Driven Spins Reveal Unexpected Stability

Author: Denis Avetisyan


New research explores how periodically driving a quantum spin chain can lead to localized states and a halting of thermalization.

The driven spin chain exhibits a phase diagram - detailed in the <span class="katex-eq" data-katex-display="false"> Q_0 - \omega_D </span> plane - where scar-induced oscillations and Floquet ETH-predicted thermalization compete, transitioning via a modulation of <span class="katex-eq" data-katex-display="false"> \omega_D </span>, and further delineated by lines representing weak and strong Hilbert space fragmentation dependent on sector-specific spin configurations.
The driven spin chain exhibits a phase diagram – detailed in the Q_0 - \omega_D plane – where scar-induced oscillations and Floquet ETH-predicted thermalization compete, transitioning via a modulation of \omega_D , and further delineated by lines representing weak and strong Hilbert space fragmentation dependent on sector-specific spin configurations.

The study demonstrates a complex interplay between Floquet scars, prethermalization, and Hilbert space fragmentation in a driven spin-one chain.

While many-body localization and thermalization paradigms dominate our understanding of closed quantum systems, the dynamics of periodically driven systems remain comparatively complex. This work, ‘Floquet scars and prethermal fragmentation in a driven spin-one chain’, investigates the non-equilibrium dynamics of a driven spin chain, revealing a rich interplay between Floquet scars, prethermalization, and Hilbert space fragmentation. Specifically, we demonstrate that the driving frequency and amplitude dictate distinct phases characterized by either oscillatory dynamics due to quantum scars, rapid thermalization, or fragmentation into ergodic and integrable subsystems. Can a deeper understanding of these fragmentation mechanisms pave the way for novel control strategies in driven quantum systems and beyond?


The Illusion of Equilibrium: Beyond Traditional Thermalization

The Eigenstate Thermalization Hypothesis (ETH) provides a framework for understanding how complex quantum systems appear to reach equilibrium, suggesting that individual energy eigenstates already embody the statistical properties of a thermal ensemble. However, this seemingly robust principle falters when systems are subjected to strong, time-dependent driving forces. Such forces can disrupt the delicate balance inherent in the ETH, preventing the system from fully exploring its available states and hindering its approach to a true thermal state. Consequently, the system may exhibit persistent, non-equilibrium behavior, displaying memory of the initial conditions or becoming trapped in transient states far from thermalization. This breakdown necessitates a refinement of traditional thermalization paradigms to accurately describe dynamics under intense driving, revealing a richer landscape of non-equilibrium phenomena in complex quantum systems.

The longstanding principle of ergodicity, which underpins much of statistical mechanics, asserts that a system will, over time, explore all accessible states in its phase space, allowing time averages to equate to ensemble averages and ultimately leading to predictable thermal equilibrium. However, this assumption falters when applied to the complex interactions within many-body systems, particularly those driven far from equilibrium. These systems, characterized by strong correlations and intricate dynamics, often exhibit behaviors that deviate significantly from ergodic predictions; complete exploration of phase space becomes hindered, and the system may become trapped in specific regions or exhibit persistent non-equilibrium features. Consequently, traditional methods reliant on the ergodic hypothesis struggle to accurately describe the transient dynamics and emergent phenomena observed in these actively driven systems, necessitating the development of novel theoretical frameworks capable of capturing the intricacies of non-ergodic behavior and predicting system evolution before full thermalization is achieved.

The behavior of complex systems prior to reaching thermal equilibrium represents a critical frontier in modern physics. Traditional analysis often focuses on systems after they have fully thermalized, effectively overlooking the rich and often non-equilibrium dynamics occurring in the interim. This transient regime, preceding complete energy distribution, is where many physically relevant processes unfold – from the initial stages of a collision to the emergence of collective behaviors. Consequently, researchers are developing novel theoretical frameworks that move beyond the limitations of ergodic assumptions and embrace techniques capable of characterizing these fleeting, pre-thermal states. These approaches, including those leveraging concepts from quantum information theory and non-equilibrium statistical mechanics, aim to provide a more complete understanding of how complex systems evolve, not just where they ultimately land.

