The phrase “band edge” shows up everywhere in semiconductor physics, photonics, and excitonics, and it is easy to read past it as jargon. It deserves more attention than that. The band edge is the boundary in energy between the states electrons normally occupy in a material and the states they can reach when they absorb energy. Almost everything excitonic — absorption, emission, transport, sensing, lasing, photovoltaics — happens at or near that boundary. If you understand the band edge, you understand why a material is transparent to one color and opaque to another, why it glows at a particular wavelength, and why excitons form in some materials and barely matter in others.
This article is a foundational explainer. It defines the band edge, explains how it arises from band structure, distinguishes the different kinds of edges, and connects the concept to excitons, devices, and the rest of the Band Edge section of this site. It assumes no more than a general familiarity with atoms, electrons, and light.
Where bands come from
Start with a single isolated atom. Its electrons occupy discrete energy levels, the familiar orbitals of introductory chemistry. Bring two atoms together and those levels split. Bring together on the order of $10^{23}$ atoms in a crystal and the splitting becomes so dense that the levels merge into continuous ranges of allowed energy — energy bands — separated by ranges where no electron states exist, called band gaps.
In a semiconductor or insulator, the most important bands are the valence band, the highest band that is essentially full of electrons at low temperature, and the conduction band, the next band up, essentially empty at low temperature. The top of the valence band is the valence band edge. The bottom of the conduction band is the conduction band edge. The energy difference between them is the band gap, usually denoted $E_g$. The band edge, in the most common usage, refers to one or both of these two boundaries.
The band gap is what makes a semiconductor a semiconductor rather than a metal or an insulator. In a metal, the valence and conduction bands overlap, or a band is only partly filled, so electrons can move freely with almost no energy cost. In an insulator, the gap is so large — several electron-volts — that ordinary thermal energy or visible light cannot push many electrons across it. In a semiconductor, the gap is intermediate, roughly a fraction of an electron-volt to a few electron-volts, so light, heat, voltage, or doping can move electrons across it in controllable numbers. That controllability is the basis of nearly all modern electronics and photonics.
A useful rule of thumb, stated here as an approximation rather than a precise conversion, is that a band gap in electron-volts corresponds to a photon wavelength in nanometers roughly by dividing about 1240 by the gap. A gap near 1.1 eV, the textbook value for silicon at room temperature, corresponds to infrared light just beyond the visible range, which is why silicon is opaque to visible light but transparent in parts of the infrared. A gap near 2 to 3 eV corresponds to visible light, which is why wide-gap materials can be transparent and why they emit in the visible when they luminesce. Exact values shift with temperature, strain, composition, and crystal structure, so treat any single number as representative, not universal.
Direct and indirect edges
Not all band edges are alike. The distinction between a direct and an indirect band gap is one of the most consequential facts about a material, and it is drawn in momentum space as well as in energy.
Electrons in a crystal carry not only energy but crystal momentum, often labeled by a wavevector $k$. The band structure $E(k)$ plots energy against momentum. In a direct-gap semiconductor, the top of the valence band and the bottom of the conduction band occur at the same momentum. An electron can jump straight across the gap by absorbing a photon, because the photon carries energy but very little momentum. Gallium arsenide is the canonical textbook example of a direct-gap material, which is one reason it has been so important for lasers and LEDs.
In an indirect-gap semiconductor, the valence band maximum and conduction band minimum sit at different momenta. An electron crossing the gap must change both its energy and its momentum, and that momentum change has to come from a phonon — a quantized vibration of the crystal lattice. The process is still possible, but it is second-order and therefore much less efficient for light emission. Silicon is the canonical indirect-gap material. That is a central reason silicon dominates electronics but is a poor light emitter, and why silicon photonics has had to work so hard to coax light out of an indirect-gap host.
The direct-indirect distinction also matters for absorption. Just above the band edge, a direct-gap material typically absorbs strongly, because the transition is allowed without phonon assistance. An indirect-gap material absorbs more weakly near the edge, with the absorption rising more gradually as phonon-assisted processes become available. Device designers live with these shapes: they set how thick an absorber must be, how sharp an LED spectrum can be, and how much material is needed to capture sunlight.
The edge is not a sharp line
Textbook diagrams draw the band edge as a clean step. Real materials are messier, and the mess matters. Thermal motion, disorder, impurities, surfaces, and strain all smear the edge. Instead of zero states inside the gap and a sudden onset of states above it, real materials show a tail of states extending into the gap — often called a band tail or Urbach tail — with the density of states falling off roughly exponentially into the gap. The steepness of that tail is a sensitive measure of disorder: a clean, well-ordered crystal has a steep edge, while an amorphous or heavily disordered material has a broad, smeared one.
