Entropy and the Second Law: Irreversibility

Core Idea

The second law can be stated as: dS_universe ≥ 0. For any real (irreversible) process, entropy of the universe increases: ΔS_total = ΔS_system + ΔS_surroundings > 0. This quantifies irreversibility; reversible processes are the limiting case where entropy generation is zero.

Explainer

From your work on entropy definition and calculation, you know that entropy measures the number of accessible microstates: S = kB ln Ω. You also know the second law as a statement about the direction of spontaneous change. Here we sharpen both into a single quantitative framework for irreversibility. The key insight is that the second law is not just a qualitative arrow — it is a precise inequality with a calculable surplus.

Consider a gas freely expanding into a vacuum (Joule expansion). The gas does no work (nothing to push against) and exchanges no heat (insulated container), so by the first law, internal energy is unchanged: ΔU = 0. Classical thermodynamics might seem silent here — no Q, no W. Yet we know this process is irreversible: the gas never spontaneously contracts. The entropy calculation resolves this immediately. The gas spreads into a larger volume, increasing the number of accessible microstates. ΔS_system = nR ln(V₂/V₁) > 0. The surroundings are untouched, so ΔS_surroundings = 0. Therefore ΔS_total > 0 — the second law correctly identifies this as irreversible and tells you exactly how irreversible it is.

The entropy generation σ = ΔS_total = ΔS_system + ΔS_surroundings is the key object. For a reversible process (like a quasi-static isothermal expansion), dS_system = δQ_rev/T and the heat transferred to the surroundings is −δQ_rev, so dS_surroundings = −δQ_rev/T, and σ = 0. For an irreversible process, less work is extracted (or more heat is dumped), meaning the surroundings gain more entropy than the system loses (or the system gains more than the surroundings lose). The surplus is σ > 0. You can think of σ as measuring the "waste" — the useful work that could have been extracted from a reversible process but wasn't. This is why irreversibility has a thermodynamic cost: every real process dissipates free energy at a rate proportional to σ.

The broader significance is that the second law gives time its direction. The microscopic laws of physics (Newton's equations, Schrödinger's equation) are time-symmetric: they look the same run forwards and backwards. Yet macroscopic processes have a definite arrow. The resolution is statistical: the time-reversed process (gas spontaneously contracting) is not forbidden by the laws of motion — it is merely overwhelmingly improbable, because the number of states with the gas expanded vastly outnumbers the states with it contracted. Entropy increasing is not a law imposed on top of mechanics; it is what happens with overwhelming probability when a system with very many degrees of freedom evolves from a low-entropy initial condition. This statistical understanding, due to Boltzmann, is one of the deepest insights in all of physics — and it sets the stage for the statistical mechanics perspective you will develop in the next courses.