Biexciton Feshbach Resonance

• Biexciton Feshbach resonance is a two-channel scattering phenomenon where exciton pairs couple with a closed biexciton state to dynamically tune interactions.
• It is realized across semiconductor platforms by coupling open channels of exciton-polaritons to biexciton states, enabling control via detuning and anisotropy.
• This resonance mechanism enhances nonlinear optical effects, facilitates polaritonic bound states, and paves the way for tunable quantum optical applications.

Biexciton Feshbach resonance is a two-channel resonant-scattering phenomenon in which an open continuum of two excitonic bosons—most commonly two excitons or two exciton-polaritons—couples to a closed-channel biexciton state. When the two-particle energy is tuned through the biexciton energy, the effective interaction is strongly renormalized: the real part acquires a dispersive sign-changing contribution, while the imaginary part is enhanced by resonant conversion into the molecular channel and its decay. In semiconductor optics, this structure appears in several distinct settings, including lower-polariton pairs coupled to a biexciton in III–V microcavities, paraexciton scattering through an orthoexciton-pair biexciton channel in Cu$_2$O, anisotropy-enabled resonances of excitons or polaritons in two-dimensional semiconductors, and heavy-polariton systems where genuine bipolariton bound states become prominent (Takemura et al., 2014, Cam, 2018, Andreev, 2023, Vermilyea et al., 2023).

1. Two-channel structure and defining criteria

A biexciton Feshbach resonance requires three ingredients. First, there must be an open channel, consisting of two scattering bosons at the relevant collision energy. Second, there must be a closed channel, namely a biexciton or biexciton-like molecular level. Third, a finite coupling must convert a pair in the open channel into the closed-channel molecule and back. In semiconductor microcavities, the canonical realization is the opposite-spin lower-polariton pair coupled to a biexciton; in Cu$_2$O, two paraexcitons couple through exchange to a two-orthoexciton channel that supports a biexciton; in strained 2D semiconductors, the open channel can be a pair of spin-polarized excitons or polaritons, while in-plane anisotropy supplies the spin-flip coupling to the biexciton singlet (Takemura et al., 2017, Cam, 2018, Andreev, 2023).

The closed-channel object need not be a perfectly stable bound state. In Cu$_2$O, the biexciton is explicitly a quasibound, decaying Feshbach resonance, with complex energy and intrinsic width, so the resonance simultaneously modifies elastic scattering and opens inelastic loss channels (Cam, 2018). In polaritonic realizations, the biexciton linewidth likewise prevents a literal divergence of the interaction and instead produces a smooth dispersive renormalization reminiscent of optical Feshbach resonances (Takemura et al., 2014).

The resonance is inherently channel-selective. In the standard microcavity case, the biexciton is formed from two excitons with anti-parallel spins, so the resonance appears in the $\uparrow\downarrow$ or $\sigma^+\sigma^-$ channel, not in the same-spin channel. This selectivity underlies the experimentally observed distinction between the same-spin interaction constant $\alpha_1$ and the opposite-spin constant $\alpha_2$ in lower-polariton pump–probe spectroscopy (Takemura et al., 2017).

2. Microscopic realizations across excitonic platforms

The biexciton Feshbach mechanism is realized or formalized in several materially distinct systems. In all cases, the resonance is controlled by the relative placement of a two-particle continuum and a biexciton-like level, but the microscopic origin of the coupling and the external tuning knob differ substantially (Takemura et al., 2014, Cam, 2018, Navadeh-Toupchi et al., 2018, Andreev, 2023, Vermilyea et al., 2023).

Platform	Open channel	Closed channel / tuning
III–V microcavity polaritons	Two lower polaritons with opposite spins	Quantum-well biexciton; tuned by cavity–exciton detuning
Cross-polariton resonance	One lower and one upper polariton with opposite spins	Biexciton; tuned so $E_{LP}+E_{UP}\approx E_B$
Cu$_2$O paraexcitons	Two paraexcitons	Two-orthoexciton biexciton channel; tuned by stress through $\Delta(S)$
Strained 2D semiconductors	Two spin-polarized excitons or polaritons	Biexciton singlet; tuned by transverse magnetic field and anisotropy
Heavy exciton-polaritons	Two lower heavy polaritons	Bipolariton / biexciton-like bound state; tuned by detuning $_2$0

In the original polaritonic realization, lower-branch exciton-polaritons in a planar microcavity provide the open channel, while a biexciton in the quantum well is the closed-channel molecule. The relevant tuning parameter is the cavity–exciton detuning $_2$1, which shifts the lower-polariton energy $_2$2 and therefore the two-polariton energy $_2$3 relative to the biexciton energy $_2$4 (Takemura et al., 2014).

