Dimension-raising phase transitions in driven magnets and condensates
Listen
Radio episode about this paper
Transcript
Introduction to the show: ident: Condensed Matter. The day in condensed matter: materials, phases and the measurements under them.
Kai: I'm Kai.
Mira: And I'm Mira, here with Lev.
Kai: We're looking at the paper: "Dimension-raising phase transitions in driven magnets and condensates".
Mira: Here's the gist of the paper. Dimension-raising phase transitions in driven magnets and condensates proposes a periodically driven system whose dimensionality is an emergent property that can be tunable, thus enables us to realize not only many-body phases with arbitrary dimensions, but also phase transitions, instead of crossovers, between phases with various dimensions.
Kai: Emergent Dimensionality Mechanism The core idea is to use a periodically driven system whose instantaneous Hamiltonian keeps the one-dimensional (1D) feature at any given time.
Mira: This system exhibits an emergent two-dimensional (2D) phase when the driving frequency exceeds a critical value, at which a dimension-raising phase transition takes place.
Kai: The authors find that the nonequilibrium feature of the system could qualitatively change the finite temperature critical behavior of the emergent 2D phase and make it different from its equilibrium counterpart.
Title and authors: Kai: So we're talking about "Dimension-raising phase transitions in driven magnets and condensates" by Zhizhen Chen and Zi Cai. The title really tells you that this isn't just about finding new phases, but about controlling the structure of the system through driving.
Mira: I think the authors are pushing beyond just static models where you look at one set of parameters, they are showing how time dependence can introduce entirely new types of transitions between different dimensional states.
Lev: From an engineering standpoint, what they’re doing is moving from a fixed Hamiltonian to one that evolves periodically, which is exactly what you’d need if you were trying to design a system with dynamically tunable properties.
Kai: The paper investigates this by taking an interacting rotor model and setting up a driving protocol where the instantaneous Hamiltonian is always 1D at any moment, but then it jumps to 2D when the driving frequency hits a certain threshold <ref:2503.09437#pg1>.
Mira: And they specifically look at how this nonequilibrium feature changes the finite temperature critical behavior of that emergent two-dimensional phase compared to its equilibrium counterpart.
Lev: That requires careful handling because in a nonequilibrium system, things like free energy aren't well-defined in the usual way, so they have to rely on equal-time correlation functions at a specific time t.
Kai: That's where the math gets tricky; they define this correlation function C of r t as one L squared times X i at time t, which is how you measure that non-equilibrium behavior.
Mira: And they use these correlation functions to show that even though the underlying system looks 1D at any instant, you can see a clear transition to 2D critical behavior when the driving frequency exceeds a critical value <ref:2503.09437#pg1,when the driving frequency exceeds a critical value>.
Lev: If we were to run this on real hardware, we'd have to worry about how quickly the system relaxes after each time step of that driving protocol. The paper mentions the relaxation time tau diverging in the low temperature phase of the unperturbed Hamiltonian, which makes it hard to use simple approximations like Magnus expansion when you drive fast.
The paper's summary: Kai: So to summarize what they found, they have this interacting rotor model where bonds switch between horizontal and vertical configurations periodically. This setup results in a nonequilibrium phase transition from an asymmetric disordered state into an emergent isotropic two-dimensional critical phase when the driving frequency is high enough.
Mira: That transition is significant because it can qualitatively shift the nature of the finite temperature transition, moving it from a Berezinski-Kosterlitz-Thouless type to a first order transition, depending on how fast you drive.
Lev: A first order transition means there's phase coexistence between a metastable state and a stable state, which we saw reflected in their hysteresis loops when they plotted the correlation length against the driving period Td.
Kai: They also showed that if you drive faster, the area of this hysteresis loop shrinks and eventually disappears at some point, leading back to a BKT phase transition where things are more conventional again.
Mira: And they have this higher-dimensional generalization where they use a time dependent Hamiltonian with couplings in four orthogonal directions, which allows them to potentially realize many-body phases with arbitrary dimensions.
Lev: Realizing four dimensions would be tough on current simulators because of the complexity, but if you could get that 4D XY model to emerge under fast driving, that would imply a much richer phase diagram than what we usually see in equilibrium systems.
The paper's improvements: Kai: One key improvement they point out is moving away from dimensional crossover phenomena. Instead of just seeing a transition between two dimensions, the high-dimensional physics emerges spontaneously via this phase transition itself.
