Application of CVD Graphene in Organic Photovoltaics as Transparent Conducting Electrodes by Hyesung Park Submitted to the Department of Electrical Engineering and Computer Science in partial fulfillment of the requirements for the degree of 1 MA SSACHUSETTS E INSTrIiffE 77 C"N"LOGY Doctor of Philosophy at the MASSACHUSETTS INSTITUTE OF TECHNOLOGY LIBRARIES JUNE 2012 ©2012 Massachusetts Institute of Technology. All rights reserved Author: Departinent of Electrical Engineering and Computer Science May 23, 2012 Certified by: Jing Kong ITT Career Development Associate Professor of Engineering and Computer Science Thesis Supervisor Accepted by: __...-_ Leslie A. Kolodziejski Chair, Department Committee on Graduate Students (This page intentionally left blank) 2 Application of CVD Graphene in Organic Photovoltaics as Transparent Conducting Electrodes by Hyesung Park Submitted to the Department of Electrical Engineering and Computer Science May 23, 2012 in partial fulfillment of the requirements for the degree of Doctor of Philosophy Abstract Graphene, a hexagonal arrangement of carbon atoms forming a one-atom thick planar sheet, has gained much attention due to its remarkable physical properties. Apart from the micromechanical cleavage of highly ordered pyrolytic graphite (HOPG), several alternate methods have been explored to achieve reliable and repeatable synthesis of large-area graphene sheets. Among these, the chemical vapor deposition (CVD) process has been demonstrated as an efficient way of producing continuous, large area graphene films and the synthesis of graphene sheets up to 30-inch has been reported. Similar to graphene research, solar cells based on organic materials have also drawn significant attention as a possible candidate for the generation of clean electricity over conventional inorganic photovoltaics due to the interesting properties of organic semiconductors such as high absorption coefficients, light weight and flexibility, and potentially low-cost, high throughput fabrication processes. Transparent conducting electrodes (TCE) are widely used in organic photovoltaics, and metal oxides such as indium tin oxide (ITO) have been commonly used as window electrodes. Usually used as thin films, these materials require low sheet resistance (Rsh) with high transparency (T). Currently the dominant material used in the industry standard However, these materials are not ideal options for organic photovoltaic is ITO. applications due to several reasons: (1) non-uniform absorption across the visible to near infrared region; (2) chemical instability; (3) metal oxide electrodes easily fracture under large bending, and they are not suitable for flexible solar cell applications; (4) limited availability of indium on the earth leading to increasing costs with time. Therefore, the need for alternative/replacement materials for ITO is ever increasing and ideally need to be developed with the following characteristics: low-cost, mechanically 3 robust, transparent, electrically conductive, and ultimately should demonstrate comparable or better performance compared to ITO-based photovoltaic devices. With superior flexibility and good electrical conductivity, as well as abundance of source material (carbon) at lower costs compared to ITO, in this thesis, we propose that the CVD graphene can be a suitable candidate material as TCE in organic photovoltaic applications, satisfying the aforementioned requirements. Thesis Supervisor: Jing Kong Title: ITT Career Development Associate Professor of Electrical Engineering and Computer Science 4 (This page intentionally left blank) 5 Acknowledgements Here I am with my thesis completed. Looking back upon the time when I first started my journey as a graduate student at MIT and now facing myself at the present moment, I am trying to think myself what I have truly learned during the past several years and whether I have found what I truly like to do in the future. Well, except the fact that I was fortunately able to gain some knowledge about my research work, they still remain as open questions. However, if I had to tell how I could successfully finish up my thesis work, I think I can give one solid answer for that. This work would not have been possible to complete without the help of various people I met at MIT. First of all, I am gratefully thankful to my advisor Professor Jing Kong. Not to mention for giving me this wonderful opportunity to work under her research group, she brought me into the world of graphene: an area that I was unaware of, an area that seems to offer so many opportunities to modern science and technology, and an area where so many amazing things could happen. I truly appreciate her guidance and valuable suggestions to move forward, and most of all her patience during the times when I thought I was completely lost and had absolutely nowhere to go. She has been more than an academic or research advisor to me, and there is absolutely no words enough to express how much I am thankful to her. I am also very grateful to my committee members Professors Mildred Dresselhaus and Vladimir Bulovic for their insightful advices. I would also like to thank my labmates from Kong's group whose presence I enjoyed a lot: Alfonso Reina who first taught me how to make this amazing graphene in less than three hours. Mario Hofmann, Daniel Nezich, and Allen Hsu for their helpful discussions. I am also thankful for other group members, Wenjing Fang, Kikang Kim, Yumeng Shi, Yongcheol Shin, Soomin Kim, Sungmi Jung, Yi Song, Minseok Choi, Roman Caudillo, Helen Zing. I also thank my collaborators from Professor Vladimir Bulovid's group, Jill Rowehl, Patrick Brown, Geoffrey J. Supran, Joel Jean, Trisha Andrew. Miles Barr, Rachel Howden from Professor Karen Gleason's group. Sehoon Jang, Matthew Smith from Professor Silvija Gradebak's group. A special thank goes out to Dahyun Oh. I will never be able to forget those days we shared together at MIT. This work would have not been complete without you. Finally, I am greatly indebted to my parents, Hakkil Park, Sooknim Cha, and my sister Hyochun Park for their love, support, and encouragement. Thank you for always believing and having faith in me. I am and can only be here because of you. Hyesung Park Cambridge, Massachusetts May, 2012 6 (This page intentionally left blank) 7 Contents A bstract.............................................................................................................................. 3 A cknow ledgem ents ....................................................................................................... 6 C ontents ............................................................................................................................. 8 List of Tables and Figures.......................................................................................... 11 1. Introduction................................................................................................................. 22 1.1 C urrent Status of Photovoltaic Energy ................................................... 22 1.2 Scope of the Thesis..................................................................................... 23 2. O rganic Photovoltaics............................................................................................... 26 2.1 Background ................................................................................................ 26 2.2 Inorganic vs. O rganic Solar C ells .......................................................... 27 2.3 Principles of O peration ............................................................................. 31 2.3.1 O rganic Sem iconductor ....................................................................... 31 2.3.2 O peration of O rganic Solar C ells ...................................................... 33 2.3.3 Device Architecture ............................................................................. 36 2.4 O rganic Solar C ell Characterization ...................................................... 40 2.4.1 Q uantum Efficiency ................................................................................ 41 2.4.2 Equivalent Circuit Model and Key Parameters of OPV Cells ..... 44 2.5 Solar Radiation ......................................................................................... 