betaFeedback

WorksPhysics and Astronomy

Heat transfer mediated by the Berry-phase in non-reciprocal many-body systems

Abstract

We investigate the adiabatic evolution of thermal state in non-reciprocal many-body systems coupled to their environment and subject to periodic drivings. In such systems we show that besides the dynamical phase a geometrical phase can exist and it drives the relaxation dynamic of the system. On the contrary to the dynamical phase which always pushes the system toward its equilibrium state we show that the geometric phase can speed up or reduce the speed of relaxation process. These results could have applications in the field of thermal management of complex systems.

I Introduction

Understanding and controlling the time evolution of the thermal state of a system in non-equilibrium situation is of tremendous importance in physics. Many strategies have been implemented to date to actively control this evolution using an external driving. Hence, by modulating some intensive quantities, such as the temperature or the chemical potential, an additional flux to the primary flux induced by a temperature bias can be generated and used to control heat exchanges. This shuttling effect[1, 2, 3, 4] results from the variation of the local curvature of flux with respect to these parameters. When the system displays a negative differential thermal resistance (i.e. a negative curvature of flux), this effect can contributes to inhibit the primary flux and even can pump heat from the cold to the hot part of the system. A slow cycling modulation of control parameters near-topological singularities[5, 6] such as exceptional points can also be used to enhance or inhibit energy exchanges within a system. Finally, the spatio-temporal modulation of thermal properties, such as the thermal conductivity or the specific heat, in systems can give rise[7] to an effective convective flux which superimposes to the diffusive flux. This leads to an apparent change of heat transport regime which can be exploited to control heat flows in solids networks at mesoscopic and macroscopic scales. Beside these developpements the concept of geometric phase theorized by Berry[8] has been exploited to develop novel pumping strategies in quantum and classical systems. Inspired by the Thouless charge pumping[9] heat pumping in solid-state systems has been proposed[10] to control heat flux in numerous classical and quantum systems[11, 12].

In this work, we introduce a general theory to describe the temporal evolution of thermal state of arbitrary non-reciprocal many-body systems[13, 14, 15] close to their equilibrium state under the action of external periodic drivings. In the adiabatic limit we show that this modulation can be used to slightly modify the relaxation dynamics.

We show that in this limit a geometrical phase can exist which adds to the dynamical phase. We discuss the necessary conditions for the existence of the geometrical phase and provide a general example in a two-body system interacting with an external bath as well an example for near-field heat transfer in a many-body system.

II Time evolution of thermal state in N-body systems

To start let us consider a generic many-body system made with NN bodies in mutual interaction and in interaction with an external bath at temperature TbT_b. The time evolution of thermal state T=(T1,…,TN)\mathbf{T}=(T_1,\ldots,T_N) of this system under temporal driving is governed by an energy balance (master) equation of the general form

CidTidt=∑j≠iPj→i(T;Tb,t),i=1,…,N.(1)C_i\frac{dT_i}{dt}=\sum_{j\ne i}\mathcal{P}_{j\to i}(\mathbf{T};T_b,t),\quad i=1,\ldots,N. \tag*{(1)}

Here Pj→i\mathcal{P}_{j\to i} denotes the power received by the ithi^{\text{th}} element from the jthj^{\text{th}} element within the system and CiC_i its heat capacity. Close to the equilibrium state Teq=(Tb,…,Tb)t\mathbf{T}_{\mathrm{eq}}=(T_b,\ldots,T_b)^t the net power can be linearized and expressed in term of pairwise thermal conductance

Gij=lim⁡Tj→TiPj→iTj−Ti(2)G_{ij}=\lim_{T_j\to T_i}\frac{\mathcal{P}_{j\to i}}{T_j-T_i} \tag*{(2)}

and of conductance

Gib=lim⁡Tb→TiPb→iTb−Ti(3)G_{ib}=\lim_{T_b\to T_i}\frac{\mathcal{P}_{b\to i}}{T_b-T_i} \tag*{(3)}

between the bath and each element. In this approximation equations (1) can be recast in the matrix form

