Polaron-to-Molecule Transition in Quantum Gases

• Polaron-to-molecule transition is a crossover from a fermionic quasiparticle dressed by particle-hole excitations to a dimer-like bound state, dependent on interaction strength and dimensionality.
• The transition in 3D is marked by a first-order branch crossing with decay rates scaling as Δω^(9/2), while in 1D it exhibits a smooth crossover validated by exact Bethe ansatz solutions.
• Momentum-resolved studies reveal dual minima at Q=0 and Q=kF, linking experimental observations in confined systems with theoretical predictions across varying Fermi gas settings.

The polaron-to-molecule transition is the interaction-driven change in the lowest-energy impurity state of a strongly imbalanced quantum gas, most commonly a single spin-$\downarrow$ atom immersed in a Fermi sea of spin-$\uparrow$ fermions. At weak attraction, the impurity is a polaron: a fermionic quasiparticle dressed by particle-hole excitations of the medium. At strong attraction, it becomes a molecule: a dimer-like bound state of the impurity and one majority fermion, itself dressed by the surrounding medium. In three and two dimensions, the zero-temperature single-impurity problem is commonly described by a branch crossing between these two many-body states, whereas in one dimension exact Bethe-ansatz results show a smooth crossover rather than a sharp transition (Bruun et al., 2010, Vlietinck et al., 2014, Mao et al., 2016).

1. Microscopic formulation of the impurity problem

The canonical setting is a maximally imbalanced Fermi gas with one minority atom of mass $m_\downarrow$ interacting via a short-range $s$-wave attraction with a majority Fermi sea of mass $m_\uparrow$. In three dimensions, the majority density and Fermi energy are

$$n_\uparrow=\frac{k_F^3}{6\pi^2},\qquad \mu_\uparrow=\epsilon_F=\frac{k_F^2}{2m_\uparrow},$$

and the interaction is parametrized by the scattering length $a$. Identical-spin interactions are neglected in the standard impurity limit (Bruun et al., 2010).

The impurity branch is defined from the pole of the impurity Green’s function. In the 3D formulation of Bruun and Massignan, the impurity propagator is

$$G_\downarrow(\mathbf p,z)^{-1}=G_\downarrow^0(\mathbf p,z)^{-1}-\Sigma_P(\mathbf p,z),$$

with the self-energy expanded in the number of holes in the Fermi sea,

$\Sigma_P=\Sigma_P^{(1)}+\Sigma_P^{(2)}+\ldots.$

The one-hole term $\Sigma_P^{(1)}$ is the ladder approximation and gives an accurate polaron energy, but the metastable decay rate appears only once two-hole processes are included (Bruun et al., 2010).

The molecular branch is defined from the pole of a dimer propagator. In the same framework, the molecule is represented by an open-channel Feshbach dimer operator and a propagator $\uparrow$0, whose pole yields a residue $\uparrow$1 and effective mass $\uparrow$2. In two dimensions, the same distinction is expressed in terms of the poles of the impurity propagator $\uparrow$3 and the dressed interaction $\uparrow$4: “From the poles of $\uparrow$5 and $\uparrow$6 we can extract the polaron and the molecule energy, respectively” (Vlietinck et al., 2014).

In 2D, the natural interaction parameter is the two-body binding energy $\uparrow$7, or equivalently

$\uparrow$8

since a bound state exists for any attraction. In 1D, the exact Yang–Gaudin solution organizes the impurity physics through the decomposition

$\uparrow$9

which remains valid throughout the attractive regime and provides an exact interpolation from a weakly dressed impurity to a tightly bound dimer (Vlietinck et al., 2014, Mao et al., 2016).

