---
title: 'Dipole Condensates: Quantum Fluids & Lattice Systems'
url: https://www.emergentmind.com/topics/dipole-condensates
type: topic
---

# Dipole Condensates: Quantum Fluids & Lattice Systems

Searching arXiv for recent and relevant papers on dipole condensates to ground the article in current literature.
arxiv_search query: "dipole condensates Bose-Einstein condensate dipolar droplets supersolids tilted Bose-Hubbard chains"
Dipole condensates are quantum-condensed states in which dipolar degrees of freedom are central to the order parameter, collective modes, or dominant interactions. In the literature, the term is used in at least two technically distinct senses. In dipolar Bose–Einstein condensates, it refers to atomic or molecular condensates governed by long-range, anisotropic magnetic or electric dipole–dipole interactions, often supplemented by beyond-mean-field stabilization terms or conservative three-body interactions [2111.06187]. In strongly constrained lattice systems, it denotes condensation of neutral particle–hole dipoles, characterized by two-body off-diagonal order such as $\langle b_i^\dagger b_{i+\mathbf s}\rangle\neq 0$ even when $\langle b_i\rangle=0$ [2306.14214]. Both usages emphasize that dipolar objects, rather than conventional scalar condensate order alone, control the macroscopic physics.

## 1. Terminological scope and definitions

A spinor dipolar condensate is a spinor Bose–Einstein condensate in which the long-range, anisotropic magnetic dipole–dipole interaction plays a role on the same footing as the $s$-wave collisions. For spin-1 gases, the order parameter is $\psi_\alpha(\mathbf r,t)$ with $\alpha=+1,0,-1$, and the magnetic dipole–dipole interaction enters through
\[
V_{dd}(\mathbf r-\mathbf r')=\frac{\mu_0(g_e\mu_B)^2}{4\pi}\,
\frac{F_i(\mathbf r)F_j(\mathbf r')(\delta_{ij}-3e_ie_j)}{|\mathbf r-\mathbf r'|^3},
\]
with local spin density $F_i(\mathbf r)=\Psi^\dagger(\mathbf r)\hat F_i\Psi(\mathbf r)$ [1010.1075].

A different definition appears in fracton-related and tilted-lattice settings. There a dipole condensate is defined by off-diagonal long-range order in a two-body correlator,
\[
\langle D_{i,\mathbf s}\rangle \equiv \langle b_i^\dagger b_{i+\mathbf s}\rangle \neq 0
\quad (\mathbf s\neq 0),
\qquad
\langle b_i\rangle=0,
\]
so that a bound particle–hole pair is the condensing object [2306.14214]. In the one-dimensional dipole-conserving Bose–Hubbard chain, the natural condensed operator is $d_n\equiv b_n b_{n+1}$ or equivalently the nearest-neighbor dipole correlator $\langle b_n^\dagger b_{n+1}\rangle$, while single-boson correlators remain exponentially decaying [2210.02470].

The coexistence of these usages is not merely terminological. In the first case, the condensate is made of particles carrying permanent or induced dipole moments; in the second, the condensed object is itself a dipole. This suggests that “dipole condensate” is best treated as a family resemblance term spanning dipolar quantum fluids and condensates of dipolar quasiparticles rather than a single microscopic phase.