C(t)dTdt=G^(t)T+S(t),(4)\mathcal{C}(t)\frac{d\mathbf{T}}{dt}=\widehat{\mathcal{G}}(t)\mathbf{T}+\mathbf{S}(t), \tag*{(4)}

where G^(t)\hat{\mathbf{G}}(t) denotes the conductances matrix with

G^ij=−∑j≠i(Gij+Gib)δik−(1−δik)Gik,(5)\hat{G}_{ij}=-\sum_{j\ne i}(G_{ij}+G_{ib})\delta_{ik}-(1-\delta_{ik})G_{ik}, \tag*{(5)}

C=diag⁡(C1,…,CN)\mathbf{C}=\operatorname{diag}(C_1,\ldots,C_N) is the heat capacity matrix and S(t)\mathbf{S}(t) is the source vector induced by the bath (i.e. Si=GibTbS_i=G_{ib}T_b). The time evolution of thermal state is given by the following expression

T(t)=U(t0,t)T0+∫t0tU(τ,t)C−1(τ)S(τ) dτ.(6)\mathbf{T}(t)=\mathbf{U}(t_0,t)\mathbf{T}_0+\int_{t_0}^{t}\mathbf{U}(\tau,t)\mathbf{C}^{-1}(\tau)\mathbf{S}(\tau)\,d\tau. \tag*{(6)}

Here U\mathbf{U} denotes the propagator of differential system (4) and T0=(T1(t0),…,TN(t0))t\mathbf{T}_0=(T_1(t_0),\ldots,T_N(t_0))^t is the initial thermal state. In reciprocal systems (i.e. G^ij=G^ji\hat{G}_{ij}=\hat{G}_{ji}) the fundamental matrix can be expressed in term of the exponential of the conductance matrix so that the thermal state (6) can easily be calculated. On the other hand, in non-reciprocal systems (i.e. G^ij≠G^ji\hat{G}_{ij}\ne\hat{G}_{ji}), the situation is more tricky and no simple expression of thermal state can be derived. Nevertheless, following Garrisson and Wright[16] we can, in the adiabatic appoximation, seek a solution of this system by expanding it over the basis {φ⊥}n\{\varphi_{\perp}\}_n of eigenstates of G^\hat{\mathbf{G}}. To proceed we first recast the master equation as

dT~dt=G~(t)T~,(7)\frac{d\widetilde{\mathbf{T}}}{dt}=\widetilde{\mathbf{G}}(t)\widetilde{\mathbf{T}}, \tag*{(7)}

by setting T~=(T,Tb)t\widetilde{\mathbf{T}}=(\mathbf{T},T_b)^t and

G~=(G^⋮G1b⋮GNb⋯⋯⋯(0 ⋯ 0)⋮0).(8)\widetilde{\mathbf{G}}= \begin{pmatrix} \hat{\mathbf{G}} & \vdots& \begin{matrix}G_{1b}\\ \vdots\\G_{Nb}\end{matrix}\\ \cdots& \cdots& \cdots\\ (0\ \cdots\ 0) & \vdots& 0 \end{pmatrix}. \tag*{(8)}

Note, that for convenience G^\hat{\mathbf{G}} is now the conductance matrix normalized by the heat capacities. Then we seek a solution of Eq. (7) using the following expansion

T~(t)=∑i=1N+1ci(t)eγdi(t)φi(t),(9)\widetilde{\mathbf{T}}(t)=\sum_{i=1}^{N+1}c_i(t)e^{\gamma_{di}(t)}\varphi_i(t), \tag*{(9)}

where γdi(t)=∫0tλi(τ)dτ\gamma_{di}(t)=\int_0^t\lambda_i(\tau)d\tau is the dynamical phase associated the eigenvalue λi\lambda_i of G~\widetilde{\mathbf{G}} while φi\varphi_i is its eigenvector. It is easy to check that λN+1=0\lambda_{N+1}=0 is an eigenvalue with the corresponding eigenvector φN+1=1N+1=(1,…,1)t\varphi_{N+1}=\mathbf{1}_{N+1}=(1,\ldots,1)^t. It turns out that the temperatures vector takes the form

