Statistical Mechanics: Phase Transitions & Graduate Formalism

Statistical Mechanics

Critical phenomena, fluctuations, non-equilibrium dynamics, and advanced field-theoretic methods.

Phase Transitions

Classification

First Order

Discontinuous change in first derivative of free energy

Ice ↔ Water Boiling

Second Order

Continuous first derivative, discontinuous second derivative

Magnetization Superconductivity

Critical Phenomena

Near a critical point, observables follow power laws in the reduced temperature $t = (T-T_c)/T_c$:

Exponent	Observable	Scaling
$\alpha$	Specific heat $C$	$C \sim \lvert t\rvert^{-\alpha}$
$\beta$	Order parameter $m$	$m \sim \lvert t\rvert^{\beta}$ (for $t < 0$)
$\gamma$	Susceptibility $\chi$	$\chi \sim \lvert t\rvert^{-\gamma}$
$\nu$	Correlation length $\xi$	$\xi \sim \lvert t\rvert^{-\nu}$

Universality

Systems with the same dimensionality and symmetry share identical critical exponents (examples of universality classes: 2D Ising, 3D XY, percolation). The exponents are not independent but obey scaling relations:

Rushbrooke: $\alpha + 2\beta + \gamma = 2$
Widom: $\gamma = \beta(\delta - 1)$
Fisher: $\gamma = \nu(2 - \eta)$

Fluctuations

Gaussian Fluctuations

For energy fluctuations: $\langle (\Delta E)^2 \rangle = k_B T^2 C_V$

For particle number: $\langle (\Delta N)^2 \rangle = k_B T \left(\frac{\partial N}{\partial \mu}\right)_{T,V}$

Fluctuation-Dissipation Theorem

Connects response functions to equilibrium fluctuations: $\chi(\omega) = \frac{1}{k_B T} \int_0^{\infty} \langle A(t)A(0) \rangle e^{i\omega t} dt$

Einstein’s Relations

Diffusion: $D = \mu k_B T$ (mobility $\mu$)
Conductivity: $\sigma = ne^2\tau/m$ (Drude model)

Non-equilibrium Statistical Mechanics

Boltzmann Equation

Evolution of distribution function: $\frac{\partial f}{\partial t} + \mathbf{v} \cdot \nabla_r f + \frac{\mathbf{F}}{m} \cdot \nabla_v f = \left(\frac{\partial f}{\partial t}\right)_{coll}$

H-theorem

Boltzmann’s H-function decreases: $H = \int f \ln f d^3v$ $\frac{dH}{dt} \leq 0$

Linear Response Theory

For small perturbation $F(t)$: $\langle A(t) \rangle = \langle A \rangle_0 + \int_{-\infty}^t \chi(t-t') F(t') dt'$

Kubo formula for conductivity: $\sigma = \lim_{\omega \to 0} \frac{1}{\omega} \int_0^{\infty} dt e^{i\omega t} \langle J(t)J(0) \rangle$

Applications

Condensed Matter Physics

Electronic properties of solids
Superconductivity (BCS theory)
Quantum Hall effects
Topological phases

Soft Matter

Polymer physics
Liquid crystals
Colloids
Biological membranes

Cosmology

Early universe thermodynamics
Dark matter freeze-out
Cosmic microwave background

Quantum Information

Thermal states in quantum computing
Entanglement entropy
Quantum thermodynamics

Computational Methods

Monte Carlo

Metropolis algorithm
Cluster algorithms (Wolff, Swendsen-Wang)
Quantum Monte Carlo

Molecular Dynamics

Verlet algorithm
Nosé-Hoover thermostat
Parrinello-Rahman barostat

Density Functional Theory

Hohenberg-Kohn theorem: Ground state density determines all properties.

