High performance bulk-heterojunction organic solar cells fabricated with non-halogenated solvent processing

Organic Electronics 12 (2011) 1465–1470

Contents lists available at ScienceDirect

Organic Electronics journal homepage: www.elsevier.com/locate/orgel

High performance bulk-heterojunction organic solar cells fabricated with non-halogenated solvent processing Choong-Do Park a,b, Trever A. Fleetham b, Jian Li b,⇑, Bryan D. Vogt b,⇑ a b

Department of Chemistry and Biochemistry, Arizona State University, Tempe, AZ, USA Flexible Display Center, Arizona State University, Tempe, AZ, USA

a r t i c l e

i n f o

Article history: Received 12 March 2011 Received in revised form 23 May 2011 Accepted 25 May 2011 Available online 12 June 2011 Keywords: Bulk heterojunction Photovoltaic Hansen solubility parameter Mesitylene Acetophenone

a b s t r a c t High efficiency bulk heterojunction (BHJ) organic solar cells based upon poly(3-hexylthiophene) (P3HT)-[6,6]-phenyl-C61-butyric acid methyl ester (PCBM) are typically solution processed using halogenated solvents like dichlorobenzene (DCB). In this report, we tune the quality of solvent systems using more environmentally friendly, halogen-free solvents to mimic that of DCB based on Hansen solubility parameters (HSPs). A mixture of acetophenone (AP) and mesitylene (MS) can nearly match the HSPs for DCB. Solar cells fabricated using this solvent mixture can exhibit power conversion efficiency (PCE) comparable to that from DCB with a similar surface morphology. It is critical to control the drying process in the mixed solvent system due to significant volatility difference between AP and MS in order to achieve comparable morphology and PCE. This report illustrates a route to fabricate high efficiency BHJ solar cells from halogen-free solvent systems that could eliminate potential hurdles for manufacturing scale-up of organic BHJ solar cells. Ó 2011 Elsevier B.V. All rights reserved.

1. Introduction Since initial reports of bulk-heterojunction (BHJ) organic solar cells [1,2], tremendous efforts have been focused on improving the device performance through processing control and new materials development [3,4]. The workhorse system for these BHJ cells has been poly(3-hexylthiophene) (P3HT)-[6,6]-phenyl-C61-butyric acid methyl ester (PCBM) blends [1]. Processing modifications such as controlling the drying by ‘solvent annealing’ [5], thermal annealing [6] and solvent additives [7] or mixtures [8] have been utilized to significantly enhance performance of these BHJ cells. However, halogenated solvents, typically dichlorobenzene (DCB), are used in the solution processing of these devices due to its superior performance in comparison to any single halogen-free solvent [9]. The device per⇑ Corresponding authors. Address: Flexible Display Center, 7700 S. River Pkwy, Tempe, AZ 85284, USA. Tel.: +1 480 727 8631 (B.D. Vogt), tel.: +1 480 727 8938 (J. Li). E-mail addresses: [email protected] (J. Li), [email protected] (B.D. Vogt). 1566-1199/$ - see front matter Ó 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.orgel.2011.05.020

formance depends upon the kinetically arrested morphology of the blend, which develops during film formation [10]. From the prospective of manufacturing, these halogenated solvents are generally not acceptable due to cost, safety, and environmental concerns; the requirement for halogenated solvents could act to limit the implementation of high efficiency organic BHJ solar cells into largescale production. In view of this desire to eliminate halogenated solvents, an alterative processing route based upon dispersion of pre-formed P3HT nanocrystals in p-xylene has been shown to exhibit reasonable performance by forming a desired morphology in the P3HT/PCBM film [11]. However, this processing method is complicated with repeated filtration steps to isolate the P3HT nanocrystals prior to the device fabrication. As the morphology is dictated by the film formation process and is one of key factors governing device performance, the thermodynamics of the solution and kinetics of drying are key factors in the film formation process and ultimately the device performance. Recently, Loo and coworkers have shown that modification of the solution thermodynamics by