T~(t)=∑i=1Nci(t)eγdi(t)φi(t)+cN+11N+1.(10)\widetilde{\mathbf{T}}(t)=\sum_{i=1}^{N}c_i(t)e^{\gamma_{di}(t)}\varphi_i(t)+c_{N+1}\mathbf{1}_{N+1}. \tag*{(10)}

Since T~→(Teq,Tb)t\widetilde{\mathbf{T}}\to(\mathbf{T}_{eq},T_b)^t after a sufficiently long time (G^\hat{\mathbf{G}} being strictly definite negative) we can write the solution

III Adiabatic limit and geometrical phase

T~(t)=∑i=1Nci(t)eγdi(t)φi(t)+Tb1N+1,(11)\widetilde{\mathbf{T}}(t)=\sum_{i=1}^{N}c_i(t)e^{\gamma_{di}(t)}\varphi_i(t)+T_b\mathbf{1}_{N+1}, \tag*{(11)}

Taking the projection of this temperature vector on the subspace e1⊗…⊗eN\mathbf{e}_1\otimes\ldots\otimes\mathbf{e}_N generated by the basis vectors ei\mathbf{e}_i of the canonic base, we get finally the thermal state

T(t)=∑i=1Nci(t)eγdi(t)φ⊥i(t)+Tb1N,(12)\mathbf{T}(t)=\sum_{i=1}^{N}c_i(t)e^{\gamma_{di}(t)}\varphi_{\perp i}(t)+T_b\mathbf{1}_N, \tag*{(12)}

where according to the form of the block matrix defined in (8) φ⊥i\varphi_{\perp i} is the ithi^{\mathrm{th}} eigenvector of G^\hat{\mathbf{G}}.

Inserting the solution in Eq. (12) into the master equation (4) we obtain, after removing the subscript ⊥\perp for readability reasons, the relations

∑i=1N∂t(cieγdiφi)=∑i=1Nci(t)eγdiG^φi+TbG^1N+S.(13)\sum_{i=1}^{N}\partial_t(c_i e^{\gamma_{di}}\varphi_i)=\sum_{i=1}^{N}c_i(t)e^{\gamma_{di}}\hat{\mathbf{G}}\varphi_i+T_b\hat{\mathbf{G}}\mathbf{1}_N+\mathbf{S}. \tag*{(13)}

As TbG~1N=−ST_b\widetilde{\mathbf{G}}\mathbf{1}_N=-\mathbf{S} and G^φi=λiφi\hat{\mathbf{G}}\varphi_i=\lambda_i\varphi_i we have equivalently

∑i=1N(c˙ieγdiφi+cieγdiφ˙i)=0.(14)\sum_{i=1}^{N}(\dot{c}_i e^{\gamma_{di}}\varphi_i+c_i e^{\gamma_{di}}\dot{\varphi}_i)=0. \tag*{(14)}

Multiplying this relation by ψj\psi_j the jthj^{\mathrm{th}} eigenvector of the transpose matrix G^t\hat{\mathbf{G}}^t and using the biorthogonality relations ψi⋅φj=δij\psi_i\cdot\varphi_j=\delta_{ij} between the eigenvectors of G^\hat{\mathbf{G}} and G^t\hat{\mathbf{G}}^t we find

c˙jeγdj+∑i=1Ncieγdiψj⋅φ˙i=0.(15)\dot{c}_j e^{\gamma_{dj}}+\sum_{i=1}^{N}c_i e^{\gamma_{di}}\psi_j\cdot\dot{\varphi}_i=0. \tag*{(15)}

Now, assuming a slow (adiabatic) variation of external control parameters with respect to time this system simplifies into

c˙j+cjψj⋅φ˙j=0,(16)\dot{c}_j+c_j\psi_j\cdot\dot{\varphi}_j=0, \tag*{(16)}

whose the solution reads

cj(t)=αjeγgj(t),(17)c_j(t)=\alpha_j e^{\gamma_{gj}(t)}, \tag*{(17)}

with

γgj(t)=−∫0tdτ ψj⋅φ˙j(τ)(18)\gamma_{gj}(t)=-\int_0^t d\tau\,\psi_j\cdot\dot{\varphi}_j(\tau) \tag*{(18)}

the so called geometric phase. Hence, in the adiabatic limit the thermal state can be written as

