Skip to content
Miko-pedia Miko-pedia
← Back to map
Statistical Mechanics · AI note · undergrad · #statmech

Complete Thermodynamics

A concise guide to thermodynamic postulates, equilibrium, potentials, Legendre transforms, stability, phase transitions, material response, and the bridge to statistical mechanics.

· 13 min
View PDF

Part I: General Principles of Classical Thermodynamics

Core ideas

Classical thermodynamics describes macroscopic equilibrium states without tracking microscopic motion. Its central variables are energy, entropy, volume, particle number, temperature, pressure, and chemical potential. The laws constrain possible processes and define useful state functions.

For review, be able to state the laws, distinguish state functions from path quantities, and identify equilibrium variables. Keep the physical question visible: identify the degrees of freedom, the conserved quantities, the approximation being made, and the observable that would be measured.

Mathematical spine

dU=TdSPdV+μdNdU=T\,dS-P\,dV+\mu\,dN

Worked example

Consider 2mol2\,\text{mol} of a monatomic ideal gas at T=300KT=300\,\text{K}. Its internal energy is

U=32nRT=32(2)(8.314)(300)=7.48kJ.U=\frac{3}{2}nRT=\frac{3}{2}(2)(8.314)(300)=7.48\,\text{kJ}.

If the gas expands isothermally to double its volume, the entropy change is

ΔS=nRlnV2V1=(2)(8.314)ln2=11.5 J/K.\Delta S = nR\ln\frac{V_2}{V_1} = (2)(8.314)\ln 2 = 11.5\ \text{J/K}.

Because TT is fixed, ΔU=0\Delta U=0 and the heat absorbed equals the work done by the gas: Q=W=TΔS=(300)(11.5)=3.46kJQ=W=T\Delta S=(300)(11.5)=3.46\,\text{kJ}.

Problems with Solutions

Problem 1. A gas expands from V1=1.0LV_1=1.0\,\text{L} to V2=2.0LV_2=2.0\,\text{L}. Calculate the work done (a) reversibly and isothermally at T=300KT=300\,\text{K}, and (b) irreversibly against a constant external pressure Pext=1.0atmP_{\text{ext}}=1.0\,\text{atm}. Take n=0.040moln=0.040\,\text{mol}.

Solution. (a) Wrev=nRTln(V2/V1)=(0.040)(8.314)(300)ln2=69.1JW_{\text{rev}}=-nRT\ln(V_2/V_1)=-(0.040)(8.314)(300)\ln 2=-69.1\,\text{J}.
(b) With Pext=1.0atm=1.013×105PaP_{\text{ext}}=1.0\,\text{atm}=1.013\times10^5\,\text{Pa} and ΔV=1.0×103m3\Delta V=1.0\times10^{-3}\,\text{m}^3, Wirr=PextΔV=101.3JW_{\text{irr}}=-P_{\text{ext}}\Delta V=-101.3\,\text{J}. The reversible path yields the maximum work (least negative).

Problem 2. A sealed rigid container holds n=3moln=3\,\text{mol} of gas. Heat is added, raising the temperature from 300K300\,\text{K} to 400K400\,\text{K}. Given CV=52RC_V=\frac{5}{2}R for this gas, find ΔU\Delta U, QQ, and WW.

Solution. Because VV is constant, W=0W=0. Then ΔU=Q=nCVΔT=(3)(52×8.314)(100)=6.24kJ\Delta U=Q=nC_V\Delta T=(3)(\frac{5}{2}\times8.314)(100)=6.24\,\text{kJ}.

Section summary. Thermodynamics is a macroscopic theory of energy, entropy, and equilibrium.

Introduction: The Nature of Thermodynamics and the Basis of Thermostatistics

Core ideas

Thermodynamics is powerful because many microscopic details are irrelevant at equilibrium. Thermostatistics supplies the microscopic meaning: entropy measures multiplicity and temperature measures how entropy changes with energy.

For review, be able to connect thermodynamic and statistical entropy, explain why few variables describe equilibrium, and identify the thermodynamic limit. Keep the physical question visible: identify the degrees of freedom, the conserved quantities, the approximation being made, and the observable that would be measured.

Mathematical spine

S=kBlnΩ,dS0for isolated systemsS=k_B\ln\Omega,\qquad dS\ge0\quad\mathrm{for\ isolated\ systems}

Worked example

A system of N=4N=4 distinguishable particles can each occupy one of three single-particle energy levels. The total number of accessible microstates is Ω=34=81\Omega = 3^4 = 81. The statistical entropy is

S=kBlnΩ=(1.38×1023J/K)ln81=6.3×1023J/K.S = k_B\ln\Omega = (1.38\times10^{-23}\,\text{J/K})\ln 81 = 6.3\times10^{-23}\,\text{J/K}.

If a constraint is removed and the number of accessible microstates doubles to Ω=162\Omega'=162, the entropy increases by ΔS=kBln29.6×1024J/K\Delta S = k_B\ln 2 \approx 9.6\times10^{-24}\,\text{J/K}, illustrating the second law at the microscopic level.