1466

C.-D. Park et al. / Organic Electronics 12 (2011) 1465–1470

addition of non-volatile additives to DCB can significantly improve the performance of P3HT:PCBM devices [7]. Thus, controlling the interactions between P3HT, PCBM, and solvent during the drying process appears to be an effective route to modulating device performance and could potentially enable fabrication of high performance devices without the use of halogenated solvents. The use of solvent mixtures of differing compositions provides one facile route to modulate the solvent quality. In an attempt to rationally select candidate solvent mixtures, we have utilized Hansen solubility parameters (HSP), which describe the total cohesion energy, E by three contributions: the dispersion interactions, Ed, permanent dipolepermanent dipole molecular interactions, Ep, and the hydrogen bonding molecular interactions, Eh. This energy is directly related to the solubility parameters (d), which are the square root of the cohesion energy density (E/V) [12]

d2 ¼ d2d þ d2p þ d2h

ð1Þ

The application of solubility parameters have been utilized in understanding the solubility of polymers [13] and pharmaceuticals [14] in solvent mixtures. Fig. 1 illustrates the HSP for a number of different solvents. The solvents that can dissolve the P3HT and PCBM mixture (chloroform [CHCl3], ortho-dichlorobenzene [o-DCB], chlorobenzene [CB], biphenyl, mesitylene [MS], carbon tetrachloride [CCl4], diethylbenzene, ethylbenzene, toluene and xylene) are grouped together in HSP space that define the interaction radius (R0) wherein solvents dissolve the polymer and solvents outside of the volume defined by R0 will not dissolve the polymer. As DCB tends to yield devices with better performance, a mixture that matches its solubility parameters (dd = 19.2, dp = 6.3, dh = 3.3 MPa1/2) would be desired based upon these solubility arguments. As illustrated by the line in Fig. 1, a mixture of MS and AP can closely mimic DCB. In this work, we describe the use of MS/AP mixtures to fabricate BHJ solar cells based upon P3HT and PCBM blends. With proper choice of relative MS/AP composition, device efficiencies approaching that obtained from those cast from DCB can be obtained.

2. Experimental details 2.1. Materials and device fabrication Bulk-heterojunction solar cells based on P3HT/PCBM blend were fabricated in the following simple device stack structure: Indium tin oxide(ITO)/poly(3,4-ethylenedioxythiophene):poly(styrenesulfonate)(PEDOT:PSS)/P3HT:PCBM/ bathocuproine(BCP)/Al. Regioregular P3HT (Rieke Metals, 4002E) and PCBM (99.5%, Solenne) were dissolved in a 1:1 ratio to 1.0 wt.% in mixtures of MS (99.0%, Sigma–Aldrich) and AP (99%, Sigma–Aldrich) by heating and stirring for 2 h at 75 °C in a nitrogen atmosphere. These solutions were then cooled slowly to ambient temperature. For comparison devices, the P3HT:PCBM blend (1.5 wt.%, 1:1 weight ratio) was dissolved in DCB (99%, Sigma–Aldrich) and stirred for 24 h at 40 °C. Before the fabrication of the devices, the substrates were cleaned by scrubbing with detergent and ultrasonic treatment in deionized water, acetone and isopropyl alcohol sequentially and UV-ozone (Jelight model 42) treatment for 40 min. The cleaned ITO-coated glass substrates were modified by spin-coating a thin layer (40 nm) of PEDOT:PSS (Clevios, 4083) and cured at 200 °C for 30 min. The polymer blended films were spin-coated in a nitrogen-filled glove box onto the PEDOT:PSS modified substrate. The as-cast films were dried overnight in nitrogen atmosphere at ambient temperature. After drying, the samples were thermally annealed at 140 °C for 30 min. Finally, BCP (14 nm) and Al cathode (100 nm) layers were thermally deposited on top of the active layer and each device had an active area of 0.04 cm2. 2.2. Device characterization The current density–voltage characteristics were measured using a solar simulator (Newport, 150 W) with an AM 1.5G filter in a nitrogen filled glovebox at room temperature. The intensity of illumination was set for 100 mW/cm2 using a Si reference Hamamatsu cell (Model C24 S1787-04) calibrated by the National Renewable Energy Laboratory (NREL), USA. The spectral mismatch of the solar simulator was estimated to be less than 10%. External quantum efficiency data were measured with monochromatic light by varying the excitation wavelength from 300 to 850 nm at intervals of 10 nm, under a white bias light of 100 mW cm2 (Optronics Laboratories, Inc.). The surface morphology of the thin films was measured using atomic force microscopy (Park Systems XE-150) operating in non-contact mode with 2 lm  2 lm scan size. The film thickness was measured using a Variable Angle Spectroscopic Ellipsometer (VASE M-2000, J.A. Woollam Co.). 3. Results and discussion

