Weak Driving in Quantum Systems

• Weak driving is a regime where external forces are much smaller than intrinsic scales, allowing approximations like the rotating-wave approximation and perturbation theory.
• It facilitates analytical and numerical models of quantum and classical systems by enabling a perturbative approach to phenomena such as multiphoton absorption and transport anomalies.
• Applications span two-level systems, superconducting qubits, spin chains, and optomechanical devices, revealing insights into nonequilibrium states, localization, and quantum speed limits.

Weak driving denotes a regime in which an external field, modulation, or boundary forcing is small relative to the intrinsic scales of the driven system, but the precise small parameter is model-dependent. In the generalized semi-classical Rabi model it is the condition $A_d \ll \omega_a$, with the experimental analogue $\Omega_{1,2} \ll \omega_a^\ast$; in the adiabaticity-based categorization of a driven two-level system it is the perturbative multiphoton absorption regime with $\gamma > 1$; in the boundary-driven XXZ chain it is the limit $\kappa \ll 1$ at finite $\Gamma$ (Dai et al., 2017, Heide et al., 2021, Popkov et al., 2012). Across the literature, weak driving is therefore not a single approximation but a family of asymptotic regimes in which the drive is small compared with a natural frequency, gap, coupling, or reservoir-induced scale. A recurrent theme is that weak driving often validates reduced descriptions such as the rotating-wave approximation or perturbation theory, yet it can also generate nontrivial transport anomalies, long-lived nonequilibrium states, localization changes, or sharp control limits (Riwar et al., 2014).

1. Regime definitions and diagnostic parameters

Weak driving is specified by different control parameters in different settings. In driven two-level systems, the relevant comparison is typically between drive amplitude and transition frequency. In transport problems it can be a small boundary twist. In turbulence it is the inequality $\omega_{nl} \ll \omega_l$. In quasiperiodically driven localization problems it is small modulation amplitude $\mu \ll 1$. In weak-driving quantum control, the emphasis is on low-amplitude realizations rather than arbitrary high-energy pulses (Dai et al., 2017, Popkov et al., 2012, TenBarge et al., 2013, Hatami et al., 2016, Janković et al., 20 Jun 2026).

Setting	Weak-driving criterion	Reported behavior
Semi-classical Rabi model	$A_d \ll \omega_a$, $\Omega_{1,2} \ll \omega_a^\ast$	RWA valid; sinusoidal Rabi oscillations
Driven TLS with adiabaticity parameter	$\gamma > 1$	Perturbative multiphoton absorption
XXZ chain with boundary twisting	$\Omega_{1,2} \ll \omega_a^\ast$0	NESS anomalies at $\Omega_{1,2} \ll \omega_a^\ast$1
Alfvénic turbulence antenna	$\Omega_{1,2} \ll \omega_a^\ast$2	Weak turbulence from low $\Omega_{1,2} \ll \omega_a^\ast$3
Quasiperiodic Anderson problem	$\Omega_{1,2} \ll \omega_a^\ast$4	Single-channel or multi-channel localization regimes
Weak-driving quantum control	Low-amplitude realizations	Critical gate time $\Omega_{1,2} \ll \omega_a^\ast$5 and energy divergence

A plausible implication is that “weak driving” should not be treated as synonymous with “linear response” unless the paper explicitly makes that identification. Some works do so, as in the perturbative multiphoton regime of a TLS, while others report strong qualitative consequences despite small drive amplitudes (Heide et al., 2021, Riwar et al., 2014).

2. Two-level, Rabi, and cavity-mediated dynamics

The generalized semi-classical Rabi Hamiltonian considered in the superconducting-transmon simulation is

$\Omega_{1,2} \ll \omega_a^\ast$6

with weak driving defined by $\Omega_{1,2} \ll \omega_a^\ast$7. In that regime the effect of the counter-rotating terms is negligible, the dynamics are well described by the rotating-wave approximation, and the effective resonant evolution yields standard Rabi oscillations

$\Omega_{1,2} \ll \omega_a^\ast$8

The experiment explicitly used $\Omega_{1,2} \ll \omega_a^\ast$9 MHz and $\gamma > 1$0 GHz, and reported sinusoidal population oscillations together with Bloch-sphere trajectories that are circular arcs in the $\gamma > 1$1-$\gamma > 1$2 plane; the trajectories for different initial phases coincide, showing that the initial phase has negligible effect in the weak-driving regime (Dai et al., 2017).

A distinct but related classification appears in the light-matter interaction analysis of a two-level system. There, weak driving is the perturbative multiphoton absorption regime for small driving field strengths, characterized by the Keldysh parameter

$\gamma > 1$3

with $\gamma > 1$4. Intraband motion is negligible, interband transitions dominate, and resonances occur at integer photon order $\gamma > 1$5, where $\gamma > 1$6. The paper presents this regime as the Rabi-physics side of a broader continuum that connects to Landau-Zener physics when the field is increased (Heide et al., 2021).

