跳转至

9.24

\(u, v\) 为独立随机变量

\[ \begin{aligned} \braket{uv} &= \int \int uv P_u(u) P_v(v) \, \mathrm{d}u \, \mathrm{d}v \\ &= \int u P_u(u) \, \mathrm{d}u \int v P_v(v) \, \mathrm{d}v \\ &= \braket{u} \braket{v} \end{aligned} \]

随机行走

热平衡

  • 两个系统温度一样,净热流为零

热力学第零定律

A 与 B 热平衡,B 与 C 热平衡,则 A 与 C 热平衡。

激光管中正在发射激光的气体:\(< 0 \,\text{K}\)

Microstates and macrostates

  • 微观态:对系统基本组成的描述
  • 宏观态:对系统整体性质的描述

最可能的宏观态:微观态数目最多

如果一个系统的能量为 \(E\),则它的微观态数目为 \(\Omega(E)\). 即微观态数目是能量的函数

温度的统计定义

设有两个与环境孤立的热系统,它们相互有热接触,能量分别为 \(E_1, E_2\). 第一个系统的微观状态数是 \(\Omega_1(E_1)\),第二个系统的微观状态数是 \(\Omega_2(E_2)\)。整个系统的微观状态数就是 \(\Omega = \Omega_1(E_1) \, \Omega_2(E_2)\)

Crucial insight

A system will appear to choose a macroscopic configuration that maximizes the number of microstates.

based on the following assumptions:

  • Each one of the possible microstates of a system is equally likely to occur.(等概率假设)
  • The system's internal dynamics are such that the microstates of the system are continually changing.(事实)
  • Given enough time, the system will explore all possible microstates and spend an equal time in each of them.(各态历经假设)

👉 系统等概率地、连续地、等时地遍历每一种微观态。

对其中一个系统做出无限小的变化,对 \(E_1\) 求极值

\[ \frac{\partial}{\partial E_1} (\Omega_1(E_1) \, \Omega_2(E_2)) = 0 \]

\[ \Omega_2 (E_2) \frac{\partial \Omega_1(E_1)}{\partial E_1} + \Omega_1 (E_1) \frac{\partial \Omega_2(E_2)}{\partial E_2} \cdot \frac{\partial E_2}{\partial E_1} = 0 \]

总能量不变,有 \(\mathrm{d} E = \mathrm{d} E_1 + \mathrm{d} E_2 = 0\),所以 \(\frac{\partial E_2}{\partial E_1} = -1\). 代入上式得

\[ \frac{\partial \ln \Omega_1(E_1)}{\partial E_1} = \frac{\partial \ln \Omega_2(E_2)}{\partial E_2} \]

满足这个条件时,微观状态数达到最大。

定义温度 \(T\)

\[ \begin{equation} \label{eq:temp_stat_def} \frac{1}{k_B T} := \frac{\partial \ln \Omega(E)}{\partial E} \end{equation} \]

其中 \(k_B\) 是玻尔兹曼常数,\(k_B = 1.3807 \times 10^{-23} \, \text{J/K}\).

该定义是否与热力学自洽?

系综

imagine repeating an experiment to measure a property of a system

Josiah Willard Gibbs, 1878:ensemble

系统在不同时刻的代表点,被想象成许多有相同宏观性质的系统在同一时刻,但处于不同微观状态的代表点。Gibbs 把这些想象的系统的集合称为统计系综,简称系综.

相空间

对于一个有 \(N\) 个粒子的系统,有 \(3N\) 个广义坐标 \(\vec{r}^N\)\(3N\) 个广义动量 \(\vec{p}^N\),则系统的状态可以用 \(6N\) 维相空间中的一个点来表示。

经典的做法:求解出相空间中的轨迹

Gibbs:考虑相空间中的一个区域,包含了系统可能处于的所有状态

  • Microcanonical ensemble(微正则系综):an ensemble of systems that each have the same fixed energy.
  • Canonical ensemble(正则系综):an ensemble of systems, each of which can exchange its energy with a large reservoir of heat(热库). As we shall see, this fixes the temperature of the systems in the ensemble.
  • Grand canonical ensemble(巨正则系综):an ensemble of systems, each of which can exchange both energy and particles with a large reservoir.

正则系综

一个大的系统(热库 reservoir/热浴 heat bath)\(E - \epsilon\) 与一个小系统(之后直接称作系统 system)\(\epsilon\) 接触,整体视作孤立系统。假设系统的微观状态数为 \(1\).

系统能量处于 \(\epsilon\) 的概率

\[ P(\epsilon) \propto \Omega(E - \epsilon) \times 1 \]

\(\ln \Omega(E - \epsilon)\)\(\epsilon = 0\) 附近做泰勒展开

\[ \ln \Omega(E - \epsilon) = \ln \Omega(E) - \epsilon \frac{\partial \ln \Omega(E)}{\partial E} + \cdots \]

代入温度的定义式(\ref{eq:temp_stat_def})得

\[ \ln \Omega(E - \epsilon) = \ln \Omega(E) - \frac{\epsilon}{k_B T} + \cdots \]