Defects add their own states inside the gap. A missing atom, a substitutional impurity, or a dangling bond at a surface can create a localized energy level that sits between the valence and conduction edges. These mid-gap states can trap carriers, provide nonradiative recombination paths that turn energy into heat instead of light, or — when engineered deliberately as dopants — supply the free electrons or holes that make a semiconductor n-type or p-type. The same physics that makes the edge imperfect is what makes doping possible.
Surfaces and interfaces deserve special mention. At the boundary of a crystal, the periodic lattice ends abruptly, and the band structure is perturbed. Surface states can pin the Fermi level, bend the bands near the surface, and change how a material absorbs or emits. In nanostructures — quantum dots, nanowires, two-dimensional monolayers — surfaces and interfaces are a large fraction of the whole, so edge effects that are minor in bulk crystals become dominant. Much of modern excitonics is, in effect, band-edge engineering at surfaces and interfaces.
What happens at the edge: absorption and emission
When a photon with energy below the band gap strikes a semiconductor, it generally passes through: there is no allowed transition for it to drive, so the material is transparent at that wavelength. When the photon energy reaches or exceeds the gap, it can promote an electron from the valence band to the conduction band, leaving behind a hole — a missing electron that behaves like a positively charged particle. That electron-hole pair is the fundamental excitation of a semiconductor. If the electron and hole remain bound to each other by their mutual Coulomb attraction, the pair is an exciton, the central object of this publication.
The energy of that exciton is slightly below the band gap, because the binding energy lowers the total. In absorption spectra, this often appears as a sharp peak just below the nominal edge, sometimes called the excitonic peak. In materials where the exciton binding energy is small compared to room-temperature thermal energy — the classic case is bulk gallium arsenide — excitons are easily ionized by heat, and the excitonic peak may be visible only at low temperature. In materials where the binding energy is large — organic semiconductors, monolayer transition metal dichalcogenides, and many two-dimensional systems — excitons survive at room temperature and dominate the optical response. The band edge therefore does not just set the absorption onset; it sets the stage on which excitons live or die.
Emission is the reverse process. An electron in the conduction band can fall back to the valence band and recombine with a hole, emitting a photon whose energy is close to the gap (or to the exciton energy, if the pair is bound). In a direct-gap material, this radiative recombination is efficient, which is why direct-gap semiconductors make good LEDs and lasers. In an indirect-gap material, radiative recombination requires phonon assistance and competes poorly with nonradiative paths, so emission is weak. The color of the emitted light, the sharpness of the spectrum, and the efficiency of the device all trace back to the character of the edge.
Tuning the edge
One of the most powerful ideas in semiconductor physics is that the band edge is not fixed. It can be tuned. Composition, strain, dimensionality, and electric fields all move it, and each tuning knob has become a technology.
Alloying is the oldest knob. Mixing two semiconductors — for example, varying the composition of an aluminum gallium arsenide alloy — shifts the gap continuously between the endpoints. This is how engineers design heterostructures with precisely chosen band offsets, confining electrons and holes in wells, wires, and dots. Quantum confinement is a second knob: when a semiconductor is shrunk to dimensions comparable to the exciton size, typically nanometers, the allowed energies shift upward and the effective gap widens. Quantum dots of the same material glow in different colors depending on their size, not their chemistry, because confinement moves the edge.
Strain is a third knob. Stretching or compressing a lattice changes the spacing between atoms and therefore the band structure. Strained layers are routine in modern transistors and lasers. Electric fields provide a fourth knob: a strong field tilts the bands and can shift the effective absorption edge, an effect exploited in electro-absorption modulators. Temperature is a fifth: most semiconductors show a band gap that shrinks modestly as temperature rises, which is why device spectra shift with heating. None of these effects requires changing the material itself; they all work by moving the edge.
Two-dimensional materials have made edge tuning especially vivid. In monolayer transition metal dichalcogenides such as molybdenum disulfide and tungsten diselenide, the band gap changes from indirect in the bulk to direct in the monolayer, and the exciton binding energy is large — on the order of hundreds of milli-electron-volts in textbook treatments — because reduced screening strengthens the electron-hole attraction. Stacking, twisting, and gating these layers moves the edge further and creates new excitonic species. A single materials family thus spans much of the physics this site covers, simply by manipulating where the edge sits.
The edge in devices
The band edge is not an abstraction that stays in textbooks. It determines device behavior in concrete ways. In a solar cell, the gap sets a trade-off: a smaller gap captures more of the solar spectrum but delivers less voltage per photon, while a larger gap captures less light but delivers more voltage. The classic single-junction efficiency analysis — the Shockley-Queisser treatment — is, at its heart, a statement about how the band edge interacts with the broad solar spectrum. Real cells add layers of complexity, but the edge sets the first-order limits.
