Conjugated Thiophene Dendrimer with an Electron-Withdrawing Core

Nov 30, 2010 - David S. Ginley,‡ Sean E. Shaheen,| Garry Rumbles,‡ and Nikos Kopidakis*,‡. Department of Physics, Colorado School of Mines, 1500...
1 downloads 0 Views 2MB Size
Download PDF
J. Phys. Chem. C 2010, 114, 22269–22276

22269

Conjugated Thiophene Dendrimer with an Electron-Withdrawing Core and Electron-Rich Dendrons: How the Molecular Structure Affects the Morphology and Performance of Dendrimer:Fullerene Photovoltaic Devices William L. Rance,*,† Benjamin L. Rupert,‡ William J. Mitchell,‡ Muhammet E. Ko¨se,§ David S. Ginley,‡ Sean E. Shaheen,| Garry Rumbles,‡ and Nikos Kopidakis*,‡ Department of Physics, Colorado School of Mines, 1500 Illinois Street, Golden, Colorado 80401, United States, National Renewable Energy Laboratory, 1617 Cole BouleVard, Golden, Colorado 80401, United States, Department of Chemistry and Biochemistry, North Dakota State UniVersity, Fargo, North Dakota 58108, United States, and Department of Physics and Astronomy, UniVersity of DenVer, DenVer, Colorado 80208, United States ReceiVed: July 22, 2010; ReVised Manuscript ReceiVed: October 5, 2010

The combination of electron-rich and electron-poor moieties in conjugated molecules is frequently utilized in order to red shift the absorption spectrum and improve photon harvesting in bulk heterojunction photovoltaic devices. In this study we characterize a conjugated thiophene dendrimer that has an electron-withdrawing core and electron-rich dendrons in order to investigate the effects of this design approach on the salient properties that influence the performance of photovoltaic devices with this dendrimer donor. Beside the absorption onset, these properties are the morphology of dendrimer:fullerene films and the dynamics of photoinduced carrier generation and loss. For comparison we also characterize a control dendrimer with the same structure but without the electron-withdrawing core. In addition to lowering the band gap by ca. 0.5 eV, the electron-withdrawing core also planarizes the dendrimer resulting in enhanced order in bulk heterojunction films. We observe longer photocarrier lifetimes in this ordered structure compared to the films of the predominantly amorphous control. The characterization of dendrimer:fullerene bulk heterojunction photovoltaic devices shows no voltage loss despite the decreased absorption onset. The properties of the device are consistent with the improved photocarrier lifetimes, but they are limited by a low short-circuit photocurrent density. We attribute this to electron confinement in the core that hinders transfer to the fullerene acceptor. 1. Introduction Organic photovoltaics (OPV) have attracted much interest recently, with power conversion efficiencies exceeding 7% in polymer:fullerene bulk heterojunctions.1 Extensive research in photoactive polymers has led to macromolecular designs that offer control of important properties that pertain to the performance of polymer:fullerene-based OPV. Alternating donoracceptor copolymers, also termed “push-pull” structures, have been designed to have low absorption onset, thereby improving the harvesting of solar photons.2,3 Design of the backbone and side chains4-6 as well as processing additives7 has offered control of the morphology of films that has led to improved charge carrier collection. It has also been recently recognized that the energetics at the polymer:fullerene interface, measured by the free energy difference for free-carrier production following exciton dissociation, is an important factor that may need to be considered in the design of next-generation materials for OPV.8 While polymers have been at the forefront of solution-based OPV, solution processable small molecules are increasingly successful as donors, demonstrating device efficiencies up to 4.4% when blended with fullerene acceptors.9 Small molecules are attractive because of their well-defined structure that allows for the in-depth understanding of their structure-property * To whom correspondence should be addressed. E-mail: wrance@ mines.edu (W.L.R); [email protected] (N.K.). † Colorado School of Mines. ‡ National Renewable Energy Laboratory. § North Dakota State University. | University of Denver.

correlations that is crucial to the design of molecules tailored for OPV applications.10 π-Conjugated dendrimers are a particular class of molecules with a highly branched structure that offers the possibility to “fill space” with electronically active groups without obstruction from the solubilizing groups.11 Previously we have reported on the synthesis,12,13 spectroscopy,14,15 charge transport,16,17 and photovoltaic properties18 of conjugated dendrimers composed of a phenyl core with conjugated thiophene dendrons (“arms”) attached at ortho, para, and meta positions on the core. These dendrimers were primarily first generation, i.e., the dendron only had one branching point, with the number of R-linked thiophenes in the arm ranging from three to five, making these structures two-dimensional. Due to the high band gaps (>2 eV) and the poor morphology of these early version of dendrimers, device efficiency of bulk heterojunction blends using phenyl-C61-butyric acid methyl ester (PCBM) as the acceptor was limited to 1.3% under 1 sun.18 In an effort to improve overlap with the solar spectrum, we recently reported on a series of thiophene dendrimers that were modified with electron-donating or -withdrawing groups in order to lower the band gap.13 Shown in Figure 1, the dendrimer 3G12S-CN (2) uses an electron-withdrawing cyanobenzene core and thiophene arms to create a push-pull structure that lowers the lowest unoccupied molecular orbital (LUMO). A similar dendrimer 3G1-2S-Ac (1), also shown in Figure 1, was synthesized without the electron-withdrawing cyano groups on the core but is otherwise identical to 2. Compared to the control molecule 1, the optical absorption onset of 2 was ca. 0.5 eV lower. Cyclic voltammetry measurements showed this change to be primarily

