Notes on Statistical Thermodynamics

Transcribed from PDF Notes

April 29, 2025

Contents

1 Fundamental Concepts 2

2 Microstates, Macrostates, and Probability 2

2.1 Example: Tossing 4 Distinguishable Coins . . . . . . . . . . . . . . . . . . . . . . 2

3 Stirling’s Approximation 3

4 Distribution of Microstates 3
4.1 Example: N=10 Distinguishable Particles in a 2-State System . . . . . . . . . . . 3

5 Multistate Systems (Distinguishable Particles) 4

6 Energy Distributions (Distinguishable Particles) 5

6.1 Example 1: N=3 particles, U=3ε . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
6.2 Example 2: N=3 particles, U=4ε . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

7 Entropy (Boltzmann Definition) 5

7.1 Example: Entropy of Coin Toss Macrostates . . . . . . . . . . . . . . . . . . . . . 6
7.2 Example: Entropy Change for Gas Expansion (N=6) . . . . . . . . . . . . . . . . 6

8 Distinguishable Particles in Boxes 6

8.1 Example: 6 Molecules in 2 Halves . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
8.2 Example: 6 Distinguishable Particles in 3 Equal Partitions . . . . . . . . . . . . . 6
8.3 Even Distribution (No Empty Boxes) . . . . . . . . . . . . . . . . . . . . . . . . . 7

9 Entropy of an Ideal Gas (Simple Model) 7

10 Entropy Change, Probability, and Impossibility 8

10.1 Entropy Change for Volume Change . . . . . . . . . . . . . . . . . . . . . . . . . 8
10.2 Impossibility and Improbability . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

11 Degeneracy 8

12 The Boltzmann Factor and Partition Function 8

12.1 Boltzmann Distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
12.2 The Partition Function (q) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
12.3 Accounting for Degeneracy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
12.4 Example Calculations (T=300K) . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

1
13 The Canonical Partition Function (Q) 9
13.1 Distinguishable, Non-Interacting Particles . . . . . . . . . . . . . . . . . . . . . . 10
13.2 Indistinguishable, Non-Interacting Particles . . . . . . . . . . . . . . . . . . . . . 10

14 Populations in a Two-Level System 10

15 Ideal Monoatomic Gas 11

15.1 Translational Partition Function . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
15.2 Thermodynamic Properties from Q . . . . . . . . . . . . . . . . . . . . . . . . . . 11

16 Ensembles and Entropy 12

16.1 Entropy of an Ensemble . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
16.2 Relation to Partition Function (Canonical Ensemble) . . . . . . . . . . . . . . . . 13

17 Molecular Degrees of Freedom: Vibration 13

17.1 Harmonic Oscillator Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
17.2 Vibrational Partition Function (qvib ) . . . . . . . . . . . . . . . . . . . . . . . . . 13
17.3 Interpretation and Population . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
17.4 Vibrational Energy and Heat Capacity . . . . . . . . . . . . . . . . . . . . . . . . 14

18 Molecular Degrees of Freedom: Rotation 14

18.1 Rigid Rotor Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
18.2 Rotational Partition Function (qrot ) . . . . . . . . . . . . . . . . . . . . . . . . . 14
18.3 Rotational Energy and Heat Capacity . . . . . . . . . . . . . . . . . . . . . . . . 15
18.4 Rotational Spectra . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

19 Molecular Degrees of Freedom: Electronic 15

19.1 Electronic Partition Function (qelec ) . . . . . . . . . . . . . . . . . . . . . . . . . 15
19.2 Electronic Energy and Heat Capacity . . . . . . . . . . . . . . . . . . . . . . . . 16

2
1 Fundamental Concepts
• Statistical Thermodynamics provides the link between the microscopic properties of
matter (positions, energies, momenta of individual particles) and its bulk macroscopic
properties (temperature, pressure, entropy, heat capacity).

• A thermodynamic system can be described by:

– Macrostate: The overall state defined by macroscopic parameters (e.g., N , V , T ,

U , P ).
– Microstate: A specific, detailed arrangement of all particles within the system (e.g.,
the exact position and momentum, or the specific quantum state, of every particle).

• A single macrostate typically encompasses a vast number of possible microstates (W ).

• Energy Levels vs. Energy States:

– Macrostate Specification (Energy Level): Specifies the number of particles (Ni )

occupying each allowed energy level (εi ).
– Microstate Specification (Energy State): Specifies the exact energy state each
individual particle occupies. If energy levels are degenerate, multiple states can have
the same energy level.

