Statistical physics

Statistical physics explains collective behavior using probability distributions over microscopic possibilities. Its central surprise is that incomplete microscopic knowledge can still produce precise predictions for internal energy, entropy and heat capacity. The bridge is statistical mechanics: specify the available microstates, assign their probabilities, then compute the expected values of measured quantities.

A two-level system already displays the idea. Warming it makes its higher energy more likely; its heat capacity first rises, then falls as the two probabilities approach equality. With two distinguishable units, merely counting microstates explains why the middle energy has twice the state degeneracy. These small examples expose the machinery before the much larger sums needed for a photon gas or an Ising model.

The appeal is economical: the same microscopic energies and probabilities explain both an average reading and its variance. A measurement's spread therefore becomes useful physical information rather than an inconvenience to discard.

Approach of a physical system to thermal equilibrium, usually through interactions with other degrees of freedom. Equality of a reduced density matrix with a quantum Gibbs state at one instant does not alone demonstrate a thermalization mechanism or equilibration over time.

An equilibrium statistical ensemble with fixed particle number, pressure and temperature, allowing volume and energy to fluctuate. With a temperature-independent volume measure, its joint weight is proportional to $e^{-\beta (E+PV)}$. Its normalization and any temperature-dependent prefactor must be specified when taking thermodynamic derivatives.

With temperature-independent microstate enthalpies and measure, differentiating the normalized weight $e^{-\beta H}/\Delta$ gives ${\partial }_{\beta }⟨H⟩=-\mathrm{Var}(H)$. Since $d\beta /dT=-1/(k_B{T}^2)$, the displayed response follows. A stable finite-system equilibrium distribution therefore has nonnegative heat capacity at constant pressure in this setting.

Integrating a canonical partition function over volume gives the displayed continuous-volume normalization. For a temperature-independent measure and Hamiltonian, $-{\partial }_{\beta }\ln \Delta =⟨E+PV⟩$. Multiplying by a volume-independent prefactor leaves normalized volume probabilities unchanged, but a temperature-dependent prefactor changes this derivative.

For a classical monatomic ideal gas, direct volume integration gives $\Delta ={\Lambda }^{-3N}(\beta P{)}^{-N-1}$. Consequently the continuous-volume ensemble has $⟨H⟩=(5N/2+1)k_{B}T$. Defining a dimensionless thermodynamic function using $\tilde{\Delta }=\beta P\Delta$ instead gives $-{\partial }_{\beta }\ln \tilde{\Delta }=(5N/2)k_{B}T$. The normalized volume distribution is identical, because the prefactor cancels in probabilities. Its temperature derivative nevertheless contributes $-k_{B}T$ to the derived enthalpy and $-k_B$ to its heat capacity. The two conventions agree at leading extensive order, but cannot be mixed in an exact finite-$N$ fluctuation identity.

A lattice gas describes particle positions by occupations of fixed cells. In the single-occupancy model, $n_{\alpha }\in \{0,1\}$ and $N={\sum }_{\alpha }{n}_{\alpha }$. Its grand canonical partition function sums configurations with weight $\exp [-\beta (U-\tilde{\mu }N)]$. Discretizing a classical position integral into cells of volume $b$ gives $\tilde{\mu }=\mu +k_{B}T\ln (b/{\lambda }^3)$, with thermal de Broglie wavelength $\lambda$.

For energy $U=-4J{\sum }_{⟨\alpha {\alpha }^{\prime }⟩}{n}_{\alpha }{n}_{{\alpha }^{\prime }}$ on a regular lattice of coordination $q$, the substitution $n_{\alpha }=(1+{\sigma }_{\alpha })/2$ maps the model to the Ising model with coupling $J$ and effective field $h=qJ+\tilde{\mu }/2$. The mean Ising spin is a density imbalance, $⟨\sigma ⟩=2bn-1$. For $J>0$ below the critical temperature, the field-driven Ising transition becomes gas-liquid phase coexistence.

For $N$ indistinguishable adsorbates on $M$ equivalent sites, at most one per site and no lateral interactions, $Z_N=\left(\genfrac{}{}{0ex}{}{M}{N}\right)e^{-N{\epsilon }_s/(k_{B}T)}$. The binomial coefficient already accounts for indistinguishability; an additional division by $N!$ would count it twice.

A hard sphere has an impenetrable core and otherwise no pair interaction. Two identical spheres of diameter $d$ cannot have centre separation below $d$. Even without attractive potential energy, crowding can produce fluid–crystal coexistence because available configuration space and entropy differ between the phases.

The two-dimensional counterpart of a hard sphere: two equal discs of radius $r$ cannot have centre separation less than $2r$. Their pair-excluded area is $4\pi r^2$, but the second-virial excluded-area parameter per particle is half that value because a pair would otherwise be counted twice.

A circular free-area model for a disc surrounded by fixed neighbours a centre-to-centre distance $R$ away. The inscribed circle available to the moving centre has radius $R-2r$. It is an approximation to the actual curved cage boundary, not the area of a disc whose radius is $R-r$.

At low density, sequential insertion gives a configurational integral approximated by ${\prod }_{j=0}^{N-1}(a-4j\pi r^2)$. Expanding its logarithm to first order in occupied area gives $N\ln a-2\pi r^{2}N(N-1)/a$. Replacing $N-1$ by $N$ and matching a free-volume expression gives $(a-Nb{)}^N$, with $b=2\pi r^2$. This matching is controlled only at low density; its pole at packing fraction $1/2$ is not a physical close-packing limit.