10.1021/jp106850f  2010 American Chemical Society Published on Web 11/30/2010

22270

J. Phys. Chem. C, Vol. 114, No. 50, 2010

Rance et al.

Figure 1. Chemical structure of 3G1-2S-CN (2) and 3G1-2S-Ac (1) dendrimers. (inset) Energy levels of dendrimers13 and PCBM34 determined from cyclic voltammetry.

attributable to lowering of the LUMO level.13 Previous density functional theory (DFT) calculations of the structure of 2 have shown that, with the exception of the β-linked thiophene at the end of the dendron, the molecule is planar and therefore π-stacking should be favored in films.15 The planar structure of dendrimer 2 can be attributed solely to the cyanobenzene core, since DFT calculations performed for 1 showed a bowl-shaped structure. Calculations of the frontier orbitals and transition densities also showed that the lower optical absorption onset is attributable to an intermolecular charge-transfer transition due to the combination of the electron-rich thiophene arms and electron-poor cyanobenzene core.15 Here we show that in addition to red shifting the absorption onset, the change of the electronic structure of the core of dendrimer 2 influences a multitude of the essential properties of the bulk heterojunction device. We report the influence that this push-pull molecular structure has on morphology and photocarrier dynamics of both neat dendrimer films and bulk heterojunction films with PCBM acting as the acceptor. The OPV device performance of 1 and 2 with PCBM is also reported. We show that the planar structure of 2 promotes order in both neat and blend films, and also aids in the formation of PCBM crystallites in bulk heterojunctions. The improved order in films of 2 is shown to result in higher photoconductivity and longer photocarrier lifetimes compared to 1. These improvements to morphology and carrier dynamics result in a higher efficiency of 1.12% in OPV devices of 2:PCBM. Despite the decrease of the band gap of 2 by ca. 0.5 eV, devices retain a high VOC of 0.93 V. While the external quantum efficiency spectrum of 2:PCBM is red-shifted compared to that of 1:PCBM, its absolute values are not enhanced despite the

improved morphology. We attribute this to electron confinement to the cyanobenzene core that may be impeding transfer to PCBM. 2. Experimental Section Both dendrimers were synthesized by a convergent route with the dendrons and core separately synthesized and coupled to form the dendrimer in the final step.13 Both dendrimers have almost identical structures except for the addition of three cyano groups attached at meta positions around the core of 2. The dendron structure remains the same with three thiophene dendrons attached to a phenyl core using acetylene groups. Acetylene groups were used as spacers to reduce crowding around the core. Each dendron was terminated with hexyl groups to improve solubility. For blend measurements, PCBM was purchased from Nano-C and was used as received. For device and X-ray diffraction (XRD) film preparation, 20 mg/mL by weight solutions were used for both dendrimer and dendrimer:PCBM films. The blending ratio of 1:4 dendrimer: PCBM by weight was used for all blend films, since both 2 and other previously studied dendrimers18 were optimized at this higher PCBM weight ratio. A toluene:chlorobenzene 1:1 by volume mixture was used as the solvent, and solutions were allowed to stir for 24 h at 60 °C. To produce thicker films for contactless time-resolved microwave conductivity (TRMC) measurements, 40 g/L solutions were used. Quartz substrates (Gem Dugout) cut 6° from (0001) were used as film substrates for XRD measurements, while quartz slides (Allen Scientific Glass) were used for TRMC measurements. Device substrates consisted of patterned indium tin oxide (ITO) on glass. All substrates were similarly cleaned through sonication in acetone and isopropyl alcohol baths, followed by an O2 plasma cleaning step.

Dendrimer:Fullerene Photovoltaic Devices

J. Phys. Chem. C, Vol. 114, No. 50, 2010 22271

Figure 2. Normalized absorbance of 1 and 2 in THF solution and film.

