High-efficiency polymer solar cells by blade coating in chlorine-free solvents

(1)

Letter

High-efﬁciency polymer solar cells by blade coating in

chlorine-free solvents

Pei-Ting Tsai

a

, Chia-Ying Tsai

b

, Chun-Ming Wang

a

, Yu-Fan Chang

c

, Hsin-Fei Meng

a,⇑

,

Zhi-Kuan Chen

d,⇑

, Hao-Wu Lin

e,⇑

, Hsiao-Wen Zan

c

, Sheng-Fu Horng

f

, Yi-Chun Lai

c

,

Peichen Yu

c

a_{Institute of Physics, National Chiao Tung University, Hsinchu 300, Taiwan} b

Institute of Photonics Technologies, National Tsing Hua University, Hsinchu 300, Taiwan c

Department of Photonics and the Institute of Electro-Optical Engineering, National Chiao Tung University, Hsinchu 300, Taiwan d

Institute of Advanced Materials, Nanjing University of Technology, Nanjing 210009, China e

Department of Materials Science and Engineering, National Tsing Hua University, Hsinchu 300, Taiwan f

Department of Electrical Engineering, National Tsing Hua University, Hsinchu 300, Taiwan

a r t i c l e

i n f o

Article history:

Received 26 August 2013

Received in revised form 27 December 2013 Accepted 21 January 2014

Available online 4 February 2014 Keywords:

PBDTTT-C-T Chlorine-free solvents Blade coating Polymer solar cells

a b s t r a c t

High-performance polymer cells are typically fabricated by employing toxic solvents such as dichlorobenzene and chlorobenzene. In this study, blade coating with the chlorine-free solvents toluene and xylene is applied to polymer solar cells that contained the low band-gap polymer PBDTTT-C-T blended with [6,6]-phenyl-C71-butyric acid methyl ester ([70]PCBM). The highest efficiencies of the cells fabricated in toluene and xylene solutions were 6.09% and 6.11%, respectively. Atomic force microscopy images show that the films formed by blade coating using toluene and xylene were extremely smooth, with roughness of only 1 nm. This blade coating has a rapid-drying in a few seconds without the long-running thermal or solvent annealing. The possibility of high-volume environmentally-friendly fabrication of efficient polymer solar cells with minimal material waste is thus demonstrated using a combination of chlorine-free solvents and blade coating.

1. Introduction

Polymer solar cells based on a blend of conjugated polymers as electron donors and fullerene derivatives that serve as electron acceptors have exhibited a ever-increasing power conversion efficiency over the past decade. Recently efficiency over 8% is reported for low band-gap polymers [1–9]. However, most high-efficiency polymer solar cells have been fabricated by spin coating using chlorine-containing solvents such as dichlorobenzene,

because of their high boiling point and slow evaporation. The desired donor–acceptor phase separation in the tens of nanometer scale is usually achieved by using a slow sol-vent evaporation method, i.e. solsol-vent annealing, after spin coating in dichlorobenzene solvent[10]. In addition, mate-rial solubility and molecular compatibility are important considerations in designing polymer composites. The solu-bility of the donor polymers have a profound effect on the film morphology and solar cell performance[11–18]. Most low band-gap polymers exhibit poor solubility in com-monly used chlorine-free solvents such as toluene and xy-lene. Therefore, uniform film with proper thickness is difficult to fabricate using chlorine-free solvents, and the solar cells by them usually exhibit unsatisfactory efficien-cies [19,20]. However, solvents containing chlorine are

http://dx.doi.org/10.1016/j.orgel.2014.01.018

⇑Corresponding authors. Tel.: +886 3 5731955 (H.-F. Meng), tel.: +86 25 83587982 (Z.-K Chen), tel.: +886 3 5715131x33879 (H.-W. Lin).

E-mail addresses:[email protected](H.-F. Meng),iamzkchen@ njut.edu.cn(Z.-K. Chen),[email protected](H.-W. Lin).

Contents lists available atScienceDirect

Organic Electronics

(2)

highly toxic and environmentally hazardous. This dilemma is a major obstacle to the potential mass production of polymer cells despite of the advantage of low-cost large-area solution fabrication. Recently, a 6.1% power conver-sion efﬁciency has been reported for polymer solar cells spin coated in a halogen-free solvent[21]. However, using spin coating alone is incompatible with high-volume pro-duction because it produces large amounts of material waste and is incompatible with the roll-to-roll process. Therefore, developing a chlorine-free polymer solar cell process involving small amounts of material waste and the compatibility to roll-to-roll fabrication is crucial.

