betaFeedback

WorksPhysics and Astronomy

Discrete time liquids

Abstract

Discrete time crystals exhibit robust subharmonic oscillations, spontaneously breaking the discrete time-translation symmetry of a periodic drive. Since their emergence, the question of whether other types of temporal many-body order can exist has attracted considerable attention. Here we identify discrete time liquids – a new form of spatiotemporal order characterized by short-range temporal coherence as a result of time-solitons. We study a two-dimensional spin system coupled to a finite-temperature cavity, whose fluctuations provide a noisy, approximately periodic drive with only finite temporal coherence. Remarkably, the spins nevertheless form a metastable state with robust period-doubled dynamics over an extended parameter regime. Unlike a time crystal, this state has only finite temporal coherence: individual trajectories stochastically switch between the two symmetry-related temporal configurations of the period-doubled response. We identify these switches as time-solitons: topological defects in temporal order that preserve local period doubling while destroying long-range temporal coherence, thereby producing the time liquid state. We propose a minimal model which reproduces the distribution of temporal correlations of the many-body system, suggesting a general description of the time liquid as a time crystal with a finite density of time-solitons. Because these solitons can occur at arbitrary times, ensemble averaging restores continuous time-translation symmetry, raising the intriguing possibility that discrete time liquids can exist in thermal equilibrium.

Introduction

Fluctuations are usually regarded as an enemy of order. In spatially ordered phases, they blur static patterns and can forbid long-range order in low dimensions [1, 2]; in dynamical systems, they dephase coherent motion and broaden sharp spectral features [3, 4]. Nevertheless, fluctuations can also play a constructive role. In order by disorder, fluctuations select ordered states that are not preferred at the mean-field level [5, 6]. In stochastic and coherence resonances, fluctuations optimize the response of nonlinear dynamical systems [7, 8, 9]. Both of these cases show that fluctuations can also create regularity rather than merely destroy it. This idea has recently expanded into the field of cavity-matter hybrids [10, 11, 12, 13]. There, the goal is to generate new phases of matter by amplifying electromagnetic fluctuations around a two-dimensional heterostructure through nanopatterning of the electromagnetic environment. In this article, we find that fluctuations not only generate new spatially coherent phases of matter, but can also give rise to many-body temporal order, producing what we term discrete time liquids.

Over the past decade, quantum many-body systems have been shown to support robust phases of matter with temporal order known as time crystals. In a conventional discrete time crystal (DTC), an externally imposed periodic drive defines a discrete time-translation symmetry, which the many-body system spontaneously breaks by locking local observables to a robust subharmonic response [14]. Such order has been established in many-body-localized, prethermal, classical Floquet and driven-dissipative systems [15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27]. In most realizations, however, the temporal structure is supplied by an external drive that is, to a good approximation, perfectly periodic. A schematic

time-crystal evolution for ferromagnetically interacting spins is generated over one period of unit length by

U^=eiJ∑⟨ij⟩S^izS^jzeiπ∑iS^ix,(1)\hat{U}=e^{iJ\sum_{\langle ij\rangle}\hat{S}^{z}_{i}\hat{S}^{z}_{j}}e^{i\pi\sum_i\hat{S}^{x}_{i}}, \tag*{(1)}

which combines ferromagnetic interactions with a global spin flip. The periodic drive therefore provides a coherent clock against which the subharmonic response is defined. This raises an intriguing question: can temporal order emerge out of noise? To study this question we consider the consequences of coupling a set of spins to a finite-temperature cavity mode (see Figure 1). Thermal fluctuations of the cavity generate an approximately periodic signal centered at the cavity frequency. Fluctuations in its amplitude and phase, however, cause this signal to lose temporal coherence beyond a characteristic autocorrelation time. One might therefore expect any induced temporal order to disappear on the same timescale. Instead, we find that the spins enter a long-lived metastable state with a sharp period-doubled response at half the cavity frequency (Figure 1(b)). This response persists far beyond the coherence time of the cavity itself (Figure 3).

Time liquid in a cavity

Figure 1. Time liquid in a cavity. (a) Schematic of the setup and three regimes reached by increasing bath temperature. A spin lattice is coupled to a cavity mode of frequency ωc=2πτosc\omega_c = \frac{2\pi}{\tau_{\mathrm{osc}}} and damping rate γ=2τcav\gamma= \frac{2}{\tau_{\mathrm{cav}}}, which is in contact with a thermal bath at temperature TbathT_{\mathrm{bath}} (color bar). The thermal fluctuations of the oscillator coordinate y(t)y(t) act on the spins as an effective drive. Each row illustrates the dynamics at the temperature of the corresponding dashed line in (b). At low temperature (blue), the effective drive y(t)y(t) is too weak to disturb the spins, which precess freely. At intermediate temperature (orange), the spins lock to a sharp subharmonic response at half the cavity frequency. At high temperature (red), the drive is too strong and the dynamics becomes disordered. Period-doubled order therefore emerges only at intermediate bath temperature. (b) Spin response amplitude ∣F[⟨Mz⟩](ν)∣\left|\mathcal{F}\left[\langle M^z\rangle\right](\nu)\right| as a function of frequency ν/π\nu/\pi and bath temperature Tbath/T0T_{\mathrm{bath}}/T_0, at ε0=0.1\varepsilon_0=0.1, τosc=0.2\tau_{\mathrm{osc}}=0.2, and γ=10−4\gamma=10^{-4}. The subharmonic response at ν/π=1\nu/\pi=1 appears within a finite window of bath temperature and vanishes at both small and large TbathT_{\mathrm{bath}}. The period doubling is thus induced by thermal noise. (c) Phase diagram as a function of the cavity coherence time τcav/τosc\tau_{\mathrm{cav}}/\tau_{\mathrm{osc}} and the bath temperature Tbath/T0T_{\mathrm{bath}}/T_0, at ε0=0.05\varepsilon_0=0.05 and τosc=0.01\tau_{\mathrm{osc}}=0.01. We identify four regimes. At low temperature, the spins precess freely because the effective drive is too weak to disturb them. At high temperature, the dynamics is disordered and the period doubling is lost within a few cavity periods. In between, the system forms a DTC or a DTL, according to whether the correlation time τcorr\tau_{\mathrm{corr}} is comparable to or much shorter than the thermalization time τth\tau_{\mathrm{th}} (see “Phase classification” in Methods). The DTC appears in an intermediate temperature window that’s getting narrower as the coherence time decreases and disappears below τcav/τosc≈130\tau_{\mathrm{cav}}/\tau_{\mathrm{osc}}\approx130. The DTL, in contrast, persists at every coherence time shown. A short coherence time of the cavity mode therefore turns the DTC into a DTL. Panels (b) and (c) share L=100L=100, J=5J=5, Mosc=kB=λ=1M_{\mathrm{osc}}=k_{\mathrm{B}}=\lambda=1, θinit=0.2\theta_{\mathrm{init}}=0.2, Np=104N_p=10^4 stroboscopic periods, and Ntraj=100N_{\mathrm{traj}}=100 noise realizations. See text for details.

Time-solitons and time liquid lifetime

Figure 3. Time-solitons and time liquid lifetime. (a) Single-trajectory stroboscopic magnetization Mz(nτosc)M^z(n\tau_{\mathrm{osc}}) (blue) and its period-doubled component (−1)nMz(nτosc)(-1)^n M^z(n\tau_{\mathrm{osc}}) (red), at τcav/τosc=20\tau_{\mathrm{cav}}/\tau_{\mathrm{osc}}=20 (top) and 200 (bottom). The dashed line marks τcav\tau_{\mathrm{cav}}. The period-doubled component is repeatedly suppressed and revives with the opposite sign. Each such event is a time-soliton, the εn=π\varepsilon_n=\pi event of Eq. (2). (b) Thermalization time τth\tau_{\mathrm{th}} (circles), noise averaged single-trajectory correlation time ⟦τcorr⟧\llbracket\tau_{\mathrm{corr}}\rrbracket (diamonds), and ensemble correlation time τcorrens\tau_{\mathrm{corr}}^{\mathrm{ens}} (squares), as a function of the coherence time τcav\tau_{\mathrm{cav}}. The heatmap gives the fraction of noise realizations per correlation time τcorr\tau_{\mathrm{corr}} at fixed τcav\tau_{\mathrm{cav}}. The three time scales converge at long τcav\tau_{\mathrm{cav}}, where solitons are rare. They separate as τcorrens≤⟦τcorr⟧≤τth\tau_{\mathrm{corr}}^{\mathrm{ens}}\leq\llbracket\tau_{\mathrm{corr}}\rrbracket\leq\tau_{\mathrm{th}} at short τcav\tau_{\mathrm{cav}}, where solitons proliferate. Solitons thus destroy long-range temporal coherence. (c) Windowed subharmonic weight S(m)S(m) of Eq. (11), measured in a window opened at period mm, for the two coherence times of (a). The dashed and dash-dotted lines indicate τcav\tau_{\mathrm{cav}} and τth\tau_{\mathrm{th}}. Shaded areas mark the standard error over realizations. The weight stays finite in windows opened far beyond τcav\tau_{\mathrm{cav}} and falls to zero only near τth\tau_{\mathrm{th}}. Period doubling is therefore re-established after every time-soliton and lost only when the lattice thermalizes. All panels share L=100L=100, J=5J=5, Mosc=kB=λ=1M_{\mathrm{osc}}=k_{\mathrm{B}}=\lambda=1, θinit=0.2\theta_{\mathrm{init}}=0.2, Np=104N_p=10^4, Ntraj=100N_{\mathrm{traj}}=100, τosc=ε0=0.001\tau_{\mathrm{osc}}=\varepsilon_0=0.001, and Tbath/T0=1T_{\mathrm{bath}}/T_0=1. See text for details.

Despite the presence of a sharp subharmonic response, the state lacks the long-range temporal coherence of a time crystal. Instead, it can be viewed as a DTC containing a finite density of time-solitons. Individual trajectories exhibit long intervals of period-doubled dynamics, interrupted by stochastic switches between the two equivalent temporal configurations. We identify these switches as time-solitons: topological defects in temporal order that locally shift the phase of the period-doubled response. Their finite density destroys long-range temporal coherence while preserving local time-crystalline order. We term this defect-rich state a discrete time liquid (DTL), in analogy with the melting of a spatial crystal whose long-range order is destroyed by the proliferation of topological defects [28]. This picture motivates a minimal stochastic deformation of the DTC evolution in Eq. (1),

U^n=eiJ∑⟨ij⟩S^izS^jzei(π+εn)∑iS^ix,εn={πprob. p0prob. 1−p,(2)\hat{U}_n=e^{iJ\sum_{\langle ij\rangle}\hat{S}_i^z\hat{S}_j^z}e^{i(\pi+\varepsilon_n)\sum_i\hat{S}_i^x},\qquad\varepsilon_n=\begin{cases}\pi& \text{prob. }p\\0 & \text{prob. }1-p,\end{cases} \tag*{(2)}

where pp sets the density of time-solitons: with probability pp, the usual π\pi spin flip is replaced by a 2π2\pi rotation, producing a switch between the two temporal configurations. We show that this minimal stochastic-switching model reproduces not only the qualitative phenomenology of the cavity-spin system, but also quantitatively matches the distribution of its temporal correlations.

The DTL regime emerges only once thermal fluctuations become strong enough to switch the interaction and the transverse field on and off out-of-phase within a single cavity cycle. This leads to a threshold temperature for the onset of period-doubled order. Counterintuitively, temporal order therefore appears upon heating, providing yet another example in which noise promotes order rather than destroying it. Many-body interactions, in turn, stabilize the resulting state, allowing the DTL phase to persist over a broad temperature window, as shown in the phase diagram (Figure 1(c)).

Spins coupled to a thermal cavity mode

We consider a two-dimensional spin lattice coupled to a single damped cavity mode, which we model as a harmonic oscillator in contact with a thermal bath at temperature TbathT_{\mathrm{bath}} (sketched in Figure 1(a)). Crucially, no external periodic drive is applied, so that the thermal fluctuations of the bath are the only time-dependent input to the total system. To access long-time dynamics at large system sizes, we treat the spins classically and consider the spin Hamiltonian:

Hspin(t)=−Jeff(t)∑⟨ij⟩SizSjz−geff(t)∑iSix,(3)H_{\mathrm{spin}}(t)=-J_{\mathrm{eff}}(t)\sum_{\langle ij\rangle}S_i^zS_j^z-g_{\mathrm{eff}}(t)\sum_iS_i^x, \tag*{(3)}
Jeff(t)=J[1+λy(t)],geff(t)=g0[1−λy(t)],(4)J_{\mathrm{eff}}(t)=J\left[1+\lambda y(t)\right],\qquad g_{\mathrm{eff}}(t)=g_0\left[1-\lambda y(t)\right], \tag*{(4)}

where ⟨ij⟩\langle ij\rangle denote nearest-neighbor pairs on an L×LL\times L square lattice with periodic boundary conditions, SiαS_i^\alpha is the α\alpha-component of the classical unit spin at site ii, JJ is the bare Ising coupling, g0g_0 is the bare transverse-field strength, y(t)y(t) is the cavity oscillator coordinate, and λ\lambda is the spin–cavity coupling. The cavity coordinate obeys the Langevin equation [29]:

y¨+γy˙+ω02y=ξ(t),(5)\ddot{y}+\gamma\dot{y}+\omega_0^2y=\xi(t), \tag*{(5)}

where γ\gamma is the damping rate, ω0\omega_0 is the natural frequency of the cavity, and ξ(t)\xi(t) is Gaussian white noise with zero mean and correlator ⟨ξ(t)ξ(t′)⟩=Dδ(t−t′)\langle\xi(t)\xi(t')\rangle=D\delta(t-t'). The noise strength D=2γkBTbath/(MoscL2)D=2\gamma k_{\mathrm{B}}T_{\mathrm{bath}}/(M_{\mathrm{osc}}L^2) is fixed by the bath temperature, with MoscM_{\mathrm{osc}} the oscillator mass and LL the linear system size (see the microscopic derivation in the Supplemental Material (SM) [29]). In the underdamped regime γ≪ω0\gamma\ll\omega_0, the cavity mode acts as a spectral filter for the white noise. As a result, the thermal fluctuations of the bath generate an approximately periodic signal y(t)y(t) around the damped frequency ωc=ω02−(γ/2)2≃ω0\omega_c=\sqrt{\omega_0^2-(\gamma/2)^2}\simeq\omega_0, with period τosc=2π/ωc\tau_{\mathrm{osc}}=2\pi/\omega_c. This signal therefore acts on the spins as an effective, noisy periodic drive. Unlike the externally imposed periodic drive of a Floquet DTC [14, 16, 18], its amplitude is set by the bath temperature, and its phase stays coherent only over a finite coherence time of the cavity mode. Throughout, “drive” refers to this effective drive. In the following, we show that it can induce temporal order in the spin lattice.

Period doubling induced by thermal noise

We first establish the connection to the standard DTC of Eq. (1). For times short compared with the coherence time of the mode, τcav=2/γ\tau_{\mathrm{cav}} = 2/\gamma [29], the oscillator coordinate takes the form y(t)≃Rcos⁡(ωct+φ)y(t) \simeq R \cos(\omega_{\mathrm{c}}t + \varphi), with a slowly varying amplitude RR and phase φ\varphi. The typical amplitude grows with the bath temperature, ⟨R2⟩=2Tbath/T0\langle R^2 \rangle= 2T_{\mathrm{bath}}/T_0, where T0≡Moscωc2L2/kBT_0 \equiv M_{\mathrm{osc}}\omega_{\mathrm{c}}^2L^2/k_{\mathrm{B}} is a reference temperature [29]. When R=λ=1R = \lambda= 1, the interaction and the transverse field in Eq. (4) are switched on out-of-phase. The dynamics over one period then approximately reproduce the kicked evolution of Eq. (1), the model underlying Floquet DTCs [14, 16, 18]. We thus set g0=(π+ε0)/τoscg_0 = (\pi+ \varepsilon_0)/\tau_{\mathrm{osc}}, so that the spins rotate by π+ε0\pi+ \varepsilon_0 per period, with a fixed deviation ε0\varepsilon_0 from the ideal π\pi kick. The bath temperature then controls how closely the thermal fluctuations realize this kicked evolution: a weak drive cannot separate the interaction from the transverse field in time, whereas a strong drive is too erratic. Period-doubled order is therefore expected to emerge only at intermediate bath temperatures.

