Mixed-quantum-dot solar cells

Abstract

Colloidal quantum dots are emerging solution-processed materials for large-scale and low-cost photovoltaics. The recent advent of quantum dot inks has overcome the prior need for solid-state exchanges that previously added cost, complexity, and morphological disruption to the quantum dot solid. Unfortunately, these inks remain limited by the photocarrier diffusion length. Here we devise a strategy based on n- and p-type ligands that judiciously shifts the quantum dot band alignment. It leads to ink-based materials that retain the independent surface functionalization of quantum dots, and it creates distinguishable donor and acceptor domains for bulk heterojunctions. Interdot carrier transfer and exciton dissociation studies confirm efficient charge separation at the nanoscale interfaces between the two classes of quantum dots. We fabricate the first mixed-quantum-dot solar cells and achieve a power conversion of 10.4%, which surpasses the performance of previously reported bulk heterojunction quantum dot devices fully two-fold, indicating the potential of the mixed-quantum-dot approach.

Fig. 1

Preparation of mixed-quantum-dot solids using solution ligand exchange. a Schematic of the solution ligand exchange process. b Calculated HOMO and c LUMO states of two coupled dots, showing charge transfer between MAPbI₃- and TG-capped QDs in the mixed film. d Projected density of states demonstrate the offset of the band position between MAPbI₃- and TG-capped QDs

Fig. 2

Characterizations of ligand attachment and transient absorption photophysics. a Fourier transform infrared (FT-IR) spectra of PbS QDs passivated by (i) MAPbI₃ (ii) TG, and (iii) two types of ligands (mixed-QD sample). b ¹H NMR spectra of quantum dot ligands in solution (red) and bound to PbS quantum dot surface (black). (i) Oleic acid and lead oleate-capped QDs in C₆D₆ benzene. (ii) TG and TG-capped QDs in d₆-DMSO. (iii) MAI + PbI₂ solution sample and MAI + PbI₂-capped QDs in d₇-DMF. For (ii) and (iii), signals between 0 and 2 ppm are from residual oleate bound to the QDs after the ligand exchange. Peaks from trace nondeuterated solvents are marked by asterisks. c Schematic representation of spectral diffusion. Dashed lines indicate the position of time traces for panels d–f relative to the peak maximum of the exciton bleach. d–f TA spectral time traces of bleaching signals for d TG-capped QDs, e MAPbI₃-capped QDs, and f mixed-QD films. We track spectral diffusion by the point of crossover between the decay of high-energy wavelengths and the rise of lower energy wavelengths, encircled in black

Fig. 3

Morphology of the mixed-quantum-dot structure for balanced carrier extraction. a Cross-section of a mixed film with different QD donor/acceptor ratios. Blue and red spheres represent acceptor and donor QDs, respectively. b Distribution of path lengths for a 1:1 A- to D-type QD ratio. For this configuration, the shortest path (normalized to thickness) is 1.1, and the median length of 1.14. c Slices of mixed CQD solids with different D/A ratios representing the dominant presence of each type of QD. d Normalized path length as a function of A- to D-type QD ratio within a mixed QD film. Error bars represent the standard deviation of the 10 shortest different path lengths for each A/D configuration. e Balanced extraction (β = 1) requires a compromise between electron and hole mobilities and their respective path-lengths. A maximum power conversion efficiency is expected when the carrier extraction is balanced (Supplementary Fig. 14)

Fig. 4

Electron microscopy analysis of mixed-quantum-dot solid and device. a STEM-HAADF image of freshly drop-cast mixed-QD ink on an ultrathin graphite-based TEM grid (sample was not annealed). QDs with amorphous MAPbI₃ shell (b) and without surface shelling (c). d Cross-section STEM image of FIB-processed mixed QD-based photovoltaic device. Inset: EELS mapping of iodine in the highlighted region (200 nm by 200 nm) indicating the homogenous dispersion of MAPbI₃-passivated QD within the mixed QD active layer (major edge of iodine M_4,5 signal starting at around 625 eV, signal intensity unit: a.u., Supplementary Fig. 19)

Fig. 5

Optoelectronic behaviors of mixed-quantum-dot solids and device performance. a Photoluminescence spectra of TG-, MAPbI₃-capped, and mixed QD films confirming the quenching effect on mixed QD sample because of efficient charge separation between donor/acceptor QDs. b Power conversion efficiency (PCE) values of mixed QD devices with different mass ratios of D- (TG-capped) and A-type (MAPbI₃-capped) QDs. c Current−voltage (J−V) characteristics under simulated AM 1.5 illumination for BHJ device and control A-type only (MAPbI₃-capped) PbS device. d Thickness-dependent PCE and e thickness dependent of BHJ and control perovskite-passivated QD devices. f EQE spectra of champion-mixed QD device and perovskite-passivated QD control sample. Error bars in b, d, and e represent the standard deviation of several devices