Fig. 1. Hansen solubility parameters (HSP) for selected solvents and solvent mixtures. Solvents in red are considered as good solvents for P3HT/PCBM blend and in blue are non-solvents. (For interpretation of the references in colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 2 shows the device performance with the use of pure MS and various MS/AP solvent mixtures. The J–V curves show a significant increase in the magnitude of short circuit current (Jsc) with the addition of AP to the MS solution. However, further increase of the AP concen-

C.-D. Park et al. / Organic Electronics 12 (2011) 1465–1470

tration from 20 vol.% to 30 vol.% lead to a negligible increase in Jsc and a small decrease in open-circuit voltage (Voc). These changes in performance can be quantified by comparing the power conversion efficiency (PCE) of the devices as illustrated in Fig. 2b. The device fabricated from the 80–20 vol.% MS–AP mixture has a larger PCE than both 90–10 and 70–30 vol.% MS–AP mixtures. This result suggests that there is an optimal concentration for device performance. However, it may be noted that the composition of the solvent mixture (80–20 vol.% MS–AP) is far removed from the ideal composition (27–73 vol.% MS–AP with dd = 19.2, dp = 6.3, dh = 2.9 MPa1/2) that matches DCB based on the HSP selection rules. One explanation could be solvent electron density, which has been shown to be critical for surfactant solubility [15], is not explicitly addressed using HSP selection rules. Moreover, HSP does not account for any electrostatic interactions that will be important in this system. However, the normal boiling point difference between MS (165 °C) and AP (202 °C) results in significant concentrating of the AP during solvent evaporation in the film formation process. Thus during film formation, the solvent quality (from HSP arguments) would be evolving due to differential evaporation of the components in the solvent mixture. Thus, it would be expected that as the film forms the solvent quality will trend towards, and hopefully

Fig. 2. (a) The J–V characteristics under illumination of 100 mW cm2 (AM 1.5G) for devices fabricated from pure mesitylene and various MS/AP mixtures and (b) PCE values of devices fabricated from pure mesitylene and various MS/AP mixtures as a function of the amount of acetophenone (vol.%) in the solvent mixtures.

1467

approach, that of DCB, due to comparably low boiling temperature of MS. Nonetheless, this mixed solvent system of MS–AP does offer a route to improved device performance. Similarly, Bouman et al. [16] illustrated the potential of mixed solvent systems using chloroform–methanol mixtures and these mixtures have been shown to be effective in P3HT/ PCBM system [8], but the origins of the improvement are not clear and halogenated solvent was still utilized. In this case, the HSP for methanol (dd = 15.1, dp = 12.3, dh = 22.3 MPa1/2) would not suggest an improvement in the match of the solvent quality to a system with wellknown good performance such as DCB. Thus, the solubility parameters of the solvent are not be the only factors impacting performance; the wettability, solvent partitioning, solution aggregation and other factors will also be important. However, simply adding a poor solvent will not necessarily improve the performance. We have examined a mixture of MS with cyclohexanone; from the HSP arguments, it should be possible to form a mixture that is between chloroform and DCB (see Fig. 1). For all of these mixtures, the performance of the device was degraded in comparison to the pure MS. The overall solubility of the P3HT and PCBM decreased with the added cyclohexanone, but the cyclohexanone is also more volatile than MS. These two factors likely explain the degraded performance of this mixture as the solvent quality during film formation will evolve towards pure MS. Despite these differences, the performance of the device fabricated with 80–20 vol.% MS/AP mixture appears to be tremendously improved over typical devices fabricated from solely non-halogenated solvents. As this mixture provides the best performance with this solvent pair, it would be meaningful to compare its performances to analogous devices fabricated from the standard DCB. Fig. 3 illustrates the difference in device performance obtained for the P3HT:PCBM blend depending upon the casting solvent used. First, the J–V curve illustrates the significant difference in the short circuit current for the device fabricated from MS in comparison to DCB as would be expected [9]. The performance of the device fabricated from DCB can be approached by addition of‘ 20 vol.% AP to the MS with JSC increased from 4.7 mA cm2 to 8.9 mA cm2, RSA decreased from 3.4 O cm2 to 2.8 O cm2, and the fill factor is increased from 59% to 66%. These changes in J–V characteristics results in a significant increase in PCE from 1.6% to 3.4% for MS and MS–AP mixture. Additionally, the external quantum efficiency (EQE) increases from a maximum of 35% at 500 nm for the device from MS by a factor of two up to 69% for the mixed solvent as shown in Fig. 3b. These are significant improvements in the device performance, although the performance for the pure MS is known to be poor. Thus, we have also compared the performance of the device formed from the mixed solvent system to those cast from DCB. From Fig. 3a, the short circuit current and open circuit voltage are both slightly larger for the 85 nm thick device from DCB, but at similar thickness (65 nm) the device performance is nearly indistinguishable between the device from DCB and from the MS–AP mixture. Similarly, the EQE for the thicker DCB sample is slightly larger at wavelengths exceeding 500 nm.

