Exciton Bose Solid: Phases and Mechanisms

• Exciton Bose solids are defined as phases where bosonic excitons form spatially ordered states, ranging from crystals to incompressible supersolids.
• Key theoretical frameworks include the extended Bose–Hubbard model, nonlocal Gross–Pitaevskii dynamics, and reduced-density-matrix analyses to characterize order parameters.
• Experimental platforms such as quantum wells and moiré systems reveal diagnostic signatures like Umklapp peaks, exciton resistance anomalies, and phase coherence.

Searching arXiv for the cited and closely related papers on exciton Bose solids, exciton crystals, supersolids, and exciton condensates. An exciton Bose solid is a phase of bosonic excitons in a solid-state environment in which excitonic degrees of freedom acquire solid-like spatial order rather than remaining a spatially uniform Bose fluid. In the literature, the term is used in more than one, partly overlapping sense. In the broadest usage, it denotes a strongly interacting Bose subsystem formed by excitons in a crystalline host, often in a regime proximate to condensation, Mott localization, or crystalline ordering. In stricter usage, it denotes an exciton crystal or an incompressible supersolid with exactly one boson per emergent lattice site, or a lattice supersolid in which translational symmetry breaking coexists with off-diagonal long-range order (Zerba et al., 2024, Conti et al., 27 Jul 2025, Lagoin et al., 2024). The modern subject spans dipolar interlayer excitons in semiconductor bilayers, moiré excitonic insulators, electrostatically defined exciton lattices, and hybrid Bose–Fermi systems in which excitons couple to carriers or electron gases (Matuszewski et al., 2011, Lagoin et al., 2021, Qi et al., 27 Jan 2026).

1. Conceptual scope and definitions

The phrase “exciton Bose solid” does not designate a single universally fixed phase. In one line of work, it refers to an exciton supersolid: a Bose–Einstein condensed state of indirect excitons that simultaneously develops a crystalline density modulation in real space (Matuszewski et al., 2011). In another, it denotes a bosonic lattice phase of excitons, especially a Mott-insulating or density-ordered state in an extended Bose–Hubbard description (Lagoin et al., 2021). In more recent work on double-layer excitons, it denotes an incompressible supersolid with exactly one boson per lattice site, distinguished from the compressible cluster supersolids seen in dipolar quantum gases (Conti et al., 27 Jul 2025). Experimental work on moiré excitonic insulators uses the more direct language of an exciton crystal, identifying a crystalline, insulating state of interacting bosonic excitons stabilized in thermal equilibrium (Qi et al., 27 Jan 2026). Work on nanoscopic exciton arrays uses the term lattice supersolid for a state that breaks translational symmetry while exhibiting off-diagonal long-range order (Lagoin et al., 2024).

A useful minimal distinction is therefore between three cases. A uniform exciton condensate is a Bose-condensed exciton fluid without spatial symmetry breaking. An exciton Bose solid in the narrow sense is a phase with diagonal long-range order in exciton density, whether or not global phase coherence survives. An exciton supersolid is the coexistence regime in which crystalline density order and excitonic condensation are both present (Matuszewski et al., 2011, Lagoin et al., 2024).

Several papers also stress that the term can be used more loosely. In transition-metal dichalcogenide heterostructures, a dense interlayer-exciton gas coupled to carriers is described as a Bose subsystem in a solid, and “exciton Bose solid” is then a conceptual umbrella covering Bose-liquid physics, proximity to condensation, and proximity to more ordered crystalline phases, even when the explicit calculation treats a homogeneous exciton gas (Zerba et al., 2024). From the reduced-density-matrix perspective, any ordered excitonic phase is characterized first by off-diagonal long-range order in the particle–hole sector; whether it is fluid-like or solid-like depends on the spatial structure of the dominant excitonic eigenmode (Torres et al., 2024).