3. Graphene ..................................................................................................................... 3.1 3.1.1 3.2 Background ................................................................................................ 52 55 55 Synthesis M ethods................................................................................ 56 Physical Properties of G raphene............................................................. 58 3.2.1 Electrical Properties ........................................................................... 58 3.2.2 O ptical Properties ................................................................................ 61 Possible Application of Graphene in Opto-electronic Devices .............. 67 3.3 8 3.4 Justification of Graphene as Transparent Conducting Electrodes in Organic Photovoltaics Applications ....................................................................... 68 4. Graphene Synthesis via Chemical Vapor Deposition (CVD).............................. 75 4.1 Graphene Synthesis and Electrode Fabrication .................................... 75 4.2 G row th Mechanism ................................................................................. 78 4.3 Characterization of CVD Graphene ....................................................... 84 5. Graphene Electrode and Organic Solar Cell Fabrication.................................... 5.1 91 Graphene as Transparent Conducting Window Electrodes in OPV Cells: R eq u irem en ts .............................................................................................................. 91 5.2 Graphene Electrode Fabrication.............................................................. 93 5.3 Organic Solar Cell Fabrication and Measurements.............................. 95 5.4 Issues Related to the PMMA Transfer Method...................................... 97 6. CVD Graphene based Organic Solar Cells ............................................................ 6.1 105 Doped Graphene Electrodes for Organic Solar Cells ............................. 105 6.1.1 A im of the W ork .................................................................................... 106 6.1.2 D evice D escription ................................................................................ 106 6.1.3 R esults and D iscussions ........................................................................ 108 6.1.4 Doping of Graphene with AuCl 3............................... 115 6.1.5 C on clu sion s............................................................................................ 116 Transition Metal Oxide Hole Transporting Layer .................................. 117 6.2.1 A im of the W ork .................................................................................... 117 6.2.2 D evice Description ................................................................................ 118 6.2.3 Effect of Graphene Surface Morphology and the MoO 3 HTL ......... 119 6.2.4 Effect of Oxygen Plasma on MoO 3 HTL............................................. 125 6.2.5 Effect of Cathode Work Function ....................................................... 133 6.2.6 C on clu sion s............................................................................................ 135 Vapor Printed PEDOT Hole Transporting Layer................................... 135 6.3.1 A im of the W ork .................................................................................... 136 6.3.2 D evice D escription ................................................................................ 137 6.3.3 V apor Printed PED O T ......................................................................... 139 6.3.4 Vapor Printed PEDOT vs PEDOT:PSS ............................................. 141 6.3.5 W ork F unction ...................................................................................... 143 6.2 6.3 9 . . . .. . . . . . . . . . . .. . . . . . . . . . . 6.3.6 R esults and D iscussions ........................................................................ 143 6.3.7 Organic Solar cells from APCVD Graphene................. 148 6.3.8 Conclusions............................................................................................ 153 6.4 Non-destructive Interface Engineering of Graphene for Universal Applications in O PV and O LED s............................................................................ 153 6.4.1 Aim of the Work.................................................................................... 154 6.4.2 Non-destructive Surface Modification of Graphene.......................... 154 6.4.3 Results and D iscussions: O PV ............................................................. 162 6.4.4 R esults and Discussions: O LED .......................................................... 170 6.4.5 O LED Device characterization............................................................ 174 6.4.6 Conclusions............................................................................................ 174 7. Conclusions................................................................................................................ 177 Publications ................................................................................................................... 182 Bibliography .................................................................................................................. 184 10 List of Tables and Figures Figure 1.1 Global cumulative installed PV capacity, 2008 [1]. Total 13.9 GW. ........ 22 Table 1.1 "A vision for PV technology for 2030 and beyond", Report by PV TRAC, July 23 2004. * estim ated costs. ............................................................................. Figure 2.1 Typical structures of organic semiconductors [12]. The delocalized electron systems can be identified by the alternating single and double carbon-carbon bonds. The bottom panel is a schematic diagram of a PPV chain segment. To form a molecule, an s-orbital for each carbon atom becomes mixed with two of the p-orbitals (sp 2 -hybridization), resulting in 5 bonds that hold the molecule together, leaving one electron per carbon atom in a pz-orbital. The overlap ofp,-orbitals of neighboring carbon atoms allows the delocalization of those electrons into ic-orbitals along a polymer backbone........................... 27 Figure 2.2 Schematic description of the electronic wave function illustrating the fundamental differences between conventional semiconductors (CSC) and the excitonic semiconductor (XSC). When y > 1, that is the wave function sits inside the potential well, excitonic behavior is expected [18]..................... 29 Figure 2.3 Progress of PV research efficiencies by design type from NREL [2012] [19]. 30 ......................................................................................................................... Figure 2.4 Schematic illustration of small molecule deposition under vacuum, organic on 31 m etal (left) or m etal on organic (right) [20]................................................. Figure 2.5 Schematic description of charge separation of a photo-generated exciton at the 33 donor/acceptor heterojunction [26]............................................................. Figure 2.6 Schematic illustration (A) and energy level diagram (B) of the charge separation process in a donor/acceptor heterojunction architecture. In (B), photocurrent generation consists of several processes corresponding to the following sequential steps: (1) photon absorption. (2) charge separation. (3) charge transfer. (4) geminate recombination (before charge separation). (5) interfacial recombination. (1)/(4) and (3)/(5) are the competing processes for efficient photocurrent generation [6,15]. .................................................... 35 Figure 2.7 Typical donor/acceptor heterojunction structures in organic solar cells [15]: (a) a bi-layer heterojunction and (b) a bulk heterojunction. ........................ 37 Figure 2. 8 Small molecules which are typically vacuum-evaporated: ZnPc (zincphthalocyanine), Me-Ptcdi (N,N'-dimethylperylene-3,4,9,10-dicarboximide), the buckminster fullerene C60 , copper phthalocyanine (CuPc), and 3,4,9,10perylenetetracarboxylic bisbenzimidazole (PTCBI) [7,15]. ....................... 