Problems with Solutions

Problem 1. A fair coin is flipped 1010 times. How many microstates correspond to exactly 55 heads and 55 tails? What is the entropy of this macrostate?

Solution. The multiplicity is Ω=(105)=252\Omega = \binom{10}{5} = 252. The entropy is S=kBln2527.83kB1.08×1022J/KS = k_B\ln 252 \approx 7.83\,k_B \approx 1.08\times10^{-22}\,\text{J/K}.

Problem 2. An isolated Einstein solid has 33 oscillators and 44 energy quanta. Using Ω=(N+q1q)\Omega = \binom{N+q-1}{q}, compute the entropy.

Solution. Here N=3N=3, q=4q=4, so Ω=(64)=15\Omega=\binom{6}{4}=15. Then S=kBln152.71kB3.74×1023J/KS=k_B\ln 15\approx 2.71\,k_B\approx 3.74\times10^{-23}\,\text{J/K}.

Problem 3. Connect the microscopic and macroscopic pictures: for an ideal gas, show that S=kBlnΩS=k_B\ln\Omega is consistent with ΔS=nRln(Vf/Vi)\Delta S = nR\ln(V_f/V_i) when the volume doubles at fixed energy.

Solution. When VV doubles, each particle has twice as many position states, so Ωf/Ωi=2N\Omega_f/\Omega_i = 2^N. Then ΔS=kBln(2N)=NkBln2=nRln2\Delta S = k_B\ln(2^N)=Nk_B\ln 2 = nR\ln 2, matching the thermodynamic result.

Section summary. Thermodynamics summarizes microscopic complexity through state variables.

The Problem and the Postulates

Core ideas

The basic problem is to determine equilibrium states and allowed processes from a small set of postulates. A simple system has an entropy function S(U,V,N,)S(U,V,N,\ldots) that is extensive, concave, and maximized at equilibrium for isolated constraints.

For review, be able to use entropy maximization, extensivity, and concavity to infer equilibrium conditions. Keep the physical question visible: identify the degrees of freedom, the conserved quantities, the approximation being made, and the observable that would be measured.

Mathematical spine

S=S(U,V,N),S(λU,λV,λN)=λS(U,V,N)S=S(U,V,N),\qquad S(\lambda U,\lambda V,\lambda N)=\lambda S(U,V,N)

Worked example

Two identical copper blocks, each with heat capacity C=500J/KC=500\,\text{J/K}, are initially at T1=300KT_1=300\,\text{K} and T2=400KT_2=400\,\text{K}. They are brought into thermal contact in an isolated enclosure. The total energy is Utot=C(T1+T2)=3.50×105JU_{\text{tot}}=C(T_1+T_2)=3.50\times10^5\,\text{J}. Entropy maximization gives the final temperature Tf=(T1+T2)/2=350KT_f=(T_1+T_2)/2=350\,\text{K}. The entropy increase is

ΔS=ClnTf2T1T2=500ln3502300×400=500ln(1.0208)=10.3J/K.\Delta S = C\ln\frac{T_f^2}{T_1T_2} = 500\ln\frac{350^2}{300\times400}=500\ln(1.0208)=10.3\,\text{J/K}.

Problems with Solutions

Problem 1. Show that S(U,V,N)=NkBln(aU3V2/N5)S(U,V,N)=Nk_B\ln(aU^3V^2/N^5) is extensive, i.e., S(λU,λV,λN)=λS(U,V,N)S(\lambda U,\lambda V,\lambda N)=\lambda S(U,V,N).

Solution. Substitute:

S(λU,λV,λN)=λNkBln ⁣(a(λU)3(λV)2(λN)5)=λNkBln ⁣(aλ5U3V2λ5N5)=λS(U,V,N).S(\lambda U,\lambda V,\lambda N) = \lambda N k_B\ln\!\left(\frac{a(\lambda U)^3(\lambda V)^2}{(\lambda N)^5}\right) = \lambda N k_B\ln\!\left(\frac{a\lambda^5 U^3V^2}{\lambda^5 N^5}\right) = \lambda S(U,V,N).

Problem 2. Two systems have entropies S1=cU1V1S_1=c\sqrt{U_1V_1} and S2=cU2V2S_2=c\sqrt{U_2V_2} with equal volumes V1=V2=VV_1=V_2=V. They exchange energy with U1+U2=UU_1+U_2=U fixed. Find the equilibrium energy distribution.

Solution. Maximize S=cV1/2(U1+UU1)S=cV^{1/2}(\sqrt{U_1}+\sqrt{U-U_1}). Setting dS/dU1=0dS/dU_1=0 gives U11/2=(UU1)1/2U_1^{-1/2}=(U-U_1)^{-1/2}, so U1=U2=U/2U_1=U_2=U/2.

Section summary. The entropy postulate defines equilibrium and thermodynamic stability.

The Conditions of Equilibrium

Core ideas

Two systems in contact exchange energy, volume, or particles until entropy is maximized. This produces equality of temperature, pressure, and chemical potential for the corresponding allowed exchanges.

