Topological Condensate: Mechanisms & Models

• Topological condensate is a class of quantum states where the macroscopic order parameter exhibits global invariants like phase winding, vortex pairing, or connectivity.
• These condensates span diverse systems including Bose–Einstein, exciton–polariton, and synthetic setups, with observable effects such as coherence changes and threshold behaviors.
• Analytical and numerical methods—ranging from Gross–Pitaevskii to Bogoliubov–de Gennes models—are used to probe their dynamic transitions and topologically protected modes.

“Topological condensate” is not a single universally fixed term in the current literature. In the cited works, it denotes several related constructions in which a condensate is classified or constrained by topology: a Bose–Einstein condensate whose macroscopic phase carries a winding number (Lyu et al., 2023), a two-dimensional exciton–polariton condensate with Berezinskii–Kosterlitz–Thouless quasi-long-range order stabilized by vortex–antivortex pairing (Caputo et al., 2016), a polariton condensate occupying topologically protected boundary modes of a higher-order lattice (Bennenhei et al., 2024), a condensate whose Bogoliubov quasiparticle bands carry non-zero Chern numbers (Jalali-Mola et al., 2023), and exciton condensates whose order parameters and boundary responses inherit the topology of topological-insulator or HgTe electronic structures (Budich et al., 2013). The phrase therefore refers not to one invariant, but to a family of topological structures attached to condensates: order-parameter winding, defect topology, bulk-boundary correspondence, quasiparticle band topology, or anomaly-controlled condensate formation.

1. Conceptual scope and principal invariants

In one class of usage, topology refers to the condensate phase itself. For a condensate on a ring with phase field $\phi(x)$, the winding number is

$$W=\frac{1}{2\pi}\oint \nabla \phi\cdot d\ell =\frac{1}{2\pi}\big[\phi(L)-\phi(0)\big],$$

and, for a discrete periodic array with site phases $\phi_n$,

$$W=\frac{1}{2\pi}\sum_{n=1}^{N}(\phi_{n+1}-\phi_n)\quad (\mathrm{mod}\ \mathbb{Z}),$$

so distinct phase configurations are separated by singular configurations where the condensate density vanishes (Lyu et al., 2023). In this sense a topological condensate is a condensate whose macroscopic order parameter carries nontrivial circulation or winding.

A second usage identifies topology with topological defect physics rather than with a quantized winding of a strictly coherent phase field. In two-dimensional exciton–polaritons, the “topological condensate” is the quasi-ordered phase of the Berezinskii–Kosterlitz–Thouless scenario, in which vortex–antivortex pairing restores algebraic first-order coherence,

$$g^{(1)}(r)\sim r^{-\alpha}, \qquad g^{(1)}(t)\sim t^{-\alpha}, \qquad 0<\alpha\le \frac14,$$

without true long-range order at finite temperature (Caputo et al., 2016).

A third usage attaches topology to the single-particle or quasiparticle spectrum in which condensation occurs. In higher-order SSH polariton lattices, condensation into 1D and 0D protected modes is classified by bulk polarizations $(P_x,P_y)\in\{0,\tfrac12\}$ derived from Wilson loops (Bennenhei et al., 2024). In the Kagome $\pi$-flux system, the condensate itself may be topologically trivial while its Bogoliubov excitations are topological, characterized by non-zero Chern numbers and chiral edge modes (Jalali-Mola et al., 2023).

A fourth usage is excitonic. In TI and HgTe bilayers, the condensate is “topological” because the interlayer exciton order parameter gaps Dirac-like electronic states in a way that yields fractionalized vortices, helical edge modes, or Majorana boundary channels, depending on symmetry and proximitization (Seradjeh, 2012). These variants are not interchangeable. Accordingly, any precise use of the term requires specifying whether the topology refers to the condensate order parameter, its defects, its occupied modes, its Bogoliubov spectrum, or its embedding in a topological electronic background.