To test this expectation, we start from the extremely underdamped regime, τcav/τosc=105\tau_{\mathrm{cav}}/\tau_{\mathrm{osc}} = 10^5, which is longer than the observation window of Np=104N_p = 10^4 stroboscopic periods. The phase φ\varphi is then effectively fixed in each realization, and the amplitude RR is the dominant stochastic variable. To quantify the response, we consider the stroboscopic spin magnetization Mz(nτosc)=L−2∑iSiz(nτosc)M^z(n\tau_{\mathrm{osc}}) = L^{-2}\sum_i S_i^z(n\tau_{\mathrm{osc}}), n∈Zn \in\mathbb{Z}, and the Fourier transform of its noise average, F~[⟨ ⁣⟨Mz⟩ ⁣⟩](ν)\widetilde{F}[\langle\!\langle M^z\rangle\!\rangle](\nu). Here, ⟨ ⁣⟨⋯ ⟩ ⁣⟩\langle\!\langle\cdots\rangle\!\rangle denotes the average over noise realizations. The transform F~[f](ν)=τosc∑n=0Np−1e−iνnf(nτosc)\widetilde{F}[f](\nu) = \tau_{\mathrm{osc}}\sum_{n=0}^{N_p-1}e^{-i\nu n}f(n\tau_{\mathrm{osc}}), is the discrete counterpart of ∫dt e−iωtf(t)\int dt\,e^{-i\omega t}f(t), with ν=ωτosc\nu= \omega\tau_{\mathrm{osc}} the dimensionless frequency, so that F~\widetilde{F} carries units of time. We initialize the lattice close to full polarization along +z+z, tilting each spin by a small random angle within a cone of half-angle θinit=0.2\theta_{\mathrm{init}} = 0.2. This tilt mimics the thermal fluctuations of a physical spin system and shows the robustness of the temporal order to different initial conditions. We then solve the coupled dynamics numerically (see “Initial conditions and simulation parameters” in Methods). Indeed, the subharmonic response at ν=π\nu= \pi emerges within a finite window of bath temperature (see the bright vertical stripe in Figure 1(b)). It is absent at low TbathT_{\mathrm{bath}}, where side peaks reflect the precessional dynamics, and it is washed out at high TbathT_{\mathrm{bath}}, where the noise is too strong. More intuitively, at high TbathT_{\mathrm{bath}} the local field Bi(t)\mathbf{B}_i(t) inherits a large amplitude from y(t)y(t). A tiny spatial difference in the spin configuration then produces a large difference in the local field, which in turn amplifies the spatial difference after a short time. As a result, at high TbathT_{\mathrm{bath}} the spin lattice develops chaotic dynamics and loses temporal order (top row of Figure 1(a)).

We note that the period-doubled response at intermediate bath temperature is absent from the drive, which carries weight only near ν=2π\nu= 2\pi [29]. The response is also robust against randomness in the initial spin configuration and against perturbations of the kick deviation ε0\varepsilon_0 [29]. The period-doubled signal is therefore generated by the spin lattice, not inherited from the drive. To understand its origin, we compare the stochastic dynamics with a deterministic reference drive in Figure 2.

Thermal fluctuations as an effective Floquet drive

Figure 2. Thermal fluctuations as an effective Floquet drive. Columns compare a deterministic reference drive (a,d,g), in which the oscillator is replaced by a sinusoid, with single realizations of the bath-driven dynamics at a small temperature Tbath/T0=0.01T_{\mathrm{bath}}/T_0 = 0.01 (b,e,h, shaded in blue) and at a temperature inside the ordering window Tbath/T0=0.9T_{\mathrm{bath}}/T_0 = 0.9 (c,f,i, shaded in orange). (a–c) Oscillator coordinate y(t)y(t), which sets the instantaneous interaction and transverse field, with gray dots marking its stroboscopic values. The deterministic drive is a sinusoid of fixed amplitude (a), whereas the stochastic drive is a small (b) or a large (c) fluctuating signal. The bath temperature thus sets the amplitude of the drive. (d–f) Corresponding spin response Mz(t)M^z(t). The deterministic drive gives long-lived period doubling (d). The weak stochastic drive produces a response close to the free precession of the undriven spin lattice, with no period doubling left after 10310^3 periods (e). The large stochastic drive sustains period doubling beyond 10310^3 periods (f). A stochastic drive of sufficient amplitude therefore reproduces the long-lived period doubling of the deterministic drive. (g) Correlation time τcorr/τosc\tau_{\mathrm{corr}}/\tau_{\mathrm{osc}} of Eq. (6) for the deterministic drive, over the plane of kick deviation ε0\varepsilon_0 and drive amplitude R0R_0. The cross marks the drive involved in (a,d), with ε0=0.1\varepsilon_0 = 0.1 and R0=1R_0 = 1. Long-lived period doubling appears in a broad region in (ε0,R0)(\varepsilon_0, R_0), so the DTC does not require fine tuning of the drive amplitude. (h,i) Per-cycle effective drive parameters {(εcycle(n),Rcycle(n))}n\{(\varepsilon_{\mathrm{cycle}}(n), R_{\mathrm{cycle}}(n))\}_n of the realizations in (b) and (c) (see “Per-cycle effective drive parameters” in Methods), overlaid on the τcorr/τosc=2×103\tau_{\mathrm{corr}}/\tau_{\mathrm{osc}} = 2 \times10^3 contour of (g). This contour separates the DTC band from the surrounding region of short-lived period doubling. The per-cycle effective drive parameters lie below the band at Tbath/T0=0.01T_{\mathrm{bath}}/T_0 = 0.01 (h) and fall within it at Tbath/T0=0.9T_{\mathrm{bath}}/T_0 = 0.9 (i). Period doubling therefore appears when the per-cycle cloud overlaps the DTC band, which explains the …

To this end, we replace the oscillator by the deterministic drive y(t)=R0cos⁡(ωct+ϕ0)y(t) = R_0\cos(\omega_{\mathrm{c}}t + \phi_0) with fixed amplitude R0R_0 and phase ϕ0\phi_0, i.e., the monochromatic limit of the stochastic signal (see Figure 2(a)). This reference drive plays the role of the externally imposed drive of a Floquet DTC and is not part of the cavity-spin system. The lifetime of the period doubling depends on ϕ0\phi_0, but this dependence weakens as the drive frequency increases [29]. We set ϕ0=0\phi_0 = 0, where the lifetime is longest. To measure how long the period doubling persists, we use the normalized staggered magnetization s(n)≡(−1)nMz(nτosc)/Mz(0)s(n) \equiv(-1)^n M^z(n\tau_{\mathrm{osc}})/M^z(0), which stays at s(n)=1s(n) = 1 for an ideal period-doubled response. We then define the correlation time

τcorr=τosc∣∑n≥0s(n)∣(6)\tau_{\mathrm{corr}} = \tau_{\mathrm{osc}}\left|\sum_{n\geq0}s(n)\right| \tag*{(6)}

over which a trajectory keeps the memory of its initial subharmonic phase. For an exponential decay, s(n)=e−nτosc/τcorrs(n) = e^{-n\tau_{\mathrm{osc}}/\tau_{\mathrm{corr}}} with τcorr≫τosc\tau_{\mathrm{corr}} \gg\tau_{\mathrm{osc}}, Eq. (6) returns the decay time. Equivalently, τcorr=∣F~[Mz](π)∣/∣Mz(0)∣\tau_{\mathrm{corr}} = |\widetilde{F}[M^z](\pi)|/|M^z(0)| is the height of the subharmonic peak of the Fourier transform used in Figure 1(b), normalized by the initial magnetization. Evaluating τcorr\tau_{\mathrm{corr}} over a grid of (ε0,R0)(\varepsilon_0,R_0) gives the DTC phase diagram of Figure 2(g), in which the bright band marks a long-lived subharmonic response, with typical dynamics shown in Figure 2(d). As in prethermal DTCs [20, 30, 31], τcorr\tau_{\mathrm{corr}} grows exponentially with the drive frequency ωc\omega_{\mathrm{c}} [29].

With the reference phase diagram of Figure 2(g) at hand, we now turn to the stochastic drive. At a low temperature, Tbath/T0=0.01T_{\mathrm{bath}}/T_0 = 0.01, the drive amplitude is small (Figure 2(b)), so that the drive cannot separate the interaction from the transverse field in time. The response then stays close to the free precession of the undriven lattice (Figure 2(e)), which gives the side peaks in Figure 1(b). To compare such a stochastic drive with the deterministic one, we assign to every cavity cycle nn an effective kick deviation εcycle(n)\varepsilon_{\mathrm{cycle}}(n) and an effective drive amplitude Rcycle(n)R_{\mathrm{cycle}}(n) (see “Per-cycle effective drive parameters” in Methods). Each realization thereby becomes a cloud of points in the plane of Figure 2(g). At low temperature, this cloud lies far below the DTC band (Figure 2(h)). Increasing the temperature pushes the cloud upward, since the drive amplitude follows a Rayleigh distribution with ⟨R2⟩=2Tbath/T0\langle R^2\rangle= 2T_{\mathrm{bath}}/T_0 [29]. This distribution is broad, so that the amplitude differs strongly between realizations. The DTC band of Figure 2(g), however, is broad in R0R_0 as well, a consequence of the many-body interactions. In an intermediate temperature window the two therefore overlap substantially, and the period doubling survives without any fine tuning of its amplitude. Indeed, at Tbath/T0=0.9T_{\mathrm{bath}}/T_0 = 0.9 the drive is a strong fluctuating signal (Figure 2(c)) whose cloud sits inside the DTC band (Figure 2(i)). The period doubling then persists beyond a thousand cycles (Figure 2(f)).

doubling grows exponentially with the cavity frequency [29], we take a high cavity frequency, τosc=0.001\tau_{\mathrm{osc}} = 0.001, and a small kick deviation, ε0=0.001\varepsilon_0 = 0.001. We fix the bath temperature at Tbath/T0=1T_{\mathrm{bath}}/T_0 = 1, well inside the ordered region of Figure 2(g). We then analyze the period-doubled response as a function of the cavity coherence time τcav\tau_{\mathrm{cav}}. Figure 3(a) shows the single-trajectory dynamics at two coherence times. The period-doubled component (−1)nMz(nτosc)(-1)^n M^z(n\tau_{\mathrm{osc}}) is repeatedly suppressed and revives with the opposite sign. Once the phase of y(t)y(t) has drifted by half a period, the lattice rebuilds its subharmonic response against the new phase. The two period-doubled configurations are thereby exchanged. We identify each such

Time-solitons and the discrete time liquid

We now consider the consequence of a short coherence time of the cavity mode on the temporal order. This finite coherence time has no counterpart in an externally imposed periodic drive. To see this effect, we first make the spin order as robust as possible. Since the lifetime of the period

@endsection event as a time-soliton, the εn=π\varepsilon_n=\pi event of Eq. (2), occurring at a time set by the noise. What limits the temporal order is therefore the proliferation of time-solitons, not the loss of period doubling.

To distinguish different ways of losing the temporal order, we use the normalized staggered magnetization s(n)s(n) of Eq. (6), which satisfy s(0)=1s(0)=1 in every realization. We introduce the thermalization time and two correlation times before and after ensemble averaging. We define the correlation time of the ensemble-averaged response as:

τcorrens≡τosc∥∑n≥0⟦s(n)⟧∥,(7)\tau_{\mathrm{corr}}^{\mathrm{ens}} \equiv\tau_{\mathrm{osc}}\left\lVert\sum_{n\geq0}\llbracket s(n)\rrbracket\right\rVert, \tag*{(7)}

where we average the staggered magnetization first and then compute its correlation time. Then we define the averaged single-trajectory correlation time, ⟦τcorr⟧\llbracket\tau_{\mathrm{corr}}\rrbracket, where we compute the correlation time τcorr\tau_{\mathrm{corr}} of Eq. (6) for each individual trajectory, and then average over trajectories. Finally, we define the thermalization time τth\tau_{\mathrm{th}} as the overall decay of magnetization irrespective of the sign, which captures the loss of spatial order due to the slow heating,

τth≡τosc∑n≥0∥⟦s(n)⟧∥.(8)\tau_{\mathrm{th}}\equiv\tau_{\mathrm{osc}}\sum_{n\geq0}\left\lVert\llbracket s(n)\rrbracket\right\rVert. \tag*{(8)}

Here, τth\tau_{\mathrm{th}} is insensitive to the subharmonic phase and measures how long the period doubling survives. In contrast, ⟦τcorr⟧\llbracket\tau_{\mathrm{corr}}\rrbracket measures how long one realization spends in its dominant crystalline configuration, and τcorrens\tau_{\mathrm{corr}}^{\mathrm{ens}} measures how long the ensemble shares a phase. The three obey the inequality τcorrens≤⟦τcorr⟧≤τth\tau_{\mathrm{corr}}^{\mathrm{ens}}\leq\llbracket\tau_{\mathrm{corr}}\rrbracket\leq\tau_{\mathrm{th}}, with equality when no soliton occurs within the

prethermal window. In the presence of time-solitons, the three time scales separate.

Figure 3(b) shows the three time scales defined above as a function of the cavity coherence time τcav\tau_{\mathrm{cav}}. We see that the thermalization time depends only weakly on τcav\tau_{\mathrm{cav}}. As in DTCs, the thermalization time is mainly controlled by the cavity frequency. We show its scaling with the cavity frequency in the SM [29]. As mentioned earlier, the single-trajectory correlation time ⟦τcorr⟧\llbracket\tau_{\mathrm{corr}}\rrbracket in Figure 3(b) lies in between the other two scales. Its distribution over trajectories is shown by the green density plot in Figure 3(b).

To check quantitatively whether this distribution of correlation times can be captured by a simpler model that contains only uncorrelated time-solitons, we propose a classical version of the quantum toy model of Eq. (2), described as:

s(n)=e−nτosc/τth×σ(n),(9)s(n)=e^{-n\tau_{\mathrm{osc}}/\tau_{\mathrm{th}}}\times\sigma(n), \tag*{(9)}

where σ(n)=±1\sigma(n)=\pm1 is a random variable that flips sign with probability pp in each period:

σ(n)=σ(n−1)×{−1prob. p+1prob. 1−p.(10)\sigma(n)=\sigma(n-1)\times\begin{cases} -1 & \text{prob. }p\\ +1 & \text{prob. }1-p. \end{cases} \tag*{(10)}

Using this model, we show in the Methods that the density of time-solitons, ρsol\rho_{\mathrm{sol}}, is determined by the flip probability, ρsol=p/τosc\rho_{\mathrm{sol}}=p/\tau_{\mathrm{osc}}, and that the ensemble correlation time is τcorrens=1/(2ρsol+τth−1)\tau_{\mathrm{corr}}^{\mathrm{ens}}=1/(2\rho_{\mathrm{sol}}+\tau_{\mathrm{th}}^{-1}). Remarkably, despite its simplicity this model reproduces the full probability distribution of the single-trajectory correlation time of the many-body cavity-spin system, as shown in Figure 4. To obtain this agreement, we only match τth\tau_{\mathrm{th}} and τcorrens\tau_{\mathrm{corr}}^{\mathrm{ens}} between the toy model and the full simulation (see SM [29] for more evidence at other values of τcav\tau_{\mathrm{cav}}). This agreement provides confidence that the time-soliton description captures the dynamics of the DTL. Finally, we show in Figure 3(b) that the correlation time of the averaged subharmonic response matches the coherence time of the cavity mode, τcorrens≃τcav\tau_{\mathrm{corr}}^{\mathrm{ens}}\simeq\tau_{\mathrm{cav}} in the limit of τcav≪τth\tau_{\mathrm{cav}}\ll\tau_{\mathrm{th}}. This implies that the cavity lifetime directly specifies the density of time-solitons as ρsol≃1/(2τcav)\rho_{\mathrm{sol}}\simeq1/(2\tau_{\mathrm{cav}}). In the other limit of τcav≫τth\tau_{\mathrm{cav}}\gg\tau_{\mathrm{th}}, τcorrens\tau_{\mathrm{corr}}^{\mathrm{ens}} saturates to the thermalization time τth\tau_{\mathrm{th}}; here, the spatial order is lost before a single time-soliton has time to occur. The system then keeps a single subharmonic phase throughout its lifetime, which is the defining property of a prethermal DTC. The DTC and the DTL are therefore distinguished by the number of time-solitons within the lifetime, ρsolτth\rho_{\mathrm{sol}}\tau_{\mathrm{th}}.