Crystallization favored by greater total accessible configurational entropy, despite increased spatial order. Ordered cages at sufficiently high packing fraction can allow more positional/free-volume configurations than a disordered crowded fluid.

Linear response theory describes a system's small change in an observable to first order in an applied perturbation. The approximation requires a stated reference state and a range of sufficiently small perturbations.

The random phase approximation treats the independent-particle response to the self-consistent total field, including the induced mean Coulomb potential. With a positive static number-response coefficient ${\Pi }_0=-\delta n/\delta U$, a potential $U=-e\varphi$ gives $\delta n=e{\Pi }_0{\varphi }_{\mathrm{tot}}$. Combining induced charge $-e\delta n$ with Poisson equation gives $ϵ=1+e^2{\Pi }_0/({ϵ}_0{q}^2)$. Exchange-correlation vertex corrections beyond this self-consistent response are omitted. It is a response approximation, not a literal assumption that every microscopic electron phase is independently randomized.

For an integrable linear response kernel $K$ and input $V$, the convolution theorem gives $\hat{I}(\omega )=\hat{K}(\omega )\hat{V}(\omega )$. The integrated output is therefore zero for every integrable input if and only if $\hat{K}(0)=\int K(t)dt=0$. For a sum of causal damped sines, $K(t)=\Theta (t){\sum }_j{a}_j{e}^{-{\gamma }_{j}t}\sin (p_{j}t)$, the condition is ${\sum }_j{a}_j{p}_j/({\gamma }_j^2+p_j^2)=0$. If all $a_j$ are positive, nonzero terms of both signs are needed; a term with $p_j=0$ vanishes identically.

Causality constrains the real and imaginary parts of a frequency-dependent linear response. For a dielectric with ${ϵ}_r(\omega )\to 1$ at large frequency, the $e^{-i\omega t}$ convention givesThe integral is a Cauchy principal value. Absorption and optical dispersion therefore cannot be specified independently.

Thermal activation uses thermal fluctuations to cross an energy barrier. The crossing rate depends on the barrier, the temperature and the system's dynamics.

For gapped excitations with $C(T)\simeq A\exp [-\Delta /(k_{B}T)]$ and approximately constant prefactor, two heat capacity measurements at $T_2>T_1$ give the displayed activation-energy estimate. A more general prefactor $A{T}^p$ changes its numerator to $\ln [C(T_2)/C(T_1)]-p\ln (T_2/T_1)$. The two-point estimate measures an effective exponential scale; it does not independently determine the prefactor or establish a detailed excitation density of states.

An energy landscape assigns an energy to each possible state and describes its variation between neighbouring states. Its local minima and intervening energy barriers help distinguish favourable states from states that a specified dynamics can reach quickly. In a Hopfield network, the assigned energy is an objective function and need not be a conserved physical quantity.

At fixed environmental constraints, integrate out unobserved coordinates to obtain an equilibrium probability density for a chosen reaction coordinate $x$. Relative to a specified coordinate measure, the displayed free energy reproduces that density. A change of coordinate can add a Jacobian term, so the profile is not independent of its coordinate convention. Minima identify probable basins, while barriers relative to an initial basin contribute to activated kinetics; a kinetic prefactor still requires dynamics. A conjugate force tilts the profile by $-Fx$.

A tree showing how basins around local minima connect as an energy threshold is raised. Branching heights encode path barriers, while horizontal placement is mainly presentational.

An energy barrier is an increase in energy that must be overcome along an allowed route between states. Its value depends on which paths or single-step moves are allowed; specifying only the energies of the endpoints does not determine the barrier.

In an appropriate spinful one-dimensional interacting conductor, low-energy density and spin disturbances form distinct collective modes, generally with different velocities. A bare injected electron can excite both modes. Boundary conditions and attached leads still matter for electrical conductance; a bulk interaction parameter alone need not set the two-terminal value of a clean contacted wire.

Mean-field theory replaces fluctuating interactions with a self-consistent average field. For example, an Ising spin feels a field determined by the average neighboring spin magnetization; a self-consistency equation approximates the interacting Ising model. Neglecting correlated fluctuations can fail near a continuous phase transition. The mean-field approximation in Landau theory is a particular use of this general framework.

A self-consistent quadratic approximation to an interacting many-particle system that retains both normal one-particle densities and anomalous pair averages. Normal contractions supply Hartree and Fock fields, while anomalous contractions supply the pair potential. A Bogoliubov transformation preserving the appropriate fermion anticommutators or boson commutators diagonalizes the resulting Hamiltonian; the densities must then be recomputed to satisfy self-consistency and the particle-number constraint.

The Weiss molecular field replaces exchange coupling to neighbouring moments by an effective local field proportional to the average magnetization. For two-state moments and $B_E={\mu }_0\lambda M$, self-consistency gives $m=\tanh (m/t)$ at zero applied field, where $m=|M|/(n\mu )$ and $t=T/T_C$, with $k_B{T}_C={\mu }_0\lambda n{\mu }^2$. This field represents a quantum exchange energy, not generally the actual magnetostatic field generated by the moments.