Analysis of the driven-dissipative system reveals that strong driving ( <span class="katex-eq" data-katex-display="false">\hbar\omega_{D}/Q_{0}=5 </span> ) leads to oscillatory behavior in the force and the emergence of quantum scar states with low <span class="katex-eq" data-katex-display="false">S_{L/2} </span>, indicative of non-thermalization, while weaker driving ( <span class="katex-eq" data-katex-display="false">\hbar\omega_{D}/Q_{0}=1.25 </span> ) promotes rapid thermalization and the absence of scars, with results obtained for system sizes of L=24 and L=18.
Analysis of the driven-dissipative system reveals that strong driving ( \hbar\omega_{D}/Q_{0}=5 ) leads to oscillatory behavior in the force and the emergence of quantum scar states with low S_{L/2} , indicative of non-thermalization, while weaker driving ( \hbar\omega_{D}/Q_{0}=1.25 ) promotes rapid thermalization and the absence of scars, with results obtained for system sizes of L=24 and L=18.

Fragmented Realities: The Emergence of Prethermalization

Periodically driven systems, unlike their static counterparts, are susceptible to Hilbert space fragmentation, a phenomenon where the system’s Hilbert space decomposes into disconnected sectors. This disconnection arises from the interplay between the drive and the system’s inherent properties, preventing transitions between these sectors. Consequently, the system cannot reach a full, globally equilibrated thermal state; instead, energy and information remain localized within individual sectors. The size of these fragmented regions, and thus the degree of hindered thermalization, is directly related to the characteristics of the driving force and the system’s conserved quantities, resulting in a non-ergodic behavior despite the time-dependent drive.

The formation of fragmented Hilbert spaces is directly linked to the presence of conserved quantities within the periodically driven system. These conserved quantities, whether related to particle number, spin, or other symmetries, impose constraints on the system’s dynamics, effectively partitioning the overall Hilbert space into disconnected sectors. Each sector represents a subspace defined by a specific value of the conserved quantity; transitions between these sectors are suppressed or forbidden by the conservation laws. Consequently, the accessible phase space is not the entire Hilbert space, but rather a collection of isolated fragments, each with a reduced dimensionality and modified dynamics. This limitation on accessible states is fundamental to understanding the emergence of prethermalization, as the system explores only a portion of the total state space.

Hilbert space fragmentation leads to prethermalization, a non-equilibrium state characterized by significantly slower dynamics than would be observed in a fully thermalized system. This transient regime persists for extended periods due to the restricted accessible phase space created by the fragmentation. Quantification of this fragmentation reveals a characteristic growth rate determined by the largest eigenvalue of the system’s transfer matrix, experimentally determined to be 3.1149. This eigenvalue directly correlates with the rate at which the system explores its fragmented Hilbert space and establishes the timescale for prethermalization before any potential transition to true equilibrium occurs.

The degree of Hilbert space fragmentation increases with system size, exhibiting a linear scaling relationship where the growth of fragmented sectors is proportional to L/3. This indicates that as the system length, denoted by L, increases, the number of disconnected sectors within the Hilbert space grows proportionally, but at a reduced rate. This scaling behavior has been empirically observed and is a key characteristic of prethermalization dynamics in periodically driven systems, influencing the timescale over which the system remains in a non-equilibrium transient state before potentially reaching full thermalization, or remaining indefinitely fragmented.

The prethermalization timescale <span class="katex-eq" data-katex-display="false">T_p</span> scales linearly with the system size and depends on the driving strength <span class="katex-eq" data-katex-display="false">Q_0</span>, as demonstrated by the behavior of <span class="katex-eq" data-katex-display="false">S(mT)/S_p</span> and <span class="katex-eq" data-katex-display="false">JT_p/\hbar</span> at a fixed driving frequency <span class="katex-eq" data-katex-display="false">\omega_D = \omega_1^*</span>.
The prethermalization timescale T_p scales linearly with the system size and depends on the driving strength Q_0, as demonstrated by the behavior of S(mT)/S_p and JT_p/\hbar at a fixed driving frequency \omega_D = \omega_1^*.

Decoding the Transient: New Observables for a Non-Equilibrium World

Exact Diagonalization (ED) was utilized to model the time evolution of a Kitaev chain subject to periodic driving. This numerical method directly solves the time-dependent Schrödinger equation for the system’s Hilbert space, allowing for precise calculation of the system’s wavefunction at any given time. The Kitaev chain, a one-dimensional model exhibiting topological phases, was chosen due to its analytical solvability which aids in validating the ED results. By simulating the dynamics using ED, we obtain time-dependent expectation values of relevant observables, providing a benchmark for comparison with theoretical predictions such as those derived from Floquet Perturbation Theory. The size of the simulated system was limited by computational resources; simulations were performed with up to N=20 sites to ensure sufficient accuracy and system size scaling was investigated to confirm reliability.