38 Figure 2. 9 Solution processed conjugated polymers. Upper row: p-type donor polymers: MDMO-PPV (poly[2-methoxy-5-(3,7-dimethyloctyloxy)]-1,4phenylenevinylene), P3HT (poly(3-hexylthiophene-2,5-diyl), and PFB (poly(9,9'-dioctylfluorene-co-bis-N,N'-(4-butylphenyl)-bis-N,N'-phenyl- 1,4- 11 phenylenediamine). Lower row: n-type acceptor polymers: CN-MEH-PPV (poly-[2-methoxy-5-(2'-ethylhexyloxy)- 1,4-(1 -cyanovinylene)-phenylene), PCBM (1-(3-methoxycarbonyl) propyl-1-phenyl[6,6]C61), and F8TB (poly(9,9'-dioctylfluoreneco-benzothiadiazole) [7]. .................................. 39 Figure 2.10 Schematic illustration of the photocarrier generation process upon the incident light in a single organic layer. The dips in the energy level diagram describe the binding energy of the exciton [5]............................................ 40 Figure 2.11 Schematic illustration of photocarrier generation requirements in donor/acceptor heterojunction organic layers for Eex > IPD - EAA (left) and for Eex < IPD - EAA (right) [5]................................................................ 41 Figure 2.12 Schematic illustration of the photocurrent generation process upon the incident light entering the donor/acceptor heterojunction organic layers [5]. 43 Figure 2.13 Schematic description of a solar cell connected to an external load...... 44 Figure 2.14 Current density-Voltage characteristics of an ideal diode under the dark and the light conditions. Ideally, the total net current can be obtained by shifting the dark current by a constant amount which is equal to the short circuit photocurrent [35]. ...................................................................................... 45 Figure 2.15 Equivalent circuit model of a solar cell with parasitic series (Rs) and parallel (Rp) resistances accounting for the non-ideality of a diode. Jdark and jph are, respectively, the dark current density and photocurrent density, and J(Rp) is the current density due to Rp, describing current flowing through Rp [29]. ......................................................................................................................... 46 Figure 2.16 Effects of the series (Rs) and parallel (Rp) resistances to the current-voltage characteristics [35] ..................................................................................... 46 Figure 2.17 Quantum efficiency of a GaAs solar cell compared with the solar photon flux density in arbitrary untis [35]. ............................................................ 48 Figure 2. 18 Key photovoltaic parameters in organic solar cells................................. 51 Figure 2.19 Illustration of the current density-voltage characteristics with a corresponding equivalent circuit model [39].............................................. 52 Figure 2.20 Spectral power density of sunlight at different radiation spectra [35]. ........ 53 Figure 3.1 Graphene is a 2D building block for other carbon materials of all other dimensionalities. It can be (a) wrapped up into OD buckyballs, (b) rolled into ID nanotubes, and (c) stacked into 3D graphite [40]. ................................ 56 Figure 3. 2 Graphene synthesis from various routes [55]. (a) Graphene flakes on a Si0 2 wafer prepared by the scotch tape method. (b) Left: Graphene suspension prepared from microcrystals obtained by the ultrasound cleavage of graphite in chloroform. Right: The suspension printed on a flexible substrate. (c) CVD graphene grown on a Ni thin film and transferred on a Si wafer. (d) Graphene grown by UHV annealing of single-crystal SiC......................... 57 Figure 3.3 (a) Graphene lattice structure and (b) the first Brillouin zone of graphene [5 6 ,5 7 ]....................................................................................................... . . 58 Figure 3.4 (a) Scanning electron microscopy (SEM) image of a suspended graphene sheet in a field effect transistor. (b) Field-effect measurements indicating a mobility greater than 200,000 cm 2 /(V-s) [45]. (c) Schematic diagram of the 12 Figure Figure Figure Figure Figure Figure Figure Figure graphene band structures for electrons and holes, and the corresponding 59 am bipolar field effect [40] .......................................................................... 3.5 Schematic illustration of the graphene band structure [63]........................ 60 62 3. 6 The geometry of graphene for optical analysis.[68].................................. 3. 7 Color contrast of a single layer graphene on Si/Si0 2 substrate as a function of 63 SiO 2 thickness and wavelength.[68]........................................................... on Si wafers of graphene with varying thicknesses 3. 8 Optical images of graphite with different thicknesses of Si0 2 and under illumination at different wavelengths. The traces in (b) indicate contrast changes in a stepwise manner for 1 - 3 layers suggesting that optical contrast can be applied to identify the number of layers in the graphene on a given substrate.[68]........................ 64 3. 9 Transmittance of a single and bilayer graphene suspended on a porous membrane. Optical absorption is measured as 2.3% per each layer. The inset shows the sample platform with several apertures where graphene flakes are . . 66 p laced .[7 1] ................................................................................................ 3. 10 Rayleigh image of a graphite flake with a different number of graphene . . 66 layers.[72 ] .................................................................................................. 3. 11 Illustration of graphene based opto-electronic devices.[76] Schematic diagram of (a) inorganic, (b) organic, and (c) dye-sensitized solar cells. Schematics of (d) organic LED (light-emitting diode) and (e) photodetector. (f-g) Schematic of capacitive touch screen (f) and resistive graphene touch 68 screen (g)[77]. (h-j) Graphene based smart windows................................ 3. 12 Properties of several transparent conducting electrodes.[76] (a) Transmittance of different materials. (b) Logarithmic plot of thickness dependence of sheet resistance. (c-d) Transmittance as a function of sheet resistance from different materials (c) and of graphene prepared by various methods (d). In (d), LPE, RGO, PAHs, and MC stand for liquid-phase exfoliation of pristine graphene, reduced graphene oxide, polyaromatic 71 hydrocarbons, and micromechanical cleavage. ........................................... Figure 4. 1 (a) Schematic diagram of the graphene synthesis and transfer process. The last part of the transfer procedure is repeated to produce multi-layer graphene stacks for LPCVD graphene. (b) Photographs of copper foils (before and after LPCVD graphene growth. The unit for the length scale is shown at the 76 bottom and is in inch................................................................................... Figure 4. 2 Illustration of the graphene growth process at different stages: (a) LPCVD 77 and (b) A PC V D .......................................................................................... Figure 4. 3 Photograph of a patterned graphene films on a quartz substrate (0.5 inch x 0 .5 inch ). ................................................................................................... . . 78 Figure 4. 4 Binary phase diagram of carbon and four different transition metals:[96] (a) C-Ni, (b) Co-C, (c) Fe-C, (d) Cu-C, showing different phase diagrams. ....... 80 Figure 4. 5 (a) Growth kinetics in the CVD graphene synthesis under steady state gas flow of methane and hydrogen. (b) Simplified description of CVD synthesis of graphene on a copper substrate consisting three major steps: removing native oxide on the copper substrate by hydrogen, nucleation/growth of 13 graphene flakes, and coalescence of the graphene flake into a continuous film .[9 0 ,9 6] ................................................................................................ . .81 Figure 4. 6 SEM images of graphene grown at different time stages: (a) Formation of nucleation sites, (b) Growth of graphene domains, (c) Coalescence into a continuous graphene sheet. In (c), dark areas indicated by the blue circle are region of a few layered graphene.[96]......................................................... 83 Figure 4. 