For $m=\tanh (m/t)$ and positive temperature, the positive solution exists uniquely for $0<t<1$: the slope at zero is $1/t>1$, while the hyperbolic tangent is concave and below one. For $t\ge 1$ only $m=0$ remains. Expanding to cubic order near $t=1$ gives $m^2\simeq 3t^2(1-t)$ and $m\simeq t\sqrt{3(1-t)}$, with mean-field critical exponent $1/2$. The prefactor $t$ follows from the truncated equation; it is not controlled beyond leading critical order.

A thermal fluctuation is the randomness of an observable in an equilibrium Boltzmann distribution. If the energy contains $-hA$, differentiation of its partition function gives ${\partial }_h⟨A⟩=\beta (⟨A^2⟩-⟨A{⟩}^2)$. Thus the fluctuation variance is related to the response to a conjugate source. In a field theory, the two-point version is the connected correlation function.

For fixed particle number and temperature, a bubble with surface tension $\Gamma$ and area $\alpha V^{2/3}$ has effective availability $A(V)=F_{\mathrm{gas}}(T,V,N)+\Gamma \alpha V^{2/3}+p_{R}V$. Its stationary condition is $p_{\mathrm{gas}}-p_R=(2/3)\Gamma \alpha V^{-1/3}$. Its curvature is $A^{\prime \prime }=V^{-1}[{\kappa }_T^{-1}-(2/9)\Gamma \alpha V^{-1/3}]$. Positive curvature is required for stable Gaussian fluctuation approximation. The variance follows from availability curvature and equilibrium fluctuations. For an ideal gas, the restoring denominator becomes $p_R+(4/9)\Gamma \alpha V^{-1/3}>0$ for positive $p_R$. Treating the internal pressure as externally fixed instead of fixing the gas particle number changes the stability problem.

A photon gas is a collection of photons described statistically as radiation. Here the equilibrium model is homogeneous, isotropic electromagnetic radiation in vacuum at temperature $T>0$: each photon has energy $E=c|\mathbf{p}|=h\nu$, with momentum $\mathbf{p}$ and frequency $\nu$, and there are two independent transverse states of electromagnetic polarization. An absorbing cavity with walls at a common temperature provides a way to establish this thermal equilibrium. The bulk formulas assume cavity dimensions much larger than the typical wavelength $hc/(k_{B}T)$, so the electromagnetic mode density in a box can replace individual cavity modes by a continuum. ^[ref]

The crucial constraint is whether photon emission and photon absorption, or other number-changing processes, can equilibrate the photon population. When they can, photon number is not an independently conserved quantity and the equilibrium photon chemical potential is ${\mu }_{\gamma }=0$. For each polarization and frequency, the mean occupation number then follows the Bose-Einstein distribution:If number-changing processes are ineffective but number-conserving interactions redistribute energy, a constrained Bose-Einstein distribution can instead have $f=[\exp ((h\nu -{\mu }_{\gamma })/(k_{B}T))-1{]}^{-1}$ with chemical potential ${\mu }_{\gamma }\le 0$ fixed by the number constraint. Thus thermal equilibrium with respect to energy exchange alone does not establish ${\mu }_{\gamma }=0$. The zero value can also persist in an already prepared Planck photon distribution during suitable subsequent evolution; continual photon emission is not required to preserve it. ^[ref]

Combining $f$ with the two-polarization density of states, $g_{\nu }/V=8\pi {\nu }^2/c^3$, gives the Planck photon distribution and the spectral photon energy density:Here $B_{\nu }$ is the spectral radiance in Planck's law; $u_{\nu }$ is stored energy density per unit frequency, not outgoing power per area. Their different units are connected by the full-sphere solid angle and the speed of light $c$. The Boltzmann constant is $k_B$, and $h$ is the Planck constant. ^[ref]

The equilibrium equation of state is particularly simple:The radiation constant $a_R$ is the same constant called $a_{\mathrm{rad}}$ in black-body radiation; $\sigma$ is the Stefan-Boltzmann constant. An outline of the energy density calculation substitutes $x=h\nu /(k_{B}T)$ and uses ${\int }_0^{\infty }{x}^3/(e^x-1)dx={\pi }^4/15$. The radiation pressure follows because the angular average of the squared directional component of photon momentum is $1/3$ for isotropy; the thermodynamic derivation of blackbody radiation pressure supplies an independent route. These formulas describe the radiation contribution, not the cavity walls. The isotropic massless relation $p=u/3$ survives a different isotropic spectrum, whereas the coefficient in $u=a_R{T}^4$ requires the zero-chemical potential Planck photon distribution.

