An exciton is what happens when a semiconductor absorbs light and does not immediately forget it. A photon disappears, an electron is promoted, and the empty state it leaves behind — the hole — stays attracted to that electron. The two orbit each other, bound by Coulomb attraction, moving together through the crystal as a single neutral entity. That entity is the exciton.
It is not a fundamental particle. It is a quasiparticle, a useful description of a collective excitation inside a material. But the description earns its keep: excitons govern how many semiconductors glow, how they absorb, how they respond to electric fields, and how light can be slowed, stored, or converted inside a device. If you want to understand excitonic devices, photonics, or quantum optics in solids, you start here.
From photon to electron-hole pair
In a semiconductor at equilibrium, the valence band is largely full of electrons and the conduction band is largely empty, separated by a band gap. When a photon with energy near or above the gap arrives, it can promote an electron from the valence band into the conduction band. Conservation of energy and momentum constrain which transitions are allowed, but the basic picture is simple: light creates an electron in the conduction band and a hole in the valence band.
Left alone, the electron and hole would be free carriers, able to drift apart under an electric field and contribute to current. In many materials, however, they do not get far before their mutual attraction matters. The electron carries negative charge; the hole behaves as a positive charge. They attract. In a bulk crystal with strong dielectric screening, that attraction may be weak and the pair may quickly dissociate, especially at room temperature. In materials with weaker screening — in quantum wells, wires, and especially in atomically thin two-dimensional crystals — the attraction is much stronger, and the bound state becomes the dominant way the material handles light.
The exciton is therefore not an exotic exception. It is the natural intermediate between light in and light out. Absorption creates it; emission annihilates it; transport, scattering, and interactions happen while it lives.
A hydrogen atom inside a crystal
The most useful mental model is the hydrogen atom, transplanted into a semiconductor. In hydrogen, a proton binds an electron. In an exciton, a hole binds an electron. The mathematics is closely analogous, with two important modifications. First, the electron and hole do not have their free-space masses. Inside a crystal, their response to forces is captured by effective masses that reflect the curvature of the bands they live in. Second, the Coulomb attraction is screened by the material itself. The dielectric constant reduces the force between the charges, and also expands the size of the orbit.
The result is a hydrogen-like series of bound states, often labeled 1s, 2s, 3s, and so on, with binding energies that scale roughly as one over the square of the principal quantum number in the simplest model. The ground state, 1s, is the most tightly bound and the most commonly observed. Higher states are more weakly bound, larger, and more fragile. In absorption spectra, this series appears as a set of resonances just below the free-particle band gap — a Rydberg-like ladder that is one of the clearest signatures that excitons are present.
The binding energy — the energy needed to pull the electron and hole apart into free carriers — is the single most important number for an exciton. It sets whether excitons survive at a given temperature, whether they can be split by an electric field, and whether a device sees excitonic effects or free-carrier effects. In conventional bulk semiconductors such as gallium arsenide, binding energies are on the order of a few milli-electron-volts, far below room-temperature thermal energy, so excitons are typically a low-temperature phenomenon in bulk form. In monolayer transition metal dichalcogenides, binding energies are reported to be hundreds of milli-electron-volts, an order of magnitude larger, which is why excitonic features dominate their optical spectra even at room temperature. The precise values depend on material, thickness, substrate, and dielectric environment, and researchers routinely report ranges rather than single universal numbers. The qualitative contrast — weak binding in bulk, strong binding in two dimensions — is well established.
The size of the exciton, often called the Bohr radius, follows the same logic in reverse. Strong screening and light effective masses produce large, diffuse excitons, sometimes called Wannier-Mott excitons, that extend over many lattice constants. Weak screening and heavier masses produce small, tightly localized excitons, closer to the Frenkel limit familiar from molecular crystals and organic semiconductors. Most inorganic semiconductors sit toward the Wannier-Mott end; most organic materials sit toward the Frenkel end. Two-dimensional semiconductors sit in an interesting middle ground: strongly bound, yet still extended over several nanometers, large enough to feel the crystal but small enough to be robust.
Bright, dark, and momentum
Not all excitons can emit light. For an exciton to recombine radiatively — to annihilate and produce a photon — the electron and hole must satisfy spin and momentum selection rules. When they do, the exciton is called bright. When they do not, it is called dark.
The simplest version of the rule concerns momentum. A photon carries very little momentum compared to crystal momenta, so direct recombination is efficient only when the electron and hole have nearly the same crystal momentum. In a direct-gap semiconductor, the conduction-band minimum and valence-band maximum occur at the same point in momentum space, and the lowest-energy exciton can be bright. In an indirect-gap material, they occur at different points, and the lowest-energy exciton needs a phonon to supply the missing momentum, making radiative recombination slower and weaker.
Spin adds a second layer. The electron and hole each carry spin, and the combined spin state determines whether the optical transition is allowed. Singlet-like configurations are typically bright; triplet-like configurations are typically dark, or only weakly allowed through mixing. In materials with strong spin-orbit coupling — notably the transition metal dichalcogenides — the spin and valley structure produces a richer set of bright and dark states, and dark excitons can lie below bright ones in energy, acting as reservoirs that trap population and shape how light is emitted over time.
Dark excitons are not irrelevant because they do not glow. They store excitation, they affect lifetimes, and they can be brightened by magnetic fields, strain, or coupling to cavities and plasmons. In some experiments, understanding the dark states is the key to understanding why the bright emission behaves the way it does.
