Improvement of Inverted OPV Performance by Enhancement of ZnO Layer Properties as an Electron Transfer Layer1

Available online at www.sciencedirect.com

ScienceDirect Materials Today: Proceedings 3 (2016) 758 – 771

12th International Conference on Nanosciences & Nanotechnologies & 8th International Symposium on Flexible Organic Electronics

Improvement of inverted OPV performance by enhancement of ZnO layer properties as an electron transfer layer Õ C.A. Polyzoidisa, C. Kapnopoulosa, E. D. Mekeridisb, L. Tzounisa, S. Tsimiklia,b, C. Gravalidisa, A. Laskarakisa, S. Logothetidisa * a

Lab for Thin Films Nanosystems and Nanometrology, Physics Department, Aristotle University of Thessaloniki, Thessaloniki 54124, Greece b Organic Electronic Technologies P.C. (OET), Antoni Tritsi 21B, GR-57001 Thessaloniki, Greece

Abstract

In the present study we focus on the optimization of NP concentration of ZnO used as an ETL for the fabrication of fully printed inverted OPVs by lab-scale Sheet-to-Sheet gravure technique. The inverted OPV architecture consistσ of the layer sequence: PET/ITO/ZnO/P3HT:PCBM/PEDOT:PSS/Ag. By diversifying ZnO nanoparticle concentration, we track the optimum concentration for better OPV efficiencies and try to correlate concentration to the electrical characteristics of the OPV and other ETL characteristics such as thickness, surface morphology and roughness, hydrophilicity etc. Further goal of this work is to achieve a cost-efficient scalability of flexible organic photovoltaics (OPVs), the optimization of the ZnO NPs and of the printing processes.

©2016 2015Elsevier The Authors. by Elsevier Ltd. © Ltd. AllPublished rights reserved. Selectionand andpeer-review peer-review under responsibility ofConference the Conference Committee Members of NANOTEXNOLOGY2015 (12th Selection under responsibility of the Committee Members of NANOTEXNOLOGY2015 International Conference ononNanosciences Nanotechnologies International Symposium on Flexible (12th International Conference Nanosciences & Nanotechnologies && 8th8th International Symposium on Flexible OrganicOrganic Electronics) Electronics). Keywords: Flexible organic photovoltaics, Gravure printing, Module, P3HT:PCBM bulk heterojunction, Spectroscopic ellipsometry

Õ

This is an open-access article distributed under the terms of the Creative Commons Attribution-Non Commercial-Share Alike License, which permits non-commercial use, distribution, and reproduction in any medium, provided the original author and source are credited. * Corresponding author. Tel.: +30 2310998174. E-mail address: [email protected]

2214-7853 © 2016 Elsevier Ltd. All rights reserved. Selection and peer-review under responsibility of the Conference Committee Members of NANOTEXNOLOGY2015 (12th International Conference on Nanosciences & Nanotechnologies & 8th International Symposium on Flexible Organic Electronics) doi:10.1016/j.matpr.2016.02.007

