Inorganic–Organic Hybrid Solar Cell: Bridging Quantum Dots to

LETTER pubs.acs.org/NanoLett

Inorganic Organic Hybrid Solar Cell: Bridging Quantum Dots to Conjugated Polymer Nanowires Shenqiang Ren,† Liang-Yi Chang,‡ Sung-Keun Lim,† Jing Zhao,‡ Matthew Smith,† Ni Zhao,§ Vladimir Bulovic,§ Moungi Bawendi,‡ and Silvija Gradecak*,† †

Department of Materials Science and Engineering, ‡Department of Chemistry, §Department of Electrical Engineering and Computer Science, Massachusetts Institute of Technology, 77 Massachusetts Avenue, Cambridge, Massachusetts 02139, United States

bS Supporting Information ABSTRACT: Quantum dots show great promise for fabrication of hybrid bulk heterojunction solar cells with enhanced power conversion efficiency, yet controlling the morphology and interface structure on the nanometer length scale is challenging. Here, we demonstrate quantum dot-based hybrid solar cells with improved electronic interaction between donor and acceptor components, resulting in significant improvement in short-circuit current and open-circuit voltage. CdS quantum dots were bound onto crystalline P3HT nanowires through solvent-assisted grafting and ligand exchange, leading to controlled organic inorganic phase separation and an improved maximum power conversion efficiency of 4.1% under AM 1.5 solar illumination. Our approach can be applied to a wide range of quantum dots and polymer hybrids and is compatible with solution processing, thereby offering a general scheme for improving the efficiency of nanocrystal hybrid solar cells. KEYWORDS: Bulk heterojunction solar cell, nanowire, quantum dots, chemical grafting

S

olution-processed nanostructured hybrid solar cells consisting of organic and inorganic components have attracted considerable interest due to the combined advantages of both material classes.1,2 In hybrid organic inorganic bulk heterojunction (BHJ) solar cells,3,4 electron-donating conjugated polymers are blended with inorganic semiconductor nanomaterials, taking advantage of their solution processability.5,6 So far, a variety of nanomaterials such as GaAs,7 CdSe,8 PbS,9 and ZnO10,11 nanowires or nanocrystals have been investigated for applications in hybrid solar cells. To improve BHJ solar cell device efficiency, several design criteria have to be taken into consideration. The domain size of the organic donor and inorganic acceptor materials should be comparable or smaller than the exciton diffusion length to increase the probability of exciton dissociation across the heterojunction. For many conjugated polymers, the exciton diffusion length is less than 10 nm,12 but in standard BHJ nanostructured solar cells processed from solution, device morphology cannot be precisely controlled on such a small scale. In addition, efficient transport of charge carriers to their respective electrodes before recombination increases the device current and voltage. Therefore, high electron and hole mobilities, controllable nanomorphology, and a well-structured interface are critical performance factors for developing efficient hybrid solar cells.13 Quantum dots (QDs) are particularly well-suited for applications in hybrid BHJ solar cells because their absorption can be tuned to cover a broad spectral range, they can have relatively high electron mobility,14 and have good photo- and chemical stability. However, achieving a controllable bicontinuous percolation network and a well-defined interface between QDs and the r 2011 American Chemical Society

