Dangling Bond Defects: The Critical Roadblock to Efficient Photoconversion in Hybrid Quantum Dot Solar Cells

Inorganic−organic hybrid materials based on silicon quantum dots (SiQDs) have been utilized for photovoltaic applications but suffer from rapid charge recombination and low carrier mobility. We present an ab initio investigation of charge dynamics to pinpoint the source of this severe problem, and our results indicate that such devices show great promise provided that dangling bond (DB) defects can be sufficiently removed. Without DBs, the predicted charge transfer (CT) rate is much higher than that of photoluminescence (PL), while the electron hopping (EH) proceeds more quickly than interfacial charge recombination (CR). In contrast, one DB in a SiQD leads to a dramatic enhancement, by 10 orders of magnitude, in the CR rate and a reduction of the EH rate by 4 orders of magnitude, so that the diffusion of carriers to electrodes becomes extremely difficult. Although other factors, such as dot size distribution and oxidation, also play a deleterious role in device performance, their effects are deemed much less important than the critical role played by dangling bonds. ■ INTRODUCTION SiQD-based inorganic−organic hybrid solar cells are an attractive candidate for photovoltaic applications as they combine the advantages of their constituents. While the conjugated polymers have strong absorption in the visible range, SiQDs offer tunable energy levels, high stability and the potential for multiple exciton generation and hot carrier collection. Unfortunately, the solar conversion efficiency of the current SiQD/P3HT blend devices is only 1.1%, much lower than the typical value of 5% for their purely organic counterparts. To understand the origin of such limited performance, a novel broadband (UVVisNIR) transient absorption spectroscopy was recently employed to probe the relevant charge dynamics. The results indicate that in regioregular P3HT/SiQD films, the inherent photovoltaic conversion process is photoexcitation on P3HT followed by an ultrafast electron transfer from P3HT to the SiQD. Such exciton dissociation would primarily result in free charge carriers, but in reality only a small proportion of these carriers are able to diffuse to the electrode because of overwhelming nonradiative recombination at the interface. Although the charge transfer (CT) rate of 7 × 10 s−1 is extremely high, the ultrafast charge recombination (CR), with a rate of 1× 10 s−1, along with the low electron mobility (10−6 ∼ 10−5 cm V−1 s−1) causes the poor performance of the SiQD/P3HT solar cells. It is widely believed that a more precise control of the dot size distribution and a lower density of trap states are required, but their influences on the charge dynamics are still unclear, in particular, which factor is most critical to the deteriorated photoconversion? This has motivated our ab initio investigation to identify the main source responsible for the observed rapid recombination and low electron mobility, which would allow attention to be focused on the key roadblock to device performance. Four possible reasons for the low photoconversion efficiency have been postulated: (1) vibronically induced transitions to the ground state due to conical intersection at the degenerate points of the potential energy surfaces; (2) dot size distribution; (3) oxidation; (4) surface dangling bonds. However, quantitative support for these possibilities is lacking, motivating the current analysis. As detailed in the subsequent section, the associated computational methodology calls for phonon-assisted transport theory beyond the standard Marcus and MJL (Marcus, Levich, and Jortner) theories, and excitedstate electronic structure beyond those estimated from density functional theory (DFT) are required. The development of a successful computational approach is itself an important aspect of our investigation. Four rate estimates, depicted in Figure 1, are carried out using first-principles analysis: (Figure 1a) CT of singlet exciton; (Figure 1b) photoluminescence (PL) from P3HT; (Figure 1c) CR at the interface; (Figure 1d) electron hopping (EH) between two dots. Since phonon-assisted hopping is the Received: July 19, 2013 Revised: December 9, 2013 Published: December 9, 2013 Article