7 Direct growth of graphene patterns from pre-patterned Ni structures (a) An optical image of a pre-patterned Ni film on Si/SiO 2. CVD graphene is grown on the surface of the Ni pattern. (b) An optical image of graphene is transferred from the Ni surface in (a) to other substrate (Si/SiO 2).[88].......... 84 Figure 4. 8 Monolayer graphene characterized by (a-b) optical microscopy, graphene transferred on quartz (a) and Si/SiO 2 (b), (c) AFM, graphene transferred on Si/SiO 2 , and (d) Raman spectroscopy, graphene transferred on Si/SiO 2 ........ 85 Figure 4. 9 (a) Transmittance of graphene films of 1 to 3 layers. As-grown LPCVDsynthesized graphene films are mostly single layered and each additional layer contributes approximately 2.3% opacity over the indicated range of wavelengths. The inset indicates the transmittance at 550 nm as a function of the number of the graphene layers. (b) The sheet resistance of graphene films transferred to quartz substrates as a function of the number of graphene layers. (c) Raman spectra of graphene films (1 - 3 Layers) on quartz substrates with a laser excitation at 532 nm (2.33 eV)........................................................... 87 Figure 4. 10 Optical (a-c, scale bar: 10 pm) and AFM (d-e, scale bar: 1pm) images of graphene transferred on to Si/SiO2 (300 nm) substrates synthesized under different pressure conditions: (a,b,d) APCVD. (c,e) LPCVD images are shown here as a reference for APCVD. (a-b) APCVD graphene consists of non-uniformly distributed multilayer regions on top of a mono-layer background, which can be clearly identified from the image taken at the edge of the graphene as shown in (b). (b) The dotted area indicates a region of the graphene film that is broken due to the transfer process, which further confirms the existence of the mono-layer background. (d) AFM image which illustrates the non-uniformity of APCVD graphene. The rms roughness of APCVD graphene is 1.66 nm compared to 1.25 nm for LPCVD graphene in (e ). ................................................................................................................... 88 Figure 5. 1 Schematic illustration of a typical bi-layer heterojunction small molecular organic solar cell ......................................................................................... 95 Figure 5. 2 Description of complete (a) graphene- or (b) ITO- based of OPV devices with CuPc/C 60 active layers and (c) the testing fixture. (d) Schematic illustration of the electrical connections: Graphene or ITO serves as the anode and top finger electrodes serve as the cathode............................................ 97 Figure 5. 3 Characteristics of a graphene OPV device with significant amounts of PMMA residues on the graphene surface. (a) AFM image of PMMA residues on the graphene electrode. PMMA was treated by acetone for 2 hours. (b) JV response of a device under simulated AM 1.5G illumination at 100 mW/cm 2, graphene/PEDOT:PSS/CuPc(40nm)/C60(40nm)/BCP(1 Onm)/Al(1 00nm), 14 made from graphene electrodes covered with large amount PMMA residues, showing poor diode characteristics. The inset describes the ideal behavior of the solar cell operation as a reference ......................................................... 99 Figure 5. 4 Graphene surface imaged by AFM after removal of PMMA via different routes. (a) Method (1), Immersing in acetone for 24 hours; (b) Method (2), First treated by acetone vapor followed by 24 hours immersing in acetone; (c) Method (3), Acetone vapor, 2 min acetone immersion, and 3 hours of annealing; (d) Method (4), 3 hours of annealing. Method (3) provides the 101 cleanest graphene surface. ............................................................................ Figure 5. 5 Raman spectra of graphene on 300 nm Si0 2 substrates where PMMA was removed via various methods as described in Figure 5.4. Raman spectra were obtained with a laser excitation wavelength of 532 nm (2.33 eV). .............. 102 Figure 5. 6 Current density vs. voltage characteristics of graphene OPV devices with the MoO 3 hole transporting layer fabricated via cleaning methods (3) and (4) under simulated AM 1.5G illumination at 100 mW/cm 2 . Graphene electrodes are patterned with 15 nm thick Cr metal masks and the thickness of MoO 3 layer is 20 nm. PCEs (power conversion efficiency) from methods (3) and (4) are 0.69 ± 0.02 % and 0.71 ± 0.01 %, respectively. ..................................... 103 Figure 6. 1 Schematic diagram of the organic solar cell structure: graphene or ITO/PEDOT:PSS (or other equivalent layer)/CuPc/C 6 o/BCP/Ag (or Mg/Ag). ....................................................................................................................... 1 08 Figure 6. 2 J-V characteristics of organic solar cells with different anodes under dark and simulated AMl.5G illumination at 100mW/cm 2 . Device performances of various types: (a) different PEDOT layer processing (b) ITO and (c) graphene electrodes are demonstrated. Shown in (d) is the comparison of performances of ITO with modified PEDOT:PSS by 02 plasma (IIIA) and graphene doped w ith AuCl 3 (10m M ) (IIC ). ............................................................................ 111 Figure 6.3 (a) Device schematic of a standard solar cell considered in this work: Graphene/HTL/CuPc(40nm)/C 6o(40nm)/BCP(1 Onm)/Ag(1 00nm); (b) Schematic diagram of flat band energy levels of each materials with various m etal electrodes (in units of eV )................................................................... 118 Figure 6.4 (a) Ultraviolet-visible transmittance spectra of a MoO 3 layer on a quartz substrate with varying thicknesses (20, 30, 40 nm). The UV transmittance measured at 550 nm is 93.7 %, 89.6 %, and 86.2 % with increasing thicknesses. AFM images of (b) MoO 3 (1 Onm) (c) MoO 3 (1 Onm)/Graphene (3L) on quartz substrates. (d, e) Scanning electron micrographs (SEM) of MoO 3 (20 nm) on graphene/quartz. Bright (d and inset, with in-lens detector) and dark (e, without in-lens detector) spots indicate that the graphene openings not covered by the MoO 3 film. Scale bars are 1 gm for (b, c), 20 pm for (d, e) and 200 gm for the inset of (d). The height bars in (b, c) are 20 nm. ....................................................................................................................... 120 Figure 6.5 Optical micrographs of graphene films on quartz substrates deposited with PEDOT:PSS (a) and MoO 3 (20nm, (b)). As shown in (a), spin-casting PEDOT:PSS results in only sporadic covering of PEDOT:PSS (dark dots) on the graphene surface. But the MoO 3 layer (b) is much more uniform.......... 120 15 Figure 6. 6 AFM images of the graphene morphology of different number of graphene layers: (a) 3 layers and (b) 1 layer. Cross sectional profiles of dotted regions are shown below. The RMS surface roughness is 0.6 nm and 2.0 nm for the 1 121 layer and the 3 layers graphene films, respectively...................................... Figure 6. 7 Optical micrograph of graphene films on the quartz substrate after the film is patterned with 10 nm of Cr. The vertical scratch mark is intentionally created to distinguish the graphene region from the quartz substrate. Many cracked regions in the graphene film are observed which can lower the conductivity of 122 graphene electrodes considerably. ................................................................ devices and ITO OPV graphene of characteristics vs. voltage Figure 6.8 Current density with PEDOT:PSS and MoO 3 hole transporting layers under simulated AM 1.5G illumination at 100 mW/cm 2 : (a) graphene electrodes patterned with 15, 25, and 40 nm thick Cr metal masks with PEDOT:PSS HTL. Using thinner Cr mask helps to reduce shunting pathways as confirmed by the increased shunt resistance and better diode behavior; (b) Devices using graphene electrode with varying MoO 3 HTL thicknesses (20 - 40 nm) under light. Graphene was patterned with thinner 15 nm Cr to ensure a smoother surface. ....................................................................................................................... 124 Figure 6. 