2. Phase winding, unwinding, and ring condensates

The one-dimensional exciton–polariton array studied in “Topological Unwinding in an Exciton-Polariton Condensate Array” provides the clearest order-parameter realization of a topological condensate (Lyu et al., 2023). Two archetypal states occur. The $\pi$-state has nearest-neighbor phase difference $\Delta\phi=\pi\ (\mathrm{mod}\ 2\pi)$, corresponding to a Bloch wave at the Brillouin-zone edge $k=\pi/d$ and non-zero winding. The zero-state has $$W=\frac{1}{2\pi}\oint \nabla \phi\cdot d\ell =\frac{1}{2\pi}\big[\phi(L)-\phi(0)\big],$$0, $$W=\frac{1}{2\pi}\oint \nabla \phi\cdot d\ell =\frac{1}{2\pi}\big[\phi(L)-\phi(0)\big],$$1, and $$W=\frac{1}{2\pi}\oint \nabla \phi\cdot d\ell =\frac{1}{2\pi}\big[\phi(L)-\phi(0)\big],$$2. In the linear Schrödinger equation with perfect periodicity, quasimomentum conservation forbids evolution from $$W=\frac{1}{2\pi}\oint \nabla \phi\cdot d\ell =\frac{1}{2\pi}\big[\phi(L)-\phi(0)\big],$$3 to $$W=\frac{1}{2\pi}\oint \nabla \phi\cdot d\ell =\frac{1}{2\pi}\big[\phi(L)-\phi(0)\big],$$4, and nonvanishing density prevents phase slips. In that limit, distinct winding sectors are topologically protected.

The driven-dissipative condensate dynamics is instead governed by a generalized Gross–Pitaevskii equation

$$W=\frac{1}{2\pi}\oint \nabla \phi\cdot d\ell =\frac{1}{2\pi}\big[\phi(L)-\phi(0)\big],$$5

so nonlinearity, gain, and loss can destroy the linear topological protection (Lyu et al., 2023). The paper identifies the combined mechanism: nonlinear density modulation drives dynamical instability of the $$W=\frac{1}{2\pi}\oint \nabla \phi\cdot d\ell =\frac{1}{2\pi}\big[\phi(L)-\phi(0)\big],$$6-state, while dissipation and gain saturation allow localized zeros of $$W=\frac{1}{2\pi}\oint \nabla \phi\cdot d\ell =\frac{1}{2\pi}\big[\phi(L)-\phi(0)\big],$$7 at which the phase is undefined. These zeros act as defects through which phase slips occur, each changing $$W=\frac{1}{2\pi}\oint \nabla \phi\cdot d\ell =\frac{1}{2\pi}\big[\phi(L)-\phi(0)\big],$$8 by an integer and enabling “topological unwinding” from the $$W=\frac{1}{2\pi}\oint \nabla \phi\cdot d\ell =\frac{1}{2\pi}\big[\phi(L)-\phi(0)\big],$$9-state to the zero-state.

The numerical and experimental signatures are explicitly collective. During unwinding, the momentum distribution broadens strongly and contains many $\phi_n$0 components, unlike a simple two-mode competition picture. In simulations with 200 cells, the $\phi_n$1-space full width at half maximum near $\phi_n$2 reaches $\phi_n$3, while the experiment shows a sudden increase of the zero-state peak width around pump power $\phi_n$4, consistent with the onset of collective unwinding (Lyu et al., 2023). In this usage, a topological condensate is therefore not merely a current-carrying state, but one whose relaxation channels are controlled by singular defects and integer changes of winding number.

Ring spinor condensates provide a closely related but spin-resolved extension. In annular exciton–polariton condensates, topological spin Meissner states are constant-amplitude or symmetry-broken states with component windings $\phi_n$5 for the circular polarizations $\phi_n$6, and the effective compensation field is shifted by

$\phi_n$7

The population imbalance satisfies

$\phi_n$8

so the chemical potential is field-independent inside the existence window $\phi_n$9 (Gulevich et al., 2016). This generalizes the spin Meissner effect to nontrivial topology of the condensate wavefunction and permits topological Meissner states at arbitrarily high magnetic field by choosing suitable winding numbers. The same framework identifies ring half-vortices as states with $$W=\frac{1}{2\pi}\sum_{n=1}^{N}(\phi_{n+1}-\phi_n)\quad (\mathrm{mod}\ \mathbb{Z}),$$0 or $$W=\frac{1}{2\pi}\sum_{n=1}^{N}(\phi_{n+1}-\phi_n)\quad (\mathrm{mod}\ \mathbb{Z}),$$1 and shows that, at zero magnetic field, they occur as superpositions of elementary half-vortex states; a pure half-vortex requires a non-zero field tuned so that $$W=\frac{1}{2\pi}\sum_{n=1}^{N}(\phi_{n+1}-\phi_n)\quad (\mathrm{mod}\ \mathbb{Z}),$$2 (Gulevich et al., 2016).