2 Microstates, Macrostates, and Probability

2.1 Example: Tossing 4 Distinguishable Coins
Let H = Heads, T = Tails. N = 4. The outcome can be described by macrostates defined by
the number of Heads (NH ) and Tails (NT ). Each macrostate corresponds to a certain number
of microstates (specific sequences).

Macrostate Specification Microstates (Examples) Thermo. Prob. Math. Prob.

WR = NNH


R NH NT Sequence (Coin 1, 2, 3, 4) PR = WR /Ω
4


1 4 0 HHHH 4 =1 1/16
4


2 3 1 HHHT, HHTH, HTHH, THHH 3 =4 4/16
HHTT, HTHT, HTTH, 4


3 2 2 2 =6 6/16
THHT, THTH, TTHH
4


4 1 3 HTTT, THTT, TTHT, TTTH 1 =4 4/16
4


5 0 4 TTTT 0 =1 1/16
P
Total Microstates (Ω = PWR ): 16
Total Probability ( PR ): 1

= 2N = 24 = 16.
P
• Total Number of Microstates (Ω): Ω = R WR

• Thermodynamic Probability (WR ): The number of microstates corresponding to a

N N!


given macrostate R. It’s often calculated using combinations: WR = N1 = N1 !N2 ! for a
2-outcome system. Note: WR is an integer ≥ 1.

• Mathematical Probability P (PR ): The probability of observing macrostate R. PR =

WR
Ω . Note: 0 ≤ P R ≤ 1 and PR = 1.

3
 • Average Number of Heads (N̄H ):
P
N W
P H,R R
X
R
N̄H = NH,R PR =
R R WR

For the coin toss example:

(4 × 1) + (3 × 4) + (2 × 6) + (1 × 4) + (0 × 1) 4 + 12 + 12 + 4 + 0 32
N̄H = = = =2
16 16 16
Similarly, N̄T = 2.

3 Stirling’s Approximation
Used to estimate the factorial of very large numbers (N  1), common in statistical mechanics.

• The logarithm of N ! is the sum:

N
X
ln(N !) = ln(n)
n=1

• For large N , the sum can be approximated by an integral:

Z N
ln(N !) ≈ ln(x) dx = [x ln(x) − x]N
1 = (N ln N − N ) − (1 ln 1 − 1) = N ln N − N + 1
1

• A more accurate form (Stirling’s formula):

1
ln(N !) ≈ N ln N − N + ln(2πN )
2
Or, for the factorial itself:
N
√

N
N! ≈ 2πN
e

• Common Approximation: For very large N (e.g., N ∼ NA ), the ln N term is much

smaller than N , and often the simpler form is sufficient:

ln(N !) ≈ N ln N − N

The 12 ln(2πN ) term is the correction factor, often negligible in logarithmic calculations
involving NA .

4 Distribution of Microstates
4.1 Example: N=10 Distinguishable Particles in a 2-State System
Let the states be State 1 and State 2. A macrostate is defined by N1 (number

in State

1). Then
N 10
N2 = N − N1 = 10 − N1 . The thermodynamic probability is W (N1 ) = N1 = N1 .

4
 10


N1 W = N 1
0 1
1 10
2 45
3 120
4 210
5 252 (Wmax )
6 210
7 120
8 45
9 10
10 1

The distribution of W vs N1 is sharply peaked around the most probable macrostate.

• For N even, Wmax occurs at N1 = N/2.
• For N odd, Wmax occurs at N1 = (N − 1)/2 and N1 = (N + 1)/2. The distribution is
slightly flatter near the maximum.
• The macrostate with the highest thermodynamic probability (Wmax ) is the most likely to
be observed at equilibrium.
Deviation Example: Changing N1 from 5 (where W = 252) by 10 % of N (i.e., by 1 particle)
to N1 = 4 or N1 = 6 (where W = 210).
W (N1 = 4) 210
= ≈ 0.833 = 83.3 %
W (N1 = 5) 252
A small deviation from the most probable distribution leads to a noticeable decrease in the
thermodynamic probability. For macroscopic N , this decrease becomes extremely sharp.

5 Multistate Systems (Distinguishable Particles)

Consider N distinguishable particles distributed among n states (or energy levels). A macrostate
Pnspecified by the set {N1 , N2 , . . . , Nn }, where Ni is the number of particles in state i, and
is
i=1 Ni = N .
The number of ways to achieve a specific macrostate (thermodynamic probability W ) is
given by the multinomial coefficient:

N! N!
W (N1 , N2 , . . . , Nn ) = = Qn
N1 !N2 ! . . . Nn ! i=1 Ni !

Derivation (for n=3):

N N!