9 Current density vs. voltage characteristics of graphene and ITO OPV devices with MoO 3 hole transporting layers with and without 02 plasma under simulated AM 1 .5G illumination at 100 mW/cm 2: (a) Graphene anode patterned by thinner (15 nm) Cr mask with MoO 3 (20 nm) HTL; (b) ITO anode with MoO 3 (20nm) along with PEDOT:PSS reference under light. For both (a) and (b), the effect of 02 plasma on the anode/MoO 3 surface appear to be minimal; (c) Graphene anode patterned by thicker (25 nm) Cr mask with 20 nm of MoO 3 HTL. It is observed that the use of 02 plasma on these rougher surfaces help to improve the device performance by planarizing the 12 8 su rfac e . .......................................................................................................... Figure 6. 10 The external quantum efficiency of devices with MoO 3 (20nm, with or without 02 plasma) HTL where light absorption primarily occurs in CuPc and C 60 : (a) with ITO electrodes; (b) with graphene electrodes patterned with 25 nm Cr; (c) Comparison between ITO and graphene. For the relatively rough graphene electrode, air/oxygen exposure significantly increased the quantum 13 1 effic ien cy ....................................................................................................... Figure 6. 11 Current density vs. voltage characteristics of graphene devices with MoO 3 hole transporting layers with and without 02 plasma under simulated AM 1.5G illumination at 100 mW/cm 2 : (a) with 30 nm of MoO 3 . (b) with 40 nm of MoO 3. Graphene anodes are patterned with thicker (25 nm) Cr mask. Although the FFs are slightly improved for both cases, the effect of oxygen is not as obvious as shown in the 20 nm of MoO 3 case from Figure 6.9(c)..... 132 Figure 6. 12 (a) J-V characteristics of graphene devices with varying top metal electrodes under light: graphene/PEDOT:PSS/CuPc(40nm)/C 60(40nm)/BCP(1 Onm)/metals(1 00nm). (b) Jsc, Voc, and PCE of devices with different cathodes. Al was coated over 134 M g and Ca electrodes to prevent oxidization. .............................................. 16 Figure 6. 13 Schematics outlining the fabrication process of PEDOT HTLs, and OPV devices. (a) PEDOT:PSS spin-coating vs. vapor printing of PEDOT deposition. The spin-casting layer covers the graphene and the surrounding quartz substrate while the vapor printed patterns align to produce PEDOT only on the graphene electrodes. (b) Graphene/ITO anode OPV structure: Graphene(or ITO)/PEDOT/DBP/C 60/BCP/Al. ............................................. 138 Figure 6. 14 (a) Transmittance data for the oCVD PEDOT HTL layers, measured using ultraviolet-visible spectroscopy (UV-Vis) over wavelengths from 350-800 nm. The oCVD PEDOT layers decrease in transmittance and sheet resistance with increasing thickness. The three thinnest PEDOT layers (2, 7, and 15 nm) have high transmittance values (>90% over a majority of the range), which are preferred for HTL layers. (b) Sheet resistance values for each thickness and the transmittance at 550 nm. The oCVD PEDOT sheet resistance was measured using a 4-point probe (taking the average of 10 measurements). With increasing oCVD PEDOT thickness, there are more pathways for charge transfer, so the sheet resistance (Rsh) decreases. Rsh decreases dramatically from the thinnest (2 nm) to thicker PEDOT layers (7, 15, and 40 nm). Transmittance and Rsh values of PEDOT:PSS are also shown for comparison. ....................................................................................................................... 1 40 Figure 6. 15 Comparing HTL coverage on quartz/graphene substrate. (a-c) Spin-coated PEDOT:PSS on quartz/graphene substrate, (d-f) oCVD PEDOT coating on quartz/graphene substrate: (a) schematic illustration of PEDOT:PSS spuncoated on a quartz substrate with graphene electrode. Most of the PEDOT:PSS layer is dewetted from the substrate with dark macroscopic defects visible to the naked eye. (b-c) Optical micrographs (at different magnifications) of the spin-cast PEDOT:PSS on the graphene surface illustrating the poor wettability of PEDOT:PSS on the graphene. In contrast, (d) is the schematic illustration of CVD PEDOT coated via vapor deposition on quartz/graphene substrate, where a uniform coating and patterning via shadow masking is achieved. The left side in (d) has the oCVD PEDOT coating whereas the right side is shadow masked. (e) Optical micrograph and (f) SEM image of oCVD PEDOT on graphene showing uniform coverage. 142 Figure 6. 16 J-V characteristics of representative graphene (3-layer, LPCVD)/ITO OPV devices (Graphene, ITO/PEDOT:PSS (40nm), vapor printed PEDOT (740nm)/DBP, 25nm/C 60, 40nm/BCP, 7.5nm/Al, I00nm) under simulated AM 1.5G illumination at 100 mW/cm 2 . (a) Graphene devices with PEDOT:PSS and vapor printed PEDOT (I5nm) HTL, compared with ITO/PEDOT:PSS reference device. (b) Graphene anode based cells with varying thicknesses of vapor printed PEDOT (7, 15, 40nm). (c) ITO anode devices with varying vapor printed PEDOT thicknesses (7, 15, 40nm) and a PEDOT:PSS reference. (d) Flat-band energy level diagram of the complete OPV device structure comparing DBP and CuPe electron donors. ................................................. 145 Figure 6. 17 J-V characteristics of representative graphene (3-layer, LPCVD)/ITO solar cell devices with non-ideal diode characteristics under simulated AM 1.5G illumination at 100 mW/cm 2 : Graphene, ITO/vapor printed PEDOT (740nm)/DBP, 25nm/C 60, 40nm/BCP, 7.5nm/Al, 1 00nm). (a) Graphene anode 17 Figure 6. Figure 6. Figure 6. Figure 6. Figure 6. Figure 6. Figure 6. solar cells. (b) ITO anode solar cells. Corresponding key photovoltaic parameters of each device are summarized in Table 6.7. ............................. 147 18 Optical (a-b, scale bar: 10 pm) and AFM (c-d, scale bar: 1p m, height bar: 20nm) images of graphene transferred on the SiO 2 (300 nm) substrates synthesized under different pressure conditions: (a, c) APCVD. (b, d) LPCVD images are shown for comparison. (a) APCVD graphene consists of non-uniformly distributed multilayer regions on top of the mono-layer background. (c) AFM image further illustrates the non-uniformity of APCVD graphene. The rms roughness of APCVD graphene is 1.66 nm compared to 1.17 nm for LPCV D graphene in (d). ........................................................... 150 19 (a) J-V characteristics of representative graphene (APCVD) OPV devices (Graphene/vapor printed PEDOT, 15nm/DBP, 25nm/C60 , 40nm/BCP, 7.5nm/Al, 100nm) along with ITO/PEDOT:PSS reference device under simulated AM 1.5G illumination at 100 mW/cm 2 . (b) Comparison of graphene-based device performances, where graphene electrodes are prepared under either LPCVD or APCVD conditions................................................. 152 20 (a) Molecular structures of PEDOT:PEG (PC) and PEDOT:PSS. (b) UVVis (ultraviolet-visible spectroscopy) spectra of each polymeric layer and the bi-layer spun-cast on quartz substrates. Both layers were spun at 4000 rpm resulting in the film thicknesses of ~45 and 25 nm, respectively. Varying the spin speed (1500 - 5000 rpm) of PEDOT:PEG (PC) did not have a significant variation on the transmittance (%T, < 5%) and film thicknesses (< 10%) as well as the sheet resistance (2.08+0.3 MQ/sq), possibly due to the originally large particle sizes of the polymer (- 600 nm in suspension). Interestingly, there was a slight increase in the %T particularly in the lower wavelength region, presumably caused by the interference at the interface. The inset shows macroscopic images of each polymer spun-cast on quartz substrates. ....................................................................................................................... 1 56 21 Characterizations of the bi-layer hole transporting layer structure. (a-d) SEM micrographs of graphene (a), graphene/PEDOT:PEG(PC) (b), graphene/PEDOT:PSS (c), graphene/PEDOT:PEG(PC)/PEDOT:PSS (d) all on quartz substrates. Dewetted PEDOT:PSS on the graphene surface is clearly observed whereas the graphene surface is completely coated by the PEDOT:PEG (PC). (e-g) AFM micrographs illustrating the surface morphologies of polymeric layers on quartz substrates: (e) PEDOT:PEG(PC), (f) PEDOT:PSS, (g) PEDOT:PEG(PC)/PEDOT:PSS. The rms (root mean square) roughness of PEDT:PEG (PC) layer has been reduced from 36.4 nm to 27.0 nm upon the deposition of PEDOT:PSS (rms: 0.83 nm). Both SEM and AFM images confirms that the rather rough surface of the PEDOT:PEG (PC) is smoothened by the PEDOT:PSS layer. ............................................ 