Skip to content
Tutorials Center
My learning

Topics

Lesson 25 of 39 · Quantum Statistics: Fermi-Dirac and Bose-Einstein

The Degenerate Fermi Gas at Finite Temperature

Compute the chemical potential and electronic heat capacity of a degenerate Fermi gas, derive the linear-in-T law with the Sommerfeld idea, and check it by numerical integration.

At room temperature a metal's electrons sit in a Fermi-Dirac distribution that is only slightly smoothed compared with the perfect step of absolute zero. That small smoothing is the whole story of the electron's contribution to the heat capacity, and it explains a puzzle from the early 1900s: why electrons, which classical theory said should carry (3/2) kT each, add almost nothing to a metal's specific heat.

Only the surface of the Fermi sea matters

When the temperature rises from zero, the step in f(ε) = 1/(e^((ε-μ)/kT) + 1) blurs over an energy range of a few kT around μ. Electrons deep below μ have no empty states to jump to (every state nearby is full), so they cannot absorb thermal energy. Only electrons within about kT of the Fermi surface participate. Their fraction is ~ T/T_F, and each gains about kT, so

Text
ΔU ~ N (T/T_F) · kT  ∝  T²      →      C = dU/dT ∝ T

This is the physical reason for a heat capacity linear in T and small compared with the classical (3/2) N k.

The Sommerfeld results

Sommerfeld made this precise with an expansion that is accurate when kT ≪ E_F. Integrals of the form ∫ φ(ε) f(ε) dε equal the zero-temperature integral plus corrections of order (kT)². For the three-dimensional free gas, the results are:

Text
μ(T) = E_F [ 1 - (π²/12) (kT/E_F)² + ... ]
U(T) = (3/5) N E_F [ 1 + (5π²/12) (kT/E_F)² + ... ]
C_V  = (π²/2) N k (T/T_F)  =  γ T,     γ = (π²/2) N k / T_F

For one mole, N k = R. Notice that μ decreases slightly with T, but only by a part in 10⁵ at room temperature in a metal.

The ratio of the electronic heat capacity to the classical value is C_V / ((3/2) N k) = (π²/3)(T/T_F). For copper at 300 K this is about 0.012, so electrons contribute only about 1 percent of what classical theory predicted.

Computing μ(T) exactly

The exact μ(T) comes from requiring that the number of particles is conserved: N = ∫ g(ε) f(ε; μ, T) dε. In units where E_F = 1 and g ∝ √ε, this becomes ∫₀^∞ √ε f dε = 2/3. The code solves this for μ at several temperatures and compares with the first Sommerfeld correction.

Python Chemical potential of a Fermi gas versus temperature

Press Run (or Ctrl + Enter) to see the result here.

Program input · what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

For small kT/E_F the expansion is excellent; as T approaches T_F, μ crosses zero and becomes negative, which is what you expect when the gas becomes classical (μ is large and negative in a dilute classical gas).

The linear heat capacity

Next we compute U(T) numerically with the exact μ(T) and differentiate to get C_V. In units with N k = 1 and T measured in T_F, the Sommerfeld law reads C/(Nk) = (π²/2) t. Finally the code converts to a molar coefficient γ for copper using T_F = E_F/k from the previous lesson.

Python Electronic heat capacity: numerical derivative versus (π²/2) t

Press Run (or Ctrl + Enter) to see the result here.

Program input · what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

The numerical derivative follows (π²/2) t closely for small t, with deviations of about 3 percent at t = 0.1 that grow to about 15 percent at t = 0.2, where the next term of the expansion matters. For copper the free-electron model gives γ of about 0.5 mJ/(mol·K²). Measured values are closer to 0.7 mJ/(mol·K²); the difference is attributed to electron-lattice and electron-electron interactions, which are folded into a thermal effective mass m* slightly larger than m_e.

Why this matters

  • Heat capacity of metals at low temperature is written C = γT + A T³, the first term from electrons and the second from phonons (the Debye law from earlier). Plotting C/T against T² gives a straight line whose intercept is γ.
  • The same physics governs electrons in white dwarfs, neutrons in neutron stars and ultracold fermionic atoms (with different densities).
  • In the grand picture, the Fermi sea is a nearly rigid object whose low-energy excitations (particle-hole pairs) are a gas of their own, a first example of the quasiparticle idea.

Check your understanding

Quick check · 4 questions

of 4 answered You scored / 4 · perfect!

  1. 1. Why is the electronic heat capacity of a metal much smaller than 3Nk/2?

    Deep states are completely filled and cannot change, so only a fraction of order T/T_F of electrons contribute.
  2. 2. The Sommerfeld result for the electronic heat capacity is

    C = (π²/2) N k (T/T_F).
  3. 3. How does μ change as the temperature of a degenerate Fermi gas is raised slightly?

    μ = E_F[1 - (π²/12)(kT/E_F)²], so it decreases by a tiny amount.
  4. 4. A plot of C/T versus T² for a metal at low temperature gives the electronic coefficient γ as

    C/T = γ + A T², so the intercept is γ and the slope is the phonon coefficient A.

Summary

  • Thermal smearing of the Fermi step affects only electrons within about kT of E_F.
  • Sommerfeld: μ = E_F[1 - (π²/12)(kT/E_F)²], C_V = (π²/2) N k T/T_F = γT.
  • Electrons contribute a few percent or less of the classical heat capacity at room temperature; at low temperature the γT term dominates the lattice T³ term.

Next: Blackbody Radiation: The Photon Gas