Thermodynamics

Boltzmann (Gibbs) Distribution

Maxwell-Boltzmann Distribution

Lana Sheridan

De Anza College

May 8, 2020
Last time

• heat capacities for ideal gases

• adiabatic processes
Overview

• adiabatic process in weather

• the Boltzmann distribution (distribution of energies)

• the Maxwell-Boltzmann distribution (distribution of speeds)

Reminder: Adiabatic Process in Ideal Gases

For an adiabatic process (Q = 0):

PV γ = const.

and:

TV γ−1 = const.

(Given the first one is true, the second follows immediately from
the ideal gas equation, P = nRT V .)
Weather and Adiabatic Process in a Gas

On the eastern side of the Rocky Mountains there is a

phenomenon called chinooks.

These eastward moving wind patterns cause distinctive cloud

patterns (chinook arches) and sudden increases in temperature.
Weather and Adiabatic Process in a Gas
As the air rises from the ocean it expands in the lower pressure at
altitude and cools. The water vapor condenses out of the air and
falls as precipitation.

As the air passes over the mountain it absorbs the latent heat from
the water condensation, then it stops cooling. As it descends, it is
compressed (nearly) adiabatically as the ambient pressure
increases. The air temperature rises!
Temperature and the Distribution of Particles’
Energies

In a gas at temperature T , we know the average translational KE

of the molecules.

However, not all of the molecules have the same energy, that’s just
the average.

How is the total energy of the gas distributed amongst the

molecules?
Temperature and the Distribution of Particles’
Energies

Ludwig Boltzmann first found the distribution of the number of

particles at a given energy given a thermodynamic system at a
fixed temperature.

Assuming that energy takes continuous values we can say that the
number of molecules per unit volume with energies in the range E
to E + dE is:
Z E +dE
N[E ,E +dE] = nV (E ) dE
E
Where
nV (E ) = n0 e −E /kB T

and n0 is a constant setting the scale: when E = 0, nV (E ) = n0 .

The Boltzmann Distribution
This particular frequency distribution:

nV (E ) ∝ e −E /kB T

is called the Boltzmann distribution or sometimes the Gibbs

distribution (after Josiah Willard Gibbs, who studied the behavior
of this distribution in-depth).

This distribution is even easier to understand for discrete energy

levels.
The Boltzmann Distribution
This particular frequency distribution:

nV (E ) ∝ e −E /kB T

is called the Boltzmann distribution or sometimes the Gibbs

distribution (after Josiah Willard Gibbs, who studied the behavior
of this distribution in-depth).

This distribution is even easier to understand for discrete energy

levels.

The probability for a given particle to be found in a state with

energy Ei drawn from a sample at temperature T :
1 −Ei /kB T
p(Ei ) = e
Z
where Z is simply a normalization constant to allow the total
probability to be 1. (The partition function.)
The Boltzmann Distribution

1 −Ei /kB T
p(Ei ) = e
Z

If we know the energies of two states E1 and E2 , E2 > E1 , we can

find the ratio of the number of particles in each:

nV (E2 )
= e −(E2 −E1 )/kB T
nV (E1 )

States with lower energies have more particles occupying them.

The Boltzmann Distribution

Lower temperature Higher temperature

1
Figure from the website of Dr. Joseph N. Grima, University of Malta.
(Somewhat Contrived) Example
Suppose a type of atom has only 2 energy states, separated in
energy by 12.0 eV.1 For a very large sample of these atoms, at
what temperature would 1% of the atoms in the sample be in the
excited (higher energy) state?

