Journal ofMaterials Chemistry A

COMMUNICATION

Publ

ishe

d on

28

Aug

ust 2

014.

Dow

nloa

ded

by A

alto

Uni

vers

ity o

n 29

/10/

2014

21:

23:1

9.

View Article OnlineView Journal

A durable SWCN

aNew Energy Technologies Group, Departm

School of Science, P.O Box 15100, FI-0007

aalto.; michael.graetzel@ep.chbLaboratory of Photonics and Interfaces, Ec

(EPFL), CH G1 551, Station 6, CH-1015 LaucEngineering Materials Group, Department

Aalto University School of Science, FI-00076

Cite this: DOI: 10.1039/c4ta03730h

Received 19th July 2014Accepted 28th August 2014

DOI: 10.1039/c4ta03730h

www.rsc.org/MaterialsA

This journal is © The Royal Society of

T/PET polymer foil based metalfree counter electrode for flexible dye-sensitizedsolar cells

Syed Ghufran Hashmi,a Thomas Moehl,b Janne Halme,*a Ying Ma,a Tapio Saukkonen,c

Aswani Yella,b Fabrizio Giordano,b Jean David Decoppet,b Shaik. M. Zakeeruddin,b

Peter Lunda and Michael Gratzel*b

An ITO free, highly conductive PET foil is fabricated by depositing

aqueous single-walled carbon nanotube (SWCNT) ink that exhibits

remarkable durability when exposed to severe mechanical stability

tests. Excellent adhesion of the SWCNT film on PET was obtained by

aging the ink overnight at 50 �C before deposition. A counter elec-

trode for a dye-sensitized solar cell was fabricated by electro-poly-

merizing the PEDOT polymer over the SWCNT film which gave 7%

solar cell efficiency and low (0.4 U cm2) charge transfer resistance.

Dye-sensitized solar cells (DSSCs) have been intensively studiedas prospective alternatives to conventional p–n junction solarcells, because of the potential for their fabrication at low cost.1

The advantage of DSSCs over silicon-based solar cells is thepossibility for the roll-to-roll mass production when the light-weight polymer (usually indium doped tin oxide-polyethyleneterephthalate (ITO-PET) and indium doped tin oxide poly-ethylene naphthalate (ITO-PEN)) is used as the exiblesubstrate.2–4 Flexible solar panels offer many advantages overclassic rigid solar panels: (i) they can be installed easily onbuildings with any shape of surfaces due to their exible nature;(ii) their light weight could enable them as a mobile powersource for various portable electronic devices; (iii) the cost ofsolar cells would be signicantly reduced by using plasticsubstrates instead of the transparent conductive oxide (TCO)glass substrates, since the current production of TCO glassinvolves expensive vacuum processes. Therefore, the develop-ment of exible DSSCs has attracted extensive research interestin recent years. To promote the exible DSSCs in the competi-tive photovoltaic market, the foremost prerequisite is thedurability of these types of exible solar cells, thus robust

ent of Applied Physics, Aalto University

6 Aalto, Finland. E-mail: janne.halme@

ole Polytechnique Federale de Lausanne

sanne, Switzerland

of Engineering Design and Production,

Aalto, Finland

Chemistry 2014

fabrication of materials on the polymer sheets associated withthe photo electrode (PE) and counter electrode (CE) is neces-sary. Unfortunately, as plastic substrates limit the thermaltreatment to around 150 �C, most of the conventional maturetechniques for PE and CE lm fabrication on glass substrates,e.g. screen-printing followed by sintering for the TiO2 lm,5 orthermal platinization for Pt CE2,6 can no longer be utilized. Toovercome this problem, some low-temperature depositionprocesses for both PE and CE need to be explored that couldgive the same robustness of the materials over the polymersheets like on glass. Lately, some studies have been carried outon fabrication of an efficient photo anode for exible DSSCs,typically a TiO2 lm either on ITO-PET or ITO-PENsubstrates.7–11 For instance, Weerasinghe and co-workersreported the fabrication of well-adhered TiO2 electrodes usingbinder-free TiO2 slurry obtained by ball-milling and acidictreatments.4,8 The deposited layer was further treated with coldiso-static pressing (CIP) to improve the mechanical stability ofTiO2 nanoparticles over ITO-PEN4 which resulted a conversionefficiency of 6.3%. Moreover, Miyasaka and co-workers showedhigh adhesion of the low temperature TiO2 binder free lm onITO-PEN which exhibited �6% efficient exible DSSCs.12

