Entropy — What It Actually Measures
Almost every student first meets entropy as "a measure of disorder". That sentence is easy to remember and it will get you through a one-mark definition, but it is not what entropy measures, and it will actively mislead you the first time a real problem tests it. This article gives the honest version — and it is not harder, only more precise.
The two definitions that matter
Entropy has one statistical definition and one thermodynamic definition, and they describe the same quantity from two directions.
W is the number of microstates — the number of distinct ways the particles and their energy can be arranged while the measurable properties of the sample stay the same. kB is the Boltzmann constant, 1.380649 × 10⁻²³ J K⁻¹ (an exactly defined value in the SI since 2019). Multiply by the Avogadro constant and you get the gas constant, R = NAkB ≈ 8.314 J K⁻¹ mol⁻¹.
Read that second one carefully: entropy change is heat divided by temperature, using the heat for a reversible path. Units are J K⁻¹, or J K⁻¹ mol⁻¹ for a molar quantity. Enthalpies are quoted in kJ and entropies in J — mixing the two is the commonest arithmetic error in this chapter.
So what is entropy really measuring?
Entropy measures how many ways the system's energy can be spread out. Two things increase W: more space for the particles (expand a gas and each molecule has more positions available), and more closely spaced energy levels to populate (heat the sample, or move to a phase or molecule with more vibrational and rotational levels).
This is why the entropy of a substance rises in the order solid < liquid < gas, why bigger and floppier molecules have higher standard entropies than small rigid ones, and why a reaction that increases the number of moles of gas almost always has a positive ΔS.
Why "disorder" is a bad word for it
"Disorder" is a visual, human judgement. Entropy is a count. The two come apart in cases you will actually be examined on:
- Water freezing below 0 °C is spontaneous, and the ice is far more ordered than the liquid. "Disorder increases" gives you no way to explain this. The energy accounting does — see the worked example below.
- The hydrophobic effect. Oil droplets in water merge into one blob, which looks tidier. They do so because the water molecules that had been forced into an ordered cage around each droplet are released — the total entropy rises while the picture looks more ordered.
If your syllabus defines entropy as disorder, write that, then add: "more precisely, a measure of the dispersal of energy over the accessible microstates."
Worked example 1 — entropy of a phase change
At the transition temperature the two phases are in equilibrium, so the change is reversible and the formula is simply:
Melting ice. ΔHfus = 6.01 kJ mol⁻¹ at 273.15 K.
ΔSfus = 6010 J mol⁻¹ ÷ 273.15 K = +22.0 J K⁻¹ mol⁻¹
Boiling water. ΔHvap = 40.7 kJ mol⁻¹ at 373.15 K.
ΔSvap = 40700 ÷ 373.15 = +109.1 J K⁻¹ mol⁻¹
Boiling produces about five times the entropy change of melting: melting only loosens a lattice, whereas boiling releases molecules into a far larger volume.
A useful cross-check. Trouton's rule says most normal liquids have ΔSvap ≈ 85–88 J K⁻¹ mol⁻¹. Benzene obeys it well: 30800 ÷ 353.25 = 87.2 J K⁻¹ mol⁻¹. Water's 109 is far above the rule because extensive hydrogen bonding makes liquid water unusually ordered to begin with, so breaking it up gains more entropy.
Worked example 2 — heating and expansion
Isothermal expansion of an ideal gas.
ΔS = nR ln(V₂ ÷ V₁). Take 1 mol expanding from 1.00 L to 10.0 L:
ΔS = 1 × 8.314 × ln(10) = 8.314 × 2.3026 = +19.14 J K⁻¹
Heating a liquid. ΔS = nCp ln(T₂ ÷ T₁). For 1 mol of water
(Cp ≈ 75.3 J K⁻¹ mol⁻¹) warmed from 298.15 K to 348.15 K:
ΔS = 75.3 × ln(1.1677) = 75.3 × 0.15504 = +11.67 J K⁻¹
Cross-check against the microstate picture. Double the volume of 1 mol of gas and each molecule has twice as many positions, so W grows by 2N for N molecules. Then ΔS = kB ln(2N) = N kB ln 2 = R ln 2 = 8.314 × 0.6931 = +5.76 J K⁻¹ mol⁻¹ — and nR ln(V₂/V₁) with V₂/V₁ = 2 gives the same 5.76 J K⁻¹. Two different routes, one answer: that is why "number of microstates" and "heat over temperature" really are the same quantity.
