Euler Relation and Gibbs–Duhem Equation

Water boils at exactly one temperature for a given pressure — you cannot choose them independently. Ice, water, and steam coexist at exactly one point, the triple point, with no freedom at all. Why should nature be so rigidly constrained? The answer lives in a single deceptively simple equation, the Gibbs–Duhem relation, which says the intensive variables of a system — temperature, pressure, chemical potential — are not independent. Change one and the others must adjust in lockstep. This constraint, and the Euler relation it descends from, is the reason phase diagrams have the shape they do. Both relations are pure consequences of extensivity: the fact that doubling a system doubles its energy, entropy, volume, and particle number together. Mathematically, extensivity makes the fundamental relation $U(S,V,N)$ a first-order homogeneous function, and Euler's theorem on homogeneous functions then forces a remarkable finite (not differential) identity: $U = TS - PV + \mu N$ . The whole internal energy is reconstructed from the intensive variables times their extensive partners. This is the Euler relation, and it is startling — a genuine equation of state that follows from additivity alone. Take the differential of the Euler relation and subtract the fundamental relation $\mathrm{d}U = T\,\mathrm{d}S - P\,\mathrm{d}V + \mu\,\mathrm{d}N$ , and almost everything cancels. What survives is the Gibbs–Duhem equation, $S\,\mathrm{d}T - V\,\mathrm{d}P + N\,\mathrm{d}\mu = 0$ : a constraint linking the changes of the three intensive variables. It says that for a single-component system only two of the three intensive variables can be chosen freely; the third is determined. This is the seed of the Gibbs phase rule, the counting law $f = c - p + 2$ that governs how many phases can coexist. These relations are not abstract curiosities. The Gibbs–Duhem equation underlies the Clausius–Clapeyron slope of coexistence curves, the Duhem–Margules relation for vapor pressures of mixtures, and the thermodynamic consistency tests applied to every measured phase diagram. When you see that liquid–vapor coexistence is a curve in the $P$ – $T$ plane rather than an area, and that the triple point is a single point, you are looking directly at the arithmetic of Euler and Gibbs–Duhem.

Key takeaways

Extensivity (first-order homogeneity of $U(S,V,N)$ ) forces the Euler relation $U = TS - PV + \mu N$ — a finite identity, remarkable because it reconstructs $U$ from the intensive variables.
Taking the differential of Euler and subtracting the fundamental relation leaves the Gibbs–Duhem equation $S\,\mathrm{d}T - V\,\mathrm{d}P + N\,\mathrm{d}\mu = 0$ .
Gibbs–Duhem says the intensive variables are not independent: for one component, only 2 of the 3 ( $T, P, \mu$ ) can be chosen freely.
This is the seed of the Gibbs phase rule $f = c - p + 2$ : coexistence of $p$ phases removes degrees of freedom (triple point $= 0$ ).
For a single component, the Euler relation also gives $G = \mu N$ : the chemical potential is the molar Gibbs energy.
Key idea: extensivity forces the Euler relation $U = TS - PV + \mu N$ , whose differential is the Gibbs–Duhem constraint $S\mathrm{d}T - V\mathrm{d}P + N\mathrm{d}\mu = 0$ , reducing the free intensive variables and setting the geometry of phase diagrams.

The core insight

Because the fundamental relation $U(S,V,N)$ is first-order homogeneous (extensivity), Euler's theorem forces the finite Euler relation $U = TS - PV + \mu N$ . Its differential minus the fundamental relation gives the Gibbs–Duhem equation $S\,\mathrm{d}T - V\,\mathrm{d}P + N\,\mathrm{d}\mu = 0$ , which makes the intensive variables interdependent and is the seed of the phase rule $f = c - p + 2$ .