## 2. Field-theoretic and mean-field descriptions of dipolar condensates

For dipolar Bose gases, the standard starting point is an extended Gross–Pitaevskii functional. In molecular condensates with strong dipole–dipole interactions,
\[
E[\psi]=\int d^3r\left\{
\frac{\hbar^2}{2m}|\nabla\psi(\mathbf r)|^2
+\frac{g}{2}|\psi(\mathbf r)|^4
+\frac12\int d^3r'|\psi(\mathbf r)|^2V_{dd}(\mathbf r-\mathbf r')|\psi(\mathbf r')|^2
+E_{\mathrm{LHY}}[\psi]
\right\},
\]
with
\[
g=\frac{4\pi\hbar^2a_s}{m},\qquad
V_{dd}(r)=\frac{\mu_0\mu^2}{4\pi}\frac{1-3\cos^2\theta}{r^3},\qquad
a_{dd}=\frac{\mu_0\mu^2m}{12\pi\hbar^2},\qquad
\epsilon_{dd}=\frac{a_{dd}}{a_s}.
\]
The Lee–Huang–Yang term is treated in local-density approximation as
\[
E_{\mathrm{LHY}}[\psi]=\int d^3r\,\frac25 g_{qf}|\psi(\mathbf r)|^5,
\qquad
g_{qf}=\frac{32}{3\sqrt\pi}\,g\,a_s^{3/2}\left(1+\frac32\epsilon_{dd}^2\right)
\]
[2111.06187].

Closely related formulations appear in strongly dipolar atomic condensates. In the generalized nonlocal nonlinear Schrödinger equation of Wächtler and Santos, the beyond-mean-field term is written through
\[
\Delta\mu(n,\epsilon_{dd})=
\frac{32}{3\sqrt\pi}\,g\,n\,\sqrt{n\,a^3}\,F(\epsilon_{dd}),
\]
with
\[
F(\epsilon_{dd})=\frac12\int_0^\pi d\theta\,\sin\theta\,
\left[1+\epsilon_{dd}(3\cos^2\theta-1)\right]^{5/2},
\]
and the dipolar mean field
\[
\Phi_{dd}(\mathbf r)=\int d^3r'\,V_{dd}(\mathbf r-\mathbf r')\,n(\mathbf r')
\]
[1601.04501].

Flattened and quasi-two-dimensional geometries admit reduced descriptions with effective interaction parameters. For harmonic confinement along one direction, Baillie and Blakie introduce
\[
\nu=4\pi n a_z\left[a+(3\cos^2\alpha-1)a_{dd}\right],\qquad
\nu_d=4\pi n a_z a_{dd}\cos^2\alpha,
\]
which control the axial condensate profile, roton structure, and mean-field stability [1407.4252].

Not all stabilization mechanisms are fluctuation-driven. A separate line of work augments the dipolar Gross–Pitaevskii theory by a conservative three-body interaction $\kappa_3|\psi|^6$, giving
\[
E[\psi]=\int d^3r\left[
\frac{\hbar^2}{2m}|\nabla\psi|^2+V_{\rm trap}|\psi|^2+\frac g2|\psi|^4
\right]
+\frac12\int d^3r\,d^3r'\,U_{dd}(\mathbf r-\mathbf r')|\psi(\mathbf r)|^2|\psi(\mathbf r')|^2
+\frac{\kappa_3}{6}\int d^3r\,|\psi|^6
\]
[1601.05869]. A Gaussian-state theory goes further by replacing the usual coherent-state ansatz with a multimode squeezed coherent state and adding an explicit three-body Hamiltonian
\[
H_{3B}=\frac{g_3}{3!}\int d^3r\,\hat\psi^{\dagger 3}(\mathbf r)\hat\psi^3(\mathbf r),
\]
thereby treating squeezing and coherent occupation self-consistently [2002.11298].

## 3. Self-bound droplets, supersolids, and interaction-driven liquid behavior

A central result of the dipolar-condensate literature is that mean-field collapse need not be the endpoint of strong dipolar attraction. On the mean-field level a purely dipolar condensate with $\epsilon_{dd}>1$ would collapse, but the LHY term provides a repulsive $n^{5/2}$ pressure which can stabilize a finite-size self-bound droplet [2111.06187]. In molecular Bose–Einstein condensates, the critical molecule number obeys a universal curve
\[
N_{\rm crit}=F(\epsilon_{dd}),
\]
which diverges as $\epsilon_{dd}\to1^+$ and falls rapidly for larger $\epsilon_{dd}$. The droplet central density saturates as
\[
n_{\rm sat}\sim \left(\frac{a_s^3}{a_{dd}^5}\right)\times O(1),
\]
and numerically follows $n_{\rm sat}\propto a_{dd}^{-B}$ with $B\approx0.6$–$0.7$ over a wide range, while the droplet sizes satisfy $R\propto (N a_{dd})^{1/2}$ in the large-$N$, saturated regime [2111.06187].

