Solar Energy Materials & Solar Cells 96 (2012) 244–250

Contents lists available at SciVerse ScienceDirect

Solar Energy Materials & Solar Cells

0927-02

doi:10.1

n Corr

E-m

lars.mu

journal homepage: www.elsevier.com/locate/solmat

Semi-transparent small molecule organic solar cells with laminatedfree-standing carbon nanotube top electrodes

Yong Hyun Kim a,n, Lars Muller-Meskamp a,n, Alexander A. Zakhidov b, Christoph Sachse a, Jan Meiss b,Julia Bikova c, Alexander Cook c, Anvar A. Zakhidov c, Karl Leo a,b

a Dresdner Innovationszentrum fur Energieeffizienz, Institut fur Angewandte Photophysik, Technische Universitat Dresden, 01062 Dresden und Fraunhofer IPMS, 01109 Dresden,

Germanyb Institut fur Angewandte Photophysik, Technische Universitat Dresden, 01062 Dresden, Germanyc Alan G. MacDiarmid NanoTech Institute, University of Texas at Dallas, Richardson, TX 75083, USA

a r t i c l e i n f o

Article history:

Received 29 August 2011

Received in revised form

26 September 2011

Accepted 2 October 2011Available online 19 October 2011

Keywords:

Organic solar cells

Semi-transparent devices

Transparent conductive electrodes

Carbon nanotubes

Lamination

Orthogonal liquids

48/$ - see front matter & 2011 Elsevier B.V. A

016/j.solmat.2011.10.001

esponding authors. Tel.: þ49 351 463 32748

ail addresses: yong.hyun.kim@iapp.de (Y.H. K

eller-meskamp@iapp.de (L. Muller-Meskamp)

a b s t r a c t

We have developed semi-transparent n–i–p organic solar cells (OSC) with free-standing multi-wall

carbon nanotube (f-CNT) sheets as transparent top electrodes. By a simple and damage-free room

temperature orthogonal liquid solution assisted self-laminating process, f-CNT top electrodes are

successfully deposited on top of ZnPc:C60 bulk heterojunction small molecule OSCs. The cells show high

fill factors above 58% as well as efficiencies up to 1.5% and greater long-term stability compared to the

device having a metal electrode. For the given cell structure with f-CNT semi-transparent electrodes,

the influence of an optical spacer on light absorption is studied by a systematic variation of the hole

transport layer thickness, supported by optical simulations. The results strongly indicate that OSCs with

f-CNT top electrodes and optimized thin film stack are highly promising for semi-transparent OSCs,

which can be used in tandem devices, in tinted smart windows, and similar applications by a simple

and damage-free process in roll-to-roll configuration that can be scaled to large area manufacturing.

& 2011 Elsevier B.V. All rights reserved.

1. Introduction

Organic solar cells (OSC) have attracted significant attention asa promising source of renewable energy. In the recent years,much effort has been devoted to the research of OSCs architec-tures and processing and their efficiency has grown significantly.The efficiencies of both thermally evaporated small molecule andwet-chemically processed polymer based OSCs have beenincreased to 8.3% [1], with further potential for improvement. Inaddition to a conventional bulk heterojunction structure, investi-gated widely due to an efficient exciton dissociation, novelconcepts such as a layer-by-layer assembly [2], block-copolymernanowires [3], quantum dots [4], and hybrid structures [5], arealso being investigated extensively. Among numerous OSCsadvantages such as potentially low cost, low material consump-tion, flexibility, and suitability for roll-to-roll mass production, anunusual and promising property is the possible semi-transpar-ency of OSCs devices. Semi-transparent thin film OSCs tuned for aparticular visible spectrum absorption band can be used for high

ll rights reserved.

; fax: þ49 351 463 37065.

im),

.

efficiency tandem stack solar cells, which can be both of conven-tional tandem type connected in series [6] and also of a parallelconnection type [7], particularly enabled by carbon nanotubes(CNT), transparent sheet common electrode connecting two ormore cells in multijunction OSCs. Furthermore, semi-transparentOSCs allow the installation of solar cells on windows or otherarchitectural building elements, facilitating an efficient energymanagement of a building or a car by providing artisti-cally colored and tinted light, shadow, and at the same timeelectricity.

