← all systems · see it in The System
Exciton
A bound state of an electron and an electron hole held together by their Coulomb attraction; an electrically neutral quasiparticle, the elementary excitation made when a solid, molecule or pigment aggregate takes in light, that carries energy without carrying net charge until it gives the energy up as light, heat, free charges or a hand-off to a neighbour. It runs from small Frenkel excitons held on one site, through charge-transfer excitons, to large hydrogen-like Wannier-Mott excitons spread over many unit cells, with mixed kinds between.
Properties published scienceour proposal
Sourced fields, and our own readings.
| boundary | One bound electron-hole pair (or one correlated excitation of a molecule or pigment aggregate) from the act that makes it (absorption of a photon, capture of an injected electron and hole) to the act that ends it (emission, decay to heat, dissociation into free carriers, hand-off to another site). Inside: the electron, the hole, their binding and relative motion, the pair's centre-of-mass motion and spin. Outside: the host lattice and its phonons, the dielectric surroundings that screen the attraction, the host's magnetic order and magnons, light and cavity modes, defects, other excitons. The boundary is soft in size: from about a nanometre (ground state in Cu2O; CrSBr's XA, mostly on-site on chromium and about 1.2 nm (some 2.5 lattice spacings) along b) to more than 2 micrometres (n=25 in Cu2O) (src:kazimierczuk-2014, src:smiertka-2026). |
| nested level | Built from an electron (conduction band, or LUMO) and a hole (valence band, or HOMO) bound by the screened Coulomb force; embedded in a host (semiconductor, insulator, molecular crystal, pigment-protein complex, 2D magnet) whose bands, screening, lattice and spins set what it is. One kind of quasiparticle; with photons it makes exciton-polaritons, with other excitons biexcitons and electron-hole liquids. |
| scale | molecular |
Measures (52) published science
| exciton Rydberg energy, GaAs | 4.2 meV |
| exciton Rydberg energy, Cu2O (yellow series) | 92 meV |
| highest Rydberg exciton seen, Cu2O | n = 25 principal quantum number |
| Rydberg exciton extension at high n, Cu2O | > 2 micrometres |
| Frenkel exciton binding energy, typical | 0.1 to 1 eV |
| Wannier-Mott binding energy, typical | ~1-50 meV bulk; ~0.1-0.5 eV in 2D meV / eV |
| exciton binding energy, monolayer TMDs | about 0.5 eV |
| exciton binding energy, monolayer WS2 | 0.32 eV |
| exciton binding energy, MoSe2 on bilayer graphene | 0.55 eV |
| exciton binding energy, interlayer excitons in TMD heterostructures | about 100 meV |
| exciton binding energy, MAPbI3 perovskite, low-temperature phase | 16 +/- 2 meV |
| exciton binding energy, CrSBr XA (Frenkel-like) | about 0.7 eV |
| exciton binding energy, CrSBr XB (Wannier-like) | about 0.3 eV |
| CrCl3 exciton: E_X, E_g, E_b | 1.5, 4.5, 3.0 eV |
| CrSBr exciton energies | XA 1.38, XB 1.8 (at 0 T) eV |
| CrSBr redshift AFM to FM | XB about 100 (theory 95); XA about 10 (theory 7) meV |
| CrSBr XB redshift against temperature | 110 at 2 K to ~85 at 110 K meV |
| CrSBr exciton length along b | XB 4.5, XA 1.2 nm |
| CrSBr diamagnetic coefficients | sigma_B 0.22 +/- 0.02, sigma_A 0.05 +/- 0.01 meV/T^2 |
| CrSBr band gap (QSGW with ladders) | about 2.07 eV |
| CrSBr self-hybridised polariton Rabi splitting | > 200 meV |
| NiPS3 exciton energy and linewidth | 1.475 eV; 0.35 meV eV; meV |
| CrI3 exciton linewidth at lowest temperatures | > 100 meV |
| CrSBr magnon coherence time and propagation | > 5 ns; > 15 micrometres ns; micrometres |
| radiative lifetime of excitons, monolayer WSe2 at low temperature | about 200 fs |
| molecular exciton lifetime, typical | nanoseconds ns |
| LH1 and LH2 fluorescence lifetimes (mutant membranes) | about 1000 ps |
| antenna transfer steps | < 100 fs to several ps fs-ps |
| B800 to B850 transfer, living R. sphaeroides | 200 fs to 1 ps fs-ps |