A distinct but closely related realization is the cross Feshbach resonance, where the open channel is a lower-polariton plus an upper-polariton pair with opposite spins. There the resonance condition is $_2$5, and the biexciton binding energy extracted from the experiment is $_2$6 (Navadeh-Toupchi et al., 2018).

In Cu$_2$7O, the microscopic structure differs fundamentally. Two paraexcitons do not scatter in an isolated single channel: due to constituent exchange, a paraexciton pair couples to an orthoexciton pair with total spin $_2$8. The para–para sector is the open channel, the ortho–ortho sector is the closed channel, and the bare closed channel supports one biexciton bound state with $_2$9. The channel coupling produces a strongly coupled resonance with $_2$0 (Cam, 2018).

In strained 2D semiconductors, the closed channel is again the biexciton singlet, but the coupling arises from anisotropy-induced splitting of the radiative exciton doublet. The single-boson Hamiltonian contains an effective in-plane field $_2$1, and the corresponding spin-flip matrix element $_2$2 couples a spin-polarized two-boson continuum to the biexciton. The magnetic field controls the detuning, while strain controls the width $_2$3 through $_2$4 (Andreev, 2023).

The bilayer-semiconductor theory of electrically tunable exciton–electron Feshbach resonances does not treat biexcitons explicitly, but its two-channel 2D formalism is directly mappable: replacing the exciton–electron molecule by a biexciton and the electron tunneling by excitonic hybridization yields a biexciton Feshbach framework with the same threshold structure, T-matrix equations, and tunable 2D scattering parameters. This suggests that electrically biased bilayers furnish a generic 2D template for biexciton resonances even when the original calculation is performed in a trionic channel (Kuhlenkamp et al., 2021).

3. Scattering theory, detuning, and effective interactions

The defining theoretical object is the two-body scattering T-matrix. In the lower-polariton–biexciton problem, the experimentally relevant consequence is a complex renormalization of the opposite-spin interaction constant. For resonantly pumped lower polaritons, the mean-field two-channel model gives

$_2$5

and

$_2$6

where $_2$7 is the background opposite-spin interaction, $_2$8 is the LP-pair–biexciton coupling, and $_2$9 is the biexciton decay rate. The real part $\uparrow\downarrow$0 is dispersive and changes sign across resonance; the imaginary part $\uparrow\downarrow$1 peaks at resonance and governs dissipative nonlinearity (Takemura et al., 2017).

A closely related description appears in the earlier polaritonic experiment through the effective interaction

$\uparrow\downarrow$2

with $\uparrow\downarrow$3. This is the standard Feshbach form: the real part controls the interaction-induced redshift or blueshift, while the imaginary part describes two-body loss into the biexciton (Takemura et al., 2014).

In Cu$\uparrow\downarrow$4O, the same logic appears in scattering language. The paraexciton scattering length is

$\uparrow\downarrow$5

with a positive background value

$\uparrow\downarrow$6

and a resonant correction that is attractive when the biexciton lies below threshold. Under stress, the ortho–para splitting $\uparrow\downarrow$7 is reduced, the detuning decreases, and the total scattering length crosses zero at a critical stress $\uparrow\downarrow$8 (order of magnitude), becoming negative for $\uparrow\downarrow$9 (Cam, 2018).

The 2D character of these systems is not incidental. In the bilayer two-channel formalism, the low-energy on-shell T-matrix obeys

$\sigma^+\sigma^-$0

so the resonance is encoded not only in a tunable scattering length but also in an effective range $\sigma^+\sigma^-$1. In the mapped biexciton problem, the same structure would govern open-channel exciton–exciton scattering, with electric field or hybridization tuning both $\sigma^+\sigma^-$2 and $\sigma^+\sigma^-$3. This suggests that narrow-resonance behavior and large effective-range corrections are generic in 2D semiconductor realizations (Kuhlenkamp et al., 2021).