This study focuses on the low band-gap polymer PBDTTT-C-T,[22]which exhibit an unusually high solubil-ity in the chlorine-free non-toxic solvents toluene and xylene. Blade coating, rather than conventional spin coating, is used to deposit the active layer[23–26]. Blade coating has the advantage of exhibiting large-area uniformity, small amount of material waste, preventing of inter layer dissolution, and being compatible with the roll-to-roll pro-cess. Blade coating involves a rapid drying process that prevents the fabrication throughput from being slowed by the conventional solvent annealing process. A high effi-ciency of 6.1% was achieves by blade coating for solar cells comprising a combination of PBDTTT-C-T and [70]PCBM that were dissolves in toluene and xylene. In addition, the performance of these devices was insensitive to the solvent regardless of whether chlorine was included. By contrast, the archetypical polymer poly(3-hexylthiophene) (P3HT) and another low band-gap polymer poly{(benzo-2,1,3-thiadiazol-4,7-diyl)-alt-(30_,40 0 di(2-octyldodecyl)-2,20_;50_,200_;500_,2000_{-quaterthiophen-5,5}00 0_{-diyl)}(POD2T-DTBT)} were much less soluble in chlorine-free solvent than in dichlorobenzene. In sharp contrast to PBDTTT-C-T these solar cells exhibit much lower efficiency using the former solvents. The capability of using donor polymers exhibiting high solubility in chlorine-free solvents to achieve high-throughput and environmental friendly production of effi-cient polymer cells is demonstrated.

2. Experimental

Fig. 1 shows the energy band diagrams, detailing the work functions of the organic solar cells and the chemical structures of the active layer materials used in this study. The energy level of the lowest unoccupied molecular orbital (LUMO) and highest occupied molecular orbital (HOMO) relative to the vacuum are labeled[17,22]. PBDTTT-C-T was purchased from Solarmer materials Inc. The PBDTTT-C-T and POD2T-DTBT exhibited low band-gaps of 1.58 eV and 1.59 eV, respectively, whereas P3HT exhibited a band-gap of 1.9 eV[23].Fig. 2a illustrates the blade coating method andFig. 2b is the picture of the auto-blade ma-chine. The blade coater is cylindrically shaped and the gap between the cylinder and the substrate was 120

l

m. The solution was delivered to the gap by using a pipette and the motion of the blade formed a wet ﬁlm. The blade speed ranged from 20 to 400 mm/s. The blade coating procedure was executed on a hot plate. After coating the wet ﬁlm, the hot air was applied to evaporate the remaining solvent in approximately 1–10 s. The thickness of the remaining

dry ﬁlm could be controlled by adjusting the blade speed. The blade coating was performed in a self-fabricated ma-chine, which was controlled using a linear motor.

Polymer solar cells were fabricated on pre-patterned in-dium-tin-oxide (ITO) glass with the device structure of ITO/PEDOT:PSS (poly-(3,4-ethylenedioxythiophene):poly-(styrenesulfonate) (CLEVIOS™ PVP AI4083, purchased from HC Starck)/PBDTTT-C-T:[70]PCBM/Ca/Al [70]PCBM was purchased from Solenne. The ITO-coated glass sub-strates were treated in an ultrasonic bath for 60 min in ace-tone, subsequently rinsed three times with deionized water and cleaned using UV ozone cleaner for 30 min. The 40 nm PEDOT:PSS layer was spin coated at 2200 rpm on the ITO substrate and baked at 200 °C in air for 15 min. To prepare the PBDTTT-C-T and [70]PCBM solu-tions, PBDTTT-C-T and [70]PCBM powders were mixed. The weight ratio of the PBDTTT-C-T to [70]PCBM was ﬁxed at 1:1.5. The mixed powder was dissolved using various solvents such as toluene, chlorobenzene, and xylene.

To perform blade coating, approximately 3% (1,8-diio-dooctane (DIO)/solvents (toluene, chlorobenzene, and xylene), v/v) DIO was used as an additive and was included to improve photovoltaic results. The mixed solution was heated on a hot plate at 80 °C. The active layers were coated using an auto-blade machine, as shown inFig. 2. Rapid-dry-ing blade coatRapid-dry-ing was performed usRapid-dry-ing a hot plate at 80– 90 °C. The solution (70

l

L) was coated using the auto-blade machine at approximately 260 mm/s to cover the 4-mm2 active area of the device with wet film. The thickness of the active layer was 100–120 nm. Hot air produced by a hair dryer containing a diffuser was applied to enhance the drying and uniformity. Dry films formed in approxi-mately 3 s for toluene and chlorobenzene, and 10 s for xy-lene. The air temperature was approximately 70 °C. To perform spin coating, approximately 3% DIO was a added to the main solvents. The mixed solution was then heated on a hot plate at 80 °C. The thickness of the active layer was 100–120 nm when spin coated at a spin rate of 1000 rpm for 40 s. The wet film was subsequently dried slowly for 2 h during the solvent annealing process. Finally, a Ca(35 nm)/Al(100 nm) electrode was formed on the top of the active layer by using thermal evaporation in a vacuum chamber with a base pressure below 3.0 106_Torr. Regarding the POD2T-DTBT devices, only Al was evapo-rated without Ca film. All of the devices were packaged in a glove box and measured in an ambient environment.