However, the trend is not the same for the counter electrode(CE) and the mechanical stabilities of its associated materialsover polymer substrates are either rarely reported or were foundproblematic. For instance the carbonaceous materials in theirvarious forms such as activated carbon,13–15 carbon black,16,17

and graphene18,19 have widely been tested in DSSCs both onrigid uorine doped tin oxide (FTO) coated glass and exibleITO-PET or ITO-PEN polymer substrates as a potentially inex-pensive alternative catalyst material to replace the traditionalexpensive platinum catalyst layer. Nevertheless all the above-mentioned reports do not speak about the mechanical stabili-ties of the deposited material. Also in our previous experiments,the replacement of ITO with single walled carbon nanotubes(SWCNTs) over the PET polymer was demonstrated by utilizinga dry transfer process and a metal free counter electrode wasfabricated which exhibited a 4% efficiency when used in the

J. Mater. Chem. A

Journal of Materials Chemistry A Communication

Publ

ishe

d on

28

Aug

ust 2

014.

Dow

nloa

ded

by A

alto

Uni

vers

ity o

n 29

/10/

2014

21:

23:1

9.

View Article Online

DSSC.20 However the adhesion of this conductive SWCNT lmwas also not satisfactory. Additionally the detachment of carboncomposite nanoparticles and aking off the carbon compositescatalyst lm over polymer sheets were clearly evident in ourother study2 which affected mainly the ll factor (FF) of the IVcurves and reduced the overall efficiency.2

As a remedy to these adhesion problems, a water-basedSWCNT ink was recently developed that can be simply spreadon the substrate and dried at 100–120 �C for 5 min to create ca.8–10 mm thick SWCNT network lms with a sheet resistance ofca. 22 U sq�1.21 Notable for this ink was its excellent adhesion toPVC: the SWCNT lms deposited on PVC sheets withstood bothrepetitive bending at least down to 2.5 cm radius of curvatureand repetitive tape adhesions tests with only a few percentincrease in sheet resistance. When the lms were further spin-coated with a thin layer of PEDOT-TsO conducting polymer thatfunctioned as an electrocatalyst, the PVC/SWCNT/PEDOT lmscould be used as a well performing Pt- and ITO-free counterelectrode in DSSC.21 Nevertheless, in that work satisfactoryadhesion of the SWCNT was observed only on PVC, presumablydue to preferential wettability properties of PVC compared toother polymers.

Here we report that with the above-mentioned ink, excellentadhesion of the SWCNT lms can be achieved also to PETsubstrates, provided that the ink is let to age overnight atelevated temperature before deposition. We thus repeated theprevious study (ref. 21), with small modications to the ink andsample preparation, and communicate the results here.

The main difference from ref. 21 was an overnight aging ofthe ink at 50 �C, which further increased the viscosity of the inkand therefore its wetting properties, presumably by completingthe micelle formation of the surfactant (dodecyl benzenesulfonic acid sodium salt, SDBS) over the SWCNT. It wasobserved that similar viscosity increase occurs also at roomtemperature but takes longer time (several days).

Fig. 1 (a–d): Demonstration of the process (a) tape mask was framedto obtain the patterns. (b) Ink spread through commonly availablepipette. (c) Solvent drying at 120 �C. (d) SWCNT conductive patterns ofvarious widths.

J. Mater. Chem. A

The deposition scheme is shown in Fig. 1(a–d) where theconductive patterns of the SWCNT were obtained by rstmasking the substrate with tape (Fig. 1a) followed by spreadingthe aqueous SWCNT ink over active areas with a commonlyavailable disposable pipette (Fig. 1b). Then the substrate wasplaced on a preheated hotplate at 120 �C. Upon solvent drying inless than 5 minutes (Fig. 1c), the nonconductive PET polymerfoil was rapidly transformed into a highly conductive substrate(Fig. 1d).

The mechanical stability of the deposited SWCNT wasexamined with bending and tape adhesion tests as in ref. 21,but with the difference that also a stronger adhering tape (3MMagic) was used and the samples were bent even down to 1 cmradius of curvature. The tests conrmed excellent adhesion ofthe SWCNT to PET.