T(t)=∑i=1Nαieγdi(t)eγgi(t)φi(t)+Tb1N.(19)\mathbf{T}(t)=\sum_{i=1}^{N}\alpha_i e^{\gamma_{di}(t)}e^{\gamma_{gi}(t)}\varphi_i(t)+T_b\mathbf{1}_N. \tag*{(19)}

This expression is the classical analog of the Berry’s result to arbitrary non-reciprocal many body systems. The constants αi\alpha_i can be readily calculated from the initial thermal state T0\mathbf{T}_0.

A SIMPLE TOY MODEL

To show to what extent the flux is affected by the geometrical phase we consider below a generic system as sketched in (Figure 1-a) made of two objets in mutual interaction and in interaction with the external bath. In this case the conductance matrix takes the form

Effective magnetic field, trajectories, and cumulative geometrical phases

Figure 1. (a) Non-vanishing component of the effective magnetic field in the parameter space (x,y)(x,y). The inset is a sketch of studied system made with two solids in mutual interaction with asymmetric conductances and in interaction with an external bath. (b) Trajectory in the parameter space associated to two different eigenvalues of G^\hat{G} when Λ=τ\Lambda=\tau and (c) Λ=10τ\Lambda=10\tau. (d) Cumulative geometric phases with τ=1 s\tau=1\ \mathrm{s}.

G^=(−aG12G21−b),(20)\hat{G}=\begin{pmatrix} -a & G_{12} \\ G_{21} & -b \end{pmatrix}, \tag*{(20)}

with a=G12+G1ba=G_{12}+G_{1b} and b=G21+G2bb=G_{21}+G_{2b}. The eigenvectors of G^\hat{G} and of its transpose are

φi=(1G21b+λi)andψi=1βi(1G12b+λi),(21)\varphi_i=\begin{pmatrix} 1 \\ \frac{G_{21}}{b+\lambda_i} \end{pmatrix}\quad\text{and}\quad\psi_i=\frac{1}{\beta_i}\begin{pmatrix} 1 \\ \frac{G_{12}}{b+\lambda_i} \end{pmatrix}, \tag*{(21)}

where

λi=12tr⁡G^±[14tr⁡2G^−det⁡G^]1/2(22)\lambda_i=\frac{1}{2}\operatorname{tr}\hat{G}\pm\left[\frac{1}{4}\operatorname{tr}^{2}\hat{G}-\det\hat{G}\right]^{1/2} \tag*{(22)}

denotes the (real) eigenvalues of G^\hat{G}, βi\beta_i being calculated from the normalization relation φi⋅ψi=1\varphi_i\cdot\psi_i=1. By setting xi=G21b+λix_i=\frac{G_{21}}{b+\lambda_i} and yi=G12G21y_i=\frac{G_{12}}{G_{21}} it follows that the geometric phase accumulated between the initial instant up to time tt is

γgi(t)=−∫t0tdτxix˙iyi1+xi2yi.(23)\gamma_{gi}(t)=-\int_{t_0}^{t}\mathrm{d}\tau\frac{x_i\dot{x}_i y_i}{1+x_i^2y_i}. \tag*{(23)}

Now, suppose that G^(t)=G^(R(t))\hat{G}(t)=\hat{G}(\mathbf{R}(t)), where R=(R1,R2)t\mathbf{R}=(R_1,R_2)^t denotes a set of control parameters and that there is a time TT for which R(t0+T)=R(t0)\mathbf{R}(t_0+T)=\mathbf{R}(t_0). Hence, we consider a closed trajectory in parameter space. Then the accumulated geometrical phase in Eq. (23) during this interval TT can be written as

