Fig. 5: Coupled oscillator model of polariton electroabsorption.
Polariton dispersion relations for the a negatively- and b positively-detuned 40 wt% SQ:NPB microcavities. The solid markers are determined from fitting the experimental reflectivity minima in each case (error bars reflect the standard error in fitted peak position) and the solid lines are calculated using the coupled oscillator Hamiltonian given in Eq. (1) from the text. The green dashed lines denote the bare cavity mode and SQ exciton energies, while the gray dashed lines reflect the energy of the proposed lower H-aggregate state. Note that there is ostensibly a third, H-aggregate-like branch in the polariton dispersion of each cavity (not shown) that is predicted by Eq. (1) and closely coincides with the lower H-aggregate state (i.e. the gray dashed line); this branch is not observed experimentally because its amplitude is very small (owing to weak coupling of the H-aggregate state to the cavity) and thus is dominated by the much stronger LP feature. c Field-induced LP and UP energy shifts derived for the negatively-detuned cavity based on the reflectivity derivative scaling analysis from Fig. 4a; error bars reflect the uncertainty in the energy shift required to match each experimental EA lineshape. The solid lines show the predicted shifts from Eq. (1) assuming that the second state in the model (\({E}_{2}\)) corresponds to an upper-excited state of SQ located at 3 eV as sketched in the inset. d Corresponding results for the positively-detuned cavity; all of the modeling results assume an applied field strength of 105 V cm−1. The same experimental data are shown in e, f, but in this case, the second state in the model is assumed to correspond to the lower H-aggregate state at 1.76 eV as shown in the inset of e. This model explains the growing Stark shift of the LP as it approaches the H-aggregate in the negatively-detuned cavity (a, e), and the bump in LP shift when it crosses the H-aggregate in the positively-detuned cavity (b, f). All of the parameters used for the two modeling scenarios (c, d versus e, f) are summarized in the Supplementary Information.