Kohn-Sham equations: $\left[-\frac{\hbar^2}{2m}\nabla^2 + v_{\text{eff}}\left[n\right](r)\right]\psi_i(r) = \epsilon_i\psi_i(r)$

Advanced Topics

Renormalization Group

Block spin transformations
Fixed points and universality
Epsilon expansion
Functional renormalization

Conformal Field Theory

At critical points, systems exhibit conformal symmetry.

Central charge characterizes universality class.

AdS/CFT Correspondence

Connects strongly coupled field theories to weakly coupled gravity.

Applications to quark-gluon plasma and condensed matter.

Graduate-Level Mathematical Formalism

The remaining sections collect the field-theoretic and many-body machinery of modern statistical mechanics in a terse, reference-style format. They are reference material rather than a self-contained exposition, and can be skipped on a first read.

Information Theory and Statistical Mechanics

Shannon entropy: $S = -k_B \sum_i p_i \ln p_i$

Maximum entropy principle: The equilibrium distribution maximizes entropy subject to constraints.

Canonical ensemble from MaxEnt: Maximize S subject to:

Normalization: $\sum_i p_i = 1$
Energy constraint: $\sum_i p_i E_i = \langle E \rangle$

Using Lagrange multipliers: $p_i = \frac{e^{-\beta E_i}}{Z}$

Jaynes’ principle: Statistical mechanics as inference theory

Relative entropy (Kullback-Leibler divergence): $D_{KL}(p||q) = \sum_i p_i \ln\left(\frac{p_i}{q_i}\right) \geq 0$

Advanced Ensemble Theory

Generalized Ensembles

Tsallis statistics: $S_q = \frac{k_B (1 - \sum_i p_i^q)}{q - 1}$

Pressure ensemble: (NPT) $\Delta(N,P,T) = \int_0^{\infty} dV \, e^{-\beta PV} Z(N,V,T)$

Isothermal-isobaric partition function: $\Delta = \frac{k_B T}{P} Z(N,\langle V \rangle,T) e^{\beta P \langle V \rangle}$

Jarzynski Equality and Fluctuation Theorems

Jarzynski equality: $\langle e^{-\beta W} \rangle = e^{-\beta \Delta F}$

Crooks fluctuation theorem: $\frac{P_F(W)}{P_R(-W)} = e^{\beta(W - \Delta F)}$

Work distribution: Gaussian near equilibrium $P(W) \approx (2\pi\sigma^2)^{-1/2} \exp\left[-\frac{(W - \langle W \rangle)^2}{2\sigma^2}\right]$

Path Integral Formulation

Quantum partition function: $Z = \text{Tr}(e^{-\beta H}) = \int \mathcal{D}[q] \exp(-S_E[q]/\hbar)$

Euclidean action: $S_E = \int_0^{\beta\hbar} d\tau \left[\frac{m\dot{x}^2}{2} + V(q)\right]$

Feynman-Kac formula: Connection to diffusion $\langle q_f|e^{-\beta H}|q_i\rangle = \int_{q(0)=q_i}^{q(\beta\hbar)=q_f} \mathcal{D}[q] \, e^{-S_E[q]/\hbar}$

Effective action at finite temperature: $\Gamma[q_c] = -k_B T \ln Z[J] + \int d\tau \, J(\tau)q_c(\tau)$

Field Theoretic Methods

Hubbard-Stratonovich Transformation

For interaction term: $\exp\left[\frac{\beta}{2} \sum_{ij} J_{ij}s_i s_j\right] = \int \mathcal{D}[\phi] \exp\left[-\frac{\beta}{2} \sum_{ij} \phi_i(J^{-1})_{ij}\phi_j + \beta\sum_i \phi_i s_i\right]$

Replica Method

For disordered systems: $\langle \ln Z \rangle = \lim_{n\to 0} \frac{\langle Z^n \rangle - 1}{n}$

Replica symmetry breaking: Order parameter $q_{ab}$

Functional Integral Representation

Grand canonical ensemble: $\Xi = \int \mathcal{D}[\psi^*, \psi] \exp(-S[\psi^*, \psi])$

