You are on page 1of 8

School of Aerospace Engineering

Statistical Thermo. and TD Properties
• Still need to find β and way to connect Statistical
Mechanics to Thermodynamics
– e.g., Q, Ni/N → p, T, S, etc.
• Following Boltzmann… will look for analogy between
entropy and randomness (disorder)
• Approach
– examine some closed systems
– motivate Boltzmann’s relation

AE/ME 6765
Statistical Thermodynamic Rleations-1

Copyright © 2009 by Jerry M. Seitzman. All rights reserved.

School of Aerospace Engineering

Entropy and Volume Change
• Perfect gas in insulated rigid box 1 2
divided into two equal cells by
removable partition
– start with all gas in cell #1
– remove partition, end with gas in both cells
• Gibbs for process
pV = nR T = NkT
dE in stat mech
1 p 1 0 dV entropy increased with more
dS = dU + dV = dE + Nk possible locations of finding
T T T V particles → increased disorder
dV V
= Nk ΔS = Nk ln 2 = Nk ln 2 = k ln 2 N
V V1
AE/ME 6765
Statistical Thermodynamic Rleations-2

Copyright © 2009 by Jerry M. Seitzman. All rights reserved.

1

constant for • General form required S = a ln Ω + bN given system see V&K p.e. 2 . Seitzman. All rights reserved. AE/ME 6765 Statistical Thermodynamic Rleations-4 Copyright © 2009 by Jerry M.. School of Aerospace Engineering Entropy and Microstates • Examples show there is relation between entropy and number of “places” particles can be – i. Seitzman. AE/ME 6765 Statistical Thermodynamic Rleations-3 Copyright © 2009 by Jerry M. 114 S = k ln Ω + S 0 Boltzmann’s const. between S and Ω. or S=Φ(Ω) • What should function Φ look like? – look at how systems combine • Entropy is extensive – S1+2=S1+S2 S1 S2 Ω1` Ω2` • Number of microstates are multiplicative – Ω1+2=Ω1×Ω2 absolute const. School of Aerospace Engineering Entropy and Energy Change • Now fix volume and double energy Q (and temperature) – through heat transfer • Gibbs for process 0 1 dV dS = dE + Nk T V dT c dE entropy increased with more = Cv = Nk v possible (available) energy T R E levels → increased disorder cv cv E2 cv N ΔS = Nk ln = Nk ln 2 = k ln 2 R R E1 R cv=const. All rights reserved.

V 1T p T −μ T AE/ME 6765 Statistical Thermodynamic Rleations-6 Copyright © 2009 by Jerry M. All rights reserved. based on postulate: equal a priori probability AE/ME 6765 Statistical Thermodynamic Rleations-5 Copyright © 2009 by Jerry M. 3 . All rights reserved. Seitzman. School of Aerospace Engineering Boltzmann’s Relation S = k ln Ω + S 0 • Analagous to 3rd Law. we know we can get Ω ln Ω ≅ ln Wmax from most probable macrostate ⎛ Q⎞ • Using Boltzmann limit ln Wmax = N ⎜1 + ln ⎟ + β E ⎡ ⎛ Q⎞ ⎤ ⎝ N⎠ S = k ⎢ N ⎜1 + ln ⎟ + β E ⎥ Q ≡ ∑ g i e − βε i ⎣ ⎝ N⎠ ⎦ i Q S = kN ln + kN + kβ E N • Can now get β by comparing to TD state relation (Gibb’s) ∂S ∂S ∂S dS = dE + dV + dN ∂E V . let S=0 for perfectly ordered system (Ω=1) – S0=0 Boltzmann’s • So S = k ln Ω Relation – our approach was not rigorous derivation – we used perfect gas – but true in general – also. N ∂V E . Seitzman. School of Aerospace Engineering Thermodynamic Relations • For large N. N ∂N E .