In trapped pancake geometries the same framework yields a phase diagram in $(\epsilon_{dd},N)$ with three main regimes: dilute BEC for $\epsilon_{dd}\lesssim1$, roton-softened gauge-broken states with “blood-cell” shapes once a roton gap closes, and symmetry-broken supersolids and isolated droplet arrays at still larger $\epsilon_{dd}$. The empirical boundaries are
\[
\epsilon_{dd}N^{1/2}\simeq C_1,\qquad
\epsilon_{dd}N^\alpha\simeq C_2,\quad \alpha\approx \frac12-\frac23,
\]
with constants set by trap geometry. In pancake geometry the roton minimum appears for $k\parallel x,y$ when $\epsilon_{dd}>\epsilon_{dd}^c\approx1.3$, and density modulation may be quantified by
\[
C=\frac{n_{\max}-n_{\min}}{n_{\max}+n_{\min}}.
\]
The Leggett bound gives
\[
f_s\ge \left[\frac1V\int d^3r\,\frac{n_{\max}}{n(\mathbf r)}\right]^{-1},
\]
and numerical solutions yield $f_s\sim0.5$–$0.9$ deep in the supersolid regime [2111.06187].

For dysprosium parameters, arrested collapse produces filament-like droplets. Wächtler and Santos identify a minimal droplet atom number $N_{\min}\simeq9\times10^2$ at $a=70\,a_B$, and a larger threshold $N_0\sim1.5\times10^3$ where the internal energy passes through zero. Numerical solutions give axial half-length $Z\approx2\,\mu$m, transverse radius $R_\perp\approx0.3\,\mu$m, aspect ratio $Z/R_\perp\gtrsim6$, and peak density $n_{\rm peak}\sim1$–$2\times10^{21}\,{\rm m}^{-3}$ [1601.04501].

The stabilization mechanism is not unique across models. In the trapped theory with conservative three-body repulsion, two distinct fluid states appear: a low-density phase and a high-density droplet phase, separated by a first-order transition terminating at a critical point [1601.05869]. In Gaussian-state theory for dysprosium, self-bound states are instead split into an expanding gas, a self-bound gas, and a self-bound liquid, with the self-bound gas–self-bound liquid boundary marked by an abrupt jump in the coherent fraction $f_c$ [2002.11298].

Interaction anisotropy can also change droplet dimensionality. In a doubly dipolar condensate with both magnetic and electric dipole moments, varying the angle $\alpha$ between the two dipoles produces a crossover from quasi-1D cigar droplets elongated along the effective polarization axis to quasi-2D pancake droplets flat in the plane perpendicular to the soft direction. For dysprosium at typical fields this occurs near $\alpha_a\approx1.24\,{\rm rad}$ [1907.08190]. At still stronger dipolarity, a harmonically confined condensate with a central Gaussian hill can host a hollow cylindrical quasi-one-dimensional metastable droplet with ring topology, with a ring branch appearing as the second excited stationary state for $a_{dd}=2000\,a_0$, $a_s=100\,a_0$, and $N$ between $2\,000$ and $8\,000$ [2507.10396].