Action for bosons: $S = \int_0^{\beta} d\tau \int d^dr \left[\psi^*(\partial_\tau - \mu)\psi + \frac{\hbar^2}{2m}|\nabla\psi|^2 + U(\psi^*\psi)\right]$

Critical Phenomena: Advanced Treatment

Scaling Theory

Scaling hypothesis: Near $T_c$, singular part of free energy: $f_s(t, h) = b^{-d}f_s(b^{y_t}t, b^{y_h}h)$

Where $y_t = 1/\nu$, $y_h = d - \beta/\nu$

Scaling relations derivation:

From $f_s$: $\alpha = 2 - d\nu$
From $m = -\partial f/\partial h$: $\beta = (d - y_h)\nu$
From $\chi = \partial^2f/\partial h^2$: $\gamma = (2y_h - d)\nu$

Data collapse: Plot $m/

^\beta$ vs $h/

^{\beta\delta}$

Renormalization Group: Field Theory

$\phi^4$ theory action: $S = \int d^dx \left[\frac{1}{2}(\nabla\phi)^2 + \frac{r}{2}\phi^2 + \frac{u}{4!}\phi^4\right]$

RG flow equations (one-loop): $\frac{dr}{dl} = (2 - \eta)r + Au \frac{r^2}{1 + r}$ $\frac{du}{dl} = \varepsilon u - Bu^2 + \frac{Cu^3}{(1 + r)^2}$

Fixed points:

Gaussian: $(r^, u^) = (0, 0)$
Wilson-Fisher: $(r^, u^) = (-\varepsilon/A, \varepsilon/B)$

Critical exponents ($\varepsilon$-expansion): $\nu = \frac{1}{2} + \frac{\varepsilon}{12} + O(\varepsilon^2)$ $\eta = \frac{\varepsilon^2}{54} + O(\varepsilon^3)$

Conformal Field Theory at Criticality

Conformal algebra in 2D: Virasoro algebra $[L_m, L_n] = (m - n)L_{m+n} + \frac{c}{12} m(m^2 - 1)\delta_{m+n,0}$

Central charge: Characterizes universality class

Ising: $c = 1/2$
XY model: $c = 1$
Potts model (q states): $c = 1 - 6/[q(q+1)]$

Operator product expansion: $\phi_i(z)\phi_j(0) = \sum_k C_{ijk}z^{h_k-h_i-h_j}\phi_k(0)$

Exact Solutions

2D Ising Model (Onsager Solution)

Transfer matrix method: $Z = \text{Tr}(T^N)$

Critical temperature: $\sinh\left(\frac{2J}{k_B T_c}\right) = 1$

Free energy per site: $f = -k_B T \ln(2\cosh(2\beta J)) - \frac{k_B T}{2\pi} \int_0^\pi d\theta \, \ln\left[1 + \sqrt{1 - \kappa^2\sin^2\theta}\right]$

Where $\kappa = 2\sinh(2\beta J)/\cosh^2(2\beta J)$

Magnetization ($T < T_c$): $m = \left[1 - \sinh^{-4}(2\beta J)\right]^{1/8}$

Bethe Ansatz

1D Heisenberg chain: $H = J\sum_i \boldsymbol{\sigma}_i \cdot \boldsymbol{\sigma}_{i+1}$

Bethe equations: $k_j L = 2\pi I_j - \sum_{k\neq j} \theta(k_j - k_k)$

Ground state energy: $\frac{E_0}{N} = -J \ln 2 + \frac{J}{4}$

Non-equilibrium Field Theory

Keldysh Formalism

Contour ordering: Forward and backward branches

Green’s functions: $G^{++}(t,t') = -i\langle T\phi(t)\phi(t')\rangle$ $G^{--}(t,t') = -i\langle \tilde{T}\phi(t)\phi(t')\rangle$ $G^{+-}(t,t') = -i\langle \phi(t')\phi(t)\rangle$ $G^{-+}(t,t') = -i\langle \phi(t)\phi(t')\rangle$