Device and TRMC films were spin-coated from solution at 800 rpm for 60 s, while XRD films were drop cast from solution. To complete OPV device structure, a 30 nm layer of poly(3,4ethylenedioxythiophene):poly(styrenesulfonate) (PEDOT:PSS) was spin-coated on the ITO prior to the deposition of the active layer, and a 100 nm aluminum layer was thermally evaporated to act as the top contact, producing a device area of 0.11 cm2. UV-vis measurements were performed using a HewlettPackard 8453 UV-vis spectrophotometer. XRD measurements were carried out on a Scintag PTS 4-circle goniometer (Bragg-Brentano geometry) with Cu KR radiation. A liquid nitrogen cooled germanium detector was used, and each sample was scanned from 2° to 30° with a 0.10°/s scan rate. Solar cells were measured in a nitrogen atmosphere with a Keithley 236 source measurement unit under a 100 mW/cm2 tungsten halogen bulb. Two calibrated Hamamatsu S1787-04 Si diodes were used to monitor the light intensity, and an Ocean Optics USB4000 spectrometer was used to determine the spectral mismatch. For external quantum efficiency (EQE) measurements, devices were illuminated by a chopped 300 W Xe lamp filtered by a monochromator and referenced against a calibrated silicon photodiode. The photoconductance of each dendrimer and dendrimer: PCBM blend was measured by flash-photolysis TRMC, a pump-probe photoconductivity probe in which free carriers are generated in the organic film by a laser pulse and monitored with a steady-state microwave probe beam. Measurements were carried out in a similar manner as has been described previously.19,20 Each sample was placed in a resonance cavity at the end of an X-band microwave waveguide that allowed for optical excitation on one end through a grating. An optical parametric oscillator (OPO) pumped by a Q-switched Nd:YAG laser provided excitation pulses of 5 ns duration. The excitation wavelength was chosen to be near the absorption maximum of 1 and 2 at 450 and 550 nm, respectively. The laser power was controlled through various combinations of neutral density filters. Upon excitation, mobile carriers were produced within the film, leading to a drop in the measured microwave power within the cavity. This change in microwave power, ∆P, is related to the change in photoconductance, ∆G, by

∆G ) -

1 ∆P K P

(1)

where K is an experimentally determined calibration factor.21 This measured photoconductance is the convolution of the actual photoconductance with the response of the micowave cavity.20 The peak of the photoconductance, called the end-of-pulse photoconductance ∆GEOP, can then be related to the sum of the electron and hole mobilities in the film, ∑µ, by

∆GEOP ) βqeI0FAφΣµ

(2)

where β is the ratio of the interior dimensions of the waveguide, qe is the electronic charge, I0 is the incident laser power, FA is the fraction of absorbed photons, and φ is the yield for freecarrier generation per absorbed photon.22 While the TRMC measurement is sensitive to all free carriers in the film, in systems with a large mismatch in mobilities the most mobile carrier dominates the photoconductivity signal. 3. Results and Discussion The normalized absorbances of 1 and 2 in both film and tetrahydrofuran (THF) solution are shown in Figure 2. A single absorption peak is seen in the solution spectrum of 1, with a slight red shift in the film yielding an optical band gap of 2.4 eV. The solution spectrum of 2 shows two absorption bands. The short wavelength band coincides with the absorption band of 1, and this absorption band in both dendrimers can be attributed to transitions that take place solely within the thiophene arm due to the meta-linked dendrons limiting conjugation through the core.15,23 The longer wavelength band in 2 is attributed to an intermolecular charge-transfer transition between the dendron and the core and is the optical manifestation of the presence of the electron-poor cyanobenzene core.15 This is further supported by the low photoluminescence quantum yield previously observed in dendrimer 2, since emission is typically less efficient in charge-transfer states.13 The absorption spectrum of 2 in film is further red shifted indicating ordering in the solid state, consistent with XRD measurements in films (see below), although a contribution from a different dielectric environment in the film cannot be ruled out. The optical band gap can be estimated from these results to be 1.8 eV, in agreement with previous work.13 Further evidence of ordering of dendrimer 2 is seen in XRD measurements of neat films, shown in Figure 3. At low scattering

22272

J. Phys. Chem. C, Vol. 114, No. 50, 2010

Figure 3. XRD profiles of both dendrimers and dendrimer:PCBM blends at (a) low angles and (b) high angles.

angles (Figure 3a), no ordering is seen in the neat film of 1, confirming that the bowl-shaped structure predicted from DFT calculations hinders dendrimer ordering. The neat film of 2 clearly forms ordered domains with a large diffraction peak at 2θ ) 2.6°, which corresponds to a d-spacing of 3.4 nm. There are also higher angle peaks in 2 at 3.5° and 5.5°, which

Rance et al. correspond to d-spacings of 2.5 and 1.6 nm, respectively. These observations verify that the planar structure of 2 previously predicted by DFT calculations leads to order in thin films. Also the d-spacing suggests 2 has an edge-on alignment similar to what has been seen in other small molecules.24 Figure 3a also shows XRD results from dendrimer:PCBM (1:4) bulk heterojunction films at low scattering angles. No ordering is seen in the blend of 1:PCBM, but a small peak is still seen in 2:PCBM at 2.4°, indicating that partial dendrimer ordering is retained despite the heavy PCBM loading. The scattering at higher angles is shown in Figure 3b along with a drop-cast PCBM film. Both dendrimers show no ordering in neat films throughout the wide angle range. The 2:PCBM blend does show a number of peaks that correspond to peaks seen in the PCBM film. This suggests that the self-ordering properties of 2 increase the degree of phase separation in blends and promote crystallization of the PCBM-rich domains. No peaks are seen in the 1:PCBM (1:4) blend despite the equivalent PCBM loading, indicating an amorphous blend structure with a lesser degree of phase separation than is the case in 2:PCBM. The differences in the film morphology of the two dendrimers have an impact on the photoconductivity of pure films and blends with PCBM, as measured by TRMC. Representative transients of the neat films and of bulk heterojunction dendrimer: PCBM (1:4) blends are shown in Figure 4. The photoconductance at the same absorbed photon flux of the neat films of 1 and 2, shown in Figure 4a, is mainly attributed to mobile holes as the thiophene-based dendrons are likely to show the same preferential transport for holes as has been seen in thiophenebased polymers. Also, calculations of the reorganization energy of dendrimer 2 have shown that the reorganization energy for hole transfer is a factor of 7 lower than that for electron transfer, which leads us to conclude that the TRMC photoconductance is dominated by holes in the pristine film of 2.16 The yield for carrier generation, φ, is expected to be small in both dendrimers compared to the dendrimer:PCBM blends. For comparison, in thiophene-based polymers a yield of ca. 2% was determined for neat polymer films while it is ca. 75% in blends with PCBM.22,25 The 5-fold increase in the photoconductance of pristine 2 can predominantly be attributed to two factors: (1)

Figure 4. Photoconductance transients of (a) neat dendrimers and (b) dendrimer:PCBM blends at the same absorbed photon flux (Ι0FA ) 1.5 × 1015 photons/cm2/pulse for (a) and I0FA ) 9 × 1012 photons/cm2/pulse for (b)), and normalized photoconductance transients of (c) neat dendrimers and (d) dendrimer:PCBM blends.

Dendrimer:Fullerene Photovoltaic Devices

J. Phys. Chem. C, Vol. 114, No. 50, 2010 22273

Figure 5. The φ∑µ product for both neat dendrimers and dendrimer:PCBM blends as a function of photon flux.

the increased local order in dendrimer 2 and/or (2) the low reorganization energy for hole transfer in 2 compared to other nonsubstituted thiophene dendrimers.16 Currently we are unable to differentiate between these two mechanisms; however, it is clear that both arise from the planar structure of dendrimer 2, while the bowl-shaped structure of dendrimer 1 does not allow packing of the molecules in an ordered film, thereby reducing the coupling of π-orbitals and lowering the hole mobility that is measured by TRMC. Another factor that may contribute to the increase in ∆G of 2 is an increased carrier generation yield, φ, from the cyanobenzene core, consistent with the low photoluminescence quantum yield in this dendrimer.13 The magnitude of the TRMC signal for neat dendrimer 2 (shown in Figure 5) is very similar to that in regioregular poly(3hexylthiophene).25 Since transport in both systems occurs via an ordered array of oligothiophenes, a drastic change in φ would manifest itself in the magnitude of the signal, contrary to observations. A comparison of the normalized photoconductance decay for the neat films (Figure 4c) shows that the photoconductivity decays after the initial carrier generation are identical, indicating similar photocarrier lifetimes in the two dendrimers despite the higher overall photoconductivity in 2. In the dendrimer:PCBM 1:4 blends there is a larger difference in the photoconductance, shown in Figure 4b. At the same absorbed photon flux the 2:PCBM blend has a photoconductivity ca. 10 times higher than the 1:PCBM blend. The crystalline PCBM phase that is only present in 2:PCBM blends increases ∆G due to the high electron mobility in this phase, as has been reported previously.26 The photoconductivity signal from 1:PCBM includes contributions from both holes in 1 and electrons in PCBM, but poor phase separation present in this blend most likely reduces the electron mobility in the PCBM. When comparing the normalized photoconductivity decays of the blends (Figure 4d), the carrier lifetime is unchanged in 1:PCBM compared to the neat film of 1, while the carriers are longer lived in 2:PCBM. We attribute this to the higher degree of order of the dendrimer and fullerene phases in 2:PCBM. As has been observed in other systems, ordering may provide a pathway for separated carriers to escape the interface, thereby enhancing their lifetimes.27,28 Using eq 2, the product of the yield for carrier generation and the sum of electron and hole mobilities, φ∑µ, was calculated

from ∆GEOP for the neat dendrimer films and the dendrimer: PCBM blends and is plotted in Figure 5 versus absorbed photon flux. The decline seen in φ∑µ as the absorbed photon flux increases has been well documented in previous work and is attributed to higher-order loss mechanisms that set in at high light intensities.20,22 At low light intensities the contribution of these higher-order loss processes diminish, and the φ∑µ product becomes independent of light intensity. This regime is reached for the blends, but not for the pristine dendrimers due to their lower photoconductance signal that does not permit measurements at very low absorbed photon flux. Overall, there is an increase in the φ∑µ product in the cyanomodified dendrimer 2 in both neat and blend films. At the lowest absorbed photon flux in the neat 2 film the φ∑µ is 10-4 cm2/ (V s), which is similar to published values for poly(3hexylthiophene) of 3.5 × 10-4 cm2/(V s).22 While there are significant differences between 2 and poly(3-hexylthiophene), a comparison between them may still be instructive: both materials show order in the solid state and their electronically active components are conjugated segments composed of planar oligothiophenes. When comparing the neat films to the dendrimer:PCBM blends, the φ∑µ product is higher for the blends at all photon fluxes, indicating enhanced free-carrier production due to the presence of the acceptor, as expected.20,29 The magnitude of φ∑µ at low light intensities is higher by ca. a factor of 10 in 2:PCBM than in 1:PCBM. We attribute this large change in φ∑µ and the longer-lived carriers seen in the decay transients to the morphology of the blend films, as already mentioned. The absence of order in dendrimer 1 indicates a completely amorphous structure that may involve fine mixing of the dendrimer and fullerene in the blend, increasing recombination due to the closer proximity of the hole in the dendrimer and the electron in the fullerene phase, consistent with the shorter lifetimes observed. The lower φ∑µ product for 1:PCBM is attributed to low carrier mobilities caused by the absence of order in either dendrimer and fullerene phase. In a recent report, Grzegorczyk et al. showed that ∑µ, measured by TRMC in poly(3-hexylthiophene):PCBM bulk heterojunctions, is dominated by the electron mobility in PCBM, which for highly ordered films has a minimum value of ca. 0.07 cm2/(V s).30 A value of >0.04 cm2/(V s) has also been measured for the ∑µ of PCBM powders using pulse radiolysis.26 For

22274

J. Phys. Chem. C, Vol. 114, No. 50, 2010

Rance et al.

Figure 6. Current density versus voltage (J-V) for 1:PCBM and 2:PCBM 1:4 blend devices.

Figure 7. EQE and absorbance of 1:PCBM and 2:PCBM (1:4) blends. (inset) Normalized EQE of 1:PCBM 1:1 and 1:4 devices.

comparison, the hole mobility in poly(3-hexylthiophene) has been quoted at 0.014 cm2/(V s).31 Assuming that ∑µ in 2:PCBM is dominated by carriers in the crystalline PCBM domains, and taking the value for their mobility to be in the range 0.04-0.07 cm2/(V s),30 the data from Figure 5 allow us to estimate the yield for free-carrier generation per photon absorbed, φ, to be in the range 0.28-0.49 and possibly even lower when the highest range of the ∑µ in PCBM crystallites (up to 0.3 cm2/ (V s))26 is considered. By contrast, the yield in poly(3hexylthiophene) with just 1% PCBM is 0.75 and increases in blends with higher PCBM content.20 We attribute the low yield for 2:PCBM to a combination of two factors: (1) confinement of the electron to the dendrimer core and (2) phase separation between 2 and PCBM that prevents the PCBM from accessing the dendrimer core. Calculations have shown that the excited state of 2 has significant charge-transfer character, with negative charge confined to the cyanobenzene core and positive charge delocalized over the conjugated oligothiophene arm.15 A similar “charge trapping” process at a dendrimer core has previously been identified for other dendritic architectures.32 Confinement of the electron to the core, combined with the physical separation between dendrimer and PCBM domains, may make the fullerene

inaccessible to the electron, thereby hampering efficient electron transfer and limiting φ. The above analysis to estimate φ is not feasible for 1:PCBM as in this case no ordered PCBM domains are observed and it is plausible that the mobility of carriers in that blend will be much lower than the value quoted for ordered PCBM. Indeed it has been shown that the ∑µ in disordered poly(3-hexylthiophene):PCBM bulk heterojunctions deposited from chloroform and not annealed can be as low as 0.004 cm2/(V s).30 In bulk heterojunction dendrimer:PCBM (1:4) devices, dendrimer 2 demonstrates improved J-V characteristics compared to 1, shown in Figure 6. A fairly high open circuit voltage (VOC) of 0.93 V is seen in 2:PCBM, which is higher than the 0.75 V seen in 1:PCBM. The similar highest occupied molecular orbital energy levels obtained in each dendrimer from cyclic voltammetry13 would predict similar maximum attainable VOC values in devices with PCBM.33 We attribute the VOC loss in 2:PCBM to increased recombination, consistent with the shorter carrier lifetimes seen by TRMC. A slightly improved short circuit current (JSC) is also seen in 2:PCBM, with the JSC increasing from 2.0 mA/cm2 in 1:PCBM to 2.5 mA/cm2 in 2:PCBM. However, the magnitude of the

Dendrimer:Fullerene Photovoltaic Devices difference in band gap between dendrimers 1 and 2 would predict a much larger difference in JSC. We attribute this to electron confinement in the cyanobenzene core as discussed above. The largest improvement in devices is seen in the fill factor, where the fill factor in 1:PCBM is only 28%, which is likely due to high carrier recombination. In 2:PCBM the fill factor increases to 47%, leading to an overall power conversion efficiency of 1.12%, compared to 0.40% in 1:PCBM. We attribute the higher fill factor to the improved morphology of the 2:PCBM bulk heterojunction. The EQE of each blend device is shown in Figure 7. The 1:PCBM device shows a peak corresponding to the primary absorption of the dendrimer, and a long tail at longer wavelengths that is attributed to contribution from PCBM due to the high loading ratio used. The slight shoulder seen at 560 nm in 1:PCBM is also attributed to PCBM, as a comparison of the normalized EQE of separate 1:1 and 1:4 blend devices of 1:PCBM, shown in the inset of Figure 7, which shows the reduction of that peak in the lower PCBM loaded sample. A broader photoresponse is seen in the 2:PCBM device, indicating that the short and long wavelength absorption regions of dendrimer 2 contribute almost equally to device performance. The peak EQE for 1:PCBM is 26% at 450 nm, while the highest EQE in 2:PCBM is 17% at 435 nm. The lower EQE in 2:PCBM is also consistent with the reduced φ due electron confinement in the core. The integrated photocurrents of 1:PCBM and 2:PCBM were 2.1 and 2.3 mA/cm2, respectively, verifying experimental JSC values. 4. Conclusions We have shown how a systematic modification of a thiophene dendrimer molecular system can modify the properties from the molecular to the device level. In particular, attachment of cyano groups to the dendrimer core created an intramolecular “push-pull” electronic structure that increased the overlap with the solar spectrum and planarized the molecule, leading to improved morphology both in pristine films and in blends with PCBM. The improved local order when the planar dendrimer was used resulted in longer photocarrier dynamics in the blends and a higher overall photoconductance signal. Power conversion efficiencies were greatly improved in 2:PCBM relative to 1:PCBM, due largely to increased fill factor and VOC stemming from the increased degree of order. The lower than expected JSC of the planar dendrimer is attributed to electron confinement in the cyanobenzene core that impedes transfer to PCBM. This implies that future push-pull designs for dendrimers may benefit from placing the electron-withdrawing moiety at the periphery of the molecule where electronic coupling to PCBM may be favored. Acknowledgment. This work was supported by the U.S. Department of Energy’s Solar Energy Technologies Program under Contract No. DE-AC36-08GO28308 with NREL. G.R. gratefully acknowledges the Division of Chemical Sciences, Geosciences, and Biosciences, Office of Basic Energy Sciences of the U.S. Department of Energy for funding the development of the transient microwave conductivity measurement. References and Notes (1) Liang, Y.; Xu, Z.; Xia, J.; Tsai, S.-T.; Wu, Y.; Li, G.; Ray, C.; Yu, L. For the Bright Future-Bulk Heterojunction Polymer Solar Cells with Power Conversion Efficiency of 7.4%. AdV. Mater. 2010, 22 (20), E135. (2) Ashraf, R. S.; Hoppe, H.; Shahid, M.; Gobsch, G.; Sensfuss, S.; Klemm, E. Synthesis and properties of fluorene-based polyheteroarylenes for photovoltaic devices. J. Polym. Sci., A: Polym. Chem. 2006, 44 (24), 6952–6961.