What an exciton does while it lives
An exciton is born with the momentum and energy of the photon that created it, plus whatever the material adds through scattering. It then moves, scatters, and interacts until it recombines or dissociates.
Its lifetime is set by competing rates. Radiative recombination produces a photon. Nonradiative recombination loses the energy as heat, often at defects, surfaces, or through Auger-like processes where the energy is transferred to another carrier. Dissociation splits the pair into free carriers, which may then be collected as current in a solar cell or photodetector, or may recombine separately.
Transport matters too. Because the exciton is neutral, it does not drift in a uniform electric field the way a free electron does. It diffuses, driven by density gradients and by energy gradients in the material. Strain, thickness variations, and electrostatic potentials can funnel excitons toward particular regions. That funneling is both a tool — researchers use it to concentrate excitons — and a complication, because disorder in real materials creates a landscape of hills and valleys that excitons must navigate.
Interactions between excitons add further richness. At low density, excitons behave roughly like independent bosons. At higher density, they scatter off each other, shift each other’s energies, and can form bound complexes such as trions and biexcitons, which are covered in the companion article on the exciton zoo. At very high density, excitons can overlap, lose their identity, and dissolve into an electron-hole plasma. The transition from a gas of excitons to a plasma — the Mott transition — is a classic problem in semiconductor physics, and it sets a practical limit on how dense an exciton population can be before the quasiparticle picture breaks down.
Why excitons matter for devices
Excitons matter because they are the bridge between optics and electronics in semiconductors. Any device that converts light to electricity or electricity to light must pass through the electron-hole pair, and in many materials that pair is excitonic for at least part of its life.
In light-emitting diodes and lasers, excitons are the precursors to photons. In photodetectors and solar cells, they are the intermediate that must be split before current can be collected — a step that is efficient when binding is weak or when interfaces provide an energetic driving force, and a challenge when binding is strong. In modulators and switches, excitons provide strong, tunable optical resonances that can be shifted by electric fields, strain, or charge density. In quantum optics, single excitons in quantum dots act as single-photon emitters, and excitons coupled to cavities form polaritons that blend light and matter.
The two-dimensional materials have sharpened this relevance. Because their excitons are strong at room temperature and sensitive to their environment, they offer a platform where excitonic physics is not a cryogenic curiosity but a room-temperature engineering variable. Encapsulation, gating, twisting, and stacking all tune the exciton. That tunability is the reason excitons have moved from textbook chapter to device roadmap.
How we know they are there
The most direct evidence comes from spectroscopy. In absorption, excitons appear as sharp peaks below the band gap. In photoluminescence, they appear as emission peaks whose energy, linewidth, and temperature dependence track the bound state. In reflection, they produce distinctive derivative-like features. When an exciton series is resolved — 1s, 2s, 3s — the spacing and intensities provide a quantitative fingerprint that is difficult to explain without bound electron-hole pairs.
Other signatures corroborate. The binding energy can be inferred from the separation between the exciton peak and the onset of free-carrier absorption. Electric-field tuning — the quantum-confined Stark effect — shifts exciton energies in a characteristic way. Magnetic fields split and shift exciton states according to their angular momentum and magnetic moment. Time-resolved measurements reveal lifetimes and diffusion that differ from free carriers. No single measurement proves the exciton picture alone, but together they form a consistent, predictive account.
What the exciton is not
It is worth clearing up three common misunderstandings. First, an exciton is not a stable particle you can put in a bottle. It is a transient excitation that lives for picoseconds to nanoseconds in most materials, longer in special cases, and then disappears. Second, an exciton is not the same as an electron-hole plasma. The plasma is the dissociated, high-density limit; the exciton is the bound, low-density limit. Third, an exciton is not just a fancy name for absorption. Absorption can produce free carriers directly. The exciton is a specific, bound outcome of absorption, with its own spectrum, dynamics, and interactions.
The road ahead
The foundational picture — hydrogen-like binding, bright and dark states, a competition between recombination and dissociation — has been stable for decades. What has changed is the material landscape and the control available. Two-dimensional crystals, moiré superlattices, and cavity structures have made excitons stronger, more tunable, and more accessible at practical temperatures. They have also revealed new species and new phases that the simple hydrogen model does not capture.
That is why this series starts here and then branches: to trions and biexcitons, to interlayer and moiré excitons, to valley physics, to polaritons, and to the measurement tools that let us watch excitons move and lose coherence. The exciton itself is simple to state — an electron and a hole, bound. Its consequences are the rest of the field.
Sources and further reading
For the hydrogen-like model, binding energies, and Wannier-Mott versus Frenkel limits, see standard semiconductor optics textbooks such as those by Klingshirn, by Yu and Cardona, and by Fox on optical properties of solids. For excitons in two-dimensional transition metal dichalcogenides, see review articles in Reviews of Modern Physics, Nature Reviews Materials, and Nature Photonics on excitons in monolayer semiconductors, which survey binding energies, bright and dark states, and room-temperature excitonic spectra. For the Mott transition and electron-hole plasma physics, see review literature on dense electron-hole systems in semiconductors. For single-photon emission from quantum-dot excitons and cavity polariton physics, see reviews in Nature and Reviews of Modern Physics on solid-state quantum optics and microcavity polaritons. Where this article gives orders of magnitude rather than single values — for example, binding energies in bulk versus monolayer materials — that reflects the spread reported across materials, substrates, and measurement methods in that review literature, not a single canonical measurement.