Fig. 2 represents the results of a SWCNT loaded PET foilstripe (L¼ 5 cm andW¼ 1 cm) which was consecutively bent (10times) over different bending radii (ranged from 2.5 cm to 1 cm)and relaxed and its sheet resistance (RSH) was regularlymeasured. This conductive PET foil exhibited remarkable elas-ticity and its RSH almost remained unchanged (�25 U Sq�1)when sequentially bent over 2.5 cm, 2 cm and 1.5 cm bendingradii (Fig. 2a–c). The marginal change in RSH (<5%) wasobserved when the stripe was further bent over a severely small(1 cm) bending radius. Hence the overall change in the RSH fromthe very rst relax mode value (�25 U Sq�1) of 2.5 cm bendingradius till the last relaxed value of 1 cm (�25.7 U Sq�1) was only2.8% showing the excellent elasticity of the deposited SWCNT.

The further strength of the deposited SWCNT over PETpolymer foil was also examined with a surface adhesion testwith two different types of pressure sensitive tapes (3MRemovable and 3M Magic). The SWCNT loaded PET polymersubstrate (W¼ 2 cm and L¼ 2.1 cm) was rolled down with a 2 kgmetallic roller disk under the pressure sensitive tapes and the

Fig. 2 Bending test of SWCNT coated PET foil over (a) 2.5 cm bendingradius, (b) 2 cm bending radius, (c) 1.5 cm bending radius and (d) 1 cmbending radius.

This journal is © The Royal Society of Chemistry 2014

Communication Journal of Materials Chemistry A

Publ

ishe

d on

28

Aug

ust 2

014.

Dow

nloa

ded

by A

alto

Uni

vers

ity o

n 29

/10/

2014

21:

23:1

9.

View Article Online

tapes were pulled out at 90� as shown in Fig. 3(a–d). The effect of‘pressing’ was seen when the tested substrate experienced theheavy rolling under tape A (3M Removable, Fig. 3a and b) whichmay have initially increased the contact adhesion between theSWCNT network and resulted an initial decrement (2.4%) inoverall RSH of the substrate with the rst two tape adhesions(Fig. 3b). This decrement was maintained with further threemore consecutive tape adhesions and pulling and no signicantchanges were observed (Fig. 3b).

To conrm the extreme strength, the adhesion of theSWCNT/PET substrate was further tested under more stickytape (3M Magic) which was again consecutively rolled downwith the 2 kg metallic roller disk and pulled at 90� (ve times)and its RSH was measured (Fig. 3c and d). This introduced agradual but marginal increase (2%) in the RSH values (Fig. 3d).However the net effect on the change in RSH aer 10 times heavyrolling and tape pulling was <1% which means almost nochange from the initial RSH value. These results certied thesuperb adhesion of the deposited SWCNT over the PET polymer.

The mechanism of this remarkable adhesion of the SWCNThas already been explained in our previous report21 i.e. the goodwetting of the exposed surface (PET polymer here) which can beobtained by reducing the surface tension of the ink formulationand increasing the interaction forces between the substrate andthe ink.21 The added surfactant sodium dodecyl benzenesulfonic acid sodium salt (SDBS) plays a key role in many ways;it increases the solubility of dispersed SWCNTs in deionizedwater (DIW), separates the individual SWCNTs22 and reducesthe surface tension of the ink rheology.21,22

Therefore, inspired by these fascinating mechanical stabilityresults, this highly conductive substrate was tested as a counterelectrode in the dye-sensitized solar cell where the conductivity/sheet resistance requirement of the substrates typically ranges

Fig. 3 Tape adhesion test (a) Tape pulling (3M Removable) at 90� (b)RSH versus number of tape pulling (3M Removable). (c) Tape pulling(3M Magic) at 90� (d) RSH versus number of tape pulling (3M Magic).

This journal is © The Royal Society of Chemistry 2014

from 5 U Sq�1 to 60 U Sq�1.23 However, the SWCNTs alone werenot found catalytic enough in our earlier experiments,20,21

therefore the electro-polymerization step over SWCNTs20 wasperformed (See Experimental section) to improve the catalyticproperties of SWCNTs.