For review, be able to derive thermal, mechanical, and diffusive equilibrium conditions from entropy maximization. Keep the physical question visible: identify the degrees of freedom, the conserved quantities, the approximation being made, and the observable that would be measured.

Mathematical spine

T1=T2,P1=P2,μ1=μ2T_1=T_2,\qquad P_1=P_2,\qquad \mu_1=\mu_2

Worked example

A rigid insulated tank is divided by a conducting wall. Side A contains 1mol1\,\text{mol} of monatomic ideal gas at TA=400KT_A=400\,\text{K}; side B contains 2mol2\,\text{mol} of the same gas at TB=250KT_B=250\,\text{K}. The wall allows heat but not particles or volume exchange. Thermal equilibrium requires TA=TB=TfT_A=T_B=T_f. Energy conservation gives

nACVTA+nBCVTB=(nA+nB)CVTf    Tf=(1)(400)+(2)(250)3=300K.n_A C_V T_A + n_B C_V T_B = (n_A+n_B)C_V T_f \;\Rightarrow\; T_f = \frac{(1)(400)+(2)(250)}{3}=300\,\text{K}.

Pressures need not equalize because the wall is rigid.

Problems with Solutions

Problem 1. Two ideal gases are separated by a movable, thermally conducting piston. Initially P1=2atmP_1=2\,\text{atm}, V1=1LV_1=1\,\text{L}, T1=300KT_1=300\,\text{K} and P2=1atmP_2=1\,\text{atm}, V2=2LV_2=2\,\text{L}, T2=300KT_2=300\,\text{K}. Find the equilibrium pressure and temperature.

Solution. Moles are n1=P1V1/RT1=0.081moln_1=P_1V_1/RT_1=0.081\,\text{mol} and n2=0.081moln_2=0.081\,\text{mol}. Since TT is already equal and the piston conducts, Tf=300KT_f=300\,\text{K}. Mechanical equilibrium gives a common PfP_f. Total volume V=3LV=3\,\text{L}, so Pf=(n1+n2)RTf/V=2.03atmP_f=(n_1+n_2)RT_f/V=2.03\,\text{atm}.

Problem 2. Explain why μ1=μ2\mu_1=\mu_2 is the condition for diffusive equilibrium, using entropy maximization with N1+N2N_1+N_2 fixed.

Solution. dS=(S1/N1)dN1+(S2/N2)dN2=(μ1/T+μ2/T)dN1=0dS = (\partial S_1/\partial N_1)dN_1 + (\partial S_2/\partial N_2)dN_2 = (-\mu_1/T+\mu_2/T)dN_1=0, hence μ1=μ2\mu_1=\mu_2.

Section summary. Equilibrium means no entropy gain remains under allowed exchanges.

Some Formal Relationships, and Sample Systems

Core ideas

The fundamental thermodynamic relation generates equations of state and identities. Ideal gases, paramagnets, and elastic systems illustrate how different work terms enter and how measurable response follows from derivatives.

For review, be able to use Euler and Gibbs-Duhem relations, compute simple equations of state, and track conjugate pairs. Keep the physical question visible: identify the degrees of freedom, the conserved quantities, the approximation being made, and the observable that would be measured.

Mathematical spine

U=TSPV+μN,SdTVdP+Ndμ=0U=TS-PV+\mu N,\qquad S\,dT-V\,dP+N\,d\mu=0

Worked example

For 1mol1\,\text{mol} of monatomic ideal gas, the fundamental relation is U=32RTU=\frac{3}{2}RT and the entropy is S=R[ln(V/NvQ3)+3/2]S = R[\ln(V/Nv_Q^3)+3/2], where vQv_Q is the quantum volume. The Euler equation states U=TSPV+μNU=TS-PV+\mu N. Using PV=RTPV=RT and the explicit forms of UU and SS, one finds the chemical potential μ=RTln(kBT/PvQ3)\mu = -RT\ln(k_BT/Pv_Q^3), demonstrating consistency of the formal structure.

Problems with Solutions

Problem 1. Starting from U=TSPV+μNU=TS-PV+\mu N, derive the Gibbs-Duhem relation SdTVdP+Ndμ=0S\,dT-V\,dP+N\,d\mu=0.

Solution. Take the differential: dU=TdS+SdTPdVVdP+μdN+NdμdU=T\,dS+S\,dT-P\,dV-V\,dP+\mu\,dN+N\,d\mu. Compare with dU=TdSPdV+μdNdU=T\,dS-P\,dV+\mu\,dN. Subtracting yields SdTVdP+Ndμ=0S\,dT-V\,dP+N\,d\mu=0.

Problem 2. A system has fundamental relation S=aUVNS=a\sqrt{UVN} with a>0a>0. Find its equation of state P(T,V,N)P(T,V,N).