Weak driving can also be used as a control resource rather than merely a perturbative limit. In the cavity-mediated spin-squeezing proposal, a weak detuned parametric two-photon drive of amplitude $\gamma > 1$7 reshapes the effective spin-spin interaction. Without a drive, the induced interaction is one-axis twisting; with a weak optimized drive, the effective Hamiltonian can be tuned to the ideal two-axis twist interaction. For $\gamma > 1$8, the resulting effective interaction yields Heisenberg-limited squeezing with $\gamma > 1$9, whereas the undriven one-axis twisting case gives $\kappa \ll 1$0 (Groszkowski et al., 2020).

3. Open-system response, measurements, and superconducting junctions

In open quantum systems under repeated measurements, weak system-environment coupling permits a general expression for the effective decay rate,

$\kappa \ll 1$1

where the generalized filter function $\kappa \ll 1$2 contains the effect of the driving field. The central conclusion is that driving fields change the decay rate, and hence the quantum Zeno and anti-Zeno behavior, both qualitatively and quantitatively. In the short-$\kappa \ll 1$3 limit the decay approaches the universal Zeno limit, but away from that limit the drive reshapes the filter function and can produce multiple Zeno–anti-Zeno crossovers (Majeed et al., 2020).

A complementary result concerns periodically driven open systems with localized bound states. The exact analysis finds that periodic driving can preserve the existed localized bound states or destroy some of them but cannot generate new localized bound states. In the weak-driving case, the energy-transfer picture is nearly precise: if a localized bound state can absorb or release one or several energy quanta, i.e. $\kappa \ll 1$4, then it dissipates; otherwise it survives as a time-dependent localized bound state. The paper therefore distinguishes weak periodic driving from static energy shifts, which can either generate or destroy localized bound states (Xiong et al., 2021).

Short superconducting junctions provide a more striking example of weak drive with strong consequences. Under a dc and ac phase bias $\kappa \ll 1$5, with $\kappa \ll 1$6, the Andreev bound-state energy is

$\kappa \ll 1$7

The ac modulation induces ionization and refill processes between bound states and the continuum. When parity-conserving annihilation processes are much faster than parity relaxation, even a weak driving may lead to a large deviation of the supercurrent from its equilibrium value, and the effect is accompanied by a quasiparticle current that may lead to a measurable charge imbalance. After switching off the ac drive, the supercurrent relaxes to a stationary nonequilibrium state on timescales where parity relaxation is negligible; in diffusive junctions the long-time decay acquires an algebraic $\kappa \ll 1$8 tail because highly transparent channels have $\kappa \ll 1$9 (Riwar et al., 2014).

4. Transport anomalies and localization under weak drive

In the driven XXZ spin-$\Gamma$0 chain, weak driving is implemented by a small twisting gradient parameter $\Gamma$1 with finite $\Gamma$2, while boundary reservoirs impose perpendicular spin orientations in the $\Gamma$3 plane. The nonequilibrium steady state exhibits an anomaly at the isotropic point $\Gamma$4. For odd $\Gamma$5, the spin current scales as $\Gamma$6 at general $\Gamma$7; for even $\Gamma$8, the current scales as $\Gamma$9 at $\omega_{nl} \ll \omega_l$0 but as $\omega_{nl} \ll \omega_l$1 for $\omega_{nl} \ll \omega_l$2. The limits $\omega_{nl} \ll \omega_l$3 and $\omega_{nl} \ll \omega_l$4 do not commute for even $\omega_{nl} \ll \omega_l$5, which yields a current discontinuity. The energy current develops a twin-peak anomaly at the isotropic point, and the character of the singularity depends qualitatively on whether the system size is even or odd (Popkov et al., 2012).

In a different class of models, weak quasiperiodic driving modifies but does not eliminate Anderson localization in one dimension. The driven onsite modulation has amplitude $\omega_{nl} \ll \omega_l$6, and the weak-driving discussion focuses on $\omega_{nl} \ll \omega_l$7. In the single-channel regime, defined by

$\omega_{nl} \ll \omega_l$8

the problem is essentially equivalent to the undriven case. In the multi-channel regime, obtained for slow driving, the localization length increases substantially and displays two scaling regimes. The weak-driving plateau is reported as

$\omega_{nl} \ll \omega_l$9

while at even smaller $\mu \ll 1$0 the localization length grows according to a distinct logarithmically corrected expression. The paper emphasizes that the localization length stays finite for a finite number of frequency components (Hatami et al., 2016).

Weak and strong Floquet driving are also separated sharply in photonic topological systems. The low-amplitude $\mu \ll 1$1 region is identified as weak driving, and the associated single edge mode propagates with chirality matching the helix chirality. Increasing the driving amplitude moves the system into a strong-driving region with $\mu \ll 1$2, where the edge-mode propagation direction reverses even though the helix chirality is fixed. This establishes that weak and strong driving can correspond to distinct topological phases rather than to a smooth quantitative crossover only (1711.02477).

