1.State Key Laboratory of Advanced Technology for Materials Synthesis and Processing, Wuhan University of Technology, Wuhan 430070, China 2.College of Science, Wuhan University of Technology, Wuhan 430070, China 3.Wuhan National High Magnetic Field Center, Huazhong University of Science and Technology, Wuhan 430074, China
Fund Project:Project supported by National Key Research and Development Program of China (Grant No. 2018YFE0111500) and the National Natural Science Foundation of China (Grant Nos. 11574243, 11834012)
Received Date:02 November 2020
Accepted Date:10 December 2020
Available Online:10 May 2021
Published Online:20 May 2021
Abstract:Thermoelectric refrigeration technology is an environment-friendly refrigeration technology with broad application prospects. The Peltier effect plays a central role in the thermoelectric refrigeration process, however, the Peltier coefficient is difficult to measure. So in the actual application process, first, the Seebeck coefficient is usually obtained, and then the Peltier coefficient is achieved by the Kelvin's second relation indirectly. It should be noted that the Kelvin's second relation is obtained under linear conditions (Ohm's law, Fourier's law, etc.), while in practice, nonlinear current-voltage relationships (Schottky junction, pn junction, etc.) and nonlinear heat transport relations are common. And quantum effect plays a leading role in the nano-scaled region, then the Peltier effect must consider the influence of nonlinearity, and the applicability of the Kelvin's second relation must also be reconsidered. This paper first summarizes the theoretical derivation of Peltier coefficient and the Kelvin’s second relation by different methods, then discusses the hypothetical conditions used in the derivation process, and points out that the Kelvin’s second relation can be established only under the hypothetical linear conditions. Then, several experimental methods of determining the Peltier coefficient are summarized. It is found that there are still many problems encountered in the measurement of Peltier coefficient, and the Kelvin’s second relation has not been proved accurately by practical experiments. Various side effects (Fourier effect, Thomson effect, Joule effect and Seebeck effect) in the measurement process affect the temperature distribution of the system directly or indirectly, making it difficult to measure Peltier heat. After that, the theoretical work of nonlinear Peltier effect is briefly introduced. In the process of thermal transport and electrical transport on a microscopic scale, quantum effect plays a leading role, and the nonlinear part of the Peltier coefficient gradually emerges. These studies show the cognition of researchers that the Peltier effect gradually changes from linear to nonlinear. The nonlinear Peltier effect not only exists objectively, but also is very important in the practical applications. However, the current research on the nonlinear Peltier effect is still at the theoretical level, and there is almost no experimental work. Finally, we discuss the research strategy and feasible research direction of Peltier effect under nonlinear conditions. An integrated study of the relationship among various heterojunction band structures, interface properties and interface effects is helpful in comprehensively understanding the Peltier effect. With the continuous improvement of experimental conditions and theoretical research, the study of nonlinear Peltier effect is expected to realize a new breakthrough. Keywords:nonlinearity/ Peltier effect/ Kelvin’s second relation/ thermoelectric refrigeration