2. Zero-temperature branch crossing and the 3D first-order picture

In three dimensions with equal masses, earlier work placed the zero-temperature crossing near

$m_\downarrow$0

and later functional-renormalization-group and variational calculations gave $m_\downarrow$1 and $m_\downarrow$2, respectively (Bruun et al., 2010, Schmidt et al., 2011, Peng et al., 2021). Defining

$m_\downarrow$3

the polaron is the ground state for $m_\downarrow$4, while the molecule is the ground state for $m_\downarrow$5 (Bruun et al., 2010).

The distinctive 3D result of Bruun and Massignan is that ordinary two-body conversion between the two branches is forbidden by energy and momentum conservation in the presence of the filled Fermi sea. Near the crossing, a zero-momentum polaron cannot simply convert into a molecule plus one hole. The leading allowed decay must instead create an additional particle-hole pair, so both $m_\downarrow$6 and $m_\downarrow$7 are effectively three-body rearrangements of the Fermi sea (Bruun et al., 2010).

This kinematic restriction produces an anomalous threshold law. For the metastable polaron on the molecular side,

$m_\downarrow$8

and near the transition the available phase space scales as $m_\downarrow$9. Because the two-hole final state must be antisymmetrized, the matrix element vanishes at threshold and contributes one additional power of $s$0. The result is

$s$1

This is much more strongly suppressed than ordinary Fermi-liquid damping, $s$2, and implies that the matrix element between equal-energy polaron and molecule states vanishes at the transition (Bruun et al., 2010).

The same conclusion was reproduced in a unified FRG treatment with full self-energy feedback, which calculated the attractive and repulsive polaron branches together with the molecule spectral function and explicitly confirmed the scaling

$s$3

for the excited molecule close to threshold (Schmidt et al., 2011). In this standard 3D interpretation, the polaron-to-molecule transition is first order because there is no avoided crossing: the branches cross, their identities remain distinct, and the metastable excitation on either side is anomalously long lived. Bruun and Massignan estimated decay times in the range $s$4, slow enough to be experimentally accessible (Bruun et al., 2010).

3. Momentum-space structure and competing interpretations

A major reinterpretation of the transition was developed in momentum-resolved variational approaches. Instead of comparing a zero-momentum polaron to a formally separate zero-momentum molecule, these works scan the full impurity dispersion $s$5 and identify two distinct minima: one at $s$6 and another at $s$7. In this picture, the conventional molecular ansatz is a restricted subset of a finite-momentum impurity sector, and the transition is the crossing of the $s$8 and $s$9 minima (Cui, 2020, Peng et al., 2021).

Within the one-particle-hole variational framework of Li and Das Sarma, the $m_\uparrow$0 branch appears around

$m_\uparrow$1

becomes degenerate with the $m_\uparrow$2 branch at

$m_\uparrow$3

and remains locally stable up to about

$m_\uparrow$4

This yields a coexistence window

$m_\uparrow$5

in which both branches are local minima of the dispersion (Cui, 2020). The later V-2ph calculation of Wang et al. sharpened this picture, finding in 3D a double-minimum structure roughly for

$m_\uparrow$6

and a crossing at

$m_\uparrow$7

close to Monte Carlo and FRG benchmarks (Peng et al., 2021).

In the strong-coupling limit of this momentum-sector interpretation, the $m_\uparrow$8 branch becomes molecule-like, and the ground-state manifold acquires an $m_\uparrow$9 degeneracy in 3D or an $$n_\uparrow=\frac{k_F^3}{6\pi^2},\qquad \mu_\uparrow=\epsilon_F=\frac{k_F^2}{2m_\uparrow},$$0 degeneracy in 2D because any direction on the Fermi surface can be chosen (Peng et al., 2021). A recent SOC-based proposal makes this momentum distinction operational: small SOC momentum probes the polaron near $$n_\uparrow=\frac{k_F^3}{6\pi^2},\qquad \mu_\uparrow=\epsilon_F=\frac{k_F^2}{2m_\uparrow},$$1, while SOC momentum $$n_\uparrow=\frac{k_F^3}{6\pi^2},\qquad \mu_\uparrow=\epsilon_F=\frac{k_F^2}{2m_\uparrow},$$2 adiabatically prepares the molecule near $$n_\uparrow=\frac{k_F^3}{6\pi^2},\qquad \mu_\uparrow=\epsilon_F=\frac{k_F^2}{2m_\uparrow},$$3, thereby visualizing the momentum difference that underlies the first-order transition (Shi et al., 30 Dec 2025).