Solution. Use 1/T=(S/U)V,N=aVN/(2U)1/T=(\partial S/\partial U)_{V,N}=a\sqrt{VN}/(2\sqrt{U}) and P/T=(S/V)U,N=aUN/(2V)P/T=(\partial S/\partial V)_{U,N}=a\sqrt{UN}/(2\sqrt{V}). Eliminating UU gives U=a2T2VN/4U=a^2T^2VN/4, and therefore P=U/V=a2NT2/4P=U/V=a^2NT^2/4.

Section summary. Formal identities reduce many derivatives to a consistent structure.

Reversible Processes and the Maximum Work Theorem

Core ideas

A reversible process proceeds through equilibrium states and produces no entropy internally. It gives the maximum useful work between specified states; irreversible processes lose availability through entropy production.

For review, be able to compute reversible work, distinguish heat and work, and connect free energy decreases to maximum work. Keep the physical question visible: identify the degrees of freedom, the conserved quantities, the approximation being made, and the observable that would be measured.

Mathematical spine

dS=δQrevT,Wmax=ΔF(T,V,N fixed)dS=\frac{\delta Q_{\rm rev}}{T},\qquad W_{\rm max}=-\Delta F\quad(T,V,N\ \mathrm{fixed})

Worked example

One mole of ideal gas expands isothermally at T=300KT=300\,\text{K} from V1=10LV_1=10\,\text{L} to V2=20LV_2=20\,\text{L}. Because TT is fixed, the Helmholtz free energy change is

ΔF=nRTlnV2V1=(1)(8.314)(300)ln2=1.73kJ.\Delta F = -nRT\ln\frac{V_2}{V_1} = -(1)(8.314)(300)\ln 2 = -1.73\,\text{kJ}.

The maximum non-expansion work obtainable from this change is Wmax=ΔF=1.73kJW_{\text{max}}=-\Delta F=1.73\,\text{kJ}. If the expansion is performed irreversibly against a constant external pressure Pext=P2=1.24atmP_{\text{ext}}=P_2=1.24\,\text{atm}, the actual work extracted is only

W=Pext(V2V1)=(1.24×105Pa)(10×103m3)=1.24kJ,|W| = P_{\text{ext}}(V_2-V_1) = (1.24\times10^5\,\text{Pa})(10\times10^{-3}\,\text{m}^3) = 1.24\,\text{kJ},

which is less than the reversible maximum.

Problems with Solutions

Problem 1. Calculate the efficiency of a Carnot engine operating between TH=600KT_H=600\,\text{K} and TC=300KT_C=300\,\text{K}.

Solution. η=1TC/TH=1300/600=0.50\eta = 1 - T_C/T_H = 1 - 300/600 = 0.50, or 50%50\%.

Problem 2. A gas with γ=5/3\gamma=5/3 undergoes adiabatic reversible expansion from (P1=2atm,V1=1L)(P_1=2\,\text{atm},V_1=1\,\text{L}) to V2=2LV_2=2\,\text{L}. Find the final pressure and temperature, assuming n=1moln=1\,\text{mol}.

Solution. For a reversible adiabat, P1V1γ=P2V2γP_1V_1^\gamma=P_2V_2^\gamma, so P2=P1(V1/V2)γ=2(1/2)5/3=0.794atmP_2=P_1(V_1/V_2)^\gamma=2(1/2)^{5/3}=0.794\,\text{atm}. The initial temperature is T1=P1V1/nR=(2×1.013×105)(103)/8.314=24.4KT_1=P_1V_1/nR=(2\times1.013\times10^5)(10^{-3})/8.314=24.4\,\text{K}. Then T2=T1(V1/V2)γ1=24.4(1/2)2/3=15.4KT_2=T_1(V_1/V_2)^{\gamma-1}=24.4(1/2)^{2/3}=15.4\,\text{K}.

Problem 3. Show that for any isothermal process, the work done BY the system is maximized when the process is reversible.

Solution. From the first and second laws, dU=δQδWdU=\delta Q-\delta W and dSδQ/TdS\ge\delta Q/T. For isothermal, dU=0dU=0 so δW=δQTdS\delta W=\delta Q\le T\,dS. Thus WTΔS=ΔF+ΔU=ΔFW\le T\Delta S = -\Delta F + \Delta U = -\Delta F (since ΔU=0\Delta U=0). The upper bound is attained only when equality holds, i.e., reversibly.

Section summary. Reversibility sets ideal bounds on work and efficiency.

Alternative Formulations and Legendre Transformations

Core ideas

Different experiments control different variables, so different potentials are useful. Legendre transforms replace inconvenient extensive variables by their conjugate intensive variables, producing FF, HH, and GG.

For review, be able to choose the correct thermodynamic potential and natural variables, and transform between potentials. Keep the physical question visible: identify the degrees of freedom, the conserved quantities, the approximation being made, and the observable that would be measured.