The fabrication of novel efficient transparent conductive elec-trodes for both top and bottom contact is of great importance forsemi-transparent OSCs. Especially for sensitive evaporated smallmolecule OSCs, the deposition process of the top electrode needsto be very gentle without damaging underlying organic layers,which can be easily shorted due to o100 nm typical thickness,giving rise to many practical difficulties. So far, metal oxides [8,9],conducting polymers [10,11], silver nanowires [12,13], and thinmetal layers [14,15] have been employed as a top electrode forOSCs and organic light emitting diodes (OLED). However, avoidingdamage to underlying organic layers during the deposition ofinorganic materials as well as complex, expensive, slow, andsmall scale deposition processes of top electrodes are technolo-gical issues, which are still challenging. The development

Y.H. Kim et al. / Solar Energy Materials & Solar Cells 96 (2012) 244–250 245

of vacuum evaporation free and large area technology for OSCstop electrodes remains one of major challenges in this field,particularly for large scale roll-to-roll processing on flexiblesubstrates.

CNTs are promising as transparent conductive electrodes foroptoelectronic devices because of their excellent electrical andoptical performance as well as mechanical flexibility, chemicalstability, and low-cost roll-to-roll compatible manufacturingprocess. The application of CNTs as electrodes for OSCs and OLEDshas been focused on the use of single-wall CNTs (SWCNT) via thesolution based processes and as bottom electrodes for opaquedevices [16–20]. The roll-to-roll compatible free-standing multi-wall CNT (f-CNT) sheets reported by Zhang et al., which are highlyconductive, transparent, and show high mechanical strength, area very promising top electrode material [21]. One of the greatadvantages of the dry span f-CNT sheets is the scalable fabricationof meter-long sheets with uniform electrical and optical proper-ties. Those f-CNT sheets offer unique applications and weresuccessfully adopted as a bottom electrode for OSCs and OLEDs,showing reasonable performances [22,23]. Moreover it has beendemonstrated that f-CNT sheets, which have inherently highporosity, can provide three-dimensional (3-D) charge collectionfrom the volume of the bulk heterojunction OSCs, with twiceincreased photo-current [22]. Additionally, Tanaka et al. have firstreported parallel tandem architecture of OSCs with f-CNT as anintermediate electrode, showing increased total photo-currentalthough at yet small efficiency of 0.31% [24]. Parallel tandemarchitectures require the bottom cell to be semi-transparent, andone of subcells to be inverted in such a way that the commonintermediate CNT electrode collects charges of same sign, addingto total current, contrary to the conventional series tandemstructure in which opposite charge carriers recombine at theinterlayer [25,26].

In this study, we report on such inverted semi-transparentsmall molecule n–i–p OSCs with f-CNT top electrodes, which canbe used in both types of tandem stack improving their perfor-mance. The top f-CNT electrodes are deposited on OSCs using aroom temperature orthogonal liquid solution assisted self lami-nating process, which is simple and damage-free for organiclayers. The OSCs with f-CNT top anode electrodes (f-CNT cell)show very low leakage currents, high fill factors and promisingefficiencies, transparencies, and long-term stability. The OSCsparameters are systematically studied, with a variation ofp-doped hole transport layer (HTL) thickness adjacent to topf-CNT anodes. The optimal optical field distribution in these OSCsis investigated by optical simulations based on the transfer-matrix formalism. These results show that f-CNT cells can serveas highly promising semi-transparent OSCs. In this report, free-standing dry drawn CNT sheets are applied as a top electrode forOSCs, providing a basis for parallel tandem OSCs or transparentapplications.

2. Experimental

Small molecule OSCs are prepared by thermal evaporation in ahigh vacuum system (K.J. Lesker, U.K.). The base pressure of thechamber is kept around 10�8 mbar. The layer sequence for solarcells is inverted, as compared to usual direct OSCs with topAl cathode, since it has a bottom ITO cathode with n-dopedelectron transport layer (ETL) and a top f-CNT anode as follows(bottom to top): ITO as a bottom cathode/10 nm tetrakis(1,3,4,6,7,8-Hexahydro-2H-pyrimido[1,2-a]pyrimidinato)ditungsten(II) (W2(hpp)4) doped C60 as an ETL/30 nm C60 as an additionalabsorber layer and ETL/30 nm mixed zinc phthalocyanine(ZnPc):fullerene C60 (volume ratio of 1:2) as a photoactive absorber

layer/X nm 10 wt.% 2,20-(perfluoronaphthalene-2,6-diylidene)dima-lononitrile (F6TCNNQ) doped N,N0-((diphenyl-N,N0-bis)9,9,-dime-thyl-fluoren-2-yl)-benzidine (BF-DPB) as a hole transport layer(HTL)/1 nm p-type dopant as a hole injection layer/a top anode.The stack is visualized in Fig. 1(a).