Distribution of the single-trajectory correlation time

Figure 4. Distribution of the single-trajectory correlation time. Probability p(τcorr)p(\tau_{\mathrm{corr}}) of the single-trajectory correlation time τcorr\tau_{\mathrm{corr}} of Eq. (6), for the cavity-spin system (green histogram) and for the toy model of Eqs. (9) and (10) (dashed line), at τcav/τosc=20\tau_{\mathrm{cav}}/\tau_{\mathrm{osc}}=20. The toy parameters are not fitted: τth\tau_{\mathrm{th}} is the value measured on the lattice, and the soliton density ρsol\rho_{\mathrm{sol}} follows from the measured τcorrens\tau_{\mathrm{corr}}^{\mathrm{ens}}, which gives ρsolτth≈37\rho_{\mathrm{sol}}\tau_{\mathrm{th}}\approx37 (top). The vertical lines mark τcorrens\tau_{\mathrm{corr}}^{\mathrm{ens}}, ∥τcorr∥\lVert\tau_{\mathrm{corr}}\rVert, and τth\tau_{\mathrm{th}} from the line cut at τcav/τosc=20\tau_{\mathrm{cav}}/\tau_{\mathrm{osc}}=20 in Figure 3(b). The toy model reproduces the probability distribution and thus captures the dynamics of the DTL in cavity-spin system. Parameters are as in Figure 3. See text for details.

We further test the time-soliton picture in the cavity-spin system. Equations (9) and (10) predict that a time-soliton affects only the phase of the subharmonic response and leaves its amplitude intact, up to an overall extremely slow thermalization decay. To test this, we measure the local period-doubled order in a window of Wcav=⌊τcav/τosc⌉W_{\mathrm{cav}}=\lfloor\tau_{\mathrm{cav}}/\tau_{\mathrm{osc}}\rceil periods opened at the stroboscopic period mm,

S(m)=1Wcav∣∑k=0Wcav−1[s(m+k)s(m)]∣.(11)S(m)=\frac{1}{W_{\mathrm{cav}}}\left|\sum_{k=0}^{W_{\mathrm{cav}}-1}\left\lbrack s(m+k)s(m)\right\rbrack\right|. \tag*{(11)}

Because it averages the product of two magnetizations, S(m)S(m) never refers to the subharmonic phase of any single realization. It therefore still registers period doubling after the solitons have scrambled that phase across the ensemble. As shown in Figure 3(c), the weight S(m)S(m) stays large in windows opened far beyond τcav\tau_{\mathrm{cav}} and falls only near the thermalization time τth\tau_{\mathrm{th}}. At this time, the lattice has absorbed enough energy to melt the spatial order and completely lose the period doubling. Time-solitons only reset the phase of the subharmonic response and leave its amplitude approximately intact for times between τcav\tau_{\mathrm{cav}} and τth\tau_{\mathrm{th}}.

Experimental observability

In ultracold atoms, our proposal is closely related to optically driven Bose gases coupled to an optical cavity [32]. In this geometry, an off-resonant laser typically pumps the Bose–Einstein condensate directly, but the cavity mode can instead be pumped. Driving the cavity directly with white noise at ultralow temperature would then realize our model, with the noise amplitude playing the role of the bath temperature in our calculation.

In condensed matter, a promising platform involves surface phonon polaritons (SPPs), which host ultrastrong electric-field fluctuations [10, 11, 12, 33, 34]. Although noise from SPPs involves multiple modes, the fluctuations are laterally coherent over a distance dd for a two-dimensional layer located a distance dd from the interface. Our model could therefore be realized using a two-dimensional array of dipoles with ferroelectric interactions at this separation. The electric-field fluctuations experienced by the dipoles scale as ⟨E2⟩SPP∼ℏωSPPϵ0d3(1+2nb(ℏωSPPkBT))⟶kBTϵ0d3\langle E^{2}\rangle_{\mathrm{SPP}}\sim\frac{\hbar\omega_{\mathrm{SPP}}}{\epsilon_{0}d^{3}}\left(1+2n_{\mathrm{b}}\left(\frac{\hbar\omega_{\mathrm{SPP}}}{k_{\mathrm{B}}T}\right)\right)\longrightarrow\frac{k_{\mathrm{B}}T}{\epsilon_{0}d^{3}} [10, 11, 12, 33, 34], where EE is the electric field, ωSPP\omega_{\mathrm{SPP}} the SPP frequency, ϵ0\epsilon_{0} the vacuum permittivity, kBk_{\mathrm{B}} Boltzmann’s constant, TT temperature, nbn_{\mathrm{b}} the Bose–Einstein distribution and ℏ\hbar the reduced Planck constant. These fluctuations can be comparable to the dipoles’ internal dynamics, as required by our model [12].

For both platforms, the main challenge is engineering the prescribed coupling of the cavity to both the interaction and kick terms. As in DTC experiments, we expect our model to be one example of a broader class exhibiting DTL order. In particular, we expect that any DTC model featuring a probabilistic switch as described in Eq. (2) would lead to DTL phenomenology.

In both platforms, the main observable is the locking of spin–spin correlation functions to half the cavity frequency as a function of fluctuation amplitude. Cavity-coupled cold gases are particularly attractive because individual trajectories can be imaged, allowing time-solitons to be resolved and their distribution functions measured.

Outlook

In this article, we demonstrate a new form of temporal order, in which robust subharmonic frequency locking coexists with only short-range temporal coherence. This combination arises from the presence of temporal topological defects, which disrupt long-range coherence but preserve the characteristic subharmonic response.

Our results establish a connection between many-body temporal order and cavity–matter hybrids. Considerable efforts have been devoted to reaching the ultrastrong light–matter coupling regime [35], yet the novel collective phenomena that may emerge in this regime remain largely unexplored. Here, we show that even a relatively simple system of interacting two-dimensional spins coupled to a cavity can exhibit unconventional dynamics governed by temporal order. This suggests that temporal ordering phenomena may play an important role in more complex cavity–matter systems, including electrons in a material lattice coupled ultrastrongly to light.

From the perspective of temporal order, our results point towards a broader landscape of discrete time liquids (DTLs). The fact that the complex spin–cavity dynamics can be captured by a simple toy model governed by a single parameter, the time-soliton density, suggests that the phase studied here may belong to a broader universality class. More generally, DTLs may emerge through different mechanisms and exhibit distinct universality classes, potentially determined by the symmetries and correlations of their temporal defects. Two particularly promising directions for realizing such distinct classes are fluctuating quasiperiodic driving and spin–cavity backaction [36]. In the former, multiple incommensurate cavity frequencies could introduce additional synthetic dimensions [37, 38], potentially allowing for higher-dimensional temporal topological defects beyond the isolated time-solitons considered here. In the latter, backaction could mediate interactions between time-solitons, introducing correlations that enhance or suppress temporal coherence. Both mechanisms could thus give rise to DTLs with qualitatively different defect structures and collective dynamics. The possible existence of distinct DTL universality classes further raises the question of whether sharp transitions between temporal phases exhibit universal critical behaviour, analogous to phase transitions in spatially ordered systems.

Finally, the most striking aspect of DTLs, both quantum and classical, is the idea that a DTC with a finite density of topological time defects can restore time-translation symmetry on average. This restoration opens a pathway for temporal order to emerge in thermal equilibrium, thereby circumventing the standard no-go theorems [39, 40] that prohibit conventional time crystals in equilibrium in closed isolated systems. At the same time, the resulting phase retains many of the defining features of time-crystalline behavior, effectively bringing temporal order into the realm of equilibrium thermodynamics.

Methods

Model and equations of motion

We numerically simulate the dynamics of the two-dimensional classical spin model with Hamiltonian of Eq. (3), with the oscillator coordinate y(t)y(t) generated from the Langevin equation of Eq. (5). The resulting classical spin dynamics obeys

S⃗˙i=S⃗i×B⃗i(t),B⃗i(t)=(geff(t),0,Jeff(t)∑j∈nn(i)Sjz)(12)\dot{\vec{S}}_i=\vec{S}_i\times\vec{B}_i(t),\qquad\vec{B}_i(t)=\left(g_{\mathrm{eff}}(t),0,J_{\mathrm{eff}}(t)\sum_{j\in\mathrm{nn}(i)}S_j^z\right) \tag*{(12)}

where B⃗i(t)\vec{B}_i(t) is the local field and nn(i)\mathrm{nn}(i) denotes the four nearest neighbors of site ii on the periodic square lattice. We use units with Mosc=kB=1M_{\mathrm{osc}}=k_B=1 and keep L=100L=100, J=5J=5 and λ=1\lambda=1 fixed throughout. The backaction of the spins on the oscillator is neglected, so that y(t)y(t) can be generated independently of the spin configuration. Consequences of the back-action are discussed in the SM [29].

For the stochastic simulations, we adjust the bare oscillator frequency ω0\omega_0 for each damping rate γ\gamma such that the damped oscillation frequency satisfies

ωc=ω02−(γ2)2=2πτosc.(13)\omega_c=\sqrt{\omega_0^2-\left(\frac{\gamma}{2}\right)^2}=\frac{2\pi}{\tau_{\mathrm{osc}}}. \tag*{(13)}

This choice fixes the oscillation period at τosc\tau_{\mathrm{osc}} throughout the damping-rate sweeps. We generate an independent realization of y(t)y(t) from Eq. (5) for each realization of the Gaussian noise. For the deterministic reference calculations behind Figure 2, we instead use the sinusoidal signal y(t)=R0cos⁡(ωct)y(t)=R_0\cos(\omega_ct) with fixed amplitude R0R_0.

Numerical integration

We integrate Eq. (12) by a splitting between the transverse-field and interaction terms of Eq. (3). Within each time step dtdt, every spin is rotated about the xx axis by geffdt/2g_{\mathrm{eff}}dt/2, then about its local zz axis by Jeff∑j∈nn(i)SjzdtJ_{\mathrm{eff}}\sum_{j\in\mathrm{nn}(i)}S_j^zdt, and then about the xx axis by geffdt/2g_{\mathrm{eff}}dt/2 again. The Langevin equation, Eq. (5), is propagated on the same grid, with the noise sampled as ξn=D/dtηn\xi_n=\sqrt{D/dt}\eta_n, ηn∼N(0,1)\eta_n\sim\mathcal{N}(0,1), and held constant within the step. The noise is generated independently at each step and for every noise realization. With small enough dtdt, this trotterization method gives converged dynamics as we show in the SM [29].

Initial conditions and simulation parameters

We initialize each spin independently and uniformly within a spherical cap of half-angle θinit\theta_{\mathrm{init}} about the +z+z axis:

S⃗i(0)=(sin⁡θicos⁡ϕi,sin⁡θisin⁡ϕi,cos⁡θi),(14)\vec{S}_i(0)=\left(\sin\theta_i\cos\phi_i,\sin\theta_i\sin\phi_i,\cos\theta_i\right), \tag*{(14)}

where ϕi∼U[0,2π)\phi_i\sim U[0,2\pi) and cos⁡θi∼U[cos⁡θinit,1]\cos\theta_i\sim U[\cos\theta_{\mathrm{init}},1]. The cap half-angle θinit\theta_{\mathrm{init}} sets how far the initial state departs from the perfectly polarized configuration, which is recovered at θinit=0\theta_{\mathrm{init}}=0. We use θinit=0.2\theta_{\mathrm{init}}=0.2 in each figure of the main text, so that the lattice starts close to the ordered configuration of the Ising interaction. Before initiating the spin dynamics, we evolve the oscillator from y=y˙=0y=\dot{y}=0 for a time 5/γ5/\gamma to approach its stationary state.

We resolve the cavity period τosc\tau_{\mathrm{osc}} with at least 20 time steps, so that dt=5×10−3dt=5\times10^{-3} at τosc=0.2\tau_{\mathrm{osc}}=0.2, dt=5×10−4dt=5\times10^{-4} at τosc=0.01\tau_{\mathrm{osc}}=0.01, and dt=5×10−5dt=5\times10^{-5} at τosc=10−3\tau_{\mathrm{osc}}=10^{-3}. We record Np=104N_p=10^4 periods τosc\tau_{\mathrm{osc}} in every run. Noise averages are taken over Ntraj=100N_{\mathrm{traj}}=100 independent noise realizations.

Per-cycle effective drive parameters

To compare a stochastic drive with the deterministic drive y(t)=R0cos⁡(ωct)y(t) = R_0 \cos(\omega_{\mathrm{c}}t), we assign to cycle nn the effective kick deviation and amplitude

εcycle(n)=ε0−g0λ∫nτosc(n+1)τoscy(t) dt,(15)\varepsilon_{\mathrm{cycle}}(n) = \varepsilon_0 - g_0\lambda\int_{n\tau_{\mathrm{osc}}}^{(n+1)\tau_{\mathrm{osc}}} y(t)\,\mathrm{d}t, \tag*{(15)}
Rcycle(n)=2τosc∫nτosc(n+1)τoscy(t)2 dt.(16)R_{\mathrm{cycle}}(n) = \sqrt{\frac{2}{\tau_{\mathrm{osc}}}\int_{n\tau_{\mathrm{osc}}}^{(n+1)\tau_{\mathrm{osc}}} y(t)^2\,\mathrm{d}t}. \tag*{(16)}

Here, εcycle(n)\varepsilon_{\mathrm{cycle}}(n) is the deviation of the spin rotation accumulated in cycle nn from π\pi. The amplitude Rcycle(n)R_{\mathrm{cycle}}(n) is the root-mean-square of y(t)y(t) over the cycle, normalized such that Rcycle=RR_{\mathrm{cycle}} = R for a pure tone y=Rcos⁡(ωct+φ)y = R\cos(\omega_{\mathrm{c}}t+\varphi). A deterministic drive therefore maps to the single point (ε0,R0)(\varepsilon_0, R_0) in every cycle. For the stochastic drive, ⟨y⟩=0\langle y\rangle= 0 gives ⟨εcycle⟩=ε0\langle\varepsilon_{\mathrm{cycle}}\rangle= \varepsilon_0. In the underdamped limit, Rcycle≃RR_{\mathrm{cycle}}\simeq R follows the Rayleigh distribution of the drive amplitude. We analytically derive this result in the SM [29].

Toy model of time-solitons

In the toy model of Eqs. (9) and (10), a time-soliton occurs independently in each period with probability pp. Neighboring solitons are therefore separated by 1/p1/p periods on average, and the density of time-solitons is ρsol=p/τosc\rho_{\mathrm{sol}} = p/\tau_{\mathrm{osc}}. In the limit of τth≫τosc\tau_{\mathrm{th}}\gg\tau_{\mathrm{osc}} and ρsolτosc≪1\rho_{\mathrm{sol}}\tau_{\mathrm{osc}}\ll1, Eq. (7) gives

τcorrens=12ρsol+τth−1.(17)\tau_{\mathrm{corr}}^{\mathrm{ens}} = \frac{1}{2\rho_{\mathrm{sol}}+\tau_{\mathrm{th}}^{-1}}. \tag*{(17)}

We fix pp by this relation with the values of τth\tau_{\mathrm{th}} and τcorrens\tau_{\mathrm{corr}}^{\mathrm{ens}} measured in the cavity-spin system, so that the toy model has no free parameter.