In a photodetector, the gap sets the longest wavelength the detector can see. A detector meant for visible light needs a gap in the visible; one meant for infrared needs a smaller gap. In a light-emitting diode or laser, the gap sets the emission color. In a transistor, the gap — together with doping and band alignment — sets leakage, breakdown, and switching behavior. Even in devices that are not nominally “optical,” the edge matters: it governs how many carriers are thermally generated, how easily they tunnel, and how the device behaves at high temperature.
Heterostructures make the edge a design variable. When two semiconductors are joined, their band edges line up with offsets that can trap electrons on one side and holes on the other, or confine both in the same layer. Quantum wells, superlattices, and two-dimensional heterostructures are all built by arranging band edges in space. The exciton, which lives at the edge, inherits whatever landscape the designer builds: it can be confined, funneled, split, or made to interact with neighbors, depending on how the edges are arranged.
Why the band edge names this section
This site’s overflow section is called Band Edge for a reason. The four desks cover excitonic research, artificial intelligence and superintelligence, quantum computing with excitonic overlap, and devices and photonics. The Band Edge covers what does not fit neatly into any one of those — solar and energy, bio and sensing, special reports, and the methods by which the field moves. The name is a deliberate double meaning. Physically, excitons live at the band edge. Metaphorically, this section lives at the edge of the field: where excitonics meets neighboring disciplines, where applications are still forming, and where the frontier is being defined.
That is also why this article comes first in the section. Every other Band Edge piece — on solar cells, sensors, bio-photonics, and methods — assumes the picture built here: bands, gaps, edges, tails, excitons just below the edge, and the tuning knobs that move the edge. If a later article refers to a “direct-gap absorber” or a “band-tail state” or an “exciton peak just below the edge,” it is referring to this picture.
What to watch for when you read about edges
A few habits will make reading about band edges more reliable. First, ask whether a quoted gap is direct or indirect, and at what temperature. A gap quoted without those qualifiers is incomplete. Second, ask whether a claimed absorption or emission feature sits at the edge, below it (excitonic or defect-related), or well above it (band-to-band continuum). Third, be skeptical of a perfectly sharp edge in any real spectrum: tails, defects, and disorder always broaden it, and the amount of broadening is itself information about material quality. Fourth, remember that a gap is a property of a specific material in a specific form — bulk, thin film, monolayer, dot — not of a chemical formula alone. The same composition can have very different edges in different geometries.
Finally, remember that the band edge is a model, not a photograph. Band diagrams compress a complex quantum-mechanical reality into a readable picture. They are extraordinarily useful, and they are also simplifications: they hide electron-electron interactions, many-body effects, and the full momentum dependence of real bands. Excitons themselves are a reminder of that simplification, because they arise precisely from the Coulomb interaction that a single-particle band diagram leaves out. The edge is the right place to start. It is not the whole story, and the rest of this site exists to fill in what the diagram leaves out.
Sources and further reading
This article is a synthesis of standard semiconductor and excitonics textbook material. For readers who want to go deeper, the following are well-established entry points, cited by topic rather than by a single paper, because the underlying facts are textbook-level and appear across many sources.
- Introductory semiconductor band structure, direct vs. indirect gaps, and doping: any standard undergraduate semiconductor device text — for example, treatments in the style of S. M. Sze’s semiconductor device texts or B. G. Streetman’s solid-state electronic devices — and introductory solid-state physics texts in the style of Kittel or Ashcroft and Mermin. These cover valence and conduction bands, band gaps, and the momentum-space picture.
- Absorption edges, Urbach tails, and disorder: review literature on Urbach tails and band-tail states in semiconductors, and the broader literature on amorphous and disordered semiconductors. The exponential-tail picture is standard in that literature.
- Excitons and the excitonic peak just below the band edge: review articles on excitons in semiconductors, including the literature on excitons in two-dimensional transition metal dichalcogenides, where large binding energies and room-temperature excitons are a central theme. Standard reviews in journals such as Nature Reviews Materials, Reviews of Modern Physics, and Nature Photonics cover this ground.
- Solar-cell efficiency and the band-gap trade-off: the Shockley-Queisser treatment of single-junction efficiency limits, as covered in photovoltaics textbooks and review articles on solar-cell physics.
- Quantum confinement and size-tunable gaps: review literature on semiconductor quantum dots and nanocrystals, including the standard treatments of confinement shifting the effective gap.
- Strain, alloying, and heterostructure band alignment: the heterostructure and quantum-well literature, as covered in semiconductor heterostructure texts and device-physics reviews.
No single experimental value in this article is presented as a new measurement. Representative gap values (for example, silicon near 1.1 eV) are textbook figures at room temperature and vary with temperature, strain, and crystal form. Where a claim involves a frontier result rather than textbook physics, later articles in this section flag it explicitly.