3. Defect-mediated order and the role of connectivity

In two-dimensional exciton–polariton condensates, the decisive topological objects are vortices rather than winding sectors of a globally phase-rigid order parameter. The work “Topological order and equilibrium in a condensate of exciton-polaritons” reports a Berezinskii–Kosterlitz–Thouless transition in a reservoir-free region of a long-lifetime polariton system, with simultaneous measurements of spatial and temporal first-order coherence (Caputo et al., 2016). Below threshold, coherence decays exponentially or Gaussian-like; near threshold the decay is stretched exponential; in the quasi-ordered phase, both spatial and temporal correlations exhibit algebraic tails with matching exponents, specifically $$W=\frac{1}{2\pi}\sum_{n=1}^{N}(\phi_{n+1}-\phi_n)\quad (\mathrm{mod}\ \mathbb{Z}),$$3 and $$W=\frac{1}{2\pi}\sum_{n=1}^{N}(\phi_{n+1}-\phi_n)\quad (\mathrm{mod}\ \mathbb{Z}),$$4 at density $$W=\frac{1}{2\pi}\sum_{n=1}^{N}(\phi_{n+1}-\phi_n)\quad (\mathrm{mod}\ \mathbb{Z}),$$5, consistent with the BKT bound $$W=\frac{1}{2\pi}\sum_{n=1}^{N}(\phi_{n+1}-\phi_n)\quad (\mathrm{mod}\ \mathbb{Z}),$$6 (Caputo et al., 2016). Stochastic simulations directly show the disappearance of free vortices and the completion of vortex–antivortex pairing in the algebraic phase. Here the topological condensate is the defect-ordered phase itself.

Connectivity can also be the relevant topological variable. In quasi-2D atomic condensates with toroidal and bow-shaped geometries, the topology concerns whether the condensate density is multiply connected or simply connected. Under a gradual offset between a harmonic trap and a Gaussian barrier, the condensate density and quantum depletion follow the topology change from a torus to a simply connected profile, whereas the thermal cloud remains multiply connected because of the structure of the low-energy quasiparticles (Roy et al., 2016). This usage is distinct from BKT order and from winding-number classifications: topology refers to the connectivity of the condensate support in real space.

These two cases illustrate a broader point. A topological condensate need not have a quantized current or a topological band. The controlling topological data may instead be the pairing or unbinding of vortices, or the connectedness of the density profile that supports the condensate and its low-energy modes.

4. Condensation in topological modes and synthetic geometries

A different and increasingly common meaning identifies a topological condensate with macroscopic occupation of topologically protected single-particle modes. In the room-temperature organic higher-order SSH polariton lattice, the relevant invariants are quantized bulk polarizations and their associated boundary states (Bennenhei et al., 2024). The 2D lattice is built from four unit-cell types with $$W=\frac{1}{2\pi}\sum_{n=1}^{N}(\phi_{n+1}-\phi_n)\quad (\mathrm{mod}\ \mathbb{Z}),$$7, $$W=\frac{1}{2\pi}\sum_{n=1}^{N}(\phi_{n+1}-\phi_n)\quad (\mathrm{mod}\ \mathbb{Z}),$$8, $$W=\frac{1}{2\pi}\sum_{n=1}^{N}(\phi_{n+1}-\phi_n)\quad (\mathrm{mod}\ \mathbb{Z}),$$9, and $$g^{(1)}(r)\sim r^{-\alpha}, \qquad g^{(1)}(t)\sim t^{-\alpha}, \qquad 0<\alpha\le \frac14,$$0, and the mismatch of these polarizations across domain walls creates protected 1D line modes and a 0D corner mode (Bennenhei et al., 2024). Under strong optical pumping, bosonic condensation is observed into these topological modes. For the 1D defect mode, the threshold is $$g^{(1)}(r)\sim r^{-\alpha}, \qquad g^{(1)}(t)\sim t^{-\alpha}, \qquad 0<\alpha\le \frac14,$$1, with linewidth collapse to the instrumental limit of $$g^{(1)}(r)\sim r^{-\alpha}, \qquad g^{(1)}(t)\sim t^{-\alpha}, \qquad 0<\alpha\le \frac14,$$2 and coherence extending over $$g^{(1)}(r)\sim r^{-\alpha}, \qquad g^{(1)}(t)\sim t^{-\alpha}, \qquad 0<\alpha\le \frac14,$$3 along the protected channel; for the 2D corner mode, the threshold is $$g^{(1)}(r)\sim r^{-\alpha}, \qquad g^{(1)}(t)\sim t^{-\alpha}, \qquad 0<\alpha\le \frac14,$$4 with a blueshift of $$g^{(1)}(r)\sim r^{-\alpha}, \qquad g^{(1)}(t)\sim t^{-\alpha}, \qquad 0<\alpha\le \frac14,$$5 (Bennenhei et al., 2024). In this setting, “topological condensate” means a coherent polariton population localized in modes enforced by bulk-boundary correspondence.