This momentum-sector view is not the only interpretation in the literature. Edwards argued that the standard “polaron” and “molecule” variational families are nested, with

$$n_\uparrow=\frac{k_F^3}{6\pi^2},\qquad \mu_\uparrow=\epsilon_F=\frac{k_F^2}{2m_\uparrow},$$4

and therefore should not be regarded as distinct phases separated by a true sharp transition. In that account, the impurity evolves continuously from weakly polaronic to strongly molecular character within a single correlated branch (Edwards, 2013). The modern literature therefore contains both a branch-crossing/first-order interpretation and a nested-variational/crossover interpretation, with the former dominating recent momentum-resolved and decay-rate analyses.

4. Dimensionality, confinement, and geometry

Dimensionality decisively changes the character of the phenomenon. In two dimensions, the existence of a two-body bound state for any attraction does not eliminate the many-body competition between a dressed impurity and a dressed dimer. Diagrammatic Monte Carlo found a clear ground-state crossing at

$$n_\uparrow=\frac{k_F^3}{6\pi^2},\qquad \mu_\uparrow=\epsilon_F=\frac{k_F^2}{2m_\uparrow},$$5

in agreement with a variational estimate $$n_\uparrow=\frac{k_F^3}{6\pi^2},\qquad \mu_\uparrow=\epsilon_F=\frac{k_F^2}{2m_\uparrow},$$6 and experiment $$n_\uparrow=\frac{k_F^3}{6\pi^2},\qquad \mu_\uparrow=\epsilon_F=\frac{k_F^2}{2m_\uparrow},$$7, while the full DiagMC series showed that the polaron $$n_\uparrow=\frac{k_F^3}{6\pi^2},\qquad \mu_\uparrow=\epsilon_F=\frac{k_F^2}{2m_\uparrow},$$8 factor almost coincides with the one-particle-hole result over the full interaction range (Vlietinck et al., 2014). Earlier 2D variational work had missed the transition when the molecule was left undressed, but allowing one particle-hole excitation in the molecular ansatz restored a genuine polaron-to-molecule transition (Parish, 2011).

In one dimension, the situation is different. The exact Bethe-ansatz solution shows that the impurity changes continuously from a weakly dressed polaron-like excitation to a tightly bound dimer as attraction is increased, with no nonanalytic jump in the spectrum. The binding energy evolves smoothly, and the effective mass increases continuously from $$n_\uparrow=\frac{k_F^3}{6\pi^2},\qquad \mu_\uparrow=\epsilon_F=\frac{k_F^2}{2m_\uparrow},$$9 toward $a$0, reflecting crossover rather than a sharp transition (Mao et al., 2016). The unified V-2ph analysis reaches the same conclusion: in 1D the ground state always remains in the $a$1 sector and there is no $a$2 level crossing (Peng et al., 2021).

Quasi-2D confinement introduces an additional scale $a$3 and destroys strict universality. In that setting, the transition exists but depends on density, confinement, and mass ratio. For the equal-mass $a$4 case with

$a$5

the calculated transition occurs at

$a$6

and the transition curves do not saturate even at $a$7, indicating persistent non-universality (Shi et al., 2021).