To determine the characteristics of the solar cell de-vices, the power conversion efficiency (PCE) was measured using a solar simulator (XES-301S, SAN-EI) under AM1.5G of irradiation. The incident photon conversion efficiency (IPCE) is defined as the average number of carriers per incident photon. To measure the IPCE, a 300-W Xenon lamp (Newport 66984) and a monochromator (Newprot 74112) were used as light sources. The beam spot on the sample was a square, and the spot size was 3 mm2_{. A} calibrated silicon photodetector with a known spectral re-sponse (Newport 818-UV). The IPCE was measure using a lock-in amplifier (Standard Research System, SR830), an optical chopper unit (SR540) operating at a 260-Hz chop-ping frequency, and a 1Xresistor in a shunt connection to convert the photocurrent to voltage. There is a sharp

(3)

peak around 825 nm in our solar simulator as shown in Fig. 3. Aside from this peak the simulator power spectrum was quite similiar to the AM1.5 reference solar spectrum. The 825 nm peak contributed to some extra photo-current when the absorption material has low band gap as in our case. The IPCE shown below extends in the long wave-length region beyond 825 nm. The extra photo-current due to the 825 nm peak could be estimated by integrating the product of this peak and the IPCE spectrum. In all the data below the short-circuit current is deducted by this ex-tra current and the power conversion efﬁciency is adjusted correspondingly. To analyze the ﬁlms fabricated using var-ious deposition methods, the surface morphology and

phase were monitored using an atomic force microscope (AFM, Dimension 3100, Digital Instruments). The high-resolution morphology was monitored using a variable temperature scanning probe microscope (VT SPM, SPA-300HV). The absorption spectrum was measured using a UV–visible spectrophotometer (HP8453).

3. Results and discussions 3.1. Spin coating

Before presenting the results of using blade coating, the data regarding the solar cells fabricated using conventional Fig. 1. (a) Energy band diagrams and (b) the chemical structures in this work.

(4)

spin coating are presented. The current density–voltage (J–V) characteristics and photovoltaic parameters exhib-ited by the four solvents, dichlorobenzene, chlorobenzene, toluene, and xylene, when using a solar simulator are shown inFig. 4andTable 1. Dichlorobenzene and chloro-benzene contain chlorine, and toluene and xylene do not. The boiling points of these solvents are 180 °C, 131 °C, 138 °C and 111 °C respectively. The dichlorobenzene de-vice exhibited the highest power conversion efficiency, which was 6.42%. This value is only slightly lower than the value produces in a previous study (7.59%) in which the same process was used. Because it exhibited a high boiling point and underwent a slow evaporation process, the blend film composed of dichlorobenzene became dry after approximately 2 h of solvent annealing, in which the polymer donor and the fullerene acceptor performed the phase separation. The other three solvents produced lower efficiency, most likely because the evaporation of dichlorobenzene occurred more quickly than did that of the other solvents. Nevertheless, the efficiency was over

6% when toluene and xylene were used, which is high for chlorine-free solvents. This result was attributed to the excellent solubility of PBDTTT-C-T in these two solvents. As previously mentioned, the intrinsic problem of spin coating is the large amount of material used and the incompatibility of this technique with continual processes occurring in a large area. Therefore, blade coating eas used for the same active layer blend of PBDTTT-C-T and [70]PCBM.

3.2. Blade coating

Fig. 5andTable 1shows the performance of the solar cells fabricated using blade coating in various solvents. All of the solar cells produced efficiency approximately 6% regardless of whether the solvents contained chlorine. The device constructed using toluene and xylene exhibited high efficiency of 6.09% and 6.11%, respectively. The device fabricated using chlorobenzene exhibited an efficiency

(a)

(b)

Blade coater

Hot plate

Fig. 2. (a) The blade coating method and (b) the auto-blade machine.

300 400 500 600 700 800 900 1000 1100 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 I-V spectra AM1.5G

Spectral irradiance

(W/m

2

nm

)

Wavelength (nm)

Fig. 3. AM1.5 reference solar spectrum and solar simulator spectrum.

0.0 0.0 0.20.2 0.40.4 0.60.6 0.80.8 -20 -20 -15 -15 -10 -10 -5 -5 0 5 10 10 15 15 20 20 Toluene Toluene Chlorobenzene Chlorobenzene Xylene Xylene Dichlorobenzene Dichlorobenzene Voltage (V) Voltage (V) J J ( mA /cmm A / c m 2 )

PBDTTT-C-T:[70]PCBM in different solvents by spin PBDTTT-C-T:[70]PCBM in different solvents by spin

Fig. 4. J–V curves of PBDTTT-C-T/[70]PCBM solar cells with active layers dissolved in different solvents using spin coating method.

(5)

5.76%. The chlorobenzene device exhibited a lower effi-ciency because it produced lower short-circuit current JSC than the other devices did, as shown inFig. 5a. The dark J–V curves are shown inFig. 5b using log scale. Under re-verse bias, the leakage current of the devices fabricated using chlorobenzene was higher than that of the devices fabricated using by toluene and xylene. Blade coating in xylene gives the highest efficiency, and its external quan-tum efficiency (IPCE) for the device is shown in Fig. 5c. After multiplying the IPCE with AM1.5 solar spectrum and integration over the wavelength, the total photo-current is 12.81 mA/cm2_{. This is about 10% lower than the} measured Jscof 14.14 mA/cm2. In order to avoid possible overestimating the Jsc, we scaled it down by a factor x = 0.906 to match the integrated IPCE inTables 1 and 3, which summarizes the performance. For devices

conditions other than blade coating in xylene, we used the same factor x for the down scaling.