The scanning electron microscopy (SEM) images of thedeposited SWCNT over PET foil and the electro-polymerizedSWCNT layer are shown in Fig. 4(a and b). Fig. 4a reveals ahighly porous but random network of SWCNT bundles homo-genously distributed over PET. No visible agglomeration wasdetected in the SWCNT suggesting a stable dispersion ofnanotubes in the solvent. The typical length of an individualnanotube seems to be in the range 1–2 mm (Fig. 4a). Moreover,some of the surfactant particles (50–150 nm) which were addedto promote the solubility as well as separation of the individualnanotubes settled in the voids (<100 nm) of the porous nano-tubes network (Fig. 4a). On the other hand, the polymerizationstep gave very thin coverage of the PEDOT polymer over theSWCNT which also penetrated into the voids of the nanotubenetwork (Fig. 4b).

Finally, to verify the applicability of the PET/SWCNT/PEDOTlms as DSSC counter electrodes, test solar cells were prepared

Fig. 4 SEM images of (a) single walled carbon nanotube network withsurfactant particles (b) PEDOT coverage over the SWCNT network.

J. Mater. Chem. A

Fig. 5 IV curve of the TPCE and PET/SWCNT/PEDOT CE based DSSCand measured under one sun equivalent light intensity. TPCE ¼ ther-mally platinized counter electrode. PET/SWCNT/PEDOT CE ¼ poly-ethylene terephthalate/single walled carbon nanotube/PEDOTcounter electrode.

Journal of Materials Chemistry A Communication

Publ

ishe

d on

28

Aug

ust 2

014.

Dow

nloa

ded

by A

alto

Uni

vers

ity o

n 29

/10/

2014

21:

23:1

9.

View Article Online

and their photovoltaic performance was measured. The resultsare presented in Fig. 5 and Table 1.

Note that rigid FTO-glass was used instead of exible plasticsince the objective of the experiment was not to realize a fullyexible DSSC, but to characterize the performance of theSWCNT/PEDOT CE in a DSSC prepared otherwise from wellperforming and well known materials. Glass substrates facili-tated also reliable lling and sealing of the electrolyte, and thehigh photocurrent densities generated by standard FTO-glassbased PEs allowed challenging the charge transfer capability ofthe CEs at conditions relevant for high efficiency cells. In ahypothetical but plausible manufacturing scenario, these ex-ible plastic CEs foils, thanks to their high mechanical dura-bility, could be manufactured with fast and cost-effective roll-to-roll techniques and laminated together with standard highperformance FTO-glass based photoelectrodes, as a compro-mise between low manufacturing costs and high energyconversion efficiency.

The fabricated exible PET/SWCNT/PEDOT counter elec-trode based DSSC exhibited excellent photovoltaic characteris-tics such as high ll factor (�70%), high short circuit currentdensity (JSC ¼ 14.4 mA cm�2) and open circuit voltage (VOC ¼702 mV) and efficiency (7%) that are reasonably close to thephotovoltaic parameters of the 8.5% efficient thermally

Table 1 IV and EIS parameters of each type of DSSCmeasured under onecalculated based on the photoelectrode area of 0.16 cm2. Series and coresistance (RCELL) is determined by the slope of IV curves at VOC

Cell typea JSC (mA cm�2) VOC (mV) F

TPCE 14.9 738 7PET/SWCNT/PEDOT CE 14.4 702 7

a TPCE ¼ Thermally platinized counter electrode. PET/SWCNT/PEDOT CEcounter electrode.

J. Mater. Chem. A

platinized counter electrode (TPCE) based DSSC prepared usingsame procedures (Fig. 5, Table 1). (We mention here that in thisstudy, these procedures differed from the typical preparation ofoptimized ca. 11% efficient TPCE DSSCs,24 as described in theExperimental section, and thus a lower efficiency was expected).Nevertheless, higher cell resistance (RCELL ¼ 58 U, corre-sponding to the slope of the IV curve at VOC) was observed inPET/SWCNT/PEDOT CE based DSSC compared to the RCELL (27U) of the TPCE DSSC (Table 1) which is an expected resultprimarily due to the four times lower RSH (7 U Sq�1) of FTOcoated glass compared to RSH (25–35 U Sq�1) of SWCNTs. Thelower RSH of the FTO glass/thermally platinized counter elec-trode (TPCE) is in line with the higher FF (76%, Fig. 5, Table 1)compared to PET/SWCNT/PEDOT.