1. Choose N1 particles from N for state 1: N1 = N1 !(N −N1 )! ways.
N −N1

(N −N1 )!
2. Choose N2 particles from remaining (N − N1 ) for state 2: N2 = N2 !(N −N1 −N2 )! ways.

3. Choose N3 particles from remaining (N − N1 − N2 ) for state 3: N3 = N − N1 − N2 .

N3


N3 = 1 way.

−N1
N3
(N −N1 )!
Total ways W = NN1 N N 2
N! N!
N3 = N1 !(N −N1 )! × N2 !(N3 )! × 1 = N1 !N2 !N3 ! . P
The probability of a macrostate R (specified by {Ni }R ) is PR = WR /Ω, where Ω = R WR
is the total number of microstates. For distinguishable particles distributed among n equally
likely states, Ω = nN .

5
6 Energy Distributions (Distinguishable Particles)
Constraints:
P
1. Conservation of Particles: i Ni = N
P
2. Conservation of Energy: i Ni εi = U (Total Energy)

6.1 Example 1: N=3 particles, U=3ε

Energy levels: 0, ε, 2ε, 3ε. Particles A, B, C are distinguishable.

Macrostate Occupation Numbers Microstate Examples Ther

(R) N0 N1 N2 N3 (Particle → Energy Level)
1 2 0 0 1 A→ 3ε, B→ 0, C→ 0; A→ 0, B→ 3ε, C→ 0; ...
2 1 1 1 0 A→ 0, B→ ε, C→ 2ε; A→ 0, B→ 2ε, C→ ε; ... (all 6 permutations)
3 0 3 0 0 A→ ε, B→ ε, C→ ε
Total Microstates (Ω)

Note: WR here counts the number of ways to assign distinguishable particles A, B, C to the
specified energy levels.

6.2 Example 2: N=3 particles, U=4ε

Energy levels: 0, ε, 2ε, 3ε. Particles A, B, C are distinguishable.

Macrostate Occupation Numbers Microstate Examples Thermo. Prob. (WR

(R) N0 N1 N2 N3 (Particle → Energy Level) WR = QNN!i !
3!
1 1 1 0 1 A→ 0, B→ ε, C→ 3ε; ... (all 6 permutations) 1!1!0!1! = 6
3!
2 1 0 2 0 A→ 0, B→ 2ε, C→ 2ε; ... (3 distinct assignments) 1!0!2!0! = 3
3!
3 0 2 1 0 A→ ε, B→ ε, C→ 2ε; ... (3 distinct assignments) 0!2!1!0! = 3
Total Microstates (Ω) 12

7 Entropy (Boltzmann Definition)

Boltzmann defined entropy (S) in terms of the thermodynamic probability (W ) of a macrostate:

S = kB ln W

where kB = R/NA ≈ 1.38 × 10−23 J K−1 is the Boltzmann constant.

• W is the number of microstates corresponding to the observed macrostate.

• Entropy is a measure of the disorder or randomness of a system. A higher W implies more

ways to arrange the particles, hence higher disorder and higher entropy.

• The state with W = 1 (perfectly ordered state) has S = kB ln 1 = 0.

• The equilibrium state corresponds to the macrostate with the maximum W (and thus
maximum S).

• The Second Law of Thermodynamics: Spontaneous processes in isolated systems tend

to increase entropy (W increases). Systems naturally evolve from less probable (more
1
ordered) states to more probable (more disordered) states. S ∝ Order .

6
7.1 Example: Entropy of Coin Toss Macrostates
Using the results from the 4-coin toss:

Macrostate (NH , NT ) WR Entropy S = kB ln WR

(4, 0) 1 0
(3, 1) 4 kB ln 4 ≈ 1.39kB
(2, 2) 6 kB ln 6 ≈ 1.79kB
(1, 3) 4 kB ln 4 ≈ 1.39kB
(0, 4) 1 0

The macrostate (2, 2) is the most disordered and has the highest entropy. Macrostates (4, 0)
and (0, 4) are the most ordered and have zero entropy.

7.2 Example: Entropy Change for Gas Expansion (N=6)

Consider 6 molecules transitioning from the most ordered state
(all in one half of a box, W = 1)
to the most disordered state (3 molecules in each half, W = 63 = 20).

∆S = Sfinal − Sinitial = kB ln Wfinal − kB ln Winitial

∆S = kB ln 20 − kB ln 1 = kB ln 20 ≈ 2.996kB
This positive entropy change reflects the spontaneous tendency for the gas to expand and become
more disordered.