C.A. Polyzoidis et al. / Materials Today: Proceedings 3 (2016) 758 – 771

1. Introduction Organic Photovoltaic devices have been a field of extensive study for many years. Yet, several attempts to commercialize OPV technology have failed. In solar energy market, inorganic photovoltaics still take the lead against OPVs due to their much better overall efficiency and operational lifetime. In order to achieve the costefficient scalability of flexible organic photovoltaics (OPVs), the optimization of the materials and printing processes is necessary. A normal OPV architecture is undesirable for a large-scale production due to the vacuum processing steps that are mainly required for the deposition of top metal electrode [1]-[4]. Also, the use of high workfunction anode materials (e.g. Ag ink) in inverted architectures enhances device stability. On the other hand, mass production of OPV devices dictates the adoption of fast, non-expensive and scalable fabrication techniques [5][6]. Gravure printing may accommodate this need, since it can easily be implemented in a roll-to-roll (R2R) production line, consumes low power and few materials, not to mention its high throughput printing and resolution [7]-[10], hence ensuring high-quality scalability. In order to print inverted OPVs of high functionality and stability, several characteristics have to be optimized such as surface morphology, thickness, roughness, good adhesion between neighboring layers, as well as the electrical characteristics of the printed OPV devices. The present study aims to optimize several parameters that enhance the layer quality and electrical characteristics of a fully gravure-printed OPV on a laboratory scale, while focusing on the ZnO Electron Transfer Layer (ETL). To that direction, instead of trying power-consuming techniques like plasma treatment [11][12][13], or applying additional chemicals on top of a layer surface like KOH, NaOH, or Cs2CO3 [14][15], an investigatory printing of ZnO nanoparticle suspensions with differentiated concentrations on top of Polyethylene-Terephthalate (PET) and Indium Tin Oxide (ITO) substrate was suggested and performed. A very cheap and commercially available ZnO suspension was chosen to be under study. As next step for this work, the effect of nanoparticle concentration’s variation on the electrical characteristics of an inverted OPV is studied. Diversified concentration is the only parameter that does not remain constant during the printing steps, so that its effects can be better discerned and understood. The sequence and the materials of the device’s subsequent layers is as follows: PET/ITO/ZnO/P3HT:PC60BM/PEDOT: PSS/Ag Every different nanoparticle concentration results in different electrical parameters of the Photovoltaic Device. After assessing the respective results, it is concluded that OPV performance gets enhanced when nanoparticle concentration rises up to a certain critical concentration percentage, above which, ZnO layer thickness gets too high values to allow high performances. During this step, the impact of ZnO layer and surface properties on OPV performance gets under investigation. On top of that, the optimum concentration that maximizes OPV efficiency results in a ZnO layer thickness that needs to be compared with the according literature values. Subsequently, next goal of this study is to achieve optimum electrical OPV characteristics with lower ZnO layer thicknesses. Filtering ZnO suspensions prior to printing is a simple solution to that direction, since it successfully reduces ZnO layer roughness and slightly reduces layer thickness. Similar OPV devices were gravure-printed, yet at lower, filtered concentrations in order to track the correlation of OPV performance to modified ZnO layer thickness and roughness, as well as to find the optimum ZnO concentration. From corresponding J-V curves it is checked whether the optimized filtered concentrations yield low layer thickness and maximize OPV efficiency, hence fulfilling literature criteria. As a final step, since suitability of ZnO dispersion for Roll-to-Roll production is the long-term goal of this study, the impact of material cost and consumption with regard to gravure printing is under discussion. Thus, differences in corresponding OPV characteristics per each ZnO concentration, either filtered or non-filtered are spotted and distinct cases that maximize OPV efficiency are selected. On the other hand, optimized ZnO concentrations with this suspension are compared with the case of another commercially available suspension which yields even lower thicknesses and also repeatedly high OPV performances. What is concluded is that the alternative ZnO may satisfy thickness criteria and be a better choice on a laboratory scale. Yet, the first ZnO suspension is to be preferred for R2R production despite very high layer thicknesses, since it gives comparable results and the production of such a ZnO layer is much cheaper.

759

760

C.A. Polyzoidis et al. / Materials Today: Proceedings 3 (2016) 758 – 771

2. Experimental part 2.1 Materials and inks The substrate for the fabrication of inverted OPVs was chosen to be the commercially available PET/ITO. The PET is heat stabilized while having a thickness of 175 μm. The ITO has a sheet resistance of ~50 Ω/sq. Chemical etching with aqua regia was applied on ITO, so that the transparent cathode electrodes are formed. Substrate cleaning follows, during which PET/ITO undergoes four sequential ultrasonication steps with isopropyl alcohol, acetone, ethanol and deionized water, respectively. For the Electron Transport Layer (ETL), a Zinc oxide (ZnO) nanoparticle ink was chosen (NanoSunguard™ in ethanol II purchased by Sigma Aldrich, 40% wt). The active layer comprises a bulk heterojunction film from the fullerene-based electron acceptor [6,6]-phenyl C61 butyric acid methyl ester (PCBM, technical grade, 99,5%, Solenne BV) and the polymeric donor Poly(3-hexylthiophene) (P3HT, Advent). Blending was done in a concentration of 160 mg/mL (16.0% w/v), having a ratio of 0.8:1 respectively. Ortho-dichlorobenzene (o-DCB) was chosen to be the solvent for the active blend. For the hole transport layer (HTL), a formulation of PEDOT:PSS was applied (CCP105D, Clevios TM, Heraeus). Finally, an Ag nanoparticle ink was applied for the top contact anode electrodes (ANP ink from Sigma Aldrich, ρ = 5-30 μΩ cm).

