Metals, Semiconductors and Insulators
Band filling, gap size and the classification of solids
Lesson 3893 of 4,500 · Solid-State and Materials Chemistry
Learning objectives
- Classify simple band structures by filling and gap
- Explain how temperature and doping alter semiconductor conduction
- Identify cases where a one-electron band classification is incomplete
Introduction
Copper conducts electricity readily, silicon can be controlled by tiny impurity concentrations, and diamond is normally an electrical insulator. Band theory connects these differences to allowed electronic states and how many electrons occupy them. Yet the labels metal, semiconductor and insulator describe behaviour under stated conditions , not merely one number on a diagram. A material may have electronic bands but transport can also depend on carrier mobility, disorder, defects and strong interactions.
Core explanation
In a simple one-electron band picture, a metal has occupied and unoccupied electronic states arbitrarily close in energy at its chemical potential. This can occur when a band is partly filled or when filled and empty bands overlap. A weak field can redistribute occupancy near the Fermi surface, producing a current if carriers are mobile. A large density of states at the Fermi level is not automatically a high conductivity: mobility and scattering also matter. Heating a conventional metal usually raises phonon scattering and often increases resistivity, even though its carrier density changes comparatively little.
A gapped material has a filled valence band and an empty conduction band at zero temperature in the ideal picture. If the gap is moderate, thermal energy, light or dopants can create mobile electrons in the conduction band and holes in the valence band; this is a semiconductor. A much larger gap generally produces very few electronic carriers under ordinary conditions and is called an insulator. There is no single universal energy-gap cutoff independent of temperature, device requirements and impurities. A narrow-gap material can act more metallic when heated, and a very pure semiconductor at low temperature can conduct extremely poorly.
The intrinsic semiconductor's electron and hole concentrations both rise with temperature. Under a nondegenerate approximation they contain a leading activated factor related to exp[−E g/(2k BT)] with additional density-of-states and temperature factors. One should not claim the exact carrier number is simply exp(−E g/k BT). Dopants can instead supply carriers at much lower energies than crossing the full gap. Donors usually contribute electrons, making n-type material; acceptors create holes, making p-type material. At high dopant concentrations, impurity bands, screening and degeneracy complicate the elementary picture.
Counting states matters as much as naming a gap. Suppose a band from N spatial orbitals can accommodate 2N electrons because of spin. N electrons half-fill it in the simplest noninteracting model. With 2N electrons it is full; if the next band is separated by a gap and no other bands overlap, the ideal zero-temperature system is insulating in this band picture. The same element can form different structures with different band widths and fillings. The MIT solid-state chemistry lecture develops this orbital-to-band reasoning, and OpenStax's band-theory chapter compares conductor, semiconductor and insulator diagrams.
Real solids challenge the neat classification. A Mott insulator can have a partly filled band in a simple calculation yet fail to conduct because strong electron-electron repulsion localises charge. Disordered materials may have localised electronic states at energies where the DOS is nonzero. Ionic solids can conduct by moving ions even if their electronic band gap is large. Superconductors have zero dc electrical resistance below a transition temperature but are not described by the ordinary metal/insulator comparison alone. Therefore specify whether a statement concerns electronic or ionic conduction and whether it relies on a one-electron model.
Optical behaviour gives complementary evidence. A wide-gap electronic insulator may be transparent to visible light if other absorptions are absent; a semiconductor absorbs photons above suitable transition energies. Direct and indirect gaps can produce different optical onsets because momentum conservation affects transition probability. Colour does not determine conductivity uniquely: defects can colour an insulating crystal without turning it into a good electronic conductor.
Step-by-step reasoning
1. Count electrons and states in the relevant bands at the specified temperature. 2. Locate the chemical potential relative to band edges and any band overlaps. 3. If a gap exists, estimate whether heat, light or dopants can produce carriers. 4. Consider mobility and scattering, not carrier count alone. 5. Ask whether correlation, disorder or ionic motion invalidates a simple electronic-band conclusion.