159 22 SEM image of PEDOT:PSS on graphene/quartz substrates around the edge of the defect region. ...................................................................................... 160 23 SEM micrographs of PEDOT:PEG(PC) (a), PEDOT:PSS (b), PEDOT:PEG(PC)/PEDOT:PSS (c) on quartz substrates. ............................ 161 24 J-V measurements of devices with graphene electrodes using either one of the polymer layers. (a) Due to the inappropriate alignment of the energy level 18 from the PEDOT:PEG (PC), non-rectifying diode behavior is observed, even though complete coverage of the buffer layer was achieved. (b) PEDOT:PSS HTL with its matching WF at the interface still resulted in an almost resistorlike device character from the inadequate wetting on the graphene surface. 163 Figure 6. 25 Descriptions of typical organic solar cells with a graphene electrode and the device performance. (a) Schematic diagram outlining the graphene anode OPV architecture: graphene/PEDOT:PEG(PC), 45nm/PEDOT:PSS, 25nm/DBP, 25nm/C6 0 , 40nm/BCP, 8.5nm/Al, I00nm. (b) Cross-sectional TEM image (left panel) of the complete device described in (a) with EDS elemental line scan overlaid onto a schematic of the device architecture (right panel). Solid lines indicate interfaces identified using TEM, dashed lines indicate expected location of interfaces not quite resolvable by TEM or EDS but partly resolvable from the difference in the color contrast. (c) Flat-band energy level diagram of the complete structure. (d) Current density vs. voltage (J-V) characteristics of a representative graphene OPV device compared with an ITO reference cell under simulated AM 1.5G illumination at 100 mW/cm 2 illustrating comparable performances. (e) J-V characteristics of ITO-based devices with different configurations of HTL, i.e. PEDOT:PSS alone and PEDOT:PEG (PC)/PEDOT:PSS, showing similar performances. 168 Figure 6. 26 Electroluminescence performance of organic light-emitting diodes using graphene anodes. (a) Schematic of the device structure: graphene/PEDOT: PEG(PC), 45 nm/PEDOT:PSS, 25 nm/spiro-TPD, 50 nm/Alq 3, 50 nm/Ag/Mg, 50 nm/Ag, 70 nm. (b) Electroluminescence (EL) spectrum and photograph of the graphene-based OLED at 4.5 V (8.69 mA/cm 2 ) applied bias, exhibiting uniform green emission characteristic of Alq 3. (c) Current density (J,circles) and luminance (L, squares) versus applied forward bias (V) for OLEDs based on graphene (red) and ITO (black). In both cases two conduction regimes are observed, each described by J oc Vm + 1: Ohmic m = 0 or space-charge-limited conduction m = 1 below EL turn-on at 2.4 V; and trap-limited conduction m > 2 at higher voltages. (d) External quantum efficiencies (EQEs) of OLEDs using graphene (red) and ITO (black) anodes, as a function of J. For the graphene devices, a maximum EQE of~0.27 % and a luminance of -77 cd/m 2 were reproducibly reached at ~5 V (17.6 mA/cm 2 ), beyond which failure was typically observed. Nevertheless, the similarity in J,L and EQE OLED performance when ITO is replaced with graphene attests to the potential of this approach. L and EQE are calculated based on emission from the front faces of the OLEDs, as 172 described in C hapter 6.4.5. ........................................................................... Table 6. 1 Summary of photovoltaic performance parameters of OPV devices from 1 13 F ig u re 6 .2 ...................................................................................................... Table 6. 2 Summary of photovoltaic parameters with PEDOT:PSS and MoO 3 HTLs for 125 devices from Figure 6.8(b)............................................................................ Table 6. 3 Summary of photovoltaic parameters (Figure 6.9) with 20nm MoO 3 HTL with or without air/oxygen exposure. Graphene electrodes are patterned with 15 12 9 n m C r. . ......................................................................................................... 19 Table 6. 4 Summary of photovoltaic parameters (Figure 6.9) of graphene devices having 20nm MoO 3 with or without air/oxygen exposure. Graphene electrodes are patterned with 25 nm C r. .............................................................................. 129 Table 6. 5 Summary of photovoltaic parameters of various metal cathodes devices in F igu re 6 .12 . ................................................................................................... 13 5 Table 6. 6 Optical transmittance (%T) and sheet resistance (Rsh) of oCVD PEDOT with varying thicknesses described in Figure 6.14. PEDOT:PSS values are also show n for com parison................................................................................... 141 Table 6. 7 Summary of key photovoltaic parameters of graphene/ITO devices from F igu re 6 .17 . ................................................................................................... 14 8 Table 6. 8 Key photovoltaic parameters for devices described in Figure 6.25(d)......... 169 20 (This page intentionally left blank) 21 CHAPTER 1: INTRODUCTION Chapter 1 1. Introduction 1.1 Current Status of Photovoltaic Energy Recent interest in solar energy is motivated by the desire for clean, green, and renewable power. Photovoltaic (PV) electric generation is a rapidly growing market, averaging about 40 - 50 % per year since 2000, and the cumulative installed PV capacity worldwide has reached ~13.9 GW (giga-watts) as of 2008 reported by NREL (National Renewable Energy Laboratory) (Figure 1.1) [1]. However, solar energy represents still less than 1 % of the total global energy generation and thus there is great potential for improvement. France, 018GW, 1% Rest ofWadd, 097GW 7% 0.46GW 3% SouthKarea, 0.36,3% Geanany, 5.3 GW, 38% .1 GW, 8% Japam, 2. I1GW 15% 34GW,24 Figure 1.1 Global cumulative installed PV capacity, 2008 [1]. Total= 13.9 GW. Commercial photovoltaic modules are mostly based on silicon, that is wafer based crystalline silicon either being single-crystals or poly-crystals. 22 Other available CHAPTER 1: INTRODUCTION technologies are based on thin films, such as amorphous silicon, cadmium telluride (CdTe), or copper-indium/gallium-selenide/sulphide (CIGS). However, silicon modules are still the dominant source of PV supply. The annual expansion of production might help reducing the module prices, but the high cost of current PV cells based on silicon (-$4/Watt peak) [2] limits widespread real-life application, and up to now, PV energy is the most expensive source of the world energy supply (Table 1.1) [3]. Therefore, the need to reduce cost through economic measures will require new fabrication methods and materials, such as roll-to-roll processing on arbitrary substrates, and organic materials (i.e., polymers or small molecules) as active solar energy harvesting media rather than inorganic silicon. Hydroelectricity Bio-energy Wind energy Geothermal energy Marine energy Solar thermal energy Photovoltaic energy Total renewable electricity production Total World electricity production in 2003 (TWh) Electricity generation costs in 2003 (E cents/kWh) 2631 175 75 50 0.8 0.5 2.5 2-8 5-6 4-12 2-10 8-15* 12-18 26-65 2969 Total electricity consumption: 16,700 consumption World estimated technical annual generation potential (x 100 TWh) 14 77-124 178 1,400 N/A >400 >2100 Current global energy consumption: 12 Table 1.1 "A vision for PV technology for 2030 and beyond", Report by PV TRAC, July 2004. * estimated costs. 