## 4. Geometry, polarization, and interaction engineering

Dipolar condensates are unusually sensitive to geometry because the kernel $(1-3\cos^2\theta)/r^3$ couples density or spin structure directly to trap anisotropy and polarization orientation. In a quasi-two-dimensional condensate with dipoles tilted in the $xz$ plane by angle $\alpha$, the effective 2D momentum-space kernel is
\[
U_{2D}(k,\alpha)=\frac{2g_d}{3l_z}f(q,\theta_k),
\]
where
\[
f(q,\theta_k)=\sqrt{2\pi}(3\cos^2\alpha-1)
+3\pi e^{q^2/2}q\,{\rm erfc}(q/\sqrt2)\,[\sin^2\alpha\cos^2\theta_k-\cos^2\alpha].
\]
The long-wavelength effective coupling becomes
\[
g_{\rm eff}(\alpha)=g+\frac{4\pi}{3}g_d(3\cos^2\alpha-1).
\]
At the magic angle $\alpha_m=\arccos(1/\sqrt3)\approx54.7^\circ$, one has $g_{\rm eff}=0$; for $\alpha>\alpha_m$, phonon instability sets in, while for $\alpha$ just below $\alpha_m$ an anisotropic roton minimum at finite $k_y$ appears [1608.06616].

The post-instability dynamics depends on which branch softens. Post-phonon-instability dynamics yields stripe patterns with dislocations and eventual breakup into anisotropic bright solitons, whereas post-roton-instability dynamics yields transient, defect-free stripes followed by local collapse at stripe peaks. Slowly ramping parameters back into the stable roton region can make the stripes persist for much longer, and weak dissipation causes the stripe amplitude slowly to decay while stripe–stripe hopping is observed before the condensate returns to uniformity [1608.06616].

Multilayer and disconnected-condensate settings provide additional control knobs. In a multilayer stack formed by a strong optical lattice, effective intra- and interlayer dipole–dipole interaction potentials can be derived analytically; the interlayer term changes the transverse aspect ratio of the ground state in central layers and reduces the excitation energy of local perturbations, affecting the development of a roton minimum [1201.6176]. In physically disconnected dipolar condensates, a trapped control condensate can engineer the density of a target condensate through inter-condensate dipole–dipole interactions. With two control condensates separated by $z_0$, a critical spacing $z_c\approx40\,l_\perp$ marks where the central minimum at $z=0$ turns into a small local maximum, producing single-peak, three-peak, and four-peak target densities; a coherence measure
\[
\mathcal C_{i,i+1}=
\frac{n_{\max}^{(i,i+1)}-n_{\min}^{(i,i+1)}}
{n_{\max}^{(i,i+1)}+n_{\min}^{(i,i+1)}}
\]
diagnoses whether adjacent peaks are phase locked or effectively isolated [2309.13000].

The dipolar interaction itself can be induced and enhanced. In an off-resonantly illuminated neutral-atom condensate, azimuthal averaging of the retarded induced interaction gives
\[
\bar U_{\rm DD}(r,\theta)
=\frac{d^2}{r^3}
\left[
(1-3\cos^2\theta)\bigl(\cos kr+kr\sin kr\bigr)
-\sin^2\theta\,(kr)^2\cos kr
\right]J_0(k\rho),
\]
which reduces to the familiar $1/r^3$ form for $kr\ll1$ and crosses over to an oscillatory $\cos(kr)/r$ interaction for $kr\gg1$. The full calculation yields an enhancement by up to thirty-fold at short wavelengths [1610.05477].

An alternative engineering route exploits geometry rather than optical retardation. For cigar-shaped condensates loaded into a one-dimensional lattice, projection onto Wannier–BEC modes yields intersite dipolar matrix elements
\[
U_{d,ij}\approx\frac{G}{(x_i-x_j)^2},
\]
so that the continuum dilute-fluctuation limit becomes a Calogero–Sutherland Hamiltonian with inverse-square interactions [1108.4107].