1468

C.-D. Park et al. / Organic Electronics 12 (2011) 1465–1470

Fig. 3. (a) The J–V characteristics under illumination of 100 mW cm2 (AM 1.5G) and (b) external quantum efficiency (EQE) measurements data for devices fabricated from DCB, MS, and 80 vol.% MS-20 vol.% AP mixture.

Table 1 summarizes the performance characteristics for devices from the three different solvents: pure MS, 80–20 MS–AP and pure DCB. One important point to note is that the performance is highly reproducible with at least 6 devices tested for each to calculate the standard deviations in performance characteristics. For the mixed solvent system, the PCE varies between 3.2% and 3.8%; for comparison, the performance for the 85 nm thick devices from DCB varies between 3.6% and 4.3%. However, it is well known that the efficiency for P3HT:PCBM solar cells is thickness sensitive for thin film less than 150 nm thick [17], which is consistent with the decrease in performance for the 65 nm

thick devices from DCB. Difficulties in obtaining similar drying and spin coating conditions limited the ability to fully match the film thickness for the different solvent systems. However for thin films, the performance from the mixture closely mimics that obtained for DCB at similar active layer thickness. In addition to HSP arguments, one could also imagine that the relative solubility of P3HT and PCBM in each component of the solvent mixture would provide a route to control the structure. Recently, a mixture of pyridine and chlorobenzene, which good solvents for carboxylated polythiophenes and PCBM, respectively, has been utilized to fabricate OPV devices as the polymer is only marginally soluble in chlorobenzene [19]. For the MS–AP mixture, MS readily dissolves both P3HT and PCBM; the solubility of the components in AP is significantly less with PCBM soluble to >10 mg/mL and P3HT nearly insoluble in AP (<0.05 mg/mL). We hypothesize that the improved device performance of the MS–AP mixed solvent system originates predominately from two contributions: a decrease in the series resistance (refer to Table 1) and the change in morphology of the active layer caused by the solvent quality upon phase separation of the PCBM and P3HT. The slower evaporation rate of the high boiling point component (AP) during spin coating and drying process may also facilitate reorganization in the blend to a more appropriate morphology for charge transport and exciton dissociation. However, the changes in device performance can generally be attributed to the morphology of the P3HT:PCBM that is dependent on the processing conditions [8,18]. To begin to understand the morphology of the blends processed with different solvents, the surface after annealing is interrogated using AFM. Fig. 4 illustrates the topography and phase images of the P3HT:PCBM films spun from the different solvents. It is clear from these micrographs that the morphology of the blend formed from the MS–AP mixed solvent closely resembles that of the film cast from DCB. For the film cast from MS, the surface is rough with large domains clearly visible in the micrograph. Conversely, no large domains are observed for the film formed from the solvent mixture or from DCB. We attribute this morphological change in the mixed solvent system to the solvent quality evolution during drying due to the lower vapor pressure of the AP that changes the thermodynamic interactions in the solution. Both the intramolecular interactions and interactions between P3HT and PCBM are

Table 1 Summary of device performance for various BHJ solar cell devices in the work.