Also the total ohmic series resistance RS, measured by elec-trochemical impedance spectroscopy (EIS) at VOC under illu-mination (data not shown), was higher in the PET/SWCNT/PEDOT cells (30U, Table 1) than in the TPCE cells (9U, Table 1).On the other hand, the charge transfer resistance at the counterelectrode RCT was similar in both cases: 0.4 U cm2 with PET/SWCNT/PEDOT and 0.5 Ucm2 with TPCE (data not shown). Thisindicates that both CE materials had similar electro-catalyticperformance for the tri-iodide reduction reaction. The lowcharge transfer resistance of the PEDOT catalyst layer has beenalready reported in the literature.25 The result also conrms thatthe FF was lower in the PET/SWCNT/PEDOT cells mainly due tothe higher RSH as anticipated based on the direct RSH

measurements discussed above, and not due to catalyticperformance limitations.

This study also conrms the results of our previous report21

where similar characteristics such as high adhesion and goodRSH of the deposited SWCNT were achieved on a PVCsubstrate.21 Also there, a lower FF was seen due to slightlyhigher RSH compared to the DSSC with TPCE. In the future, theobjective of the research will be to decrease RSH by adjusting theconcentration of the surfactant sodium dodecyl benzenesulfonic acid sodium salt (SDBS) particles without compro-mising the adhesion characteristics of SWCNTs. Increasing theSWCNT lm thickness from the present ca. 5–10 mm should inprinciple decrease RSH and optimization of the amount ofPEDOT could potentially decrease RCT. Future studies shouldalso address the adhesion of the PEDOT or the alternativecatalyst material on the SWCNT lm, although we do not expectit to be as problematic and critical as achieving good adhesionof the much thicker SWCNT lm on the substrate.

sun equivalent light intensity. The short circuit current density JSC wasunter electrode resistance (RS and RCE) are measured by EIS. Total cell

F (%) h (%) RCELL (U) RS (U) RCT (U cm2)

6 8.5 27 9 0.50 7.0 58 30 0.4

¼ Polyethylene terephthalate – single walled carbon nanotube – PEDOT

This journal is © The Royal Society of Chemistry 2014

Communication Journal of Materials Chemistry A

Publ

ishe

d on

28

Aug

ust 2

014.

Dow

nloa

ded

by A

alto

Uni

vers

ity o

n 29

/10/

2014

21:

23:1

9.

View Article Online

Hence this highly conductive SWCNT coated PET foil incombination with PEDOT unveils amazing characteristics interms of very high elasticity, exibility, durability and remark-able potential when tested as the counter electrode in the DSSC.In the future development, a rolling or pressing step couldperhaps also be used to further reduce the overall sheet resis-tance of the substrate as by the results from the tape adhesiontest.

In summary, we demonstrated rapid transformation of acommonly available PET polymer foil into a highly conductingITO free substrate through an easy and quick deposition tech-nique of an aqueous SWCNT ink. This metal free SWCNT coatedexible PET foil exhibited a marginal (2.8%) change in the sheetresistance when subjected to severe repetitive bending testsdown to 1 cm bending radius. The strong mechanical stabilitywas also conrmed with repetitive tape adhesion tests thatshowed a marginal net change in the sheet resistance. Thissubstrate was then tested for counter electrode application inthe dye-sensitized solar cell yielding high efficiency (7%) andhigh catalytic activity in terms of very low charge transferresistance (RCT ¼ 0.4 U cm2).

In conclusion, thanks to an additional aging step in the inkpreparation, the SWCNT lms prepared from a water-basedSWCNT ink adhere well not only to PVC but also to PET polymerfoil that is a more commonly used substrate material in exiblesolar cells and printed electronic devices. The combination ofgood adhesion and low sheet resistance makes these SWCNTlms an interesting electrode candidate for various electro-chemical devices besides the dye-sensitized solar cells demon-strated here.

Experimental setupMaterials

Single walled carbon nanotubes (purity > 90%, bundle length 1–5 mm, diameter 4–5 nm) were purchased fromCarbon Solutions.Surfactant sodium dodecyl benzene sulfonic acid sodium saltwas purchased from Fluka Chemica. Monomer 3,4-ethyl-enedioxythiophene (EDOT), lithium perchlorate (LiClO4).Chloroplatinic acid hydrate (H2PtCL4$6H2O) and all thesolvents were obtained from Sigma Aldrich. The C101 dye wassynthesized as reported previously.24

Counter electrodes (CE)

Thermal platinization. Thermal platinization on FTO coatedglass (TEC 7) was performed by drop casting (1 drop) 10 mMchloroplatinic acid hydrate (H2PtCl4$6H2O) solution (2 prop-anol) followed by sintering the substrates at 410 �C for 20minutes. The substrates were then cooled down to roomtemperature and were sealed in a plastic box before use.