A related but geometrically distinct construction is the synthetic Hall cylinder for ultracold atoms. There the condensate lives on a synthetic cylindrical manifold built from one real spatial dimension and a cyclic spin dimension, threaded by a synthetic radial flux (Li et al., 2018). The band structure is protected by a nonsymmorphic symmetry $$g^{(1)}(r)\sim r^{-\alpha}, \qquad g^{(1)}(t)\sim t^{-\alpha}, \qquad 0<\alpha\le \frac14,$$6, with $$g^{(1)}(r)\sim r^{-\alpha}, \qquad g^{(1)}(t)\sim t^{-\alpha}, \qquad 0<\alpha\le \frac14,$$7, and exhibits symmetry-enforced crossings that disappear in the planar “unzipped” counterpart (Li et al., 2018). Transport of the condensate through these bands yields Bloch oscillations with period multiplier $$g^{(1)}(r)\sim r^{-\alpha}, \qquad g^{(1)}(t)\sim t^{-\alpha}, \qquad 0<\alpha\le \frac14,$$8, interpreted as a Möbius-strip topology in momentum space. Breaking the symmetry with RF couplings opens a gap and restores $$g^{(1)}(r)\sim r^{-\alpha}, \qquad g^{(1)}(t)\sim t^{-\alpha}, \qquad 0<\alpha\le \frac14,$$9 (Li et al., 2018). Here the condensate is “topological” because its band transport is inseparable from a nontrivial synthetic geometry and symmetry-protected crossings.

Both cases concern topological protection of the modes available to condensation, not topology of the condensate phase field itself. The former is a bulk-band or higher-order topological classification; the latter is an order-parameter classification.

5. Topological quasiparticles, orbital order, and dynamical topology

A condensate can also be called topological when the nontrivial topology resides in its excitations or in its collective dynamics. In the Kagome $(P_x,P_y)\in\{0,\tfrac12\}$0-flux Bose system, mean-field interactions admit degenerate condensates at the $(P_x,P_y)\in\{0,\tfrac12\}$1 and $(P_x,P_y)\in\{0,\tfrac12\}$2 points, but only the $(P_x,P_y)\in\{0,\tfrac12\}$3-point condensate has a Bogoliubov–de Gennes spectrum with non-zero Chern numbers and chiral edge states (Jalali-Mola et al., 2023). The distinction is symmetry-based: for the $(P_x,P_y)\in\{0,\tfrac12\}$4 condensate, complex-valued condensate amplitudes break TRS in the BdG Hamiltonian while preserving particle-hole-like symmetry; for the $(P_x,P_y)\in\{0,\tfrac12\}$5 condensate, TRS is preserved and the quasiparticle bands remain topologically trivial (Jalali-Mola et al., 2023). With increasing nearest-neighbor interaction $(P_x,P_y)\in\{0,\tfrac12\}$6, the $(P_x,P_y)\in\{0,\tfrac12\}$7-condensate BdG bands undergo a topological phase transition near $(P_x,P_y)\in\{0,\tfrac12\}$8, changing from $(P_x,P_y)\in\{0,\tfrac12\}$9 to $\pi$0 (Jalali-Mola et al., 2023). In this usage, the condensate order parameter is not itself the topological invariant; the topology belongs to the quasiparticle spectrum built on top of it.

Time can also be the topological dimension of interest. In the low-energy 1D BEC studied in “Dynamical Topological Phase Transition in a Bose-Einstein Condensate,” a sudden quench from a non-displaced harmonic-fluid Hamiltonian to a displaced one produces mode-resolved geometric phases

$\pi$1

whose collective synchronization at the global period $\pi$2 generates nonanalyticities in the total geometric phase and in the dynamical topological order parameter

$\pi$3

The Loschmidt amplitude exhibits Fisher-zero crossings at $\pi$4, and the resulting dynamical criticality has universal exponent $\pi$5 (Abdi, 2019). This is not equilibrium topological order, but a topological classification of collective time evolution.

A further extension is the higher-band atom-cavity system that forms chiral $\pi$6 order and then, with increasing pump strength, enters a self-organized topological superfluid state (Kleine-Pollmann et al., 3 Mar 2026). In that system a superradiant checkerboard density wave rectifies staggered orbital chirality into a global topological superfluid, and both truncated-Wigner simulations and mean-field analysis identify the transition as first order, with a discontinuous onset of cavity photons and an approximate threshold at $\pi$7 for the stated parameter set (Kleine-Pollmann et al., 3 Mar 2026). The relevant topology is carried by the orbital order and the Bogoliubov bands of the rectified chiral state.

