Entropy and spontaneity

Why does a hot cup of tea always cool down and never spontaneously boil by pulling heat back out of the cooler room? Why does a gas released into a vacuum spread out and never voluntarily crowd back into one corner? Both are one-way streets, and the sign on the street is entropy. Entropy, SS, is the thermodynamic quantity that tracks how many microscopic arrangements are consistent with what we see macroscopically, and the Second Law says that for an isolated system it can only increase or stay the same — never decrease. That single statistical fact is the deepest reason chemical and physical processes have a preferred direction. In CY11001 you meet entropy from two angles at once: a statistical picture built on counting microstates, and a classical picture built on the Clausius inequality, ΔS≥∫δq/T\Delta S \ge \int \delta q/T. Together they let you calculate ΔS\Delta S for gas expansions, heating, phase changes and mixing, and to judge spontaneity by comparing system and surroundings.

The core insight

Entropy counts accessible microstates (S=kBln⁡WS = k_B\ln W), and the Clausius inequality ΔSuniv≥0\Delta S_{univ}\ge 0 turns that counting fact into a quantitative test for whether a process can happen spontaneously.

The core idea

Two languages, one quantity. Statistically, entropy measures the number of microscopic arrangements (WW, the microstates) consistent with a given macroscopic state. A spontaneous process moves the system toward macrostates reachable by overwhelmingly more microstates — that is the microscopic reason isolated systems drift toward higher entropy.

S=kBln⁡WS = k_B \ln W

where: SS is entropy, kBk_B is Boltzmann's constant, and WW is the number of microstates. CY11001 uses this as a conceptual anchor, not a full partition-function derivation — that machinery is out of scope here.

The classical, testable form. Classically, entropy is defined through heat exchanged along a reversible path, and the Second Law is stated as the Clausius inequality. For any process taking a system from state 1 to state 2,

ΔS≥∫12δqT\Delta S \ge \int_1^2 \frac{\delta q}{T}

where: δq\delta q is heat exchanged along the actual path and TT is the temperature at which it is exchanged; equality holds for a reversible path, strict inequality for an irreversible one. For an isolated system, δq=0\delta q = 0 throughout, so this collapses to ΔS≥0\Delta S \ge 0: the entropy of an isolated system never decreases. Because SS is a state function, ΔS\Delta S between two fixed states does not depend on which path you took to get there — you may always compute it along a convenient reversible path even when the real process was irreversible, then evaluate ∫δqrev/T\int \delta q_{rev}/T along that imagined path.

Reading spontaneity correctly. A process is spontaneous only when the universe's entropy rises, ΔSuniv=ΔSsys+ΔSsurr≥0\Delta S_{univ} = \Delta S_{sys} + \Delta S_{surr} \ge 0 — not when ΔSsys\Delta S_{sys} alone is positive. Water freezing below 0 degrees Celsius has ΔSsys<0\Delta S_{sys} < 0, yet it is spontaneous because the surroundings gain more entropy than the system loses.

The standard toolbox (ideal gas, nn moles). These formulas cover essentially every ΔS\Delta S calculation you will do in this unit:

  • Isothermal expansion or compression: ΔS=nRln⁡V2V1=−nRln⁡P2P1\Delta S = nR\ln\dfrac{V_2}{V_1} = -nR\ln\dfrac{P_2}{P_1}
  • Heating at constant volume: ΔS=nCV,mln⁡T2T1\Delta S = nC_{V,m}\ln\dfrac{T_2}{T_1}
  • Heating at constant pressure: ΔS=nCP,mln⁡T2T1\Delta S = nC_{P,m}\ln\dfrac{T_2}{T_1}
  • General process, both TT and VV changing: ΔS=nCV,mln⁡T2T1+nRln⁡V2V1\Delta S = nC_{V,m}\ln\dfrac{T_2}{T_1} + nR\ln\dfrac{V_2}{V_1}
  • Phase transition at constant T,PT,P (reversible, since the two phases coexist in equilibrium): ΔStrs=ΔHtrsTtrs\Delta S_{trs} = \dfrac{\Delta H_{trs}}{T_{trs}}
  • Mixing of ideal gases at constant T,PT,P: ΔSmix=−nR∑ixiln⁡xi\Delta S_{mix} = -nR\sum_i x_i \ln x_i, always positive since each mole fraction 0<xi<10<x_i<1
Figure
A closed system undergoing an irreversible change compared with the surroundings, showing entropy of the system and surroundings added to test the sign of the universe total
Spontaneity test: add system and surroundings entropy system delta S sys can be negative e.g. water freezing surroundings delta S surr large reservoir absorbs or releases heat add delta S universe = delta S sys plus delta S surr spontaneous only if this sum is greater than or equal to zero
Derivation ASurroundings entropy from heat flow at constant temperature