For the reference solar cell (Ag cell), the thickness of the HTL is50 nm and a thin metal layer (Al 1 nm and Ag 14 nm) isevaporated as a top electrode. Subsequently, 60 nm of tris(8-hydroxy-quinolinato)-aluminum (Alq3) is deposited on thetop as a capping layer increasing the transparency of the Ag cell.The sheet resistance and transparency of this type of referencemetal electrode are around 5 O sq�1 and 64.7% (at 550 nm, withAlq3), respectively.

For the f-CNT cells, the HTL thicknesses are varied from 20 to80 nm. The f-CNT sheets (sheet resistance: �250 O sq�1 andtransparency: �38% (at 550 nm)) are prepared from a multi-walled CNT forest produced by a chemical vapor deposition (CVD)process at the Alan G. MacDiarmid NanoTech Institute asdescribed previously [21]. The as-prepared f-CNT sheets aredirectly deposited by manual lamination assembly on top oforganic stacks in a nitrogen filled glove box without air exposure.Afterwards, the f-CNT cells are subsequently immersed intoorthogonal liquid hydrofluoroether (HFE) for several seconds todensify the f-CNT sheets (from typically 15–20 mm wide aerogelto 50–100 nm f-CNT film) on top of the solar cells. As shownbefore, such densification significantly improves the f-CNT elec-trode performance by increased transparency and improvedconductivity due to dense tube to tube interconnects withoutany device degradation [27]. The densified f-CNT sheet used inthis study turns into around 80 nm thick porous film. The solarcell active areas are 2.56–6.91 mm2 measured using an opticalmicroscope.

The current–voltage characteristics are examined with asource measurement unit (SMU 236, Keithley Instruments) underan AM 1.5G sun simulator (Steuernagel SC1200) in a nitrogenfilled glove box. The incident light intensity is monitored by asilicon reference diode, calibrated by Fraunhofer ISE (Freiburg,Germany). The values of short circuit current density are normal-ized to 100 mW cm�2 and are not corrected for spectral mis-match. The samples were reproduced several times, showingnegligible measurement deviation. External quantum efficiencymeasurements (details in Supplementary material) are carriedout using color filters in the same setup. The transparency of theelectrodes and solar cells is measured using a spectrophotometer(Perkin-Elmer Lambda 900), which includes the optical loss of theglass substrate (�8%). Optical simulations are done with atransfer-matrix formalism based simulation program [28,29].Optical constants of ITO, f-CNT, and organic layers are obtainedby transmission and reflection measurements, calculated fromseveral samples with different thicknesses on glass. The atomicforce microscopy (AFM) images of f-CNT are recorded in tappingmode (AIST-NT Combiscope).

3. Results and discussion

ZnPc:C60 bulk heterojunction based small molecule OSCs withn–i–p configuration are fabricated with top f-CNT electrodes.A schematic drawing of the fabrication process for f-CNT cells, aphotograph of the semi-transparent cell, atomic force microscopy(AFM), and scanning electron microscope (SEM) image of anf-CNT sheet on top of the device after densification are shownin Fig. 1. It is known that f-CNTs have predominant orientationwith respect to the direction of drawing. Laminated nanotubes onthe device keep their orientation, as visible in microscope images[21]. The averaged transparency of the f-CNT cell (HTL of 50 nm)

Fig. 1. (a) Fabrication process for f-CNT cell, (b) a photograph of f-CNT cell, (c) AFM (4�4 mm), and (d) SEM image of f-CNT sheet on the top of solar cell.

Y.H. Kim et al. / Solar Energy Materials & Solar Cells 96 (2012) 244–250246

in the visible region is about 27% (see Supplementary material).Fig. 2 shows typical current–voltage (I–V) characteristics of OSCsas a function of the different HTL thicknesses and illuminationfrom bottom (ITO side) and top (f-CNT side). The characteristicvalues of all devices are shown in Table 1. A thin metal layer (Al1 nm/Ag 14 nm) is employed as the top electrode for the refer-ence cell (Ag cell), using a similar stack to that reported for thesemi-transparent cells by Meiss et al. [15]. They showed opti-mized semi-transparent OSCs with thin Ag layer with powerconversion efficiencies (PCE) up to 2.2%.

