© 2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim 5451 www.advmat.de www.MaterialsViews.com COMMUNICATION wileyonlinelibrary.com Adv. Mater. 2011, 23, 5451–5455 Dr. H.-C. Chen, Dr. C.-W. Lai, Dr. I.-C. Wu, H.-R. Pan, Dr. I.-W. P. Chen, Y.-K. Peng, C.-L. Liu, Prof. C.-h. Chen, Prof. P.-T. Chou Department of Chemistry National Taiwan University E-mail: chop@ntu.edu.tw DOI: 10.1002/adma.201102775 Organic photovoltaics (OPVs) have attracted considerable interest as promising candidates for a new renewable energy source because of the facile tailored organic molecules, the highly tunable optical and electronic properties, as well as their potential for low cost and ease of fabrication. [1–8] The best performing bulk hetero- junction (BHJ) solar cells (power conversion efficiency, PCE > 7%) employ an optimized blend of a polymeric donor and a fullerene acceptor. [6–8] However, because of the low extinction coefficient of the fullerene derivatives, only a limited part of the solar spec- trum is generally harvested. Compared to fullerene derivatives, significant attention has been paid to the use of semiconductor nanocrystals (NCs) in terms of their long-term stability, sizable absorption ranging from the UV to the near infrared, electronic and geometric properties, and a tunable band structure. [9,10] Hybrid solar cells (HSCs) based on inorganic NCs and con- jugated polymers have the advantages of being morphologically more stable and being able to utilize the high electron mobility of the inorganic phase to overcome charge-transport limitations associated with the organic materials. [11–18] However, there is still no report of a higher performance observed in HSCs other than organic BHJ solar cells because of the complicated surface/ interface of NCs. These NCs are stabilized by non-conductive surfactants consisting of long alkyl chains, such as trioctylphos- phine oxide (TOPO) or n-octadecylphosphonic acid (ODPA) spe- cies, which seriously hamper the generation, separation, and transport of charge carriers. With strategies aimed at lifting this limitation, NCs have undergone pyridine ligand treatment to remove the long TOPO or ODPA ligand chains to achieve high charge separation yields and/or efficient charge transport. [18,19] For better interaction, a number of factors, including the NCs’ size and shape, [9,12,16] thermal annealing, [11–18] solvent–vapor annealing, [14,20] as well as directly grafting a conjugated polymer onto quantum dots, [21] have been identified experimentally as significantly affecting the nanometer-scale morphology. Recently, a new strategy for improving the compatibility between NCs and conjugated polymers was reported by functionalizing the polymers with substituents either as terminal or as side groups capable of binding to the NC’s surface. [15,22–24] Herein, we report the fabrication of HSCs with an inverted structure, employing a functional conjugated polymer of mono- aniline-capped poly[(4,4´-bis(2-ethylhexyl)-dithieno[3,2- b:2´,3´- d] silole)-2,6-diyl- alt-(2,1,3-benzothiadiazole)-4,7-diyl] (PSBTBT- NH 2 ) as a donor (D) and cadmium telluride (CdTe) NCs com- prising a tetrapod or nanorod-shape as an acceptor (A). We then demonstrate that a PV device efficiency can be substan- tially boosted by utilizing a unique mono-aniline end group of PSBTBT-NH 2 as a strong anchor to attach onto the CdTe NCs’ surfaces. Simultaneously, benzene-1,3-dithiol (BDT) solvent– vapor annealing is exploited to improve the charge separation at the D/A interface and efficient charge transportation in the HSCs. We also fully utilize the positive gain introduced by a 30-nm-thick fullerene (C 60 ) layer and cathode/anode buffer layers to promote the device performance. As a result, a record high PCE of up to 3.2% has been consistently achieved. The synthetic scheme of functional PSBTBT-NH 2 , along with an inverted device architecture and relative energy level diagram is shown in Figure 1. Detailed experimental procedures for the synthesis are described in the Supporting Information. In the CdTe:PSBTBT-NH 2 architecture, a MoO 3 layer was used as a hole extraction interlayer [25,26] between the active layer and a Pd electrode. The ZnO film was used as an electron transport buffer layer, [27,28] whereas the C 60 acted as a buffer layer to reduce the recombination of charges, passivate inorganic surface trap sites, and improve the exciton dissociation efficiency at the polymer/ metal oxide interface. [27–29] CdTe NCs were intended to func- tion as the electron transporting phase while the PSBTBT-NH 2 was expected to serve as the hole transporting component. The highest occupied molecular orbital (HOMO, –5.10 eV) and the lowest unoccupied molecular orbital (LUMO, –3.65 eV) levels of PSBTBT-NH 2 , together with the valence (–5.78 eV) and conduc- tion (–4.07 eV) bands of CdTe relative to the vacuum level, are schematically shown in Figure 1c. The HOMO and the LUMO levels for both PSBTBT-NH 2 and CdTe NCs are in positions such that electron injection is favorable toward ZnO while hole transport is directed toward the Pd electrode. It is worth noting that the work function of MoO 3 correctly matches Pd to enable barrier-less hole injection, resulting in a better Ohmic contact. Figure 2a,e reveal transmission electron microscopy (TEM) images and the corresponding selected-area electron diffraction (SAED) patterns of the pristine CdTe tetrapods and CdTe nano- rods. The as-prepared tetrapod NCs possess four arms, each of which is 4.9 ± 0.6 nm in diameter and 112.3 ± 24.5 nm in length, whereas the nanorod NCs are measured to be 4.5 ± 0.7 nm in width and 55.1 ± 7.2 nm in length. The SAED patterns of the Hsieh-Chih Chen, Chih-Wei Lai, I-Che Wu, Hsin-Ru Pan, I-Wen P. Chen, Yung-Kang Peng, Chien-Liang Liu, Chun-hsien Chen, and Pi-Tai Chou* Enhanced Performance and Air Stability of 3.2% Hybrid Solar Cells: How the Functional Polymer and CdTe Nanostructure Boost the Solar Cell Efficiency