CNT ink formulation. 0.05 g of single walled carbon nano-tubes were rst added with 0.1 g of surfactant sodium dodecylbenzene sulfonic acid sodium salt in 10 ml of deionized water(DIW). The solution was then sonicated with a tip sonicator(Branson model 450D, pulse on 1 s, pulse off 2 s) for 30 minutesand a well dispersed ink was obtained at the end of the process.

This journal is © The Royal Society of Chemistry 2014

The ink was additionally aged overnight at 50 �C to increase theviscosity of the solution. Except for the aging, this ink formu-lation followed ref. 21 and was not optimized further.

Dip coating of polyethylene terephthalate (PET) polymersubstrates in SWCNT ink.Highly conductive SWCNT layers overPET foil were obtained by directly dipping the PET polymersubstrates into the SWCNT ink. The substrates were then takenout from the SWCNT ink contained and dried at 120 �C on apreheated hotplate for 5 minutes. Upon solvent evaporation in 5minutes, highly adhesive SWCNT patterns (5 mm) wereobtained. The thickness of the SWCNT lm was determinedthrough a cross sectional SEM image. This ink depositionprocedure and the obtained lm thickness were similar to ref.21 and were not optimized further.

Electro-polymerization of SWCNT coated PET substrates.Monomer 3,4-ethylenedioxythiophene (EDOT) was electro-polymerized at +1 volt (as described in our previous work20) overthe SWCNT loaded PET substrate via chronocoulometry in anaqueous solution containing 0.3 M LiClO4 and 0.3 mM EDOT.The deposition rate and charge was kept constant for 60 s andaround 70 mC cm�2, respectively. Note that in ref. 21 thePEDOT deposition was carried out by spin coating with essen-tially similar results.

Photoelectrodes (PE)

The photo electrodes for this experiment were prepared asdescribed before.26 In short the FTO coated glass substrates(Sheet resistance 15 U Sq�1, Pilkington) were rst cleaned withmild detergent and then were sequentially sonicated in acetoneand ethanol solutions (15 min each) and then dried withcompressed air. Further cleaning of the substrate was per-formed for 25 minutes in UV-O3 cleaner (UVO-Cleaner, Model256-220, Jelight Company Inc, USA) and then the substrate wasdirectly immersed in 40 mM TiCl4 solution and heated in apreheated oven at 70 �C for 30 minutes in a closed container.Then all the substrates were sequentially rinsed with deionizedwater and ethanol solutions and were dried with compressedair. 8 mm and 5 mm thick layer of nanocrystalline TiO2 particles(20 nm) and scattering particles (400 nm) respectively weresequentially deposited via screen printing on the TiCl4 treatedsubstrates and heated at 110 �C for 5 minutes. The depositedlayers were then sintered at 500 �C for 30 minutes and werecooled down to room temperature. Aer that the TiCl4 treat-ment step was repeated as described above and the substrateswere then sintered again at 500 �C for 30 minutes and werecooled down to room temperature to complete the sequence.Then the photo electrodes were sensitized with dye C 101. Theactive area of the photoelectrode was (0.16 cm2). The photo-electrodes are not the focus of this work, but were employed as astandard component in the test cells that we prepared tocharacterize the performance of the PET/SWCNT/PEDOT CEs inthe DSSC.

Cell assembly

The PE and CE were separated with a thick (35–70 mm) Bynelframe foil that also denes the cell channel. A quick drying

J. Mater. Chem. A

Journal of Materials Chemistry A Communication

Publ

ishe

d on

28

Aug

ust 2

014.

Dow

nloa

ded

by A

alto

Uni

vers

ity o

n 29

/10/

2014

21:

23:1

9.

