The conventional giant oscillator strength model of the two-photon (\ensuremath{\gamma}) generation of an excitonic molecule (m) attributes this process to the x-\ensuremath{\gamma} optical conversion (where x is the exciton) i.e, to the scheme \ensuremath{\gamma}+\ensuremath{\gamma}\ensuremath{\rightarrow}x+\ensuremath{\gamma}\ensuremath{\rightarrow}m. Recenty [A.L. Ivanov and H. Haug, Phys. Rev. B 48, 1490 (1993)], it was argued that a more adequate description can be done within the bipolariton model which follows the other scenario: \ensuremath{\gamma}+\ensuremath{\gamma}\ensuremath{\rightarrow}x+x\ensuremath{\rightarrow}m, where the Coulombic resonant coupling of the x components of the two interacting polaritons gives rise to the m formation. In the present work, we develop and analyze for the deuteron potential of an x-x interaction an exactly solvable bipolariton model of the two-\ensuremath{\gamma} m generation. This model treats an m optical creation in terms of the polariton-polariton resonant scattering and includes both the x-y polariton coupling and the x-x attraction beyond a low-order perturbation theory. The bipolariton model and the giant oscillator strength model give different descriptions of the third-order m nonlinear optical susceptibility ${\mathrm{\ensuremath{\chi}}}^{(3)}$ and of the two-\ensuremath{\gamma} m absorption. With the high-precision measurements in CuCl of the m radiative width ${\mathrm{\ensuremath{\Gamma}}}^{\mathit{m}}$, ${\mathrm{\ensuremath{\chi}}}^{(3)}$, and the two-\ensuremath{\gamma} m absorption coefficient ${\mathit{K}}^{(2)}$, we make a systematic comparison between the experiments and the two models, which allows us to unambiguously discriminate both models in favor of the bipolariton one.
Read full abstract