Keldysh rotation: $G^R = G^{++} - G^{+-}$ $G^A = G^{++} - G^{-+}$ $G^K = G^{++} + G^{--} - G^{+-} - G^{-+}$

Langevin Dynamics

Stochastic equation: $\partial_t\phi = -\Gamma\frac{\delta F}{\delta\phi} + \eta$

Noise correlations: $\langle \eta(x,t)\eta(x',t') \rangle = 2\Gamma k_B T\delta(x-x')\delta(t-t')$

Martin-Siggia-Rose formalism: Path integral with response field $Z = \int \mathcal{D}[\phi, \tilde{\phi}] \exp(iS[\phi, \tilde{\phi}])$

Quantum Many-Body Systems

Fermi Liquid Theory

Quasiparticle concept: Landau parameters $f^s$, $f^a$

Effective mass: $\frac{m^*}{m} = 1 + \frac{F_1^s}{3}$

Compressibility: $\frac{\kappa}{\kappa_0} = (1 + F_0^s)^{-1}$

Collective modes: Zero sound velocity $s = v_F\sqrt{1 + \frac{F_0^s}{3}}$

BCS Theory of Superconductivity

BCS Hamiltonian: $H = \sum_k \varepsilon_k c^\dagger_{k\sigma}c_{k\sigma} - g\sum_{kk'} c^\dagger_{k\uparrow}c^\dagger_{-k\downarrow}c_{-k'\downarrow}c_{k'\uparrow}$

Gap equation: $\Delta = g\sum_k \frac{\Delta}{2E_k} \tanh(\beta E_k/2)$

Where $E_k = \sqrt{\varepsilon_k^2 +

\Delta

^2}$

Critical temperature: $k_B T_c = 1.14\hbar\omega_D \exp(-1/N(0)g)$

Luttinger Liquids (1D)

Bosonization: Fermion operators to Boson fields $\psi(x) \sim \exp[i\phi(x)]$

Luttinger parameter: $K < 1$ repulsive, $K > 1$ attractive

Power-law correlations: $\langle \psi^\dagger(x)\psi(0) \rangle \sim x^{-1/(2K)}$

Modern Developments

Tensor Network Methods

Matrix Product States (MPS): $|\psi\rangle = \sum_{s_1...s_N} \text{Tr}(A^{s_1}...A^{s_N})|s_1...s_N\rangle$

DMRG algorithm: Variational optimization of MPS

Area law entanglement: $S \sim L^{d-1}$ for ground states

Machine Learning in Statistical Mechanics

Neural network representation of states: $\psi(s) = \exp\left[\sum_i a_i s_i + \sum_{ij} W_{ij}h_i(s)s_j + ...\right]$

Variational Monte Carlo with NNs: $E = \frac{\langle\psi|H|\psi\rangle}{\langle\psi|\psi\rangle}$

Unsupervised learning of phases:

Principal component analysis
Autoencoders
Diffusion maps

Quantum Thermalization

Eigenstate Thermalization Hypothesis (ETH): $\langle E_n|O|E_m\rangle = O(E)\delta_{nm} + e^{-S(E)/2}f_O(E,\omega)R_{nm}$

Many-body localization: Failure of thermalization

Floquet systems: Time-periodic Hamiltonians

Stochastic Processes and Field Theory

Doi-Peliti Formalism

Creation/annihilation operators for classical particles: $a^\dagger|n\rangle = |n+1\rangle$ $a|n\rangle = n|n-1\rangle$

Master equation to “Schrodinger” equation: $\partial_t|\psi\rangle = H|\psi\rangle$

Coherent state path integral: $P(n,t) = \int \mathcal{D}[\phi^*,\phi] \exp(-S[\phi^*,\phi])$