γgi(T)=∮∂SdR⋅Ai,(24)\gamma_{gi}(T)=\oint_{\partial S}\mathrm{d}\mathbf{R}\cdot\mathbf{A}_i, \tag*{(24)}

where

Ai=(∇Rφi)t⋅ψi(25)\mathbf{A}_i=(\nabla_{\mathbf{R}}\varphi_i)^t\cdot\psi_i \tag*{(25)}

is analog to a vector potential whose associated magnetic field is Bi=∇×Ai\mathbf{B}_i=\nabla\times\mathbf{A}_i. Here ∂S\partial S is the path of the trajectory in parameter space. Using the Stoke theorem with the oriented surface SS enclosed by the countour ∂S\partial S generated by the change of parameter R\mathbf{R} during the interval [t0,t0+T][t_0,t_0+T] we can express the geometric phase in term of Bi\mathbf{B}_i as

γgi(T)=∫SdS⋅Bi(26)\gamma_{gi}(T)=\int_S\mathrm{d}\mathbf{S}\cdot\mathbf{B}_i \tag*{(26)}

where dS\mathrm{d}\mathbf{S} is the infinitesimal oriented surface element on SS.

These expressions allow us to determine the vector potential and magnetic field in our two-body system for any trajectory R\mathbf{R}. With Eq. (23) for one period TT we find

Ai=xiyiβi(∇Rxi).(27)\mathbf{A}_i=\frac{x_i y_i}{\beta_i}(\nabla_{\mathbf{R}}x_i). \tag*{(27)}

Note that this expression is not gauge invariant as it is typical for vector potentials. For example, also the expression

Ai′=−xi22βi(∇Ryi)(28)\mathbf{A}'_i=-\frac{x_i^2}{2\beta_i}(\nabla_{\mathbf{R}}y_i) \tag*{(28)}

is a valid representation of the vector potential. On the other hand, the magnetic field is gauge invariant, and consequently we obtain for both Ai\mathbf{A}_i and Ai′\mathbf{A}'_i the magnetic field

B=−(∇Rxi)×(∇Ryi)xiβi2.(29)\mathbf{B}=-\left(\nabla_{\mathbf{R}}x_i\right)\times\left(\nabla_{\mathbf{R}}y_i\right)\frac{x_i}{\beta_i^2}. \tag*{(29)}

Interestingly, from this expression it is obvious that B=0\mathbf{B}=\mathbf{0} if yi=1y_i=1, i.e. if G12=G21G_{12}=G_{21}. Hence, non-reciprocity is a necessary condition for the appearance of a Berry phase. Furthermore, the magnetic field vanishes if ∇Rxi\nabla_{\mathbf{R}}x_i and ∇Ryi\nabla_{\mathbf{R}}y_i are parallel or antiparallel. This happens for a constant background conductance if ∇RG12\nabla_{\mathbf{R}}G_{12} and ∇RG21\nabla_{\mathbf{R}}G_{21} are parallel or antiparallel, i.e. if for example G12G_{12} and G21G_{21} change in phase with the external parameter change. Therefore, if G12G_{12} and G21G_{21} are in phase or phase-shifted by π\pi, then there is no Berry phase.

The above statements are very general. Now, we take a concrete path in the parameter space by choosing the two natural parameters xix_i and yiy_i for a path embedded in R3\mathbb{R}^3, i.e. R(t)=(xi(t),yi(t),0)t\mathbf{R}(t)=(x_i(t),y_i(t),0)^{t}. Then the magnetic field is

B=−xiβi2ez.(30)\mathbf{B}=-\frac{x_i}{\beta_i^2}\mathbf{e}_z. \tag*{(30)}

In Figure 1 we plot the non-vanishing component of Bi\mathbf{B}_i in the space of control parameters as well as the trajectories defined by the parametric curve R=(xi(t),yi(t))t\mathbf{R}=(x_i(t),y_i(t))^{t}. It is clear from expression (26) that the geometric phases γgi(T)\gamma_{gi}(T) depends intimately on the shape of closed contour generated by these trajectories during one oscillation period.