polymer matrix remains challenging. In addition, the inorganic QDs tend to phase separate from nonpolar conjugated polymers at higher loading concentrations, decreasing interfacial area,15 and this phase separation of the photoactive films can be only partly mitigated by selection of solvent and processing conditions. Therefore, the next generation of hybrid QD photovoltaics will require a strategy for controlling phase separation, increasing interfacial areas, and improving the optoelectronic interactions of inorganic QDs and organic polymers. Here, we demonstrate a facile method to synthesize and control the nanoscale morphology of poly(3-hexylthiophene) (P3HT) nanowires (NWs) and CdS QDs hybrids using chemical grafting16 and ligand exchange methods.17 Solvent-assisted chemical grafting and ligand exchange were used to control the interface of P3HT/CdS nanohybrids and the CdS QD interparticle distance, respectively. The effects of controlling morphology on the optical and electrical properties are discussed. We find that the formation of an interpenetrating and percolating P3HT/CdS BHJ network gives rise to efficient charge separation and charge transport, significantly improving photovoltaic performance. The synthesis of CdS QDs followed procedures reported by Peng et al.18 After the synthesis the CdS QDs were centrifuged, precipitated with isopropanol, and redispersed in hexane. P3HT NWs were prepared using solution-phase self-assembly. Nonpolar P3HT was initially completely solvated in a nonpolar Received: July 15, 2011 Revised: August 16, 2011 Published: August 22, 2011 3998

dx.doi.org/10.1021/nl202435t | Nano Lett. 2011, 11, 3998–4002

Nano Letters

LETTER

Figure 1. (a) The schematic device architecture and (b) the corresponding flat-band diagram.

solvent, 1,2-dichlorobenzene (1,2-DCB). By adding nonsolvent cyclohexanone into P3HT, the interaction among P3HT polymer chains becomes dominant and generates a randomly aggregated state. In this stage, the P3HT solution forms a dark purple gel whose viscosity strongly decreases under manual agitation and results in a formation of P3HT NWs. To form the photoactive layer of the hybrid solar cell, P3HT NWs were blended with CdS QDs (Figure 1a). We selected CdS QDs due to the large band offset (1.6 eV) between the HOMO level of P3HT and the conduction band edge of CdS, which can improve the open circuit voltage (VOC) (Figure 1b).19 Hybrid solar cells consisting of P3HT/CdS nongrafted and grafted nanowire structures were prepared by a solvent-assisted method. For the nongrafting process, the same solvent 1,2-dichlorobenzene (1,2-DCB) was used for the CdS QDs and the P3HT NWs. In the chemical grafting process, the 1D coaxial nanowire structures were formed by first dissolving the P3HT and CdS into 1,2-DCB and octane, respectively, and then mixing these two solutions together. In both cases, the P3HT/CdS blended solution was spun onto an ITO patterned glass slide to form a 140 nm thick film. The coated devices were heat-treated at 175 °C for 10 min and quickly cooled to room temperature. Finally, the 10 nm thick bathocuproine (BCP) hole blocking layer and the top Mg/Ag electrode were evaporated and the final device area was defined as the overlap between the top and bottom electrodes. Current voltage (J V) characteristics of the devices were measured in a nitrogen atmosphere glovebox with a Keithley 6487 source meter. The light response was measured under 100 mW/cm2 illumination from an AM1.5 solar simulator. Transmittance and absorbance spectra were measured with a Cary 5000 UV vis NIR dual-beam spectrophotometer. A JEOL 2010 FEG analytical transmission electron microscope (TEM) operated at 200 kV and equipped with an energy dispersive spectrometer (EDS) was used for structural and chemical analysis. Device surface morphology was investigated using a Digital Instruments Dimension 3000 atomic force microscope (AFM) operated in tapping mode. The photoactive material of our hybrid solar cell consists of the electron-donor P3HT NWs and electron-acceptor CdS QDs. Figure 2a,b shows representative TEM images of P3HT NWs and colloidal 4 nm CdS QDs capped by oleic acid ligands (for TEM imaging only), respectively. As prepared P3HT NWs have an aspect ratio of 100 with an average length of 1000 ( 150 nm and diameter of 10 ( 2 nm. Figure 2c shows the UV vis absorbance spectra of P3HT and CdS components. In comparison with the pristine P3HT phase, a significant red-shift and an additional vibronic feature at 600 nm are observed in the absorbance spectra of P3HT NWs (Figure 2c). These results

Figure 2. TEM images of (a) P3HT nanowires and (b) CdS quantum dots. (c) Absorbance of pristine P3HT, P3HT nanowires, and CdS quantum dots.