8 Distinguishable Particles in Boxes

8.1 Example: 6 Molecules in 2 Halves
6


Macrostate defined by (NI , NII ), where NI + NII = 6. W (NI , NII ) = NI .

W = N6I


Macrostate NI NII S = kB ln W
1 0 6 1 0
2 1 5 6 1.79kB
3 2 4 15 2.71kB
4 3 3 20 2.96kB
5 4 2 15 2.71kB
6 5 1 6 1.79kB
7 6 0 1 0

8.2 Example: 6 Distinguishable Particles in 3 Equal Partitions

A microstate specifies which partition each of the 6 particles is in. Total microstates = 36 = 729.
Let’s analyze by macrostate (distribution) (N1 , N2 , N3 ).

• Arrangement: How the partitions are assigned the numbers (e.g., (6,0,0) can be arranged
as [6 0 0], [0 6 0], [0 0 6]).

• WR (Ways per Arrangement): For a given arrangement, the number of ways to assign
the 6 distinguishable particles, WR = N1 !N6!2 !N3 ! .

• WTotal (Macrostate) = (Number of Arrangements) ×WR .

Examples of Macrostates:

7
 6!
• Macrostate (6,0,0): 3 arrangements ([600],[060],[006]). WR = 6!0!0! = 1. WTotal = 3 × 1 =
3.
6!
• Macrostate (5,1,0): 6 arrangements ([510],[501],[150],[051],[105],[015]). WR = 5!1!0! = 6.
WTotal = 6 × 6 = 36.
6!
• Macrostate (4,1,1): 3 arrangements ([411],[141],[114]). WR = 4!1!1! = 30. WTotal =
3 × 30 = 90.
6!
• Macrostate (3,2,1): 6 arrangements. WR = 3!2!1! = 60. WTotal = 6 × 60 = 360.
6!
• Macrostate (4,2,0): 6 arrangements. WR = 4!2!0! = 15. WTotal = 6 × 15 = 90.
6!
• Macrostate (3,3,0): 3 arrangements. WR = 3!3!0! = 20. WTotal = 3 × 20 = 60.
6!
• Macrostate (2,2,2): 1 arrangement ([222]). WR = 2!2!2! = 90. WTotal = 1 × 90 = 90.

Total Microstates Ω = WTotal = 3 + 36 + 90 + 360 + 90 + 60 + 90 = 729 = 36 .

P
Generalization: For N distinguishable particles placed into n distinct boxes (states), the
total number of microstates is Ω = nN .

8.3 Even Distribution (No Empty Boxes)

If N distinguishable particles are distributed into N distinct boxes such that each box contains
exactly one particle (even distribution, no empty boxes allowed), the total number of ways is
W = N !.

9 Entropy of an Ideal Gas (Simple Model)

Consider one mole (N = NA ) of a gas evenly distributed in a volume V . If we imagine dis-
cretizing the volume into NA tiny cells, each occupied by one molecule, the number of ways to
arrange the distinguishable molecules is W = NA !.

S = kB ln W = kB ln(NA !)

Using Stirling’s approximation (ln N ! ≈ N ln N − N ):

ln(NA !) ≈ NA ln NA − NA

S ≈ kB (NA ln NA − NA ) = R ln NA − R
Calculating ln NA = ln 6.022 × 1023 ≈ 54.75.

S ≈ R(54.75 − 1) = 53.75R

S ≈ 53.75 × 8.314 J mol−1 K−1 ≈ 447 J mol−1 K−1

(This is a highly simplified model, ignoring volume dependence, momentum, etc. The Sackur-
Tetrode equation provides a more rigorous result.)

8
10 Entropy Change, Probability, and Impossibility
10.1 Entropy Change for Volume Change
Consider N molecules isothermally changing volume from V1 to V2 . The probability ratio of
finding the molecules in V2 compared to V1 is related to the ratio of available microstates (phase
space volume):
 N
W2 V2
p= =
W1 V1
The change in entropy is given by thermodynamics (for ideal gas):
 
V2
∆S = N kB ln
V1
Comparing these, we find a relationship between entropy change and the probability ratio:
 N
V2
∆S = kB ln = kB ln p
V1

10.2 Impossibility and Improbability

A process is considered ”impossible” if its probability is so astronomically small that it is never
expected to occur naturally within the age of the universe. Example: Spontaneous compression
of N gas molecules from V1 to V2 = V1 /2.