∆E = E2 − E1 = 12 eV
(Somewhat Contrived) Example
Suppose a type of atom has only 2 energy states, separated in
energy by 12.0 eV.1 For a very large sample of these atoms, at
what temperature would 1% of the atoms in the sample be in the
excited (higher energy) state?

nV (E2 ) 1
=
nV (E1 ) 99

1
This does not describe any real atom.
(Somewhat Contrived) Example
Suppose a type of atom has only 2 energy states, separated in
energy by 12.0 eV.1 For a very large sample of these atoms, at
what temperature would 1% of the atoms in the sample be in the
excited (higher energy) state?

nV (E2 ) 1
=
nV (E1 ) 99
1
e −(E2 −E1 )/kB T =
99
−(E2 − E1 )
= − ln(99)
kB T
(E2 − E1 )
T =
kB ln 99
12 eV × 1.602 × 10−19 J/eV
=
(1.38 × 10−23 ) ln 99
= 30, 300 K (3 sig figs)
1
This does not describe any real atom.
(Somewhat Contrived) Example
Suppose a type of atom has only 2 energy states, separated in
energy by 12.0 eV.2 For a very large sample of these atoms,

∆E = E2 − E1 = 12 eV
At what temperature would the number of atoms in each state be
equal?

2
This does not describe any real atom.
(Somewhat Contrived) Example
Suppose a type of atom has only 2 energy states, separated in
energy by 12.0 eV.2 For a very large sample of these atoms,

∆E = E2 − E1 = 12 eV
At what temperature would the number of atoms in each state be
equal?

nV (E2 )
= 1
nV (E1 )

2
This does not describe any real atom.
(Somewhat Contrived) Example
Suppose a type of atom has only 2 energy states, separated in
energy by 12.0 eV.2 For a very large sample of these atoms,

∆E = E2 − E1 = 12 eV
At what temperature would the number of atoms in each state be
equal?

nV (E2 )
= 1
nV (E1 )
e −(E2 −E1 )/kB T = 1
−(E2 − E1 )
= 0
kB T
T → ∞

2
This does not describe any real atom.
Aside: Lasers
Lasers emit coherent light. One photon interacts with an atom and
causes another to be emitted with the same phase.

This starts a cascade.

Inside a laser cavity there are atoms that are in a very strange
state: a higher energy level is more populated than a lower one.
This is called a “population inversion”.
Aside: Lasers

This is necessary for the photon cascade. Since:

nV (E2 )
= e −(E2 −E1 )/kB T , E2 > E1
nV (E1 )

we can associate a “negative temperature”, T , to these two energy

states in the atoms.
Maxwell-Boltzmann speed distribution

The Boltzmann distribution for energy can be leveraged to find a

distribution of the speeds of the molecules.

This is the Maxwell-Boltzmann speed distribution.

The number of molecules with speeds between v and v + dv is

Z v +dv Z v +dv  3/2
m0 2 /2k T
Nv dv = 4πN v 2 e −m0 v B
dv
v v 2πkB T
 Maxwell-Boltzmann speed distribution
The number of molecules with speeds between v and v + dv is
Z v +dv Z v +dv
641
3/2
21.5 Distribution of Molecular Speeds m0

2
Nv dv = 4πN v 2 e −m0 v /2kB T dv
v v 2πkB T
of speeds of N gas
The number of molecules
having speeds ranging from v
to v ! dv equals the area of
(21.41) N the tan rectangle, Nv dv.
v

vmp
stant, and T is the v avg
mann factor e 2E/k BT
vrms

hat lower than the

distribution curve
Nv

(21.42) v
dv
Maxwell-Boltzmann speed distribution (Skipping)

The energy of a molecule can be written:

E = Ktrans +  + U

where
p2
• translational kinetic energy, Ktrans = 2m
0

•  includes any rotational or oscillational energy

• U is potential energy (if relevant) that depends on the

location of the molecule

Since we only want to know about the distribution of speeds, we

will need to get rid of any dependence on  and U.
Aside: Reminder about probability distributions
(Skipping)
Suppose I have a probability distribution over two variables, x and
y:
p(x, y )

If the two variables are independently distributed then:

p(x, y ) = p(x)p(y )
Aside: Reminder about probability distributions
(Skipping)
Suppose I have a probability distribution over two variables, x and
y:
p(x, y )

If the two variables are independently distributed then:

p(x, y ) = p(x)p(y )

We can eliminate the dependence on x by just summing over x:

X X X 1
>

p(x, y ) = p(x)p(y ) = p(y ) p(x)
 = p(y )
x x

x

This is how we eliminate rotational and vibrational motion.