The bottom illuminated Ag reference cell in our experimentshows a PCE of 1.9% (short circuit current (JSC)¼5.6 mA cm�2, fillfactor (FF)¼61.3%, and open circuit voltage (VOC)¼0.57 V).All f-CNT cells show high FFs (58�59%) comparable to that ofthe reference Ag cell but somewhat lower JSC. The highest

observed PCE of f-CNT cells is 1.5% with a JSC of 4.7 mA cm�2

for an HTL thickness of 20 nm and bottom illumination. With topillumination, the f-CNT cell with an HTL thickness of 50 nm showsa PCE of 0.9% and a FF of 59.4%, comparable to that of the Ag cell(PCE of 1.0% and FF of 60.4%). All f-CNT cells show very lowleakage current, showing saturation factors [J(�1 V) JSC

�1] below1.2, which are even lower than those of highly optimized ZnPc:C60

bulk heterojunction OSCs [14,30]. This result indicates thatneither the lamination process of top f-CNT sheets nor thedensification in orthogonal hydrofluoroether (HFE) fluid causeany sizable shunt paths in the device. Electrical shortage of OSCscaused by a rough CNT surface is a very common problem fordevices with CNT electrodes, which is typically solved by smooth-ing planarization layers such as PEDOT:PSS [17,23], which are notapplied here.

Y.H. Kim et al. / Solar Energy Materials & Solar Cells 96 (2012) 244–250 247

The high FFs as well as low leakage currents, observed for allf-CNT cells, with both bottom and top illumination, indicate fourpoints: (i) f-CNT sheets are uniformly deposited on all deviceswithout major mechanical deviations. (ii) The conductivity off-CNT is sufficiently high for the top electrode not to cause a largeloss in photo-generated current. (iii) No charge carrier extractionbarriers are formed at the f-CNT/organic interfaces. (iv) The roughf-CNT surface does not cause significant structural defects on theunderlying organic layers during deposition. Therefore, lower JSC

of the f-CNT cells compared to the Ag cell is attributed mostly to

-0.25

-6

-4

-2

0

2

4 f-CNT cell with HTL 20nm f-CNT cell with HTL 50nm f-CNT cell with HTL 80nm Ag cell

curr

ent d

ensi

ty [m

A c

m-2

]cu

rren

t den

sity

[mA

cm

-2]

voltage [V]

-6

-4

-2

0

2

4 f-CNT cell with HTL 20nm f-CNT cell with HTL 50nm f-CNT cell with HTL 80nm Ag cell

0.00 0.25 0.50 0.75

-0.25voltage [V]

0.00 0.25 0.50 0.75

Fig. 2. I�V characteristics of Ag reference cell and f-CNT cells with different HTL

thicknesses (20, 50, and 80 nm) during (a) bottom and (b) top illumination.

Table 1Characteristics of semi-transparent OSCs with different HT

directions, which were measured under an AM 1.5G i

experiments were reproduced at least three times, with no

in the efficiency column.

VOC (V) JSC (mA c

Bottom illumination

f-CNT cell with HTL 20 nm 0.57 4.5

f-CNT cell with HTL 50 nm 0.56 3.8

f-CNT cell with HTL 80 nm 0.56 3.7

Ag cell 0.57 5.4

Top illumination

f-CNT cell with HTL 20 nm 0.52 2.8

f-CNT cell with HTL 50 nm 0.53 2.8

f-CNT cell with HTL 80 nm 0.53 2.4

Ag cell 0.55 3.2

lower light absorption in the bulk heterojunction. Considering theimportant role of the top electrode, which reflects unabsorbedlight back to the absorber layer, more reflective thin Ag electrodeenables the Ag cell to generate more photo-current than the f-CNTcell when illuminated from the bottom (ITO) side. However,during top illumination, low reflectivity of the ITO electrode leadsto a negligible and equal reflection for both cells, therefore only asmall performance difference between the Ag cell and the f-CNTcells is observed, which can be attributed to different

L thicknesses when illuminated from bottom and top

llumination and normalized to 100 mW cm�2. The

relevant deviation, the measurement ranges are given

m�2) FF (%) PCE (%) J (�1 V) JSC�1

58.2 1.570.03 1.12

59.5 1.370.04 1.15

59.4 1.270.03 1.17

61.3 1.970.09 1.14

59.1 0.970.08 1.12

59.4 0.970.06 1.13

60.7 0.870.05 1.17

60.4 1.070.04 1.19

CNT

HTLab

sorb

ed p

hoto

n ra

te [s

-1 m

-3]

abso

rbed

pho

ton

rate

[s-1

m-3

] layer position

HTL 20nm HTL 50nm HTL 80nm

ZnPc: C60

C60, ETL

ITO

illumination direction

ZnPc:C60

C60, ETL ITOHTL

CNT HTL 20nm HTL 50nm HTL 80nm

layer position

illumination direction

1.2x1028

1.0x1028

8.0x1027

6.0x1027

4.0x1027

2.0x1027

6x1027

5x1027

4x1027

3x1027

2x1027

1x1027

Fig. 3. Simulated absorbed photon rates profile of f-CNT cells with different HTL

thicknesses (20, 50, and 80 nm) during (a) bottom and (b) top illumination under

an AM 1.5G illumination condition.