2.2 Fabrication techniques applied Sheet-to-sheet gravure printing was applied. All OPV modules were fabricated with the RK printing proofer (RK printing instruments). It provides a printing speed range between 1 and 22 m/min. With regard to the engraved plate, the cells of its surface pattern had a tone at 100% and a line density of 120 lines/cm. They were of reverse tetrahedral pyramidal shape and had a nominal volume of 12.7 ml/m 2. Finally, the engraved cells form rectangular stripes with dimensions of 7 mm (width) and 46 mm (length), hence resulting in printed patterns with almost identical shape and size.

Fig 1: Schematic illustration of the used sheet-to-sheet gravure printing system

C.A. Polyzoidis et al. / Materials Today: Proceedings 3 (2016) 758 – 771

2.3 OPV module fabrication As it has been previously mentioned, ITO underwent an etching process with aqua regia, during which only 8 ITO stripes were patterned. Every stripe constitutes the transparent bottom cathode electrode. In order to secure a successful adhesion of the subsequent ZnO layer to the patterned ITO surface, an oxygen-plasma treatment followed for 5 min (the pressure was 1.2 mbar and the applied power 40 W), hence removing any residual impurities and increasing the free energy of the surface. Subsequently, ZnO nanoparticle suspension was gravure-printed. The speed was chosen to remain the same for all printable layers, namely 18 m/min, since a roll-to-roll large-scale production line would require a constant web speed for all simultaneously printed/coated layers later on. Wet ZnO film was annealed at 140 °C for 1min. The photoactive blend was printed with the same speed. In order for the HTL material (CCP 105D) to be printed effectively, dry active layers were subjected to mild oxygen-plasma treatment for 20 sec (the pressure was 1.12 mbar and the applied power was 10 W). After the PEDOT:PSS layer, the Ag nanoparticle suspension was printed and dried. Annealing temperature (140ο C) and time (60 sec) remained the same for each printed layer. With regard to the final fabricated device, each one consists of 8 cells that are interconnected in series. The total active area of the OPV module is 8 cm2, namely 1 cm2 per each single cell. Each printed cell has a width of 7 mm and a length of 46 mm. The total device area is up to 45 cm2.

Fig. 2: Layout of gravure printed photovoltaic module with 8 serially interconnected cells.

Fig. 3: Side view of the serial interconnection for the OPV cells

2.4. Layer and electrical characterization of OPV device In order to examine the morphology of the ZnO layer on the microscale an optical microscope was used. On the other hand, nanotopography and the roughness of the surface, both Peak to Valley (P2V) and Root mean square (RMS), were measured and depicted with Atomic Force Microscopy (SOLVER, P47H from NT-MDT). The printing quality of the ETL was additionally studied with Field-emission Scanning Electron Microscopy (NEON 40, Carl Zeiss AG) operating at an accelerating voltage of 1 kV. Moreover, the wettability of the ZnO layers was characterized via static contact angle measurements (CAM 200, KSV Instruments) by applying the pendant droplet