To obtain a more comprehensive insight into the origin of this moderate dependence on solvents, the absorption spectra of the pure PBDTTT-C-T film and the blend film blade coated using various solvents were measured, and the results are shown inFig. 6. In pure PBDTTT-C-T film, the xylene film produced a pronounced long-wavelength shoulder peak at 705 nm, whereas the other two films did not produced a clear peak as shown inFig. 6a. The cause of this shoulder peak is generally attributed to the polymer aggregates in which the ordered chain packing produces a low amount of exciton energy. Xylene exhibited the highest boiling point among the three solvents and this strong aggregate formation may have resulted from using a longer drying time under the same heating conditions used Table 1

Photovoltaic parameters for PBDTTT-C-T:[70]PCBM solar cells with active layers dissolved in different solvents by using spin-coating and blade-coating.

Cell Jsc(mA/cm2) Voc (V) Fill factor PCE (%)

Toluene (spin) 13.277 ± 0.226 0.799 ± 0.002 0.535 ± 0.006 5.679 ± 0.178 Chlorobenzene (spin) 12.094 ± 0.122 0.784 ± 0.002 0.496 ± 0.004 4.712 ± 0.069 Xylene (spin) 14.471 ± 0.321 0.784 ± 0.001 0.481 ± 0.005 5.459 ± 0.064 Dichlorobenzene (spin) 13.608 ± 0.127 0.799 ± 0.001 0.586 ± 0.004 6.358 ± 0.059 Toluene (blade) 13.273 ± 0.140 0.794 ± 0.002 0.565 ± 0.016 5.954 ± 0.132 Chlorobenzene (blade) 11.793 ± 0.39 0.799 ± 0.002 0.594 ± 0.016 5.550 ± 0.207 Xylene (blade) 12.813 ± 0.260 0.809 ± 0.003 0.577 ± 0.007 5.983 ± 0.129 0.0 0.2 0.4 0.6 0.8 -20 -15 -10 -5 0 5 10 15 20 Voltage (V) J ( mA /cm 2 ) Toluene Toluene Chlorobenzene Chlorobenzene Xylene Xylene

PBDTTT-C-T:[70]PCBM in different solvents by blade PBDTTT-C-T:[70]PCBM in different solvents by blade

(a)

-1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 1E-4 1E-3 0.01 0.1 1 10 100 Voltage (V) Toluene Chlorobenzene Xylene

The dark current of PBDTTT-C-T:[70]PCBM The dark current of PBDTTT-C-T:[70]PCBM

(b)

(c)

300 400 500 600 700 800 0 20 40 60 80 100 E.Q.E (%) Wavelengh (nm) PBDTTT-C-T:[70]PCBM @Xylene IPCE PBDTTT-C-T:[70]PCBM @Xylene IPCE J ( mA /cm 2)

Fig. 5. (a) The J–V curves and (b) the dark J–V curves using log scale of PBDTTT-C-T/[70]PCBM solar cells with active layers dissolved in different solvents using blade coating method. (c) IPCE spectrum of PBDTTT-C-T/[70]PCBM @xylene device.

(6)

in the blade coating process. However, this major difference in the absorption spectra was not visible in the blend ﬁlm analysis, as shown inFig. 6b, before and after

normalization. Chlorobenzene, toluene, and xylene pro-duced similar spectra in the 600–750 nm range and exhib-ited clearly visible shoulder peaks. The chlorobenzene ﬁlm

400 500 600 700 800 900 Blend (PBDTTT-C-T/[70]PCBM) Normalized (a.u.) Wavelength (nm) PBDTTT-C-T:[70]PCBM @Toluene PBDTTT-C-T:[70]PCBM @Chlorobenzene PBDTTT-C-T:[70]PCBM @Xylene 400 500 600 700 800 900 Normalized (a.u.) Wavelength (nm) PBDTTT-C-T @Toluene PBDTTT-C-T @Chlorobenzene PBDTTT-C-T @Xylene Pure donor (PBDTTT-C-T)

(a)

(b)

Fig. 6. The absorption spectra of ﬁlms prepared with different solvents. (a) The pure PBDTTT-C-T and (b) the blend of PBDTTT-C-T/[70]PCBM.

(c) Xylene (Blade)

(b) Chlorobenzene (Blade)

(a) Toluene (Blade)

10 He ig h t ( n m) 0 3 Deg ree ( o) 0

250 nm

Phase

Phase Phase

R

_rms

= 1.18 nm

R

_rms

= 1.67 nm

R

_rms

= 0.99 nm

0 phase 10 He ig h t ( n m)

(d) Toluene (Spin)

250 nm

(e) Chlorobenzene (Spin)

250 nm

(f) Xylene (Spin)

250 nm

0 3

250 nm

Phase

Phase Phase

R

rms

= 1.31 nm

R

rms

= 1.99 nm

R

rms

= 1.81 nm

Deg

ree (

o)

(7)

exhibited a slightly lower peak than the other solvents. Therefore, this negligible difference in aggregate formation was unlikely to have caused the differences in JSC, as shown inFig. 5a.

