Light Bouncing Inside Thin Films Can Fake Molecular Signals in Perovskite Solar Cells
A Japanese physicist says a sharp infrared signal long read as molecular change can come from optics alone, and proposes a dilution test to tell the two apart.
A narrow infrared signal that researchers have read as evidence of molecular change in perovskite solar materials may often come from something simpler: light bouncing around inside the thin film. That is the argument of a new analysis by Jun Nishida, an assistant professor at the Institute for Molecular Science in Japan, published as an Editor's Pick in The Journal of Chemical Physics.
The technique at issue is time-resolved infrared spectroscopy. A visible light pulse excites the material, and an infrared pulse then probes its molecular vibrations. It is widely used to study how light-generated charges interact with molecules in solar materials. In lead halide perovskites, several studies reported vibrational absorption that rises sharply after excitation. That was puzzling, because the organic cations involved are not directly bonded to the inorganic framework that carries the charges.
Nishida proposes that much of the signal is optical rather than chemical. He calls the mechanism carrier-induced photonic resonance, or CIPR. When light creates free charges in the film, they change how infrared light reflects and transmits through the stacked layers of film and substrate. Those changes can create sharp spectral features that look like vibrational changes even when the vibrations themselves have not moved.
To test the idea, Nishida reanalyzed earlier measurements on a perovskite film using multilayer transmission theory, which accounts for multiple reflections at every interface. His model included only the broad optical response of the charges and assumed no change at all in the organic cations' vibrations. It still reproduced most of the narrow, blue-shifted signal seen in experiments, including its position and amplitude.
Depending on the spectral character of the charge response, the effect can imitate increased absorption by excited molecules, a shift in a vibrational peak, or the bleaching of an existing absorption band. The model also predicts that an adjacent silica layer, which is not excited, can produce extra resonances that could be mistaken for charge trapping or charge transfer into that layer. Nishida concludes that the spectral shape of an infrared signal alone cannot confirm a molecular change, and that light propagation through the entire stack must be considered.
He suggests a control experiment: measure perovskite nanocrystals dispersed in a dilute solution instead of a densely packed film. In his calculations CIPR weakens at lower concentrations, while the longer optical path keeps the genuine molecular signal strong. The test has not yet been run.
The picture is less clear at the nanoscale. Nishida also revisited ultrafast infrared nanospectroscopy, which uses a sharp metallic tip to reach resolution below 100 nanometers. There, the predicted optical contribution depended heavily on how the tip and sample interaction was modeled. In the models considered most realistic, the optical effect explained only a minor part of the observed resonance, which suggests a real molecular shift may also contribute. One possible cause is light-induced expansion of the perovskite lattice. The gap between the conventional and nanoscale results still needs work.
The paper, "Photonic effects in visible-pump infrared-probe vibrational spectroscopies on photovoltaic thin films," appears in volume 165, issue 12 (DOI 10.1063/5.0350300). The author argues it gives researchers a firmer basis for reading infrared measurements of perovskites and other photovoltaic thin films.
