Opening the book…
The free-electron model explains why metals conduct, and it even gives a sensible resistivity, but it cannot explain why diamond does not. Every solid is full of electrons. What decides whether they can carry a current?
Resistance in a metal. In Chapter 20 we pictured the conduction electrons of a metal colliding constantly with its ions, each collision erasing the small drift velocity the electric field had given the electron. If \(\tau\) is the mean free time between collisions, Drude’s model gave the resistivity
\[\rho = \frac{m}{ne^2\tau}\]
(equation (20.7)). We can now ask what that model leaves unexplained.
Example 42.4: How far does an electron go between collisions?
Copper has a resistivity of \(1.68\times 10^{-8}\,\mathrm{\Omega\cdot m}\) at room temperature. Using \(n = 8.49\times 10^{28}\,\mathrm{m^{-3}}\) and the Fermi speed of \(1.57\times 10^{6}\,\mathrm{m/s}\), find (a) the mean time between collisions and (b) the mean free path of a conduction electron. Compare it with the spacing of copper atoms, \(0.26\,\mathrm{nm}\).
Solution
(a) Solving \(\rho = m/(ne^2\tau)\) for \(\tau\),
\[\tau = \frac{m}{ne^2\rho} = \frac{9.109\times10^{-31}}{(8.49\times10^{28})(1.602\times10^{-19})^2(1.68\times10^{-8})} = \boxed{2.5\times 10^{-14}\,\mathrm{s}}.\]
(b) The electrons that carry current are those near the Fermi energy, moving at \(v_F\), so
\[\lambda = v_F\tau = (1.57\times10^{6})(2.5\times10^{-14}) = 3.9\times 10^{-8}\,\mathrm{m} = \boxed{39\,\mathrm{nm}},\]
about \(150\) atomic spacings. A classical electron weaving between the ions would hit one almost every spacing. The quantum explanation is that an electron is a wave, and a wave passes through a perfectly regular lattice without scattering at all, just as light passes undeflected through a perfect crystal. Electrons scatter only from what spoils the regularity: impurities, defects, and the thermal vibrations of the ions. Since the vibrations grow with temperature, the resistivity of a metal rises as it warms.
Energy bands. To see why some solids do not conduct at all, we need to take the ions seriously. Bring two identical atoms together, and each of their atomic energy levels splits into two slightly different levels, because the electron’s wave function can combine across the two atoms in two ways. Bring together six atoms, and each level splits into six. In a solid of \(N\) atoms, with \(N\) around \(10^{23}\), each atomic level splits into \(N\) levels so closely spaced that they form an effectively continuous energy band (Figure 42.3). The outer levels, whose wave functions overlap most, spread into the widest bands. Between the bands lie band gaps, ranges of energy in which no electron states exist at all.
Each band holds \(2N\) electrons for each atomic orbital it came from (the factor \(2\) is again spin), and the electrons fill the bands from the bottom up. Everything now depends on how the filling ends.
- Conductors. If the highest occupied band is only partly full, as in copper, there are empty states immediately above the filled ones. An applied field can nudge electrons into slightly higher-energy states moving in the direction of the force, and a current flows. The free electrons of Section 42.2 are the electrons of this partly filled band.
- Insulators. If the highest occupied band, the valence band, is completely full and the next band up, the conduction band, is separated from it by a wide gap, then no electron has an empty state within reach. A field cannot change any electron’s motion, because every nearby state is taken, and no current flows. In diamond the gap is \(5.5\,\mathrm{eV}\), two hundred times \(k_BT\) at room temperature.
- Semiconductors. A semiconductor is an insulator with a small gap, about \(1\,\mathrm{eV}\) (Table 42.2). At absolute zero it is a perfect insulator, but at room temperature a few electrons are thermally excited across the gap into the conduction band, where they can carry current.