In Figure 1-b and Figure 1-c we show some typical trajectories followed by these parameters during one cycle of modulation and the corresponding cumulative geometric phase (Figure 1-d) when the exchanges conductances obey to the following variation laws

G12=G+δGcos⁡(2πΛt),(31)G_{12}=G+\delta G\cos\left(\frac{2\pi}{\Lambda}t\right), \tag*{(31)}
G21=H+δHcos⁡(2πΛt+θ),(32)G_{21}=H+\delta H\cos\left(\frac{2\pi}{\Lambda}t+\theta\right), \tag*{(32)}
G1b=g+δgcos⁡(2πτt),(33)G_{1b}=g+\delta g\cos\left(\frac{2\pi}{\tau}t\right), \tag*{(33)}
G2b=h+δhcos⁡(2πτt+θ).(34)G_{2b}=h+\delta h\cos\left(\frac{2\pi}{\tau}t+\theta\right). \tag*{(34)}

The numerical results show clearly the sensitivity of geometric effect to the multiperiodicity of driving. When Λ∼τ\Lambda\sim\tau and Λ→∞\Lambda\to\infty the integration surface reduces to a single point so that the Berry phase vanishes. On the contrary, with a finite period Λ\Lambda the closed contour delimits a non-vanishing area and this contour is simple (without crossing point) provide that Λ=τ\Lambda=\tau whereas in the case of multiperiodic driving (i.e. Λ≠τ\Lambda\ne\tau) this contour presents, in general, several crossing points and it can be decomposed into several loops which are traveled either in clockwise or anti-clockwise direction. In the first case if a loop is in the first quadrant (i.e. xi>0x_i>0 and yi>0y_i>0) and the contour is browsed in clockwise direction resulting in a positive contribution to the geometric phase γi\gamma_i. In other words, during this period the geometric phase tends to insulate the different parts of the system. On the other hand, if the loop is browsed in the opposite direction the generated geometric phase is negative and the relaxation process is accelerated. This phenomenon is analog to the coiled light mechanism discovered by Chiao et al. [17, 18], where the parameters space corresponds to the light polarization state.

Finally, the relaxation dynamic is plotted in Figure 2 when Λ=10τ\Lambda=10\tau and Λ=τ\Lambda=\tau both using the approximate adiabatic solution in Eq. (19) and a numerical exact solution of the energy balance equation in Eq. (1) using a Runge-Kutta (RK) type method. When the oscillation period Λ\Lambda of main coupling channel is sufficiently large compared to the relaxation time of system we see (Figure 2-a) that the relaxation dynamics is properly described by the adiabatic approximation (12). On the other hand, for shorter oscillations we observe in Figure 2-b a deviation between this approximation and the exact evolution calculated with the RK method. Notice that, the cumulated dynamic phases (always negative) being, at large time scale, larger than the cumulated geometric phase as shown in Figure 2-c and Figure 2-d it always pushes the system toward its equilibrium state. On the other hand, the geometrical phase, which can be either positive or negative, has an oscillatory character and its accumulation after one or several periods can be either positive or negative so that they can speed up or speed down the relaxation process. Unfortunately, this geometric effect cannot be persistent in time and the system finishes at the end to be driven only by the dynamical phase. In our toy model the two Berry phases γgi(Λ)\gamma_{gi}(\Lambda) become much smaller than the dynamical phases γdi(Λ)\gamma_{di}(\Lambda) after one period Λ=10τ\Lambda= 10\tau but during the first cycle they can be of the same order of magnitude and even much larger as visualized in Figure 3. Optimization procedures will certainly be able to find the maximal Berry phases in this simple system and more generally in arbitrary many-body systems. However, this problem goes far beyond the scope of the present work.

Image not extracted

Figure 2. Approximate (adiabatic) and exact relaxation dynamics of hot bodies in a two body system driven by the conductance matrix (31)-(34) with G=1G=1, H=0.8H=0.8, g=0.5g=0.5, h=0.3h=0.3, δG=0.5\delta G=0.5, δH=0.4\delta H=0.4, δg=δh=0.1\delta g=\delta h=0.1 and θ=π/2\theta=\pi/2. The initial temperature is T1(0)=400 KT_1(0)=400\ K while the external bath is set at Tb=300 KT_b=300\ K in the case (a) Λ=10τ\Lambda=10\tau and (b) Λ=τ\Lambda=\tau with τ=1 s\tau=1\ \mathrm{s}. (c)-(d) Cumulated geometric and dynamic phases.