The single-trajectory correlation time is τcorr=∣Z∣\tau_{\mathrm{corr}} = \lvert Z\rvert, with Z≡τosc∑k≥0akσ(k)=τosc(1+2W)/(1−a)Z\equiv\tau_{\mathrm{osc}}\sum_{k\geq0}a^k\sigma(k)=\tau_{\mathrm{osc}}(1+2W)/(1-a), a≡e−τosc/τtha\equiv e^{-\tau_{\mathrm{osc}}/\tau_{\mathrm{th}}}, and W≡∑i≥1(−1)iakiW\equiv\sum_{i\geq1}(-1)^i a^{k_i}. Here, k1<k2<⋯k_1<k_2<\cdots are the periods at which solitons occur. Since W=−ak1(1+W′)W=-a^{k_1}(1+W'), with W′W' sharing the same probability distribution with WW, the cumulative probability distribution P(W≤w)P(W\leq w) obeys

P(W≤w)=∑g≥1(1−p)g−1p[1−P(W≤−wa−g−1)].(18)P(W\leq w)=\sum_{g\geq1}(1-p)^{g-1}p\left[1-P\left(W\leq-wa^{-g}-1\right)\right]. \tag*{(18)}

We solve Eq. (18) iteratively and construct the probability distribution of ZZ and τcorr\tau_{\mathrm{corr}} (more details are found in the SM [29]).

Phase classification

Each point in the map of Figure 1(c) is classified by the correlation time ∥τcorr∥\lVert\tau_{\mathrm{corr}}\rVert of Eq. (6), the thermalization time τth\tau_{\mathrm{th}} of Eq. (8), and the precession weight

Pprec=max⁡(0,−min⁡kC(k)),C(k)≡∑m∥s(m+k)s(m)∥∑m∥s(m)2∥.(19)P_{\mathrm{prec}}=\max\left(0,-\min_k C(k)\right),\qquad C(k)\equiv\frac{\sum_m\lVert s(m+k)s(m)\rVert}{\sum_m\lVert s(m)^2\rVert}. \tag*{(19)}

where mm runs over a reference window of 100 periods starting at m=50m=50. The weight PprecP_{\mathrm{prec}} approaches unity when C(k)C(k) oscillates down to −1-1 and vanishes otherwise. We first identify the boundary of free precession as the lowest temperature at which PprecP_{\mathrm{prec}} has fallen to its high-temperature floor. We then label the remaining points as disordered when ∥τcorr∥<10τosc\lVert\tau_{\mathrm{corr}}\rVert<10\tau_{\mathrm{osc}}. All of the remaining points show local temporal order. We label them a DTC if ∥τcorr∥/τth\lVert\tau_{\mathrm{corr}}\rVert/\tau_{\mathrm{th}} exceeds 0.418, and a DTL otherwise. More details on the phase classification approach can be found in the SM [29].

Data availability.—The data and code associated with this manuscript are deposited in DOI: 10.5281/zenodo.23168407. Automatic access is restricted; access credentials will be provided to human researchers upon reasonable request via Zenodo for academic and verification purposes.

Acknowledgments.—We thank Duillio De Santis, Hope Bretcher and Sambuddha Chattopadhyay for valuable discussions. The authors used Claude Opus 5.5 during the preparation of this manuscript to assist with code optimization and to improve the readability of the manuscript. The authors assume full responsibility for the scientific content, accuracy, and integrity of this paper.

Funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) – Project number 579633640 (Emmy Noether Programme, “Light–Matter Hybrids”). Funded by the European Union (ERC, QuSimCtrl, 101113633). Views and opinions expressed are, however, those of the authors only and do not necessarily reflect those of the European Union or the European Research Council Executive Agency. Neither the European Union nor the granting authority can be held responsible for them. Z.W. acknowledges financial support from the China Scholarship Council (CSC, No. 202506750006) and the NKRPDCC (Grant No. 2022YFA1405304).

Author contributions.—M.H.M. conceived the research with M.B.; Z.W. set up the theoretical framework and performed the numerical simulations with support from P.M.S., M.B., and M.H.M. M.H.M. supervised the project with support from M.B. All authors contributed to interpreting the data. Z.W. and M.H.M. wrote the manuscript with input from P.M.S. and M.B.

Competing Interests Statement.—The authors declare no competing interests.

This Supplemental Material collects the derivations and diagnostics behind the main-text results. Sections S1 and S2 derive oscillator Langevin equation, Eq. (5), by eliminating the bath and give the stationary statistics of the filtered drive. Sections S3 and S4 show the robustness of the response and compare the drive and response spectra behind Figure 1. Sections S5 and S6 give the prethermal lifetime scaling at long and at short coherence times of the cavity mode. Section S7 shows how the drive phase affects the lifetime of the deterministic DTC of Figure 2. Section S8 derives the time scales of the toy model of Eqs. (9) and (10) and the distribution of single-trajectory correlations times behind Figure 4. Section S9 gives the numerical details and checks behind the phase diagram of Figure 1(c). Sections S10 and S11 discuss the spin backaction on the oscillator and the numerical methods.

Langevin equation from thermal bath

We derive the oscillator Langevin equation by eliminating the bath, following the standard treatment of a coordinate coupled to a harmonic reservoir [41].

Bath model and Langevin equation

The reservoir is a set of harmonic modes (coordinates XαX_\alpha, momenta PαP_\alpha, masses mαm_\alpha, frequencies ωα\omega_\alpha) coupled linearly to xx with strengths cαc_\alpha, plus a counterterm cancelling the bath-induced frequency shift,

Hbath=∑α[Pα22mα+12mαωα2Xα2],(20)H_{\mathrm{bath}} = \sum_\alpha\left[\frac{P_\alpha^2}{2m_\alpha} + \frac{1}{2}m_\alpha\omega_\alpha^2X_\alpha^2\right], \tag*{(20)}
Hosc−bath=−x∑αcαXα+x2∑αcα22mαωα2.(21)H_{\mathrm{osc-bath}} = -x\sum_\alpha c_\alpha X_\alpha+ x^2\sum_\alpha\frac{c_\alpha^2}{2m_\alpha\omega_\alpha^2}. \tag*{(21)}

The bare oscillator is Hosc=p2/2Mosc+12Moscω02x2H_{\mathrm{osc}} = p^2/2M_{\mathrm{osc}} + \frac{1}{2}M_{\mathrm{osc}}\omega_0^2x^2 (pp conjugate to xx, mass MoscM_{\mathrm{osc}}, natural frequency ω0\omega_0), coupled to the spins through λyH(−)\lambda y H^{(-)}. Here x=x/L\mathbf{x}=x/L is the intensive coordinate normalized in Sec. C, λ\lambda the coupling strength, and H(−)H^{(-)} (Eq. (S.11)) the extensive spin combination modulating JeffJ_{\mathrm{eff}} and geffg_{\mathrm{eff}}. Hamilton’s equations give

Moscx¨+Moscω02x+x∑αcα2mαωα2=−λLH(−)(t)+∑αcαXα,(22)M_{\mathrm{osc}}\ddot{x}+M_{\mathrm{osc}}\omega_0^2x+x\sum_\alpha\frac{c_\alpha^2}{m_\alpha\omega_\alpha^2}=-\frac{\lambda}{L}H^{(-)}(t)+\sum_\alpha c_\alpha X_\alpha, \tag*{(22)}
mαX¨α+mαωα2Xα=cαx.(23)m_\alpha\ddot{X}_\alpha+m_\alpha\omega_\alpha^2X_\alpha=c_\alpha x. \tag*{(23)}

Each bath mode (Eq. (S.4)) is driven by x(t)x(t); integrating from an initial time tit_i,

Xα(t)=Xα(h)(t)+cαmαωα∫titdt′ sin⁡[ωα(t−t′)]x(t′).(24)X_\alpha(t)=X_\alpha^{(\mathrm{h})}(t)+\frac{c_\alpha}{m_\alpha\omega_\alpha}\int_{t_i}^{t}dt'\,\sin[\omega_\alpha(t-t')]x(t'). \tag*{(24)}
Equation not extracted

with the free evolution of mode α\alpha,

Xα(h)(t)=Xα(ti)cos⁡[ωα(t−ti)]+Pα(ti)mαωαsin⁡[ωα(t−ti)],(25)X_{\alpha}^{(\mathrm{h})}(t)=X_{\alpha}(t_i)\cos[\omega_{\alpha}(t-t_i)]+\frac{P_{\alpha}(t_i)}{m_{\alpha}\omega_{\alpha}}\sin[\omega_{\alpha}(t-t_i)], \tag*{(25)}

set by the initial data Xα(ti)X_{\alpha}(t_i), Pα(ti)P_{\alpha}(t_i). Inserting Eq. Eq. (S.5) into Eq. Eq. (S.3) and integrating the memory term by parts, the instantaneous piece ∝x(t)\propto x(t) cancels the counterterm, leaving the generalized Langevin equation

Moscx¨+Moscω02x+∫titdt′ Γ(t−t′)x˙(t′)=Ξ(t)−λLH(−)(t),(26)M_{\mathrm{osc}}\ddot{x}+M_{\mathrm{osc}}\omega_0^2x+\int_{t_i}^{t}dt'\,\Gamma(t-t')\dot{x}(t')=\Xi(t)-\frac{\lambda}{L}H^{(-)}(t), \tag*{(26)}

with kernel Γ(t)=∑α(cα2/mαωα2)cos⁡(ωαt)\Gamma(t)=\sum_{\alpha}(c_{\alpha}^2/m_{\alpha}\omega_{\alpha}^2)\cos(\omega_{\alpha}t) and bath force Ξ(t)=∑αcαXα(h)(t)\Xi(t)=\sum_{\alpha}c_{\alpha}X_{\alpha}^{(\mathrm{h})}(t). A boundary term −x(ti)Γ(t−ti)-x(t_i)\Gamma(t-t_i) has been dropped (x(ti)=0x(t_i)=0).

Markovian limit and fluctuation–dissipation

When the bath correlation time is short, x˙\dot{x} is nearly constant over the memory range of Γ\Gamma and the integral becomes local,

∫titdt′ Γ(t−t′)x˙(t′)≃x˙(t)∫0∞ds Γ(s)≡Moscγx˙(t),(27)\int_{t_i}^{t}dt'\,\Gamma(t-t')\dot{x}(t')\simeq\dot{x}(t)\int_{0}^{\infty}ds\,\Gamma(s)\equiv M_{\mathrm{osc}}\gamma\dot{x}(t), \tag*{(27)}

defining the damping rate γ\gamma; equivalently Γ(t)→2Moscγ δ(t)\Gamma(t)\to2M_{\mathrm{osc}}\gamma\,\delta(t), the factor of two from the one-sided integral (∫0∞ds δ(s)=12\int_{0}^{\infty}ds\,\delta(s)=\frac{1}{2}). Equation Eq. (S.7) becomes

x¨+γx˙+ω02x=Ξ(t)Mosc−λMoscLH(−)(t).(28)\ddot{x}+\gamma\dot{x}+\omega_0^2x=\frac{\Xi(t)}{M_{\mathrm{osc}}}-\frac{\lambda}{M_{\mathrm{osc}}L}H^{(-)}(t). \tag*{(28)}

In thermal equilibrium at temperature TbathT_{\mathrm{bath}} the modes are independent, with ⟨Xα(ti)2⟩=kBTbath/(mαωα2)\langle X_{\alpha}(t_i)^2\rangle=k_{\mathrm{B}}T_{\mathrm{bath}}/(m_{\alpha}\omega_{\alpha}^2), ⟨Pα(ti)2⟩=mαkBTbath\langle P_{\alpha}(t_i)^2\rangle=m_{\alpha}k_{\mathrm{B}}T_{\mathrm{bath}}, and zero means and cross-correlation (kBk_{\mathrm{B}} Boltzmann’s constant). Then Ξ\Xi is zero-mean Gaussian and, from Eq. Eq. (S.6), obeys the classical fluctuation–dissipation relation ⟨Ξ(t)Ξ(t′)⟩=kBTbathΓ(t−t′)\langle\Xi(t)\Xi(t')\rangle=k_{\mathrm{B}}T_{\mathrm{bath}}\Gamma(t-t'). In the Markovian limit it is white,

⟨Ξ(t)Ξ(t′)⟩=2MoscγkBTbath δ(t−t′).(29)\langle\Xi(t)\Xi(t')\rangle=2M_{\mathrm{osc}}\gamma k_{\mathrm{B}}T_{\mathrm{bath}}\,\delta(t-t'). \tag*{(29)}

Equations Eqs. (S.9)–(S.10) govern the extensive coordinate xx; its intensive form follows in § 8 C.

Thermodynamic limit

The oscillator is a single cavity mode, x∝a^+a^†x\propto\hat{a}+\hat{a}^{\dagger}, whereas the spin combination it couples to,

H(−)=−∑⟨ij⟩JSizSjz+g0∑iSix,(30)H^{(-)}=-\sum_{\langle ij\rangle}J S_i^zS_j^z+g_0\sum_i S_i^x, \tag*{(30)}

the part of HspinH_{\mathrm{spin}} (Eq. Eq. (3)) multiplying xx, is extensive: H(−)=O(N)H^{(-)}=O(N) (N=L2N=L^2 spins). At fixed photon density ⟨a^†a^⟩/N=O(1)\langle\hat{a}^{\dagger}\hat{a}\rangle/N=O(1) the mode amplitude scales as x∼⟨a^†a^⟩∼Lx\sim\sqrt{\langle\hat{a}^{\dagger}\hat{a}\rangle}\sim L, so we use the intensive coordinate and per-spin force

y≡xL,h(−)(t)≡H(−)(t)N,(31)y\equiv\frac{x}{L},\qquad h^{(-)}(t)\equiv\frac{H^{(-)}(t)}{N}, \tag*{(31)}

both O(1)O(1), with h(−)(t)h^{(-)}(t) the instantaneous configuration. The coupling λyH(−)=O(N)\lambda yH^{(-)}=O(N) is then extensive while the modulation λy\lambda y stays O(1)O(1). Dividing Eq. Eq. (S.9) by LL (with H(−)=L2h(−)H^{(-)}=L^2h^{(-)}) gives

y¨+γy˙+ω02y=Ξ(t)MoscL−λMosch(−)(t).(32)\ddot{y}+\gamma\dot{y}+\omega_0^2y=\frac{\Xi(t)}{M_{\mathrm{osc}}L}-\frac{\lambda}{M_{\mathrm{osc}}}h^{(-)}(t). \tag*{(32)}

with O(1)O(1) backaction. The intensive noise ξ≡Ξ/(MoscL)\xi\equiv\Xi/(M_{\mathrm{osc}}L) is white by Eq. Eq. (S.10), the 1/L1/L rescaling its strength by 1/L21/L^2:

⟨ξ(t)ξ(t′)⟩=D δ(t−t′),D=2γkBTbathMoscL2,(33)\langle\xi(t)\xi(t')\rangle=D\,\delta(t-t'),\qquad D=\frac{2\gamma k_{\mathrm{B}}T_{\mathrm{bath}}}{M_{\mathrm{osc}}L^2}, \tag*{(33)}

the noise strength below Eq. Eq. (5). Equivalently, the drive amplitude σR2=kBTbath/(Moscω02L2)=Tbath/T0\sigma_R^2=k_{\mathrm{B}}T_{\mathrm{bath}}/(M_{\mathrm{osc}}\omega_0^2L^2)=T_{\mathrm{bath}}/T_0 is tuned in the main text, with the reference temperature T0≡Moscω02L2/kBT_0\equiv M_{\mathrm{osc}}\omega_0^2L^2/k_{\mathrm{B}} used on the figure axes. Dropping the backaction recovers Eq. Eq. (5), y¨+γy˙+ω02y=ξ(t)\ddot{y}+\gamma\dot{y}+\omega_0^2y=\xi(t).