Mathematical spine

F=UTS,H=U+PV,G=UTS+PVF=U-TS,\qquad H=U+PV,\qquad G=U-TS+PV

Worked example

For a simple system with U(S)=αS2U(S)=\alpha S^2 at fixed volume, the temperature is T=(U/S)V=2αST=(\partial U/\partial S)_V=2\alpha S. Solving for SS gives S=T/(2α)S=T/(2\alpha). The Helmholtz free energy is the Legendre transform

F(T)=UTS=α(T2α)2T(T2α)=T24αT22α=T24α.F(T) = U - TS = \alpha\left(\frac{T}{2\alpha}\right)^2 - T\left(\frac{T}{2\alpha}\right) = \frac{T^2}{4\alpha} - \frac{T^2}{2\alpha} = -\frac{T^2}{4\alpha}.

This expresses the same physics in the natural variables (T)(T) rather than (S)(S).

Problems with Solutions

Problem 1. From the Euler equation U=TSPV+μNU=TS-PV+\mu N, show that G=μNG=\mu N.

Solution. By definition G=UTS+PVG=U-TS+PV. Substituting the Euler relation gives G=(TSPV+μN)TS+PV=μNG=(TS-PV+\mu N)-TS+PV=\mu N.

Problem 2. Given U(S,V)=aS4/V2U(S,V)=aS^4/V^2, find F(T,V)F(T,V).

Solution. T=(U/S)V=4aS3/V2T=(\partial U/\partial S)_V=4aS^3/V^2, so S=(TV2/4a)1/3S=(TV^2/4a)^{1/3}. Then

F=UTS=aS4V2TS=34(T4V24a)1/3.F=U-TS=\frac{aS^4}{V^2}-TS=-\frac{3}{4}\left(\frac{T^4V^2}{4a}\right)^{1/3}.

Problem 3. Why are the natural variables of FF temperature and volume?

Solution. The differential is dF=SdTPdV+μdNdF = -S\,dT - P\,dV + \mu\,dN. Because dFdF is expressed in terms of dTdT, dVdV, and dNdN, these are the natural variables, meaning all other thermodynamic quantities can be obtained as derivatives with respect to them.

Section summary. Thermodynamic potentials are adapted to controlled variables.

The Extremum Principle in the Legendre Transformed Representations

Core ideas

Equilibrium can be found by minimizing the appropriate potential: UU at fixed entropy, FF at fixed temperature and volume, GG at fixed temperature and pressure. These criteria are equivalent forms of the entropy maximum principle.

For review, be able to apply minimum principles to common constraints and interpret metastability. Keep the physical question visible: identify the degrees of freedom, the conserved quantities, the approximation being made, and the observable that would be measured.

Mathematical spine

(T,V,N):F minimum,(T,P,N):G minimum(T,V,N):\quad F\ \mathrm{minimum},\qquad (T,P,N):\quad G\ \mathrm{minimum}

Worked example

A substance at T=300KT=300\,\text{K} and P=1atmP=1\,\text{atm} can exist as a liquid or a vapor. At this point, Gliq=10.0kJ/molG_{\text{liq}}=-10.0\,\text{kJ/mol} and Gvap=9.5kJ/molG_{\text{vap}}=-9.5\,\text{kJ/mol}. Because GG is minimized at fixed TT and PP, the liquid is the stable phase. If a small fluctuation created a vapor bubble, ΔG=GvapGliq=+0.5kJ/mol>0\Delta G = G_{\text{vap}}-G_{\text{liq}}=+0.5\,\text{kJ/mol}>0, so the bubble would collapse.

Problems with Solutions

Problem 1. Show that at fixed TT and VV, the Helmholtz free energy FF is minimized at equilibrium.

Solution. For a system in contact with a thermal reservoir at TT, the total entropy change is ΔStot=ΔSsys+ΔSres\Delta S_{\text{tot}}=\Delta S_{\text{sys}}+\Delta S_{\text{res}}. The reservoir gives heat TΔSres=ΔUsys-T\Delta S_{\text{res}}=\Delta U_{\text{sys}} (if only heat is exchanged). Thus ΔStot=ΔSsysΔUsys/T=ΔFsys/T\Delta S_{\text{tot}}=\Delta S_{\text{sys}}-\Delta U_{\text{sys}}/T=-\Delta F_{\text{sys}}/T. Since ΔStot0\Delta S_{\text{tot}}\ge 0, we have ΔFsys0\Delta F_{\text{sys}}\le 0, so FF decreases and is minimized.

Problem 2. A system has F(T,V)=NkBTln(V/V0)+f(T)F(T,V)=-Nk_BT\ln(V/V_0)+f(T). Find the pressure and show that at fixed TT, FF is minimized when the system adopts the volume consistent with the external pressure.

Solution. P=(F/V)T=NkBT/VP=-(\partial F/\partial V)_T=Nk_BT/V. Consider a fluctuation δV\delta V at fixed TT. Then δF=(NkBT/V)δV+12(NkBT/V2)(δV)2+\delta F = -(Nk_BT/V)\delta V + \frac{1}{2}(Nk_BT/V^2)(\delta V)^2+\cdots. The first-order term vanishes at the equilibrium volume where Pext=NkBT/VP_{\text{ext}}=Nk_BT/V. The second derivative 2F/V2=NkBT/V2>0\partial^2F/\partial V^2=Nk_BT/V^2>0 confirms a minimum.