2. Physical platforms and excitonic degrees of freedom

The central platforms are spatially indirect or interlayer excitons, because their long lifetimes and dipolar interactions make strongly correlated bosonic behavior experimentally realistic. In GaAs coupled quantum wells and wide quantum wells, electrons and holes are spatially separated by an electric field, producing indirect excitons with long radiative lifetimes, permanent electric dipole moments, and quantum degeneracy temperatures of order a few kelvin (High et al., 2011, Alloing et al., 2013). In van der Waals heterostructures, type-II band alignment in TMD bilayers places electrons and holes in different monolayers, yielding interlayer excitons with large binding energies, strong dipolar interactions, and densities up to $10^{12}\,\mathrm{cm^{-2}}$ (Wang et al., 2021), while more recent equilibrium electron–hole bilayers in MoSe$_2$/hBN/WSe$_2$ realize excitonic-insulator regimes with a charge gap of about $30\,\mathrm{meV}$ in the dilute limit and a Mott density near $0.75\times 10^{12}\,\mathrm{cm}^{-2}$ (Qi et al., 16 Mar 2026).

The same ingredients also support explicitly lattice-based exciton physics. In electrostatically patterned GaAs double quantum wells, indirect excitons are confined in a programmable two-dimensional lattice with period $a=400$ or $800\,\mathrm{nm}$ and depth $V_0\simeq1.5\,\mathrm{meV}$, realizing a multi-orbital Bose–Hubbard regime (Lagoin et al., 2021). In moiré systems, the underlying lattice is supplied by the moiré potential itself. A WS$_2$/WSe$_2$ moiré superlattice can host correlated hole Mott and generalized Wigner-crystal states, and when Coulomb-coupled to a MoSe$_2$0 layer it yields a tunable extended Bose–Hubbard model of interlayer excitons and charge carriers in thermal equilibrium (Qi et al., 27 Jan 2026). A distinct dual-moiré realization uses WS$_2$1/bilayer WSe$_2$2/WS$_2$3, where interlayer excitons are formed from an electron in one WS$_2$4 moiré lattice and an empty site in the other, producing a Bose–Fermi mixture of excitons and holes (Zeng et al., 2022).

Hybrid Bose–Fermi systems provide another route. In coupled quantum wells, a nearby two-dimensional electron gas mediates a momentum-dependent effective interaction between indirect excitons and can drive a roton instability of the exciton condensate (Matuszewski et al., 2011). In TMD heterostructures with a hole-doped layer coupled to an exciton-hosting bilayer, resonant conversion $_2$5 between holes, excitons, and trions produces a solid-state Feshbach resonance and a tunable Bose–Fermi mixture in which the excitonic subsystem is a dense Bose gas embedded in a crystal (Zerba et al., 2024).

3. Microscopic descriptions and order parameters

Three theoretical languages recur. The first is the extended Bose–Hubbard framework, appropriate when excitons are localized by a lattice potential or moiré pattern. In the GaAs exciton-lattice work, the full multi-orbital Bose–Hubbard Hamiltonian contains hopping $_2$6, on-site interaction $_2$7, and a chemical potential $_2$8, with the key regime being $_2$9, where $_2$0 is the spacing between Wannier states (Lagoin et al., 2021). The paper further emphasizes that extending the model to nearest-neighbor interactions is necessary to realize phases that spontaneously break lattice symmetry, such as checkerboards, stripes, or lattice supersolids (Lagoin et al., 2021). The moiré exciton-crystal experiment makes this viewpoint explicit by stating that the device constructs a tunable extended Bose–Hubbard model with electrical control over exciton and charge doping in thermal equilibrium (Qi et al., 27 Jan 2026).

The second is the Gross–Pitaevskii and nonlocal mean-field framework, used for continuum dipolar excitons. In the hybrid exciton–electron system, integrating out the electron gas gives a momentum-dependent effective exciton–exciton interaction $_2$1, and the exciton collective mode obeys a Bogoliubov-type dispersion

$_2$2

so a roton minimum appears when $_2$3 becomes negative over an interval of $_2$4 (Matuszewski et al., 2011). In the recent theory of an excitonic incompressible Bose solid, the standard GP formalism is extended to include strong two-particle correlations and exclude exciton self-interactions. The effective interaction $_2$5 is set equal to $_2$6 for $_2$7 and to zero for $_2$8, and the resulting nonlocal mean-field potential subtracts self-interaction within each unit cell, generating an effective well that localizes one exciton per site (Conti et al., 27 Jul 2025). The paper’s central claim is that this is the ingredient needed to describe a true Bose solid with exactly one boson per site (Conti et al., 27 Jul 2025).

