The importance of dye chemistry and TiCl4 surface treatment in the behavior of Al2O3 recombination barrier layers deposited by atomic layer deposition in solid-state dye-sensitized solar cells

• Published in: Physical chemistry chemical physics : PCCP 2012-01, Vol.14 (35), p.12130-12140
• Main Authors: BRENNAN, Thomas P, BAKKE, Jonathan R, MCGEHEE, Michael D, BENT, Stacey F, DING, I-Kang, HARDIN, Brian E, NGUYEN, William H, MONDAL, Rajib, BAILIE, Colin D, MARGULIS, George Y, HOKE, Eric T, SELLINGER, Alan
• Format: Article
• Language: English
• Subjects: Chemistry; Colloidal state and disperse state; Exact sciences and technology; General and physical chemistry; Porous materials
• ISSN: 1463-9076; 1463-9084
• DOI: 10.1039/c2cp42388j

Atomic layer deposition (ALD) was used to fabricate Al(2)O(3) recombination barriers in solid-state dye-sensitized solar cells (ss-DSSCs) employing an organic hole transport material (HTM) for the first time. Al(2)O(3) recombination barriers of varying thickness were incorporated into efficient ss-DSSCs utilizing the Z907 dye adsorbed onto a 2 μm-thick nanoporous TiO(2) active layer and the HTM spiro-OMeTAD. The impact of Al(2)O(3) barriers was also studied in devices employing different dyes, with increased active layer thicknesses, and with substrates that did not undergo the TiCl(4) surface treatment. In all instances, electron lifetimes (as determined by transient photovoltage measurements) increased and dark current was suppressed after Al(2)O(3) deposition. However, only when the TiCl(4) treatment was eliminated did device efficiency increase; in all other instances efficiency decreased due to a drop in short-circuit current. These results are attributed in the former case to the similar effects of Al(2)O(3) ALD and the TiCl(4) surface treatment whereas the insulating properties of Al(2)O(3) hinder charge injection and lead to current loss in TiCl(4)-treated devices. The impact of Al(2)O(3) barrier layers was unaffected by doubling the active layer thickness or using an alternative ruthenium dye, but a metal-free donor-π-acceptor dye exhibited a much smaller decrease in current due to its higher excited state energy. We develop a model employing prior research on Al(2)O(3) growth and dye kinetics that successfully predicts the reduction in device current as a function of ALD cycles and is extendable to different dye-barrier systems.