indicate that 1D self-organized P3HT NWs have improved structural ordering and the π π interactions between polymer chains lead to superior electronic properties,20 as supported by a two-order-of-magnitude increase of the hole mobility in P3HT NWs (not shown). Because the diameters of P3HT NWs are below the exciton diffusion length of the P3HT phase, we studied the effect of the grafting process on exciton dissociation between the P3HT and CdS QD phases by comparing a randomly blended system (nongrafting process) with 1D coaxial nanostructures formed by decorating P3HT NWs with CdS QDs (grafting process). A key to the grafting process is to dissolve both QDs and NWs in compatible solvents, allowing them to interact with each other under conditions that preserve the stability and photophysical properties of both components. In the present study, the P3HT NWs and CdS QDs were dissolved in 1,2-DCB and in octane, respectively. Figure 3a shows a TEM image of the randomly blended P3HT/CdS system; the CdS QDs are randomly distributed and there is no ordering interaction between P3HT NWs and CdS QDs. The chemically grafted CdS QDs increase the organic inorganic interaction and maximize the interfacial area with P3HT NWs by the formation of P3HT/CdS coaxial nanowire hybrids (Figure 3b). This architecture carries advantages over a randomly distributed system because phase segregation is minimized and the P3HT-QD interface is maximized, improving dissociation of photogenerated excitons in P3HT, 3999

dx.doi.org/10.1021/nl202435t |Nano Lett. 2011, 11, 3998–4002

Nano Letters which occurs due to the conduction band offset at the interface. Figure 3c shows the absorbance and photoluminescence (PL) of films produced from nongrafted and grafted P3HT NWs with 80 wt % CdS. The vibronic absorption feature at 605 nm in the grafted film is clearly resolved, indicating that the crystalline P3HT NWs are preserved after CdS QDs grafting. The PL spectral intensities of the P3HT/CdS grafted hybrid films are significantly quenched, presumably due to the efficient nonradiative channel for charge transfer from P3HT to CdS phase. In order to understand the bonding structure of the CdS QDgrafted P3HT NWs, X-ray photoelectron spectroscopy (XPS) spectra of P3HT/CdS hybrids with and without grafting were obtained (Figure 3a,b, respectively). We focus on the XPS spectra of S 2p peaks to more precisely identify the bond formed during chemical grafting. Figure 4a shows the S 2p peaks for nongrafted samples where the S bonding energies is well matched through peak deconvolution with the independent P3HT and CdS components. In comparison, XPS spectra of the grafted CdS/P3HT (Figure 4b) reveals a new intermediate

Figure 3. TEM images of P3HT/CdS hybrid films synthesized (a) without grafting and (b) using grafting process by solvent exchange. The inset images show schematic representations of each; the nongrafting and grafting method is used to control the interface between CdS QDs (yellow spheres) and P3HT NWs (purple lines). (c) Optical properties of P3HT/CdS hybrid thin films; absorbance of the grafting sample and PL spectra of the nongrafted and grafted films.

LETTER

bonding energy, which can be assigned to the C S Cd bond connecting the P3HT polymer backbone and CdS QDs21 forming a close contact after the grafting. These findings are further supported by transient PL measurements of P3HT/CdS hybrids with and without grafting (Figure 4c), which show that grafted samples have a shorter exciton lifetime and thus faster exciton dissociation. Taken together, these results can be understood as follows: the grafting of CdS QDs onto the surface of highly ordered P3HT NWs increases the interaction between these two materials and creates a large number of interfaces for charge transfer,22 which results in a shorter PL lifetime, quenched PL intensities, and increased charge transfer rates. It was recently demonstrated that the performance of solar cells based on PbS QDs can be improved by ligand-exchange treatments involving thiols.23 The beneficial effect of thiols has been attributed to a shortened QD spacing, improved crosslinking of QDs, and improved surface passivation. The photoconductivity of different QDs has been shown to be largely enhanced with a consequence of the ligand exchange treatment.17 We performed the ligand exchange in our organic P3HTinorganic CdS hybrid system to shorten the interdistance between CdS QDs and the P3HT NWs. The CdS QDs were

