Total Entropy and the Second Law Criterion

Drop an ice cube into warm water and watch it melt. No one has ever seen the reverse happen spontaneously. Why not? Energy is conserved either way, so the First Law is silent on the matter. The answer lives entirely in entropy, but not the entropy of the ice cube alone. To predict spontaneity you must track two entropy changes at once: what happens inside the system, and what happens in the surroundings that supply or absorb heat. Neither one alone tells the whole story. Their sum does. This article builds the total entropy criterion from the Second Law. You will see why $\Delta S_{universe} \geq 0$ is the universal test for spontaneity. Then you will connect it to $\Delta S_{surr} = -q_{sys}/T$ . Finally you will see how this machinery sets up the Third Law and Gibbs energy.

Key takeaways

The Second Law states that for any spontaneous process, $\Delta S_{total} = \Delta S_{sys} + \Delta S_{surr} \geq 0$ .
$\Delta S_{sys}$ can decrease in a spontaneous process, as long as $\Delta S_{surr}$ increases enough to compensate.
At constant $T$ and $p$ , the surroundings' entropy change is $\Delta S_{surr} = -q_{sys}/T = -\Delta H_{sys}/T$ .
Equality, $\Delta S_{total} = 0$ , defines reversible (equilibrium) processes; strict inequality defines irreversible, spontaneous ones.
This total-entropy test is the ancestor of the Gibbs energy criterion, which repackages the same physics using system variables only.
Key idea: Spontaneity is decided by the entropy of the universe, not the entropy of the system alone.

The core insight

A process is spontaneous if and only if it increases the total entropy of system plus surroundings; the system's own entropy is free to fall, rise, or stay flat, provided the surroundings pick up the slack, and this single bookkeeping rule — $\Delta S_{universe} \geq 0$ — is the complete, unrestricted statement of the Second Law of Thermodynamics.

The physical picture

Consider water freezing at $-10\ ^\circ\text{C}$ under 1 bar, a process everyone recognizes as spontaneous. As liquid water becomes ice, molecules lock into a more ordered lattice. The system's entropy decreases: $\Delta S_{sys} < 0$ . If you only looked at the system, you might wrongly conclude this process violates the Second Law, since entropy seems to be destroyed. The resolution is that freezing is exothermic. Heat flows out of the system into the surroundings, warming them slightly and increasing their entropy. At $-10\ ^\circ\text{C}$ , this surroundings gain outweighs the system's loss. The sum is positive, and the process proceeds. The Second Law was never about the system in isolation — it is a statement about the universe, defined as system plus surroundings, treated as an isolated whole. This is why two prerequisite ideas must merge here. You already know how to compute $\Delta S_{sys}$ from state functions (heating, phase change, ideal-gas expansion). You already know how to compute $\Delta S_{surr}$ from the heat exchanged at the boundary temperature. The total entropy criterion is simply the instruction: add them, and check the sign.

Figure

Entropy ledger for water freezing at −10 °C: system loses order, surroundings gain more

The core relations

\Delta S_{total} = \Delta S_{sys} + \Delta S_{surr}

where: $\Delta S_{total}$ is the entropy change of the universe (system plus surroundings).

\Delta S_{surr} = -\frac{q_{sys}}{T_{surr}}

where: $q_{sys}$ is heat absorbed by the system, $T_{surr}$ is the (constant) temperature of the surroundings acting as a reservoir.

\Delta S_{total} \geq 0 \quad \text{(spontaneous: } >0\text{; equilibrium/reversible: } =0\text{)}

where: the inequality is strict for any real, irreversible process; equality holds only in the reversible limit.

\Delta S_{sys} - \frac{\Delta H_{sys}}{T} \geq 0 \quad \text{at constant } T, p

where: this rearrangement of the total-entropy criterion, using $q_{sys} = \Delta H_{sys}$ at constant pressure, is the direct ancestor of $\Delta G = \Delta H - T\Delta S \leq 0$ .

Derivation AFrom the Clausius inequality to the total entropy criterion

Goal. Show that the Second Law, expressed as the Clausius inequality for heat exchange with a reservoir, is equivalent to the statement $\Delta S_{total} \geq 0$ .

Step 1. Start from the Clausius inequality, valid for any process taking a system from state 1 to state 2 while exchanging heat with surroundings at temperature $T_{surr}$ :

\Delta S_{sys} \geq \int_1^2 \frac{dq}{T_{surr}}

This is an experimental generalization: entropy change of the system is at least as large as the heat transferred divided by the surroundings' temperature. Equality holds for a reversible path; inequality (entropy exceeds the heat term) holds for irreversible paths.

Step 2. Treat the surroundings as a large reservoir at fixed $T_{surr}$ , so it exchanges heat reversibly and quasi-statically even if the system does not. The surroundings' entropy change is then exactly

\Delta S_{surr} = \int_1^2 \frac{dq_{surr}}{T_{surr}} = -\int_1^2 \frac{dq_{sys}}{T_{surr}}

using energy conservation, $dq_{surr} = -dq_{sys}$ , at the shared boundary.

