KALMAN KNIZHNIK - DERIVATION OF THE STEFAN-BOLTZMANN LAW

Here we derive the Stefan-Boltzmann law, which describes the energy density spectrum of blackbody
radiation. We begin by modeling the energy spectrum of a photon gas as a simple harmonic
oscillator, with energy levels given by
1
En = h̄ω(n + ) (1)
2
The partition function for the simple harmonic oscillator is easily derived from this energy spectrum:
∞ ∞
1 e−βh̄ω/2
e−βh̄ω(n+ 2 ) = e−βh̄ω/2 e−βh̄ωn =
X X
Z= (2)
n=0 n=0
1 − e−βh̄ω

The average thermal energy of the photon gas is, then,

d(lnZ) h̄ω h̄ω
hU (ω)i = − = + βh̄ω (3)
dβ 2 e −1
To obtain the total energy, we need to integrate equation 3 over all possible frequencies. Thus,
Z ∞
U= hU (ω)ig(ω)dω (4)
0

where g(ω) is the multiplicity of the energy for a given frequency. In other words, there may be
several frequencies that each have the same energy. To derive what g(ω) is, we need to calculate
the density of states of the wavenumber k:
V 2 V k2
g(k) = 2 × 4πk = (5)
(2π)3 π2
where the factor of 2 corresponds to the two possible polarizations of the photons, the factor
of V is meant to restrict the number of wave numbers to a box of volume V , and the (2π)3 is
generally conventional. Finally, the 4πk 2 corresponds to the area of a sphere of radius k. Thus,
g(k) represents the number of wave vectors that correspond to the value k. To obtain g(ω) from
this, we note that g(ω)dω = g(k)dk, and so

dk V k2 V ω2
g(ω) = g(k) = 2 = 2 3 (6)
dω π c π c
where in the last two equalities I have used the fact that ω = ck. We can now rewrite equation 4
as
h̄ω V ω 2
Z ∞
h̄ω
U= [ + βh̄ω ]dω (7)
0 2 e − 1 π 2 c3
We immediately see a problem, in that the first integral, which corresponds to the energy of the
vacuum, obviously diverges. Oops. Fortunately, some smart people decided that we can redefine
our energy scale such that the energy of the vacuum is zero, meaning that we are ignoring this
infinite result. Odd, but it works. This integral now reduces to
ω3
Z ∞
h̄V
U= dω (8)
π 2 c3 0 eβh̄ω − 1
If we rewrite this in terms of the energy density, this is:
ω3
Z ∞ Z ∞
U h̄
u≡ = uω dω ≡ dω (9)
V 0 π c3
2
0 eβh̄ω−1
where
h̄ ω3
uω = (10)
π 2 c3 eβh̄ω−1
is known as a blackbody distribution. We can integrate equation 9 by changing variables, and
letting x = βh̄ω. Then dω = dx/βh̄. The integral becomes:

x3 k4 T 4 π 2 kB
3
Z ∞
1 1
u= dx = 2B 3 2 Γ(4)ζ(4) = T4 (11)
π 2 c3 h̄2 β 4 0 ex−1 π c h̄ 2
15h̄ c3
This equation tells us that the energy density of a photon gas depends only on the temperature,
and specifically on the fourth power of the temperature. We are almost at the Stefan-Boltzmann
law. To get there, we quote a result from kinetic theory, namely that the flux of particles striking
a unit area of a container is
1 nc
Φ = nhvi = (12)
4 4
where I have replaced v with c, and using the fact that power emitted is the energy h̄ω times the
flux, we get
nh̄ω
P = h̄ωΦ = c (13)
4
But nh̄ω = u is the energy density (remember that here n is not the quantum number but the
number density). Thus, we arrive at the Stefan-Boltzmann Law

uc π 2 kB
3
P = = T 4 ≡ σT 4 2 (14)
4 60h̄2 c2
where σ is the Stefan-Boltzmann constant.

Alternatively, we can skip the whole blackbody argument, and simply quote the result, also from
kinetic theory, that the pressure of a gas of particles with number density n is
1 1
p = nmhv 2 i ⇒ pγ = nmc2 (15)
3 3
and we can finterpret mc2 as the energy as the energy of a single photon. Then nmc2 is the energy
density of a photon gas, so
u
pγ = (16)
3
Now we use the first law of thermodynamics, dU = T dS − pdV to obtain

∂U ∂S ∂p
=T −p=T −p (17)
∂V T ∂V T ∂T V
where the last equality follows from a Maxwell relation. Now, the left hand side is simply the
energy density u, and plugging this in above, along with equation 16 gives

T ∂u u du dT
u= − ⇒ = (18)
3 ∂T V 3 4u T

leading immediately to u = AT 4 , with A as some constant of integration. We then use equation

13, noting that nh̄ω ≡ u, to arrive, once again, at
1 1
P = uc = ( Ac)T 4 ≡ σT 4 2 (19)
4 4
which is again the Stefan-Boltzmann law. This derivation does not tell us what the constant is,
but it is good enough to tell us the temperature dependence of the power emitted by a blackbody.