For a computed example with stipulated $T=2.725\mathrm{K}$, the photon energy density is $u\simeq 4.17174\times 10^{-14}\mathrm{J}{\mathrm{m}}^{-3}$ and the radiation pressure is $p\simeq 1.39058\times 10^{-14}\mathrm{Pa}$. The photon number density in the International System of Units isHere $\zeta$ is the Riemann zeta function. The mean photon energy differs from the energy at the peak of a spectral curve: counting photons and weighting them by energy emphasize different frequencies. The calculations use the exact International System of Units values of the Planck constant, Boltzmann constant and speed of light. ^[ref]

A motivating observation is the cosmic microwave background spectrum measured by the Far Infrared Absolute Spectrophotometer (FIRAS) aboard COBE. Its apparatus used a polarizing Michelson interferometer, with one input viewing a roughly seven-degree sky patch and the other an internal blackbody. Scanning mirrors changed the optical path length difference; heat-sensitive detectors recorded the varying power, and Fourier transform spectroscopy recovered the spectral difference. Operation outside the terrestrial atmosphere removed atmospheric emission and absorption, while shielding and cooling the optics in a helium cryostat suppressed the instrument's own radiation. ^[ref]

Absolute calibration came from moving an external blackbody into the sky aperture, filling the same beam used for the sky. This was an iron-loaded epoxy absorber shaped as a re-entrant cone. Repeated reflections made it a close approximation to a blackbody, while thermometers supplied its temperature. Mather and colleagues explicitly evaluated reflections, temperature gradients, radiation leaking through the material and leakage around its edge. These are limits on how faithfully the calibrator realizes thermal equilibrium radiation, rather than details that can be eliminated merely by calling it black. ^[ref]

During sky observations the internal blackbody temperature was adjusted to nearly null the spectral difference, reducing gain and drift errors. Fixsen and colleagues' 1996 analysis used calibrated spectra over wavenumbers $2$ to $21\mathrm{c}{\mathrm{m}}^{-1}$, about $60$ to $630\mathrm{GHz}$. They separated the sky's spatially uniform component, its dipole and Galactic emission, and varied sky cuts and foreground templates as controls. The resulting cosmic microwave background spectrum agreed with a single Planck function, with reported RMS deviations below 50 parts per million of its peak brightness. This is a residual statistic over the analyzed spectrum, not a pointwise relative error bound at every frequency. ^[ref]

The observation motivates using a photon gas as a physical model rather than merely a formal counting exercise. Its interpretation has a clear boundary: agreement with Planck's law constrains departures from an equilibrium spectral shape within the calibrated band and foreground assumptions; it does not directly measure radiation pressure or entropy density of a photon gas. Nor does it establish that today's cosmic microwave background is continually re-equilibrating with matter. A previously thermal spectrum can retain its Planck function shape as expansion of the universe lowers every photon frequency in the same proportion. The usefulness of the model is that one measured spectral temperature, together with its equilibrium assumptions, predicts energy density, radiation pressure, photon number density and entropy density of a photon gas consistently.

A cavity with two transverse photon polarizations and standing-wave spacing $\pi /L$ has energy density of states $2L/(\pi \hslash c)$. In the continuum regime, integrating the Bose occupation at zero chemical potential gives $U=bL{T}^2$, where $b=\pi k_B^2/(3\hslash c)$. The thermal quantities are $C_L=S=2bLT$, Helmholtz free energy $A=-bL{T}^2$ and generalized pressure $P=b{T}^2=U/L$. The last quantity has units of force. These thermal formulas exclude a separate boundary-dependent vacuum energy and require temperature above the mode-spacing scale.

At fixed temperature, a photon gas has constant energy density $u$ and radiation pressure $p=u/3$. An increase $\Delta V$ in volume requires both $\Delta U=u\Delta V$ and expansion work $p\Delta V$, giving absorbed heat $4u\Delta V/3$.

The entropy density of a photon gas is its entropy per physical volume, $s=S/V$. For homogeneous equilibrium vacuum radiation with two states of electromagnetic polarization and zero photon chemical potential,The first equality is the entropy density at zero chemical potential, obtained from the thermodynamic Euler relation $U=TS-pV+{\mu }_{\gamma }{N}_{\gamma }$ with internal energy $U=uV$. The second uses the photon gas equation of state, $u=a_R{T}^4$ and $p=u/3$. If an independently constrained photon number gives a nonzero chemical potential, the general entropy density is $(u+p-{\mu }_{\gamma }{n}_{\gamma })/T$ and $u=a_R{T}^4$ need not hold. This distinction explains why the absence of a conserved number matters for entropy as well as for the spectral shape. ^[ref]

A short first law of thermodynamics check, for neighboring equilibrium states with ${\mu }_{\gamma }=0$, isThe thermodynamic Euler relation fixes the additive normalization. Equivalently, this radiation entropy tends to zero with temperature at fixed volume; no contribution from the containing walls is included. Interpreting $TdS$ as transferred heat and $pdV$ as boundary work requires a reversible thermodynamic process. For a reversible thermodynamic process that is also an adiabatic process, there is no entropy production or entropy exchange, soAn irreversible adiabatic process can produce entropy even without heat exchange, so the word adiabatic alone is insufficient to infer $T^{3}V=\text{constant}$.

For an analytic example, expand an initially equilibrated photon gas reversibly from $V_0$ to $8V_0$, keeping only the same two photon states of electromagnetic polarization. Its temperature halves, its energy density becomes $u_0/16$, its entropy density of a photon gas becomes $s_0/8$, and its total entropy stays $S_0$. The internal energy halves rather than remaining fixed: the missing half leaves as work against radiation pressure,This is a stipulated reversible radiation model, not a report of a piston experiment. Maintaining a real cavity along this path requires controlling wall heat exchange and including the wall heat capacity in any actual apparatus balance.

For a computed scale at stipulated $T=2.725\mathrm{K}$, $s\simeq 2.04122\times 10^{-14}\mathrm{J}{\mathrm{K}}^{-1}{\mathrm{m}}^{-3}$. Combining this with the photon number density givesThis is the total equilibrium radiation entropy divided by its mean photon number, not a separately assigned entropy carried by each labeled particle. During the ideal expansion, $n_{\gamma }\propto T^3$ and $T^{3}V$ is fixed, so the mean number $N_{\gamma }=n_{\gamma }V$ also stays constant. This follows from the path; it was not imposed as the independent number constraint used to establish chemical potential.

The beautiful connection is that expansion lowers each typical photon energy and dilutes the radiation in just the proportions needed to retain its equilibrium spectral shape and total entropy. For the cosmic microwave background, homogeneous expansion of the universe has physical volume proportional to the cube of the scale factor; fixed-species cosmological entropy conservation therefore gives $T\propto a^{-1}$. This describes preservation of an already established Planck photon distribution and connects the FIRAS spectral observation in photon gas to an entropy prediction; FIRAS did not perform a direct entropy measurement. When other equilibrated species disappear and transfer energy into radiation, use the effective number of relativistic entropy degrees of freedom, $g_{\ast s}$: the conserved quantity for the combined sector is $g_{\ast s}{T}^{3}V$, under the corresponding equilibrium and no-entropy production assumptions. A changing species population, separate decoupled sectors or entropy production requires an appropriate entropy balance rather than the pure photon gas formula. ^[ref]

Statistical mechanics connects microstates to macrostates through probability distributions. Its microcanonical ensemble, canonical ensemble and grand canonical ensemble specify different constraints on exchanges with the surroundings. An ensemble average describes repeated preparations under those constraints; it need not be the value in every individual preparation.

This makes the subject useful whenever tracking every microscopic detail would obscure the question being asked. To predict a heat capacity, we need how the expected value of energy changes with temperature, rather than a record of every unit's state. The partition function packages that calculation into one reusable object.

In his 1902 book, J. Willard Gibbs systematically developed the canonical ensemble. His preface contrasts following one system through time with studying a collection of systems with different microscopic conditions. Chapter IV defines the canonical ensemble through an exponential function of energy and establishes its normalization. This supplied a general strategy for predicting collective measurements by averaging over possible microscopic conditions, with the partition function organizing the calculation. ^{[ref] [ref]}

Boundaries or inclusions change the available thermal or quantum fluctuations and thereby the free energy. Their force follows from $F=-\partial \mathcal{F}/\partial d$ under specified control conditions. Electromagnetic dispersion, depletion and critical fluctuations are distinct realizations, with signs and ranges set by their physical modes and boundary conditions.

Confining a near-critical fluctuating order parameter changes its free energy. Its range is governed by the growing correlation length, and its sign depends on the boundary preferences. This is a thermal collective effect, distinct from the zero-temperature ideal-conductor electromagnetic formula.

Two parallel ideal conducting plates at zero temperature have separation-dependent electromagnetic energy per area $-{\pi }^2\hslash c/(720d^3)$ and pressure $-{\pi }^2\hslash c/(240d^4)$. The negative sign denotes attraction in this geometry. Finite temperature, material response and boundary geometry change the ideal result; an absolute divergent vacuum energy is not the measured quantity.

For a massless real scalar in one spatial dimension, applying $e^{-\omega /(\pi \Lambda )}$ to both the Dirichlet mode sum and the unbounded-line integral gives the same leading bulk energy $\pi L{\Lambda }^2/2$. Their finite difference is $-\pi /(24L)$ in units $\hslash =c=1$, producing the attractive force $-\pi /(24L^2)$. This scalar boundary force is distinct from the retarded atom-surface Casimir-Polder interaction.

classical statistical mechanics assigns probability distributions to classical states. For a Hamiltonian particle system the canonical weight is $e^{-\beta H}$ on phase space. The weak-exchange, high-temperature limit of a quantum gas supplies the measure $d^{d}x{d}^{d}p/(2\pi \hslash {)}^d$ per particle and the indistinguishable-particle factor $1/N!$.

quantum statistical mechanics describes equilibrium and fluctuations with density matrices and traces over a quantum Hilbert space. The canonical equilibrium state is $e^{-\beta H}/Z$, while particle exchange introduces $H-\mu N$ in the grand canonical ensemble. Noncommutativity and particle exchange distinguish it from a classical phase-space distribution.

The ergodic hypothesis identifies long-time averages of equilibrium observables along typical trajectories with averages over the invariant equilibrium ensemble at the same conserved quantities. In the microcanonical formulation, the ensemble is normalized phase-space measure on the energy surface. In a measure-preserving description, ergodicity means invariant measurable sets have measure zero or one; the Birkhoff ergodic theorem then equates time averages and ensemble expectations almost everywhere for integrable observables. This explains why statistical-mechanical ensembles can predict long-time measurements. A particular mechanical model still requires its own ergodicity assumptions; integrable idealized systems need not be ergodic.

An ice bath can absorb heat while its temperature barely changes: the heat converts ice into water. The latent heat of fusion quantifies this melting budget. For a pure substance at its equilibrium melting temperature $T_m(p)$ and specified pressure $p$, define it here per unit mass:Here $h_l$ and $h_s$ are the specific enthalpies of the coexisting bulk liquid and solid, and $H$ is enthalpy. For ordinary melting, $L>0$, in $\mathrm{J}\mathrm{k}{\mathrm{g}}^{-1}$; an amount $m$ has a total fusion enthalpy $mL$ in joules. A molar fusion enthalpy instead has units $\mathrm{J}\mathrm{mo}{\mathrm{l}}^{-1}$ and equals $ML$, where $M=m/n$ is mass divided by amount of substance in $\mathrm{kg}\mathrm{mo}{\mathrm{l}}^{-1}$. Always check the normalization before inserting a tabulated value into an energy balance.^{[ref] [ref]}

For a closed system converting completely from solid to liquid at fixed pressure, with only pressure-volume work and negligible changes in bulk kinetic energy and potential energy, the first law of thermodynamics gives $Q_p=\Delta H=mL$. Both endpoints are at $T_m(p)$: warming initially colder ice or subsequently warming the resulting water requires additional sensible heat. Since $H=U+pV$, the corresponding specific internal energy jump is $\Delta u=L-p\Delta v$, where $v=V/m$; $L$ includes the work associated with the change in volume. Heat is transferred energy, whereas enthalpy is a state function. ^[ref]

Let $x=m_l/m$ be the liquid phase fraction by mass, rather than by volume. Neglecting interfacial contributions to enthalpy, equilibrium phase coexistence at fixed pressure givesIn the model above, reaching $T_m$ takes $2.10\mathrm{kJ}$ of sensible heat; the plateau then absorbs $33.40\mathrm{kJ}$, ending at a cumulative $35.50\mathrm{kJ}$. At $18.80\mathrm{kJ}$, half the mass is liquid and the temperature is still $T_m$. The final warming takes another $4.18\mathrm{kJ}$. The rounded $L$ and $c_l$ are inspired by the introductory ice-melting calorimetry values in Case Western's demonstration; constant properties and exact equilibrium throughout are stipulated model assumptions. ^[ref]

Within one thermodynamic phase, specific heat capacity at constant pressure is $c_p=(\partial h/\partial T{)}_p$, in $\mathrm{J}\mathrm{k}{\mathrm{g}}^{-1}{\mathrm{K}}^{-1}$; a sample's heat capacity at constant pressure is $C_p=m{c}_p$, in $\mathrm{J}{\mathrm{K}}^{-1}$. The plateau is a finite enthalpy change at zero temperature change through varying $x$, so neither a finite single-phase $c_p$ nor $Q=m{c}_p\Delta T$ represents it. The heating curve needs separate sensible heat intervals and a fusion contribution. ^[ref]

The plateau assumes a pure substance, a stable bulk solid-liquid phase coexistence pair, fixed pressure, and sufficiently slow, uniform heating to remain near thermodynamic equilibrium. The total mass and composition must remain fixed, with negligible evaporation and no other phase transition. A fast heater can produce temperature gradients, so a local thermometer need not read the equilibrium $T_m$. Neither $T_m$ nor $L$ is universal across pressures or compositions: common-pressure melting-point slope describes the pressure dependence of $T_m$, while mixtures can melt over a temperature range as their coexisting compositions change. Their liquidus must be consulted instead of assuming a pure-substance plateau. ^[ref]

This makes ice useful as a temperature buffer: under these conditions it can absorb $m_{s}L$ before the remaining solid mass $m_s$ disappears, after which the water warms. If a stipulated constant net input power $P$ goes entirely into melting, that buffer lasts $m_{s}L/P$. Reversing the equilibrium conversion releases the same fusion enthalpy during freezing. The beautiful thermodynamic connection is $L=T_m(s_l-s_s)$, with $s$ the specific entropy and $T_m$ absolute temperature: equal Gibbs free energies at phase coexistence allow different enthalpies and entropies. The nearly stationary thermometer hides a finite, measurable change of state. ^[ref]

Ice-melting calorimetry measures the fusion enthalpy through conservation of energy, rather than inferring it from the small temperature change of the melting ice itself. In a mixing experiment, warm water and the container cool while added ice melts and the resulting water warms. Their combined energy balance separates fusion from sensible heat.

Weigh the container and stirrer, then weigh the added warm water to obtain $m_w$. Let the container and water reach a common initial temperature $T_i$. Prepare ice in a clean ice-water mixture near $T_m$, rather than assuming freezer ice is already at $T_m$; drain and blot its surface immediately before transfer. Add it without splashing, stir gently, and record temperature before addition and until after the last ice disappears. Reweigh to find the added mass $m_i$. Wisconsin's laboratory guide specifies drying each piece, weighing before and after, and recording the temperature history; it places the initial and final temperatures on opposite sides of the jacket temperature to reduce net environmental exchange. ^[ref]

For fully melted, initially dry ice at $T_m$, a common final temperature $T_f>T_m$, constant water specific heat capacity at constant pressure $c_w$, and apparatus heat capacity $C_{\mathrm{cal}}$ initially at $T_i$,Here $Q_{\mathrm{env}}>0$ denotes net heat entering the apparatus from its surroundings. The left side supplies the fusion enthalpy and the sensible heat on the right; the water produced by melting must be warmed too. The balance extends Case Western's mixing demonstration by including the apparatus and environmental terms. If the specific heat capacities vary appreciably, replace their products with the corresponding temperature integrals. ^[ref]