Figure 5. Tapping mode AFM images of P3HT/CdS hybrid thin films (a) before and (b) after the ligand exchange. (c) TEM image of curved P3HT NWs after the ligand exchange. (d) Charge carrier recombination rate constants krec determined by transient open circuit voltage decay measurements at different illumination levels.

Figure 4. XPS spectra of high resolution S 2p in (a) nongrafted and (b) grafted P3HT/CdS hybrid films. Black full line corresponds to the experimental result that was deconvoluted (dashed lines) to yield the final fit (red full line). (c) Time-resolved photoluminescence spectra of the nongrafted and grafted films. 4000

dx.doi.org/10.1021/nl202435t |Nano Lett. 2011, 11, 3998–4002

Nano Letters

Figure 6. (a) Cross-sectional TEM image of P3HT/CdS hybrid solar cells. Inset image shows the EDS elemental mapping (red, Cd; green, In; and blue, Ag). (b) Current voltage characteristics of P3HT/CdS hybrid solar cells from nongrafting, grafting and the subsequent ligand exchange. (c) The photovoltaic performance summary (JSC and VOC) of P3HT/CdS after the chemical grafting and ligand exchange, as a function of the CdS weight concentration.

originally capped with n-butylamine surfactants, which were removed through an ethanedithiol (EDT) treatment in acetonitrile solvent. Because acetonitrile is a relatively good solvent for ligands, but a poor solvent for P3HT NWs, the solvent quality induces ordering in the polymer phase and change the degree of phase separation. The surface morphology before and after the ligand exchange was investigated using AFM and TEM (Figure 5a c, respectively) and it was found that the P3HT/ CdS film forms small agglomerates and, as a consequence, CdS QDs get closer after the ligand exchange. Next, we examined the effect of the controlled nanoscale morphology on electrical properties. To directly measure charge recombination, we fabricated a device after grafting and ligandexchange treatment, and performed small perturbation transient VOC decay measurements.24,25 In this measurement, a lowintensity pulsed laser light at 520 nm wavelength is used to induce a small perturbation to the VOC by transiently generating additional electrons and holes. The resulting additional transient photovoltage (V) then decays with a lifetime that is determined by the recombination rate constant (krec) of the electrons and holes. In Figure 5d, krec is shown as a function of the solar simulator light intensity. For all devices, we observe that the ligand exchange-treated device exhibits krec approximately two times higher than that of the nontreated device, indicating that photogenerated charge carrier concentrations are significantly higher than that of the nontreated device. The ligand exchange process reduces the distance among CdS QDs and thus increases the free charge carriers due to efficiently separated charge transfer states, consequently increasing the charge recombination rate constant. The charge carrier concentration and charge transport is highly dependent on the carrier mobility of the materials, which also strongly influences the photovoltaic performance. Therefore, both electron and hole mobility in CdS and P3HT components should be matched and high enough to extract charges efficiently and prevent build up of space charge. We have tested the electron-only and hole-only devices with grafting