Where this lives today

The Euler and Gibbs–Duhem equations are fundamental tools throughout equilibrium thermodynamics, especially for multicomponent systems and phase equilibria. In solution chemistry and chemical engineering, the multicomponent Gibbs–Duhem equation $\sum_i x_i\,\mathrm{d}\mu_i = 0$ (at fixed $T, P$ ) is used daily as a thermodynamic consistency test: measured activity coefficients of the components of a mixture must satisfy it, and any dataset that violates it is rejected as experimentally flawed. The Duhem–Margules relation it yields predicts one component's vapor pressure from the other's — essential in distillation design. In materials science, the phase rule $f = c - p + 2$ is the organizing principle behind every phase diagram, from binary alloys to multicomponent ceramics and geological systems; the CALPHAD community uses it to know exactly how many phases can coexist and how many variables label each region. In geochemistry and petrology, Gibbs–Duhem and the phase rule tell a geologist how many minerals can coexist in a rock at a given depth, turning a mineral assemblage into a thermobarometer that reads pressure and temperature deep in the Earth. And the Clausius–Clapeyron relation that Gibbs–Duhem generates governs everything from the design of steam cycles to the humidity physics of the atmosphere — the same equation that sets how much water vapor warming air can hold, a keystone of climate modeling.

The physical picture

Everything here flows from one homely fact: thermodynamics is additive. Put two identical boxes of gas side by side and remove the partition, and the combined system has twice the energy, twice the entropy, twice the volume, twice the particles. The intensive variables — temperature, pressure, chemical potential — stay the same. This is extensivity, and it is a strong constraint: it says the fundamental relation $U(S,V,N)$ must be a first-order homogeneous function, scaling exactly linearly with system size. Euler's theorem on homogeneous functions turns this scaling property into an algebraic identity. For any function that scales linearly, the sum of each variable times its partial derivative equals the function itself. Applied to $U$ , with the derivatives being $T$ , $-P$ , $\mu$ , this reads $TS - PV + \mu N = U$ . What is remarkable is that this is a finite relation — not a differential — that reconstructs the total energy from products of conjugate variables. Extensivity alone, no other law, gives you an equation of state.

Figure

Extensivity: doubling the system doubles extensive variables, leaves intensive ones fixed

Now differentiate the Euler relation and compare the result with the fundamental thermodynamic relation $\mathrm{d}U = T\,\mathrm{d}S - P\,\mathrm{d}V + \mu\,\mathrm{d}N$ . Differentiating $U = TS - PV + \mu N$ generates additional terms involving the differentials of the intensive variables, namely $S\,\mathrm{d}T$ , $-V\,\mathrm{d}P$ , and $N\,\mathrm{d}\mu$ . Consistency with the fundamental thermodynamic relation requires these additional terms to cancel, yielding the Gibbs–Duhem equation.

Figure

Gibbs–Duhem constrains phase diagrams: coexistence is a curve, the triple point a single point

The geometric consequence is immediate. For a one-component system, the Gibbs–Duhem equation leaves only two independent intensive variables, so a single equilibrium phase occupies a two-dimensional region in the $P$ – $T$ plane. Demand that two phases coexist ( $\mu_\alpha = \mu_\beta$ ) and you impose one more equation, dropping to one freedom — a curve. Three phases ( $\mu_\alpha = \mu_\beta = \mu_\gamma$ ) leave zero freedom — a point, the triple point. That is the phase rule $f = c - p + 2$ , and it is nothing but Gibbs–Duhem counting.

The core equations

Extensivity is the statement that the fundamental relation is first-order homogeneous:

U(\lambda S, \lambda V, \lambda N) = \lambda\, U(S,V,N) \qquad (1)

where: $\lambda > 0$ — arbitrary scale factor; the intensive variables $T, P, \mu$ depend only on ratios (densities) and are unchanged by scaling. Euler's theorem on first-order homogeneous functions applied to (1) yields the finite Euler relation:

U = TS - PV + \mu N \qquad (2)

where: $T = (\partial U/\partial S)_{V,N}$ , $-P = (\partial U/\partial V)_{S,N}$ , $\mu = (\partial U/\partial N)_{S,V}$ . This expresses the total internal energy as a sum of products of conjugate intensive and extensive variables—the Euler relation implied solely by extensivity. Taking the differential of (2) gives, by the product rule,

