Publication date: 3rd July 2020
Strong interaction of charge carriers with a soft and polar lattice challenges our understanding and exploitation of charge-carrier dynamics in lead halide perovskites. For example, surprisingly slow electron cooling in metal-halide perovskites were reported recently [1] and may enable efficient hot-electron charge extraction [2], hot-electron plasmonics and catalysis [3], or carrier multiplication. Colloidal perovskite nanocrystals add yet additional knobs of control to leverage high efficiencies: their facile tuning of composition, size, and shape, and a synthetically very accessible surface all may be used to optimize charge-carrier lifetimes. Here, we combine state-of-the-art ultrafast photoluminescence and absorption spectroscopy and nonadiabatic molecular dynamics simulations to investigate charge-carrier cooling in CsPbBr3 nanocrystals over a very broad size regime, from 0.8 nm to 12 nm. Contrary to the prevailing notion that polaron formation slows down charge-carrier cooling in lead-halide perovskites [1], no suppression of carrier cooling is observed in CsPbBr3 nanocrystals except for a slow cooling of electrons in the vicinity (within ~ 0.1 eV) of the conduction band edge. Instead, we suggest that the observed cooling process may be rationalized by fast phonon-mediated intra-band transitions driven by strong and size-dependent electron-phonon coupling. Our ab initio simulations allow a direct comparison to a variety of time-resolved spectroscopies, intuitively explains the persistent ‘warm’ electrons, and yields the spectrum of phonons modulating the excited electron and hole states. The presented experimental and computational methods may easily be extended to a wider range of material systems and thus guide the development of devices utilizing hot charge carriers.
I.I. acknowledges The Netherlands Organization of Scientific Research (NWO) for financial support through the Innovational Research Incentive (Vidi) Scheme (723.013.002), and S.C.B. acknowledges financial support through the Innovational Research Incentives (Veni) Scheme (722.017.011). The computational work was carried out on the Dutch national e-infrastructure with the support of the SURF Cooperative. P.G. acknowledges BOF UGent for support. J.M.H. and K.C. acknowledge support from the Marsden Fund.