Using the composite boson (coboson) many-body formalism, we determine under which conditions "trion-polariton" can exist. Dipolar attraction can bind an exciton and an electron into a trion having an energy well separated from the exciton energy. Yet, the existence of long-lived "trionpolariton" is a priori implausible not only because the photon-trion coupling, which scales as the inverse of the sample volume, is vanishingly small, but mostly because this coupling is intrinsically "weak". Here, we show that a moderately dense Fermi sea renders its observation possible: on the pro side, the Fermi sea overcomes the weak coupling by pinning the photon to its momentum through Pauli blocking; it also overcomes the dramatically poor photon-trion coupling by providing a volume-linear trion subspace to which the photon is coherently coupled. On the con side, the Fermi sea broadens the photon-trion resonance due to the fermionic nature of trions and electrons; it also weakens the trion binding by blocking electronic states relevant for trion formation. As a result, the proper way to observe this novel polariton is to use doped semiconductor having long-lived electronic states, highly-bound trion and Fermi energy as large as a fraction of the trion binding energy.
PACS numbers:Exciton-polaritons have recently attracted very much attention because of claimed observations of BoseEinstein condensation [1][2][3][4]. When the coupling between a photon and an exciton is strong, exciton-polaritons [5,6] are formed. By contrast, when this coupling is weak, photons are absorbed. In the former case, the exciton lifetime is long compared to the Rabi oscillation period and the Q exciton recombines into the same Q photon (see Fig. 1(a)). When it is short, the exciton changes momentum before recombination occurs and the initial photon Q cannot be re-emitted; it gets absorbed along the Fermi golden rule, the small exciton lifetime producing a broadening of the exciton discrete level, which plays the role of a continuum [7].We here consider another semiconductor bound state, the trion, and determine under which conditions this composite fermion can strongly couple to photon to form a polariton. Many objections lead us to first reject the idea: (i) the trion coupling to photon is intrinsically weak because the emitted photon can have a momentum different from its initial value, even when the trion lifetime is long; (ii) the photon-trion coupling is vanishingly small because it scales as the inverse of the sample volume; (iii) an increase of the number of electrons available for pairing broadens the photon-trion resonance due to the fermionic nature of trions and electrons; (iv) it also reduces the trion binding by Coulomb screening and by Pauli blocking the electronic states relevant to the formation of bound trion. Despite all these objections, we will show that there exists a narrow window in which "trion-polariton" can be formed.Semiconductor trions [8][9][10][11][12][13][14] are made of two conduction-electrons and one valence-hole, ...