\mathrm{d}U = (T\,\mathrm{d}S + S\,\mathrm{d}T) - (P\,\mathrm{d}V + V\,\mathrm{d}P) + (\mu\,\mathrm{d}N + N\,\mathrm{d}\mu) \qquad (3)

where: every term is differentiated, including the intensive variables — unlike the fundamental relation, which holds them implicit. Subtracting the fundamental fundamental thermodynamic relation $\mathrm{d}U = T\,\mathrm{d}S - P\,\mathrm{d}V + \mu\,\mathrm{d}N$ from Eq. (3) cancels the $T\,\mathrm{d}S$ , $-P\,\mathrm{d}V$ , and $\mu\,\mathrm{d}N$ terms, leaving the Gibbs–Duhem equation:

S\,\mathrm{d}T - V\,\mathrm{d}P + N\,\mathrm{d}\mu = 0 \qquad (4)

where: the three intensive differentials are not independent; dividing by $N$ gives the molar form $s\,\mathrm{d}T - v\,\mathrm{d}P + \mathrm{d}\mu = 0$ , i.e. $\mathrm{d}\mu = -s\,\mathrm{d}T + v\,\mathrm{d}P$ . The counting consequence is the Gibbs phase rule:

f = c - p + 2 \qquad (5)

where: $f$ — number of independent intensive variables (degrees of freedom); $c$ — number of components; $p$ — number of coexisting phases. Each phase supplies one Gibbs–Duhem constraint; each coexistence condition removes a freedom. For a single component, the Euler relation collapses $G = U - TS + PV = \mu N$ , giving the useful identity:

\mu = \frac{G}{N} = g \qquad (6)

where: $g = G/N$ — molar Gibbs free energy. The chemical potential is the Gibbs energy per particle — the quantity minimized in phase and chemical equilibrium.

Derivation AEuler relation from extensivity, then Gibbs–Duhem as its differential

The goal is to establish the two relations in sequence and expose their logical dependence: extensivity forces Euler, and Euler's differential (minus the fundamental relation) forces Gibbs–Duhem. The Gibbs–Duhem equation is therefore not an independent postulate. It follows directly by differentiating the Euler relation and using the fundamental thermodynamic relation, so it holds for every extensive thermodynamic system. We carry it through algebraically and verify the cancellation symbolically.

Step 1. Extensivity means $U(\lambda S, \lambda V, \lambda N) = \lambda U$ . Differentiate both sides with respect to $\lambda$ and set $\lambda = 1$ : the chain rule gives $S\,U_S + V\,U_V + N\,U_N = U$ . Inserting $U_S = T$ , $U_V = -P$ , $U_N = \mu$ yields the Euler relation $TS - PV + \mu N = U$ . Why: Euler's homogeneous-function theorem converts the scaling property directly into a finite identity — no thermodynamic law beyond additivity is used.

Step 2. Differentiate the Euler relation $U = TS - PV + \mu N$ using the product rule on each term: $\mathrm{d}U = T\,\mathrm{d}S + S\,\mathrm{d}T - P\,\mathrm{d}V - V\,\mathrm{d}P + \mu\,\mathrm{d}N + N\,\mathrm{d}\mu$ . Why: because Euler is a finite relation among all six quantities, its differential contains both the "extensive-differential" terms and the "intensive-differential" terms.

Step 3. Subtract the fundamental relation $\mathrm{d}U = T\,\mathrm{d}S - P\,\mathrm{d}V + \mu\,\mathrm{d}N$ . The $T\,\mathrm{d}S$ , $-P\,\mathrm{d}V$ , $\mu\,\mathrm{d}N$ terms cancel identically, leaving $0 = S\,\mathrm{d}T - V\,\mathrm{d}P + N\,\mathrm{d}\mu$ — the Gibbs–Duhem equation. Why: the fundamental relation already accounts for the extensive-differential terms, so consistency demands the intensive-differential terms sum to zero.

Step 4. Read off the consequence: with three intensive variables ( $T, P, \mu$ ) linked by one equation, only two are independent for a single component. This counting leads directly to the Gibbs phase rule, $f=c-p+2$ : each additional coexistence condition removes one thermodynamic degree of freedom. Why: the counting of free intensive variables is exactly the counting of Gibbs–Duhem constraints versus coexistence conditions.