| LH2 to LH1 transfer | about 8 ps |
| LH1 to reaction centre and charge separation | about 110 ps |
| exciton-exciton annihilation, effective, in cells | about 30 ps |
| B850 exciton delocalisation | 16 +/- 4 BChl (nonlinear spectroscopy) vs ~4 BChl sites vs 8 +/- 2 (IPR with disorder, computed) molecules |
| exciton formation spin statistics (injected charges) | 25% singlets, 75% triplets fraction |
| maximum internal quantum efficiency: fluorescent vs phosphorescent/TADF OLED | up to 25% vs up to (nearly) 100% percent |
| photoluminescence yield, TMD monolayers | < 1% (95% after superacid treatment of MoS2) percent |
| electroluminescence quantum yield, TMD LEDs at room temperature | about 1 percent |
| absorption at lowest exciton, TMD monolayer | > 15 percent |
| exciton linewidth broadening at room temperature, TMDs | 30-40 meV |
| exciton diffusion length, common organic semiconductors (2000s) | < 10 nm |
| exciton diffusion coefficient, organic semiconductors, typical | order of 10^-4 cm^2/s |
| linear exciton slope A, monolayer hBN | 2.77 (Gamma-M), 1.94 (Gamma-K) eV.Angstrom |
| massless exciton group velocity, hBN | 0.003 c |
| parabolic exciton mass, hBN | 1.93 (Gamma-M), 8.74 (Gamma-K) m_e |
| direct optical gap, freestanding monolayer hBN | 6.6 eV |
| hBN dipole-forbidden exciton XD1 (GW@PBE) | 6.26 eV; binding 1.65 eV |
| linear exciton dispersion range in TMDs | about 10% of the Brillouin zone fraction |
| CrSBr redshift AFM to FM (review) | high-energy exciton (~1.77 eV) approaching 90; low-energy exciton (~1.36 eV) roughly 15 meV |
| exciton Mott density, TMD monolayers (theory) | 3 x 10^12 to 1 x 10^13 cm^-2 |
| singlet-fission gain in carotenoid-to-BChl transfer | up to 20 percent |
| carotenoid-BChl triplet-pair decay window (delayed luminescence) | nanoseconds to milliseconds time |
| Hubbard exciton lifetime, Sr2IrO4 | about 1 ps |
Open contentions (10) open contention
Where sources or readings disagree. Both stay on the map.
- C-806 Does photosynthesis use quantum effects to move energy?
booked: The map follows the 2017 finding that energy hops between small clusters and the quantum coherence fades in about 60 femtoseconds with no shown role, and keeps the long-lived coherence view alongside.
the other reading: the exciton tries several paths at once (wave-like coherence) - C-3618 Across how many pigment molecules does an exciton spread in a light-harvesting ring?
booked: The map keeps the competing readings; none is ruled.
the other reading: a single size; one measure is right and the others wrong - C-3619 Does splitting one excitation into two help photosynthesis harvest light?
booked: The map keeps the competing readings; none is ruled.
the other reading: the role of singlet fission in these organisms remains poorly understood (the same paper's introduction); carotenoid triplets in antennas as a loss or photoprotective side channel (the specialist's reading of the older literature; unsourced here) - C-3620 How large is the energy gap of the layered magnet CrSBr, and so what kind of excitons does it hold?
booked: The map keeps the competing readings; none is ruled.
the other reading: about 1.5 eV (DFT-based GW), which would make both Wannier-Mott-like; about 2.2 eV (plasmon-pole GW) - C-3621 How far do excitons travel in organic materials, and which measurement can be trusted?
booked: The map keeps the competing readings; none is ruled.
the other reading: static quenching underestimates LD; time-resolved annihilation overestimates LD; steady-state annihilation proposed as more accurate - C-3622 What is the energy gap of a single sheet of boron nitride?
booked: The map keeps the competing readings; none is ruled.
the other reading: 6.1 eV as the intrinsic monolayer gap; 5.95 eV earlier GW-BSE; 6.23-6.31 eV the Liu theory - C-3623 What kind of exciton lives in the magnet NiPS3?
booked: The map keeps the competing readings; none is ruled.