Section summary. Equilibrium is an extremum principle in the right variables.

Stability of Thermodynamic Systems

Core ideas

Stable systems resist small fluctuations. Concavity of entropy and convexity of potentials imply positive heat capacities, compressibilities, and susceptibilities. Instability signals phase separation or a transition.

For review, be able to use second-derivative tests, identify stable response signs, and interpret spinodal regions. Keep the physical question visible: identify the degrees of freedom, the conserved quantities, the approximation being made, and the observable that would be measured.

Mathematical spine

CV>0,κT=1V(VP)T>0C_V>0,\qquad \kappa_T=-\frac1V\left(\frac{\partial V}{\partial P}\right)_T>0

Worked example

For an ideal gas, CV=32nRC_V=\frac{3}{2}nR and the isothermal compressibility is κT=1/P\kappa_T=1/P. Both are strictly positive for all equilibrium states. Suppose a hypothetical gas had P=nRT/(Vnb)an2/V2P=nRT/(V-nb)-an^2/V^2 (van der Waals). Then

(PV)T=nRT(Vnb)2+2an2V3.\left(\frac{\partial P}{\partial V}\right)_T = -\frac{nRT}{(V-nb)^2}+\frac{2an^2}{V^3}.

The stability condition κT1=V(P/V)T>0\kappa_T^{-1}=-V(\partial P/\partial V)_T>0 becomes

nRTV(Vnb)2>2an2V2.\frac{nRTV}{(V-nb)^2} > \frac{2an^2}{V^2}.

For temperatures below the critical temperature Tc=8a/(27Rb)T_c=8a/(27Rb), there is a region of VV where this inequality fails. This unstable region lies between the spinodal points and corresponds to the unphysical part of the van der Waals isotherm.

Problems with Solutions

Problem 1. Show that CP>CVC_P>C_V for any stable system.

Solution. The general identity is CPCV=TVα2/κTC_P-C_V=TV\alpha^2/\kappa_T. Stability requires T>0T>0, V>0V>0, κT>0\kappa_T>0, and α20\alpha^2\ge0. Therefore CPCV0C_P-C_V\ge0, with equality only when α=0\alpha=0.

Problem 2. A fluid has CV=aT3VC_V=aT^3V and κT=b/T\kappa_T=b/T with a,b>0a,b>0. Are the stability conditions satisfied at all temperatures?

Solution. Yes: CV>0C_V>0 for T>0T>0, and κT>0\kappa_T>0 for T>0T>0. The system is stable.

Problem 3. For a van der Waals gas with a=0.365Pam6/mol2a=0.365\,\text{Pa\,m}^6\text{/mol}^2 and b=4.29×105m3/molb=4.29\times10^{-5}\,\text{m}^3\text{/mol}, estimate the critical temperature.

Solution. Tc=8a/(27Rb)=(8×0.365)/(27×8.314×4.29×105)304KT_c=8a/(27Rb)=(8\times0.365)/(27\times8.314\times4.29\times10^{-5})\approx 304\,\text{K}.

Section summary. Stability is encoded in the curvature of thermodynamic potentials.

First-Order Phase Transitions

Core ideas

First-order transitions occur when two phases have equal thermodynamic potential but different entropy or volume. Latent heat, coexistence curves, metastability, nucleation, and Maxwell constructions describe their behavior.

For review, be able to use equality of chemical potentials, apply Clapeyron equation, and interpret latent heat. Keep the physical question visible: identify the degrees of freedom, the conserved quantities, the approximation being made, and the observable that would be measured.

Mathematical spine

μα(T,P)=μβ(T,P),dPdT=ΔSΔV=LTΔV\mu_\alpha(T,P)=\mu_\beta(T,P),\qquad \frac{dP}{dT}=\frac{\Delta S}{\Delta V}=\frac{L}{T\Delta V}

Worked example

For the ice—water transition at 0C0^\circ\text{C}, the molar enthalpy of fusion is ΔHfus=6.01kJ/mol\Delta H_{\text{fus}}=6.01\,\text{kJ/mol} and the volume change is ΔV=VliqVice=1.6×106m3/mol\Delta V=V_{\text{liq}}-V_{\text{ice}}=-1.6\times10^{-6}\,\text{m}^3\text{/mol}. The Clapeyron equation gives the slope of the coexistence curve:

dPdT=ΔHfusTΔV=6010(273.15)(1.6×106)13.7MPa/K.\frac{dP}{dT} = \frac{\Delta H_{\text{fus}}}{T\Delta V} = \frac{6010}{(273.15)(-1.6\times10^{-6})} \approx -13.7\,\text{MPa/K}.

The negative slope means increasing pressure lowers the melting point, explaining why ice skates melt a thin layer of ice.

Problems with Solutions

Problem 1. Water boils at 100C100^\circ\text{C} with ΔHvap=40.7kJ/mol\Delta H_{\text{vap}}=40.7\,\text{kJ/mol}. Treating the vapor as ideal and neglecting the liquid volume, estimate dP/dTdP/dT at the boiling point.