Y.H. Kim et al. / Solar Energy Materials & Solar Cells 96 (2012) 244–250248

transmission properties (PCE and JSC, Ag cell: 1.0%, 3.2 mA cm�2,f-CNT cell: 0.9%, 2.8 mA cm�2). Generally, the lower JSC for topillumination, compared to bottom illumination is attributed toa lower transparency of the top electrodes than that of the bottomelectrodes and the weak reflection at the ITO interface.We believe that further enhancement of the conductivity andtransparency of CNT networks by optimized synthesis processand other means enables further improvements in efficiency andtransparency of the f-CNT cell. In fact, coating CNTs by very thinmetal layers (Au or Al) of �30 nm changes significantly sheetresistance of highly transparent MWCNTs bringing it down to20–30 O sq�1 while transparency is still quite high.

Optical simulations for f-CNT cell series with HTL thicknessesof 20, 50, and 80 nm are carried out to investigate the optical fieldeffect with respect to the HTL thickness. Fig. 3 shows the absorbedphoton rates in the f-CNT cells along the device configurations.It is shown that thinner HTL thicknesses lead to a strong increasein the absorbed photon rate of f-CNT cells in the bulk hetero-junction region when illuminated from bottom. Therefore, highlyenhanced JSC for thinner HTLs is mainly attributed to the im-proved photon absorption in the bulk heterojunction. Theseresults show that the maximum optical field intensity for bottomilluminated f-CNT cells is properly located at the bulk hetero-junction absorber layer when the HTL thickness is around 20 nm,well consistent with the experimental results. In contrast, topilluminated f-CNT cells with all different HTL thicknesses exhibitsimilar absorbed photon rates for all devices, showing smallinfluence of the HTL thickness on JSC. The HTL does not serve asan optical spacer in these OSCs due to large optical absorption ofthe top f-CNT electrode and the reversed illumination conditions.

Fig. 4 shows the simulated optical field distribution for f-CNTcells under bottom illumination. It is shown that the highestintensity optical field is properly situated in the absorber layer forthe major absorption peaks of ZnPc (630 nm) and C60 (450 nm) of

300

400

500

600

700

800

HTL 50nm

wav

elen

gth

[nm

]

illumination d

ETLC60ZnPc:C60HTL

ZnPc:C60HTL

illumination direc300

400

500

600

700

800

wav

elen

gth

[nm

]

ETLC60ZnPc:C60HTL

HTL 20nm

ZnPc:C60HTL

HTL 50nm

HTL 20nm

Fig. 4. Simulated optical field distributions in f-CNT cells with respect to different HTL

from (a) bottom and (b) top. The interfaces of layers are shown as vertical lines. The m

lines. f-CNT and ITO electrodes are omitted for better visibility.

f-CNT cells with 20 nm HTL. On the other hand, relatively low fieldintensities dominate the absorption area of an f-CNT cell with anHTL of 80 nm. These simulated results clearly demonstrate the effectof HTL thickness on JSC, indicating that the optimization of opticalfield geometry of devices is quite important to employ a topelectrode for efficient semi-transparent OSCs. It is shown that theoptimization of spacer layer thickness remarkably enhances JSC withnegligible ohmic losses, because of highly transparent and conduc-tive nature of a doped organic spacer layer [29,31]. Our resultsdemonstrate that the geometry of semi-transparent OSCs, includinga porous, laminated top electrode, can be optimized using conven-tional transfer-matrix theory, assuming top electrodes as a homo-geneous layer without optical scattering.

It should be noticed that 3-D charge collection by highlyporous f-CNT electrodes, found earlier to increase the photo-current in polymeric OSCs does not play important role in ourpresent n–i–p devices, since the liquid solution assisted lamina-tion process does not cause significant protrusion of CNT nano-leads into HTL and photoactive layer, while showing very lowleakage current as well as high FF due to low manufacturingdefects. It is expected that creating architectures with highlyporous CNTs protruding into photoactive layers assisted byadditional pressure during the lamination process will furtherenhance performance by the effects of 3-D collection by sup-pressed charge recombination and transport in built-in electricalfields, strongly enhanced by tips and sides of CNT. The f-CNT cellwith highly transparent and conductive CNT electrode with largerporosity will demonstrate this effect of enhanced charge collec-tion at higher overall transparency, which will become particu-larly important for tandems. The application of f-CNT cells withdifferent transparencies and porosity in parallel type versus inseries tandem type will be described separately.