The third is the order-parameter and reduced-density-matrix perspective. A conventional bosonic condensate is identified by a macroscopic eigenvalue of the one-body density matrix, whereas exciton condensation is identified by off-diagonal long-range order in the modified particle–hole reduced density matrix $_2$9, specifically by an eigenvalue $30\,\mathrm{meV}$0 (Torres et al., 2024). This framework does not by itself determine whether the condensate is fluid-like or solid-like; rather, the distinction lies in the spatial structure of the dominant excitonic eigenmode. A spatially periodic or structured $30\,\mathrm{meV}$1 corresponds naturally to an ordered excitonic phase, while a delocalized mode corresponds to a uniform exciton condensate (Torres et al., 2024).

A separate, more field-theoretic notion of an exciton Bose condensate appears at a Lifshitz critical point where boson–hole pairs $30\,\mathrm{meV}$2 condense while single bosons do not. That work discusses an exciton Bose condensate and an exciton Bose liquid, and treats an exciton Bose solid as a natural but not explicitly constructed descendant in which excitonic operators develop spatially periodic expectation values (Ma et al., 2018). This suggests that finite-momentum instabilities of an excitonic condensate provide a route to density-wave or crystalline exciton order.

4. Crystalline, Mott, and supersolid realizations

The most direct crystalline realization is the exciton crystal in a moiré excitonic insulator. In a MoSe$30\,\mathrm{meV}$3/WS$30\,\mathrm{meV}$4/WSe$30\,\mathrm{meV}$5 moiré electron–hole bilayer, optical spectroscopy reveals spontaneous crystallization of long-lived excitons at one exciton filling per three moiré sites, $30\,\mathrm{meV}$6, evidenced by strong Umklapp scattering peaks in the optical spectrum (Qi et al., 27 Jan 2026). Exciton transport simultaneously shows a pronounced exciton resistance peak at the same filling, consistent with suppressed exciton hopping in a crystalline phase (Qi et al., 27 Jan 2026). The paper identifies this as a generalized exciton crystal in which dipolar excitons avoid all nearest-neighbor moiré site occupations and break the symmetry of the underlying moiré lattice (Qi et al., 27 Jan 2026). This is the clearest current experimental realization of an exciton Bose solid in the narrow sense.

A closely related but more composite regime is the exciton density wave in Coulomb-coupled dual moiré lattices. There, correlated insulating states appear when the combined filling factor equals $30\,\mathrm{meV}$7 or $30\,\mathrm{meV}$8, and these are interpreted as exciton density waves in a Bose–Fermi mixture of excitons and holes (Zeng et al., 2022). Strong repulsive interactions cause the holes to form robust generalized Wigner crystals, while the exciton fluid is restricted to channels that spontaneously break translational symmetry of the lattice (Zeng et al., 2022). The authors explicitly connect this setting to Bose–Fermi mixture theories in which supersolid phases are predicted, and argue that the observed excitonic density waves should possess finite superfluid densities in the weak-disorder limit (Zeng et al., 2022). This makes them Bose-solid–like phases and plausible exciton supersolids, even though phase coherence is not directly measured in that experiment.

A lattice-route precursor is the Mott insulator of strongly interacting two-dimensional excitons. In electrostatically defined GaAs exciton lattices, Mott phases are observed with one or two excitons per site in particular Wannier states, with the energy splitting between the $30\,\mathrm{meV}$9 and $0.75\times 10^{12}\,\mathrm{cm}^{-2}$0 plateaus matching the calculated on-site repulsion $0.75\times 10^{12}\,\mathrm{cm}^{-2}$1 (Lagoin et al., 2021). The work explicitly states that this establishes a concrete platform from which one can progress to excitonic Bose solids such as stripes, checkerboards, and lattice supersolids once nearest-neighbor interactions become relevant (Lagoin et al., 2021). In that sense, the exciton Mott insulator is a bosonic lattice solid without broken translational symmetry, while the projected extended-Bose–Hubbard phases are true exciton Bose solids.