Thickness (nm) Jsc* (mA cm2) Voc (V) RSA (O cm2) FF (%) PCE (%) EQE (%) Jsc** (mA cm2) * **

MS

80–20 vol.% MS/AP mix.

DCB (thin)

DCB (thicker)

60 4.68 ± 0.76 0.56 ± 0.04 3.44 ± 0.17 58.9 ± 1.3 1.53 ± 0.16 3.45 ± 1.8 5.08 ± 0.68

60 8.88 ± 0.67 0.58 ± 0.02 2.81 ± 0.14 66.1 ± 1.4 3.38 ± 0.13 68.9 ± 1.3 9.04 ± 0.58

65 8.34 ± 0.26 0.59 ± 0.01 2.49 ± 0.12 65.3 ± 1.1 3.20 ± 0.10 62.6 ± 2.9 8.15 ± 0.15

85 9.15 ± 0.15 0.63 ± 0.01 2.73 ± 0.13 68.6 ± 2.0 3.92 ± 0.18 74.2 ± 1.2 10.4 ± 0.16

Jsc: current density measured under similated solar illumination. Jsc: current density calculated based on EQE measurement.

C.-D. Park et al. / Organic Electronics 12 (2011) 1465–1470

1469

an optimal composition in this solvent mixture; either increasing to 30 vol.% AP or decreasing to 10 vol.% AP results in a decrease in PCE. The overall device performance of the device fabricated from the non-halogenated MS–AP mixed solvent system can be comparable to that of a standard device from DCB. Further optimization of this work could enable the use of fully, non-halogenated solvent mixture systems for scale-up manufacturing of high performance BHJ organic solar cells. Acknowledgements We thank B. O’Brien and J. Torres for assistance with AFM measurements. We acknowledge financial support from the Army Research Laboratory through the Flexible Display Center at ASU under Cooperative Agreement W911NF-04-2-0005. J. Li would like to acknowledge the partial support of the Advanced Photovoltaics Center, and the National Science Foundation (CBET-0756148). References

Fig. 4. AFM topography and phase images of P3HT/PCBM blend films cast from (a,b) MS, (c,d) 80 vol.% MS-20 vol.% AP mixture, and (e,f) DCB (scale bar for topology in nm).

mitigated by the solvent initially and the solvent quality upon crystallization of one phase or phase separation will dictate how the morphology evolves. The film cast from DCB exhibits the smallest domain size of the three films, but the surface morphology for the solvent mixture is very similar to that for the DCB. This similarity in morphology is in good agreement with the J–V characteristics of the analogous devices. These results suggest that high performance organic BHJ solar cells can be fabricated with proper choice of halogen-free solvent pairs. The elimination of the use of halogenated solvents without compromising the fabrication condition could prove to be important for the potential commercialization of polymer based BHJ solar cells. Future work will focus on extending these concepts to other donor–acceptor pairs that have higher theoretical efficiencies, especially when one component is not soluble in typical solvents (i.e. DCB).

4. Conclusion In conclusion, we demonstrate the use of the non-halogenated solvent mixtures for the fabrication of BHJ organic solar cells. In comparison to a pure aromatic solvent, i.e. mesitylene, the use of a mixture containing 20 vol.% AP in MS results in more homogeneous morphology of the P3HT:PCBM film with a corresponding improvement in PCE from 1.6% to 3.4% and EQE from 35% to 69%. We find