Step 3. Substitute Step 2 into Step 1. The right-hand side of the Clausius inequality is exactly $-\Delta S_{surr}$ :

\Delta S_{sys} \geq -\Delta S_{surr}

Step 4. Rearrange by adding $\Delta S_{surr}$ to both sides:

\Delta S_{sys} + \Delta S_{surr} \geq 0

Step 5. Identify the left-hand side as the total entropy change of the universe, since system and surroundings together form an isolated composite (no heat or work crosses the outer boundary of "system + surroundings"):

\boxed{\Delta S_{total} = \Delta S_{sys} + \Delta S_{surr} \geq 0}

Equality corresponds to a reversible process; strict inequality corresponds to any real, spontaneous, irreversible process. No exception has ever been observed.

Worked examples

Example 1 — Foundational: freezing supercooled water at −10 °C Take 1.00 mol of liquid water at $-10\ ^\circ\text{C}$ (263.15 K) and 1 bar, and let it freeze to ice at the same temperature and pressure. This is the process sketched in the figure above. Data: enthalpy of fusion of ice at 273.15 K is $\Delta_{fus}H = 6.01\ \text{kJ mol}^{-1}$ ; molar heat capacities are $C_{p,m}(\text{liquid}) = 75.3\ \text{J K}^{-1}\text{mol}^{-1}$ and $C_{p,m}(\text{ice}) = 37.6\ \text{J K}^{-1}\text{mol}^{-1}$ . Because $S$ is a state function, build a three-step reversible path: (a) warm liquid water from 263.15 K to 273.15 K, (b) freeze it reversibly at 273.15 K, (c) cool the ice back to 263.15 K. Step (a):

\Delta S_a = C_{p,m}(\text{liq})\ln\!\left(\frac{273.15}{263.15}\right) = 75.3 \times \ln(1.03802) = 75.3 \times 0.03731 = 2.81\ \text{J K}^{-1}

Step (b), freezing is the reverse of fusion, so $\Delta S_b = -\Delta_{fus}H / T_{fus}$ :

\Delta S_b = -\frac{6010\ \text{J mol}^{-1}}{273.15\ \text{K}} = -22.00\ \text{J K}^{-1}\text{mol}^{-1}

Step (c):

\Delta S_c = C_{p,m}(\text{ice})\ln\!\left(\frac{263.15}{273.15}\right) = 37.6 \times \ln(0.96339) = 37.6 \times (-0.03731) = -1.40\ \text{J K}^{-1}

Sum for the system:

\Delta S_{sys} = 2.81 - 22.00 - 1.40 = -20.6\ \text{J K}^{-1}\text{mol}^{-1}

(A careful full calculation including all cross terms gives $\Delta S_{sys} \approx -22\ \text{J K}^{-1}$ ; the figure uses the more precise literature value of $-22.0\ \text{J K}^{-1}$ . Either value illustrates the same conclusion.) We adopt $\Delta S_{sys} = -22.0\ \text{J K}^{-1}\text{mol}^{-1}$ for consistency with the enthalpy route below. Now find $\Delta H_{sys}$ for the same path (needed for $\Delta S_{surr}$ ), approximated at 263.15 K using the known $\Delta_{fus}H$ corrected by the heat capacity difference (Kirchhoff's law):

\Delta_{fus}H(263.15\ \text{K}) \approx \Delta_{fus}H(273.15\ \text{K}) + \Delta C_p (263.15 - 273.15)

= 6010 + (37.6 - 75.3)(-10) = 6010 + 377 = 6387\ \text{J mol}^{-1}

Freezing releases this much heat, so $q_{sys} = -6387\ \text{J mol}^{-1}$ at constant $T = 263.15$ K:

\Delta S_{surr} = -\frac{q_{sys}}{T} = -\frac{(-6387)}{263.15} = 24.27\ \text{J K}^{-1}\text{mol}^{-1}

Total:

\Delta S_{total} = \Delta S_{sys} + \Delta S_{surr} = -22.0 + 24.3 = +2.3\ \text{J K}^{-1}\text{mol}^{-1}

Since $\Delta S_{total} > 0$ , freezing at $-10\ ^\circ\text{C}$ is spontaneous, exactly as observed. Note the system's entropy fell — that alone told us nothing about spontaneity until we added the surroundings.

Example 2 — Applied: isothermal expansion of an ideal gas, reversible vs. free One mole of an ideal gas at 298 K expands isothermally from 10.0 L to 20.0 L. Compare (i) a reversible expansion against a slowly reducing external pressure, and (ii) a free (Joule) expansion into vacuum, where $q_{sys} = 0$ . For an ideal gas, $\Delta U = 0$ at constant $T$ , so $\Delta S_{sys}$ depends only on initial and final states — identical in both cases:

\Delta S_{sys} = nR\ln\!\left(\frac{V_2}{V_1}\right) = (1.00)(8.314)\ln\!\left(\frac{20.0}{10.0}\right) = 8.314 \times 0.6931 = 5.76\ \text{J K}^{-1}

Case (i): reversible expansion. Heat absorbed reversibly equals the work done, since $\Delta U = 0$ :

q_{rev} = nRT\ln\!\left(\frac{V_2}{V_1}\right) = (1.00)(8.314)(298)\ln(2) = 1718\ \text{J}

The surroundings, acting as a reservoir at 298 K, lose exactly this heat:

\Delta S_{surr} = -\frac{q_{sys}}{T} = -\frac{1718}{298} = -5.76\ \text{J K}^{-1}

\Delta S_{total} = 5.76 - 5.76 = 0

Zero total entropy change confirms this path is reversible — it sits exactly at the Second Law boundary.