## 5. Spinor dipole condensates, magnetization dynamics, and vortical textures

In spinor condensates, dipolar interactions generate magnetization dynamics with direct analogies to condensed-matter ferromagnets. Starting from the spin-1 Gross–Pitaevskii equations, Yasunaga and Tsubota derive for the local spin density
\[
\partial_t\mathbf F=\mathbf K+\gamma_e\,\mathbf F\times \mathbf H_{\rm eff},
\]
where $\mathbf K$ is a quantum-pressure or spin-diffusion term and
\[
\mathbf H_{\rm eff}(\mathbf r)=H\,\hat z-\frac{c_{dd}}{g\mu_B}(d_x,d_y,d_z).
\]
Under the single-mode approximation this reduces to
\[
\frac{d\mathbf S}{dt}=\gamma_e\,\mathbf S\times\mathbf H_{\rm eff}^{\rm SMA},
\]
and linearization about $S=S_0\hat z$ gives
\[
\omega^2=\gamma_e^2
\left[H+(N_{dd}^x-N_{dd}^z)S_0\right]
\left[H+(N_{dd}^y-N_{dd}^z)S_0\right].
\]
For a spherical condensate $N_{dd}^x=N_{dd}^y=N_{dd}^z$, so $\omega=\gamma_eH$, whereas elliptic, pancake, and rod geometries shift the resonance through the shape dependence of $N_{dd}^i$ [1010.1075].

When rotation is added, dipolar spinor condensates support an extensive defect taxonomy. Simula and collaborators compute phase diagrams in the plane $(g_d/g,\Omega/\omega_\perp)$ and identify regions with no defects, spontaneous spin currents in a single polar-core spin vortex $(0,1)$, a single Mermin–Ho skyrmion $(1,1)$, a staggered skyrmion lattice, a staggered half-quantum vortex lattice, vortex sheets of $(1/2,1/2)$ defects, and a mixed lattice of $(1,1)$ and $(1,0)$ vortices, with collapse beyond $g_d\approx0.25$ [1007.3551].

Even without imposed rotation, semiclassical and dual descriptions predict spontaneous textures. A three-component real order parameter $\psi(\mathbf r)=(\psi_x,\psi_y,\psi_z)$ with magnetization density $M_k(\mathbf r)=\mu\,\psi_k(\mathbf r)\sqrt{n(\mathbf r)}$ yields ground states including a flare state, a spin-vortex state, and in elongated traps a spin helix state with local azimuthal angle $\theta(z)=\arctan[M_y(0,0,z)/M_x(0,0,z)]$ and wavevector $k(z)=\partial\theta/\partial z$ [1007.2027]. The same texture landscape can be reformulated through a Dirac string gas with negative string tension,
\[
T_{\rm str}=-\frac{\lambda\,d\sigma}{6},
\]
which favors closed strings, small curvature, and widely separated residual monopoles. In that language, three-dimensional condensates prefer a meron-like vortex texture, quasi-one-dimensional condensates prefer the axially polarized flare texture, and quasi-two-dimensional condensates exhibit either a meron texture or an in-plane polarized texture [1501.02019].

In dominantly dipolar scalar condensates, vortices themselves acquire dipolar signatures. With the extended Gross–Pitaevskii functional including
\[
\frac25\gamma_{QF}\int d^3r\,|\Psi(r)|^5,
\]
Bland and collaborators find that tilting the magnetic field induces magnetostriction and permits vortex nucleation by rotating the polarization direction. For $a_s=110\,a_0$, $N=1.5\times10^4$, and $\gamma=2.6$, vortices appear via a dynamical instability at
\[
\Omega_c\simeq0.8\,\omega_r
\]
for tilt angle $\theta\approx35^\circ$. They identify three dynamical regimes: $0^\circ\le \theta\lesssim20^\circ$ as a stable Thomas–Fermi branch, $20^\circ\lesssim\theta\lesssim50^\circ$ as a “Goldilocks window” for vortex penetration and stripe-lattice formation, and $50^\circ\lesssim\theta\le90^\circ$ as a “droplet-like stability” regime where strong head-to-tail attraction suppresses the instability [2303.13263].