For strained 2D excitons or polaritons, the resonant part of the scattering amplitude is written explicitly as

$\sigma^+\sigma^-$4

and in the broad-resonance limit the effective interaction becomes

$\sigma^+\sigma^-$5

The resonance is therefore simultaneously controlled by detuning $\sigma^+\sigma^-$6 and by the width parameter $\sigma^+\sigma^-$7, which is itself tunable through strain (Andreev, 2023).

4. Spectroscopic and optical manifestations

The most direct experimental signatures are optical. In the lower-polariton experiment, the biexciton lies $\sigma^+\sigma^-$8 below the exciton, the biexciton linewidth is $\sigma^+\sigma^-$9, and the Feshbach resonance is observed at cavity detuning $\alpha_1$0, where the two-lower-polariton energy crosses the biexciton. At low density, the transmitted probe shows a dispersive energy shift that changes from redshift to blueshift, together with a pronounced amplitude dip centered near resonance; at higher density, an anticrossing appears, signaling a strong-coupling Feshbach regime in which $\alpha_1$1 (Takemura et al., 2014).

The cross-polariton experiment exhibits the same phenomenology in an inter-branch channel. The cross Feshbach resonance occurs at a negative cavity detuning of about $\alpha_1$2, where the sum of one lower and one upper polariton equals the biexciton energy. The key observations are a prompt change of sign of the upper-polariton energy shift and a maximum reduction of probe transmission at the same detuning, interpreted as resonant conversion of the lower-plus-upper pair into a biexciton (Navadeh-Toupchi et al., 2018).

A later many-body T-matrix analysis unified the lower–lower and lower–upper resonances within a single impurity-in-BEC framework. Using the same biexciton parameters for both datasets, it obtained $\alpha_1$3 and $\alpha_1$4, and showed that the observed dispersive shifts are the broadened remnants of avoided crossings that would be sharper for a less lossy biexciton. The same work also found that the computed radiative decay into two polaritons is only $\alpha_1$5, far smaller than the $\alpha_1$6 linewidth needed to fit experiment, implying important additional decay channels (Bastarrachea-Magnani et al., 2019).

In Cu$\alpha_1$7O, the spectroscopic emphasis is different because the central issue is collision-induced paraexciton loss rather than pump–probe line shifts. The inelastic cross section

$\alpha_1$8

yields a two-body loss rate

$\alpha_1$9

which is proportional to the quasibiexciton decay width. Using the fitted paraexciton linewidth

$\alpha_2$0

the estimated loss rate at $\alpha_2$1 is $\alpha_2$2 and $\alpha_2$3. The same resonance thereby explains both collisional loss and the stress-dependent renormalization of effective interactions (Cam, 2018).

In heavy-polariton systems, the predicted observables are absorption and luminescence spectra containing a gapped bipolariton-related mode in addition to the gapless Goldstone branch. The line associated with the bound pair moves through the chemical potential across the polariton-superfluid–bipolariton-superfluid boundary, furnishing a many-body spectroscopic analog of the two-body biexciton crossing (Vermilyea et al., 2023).

5. Many-body regimes and phase structure

Once a biexciton resonance is embedded in a finite-density medium, the problem ceases to be purely two-body. In polariton mixtures, the minority spin component behaves as a mobile impurity interacting with a Bose-condensed majority component, and the biexciton appears as a pole in the ladder T-matrix. The impurity self-energy

$\alpha_2$4

produces resonant energy shifts of upper or lower polaritons, depending on which branch the impurity occupies. This is a Bose-polaron formulation of biexciton Feshbach physics (Bastarrachea-Magnani et al., 2019).

In Cu$\alpha_2$5O, the central many-body implication concerns paraexciton Bose–Einstein condensation. Elastic collisions dominate over biexciton-mediated loss for sub-kelvin temperatures and moderate stress, with $\alpha_2$6 near $\alpha_2$7 and $\alpha_2$8 far from $\alpha_2$9. However, once the total scattering length becomes negative, the condensate is metastable. For the experimental trap with $E_{LP}+E_{UP}\approx E_B$0, the fitted value $E_{LP}+E_{UP}\approx E_B$1 at $E_{LP}+E_{UP}\approx E_B$2 implies $E_{LP}+E_{UP}\approx E_B$3, so collapse is expected for realistic condensate occupations (Cam, 2018).