Figure 42.4 compares the three. Graphite and diamond differ in exactly this way. In diamond, each carbon atom uses all four valence electrons in covalent bonds, filling the valence band. In graphite, each carbon bonds to only three neighbors within a flat sheet, and the fourth electron of each atom spreads across the whole sheet in a band that is not full, so graphite conducts along its sheets almost like a metal.
| Material | \(E_g\) (eV) | Material | \(E_g\) (eV) |
|---|---|---|---|
| Quartz (\(\mathrm{SiO_2}\)) | \(9\) | Gallium arsenide | \(1.42\) |
| Diamond | \(5.5\) | Silicon | \(1.12\) |
| Gallium nitride | \(3.4\) | Germanium | \(0.67\) |
| Silicon carbide | \(3.3\) | Indium antimonide | \(0.17\) |
The number of electrons that heat can lift across a gap depends on a Boltzmann factor. For a pure material the density of conduction electrons is proportional to
\[e^{-E_g/2k_BT}, \tag{42.5}\]
where the factor of \(2\) arises because each excitation creates two carriers (as we will see shortly) and the energy cost is shared between them.
Example 42.5: Diamond versus silicon
Compare the factor \(e^{-E_g/2k_BT}\) at \(300\,\mathrm{K}\) for (a) diamond and (b) silicon. (c) By what factor does silicon’s value grow if it warms to \(350\,\mathrm{K}\)?
Solution
At \(300\,\mathrm{K}\), \(2k_BT = 2(8.617\times10^{-5})(300) = 0.0517\,\mathrm{eV}\).
(a) Diamond: \(e^{-5.5/0.0517} = e^{-106} \approx \boxed{6\times10^{-47}}\).
(b) Silicon: \(e^{-1.12/0.0517} = e^{-21.7} \approx \boxed{4\times10^{-10}}\).
Silicon’s factor is about \(10^{37}\) times larger. A gap only five times smaller changes the result by thirty-seven powers of ten, because the gap sits in an exponent. This is why diamond is one of the best insulators known while silicon conducts measurably.
(c) At \(350\,\mathrm{K}\), \(2k_BT = 0.0603\,\mathrm{eV}\), so the factor becomes \(e^{-1.12/0.0603} = e^{-18.6} = 8.6\times10^{-9}\), larger by
\[\frac{8.6\times10^{-9}}{3.9\times10^{-10}} \approx \boxed{22}.\]
A modest warming increases silicon’s conductivity twentyfold. That is why the resistivity of a semiconductor falls steeply as it warms, the large negative temperature coefficient of silicon listed in Chapter 20, while a metal’s resistivity rises.
Band gaps also decide how a solid looks. A photon can be absorbed by exciting an electron across the gap only if \(hf \ge E_g\). Visible photons carry between \(1.8\) and \(3.1\,\mathrm{eV}\), too little to cross diamond’s \(5.5\,\mathrm{eV}\) gap, so diamond lets visible light through and is transparent. Silicon’s gap is smaller than any visible photon’s energy, so silicon absorbs all visible light. In metals, with empty states immediately above the filled ones, electrons can absorb photons of any energy and re-radiate them, which gives metals their shine.
Checkpoint 42.4
Pure silicon is opaque to visible light but transparent to infrared light of wavelength \(1.5\,\mathrm{\mu m}\), which is why it can be used for infrared lenses. Use \(E_g = 1.12\,\mathrm{eV}\) and \(hc = 1240\,\mathrm{eV\cdot nm}\) to explain both facts.
Answers to Checkpoints
Checkpoint 42.4. A photon is absorbed only if its energy is at least \(E_g\), enough to lift an electron across the gap. The threshold wavelength for silicon is \(\lambda = hc/E_g = 1240/1.12 \approx 1100\,\mathrm{nm}\). Every visible photon (\(400\,\mathrm{}\)–\(700\,\mathrm{nm}\)) has a shorter wavelength and more energy than this, so it is absorbed and silicon is opaque. A \(1500\,\mathrm{nm}\) photon carries only \(1240/1500 = 0.83\,\mathrm{eV}\), less than the gap, so it cannot be absorbed and passes through.
This page titled 42.3: Electrical Conduction in Metals and Insulators was last modified on Thu, 12 Nov 2026 04:27:08 GMT and is shared under a CC BY 4.0 license and was authored, remixed, and/or curated by Michael Ragsdale.
The book could not be loaded just now.