J. Phys. Chem. C, Vol. 114, No. 50, 2010 22275 (3) Brabec, C.; Winder, C.; Sariciftci, N.; Hummelen, J.; Dhanabalan, A.; van Hal, P.; Janssen, R. A low-bandgap semiconducting polymer for photovoltaic devices and infrared emitting diodes. AdV. Funct. Mater. 2002, 12 (10), 709–712. (4) Mayer, A. C.; Toney, M. F.; Scully, S. R.; Rivnay, J.; Brabec, C. J.; Scharber, M.; Koppe, M.; Heeney, M.; McCulloch, I.; McGehee, M. D. Bimolecular Crystals of Fullerenes in Conjugated Polymers and the Implications of Molecular Mixing for Solar Cells. AdV. Funct. Mater. 2009, 19 (8), 1173–1179. (5) Zhu, Z.; Waller, D.; Gaudiana, R.; Morana, M.; Muehlbacher, D.; Scharber, M.; Brabec, C. Panchromatic conjugated polymers containing alternating donor/acceptor units for photovoltaic applications. Macromolecules 2007, 40 (6), 1981–1986. (6) Tu, G.; Bilge, A.; Adamczyk, S.; Forster, M.; Heiderhoff, R.; Balk, L. J.; Muehlbacher, D.; Morana, M.; Koppe, M.; Scharber, M. C.; Choulis, S. A.; Brabec, C. J.; Scherf, U. The influence of interchain branches on solid state packing, hole mobility and photovoltaic properties of poly(3hexylthiophene) (P3HT). Macromol. Rapid Commun. 2007, 28 (17), 1781– 1785. (7) Peet, J.; Kim, J. Y.; Coates, N. E.; Ma, W. L.; Moses, D.; Heeger, A. J.; Bazan, G. C. Efficiency enhancement in low-bandgap polymer solar cells by processing with alkane dithiols. Nat. Mater. 2007, 6 (7), 497–500. (8) Ohkita, H.; Cook, S.; Astuti, Y.; Duffy, W.; Tierney, S.; Zhang, W.; Heeney, M.; Mcculloch, I.; Nelson, J.; Bradley, D. D. C.; Durrant, J. R. Charge Carrier Formation in Polythiophene/Fullerene Blend Films Studied by Transient Absorption Spectroscopy. J. Am. Chem. Soc. 2008, 130 (10), 3030–3042. (9) Walker, B.; Tamayo, A.; Dang, X.-D.; Zalar, P.; Seo, J. H.; Garcia, A.; Tantiwiwat, M.; Nguyen, T.-Q. Nanoscale Phase Separation and High Photovoltaic Efficiency in Solution-Processed, Small-Molecule Bulk Heterojunction Solar Cells. AdV. Funct. Mater. 2009, 19 (19), 3063–3069. (10) Lloyd, M. T.; Anthony, J. E.; Malliaras, G. G. Photovoltaics from soluble small molecules. Mater. Today 2007, 10 (11), 34–41. (11) Ma, C.-Q.; Fonrodona, M.; Schikora, M. C.; Wienk, M. M.; Janssen, R. A. J.; Ba¨uerle, P. Solution-Processed Bulk-Heterojunction Solar Cells Based on Monodisperse Dendritic Oligothiophenes. AdV. Funct. Mater. 2008, 18 (20), 3323–3331. (12) Mitchell, W.; Kopidakis, N.; Rumbles, G.; Ginley, D.; Shaheen, S. The synthesis and properties of solution processable phenyl cored thiophene dendrimers. J. Mater. Chem. 2005, 15 (42), 4518–4528. (13) Rupert, B. L.; Mitchell, W. J.; Ferguson, A. J.; Koese, M. E.; Rance, W. L.; Rumbles, G.; Ginley, D. S.; Shaheen, S. E.; Kopidakis, N. Lowbandgap thiophene dendrimers for improved light harvesting. J. Mater. Chem. 2009, 19 (30), 5311–5324. (14) Mitchell, W. J.; Ferguson, A. J.; Kose, M. E.; Rupert, B. L.; Ginley, D. S.; Rumbles, G.; Shaheen, S. E.; Kopidakis, N. Structure-Dependent Photophysics of First-Generation Phenyl-Cored Thiophene Dendrimers. Chem. Mater. 2009, 21 (2), 287–297. (15) Kose, M.; Mitchell, W. J.; Kopidakis, N.; Chang, C. H.; Shaheen, S. E.; Kim, K.; Rumbles, G. Theoretical studies on conjugated phenylcored thiophene dendrimers for photovoltaic applications. J. Am. Chem. Soc. 2007, 129 (46), 14257–14270. (16) Kose, M. E.; Graf, P.; Kopidakis, N.; Shaheen, S. E.; Kim, K.; Rumbles, G. Exciton Migration in Conjugated Dendrimers: A Joint Experimental and Theoretical Study. ChemPhysChem 2009, 10 (18), 3285– 3294. (17) Kose, M. E.; Long, H.; Kim, K.; Graf, P.; Ginley, D. Charge Transport Simulations in Conjugated Dendrimers. J. Phys. Chem. A 2010, 114 (12), 4388–4393. (18) Kopidakis, N.; Mitchell, W. J.; Van De Lagemaat, J.; Ginley, D. S.; Rumbles, G.; Shaheen, S. E.; Rance, W. L. Bulk heterojunction organic photovoltaic devices based on phenyl-cored thiophene dendrimers. Appl. Phys. Lett. 2006, 89 (10), 103524. (19) Kroeze, J.; Savenije, T.; Vermeulen, M.; Warman, J. Contactless determination of the photoconductivity action spectrum, exciton diffusion length, and charge separation efficiency in polythiophene-sensitized TiO2 bilayers. J. Phys. Chem. B 2003, 107 (31), 7696–7705. (20) Ferguson, A. J.; Kopidakis, N.; Shaheen, S. E.; Rumbles, G. Quenching of Excitons by Holes in Poly(3-hexylthiophene) Films. J. Phys. Chem. C 2008, 112 (26), 9865–9871. (21) Savenije, T.; de Haas, M.; Warman, J. The yield and mobility of charge carriers in smooth and nanoporous TiO2 films. Z. Phys. Chem. 1999, 212, 201–206. (22) Dicker, G.; de Haas, M.; Siebbeles, L. Signature of exciton annihilation in the photoconductance of regioregular poly(3-hexylthiophene). Phys. ReV. B 2005, 71 (15), 155204. (23) Gaab, K.; Thompson, A.; Xu, J.; Martinez, T.; Bardeen, C. Metaconjugation and excited-state coupling in phenylacetylene dendrimers. J. Am. Chem. Soc. 2003, 125 (31), 9288–9289. (24) Ponomarenko, S. A.; Kirchmeyer, S.; Elschner, A.; Huisman, B.H.; Karbach, A.; Drechsler, D. Star-Shaped Oligothiophenes for SolutionProcessible Organic Field-Effect Transistors. AdV. Funct. Mater. 2003, 13 (8), 591–596.

