Thermodynamics

Clausius Inequality

Q: Why does Clausius's inequality follow from the second law?

Start from the Kelvin-Planck statement: no cyclic engine can convert heat entirely to work. Couple any candidate cycle with a reversible Carnot engine running between the same temperature reservoirs. If the candidate cycle violated ∮ dQ/T > 0, you could combine the two to extract net work from a single reservoir, contradicting Kelvin-Planck. The inequality is the integral form forced by this contradiction.

For any cyclic process: ∮ dQ_rev/T = 0 (reversible) or ∮ dQ/T ≤ 0 (irreversible)

The Clausius inequality (Rudolf Clausius, 1854) states that for any thermodynamic cycle, ∮ dQ/T ≤ 0, where Q is heat absorbed by the system and T is the temperature at which it is absorbed. Equality holds iff the cycle is reversible. The inequality is a precise mathematical statement of the second law of thermodynamics. Its corollary is the existence of entropy S as a state function: dS = dQ_rev/T, with ΔS ≥ ∫dQ/T for any process — entropy never decreases in an isolated system. Clausius derived the famous "Die Energie der Welt ist konstant; die Entropie der Welt strebt einem Maximum zu" (1865) — the heat death of the universe. Foundation of statistical mechanics (Boltzmann's S = k log W).

Inequality∮ dQ/T ≤ 0 (reversible: =)
AuthorClausius 1854
Defines entropydS = dQ_rev/T
Equivalent toSecond law of thermodynamics
Heat deathEntropy → max in closed universe
BoltzmannS = k log W

Interactive visualization

Press play, or step through manually. The visualization is yours to drive — try it before reading on.

Open visualization fullscreen ↗

Watch the 60-second explainer

A condensed visual walkthrough — narrated, captioned, under a minute.

Why Clausius matters

Defines entropy. Without ∮ dQ_rev/T = 0, there is no path-independent entropy function — the integral has to vanish around any reversible loop for the differential to be exact.
Engine efficiency limits. Combined with Clausius, no engine operating between hot and cold reservoirs can exceed Carnot efficiency η = 1 − T_c/T_h. Sets the absolute ceiling for power plants, refrigerators, and heat pumps.
Refrigeration coefficient of performance. COP_max = T_c / (T_h − T_c). A −20°C freezer in a 20°C room cannot beat COP = 6.3, even with perfect engineering. Determines compressor sizing.
Statistical mechanics bridge. Boltzmann's S = k ln W reconciles thermodynamic entropy (Clausius) with microscopic counting. The same S appears in chemistry, information theory (Shannon entropy), and black-hole thermodynamics (Bekenstein-Hawking).
Direction of time. The arrow of time is encoded in ΔS ≥ 0. Microscopic laws are time-symmetric; macroscopic irreversibility comes from the inequality.
Cosmology. Heat death is a direct corollary — no temperature gradients means no work, no life, no order. Sets the long-term fate of any closed universe.
Exergy and engineering. Industrial second-law analysis quantifies wasted work as T_0·ΔS_gen — every kilowatt of irreversibility is paid for in fuel.

The math

Reversible cycle. ∮ dQ_rev/T = 0 — exact differential, defines entropy S as a state function.
Any cycle. ∮ dQ/T ≤ 0 — equality iff every step is reversible. Otherwise, strict inequality.
Process form. ΔS ≥ ∫ dQ/T, equality reversible, > irreversible. For an isolated system Q = 0, so ΔS ≥ 0.
Boltzmann factor. dS = dQ_rev/T at the macroscopic level corresponds to S = k ln W microscopically; k = 1.38 × 10⁻²³ J/K.
Universe statement. ΔS_system + ΔS_surroundings ≥ 0, which is the cleanest form for chemistry and engineering.

Common misconceptions

"Applies only to closed systems." The inequality applies to any cycle of the system; heat may cross the boundary. Open and closed both count — what matters is summing dQ/T over the system's boundary.
"Irreversible means inefficient." Correlated but not identical. A real Otto cycle with 30% efficiency is irreversible and inefficient; a slow quasi-static expansion against atmospheric pressure can be nearly reversible yet produce no net work. Reversibility is about the path's reversal, not its yield.
"Heat death is imminent." Estimated at 10^100 years — 90 orders of magnitude beyond today's 1.4 × 10^10 year cosmic age. Stellar fuel runs out around 10^14 years; black holes evaporate around 10^67. Heat death is far past these.
"Entropy can decrease locally." Yes — your refrigerator does it. But the heat dumped to the room generates more entropy than is removed from the freezer. ΔS_universe is the strict inequality; only the system's slice can run backward.
"Reversible processes exist." Idealizations only. Every real process generates some entropy through finite-time effects, friction, or finite temperature gradients. The inequality is always strict in practice.
"S = 0 is a fundamental zero." Conventional. Pre-third-law thermodynamics had only ΔS measurable. Nernst (1906) and Planck (1911) fixed S = 0 for a perfect crystal at T = 0 by convention; absolute entropies follow.