Ageing measurements are carried out to investigate thestability of OSCs with respect to the different kinds of top

HTL 80nm

irection

0.2

0.60.4

ETLC60 ETLC60ZnPc:C60HTL

tion

HTL 80nm

ETLC60 ETLC60ZnPc:C60HTL field intensity

[arbi. unit]

1.00.8

0.0

thicknesses of 20 nm (left), 50 nm (middle), and 80 nm (right) when illuminated

ajor absorption peaks of ZnPc (630 nm) and C60 (450 nm) are shown as horizontal

0.85

0.90

0.95

1.00

0.85

0.90

0.95

1.00

00.85

0.90

0.95

1.00

0.85

0.90

0.95

1.00

norm

aliz

ed P

CE

nor

mal

ized

FF

norm

aliz

ed V

OC

f-CNT cellAg cell

ageing time [hours] ageing time [hours]100 200 300 400 500 600 700 100 200 300 400 500 600 700

nor

mal

ized

JS

C

Fig. 5. Ageing characteristics of encapsulated OSCs with an f-CNT or a thin metal top electrode, illuminated by white LEDs equivalent to 3 suns at the temperature

of �50 1C.

Y.H. Kim et al. / Solar Energy Materials & Solar Cells 96 (2012) 244–250 249

electrodes. The encapsulated devices are illuminated with whiteLEDs at the intensity of 3 suns (Ag cell: 297 mW cm�2, f-CNT cell:312 mW cm�2) and heated to �50 1C for acceleration of ageingspeed. The normalized PCE, FF, JSC, and VOC for OSCs are shown inFig. 5. The Ag cell shows a PCE decay of �6% after 700 h. Incontrast, the f-CNT cell exhibits more stable performances com-pared to Ag cell, showing PCE decay of only 3% after �700 h.Especially, the FF of the f-CNT cell is very stable, retaining itsinitial value but that of the Ag cell keeps decreasing, indicatingthat the degradation of FF for f-CNT cell is almost free fromoxidation and deterioration of the top electrode, which directlyaffect the device degradation reported by Voroshazi et al. [32] andis a common issue for thin film metal electrodes. These resultssuggest that the f-CNT devices have superior stability comparedto the devices because of the metal-free electrodes concept.

4. Conclusion

In conclusion, f-CNT top electrodes have been successfullyadopted for semi-transparent OSCs. In a simple orthogonal liquidassisted self laminating process, the f-CNT membrane could beapplied as top contact to the OSCs without any carrier substrate.The fabricated devices show relatively high FFs of up to 60% andcompetitive efficiencies of up to 1.5%. All devices show very lowleakage currents without any additional smoothing layers. The metalelectrode-free f-CNT cell shows better long-term stability comparedto the device having an oxidative metal electrode. These resultsindicate homogeneous and damage-free application of the f-CNTmembrane despite of its high roughness and the delicate interface ofthe device. Optimization of the HTL thickness as optical spacer cancontrol the field distribution in devices, which remarkably improvesthe efficiency of f-CNT cells. It is shown that the semi-transparentOSCs can be simulated and optimized using transfer-matrixapproach without optical scattering, despite the semi-transparent,porous nature of the f-CNT top electrodes. The inherent potential forimproving CNT sheets networks by optimized CVD growth process ofCNT forests with higher and narrower bundles [33] is predicted toresult in higher conductivity and transparency, which will allowfurther improvement of the transparency and efficiency of f-CNT

based OSCs. It is also expected that higher efficiency parallel tandemsolar cells will be enabled by stacking OSCs with a common f-CNTcharge collection electrode without need for current-matching oftandem cell connected in series. Preliminary results indicate thatf-CNTs networks with higher transparency and larger porosity maylead to additional photo-current enhancements. Moreover, thefabrication of large area devices is possible because of the scalableroll-to-roll processing of f-CNT sheets. These results illustrate thefeasibility of applying transparent f-CNT top electrodes with theirunique properties as charge collecting electrodes for large area andefficient semi-transparent OSCs using a simple and reliable processof lamination of free-standing CNT sheets.

Acknowledgments

The authors thank O. R. Hild and C. May for supporting theDIZEeff, A. Merten for help with optical measurements, and J. Leefor fruitful discussion. This work was funded by the EuropeanUnion (EFRE), the Fraunhofer Gesellschaft, and the Free State ofSaxony as part of the Dresdner Innovationszentrum Energieeffi-zienz. Support of DOE STTR grant on Hybrid tandems, WelchFoundation grant AT-1617 and Rice/AFRL grant via CONTACTconsortium of Texas are highly appreciated. AZ acknowledgessupport of Alexander von Humboldt Foundation.