虽然基于Peltier效应的器件已经在激光二极管冷却、汽车座椅冷却器/加热器等方面有了商业应用[3,5], 但是对Peltier效应本身的研究极少, 这主要可能有两个原因. 一是因为Peltier系数很难通过实验直接精确测量[10-16]. 要直接测量Peltier系数, 必须测量得到Peltier热$ {Q}_{\rm{ab}} $和电流I. 但是Peltier热$ {Q}_{\rm{ab}} $测量时常伴随着Joule热、Fourier热、Thomson热等多种热效应的干扰, 因此Peltier系数的直接精准测量极其困难. 第二个主要原因就是Kelvin第二关系式(Π = αT), 其中α为Seebeck系数, T为绝对温度. 该关系式将Seebeck效应和Peltier效应直接联系起来, 使得这两种热电效应可以作为一个整体来研究[17]. 而不再直接测量Peltier系数, 转而通过测量较易获得的Seebeck系数, 然后通过Kelvin第二关系式间接获得Peltier系数[13]. 但是, Kelvin第二关系式是基于线性条件(即: Ohm定律、Fourier定律等)得到的[18-20], 这意味着基于Kelvin关系式的热电理论本质上是一个线性理论. 而在热电单元器件中存在着大量的各种非线性条件, 比如存在于肖特基结和pn结中的非线性电流-电压关系、非线性热输运关系等, 此时Kelvin关系式是否依然成立, 非线性热电效应能否给热电理论和材料研究带来进一步的突破, 值得进行深入思考和研究. 从简单到复杂, 从线性到非线性是科学研究的发展规律. 如图2所示, 从线性Ohm定律到半导体异质结中的非线性电输运关系, 从顺磁到铁磁, 从介电到铁电, 从线性光学到非线性光学, 这些自然界中的非线性现象为晶体管、存储器、光通信等众多现代科技的发展奠定了基础. 由此可见, 非线性现象不仅是一种基本的自然现象, 而且对于推进人类科技文明的发展也至关重要. 而Peltier效应作为一种界面处的电热效应, 如果仅仅将其视为一种线性可逆的效应显然是不合理的, 还应充分考虑其非线性属性. 但是由于Peltier系数的测量非常困难, 导致对Peltier效应的研究很少, 所以还不清楚在不同材料体系中Peltier效应偏离线性的程度. 并且目前关于Peltier效应的研究基本都局限在线性条件(Ohm定律、Fourier定律等)内, 选择的材料都具有较大的载流子浓度, 这意味着材料具有强的线性输运性质[21], 因此导致目前Peltier效应的非线性表现不明显, 而在非线性条件以及强非线性材料的基础上是否可能出现较为显著的非线性Peltier效应? 这是需要回答的科学问题, 这也将填补甚至突破当前的基础热电科学, 而基础科学的突破往往能够带来意想不到的收获. 随着热电单元和器件的小型化, 必然伴随着微纳尺度热输运与电输运的研究, 此时量子效应将起主导作用[22-24], 若Peltier热不再与电流成线性关系, 此时Peltier系数将会受电流密度和温度梯度的影响, 因此在Peltier系数的推导中必须考虑输运过程中高阶项的影响, 而这方面的研究目前仍处于起步阶段. 图 2 线性与非线性现象的示意图 (a)线性的Ohm定律; (b)非线性的肖克莱方程式; (c)顺磁材料中磁感应强度与磁场强度的线性关系; (d)铁磁材料中磁感应强度与磁场强度的非线性关系; (e)顺电材料中电位移与电场强度的线性关系; (f)铁电材料中电位移与电场强度的非线性关系; (g)线性光学中极化强度与光场场强的线性关系; (h)非线性光学中极化强度与光场场强的非线性关系; (i)线性的Peltier效应; (j)非线性Peltier效应 Figure2. Schematic diagram of linear and nonlinear phenomena: (a) Linear Ohm's law; (b) nonlinear Shockley equation; (c) linear relationship between magnetic induction intensity and magnetic field intensity in paramagnetic materials; (d) nonlinear relationship between magnetic induction intensity and magnetic field intensity in ferromagnetic materials; (e) linear relationship between electric displacement and electric field intensity in paraelectric materials; (f) nonlinear relationship between electric displacement and electric field intensity in ferroelectric materials; (g) linear relationship between polarization intensity and optical field strength in linear optics; (h) nonlinear relationship between polarization intensity and optical field strength in nonlinear optics; (i) linear Peltier effect; (j) nonlinear Peltier effect.
图 3 由两种材料 (a, b)以及电压表组成的电路[18], $ {T}_{1} $表示高温, $ {T}_{0} $表示低温, ${T}'$表示环境温度, $ {\widetilde {\mu }}_{x} $(x为0, 1, l, r)表示不同位置的电化学势 Figure3. A circuit composed of two materials (a, b) and a voltmeter[18], $ {T}_{1} $ represents high temperature, $ {T}_{0} $ represents low temperature, ${T}'$ represents ambient temperature, and $ {\widetilde {\mu }}_{x} $(x is 0, 1, l, r) represents electrochemical potential at different positions.