4 . N ⎥⎦ Q E β = 1 kT S = kN ln + kN + N T ∴ S = S (N . Seitzman. Mech. N • ∂E ∂S ⎡ ∂ ln Q ∂ ln N ⎤ ⎡ ∂β ⎤ = kN ⎢ − ⎥ + k⎢ E +β⎥ ∂E V .N ∂E V . N ⎢⎣ ∂E V . N ⎣⎢ ∂E V . Seitzman. T . eqn. School of Aerospace Engineering Thermodynamic Relations • ∂E 1 1 − QE ∂β ⎡ ∂β ⎤ = kN + k ⎢E +β⎥ T Q N ∂E V . All rights reserved. Q ) Partition Function TD variables = ∑ g i e −ε i kT i AE/ME 6765 Statistical Thermodynamic Rleations-8 Copyright © 2009 by Jerry M. All rights reserved. N ⎣⎢ ∂E V .N ⎦⎥ ⎣⎢ ∂E V . N ⎢⎣ ∂E V . N ⎥⎦ ∂Q ∂β ∂β ∂E V . School of Aerospace Engineering Thermodynamic Relations • So look at partial derivatives of Q S = kN ln + kN + kβ E Stat. N ⎦⎥ 1T 0 1 1 ∂Q ⎡ ∂β ⎤ = kN + k⎢ E +β⎥ T Q ∂E V . N ⎦⎥ ∂β ⎡ ∂β ⎤ = − kE + k ⎢E +β⎥ ∂E V . E . N ∂Q QN i Q = −∑ ε i g i e − βε i = −∑ ε i QE = − ∑ ε i Ni = − ∂β N N N AE/ME 6765 Statistical Thermodynamic Rleations-7 Copyright © 2009 by Jerry M.

N .N 0 p ∂ ⎛ Q ⎞ = ⎜ kN ln + kN + kβ E ⎟ T ∂V ⎝ N ⎠ E .N ∂V E . N . All rights reserved.V . 118 ∂V E. β −μ ~ ∂ ⎛ ⎡ ⎛ ∂ ln Q d ln N ⎞⎟ Q⎤ ⎞ ⎜ kN ln + kN ⎟ = ⎢kN ⎜⎜ ∂N Q − + k ln ⎥ + k = ⎟ T ∂N ⎝ N ⎠ E . V ) + ∂N E .Ly.N. School of Aerospace Engineering Chemical Potential and Q ∂S − μ~ per molec. Seitzman.T ) Q ≡ ∑ gi e − βε i εi. β ∂ β ∂N E .V .tr=f(Lx. p. Seitzman. School of Aerospace Engineering Pressure and Partition Function Alternative to Helmholtz ∂S p Free Energy derivation • Examine = in V&K. β ∂V E .V E .β p ∂ ln Q Is Q=Q(V)? Yes!! = kN T ∂V E . β =1 kT ∂β E .N T Q – since S=S(E.V . • Examine = Q ∂N E .V i 0 − μ~ ∂ ⎛ Q ⎞ = ⎜ kN ln + kN + kβ E ⎟ T ∂N ⎝ N ⎠ E .V dN ⎠ N ⎥⎦ ~ 1/N μ Q 0 = − k ln T N N Q – recall e −α = ⇒ α = ln ⇒ α = − μ~ kT ∑ gi e − βε i N i AE/ME 6765 Statistical Thermodynamic Rleations-10 Copyright © 2009 by Jerry M.V.V ⎢⎣ ⎝ E .β ∂ ⎛ Q⎞ ∂ p = ⎜ kN ln ⎟ = (kN ln Q − kN ln N ) T ∂V ⎝ N ⎠ E . N .Lz) i AE/ME 6765 Statistical Thermodynamic Rleations-9 Copyright © 2009 by Jerry M. All rights reserved. 5 . N . N ∂ N E .V T S = kN ln + kN + kβE N ∂S ∂S ∂S ∂β = Q = ∑ g i e − βε i = Q(T . N ∂V E.V .β) S = kN ln + kN + kβ E N ∂S ∂S ∂S ∂β = + ∂V E. N (.

m = Nm part de ⎡ ∂ ln Q ∂ 2 ln Q ⎤ ∂ ln Q cv = cv = R ⎢2T +T 2 ⎥ e = RT 2 dT ∂T ∂T 2 ⎦ ∂T ⎣ AE/ME 6765 Statistical Thermodynamic Rleations-11 Copyright © 2009 by Jerry M. we saw =− ⇒E=− Q ∂β ∂β N – so to find E N ∂Q ∂T E=− dβ 1 Q ∂T ∂β β = 1 kT ⇒ =− 2 1 ∂Q ∂Q = − (− kT 2 ) dT kT N = NkT 2 Q ∂T Q ∂T ∂ ln Q way to find E if N.N.. ⎝ N ⎠ kT N g e −ε kT Q ≡ ∑ g i e −ε kT i – examples i = i i i N Q Partition ⎛ Q⎞ E ∂ ln Q Function Boltzmann S = kN ⎜1 + ln ⎟ + E = kNT 2 Distribution ⎝ N⎠ T ∂T v ~ For given p ∂ ln Q μ Q .) ⎛ Q⎞ E – Boltzmann (dilute) limit ln Wmax = N ⎜1 + ln ⎟ + for weakly interacting molec..state (e. T known E = NkT 2 – also cv=? ∂T R = k m part . 6 . cv = RT ⎢ +T ⎥ + OTHERS ⎢⎣ ∂T v ∂T 2 v ⎥⎦ . Q.εi) = kN = − k ln T ∂V E . E. School of Aerospace Engineering Energy and Partition Function • Let’s examine relationship between Q and other TD variables ∂Q QE N ∂Q • Recall in finding β.including cv AE/ME 6765 Statistical Thermodynamic Rleations-12 Copyright © 2009 by Jerry M.V) we can find all ⎡ 2∂ ln Q ∂ 2 ln Q ⎤ TD props.substance (gi. N T N .g. Seitzman. All rights reserved. All rights reserved. Seitzman. School of Aerospace Engineering Summary • Derived TD properties using – Boltzmann’s relation S=klnΩ generally true (a priori prob.