Appendix A. Supplementary materials

Supplementary data associated with this article can be foundin the online version at doi:10.1016/j.solmat.2011.10.001.

References

[1] M.A. Green, K. Emery, Y. Hishikawa, W. Warta, Solar cell efficiencytables (version 37), Progress in Photovoltaics: Research and Applications 19(2011) 84–92.

[2] M. Li, S. Ishihara, M. Akada, M. Liao, L. Sang, J.P. Hill, V. Krishnan, Y. Ma,K. Ariga, Electrochemical-coupling layer-by-layer (ECC-LbL) assembly, Jour-nal of the American Chemical Society 133 (2011) 7348–7351.

Y.H. Kim et al. / Solar Energy Materials & Solar Cells 96 (2012) 244–250250

[3] R. Charvet, S. Acharya, J.P. Hill, M. Akada, M. Liao, S. Seki, Y. Honsho, A. Saeki,K. Ariga, Block-copolymer-nanowires with nanosized domain segregationand high charge mobilities as stacked p/n heterojunction arrays for repea-table photocurrent switching, Journal of the American Chemical Society 131(2009) 18030–18031.

[4] B. Landi, S. Castro, H. Ruf, C. Evans, S. Bailey, R. Raffaelle, CdSe quantum dot-single wall carbon nanotube complexes for polymeric solar cells, Solar EnergyMaterials and Solar Cells 87 (2005) 733–746.

[5] W.U. Huynh, J.J. Dittmer, A.P. Alivisatos, Hybrid nanorod-polymer solar cells,Science 295 (2002) 2425–2427.

[6] J.Y. Kim, K. Lee, N.E. Coates, D. Moses, T.Q. Nguyen, M. Dante, A.J. Heeger,Efficient tandem polymer solar cells fabricated by all-solution processing,Science 317 (2007) 222–225.

[7] S. Sista, Z. Hong, M.H. Park, Z. Xu, Y. Yang, High efficiency polymer tandemsolar cells with three terminal structure, Advanced Materials 22 (2010)E77–E80.

[8] F. Li, S. Ruan, Y. Xu, F. Meng, J. Wang, W. Chen, L. Shen, Semitransparentinverted polymer solar cells using MoO3/Ag/WO3 as highly transparentanodes, Solar Energy Materials and Solar Cells 95 (2011) 877–880.

[9] H. Schmidt, H. Flugge, T. Winkler, T. Bulow, T. Riedl, W. Kowalsky, Efficientsemitransparent inverted organic solar cells with indium tin oxide topelectrode, Applied Physics Letters 94 (2009) 243302-1–243302-3.

[10] S.K. Hau, H.L. Yip, J. Zou, A.K.Y. Jen, Indium tin oxide-free semi-transparentinverted polymer solar cells using conducting polymer as both bottom andtop electrodes, Organic Electronics 10 (2009) 1401–1407.

[11] P. Matyba, H. Yamaguchi, G. Eda, M. Chhowalla, L. Edman, N.D. Robinson,Graphene and mobile ions: the key to all-plastic, solution-processed light-emitting devices, ACS Nano 4 (2010) 637–642.

[12] W. Gaynor, J.Y. Lee, P. Peumans, Fully solution-processed inverted polymersolar cells with laminated nanowire electrodes, ACS Nano 4 (2009) 30–34.

[13] J.Y. Lee, S.T. Connor, Y. Cui, P. Peumans, Semitransparent organic photovoltaiccells with laminated top electrode, Nano Letters 10 (2010) 1276–1279.

[14] J. Meiss, M. Furno, S. Pfuetzner, K. Leo, M. Riede, Selective absorptionenhancement in organic solar cells using light incoupling layers, Journal ofApplied Physics 107 (2010) 053117-1–053117-7.

[15] J. Meiss, K. Leo, M.K. Riede, C. Uhrich, W.M. Gnehr, S. Sonntag, M. Pfeiffer,Efficient semitransparent small-molecule organic solar cells, Applied PhysicsLetters 95 (2009) 213306-1–213306-3.

[16] T.M. Barnes, J.D. Bergeson, R.C. Tenent, B.A. Larsen, G. Teeter, K.M. Jones,J.L. Blackburn, J. van de Lagemaat, Carbon nanotube network electrodesenabling efficient organic solar cells without a hole transport layer, AppliedPhysics Letters 96 (2010) 243309-1–243309-3.

[17] L. Hu, J. Li, J. Liu, G. Gruner, T. Marks, Flexible organic light-emitting diodeswith transparent carbon nanotube electrodes: problems and solutions,Nanotechnology 21 (2010) 155202-1–155202-10.

