Shanghai Key Laboratory of Special Artificial Microstructure Materials and Technology, Center for Phononics and Thermal Energy Science, China-EU Joint Lab on Nanophononics, School of Physics Science and Engineering, Tongji University, Shanghai 200092, China
Fund Project:Project supported by the National Natural Science Foundation of China (Grant Nos. 11935010, 11775159).
Received Date:15 September 2021
Accepted Date:25 October 2021
Published Online:05 December 2021
Abstract:With the in-depth understanding of nano-/micro-scaled systems and the developing of the corresponding experimental techniques, the heat transport and energy conversion processes in these small systems have attracted much interest recently. In contrast to the static manipulation methods, which hinge on the steady nonequilibrium sources such as temperature bias, chemical potential difference, etc., the temporal driving methods can control small systems in nonequilibrium non-steady states with much more versatility and universality. The research on periodically driven small systems holds both fundamental and pragmatic promises. This review is based on the fundamental concept of geometry. By analyzing the geometric phase and thermodynamic length in the transport process and the energy conversion process, we provide a unified perspective for the recent researches on the thermodynamic properties of driven nonequilibrium quantum systems. Thermodynamic geometry not only is the intrinsic origin of the nontrivial transport and dissipation, but also provides us with an all-applicable theoretical framework. The discussion over the geometry would yield multiple thermodynamic constraints on the transport and energy conversion, and can naturally construct a general optimization method as well. This will conduce to a better understanding of functionality for nonequilibrium quantum many-body systems acting as thermal machines. Also, this will inspire people to design quantum thermal machines with simultaneously more ideal performance, i.e. higher efficiency, higher power and higher constancy. Keywords:nonequilibrium quantum system/ geometric phase/ thermodynamic distance/ heat pump/ quantum heat engine
2.周期性驱动量子输运中的几何相与距离在周期驱动热力学系统中, 几何相和热力学距离作为两个重要概念, 为我们提供了一种统一描述经典/量子系统的理论框架. 在几何上, 在一个可以定义任意两点距离的空间中, 可以使用Riemann空间中的度规来描述相近两点间的距离. 此无穷小距离沿着曲线的积分就是此路径首尾两点间的距离. 在一个非平坦的空间中, 一个矢量途经一条闭合路径进行平行移动, 其末了矢量与初始矢量间的夹角被称为和乐(holonomy)角, 它即是几何相的数学含义. 一个光滑空间中, 此和乐角可以通过路径上几何联络的线积分得到. 从另一个角度, 联络本身定义了曲面上向量平行移动的方式, 它使我们可以把不同位置处切空间内的向量进行比较. 在物理上, 这些数学概念也有对应的意义. 开放系统内热泵浦、功转换等过程中, 几何相描述了经由一个绝热驱动, 系统分布在回到其初始状态时, 额外累积定向转移热量/功的多少[15]. 它可以由参数空间内局域的几何联络的积分得到. 描述能量流(功/热)的几何联络可以类比孤立量子系统中的Berry联络, 它在驱动路径上的投影给出了在此参数点进行一个微小绝热驱动所引起的额外热流泵浦/功输出. 另外, 若驱动速度较小, 但并非完全绝热, 则热力学距离则描述了驱动过程中引起的耗散大小[23]. 相近参数点间的热力学距离可以定义出一个参数空间中的度规. 这样的物理图像如图1所示. 接下来, 我们给出具体的分析和推导. 图 1 周期性驱动非平衡量子输运和其中几何性质的示意图 (a) 非平衡量子系统示意图. 量子系统由一个包含多个能级的系统来表示, 它可以与多个热库相连. 热库温度($ {T}_{\mathrm{h}} $和$ {T}_{\mathrm{c}} $)和系统参数$(\lambda)$都被含时地驱动. 