Floquet Perturbation Theory (FPT) provides a means to analyze the time-dependent Schrödinger equation for periodically driven quantum systems. By expressing the time evolution operator as a sum over time-ordered integrals, FPT allows for the calculation of an effective time-independent Hamiltonian, known as the Floquet Hamiltonian H_{F}. This Hamiltonian governs the dynamics in the Floquet space, which is an extended Hilbert space accounting for the multiple energy levels arising from the periodic drive. Crucially, the Floquet Hamiltonian describes the long-time behavior of the system and enables the identification of quasi-energy bands and associated eigenstates, providing a framework for understanding the driven dynamics and any emergent prethermalization phenomena. The accuracy of the Floquet Hamiltonian depends on the order of perturbation theory retained in the calculation; higher orders generally improve the description of the driven system but increase computational complexity.

Entanglement entropy, fidelity, and correlation functions serve as quantitative metrics for characterizing the prethermal state and distinguishing it from true thermal equilibrium. Entanglement entropy, specifically the von Neumann entropy calculated from the reduced density matrix, measures the degree of quantum entanglement and can reveal non-thermal correlations persisting during prethermalization. Fidelity, defined as the overlap between the time-evolved state and the initial state, quantifies the sensitivity of the system to perturbations and diminishes rapidly as the system approaches thermalization. Finally, correlation functions, such as the two-point function of local operators, probe spatial correlations and exhibit deviations from the expected behavior in a fully thermalized system, providing further evidence of prethermalization dynamics. Analysis of these observables allows for precise quantification of deviations from thermal behavior and characterization of the prethermal state’s properties.

Upon analysis of the fragmented Hilbert space resulting from the driven Kitaev chain, distinct dynamical behaviors are observed depending on the fragment. Specifically, the largest fragment, which is integrable, exhibits oscillatory dynamics in its entanglement entropy. This entropy does not saturate to a thermal value, but instead oscillates and ultimately saturates at either 0 or \ln(2). This behavior contrasts with the non-integrable fragments, indicating a breakdown of the standard thermalization paradigm within this dominant portion of the Hilbert space and highlighting the unique properties of the driven system.

Finite-size fluctuations are evident in the plot of <span class="katex-eq" data-katex-display="false">\mathcal{F}(mT)</span> as a function of <span class="katex-eq" data-katex-display="false">m</span>, and the scaled spin distribution <span class="katex-eq" data-katex-display="false">S_{L/2}/S_p</span> reveals quasienergy dependence for Floquet eigenstates with <span class="katex-eq" data-katex-display="false">W_{\ell}=1</span>, <span class="katex-eq" data-katex-display="false">Q_0/J=40</span>, and scaled energies.
Finite-size fluctuations are evident in the plot of \mathcal{F}(mT) as a function of m, and the scaled spin distribution S_{L/2}/S_p reveals quasienergy dependence for Floquet eigenstates with W_{\ell}=1, Q_0/J=40, and scaled energies.

Beyond Equilibrium: Reshaping Quantum Control and Materials Design

Quantum systems, when disturbed, typically evolve towards thermal equilibrium – a state of maximum disorder. However, the existence of what are known as QuantumScars challenges this expectation. Supported by the mathematical framework of the Floquet Hamiltonian – which describes systems periodically driven in time – these scars are special quantum states that resist the usual tendency towards thermalization. Instead of quickly losing energy and becoming disordered, systems possessing QuantumScars exhibit persistent, coherent oscillations and unexpectedly complex dynamics. These non-equilibrium states arise because the scarred states lack the typical sensitivity to perturbations that drives thermalization, effectively shielding portions of the quantum system from disorder. This resistance to thermalization isn’t merely a curiosity; it offers a pathway to maintain quantum coherence for extended periods, which is crucial for applications in quantum technologies.

