Transformed Double-Capped Gold Nanorods in Dye Co-Sensitized Solar Cells for Semitransparent Windows

Abstract: Background: Dye sensitized solar cells (DSSCs) containing two different dyes were recently used for applications to windows. To enhance the efficiency of this type of solar cells by means of the effect of localized surface plasmon resonance (LSPR), we produced gold nanorods (GNRs) with an aspect ratio (a.r.) equal to 3:1 and tos 4:1. With an actual window application in mind, and mainly to prevent corrosion by the redox mediator in the cell, we considered the capping of GNRs before introducing them into the titanium oxide (TiO2) layer of the anode. In particular, we made a double-capping with silica and titania layers for a limited total thickness (i.e., about 6 nm), while still allowing a significant localized LSPR effect despite the increased distance between gold and dye molecules. We documented the different transformations in dimensions of the two types of capped gold nanorods (c-GNRs) due to the effect of sintering. Our aim was to evaluate the influence that these transformations would have on the photovoltaic performances of DSSCs. Methods: We added c-GNRs with a ratio of 2% in w/w to a transparent semiconductor paste, which was doctor bladed on the photoanodes of the co-sensitized solar cells made with commercially available organic sensitizers (L1 or L0) and the squaraine SQ2, which acted as a co-sensitizer. The films had a thickness of about 6 ?m and were sintered at 450°C. We used transmission electron microscopy (TEM) analysis to document the transformations, absorbance and absorptance spectra in order to control the effects of these modifications, and transmittance spectra for evaluating the see-through effects. We performed current-voltage, external quantum efficiency (EQE%) and electrochemical impedance spectroscopy (EIS) characterizations of the DSSCs. Results: The semiconductor films with c-GNRs that had GNRs with an a.r. equal to 4:1 (c-GNRs 4:1) had lower absorption and higher transmission as compared to those with GNRs a.r equal to 3:1 (c-GNRs 3:1). Only the c- GNRs 3:1, which retained a similar shape and an a.r. equal to 1.5 after sintering, produced an enhancement in the power conversion efficiency ?% (23%), current Jsc (8%), and voltage Voc (2.5%) when used in combination with the dye cocktail containing the organic dye L1. On the contrary, the presence of c-GNRs 4:1 negatively influenced the photovoltaic performances of the cells containing this dye cocktail. The same occurred for both types of c-GNRs with the dye cocktail containing L0. Conclusion: The use of c-GNRs 3:1 could actually improve the efficiency of co-sensitized DSSCs. On the other hand, the transformed dimensions of the c-GNRs 4:1 negatively influenced the photovoltaic characteristics when we used the same concentration of nanoparticles, and a semiconductor paste in small grains (i.e., about 20 nm). We attributed this fact both to a reduced penetration of the dyes in the films and to an inferior plasmonic effect.