Mira: They contrast this with dimensional reduction phenomena, which can happen if you have geometric frustration or certain types of anisotropic orbit order in the system. This paper focuses on the dynamic emergence, not just static constraints.
Lev: That distinction is important for us because it means we're not necessarily dealing with a simple geometry issue; we're dealing with how the time evolution forces the system into a higher effective dimension.
Kai: They also found that the disordered phase they observe isn't actually a 1D phase, even though it looks like it might be one, and this differs from the high-temperature phase of a standard two dimensional XY model in terms of correlation length dependence on temperature <ref:2503.09437#pg1>.
Mira: That difference is shown by a linear relation between the inverse of the correlation length and temperature in the slowly driven case, which matches the 1D XY model but not what they call their disordered phase <ref:2503.09437#pg1>.
Lev: So, even at higher temperatures where quantum fluctuations are more important, this mechanism for dimension raising doesn't seem to change much. That's a bit reassuring if you're trying to build something that needs to be robust against thermal noise.
Conclusion: Kai: So the main point of "Dimension-raising phase transitions in driven magnets and condensates" is that we can dynamically adjust the effective dimension of many-body systems by tuning how fast you drive them.
Mira: This opens up a way to realize phases with arbitrary dimensions, and importantly, it allows for phase transitions between these different dimensional states rather than just crossovers.
Lev: For us in hardware research, the implication is that we need to be mindful of the driving protocol; if we want a specific emergent dimensionality—say 2D or 4D—we have to ensure our time dependence forces that transition when the driving frequency is right.
Kai: And they suggest investigating emergent frustrated systems, like orderby-disorder phenomena, and comparing those behaviors directly against their equilibrium counterparts.
Mira: It’s a strong direction for future research because understanding how frustration arises dynamically during this process could give us new insights into complex many-body physics.
Lev: I just think the work is solid because it ties the dynamics to measurable quantities like hysteresis loops in correlation functions, which gives us concrete evidence that this nonequilibrium physics is real and observable in principle.
Zhizhen Chen, Zi Cai
T. D. Lee Institute, Shanghai Jiao Tong University · Wilczek Quantum Center and Key Laboratory of Artificial Structures and Quantum Control, School of Physics and Astronomy, Shanghai Jiao Tong University
cond-mat.stat-mech
Submitted: 2025-03-12
Updated: 2025-03-12
Comments: 6 pages
Journal ref: New J. Phys. 28, 063502 (2026)
License: http://creativecommons.org/licenses/by/4.0/
Importance score: 89/100
The gist: Dimension-raising phase transitions in driven magnets and condensates proposes a periodically driven system whose dimensionality is an emergent property that can be tunable…
Key concepts
- Emergent Dimensionality
- The core idea is that by periodically driving a system, its effective dimensionality can change. The instantaneous Hamiltonian looks 1D at any moment, but when the driving frequency is high enough, the system spontaneously transitions into an emergent 2D phase. This allows for realizing many-body phases with arbitrary dimensions.
- Nonequilibrium Phase Transition (NEPT)
- This is a transition that occurs due to the periodic driving itself, not just changes in temperature. The system moves from an asymmetric disordered state to an isotropic 2D critical phase when the driving frequency surpasses a critical value. This transition can qualitatively alter how the system behaves at finite temperatures.
- Berezinski-Kosterlitz-Thouless (BKT) Transition
- This is a specific type of phase transition often found in 2D systems, characterized by topological defects. The paper shows that the NEPT can change this transition nature, potentially causing it to become a first-order transition instead of the standard BKT behavior.
- Higher-Dimensional Generalization
- The driving protocol can be extended to four dimensions by switching couplings in four orthogonal directions over different parts of the driving period. This allows researchers to realize many-body phases with effective dimensions higher than three, such as a 4D XY model under fast driving.