Solution. ΔVVgas=RT/P=(8.314)(373.15)/(1.013×105)0.0306m3/mol\Delta V\approx V_{\text{gas}}=RT/P=(8.314)(373.15)/(1.013\times10^5)\approx 0.0306\,\text{m}^3\text{/mol}. Then

dPdT=40700(373.15)(0.0306)3.56×103Pa/K=0.035atm/K.\frac{dP}{dT}=\frac{40700}{(373.15)(0.0306)}\approx 3.56\times10^3\,\text{Pa/K}=0.035\,\text{atm/K}.

Problem 2. Two phases α\alpha and β\beta coexist at (T0,P0)(T_0,P_0). If μα(T,P)=μβ(T,P)\mu_\alpha(T,P)=\mu_\beta(T,P) defines the coexistence curve, explain why crossing it discontinuously changes SS and VV.

Solution. Along the curve dμα=dμβd\mu_\alpha=d\mu_\beta. Using dμ=SdT+VdPd\mu=-S\,dT+V\,dP (per mole), the slope is dP/dT=ΔS/ΔVdP/dT=\Delta S/\Delta V. A first-order transition has ΔS0\Delta S\neq0 and ΔV0\Delta V\neq0, so SS and VV jump discontinuously across the curve.

Problem 3. Sketch a Maxwell construction on a van der Waals PP-VV isotherm below TcT_c and explain its physical meaning.

Solution. The oscillating part of the isotherm violates κT>0\kappa_T>0. The Maxwell construction replaces it with a horizontal line at PeqP_{\text{eq}} such that the areas above and below the line are equal. This ensures μα=μβ\mu_\alpha=\mu_\beta and describes liquid-vapor coexistence.

Section summary. First-order transitions are phase coexistence with discontinuous first derivatives.

Properties of Materials

Core ideas

Material properties are thermodynamic derivatives: heat capacities, compressibilities, expansion coefficients, susceptibilities, and elastic moduli. Maxwell relations connect derivatives that may be easier to measure.

For review, be able to derive response functions, use Maxwell relations, and relate microscopic behavior to macroscopic coefficients. Keep the physical question visible: identify the degrees of freedom, the conserved quantities, the approximation being made, and the observable that would be measured.

Mathematical spine

(SV)T=(PT)V,CPCV=TVα2κT\left(\frac{\partial S}{\partial V}\right)_T=\left(\frac{\partial P}{\partial T}\right)_V,\qquad C_P-C_V=TV\frac{\alpha^2}{\kappa_T}

Worked example

For one mole of ideal gas, the Maxwell relation (S/V)T=(P/T)V(\partial S/\partial V)_T=(\partial P/\partial T)_V gives

(SV)T=(TnRTV)V=nRV.\left(\frac{\partial S}{\partial V}\right)_T = \left(\frac{\partial}{\partial T}\frac{nRT}{V}\right)_V = \frac{nR}{V}.

Integrating at fixed TT yields S(T,V)=nRlnV+f(T)S(T,V)=nR\ln V+f(T), where f(T)f(T) is an integration function. This confirms the logarithmic volume dependence of the entropy of an ideal gas without needing statistical mechanics.

Problems with Solutions

Problem 1. Derive CPCV=TVα2/κTC_P-C_V=TV\alpha^2/\kappa_T for an ideal gas and verify that CPCV=nRC_P-C_V=nR.

Solution. For an ideal gas, α=1/T\alpha=1/T and κT=1/P\kappa_T=1/P. Thus

CPCV=TV(1/T)21/P=PVT=nR.C_P-C_V = TV\frac{(1/T)^2}{1/P} = \frac{PV}{T} = nR.

Problem 2. Mercury at 300K300\,\text{K} has α=1.81×104K1\alpha=1.81\times10^{-4}\,\text{K}^{-1}, κT=3.82×1011Pa1\kappa_T=3.82\times10^{-11}\,\text{Pa}^{-1}, and Vm=1.48×105m3/molV_m=1.48\times10^{-5}\,\text{m}^3\text{/mol}. Estimate CP,mCV,mC_{P,m}-C_{V,m}.

Solution. Using CPCV=TVα2/κTC_P-C_V=TV\alpha^2/\kappa_T:

CP,mCV,m=(300)(1.48×105)(1.81×104)23.82×10113.81J/molK.C_{P,m}-C_{V,m} = \frac{(300)(1.48\times10^{-5})(1.81\times10^{-4})^2}{3.82\times10^{-11}} \approx 3.81\,\text{J/mol$\cdot$K}.

Problem 3. Use the Maxwell relation (T/P)S=(V/S)P(\partial T/\partial P)_S=(\partial V/\partial S)_P together with CP=T(S/T)PC_P=T(\partial S/\partial T)_P to show that for an adiabatic process (T/P)S=TαV/CP(\partial T/\partial P)_S=T\alpha V/C_P.