the other reading: Zhang-Rice-type (singlet-triplet many-body); Hund's exciton; localised on Ni with negligible Zhang-Rice part - C-3624 How tightly bound are excitons in perovskite solar-cell material?
booked: The map keeps the competing readings; none is ruled.
the other reading: 37-50 meV from earlier low-field fits - C-3625 Where does the light go in freshly made atom-thin semiconductors?
booked: The map keeps the competing readings; none is ruled.
the other reading: defect states trap excitons before they recombine radiatively, and the superacid shields them (Mueller and Malic 2018 review, "presumably") - C-3626 How tightly are excitons bound in ordinary bulk crystals?
booked: The map keeps the competing readings; none is ruled.
the other reading: a few meV (GaAs, room-temperature perovskites) to above 100 meV (Cu2O 1s, central-cell corrections) - the specialist's range, ZnO ~60 meV, unsourced here
Cross-scale patterns seen here (18) our proposal
Shapes that recur when the same kind of problem is solved at different scales, and how each shows in Exciton. Each links to every other system that shows it.
- namespace completeness (the map predicts its gaps) (structural)
The hydrogen-like series predicts its own missing members and where they sit (E = -R/n^2): followed to n = 25 in Cu2O, to 5s in WS2 where it extrapolates the gap; in 2D the series bends away from hydrogen in a way the screened potential predicts. The shelf of kinds (Frenkel, charge-transfer, Zhang-R - same role, different builds (structural)
One job, carrying absorbed light energy without moving charge to where it is turned into charge or light, done by unrelated builds: bacteriochlorophyll rings on proteins, organic films, inorganic monolayers, magnetic crystals. - same shape, opposite sign (structural)
One electric-dipole transition run both ways: absorption makes the exciton and radiative recombination unmakes it, with the same oscillator strength setting both (linewidth and radiative lifetime). In devices, an LED runs charge into light and a solar cell light into charge through the same bound pa - symmetry breaking (name the cut) (structural)
A parity-forbidden d-d exciton is dark until hybridisation and the crystal field break the symmetry and brighten it; a zigzag spin texture breaks rotational symmetry and sets the exciton's linear polarization axis, which turns with the Neel vector; broken time-reversal symmetry gives circular dichro - templating (structural)
The hydrogen atom as a module swapped into crystals: excitons are the condensed-matter analogues of hydrogen, with effective masses and screening put in, and still work (Rydberg series, blockade like Rydberg atoms). The same electron-hole mould runs in a leaf's antenna, an OLED, a TMD monolayer and - The carry (structural)
Overflow held and passed on later: in purple bacteria a carotenoid-BChl triplet pair acts as a reservoir ("energy buffering") that may spread the arrival of excitons to match the reaction centre's processing capacity (Wang 2026). The wait: the antenna funnel settles at its slowest hand-off, LH1 to t - The Payment to Entropy (structural)
Every hand-off of an exciton pays a toll. Form (spin): injected charges make 75% triplets, so a fluorescent OLED tops out at 25% internal quantum efficiency until TADF or phosphorescence recovers them. Form (thermal and trap): as-made TMD monolayers give under 1% of their excitons back as light, the - the whole sets the local reference (structural)
Each layer's exciton energy is read against the spin alignment of its neighbours (shift proportional to cos^2(theta/2)); the surroundings set each exciton's binding by hundreds of meV; surface excitons in CrSBr sit below bulk ones because they are less screened. - trade one property for another (structural)
Binding bought with sensitivity and extraction: the tightly bound Frenkel-like XA (0.7 eV, 1.2 nm) barely notices magnetic order or phonons, the looser Wannier-like XB (0.3 eV, 4.5 nm) is ten times more sensitive; in devices a large binding keeps excitons alive at room temperature but costs open-cir - Umwelt: each system carves its own slice of one shared medium (structural)
Two excitons in one crystal carve different slices of the same surroundings: in CrSBr the Wannier-like XB hears the interlayer spin alignment and out-of-plane Ag phonons, the Frenkel-like XA barely hears either; the dark acoustic magnon is heard by neither while the optical one is. - who opens the door (structural)
Where an exciton ends: in a bare crystal on whatever trap or phonon it meets (open field); in a leaf or purple bacterium at a reaction centre the cell built and placed inside LH1 ahead of time (a port); in a device at a junction built by engineers (a port installed from outside). - holding / entrainment (partial)
