Field Note: The Effective Mass

Field Note · Trolla · Condensed matter physics

Effective mass is the simplest useful lie in condensed matter physics. You take an electron, you apply a field, you measure acceleration, and you pretend the electron is a bare particle with a different mass. It works so well that nobody complains.

The definition that actually matters

When a particle moves in a periodic potential, its dispersion $E(\mathbf{k})$ is generally nonlinear. Near a band extremum, you approximate:

$$E(\mathbf{k}) \approx E_0 + \frac{\hbar^2}{2} (\mathbf{k} - \mathbf{k}_0)^T \cdot \mathbf{M}^{-1} \cdot (\mathbf{k} - \mathbf{k}_0)$$

The tensor $\mathbf{M}$ is the effective mass tensor. Its components encode how the particle responds to forces along different crystallographic directions. In a cubic crystal along a high-symmetry direction, $\mathbf{M}$ reduces to a scalar $m^*$, and the equation of motion takes the familiar form:

$$\mathbf{F} = m^* \mathbf{a}$$

This is not Newton's second law. Newton's second law applies to a bare particle. What you have here is a quasiparticle — the electron dressed by the lattice potential — and $m^*$ is a parameter in an equation that mimics Newton's law. The distinction matters when you push the system hard enough that the parabolic approximation breaks down, but for most purposes in semiconductor physics, the parabolic band is an excellent approximation near the band edge.

When $m^*$ is negative

Here is the thing that still strikes me as magical every time I teach it: the effective mass can be negative. An electron near the top of a valence band responds to an electric field as if it had negative mass. The physical explanation is that the electron is actually a hole — the absence of an electron — and holes carry positive charge. But you do not need to invoke holes explicitly. The band structure itself gives you a negative curvature, and the negative curvature gives you a negative effective mass, and the equations of motion do the rest.

Negative effective mass is not a mathematical artifact. It is measurable. In cyclotron resonance, the sign of $m^$ determines the direction of precession. In Hall measurements, the sign of the Hall coefficient tells you whether the carriers are electron-like ($m^ > 0$) or hole-like ($m^* < 0$).

The $kp$ connection

The effective mass is not an independent parameter — it is determined by the band structure, which is in turn determined by the crystal potential. The $kp$ method (not $k \cdot p$ as the notation suggests — it is $k$-vector times $p$-matrix-element, but nobody writes it that way) relates $m^*$ to interband matrix elements:

$$\frac{m_e}{m^*} = 1 + \frac{2}{m_e} \sum_{n \neq c} \frac{|\langle c | \mathbf{p} | n \rangle|^2}{E_c - E_n}$$

This formula, due to Kane, makes explicit what effective mass is: a measure of how strongly the band you are in couples to all the other bands. A large interband coupling means a small effective mass (the electron is "light" because it can easily tunnel into other bands and accelerate). A small interband coupling means a large effective mass (the electron is "heavy" because it is trapped in its band).

Experimental determination

There are three standard methods:

1. Cyclotron resonance: Apply a magnetic field perpendicular to an electric field. The absorption peak frequency $\omega_c = eB/m^$ gives $m^$ directly. This is the most direct measurement.

2. Shubnikov–de Haas oscillations: Quantum oscillations in magnetoresistance reveal the extremal cross-section of the Fermi surface, from which $m^*$ follows via the temperature dependence of the oscillation amplitude.

3. Optical absorption: The absorption edge in a direct-gap semiconductor shifts with $m^*$ because the joint density of states depends on the reduced effective mass of electron-hole pairs.

When the effective mass concept breaks down

The effective mass approximation works when:

• You are near a band extremum (parabolic band).
• The wavepacket is larger than the lattice spacing (long-wavelength limit).
• External fields vary slowly compared to the lattice period.

It breaks down when:

• You are near a band anticrossing (non-parabolic dispersion).
• The field is strong enough to induce interband tunneling (Bloch oscillations).
• The crystal is disordered enough that localization sets in (the concept of a well-defined $m^*$ loses meaning).

In all of these cases, the failure is not a failure of physics. It is a failure of approximation. The underlying Schrödinger equation is still correct; the effective mass parameter is just no longer adequate.