Solution. Start from (T/P)S=(V/S)P(\partial T/\partial P)_S = (\partial V/\partial S)_P. Write

(VS)P=(VT)P(TS)P=VαTCP=TVαCP.\left(\frac{\partial V}{\partial S}\right)_P = \left(\frac{\partial V}{\partial T}\right)_P\left(\frac{\partial T}{\partial S}\right)_P = V\alpha \cdot \frac{T}{C_P} = \frac{TV\alpha}{C_P}.

Section summary. Response functions summarize how materials react to changed conditions.

Part II: Statistical Mechanics

Core ideas

Statistical mechanics explains where entropy, temperature, equations of state, and fluctuations come from. It also extends thermodynamics to finite systems, fluctuations, and microscopic models.

For review, be able to connect partition functions to free energies and explain why thermodynamic laws emerge statistically. Keep the physical question visible: identify the degrees of freedom, the conserved quantities, the approximation being made, and the observable that would be measured.

Mathematical spine

Z=seβEs,F=kBTlnZZ=\sum_s e^{-\beta E_s},\qquad F=-k_BT\ln Z

Worked example

A quantum system has two energy levels, E1=0E_1=0 and E2=ϵ=0.100eVE_2=\epsilon=0.100\,\text{eV}. At T=300KT=300\,\text{K}, kBT0.02585eVk_BT\approx0.02585\,\text{eV}. The partition function is

Z=eβE1+eβE2=1+e0.100/0.02585=1+e3.8681.0209.Z = e^{-\beta E_1}+e^{-\beta E_2} = 1+e^{-0.100/0.02585} = 1+e^{-3.868} \approx 1.0209.

The probability of occupying the excited state is

p2=eβϵZ0.02091.02090.0205.p_2 = \frac{e^{-\beta\epsilon}}{Z} \approx \frac{0.0209}{1.0209} \approx 0.0205.

The average energy is U=p2ϵ2.05×103eVU=p_2\epsilon\approx 2.05\times10^{-3}\,\text{eV}, and the Helmholtz free energy is

F=kBTlnZ=(0.02585)ln(1.0209)5.36×104eV.F = -k_BT\ln Z = -(0.02585)\ln(1.0209) \approx -5.36\times10^{-4}\,\text{eV}.

Problems with Solutions

Problem 1. A three-level system has energies 00, ϵ\epsilon, and 2ϵ2\epsilon. Find the partition function ZZ, the average energy UU, and the heat capacity CVC_V as functions of TT.

Solution. Z=1+eβϵ+e2βϵZ=1+e^{-\beta\epsilon}+e^{-2\beta\epsilon}.

U=lnZβ=ϵ(eβϵ+2e2βϵ)1+eβϵ+e2βϵ.U = -\frac{\partial\ln Z}{\partial\beta} = \frac{\epsilon(e^{-\beta\epsilon}+2e^{-2\beta\epsilon})}{1+e^{-\beta\epsilon}+e^{-2\beta\epsilon}}.

Then CV=(U/T)V=kBβ2(E2E2)C_V=(\partial U/\partial T)_V = k_B\beta^2(\langle E^2\rangle-\langle E\rangle^2), where E2=(ϵ2eβϵ+4ϵ2e2βϵ)/Z\langle E^2\rangle=(\epsilon^2e^{-\beta\epsilon}+4\epsilon^2e^{-2\beta\epsilon})/Z.

Problem 2. NN independent two-level systems (each with energies 00 and ϵ\epsilon) form a solid. Find the total partition function and free energy.

Solution. For one system, Z1=1+eβϵZ_1=1+e^{-\beta\epsilon}. Because the systems are independent and distinguishable (localized), Z=(Z1)N=(1+eβϵ)NZ=(Z_1)^N=(1+e^{-\beta\epsilon})^N. The free energy is

F=kBTlnZ=NkBTln(1+eϵ/kBT).F = -k_BT\ln Z = -Nk_BT\ln(1+e^{-\epsilon/k_BT}).

Problem 3. Show that the statistical entropy S=(F/T)VS=-(\partial F/\partial T)_V yields the familiar expression S=kB(lnZ+βU)S=k_B(\ln Z+\beta U).

Solution. F=kBTlnZF=-k_BT\ln Z, so

S=(FT)V=kBlnZ+kBT(lnZT)V.S = -\left(\frac{\partial F}{\partial T}\right)_V = k_B\ln Z + k_BT\left(\frac{\partial\ln Z}{\partial T}\right)_V.

Since β=1/kBT\beta=1/k_BT, lnZ/T=(lnZ/β)(dβ/dT)=(U)(1/kBT2)=U/kBT2\partial\ln Z/\partial T = (\partial\ln Z/\partial\beta)(d\beta/dT)=(-U)(-1/k_BT^2)=U/k_BT^2. Therefore

S=kBlnZ+kBTUkBT2=kB(lnZ+βU).S = k_B\ln Z + k_BT\cdot\frac{U}{k_BT^2} = k_B(\ln Z+\beta U).

Section summary. Statistical mechanics is the microscopic foundation of thermodynamics.