Image not extracted

Figure 3. Geometrical and dynamical phases for the model as in Figure 2-c for Λ=10τ\Lambda= 10\tau but with G=0.01G = 0.01, H=0.1H = 0.1, g=0.01g = 0.01, h=0.01h = 0.01, δG=0.005\delta G = 0.005, δH=0.05\delta H = 0.05, δg=0.1g\delta g = 0.1g, δh=0.1h\delta h = 0.1h, and θ=π/2\theta= \pi/2.

conclusion

In summary, although the physics of non-reciprocal systems remain today largely elusive, the results introduced in this work highlight the peculiarities of relaxation process for this systems when they are driven by periodic external actuations. On the contrary to reciprocal systems, the presence of a geometrical phase superimposes to the dynamical phase and has the potential to significantly alter the relaxation dynamic of systems. We have shown that this phase can be used either to accelerate or reduce the speed of relaxation. We hope that these preliminary results will stimulate research on the thermal control of non-reciprocal systems. On a theoretical point of view it would be interesting to explore the role played by the dissipation mechanisms induced by the external driving as well as the potential of multispectral drivings on the relaxation dynamics. The non-adiabatic control of these systems remains also a challenging problem. These problems will be addressed in subsequent studies.

Acknowledgments

P.B.-A. acknowledges support from the Agence Nationale de la Recherche in France through the NBodheat project (ANR-21-CE30-0030-01). S.-A. B. acknowledges support from Heisenberg Programme of the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) under the project No. 404073166. This research was supported in part by the National Science Foundation under Grant No. NSF PHY-1748958.

References

  1. [1]N. Li, P. Hänggi and B. Li, Europhys. Lett. 84, 40009 (2008).
  2. [2]N. Li, F. Zhan, P. Hänggi and B. Li, Phys. Rev. E 80, 011125, (2009).
  3. [3]I. Latella, R. Messina, J. M. Rubi and P. Ben-Abdallah, Phys. Rev. Lett. 121, 023903 (2018).
  4. [4]R. Messina and P. Ben-Abdallah, Phys. Rev. B 101, 165435 (2020).
  5. [5]H. Li, L. J. Fernández-Alcázar, F. Ellis, B. Shapiro and T. Kottos, Phys. Rev. Lett., 123, 165901 (2019).
  6. [6]G. Xu, Y. Li, W. Li, S. Fan and C.-W. Qiu, Phys. Rev. Lett. 127, 105901 (2021).
  7. [7]D.Torrent, O. Poncelet and J.-C. Batsale, Phys. Rev. Lett., 120, 255001 (2018).
  8. [8]M. V. Berry, Proc. Math. Phys. Eng. Sci. 392, 45 (1984).
  9. [9]D. J. Thouless, Phys. Rev. B 27, 6083, (1983).
  10. [10]Z. Wang, L. Wang, J. Chen and C. Wang, J. Ren, Front. Phys. 17 (1): 132021 (2022).
  11. [11]J. Ren, S. Liu and B. Li, Phys. Rev. Lett. 108, 210603 (2012).
  12. [12]T. Chen, X.-B. Wang and J. Ren, Phys. Rev. B 87, 144303 (2013).
  13. [13]P. Hänggi and F. Marchesoni, Rev. Mod. Phys. 81, 387 (2009).
  14. [14]N. Li, J. Ren, L. Wang, G. Zhang, P. Hänggi and B. Li, Rev. Mod. Phys. 84, 1045 (2012).
  15. [15]S.A. Biehs, R. Messina, P.S. Venkataram, A.W. Rodriguez, J.C. Cuevas, and P. Ben-Abdallah, Rev. Mod. Phys. 93, 025009 (2021).
  16. [16]J. C. Garrison and E. M. Wright, Phys. Lett. A, 128, 3,4 (1988).
  17. [17]R. Y. Chiao and Y.-S. Wu, Phys. Rev. Lett., 57 (8). 933-936 (1986).
  18. [18]A. Tomita and R. Y. Chiao, Phys. Rev. Lett., 57 (8). 937-940 (1986).

Paper details

vlnc:W_952atrka
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

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

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents