Helmholtz Free Energy

Definition

The Helmholtz free energy of a system is defined as

F = U − T·S

where U is the internal energy (J), T is the absolute temperature (K), and S is the entropy (J/K), so the product TS also carries units of joules and F is an energy. The symbol is usually F (from the German freie Energie) or sometimes A (from Arbeit, work) in chemistry texts.

The physical reading is a competition between two terms. U is the total energy locked in the system's microscopic degrees of freedom. TS is the part of that energy that is thermally "unavailable" — energy that must stay tied up in maintaining the system's entropy at temperature T. What is left over, F, is the energy that is free to be converted into work while the temperature is held fixed by contact with a reservoir.

Why it matters

For an isolated system the equilibrium rule is simple: entropy is maximized. But almost nothing in a lab or in nature is truly isolated — most systems sit in a thermostatted bath, exchanging heat freely with their surroundings while their volume is pinned by rigid walls. For those conditions the relevant extremal principle is not "maximize S" but "minimize F." The Helmholtz free energy is precisely the quantity nature minimizes when temperature and volume are the controlled variables.

That makes F the workhorse of statistical mechanics. The canonical ensemble — the model of a system exchanging energy with a heat bath — hands you F directly through F = −kT ln Z, and every measurable thermodynamic property is a derivative of F. It is also central to phase transitions: the phase with the lower Helmholtz free energy at a given (T, V) is the stable one, and the crossing of two F curves marks a first-order transition. Elasticity of polymers, the equation of state of gases, magnetization of spin systems, and the melting of crystals are all read off from F.

How it works, step by step

1. Start from the second law with a reservoir. Put the system in thermal contact with a reservoir at temperature T. The reservoir is huge, so its temperature never changes. The combined entropy of system plus reservoir cannot decrease: dS_sys + dS_res ≥ 0.

2. Express the reservoir's entropy change. Heat δQ flowing into the system leaves the reservoir, so dS_res = −δQ/T. Substituting gives dS_sys − δQ/T ≥ 0, i.e. δQ ≤ T dS_sys.

3. Bring in the first law. Energy conservation says dU = δQ − δW, where δW is work done by the system. Replace δQ using step 2: dU ≤ T dS_sys − δW, which rearranges to δW ≤ T dS_sys − dU.

4. Recognize the free energy. At constant temperature, T dS − dU = −d(U − TS) = −dF. So δW ≤ −dF, and integrating over a finite isothermal process gives W ≤ −ΔF. The most work you can ever extract equals the drop in Helmholtz free energy, achieved only in the reversible limit.

5. Specialize to no work (fixed volume). If the walls are rigid and the system does no work, δW = 0, and the inequality collapses to dF ≤ 0. F can only fall, and it stops falling at the minimum — that stationary point is thermal equilibrium.

The differential and the Maxwell relation

Starting from F = U − TS and the fundamental relation dU = T dS − P dV + μ dN, the entropy terms cancel and you are left with the exact differential in F's natural variables T, V, and N:

dF = −S dT − P dV + μ dN

Here P is pressure (Pa), μ is the chemical potential (J per particle, or J/mol), and N is particle number (or moles). Reading off the partial derivatives:

Quantity	Obtained from F	Held fixed
Entropy S	S = −(∂F/∂T)V,N	V, N
Pressure P	P = −(∂F/∂V)T,N	T, N
Chemical potential μ	μ = (∂F/∂N)T,V	T, V
Internal energy U	U = F + TS = F − T(∂F/∂T)V	V, N

Because dF is an exact differential, the mixed second partial derivatives are equal, which yields the Helmholtz Maxwell relation

(∂S/∂V)_T = (∂P/∂T)_V

This lets you replace an entropy derivative (hard to measure) with a pressure derivative (easy to measure) — one of the everyday payoffs of the free-energy formalism.

Legendre transform of internal energy

Helmholtz free energy is not an independent postulate; it is the Legendre transform of the internal energy U(S, V, N) that swaps the entropy S for its conjugate variable, the temperature T = (∂U/∂S)_V,N. The transform subtracts the product of the old variable and its conjugate:

F(T, V, N) = U − S·(∂U/∂S) = U − T·S

The point of the transform is to change the independent variable from the awkward, hard-to-fix S to the easily controlled T without losing any information. The four common potentials are all Legendre transforms of one another:

Potential	Definition	Natural variables	Minimized when
Internal energy U	U	S, V, N	S, V fixed (isolated)
Enthalpy H	H = U + PV	S, P, N	S, P fixed
Helmholtz free energy F	F = U − TS	T, V, N	T, V fixed
Gibbs free energy G	G = U + PV − TS	T, P, N	T, P fixed

Reading the table left to right: swap S for T to go from U to F; add PV to go from F to G. This is why G = F + PV, and why chemists (working at fixed atmospheric pressure) reach for G while physicists (working at fixed volume, or computing partition functions) reach for F.

The bridge to statistical mechanics

The deepest identity of the Helmholtz free energy lives in the canonical ensemble. A system in contact with a bath at temperature T occupies microstate i of energy E_i with probability p_i = exp(−E_i/kT) / Z, where the normalizing sum

Z = Σ_i exp(−E_i / kT)

is the canonical partition function, k = 1.381 × 10⁻²³ J/K is the Boltzmann constant, and the sum runs over all microstates. The Helmholtz free energy is then

F = −k·T·ln Z

Once you have ln Z as a function of T and V, all of thermodynamics follows by differentiating: S = −(∂F/∂T)_V, P = −(∂F/∂V)_T, and U = kT²(∂ ln Z/∂T)_V. For an ideal monatomic gas of N atoms the single-particle partition function gives F = −NkT[ln(nV/N) + 1] with n = (2πmkT/h²)^3/2, from which P = −(∂F/∂V)_T = NkT/V reproduces the ideal gas law PV = NkT. The abstract F = −kT ln Z thus contains PV = nRT as a special case.

Worked example: isothermal expansion of an ideal gas

Take 1 mole of an ideal gas at T = 300 K and expand it reversibly and isothermally from V₁ = 1.0 L to V₂ = 2.0 L. Because U of an ideal gas depends only on temperature, ΔU = 0 for this isothermal step. The change in Helmholtz free energy is therefore ΔF = ΔU − TΔS = −TΔS.

The entropy change of an isothermal ideal-gas expansion is ΔS = nR ln(V₂/V₁) = (1 mol)(8.314 J/(mol·K)) ln 2 = 5.76 J/K. So

ΔF = −T·ΔS = −(300 K)(5.76 J/K) = −1.73 kJ

The maximum work the gas can deliver in this isothermal expansion is W_max = −ΔF = +1.73 kJ. Because the process is reversible, the gas actually does exactly this much work, and it must absorb 1.73 kJ of heat from the reservoir (Q = W since ΔU = 0). Had the expansion been carried out irreversibly — say a sudden free expansion into vacuum — the gas would do less work (zero, in the free-expansion limit), while ΔF is unchanged because F is a state function. The gap between −ΔF and the work actually done is exactly the dissipation.

A note on history

The quantity was introduced by Hermann von Helmholtz in his 1882 paper Die Thermodynamik chemischer Vorgänge ("The Thermodynamics of Chemical Processes"). Helmholtz split the internal energy into a "free energy" that could be converted to work and a "bound energy" (the TS term) tied up with the entropy — the language that survives today. It built on the Clausius formulation of entropy (1865) and was contemporaneous with the work of Willard Gibbs, whose analogous constant-pressure potential G became the standard tool of physical chemistry. The two potentials, F and G, remain the twin pillars of equilibrium thermodynamics: F for fixed volume, G for fixed pressure.

Common misconceptions

• "Free energy is the energy the system has left." F is not stored energy in the naïve sense; it is a potential whose decrease bounds extractable work at constant T. Its absolute value even carries the same reference-state ambiguity as U and S — only ΔF is physically meaningful.
• "F is always the maximum useful work." −ΔF bounds the total work at constant T, including any work done pushing back the atmosphere. If you want the non-expansion (useful) work at constant T and P, use −ΔG instead. Mixing up F and G is the single most common error.
• "Minimizing F contradicts maximizing entropy." They are the same principle under different constraints. Minimizing F = U − TS at fixed T and V is exactly equivalent to maximizing the total entropy of system plus reservoir. The −TS term in F is the reservoir's entropy accounting.
• "F and G differ by a big amount." They differ only by PV. For condensed phases at ordinary pressures PV is tiny compared to U, so F ≈ G for solids and liquids; the distinction matters mostly for gases and high-pressure work.
• "You can pick T and S independently in F." The natural variables of F are T and V, not T and S. Choosing S as an independent variable belongs to U; the whole point of the Legendre transform was to eliminate S in favor of T.
• "F = −kT ln Z needs a formula for Z." The identity is exact for any Z, even a numerical sum over energy levels. You never need a closed form — a computed partition function gives a computed free energy directly.