Stationary statistics of the filtered drive

We derive the Rayleigh-amplitude, uniform-phase form of y(t)y(t). Without backaction Eq. Eq. (5) is linear; with y(ti)=y˙(ti)=0y(t_i)=\dot{y}(t_i)=0,

y(t)=∫titdt′ G(t−t′)ξ(t′),G(t)=e−γt/2ωcsin⁡(ωct)Θ(t),(34)y(t)=\int_{t_i}^{t}dt'\,G(t-t')\xi(t'),\qquad G(t)=\frac{e^{-\gamma t/2}}{\omega_c}\sin(\omega_ct)\Theta(t), \tag*{(34)}

with ωc=ω02−γ2/4\omega_c=\sqrt{\omega_0^2-\gamma^2/4}. For γ≪ω0\gamma\ll\omega_0, coarse-grained over ω0−1≪t≪γ−1\omega_0^{-1}\ll t\ll\gamma^{-1}, we write yy in quadratures,

y(t)=C(t)cos⁡ωct+S(t)sin⁡ωct,(35)y(t)=C(t)\cos\omega_ct+S(t)\sin\omega_ct, \tag*{(35)}

where C,SC,S are slowly varying (scale γ−1\gamma^{-1}) Gaussian processes. In the stationary limit t−ti≫γ−1t-t_i\gg\gamma^{-1} they are independent, zero-mean, and of equal variance,

⟨C2⟩=⟨S2⟩=σR2,⟨CS⟩=0,(36)\langle C^2\rangle=\langle S^2\rangle=\sigma_R^2,\qquad\langle CS\rangle=0, \tag*{(36)}

with

σR2=D2γω02=kBTbathMoscω02L2,(37)\sigma_R^2=\frac{D}{2\gamma\omega_0^2}=\frac{k_{\mathrm{B}}T_{\mathrm{bath}}}{M_{\mathrm{osc}}\omega_0^2L^2}, \tag*{(37)}

using Eq. Eq. (S.14). This variance is exact, fixed by equipartition: ⟨x2⟩=kBTbath/(Moscω02)\langle x^2\rangle=k_{\mathrm{B}}T_{\mathrm{bath}}/(M_{\mathrm{osc}}\omega_0^2) and ⟨y2⟩=⟨x2⟩/L2\langle y^2\rangle=\langle x^2\rangle/L^2; ωc=ω0\omega_c=\omega_0 sets the oscillator, not the amplitude scale. With C=Rcos⁡φC=R\cos\varphi, S=−Rsin⁡φS=-R\sin\varphi,

y(t)≃Rcos⁡(ωct+φ),(38)y(t)\simeq R\cos(\omega_ct+\varphi), \tag*{(38)}

with φ\varphi uniform on [0,2π)[0,2\pi) and the amplitude R=C2+S2R=\sqrt{C^2+S^2} Rayleigh-distributed,

fR(R)=RσR2e−R2/2σR2.(39)f_R(R)=\frac{R}{\sigma_R^2}e^{-R^2/2\sigma_R^2}. \tag*{(39)}

The drive strength ⟨R2⟩=2σR2∝Tbath\langle R^2\rangle=2\sigma_R^2\propto T_{\mathrm{bath}} thus vanishes as Tbath→0T_{\mathrm{bath}}\to0 and grows without bound at large TbathT_{\mathrm{bath}}. These two limits bounding the ordering window of Figure 1(b). Over times ≪γ−1\ll\gamma^{-1}, RR and φ\varphi are frozen, so the spins see a quasi-monochromatic drive of random amplitude and phase.

Since the Green function of Eq. (5) carries the envelope e−γt/2e^{-\gamma t/2}, the stationary correlation of the mode decays as ⟨y(t)y(0)⟩∝e−γt/2cos⁡(ωct)\langle y(t)y(0)\rangle\propto e^{-\gamma t/2}\cos(\omega_c t). The time that the mode keeps the memory of its phase is therefore

τcav=2γ,(40)\tau_{\mathrm{cav}}=\frac{2}{\gamma}, \tag*{(40)}

which we call coherence time in the main text.

Robustness of the noise-induced temporal order

The main text states that the period-doubled response of Figure 1(b) is robust against randomness in the initial spin configuration and against perturbations of the kick deviation ε0\varepsilon_0. We verify both claims by repeating the Figure 1(b) measurement at the optimal bath temperature Tbath/T0=1T_{\mathrm{bath}}/T_0=1 while separately scanning ε0\varepsilon_0 and the initial-state cap half-angle θinit\theta_{\mathrm{init}}. Figure S1 shows the resulting subharmonic response. The bright ridge at ν/π=1\nu/\pi=1 survives over a finite window of the kick deviation (Figure S1). It also persists across a large swept range of initial-state randomness, decaying only gradually as θinit\theta_{\mathrm{init}} approaches the fully random cap θinit=π\theta_{\mathrm{init}}=\pi (Figure S1). The

Fourier amplitude of the noise-averaged response against $\nu/\pi$ as $\varepsilon_0$ and $\theta_{\mathrm{init}}$ are varied

Figure S1. Robustness of the noise-induced period doubling. Fourier amplitude ∣F~[⟦Mz⟧](ν)∣|\widetilde{F}[\llbracket M^z\rrbracket](\nu)| of the noise-averaged response (the Figure 1(b) observable) against ν/π\nu/\pi, at the optimal bath temperature Tbath/T0=1T_{\mathrm{bath}}/T_0=1, as (a) the static kick deviation ε0\varepsilon_0 and (b) the initial-state randomness θinit\theta_{\mathrm{init}} are varied. The subharmonic ridge at ν/π=1\nu/\pi=1 persists over a finite band of ε0\varepsilon_0 and across a large range of initial-state disorder. The period doubling is therefore robust to both drive-parameter and initial-configuration perturbations. Parameters: L=100L=100, J=5J=5, τosc=0.2\tau_{\mathrm{osc}}=0.2, γ=10−4\gamma=10^{-4}, Np=104N_p=10^4, and Ntraj=100N_{\mathrm{traj}}=100. Panel (a) fixes θinit=0.2\theta_{\mathrm{init}}=0.2, and panel (b) fixes ε0=0.1\varepsilon_0=0.1.

noise-induced DTC therefore requires neither fine tuning of the effective kick nor a specially prepared initial state, consistent with the prethermal-DTC mechanism of Figure 2.

Comparison between cavity response and period-doubled spin response

The period-doubled signal is generated by the spin lattice, not inherited from the drive. Using the same Fourier amplitude of the noise-averaged signal, ∣F~[⋅]∣|\widetilde{F}[\cdot]| (the Figure 1(b) observable), Figure S2 compares drive and response. The drive has weight only near its frequency ωc\omega_c (ν/π=2\nu/\pi=2), whereas the response develops a sharp subharmonic at ν/π=1\nu/\pi=1 that is absent from the drive. The full frequency-temperature dependence of this subharmonic, ∣F~[⟦Mz⟧](ν)∣|\widetilde{F}[\llbracket M^z\rrbracket](\nu)| over (ν/π,Tbath)(\nu/\pi,T_{\mathrm{bath}}), is the emergence map of Figure 1(b), whose bright ridge at ν/π=1\nu/\pi=1 is the period-doubled response.

Fourier amplitude of the noise-averaged drive and response against $\nu/\pi$

Figure S2. Drive versus response spectra. Fourier amplitude ∣F~[⋅](ν)∣|\widetilde{F}[\cdot](\nu)| of the noise-averaged drive ⟦y⟧\llbracket y\rrbracket (blue) and response ⟦Mz⟧\llbracket M^z\rrbracket (orange) against ν/π\nu/\pi (ν=ωτosc\nu=\omega\tau_{\mathrm{osc}}), with the parameters of Figure 1(b) at Tbath/T0=1T_{\mathrm{bath}}/T_0=1. The drive carries weight at its frequency ωc\omega_c (ν/π=2\nu/\pi=2) but none at ν/π=1\nu/\pi=1. The response develops a sharp subharmonic peak at ν/π=1\nu/\pi=1, half the cavity frequency, which is absent from the drive.

Prethermal lifetime scaling

As in prethermal discrete time crystals [20, 30, 31], the lifetime of the noise-induced period doubling grows exponentially with the cavity frequency. We test this scaling by fixing the operating point and sweeping the period τosc\tau_{\mathrm{osc}}, hence ωc=2π/τosc\omega_c=2\pi/\tau_{\mathrm{osc}}. For a deterministic drive at a representative point of the DTC band of Figure 2(g) (ε0=0\varepsilon_0=0, R0=1R_0=1), the correlation time τcorr\tau_{\mathrm{corr}} (Eq. (6)), averaged over initial spin configurations, grows exponentially by two orders of magnitude, τcorr/τosc∝ecωc\tau_{\mathrm{corr}}/\tau_{\mathrm{osc}}\propto e^{c\omega_c} with c≃0.29c\simeq0.29 (Figure S3). This is the exponential prethermal scaling invoked in the main text for both Figure 2 and Figure 3.

Prethermal lifetime scaling

Figure S3. Prethermal lifetime scaling. Correlation time in units of the period τosc\tau_{\mathrm{osc}}, τcorr/τosc\tau_{\mathrm{corr}}/\tau_{\mathrm{osc}} (Eq. (6)), against the frequency ωc=2π/τosc\omega_c = 2\pi/\tau_{\mathrm{osc}}, with dashed exponential guides ∝ecωc\propto e^{c\omega_c}. (a) Deterministic drive at a representative point of the Figure 2(g) DTC region (ε0=0\varepsilon_0 = 0, R0=1R_0 = 1), averaged over initial-state realizations (error bars: standard deviation). The lifetime grows exponentially by two orders of magnitude, with c≃0.29c \simeq0.29. (b) Stochastic drive at fixed coherence time τcav/τosc=2×105\tau_{\mathrm{cav}}/\tau_{\mathrm{osc}} = 2 \times10^5. The green density shows the per-realization distribution P(τcorr/τosc)P(\tau_{\mathrm{corr}}/\tau_{\mathrm{osc}}) at each frequency, with the noise-averaged mean \left\llbracket\tau_{\mathrm{corr}}\right\rrbracket (black squares) and the median (orange circles). Both grow exponentially, with c≃0.19c \simeq0.19 and 0.230.23. Parameters: L=100L = 100, J=5J = 5, λ=1\lambda= 1, and θinit=0.2\theta_{\mathrm{init}} = 0.2. Panel (b) uses Tbath/T0=1T_{\mathrm{bath}}/T_0 = 1 and Ntraj=100N_{\mathrm{traj}} = 100.

The same scaling survives the stochastic drive. We hold the coherence time of the cavity mode fixed in units of cavity cycles, τcav/τosc=2×105\tau_{\mathrm{cav}}/\tau_{\mathrm{osc}}=2\times10^5, so that every frequency is compared at equal quality factor. We then record the single-trajectory correlation time of each noise realization. Figure S3 shows the resulting distribution against ωc\omega_c. A few long-lived realizations pull the noise-averaged mean ⟦τcorr⟧\llbracket\tau_{\mathrm{corr}}\rrbracket well above the typical (median) value, Nevertheless, both the mean and the median grow exponentially, with rates c≃0.19c \simeq0.19 and 0.230.23 comparable to the deterministic value. The prethermal protection generated by the narrowband stochastic drive therefore obeys the same exponential frequency scaling as its deterministic counterpart.

Thermalization time at short coherence time

Section S5 shows the exponential growth of the correlation time at a long coherence time, where no time-soliton occurs. Here, we test it in the DTL regime of Figure 3, where time-solitons separate the correlation times from τth\tau_{\mathrm{th}}. We fix the coherence time at τcav/τosc=20\tau_{\mathrm{cav}}/\tau_{\mathrm{osc}} = 20 and sweep the cavity period from τosc=0.01\tau_{\mathrm{osc}} = 0.01 to 0.0010.001, i.e., from 200π200\pi to 2000π2000\pi, with all other parameters as in Figure 3. For each noise realization, we evaluate τosc∑n=0Np∣s(n)∣\tau_{\mathrm{osc}} \sum_{n=0}^{N_p} |s(n)|, whose noise average is the thermalization time τth\tau_{\mathrm{th}} of Eq. (8). Figure S4 shows that τth/τosc\tau_{\mathrm{th}}/\tau_{\mathrm{osc}} grows

Thermalization time at short coherence time

Figure S4. Thermalization time at short coherence time. Thermalization time, τth/τosc\tau_{\mathrm{th}}/\tau_{\mathrm{osc}} (Eq. (8)), as a function of the cavity frequency ωc=2π/τosc\omega_c = 2\pi/\tau_{\mathrm{osc}}, at the fixed coherence time τcav/τosc=20\tau_{\mathrm{cav}}/\tau_{\mathrm{osc}} = 20 of the DTL in Figure 4. The green density shows the distribution of the single-realization value τosc∑n∣s(n)∣\tau_{\mathrm{osc}} \sum_n |s(n)| at each frequency, with its mean τth\tau_{\mathrm{th}} (black squares) and its median (orange circles). The cavity frequency controls the lifetime of the period doubling also in the DTL. Parameters: L=100L = 100, J=5J = 5, λ=1\lambda= 1, ε0=0.001\varepsilon_0 = 0.001, θinit=0.2\theta_{\mathrm{init}} = 0.2, Tbath/T0=1T_{\mathrm{bath}}/T_0 = 1, Np=104N_p = 10^4, and Ntraj=100N_{\mathrm{traj}} = 100.

monotonically, where the distribution at each frequency is narrow. The growth is slower than a single exponential in ωc\omega_c but still shows a strong frequency dependence.

Drive-phase dependence of the DTC lifetime

To understand the mechanism, the main text replaces the oscillator with a single deterministic tone y(t)=R0cos⁡(ωct+ϕ0)y(t) = R_0 \cos(\omega_c t + \phi_0) and fixes the phase to ϕ0=0\phi_0 = 0 (Figure 2(a,Figure 2d,Figure 2g)). The full stochastic dynamics (Figure 2(b,Figure 2c,Figure 2e,Figure 2f,Figure 2h,Figure 2i)), however, sees a filtered drive y(t)≃Rcos⁡(ωct+φ)y(t) \simeq R \cos(\omega_{\mathrm{c}}t+\varphi). Its phase φ\varphi is uniform on [0,2π)[0,2\pi) and takes a different value in each noise realization (Sec. S2). Here, we show how strongly the period doubling of the deterministic model depends on this phase.

Figure S5 shows the correlation time τcorr\tau_{\mathrm{corr}} of the deterministic response versus ϕ0\phi_0, normalized to its value at ϕ0=0\phi_0=0. The correlation time is largest near ϕ0=0\phi_0=0 and vanishes near ϕ0=π\phi_0=\pi, where the stroboscopic sampling times t=nτosct=n\tau_{\mathrm{osc}} fall half a drive period out of step with the kick sequence. The drive phase therefore strongly affects the lifetime of the period doubling at both frequencies. This phase dependence, however, weakens as the drive frequency increases. We take ϕ0=0\phi_0=0 as the reference phase in the deterministic model of Figure 2(a), since it gives the longest correlation time. The fraction of phases that reach this lifetime grows with the drive frequency.

Drive-phase dependence of the DTC lifetime

Figure S5. Drive-phase dependence of the DTC lifetime. Correlation time τcorr\tau_{\mathrm{corr}} of Eq. (6), normalized to its value at ϕ0=0\phi_0 = 0, of the deterministic drive y(t)=R0cos⁡(ωct+ϕ0)y(t) = R_0 \cos(\omega_c t + \phi_0) as a function of the initial phase ϕ0\phi_0, for a slow (ωc=10π\omega_c = 10\pi) and a fast (ωc=200π\omega_c = 200\pi) drive. At both frequencies, τcorr\tau_{\mathrm{corr}} is largest near ϕ0=0\phi_0 = 0 and vanishes near ϕ0=π\phi_0 = \pi. The drive phase therefore strongly affects the DTC lifetime. As ωc\omega_c increases, the plateau of maximal τcorr\tau_{\mathrm{corr}} around ϕ0=0\phi_0 = 0 widens. The phase dependence thus weakens at higher frequency. Parameters: L=100L = 100, J=5J = 5, ε0=0\varepsilon_0 = 0, R0=1R_0 = 1, θinit=0.2\theta_{\mathrm{init}} = 0.2, Np=104N_p = 10^4.