Worked example 3 — ΔS for a reaction
Because of the third law (a perfect crystal at 0 K has S = 0), entropy has an absolute scale, unlike enthalpy. So tables list absolute standard entropies S°, not "entropies of formation", and you use them directly:
N₂(g) + 3H₂(g) → 2NH₃(g) at 298 K, using S° = 191.61 (N₂), 130.68 (H₂), 192.77 (NH₃) J K⁻¹ mol⁻¹:
Products: 2 × 192.77 = 385.54
Reactants: 191.61 + (3 × 130.68) = 191.61 + 392.04 = 583.65
ΔS° = 385.54 − 583.65 = −198.1 J K⁻¹ mol⁻¹
Sanity check the sign before you trust the arithmetic: 4 moles of gas become 2 moles of gas, so the gas is being confined and ΔS must be negative. It is. Standard entropy values differ by a unit or so between data tables, so quote the source you used.
Worked example 4 — why ice forms, when ΔS of the system is negative
The second law does not say the system's entropy must rise. It says the entropy of the universe must rise:
Freezing water: ΔH = −6.01 kJ mol⁻¹ and ΔSsys ≈ −22.0 J K⁻¹ mol⁻¹ (we treat both as roughly constant with temperature, which is an approximation).
At −10 °C (263.15 K): ΔSsurr = +6010 ÷ 263.15 = +22.84
ΔSuniverse = −22.00 + 22.84 = +0.84 J K⁻¹ mol⁻¹ → spontaneous.
At +10 °C (283.15 K): ΔSsurr = +6010 ÷ 283.15 = +21.23
ΔSuniverse = −22.00 + 21.23 = −0.78 J K⁻¹ mol⁻¹ → not spontaneous
(so ice melts instead).
The system becomes more ordered either way. What changes is how much entropy the released heat buys in the surroundings — and colder surroundings buy more per joule, because ΔS = q/T.
Multiplying ΔSuniverse by −T gives ΔH − TΔS = ΔG. So ΔG < 0 and ΔSuniverse > 0 are the same statement, written from the system's point of view: Gibbs energy is the second law made convenient.
Common mistakes that cost marks
- Mixing kJ and J. ΔH is in kJ mol⁻¹, S° in J K⁻¹ mol⁻¹. Convert before you subtract, every single time.
- Using ΔH/T at the wrong temperature. For a phase change it is valid only at the transition temperature, where the change is reversible.
- Forgetting the surroundings. An exothermic reaction with negative ΔSsys can still be spontaneous. Always ask about the universe.
- Using entropies of formation. Tables give absolute S°, so do not subtract "elements in standard states" the way you do for ΔHf°. Elements have non-zero standard entropy.
- Assuming ΔS is always positive for a spontaneous change. The freezing example above is the counter-example examiners love.
Where this appears in exams
| Exam | Typical question |
|---|---|
| CBSE/ICSE Class 11–12 | Define entropy; predict the sign of ΔS; ΔS of fusion and vaporisation |
| JEE/NEET | ΔS for expansion and heating; spontaneity from ΔG = ΔH − TΔS |
| IIT-JAM / CUET-PG | Entropy of mixing; Trouton's rule and its exceptions; third-law entropies |
| GATE / CSIR-NET | Statistical thermodynamics, partition functions, residual entropy |
Entropy is usually only half a question — the mark comes from combining it with enthalpy to decide spontaneity. The Gibbs Free Energy calculator takes ΔH, ΔS and T, handles the kJ/J conversion, and returns ΔG with its sign, so you can check your hand working and find the temperature at which ΔG changes sign.
Open the Gibbs Free Energy (ΔG) Calculator →Thermodynamics rewards students who understand the reasoning rather than memorise the signs. ABC Chemistry runs Class 11–12 chemistry coaching at its Gurugram centre and online classes across India — details at abcchemistry.in.