Active Matter

Toner-Tu equations: Flocking $\partial_t\rho + \nabla\cdot(\rho\mathbf{v}) = 0$ $\partial_t\mathbf{v} + \lambda(\mathbf{v}\cdot\nabla)\mathbf{v} = \alpha\mathbf{v} - \beta|\mathbf{v}|^2\mathbf{v} - \nabla P + \nu\nabla^2\mathbf{v} + \mathbf{f}$

Motility-induced phase separation: $\partial_t\rho = \nabla\cdot[(D(\rho) + D_t)\nabla\rho]$

Advanced Computational Methods

Quantum Monte Carlo

Path integral Monte Carlo: $\rho(R,R';\beta) = (2\pi\lambda\beta)^{-3N/2}\sum_P (\pm)^P \exp\left[-\beta\sum_i V(R_i)\right]$

Sign problem: Fermionic systems, frustrated magnets

Continuous-time algorithms: Worm algorithm, CT-QMC

Machine Learning Acceleration

import torch
import torch.nn as nn

class VariationalWavefunction(nn.Module):
    def __init__(self, L, hidden_dim=100):
        super().__init__()
        self.L = L
        self.net = nn.Sequential(
            nn.Linear(L, hidden_dim),
            nn.ReLU(),
            nn.Linear(hidden_dim, hidden_dim),
            nn.ReLU(),
            nn.Linear(hidden_dim, 2)  # Real and imaginary parts
        )
    
    def forward(self, states):
        """states: (batch_size, L) binary spin configurations"""
        out = self.net(states.float())
        log_amp = out[:, 0]
        phase = out[:, 1]
        return log_amp, phase
    
    def sample(self, n_samples):
        """Metropolis sampling from |psi|^2"""
        states = torch.randint(0, 2, (n_samples, self.L))
        # Implement Metropolis-Hastings...
        return states

Research Frontiers

Quantum Information and Statistical Mechanics

Entanglement entropy scaling:

Volume law: S ∼ L^d (thermal, excited states)
Area law: S ∼ L^{d-1} (ground states)
Logarithmic: S ∼ log L (1D critical)

Tensor network representations:

MPS, PEPS, MERA
Entanglement renormalization

Non-equilibrium Quantum Systems

Prethermalization: Quasi-stationary states

Dynamical phase transitions: Non-analytic behavior in Loschmidt echo

Floquet engineering: Designer Hamiltonians

Machine Learning and Physics

Reverse engineering Hamiltonians: Learning from data

Accelerating simulations: Neural network quantum states

Discovering order parameters: Unsupervised learning

Topological Phases

Symmetry-protected topological phases:

Classification by cohomology
Edge states

Topological order:

Anyonic excitations
Topological entanglement entropy

Many-Body Localization

Phenomenology:

Area law entanglement
Emergent integrability
l-bits (localized integrals of motion)

Transitions:

Thermal to MBL
MBL to ergodic

References and Further Reading

Classic Textbooks

Pathria & Beale - Statistical Mechanics
Kardar - Statistical Physics of Particles & Fields
Landau & Lifshitz - Statistical Physics (Parts 1 & 2)
Huang - Statistical Mechanics

Advanced Monographs

Altland & Simons - Condensed Matter Field Theory
Sachdev - Quantum Phase Transitions
Nishimori & Ortiz - Elements of Phase Transitions and Critical Phenomena
Täuber - Critical Dynamics

Specialized Topics

Gogolin, Nersesyan & Tsvelik - Bosonization and Strongly Correlated Systems
Schollwöck - The density-matrix renormalization group in the age of matrix product states
Eisert, Cramer & Plenio - Colloquium: Area laws for the entanglement entropy
Carleo & Troyer - Solving the quantum many-body problem with artificial neural networks

Recent Reviews

Nandkishore & Huse - Many-body localization and thermalization (2015)
Calabrese, Cardy & Doyon - Special issue on quantum integrability in out of equilibrium systems (2016)
Abanin et al. - Colloquium: Many-body localization, thermalization, and entanglement (2019)
Carrasquilla - Machine learning for quantum matter (2020)