Interaction-induced gauge fields provide another route to topological ground states. A BEC with a density-dependent synthetic vector potential can develop a coreless vortex–antivortex pair in 2D and a coreless vortex ring in 3D as analytic Thomas–Fermi ground states, with persistent current driven by a synthetic magnetic field proportional to density gradients (Zheng et al., 2014). These are topological condensates in the sense that the ground-state flow texture is generated self-consistently by many-body gauge feedback rather than by imposed rotation.

6. Excitonic topological condensates in topological materials

In topological-insulator and HgTe systems, the term usually refers to an interlayer exciton condensate whose order parameter is tied to topological electronic structure. In TI thin films, the condensate is an intersurface coherent state with order parameter

$\pi$8

which acts as an interlayer mass in the low-energy Dirac theory (Seradjeh, 2012). For imbalanced electrons and holes, the TI thin-film exciton condensate admits a spatially modulated FFLO-like phase that preempts a first-order transition from the uniform condensate to the normal state at low temperature, and a vortex of the condensate carries a precisely fractional charge $\pi$9 (Seradjeh, 2012). Near the transition, Coulomb drag provides an experimentally accessible probe: the drag resistivity shows an upturn on cooling toward the mean-field critical temperature, with the onset of the upturn near $\pi$0 (Mink et al., 2011).

In three-dimensional TIs, the full 3D Hamiltonian is required to identify the correct pairing symmetry. The equilibrium condensate is an intersurface excitonic superfluid in which only particles with similar chirality play a significant role in condensation, and the intersurface polarizability vanishes in the condensed phase, suppressing surface current flow and leaving only intersurface current through the bulk (Kim et al., 2012). This resolves a major concern specific to 3D TIs, namely that side-surface connectivity might destroy intersurface coherence.

When such a topological exciton condensate is contacted by superconductors, the relative phase between the superconductors and the condensate becomes topologically decisive. For two TI surfaces with a superconducting phase difference satisfying

$\pi$1

the boundary hosts a pair of counter-propagating Majorana modes, protected by an emergent antiunitary symmetry $\pi$2 with $\pi$3 (Seradjeh, 2012). In this case, the condensate itself is an excitonic mass, while the topological boundary physics is Majorana.

Bilayer critical HgTe quantum wells realize a time-reversal-symmetric version. The mean-field bilayer Hamiltonian is classified by a $\pi$4 invariant, and different excitonic channels can be topologically trivial or nontrivial even though all are interaction-driven interlayer condensates (Budich et al., 2013). The channels labeled xz0 and yzz are nontrivial and support helical edge states, whereas yxx and yxy are trivial (Budich et al., 2013). The paper identifies this as a helical topological exciton condensate, notable because topology and spontaneous symmetry breaking are intertwined at the level of the order-parameter channel itself.

7. Limits of protection and the breadth of the term

One recurrent misconception is that any topology attached to a condensate automatically guarantees dynamical stability. The cited literature repeatedly shows that this is not generally correct. In the polariton array, winding-number protection fails once nonlinearity, gain, loss, and defects are admitted; the $\pi$5-state can unwind spontaneously to the zero-state through repeated phase slips (Lyu et al., 2023). In that setting, topology constrains the allowed unwinding mechanism, but does not prevent it.

An even stronger challenge appears in spin-1 spinor condensates. The AFM and FM order-parameter submanifolds have familiar homotopy classifications, but the full GP dynamics does not preserve either submanifold. A state prepared entirely within the AFM or FM manifold immediately evolves into a mixed state, so homotopy classifications internal to those submanifolds do not define dynamically invariant topological sectors. The paper further identifies the linear Zeeman effect as an efficient catalyst for extracting the alternate component and therefore as a practical tool for “topology engineering” rather than for maintaining strict topological isolation (Oh et al., 2013).

The semantic breadth of the term is also wider than bosonic order parameters. For Kähler–Dirac fermions on compact curved manifolds, a global $\pi$6 anomaly produces a condensate whose magnitude is fixed by global topology rather than by local dynamics, scaling with the Euler characteristic and the volume, and the anomaly survives discretization on arbitrary triangulations (Catterall et al., 2018). In that context, the symmetry-breaking pattern is

$\pi$7

and the condensate is topological because it is indexed by $\pi$8, not because it carries a winding or occupies a topological band (Catterall et al., 2018).

Accordingly, “topological condensate” is best understood as a family resemblance term. Across current usage, it may denote a condensate with nontrivial phase winding, a defect-ordered condensate, a condensate occupying topological boundary modes, a condensate whose Bogoliubov spectrum is topological, an exciton condensate inheriting topological electronic structure, or an anomaly-determined fermionic condensate. The common element is that some robust feature of the condensate or its excitations is classified by global topology or by a topological invariant; the specific invariant, however, must always be stated.