3.3. AFM images

To understand the morphology of the films that were blade coated using various solvents further, atomic force microscope (AFM) was used to obtain the topography and phase images of the films formed using various sol-vents. The results are shown in Fig. 7(a–c, for blade, d–f for spin) andTable 2, with the films deposited using spin coating displayed for comparison. As shown inFig. 7a–c, the root mean square roughness values were 1.18, 1.67, and 0.99 nm for the films composed of toluene, chloroben-zene, and xylene, respectively. The chlorobenzene film exhibited the highest roughness and the largest grain size of the films. The grains are shown at a scale of 0.3

l

m, were visible only in the chlorobenzene film. To demonstrate the morphological difference further, high-resolution AFM was used to measure the chlorobenzene, toluene, and xylene films. The topography diagrams are displayed in Fig. 8. The toluene and xylene films exhibited only a small grain at the scale of 50 nm. The chlorobenzene film exhibited both a small grain of 50 nm and a large grain of 0.3

l

m. This large grain structure is most likely responsible for the low Jscof the chlorobenzene device, shown inFig. 5a.

This large grain structure may contain a high concentration of grain boundaries and consequently form photo-carrier recombination centers. Furthermore, the large grains may cause the film to have poor contact with the top metal cathode. High-resolution AFM exhibited higher surface roughness than the low-resolution AFM for the sample. The possible reason is that certain fine structures in the film could not be resolved by the low-resolution AFM, for example, a dip with high depth but small area. The phase image for the chlorobenzene film exhibited a high phase contrast between the center and the edge of the grain as shown inFig. 7b. However, no high phase contrast in the toluene and xylene films was observed, as shown in Fig. 7a and c. The high phase contrast in the chlorobenzene film suggests that donor–acceptor phase separation oc-curred at the 0.3

l

m scale, which is much higher than the exciton diffusion length of approximately 20 nm. This large-scale phase separation may reduce the total bulk het-ero-junction interface concentration and lower the exciton Table 2

The Rrmsvalues of the PBDTTT-C-T/[70]PCBM ﬁlm dissolved in various solvents by using spin-coating and blade-coating.

Toluene (nm) Chlorobenzene (nm) Xylene (nm) Blade 1.18 ± 0.19 1.67 ± 0.27 0.99 ± 0.27 Spin 1.31 ± 0.63 1.99 ± 0.65 1.81 ± 0.44

(a) Toluene (Blade)

250 nm

(b) Chlorobenzene (Blade)

(c) Xylene (Blade)

Toluene R_rms= 2.18 ± 0.12 nm Chlorobenzene R_rms= 4.83 ± 0.11 nm Xylene R_rms= 2.47 ± 0.15 nm

250 nm

Fig. 8. The high resolution AFM topography images and Rrmsvalues of the PBDTTT-C-T/[70]PCBM active layers dissolved in (a) toluene, (b) chlorobenzene and (c) xylene. The upper diagrams are 2-D images and the under diagrams are 3-D images (1lm 1lm).

0.0 0.2 0.4 0.6 0.8 -20 -15 -10 -5 0 5 10 15 20 J ( mA/cm 2) Voltage (V) spin spin+annealing blade blade+annealing PBDTTT-C-T:[70]PCBM @Toluene

Fig. 9. J–V curves of PBDTTT-C-T/[70]PCBM @toluene devices, and the thermal annealing has negative effect.

(8)

dissociation probability. The smooth surface of the blend film with a roughness of approximately 1 nm is crucial for maintaining the stability of the device. A film with 1 nm roughness is similar to a typical spin-coated polymer film with a single component. Therefore, blade coating PBDTTT-C-T and [70] PCBM simultaneously produced an extremely smooth film and a high exciton dissociation probability. These results usually do not occur in spin coat-ing. The AFM images for the films fabricated using spin coating are shown inFig. 7d–f for comparison. In spin coat-ing, the roughness of the film is generally higher than that of the films fabricated using blade coating. Large grains were visible in the chlorobenzene film. As shown in Fig. 4the chlorobenzene also produced a lower level of effi-ciency than toluene and xylene did, which is similar to the trend that occurred when the devices were blade coated. The smooth surface of the blade-coated films is probably related to the rapid drying technique that was used. Simul-taneously heating the films from the bottom and the top by using a hot plate and hot wind cause the films to become dry in seconds. The xylene film took 10 s to dry, whereas the toluene and chlorobenzene films took 1–3 s. Rapid drying prevented the formation of a rough structure caused by horizontal or vertical molecular migration dur-ing the drydur-ing time.

3.4. Annealing effect

In the conventional solvent annealing fabrication of polymer solar cells using dichlorobenzene, a thermal annealing step conducted after solvent annealing is performed. The efficiency increased as the microscopic morphology was tuned when using thermal annealing. Thermal annealing, however, negatively affected the PBDTTT-C-T and [70]PCBM blend film, regardless of the solvent or the deposition methods used. The effect of per-forming thermal annealing at 140 °C for 20 min is shown in Fig. 9andTable 3. After completing the thermal annealing process, the efficiency decreased approximately 50% for all conditions. This suggests that the as-deposited films al-ready possessed a nearly optimal morphology, and that further tuning using thermal annealing at a high tempera-ture was not required.