Go deeper · Derivation B · 1 minThe chemical potential of an ideal gas and the Clausius–Clapeyron slope

Derivation BThe chemical potential of an ideal gas and the Clausius–Clapeyron slope

The goal is to put the molar Gibbs–Duhem equation to work, deriving two concrete and important results: the pressure dependence of the ideal-gas chemical potential (which underlies the law of mass action and every equilibrium constant), and the Clausius–Clapeyron slope of a coexistence curve (which is why the boiling point depends on altitude). Both follow directly from $\mathrm{d}\mu = -s\,\mathrm{d}T + v\,\mathrm{d}P$ applied to specific situations.

Step 1. Molar Gibbs–Duhem: divide (4) by $N$ to get $\mathrm{d}\mu = -s\,\mathrm{d}T + v\,\mathrm{d}P$ , where $s = S/N$ , $v = V/N$ . This is the differential of $\mu$ in its natural variables $(T,P)$ , consistent with $\mu = g = G/N$ . Why: since $\mu$ is the molar Gibbs energy, its differential is the molar version of $\mathrm{d}G = -S\,\mathrm{d}T + V\,\mathrm{d}P$ .

Step 2. For an ideal gas at fixed $T$ , $v = k_BT/P$ , so $\mathrm{d}\mu = v\,\mathrm{d}P = (k_BT/P)\,\mathrm{d}P$ . Integrate from a reference pressure $P_0$ : $\mu(T,P) = \mu_0(T) + k_BT\ln(P/P_0)$ . Why: the logarithmic pressure dependence of the chemical potential is the origin of the law of mass action and of every $\Delta G = -RT\ln K$ equilibrium relation.

Step 3. For two coexisting phases $\alpha, \beta$ , equilibrium requires $\mu_\alpha = \mu_\beta$ everywhere along the coexistence curve, so $\mathrm{d}\mu_\alpha = \mathrm{d}\mu_\beta$ . Using the molar Gibbs–Duhem for each: $-s_\alpha\,\mathrm{d}T + v_\alpha\,\mathrm{d}P = -s_\beta\,\mathrm{d}T + v_\beta\,\mathrm{d}P$ . Why: the phases must share $T$ and $P$ and remain in chemical equilibrium as you move along the curve, so their chemical potentials change together.

Step 4. Solve for the slope: $\dfrac{\mathrm{d}P}{\mathrm{d}T} = \dfrac{s_\beta - s_\alpha}{v_\beta - v_\alpha} = \dfrac{\Delta s}{\Delta v} = \dfrac{L}{T\,\Delta v}$ , the Clausius–Clapeyron relation, with $L = T\Delta s$ the latent heat. Why: the coexistence-curve slope is fixed entirely by the entropy and volume jumps between phases — a direct dividend of Gibbs–Duhem, and the reason water boils at lower temperature on a mountain.

Limiting cases

How the relations play out:

Case	Result	Meaning
Single component	$G = \mu N$ , so $\mu = g = G/N$	Chemical potential = molar Gibbs energy.
One phase (c=1, p=1)	$f = 2$	Two free intensive variables — a phase fills an area.
Two-phase coexistence	$f = 1$	Coexistence is a curve; slope = Clausius–Clapeyron $\Delta s/\Delta v$ .
Triple point (c=1, p=3)	$f = 0$	No freedom — a single point (unique $T, P$ ).
Ideal gas, fixed $T$	$\mu = \mu_0 + k_BT\ln(P/P_0)$	Logarithmic pressure dependence → law of mass action.
Non-extensive (gravity)	Euler & Gibbs–Duhem fail	Long-range forces break homogeneity; new thermodynamics needed.

Common trap“The Euler relation is just the fundamental relation integrated”

It is not, and the distinction matters. The fundamental relation $\mathrm{d}U = T\,\mathrm{d}S - P\,\mathrm{d}V + \mu\,\mathrm{d}N$ is a differential statement valid for any system. The Euler relation $U = TS - PV + \mu N$ is a finite, integrated statement that holds only because of extensivity (first-order homogeneity) — you cannot get it by naive integration of the differential, which would leave an undetermined function. The proof genuinely requires Euler's homogeneous-function theorem. This is why the Euler relation fails for systems where energy is not extensive — self-gravitating star clusters, small clusters with large surface-to-volume ratios, systems with long-range interactions — even though the differential fundamental relation still holds. Gibbs–Duhem fails with it. Extensivity is doing essential work.