Time scales of the toy model of time-solitons

Here, we show several analytic results for the toy model of Eq. (2).

Deep inside the ordered regime, the stroboscopic response of Eq. (2) is a period-doubled signal. We denote by Mz(nτ)M^z(n\tau) the stroboscopic magnetization at period nn, which should be understood as the expectation value of the operator M^z\hat{M}^z for all-polarized initial states in the quantum case. At p=0p=0, the normalized staggered magnetization s(n)=(−1)nMz(nτ)/Mz(0)s(n)=(-1)^n M^z(n\tau)/M^z(0) of the main text therefore stays at its initial value s(0)=1s(0)=1, while for 0<p<1/20<p<1/2 it flips sign with probability pp at each period. Without loss of generality, we write the normalized staggered magnetization s(n)s(n) of a single realization as

s(n)=e−nτ/τthσ(n),σ(n)=±1,(41)s(n)=e^{-n\tau/\tau_{\mathrm{th}}}\sigma(n),\qquad\sigma(n)=\pm1, \tag*{(41)}

where τ\tau is the period of the drive, so that period nn occurs at the time t=nτt=n\tau. The factor e−nτ/τthe^{-n\tau/\tau_{\mathrm{th}}} is the prethermal envelope we introduced to capture the decay of the period doubling due to thermalization with a lifetime

τth≫τ.(42)\tau_{\mathrm{th}}\gg\tau. \tag*{(42)}

Here, σ(n)\sigma(n) is the subharmonic phase, which takes the two values ±1\pm1 of the two period-doubled configurations. A time-soliton is the event εn=π\varepsilon_n=\pi in the drive of Eq. (2), which exchanges the sign of σ(n)\sigma(n), so that σ\sigma follows the process

σ(n)=σ(n−1)×{−1prob. p+1prob. 1−p.(43)\sigma(n)=\sigma(n-1)\times \begin{cases} -1 & \text{prob. }p\\ +1 & \text{prob. }1-p. \end{cases} \tag*{(43)}

The solitons are independent between periods. The gap kk between two neighboring solitons therefore has distribution P(k)=(1−p)k−1pP(k)=(1-p)^{k-1}p, with mean ⟦k⟧=1/p\llbracket k\rrbracket=1/p periods, where ⟦⋅⟧\llbracket\cdot\rrbracket denotes the average over the random realizations of the solitons. The typical time between two solitons is therefore ⟦k⟧τ\llbracket k\rrbracket\tau, and the density of time-solitons is its inverse,

ρsol≡1⟦k⟧τ=pτ.(44)\rho_{\mathrm{sol}}\equiv\frac{1}{\llbracket k\rrbracket\tau}=\frac{p}{\tau}. \tag*{(44)}

Because the solitons are independent, the correlator of σ(n)\sigma(n) follows immediately. Over a time separation kk, the sign is reversed once per soliton, so that σ(m)σ(m+k)=(−1)Nk\sigma(m)\sigma(m+k)=(-1)^{N_k}, with NkN_k the number of solitons in this interval. The averaged correlator thus is

⟦σ(m)σ(m+k)⟧=[(1−p)−p]k=(1−2ρsolτ)k.(45)\llbracket\sigma(m)\sigma(m+k)\rrbracket=[(1-p)-p]^k=(1-2\rho_{\mathrm{sol}}\tau)^k. \tag*{(45)}

Three time scales

We can also define the three time scales of the main text in the context of the toy model. We consider

\tau_{\mathrm{th}}=\tau\left\llbracket\sum_{k\geq0}|s(k)|\right\rrbracket, \tag*{(46)}
\llbracket\tau_{\mathrm{corr}}\rrbracket\equiv\tau\left\llbracket\left|\sum_{k\geq0}s(k)\right|\right\rrbracket, \tag*{(47)}
τcorrens≡τ∣∑k≥0⟦s(k)⟧∣.(48)\tau_{\mathrm{corr}}^{\mathrm{ens}}\equiv\tau\left|\sum_{k\geq0}\llbracket s(k)\rrbracket\right|. \tag*{(48)}

where the prefactor τ\tau converts the sum over periods into a time. Two of the three can be derived directly, as we show below.

Using ∣s(k)∣=e−kτ/τth|s(k)|=e^{-k\tau/\tau_{\mathrm{th}}}, Eq. (S.27) is nothing but

\tau\left\llbracket\sum_{k\geq0}|s(k)|\right\rrbracket=\frac{\tau}{1-e^{-\tau/\tau_{\mathrm{th}}}}. \tag*{(49)}

In the limit of Eq. (S.23), it reduces to

lim⁡τth≫ττ1−e−τ/τth=τth,(50)\lim_{\tau_{\mathrm{th}}\gg\tau}\frac{\tau}{1-e^{-\tau/\tau_{\mathrm{th}}}}=\tau_{\mathrm{th}}, \tag*{(50)}

which confirms that the integrated definition in Eq. (S.27) returns the prethermal scale τth\tau_{\mathrm{th}}.

Using the correlator of Eq. (S.26), we can obtain

τcorrens=τ∑k≥0e−kτ/τth(1−2ρsolτ)k=τ1−(1−2ρsolτ)e−τ/τth.(51)\tau_{\mathrm{corr}}^{\mathrm{ens}}=\tau\sum_{k\geq0}e^{-k\tau/\tau_{\mathrm{th}}}(1-2\rho_{\mathrm{sol}}\tau)^k=\frac{\tau}{1-(1-2\rho_{\mathrm{sol}}\tau)e^{-\tau/\tau_{\mathrm{th}}}}. \tag*{(51)}

In the limit of Eq. (S.23) and for ρsolτ≪1\rho_{\mathrm{sol}}\tau\ll1, it reduces to

lim⁡τth≫τ, ρsolτ≪1τcorrens=12ρsol+τth−1.(52)\lim_{\substack{\tau_{\mathrm{th}}\gg\tau,\ \rho_{\mathrm{sol}}\tau\ll1}}\tau_{\mathrm{corr}}^{\mathrm{ens}}=\frac{1}{2\rho_{\mathrm{sol}}+\tau_{\mathrm{th}}^{-1}}. \tag*{(52)}

Moreover, when the solitons dominate, ρsol≫τth−1\rho_{\mathrm{sol}}\gg\tau_{\mathrm{th}}^{-1}, this reduces to

lim⁡τth−1≪ρsol≪τ−1τcorrens=12ρsol.(53)\lim_{\tau_{\mathrm{th}}^{-1}\ll\rho_{\mathrm{sol}}\ll\tau^{-1}}\tau_{\mathrm{corr}}^{\mathrm{ens}}=\frac{1}{2\rho_{\mathrm{sol}}}. \tag*{(53)}

This is the limit where the soliton density alone sets the correlation time.

The three time scales obey τcorrens≤⟦τcorr⟧≤τth\tau_{\mathrm{corr}}^{\mathrm{ens}}\leq\llbracket\tau_{\mathrm{corr}}\rrbracket\leq\tau_{\mathrm{th}}, because the modulus taken inside a summation or ensemble average would give a result equal or larger than the modulus taken outside them. In the DTC regime, ρsolτth≪1\rho_{\mathrm{sol}}\tau_{\mathrm{th}}\ll1, no soliton typically appears before thermalization. The modulus then act trivially, and three time scales collapse to one,

τcorrens≃⟦τcorr⟧≃τth,ρsolτth≪1.(54)\tau_{\mathrm{corr}}^{\mathrm{ens}}\simeq\llbracket\tau_{\mathrm{corr}}\rrbracket\simeq\tau_{\mathrm{th}},\qquad\rho_{\mathrm{sol}}\tau_{\mathrm{th}}\ll1. \tag*{(54)}

In the DTL regime, ρsolτth≫1\rho_{\mathrm{sol}}\tau_{\mathrm{th}}\gg1, they separate, and τcorrens\tau_{\mathrm{corr}}^{\mathrm{ens}} becomes directly relevant to the soliton density via Eq. (S.34).

Distribution of the single-trajectory correlation

The correlation time of single trajectory, before the modulus of Eq. (S.28) is taken, is the random variable

Z=τ∑k≥0s(k)=τ∑k≥0akσ(k),a≡e−τ/τth.(55)Z=\tau\sum_{k\geq0}s(k)=\tau\sum_{k\geq0}a^k\sigma(k),\qquad a\equiv e^{-\tau/\tau_{\mathrm{th}}}. \tag*{(55)}

a sum of the signs σ(k)=±1\sigma(k)=\pm1 weighted by the decaying envelope. Here, we derive the distribution for this random variable.

Let k1<k2<k3<…k_1<k_2<k_3<\ldots be the period where the solitons appear in one trajectory. Between two neighboring solitons the sign does not change, so σ(k)=+1\sigma(k)=+1 for k<k1k<k_1, then σ(k)=−1\sigma(k)=-1 for k1≤k<k2k_1\leq k<k_2, and so on, the sign alternating at each soliton. The sum in Eq. (S.36) can be written as

Z=τ[∑k=0k1−1ak−∑k=k1k2−1ak+∑k=k2k3−1ak−…].(56)Z=\tau\left[\sum_{k=0}^{k_1-1}a^k-\sum_{k=k_1}^{k_2-1}a^k+\sum_{k=k_2}^{k_3-1}a^k-\ldots\right]. \tag*{(56)}

Each block is a partial geometric series,

∑k=mn−1ak=am−an1−a.(57)\sum_{k=m}^{n-1}a^k=\frac{a^m-a^n}{1-a}. \tag*{(57)}

so that the block running from kik_i to ki+1k_{i+1} contributes (−1)iτ(aki−aki+1)/(1−a)(-1)^i\tau(a^{k_i}-a^{k_{i+1}})/(1-a). Summing them together, we obtain

Z=τ1−a[1−2ak1+2ak2−2ak3+…].(58)Z=\frac{\tau}{1-a}\left[1-2a^{k_1}+2a^{k_2}-2a^{k_3}+\ldots\right]. \tag*{(58)}

It is convenient to separate that leading term by writing

Z=τ1+2W1−a,W≡∑i≥1(−1)iaki,(59)Z=\tau\frac{1+2W}{1-a},\qquad W\equiv\sum_{i\geq1}(-1)^ia^{k_i}, \tag*{(59)}

so that all the randomness sits in WW, and W=0W=0 for a trajectory without solitons.

A self-consistent equation for the distribution of W

By Eq. (S.25), each gap is an independent random variable GG with

P(G=g)=(1−ρsolτ)g−1ρsolτ,g=1,2,…(60)P(G=g)=(1-\rho_{\mathrm{sol}}\tau)^{g-1}\rho_{\mathrm{sol}}\tau,\qquad g=1,2,\ldots \tag*{(60)}

Write the soliton positions as sums of the gaps, ki=g1+g2+⋯+gik_i=g_1+g_2+\cdots+g_i with each gig_i distributed as in Eq. (S.41), and condition on the first gap, k1=g1≡Gk_1=g_1\equiv G. We can express

W=∑i≥1(−1)iaki=∑i≥1(−1)iak1aki−k1=−aG[1−ak2−k1+ak3−k1−…].(61)W=\sum_{i\geq1}(-1)^ia^{k_i}=\sum_{i\geq1}(-1)^ia^{k_1}a^{k_i-k_1} =-a^G\left[1-a^{k_2-k_1}+a^{k_3-k_1}-\ldots\right]. \tag*{(61)}

Consider now the bracket. Dropping its leading 11, what remains is

W′≡∑j≥1(−1)jakj+1−k1,(62)W'\equiv\sum_{j\geq1}(-1)^ja^{k_{j+1}-k_1}, \tag*{(62)}

and Eq. (S.42) reads

W=−aG(1+W′).(63)W=-a^G(1+W'). \tag*{(63)}

This means that if the first soliton appears after G=gG=g periods (an event of probability Eq. (S.41)), then W=−ag(1+W′)W=-a^g(1+W'). This allows us to write

P(W≤w)=∑g≥1P(G=g)P(−ag(1+W′)≤w).(64)P(W\leq w)=\sum_{g\geq1}P(G=g)P(-a^g(1+W')\leq w). \tag*{(64)}

This random variable W′W' has exactly the form of Eq. (S.40), since we have kj+1−k1=g2+⋯+gj+1k_{j+1}-k_1=g_2+\cdots+g_{j+1}, i.e. sums of the gaps g2,g3,…g_2,g_3,\ldots, in the same way that the positions entering WW are sums of g1,g2,…g_1,g_2,\ldots. To use this property, we focus on Eq. (S.45), and rewrite the −ag(1+W′)≤w-a^g(1+W')\leq w as W′≥−wa−g−1W'\geq-wa^{-g}-1, and obtain

P(W≤w)=∑g≥1P(G=g)P(W′≥−wa−g−1)=∑g≥1P(G=g)[1−P(W′≤−wa−g−1)]=∑g≥1P(G=g)[1−P(W≤−wa−g−1)].(65)\begin{aligned} P(W\leq w)&=\sum_{g\geq1}P(G=g)P(W'\geq-wa^{-g}-1)\\ &=\sum_{g\geq1}P(G=g)\left[1-P(W'\leq-wa^{-g}-1)\right]\\ &=\sum_{g\geq1}P(G=g)\left[1-P(W\leq-wa^{-g}-1)\right]. \tag*{(65)} \end{aligned}

This is a relation between P(W≤w)P(W\leq w) and P(W≤−wa−g−1)P(W\leq-wa^{-g}-1).

Numerical solution

We solve Eq. (S.46) numerically by iteration, using the following procedure.

  1. Grid. Since G≥1G\geq1 and W≤0W\leq0, Eq. (S.44) gives aG∈(0,a]a^G\in(0,a] and 1+W′∈[0,1]1+W'\in[0,1], so the support is bounded,

−a≤W<0,τ1−2a1−a≤Z<τ1−a≃τth.(66)-a\leq W<0,\qquad\tau\frac{1-2a}{1-a}\leq Z<\frac{\tau}{1-a}\simeq\tau_{\mathrm{th}}. \tag*{(66)}

and a uniform grid of NWN_W points on [−a,0][-a,0] covers it with no truncation.

  1. Gap sum. Truncate the sum over gg at gmax⁡g_{\max} with (1−ρsolτ)gmax⁡(1-\rho_{\mathrm{sol}}\tau)^{g_{\max}} below machine precision.

  1. Iterate. Starting from the uniform distribution on [−a,0][-a,0], evaluate the right-hand side of Eq. (S.46) and replace the old distribution by the result. The rescaled argument −wa−g−1-wa^{-g}-1 generally falls between two grid points, so that P(W≤⋅)P(W\leq\cdot) is obtained by linear interpolation. It may also fall outside the support of Eq. (S.47), where the distribution takes value 0 when below the support or 1 when above it.

  1. Stop. Monitor δn=max⁡W∣Pn+1(W≤w)−Pn(W≤w)∣\delta_n=\max_W\lvert P_{n+1}(W\leq w)-P_n(W\leq w)\rvert and stop once it falls below the tolerance. The distribution of ZZ then follows from the change of variable of Eq. (S.40).

Distribution of $Z$ in the toy model

Figure S6. Distribution of ZZ in the toy model. (a) Distribution P(Z)P(Z) obtained from Eq. (S.46), for two cases with ρsolτth\rho_{\mathrm{sol}}\tau_{\mathrm{th}}: τth=100\tau_{\mathrm{th}}=100 with ρsol=0.03\rho_{\mathrm{sol}}=0.03 (ρsolτth=3\rho_{\mathrm{sol}}\tau_{\mathrm{th}}=3), and τth=1000\tau_{\mathrm{th}}=1000 with ρsol=0.002\rho_{\mathrm{sol}}=0.002 (ρsolτth=2\rho_{\mathrm{sol}}\tau_{\mathrm{th}}=2). Dashed vertical lines mark the mean value from Eq. (S.32). A finite fraction of the weight lies at Z<0Z<0, coming from the trajectories with an odd number of solitons within the lifetime. (b) Convergence of the iteration, as the largest change per sweep δn\delta_n against the sweep index nn. Both panels use a grid of Nw=4001N_w=4001 points and a tolerance of 10−1410^{-14}.

