for discrete-time emergence of squeezing rather than pulsed control. In the spinor-condensate case, stroboscopic squeezing means that for large $$q \sum_m m^2 a_m^\dagger a_m = -q N_0 + \text{const.}, \quad N_0=a_0^\dagger a_0.$$2 the system exhibits squeezing only at discrete times or narrow windows in time, much like snapshots taken at the right phase of a fast oscillation (Giovanoudis et al., 7 Jul 2026). This differs from pulsed back-action-evading measurement, where the stroboscopic element is the probe sequence itself (Vasilakis et al., 2014, Brunelli et al., 2020).

Second, some works engineer nonlinear or mediator states stroboscopically rather than directly squeezing an observable in the standard linear-quadrature sense. In stroboscopic high-order nonlinearity for quantum optomechanics, the nonlinear potential is applied in short pulses synchronized to the oscillator period, and the resulting nonclassical resource is characterized by squeezing of a nonlinear quadrature,

$$q \sum_m m^2 a_m^\dagger a_m = -q N_0 + \text{const.}, \quad N_0=a_0^\dagger a_0.$$3

not by ordinary Gaussian quadrature squeezing. The variance

$$q \sum_m m^2 a_m^\dagger a_m = -q N_0 + \text{const.}, \quad N_0=a_0^\dagger a_0.$$4

is the relevant resource metric, and the paper is explicit that this is generalized or nonlinear squeezing (Rakhubovsky et al., 2019).

Third, some papers discuss stroboscopic dynamics that are relevant to squeezing without actually demonstrating it. In long-range interacting spin systems, stroboscopic aliasing generates period-$$q \sum_m m^2 a_m^\dagger a_m = -q N_0 + \text{const.}, \quad N_0=a_0^\dagger a_0.$$5 motion between $$q \sum_m m^2 a_m^\dagger a_m = -q N_0 + \text{const.}, \quad N_0=a_0^\dagger a_0.$$6 emergent dynamical fixed points. The unstable fixed points are proposed as a route to squeezing and entangled-state preparation, but squeezing is not explicitly computed or demonstrated (Kelly et al., 2020).

Fourth, some periodically timed squeezing-related schemes are explicitly not stroboscopic squeezing in the standard sense. In nonlinear atomic force microscopy, the two-frequency drive is a continuous intermodulation protocol that enhances time near the surface and produces classical phase-space squeezing, but the paper states that it is not stroboscopic squeezing in the standard experimental or quantum-optics sense (Marzlin et al., 2021).

Finally, there are neighboring uses of “stroboscopic” and “squeezed” that concern squeezed resources rather than squeezed target systems. In superconducting-qubit readout, stroboscopic longitudinal coupling is combined with injected squeezed microwave light, yielding a $$q \sum_m m^2 a_m^\dagger a_m = -q N_0 + \text{const.}, \quad N_0=a_0^\dagger a_0.$$7 increase in the room-temperature signal-to-noise ratio and slowing dephasing by a factor of $$q \sum_m m^2 a_m^\dagger a_m = -q N_0 + \text{const.}, \quad N_0=a_0^\dagger a_0.$$8 when the orthogonal phase is squeezed. This protocol is stroboscopic and uses squeezing, but it is a readout architecture rather than a squeezed-state preparation scheme for the qubit itself (Eddins et al., 2017).

These distinctions show that stroboscopic squeezing is best understood as a timing principle rather than a single method. Depending on the platform, it may mean discrete-time capture of transient squeezing, pulsed QND measurement of a fixed quadrature, alternation of squeezing and displacement operations, intermittent dressing to preserve coherent nonlinear evolution, or stroboscopic application of a nonlinear potential that produces a non-Gaussian squeezed resource.