LETTER

hybrid films (details of the device fabrication and representative results are described in the Supporting Information, Figure S1), which indicate well-matched electron and hole mobilities after the ligand exchange process. We note that the conductivity of P3HT phase is reduced by the EDT treatment, which may result from the decreased P3HT molecular ordering due to the aggregated features as shown in the TEM image (Figure 5c). Following these promising transient measurements, devices based on chemical grafting and ligand-exchange treated films were prepared, as described in Figure 1a. The bright-field TEM image of a cross-section of a typical complete device (Figure 6a) shows that CdS QDs are uniformly distributed across the 150 nm thick photoactive layer. The grafting of CdS QDs onto P3HT NWs is confirmed using EDS elemental mapping image (the inset EDS elemental mapping image of Figure 6a and Supporting Information, Figure S2). In comparison with the QD phase separation that occurs in randomly blended hybrid solar cells and leads to QD-rich top layer, the image suggests that the P3HT NW structure and chemical grafting process plays a crucial role in templating CdS QD across the vertical direction and in improving the interface between P3HT and CdS components. Figure 6b shows the current voltage characteristics of the P3HT/CdS hybrid solar devices using nongrafted films, grafted films, and subsequent EDT treated films as a photoactive layer, whereas Figure 6c shows a summary of the short circuit current density (JSC) and VOC as a function of CdS wt% for samples after the chemical grafting and ligand exchange. Devices at 80 wt % CdS QDs without chemical grafting shows average JSC of 1.9 mA/cm2, VOC of 0.8 V, and a fill factor (FF) of 0.40 under simulated AM 1.5 illumination (100 mW/cm2) and gives a power conversion efficiency of 0.6%. Our results show that a significant improvement of the photovoltaic performance is achieved in hybrid devices fabricated using chemical grafting and ligand exchange processes, giving an improved average JSC of 10.0 mA/cm2, VOC of 1.0 V, FF of 0.32 and the power conversion efficiency of 3.2%, which are measured over 100 devices (the best solar device performance gives JSC of 10.9 mA/cm2, VOC of 1.1 V, FF of 0.35 and the maximum efficiency of 4.1%). The enhancement in JSC could be attributed to both the increased charge separation efficiency in the photoactive film, enabled by the chemical grafting CdS QDs onto highly ordered P3HT NWs, and the enhanced charge collection efficiency, enabled by the bicontinuous percolated nanomorphology in hybrid solar devices that can provide more effective transport pathways for both electrons and holes. The formation of a CdS coaxial percolation network and its effect on charge transport is supported by the fact that as the CdS weight concentration is increased in the grafting samples, the JSC increases to a maximum value at 80 wt %, and then decreases for a higher loading (Figure 5c). In conclusion, we have presented a facile and simple method to synthesize and control the nanomorphology of P3HT/CdS coaxial NW BHJ hybrid solar cells. The chemical grafting of CdS QDs onto crystalline P3HT NWs results in not only the controllable dispersion of CdS QDs within the P3HT matrix, but also stronger electronic interaction between CdS QDs and P3HT NWs. The formation of bicontinuous donor/acceptor phases and a well-defined interface in the hybrid photoactive film through grafting and ligand exchange can largely enhance charge separation and transport efficiency, which are essential for hybrid inorganic organic solar cells. The results shown here represent a general approach for increasing the efficiency of hybrid solar cells 4001

dx.doi.org/10.1021/nl202435t |Nano Lett. 2011, 11, 3998–4002

Nano Letters via enhancing the interfacial interactions and controlling nanomorphology between the organic and inorganic materials.

’ ASSOCIATED CONTENT

bS

Supporting Information. Materials and experimental methods, device cross-sectional EDS elemental mapping, and characterization of the electron and hole devices. This material is available free of charge via the Internet at http://pubs.acs.org.

LETTER

(21) NIST X-ray Photoelectron Spectroscopy Database, V. N. I. o. S. a. T., Gaithersburg, 2003; http://srdata.nist.gov/xps/. (22) Zhao, L.; Pang, X. C.; Adhikary, R.; Petrich, J. W.; Lin, Z. Q. Angew. Chem., Int. Ed. 2011, 50, 3958. (23) Luther, J. M.; Law, M.; Song, Q.; Perkins, C. L.; Beard, M. C.; Nozik, A. J. ACS Nano 2008, 2, 271. (24) O’Regan, B. C.; Scully, S.; Mayer, A. C.; Palomares, E.; Durrant, J. J. Phys. Chem. B 2005, 109, 4616. (25) Goh, C.; Scully, S. R.; McGehee, M. D. J. Appl. Phys. 2007, 101, 114503.

’ AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected].

’ ACKNOWLEDGMENT We thank Geoffrey Supran for assistance with AFM imaging and Eric Jones for technical support. This work was supported by Eni S.p.A. under the Eni-MIT Alliance Solar Frontiers Program. The authors acknowledge access to Shared Experimental Facilities provided by the MIT Center for Materials Science Engineering supported in part by MRSEC Program of National Science Foundation under award number DMR - 0213282. ’ REFERENCES (1) Milliron, D. J.; Gur, I.; Alivisatos, A. P. MRS Bull. 2005, 30, 41. (2) Gur, I.; Fromer, N. A.; Chen, C. P.; Kanaras, A. G.; Alivisatos, A. P. Nano Lett. 2007, 7, 409. (3) Yu, G.; Heeger, A. J. J. Appl. Phys. 1995, 78, 4510. (4) Huynh, W. U.; Dittmer, J. J.; Alivisatos, A. P. Science 2002, 295, 2425. (5) Greenham, N. C.; Peng, X. G.; Alivisatos, A. P. Phys. Rev. B 1996, 54, 17628. (6) Zhou, Y. F.; Riehle, F. S.; Yuan, Y.; Schleiermacher, H. F.; Niggemann, M.; Urban, G. A.; Kruger, M. Appl. Phys. Lett. 2010, 96, 013304. (7) Ren, S. Q.; Zhao, N.; Crawford, S. C.; Tambe, M.; Bulovic, V.; Gradecak, S. Nano Lett. 2011, 11, 408. (8) Wang, P.; Abrusci, A.; Wong, H. M. P.; Svensson, M.; Andersson, M. R.; Greenham, N. C. Nano Lett. 2006, 6, 1789. (9) Zhao, N.; Osedach, T. P.; Chang, L. Y.; Geyer, S. M.; Wanger, D.; Binda, M. T.; Arango, A. C.; Bawendi, M. G.; Bulovic, V. ACS Nano 2010, 4, 3743. (10) Shim, M.; Guyot-Sionnest, P. J. Am. Chem. Soc. 2001, 123, 11651. (11) Beek, W. J. E.; Wienk, M. M.; Janssen, R. A. J. Adv. Funct. Mater. 2006, 16, 1112. (12) Shaw, P. E.; Ruseckas, A.; Samuel, I. D. W. Adv. Mater. 2008, 20, 3516. (13) Chen, L. M.; Hong, Z. R.; Li, G.; Yang, Y. Adv. Mater. 2009, 21, 1434. (14) Kovalenko, M. V.; Scheele, M.; Talapin, D. V. Science 2009, 324, 1417. (15) Coe-Sullivan, S.; Steckel, J. S.; Woo, W. K.; Bawendi, M. G.; Bulovic, V. Adv. Func. Mater. 2005, 15, 1117. (16) Yang, S.; Wang, C. F.; Chen, S. J. Am. Chem. Soc. 2011, 133, 8412. (17) Jarosz, M. V.; Porter, V. J.; Fisher, B. R.; Kastner, M. A.; Bawendi, M. G. Phys. Rev. B 2004, 70. (18) Peng, Z. A.; Peng, X. G. J. Am. Chem. Soc. 2001, 123, 183. (19) Leventis, H. C.; King, S. P.; Sudlow, A.; Hill, M. S.; Molloy, K. C.; Haque, S. A. Nano Lett. 2010, 10, 1253. (20) Park, Y. D.; Lee, S. G.; Lee, H. S.; Kwak, D.; Lee, D. H.; Cho, K. J. Mater. Chem. 2010, 21, 2338. 4002

dx.doi.org/10.1021/nl202435t |Nano Lett. 2011, 11, 3998–4002