761

762

C.A. Polyzoidis et al. / Materials Today: Proceedings 3 (2016) 758 – 771

method (droplet volume was a 5 μl). The optical properties and the thickness of the ZnO films were examined by Spectroscopic Ellipsometry (SE). In our case a phase modulated SE system (Horiba) was used, with a working range between 0.7–6.5 eV and with 20 meV steps. Angle of incidence was kept at 70°. For the analysis of ellipsometry measurements, TaucLorentz model with two oscillators was employed. Electrical characterization of OPV modules, namely the extraction of J-V curves, was performed by using a Keithley 2420 SMU. A Newport Oriel Solar Simulator (91191) provided an AM 1.5G spectrum and a power density of 100 mW/cm2 (1 sun illumination). Light spot size covered an area of 12 cm2. 3. Results and discussion 3.1 Characterization of morphology and roughness Optical microscopy examination of microscale topography demonstrated a homogeneously printed ETL in all cases of concentrations (4%, 5%, 7%, 10%, 15% wt). On the other hand the highest concentration under test (25% wt) resulted in a wavy surface. This may be attributed to the increased viscosity and decreased solvent content in combination with the effects of the gravure plate. Yet, the “ribbon-like” surface successfully covers the substrate, hence excluding any ZnO layer incontinuities (e.g. pin holes, etc.). This can be corroborated by following SEM image for an even lower concentration (7% wt). a

b

Fig. 4a: Optical microscopy image of ZnO layer on PET/ITO substrate (5% wt nanoparticle concentration) Fig. 4b: Image of ZnO layer with 25% wt nanoparticle concentration

C.A. Polyzoidis et al. / Materials Today: Proceedings 3 (2016) 758 – 771

Fig. 5: SEM image of printed ZnO layer (7% wt) on ITO/PET substrate. ITO was revealed by an accidental scratch on sample.

With regard to nanoscale topography and roughness, AFM measurements of ZnO samples on PET/ITO substrates have shown that a lower nanoparticle concentration results in higher roughness. An explanation can be easily found due to the fact that, in this work, nanoparticle concentration is differentiated only by the addition of solvent to the pristine commercially available suspension (40% wt). As a result, a lower nanoparticle concentration also means lower polymeric ligands’ concentration, hence becoming easier for nanoparticles to form aggregates and agglomerates. It is noteworthy that at high concentrations (i.e. 25% wt) nanoscale roughness is minimized contrary to microscale roughness.

Fig. 6: Dependence of Peak-to-Valley (left axis) and Root-Mean-Square roughness (right axis) on ZnO nanoparticle concentration.

763

764

C.A. Polyzoidis et al. / Materials Today: Proceedings 3 (2016) 758 – 771

a

b

Fig. 7a & 7b: AFM images of ZnO surface morphology (5% wt & 25% wt respectively) with dimensions 5x5 μm2 Additionally, an increase in ZnO nanoparticle concentration enhances wettability of the ETL. Since no surface treatment technique is applied on ZnO layers and the only differentiating parameter is nanoparticle concentration, it is reasonable that a rise in roughness would result in trapped air beneath the water droplet (or the wet film) [16].

Fig. 8: Dependence of ZnO surface hydrophilicity on nanoparticle concentration

3.2 Optical properties of the ZnO layer Since ETL roughness affects its optical characteristics, the following geometrical modelling is considered, in which the ZnO film consists of two hypothetical layers. The first one (with thickness d 1) is considered a homogeneous, continuously printed ZnO with no roughness. The second one (with thickness d_effective) is

C.A. Polyzoidis et al. / Materials Today: Proceedings 3 (2016) 758 – 771

considered to consist from a compound material with 50% ZnO and 50% air (void space) and its purpose is to introduce the impact of roughness, hence being related to RMS roughness. Effective Medium Theory (Bruggeman model) was, therefore, applied so that the optical parameters of the latter layer be extracted [17]. It is obvious that absorbance of the ETL does not only depend on ZnO properties but on the texture of ETL surface as well.