V1 /2 N
   N
1
p= =
V1 2
• If N = 10, p = (1/2)10 = 1/1024 ≈ 1 × 10−3 . This might occur relatively frequently on
molecular timescales.
• If N = 50, p = (1/2)50 ≈ 8.9 × 10−16 . Extremely improbable.
• If N = NA ≈ 6 × 1023 , p = (1/2)NA which is practically zero. Spontaneous compression
of macroscopic gas is impossible.
Note: The magnitude of ∆S is the same for compression (V1 → V1 /2, ∆S < 0) and expansion
(V1 /2 → V1 , ∆S > 0), but the sign reflects the direction of spontaneous change. |∆S| =
N kB ln 2.

11 Degeneracy
An energy level εi is degenerate if there is more than one distinct quantum state corresponding
to that energy level. The degree of degeneracy, gi , is the number of states having the energy
εi .

12 The Boltzmann Factor and Partition Function

12.1 Boltzmann Distribution
The probability (Pi ) of a system (or particle) being in a specific microstate i with energy Ei at
thermal equilibrium (temperature T ) is proportional to the Boltzmann factor:
Pi ∝ e−Ei /(kB T ) = e−βEi
where β = 1/(kB T ). e−βEi is dimensionless. Lower energy states are more probable.

9
12.2 The Partition Function (q)
P
To normalize the probability ( Pi = 1), we introduce the molecular partition function, q:
X
q= e−βεi
states i

The sum is over all possible quantum states accessible to a single particle. The probability of
finding a particle in state i is:
e−βεi
Pi =
q

12.3 Accounting for Degeneracy

If energy levels εj have degeneracy gj , the partition function can be written as a sum over
energy levels:
X
q= gj e−βεj
levels j

• Probability of occupying a specific state k within level j:

e−βεj
Pk (state) =
q

• Probability of occupying any state within level j:

gj e−βεj
Pj (level) =
q

The partition function q provides a measure of the number of thermally accessible states at a
given temperature.

12.4 Example Calculations (T=300K)

kB T ≈ 4.14 × 10−21 J at 300 K. Assume g = 1 unless stated.
Case 1: ε0 = 0, ε1 = 1 J, ε2 = 2 J βε1 ≈ 2.4 × 1020  1, βε2 ≈ 4.8 × 1020  1. q =
e0 + e−βε1 + e−βε2 ≈ 1 + 0 + 0 = 1. P0 ≈ 1, P1 ≈ 0, P2 ≈ 0. (Only ground state populated)
Case 2: ε0 = 0, ε1 = 1 × 10−20 J, ε2 = 2 × 10−20 J βε1 ≈ 2.415, βε2 ≈ 4.83. q = e0 +
e −2.415 + e−4.83 ≈ 1 + 0.0894 + 0.0080 ≈ 1.0974. P0 = 1/q ≈ 0.911. P1 = e−2.415 /q ≈ 0.081.
P2 = e−4.83 /q ≈ 0.007. (Mainly ground state, some excited state population)
Case 3: ε0 = 0, ε1 = 1 × 10−40 J, ε2 = 2 × 10−40 J βε1 ≈ 2.4 × 10−20 ≈ 0, βε2 ≈ 4.8 × 10−20 ≈
0. q = e0 + e0 + e0 = 1 + 1 + 1 = 3. P0 = 1/3 ≈ 0.33. P1 = e0 /3 ≈ 0.33. P2 = e0 /3 ≈ 0.33.
(Nearly equal population due to εi  kB T )
Interpretation:

• q indicates the effective number of accessible states.

• Higher T or smaller energy spacing (∆ε) increases q and excited state populations.

• The occupational probability Pj depends on both the Boltzmann factor e−βεj and the
partition function q.

13 The Canonical Partition Function (Q)

For a system of N particles.

10
13.1 Distinguishable, Non-Interacting Particles
If the N particles arePdistinguishable (e.g., fixed in a crystal lattice) and do not interact, the
total energy is E = N k=1 εk , where εk is the energy of the k-th particle in its specific state.
The system partition function Q is:
X X X X
Q= e−βE = ··· e−β(ε1 +ε2 +···+εN )
all system states states of p1 states of p2 states of pN

! !  
X X X
Q= e−βεi1 e−βεi2 ... e−βεiN 
states i1 states i2 states iN

If the particles are identical (have the same energy levels available):

Q = qN (Distinguishable identical particles)

where q is the molecular partition function.

13.2 Indistinguishable, Non-Interacting Particles

For identical, indistinguishable particles (like gas molecules), we must account for the fact that
swapping particles does not lead to a new microstate.

• Fermions: Obey Pauli exclusion principle (max one particle per quantum state). Use
Fermi-Dirac statistics.

• Bosons: No restriction on occupation number per state. Use Bose-Einstein statistics.