Worked examples

1 · The phase rule counts the geometry of a phase diagram (confidence builder)

Use the phase rule $f = c - p + 2$ to explain why, for pure water ( $c = 1$ ), single phases occupy areas, coexistence is a line, and the triple point is a single point. Then find how many phases can coexist at most.

Step 1. Single phase, $p = 1$ : $f = 1 - 1 + 2 = \mathbf{2}$ . Two intensive variables ( $T$ and $P$ ) can be varied independently, so a single phase (liquid, say) fills a two-dimensional area of the $P$ – $T$ plane.

Step 2. Two phases coexisting, $p = 2$ : $f = 1 - 2 + 2 = \mathbf{1}$ . Only one variable is free — fix $T$ and $P$ is determined, tracing a curve (the boiling or melting line). This is why the boiling point changes with altitude: move along the coexistence curve and $T$ follows $P$ .

Step 3. Three phases, $p = 3$ : $f = 1 - 3 + 2 = \mathbf{0}$ . No freedom at all — solid, liquid, and vapor coexist at exactly one $(T, P)$ , the triple point of water ( $273.16\ \mathrm{K}$ , $611\ \mathrm{Pa}$ ), a single point.

Step 4. Maximum phases: $f \ge 0$ requires $p \le c + 2 = 3$ for one component. So a pure substance can have at most three phases coexisting in equilibrium — never four at a single point. The entire architecture of the phase diagram is dictated by Gibbs–Duhem counting.

2 · Chemical potential shift and the boiling-point drop on a mountain (exam level)

(a) Find the change in chemical potential of $1\ \mathrm{mol}$ of an ideal gas when its pressure is doubled at $T = 300\ \mathrm{K}$ . (b) Estimate how much the boiling point of water drops on a mountain where the pressure is $0.7\ \mathrm{atm}$ , using Clausius–Clapeyron with latent heat $L = 40.7\ \mathrm{kJ/mol}$ .

Step 1. Chemical potential shift: from $\mu = \mu_0 + RT\ln(P/P_0)$ (per mole), doubling gives $\Delta\mu = RT\ln 2 = (8.314)(300)(0.693) = \mathbf{1729\ J/mol}$ . Increasing the pressure raises the chemical potential, reflecting the larger Gibbs free energy per mole of the compressed gas.

Step 2. Clausius–Clapeyron for liquid–vapor, approximating $\Delta v \approx v_{\text{vap}} = RT/P$ (vapor volume dominates): $\dfrac{\mathrm{d}P}{\mathrm{d}T} = \dfrac{L}{T\,\Delta v} = \dfrac{LP}{RT^2}$ , so $\dfrac{\mathrm{d}\ln P}{\mathrm{d}T} = \dfrac{L}{RT^2}$ .

Step 3. Assuming the latent heat $L$ is approximately temperature-independent, integration gives $\ln(P/P_0) = -\dfrac{L}{R}\left(\dfrac1T - \dfrac1{T_0}\right)$ . Set $P/P_0 = 0.7$ , $T_0 = 373\ \mathrm{K}$ , $L/R = 40\,700/8.314 = 4895\ \mathrm{K}$ : $\ln 0.7 = -0.357 = -4895(1/T - 1/373)$ .

Step 4. Solve: $1/T - 1/373 = 0.357/4895 = 7.29\times10^{-5}$ , so $1/T = 2.681\times10^{-3} + 7.29\times10^{-5} = 2.754\times10^{-3}$ , giving $T = \mathbf{363\ K} = 90^\circ\mathrm{C}$ . The boiling temperature decreases by about $10\,{\rm K}$ , illustrating quantitatively how the Clausius--Clapeyron equation predicts the pressure dependence of the boiling point.