1.2 Scope of the Thesis Organic photovoltaics (OPV) have gained much attention as a possible candidate for the generation of clean electricity due to the interesting properties of organic semiconductors. In most OPV devices, ITO has been widely used as a transparent conducting window 23 CHAPTER 1: INTRODUCTION electrode. However, the need for alternative materials to ITO is ever increasing due to the limited availability of indium on earth and its rising cost. Therefore, an ITOsubstitute, that is low-cost, mechanically robust, transparent, and electrically conductive, needs to be developed. In this thesis, we propose that graphene can be a suitable candidate satisfying the general conditions required for transparent conducting electrodes in the related fields. Chapters 2 and 3 will be devoted to the background information regarding organic photovoltaic solar cells and graphene. Chapter 4 will present how these amazing two dimensional graphene films can be synthesized from the chemical vapor deposition process, "rather simply", considering the importance of the remarkable physical properties of graphene. In Chapter 5, we will learn about how one can make transparent conducing electrodes from CVD graphene and build organic solar cells upon them. Chapter 6 will present current limitations of graphene as a transparent conductor in OPV applications and provide possible solutions via several routes such as nondestructive surface modification of the graphene electrodes to make them suitable for hole transporting layer deposition or the introduction of alternative hole transporting layer materials which can avoid surface modification processes. Finally, Chapter 7 will conclude this thesis by summarizing the main discoveries of this research work. 24 CHAPTER 1: INTRODUCTION (This page intentionally left blank) 25 CHAPTER 2: ORGANIC PHOTOVOLTAICS Chapter 2 2. Organic Photovoltaics 2.1 Background Recently, solar cells based on flexible organic semiconducting materials have drawn significant attention as an alternate source of clean energy over conventional silicon based inorganic cells, offering many advantages, including light weight, flexibility, high transparency, potentially low processing cost, high-throughput manufacturing processes using roll-to-roll or spray deposition, and relatively simple fabrication methods. The possibility of ultra-thin, flexible devices and integration into appliances or building materials is of great interest. Also, the tuning of color is feasible through manipulations of chemical structures [4-7]. The field was initiated by application of small organic molecules as pigments [5,8]. The development of semiconducting polymers [9] and the application of these materials into organic solar cells resulted in remarkable improvements in the past years [10,11]. The ability of semiconducting organic materials to transport electrical current and to absorb light in the ultraviolet (UV)-visible part of the solar spectrum is primarily due to the sp 2-hybridization of carbon atoms. In conducting polymers, the electrons in the pz-orbital of each sp2-hybridized carbon atom form n-bonds with neighboring p, electrons in a linear chain of sp2 -hybridized carbon atoms, which leads to dimerization (an alternating single and double bond structure). Due to the isomeric effect, these nelectrons are delocalized in nature, resulting in high electronic polarization (see Figure 2.1) [12]. 26 .. .... .... .. . . ....... ..... . .. CHAPTER 2: ORGANIC PHOTOVOLTAICS bfnd1 Figure 2.1 Typical structures of organic semiconductors [12]. The delocalized electron systems can be identified by the alternating single and double carbon-carbon bonds. The bottom panel is a schematic diagram of a PPV chain segment. To form a molecule, an sorbital for each carbon atom becomes mixed with two of the p-orbitals (sp2 _ hybridization), resulting in a bonds that hold the molecule together, leaving one electron per carbon atom in a pz-orbital. The overlap of pz-orbitals of neighboring carbon atoms allows the delocalization of those electrons into xt-orbitals along a polymer backbone. 2.2 Inorganic vs. Organic Solar Cells Major differences between organic semiconductors and crystalline, inorganic solid-state semiconductors are the relatively low charge-carrier mobility and small diffusion length of the primary photo-excitations (excitons) in rather amorphous and disordered organic materials [5,13], which leads to a large effect on the device geometry and efficiency. These excitons usually require strong electric fields to dissociate them into free charge carriers which are the desired final products for photovoltaic conversion, due to their generally large excitonic binding energies which usually exceed those of inorganic semiconductors [14,15]. Most organic semiconductors are hole conductors and have an optical band gap of around 2 eV, which is considerably higher than that of silicon (1.1 eV). That is, upon photon absorption, free charge carriers (unbound electron-hole pairs) are generated in conventional semiconductors, whereas electrostatically bound charge 27 CHAPTER 2: ORGANIC PHOTOVOLTAICS carriers (bound electron-hole pairs, exciton) are formed in excitonic organic semiconductors. In general, it is the minority charge carriers (holes in the n-type material and electrons in the p-type material) that contribute to the photocurrent in solar cells. Therefore, in conventional inorganic solar cells, the diffusion length of the minority carriers, the built-in potential, and the bulk material properties dominate the device characteristics. On the other hand, in organic solar cells, exciton diffusion length, interfacial charge separation rate, and interfacial material characteristics dominate the device properties. An approximate analysis has been done which can be used as a guideline to distinguish between the two different phenomena through a dimensionless constant y corresponding to the following relation [16,17]: Y = ( )~ TB ( 2 41T EokBro e mef e2 (2.1) y > 1: Excitonic organicsemiconductor y < 1: Conventional inorganicsemiconductor where rc is the width of the Coulomb potential well at kBT, rB is the Bohr radius of the relevant charge carrier, q is the electronic charge, co is the permittivity of free space, kB is Boltzmann's constant, me is the mass of a free electron in a vacuum, meff is the effective mass of the electron in the semiconductor, and T is the absolute temperature. Figure 2.2 schematically illustrates this fundamental difference. Excitonic features of organic semiconductors originate from the generally low dielectric constants and weak interatomic electronic interactions of non-covalently bonded organic molecules. Therefore, the electron wave function is spatially localized in the potential well of its conjugate holes and vice versa. 28 CHAPTER 2: ORGANIC PHOTOVOLTAICS 0T CSC >~1 Al - electron -0.15 wavefuinctions c CS = 15 xsc s=4 -0.32 1 .1 -150 -100 -50 0 50 100 150 Carrier separation distance, A Figure 2.2 Schematic description of the electronic wave function illustrating the fundamental differences between conventional semiconductors (CSC) and the excitonic semiconductor (XSC). When y > 1, that is the wave function sits inside the potential well, excitonic behavior is expected [18]. Despite the fundamental differences in the mechanism of free charge carrier generation, similar central charge transport equations from a conventional inorganic p-n junction relation can be applied to excitonic organic systems after taking into account the exciton dynamics. Therefore, the current density relations for electrons and holes can be described as follows [18]: J =qpanE+ qDnVn and Dn = p - ) qppE - qDpVp and Dp= (i) T (2.2) p (2.3) where Yu and yt are respectively the electron and hole mobility, n and p are the charge carrier density, E is the electric field, and Dn and DP are the diffusion coefficient from the Einstein relationship. These features mentioned above generally limit the light harvesting of the solar spectrum and the power conversion efficiency of organic solar cells. Nonetheless, the chemical flexibility of organic semiconductors and their potential for low cost, large- 29 CHAPTER 2: ORGANIC PHOTOVOLTAICS 00 4SI VYW IlM tI YWV I I V I I I I I ~o I I E *0 I IE EE I II VV I I c~.i I I C0r ID U) I scale production still draws significant attention to these materials in research worldwide. I T' #C a1 i i Figure 2.3 compares various types of solar cells found in current PV research. -J 0- Mo 0 9 il A& i (%) AUO!3W Figure 2.3 Progress of PV research efficiencies by design type from NREL [2012] [19]. 