High Temperature / Low Density Approximation (Corrected Boltzmann Statistics):

When the number of accessible states is much larger than the number of particles (q  N ),
multiple occupancy of states is rare. We can approximate the partition function by dividing
the distinguishable result by N ! (the number of permutations of N particles):

qN
Q= (Indistinguishable identical particles, high T / low ρ limit)
N!

This correction avoids the Gibbs paradox.

14 Populations in a Two-Level System

Example: A system with energy spacing ∆E = 207 cm−1 at T = 298 K. ∆E = hcν̃ =
(6.626 × 10−34 J s)(3 × 108 m s−1 )(20 700 m−1 ) ≈ 4.114 × 10−21 J. kB T ≈ (1.38 × 10−23 J K−1 )(298 K) ≈
4.112 × 10−21 J. So, β∆E = ∆E/(kB T ) ≈ 1.0.
Case 1: Non-degenerate (g0 = 1, g1 = 1). Let E0 = 0, E1 = ∆E. q = g0 e0 + g1 e−β∆E =
1 + e−1.0 ≈ 1 + 0.368 = 1.368. P0 = g0 e0 /q = 1/1.368P≈ 0.731. P1 = g1 e−1.0 /q =
e−1.0 /1.368 ≈ 0.368/1.368 ≈ 0.269. Average Energy hEi = Ei Pi = 0 × P0 + ∆E × P1 ≈
(4.114 × 10 −21 J)(0.269) ≈ 1.107 × 10 −21 J. hEi in cm = hEi/(hc) ≈ 55.7 cm−1 .
−1

Case 2: g0 = 1, g1 = 2. q = g0 e + g1 e−β∆E = 1 + 2e−1.0 ≈ 1 + 2(0.368) = 1.736.

0

P0 = g0 e0 /q = 1/1.736 ≈ 0.576. P1 (level) = g1 e−1.0 /q = 2e−1.0 /1.736 ≈ 0.736/1.736 ≈

Ei Pi (level) = 0 × P0 + ∆E × P1 ≈ (4.114 × 10−21 J)(0.424) ≈
P
0.424. Average Energy hEi =
1.744 × 10 −21 J. hEi in cm ≈ 87.8 cm−1 .
−1

11
15 Ideal Monoatomic Gas
Particles like He, Ne, Ar. Only translational energy is significant at normal temperatures.
Etotal = Etrans + Erot + Evib + Eelec qtotal = qtrans × qrot × qvib × qelec
For monoatomic gas:

• qrot = 1 (no rotation for point mass).

• qvib = 1 (no vibration).

• qelec ≈ gel,0 (ground state electronic degeneracy, usually 1 for noble gases).

Therefore, qtotal ≈ qtrans .

15.1 Translational Partition Function

From quantum mechanics, the energy levels for a particle of mass m in a 3D box of dimensions
a × b × c are:
!
h2 n2x n2y n2z
εnx ,ny ,nz = + 2 + 2 (nx , ny , nz = 1, 2, 3, . . . )
8m a2 b c

For a cubic box (a = b = c, Volume V = a3 ):

∞ X
∞ X
∞
βh2
X  
2 2 2
qtrans = exp − (n + ny + nz )
8ma2 x
nx =1 ny =1 nz =1

∞
!3
2 n2
− βh
X
qtrans = e 8ma2

n=1

Approximating the sum by an integral (valid for large box / high T):
s
∞ Z ∞
βh2 n2 βh2 n2 2πma2
r
1 π
e− 8ma2 ≈ e− 8ma2 dn =
X
=
0 2 βh2 /(8ma2 ) βh2
n=1
s !3 3/2 3/2
2πma2 2πma2 kB T
 
2πmkB T
qtrans = = = a3
βh2 h2 h2
 3/2
2πmkB T V
qtrans = V =
h2 Λ3
q
h2
where Λ = 2πmkB T is the thermal de Broglie wavelength.

15.2 Thermodynamic Properties from Q

N
For N indistinguishable particles (ideal gas), Q = qtrans /N !.

ln Q = N ln qtrans − ln N ! ≈ N ln qtrans − (N ln N − N )
   
3 2πmkB T
ln Q = N ln + ln V − N ln N + N
2 h2

12
 Average Energy hEi:
   
∂ ln Q 2 ∂ ln Q
hEi = − = kB T
∂β N,V ∂T N,V

3
Since qtrans ∝ T 3/2 , ln qtrans = 2 ln T + const.
3
ln Q = N ( ln T ) + terms independent of T
2
 
∂ ln Q 31 3N
=N =
∂T N,V 2T 2T
 
2 3N 3 3
hEi = kB T = N kB T = nRT
2T 2 2
Heat Capacity CV :
 
∂hEi 3 3
CV = = N kB = nR
∂T N,V 2 2

Pressure hP i:  
∂ ln Q
hP i = kB T
∂V N,T
ln Q = N ln V + terms independent of V
 
∂ ln Q N
=
∂V N,T V
 
N
hP i = kB T
V
hP iV = N kB T = nRT (Ideal Gas Law)

16 Ensembles and Entropy

An ensemble is a hypothetical collection of a very large number (A) of identical systems, pre-
pared under the same macroscopic conditions (N, V, E for microcanonical; N, V, T for canonical).