由此, 可以产生系统与热库间的热量交换($ {Q}_{\mathrm{h}} $和$ {Q}_{\mathrm{c}} $)以及系统的功输出(W). (b) 此非平衡量子系统在参数空间($\boldsymbol{\varLambda }\equiv \left(T, \lambda \right) $)中的几何性质. 曲线坐标系表现出非均匀的热力学距离, 而各点的箭头表示几何联络. 几何联络在几何上对应平行移动一个微小参数时带来的和乐(holonomy)角. 热力学距离定义了一个具有度规的黎曼曲面 Figure1. A scheme of periodically driven nonequilibrium quantum transport and its geometry. (a) A diagrammatic nonequilibrium quantum system. The middle quantum system is illustrated by a multi-level system, which is coupled with several thermal reservoirs. The temperature of reservoirs ($ {T}_{\mathrm{h}} $ and $ {T}_{\mathrm{c}} $) and the mechanical parameter of the system ($ \lambda $) are simultaneously and periodically modulated. The heat exchange ($ {Q}_{\mathrm{h}} $ and $ {Q}_{\mathrm{c}} $) and work output $(W)$ are thus generated. (b) The geometry of this nonequilibrium quantum system in the space of parameters $(\boldsymbol{\varLambda }\equiv $$ \left(T, \lambda \right) )$. The curvilinear coordinate is adopted to show the inhomogeneous thermodynamic distance and the local vectors are for the geometric connection, as derived in the main text. Geometrically, the geometric connection is the holonomy angle during an infinitesimal parallel transport and the thermodynamic distance between neighboring points defines a Riemannian space with endowed metric.
其中$ {\mathrm{d}\boldsymbol{S}}^{\mu \nu } $为特定方向的面元, 反对称的几何曲率${\boldsymbol{F}}_{\mu \nu }\equiv {\partial }_{\mu }{\boldsymbol{A}}_{\nu }-{\partial }_{\nu }{{\boldsymbol{A}}}_{\mu }$为参数空间的内禀量, 不依赖于规范的选取. 如果把驱动路径反向, 热的几何相效应也反向. 因此, 在只有纯的几何相效应(动力学相效应为零)的情况下, 系统为可逆热泵浦和热机. 近年来, 大量关于受驱动非平衡量子系统中热流泵浦的研究基本都基于以上描述的理论框架. 如图2(a)所示, 最初关于几何热泵的研究是关于一个量子分子结的声子热泵[15]. 通过动态驱动分子结两端热库的温度, 可以产生一个累积的定向热流. 这个热流的一部分累积量生成函数可以和Berry相位进行对应, 由一个纯几何量给出, 即几何曲率(如图2(b))在驱动包裹的参数空间区域内的积分. 几何相热流的存在, 使得静态条件下普适成立的涨落定理被打破. 后续关于此情况下新的形式的涨落定理被推导得出, 此特殊的涨落定理需要考虑几何相热流的贡献[43]. 图 2 几何相热泵浦 (a) 最初研究的量子分子结系统, 工作介质由一个两能级系统描述, ${p}$和${k}$分别为各个占据数和跃迁速率[15]; (b) 驱动两端温度产生的热泵浦现象, 图中用颜色表示了几何曲率的大小. 几何泵浦的热量为驱动回路包裹的范围内的几何曲率积分[15]. (c), (d)自旋-玻色系统中几何相热流与系统-热库耦合强度$ \alpha $的关系[18]. (c)为无Zeeman劈裂能的情形; (d)为有Zeeman劈裂能的情形[18]. (a), (b)改编自文献[15]; (c), (d)改编自文献[18] Figure2. The geometric heat pump effect. (a) The originally studied quantum molecular junction system. The working medium is described by a quantum two-level system, with ${p}$ and ${k}$ denoting different populations and transition rates[15]. (b) The geometric curvature in the two-temperature parameter space. The color denotes the magnitude of the geometric curvature. The pumped heat is the integral of geometric curvature over the encircled area[15]. (c), (d) The geometrically pumped heat versus the coupling strength between the middle system and reservoirs in a quantum spin-boson system[18]. (c) is for the setup with no Zeeman splitting while the splitting is present in (d) [18]. (a), (b) are adapted from [15], while (c), (d) are adapted from [18].