32.2.2.Peltier系数的推导 -->
2.2.2.Peltier系数的推导
考虑由a和b两个导体构成的等温接头, 并且有电流($ -{\rm{e}}{{J}} $)通过, e为电子电荷, 如图4所示. 经过这种接头, 能量流将会变得不连续, 则在接头处能量的差异称为“Peltier热”. 这里有能量流密度$ {{w}}={{q}}+\widetilde {\mu }{{J}} $, 并且Callen认为$ \widetilde {\mu } $和J经过接头仍然是连续的, 所以有: 图 4 两种材料 (a, b)组成的界面处Peltier效应示意图[18], $ {{{w}}}_{{\rm{a}}\left({\rm{b}}\right)} $表示材料a(b)中的能量流密度, $ {{{q}}}_{\rm{a}}-{{{q}}}_{\rm{b}} $表示界面处吸收(放出)的Peltier热量 Figure4. Schematic diagram of the Peltier effect at the interface composed of two materials (a, b)[18], $ {{{w}}}_{{\rm{a}}\left({\rm{b}}\right)} $ represents the energy flow density in material a(b), and $ {{{q}}}_{\rm{a}}-{{{q}}}_{\rm{b}} $ represents the Peltier heat absorbed (released) at the interface.
随着半导体物理学的发展, 1957年Ioffe[19]和1962年Heikes等[20]利用能带理论对Seebeck系数和Peltier系数再次进行了理论推导. Ioffe利用能带理论推导了非简并且单极扩散的n型半导体中的Seebeck系数, 他将发生在半导体内部和边界上的热电现象分别考虑[19]. 如图5(a)所示, 半导体两端与金属相接触, 一端为高温端Th, 一端为低温端Tc, 其中μ表示化学势, 有$ \mu ={E}_{\rm{F}}-{E}_{\rm{c}} $, $ {E}_{\rm{F}} $为费米能级, $ {E}_{\rm{c}} $为导带底. 在半导体内部由于载流子浓度和扩散系数随温度增大, 引起电子由高温端向低温端的扩散, 电子在低温端积累, 半导体内部形成内建电场, 在电场作用下电子由低温端向高温端漂移, 最终电子的漂移和扩散将达到平衡, 并且电场的存在使能带发生倾斜. 在界面处由于接触电势也会随温度变化, 冷热两端接触电势大小不同, 因此也会对Seebeck效应产生影响, 本质上这也是化学势随温度变化的直接表现. 下面分别考虑扩散系数、载流子浓度和接触电势对Seebeck系数的影响. 图 5 Seebeck效应和Peltier效应的能带原理图 (a)由金属-n型半导体-金属结构组成的器件在温度梯度下的能带结构; (b)由金属-n型半导体结构组成的器件在无外加电场下的能带结构 Figure5. Energy band principle diagrams of Seebeck effect and Peltier effect: (a) Energy band structure of a device composed of a metal-semiconductor (n-type)-metal structure under a temperature gradient; (b) energy band structure of a device composed of a metal-semiconductor (n-type) structure without an external electric field.
表2不同方法对Seebeck系数、Peltier系数以及Kelvin第二关系式推导时的假设条件及存在的问题 Table2.The assumptions and problems in the derivation of Seebeck coefficient, Peltier coefficient and Kelvin's second relationship by different methods.
其中κ为热导率, A为热电臂的横截面积. 作者采用由N对适当掺杂的n型和p型碲化铋半导体构成的热电组件, 其中上下两个金属块可以与外界交换热量, 其余四个面全部由绝热材料包覆, 如图6所示. 当有电流通过时, 单位时间内热端金属块释放的热量为$ {{\rm{d}}Q}_{\rm{h}}/{\rm{d}}t $, 冷端金属块吸收的热量为$ {{\rm{d}}Q}_{\rm{c}}/{\rm{d}}t $. 图 6 由n型和p型碲化铋半导体构成的热电器件[14] (a)由两对n/p热电对组成的热电器件; (b)一对n/p热电对的热端结构; (c)热电器件的侧面结构, 器件四周全部被绝热材料包覆, 只有两侧可以与外界换热 Figure6. Thermoelectric devices composed of n-type and p-type bismuth telluride semiconductors[14]: (a) A thermoelectric device composed of two pairs of n/p thermoelectric pairs; (b) hot end structure of a pair of n/p thermoelectric pairs; (c) side structure of the thermoelectric device, all around the device are covered with insulating materials, and only two sides can exchange heat with the outside world.