View Article Online

epoxy (Magic Epoxy – Pakistan) was also employed at the borderof cells with a exible counter electrode based DSSC to avoid anypossible leakage of the liquid electrolyte. Then the electrolyte(Z960 i.e. 1 M DMII, 0.03 M I2, 0.05 M LiI, 0.5 M tBPy, and 0.1 MGuSCN in Acetonitrile : Valeronitrile ¼ 85 : 15) was injectedthrough the drilled hole at the PE side. The cells were thensealed with 15 mm thick Surlyn foil and a thin glass cover. Thesilver contacts were then fabricated with an ultrasonic solderingsystem (MBR electronics, Model USS-9200) on both the FTOglass based PE and CE whereas low temperature silver ink(Plano GmbH) was coated on the non-active area of SWCNT/PEDOT deposited PET polymer based counter electrodes. Weindicate here that the preparation procedure of the DSSCs forthis study were somewhat different from that of optimizedTPCE DSSCs24 as follows: (1) a less conductive and less trans-parent FTO-glass was used (TEC-15 instead of NSG-10), (2) thedrilling of the electrolyte lling holes and their sealing weredone for both cell types on the photoelectrode side instead ofthe counter electrode side, and (3) no antireection coatings orlms were used at the photoelectrode side. Note that the testcell preparation, except for the CEs, was somewhat differentthan in ref. 21, and hence, comparison of the present resultswith ref. 21 can be made only with respect to the adhesioncharacteristics and sheet resistance of the PET/SWCNT lms.The higher cell efficiency obtained here (7% vs. 5% in ref. 21) ismainly due to different PE preparation, smaller active area, andless viscous electrolyte, which gave higher efficiency also for theTPCE cell (8.5% vs. 5.2% in ref. 21).

Measurements

The photovoltaic (IV curves) measurement was performed in anarticial solar simulator under 1000 W m�2 light intensityequivalent to 1 Sun by applying external potential bias and atthe same time the current was recorded via a digital sourcemeter (Keithley, Model 2400). A 450 W Xenon lamp (Oriel) wasused as a light source which was loaded with a sunlight lter(Schott K113) to minimize mismatch between the spectrum ofthe light source and the AM 1.5G. The light was incident on thecell from the photoelectrode side. Due to the high absorbance ofthe PEs and low reectance of both type of studied CEs, theeffect of back-reection of light from the CEs is assumed to havea negligible effect on the short-circuit current density of thesecells, as has been shown experimentally for similar materials inthe ESI of ref. 21. The impedance spectra of the cells wererecorded with an electrochemical impedance spectrum analyzer(Zahner-Elektrik IM6 potentiostat) whereas Zview2 soware wasused to t the measured data with our previously reportedequivalent circuit. Scanning electron microscopes Zeiss Ultra 55FEG-SEM and JEOL JSM-7500FA were used to record the imagesof the samples.

Mechanical durability test

Bending test. The SWCNT pattern was coated on PET foil (L¼ 5 cm, W ¼ 1 cm, and thickness � 5–10 mm) and then wasconsecutively bended and relaxed for 10 times over differentbending radii (ranged from 2.5 cm to 1 cm) and the resistance

J. Mater. Chem. A

(R) of the SWCNT lm was measured. The sheet resistance wascalculated by using the formula i.e. RSH ¼ (RW)/L whereW and Lare the width and length of the lm respectively between theelectrical contacts.

Tape adhesion test. Two different types of pressure sensitivetapes (3M Removable and 3M Magic denoted as tape A and B,respectively) were rst applied (ve times for each tape) over theSWCNT coated PET foil (L¼ 2.1 cm andW¼ 2 cm). The stickingwas assured by rolling over the tape with a 2 kg metallic diskmultiple times and then the tape was pulled at 90� angle. Theresistance of the substrate was measured aer each tape pull-ing. The sheet resistance was then calculated by using the sameformula as mentioned in the bending test section.

Acknowledgements

This work was funded from National Consortium for Low CostPhotovoltaic from the Academy of Finland (Grant number13135143). Ghufran Hashmi thanks Tekniikan edistamissaatio(Finnish Foundation for Technology Promotion) for the travelgrant and Mr Janne Patakangas for the help in the bending test.MG thanks the ECR Advanced Grant Agreement no. 247404under the CE-Mesolight project and Swiss National ScienceFoundation for partial nancial support.

References

1 G. Hashmi, K. Miettunen, T. Peltola, J. Halme, I. Asghar,K. Aitola, M. Toivola and P. Lund, Renewable SustainableEnergy Rev., 2011, 15, 3717–3732.