The ability to characterize and comprehend non-thermal eigenstates presents a paradigm shift in quantum control, offering opportunities to influence system behavior prior to the traditionally assumed endpoint of thermal equilibrium. These states, unlike those defined by standard statistical mechanics, retain memory of initial conditions and exhibit unique responses to external stimuli. Researchers are discovering that by precisely tailoring interactions to target these specific eigenstates, it becomes possible to steer quantum evolution along desired pathways, suppressing unwanted decoherence and enhancing specific processes. This preemptive control transcends simply mitigating thermalization; it unlocks the potential to engineer quantum systems with sustained, non-equilibrium properties, opening doors to advanced computation, sensing, and the creation of materials exhibiting novel functionalities – all before the system succumbs to the randomness of equilibrium.

The exploration of quantum phenomena beyond simple equilibrium extends into the realm of materials science and device engineering. Researchers are beginning to leverage the principles of non-thermalization and the existence of persistent quantum states – such as those indicated by QuantumScars – to envision entirely new classes of devices. This approach allows for the prospective design of materials exhibiting tailored properties, potentially optimizing energy transport, enhancing sensing capabilities, or creating robust quantum memories. By precisely controlling the initial conditions and leveraging these non-equilibrium dynamics, it may become possible to create materials where specific quantum states are preferentially populated, leading to functionalities unattainable in traditional, thermally-equilibrated systems. This represents a paradigm shift, moving beyond simply using quantum effects and towards proactively designing them into the very fabric of future technologies.

Strong finite-size fluctuations are evident in the plot of <span class="katex-eq" data-katex-display="false">\mathcal{F}(mT)</span> as a function of <span class="katex-eq" data-katex-display="false">m</span>, and the scaled spin <span class="katex-eq" data-katex-display="false">S_{L/2}/S_p</span> exhibits quasienergy dependence <span class="katex-eq" data-katex-display="false">E_F/J</span> for Floquet eigenstates with <span class="katex-eq" data-katex-display="false">W_{\ell}=1</span>, indicating significant system behavior at <span class="katex-eq" data-katex-display="false">Q_0/J = 40</span>.
Strong finite-size fluctuations are evident in the plot of \mathcal{F}(mT) as a function of m, and the scaled spin S_{L/2}/S_p exhibits quasienergy dependence E_F/J for Floquet eigenstates with W_{\ell}=1, indicating significant system behavior at Q_0/J = 40.

The study of driven quantum systems, as demonstrated by this exploration of the spin-one chain, reveals a compelling truth: order manifests through interaction, not control. The emergence of Floquet scars and prethermal fragmentation isn’t dictated by an overarching design, but rather arises from the interplay of driving frequency, amplitude, and the system’s inherent properties. This mirrors a fundamental principle; systems don’t require architects, they self-organize. As Mary Wollstonecraft observed, “The mind is but a narrow vessel; therefore, it must be constantly supplied with new materials,” – a sentiment that applies equally to a quantum system seeking equilibrium, constantly adjusting and reorganizing its state based on external stimuli and internal dynamics. Sometimes inaction-allowing the system to evolve naturally-is the best tool for understanding its complex behavior.

Where Do We Go From Here?

The exploration of driven quantum systems, as exemplified by this work on Floquet scars and fragmentation, increasingly suggests the inadequacy of seeking overarching, system-wide control. The system is a living organism where every local connection matters; attempting to dictate global behavior often proves futile, yielding only suppressed creative adaptation. Instead, the focus shifts towards understanding how robust, localized excitations – the scars – emerge from the interplay of local rules and external drive.

A critical, and perhaps overlooked, limitation lies in the inherent difficulty of probing genuinely many-body driven systems. Current investigations, while insightful, often rely on parameter regimes that remain analytically tractable. Expanding the scope to encompass strongly interacting, disordered systems will be crucial. This requires not just more computational power, but also novel theoretical frameworks that move beyond perturbation theory and embrace the inherent complexity of emergent phenomena.

The relationship between prethermalization, fragmentation, and the longevity of quantum scars warrants further investigation. Are scars merely transient features of a prethermalized state, or do they represent a qualitatively different form of dynamical protection? Ultimately, the challenge lies in accepting that order doesn’t need architects. It arises spontaneously, a consequence of local interactions, and the task is to learn to read the patterns that emerge, not to impose them.


Original article: https://arxiv.org/pdf/2603.06111.pdf

Contact the author: https://www.linkedin.com/in/avetisyan/

See also:

2026-03-10 04:50