The supersolid route is exemplified by two distinct systems. In the hybrid exciton–electron system, reducing the exciton–electron separation $0.75\times 10^{12}\,\mathrm{cm}^{-2}$2 makes the effective exciton–exciton interaction negative over a range of momenta, creating a roton minimum; when the roton crosses zero, the homogeneous condensate becomes unstable and the ground state found by imaginary-time Gross–Pitaevskii evolution is a triangular lattice of density maxima with Bragg peaks in momentum space (Matuszewski et al., 2011). Because the state retains a nonzero condensate order parameter while breaking translational symmetry, it is an exciton supersolid, and the paper notes that this is precisely what one would call an “exciton Bose solid” in the supersolid sense (Matuszewski et al., 2011). In nanoscopic exciton arrays governed by a Dicke-Hubbard Hamiltonian, dipolar and Dicke correlations at half lattice filling induce both condensation in a single sub-radiant state and dipolar quantum order that spontaneously breaks lattice symmetry, which is presented as evidence for a lattice supersolid dissipatively prepared across $0.75\times 10^{12}\,\mathrm{cm}^{-2}$3 sites (Lagoin et al., 2024).

Finally, the 2025 Gross–Pitaevskii theory sharpens the notion of an exciton Bose solid by defining it as an incompressible supersolid. The paper argues that a true Bose solid requires exactly one boson per lattice site, generating a particle–antiparticle gap and incompressibility, yet retaining global phase coherence because the particles belong to a single Bose condensate (Conti et al., 27 Jul 2025). Within that extended GP framework, both superfluid and incompressible supersolid ground states are obtained across experimentally accessible exciton densities and interlayer separations (Conti et al., 27 Jul 2025).

5. Experimental signatures and diagnostics

The principal diagnostics fall into three classes: spectroscopic signatures of crystalline order, transport signatures of excitonic motion, and coherence signatures of condensate order.

For crystalline order, the most direct spectroscopic marker is the appearance of Umklapp peaks. In the moiré exciton crystal, a satellite peak on the high-energy side of the main $0.75\times 10^{12}\,\mathrm{cm}^{-2}$4 resonance appears only at $0.75\times 10^{12}\,\mathrm{cm}^{-2}$5, and the measured energy separation $0.75\times 10^{12}\,\mathrm{cm}^{-2}$6 agrees with the estimate $0.75\times 10^{12}\,\mathrm{cm}^{-2}$7 from the folded exciton dispersion (Qi et al., 27 Jan 2026). In the 2011 supersolid proposal, a perfect supersolid exhibits sharp Bragg peaks at reciprocal-lattice vectors, while a disordered supersolid produces concentric rings in momentum space (Matuszewski et al., 2011). In the 2024 nanoscopic array experiment, translational symmetry breaking is diagnosed as dipolar quantum order in a sub-wavelength lattice, coexisting with condensation in a sub-radiant state (Lagoin et al., 2024).

For transport, bilayer and moiré systems are especially informative. In double-layer graphene at total filling $0.75\times 10^{12}\,\mathrm{cm}^{-2}$8, perfect Coulomb drag and large counterflow conductance diagnose the exciton superfluid, while vanishing conductance at low temperature and large $0.75\times 10^{12}\,\mathrm{cm}^{-2}$9 indicates an exciton insulator that the paper argues is a solid phase driven by dipole–dipole interactions in the dilute limit (Zeng et al., 2023). Its melting restores perfect drag, suggesting reentrant superfluid-like behavior and possible quantum coherence of the low-temperature solid (Zeng et al., 2023). In the moiré exciton crystal, both two-terminal and four-terminal exciton resistance display pronounced peaks precisely at the commensurate crystalline filling $a=400$0, consistent with suppressed exciton hopping in a crystalline phase (Qi et al., 27 Jan 2026). In TMD Bose–Fermi mixtures, resonantly enhanced hole–exciton–trion scattering produces non-Drude transport, a strong tunable peak in resistivity, and a sign-changing exciton drag; these are not signatures of a Bose solid by themselves, but they provide a transport-based probe of a strongly interacting exciton Bose subsystem in a solid (Zerba et al., 2024).