Visual explanation
Draw three energy diagrams. A metal has a horizontal Fermi line cutting through a band. A semiconductor has a modest empty interval between a filled valence band and conduction band; show a thermally excited electron above and a hole below. An insulator has a wider gap, with virtually no thermal carriers under the same conditions. Beside all three draw a small arrow labelled mobility to remind the reader that available carriers still need to move.
Real-world analogy
A parking garage represents available electronic states. A completely filled lower floor cannot rearrange its cars easily if every nearby bay is taken. A partly filled floor has neighbouring spaces and cars can respond to direction signs. A separate upper floor requires a ramp of sufficient energy; a small gap is a low ramp and a large gap a high one. Congested roads outside the garage represent scattering and mobility.
Real-world example
Silicon's moderate electronic gap and controllable dopants make it useful in transistors. Phosphorus substitution can act as a donor in suitable silicon crystals, increasing mobile electrons under operating conditions. The useful switching behaviour comes from controlling carrier concentrations and junction fields; it is not explained solely by saying that silicon has a gap.
Why?
Why is a filled band usually unable to produce ordinary low-field current by itself? In equilibrium, velocities of occupied states cancel. Small shifts in occupation would require empty nearby states, but the band is already full. Exciting carriers across a gap or into another band is then necessary in the elementary model.
Common misconception
"A solid with a nonzero band gap can never conduct." Semiconductors conduct when carriers are thermally generated, photoexcited or supplied by dopants. Even a wide-gap material may conduct ions or exhibit leakage through defects; the gap alone does not specify all transport mechanisms.
Worked example
Question: Two ideal crystals each have N unit cells and one spin-degenerate band formed from one orbital per cell. Crystal A has N electrons in that band. Crystal B has 2N electrons, and its next band begins 2 eV higher. Classify their zero-temperature electronic behaviour in this simple noninteracting model.
Reasoning: Each band has N spatial states and can hold 2N electrons including spin. Crystal A has half the capacity filled, leaving nearby empty states in the same band, so it is metallic in the basic model. Crystal B fills the band completely. With its next band separated by 2 eV and no overlaps, it is an electronic band insulator at zero temperature. Actual conductivity could differ if correlations, defects or mobile ions matter.
Answer: A is metallic and B is a gapped band insulator within the stated idealisation.
Quick check
1. Does a partly filled band always guarantee high measured conductivity? Answer: No. Strong scattering, disorder or electron correlation can suppress carrier mobility or localise charge.
Exam focus
Distinguish partly filled, overlapping and gapped bands. Explain semiconductor conduction with electrons and holes and show how doping changes carrier numbers. Do not define a universal numeric gap boundary between semiconductor and insulator. State the model assumptions for any classification.
Advanced insight
Electronic conductivity can be written schematically as σ = e(nμ e + pμ h) for electrons and holes in a simple semiconductor. Carrier concentration and mobility each change with temperature, often in opposite ways. A conductivity trend alone therefore cannot identify a gap without further measurements. Hall effect, optical absorption and temperature dependence can jointly constrain the carrier picture.
Summary
Metals have accessible states near the chemical potential in basic band theory. Semiconductors and insulators have gaps but differ in how readily useful mobile carriers appear under specified conditions. Filling, gap, temperature, doping, mobility and possible correlation or ionic conduction must all be considered when assigning behaviour.
Practice questions
1. How can a band be metallic without being half filled? Answer: Any partial filling or overlap between bands can leave occupied and empty states near the chemical potential. 2. What two types of electronic carrier are generated by excitation across a semiconductor gap? Answer: Conduction-band electrons and valence-band holes. 3. Why can a wide-gap ionic solid conduct under some conditions? Answer: Ions may migrate through defects even while electronic conduction remains small. 4. What is one reason a partly filled band could nevertheless be insulating? Answer: Strong electron correlation or disorder can localise the electrons.