Seitzman. β compare to TD S expression but modified energy ∂S 1 = = kβ ⇒ β = 1 kT Same as what we pop. All rights reserved. School of Aerospace Engineering BE-FD Results • Plug back into S expression ε − μ~ E − μ~N ⎛ − i ⎞ S= m k ∑ g i ln⎜⎜1 m e kT ⎟⎟ T i ⎝ ⎠ • Pressure In Boltzmann limit ∂S p ∂ε ∂ ln Q = ⇒ p = −∑ N i i p = NkT ∂V E.N ∂V E.V T AE/ME 6765 Statistical Thermodynamic Rleations-13 Copyright © 2009 by Jerry M. School of Aerospace Engineering BE-FD Results • If we remove Boltzmann limit assumption – but retain weakly interacting molec. ⎛ g ± Ni g ± Ni ⎞ e −α − βε i ln W⎡ BE ⎤ = ∑ ⎜⎜ N i ln i ± g i ln i ⎟ N i* = gi ⎢⎣ FD ⎥⎦ i ⎝ N i g i ⎟⎠ ⎡ BE ⎤ ⎢⎣ FD ⎥⎦ 1 m e −α − βε i ⇒ ln Wmax ⎡ BE ⎤ { ( = ∑ N i* (α + βε i ) m g i ln 1 m e −α − βε i )} ⎢⎣ FD ⎥⎦ i { ∴ S = k (αN + βE ) m k ∑ g i ln 1 m e −α − βε i ( )} i – to get α.N • So properties still only function of gi. N T g found in N i BE = (ε i − μ~ ) ikT * ∂S −μ~ Boltzmann limit!! ⎡ ⎤ ⎢⎣ FD ⎥⎦ e m1 = = kα ⇒ α = − μ~ kT ∂N E . but now we no can longer employ the simpler partition function expressions • For most general case (interacting particles). distribution ∂E V . Seitzman. need to employ Gibbs method (canonical ensembles. εi. see texts by Denbigh or McQuarrie or Laurendeau) AE/ME 6765 Statistical Thermodynamic Rleations-14 Copyright © 2009 by Jerry M.N T i ∂V E. All rights reserved. 7 .

what is the statistical thermodynamic interpretation of T? AE/ME 6765 Statistical Thermodynamic Rleations-15 Copyright © 2009 by Jerry M. All rights reserved. School of Aerospace Engineering Stat. School of Aerospace Engineering Interpretation of Properties • Originally. N) • When we defined the Lagrange multipliers for the most probably macrostate and related them to entropy. our examination of the “state” of a substance (group of particles) depended only on the particles’ microscopic properties (εi. Interpretation of T • Consider Boltzmann limit population distribution Ni g e −ε i kT Ni g e −ε i kT g −(ε −ε ) kT = i ⇒ = i −ε j kT = i e i j N Q N j g je gj = − (ε i − ε j ) kT Ni gi ln Ni gi ln N0 g0 Nj gj 1/T −1 εi − ε j ⎛ N g ⎞ T= ⎜⎜ ln j j ⎟⎟ εi − ε0 k ⎝ Ni gi ⎠ k • So temperature is the measure of how the particles are distributed amount the energy levels – for a given gas (gi. Seitzman. etc. Thermo. μ. εi) in equilibrium. defined based on averaging over groups of particles • For example. V. gi) and the system’s extensive properties (E. we then found “definitions” for T. T is the ONLY thing needed to determine the energy population distribution AE/ME 6765 Statistical Thermodynamic Rleations-16 Copyright © 2009 by Jerry M. p. – thus properties like temperature and pressure can only be defined for groups of particles – they are essentially statistical quantities. 8 . Seitzman. All rights reserved.