When the photons arrive to the active perovskite layer they will be absorbed provided they have the appropriate wavelength. The wavelength of the absorbed photons is equal or smaller than that corresponds to the energy gap of the material. Also the thickness of the material must be sufficiently greater than than the penetration depth of the radiation which is =1/alpha, with alpha is the absorption coefficient of the material.
The absorbed photons generate excitons which are relatively weakly electrostatic bound. Excitons needs energy to become free electrons and holes. Because of the perovskite has relatively high dielectric constant, the excitons dissociates thermally and become readily free electron holes. This behavior is very similar to that of metallic semiconductor such as silicon.
The generated electrons will drift to the electron transport layer by the built in electric field, where they will be collected and delivered to the metallic cathode to the external circuit if there is an outer connection. The same happens for the holes where they will drift to the hole transport layer by the built electric field and delivered to the anode and then to outer circuit if a connection available.
During their collection process they last during their movement to the transport layers. If they elapse time greater than their lifetime, they will be lost by recombination. So, the thickness of the pervoskite layer must be smaller than the diffusion length of the electrons and holes for avoiding the losses by recombination in the perovskite layer. The diffusion length is the average length which the carriers move in their lifetime before they recombine.
In case of open circuit the collected electrons accumulate on the cathode while the collected holes accumulate on the anode and establish the electromotive force of the cell or the so called opencircuit voltage.
As the operating principles of the pervoskite solar cell is similar to the metallic solar cells i would like that you refer to the book chapter:Chapter Solar cells and arrays: Principles, analysis and design