2 S. G. Hashmi, J. Halme, T. Saukkonen, E. Rautama andP. Lund, Phys. Chem. Chem. Phys., 2013, 15, 17689–17695.

3 T. Miyasaka and Y. Kijitori, J. Electrochem. Soc., 2004, 151,A1767–A1773.

4 H. C. Weerasinghe, P. M. Sirimanne, G. P. Simon andY. B. Cheng, Prog. Photovoltaics, 2012, 20, 321–332.

5 S. Ito, P. Chen, P. Comte, M. K. Nazeeruddin, P. Liska,P. Pechy and M. Gratzel, Prog. Photovoltaics, 2007, 15, 603–612.

6 N. Papageorgiou, W. F. Maier and M. Gratzel, J. Electrochem.Soc., 1997, 144, 876–884.

7 T. Yamaguchi, N. Tobe, D. Matsumoto and H. Arakawa,Chem. Commun., 2007, 4767.

8 H. C. Weerasinghe, P. M. Sirimanne, G. P. Simon andY. B. Cheng, J. Photochem. Photobiol., A, 2009, 206, 64.

9 H. Lee, D. Hwang, S. M. Jo, D. Kim, Y. Seo and D. Y. Kim, ACSAppl. Mater. Interfaces, 2012, 4, 3308.

10 Y. Han, J. M. Pringle and Y. Cheng, RSC Adv., 2014, 4, 1393.11 H. C. Weerasinghe, F. Huang and Y. B. Cheng, Nano Energy,

2013, 2, 174–189.12 T. Miyasaka, M. Ikegami and Y. Kijitori, J. Electrochem. Soc.,

2007, 154, A455–A461.13 J. Chen, B. Li, J. Zheng, J. Zhao, H. Jing and Z. Zhu,

Electrochim. Acta, 2011, 56, 4624.14 J. Chen, K. Li, Y. Luo, X. Guo, D. Li, M. Deng, S. Huang and

Q. Meng, Carbon, 2009, 47, 2704.

This journal is © The Royal Society of Chemistry 2014

Communication Journal of Materials Chemistry A

Publ

ishe

d on

28

Aug

ust 2

014.

Dow

nloa

ded

by A

alto

Uni

vers

ity o

n 29

/10/

2014

21:

23:1

9.

View Article Online

15 G. Veerappan, K. Bojan and S. Rhee, Renewable Energy, 2012,41, 383.

16 J. M. Kim and S. W. Rhee, Electrochim. Acta, 2012, 83, 264–270.

17 W. Huang, X. Zhang, F. Huang, Z. Zhang, J. He and Y. Cheng,J. Nanopart. Res., 2012, 14, 838.

18 L. Wan, S. Wang, X. Wang, B. Dong, Z. Xu, X. Zhang, B. Yang,S. Peng, J. Wang and C. Xu, Solid State Sci., 2011, 13, 468.

19 J. D. Roy-Mayhew, D. J. Bozym, C. Punckt and I. A. Aksay, ACSNano, 2010, 4, 6203.

20 K. Aitola, M. Borghei, A. Kaskela, E. Kemppainen,A. G. Nasibulin, E. I. Kauppinen, P. D. Lund, V. Ruiz andJ. Halme, J. Electroanal. Chem., 2012, 683, 70.

21 S. G. Hashmi, J. Halme, Y. Ma, T. Saukkonen and P. Lund,Adv. Mater. Interfaces, 2014, 1, 1300055.

This journal is © The Royal Society of Chemistry 2014

22 M. F. Islam, E. Rojas, D. M. Bergey, A. T. Johnson andA. G. Yodh, Nano Lett., 2003, 3, 269.

23 G. Hashmi, K. Miettunen, J. Halme, I. Asghar, H. Vahlman,T. Saukkonen, Z. Huaijin and P. Lund, J. Electrochem. Soc.,2012, 159, H656–H661.

24 F. Gao, Y. Wang, D. Shi, J. Zhang, M. Wang, X. Jing,R. H. Baker, P. Wang, S. M. Zakeeruddin and M. Gratzel, J.Am. Chem. Soc., 2008, 130, 10720–10728.

25 S. Ahmad, J. Yum, Z. Xianxi, M. Gratzel, H. Butt andM. K. Nazeeruddin, J. Mater. Chem., 2010, 20, 1654–1658.

26 S. Ito, T. N. Murakami, P. Comte, P. Liska, C. Gratzel,M. K. Nazeeruddin and M. Gratzel, Thin Solid Films, 2008,516, 4613–4619.

J. Mater. Chem. A