Show that when a system exchanges heat qsysq_{sys} with a large surrounding reservoir held at fixed temperature TT, the resulting entropy change of the surroundings is simply ΔSsurr=−qsys/T\Delta S_{surr} = -q_{sys}/T, so that a full spontaneity check needs only ΔSsys\Delta S_{sys} (from the state-function formulas) plus this one surroundings term, without tracking any messy irreversibility inside the system itself.

Step 1. Treat the surroundings as a reservoir so large that whatever heat it absorbs or releases changes its temperature negligibly — it stays at essentially fixed TT throughout the process, no matter how violently irreversible the process is inside the system.

Step 2. Because the surroundings' temperature stays constant, any heat exchange with it can be modeled as happening along a reversible path at that fixed TT, even if the system itself is changing irreversibly. This is legitimate because entropy change depends only on the path taken by whichever body you are computing it for — here, the reservoir's own path is reversible by virtue of its size.

Step 3. Apply the reversible-path definition of entropy change directly to the surroundings: ΔSsurr=∫δqsurr/T=qsurr/T\Delta S_{surr} = \int \delta q_{surr}/T = q_{surr}/T since TT is constant and can be pulled out of the integral.

Step 4. Use energy conservation between system and surroundings: whatever heat flows into the system leaves the surroundings, so qsurr=−qsysq_{surr} = -q_{sys}. Substituting gives ΔSsurr=−qsys/T\Delta S_{surr} = -q_{sys}/T.

Step 5. Combine with ΔSsys\Delta S_{sys}, computed along a convenient reversible path between the same two states (using the standard toolbox formulas), to get ΔSuniv=ΔSsys−qsys/T≥0\Delta S_{univ} = \Delta S_{sys} - q_{sys}/T \ge 0 as the operational spontaneity test — this is exactly the Clausius inequality applied to the system's actual (possibly irreversible) heat exchange with a constant-temperature reservoir.

Worked example

One mole of an ideal gas (CV,m=52RC_{V,m} = \tfrac{5}{2}R) expands irreversibly against a constant external pressure of 1 atm from (T1=300 K,V1=10 L)(T_1=300\text{ K}, V_1=10\text{ L}) to (T2=300 K,V2=20 L)(T_2=300\text{ K}, V_2=20\text{ L}). Find ΔSsys\Delta S_{sys}, ΔSsurr\Delta S_{surr}, and confirm the process is spontaneous.

Step 1. Since SS is a state function, compute ΔSsys\Delta S_{sys} along a reversible isothermal path between the same two end states, ignoring the actual irreversible path entirely:

ΔSsys=nRln⁡V2V1=(1)(8.314)ln⁡2010=8.314×0.6931=5.76 J K−1\Delta S_{sys} = nR\ln\frac{V_2}{V_1} = (1)(8.314)\ln\frac{20}{10} = 8.314 \times 0.6931 = 5.76\ \text{J K}^{-1}

Step 2. Find the actual heat exchanged along the real irreversible path. Work done against the constant external pressure is w=−Pext(V2−V1)=−(101325 Pa)(20−10)×10−3 m3=−1013.25 Jw = -P_{ext}(V_2-V_1) = -(101325\ \text{Pa})(20-10)\times10^{-3}\ \text{m}^3 = -1013.25\ \text{J}, meaning the system does +1013.25 J+1013.25\ \text{J} of work on the surroundings. Since ΔT=0\Delta T = 0 for an ideal gas, ΔU=0\Delta U = 0, so qsys=−w=+1013.25 Jq_{sys} = -w = +1013.25\ \text{J} flows into the system from the surroundings.