In dipolar-exciton systems with a narrow opposite-spin resonance, the many-body consequences are even more extensive. The resonance can drive a transformation from a resonant exciton superfluid into a superfluid of biexcitons, and the transition may be either of the first or the second kind. Long-range exchange then broadens the biexciton resonance and shifts the gap minimum to a circle in momentum space, producing a second-order transition into a phase of counter-propagating linearly polarized excitonic condensates. The phase diagram contains radiative exciton, dark biexciton, and intermediate $E_{LP}+E_{UP}\approx E_B$4-X-XX phases (Andreev, 2020).

For heavy exciton-polaritons, the effective two-channel Hamiltonian of lower polaritons and bipolaritons yields a phase diagram with a polariton superfluid (PSF), a bipolariton superfluid (BSF), and phase-separation regions. The continuous PSF–BSF boundary is

$E_{LP}+E_{UP}\approx E_B$5

so the many-body transition is directly controlled by the same bound-state energy $E_{LP}+E_{UP}\approx E_B$6 that governs the two-body Feshbach-like resonance (Vermilyea et al., 2023).

A broader implication, explicit in the bilayer-semiconductor analysis, is that a biexciton resonance should feed back into optical spectra through a self-energy built from the open-channel T-matrix. Although that work treats an exciton–electron resonance rather than a biexciton one, its Keldysh–T-matrix structure suggests a direct route to exciton–exciton polarons or biexciton polarons in electrically tunable bilayers (Kuhlenkamp et al., 2021).

6. Interpretive issues, limitations, and technological directions

Several recurrent interpretive points are important. First, a biexciton Feshbach resonance is not synonymous with a stable biexciton line. In Cu$E_{LP}+E_{UP}\approx E_B$7O the biexciton is explicitly quasibound and decaying; in microcavities the biexciton linewidth can dominate the observable lineshape; and in broad-resonance regimes the experimentally visible effect can be a dispersive shift or dissipative nonlinearity rather than a sharp molecular peak (Cam, 2018, Takemura et al., 2014, Bastarrachea-Magnani et al., 2019).

Second, the terminology bipolariton is not uniform. In conventional photon microcavities, what is often called a bipolariton is frequently a radiatively renormalized biexciton, whereas the heavy-polariton theory reserves the term for a genuine bound state of two polaritons with appreciable hybrid character. This distinction matters because the heavy-polariton limit supports true two-polariton bound states that are exponentially suppressed in ordinary photon-polariton systems (Vermilyea et al., 2023).

Third, not every relevant formalism is written in biexciton language from the outset. The bilayer 2D theory of tunable exciton–electron Feshbach molecules does not include an explicit biexciton sector, yet its open/closed-channel Hamiltonian, threshold energies, and finite-range T-matrix expansion map directly onto a biexciton resonance after the replacement of the closed-channel trion-like state by a biexciton. This suggests that resonance engineering in bilayers, twist structures, or moiré systems can be analyzed within a common 2D Feshbach framework even when the microscopic constituents vary (Kuhlenkamp et al., 2021).

The technological consequences are substantial. In the confined-polariton setting, a biexciton-mediated resonance yields the effective Hamiltonian

$E_{LP}+E_{UP}\approx E_B$8

which underlies the proposed Feshbach blockade: a resonantly enhanced single-photon Kerr nonlinearity and interaction-induced suppression of two-polariton occupancy in a polariton dot (Carusotto et al., 2010). In inter-branch scattering, the cross Feshbach resonance provides the condition for converting a lower-plus-upper polariton pair into a biexciton and then into two lower polaritons, which was identified as a route toward momentum- and polarization-entangled photon pairs (Navadeh-Toupchi et al., 2018). In strained 2D semiconductors, the same basic mechanism enables controllable boson fusion, squeezing, and giant Feshbach dimers whose concurrence is $E_{LP}+E_{UP}\approx E_B$9, tunable by magnetic field through the biexciton-core weight (Andreev, 2023).

Taken together, these results establish biexciton Feshbach resonance as a unifying concept for resonant interaction control in excitonic matter: a two-channel mechanism that connects microscopic bound states, nonlinear optical response, loss, Bose-polaron physics, condensate stability, and bound-pair superfluidity across microcavities, bulk excitons, bilayers, and two-dimensional semiconductors.