Maxwell-Boltzmann speed distribution (Skipping)
Put this expression in the Boltzmann distribution:

p(r, p, ) d3 r d3 p d = Ae −E /kB T d3 r d3 p d

Eliminate dependence on position, rotation, and oscillation:

Z Z
p(r, p, ) d3 r d3 p d
 p
 Z  Z 
−p2 /2m0 kB T 0 −/kB T 00 −U/kB T 3
=Ce dp C e d C e d r
 r
Maxwell-Boltzmann speed distribution (Skipping)
Put this expression in the Boltzmann distribution:

p(r, p, ) d3 r d3 p d = Ae −E /kB T d3 r d3 p d

Eliminate dependence on position, rotation, and oscillation:

Z Z
p(r, p, ) d3 r d3 p d
 p
 Z :1 Z
 :1

 00 
−p2 /2m0 kB T 0 −/k T −U/k T
 3
=Ce dp C e d C e d r
 B B

   r
Maxwell-Boltzmann speed distribution (Skipping)
Put this expression in the Boltzmann distribution:

p(r, p, ) d3 r d3 p d = Ae −E /kB T d3 r d3 p d

Eliminate dependence on position, rotation, and oscillation:

Z Z
p(r, p, ) d3 r d3 p d
 p
 Z :1 Z
 :1

 00 
−p2 /2m0 kB T 0 −/k T −U/k T
 3
=Ce dp C e d C e d r
B  B

   r

2 /2m k T
p(p) d3 p = Ce −p 0 B
d3 p
Maxwell-Boltzmann speed distribution (Skipping)
Put this expression in the Boltzmann distribution:

p(r, p, ) d3 r d3 p d = Ae −E /kB T d3 r d3 p d

Eliminate dependence on position, rotation, and oscillation:

Z Z
p(r, p, ) d3 r d3 p d
 p
 Z :1 Z
 :1

 00 
−p2 /2m0 kB T 0 −/k T −U/k T
 3
=Ce dp C e d C e d r
B  B

   r

2 /2m k T
p(p) d3 p = Ce −p 0 B
d3 p

replace momentum with velocity components:

2 2 2
p(v) d3 v = Ce −m0 (vx +vy +vz )/2kB T dvx dvy dvz
Maxwell-Boltzmann speed distribution (Skipping)
We can find C .
The total probability must equal one.
ZZZ
2 2 2
Ce −m0 (vx +vy +vz )/2kB T dvx dvy dvz = 1

Using the identity: Z∞

2 √
e −x dx = π
−∞
the three integrals can be evaluated separately:
Z∞ r
−m0 vx2 /2kB T 2πkB T
e dvx =
−∞ m0

There are three integrals, so

 3/2
m0
C=
2πkB T
Maxwell-Boltzmann speed distribution (Skipping)

Now our distribution is:

 3/2
3 m0 2 2 2
p(v) d v = e −m0 (vx +vy +vz )/2kB T dvx dvy dvz
2πkB T

Lastly, we want an expression for how many molecules have speeds

between v and v + dv.

This means we need to get rid of the direction dependence –

transform to spherical coordinates.

dvx dvy dvz = v 2 sin θ dv dθ dφ

Maxwell-Boltzmann speed distribution (Skipping)

Z Z
p(v) d3 v
φ θ
 3/2 Z 2π Zπ
m0 2 −m0 v 2 /2kB T
= v e dv dφ sin θdθ
2πkB T 0 0

 3/2
m0 2 /2k T
= 4π v 2 e −m0 v B
dv
2πkB T

This is the probability density for 1 molecule. For N molecules:

 3/2
m0 2 /2k
Nv dv = 4πN v 2 e −m0 v BT
dv
2πkB T
Summary

• Boltzmann distribution (energies)

• Maxwell-Boltzmann distribution (speeds)

Homework
• Full-solution HW2, due today
• WebAssign, due today
Serway & Jewett (additional problems you might like to look at):
• Ch 21, onward from page 644. Probs: 52, 58, 65
• new: Ch 21, onward from page 644. Probs: 41, 42, 43