Step 3. Apply the derivation above: the surroundings lose 1013.25 J1013.25\ \text{J} at constant T=300 KT = 300\ \text{K}, so

ΔSsurr=−qsysT=−1013.25300=−3.38 J K−1\Delta S_{surr} = \frac{-q_{sys}}{T} = \frac{-1013.25}{300} = -3.38\ \text{J K}^{-1}

Step 4. Add the two contributions: ΔSuniv=ΔSsys+ΔSsurr=5.76−3.38=2.38 J K−1>0\Delta S_{univ} = \Delta S_{sys} + \Delta S_{surr} = 5.76 - 3.38 = \mathbf{2.38\ J\ K^{-1} > 0}. Since ΔSuniv>0\Delta S_{univ} > 0, the Clausius inequality is satisfied and the expansion is confirmed spontaneous.

Competitive stretch

If the same expansion were carried out reversibly instead (quasi-static, Pext≈PgasP_{ext}\approx P_{gas} at every instant), ΔSsys\Delta S_{sys} is unchanged at 5.76 J K−15.76\ \text{J K}^{-1} because it is a state function, but now ΔSsurr=−5.76 J K−1\Delta S_{surr} = -5.76\ \text{J K}^{-1} exactly, giving ΔSuniv=0\Delta S_{univ}=0 — the reversible limit, the boundary case between spontaneous and non-spontaneous.

Common traps

  • Computing ΔSsys\Delta S_{sys} from the actual irreversible heat, writing ΔS=qirrev/T\Delta S = q_{irrev}/T — wrong, because q/Tq/T only equals ΔS\Delta S along a reversible path.
  • Declaring a process non-spontaneous just because ΔSsys<0\Delta S_{sys} < 0 — spontaneity needs ΔSuniv≥0\Delta S_{univ}\ge 0, and the surroundings term can easily outweigh a negative system term.
  • Dropping the sign in ΔS=−nRln⁡(P2/P1)\Delta S = -nR\ln(P_2/P_1) — expansion means V2>V1V_2>V_1 but P2<P1P_2<P_1, so the pressure form of the same formula carries a minus sign relative to the volume form.
  • Using ΔSmix=−nR∑xiln⁡xi\Delta S_{mix}=-nR\sum x_i\ln x_i outside its ideal-gas assumptions — it does not apply once there is an enthalpy of mixing or a reaction between the components.

Ask the tutor: If the reversible-path trick for computing ΔSsys\Delta S_{sys} from an irreversible process still feels like a cheat, ask the AI tutor to re-walk the free-expansion worked example line by line and explain why it is mathematically legitimate.

If entropy is just about counting microstates, why do we bother with the heat-based Clausius definition at all

Because counting microstates directly is usually impossible for a real chemical system — you cannot enumerate every microscopic arrangement of a mole of gas molecules by hand. The heat-based definition, dS≥δq/TdS \ge \delta q/T, gives you a practical, measurable route to the same quantity using temperature and heat flow, which you can obtain from calorimetry or from state-function formulas. The two pictures describe the same SS; one is conceptual, the other is what you actually calculate with.

Why can I use a reversible path formula even when the real process was irreversible

Because entropy is a state function: ΔS\Delta S between two fixed states depends only on those states, not on the route taken to connect them. So you are free to imagine any convenient reversible path linking the same initial and final states purely as a calculation device, evaluate ∫δqrev/T\int \delta q_{rev}/T along that imagined path, and get the correct ΔSsys\Delta S_{sys} — even though the gas never actually followed that path. This trick only works for the system itself; you must separately track the surroundings using the actual heat exchanged.

Does a negative delta S sys ever mean a process cannot happen

No, not by itself. A negative ΔSsys\Delta S_{sys} only tells you the system's own disorder decreased; it says nothing about the surroundings. The only sure test for spontaneity is ΔSuniv=ΔSsys+ΔSsurr≥0\Delta S_{univ} = \Delta S_{sys} + \Delta S_{surr} \ge 0. Many familiar spontaneous processes, like water freezing or a gas being compressed by a piston connected to a falling weight, have ΔSsys<0\Delta S_{sys} < 0 while ΔSsurr\Delta S_{surr} more than compensates.

Before this

Where this leads