Origins · 1 mingibbs-duhem equation and how this was discovered

gibbs-duhem equation

concept · constraint on intensive variables

The equation is named jointly for Josiah Willard Gibbs and Pierre Duhem. Gibbs derived the underlying structure in his monumental 1875–78 memoir On the Equilibrium of Heterogeneous Substances, where he introduced the chemical potential, the fundamental relation, and the phase rule; the constraint among intensive variables is implicit throughout. Pierre Duhem, a French physicist and philosopher of science, developed and popularized the relation in the 1880s–1890s, applying it systematically to solutions and mixtures — the multicomponent form $\sum_i N_i\,\mathrm{d}\mu_i = -S\,\mathrm{d}T + V\,\mathrm{d}P$ is central to solution thermodynamics and the Duhem–Margules equation for vapor pressures. The relation encodes a deep truth: in a system at equilibrium, the intensive fields are not independent knobs but a constrained set, and this constraint is what gives phase diagrams their rigid, universal geometry. Duhem, incidentally, is equally remembered for his philosophy of science (the Duhem–Quine thesis) and for reviving interest in medieval physics.

History

1755

— Euler proves the homogeneous-function theorem that underlies the Euler relation.

1865

— Clausius writes dU = T dS − P dV, the fundamental relation Gibbs–Duhem descends from.

1875

— Gibbs, On the Equilibrium of Heterogeneous Substances, introduces μ, the fundamental relation, and the phase rule.

1876

— Gibbs states the phase rule f = c − p + 2 explicitly.

1886

— Duhem develops the constraint among intensive variables for solutions and mixtures.

1898

— Margules and Duhem give the Duhem–Margules relation for the vapor pressures of binary mixtures.

Going further

References. Callen, Thermodynamics and an Introduction to Thermostatistics, 2nd ed., ch. 3 (Euler and Gibbs–Duhem) and ch. 9 (the phase rule) · Pathria & Beale, Statistical Mechanics, 3rd ed., ch. 1 · Kardar, Statistical Physics of Particles, ch. 1 · Reif, Fundamentals of Statistical and Thermal Physics, ch. 8 · Denbigh, The Principles of Chemical Equilibrium (Gibbs–Duhem for mixtures) · Gibbs, On the Equilibrium of Heterogeneous Substances (1875–78). Key numerical benchmarks: the ideal-gas chemical potential $\mu = \mu_0 + RT\ln(P/P_0)$ gives $\Delta\mu = RT\ln 2 = 1729\ \mathrm{J/mol}$ for a pressure doubling at $300\ \mathrm{K}$ ; the Clausius–Clapeyron equation with $L = 40.7\ \mathrm{kJ/mol}$ places the boiling point of water at $0.7\ \mathrm{atm}$ at $363\ \mathrm{K}$ . Throughout, $R = 8.314\ \mathrm{J\,mol^{-1}K^{-1}}$ and the triple point of water is $273.16\ \mathrm{K}$ , $611\ \mathrm{Pa}$ .

Builds on: Thermodynamic Potentials and the Fundamental Relation · Temperature, Pressure, and Chemical Potential as Entropy Derivatives · Maxwell Relations · Second Law of Thermodynamics and Entropy. Leads to: Helmholtz Free Energy, Enthalpy, and Gibbs Free Energy · The Microcanonical Ensemble and the Postulate of Equal a Priori Probabilities · Legendre Transforms.

Why is the Euler relation a "finite" identity while the fundamental relation is a differential?

The fundamental relation $\mathrm{d}U = T\,\mathrm{d}S - P\,\mathrm{d}V + \mu\,\mathrm{d}N$ tells you how $U$ changes for small changes in $S, V, N$ — it is a statement about slopes. The Euler relation $U = TS - PV + \mu N$ tells you the actual value of $U$ at any point, reconstructed from the intensive variables and the extensive variables at that point. Getting a finite value from a differential normally requires integration and leaves an unknown constant — but here extensivity does something special. Because $U$ scales linearly with system size, Euler's theorem pins down the integration exactly, with no free constant: the finite relation is forced. This only works for homogeneous (extensive) functions, which is why the Euler relation is a genuine, constant-free equation of state while the fundamental relation is a differential.

Does the Gibbs–Duhem equation mean I can never vary T, P, and μ at once?