The magnet's optical magnon sets the exciton energy ticking at the magnon frequency (periodic shift read by a probe); the exciton is entrained, the acoustic mode leaves it still. - not-diluting wins (partial)
Bright excitons drain by emission (about 200 fs), dark ones do not and can come to outnumber them; in OLEDs 75% of excitons are dark triplets. In tungsten-based TMDs the dark state also lies lower in energy than the bright one, so dark excitons fill up by thermal population as well as by a slow drai - same toolkit, independent builds (partial)
The hydrogenic toolkit (Coulomb attraction, screening, effective masses) assembled separately in GaAs, Cu2O, perovskites and TMDs gives the same kind of Rydberg-like exciton series. - smooth to sharp in finite time (partial)
The longitudinal exciton band in a 2D material has a non-analytic V-shaped discontinuity at zero momentum: the smooth band breaks at a finite parameter (Q = 0). A kink, not an infinity. - barrier-gates-complexity (surface)
Spin order as a seal: in the AFM state excitons are confined to one layer (monolayer-like in bulk); with the seal lifted (FM) they spread across layers. And an insulating spacer between layers gives interlayer excitons a much longer life. The barrier shapes the exciton but no source shows complexity - causal ceiling (surface)
Only excitons in the light cone (near-zero momentum) can emit, and light alone cannot reach finite-momentum excitons; but an electron beam reaches them, so the ceiling is on one channel, not on what can be known. - partner-becomes-part (surface)
Exciton and photon in strong coupling become one hybrid quasiparticle that inherits properties from both; electron and hole become one bound pair. Inheritance here is of properties, not of descent.
Valence: what it trades, and with whom (18) published science
Each exchange with a neighbour, read from both sides: who gains, who pays, and what carries it.
- Quasiparticle via sys:matter · Exciton even
- Electron via sys:matter · Exciton gain
- Photon via sys:electromagnetic-field · Exciton gain
- Human (Homo sapiens) via sys:electromagnetic-field · Exciton even
- Phonon via sys:thermal-energy · Exciton loss
- Magnon via sys:matter · Exciton mixed
- Laser via sys:electromagnetic-field · Exciton loss
- Photosynthesis via sys:electromagnetic-field · Exciton loss
- Light (visible light) via sys:electromagnetic-field · Exciton gain
- Magnetism via sys:matter · Exciton mixed
- Antenna (light-harvesting) complexes via sys:electromagnetic-field · Exciton even
- Electron hole via sys:matter · Exciton gain
- Semiconductor via sys:matter · Exciton gain
- Solar cell via sys:electricity · Exciton loss
- Light-emitting diode via sys:electricity · Exciton loss
- Carotenoids via sys:electromagnetic-field · Exciton mixed
- Hexagonal boron nitride via sys:matter · Exciton gain
- Chromium sulfide bromide (CrSBr) via sys:matter · Exciton mixed
Partners and parts published science
Built from / builds (5)
- Electron
- Electron hole
- Competition for excitation in a blockade volume (Rydberg excitons in Cu2O)
- Competition for an exciton in a TMD monolayer: emission against trions and annihilation
- Competition for an antenna excitation: reaction centre, quenching, fluorescence, annihilation
Sits inside / holds (6)
- Crystal
- Antenna (light-harvesting) complexes
- Quantum dot (semiconductor nanocrystal)
- Semiconductor
- Two-dimensional material
- Van der Waals magnetic semiconductor
expressed by (4)
- Recombination/Composition
- Templating / Standardization
- Interoperability/Transduction (Bridging)
- Resource Denial/Attrition
floor filled by (14)
- Quasiparticle
- Electron
- Photon
- Electromagnetic field
- Thermal energy (heat, temperature and heat flow)
- Phonon
- Magnon
- Matter
- Photosynthesis
- Magnetism
- Antenna (light-harvesting) complexes
- Electron hole
- Semiconductor
- Solar cell
represents (1)
Kinds (19)
- Quasiparticle
- Frenkel exciton
- Wannier-Mott exciton
- Charge-transfer exciton
- Zhang-Rice exciton
- Hubbard exciton
- Interlayer (spatially indirect) exciton
- Dark exciton
- Triplet exciton
- Rydberg exciton
- Self-trapped exciton
- Massless (linearly dispersing) exciton
- Exciton-polariton
- Biexciton
- Trion (charged exciton)
- Fractional exciton
- J-aggregate (and H-aggregate)
- Bound exciton (impurity-bound)
- Multiexciton (quantum dots)
wired (5)
Notes published science
- An exciton can act like a particle of light: in a single sheet of boron nitride, the bound electron-hole pair's energy rises in straight proportion to its momentum, as light's does, so it behaves as if it had no mass at all, even though it is built from two massive particles. Predicted in 2015, seen directly in 2026 by firing electrons through the free-standing sheet (src:liu-2026, src:qiu-2015).