Fig. 9: Equivalent layer structure for ellipsometric analysis

Fig. 10: Dependence of calculated bandgap (and respective layer absorbance) on nanoparticle concentration

According to the calculated ZnO layer bandgap diagram, in the case of low nanoparticle concentration and high roughness of ETL absorbance is high, namely energy bandgap takes lower values. In the case of higher concentrations lower roughness allows for less incident light scattering on ZnO surface, hence allowing Eg to rise. According to the literature, ‫ܧ‬௚஻௎௅௄ ൌ ͵Ǥ͵͹ܸ݁ [18].

Fig. 11: Dependence of estimated thicknesses on nanoparticle concentration. Total thickness is the sum of d_effective and d_1.

765

766

C.A. Polyzoidis et al. / Materials Today: Proceedings 3 (2016) 758 – 771

3.3 Electrical characterization of OPV module Electrical characterization of the OPV devices with different ETL thicknesses was performed. The following graphs present how efficiency, short-circuit current density and open-circuit voltage change with ZnO nanoparticle concentration. What can be deduced from the graphs is that PCE gets maximized (2.19%) at concentration values around 15% wt (blue dots and trend line), while at the same time achieving an unacceptably high thickness value (122 nm). For lower concentrations, where roughness gets higher values, the quality of ZnO/active layer interface is poor, thus deteriorating carrier extraction and efficiency. Also, a thin ZnO layer of such surface quality is not an efficient hole trap, hence giving rise to leakage current due to low shunt resistance. All related numerical values are summarized in Table 1. In search of a way to shift such trend lines towards left, namely lower nanoparticle concentrations and thicknesses while retaining low roughness, filtering of ZnO suspensions (0.1 μm, PTFE) was applied. In such a way, all nanoparticles and aggregates larger than 100 nm would be excluded from being printed on PET/ITO. A different concentration range was also tried, i.e. 1%, 2%, 3%, 5%, 7%, 10% wt. The electrical characterization of the respective OPV devices yielded trend lines that correspond to lower concentrations and thicknesses (red dots and trend line). According to red trend line PCE gets maximized around 5% wt that corresponds to ~32 nm thickness, whereas maximum PCE values remain almost the same as prior to filtering ZnO suspension. All related numerical values are summarized in Table 2.

Fig. 12: Dependence of OPV PCE values on ZnO nanoparticle concentration with- (red colored) and without suspension filtering (blue colored).

C.A. Polyzoidis et al. / Materials Today: Proceedings 3 (2016) 758 – 771

Fig. 13: Dependence of open-circuit voltage values on ZnO nanoparticle concentration.

Fig. 14: Dependence of short-circuit current density values on ZnO nanoparticle concentration.

767

768

C.A. Polyzoidis et al. / Materials Today: Proceedings 3 (2016) 758 – 771

Table 1: Electrical characteristics of OPV devices with unfiltered ZnO nanoparticle suspension % wt

ZnO Thickness PCE [%] FF [%] [nm]

Voc [V]

Jsc [mA/cm2]

R_series [Ωμ cm2]

R_shunt [Ωμ cm2]

4

21

0.41

26.3

2.11

5.64

45

57

5

35

1.07

29.2

2.97

8.9

35

59

7

49

0.74

27.6

2.66

7.7

38

59

10

73

1.60

33

3.80

11.05

34

148

15

122

2.19

37

4.08

12.65

29

105

25

227

0.48

31.3

4.07

2.82

129

304

Table 2: Electrical characteristics of OPV devices with 0.1 μm filtered ZnO nanoparticle suspension % wt

ZnO Thickness [nm]

PCE [%]

FF [%]

Voc [V]

1

12

1.14

29.6

2

15

1.87

30.4

3

16

1.47

5

32

7 10

Jsc [mA/cm2]

R_series [Ωμ cm2]

R_shunt [Ωμ cm2]

3.14

9.84

21

55

3.94

12.49

61

115

30.4

3.34

11.54

27

57

1.4

33.7

3.80

10.94

26

79

42

2

31.5

3,6

12.24

32

89

71

0.50

24.8

3.52

6.09

52

49

3.4 Surface characterization of filtered ZnO AFM measurements on filtered ZnO surfaces revealed that changes in roughness are much smaller compared to unfiltered ZnO cases. However, it appears that filtering higher concentrations (10% wt) results in porous film (average pore size is 175 nm), hence a 72.5 nm peak-to-valley roughness.