Terminology
Summary
Dimension-raising phase transitions in driven magnets and condensates proposes a periodically driven system whose dimensionality is an emergent property that can be tunable, thus enables us to realize not only many-body phases with arbitrary dimensions, but also phase transitions, instead of crossovers, between phases with various dimensions. p. 0 Dimension-raising phase transitions in driven magnets and condensates
Zhizhen Chen1, 2 and Zi Cai2, ∗
1T. Lee Institute, Shanghai Jiao Tong University and Shanghai
Research Center for Quantum Sciences, Shanghai 200240, China
2 Wilczek Quantum Center and Key Laboratory of Artificial Structures and Quantum Control,
School of Physics and Astronomy, Shanghai Jiao Tong University, Shanghai 200240, China
We propose a periodically driven system whose dimensionality is an emergent property that can
be tunable, thus enables us to realize not only many-body phases with arbitrary dimensions, but
also phase transitions, instead of crossovers, between phases with various dimensions.Where the paper says it
Emergent Dimensionality Mechanism
The core idea is to use a periodically driven system whose instantaneous Hamiltonian keeps the one-dimensional (1D) feature at any given time. This system exhibits an emergent two-dimensional (2D) phase when the driving frequency exceeds a critical value, at which a dimension-raising phase transition takes place. p. 0 Despite this, an emergent two-dimensional (2D) phase appears when the driving
frequency exceeds a critical value, at which a dimension-raising phase transition takes place. We
find that the nonequilibrium feature of the system could qualitatively change the finite temperature
critical behavior of the emergent 2D phase and make it different from its equilibrium counterpart. Despite this, an emergent two-dimensional (2D) phase appears when the driving
frequency exceeds a critical value, at which a dimension-raising phase transition takes place. We
find that the nonequilibrium feature of the system could qualitatively change the finite temperature
critical behavior of the emergent 2D phase and make it different from its equilibrium counterpart.Where the paper says it
Where the paper says it
Model Formulation and Driving Protocol
The study investigates an interacting rotor model in a lattice whose connectivity is periodically switched between the patterns with either horizontal bonds or vertical bonds only. The instantaneous Hamiltonian at any given time is a set of decoupled 1D chains. p. 0 In this study, we investigate an interacting rotor model
in a lattice whose connectivity is periodically switched
between the patterns with either horizontal bonds or
vertical bonds only (Fig. 1 a), thus its instantaneous
Hamiltonian at any given time is a set of decoupled 1D
chains.Where the paper says it
-
For the 1st (2nd) half of each period, only the vertical (horizontal) coupling are switched on. p. 1
Where the paper says it
Td is the period of the driving. (2) represents a periodic driving protocol that at the 1st (2nd) half of each period, only the vertical (horizontal) coupling are switched on (see Fig.
-
For the second half of each period, horizontal coupling is switched on and vertical coupling is zero. p. 1
Where the paper says it
Td is the period of the driving. (2) represents a periodic driving protocol that at the 1st (2nd) half of each period, only the vertical (horizontal) coupling are switched on (see Fig.
Nonequilibrium Phase Transition Characteristics
The system exhibits a nonequilibrium phase transition (NEPT) from an asymmetric disordered phase to an emergent isotropic 2D critical phase when the driving frequency exceeds a critical value. The periodic driving could qualitatively change the nature of its finite temperature phase transition from the BerezinskiKosterlitz-Thouless(BKT) type to the 1st order. p. 0 By tuning the driving frequency, we find a nonequilibrium phase transition (NEPT) from an asymmetric disordered phase, to
an emergent isotropic 2D critical phase. Although the
latter resembles the quasi-long-range ordered phase of
an equilibrium 2D XY model, we find that the periodic
driving could qualitatively change the nature of its finite temperature phase transition from the BerezinskiKosterlitz-Thouless(BKT) type[29, 30] to the 1st order. The lifetime of the metastable state crucially
depends on the system size[33], and the long-time evolution will finally drive the system to the stable state,
which will suddenly jump from the asymmetric 1D state
to the emergent isotropic 2D state via a 1st-order phase
transition at a critical Td. If the lifetime of the metastable state is much longer
than the simulation time, we observe a phase coexistence
with a hysteresis behavior where the value of a physical quantity lags behind changes in the effect causing it.Where the paper says it
Where the paper says it
Higher-Dimensional Generalization
A higher-dimensional generalization of this driving protocol makes it possible to realize many-body phases with arbitrary dimensions. This is achieved by a time-dependent Hamiltonian involving couplings in four orthogonal directions, where the lattice sites are organized as L4 decoupled 1D chains with length L along each direction during specific intervals of the driving period. p. 4 5 (a) for an example with L = 2). The timedependent Jα(t) are chosen as following:
Jx(t) = J, Jy,z,w(t) = 0; nTd < t ≤ (n +
1
4
)Td
Jy(t) = J, Jx,z,w(t) = 0; (n +
1
4
)Td < t ≤ (n +
1
2
)Td
Jz(t) = J, Jx,y,w(t) = 0; (n +
1
2
)Td < t ≤ (n +
3
4
)Td
Jw(t) = J, Jx,y,z(t) = 0; (n +
3
4
)Td < t ≤ (n + 1)Td
(8)
which means that at each quarter of the driving period, the L
4
lattice sites are organized as L
3 decoupled
1D chains with length L along α = x, y, z, w respectively
(see Fig. Such a lattice reconfiguration, even though
unrealistic for the cold atom setup, is accessible in the
Rydberg atom arrays, where each atom can be moved
by the optical tweezers in a programable way. With fast
driving, we expect the emergence of a 4D XY model.Where the paper says it
Where the paper says it
Experimental Relevance and Outlook
Experimental realizations can be related to magnetic systems with easy-plane anisotropy or hardcorebosonic models in optical lattices where hopping can be dynamically modulated. The NEPT could be observed at timescales accessible in current optical lattice setups, such as a driving period Td ∼ 10ms. p. 3 (2) by periodically varying the frequency
of the trap along different directions. For an 1D optical lattice with the single-particle hopping Jh = 400Hz,
one can estimate that the NEPT could be observed at a
driving period Td ∼ 10ms, an accessible timescale in the
current optical lattice setup. The
former will lead to anisotropic TOF images elongated
along either the x or y direction, while a symmetric
TOF image will be observed in the presence of fast driving. The short-range and quasi-long-range correlations
can also be distinguished from the TOF images, where
the height of coherence peak will exhibit a hysteresis
loop around the NEPT. Since the dimensionality is an emergent property in
our driving protocols, one may wonder whether other
important concepts, such as the frustration, can emerge
in a similar way. If the lattice connectivity periodically
switches between two frustration-free configurations, the
system adiabatically follows their respective equilibrium
states for a sufficiently slow driving.Where the paper says it
Where the paper says it
Where the paper says it
Finite Temperature Phase Diagram
The phase diagram in terms of temperature T and driving period Td shows distinct regions including phase coexistence, 2D disordered phase, 1D disordered phase, and 2D critical phase. With faster driving, the area of the hysteresis loop shrinks and eventually disappears at a tricritical point after which the BKT phase transition is recovered. p. 3 The phase diagram in terms of T and Td is plotted
in Fig. 4, from which we can find that with faster driving,
the area of the hysteresis loop shrinks and finally disappears at a tricritial point, after which the BKT phase
transition is recovered[33]. its critical behavior) thus make them differ from the
prediction of the zeroth-order approximation in Eq. By slowly changing the temperature along the forwardbackward path, we also find a hysteresis loop in the C1−T
curve as shown in Fig. By slowly changing the temperature along the forwardbackward path, we also find a hysteresis loop in the C1−T
curve as shown in Fig. 3 (b), which indicates a 1st order
instead of a BKT-type of finite-temperature phase transition.Where the paper says it
Where the paper says it
Where the paper says it
Pruning redundant reasoning tokens
The authors state that the disordered phase in the former is actually an 1D phase, thus differs from the high-temperature phase of the 2D XY model, even though both of them are short-ranged correlated. The difference can be reflected in the temperature (T) dependence of the correlation lengths (ξ), as shown by a linear relation between 1/ξ and T in the slowly driven phase, agreeing with the 1D XY model but differing from the disordered phase. p. 2 This difference can be reflected in the temperature (T) dependence of the correlation lengths (ξ). In
the slowly driven phase, the inset of Fig. It is known that at sufficiently
high temperature, the thermal fluctuations overwhelm
the quantum fluctuations, thus the finite-temperature
phase transition of a quantum model usually shares the
same universality class with its classical counterpart[38]. Therefore, we expect that the nature of dimension-raising
NEPT will not be changed by quantum fluctuations, at
least for sufficiently high temperature.Where the paper says it
Where the paper says it
Where the gains come from
The mechanism differs from dimensional crossover phenomena with tunable anisotropic inter-chain or inter-layer couplings in that the high-dimensional physics emerges spontaneously via a phase transition instead of crossover. Furthermore, it is different from dimensional reduction phenomena which are induced either by the presence of anisotropic orbit order or by geometric frustration. p. 4 It differs
from intensively-studied dimensional crossover phenomena with tunable anisotropic inter-chain or inter-layer
couplings[1–7], in the sense that the high-dimensional
physics in our case emerges spontaneously via a phase
transition instead of crossover. Due to its nonequilibrium