3.5. Polymers with low solubility in toluene

Chlorinated solvents are in general stronger than the chlorine-free solvents like toluene. To illustrate the uniqueness of PBDTTT-C-T further, another two polymers dissolved in a toluene solution were studied. The first poly-mer was POD2T-DTBT, which had a low band-gap of 1.6 eV and exhibited a high efficiency of 6.26% when conventional annealing was used[27]and 6.67% when blade coating and chlorobenzene were used[25,28]. The other polymer was the archetypical polymer P3HT, which exhibited an effi-ciency of approximately 4% when solvent annealing was used. Unlike PBDTTT-C-T, both POD2T-DTBT and P3HT exhibited poor solubility in chlorine-free solvents. For example, the blend with a concentration of 19 mg/ml and a 1:1 P3HT:[60]PCBM ratio were fully dissolved in dichlo-robenzene overnight, but they required two days to dis-solve in toluene. The solubility of POD2T-DTBT in toluene was extremely low, which indicates that performing blade coating with this solution at room temperature is impossi-ble. A uniform film could be fabricated only by heating the toluene solution to 100 °C. The solar cell characteristics of POD2T-DTBT blended with [70]PCBM by heating a toluene solution are shown inFig. 10andTable 4. In contrast to PBDTTT-C-T, the toluene device exhibited a significantly low efficiency of 3.16% compared with the 5.62% exhibited Table 3

Photovoltaic parameters for PBDTTT-C-T:[70]PCBM @toluene by spin-coating and blade-coating. The thermal annealing has negative effect.

Spin 13.170 ± 0.171 0.779 ± 0.010 0.446 ± 0.013 4.626 ± 0.207 Spin + annealing 7.879 ± 0.225 0.709 ± 0.008 0.311 ± 0.010 1.706 ± 0.105 Blade 13.273 ± 0.140 0.794 ± 0.002 0.565 ± 0.016 5.954 ± 0.132 Blade + annealing 10.200 ± 0.315 0.772 ± 0.002 0.353 ± 0.003 2.782 ± 0.103 0.0 0.2 0.4 0.6 0.8 -20 -15 -10 -5 0 5 10 15 20 Chlorobenzene (25 o C) Toluene (100 oC) Voltage (V) J ( mA/cm 2 )

POD2T-DTBT:[70]PCBM in different solvents

Fig. 10. J–V curves of POD2T-DTBT/[70]PCBM solar cells with active layers dissolved in chlorobenzene at room temperature and toluene at 100 °C.

Table 4

Photovoltaic parameters for POD2T-DTBT/[70]PCBM solar cells with active layers dissolved in different solvents by blade-coating.

Chlorobenzene (25 °C) 16.610 ± 0.302 0.690 ± 0.001 0.485 ± 0.015 5.573 ± 0.045

(9)

by the chlorobenzene device. Therefore, to obtain high efﬁ-ciency when using a chlorine-free solvent, applying a molecular design for achieving high solubility is necessary.

The absorption spectra of the pure POD2T-DTBT film and the blend film are shown inFig. 11a, and b. The spectra of the pure toluene and chlorobenzene films were almost

400 500 600 700 800 900 Blend (P3HT/[60]PCBM) P3HT:[60]PCBM @Toluene P3HT:[60]PCBM @Chlorobenzene Wavelength (nm) 400 500 600 700 800 900 Normalized (a.u.) Wavelength (nm) P3HT @Toluene P3HT @Chlorobenzene Pure donor (P3HT) 400 500 600 700 800 900 Blend (POD2T-DTBT/[70]PCBM) POD2T-DTBT:[70]PCBM @Toluene POD2T-DTBT:[70]PCBM @Chlorobenzene Wavelength (nm) 400 500 600 700 800 900 Normalized (a.u.) Wavelength (nm) POD2T-DTBT @Toluene POD2T-DTBT @Chlorobenzene Pure donor (POD2T-DTBT)

(a)

(b)

(c)

(d)

Normalized (a.u.)

Fig. 11. The absorption spectra of ﬁlms prepared with different solvents. (a) The pure POD2T-DTBT, (b) POD2T-DTBT/[70]PCBM, (c) the pure P3HT and (d) P3HT/[60]PCBM. POD2T-DTBT:[70]PCBM @Chlorobenzene Rrms= 3.13 ± 0.12 nm P3HT:[60]PCBM @Chlorobenzene R_rms= 1.73 ± 0.16 nm

(a)

(b)

(c)

(d)

POD2T-DTBT:[70]PCBM @Toluene Rrms= 3.39 ± 0.14 nm P3HT:[60]PCBM @Toluene R_rms= 3.33 ± 0.11 nm 250 nm 250 nm 250 nm 250 nm

Fig. 12. AFM topography images and Rrmsvalues of the POD2T-DTBT/[70]PCBM and P3HT/[60]PBCM active layers dissolved in (ac) toluene and (bd) chlorobenzene (1lm 1lm for 2-D and 3-D images).