C.A. Polyzoidis et al. / Materials Today: Proceedings 3 (2016) 758 – 771

a

b

c

Fig. 15.a, 15.b & 15.c: AFM images of ZnO surface topography with nanoparticle concentrations of 1%, 5% and 10% wt respectively

Table 3: Thickness and roughness values for filtered ZnO suspensions of distinct concentrations ZnO NP Non-filtered ZnO 0.1 μm filtered ZnO concentration thickness Peak-to-Valley RMS thickness Peak-to-Valley RMS [% wt] [nm] [nm] [nm] [nm] [nm] [nm] 1 (not performed) 11.5 67.3 7.9 5 36.8 90.8 8.84 31.9 61.5 6.3 10 74.2 43.4 4.54 71.5 72.5 8.3

769

770

C.A. Polyzoidis et al. / Materials Today: Proceedings 3 (2016) 758 – 771

4. Conclusions Previous results have shown that by filtering the ETL nanoparticle suspension, i.e. by excluding the largest nanoparticles and agglomerates (size more than 100 nm), a PCE peak is feasible at lower NP concentrations. This shift may well be attributed to low roughness deviation in contrast with the unfiltered ZnO case, since a significant change in roughness may drastically affect the quality of ZnO/P3HT:PCBM interface. To further prove the lab- and industrial scale applicability of this study, four distinct cases of ZnO nanoparticle suspensions were compared. Fabrication and material parameters were kept the same but for ETL inks. All cases yield PCE>1.5%, which is a respectable value for printed OPVs. Cases no.1 & no.2 correspond to previous optimization processes, with- and without filtering, whereas no.3 & no.4 correspond to another commercially available suspension that has already been successfully tested in lab [19]. Case no.4 yields the highest OPV efficiency (PCE = 2.22%) and the thinnest ETL (12 nm) while also providing repeatable results. All respective values and parameters are presented in Table 4. It is important to mention that the cost values apply only to the case of LTFN, hence serving as an example. What can be seen is that case no.1 may violate the thickness limits that are indicated by literature and on the same time yield a very high PCE value (2.19%), while the fabrication cost of the ETL gets a very low value (~0.08 €). Gravure printing power consumption was deliberately not taken into account for all cases. Table 4: Comparison of ETL fabrication recipes Supplier 1 2

Sigma Aldrich Sigma Aldrich

Solvent

mean NP size

Filter ETL PCE FF Voc Jsc Cost %wt used thickness [%] [%] [V] [mA/cm2] [€/ml]

Ethanol

130 nm

No

15

122 nm

2.19

4.1

12.65

~ 0.16

~ 0.08

Ethanol

130 nm

Yes

2

15 nm

1.87 30.4 3.9

12.49

~ 0.033

~ 2.4 [*]

4

9.91

~10

~5

4.3

9.73

~10

~ 7.3 [*]

37

3

Other

N/A

<10 nm

No

1

20 nm

1.59 30.4

4

Other

N/A

<10 nm

Yes

1

12 nm

2.22

43

Cost [€/layer]

(In [*] cases the cost of filtering was added into account) To conclude, in order for an OPV manufacturer to create a commercially viable device, it must compromise production cost and efficiency in an optimum way. As it has been already shown, ultra-low thickness of ZnO layer does not necessarily yield better overall OPV efficiencies than other devices with a thicker ETL. In the present case, an effort to simultaneously optimize all three aforementioned criteria (low thickness, low cost, high PCE). To that direction, the need for an increase in efficiency was achieved by tracking the optimum nanoparticle concentration (15%wt). The need for an OPV that is not too sensitive to be fabricated under ambient (i.e. low-power consuming) conditions, was satisfied by adopting the inverted geometry, which also ensures longer lifespan of the device than the conventional one. Also, gravure printing technique secures low production cost on top of the very low ZnO suspension cost, not to mention the minimized buffer layer cost.