For coherence, classical exciton-condensation diagnostics remain important because a supersolid must retain them. A cold gas of indirect excitons in GaAs coupled quantum wells exhibits extended spontaneous coherence and phase singularities when cooled below a few kelvin, especially in the macroscopically ordered exciton state and in vortices of linear polarization, with coherence lengths far exceeding the classical thermal de Broglie scale (High et al., 2011). A trapped gas of indirect excitons in a single wide GaAs quantum well shows the complementary “gray condensate” signatures: dense excitons with anomalously weak photoluminescence, macroscopic spatial coherence, and linear polarization of the weak emitted light, consistent with a predominantly dark condensate coherently coupled to a weak bright component (Alloing et al., 2013). These are coherence markers of the fluid side of excitonic Bose matter and are directly relevant when assessing whether a crystalline exciton phase is merely solid-like or truly supersolid.

6. Relations to neighboring phases and open issues

Exciton Bose solids sit at the intersection of several neighboring phases: exciton condensates, excitonic insulators, exciton Mott states, and coupled Bose–Fermi mixtures. Uniform high-temperature exciton condensation in TMD double layers demonstrates that strong interactions and Bose coherence can coexist without crystallization; that work therefore constrains the “liquid” region of the phase diagram against which solid phases must be compared (Wang et al., 2021). Multicomponent equilibrium exciton condensates in MoSe$a=400$1/hBN/WSe$a=400$2 show that exciton Bose fluids can also carry nontrivial flavor order, with low-field two-component intravalley condensates, intermediate-field two-component intervalley condensates, and high-field single-component condensates (Qi et al., 16 Mar 2026). A plausible implication is that any future exciton Bose solid in such systems may carry not only density order but also spin–valley texture (Qi et al., 16 Mar 2026).

The relationship to excitonic insulators is similarly twofold. In the particle–hole reduced-density-matrix perspective, exciton condensation, excitonic insulators, and BEC–BCS crossover states are unified by ODLRO in $a=400$3 with a dominant eigenvalue $a=400$4 (Torres et al., 2024). This suggests that an exciton Bose solid should be understood not as a separate object from excitonic condensation, but as a spatially ordered realization of it. The two-band coherent-exciton theory makes a related point in a different language: local excitonic pairing $a=400$5 and global phase coherence of the emergent U(1) phase field are distinct scales, and the coherent phase can exhibit a gapless single-particle density of states despite strong excitonic correlations (Apinyan et al., 2016). A plausible implication is that ordered exciton solids and supersolids may require independent control over pairing, phase stiffness, and translational symmetry breaking, rather than emerging from any one of these alone.

Several open issues remain explicit in the literature. Some works identify an exciton solid from transport but lack direct imaging of crystalline order (Zeng et al., 2023). Others image or spectroscopically infer crystalline order but do not yet demonstrate superfluid coherence in the same phase (Qi et al., 27 Jan 2026, Zeng et al., 2022). The 2024 lattice-supersolid work provides evidence of coexistence of symmetry breaking and condensation, but in a driven-dissipative Dicke-Hubbard setting rather than an equilibrium semiconductor bilayer (Lagoin et al., 2024). The 2025 extended GP theory provides a concrete microscopic criterion for an incompressible Bose solid, yet remains a continuum mean-field framework whose decisive experimental test still lies ahead (Conti et al., 27 Jul 2025).

Taken together, the field now supports a coherent taxonomy. A uniform exciton condensate is established in several solid-state platforms (Wang et al., 2021, Qi et al., 16 Mar 2026). Exciton Mott and density-wave precursors are realized in programmable or moiré lattices (Lagoin et al., 2021, Zeng et al., 2022). A thermodynamically stable exciton crystal has been observed in a moiré excitonic insulator (Qi et al., 27 Jan 2026). Exciton supersolids are predicted and, in lattice or hybrid settings, supported by increasingly direct evidence (Matuszewski et al., 2011, Lagoin et al., 2024). In this sense, “exciton Bose solid” has evolved from a loose descriptor of excitonic Bose matter in solids into a technically specific family of correlated phases in which excitons behave as bosons, develop solid-like order, and, in the most stringent formulations, remain phase coherent and incompressible at the same time (Conti et al., 27 Jul 2025).