16.1 Entropy of an Ensemble

For an ensemble of A distinguishable systems, where aj systems are in microstate j, the total
number of ways to form the ensemble is:
A!
Wens = Q
j aj !

The ensemble entropy is Sens = kB ln Wens .

If Pj = aj /A is the probability of finding a system in state j, using Stirling’s approx:
X
Sens = −AkB Pj ln Pj
j

The average entropy per system in the ensemble is:

Sens X
Ssystem = = −kB Pj ln Pj (Gibbs entropy formula)
A
j

13
16.2 Relation to Partition Function (Canonical Ensemble)
In the canonical ensemble (N, V, T fixed), Pj = e−βEj /Q. Substituting into Gibbs entropy
formula:
X e−βEj
Ssystem = −kB (−βEj − ln Q)
Q
j
X e−βEj X e−βEj
= kB β Ej + kB ln Q
Q Q
j j
X
= kB βhEi + kB ln Q (since Pj = 1)
hEi
= + kB ln Q
T
Since A = −kB T ln Q (Helmholtz free energy), S = (hEi−A)/T . Also, hEi = kB T 2 (∂ ln Q/∂T )N,V .
Substituting this gives:
 
∂ ln Q
Ssystem = kB T + kB ln Q
∂T N,V

17 Molecular Degrees of Freedom: Vibration

17.1 Harmonic Oscillator Model
Energy levels for a quantum harmonic oscillator (frequency ν):
 
1
εv = v + hν (v = 0, 1, 2, . . . )
2

E0 = 12 hν is the zero-point energy (ZPE).

17.2 Vibrational Partition Function (qvib )

∞
X ∞
X ∞ 
X v
qvib = e−βεv = e−β(v+1/2)hν = e−βhν/2 e−βhν
v=0 v=0 v=0
P∞
Using the geometric series sum n=0 xn = 1/(1 − x):

e−βhν/2 e−Θvib /(2T )

qvib = =
1 − e−βhν 1 − e−Θvib /T

hν
where Θvib = is the characteristic vibrational temperature. Note: Often, energy is
kB
measured relative to the ZPE, so ε0v = vhν. In this case:
∞
0
X 1 1
qvib = e−βvhν = −βhν
= −Θ
v=0
1−e 1 − e vib /T

0
Thermodynamic results are independent of the energy zero choice. Let’s use qvib for simplicity.

14
17.3 Interpretation and Population
Θvib is typically high (1000 K to 4000 K).

• If T  Θvib : e−Θvib /T ≈ 0. qvib

0 ≈ 1. Only ground state (v = 0) is significantly populated.

• If T  Θvib : e−Θvib /T ≈ 1−Θvib /T . qvib

0 ≈ 1/(Θvib /T ) = T /Θvib . Many states populated.

Population of vibrational state v: Pv = e−βvhν /qvib

0 = e−vΘvib /T (1 − e−Θvib /T ). Population of
ground state (v = 0): P0 = 1 − e−Θvib /T .
Example (N2 ): Θvib ≈ 3395 K. At T = 300 K, Θvib /T ≈ 11.3. P0 = 1 − e−11.3 ≈
1 − 1.2 × 10−5 ≈ 0.999988. P1 = e−11.3 (1 − e−11.3 ) ≈ 1.2 × 10−5 . Very few molecules are in
excited vibrational states at room temperature.

17.4 Vibrational Energy and Heat Capacity

 0 
∂ ln qvib Θvib
hEi0vib = kB T2
= R Θ /T (per mole, rel. to ZPE)
∂T V e vib − 1
∂hEi0 Θvib 2 eΘvib /T
   
(CV )vib = =R
∂T V T (eΘvib /T − 1)2
• Low T (T  Θvib ): (CV )vib → 0. Vibrations are ”frozen out”.

• High T (T  Θvib ): (CV )vib → R. Equipartition theorem applies (2 quadratic terms,

1 2 1 2
2 kx + 2 px /µ).