Acknowledgements This research has been financially supported by the Greek State Scholarships Foundation (IKY). It has also been co-financed by Greece and the European Union (European Social Funds) through the Operational Program “'Human Resources Development” of the National Strategic Framework (NSRF) 2007-2013. This work has been partially supported by the EC Project FP7-NMP-2012-LARGE-6, grant no 310229, SMARTONICS.

C.A. Polyzoidis et al. / Materials Today: Proceedings 3 (2016) 758 – 771

References [1]

Solution-processed Organic Solar Cells, Brabec et al., MRS Bulletin, Vol. 33, July 2008

[2]

Fabrication of Polymer Solar Cells Using Aqueous Processing for All Layers Including the Metal Back Electrode, F.C. Krebs et al., Advanced Energy Materials, 1 (2011) 68-71 Printed metal back electrodes for R2R fabricated polymer solar cells studied using the LBIC technique, F.C. Krebs et al., Solar Energy Materials & Solar Cells, 95 (2011) 1348-1353 R2R-printed inverted OPV modules - towards arbitrary patterned designs, M. Valimaki, et al., Nanoscale, (2015).

[3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19]

Evolution of vertical phase separation in P3HT:PCBM thin films induced by thermal annealing, P.G. Karagiannidis et al., Materials Chemistry and Physics, 129 (2011) 1207-1213. Development of Lab-to-Fab Production Equipment Across Several Length Scales for Printed Energy Technologies, M. Hösel, H.F. Dam, F.C. Krebs, Including Solar Cells, Energy Technology, 3 (2015) 293-304. Patternable polymer bulk heterojunction photovoltaic cells on plastic by rotogravure printing, Subramanian et al., Solar Energy Materials and Solar Cells, 93 (2009) 459-464. Gravure printed flexible organic photovoltaic modules, P. Kopola et al., Solar Energy Materials and Solar Cells, 95 (2011) 1344-1347. Paper Electronics, D. Tobjörk, R. Österbacka, Advanced Materials, 23 (2011) 1935-1961. A development and evaluation of micro-gravure coater for printed electronics, J.K. Hwang, S. Bae, D.S. Kim, Japanese Journal of Applied Physics, 53 (2014) 05HC12. GravureǦPrinted ZnO in Fully RollǦtoǦRoll Printed Inverted Organic Solar Cells: Optimization of Adhesion and Performance, Välimäki et al., Volume 3, Issue 4, pages 407–413, April 2015 Oxygen Plasma Functioning of Charge Carrier Density in Zinc Oxide Thin-Film Transistors, Min-Ching Chu et al 2013 Appl. Phys. Express 6 076501 Workfunction tuning of zinc oxide films by argon sputtering and oxygen plasma: an experimental and computational study , Fang-Ling Kuo et al 2012 J. Phys. D: Appl. Phys. 45 065301 Indium tin oxide modified with sodium compounds as cathode of inverted polymer solar cells, Sun et al., J. Mater. Chem., 2011,21, 18339-18346 Tuning indium tin oxide work function with solution-processed alkali carbonate interfacial layers for high-efficiency inverted organic photovoltaic cells, Fei Chen et al 2013 Nanotechnology 24 484011 Intermolecular and Surface Forces, J. Israelachvili, Academic Press 1985-2004 In-line spectroscopic ellipsometry for the monitoring of the optical properties and quality of roll-to-roll printed nanolayers for organic photovoltaics, Logothetidis et al., Solar Energy Optical properties and electronic structure of amorphous germanium, J. Tauc et al., Phys. Stat. Sol., 15 (1966) 627 Gravure printed organic photovoltaic modules onto flexible substrates consisting of a P3HT:PCBM photoactive blend, Kapnopoulos et al., 12th International Conference on Nanosciences & Nanotechnologies & 8th International Symposium on Flexible Organic Electronics (2015)

771