4.热力学距离: 热机中的不可逆性第3节中描述的几何相贡献没有分析驱动过程中带来的不可逆性和由此带来的热力学效率的减小. 本节将运用热力学距离的概念对此进行讨论. 热力学距离最初源于宏观平衡热力学的研究, 它基于热力学势函数(内能, 熵等)在参数空间中的二阶导数, 且其发散性代表了相变的发生[48,49]. 近年来, 此概念被用来描述单一热库接触的微纳系统的热力学和统计物理性质, 与前面宏观系统的情况不同, 此时的热力学距离基于随机的香农熵($ S= $$ -\mathrm{l}\mathrm{n}p $)来定义[23]. 此时的热力学度规与信息几何中的Fisher信息矩阵直接相关, 即: 当把Fisher信息定义中的概率分布选取为Gibbs分布, 则可以得到对应的热力学度规. 根据此度规计算的两点间的热力学距离确定了这两点间态转换的最小耗散, 这与信息几何中Cramer-Rao界限具有很强的类比性. 此时的热力学距离是一个实验可直接测量的量[50], 它可以帮助人们设计进行微纳系统态转换时最优(耗散最小)的路径方案, 此方案对应着参数空间中由热力学度规确定的测地线[24,51]. 上面一系列重要的研究仅仅关注了受驱动系统的耗散性质, 而没有研究受到周期驱动的微纳系统中涉及的热机功能以及其能量转换效率. 近期, Brandner和Saito[25]的研究工作讨论了这个问题. 类似于本文第2节中的讨论, 他们用几何联络来描述慢驱动产生的功输出, 用热力学度规来描述驱动带来的耗散. 如图3(a)中给出的, 热力学几何限定了有限时间热机工作时的效率-功率权衡关系. 图中的灰色区域为热力学几何给出的不等式所禁止的, 即在一定范围内若要获得更大的效率就要牺牲一部分功率. 而图3(a)中黑线为以匀速驱动的结果, 而橙线则代表了经过优化的热机表现, 具体的优化方式将在下一节中进行描述. 另一方面, 与先前的结果[6]不同, 在此研究的具体模型中, 无论驱动的幅度如何, 量子相干性(由驱动引起)都只会损害热机的工作性能. 图 3 通过热力学距离对慢驱动热机的限制和优化 (a) 周期驱动量子热机中的功率与效率权衡. 灰色区域是根据热力学几何得出的不可能区域, 黑色线对应等速率驱动方式, 而橙色线对应经过优化的驱动方式(驱动速度随时间变化), 这两种驱动方式中驱动速度的具体形式由图(b)给出[25]. (c), (d) 对于驱动量子热机(一个谐振子)的多目标优化[26] (c)不同的优化后的驱动速度; (d) 功的相对涨落与谐振子频率的关系, 此时为效率-功涨落的多目标优化. 红线为匀速率的驱动, 而蓝线对应优化后的驱动, 灰色区域为由热力学几何给出的不可能区域[26]. (a), (b)改编自文献[25]; (c), (d)改编自文献[26] Figure3. The constraint on and optimization of slowly driven quantum heat engine using the thermodynamics distance method: (a) The efficiency-power tradeoff in driven quantum heat engine. The gray area is ruled out by the thermodynamic geometry, with the black and orange line corresponding to the constant-speed driving and optimized driving protocols (driving speed is time dependent) respectively. The driving speed in these two protocols is illustrated in Figure (b)[25]. (c), (d) A multiple target optimization of a driven heat engine composed of a harmonic oscillator: (c) the driving speed of the optimization result with different targets[26]; (d) the relative work fluctuation versus the oscillator’s frequency (system’s parameter) with a multiple target optimization. The red line is for the constant speed driving while the blue one is for the optimized protocol. The gray area is prohibited by the thermodynamic geometry[26]. (a), (b) are adapted from Ref. [25], while (c), (d) are adapted from Ref. [26].