[1] G. Yu, J. Gao, J.C. Hummelen, F. Wudl, A.J. Heeger, Polymer photovoltaic cells: enhanced efficiencies via a network of internal donor–acceptor heterojunctions, Science 270 (1995) 1789–1791. [2] N.S. Sariciftci, D. Braun, C. Zhang, V.I. Srdanov, A.J. Heeger, G. Stucky, F. Wudl, Semiconducting polymer-buckminsterfullerene heterojunctions: diodes, photodiodes, and photovoltaic cells, Applied Physics Letters 62 (1993) 585–587. [3] B.R. Saunders, M.L. Turner, Nanoparticle-polymer photovoltaic cells, Advances in Colloid and Interface Science 138 (2008) 1–23. [4] R. Kroon, M. Lenes, J.C. Hummelen, P.W.M. Blom, B. de Boer, Small bandgap polymers for organic solar cells (polymer material development in the last 5 years), Polymer Reviews 48 (2008) 531– 582. [5] G. Li, Y. Yao, H. Yang, V. Shrotriya, G. Yang, Y. Yang, ‘‘Solvent Annealing’’ effect in polymer solar cells based on poly(3hexylthiophene) and methanofullerenes, Advanced Functional Materials 17 (2007) 1636–1644. [6] Y. Kim, S.A. Choulis, J. Nelson, D.D.C. Bradley, S. Cook, J.R. Durrant, Device annealing effect in organic solar cells with blends of regioregular poly(3-hexylthiophene) and soluble fullerene, Applied Physics Letters 86 (2005) 063502. [7] C.S. Kim, L.L. Tinker, B.F. DiSalle, E.D. Gomez, S. Lee, S. Bernhard, Y.L. Loo, Altering the thermodynamics of phase separation in inverted bulk heterojunction organic solar cells, Advanced Materials 21 (2009) 3110–3115. [8] A.J. Moulé, K. Meerholz, Morphology control in solution-processed bulk-heterojunction solar cell mixtures, Advanced Functional Materials 19 (2009) 3028–3036. [9] M. Morana, P. Koers, C. Waldauf, M. Koppe, D. Muehlbacher, P. Denk, M. Scharber, D. Waller, C. Brabec, Organic field-effect devices as tool to characterize the bipolar transport in polymer-fullerene blends: the case of P3HT–PCBM, Advanced Functional Materials 17 (2007) 3274–3283. [10] A.R. Campbell, J.M. Hodgkiss, S. Westenhoff, I.A. Howard, R.A. Marsh, C.R. McNeill, R.H. Friend, N.C. Greenham, Low-temperature control of nanoscale morphology for high performance polymer photovoltaics, Nano Letters 8 (2008) 3942–3947. [11] S. Berson, R. DeBettignies, S. Bailly, S. Guillerez, Poly(3hexylthiophene) fibers for photovoltaic applications, Advanced Functional Materials 17 (2007) 1377–1384. [12] C.M. Hansen, The universality of the solubility parameter, Product R&D 8 (1969) 2–11. [13] A.I. Imre, W.A. Van Hook, Liquid–Liquid demixing from solutions of polystyrene. 1. A review. 2. Improved correlation with solvent properties, Journal of Physical and Chemical Reference Data 25 (1996) 637. [14] A. Martin, J. Newburger, A. Adjei, Extended Hildebrand solubility approach: solubility of theophylline in polar binary solvents, Journal of Pharmaceutical Sciences 69 (1980) 487–491. [15] R.F. Tabor, S. Gold, J. Eastoe, Electron density matching as a guide to surfactant design, Langmuir 22 (2005) 963–968.

1470

C.-D. Park et al. / Organic Electronics 12 (2011) 1465–1470

[16] M.M. Bouman, E.E. Havinga, R.A.J. Janssen, E.W. Meijer, Chiroptical properties of regioregular chiral polythiophenes, molecular crystals and liquid crystals science and technology, Section A: Molecular Crystals and Liquid Crystals 256 (1994) 439–448. [17] L. Gang, S. Vishal, Y. Yan, Y. Yang, Investigation of annealing effects and film thickness dependence of polymer solar cells based on poly(3-hexylthiophene), Journal of Applied Physics 98 (2005) 043704.

[18] X. Yang, J. Loos, S.C. Veenstra, W.J.H. Verhees, M.M. Wienk, J.M. Kroon, M.A.J. Michels, R.A.J. Janssen, Nanoscale morphology of highperformance polymer solar cells, Nano Letters 5 (2005) 579–583. [19] B.J. Worfolk, D.A. Rider, A.L. Elias, M. Thomas, K.D. Harris, J.M. Buriak, Bulk heterojunction organic photovoltaics based on carboxylated polythiophenes and PCBM on glass and plastic substrates, Advanced Functional Material 21 (2011) 1816–1826.