Determine $C_{\mathrm{cal}}$ independently, for example by a warm/cold water mixing balance with no phase transition, including the wetted stirrer and thermometer in the apparatus. Check the thermometer calibration, keep the lid in place between transfers, and monitor the surroundings. Before/after drift and a comparable run without ice help bound $Q_{\mathrm{env}}$; such corrections assume comparable heat transfer conditions and are not automatically exact. Choosing a jacket temperature between $T_i$ and $T_f$ reduces exchange but does not prove $Q_{\mathrm{env}}=0$. Ensure all the ice has melted and the contents are well mixed before using the measured $T_f$: with residual ice, the weighed addition is not the mass that has undergone fusion. Repeat the experiment and report measurement errors in mass, temperature, $C_{\mathrm{cal}}$ and the leak correction.

An independent way to see the plateau is David Roundy's Oregon State experiment: heat an ice-water mixture slowly with a waterproof resistor in an insulated cup, stir, and record temperature, heater electric current $I(t)$, electric potential difference $V(t)$ and elapsed time. Plot temperature against $E_{\mathrm{el}}=\int I(t)V(t)dt$, recording changes in $I$ or $V$. The initial solid mass must exclude accompanying water. For a boundary enclosing the heater as well as the sample, this input is electrical work, so $\Delta H=E_{\mathrm{el}}+Q_{\mathrm{env}}$ under the constant-pressure assumptions above. Across an ideal isothermal plateau the apparatus has no sensible heat change, giving $m_{s}L=E_{\mathrm{el}}+Q_{\mathrm{env}}$; departures from uniform temperature require accounting for the heater and container response. ^[ref]

Surface water counted as ice needs warming but has already undergone melting. Its presence therefore biases a dry-ice estimate of latent heat of fusion downward. This is why drying the added ice is a quantitative control in ice-melting calorimetry, rather than cosmetic preparation. Let $m_a$ be the total added mass at $T_m$, of which a fraction $r$ is already liquid and $(1-r)$ remains solid. The correct constant-property balance isIf an analysis incorrectly treats all $m_a$ as solid and sets $Q_{\mathrm{env}}=0$, it obtainsThus $r=0.10$ gives a $10\%$ low estimate with no heat leak. Unaccounted incoming heat also lowers the estimate; outgoing heat raises it. These signs follow directly from the displayed energy balance. Conversely, dry ice initially below $T_m$ absorbs additional sensible heat; neglecting that term overestimates $L$. These are different biases and may partially cancel, so agreement with a reference value alone does not validate the controls.

For this model with $r=0$ and $Q_{\mathrm{env}}=0$, the balance predicts $T_f-T_m\simeq 8.829\mathrm{K}$, with $16.700\mathrm{kJ}$ consumed by fusion and $1.845\mathrm{kJ}$ by warming the melted water. At $r=0.10$ with the same zero leak, it predicts $T_f-T_m\simeq 10.369\mathrm{K}$ and $L_{\mathrm{app}}=300.6\mathrm{kJ}\mathrm{k}{\mathrm{g}}^{-1}$. A net incoming $1\mathrm{kJ}$ shifts $L_{\mathrm{app}}$ downward by a further $20\mathrm{kJ}\mathrm{k}{\mathrm{g}}^{-1}$ because $m_a=0.050\mathrm{kg}$. The warmer final water in the wet case is evidence of a smaller fusion budget, not a larger true $L$.

Superfluidity is a quantum-fluid response with phase stiffness and flow without viscous dissipation in an appropriate low-speed regime. A superfluid supports a phase-gradient current; defects and excitations can limit that flow. Bose-Einstein condensation concerns macroscopic one-particle occupation and is a distinct criterion: the Gaussian phase stiffness theory can have divergent absolute-phase fluctuations while retaining a cost for gradients.

Why does cooling a liquid make it flow more freely through a tiny gap? Helium-4 supplies an experimental entrance to this question: the helium slit-flow experiment compares the motion of liquid levels on opposite sides of the transition into superfluidity. A flow measurement probes resistance to motion; a measurement of Bose-Einstein condensation probes occupation of a quantum state. The two-fluid model of superfluidity and the Landau criterion connect these questions to entropy transport and to the energy required to create an excitation.

Two weakly linked Bose-Einstein condensates form a bosonic Josephson junction. Their relative phase controls a particle current through the weak link, analogous to the pair current of a superconducting Josephson junction. For neutral reservoirs the phase evolution follows a chemical-potential difference rather than an electric voltage. Interactions can make the tunneling dynamics nonlinear, including self-trapping at sufficiently large imbalance.

Two weakly coupled condensates in a double-well potential realize one bosonic Josephson junction. The 2005 experiment measured the density distribution in situ and inferred the evolving relative phase from interference fringes. It distinguished tunneling oscillations from nonlinear suppression of large-amplitude transfer, called macroscopic quantum self-trapping.^[ref]

A repulsive optical obstacle probes dissipation by being moved through a condensate. Heating as a function of speed tests low-speed flow stability; stationary-obstacle and thermal-cloud controls help separate other heating mechanisms. The measured onset need not equal the homogeneous critical velocity of a homogeneous condensate because density gradients, obstacle shape and vortex nucleation matter.