(10)

identical. The blend film spectra displayed a distinct absorption shoulder of approximately 705 nm when POD2T-DTBT was dissolved in toluene, but this did not oc-cur in the spectra produced when chlorobenzene was used. This implies that the high concentration of POD2T-DTBT aggregate was caused by the poor solubility of the polymer in toluene. Similarly, the pure P3HT film produced an almost identical absorption spectra when toluene and chlorobenzene were used, as shown in Fig. 11c. By contrast, the blend film spectra displayed a clear red shift when POD2T-DTBT and [70]PCBM were dissolved in toluene, which may have been caused by the P3HT

aggregate that formed because of poor solubility. In contrast to POD2T-DTBT and P3HT, the absorption spectrum of the PBDTTT-C-T blend film exhibited little solvent dependence, as shown in Fig. 6. This characteristic is associated with the excellent solubility of PBDTTT-C-T in chlorine-free solvents. The high-resolution AFM images for the POD2T-DTBT and P3HT blend film when toluene and chlo-robenzene were used are shown inFig. 12for toluene and chlorobenzene. Distinct granular structures were observed in the chlorobenzene films, but not in the toluene films. This grain structure suggests that proper donor–acceptor phase separation is required to achieve a high exciton dis-sociation rate. As shown inFigs. 10 and 13, andTables 4 and 5, the toluene device exhibited much lower efficiency than did the chlorobenzene devices for the two polymers. Therefore, the poor efficiency produced when using tolu-ene is related to the lack of phase separation at the proper length scale. By contrast, the granular structures of the PBDTTT-C-T film were noticeable even when the polymer was dissolved in toluene, as shown in Fig. 8a. Therefore, the key criteria for achieving high performance using a chlorine-free solvent is the high solubility of the polymer and the formation of granular structures at the 20–50 nm length scale.

Finally, we remarked on the need for heating of the blend solution used in coating. Full dissolution of the PBDTTT-C-T is possible only with heating to 80 °C for all solvents in this work. The solution became a gel once the temperature drops back to room temperature. The coating

0.0 0.2 0.4 0.6 -15 -10 -5 0 5 10 15 Voltage (V) J ( mA /cm 2 ) Dichlorobenzene (Spin) Chlorobenzene (Blade) Toluene (Blade) P3HT:[60]PCBM in different solvents

Fig. 13. J–V curves of P3HT/[60]PCBM solar cells with active layers dissolved in various solvents.

Table 5

Photovoltaic parameters for P3HT/[60]PCBM solar cells with active layers dissolved in different solvents by blade-coating.

Dichlorobenzene (spin) 8.919 ± 0.106 0.661 ± 0.003 0.654 ± 0.008 3.856 ± 0.107 Chlorobenzene (blade) 10.317 ± 0.209 0.602 ± 0.001 0.581 ± 0.005 3.607 ± 0.040

Toluene (blade) 7.855 ± 0.145 0.589 ± 0.001 0.605 ± 0.005 2.783 ± 0.047

(a)

(b)

(c)

(d)

foggy ﬁlm glass

clear ﬁlm glass

moled white ﬁlm

clear ﬁlm

POD2T-DTBT blend 25 o_C _{POD2T-DTBT blend 100}o_C

PBDTTT-C-T blend 25 o_C _{PBDTTT-C-T blend 80}o_C

(11)

of the gel gave a foggy film with high roughness as shown in Fig. 14a. Therefore, all the devices are coated with heated solution to ensure a clear film with low roughness as shown inFig. 14b. As for the POD2T-DTBT, it was fully dissolved in toluene only with heating up to 100 °C. Once the solution was cooled down to room temperature, in-stead of gelation there are precipitates. Coating of solution at room temperature also gives foggy rough film similar to PBDTTT-C-T films. The film was mottled white as shown in Fig. 14c. Coating of toluene solution at 100 °C gives clear film as shown inFig. 14d. POD2T-DTBT can be fully dis-solved in chlorobenzene and coated at room temperature. 4. Conclusion

In summation, a high-performance polymer solar cell based on PBDTTT-C-T was successfully fabricated using blade coating in toluene and xylene solution. These sol-vents are chlorine-free. Therefore, the conventional use of toxic solvents such as dichlorobenzene and chlorobenzene in polymer solar cells can be avoided. The highest efﬁ-ciency achievable by spin coating was 6.42%, and that for blade coating was 6.11%. The good solubility of PBDTTT-C-T in the chlorine-free solvents was the primary reason for the high performance. By combining a non-toxic sol-vent with blade coating, which is suitable for roll-to-roll fabrication with minimal material waste, efﬁcient polymer solar cells can be fabricated in high volumes, using an envi-ronmentally-friendly approach.

Acknowledgements

This work was supported by the National Science Council of Taiwan under Grant Nos. 103-2120-M-009-003-CC1, 102-2221-E-007-125-MY3, 101-2112-M-009-006-MY3, 101-2112-M-007-017-MY3 and the Ministry of Economic Affairs of Taiwan under Grant No. 100-EC-17-A-07-S1-157. We thank Wen-Bin Jian for the high-resolution AFM images.

Reference

[1]Z. He, C. Zhong, S. Su, M. Xu, H. Wu, Y. Cao, Nat. Photonics 6 (2012) 591–595.

[2]Y.-W. Su, S.-C. Lan, K.-H. Wei, Mater. Today 15 (2012) 554–562. [3]L. Ye, S. Zhang, W. Ma, B. Fan, X. Guo, Y. Huang, H. Ade, J. Hou, Mater.

Views 24 (2012) 6335–6341.