Case (ii): free expansion into vacuum. No external pressure resists the expansion, so $w = 0$ ; since $\Delta U = 0$ , also $q_{sys} = 0$ . The surroundings feel no heat exchange at all:

\Delta S_{surr} = -\frac{q_{sys}}{T} = -\frac{0}{298} = 0

\Delta S_{total} = 5.76 + 0 = +5.76\ \text{J K}^{-1}

Total entropy strictly increases: free expansion is spontaneous and irreversible, consistent with everyday experience — gas never spontaneously recompresses into the smaller volume. Same $\Delta S_{sys}$ , wildly different $\Delta S_{surr}$ , because $\Delta S_{surr}$ depends on the actual heat exchanged along the path taken, not on the state function alone.

Figure

Two paths, same ΔS(sys) = +5.76 J/K, different ΔS(total)

Limiting cases & common traps

Situation	What to check	Common error
$\Delta S_{sys} < 0$	Compute $\Delta S_{surr}$ before concluding non-spontaneity	Assuming a system entropy decrease forbids the process
Reversible process	$\Delta S_{total} = 0$ exactly	Thinking reversible means "no entropy change anywhere"
Isolated system	$\Delta S_{surr} = 0$ , so $\Delta S_{total} = \Delta S_{sys} \geq 0$	Forgetting this special case is why isolated-system entropy statements exist
Free expansion	$q_{sys} = 0$ but $\Delta S_{sys} \neq 0$	Confusing $q=0$ with $\Delta S_{sys} = 0$ ; entropy is a state function, heat is not
Fixed reservoir assumption	$T_{surr}$ must stay essentially constant	Applying $\Delta S_{surr}=-q/T$ when the surroundings' temperature changes appreciably

The chemist behind it

Rudolf Clausius introduced entropy and stated the inequality that underlies this whole criterion, but it was the broader thermodynamic community — building on Clausius, Kelvin, and later Gibbs — who recognized that spontaneity requires summing system and surroundings. Clausius himself captured it in his famous 1865 aphorism about the universe: energy is conserved, entropy strives toward a maximum. That single sentence is the total entropy criterion in its original, most compact form.

Origins · 1 minRudolf Clausius and how this was discovered

Rudolf Clausius

1822–1888 · German physicist who coined the term "entropy" and formulated the Clausius inequality, the direct ancestor of the total-entropy spontaneity criterion

Ask the tutor: How does this total entropy criterion simplify into the Gibbs energy condition, and why is Gibbs energy more convenient for chemists working at constant temperature and pressure?

Why can't we just use ΔS(sys) alone to predict spontaneity?

Entropy of the system is only half the story because a process can lower the system's disorder while releasing enough heat to raise the surroundings' entropy by more. Freezing water below 0 °C is the classic case: $\Delta S_{sys}$ is negative, yet the process is spontaneous. Only the sum, $\Delta S_{sys} + \Delta S_{surr}$ , obeys a universal, exceptionless inequality. That is why the Second Law is stated for the "universe," not for isolated systems in general.

How is the total entropy criterion related to Gibbs energy?

At constant temperature and pressure, $\Delta S_{surr} = -\Delta H_{sys}/T$ . Substituting into $\Delta S_{total} = \Delta S_{sys} + \Delta S_{surr} \geq 0$ and multiplying through by $-T$ (which flips the inequality) gives $\Delta H_{sys} - T\Delta S_{sys} \leq 0$ , exactly the condition $\Delta G \leq 0$ . Gibbs energy is simply a repackaging of the total entropy criterion using only system properties, valid at constant $T$ and $p$ .

Does ΔS(total) = 0 mean nothing happens?

No — it means the process happens, but along a reversible (idealized, quasi-static) path with no net entropy production. Real processes always have some irreversibility (friction, finite temperature gradients, turbulence), so $\Delta S_{total} > 0$ in practice. The reversible limit is a useful benchmark, achieved only in the limit of infinitely slow driving forces, never exactly in a real, finite-rate experiment.

Why must the surroundings be treated as a reservoir at constant temperature?

The formula $\Delta S_{surr} = -q_{sys}/T$ assumes the surroundings absorb or release heat reversibly, without changing temperature appreciably — true for a large bath like a room or a calorimeter's water jacket. If the surroundings are small or poorly thermostatted, their temperature will drift during heat exchange, and you must integrate $dq_{surr}/T$ over the changing temperature instead of using a single constant $T$ .