The lambda point is the temperature of the transition between ordinary liquid helium-4 and its superfluid phase, named after the sharp lambda-shaped heat capacity feature. At saturated vapor pressure it lies near $2.17\mathrm{K}$. Its value varies with pressure.

The particle-number density associated with the superfluid response. In two dimensions it is a number per unit area and gives the phase stiffness ${\rho }_s={\hslash }^2{n}_s/m$ for particles of mass $m$. It differs from total density and need not equal condensate density.

In the two-fluid model of superfluidity, the superfluid fraction is ${\rho }_s/\rho$, where ${\rho }_s$ is the superfluid mass density and $\rho$ the total mass density. For a one-component fluid of particles of the same mass this also equals $n_s/n$, using the superfluid number density $n_s$ and total number density $n$. It measures the superfluid response and is distinct from the condensate fraction, which measures occupation of a single-particle state. Interactions can make the two fractions very different.

A superfluid exhibits superfluidity, with phase stiffness and a phase-gradient current. In an appropriate ordered or quasi-ordered regime it supports low-speed flow without viscous dissipation, and its long-wavelength Goldstone boson has a sound-like dispersion. True thermodynamic phase selection and Bose-Einstein condensation need separate infrared conditions.

The normal component of a superfluid describes the momentum and entropy carried by excitations rather than the phase-gradient condensate current. Its velocity need not equal the superfluid velocity. In phenomenological condensate equations, prescribing its motion supplies a reference flow for dissipation; a full two-component fluid theory would also evolve its momentum and energy.

These are two responses of one liquid, not separable kinds of atoms. Their fractions need not equal the Bose-Einstein condensation fraction.^[ref]

The two-fluid model of superfluidity gives the normal component of a superfluid and the superfluid independent velocities.

An elementary uniform-flow example makes the distinction operational. Write the mass flux per unit area and the ideal entropy flux asHere $s$ is specific entropy, in $\mathrm{J}\mathrm{k}{\mathrm{g}}^{-1}{\mathrm{K}}^{-1}$, and ${\rho }_s,{\rho }_n$ are response mass densities. The mass density ${\rho }_s$ used here gives the three-dimensional phase stiffness coefficient ${\rho }_s(\hslash /m{)}^2$ when ${\mathbf{v}}_s=(\hslash /m)\nabla \theta$ is substituted into the kinetic energy density ${\rho }_s{v}_s^2/2$. For a stipulated ${\rho }_s=3\rho /4$, ${\rho }_n=\rho /4$ and $v_n=3u$ along a channel, zero net mass flux requires $v_s=-u$. Yet $j_S=3\rho su$ is positive. Opposing velocities can therefore transport entropy while transporting no net mass.

This example explains why excellent apparent heat transport need not be ordinary diffusion of heat through a stationary liquid. In the ideal zero-net-mass-flux limit, the heat flux associated with this entropy transport is $T{j}_S$. Viscous losses, quantum vortices and external heat exchange require additional terms. The model also supports second sound.

Closely spaced disks undergo torsional oscillations in He-II. Viscous entrainment of the normal component contributes a temperature-dependent moment of inertia, while vortex-free superfluid need not rotate with the disks. The interpretation requires the normal component to follow the drive under the apparatus's spacing and frequency conditions. The experiment is associated with Elephter Andronikashvili.

Second sound is an entropy and temperature wave. At negligible thermal expansion, it has almost zero net mass flux. Lev Davidovich Landau predicted two sound modes in 1941.^[ref]

The thermomechanical effect couples a temperature difference to a pressure difference across a superleak. For the closed-chamber model in the apparatus figure, consider neighboring equilibrium states with small velocities, negligible height difference and pure helium-4. Let ${\mu }_m$ be chemical potential per unit mass, in $\mathrm{J}\mathrm{k}{\mathrm{g}}^{-1}$. The Gibbs free energy differential per unit mass givesThe superleak allows the superfluid to redistribute until the two sides have equal ${\mu }_m$. Setting $d{\mu }_m=0$ therefore givesThe sign matters: heating one side at initially equal pressure lowers its ${\mu }_m$; inflow raises its pressure until chemical balance is restored. The hotter side has the higher pressure. The integral follows the constant-${\mu }_m$ path; it need not be evaluated at fixed pressure when material properties vary appreciably.

For an original local example, stipulate constant $\rho =150\mathrm{kg}{\mathrm{m}}^{-3}$, $s=100\mathrm{J}\mathrm{k}{\mathrm{g}}^{-1}{\mathrm{K}}^{-1}$ and $T_h-T_c=0.020\mathrm{K}$. The predicted pressure difference is $300\mathrm{Pa}$. These are teaching assumptions, not helium measurements. This static thermodynamics relation predicts neither the required heater power nor the filling time: those require heat leakage, channel conductance and chamber heat capacity. An outlet would introduce kinetic energy and a flowing balance, so the static $300\mathrm{Pa}$ is not a predicted fountain height.

A superleak is a narrow or porous connection that strongly suppresses the viscous normal component of a superfluid while allowing appreciable superfluid flow. Treating ${\mathbf{v}}_n\simeq 0$ is a regime assumption, not a perfect material filter for particular atoms. The thermomechanical effect uses that distinction to maintain a temperature difference during chemical equilibration. The idealization fails when normal leakage or dissipative quantum vortex motion becomes appreciable.