Unravelling the Effects of Grain Boundary and Chemical Doping on Electron–Hole Recombination in CH3NH3PbI3 Perovskite by Time-Domain Atomistic Simulation

Advancing organohalide perovskite solar cells requires understanding of carrier dynamics. Electron-hole recombination is a particularly important process because it constitutes a major pathway of energy and current losses. Grain boundaries (GBs) are common in methylammonium lead iodine CH3NH3PbI3 (MAPbI3) perovskite polycrystalline films. First-principles calculations have suggested that GBs have little effect on the recombination; however, experiments defy this prediction. Using nonadiabatic (NA) molecular dynamics combined with time-domain density functional theory, we show that GBs notably accelerate the electron-hole recombination in MAPbI3. First, GBs enhance the electron-phonon NA coupling by localizing and contributing to the electron and hole wave functions and by creating additional phonon modes that couple to the electronic degrees of freedom. Second, GBs decrease the MAPbI3 bandgap, reducing the number of vibrational quanta needed to accommodate the electronic energy loss. Third, the phonon-induced loss of electronic coherence remains largely unchanged and not accelerated, as one may expect from increased electron-phonon coupling. Further, replacing iodines by chlorines at GBs reduces the electron-hole recombination. By pushing the highest occupied molecular orbital (HOMO) density away from the boundary, chlorines restore the NA coupling close to the value observed in pristine MAPbI3. By introducing higher-frequency phonons and increasing fluctuation of the electronic gap, chlorines shorten electronic coherence. Both factors compete successfully with the reduced bandgap relative to pristine MAPbI3 and favor long excited-state lifetimes. The simulations show excellent agreement with experiment and characterize how GBs and chlorine dopants affect electron-hole recombination in perovskite solar cells. The simulations suggest a route to increased photon-to-electron conversion efficiencies through rational GB passivation.