Setting	Representative control parameter	Character
3D single impurity	$a$8	First-order branch crossing
2D single impurity	$a$9	Ground-state crossing
1D single impurity	$$G_\downarrow(\mathbf p,z)^{-1}=G_\downarrow^0(\mathbf p,z)^{-1}-\Sigma_P(\mathbf p,z),$$0, exact BA	Smooth crossover
quasi-2D	$$G_\downarrow(\mathbf p,z)^{-1}=G_\downarrow^0(\mathbf p,z)^{-1}-\Sigma_P(\mathbf p,z),$$1 and $$G_\downarrow(\mathbf p,z)^{-1}=G_\downarrow^0(\mathbf p,z)^{-1}-\Sigma_P(\mathbf p,z),$$2	Non-universal transition

These dimension-dependent results imply that the label “polaron-to-molecule transition” covers more than one mechanism: a first-order branch crossing in 3D and 2D, a smooth crossover in 1D, and a confinement-dependent transition in quasi-2D (Vlietinck et al., 2014, Mao et al., 2016, Shi et al., 2021).

5. Spectral, thermodynamic, and dynamical signatures

The transition is observed not only in ground-state energetics but also in spectral functions, quasiparticle residues, linewidths, and thermodynamic response. In FRG, the impurity spectral function contains both attractive and repulsive polarons, while the molecular propagator yields a very small residue characteristic of a composite particle: $$G_\downarrow(\mathbf p,z)^{-1}=G_\downarrow^0(\mathbf p,z)^{-1}-\Sigma_P(\mathbf p,z),$$3 at unitarity. Across the transition, fermionic spectral weight shifts from the attractive to the repulsive polaron branch, whereas the molecule remains a low-residue bosonic excitation embedded in a broader continuum (Schmidt et al., 2011).

At finite temperature, the sharp zero-temperature transition becomes a crossover because thermodynamics samples thermally occupied excited states in the impurity spectral function. A variational finite-$$G_\downarrow(\mathbf p,z)^{-1}=G_\downarrow^0(\mathbf p,z)^{-1}-\Sigma_P(\mathbf p,z),$$4 analysis expressed the impurity free energy through the full spectral decomposition and found that the zero-$$G_\downarrow(\mathbf p,z)^{-1}=G_\downarrow^0(\mathbf p,z)^{-1}-\Sigma_P(\mathbf p,z),$$5 kink is smoothed already for roughly

$$G_\downarrow(\mathbf p,z)^{-1}=G_\downarrow^0(\mathbf p,z)^{-1}-\Sigma_P(\mathbf p,z),$$6

The same work argued that the Tan contact remains a useful remnant of the zero-temperature transition: on the polaron side $$G_\downarrow(\mathbf p,z)^{-1}=G_\downarrow^0(\mathbf p,z)^{-1}-\Sigma_P(\mathbf p,z),$$7 is non-monotonic because thermal occupation of the molecule-hole continuum increases the contact, while on the molecule side $$G_\downarrow(\mathbf p,z)^{-1}=G_\downarrow^0(\mathbf p,z)^{-1}-\Sigma_P(\mathbf p,z),$$8 decreases monotonically with $$G_\downarrow(\mathbf p,z)^{-1}=G_\downarrow^0(\mathbf p,z)^{-1}-\Sigma_P(\mathbf p,z),$$9 (Parish et al., 2020).

This finite-$\Sigma_P=\Sigma_P^{(1)}+\Sigma_P^{(2)}+\ldots.$0, finite-density smoothing has been observed directly. In a degenerate 3D $\Sigma_P=\Sigma_P^{(1)}+\Sigma_P^{(2)}+\ldots.$1 gas, Raman spectroscopy with large momentum transfer isolated the coherent quasiparticle peak and extracted the polaron energy, many-body spectral weight $\Sigma_P=\Sigma_P^{(1)}+\Sigma_P^{(2)}+\ldots.$2, and contact. As attraction increased, all three observables evolved continuously, with $\Sigma_P=\Sigma_P^{(1)}+\Sigma_P^{(2)}+\ldots.$3 decreasing smoothly toward zero rather than jumping. The data were consistent with a model in which fermionic polarons and bosonic molecules are both thermally occupied, so that each branch remains populated even where it is not the absolute ground state (Ness et al., 2020).