Distribution of the single-trajectory correlation time at a longer coherence time

Figure S7. Distribution of the single-trajectory correlation time at a longer coherence time. Same as Figure 4, at τcav/τosc=40\tau_{\mathrm{cav}}/\tau_{\mathrm{osc}}=40. Probability distribution p(τcorr)p(\tau_{\mathrm{corr}}), with τcorr=∣Z∣\tau_{\mathrm{corr}}=|Z|, for the cavity-spin system (green histogram) and for the toy model with the parameters of that same lattice point (dashed line). The toy parameters are not fitted: τth\tau_{\mathrm{th}} is the value measured on the lattice, and ρsol\rho_{\mathrm{sol}} follows from inverting Eq. (S.32) for the measured τcorrens\tau_{\mathrm{corr}}^{\mathrm{ens}}, which gives ρsolτth≈15\rho_{\mathrm{sol}}\tau_{\mathrm{th}}\approx15 (top). The vertical lines mark τcorrens\tau_{\mathrm{corr}}^{\mathrm{ens}}, ⟦τcorr⟧\llbracket\tau_{\mathrm{corr}}\rrbracket, and τth\tau_{\mathrm{th}}. The toy model again reproduces the distribution. All other parameters are as in Figure 3.

Figure S6(a) shows the distribution obtained in this way. Figure S6(b) shows that the iteration converges quickly. The distribution is broad, spanning the whole support, and carries appreciable weight on both sides of Z=0Z=0, the negative weight coming from the trajectories with an odd number of solitons within the lifetime. This is what separates the two correlation times of Eqs. (S.28) and (S.29).

The same construction can be tested against the spin lattice in cavity. For a given coherence time of the cavity mode, the lattice supplies both τth\tau_{\mathrm{th}} and τcorrens\tau_{\mathrm{corr}}^{\mathrm{ens}} directly, and inverting Eq. (S.32) for the latter fixes the soliton density ρsol\rho_{\mathrm{sol}}. The toy model is then fully determined, and its distribution can be compared with the distribution of single-trajectory correlation times measured on the lattice. Figure 4 of the main text shows the comparison deep in the DTL, at τcav/τosc=20\tau_{\mathrm{cav}}/\tau_{\mathrm{osc}}=20 (ρsolτth=37\rho_{\mathrm{sol}}\tau_{\mathrm{th}}=37). Figure S7 shows it at τcav/τosc=40\tau_{\mathrm{cav}}/\tau_{\mathrm{osc}}=40 (ρsolτth=15\rho_{\mathrm{sol}}\tau_{\mathrm{th}}=15). Both figures show the single-trajectory correlation time τcorr=∣Z∣\tau_{\mathrm{corr}}=|Z| on a logarithmic axis. In both cases the distribution is broad, spanning more than an order of magnitude in τcorr\tau_{\mathrm{corr}}, and is peaked well below the prethermal scale. The toy model captures the position of the peak and the overall width, and reproduces the slow rise on the small-τcorr\tau_{\mathrm{corr}} side.

Phase diagram

The phase diagram of Figure 1(c) is classified by three quantities, defined in the main text and in Methods. We recall them here in terms of the normalized staggered magnetization sr(n)=(−1)nMxr(nosc)/Mxr(0)s_r(n)=(-1)^nM_x^r(n_{\mathrm{osc}})/M_x^r(0) of realization r=1,…,Ntrajr=1,\ldots,N_{\mathrm{traj}}, which satisfies sr(0)=1s_r(0)=1. The three quantities are the noise-averaged single-trajectory correlation time, the thermalization time, and the precession weight,

⟦τcorr⟧=τoscNtraj∑r=1Ntraj∣∑n=0Npsr(n)∣(67)\llbracket\tau_{\mathrm{corr}}\rrbracket=\frac{\tau_{\mathrm{osc}}}{N_{\mathrm{traj}}}\sum_{r=1}^{N_{\mathrm{traj}}}\left|\sum_{n=0}^{N_p}s_r(n)\right| \tag*{(67)}
τth=τoscNtraj∑r=1Ntraj∑n=0Np∣sr(n)∣(68)\tau_{\mathrm{th}}=\frac{\tau_{\mathrm{osc}}}{N_{\mathrm{traj}}}\sum_{r=1}^{N_{\mathrm{traj}}}\sum_{n=0}^{N_p}|s_r(n)| \tag*{(68)}
Pprec=max⁡(0,−min⁡kC(k)),(69)P_{\mathrm{prec}}=\max\left(0,-\min_k C(k)\right), \tag*{(69)}

where C(k)=∑m∣s(m+k)s(m)∣∑m∣s(m)2∣C(k)=\frac{\sum_m|s(m+k)s(m)|}{\sum_m|s(m)^2|}, and mm runs over a reference window of 100 periods starting at m=50m=50. Each parameter point is then labeled in three steps. Points in the freely precessing band at low temperature, where PprecP_{\mathrm{prec}} has not yet fallen to its high-temperature floor PfloorP_{\mathrm{floor}}, show free precession. Among the remaining points, those with ⟦τcorr⟧<10τosc\llbracket\tau_{\mathrm{corr}}\rrbracket<10\tau_{\mathrm{osc}} are disordered. All other points are a DTC if ⟦τcorr⟧/τth>0.418\llbracket\tau_{\mathrm{corr}}\rrbracket/\tau_{\mathrm{th}}>0.418 and a DTL otherwise. Figure S8 shows the three quantities over the plane of Figure 1(c). The levels PfloorP_{\mathrm{floor}} and 0.4180.418 are fixed by the data, as described below, whereas the level 10τosc10\tau_{\mathrm{osc}} is a choice.

Four heat maps showing correlation time, thermalization time, their ratio, and precession weight

Figure S8. The quantities behind the classification, over the plane of Figure 1(c). (a) Correlation time τcorr\tau_{\mathrm{corr}} and (b) thermalization time τth\tau_{\mathrm{th}}. (c) Their ratio and (d) the precession weight PprecP_{\mathrm{prec}}, on linear scales. Each panel carries, as a white dashed line, the boundary that its own quantity defines: the disorder level τcorr=10τosc\tau_{\mathrm{corr}}=10\tau_{\mathrm{osc}} in (a), the DTC level τcorr/τth=0.418\tau_{\mathrm{corr}}/\tau_{\mathrm{th}}=0.418 in (c), and the edge of the freely precessing band in (d). Panel (b) defines no boundary. Comparing (a) and (b) shows the origin of the DTC island. The thermalization time τth\tau_{\mathrm{th}} falls smoothly with temperature everywhere, whereas τcorr\tau_{\mathrm{corr}} rises to meet it only within a window that widens with the coherence time of the cavity mode. Parameters as in Figure 1(c). See text for more details.

We classify with the single-trajectory scale ⟦τcorr⟧\llbracket\tau_{\mathrm{corr}}\rrbracket rather than with the ensemble scale τcorrens\tau_{\mathrm{corr}}^{\mathrm{ens}}, since the DTC and the DTL are distinguished by how long a single realization keeps its subharmonic phase.

The precession weight PprecP_{\mathrm{prec}} stays at unity as long as the drive is too weak to disturb the free precession of the spins. It then falls as the temperature increases and settles on a floor of 0.02 to 0.03, set by the residual noise of the correlator. We determine PfloorP_{\mathrm{floor}} from this floor in each column of fixed τcav\tau_{\mathrm{cav}}. Specifically, we collect the values of PprecP_{\mathrm{prec}} at the highest temperatures, from one tenth of the maximum temperature upward. We set PfloorP_{\mathrm{floor}} to their median plus five times their median absolute deviation. The free precession then extends from the lowest temperature up to the temperature at which PprecP_{\mathrm{prec}} first reaches PfloorP_{\mathrm{floor}}.

To fix the cut between the DTC and the DTL, we consider all points that show local temporal order, i.e., all points that are neither freely precessing nor disordered. The histogram of ⟦τcorr⟧/τth\llbracket\tau_{\mathrm{corr}}\rrbracket/\tau_{\mathrm{th}} over these points has two groups. Most points have @page 15/18 starts_mid=0 ends_mid=1

Distribution of $Z$ in the toy model

Figure S6 (PDF p. 15). Distribution of ZZ in the toy model. (a) Distribution P(Z)P(Z) obtained from Eq. (S.46), for two cases with ρsolτth\rho_{\mathrm{sol}}\tau_{\mathrm{th}}: τth=100\tau_{\mathrm{th}}=100 with ρsol=0.03\rho_{\mathrm{sol}}=0.03 (ρsolτth=3\rho_{\mathrm{sol}}\tau_{\mathrm{th}}=3), and τth=1000\tau_{\mathrm{th}}=1000 with ρsol=0.002\rho_{\mathrm{sol}}=0.002 (ρsolτth=2\rho_{\mathrm{sol}}\tau_{\mathrm{th}}=2). Dashed vertical lines mark the mean value from Eq. (S.32). A finite fraction of the weight lies at Z<0Z<0, coming from the trajectories with an odd number of solitons within the lifetime. (b) Convergence of the iteration, as the largest change per sweep δn\delta_n against the sweep index nn. Both panels use a grid of Nw=4001N_w=4001 points and a tolerance of 10−1410^{-14}.

Distribution of the single-trajectory correlation time at a longer coherence time

Figure S7 (PDF p. 15). Distribution of the single-trajectory correlation time at a longer coherence time. Same as Figure 4, at τcav/τosc=40\tau_{\mathrm{cav}}/\tau_{\mathrm{osc}}=40. Probability distribution p(τcorr)p(\tau_{\mathrm{corr}}), with τcorr=∣Z∣\tau_{\mathrm{corr}}=|Z|, for the cavity-spin system (green histogram) and for the toy model with the parameters of that same lattice point (dashed line). The toy parameters are not fitted: τth\tau_{\mathrm{th}} is the value measured on the lattice, and ρsol\rho_{\mathrm{sol}} follows from inverting Eq. (S.32) for the measured τcorrens\tau_{\mathrm{corr}}^{\mathrm{ens}}, which gives ρsolτth≈15\rho_{\mathrm{sol}}\tau_{\mathrm{th}}\approx15 (top). The vertical lines mark τcorrens\tau_{\mathrm{corr}}^{\mathrm{ens}}, ⟦τcorr⟧\llbracket\tau_{\mathrm{corr}}\rrbracket, and τth\tau_{\mathrm{th}}. The toy model again reproduces the distribution. All other parameters are as in Figure 3.

Figure S6(a) shows the distribution obtained in this way. Figure S6(b) shows that the iteration converges quickly. The distribution is broad, spanning the whole support, and carries appreciable weight on both sides of Z=0Z=0, the negative weight coming from the trajectories with an odd number of solitons within the lifetime. This is what separates the two correlation times of Eqs. (S.28) and (S.29).

The same construction can be tested against the spin lattice in cavity. For a given coherence time of the cavity mode, the lattice supplies both τth\tau_{\mathrm{th}} and τcorrens\tau_{\mathrm{corr}}^{\mathrm{ens}} directly, and inverting Eq. (S.32) for the latter fixes the soliton density ρsol\rho_{\mathrm{sol}}. The toy model is then fully determined, and its distribution can be compared with the distribution of single-trajectory correlation times measured on the lattice. Figure 4 of the main text shows the comparison deep in the DTL, at τcav/τosc=20\tau_{\mathrm{cav}}/\tau_{\mathrm{osc}}=20 (ρsolτth=37\rho_{\mathrm{sol}}\tau_{\mathrm{th}}=37). Figure S7 shows it at τcav/τosc=40\tau_{\mathrm{cav}}/\tau_{\mathrm{osc}}=40 (ρsolτth=15\rho_{\mathrm{sol}}\tau_{\mathrm{th}}=15). Both figures show the single-trajectory correlation time τcorr=∣Z∣\tau_{\mathrm{corr}}=|Z| on a logarithmic axis. In both cases the distribution is broad, spanning more than an order of magnitude in τcorr\tau_{\mathrm{corr}}, and is peaked well below the prethermal scale. The toy model captures the position of the peak and the overall width, and reproduces the slow rise on the small-τcorr\tau_{\mathrm{corr}} side.

Phase diagram

The phase diagram of Figure 1(c) is classified by three quantities, defined in the main text and in Methods. We recall them here in terms of the normalized staggered magnetization sr(n)=(−1)nMxr(nosc)/Mxr(0)s_r(n)=(-1)^nM_x^r(n_{\mathrm{osc}})/M_x^r(0) of realization r=1,…,Ntrajr=1,\ldots,N_{\mathrm{traj}}, which satisfies sr(0)=1s_r(0)=1. The three quantities are the noise-averaged single-trajectory correlation time, the thermalization time, and the precession weight,

⟦τcorr⟧=τoscNtraj∑r=1Ntraj∣∑n=0Npsr(n)∣(70)\llbracket\tau_{\mathrm{corr}}\rrbracket=\frac{\tau_{\mathrm{osc}}}{N_{\mathrm{traj}}}\sum_{r=1}^{N_{\mathrm{traj}}}\left|\sum_{n=0}^{N_p}s_r(n)\right| \tag*{(70)}
τth=τoscNtraj∑r=1Ntraj∑n=0Np∣sr(n)∣(71)\tau_{\mathrm{th}}=\frac{\tau_{\mathrm{osc}}}{N_{\mathrm{traj}}}\sum_{r=1}^{N_{\mathrm{traj}}}\sum_{n=0}^{N_p}|s_r(n)| \tag*{(71)}
Pprec=max⁡(0,−min⁡kC(k)),(72)P_{\mathrm{prec}}=\max\left(0,-\min_k C(k)\right), \tag*{(72)}

where C(k)=∑m∣s(m+k)s(m)∣∑m∣s(m)2∣C(k)=\frac{\sum_m|s(m+k)s(m)|}{\sum_m|s(m)^2|}, and mm runs over a reference window of 100 periods starting at m=50m=50. Each parameter point is then labeled in three steps. Points in the freely precessing band at low temperature, where PprecP_{\mathrm{prec}} has not yet fallen to its high-temperature floor PfloorP_{\mathrm{floor}}, show free precession. Among the remaining points, those with ⟦τcorr⟧<10τosc\llbracket\tau_{\mathrm{corr}}\rrbracket<10\tau_{\mathrm{osc}} are disordered. All other points are a DTC if ⟦τcorr⟧/τth>0.418\llbracket\tau_{\mathrm{corr}}\rrbracket/\tau_{\mathrm{th}}>0.418 and a DTL otherwise. Figure S8 shows the three quantities over the plane of Figure 1(c). The levels PfloorP_{\mathrm{floor}} and 0.4180.418 are fixed by the data, as described below, whereas the level 10τosc10\tau_{\mathrm{osc}} is a choice.

We classify with the single-trajectory scale ⟦τcorr⟧\llbracket\tau_{\mathrm{corr}}\rrbracket rather than with the ensemble scale τcorrens\tau_{\mathrm{corr}}^{\mathrm{ens}}, since the DTC and the DTL are distinguished by how long a single realization keeps its subharmonic phase.

The precession weight PprecP_{\mathrm{prec}} stays at unity as long as the drive is too weak to disturb the free precession of the spins. It then falls as the temperature increases and settles on a floor of 0.02 to 0.03, set by the residual noise of the correlator. We determine PfloorP_{\mathrm{floor}} from this floor in each column of fixed τcav\tau_{\mathrm{cav}}. Specifically, we collect the values of PprecP_{\mathrm{prec}} at the highest temperatures, from one tenth of the maximum temperature upward. We set PfloorP_{\mathrm{floor}} to their median plus five times their median absolute deviation. The free precession then extends from the lowest temperature up to the temperature at which PprecP_{\mathrm{prec}} first reaches PfloorP_{\mathrm{floor}}.