## 6. Dipole condensation in constrained lattice systems and multipolar generalizations

In tilted Bose–Hubbard systems, the word “dipole” refers to a mobile particle–hole composite rather than to a permanent magnetic or electric moment. For a strongly tilted chain, a Schrieffer–Wolff expansion yields an effective dipole-conserving Hamiltonian
\[
H_{\rm DBHM}=
-\sum_n\left[t_3\,b_{n-1}^\dagger b_n^2 b_{n+1}^\dagger
+t_4\,b_{n-1}^\dagger b_n b_{n+1} b_{n+2}^\dagger
+{\rm h.c.}\right]
+\frac12 U_{\rm eff}\sum_n n_n^2
+V_{\rm eff}\sum_n n_n n_{n+1}+\cdots,
\]
which commutes with both total boson number and total dipole moment [2210.02470]. Its low-energy field theory is a quantum Lifshitz theory with Gaussian Lagrangian
\[
\mathcal L_0=\frac12 K_\tau(\partial_\tau\phi)^2+\frac12 K_D(\partial_x^2\phi)^2.
\]
Unlike the conventional Bose–Hubbard model, the phase diagram contains no compressible phases and is instead dominated by various types of exotic dipolar condensates [2210.02470].

The diagnostic of these phases is the mismatch between one-body and two-body coherence. In the dipole-condensed phase,
\[
\langle d_n d_m\rangle_c\sim |n-m|^{-K_D^*},
\]
while
\[
\langle b_n b_m^\dagger\rangle\sim e^{-|n-m|/\xi_b},
\]
and the charge compressibility vanishes. The phase is therefore gapless but incompressible, with power-law order in neutral dipole operators rather than in single-boson operators [2210.02470].

A broader claim is advanced in the multipolar-condensate literature: dipole condensation can arise generically in bosonic systems through a self-proximity effect. In the Mott regime of the conventional Bose–Hubbard model,
\[
|\Psi\rangle\approx |\mathrm{MI}\rangle+\frac{t_1}{U}\sum_{\langle i,j\rangle}
(b_i^\dagger b_j+{\rm h.c.})|\mathrm{MI}\rangle,
\]
so that
\[
\langle b_i\rangle=0,\qquad
\langle D_i\rangle=\langle b_i^\dagger b_{i+1}\rangle
=2\,\frac{t_1}{U}\,n_0(n_0+1)\neq0.
\]
This “self-proximity effect” is then argued to guarantee $\langle D\rangle\neq0$ for any nonzero $t_1$, even when $\langle b\rangle$ itself vanishes [2306.14214]. The same work proposes a dipolar Josephson effect with
\[
J_D=A\sin(\theta_1-\theta_2),
\]
so that supercurrents of dipoles arise in the absence of particle flows, and extends the mechanism to a hierarchy of multipolar condensates in which kinetics of the $n$th multipole induces condensation of the $(n+1)$th [2306.14214].

Synthetic gauge-field constructions sharpen the electrodynamic aspect of this picture. In a two-component ultracold-atom chain with spin-dependent tilt,
\[
\hat D_m^\dagger=\hat b_{m,\uparrow}^\dagger \hat b_{m,\downarrow},
\qquad
E_{xw}\simeq \Delta_\uparrow-\Delta_\downarrow\equiv\Delta,
\]
so that the dipole Hamiltonian takes the form
\[
\hat K_E=-t_2\sum_m\left[e^{-i\Delta t/\hbar}\hat D_m^\dagger\hat D_{m+1}+{\rm h.c.}\right].
\]
The resulting responses include an atomic analog of perfect Coulomb drag, in which increasing intercomponent interactions leads to equal and opposite displacements of the centers of mass of the two spin components, and a dipolar Josephson relation
\[
J_D=2t_2\rho_s^D\sin(\phi_D)
\]
after a short pulse of the rank-2 electric field imprints a dipole phase twist [2509.14492].

These lattice results do not replace the dipolar-BEC meaning of dipole condensate. Rather, they delineate a second, conceptually independent branch of the subject in which the condensing object is a dipole operator and the central observables are two-body coherence, dipole transport, and multipolar Josephson effects.

Source: https://www.emergentmind.com/topics/dipole-condensates