[4]X. Guo, M. Zhang, J. Tan, S. Zhang, L. Huo, W. Hu, Y. Li, J. Hou, Adv. Mater. 24 (2012) 6536–6541.

[5]Y. Liang, Z. Xu, J. Xia, S.-T. Tsai, Y. Wu, G. Li, C. Ray, L. Yu, Adv. Mater. 22 (2012) E135–E138.

[6]H.Y. Chen, J.H. Hou, S.Q. Zhang, Y.Y. Liang, G.W. Yang, Y. Yang, L.P. Yu, Y. Wu, G. Li, Nat. Photonics 3 (2009) 649–653.

[7]S.C. Price, A.C. Stuart, L. Yang, H. Zhou, W. You, J. Am. Chem. Soc. 133 (2011) 4625–4631.

[8]C. Piliego, T.W. Holcombe, J.D. Douglas, C.H. Woo, P.M. Beaujuge, J.M.J. Fréchet, J. Am. Chem. Soc. 132 (2010) 7595–7597.

[9]Z. He, C. Zhong, X. Huang, W.Y. Wong, H. Wu, L. Chen, S. Su, Y. Cao, Adv. Mater. 23 (2011) 4636–4643.

[10]G. Li, V. Shrotriya, J. Huang, Y. Yao, T. Moriarty, K. Emery, Y. Yang, Nat. Mater. 4 (2005) 864–868.

[11]P.A. Troshin, D.K. Susarova, E.A. Khakina, A.A. Goryachev, O.V. Borshchev, S.A. Ponomarenko, V.F. Razumova, N.S. Sariciftci, J. Mater. Chem. 22 (2012) 18433–18441.

[12]P.A. Troshin, H. Hoppe, J. Renz, M. Egginger, J.Y. Mayorova, A.E. Goryachev, A.S. Peregudov, R.N. Lyubovskaya, G. Gobsch, N.S. Sariciftci, V.F. Razumov, Adv. Funct. Mater. 19 (2009) 779–788. [13]C.-K. Lee, C.-W. Pao, J. Phys. Chem. C 116 (2012) 12455–12461. [14]H. Hoppe, T. Glatzel, M. Niggemann, W. Schwinger, F. Schaefﬂer, A.

Hinsch, M.C. Lux-Steiner, N. Sariciftci, Thin Solid Films 511–512 (2006) 587–592.

[15]H. Hoppe, N.S. Sariciftci, J. Mater. Chem. 16 (2006) 45–61. [16]S.E. Shaheen, C. Brabec, N.S. Saricifci, F. Padinger, T. Fromherz, J.C.

Hummelen, Appl. Phys. Lett. 78 (2001) 841–843.

[17]M.T. Rispens, A. Meetsma, R. Rittberger, C.J. Brabec, N.S. Sariciftci, J.C. Hummelen, Chem. Commun. (2003) 2116–2118.

[18]C.M. Björström, K.O. Magnusson, E. Moons, Synth. Met. 152 (2005) 109–112.

[19]F. Reisdorffer, O. Haas, P. Le Rendu, T.P. Nguyen, Synth. Met. 161 (2012) 2544–2548.

[20]S.K. Jang, S.C. Gong, H.J. Chang, Synth. Met. 162 (2012) 426–430. [21]K.-S. Chen, H.-L. Yip, C.W. Schlenker, D.S. Ginger, A.K.-Y. Jen, Org.

Electron. 13 (2012) 2870–2878.

[22]L. Huo, S. Zhang, X. Guo, F. Xu, Y. Li, J. Hou, Angew. Chem. Int. Ed. 50 (2011) 9697–9702.

[23]S.-R. Tseng, H.-F. Meng, K.-C. Lee, S.-F. Horng, Appl. Phys. Lett. 93 (2008) 153308.

[24]C.-Y. Chen, H.-W. Chang, Y.-F. Chang, B.-J. Chang, Y.-S. Lin, P.-S. Jian, H.-C. Yeh, H.-T. Chien, E.-C. Chen, Y.-C. Chao, H.-F. Meng, H.-W. Zan, H.-W. Lin, S.-F. Horng, Y.-J. Cheng, F.-W. Yen, I.-F. Lin, H.-Y. Yang, K.-J. Huang, M.-R. Tseng, J. Appl. Phys. 110 (2011) 094501.

[25]S.-L. Lim, E.-C. Chen, C.-Y. Chen, K.-H. Ong, Z.-K. Chen, H.-F. Meng, Sol. Energy Mater. Sol. Cells 107 (2012) 292–297.

[26]Y.-F. Chang, Y.-C. Chiu, H.-C. Yeh, H.-W. Chang, C.-Y. Chen, H.-F. Meng, H.-W. Lin, H.-L. Huang, T.-C. Chao, M.-R. Tseng, H.-W. Zan, S.-F. Horng, Org. Electron. 13 (2012) 2149–2155.

[27]K.-H. Ong, L. Lim, H.-S. Tan, H.-K. Wong, J. Li, Z. Ma, L.C.H. Moh, S.-H. Lim, J.C. de Mello, Z.-K. Chen, Adv. Mater. 23 (2011) 1409–1413. [28]R. Kroon, M. Lenes, J.C. Hummelen, P.W.M. Blom, B. De Boer, Polym.