[18] A. Du Pasquier, H.E. Unalan, A. Kanwal, S. Miller, M. Chhowalla, Conductingand transparent single-wall carbon nanotube electrodes for polymer–fullerene solar cells, Applied Physics Letters 87 (2005) 203511-1–203511-3.

[19] J. van de Lagemaat, T.M. Barnes, G. Rumbles, S.E. Shaheen, T.J. Coutts,C. Weeks, I. Levitsky, J. Peltola, P. Glatkowski, Organic solar cells with carbonnanotubes replacing InO: Sn as the transparent electrode, Applied PhysicsLetters 88 (2006) 233503-1–233503-3.

[20] M.W. Rowell, M.A. Topinka, M.D. McGehee, H.J. Prall, G. Dennler,N.S. Sariciftci, L. Hu, G. Gruner, Organic solar cells with carbon nanotubenetwork electrodes, Applied Physics Letters 88 (2006) 233506-1–233506-3.

[21] M. Zhang, S. Fang, A.A. Zakhidov, S.B. Lee, A.E. Aliev, C.D. Williams,K.R. Atkinson, R.H. Baughman, Strong, transparent, multifunctional, carbonnanotube sheets, Science 309 (2005) 1215–1219.

[22] R. Ulbricht, S.B. Lee, X. Jiang, K. Inoue, M. Zhang, S. Fang, R.H. Baughman,A.A. Zakhidov, Transparent carbon nanotube sheets as 3-D charge collectors inorganic solar cells, Solar Energy Materials and Solar Cells 91 (2007) 416–419.

[23] C.D. Williams, R.O. Robles, M. Zhang, S. Li, R.H. Baughman, A.A. Zakhidov,Multiwalled carbon nanotube sheets as transparent electrodes in highbrightness organic light-emitting diodes, Applied Physics Letters 93 (2008)183506-1–183506-3.

[24] S. Tanaka, K. Mielczarek, R. Ovalle-Robles, B. Wang, D. Hsu, A. Zakhidov,Monolithic parallel tandem organic photovoltaic cell with transparent carbonnanotube interlayer, Applied Physics Letters 94 (2009) 113506-1–113506-3.

[25] T. Ameri, G. Dennler, C. Lungenschmied, C.J. Brabec, Organic tandem solarcells: a review, Energy & Environmental Science 2 (2009) 347–363.

[26] A. Hadipour, B. de Boer, P.W.M. Blom, Solution-processed organic tandemsolar cells with embedded optical spacers, Journal of Applied Physics 102(2007) 074506-1–074506-6.

[27] A.A. Zakhidov, J.K. Lee, H.H. Fong, J.A. DeFranco, M. Chatzichristidi,P.G. Taylor, C.K. Ober, G.G. Malliaras, Hydrofluoroethers as orthogonalsolvents for the chemical processing of organic electronic materials,Advanced Materials 20 (2008) 3481–3484.

[28] J. Meiss, N. Allinger, C. Falkenberg, K. Leo, M. Riede, Transparent conductivelayers for organic solar cells: simulation and experiment, Proceedings of SPIE7416 (2009) 741601–741603.

[29] R. Schueppel, R. Timmreck, N. Allinger, T. Mueller, M. Furno, C. Uhrich, K. Leo,M. Riede, Controlled current matching in small molecule organic tandemsolar cells using doped spacer layers, Journal of Applied Physics 107 (2010)044503-1–044503-6.

[30] S. Pfuetzner, J. Meiss, A. Petrich, M. Riede, K. Leo, Thick C60:ZnPc bulkheterojunction solar cells with improved performance by film depositionon heated substrates, Applied Physics Letters 94 (2009) 253303-1–253303-3.

[31] B. Maennig, J. Drechsel, D. Gebeyehu, P. Simon, F. Kozlowski, A. Werner, F. Li,S. Grundmann, S. Sonntag, M. Koch, Organic p–i–n solar cells, Applied PhysicsA: Materials Science & Processing 79 (2004) 1–14.

[32] E. Voroshazi, B. Verreet, A. Buri, R. Muller, D. Di Nuzzo, P. Heremans,Influence of cathode oxidation via the hole extraction layer in polymer:ful-lerene solar cells, Organic Electronics 12 (2011) 736–744.

[33] A.A. Kuznetsov, A.F. Fonseca, R.H. Baughman, A.A. Zakhidov, Structural modelfor dry-drawing of sheets and yarns from carbon nanotube forests, ACS Nano5 (2011) 985–993.