- One magnetic crystal, CrSBr, holds both textbook kinds of exciton at once: a tight one, sitting mostly on the chromium atoms and about 1.2 nm long (two or three lattice spacings), and a loose one about 4.5 nm long. Line up the magnet's layers with a field and the loose one shifts about 100 meV, ten times the tight one, because it can spill into the next layer only when the spins there point the same way (src:smiertka-2026).
- The record Rydberg excitons, more than 2 micrometres across (thousands of times their ground state), were found in a sliver of copper oxide cut from a natural rock from the Tsumeb mine in Namibia; a single exciton that big blocks any other from forming near it (src:kazimierczuk-2014).
Layers
The ways it takes in its world, where they meet, and how it acts back, each from L1 (where it meets the world) up.
exciton (reader's view: the host and the light it takes in, absorption, binding, the bound pair, its motion, its fate)
- L1: The host and the light falling on it: a semiconductor, insulator, molecular crystal, pigment-protein complex or 2D magnet in an electromagnetic field
- L2: Absorption: an electron promoted from valence band (HOMO) to conduction band (LUMO), leaving a hole; or injected electrons and holes captured
- L3: Binding: the screened Coulomb attraction, set by the dielectric surroundings, masses and orbitals (Frenkel, charge-transfer, Wannier-Mott and mixed)
- L4: The bound pair: energy below the gap, binding energy, radius, spin (singlet or triplet), bright or dark
- L5: Motion: centre-of-mass momentum and exciton band structure (parabolic or linear), diffusion, hopping and coherent transfer between sites, drag by magnons
- L6: Fate: light (radiative recombination), heat (phonons), free charges (dissociation), hand-off to another site or a reaction centre, splitting (singlet fission), annihilation, a push on spins
the boardroom (shared floors)
- L4: The bound pair: one electron-hole state with its energy, binding energy, radius and character on the Frenkel-to-Wannier continuum, spin and brightness; every influence ends as a change in these
- L5: Motion: the exciton band structure in centre-of-mass momentum (parabolic, or linear and massless in 2D), diffusion over nanometres, hopping between sites, drag by magnons, coherence
- L6: Fate: how the excitation ends (light, heat, charge, transfer, fission, annihilation, a push on spins); the top floor: nothing in an exciton reads, regulates or means
way in: A fast electron beam (momentum transfer)
- L1: Electrons in a scanning transmission electron microscope passing through a freestanding monolayer
- L2: Inelastic scattering transfers a momentum Q and energy to the sample, reaching momenta photons cannot
- L3: Excitons made at finite centre-of-mass momentum, up to the third Brillouin zone in one shot
way in: An applied magnetic field
- L1: A laboratory magnetic field (to 85 T in pulsed magnets)
- L2: Zeeman and diamagnetic response; in a magnet, the field cants the spins
- L3: A diamagnetic shift whose coefficient measures the exciton's size: 0.22 against 0.05 meV/T^2 for XB and XA
way in: Confined light (cavity modes, strong coupling)
- L1: Photons held in an optical microcavity, or in the crystal itself
- L2: Strong exciton-photon coupling: anticrossing of the two oscillators
- L3: A hybrid light-matter quasiparticle: Rabi splitting of a few meV (NiPS3) to over 200 meV (bulk CrSBr)
way in: Defects, impurities and disorder
- L1: Impurities, vacancies, strain pockets, spin disorder
- L2: Trapping in local potential wells; bound excitons with giant oscillator strength
- L3: Localised excitons and single-photon emitters at low temperature; a lost path to light (PL yield under 1%)
way in: Injected electrons and holes (electrical excitation)
- L1: An electric current bringing electrons and holes into an emitting layer (LED, OLED)
- L2: Electron-hole capture: injected charges with random spins meet and bind
- L3: Spin set by statistics: 25% singlets, 75% triplets
way in: Lattice vibrations and heat (phonons)
- L1: The vibrating lattice: acoustic and optical phonons, temperature