5. Collective activation, breather formation, and weak turbulence

Weak driving need not imply weak outcomes in nonlinear systems. In the linear chain of coupled nonlinear oscillators moving in a metastable cubic potential,

$\mu \ll 1$3

the homogeneous external forcing

$\mu \ll 1$4

was taken with a very weak amplitude $\mu \ll 1$5. For frequencies in the interval $\mu \ll 1$6, this weak forcing suffices to accomplish speedy escape. The mechanism is not direct barrier crossing by single units: the drive first excites phonon modes, then spatial fluctuations are amplified by a Mathieu-type instability, energy localizes into chaotic breathers, and mobile breathers merge with standing breathers until a transition-state configuration is formed. No escapes were observed in the uncoupled case $\mu \ll 1$7 under the same external drive and initial conditions, so the effect depends on cooperativity between the units of the chain (Hennig et al., 2013).

The oscillating Langevin antenna for plasma turbulence simulations provides an explicit weak-driving criterion in terms of turbulence rates. The antenna injects energy into selected Fourier modes through a parallel body current, and smaller $\mu \ll 1$8 gives weak turbulence while larger $\mu \ll 1$9 gives strong turbulence. The weak regime is defined by $A_d \ll \omega_a$0, i.e. the nonlinear energy-transfer rate is much smaller than the linear frequency. Achieving this regime requires choosing $A_d \ll \omega_a$1, while keeping $A_d \ll \omega_a$2 and typically $A_d \ll \omega_a$3. In this setting, the dynamics are quasi-linear and the nonlinear interaction time is long compared with the wave period (TenBarge et al., 2013).

A common misconception is that weak driving precludes threshold phenomena or collective amplification. The nonlinear chain and the superconducting-junction studies show the opposite: suitably tuned weak drive can trigger breather-mediated escape or large nonequilibrium current changes because the relevant amplification is supplied by the system’s internal dynamics rather than by the instantaneous drive amplitude alone (Hennig et al., 2013, Riwar et al., 2014).

6. Sensing, nonreciprocity, and weak-driving control limits

Weak driving is central to several control and sensing architectures. In the active cavity-magnon sensor, weak driving refers to a small-amplitude alternating magnetic field $A_d \ll \omega_a$4 applied to the YIG sphere. Floquet modulation then produces sidebands at $A_d \ll \omega_a$5, and in the weak-driving limit only the first-order sidebands are appreciable. The first sideband amplitude

$A_d \ll \omega_a$6

is linear in $A_d \ll \omega_a$7, so the sideband amplitude is directly proportional to the weak ac magnetic field. The PCB-based room-temperature device reports a detection limit of $A_d \ll \omega_a$8 at 80 MHz (Yang et al., 16 Apr 2026).

In cavity optomechanics, the mechanical resonator can be excited by a weak coherent parametric drive

$A_d \ll \omega_a$9

The paper identifies a regime of unidirectional isolation for $\Omega_{1,2} \ll \omega_a^\ast$0, and a transition to unidirectional amplification when $\Omega_{1,2} \ll \omega_a^\ast$1, with an amplification threshold around $\Omega_{1,2} \ll \omega_a^\ast$2. Thermal-noise terms proportional to $\Omega_{1,2} \ll \omega_a^\ast$3 limit the observation of ideal single-photon nonreciprocity and require the mechanical resonator to be cooled very close to its quantum ground state (Liu et al., 2018).

Weak-driving constraints also appear as a control-theoretic speed limit. For two exchange-coupled spins, low-amplitude realizations of the two-qubit quantum Fourier transform cease to exist below a critical gate time $\Omega_{1,2} \ll \omega_a^\ast$4, and the minimum control energy diverges on approach to this limit according to the area-pole law

$\Omega_{1,2} \ll \omega_a^\ast$5

The first term is identified as the time-optimal area cost and the second as a pole at the speed limit. The paper therefore distinguishes weak-driving realizability from full controllability: below $\Omega_{1,2} \ll \omega_a^\ast$6, high-fidelity solutions may still be found by broadband, high-amplitude methods, but not within the low-amplitude weak-driving regime (Janković et al., 20 Jun 2026).

Taken together, these results suggest that weak driving is best understood as a structurally constrained regime rather than a synonym for negligible physics. Depending on the model, it can imply RWA-valid sinusoidal oscillations, perturbative multiphoton absorption, topological behavior identical to the undriven case, or instead sharp anomalies, Floquet sideband transduction, protected nonequilibrium populations, and a distinct quantum speed limit (Dai et al., 2017, Heide et al., 2021, Janković et al., 20 Jun 2026).