18 Molecular Degrees of Freedom: Rotation

18.1 Rigid Rotor Model
For a linear molecule with moment of inertia I = µr2 . Energy levels:

~2
εJ = J(J + 1) = BJ(J + 1) (J = 0, 1, 2, . . . )
2I
where B = ~2 /(2I) is the rotational constant (in Joules). ~ = h/(2π). Often use B in Hz
(BH = B/h) or B̃ in cm−1 (B̃ = B/(hc)). Each level J has degeneracy gJ = 2J + 1.

18.2 Rotational Partition Function (qrot )

∞
X ∞
X
qrot = gJ e−βεJ = (2J + 1)e−βBJ(J+1)
J=0 J=0

B ~2
Define the characteristic rotational temperature: Θrot = = .
kB 2IkB
∞
X
qrot = (2J + 1)e−J(J+1)Θrot /T
J=0

Θrot is typically small (1 K to 100 K). At room temperature, T  Θrot .

High Temperature Approximation (T  Θrot ): Replace sum by integral:
Z ∞
qrot ≈ (2J + 1)e−J(J+1)Θrot /T dJ
0

15
Let x = J(J + 1), dx = (2J + 1)dJ.
Z ∞ " #∞
−xΘrot /T
 
−xΘrot /T e 1 T
qrot ≈ e dx = =0− =
0 −Θrot /T −Θrot /T Θrot
0

Symmetry Number (σ): For homonuclear diatomics (N2 , O2 , H2 ) or symmetric poly-

atomics, identical rotational states are reached by rotation. We must divide by the symmetry
number σ to avoid overcounting. σ = 1 for heteronuclear diatomics (CO, HCl). σ = 2 for
homonuclear diatomics (H2 , N2 ).

T
qrot ≈ (T  Θrot )
σΘrot

Example: Calculate Θrot for H2 and D2 . r ≈ 74.16 pm. µ(H2 ) ≈ 0.5 amu ≈ 8.35 × 10−28 kg.
(1.054 × 10−34 J s)2
I(H2 ) ≈ 4.60 × 10−48 kg m2 . Θrot (H2 ) = 2(4.60 × 10−48 kg m2 )(1.38 × 10−23 J K−1 )
≈ 87.5 K. µ(D2 ) ≈
1 amu ≈ 1.67 × 10−27 kg. I(D2 ) ≈ 9.20 × 10−48 kg m2 . Θrot (D2 ) ≈ 43.8 K.

18.3 Rotational Energy and Heat Capacity

Using qrot ≈ T /(σΘrot ) for T  Θrot .
ln qrot ≈ ln T − ln(σΘrot )
   
∂ ln qrot 1
hEirot = kB T 2 = kB T 2
= kB T (per molecule)
∂T V T
hEirot = RT (per mole)
 
∂hEi
(CV )rot = = R (per mole)
∂T V
This agrees with equipartition theorem (2 rotational degrees of freedom for a linear molecule,
each contributing 21 R).

18.4 Rotational Spectra

Selection rule: ∆J = ±1. Transitions J → J + 1. ∆ε = εJ+1 − εJ = B(J + 1)(J + 2) −
BJ(J + 1) = 2B(J + 1). Frequencies of absorption lines: ν = ∆ε/h = 2BH (J + 1). Lines
appear at 2BH , 4BH , 6BH , . . . . Spacing is 2BH . Intensity profile depends on population PJ ∝
(2J + 1)e−EJ /(kB T ) , showing a maximum at intermediate J.

19 Molecular Degrees of Freedom: Electronic

19.1 Electronic Partition Function (qelec )
Energy levels εel,i with degeneracies gel,i .
X
qelec = gel,i e−βεel,i = gel,0 e−βεel,0 + gel,1 e−βεel,1 + . . .
i

Electronic energy spacings are usually very large (∼ eV ∼ 104 − 105 cm−1 ). Corresponding
”electronic temperature” Θel is very high. At normal temperatures, T  Θel . Only the ground
electronic state is significantly populated. Choosing εel,0 = 0:
qelec ≈ gel,0

where gel,0 is the degeneracy of the ground electronic state (often 1 for closed-shell molecules).

16
19.2 Electronic Energy and Heat Capacity
If qelec ≈ gel,0 (constant).
 
2 ∂ ln qelec
hEielec = kB T =0 (relative to ground state)
∂T V
 
∂hEi
(CV )elec = =0
∂T V
Electronic degrees of freedom usually do not contribute to hEi or CV at ordinary temperatures.
(Note: If energy is measured relative to dissociated atoms, εel,0 = −De , then hEielec = −De ).

17