- L2: Exciton-phonon scattering; strong coupling dresses the exciton in phonons (self-trapping)
- L3: Linewidth, thermalisation and energy shifts: 30-40 meV broadening at room temperature in TMDs; XB's magnetic redshift falls with out-of-plane Ag phonons; perovskite binding falls to a few meV above 160 K
way in: Light absorbed (photons at an excitonic resonance)
- L1: Light in the host: photons near the band gap or a molecular transition
- L2: Absorption at a resonance with an oscillator strength; forbidden d-d transitions brightened only when symmetry is broken; a TMD monolayer takes more than 15% of light at its lowest exciton
- L3: What is set at birth: energy, a near-zero centre-of-mass momentum (photons carry little), bright spin state and valley or polarization
way in: Other excitons (density)
- L1: Other excitons nearby at high density
- L2: Dipole-dipole and Auger interaction; magnon-mediated attraction in CrSBr
- L3: Blockade, annihilation, binding into biexcitons, a density-dependent redshift; at high density a Mott transition to a fully ionised electron-hole plasma (predicted 3e12-1e13 cm^-2 in TMDs)
way in: The dielectric surroundings (screening of the attraction)
- L1: The surrounding material: the host's own polarisability, substrate, encapsulating layers, vacuum
- L2: Screened Coulomb interaction (Rytova-Keldysh in 2D); weak screening in thin layers
- L3: Binding energy and radius set: typically about 1-50 meV in bulk (review; Cu2O's Rydberg energy 92 meV), about 0.1-0.5 eV in 2D; tunable by hundreds of meV by the environment
way in: The host's magnetic order and magnons
- L1: Spin order in a 2D magnet (ferromagnetic, A-type or zigzag antiferromagnetic) and its spin waves (magnons, GHz)
- L2: Coupling through exchange and spin-orbit coupling: spin-forbidden interlayer hopping in the AFM state; magnetostriction as an indirect path
- L3: Energy, polarization and size set by the spins: shift proportional to cos^2(theta/2) of the interlayer canting; circular dichroism in CrI3, linear in NiPS3; the Wannier-like exciton ten times more sensitive than the Frenkel-like one
way out: A push on the host's spins (back-action on magnetic order)
- L1: The magnetic order changed: a flipped domain, coherent magnons
- L3: Exciton-magnon coupling adjusts the canting; circularly polarised light at an exciton resonance flips a magnetization
- L6: Exciton density or resonant absorption acting on spins
way out: Dissociation into free electron and hole (photocurrent)
- L1: Free carriers: a photocurrent or a stored charge separation
- L3: Tunnel ionisation or band offsets larger than the binding energy (type-II junctions; donor-acceptor interfaces); charge separation in a reaction centre
- L6: The bound pair split
way out: Energy handed on by electron exchange (Dexter transfer; how triplets move)
- L1: A neighbouring molecule in its triplet state (OLED host to guest, polymer chain to chain)
- L3: Dexter-type transfer: a correlated transfer of two electrons, coupled through wavefunction overlap between neighbours (short range); the route for triplets
- L6: The excitation leaves this molecule by exchange
way out: Energy handed to another site (Forster transfer, antenna hops, singlet fission)
- L1: A neighbour excited: the next pigment, the reaction centre's special pair, a quencher
- L3: Hopping or coherent transfer to a neighbour with matching absorption (Forster, through the near field); a singlet split into a triplet pair that fuses back (TTA) into a neighbour's singlet
- L6: The excitation leaves this site whole
way out: Nonradiative decay into phonons (heat)
- L1: Lattice heat
- L3: Phonon emission, trapping at defects, quenching (NPQ in plants)
- L6: The excitation ends without light
way out: Radiative recombination (luminescence, LEDs, lasing)
- L1: A photon in the field: narrow lines below the gap; polarization carrying the host's magnetic state
- L3: Only excitons in the light cone (vanishing momentum) and bright spin can emit: about 200 fs radiative lifetime in WSe2; singlets only in fluorescent emitters
- L6: Recombination of electron and hole