History

1824 Carnot. Sadi Carnot publishes Réflexions sur la puissance motrice du feu with the cycle and reversible-engine bound. Caloric theory; no entropy yet.
1850 Clausius. Reformulates Carnot's argument in terms of mechanical heat theory. First law: Q = ΔU + W.
1854 Clausius inequality. Statement of ∮ dQ/T ≤ 0 in Annalen der Physik.
1865 Coining "entropy". Clausius proposes the name from Greek τροπή (transformation), parallel to "energy". Famous "Energie konstant; Entropie strebt einem Maximum" line.
1872 Boltzmann H-theorem. Microscopic derivation of monotonic entropy increase from kinetic theory.
1877 Boltzmann. S = k log W on his tombstone in Vienna.
1948 Shannon. Communication-theoretic entropy H = −Σ p_i log p_i — same functional form, applied to information.
1973 Bekenstein-Hawking. Black hole entropy S = A/4 (in Planck units) — area, not volume; quantum gravity inherits the inequality.

Applications

Power plants. Steam Rankine, gas Brayton, combined cycle — all bounded above by Carnot. A 1500 K combustor against a 300 K condenser caps efficiency at 80%; real plants reach 60% (combined-cycle).
Refrigerators and heat pumps. COP capped by T_c/(T_h − T_c) and T_h/(T_h − T_c) respectively. Subzero industrial freezers and ground-source heat pumps live within 30 to 60% of these limits.
Chemistry. Reaction spontaneity is determined by ΔG = ΔH − TΔS < 0 — Gibbs free energy is built from Clausius. Equilibrium constants follow Arrhenius/Eyring.
Information theory. Landauer (1961): erasing one bit at temperature T dissipates at least kT ln 2 of energy. Direct consequence of the inequality applied to computation.
Black holes. Generalized second law: ΔS_BH + ΔS_outside ≥ 0. Bekenstein bound S ≤ 2π·k·R·E/ℏc constrains information storage.
Cosmology. Entropy of the cosmic microwave background, of black holes, of dark energy de Sitter horizons — modern accounting uses Clausius's framework.

Worked example

Setup. A heat engine runs between T_h = 600 K and T_c = 300 K. Per cycle, it absorbs Q_h = 1000 J from the hot reservoir, does W = 400 J of work, and rejects Q_c = 600 J to the cold reservoir.
Carnot bound. η_Carnot = 1 − 300/600 = 0.5 = 50%. Real engine: η = W/Q_h = 0.4 = 40%. Below Carnot, as required.
Clausius check. ∮ dQ/T = Q_h/T_h − Q_c/T_c = 1000/600 − 600/300 = 1.67 − 2.0 = −0.33 J/K. Negative — consistent with the inequality.
Reversible case. A Carnot engine doing W = 500 J would reject Q_c = 500 J. Then ∮ dQ/T = 1000/600 − 500/300 = 0. Equality holds for the reversible cycle.
Entropy generated. ΔS_universe = 0.33 J/K per cycle for the irreversible engine. Multiplied by T_0 = 300 K, this is 100 J of work permanently lost — the gap between 50% Carnot and 40% real efficiency.

Frequently asked questions

Why does Clausius's inequality follow from the second law?

Start from the Kelvin-Planck statement: no cyclic engine can convert heat entirely to work. Couple any candidate cycle with a reversible Carnot engine running between the same temperature reservoirs. If the candidate cycle violated ∮ dQ/T > 0, you could combine the two to extract net work from a single reservoir, contradicting Kelvin-Planck. The inequality is the integral form forced by this contradiction.

How does it imply entropy is a state function?

For a reversible cycle, ∮ dQ_rev/T = 0. That means the integral ∫ dQ_rev/T between any two states is path-independent — exactly the criterion for an exact differential. So we define dS ≡ dQ_rev/T, and S is a state function whose change depends only on endpoints. For irreversible processes, ΔS > ∫ dQ/T, so entropy still has a well-defined change but heat divided by T underestimates it.

What is the relation to available work (exergy)?

Exergy is the maximum useful work extractable from a system as it equilibrates with its surroundings at T_0. Exergy destruction ΔE_destroyed = T_0 · ΔS_universe ≥ 0 by Clausius. Every irreversibility costs you T_0 times the entropy generated. This is the engineering form: friction, heat leaks across finite temperature differences, mixing, and unrestrained expansion all destroy exergy proportional to the entropy they generate.

How does Boltzmann's statistical formula give S = k log W?

Boltzmann (1877) connected thermodynamic entropy to microscopic counting: S = k ln W, where W is the number of microstates consistent with the macrostate. Adding two independent systems multiplies their microstate counts (W = W1·W2) but adds their entropies (S = S1 + S2) — only the logarithm satisfies this. The constant k = 1.38 × 10⁻²³ J/K matches Clausius's macroscopic dS = dQ/T to the microscopic counting.

What is a reversible vs irreversible cycle?

Reversible: an idealized cycle that proceeds through a continuous sequence of equilibrium states, infinitely slowly, with no friction, no finite-temperature heat transfer, no mixing. The Carnot cycle is the canonical example. Irreversible: any real cycle. Friction generates heat, finite temperature gradients drive heat across them with entropy generation, and any unrestrained expansion creates entropy. Real engines lose at least 30 to 50% of Carnot efficiency to irreversibilities.

How does it predict heat death of the universe?

Clausius's 1865 statement: the energy of the world is constant; the entropy of the world tends toward a maximum. Treating the universe as a closed system, ΔS_universe ≥ 0 implies entropy monotonically increases. At maximum entropy, no temperature gradients remain, no work is extractable, and all processes cease. Modern estimates place this state at perhaps 10^100 years — far beyond stellar lifetimes, dwarfing current cosmic age (1.4 × 10^10 years) by 90 orders of magnitude.