feature, the proposed mechanism is also different from
those governing the dimensional reduction phenomena,
which are induced either by the presence of anisotropic
orbit order[8, 9], or by the geometric frustration[10–12]. The ensemble average over the noise trajectories can be performed by directly sampling over a set
of independent noise realizations. We will focus on the
long-time asymptotic behaviors of the model and their
dependence on the driving period Td and temperature T.Where the paper says it
Where the paper says it
How the training signal is built
The system dynamics are governed by a Langevin equation modeling thermal coupling, where mechanical torque fi(t) takes the form of Xα=x,y Jα(t) sin(θi − θi+⃗eα). The thermal noise ξi(t) is modeled by a white noise with strength D squared = 2kBT γI where T is the temperature of the bath. p. 1 The thermal noise ξi(t) is a zero-mean stochastic torque,
which can be modeled by a white noise: ⟨ξi(t)ξj(t
′
)⟩ξ =
D2
δijδ(t − t
′
), where ⟨·⟩ξ indicates the ensemble average over different noise trajectories. D is the strength of
the noise and satisfies the classical fluctuation-dissipation
theorem: D2 = 2kBT γI where T is the temperature of
the bath. Numerically, we solve the stochastic differential Eq. (3)
using the Heun algorithm[32] with a discrete time step
of ∆t = 10−2J
−1
.Where the paper says it
Where the paper says it
Conclusion and outlook
The paper concludes by proposing a mechanism that allows for the dynamic adjustment of the effective dimension of many-body systems. The authors suggest investigating the properties of emergent frustrated systems, such as orderby-disorder phenomenon, compared to their equilibrium counterparts. p. 4 It would be interesting to investigate the properties of such emergent frustrated systems (e. the orderby-disorder phenomenon, fractionalized excitations), and
compared them to their equilibrium counterparts. Acknowledgments. — This work is supported by the
National Key Research and Development Program
of China (2020YFA0309000,2024YFA1408303), Natural Science Foundation of China (Grant No.Where the paper says it
Where the paper says it
Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.
Improvements for AI systems
-
Dynamically tunable dimensionality realization allows for many-body phases with arbitrary dimensions by tuning a periodically driven system. This enables systems to transition between phases of different effective dimensions rather than just exhibiting crossovers.
-
The system can exhibit a nonequilibrium phase transition (NEPT) from an asymmetric disordered phase to an emergent isotropic 2D critical phase when the driving frequency exceeds a critical value. This NEPT can qualitatively change the finite temperature critical behavior, shifting it from Berezinski-Kosterlitz-Thouless (BKT) type to first order.
-
The model can be generalized to four dimensions by using a time-dependent Hamiltonian with coupling strengths switched at different quarters of the driving period. This allows for the realization of many-body phases with effective dimensions higher than three, such as a 4D XY model under fast driving.
-
The system can exhibit hysteresis in correlation functions, indicating a first order phase transition or phase coexistence between metastable and stable states depending on initial conditions. This hysteresis loop in C1 versus Td allows for the study of the lifetime of the metastable state and its dependence on system size.
-
Experimental platforms like optical tweezers with programmable reconfiguration and Rydberg atom arrays can realize this model by dynamically modulating lattice geometry or interaction strengths. These setups allow for observing dimension-raising NEPTs in systems such as hardcore bosons, where quasi-1D features transition to emergent 2D isotropic phases.
Abstract
We propose a periodically driven system whose dimensionality is an emergent property that can be tunable, thus enables us to realize not only many-body phases with arbitrary dimensions, but also phase transitions, instead of crossovers, between phases with various dimensions. We study an interacting rotor model whose instantaneous Hamiltonian keeps the one-dimensional (1D) feature at any given time. Despite this, an emergent two-dimensional (2D) phase appears when the driving frequency exceeds a critical value, at which a dimension-raising phase transition takes place. We find that the nonequilibrium feature of the system could qualitatively change the finite temperature critical behavior of the emergent 2D phase and make it different from its equilibrium counterpart. A four-dimensional (4D) generalization and experimental realizations of the proposed model based on a programmable reconfiguration technique in optical tweezers setups have also been discussed.
Sources
Related papers
- Emergent long-tail dynamics in driven magnets with dynamical frustration
- Statistical Mechanics of Dynamical System Identification
- Second law of thermodynamics in closed quantum many-body systems
- Chemical potentials from structure factors: II. Charged multi-component mixtures
- Genuine quantum scars in Floquet chaotic many-body systems
- Stochastic Thermodynamics for Autoregressive Generative Models: A Non-Markovian Perspective