22276

J. Phys. Chem. C, Vol. 114, No. 50, 2010

(25) Dicker, G.; de Haas, M.; Siebbeles, L.; Warman, J. Electrodeless time-resolved microwave conductivity study of charge-carrier photogeneration in regioregular poly(3-hexylthiophene) thin films. Phys. ReV. B 2004, 70 (4), 045203. (26) De Haas, M. P.; Warman, J. M.; Anthopoulos, T. D.; de Leeuw, D. M. The mobility and decay kinetics of charge carriers in pulse-ionized microcrystalline PCBM powder. AdV. Funct. Mater. 2006, 16 (17), 2274– 2280. (27) Grzegorczyk, W. J.; Savenije, T. J.; Heeney, M.; Tierney, S.; McCulloch, I.; van Bavel, S.; Siebbeles, L. D. A. Relationship between Film Morphology, Optical, and Conductive Properties of Poly(thienothiophene): [6,6]-Phenyl C-61-Butyric Acid Methyl Ester Bulk Heterojunctions. J. Phys. Chem. C 2008, 112 (41), 15973–15979. (28) Dayal, S.; Reese, M. O.; Ferguson, A. J.; Ginley, D. S.; Rumbles, G.; Kopidakis, N., AdV. Funct. Mater. 2010, in press. (29) Savenije, T. J.; Kroeze, J. E.; Wienk, M. M.; Kroon, J. M.; Warman, J. M. Mobility and decay kinetics of charge carriers in photoexcited PCBM/ PPV blends. Phys. ReV. B 2004, 69 (15), 1–11. (30) Grzegorczyk, W. J.; Savenije, T. J.; Dykstra, T. E.; Piris, J.; Schins, J. M.; Siebbeles, L. D. A. Temperature-Independent Charge Carrier

Rance et al. Photogeneration in P3HT-PCBM Blends with Different Morphology. J. Phys. Chem. C 2010, 114 (11), 5182–5186. (31) Dicker, G.; de Haas, M.; Warman, J.; de Leeuw, D.; Siebbeles, L. The disperse charge-carrier kinetics in regioregular poly(3-hexylthiophene. J. Phys. Chem. B 2004, 108 (46), 17818–17824. (32) Nantalaksakul, A.; Mueller, A.; Klaikherd, A.; Bardeen, C. J.; Thayumanavan, S. Dendritic and Linear Macromolecular Architectures for Photovoltaics: A Photoinduced Charge Transfer Investigation. J. Am. Chem. Soc. 2009, 131 (7), 2727–2738. (33) Veldman, D.; Meskers, S. C. J.; Janssen, R. A. J. The Energy of Charge-Transfer States in Electron Donor-Acceptor Blends: Insight into the Energy Losses in Organic Solar Cells. AdV. Funct. Mater. 2009, 19 (12), 1939–1948. (34) Scharber, M.; Wuhlbacher, D.; Koppe, M.; Denk, P.; Waldauf, C.; Heeger, A.; Brabec, C. Design rules for donors in bulk-heterojunction solar cellssTowards 10% energy-conversion efficiency. AdV. Mater. 2006, 18 (6), 789.

JP106850F