Abstract:We investigate the polarization selection rules of sharp zero-phonon lines (ZPLs) from isolated defects in hexagonal boron nitride (h-BN) and compare our findings with the predictions of a configuration coordinate model involving two electronic states. Our survey, which spans the spectral range ~550-740 nm, reveals that, in disagreement with a two-level model, the absorption and emission dipoles are often misaligned. We relate the dipole misalignment angle ( ) to the ZPL Stokes shift ( ) and find that ≈ 0° when corresponds to an allowed h-BN phonon frequency and that 0°≤ ≤ 90° when exceeds the maximum allowed h-BN phonon frequency. Consequently, a two-level configuration coordinate model succeeds at describing excitations mediated by the creation of one optical phonon but fails at describing excitations that require the creation of multiple phonons. We propose that direct excitations requiring the creation of multiple phonons are inefficient due to the low Huang-Rhys factors in In the Frank-Condon approximation, where the fast electronic rearrangement precedes the slower lattice relaxation, the transition rate from to ( * ,is proportional to where * − * are the associated Laguerre polynomials and is a measure of defect-lattice coupling called the Huang-Rhys factor. In natural units = To investigate the polarization properties of absorption, we rotate the polarization of the exciting light and monitor the total fluorescence intensity. The result of this absorption measurement is shown as the green triangles in Fig. 2b. Fixing the polarization of the exciting light to maximize the fluorescence, we determine the polarization of the emitted photons using a polarization analyzer placed in the collection path of the microscope. The result of this emission measurement is shown as the red circles in Fig. 2b. The solid lines are best fits to the data using the equationwhere 〈 〉 in each fit is the orientation of the absorption or emission dipole spectrally averaged over the collection window. As predicted by Equation 1, we find that the maxima of absorption and emission are aligned for this defect.Additionally, we have shown previously that the temperature dependence of the ZPL intensity in h-BN is well-modeled by the temperature-dependent DebyeWaller factor [33]. Thus, we conclude that the configuration coordinate model is a good description of the observed properties for the defect shown in Fig. 2.A survey of defect ZPLs that span an appreciable energy range reveals that, in contrast to the data shown in Fig. 2 For the emission measurement we fix the polarization angle of the exciting light to ( ) and record an emission spectrum for a series of positions of the polarization analyzer in the collection path. In an analogous fashion to the absorption case we obtain ( , ) and ( ) for the emitted light. For the case of emission we apply a calibration to ( ) to correct for wavelength-and polarization-dependent retardances (see Supporting Information) introduced by the collection path of the confocal microscope. To be...