Finite momentum adds another route to the transition. In a $\Sigma_P=\Sigma_P^{(1)}+\Sigma_P^{(2)}+\ldots.$4-in-$\Sigma_P=\Sigma_P^{(1)}+\Sigma_P^{(2)}+\ldots.$5 experiment on the moving Fermi polaron, the attractive branch followed the usual Fermi-liquid form at low momentum,

$\Sigma_P=\Sigma_P^{(1)}+\Sigma_P^{(2)}+\ldots.$6

but at intermediate momenta its energy became non-monotonic and the linewidth increased sharply. Ladder $\Sigma_P=\Sigma_P^{(1)}+\Sigma_P^{(2)}+\ldots.$7-matrix theory identified the threshold as the entry of the attractive polaron into the molecule-hole continuum, determined by

$\Sigma_P=\Sigma_P^{(1)}+\Sigma_P^{(2)}+\ldots.$8

For a representative dataset with $\Sigma_P=\Sigma_P^{(1)}+\Sigma_P^{(2)}+\ldots.$9, the zero-temperature kink occurred near $\Sigma_P^{(1)}$0, and the authors interpreted the onset of strong broadening as a motion-induced polaron-molecule transition (Hennebichler et al., 19 Jun 2026).

6. Special realizations and broader generalizations

The transition persists, with modified structure, in a number of specialized settings. Near an orbital Feshbach resonance in $\Sigma_P^{(1)}$1, the impurity and molecule are inherently two-channel objects involving spin exchange between open and closed channels. A variational calculation found a crossing at

$\Sigma_P^{(1)}$2

for density $\Sigma_P^{(1)}$3, and a later treatment including one particle-hole excitation in the molecular state showed that the transition moves toward the BCS side as density increases, a narrow-resonance-like feature of orbital Feshbach physics (Chen et al., 2016, Xu et al., 2017).

In a 2D $\Sigma_P^{(1)}$4 Fermi superfluid, the host medium itself undergoes a topological transition at $\Sigma_P^{(1)}$5. There the impurity competes between a polaron dressed by pair-breaking excitations and a molecule formed with one Bogoliubov quasiparticle. The molecular state becomes favorable in the topologically trivial regime $\Sigma_P^{(1)}$6, and the paper concludes that pairing in the background superfluid facilitates impurity-fermion binding relative to the noninteracting Fermi-sea case (Qin et al., 2019).

The same conceptual competition reappears beyond the single-fermion-impurity problem. In a density-matched $\Sigma_P^{(1)}$7-$\Sigma_P^{(1)}$8 Bose–Fermi mixture, the condensate fraction

$\Sigma_P^{(1)}$9

acts as an order parameter for a transition from a polaronic condensate to a molecular Fermi gas. Experimentally, $\uparrow$00 vanished near $\uparrow$01, linking the impurity-limit polaron-to-molecule competition to a finite-density transition between phases of different quantum statistics (Duda et al., 2021).

Finally, recent quantum-simulation work has framed the transition in finite systems as a spectroscopic bifurcation rather than a strict thermodynamic singularity. A digital simulation on NISQ hardware observed a smooth evolution from a dressed quasiparticle regime to a molecular bound-state branch, with a spectral bifurcation and approximately linear strong-coupling energy shift $\uparrow$02; the paper explicitly described this as a finite-size crossover connected to the standard polaron-molecule transition (Catala et al., 26 Jan 2026).

Taken together, these results establish the polaron-to-molecule transition as a unifying impurity phenomenon whose detailed realization depends on dimension, momentum sector, spectral kinematics, and medium structure. In its simplest 3D single-impurity form it is commonly understood as a first-order crossing between competing many-body branches, but confinement, topology, finite temperature, finite impurity density, or exactly solvable 1D kinematics can convert that sharp transition into a smooth crossover or re-express it as a competition between distinct momentum sectors.