The characterization of a benzo[2,1-b;3,4-b′]dithiophene containing conjugated polymer (PBTT) is demonstrated, with regard to its photovoltaic performance. X-ray diffraction measurements reveal that the thermal treatment results in an increased crystallinity within the PBTT:PCBM network and subsequent spatial rearrangement in the film. Upon stepwise annealing, the PBTT-based bulk-heterojunction solar cells show an overall conversion efficiency of 2.7 % under 1 sun light illumination. The photovoltaic devices based on PBTT show a high efficiency, maintained over one month. All these aspects suggest that the use of self-organizable materials is an efficient approach for high-performance photovoltaic applications.
During the last few years, organic solar cells have attracted a great deal of interest because of their advantages related to cost-effective manufacturing and flexible applications.1–4 Among these cells, bulk heterojunction solar cells, in which the hole-transporting conjugated polymers (donors) and electron-transporting molecules (acceptors), such as fullerenes, are blended in the active organic layer, are regarded as highly efficient systems for rapid exciton dissociation and charge transport. Upon irradiation with light, Coulomb-correlated electron–hole pairs (i.e., excitons) are generated. In order to achieve pronounced power conversion efficiencies these excitons must dissociate into electrons and holes with a high yield; furthermore, an efficient transport of charge carriers to their respective electrodes, which can be realized by forming heterojunctions of materials with different electron affinities, is also required. At this moment, the best polymer solar cells exhibit power conversion efficiencies in the range of 5–6 %.5 Attempts to achieve higher photovoltaic performance have included extending the absorption to longer wavelengths,6 increasing the absorption coefficients,7 and optimizing the charge separation and transport within the bulk heterojunction network.8 However, the concept of using self-organization to improve the efficiency of bulk heterojunction solar cells has been particularly appealing ever since efficient photovoltaic cells were fabricated from a hexabenzocoronene derivative (HBC) and perylenediimide.9 Later on, studies on photovoltaic cells fabricated from self-assembling materials have been abundant.10
Herein, we report the conjugated polymer poly[2,7-(benzo[2,1-b;3,4-b′]dithiophene)-alt-2,2′-(3,3′-didodecyl-5,5′-bithiophenyl)] (PBTT) (Figure 1), which is a rigid copolymer utilizing an unsubstituted benzodithiophene moiety to enforce self-organization of the polymer chain. In this structure architecture, the benzodithiophene moiety affords a free volume for the two side chains (dodecyl) on the thiophene units, which potentially increases the π–π stacking between polymer backbones. Thus, this polymer demonstrates high charge-carrier mobilities up to 0.5 cm2 V−1 s−1 on plastic substrates.11 As shown in Figure 1 b, an absorption maximum of PBTT occurs at 540 nm in a dichlorobenzene solution, while a PBTT film exhibits a broader absorption. In addition, the absorption of the solution mixture of PBTT and [6,6]-phenyl-C71-butyric acid methyl ester (PCBM) overlaps with nearly the whole visible region from 350 to 650 nm, which renders the blend as a good light absorber for highly efficient photovoltaic devices.
Results and Discussion
The molecular energy levels of PBTT were determined by cyclic voltammetry (CV) in a three-electrode cell with a working electrode of indium tin oxide (ITO) glass, a silver quasireference electrode (AgQRE, calibrated against the Fc/Fc+ redox couple), and a Pt counter electrode. A highest occupied molecular orbital (HOMO) level of −5.2 eV was calculated from the onset of the oxidation peak. With the optical bandgap of 2.0 eV evaluated from the onset of the absorption spectrum, the lowest unoccupied molecular orbital (LUMO) level was estimated to be −3.2 eV. For an efficient solar cell, the HOMO and the LUMO energies of donors should be more positive than those of acceptors in order to guarantee a ready charge transfer in the device.12 Therefore, PCBM, with an experimentally determined LUMO level of −4.3 eV and a HOMO level of −6.1 eV, was chosen as acceptor. In the pair, a 0.9 eV offset of HOMO levels and a 1.1 eV LUMO-level difference support an efficient hole migration from PCBM excited states to PBTT as well as a favorable electron transfer from PBTT to PCBM.
In bulk heterojunction solar cells, the overall conversion efficiency is strongly related to the interpenetrating nanoscale network morphology of the polymer and fullerene blend.13 In addition to other parameters, such as choice of solvents, donor/acceptor (D/A) ratio, casting methods, and solution concentrations, the chosen thermal annealing procedure has a significant influence on the film morphology. Here, the annealing effect was monitored by X-ray diffraction (XRD) measurements through which the related lattice constant d can be calculated using Bragg’s law, as shown in Equation (1).((1))
where λ=0.154 nm is the wavelength of incident beam, 2θ is the angle between incident and scattered X-ray wave vectors, and n is the interference order. Figure 2 displays the XRD results of a PBTT:PCBM (1:2) blend film before and after heating. An increase in the crystallinity of PBTT can be clearly seen in the XRD spectra. The annealed sample shows an increased intensity in the peak at 2θ=4.14°, which corresponds to the lamellar spacing between polymer chains. According to Equation (1), d=2.07 nm was calculated. In the case of the untreated sample, only very weak diffraction peaks were recorded. This result indicates that a spatial rearrangement of the polymer chains leads to a stronger interchain interaction upon heat treatment. The increased crystallinity within the network and subsequent spatial rearrangement in the film indicates that the thermal annealing process should have a dramatic influence on the photovoltaic performance by facilitating charge transport to the electrodes.13b
Figure 1 c displays the device architecture. The photovoltaic cells were fabricated from dichlorobenzene solutions of PBTT and PCBM blended in different weight ratios of 1:1, 1:2, and 1:3, with or without 1,8-octanedithiol (24 mg mL−1) as the processing additive.14 The best performance was obtained from the devices with a D/A ratio of 1:2, with 1,8-octanedithiol included (Table 1 and Figure 3). Because the devices show efficiency improvements at higher fullerene ratios (1:2) in comparison to the optimum ratio near 1:1 in some systems, such as P3HT:PCBM, it is likely that in this system PCBM is intercalated between the side chains of PBTT, as discussed by McGehee.15 The impact of additives can be seen in the XRD data of PBTT:PCBM (1:2) films with or without processing additive, where increased intensity in the sample with 1,8-octanedithiol included is shown (Figure 2 b).
Table 1. Photovoltaic performance of devices in different D/A ratios with or without 1,8-octanedithiol.
ISC [mA cm−2][b]
[a] D/A ratios are referred to as PBTT:PCBM, in weight ratio. [b] Devices were tested under AM 1.5 simulated solar irradiation.
The post-deposition annealing process was carried out on the same sample every two minutes in a stepwise fashion, resulting in a dramatically enhanced device performance with an increase in the open-circuit voltage (VOC), short-circuit current (ISC), and fill factor (FF) as well as a twofold increase in power conversion efficiency to 2.68 % (Figure 4). After two minutes of annealing at 120 °C, an increase in the VOC from 0.46 V to 0.68 V was observed. An enhancement in the ISC and the FF values, 6.94 to 7.61 mA cm−2 and 0.31 to 0.43, respectively, was also noted. Furthermore, the power conversion efficiency was boosted from 1.0 % to 2.3 %. Further step-by-step annealing every two minutes resulted in a gradually enhanced ISC but not in a significant change in VOC and FF. These data correspond to a slowly increased overall efficiency with the highest value of 2.7 % after a total thermal treatment period of 14 min, with an ISC of 8.53 mA cm−2, a VOC of 0.67 V, and an FF of 0.46. In addition, we noted that the experimentally obtained VOC value of 0.68 V fitted well with the expected theoretical value, which is in the range of 0.6–0.7 V.16 Continued thermal treatment after this 14 min period resulted in a trend toward slightly decreasing ISC and FF values, while the VOC value remained unchanged; thus, the same trend in overall efficiency is observed. It should be noted that at the beginning of our study on PBTT-based solar cells, PCBM was chosen as the acceptor due to its similar electronic properties to PCBM.6d The PBTT:PCBM cells, however, gave an overall efficiency of 1.2 % with a short-circuit current of only 4.7 mA cm−2, around half of those in the PBTT:PCBM cells. This result is most likely due to the higher absorption coefficients of PCBM in the visible region.17
To elucidate why the annealing improved the device performance, we first investigated the atomic force microscopy (AFM) topography of films cast from PBTT:PCBM before and after thermal treatment (Figure 5). In good agreement with the XRD results, a finer domain size and better phase-segregated bulk heterojunction morphology are formed in the blend film (Figure 5 b) due to PBTT’s enhanced aggregation properties. The morphology of the film shown in Figure 5 b is believed to be favorable for the charge transport of holes and electrons and stronger interfacial dipole moments in the corresponding separated continuous percolation pathways.9, 10a, b
Another approach to investigate the reasons for the enhanced efficiency after thermal treatment of these devices is to observe the change of shunt resistance or series resistance in the device upon heating.8a In an efficient cell, series resistance is expected to be as small and shunt resistance to be as large as possible. The shunt resistance RSH is due to the recombination of charge carriers near the dissociation site (e.g., D/A interface), while the series resistance RS considers conductivity (i.e., mobility of the specific charge carrier in the respective transport medium). Based upon the equivalent circuit of a photovoltaic cell, the current produced by the solar cell is equal to that produced by the current source, minus that which flows through the diode and that which flows through the shunt resistor. Further, by the Shockley diode equation and Ohm’s law, the current density–voltage characteristics can be described by Equation (2).((2))
where IPH is the photocurrent, ID is the diode current, ISH is the shunt current, I0 is the dark current, q is the electron charge, n is the diode ideality factor, k is Boltzmann’s constant, V is the voltage produced by the device, RS is the series resistance, RSH is the shunt resistance, and T is the absolute temperature.
According to Equation (2) and the values calculated from the I–V curves before and after thermal treatment (Figure 6 a), we determined that the shunt resistance increases from an initial value of RSH=94 Ω cm2 to RSH=288 Ω cm2 after annealing at 120 °C, while RS was observed to be similar before and after annealing. This augmented shunt resistance most likely indicates less recombination of charge carriers near the donor/acceptor interface in the bulk heterojunction films after thermal annealing and, hence, an increased power conversion efficiency of the devices.
As shown in Figure 6 b, a large increase in terms of external quantum efficiency (EQE) between 390 nm to 700 nm was achieved upon annealing, which can possibly be attributed to the spatial rearrangement of the polymer chains.18
We have demonstrated the characterization of an unsubstituted benzodithiophene-based conjugated polymer, PBTT, with regard to its photovoltaic performance. The significant effect of thermal annealing on film aggregation has been utilized as a way to optimize the heterojunction solar cell morphology favoring exciton dissociation and charge transport, which has been proven by XRD and AFM investigations of the film. It should be noted that the bulk heterojunction solar cells based on PBTT as donor and PCBM as acceptor showed a significant twofold increase in efficiency after only two minutes’ annealing. Further stepwise post-deposition annealing gave an optimal power-conversion efficiency of 2.7 % under AM1.5 simulated solar illumination of 1000 W m−2, with a short-circuit current ISC of 8.53 mA cm−2, an open-circuit voltage VOC of 0.67 V, and a fill factor FF of 0.46. The overall efficiency decreased by only 5 % over a period of 32 days, which qualifies the photovoltaic device based on PBTT as a potential source of renewable energy. Interestingly, a similar polymer poly(2,5-bis(3-tetradecylthiophen-2-yl)thieno[3,2-b]thiophene (pBTTT) published by McGehee showed a lower photovoltaic performance after annealing.19 We note that with larger π-spacer between two thiophene units, PBTT exhibits better photovoltaic performance. With these aspects, it suggests to us the introduction of longer unsubstituted planar π-spacers into the thiophene copolymer in order to obtain stronger self-organization ability in the film and better phase separation ability in the mixture with acceptor molecules. This promising photovoltaic system of PBTT and PCBM suggests that the use of self-organizable materials is an efficient approach for high-performance photovoltaic applications.
Device Fabrication and Testing: The patterned ITO substrates were first cleaned with acetone and isopropyl alcohol in an ultrasonic bath, followed by a cleaning treatment with oxygen plasma for 10 min. Subsequently, a conducting PEDOT:PSS layer with a thickness of about 40 nm (Clevios P) was spin-cast (5000 rpm) from aqueous solution. The substrates were dried at 150 °C for 10 min in air, and then moved into a glove-box for spin-casting of the photoactive layer. The dichlorobenzene solution comprising PBTT (8 mg mL−1) and PCBM (16 mg mL−1), with 1,8-octanedithiol (24 mg mL−1) as a processing additive, was then spin-cast at 700 rpm on top of the PEDOT:PSS layer. A silver layer (ca. 100 nm) was subsequently evaporated through a mask onto the surface to form the cathode. Post-deposition annealing was carried out in the glove-box every two minutes in a stepwise fashion. The effective areas of the devices were about 6 mm2, as defined by the overlap between the etched ITO and the top electrode. Accurate values of these areas were determined using a microscope for subsequent calculations. Incident light was focused on the effective area of each device through a lens. Current–wavelength curves were recorded with a Keithley 236 Source Measure Unit. A tungsten halogen lamp was employed as a light source, supplying monochromatic light from 300 to 700 nm through a TRIAX 180 monochromator. The incident light intensity was determined using a calibrated silicon diode. The maximum intensity was 1 W m−2 at ca. 600 nm. Solar light was obtained from a solar simulator (Lichttechnik, Germany) using a 575 W metal halide lamp, in combination with an ODF Filter, to produce a spectral distribution close to the global radiation AM1.5 G. The light intensity was adjusted to 1000 W m−2.
Financial support by the Dutch Polymer Institute (DPI) is gratefully acknowledged. We also thank Ashlan Musante for her help in preparing this manuscript.