For a single-component system in a single phase, yes — you can independently vary only two of the three, and Gibbs–Duhem determines the third. Physically: if you fix $T$ and $P$ for a pure substance, its chemical potential is fully determined (there is nothing left to choose). This is not a limitation of your control apparatus; it is a statement about how many independent intensive quantities the system possesses. If you have more components, each new component $i$ brings its own chemical potential $\mu_i$ and the multicomponent Gibbs–Duhem $\sum_i N_i\,\mathrm{d}\mu_i = -S\,\mathrm{d}T + V\,\mathrm{d}P$ gives one constraint among the larger set — so mixtures have more freedom, exactly as the phase rule $f = c - p + 2$ counts.

How does the phase rule come out of Gibbs–Duhem so cleanly?

Count the variables and the constraints. For $c$ components and $p$ phases, the intensive variables describing each phase are $T$ , $P$ , and the composition ( $c - 1$ independent mole fractions), so each phase has $2 + (c-1) = c + 1$ intensive variables. But each phase satisfies one Gibbs–Duhem relation, and for the phases to coexist in equilibrium every component's chemical potential must be equal across all phases, giving $c(p-1)$ equality conditions. The net number of free variables is $f = p(c+1) - p - c(p-1)$ , which simplifies to $f = c - p + 2$ . Every " $-1$ " in that count is a Gibbs–Duhem constraint or a coexistence equality — so the phase rule is literally the bookkeeping of how many Gibbs–Duhem-type constraints act on the intensive variables. That is why coexistence lowers dimensionality by exactly one per phase.

Why does the Euler relation fail for self-gravitating systems?

Because gravity is long-range, and long-range forces break extensivity. The whole derivation rests on $U(\lambda S, \lambda V, \lambda N) = \lambda U$ — energy scaling linearly with system size. But the gravitational potential energy of a self-bound system scales like $N^2$ (every particle interacts with every other), not $N$ , so doubling the system more than doubles the energy. Homogeneity fails, Euler's theorem no longer applies, and the finite Euler relation $U = TS - PV + \mu N$ is simply wrong. Gibbs–Duhem fails with it, ensembles become inequivalent, and star clusters can even have negative heat capacity. This is a deep and active area: the thermodynamics of gravitating systems and other non-extensive systems requires abandoning the comfortable machinery this article builds — a reminder that the elegance of Euler and Gibbs–Duhem is bought with the assumption of short-range interactions.

Is the multicomponent Gibbs–Duhem equation really used in practice, or is it just theory?

It is used constantly, especially as a data-quality gatekeeper. For a mixture at fixed $T$ and $P$ , Gibbs–Duhem reduces to $\sum_i x_i\,\mathrm{d}\mu_i = 0$ , equivalently $\sum_i x_i\,\mathrm{d}\ln\gamma_i = 0$ for the activity coefficients $\gamma_i$ . When chemists or engineers measure the activity coefficients of the components of a solution (from vapor-pressure or freezing-point data), those measurements must satisfy this relation. A dataset that fails the Gibbs–Duhem consistency test is discarded as experimentally unreliable — the relation catches systematic errors that would otherwise corrupt a distillation or extraction design. It also lets you compute one component's activity from the other's when direct measurement is hard. So far from being abstract, Gibbs–Duhem is a routine, load-bearing constraint in industrial thermodynamics.

What is the difference between the chemical potential μ and the molar Gibbs energy g?

For a single-component system there is no difference — they are equal, $\mu = g = G/N$ , a direct consequence of the Euler relation $G = \mu N$ . This is why the chemical potential is often just called "the molar Gibbs energy." The distinction only appears in mixtures. There, the chemical potential of component $i$ is the partial molar Gibbs energy, $\mu_i = (\partial G/\partial N_i)_{T,P,N_{j\ne i}}$ — the change in total Gibbs energy when you add one particle of species $i$ holding everything else fixed. The total Gibbs energy is then $G = \sum_i \mu_i N_i$ (the multicomponent Euler relation), so each $\mu_i$ is a partial contribution, not simply $G/N$ . So: for one component, $\mu$ and $g$ coincide; for many components, $\mu_i$ is the partial molar Gibbs energy of each species, and their $N_i$ -weighted sum is $G$ .