然后通过锁相热成像观察到小电流下温度分布的微小变化, 得到正向偏置的热功率密度$ {P}^{+} $和反向偏置的热功率密度$ {P}^{-} $, 再根据以上公式就可以得到薄膜样品的Peltier系数. 根据Breitenstein等[29]提出的方法, Jin等[30]利用热悬浮装置和锁相热成像技术, 对有机薄膜中的Peltier效应进行了研究. 如图7(a)所示, 他们在超薄悬浮的聚对二甲苯薄膜上制备了一种基于Ni-ett的器件, 由于聚对二甲苯作为绝热衬底其导热系数很低, 因此样品与基底之间的热交换比较弱. 并且因为他们将器件放置于真空环境中, 所以器件与空气的热对流也几乎可以忽略, 而器件表面对外界的热辐射则忽略不计. 之后他们通过锁相热成像技术得到器件的温度分布, 并利用PT100对温度进行校准, 然后根据Breitenstein等[29]提出的方法对Peltier热和Joule热进行分离, 可以得到Peltier效应和Joule效应对温度分布的贡献, 利用式(46)可以得到样品的Peltier系数, 他们通过与Seebeck系数相比发现其满足Kelvin第二关系式, 如表3所示. 但是测试过程中薄膜样品与基板的热交换不可避免, 如图7(d)所示, 这导致其实际测定的Peltier热和Joule热比其理论模拟的值低, 而这对Peltier系数的测量必然会产生一定的影响. 图 7 利用热悬浮装置和锁相热成像技术对有机薄膜中Peltier效应的测量[30] (a)在横向结构的薄膜热电器件上同时发生的热效应示意图; (b)由Joule热和Peltier热造成的温度分布示意图; (c)分离Joule热和Peltier热的机制示意图; (d)在电流密度为1.5 A/mm2, 施加时间为0.01 s时Peltier热和Joule热导致的温度分布 Figure7. Measurement of the Peltier effect in organic thin films using thermal levitation devices and lock-in thermal imaging technology[30]: (a) Schematic diagram of the thermal effects simultaneously occurring on the thin-film thermoelectric device with lateral structure; (b) temperature distribution caused by Joule heat and Peltier heat; (c) mechanism of separating Joule heat and Peltier heat; (d) temperature distribution caused by Peltier heat and Joule heat when current density is 1.5 A/mm2 and application time is 0.01 s.
当有磁场存在时, 由于电子的自旋在磁场下不具有时间反演对称性, Onsager倒易关系不再适用, 那么在磁性材料中, Onsager倒易关系是否适用呢? 此时Kelvin第二关系式是否依然符合实际情况呢? Avery等[15]设计了一种悬浮的磁性薄膜装置, 通过将磁性薄膜生长在Si-N膜上, Si-N膜通过支撑腿悬浮在Si框架之上, 其中$ {\kappa }_{\rm{L}} $是连接Si框架的支撑腿的热导率, $ {\kappa }_{\rm{B}} $是在磁性薄膜和Si-N膜的热导率, 如图8所示, 探究了磁性材料中热电效应的Kelvin第二关系式. 在忽略Seebeck效应、Thomson效应和与外界热交换的情况下, 对系统热量的变化进行了分析. 由于Joule热和Peltier热的存在, 样品内部以及样品与环境之间将会有温度梯度存在. 作者测量了样品与环境的温差, 并由此计算出了Peltier效应造成的最大温差$ \Delta {T}_{\varPi } $为200 mK左右, 而作者估算的测量误差为70 mK左右, 可以看到$ \Delta {T}_{\varPi } $和测量误差相差不大. 如表3所示, 尽管最后作者通过实验得到的Peltier系数和通过$ \varPi =T\alpha $计算得到Peltier系数符合得很好, 但是这也并不能充分说明该实验结果真正符合Kelvin第二关系式. 图 8 悬浮磁性薄膜的Peltier效应测试装置[15] (a)测试装置的侧面示意图; (b)测试装置的SEM图, 其中粉色表示样品, 红色和黄色表示加热器, 蓝色和绿色表示热电偶; (c)测试装置的局部放大SEM图 Figure8. Peltier effect test device of suspended magnetic film[15]: (a) Side view of the test device; (b) SEM image of the test device, in which pink represents the sample, red and yellow represent heaters, and blue and green represent thermocouples; (c) a partial enlarged SEM image of the test device.