To fix the cut between the DTC and the DTL, we consider all points that show local temporal order, i.e., all points that are neither freely precessing nor disordered. The histogram of ⟦τcorr⟧/τth\llbracket\tau_{\mathrm{corr}}\rrbracket/\tau_{\mathrm{th}} over these points has two groups. Most points have a small ratio, around 0.12, and form the DTL. A smaller group has a ratio up to 0.83, where τcorr\tau_{\mathrm{corr}} is comparable to τth\tau_{\mathrm{th}}, and forms the DTC. The two groups are separated by a dip near 0.4. We place the cut at the value that best separates the two groups, 0.418. Because the dip is shallow, the number of DTC points depends on the exact cut.

The boundary between the DTL and the disordered region is the least sharp of the three. The distribution of τcorr\tau_{\mathrm{corr}} shows no feature near the cut τcorr=10τosc\tau_{\mathrm{corr}}=10\tau_{\mathrm{osc}}. Moving the cut between 3 and 30 cavity periods changes the number of disordered points from 256 to 816. This is expected, since the local temporal order is lost gradually on heating.

Spin backaction on the oscillator

Here, we demonstrate the backaction effect by including the backaction term of § 8 C. The intensive oscillator equation then reads

y¨+γy˙+ω02y=ξ(t)−λMosch−(t).(73)\ddot{y}+\gamma\dot{y}+\omega_0^2 y=\xi(t)-\frac{\lambda}{M_{\mathrm{osc}}}h^{-}(t). \tag*{(73)}

Numerical methods

We evolve the classical spins by the Poisson-bracket equation S⃗˙i=S⃗i×B⃗i\dot{\vec{S}}_i=\vec{S}_i\times\vec{B}_i, with the instantaneous local field B⃗i=(geff(t),0,Jeff(t)∑j∈nn(i)Sjz)\vec{B}_i=(g_{\mathrm{eff}}(t),0,J_{\mathrm{eff}}(t)\sum_{j\in nn(i)}S_j^z). It is integrated by a second-order Strang splitting into a transverse-field sector (every spin rotates about x^\hat{x} by geff dtg_{\mathrm{eff}}\,dt) and an Ising sector (each spin rotates about z^\hat{z} by Jeff∑jSjz dtJ_{\mathrm{eff}}\sum_jS_j^z\,dt, with all SizS_i^z, and hence every zz-field,

The per-spin force h−(t)=H(−)(t)/Nh^{-}(t)=H^{(-)}(t)/N is defined in § 8 C, with H(−)H^{(-)} given by Eq. (S.11).

We repeat the calculation of Figure 2(h,Figure 2i) with the backaction term included and compare the per-cycle effective-drive parameters (εcycle,Rcycle)(\varepsilon_{\mathrm{cycle}},R_{\mathrm{cycle}}) in Figure S9. At both Tbath/T0=0.01T_{\mathrm{bath}}/T_0=0.01 and Tbath/T0=0.9T_{\mathrm{bath}}/T_0=0.9, the backaction reshapes the cloud, shifting its weight toward smaller RcycleR_{\mathrm{cycle}} and broadening its distribution along εcycle\varepsilon_{\mathrm{cycle}}. These results indicate that the spins can reshape the noisy drive. How this feedback modifies the phase diagram of Figure 1(c) remains an open question. frozen during the substep). Each substep is an exact rotation, so ∣S⃗i∣=1\lvert\vec{S}_i \rvert= 1 is preserved to machine precision and the scheme is stable over the 10410^4 periods used here. The oscillator coordinate y(t)y(t) is propagated on the same time step and held fixed within it. Each spin is initialized in a spherical cap of half-angle θinit\theta_{\mathrm{init}} about +z^+\hat{z} (see Methods), from a fixed random seed.

Effect of spin backaction

Figure S9. Effect of spin backaction. Per-cycle effective-drive parameters (εcycle,Rcycle)(\varepsilon_{\mathrm{cycle}}, R_{\mathrm{cycle}}) for single trajectories at (a) Tbath/T0=0.01T_{\mathrm{bath}}/T_0 = 0.01 and (b) Tbath/T0=0.9T_{\mathrm{bath}}/T_0 = 0.9. Blue circles show the backaction-free trajectories used in Figure 2(h,i), and orange triangles show the corresponding trajectories with backaction. At each temperature, the two trajectories use the same noise sequence and initial spin configuration. All other parameters are those of Figure 2.

We validate the splitting against an independent high-order adaptive ODE solver. Figure S10 shows that the subharmonic response is indistinguishable across Strang steps dt∈{0.002,0.005,0.01}dt \in\{0.002, 0.005, 0.01\} and the ODE reference over 10410^4 periods, and that the error versus the reference follows the expected second-order ∝dt2\propto dt^2 scaling at early times (inset). Over the full run, the error remains bounded at ∼5×10−4\sim5 \times10^{-4} for the smaller steps. This value is the accuracy of the reference solver itself, whose solutions at successive tolerances already differ by this much. No residual splitting error is therefore resolvable below this level, and the production step dt=0.005dt = 0.005 lies well within the converged regime.

Integrator convergence

Figure S10. Integrator convergence. Deterministic drive, τosc=0.2\tau_{\mathrm{osc}} = 0.2. Main panel: the stroboscopic subharmonic response (−1)nMz(nτosc)(-1)^n M^z(n\tau_{\mathrm{osc}}) from the exact-rotation Strang splitting at dt=0.002,0.005,0.01dt = 0.002, 0.005, 0.01 and from the adaptive ODE solver (black dashed, reference), which coincide over 10410^4 periods. Inset: the Strang error max⁡∣ΔMz∣\max\lvert\Delta M^z \rvert relative to the ODE reference versus dtdt, over the first 100 periods (squares) and over the full 10410^4-period run (circles). The early error follows the second-order ∝dt2\propto dt^2 scaling (gray line). The full-run error stays bounded and saturates near the reference’s own accuracy floor (∼5×10−4\sim5 \times10^{-4}), confirming the integrator over the whole run.

References

References

  1. [1]N. D. Mermin and H. Wagner, Absence of Ferromagnetism or Antiferromagnetism in One- or Two-Dimensional Isotropic Heisenberg Models, Physical Review Letters 17, 1133 (1966).
  2. [2]P. C. Hohenberg, Existence of Long-Range Order in One and Two Dimensions, Physical Review 158, 384 (1967).
  3. [3]M. Lax, Classical Noise. V. Noise in Self-Sustained Oscillators, Physical Review 160, 290 (1967).DOI
  4. [4]A. Pikovsky, M. Rosenblum, and J. Kurths, Synchronization: A Universal Concept in Nonlinear Sciences, Cambridge Nonlinear Science Series (Cambridge University Press, 2001).
  5. [5]J. Villain, R. Bidaux, J.-P. Carton, and R. Conte, Order as an effect of disorder, Journal of Physics 41, 1263 (1980).DOI
  6. [6]C. L. Henley, Ordering due to disorder in a frustrated vector antiferromagnet, Physical Review Letters 62, 2056 (1989).pubmed.ncbi.nlm.nih.gov/10039845
  7. [7]R. Benzi, A. Sutera, and A. Vulpiani, The mechanism of stochas- tic resonance, Journal of Physics A: Mathematical and General 14, L453 (1981).DOI
  8. [8]L. Gammaitoni, P. Hänggi, P. Jung, and F. Marchesoni, Stochastic resonance, Reviews of Modern Physics 70, 223 (1998).
  9. [9]A. S. Pikovsky and J. Kurths, Coherence Resonance in a Noise- Driven Excitable System, Physical Review Letters 78, 775 (1997).
  10. [10]H. M. Bretscher, L. Graziotto, M. H. Michael, A. Montanaro, I.-T. Lu, A. Grankin, J. W. McIver, J. Faist, D. Fausti, M. Eckstein, M. Ruggenthaler, A. Rubio, D. Basov, M. Hafezi, M. Claassen, D. M. Kennes, and M. A. Sentef, Fluctuation engineering in cavity quantum materials (2026), arXiv:2604.08666 [cond-mat.mes-hall].arxiv.org/abs/2604.08666
  11. [11]C. J. Eckhardt, A. Grankin, D. M. Kennes, M. Ruggenthaler, A. Rubio, M. A. Sentef, M. Hafezi, and M. H. Michael, Surface-mediated ultrastrong cavity coupling of two-dimensional itinerant electrons, Phys. Rev. Lett. 135, 156902 (2025).DOI
  12. [12]E. Viñas Boström, M. H. Michael, C. Eckhardt, and A. Rubio, Equilibrium nonlinear phononics by electric field fluctuations of terahertz cavities, Phys. Rev. Res. 7, 033163 (2025).
  13. [13]I. Keren, T. A. Webb, S. Zhang, J. Xu, D. Sun, B. S. Y. Kim, D. Shin, S. S. Zhang, J. Zhang, G. Pereira, J. Yao, T. Okugawa, M. H. Michael, E. Viñas Boström, J. H. Edgar, S. Wolf, M. Julian, R. P. Prasankumar, K. Miyagawa, K. Kanoda, G. Gu, M. Cothrine, D. Mandrus, M. Buzzi, A. Cavalleri, C. R. Dean, D. M. Kennes, A. J. Millis, Q. Li, M. A. Sentef, A. Rubio, A. N. Pasupathy, and D. N. Basov, Cavity-altered superconductivity, Nature 650, 864 (2026).
  14. [14]M. P. Zaletel, M. Lukin, C. Monroe, C. Nayak, F. Wilczek, and N. Y. Yao, Colloquium : Quantum and classical discrete time crystals, Reviews of Modern Physics 95, 031001 (2023).DOI
  15. [15]K. Sacha and J. Zakrzewski, Time crystals: A review, Reports on Progress in Physics 81, 016401 (2017).DOI
  16. [16]D. V. Else, B. Bauer, and C. Nayak, Floquet Time Crystals, Physical Review Letters 117, 090402 (2016).DOI
  17. [17]V. Khemani, A. Lazarides, R. Moessner, and S. L. Sondhi, Phase Structure of Driven Quantum Systems, Physical Review Letters 116, 250401 (2016).DOI
  18. [18]N. Y. Yao, A. C. Potter, I.-D. Potirniche, and A. Vishwanath, Discrete Time Crystals: Rigidity, Criticality, and Realizations, Physical Review Letters 118, 030401 (2017).DOI
  19. [19]S. Choi, J. Choi, R. Landig, G. Kucsko, H. Zhou, J. Isoya, F. Jelezko, S. Onoda, H. Sumiya, V. Khemani, C. von Keyserlingk, N. Y. Yao, E. Demler, and M. D. Lukin, Observation of discrete time-crystalline order in a disordered dipolar many-body system, Nature 543, 221 (2017).
  20. [20]D. V. Else, B. Bauer, and C. Nayak, Prethermal Phases of Matter Protected by Time-Translation Symmetry, Physical Review X 7, 011026 (2017).DOI
  21. [21]A. Kyprianidis, F. Machado, W. Morong, P. Becker, K. S. Collins, D. V. Else, L. Feng, P. W. Hess, C. Nayak, G. Pagano, N. Y. Yao, and C. Monroe, Observation of a prethermal discrete time crystal, Science 372, 1192 (2021).DOI
  22. [22]N. Y. Yao, C. Nayak, L. Balents, and M. P. Zaletel, Classical discrete time crystals, Nature Physics 16, 438 (2020).DOI
  23. [23]F. Iemini, A. Russomanno, J. Keeling, M. Schirò, M. Dalmonte, and R. Fazio, Boundary Time Crystals, Physical Review Letters 121, 035301 (2018).
  24. [24]Z. Gong, R. Hamazaki, and M. Ueda, Discrete Time-Crystalline Order in Cavity and Circuit QED Systems, Physical Review Letters 120, 040404 (2018).DOI
  25. [25]F. M. Gambetta, F. Carollo, M. Marcuzzi, J. P. Garrahan, and I. Lesanovsky, Discrete Time Crystals in the Absence of Manifest Symmetries or Disorder in Open Quantum Systems, Physical Review Letters 122, 015701 (2019).DOI
  26. [26]H. Keßler, P. Kongkhambut, C. Georges, L. Mathey, J. G. Cosme, and A. Hemmerich, Observation of a Dissipative Time Crystal, Physical Review Letters 127, 043602 (2021).
  27. [27]J. C. Schumann, I. Lesanovsky, and P. Solanki, Hierarchical time crystals (2026), 2601.09779 [quant-ph].DOI
  28. [28]B. I. Halperin and D. R. Nelson, Theory of two-dimensional melting, Phys. Rev. Lett. 41, 121 (1978).
  29. [29]See Supplemental Material.
  30. [30]F. Machado, D. V. Else, G. D. Kahanamoku-Meyer, C. Nayak, and N. Y. Yao, Long-Range Prethermal Phases of Nonequilibrium Matter, Physical Review X 10, 011043 (2020).DOI
  31. [31]B. Ye, F. Machado, and N. Y. Yao, Floquet phases of matter via classical prethermalization, Physical Review Letters 127, 140603 (2021).DOI
  32. [32]P. Kongkhambut, J. Skulte, L. Mathey, J. G. Cosme, A. Hemmerich, and H. Keßler, Observation of a continuous time crystal, Science 377, 670 (2022), https://www.science.org/doi/pdf/10.1126/science.abo3382.DOI
  33. [33]C. J. Eckhardt, S. Chattopadhyay, D. M. Kennes, E. A. Demler, M. A. Sentef, and M. H. Michael, Theory of resonantly enhanced photo-induced superconductivity, Nature Communications 15, 2300 (2024).
  34. [34]L. Grunwald, X. Cheng, E. V. Boström, M. Ruggenthaler, M. H. Michael, D. M. Kennes, and A. Rubio, Cavity control of strongly correlated electrons beyond resonant coupling (2026), arXiv:2603.18933 [quant-ph].arxiv.org/abs/2603.18933
  35. [35]G. Kipp, H. M. Bretscher, B. Schulte, D. Herrmann, K. Kusyak, M. W. Day, S. Kesavan, T. Matsuyama, X. Li, S. M. Langner, J. Hagelstein, F. Sturm, A. M. Potts, C. J. Eckhardt, Y. Huang, K. Watanabe, T. Taniguchi, A. Rubio, D. M. Kennes, M. A. Sentef, E. Baudin, G. Meier, M. H. Michael, and J. W. McIver, Cavity electrodynamics of van der waals heterostructures, Nature Physics 21, 1926 (2025).
  36. [36]In this article we neglect the backaction of the spins on the cavity mode, whose effect on the effective drive we discuss in the SM [29]. In this case, the time-solitons are set by the cavity fluctuations alone and occur independently. Once the backaction becomes relevant, the cavity mode keeps a memory of the spin dynamics and may generate effective interactions between time-solitons. In the toy model, such interactions can be provided by a flip probability conditioned on the history of the trajectory. Such interactions could lock the solitons at regular separations, as in the case of soliton crystals [42], or enhance their proliferation. They would thereby either strengthen or weaken the temporal coherence.
  37. [37]I. Martin, G. Refael, and B. Halperin, Topological Frequency Conversion in Strongly Driven Quantum Systems, Physical Review X 7, 041008 (2017).DOI
  38. [38]D. V. Else, W. W. Ho, and P. T. Dumitrescu, Long-Lived Interacting Phases of Matter Protected by Multiple Time-Translation Symmetries in Quasiperiodically Driven Systems, Physical Review X 10, 021032 (2020).
  39. [39]P. Bruno, Impossibility of Spontaneously Rotating Time Crystals: A No-Go Theorem, Physical Review Letters 111, 070402 (2013).DOI
  40. [40]H. Watanabe and M. Oshikawa, Absence of Quantum Time Crystals, Physical Review Letters 114, 251603 (2015).pubmed.ncbi.nlm.nih.gov/26197119
  41. [41]G. W. Ford, M. Kac, and P. Mazur, Statistical Mechanics of Coupled Oscillators, Journal of Mathematical Physics 6, 504 (1965).DOI

Paper details

vlnc:W_mtfcdgrf
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 3

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Claude Opus 5.5 (bare)

View statement
Looking for related papers…
Contents