30 CHAPTER 2: ORGANIC PHOTOVOLTAICS 2.3 Principles of Operation 2.3.1 Organic Semiconductor In organic photovoltaic solar cells, the primary reaction upon light absorption occurs through organic semiconducting materials, which are often referred to as the "active layer". Most of the incident solar spectrum is absorbed by the active layers which generate free charge carriers. Typical active layer materials used in OPV are small molecules or polymers, which primarily differ by their molecular weight and solubility in common solvents. Small molecules are typically deposited via vacuum-evaporation and polymers are mostly solution processed, such as by spin-coating or ink jet printing. Therefore, small molecular solar cells are expected to be more sensitive to the order of deposition between the organics and metals, i.e. organic-on-metal or metal-on-organic. For instance, as shown in Figure 2.4, when small molecules are evaporated onto the metal surface limited interaction should occur at the interface, whereas, in the opposite case, chemical reaction or inter-diffusion between the two materials is expected [20]. Organic on metal Low-T source Metal on organic High-T source 49 abrupt C 0 inera'000000000o00000 000000000000000 000000000000000 000000000000000 000000000000000 000000000000000 000000000000000 .cn L0 0 0 =0= 11; - *0 41 4m 1 141 Figure 2.4 Schematic illustration of small molecule deposition under vacuum, organic on metal (left) or metal on organic (right) [20]. 31 CHAPTER 2: ORGANIC PHOTOVOLTAICS As mentioned earlier, organic semiconductors are based on a conjugated it electron system. The conjugation system, i.e. an alternation of single and double bonds, allows it electrons to be more mobile than c7 electrons, which leads to hopping transport from site to site. Upon light absorption, the promotion of an electron from the it-orbital (bonding state) to the it*-orbital (anti-bonding state), i.e. a 7t-7c* transition, occurs. These molecular it-t* orbitals correspond to the highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO) in conjugated organic materials. The HOMOs and LUMOs respectively form the valence band (VB) and the conduction band (CB) in crystalline inorganic semiconducting materials are well semiconductors. distinguished Hence, excitonic from the solid-state organic inorganic semiconductors in several fundamental aspects, such as the following: * Due to the weak inter/intra molecular van der Waals forces in solid-state organic semiconductors, VBs and CBs are not formed and the electronics states are localized on single molecules and do not form a large band. As a consequence, excitonic binding energies are relatively large and excitons do not spontaneously dissociate into free charge carriers upon incident photo-excitation. The binding energies of electron/hole pairs for organic semiconductors are typically a few hundred meV [21-25], whereas those of crystalline inorganic semiconductors are a few meV. Therefore, thermal energy/excitation at room temperature, -25 meV, is not sufficient for the spontaneous exciton dissociation into free charge carriers. * Charge transport processes occur via hopping a mechanism between the localized states unlike transport within a band as is observed in inorganics materials. Therefore, the mobility of organic semiconductors is considerably lower than their inorganic counterparts by several orders of magnitude. The charge carrier mobility of organic materials typically ranges from 10-5 to 1 cm2 V s- 1,compared to Ie of~1400 cm V's' for silicon [13]. * Because of the low carrier mobility in organic semiconductors, the thickness of active organic layers in a photovoltaic cell is generally limited to a few hundred nanometers. However, since the absorption coefficient is relatively high in the 32 CHAPTER 2: ORGANIC PHOTOVOLTAICS UV-Vis (ultraviolet - visible) regime, ~10-7 cm~' [6], most of the light absorption from the solar spectrum can occur within ca. 100 nm thick layers. 2.3.2 Operation of Organic Solar Cells Although the simplest device structure consists of a single organic layer sandwiched between the two metal contacting electrodes, most of the recent advances in organic photovoltaic solar cells are based on advances made in the heterojunction donor-acceptor structure which was pioneered by Tang [4]. The basic idea in designing this structure was using two materials with different electron affinities and ionization potentials whose interface provides the driving force for exciton dissociation into free electrons and holes (Figure 2.5): electrons will favor the material with the large electron affinity (acceptor, ntype) and the holes will be accepted by the material with the lower ionization potential (donor, p-type). Vacuum Level Ak IP =1i EA LUMO h' Donor HOMO Acceptor Figure 2.5 Schematic description of charge separation of a photo-generated exciton at the donor/acceptor heterojunction [26]. 33 CHAPTER 2: ORGANIC PHOTOVOLTAICS In the heterojunction structure, the conversion of light into electricity can be described by the following steps and this process is schematically illustrated from Figure 2.6 [6]: " Upon the absorption of a photon, an electron from its bound state is excited to the lowest unoccupied molecular orbital (LUMO) and leaves a hole in the highest occupied molecular orbital (HOMO). This formation of an excited state forms a tightly bound electron/hole pair, which is called an exciton. * Excitons diffuse to the donor/acceptor interface where exciton dissociation occurs by the potential drop at the interface. Exciton dissociation is energetically favorable when the ionization potential and electron affinity of the donor are higher than those of the acceptor. Hence, electrons from the LUMO of the donor can transfer to the LUMO of the acceptor, and similarly holes from the HOMO of the acceptor can travel to the HOMO of the donor. " After charge separation, free charge carriers are transported within the organic semiconductor to their respective electrodes, i.e. electrons migrate to the cathode and holes are transported to the anode. 34 CHAPTER 2: ORGANIC PHOTOVOLTAICS r- (A) Donor Acceptor 2 (B) LUMO G- 0LUMO 4 HOMIO h& HMO~ cathode Figure 2.6 electron donor electron acceptor anode Schematic illustration (A) and energy level diagram (B) of the charge separation process in a donor/acceptor heterojunction architecture. In (B), photocurrent generation consists of several processes corresponding to the following sequential steps: (1) photon absorption. (2) charge separation. (3) charge transfer. (4) geminate recombination (before charge separation). (5) interfacial recombination. (1)/(4) and (3)/(5) are the competing processes for efficient photocurrent generation [6,15]. 35 CHAPTER 2: ORGANIC PHOTOVOLTAICS The efficiency of the above process can be related to the number of charge carriers collected at the electrodes that have been generated due to photo-excitation, which is often described by the incident photon to charge carrier efficiency (IPCE). The IPCE is the ratio of the number of incident photons of a given wavelength per unit time and area to the number of electrons per unit time and area leaving the device under the short circuit condition: IPCE = # of carrierscollected at electrodes # of incident photons 2.3.3 Device Architecture Efficient exciton dissociation at the heterojunction requires the donor and acceptor length scale to be within the range of the exciton diffusion length (- tens of nm), while the optimal thickness of the active layers need to be comparable to the penetration depth of the incident light (typically < 200 nm). Typical heterojunction configurations are described in Figure 2.7 [15]: showing the bi-layer heterojunction and bulk heterojunction geometry. In a bi-layer heterojunction structure, donor and acceptor organic materials are sequentially stacked through vacuum sublimation or spin-coating. In a bulk heterojunction device, active layers composed of mixtures of donor/acceptor materials are deposited by co-evaporation under vacuum or spin-coating. The bi-layer structure enables directional photogenerated charge transfer with reduced recombination losses, while the bulk heterojunction geometry ensures higher interfacial junction areas which lead to efficient exciton dissociation. 36 CHAPTER 2: ORGANIC PHOTOVOLTAICS tight 1ght Large interfacial area due to phase separation in the blend, but here percolation is needed Direct path for charge carrier to electrodes Typical donor/acceptor heterojunction structures in organic solar cells [15]: heterojunction and (b) a bulk heterojunction. bi-layer (a) a Figure 2.7 Due to the nature of structural configurations in each device type, exciton dissociation is less efficient from the bi-layer heterojunction device, whereas the percolating network for charge transport often becomes an issue in bulk heterojunction device. Figures 2.8 and 2.9 list commonly used small molecules and polymers in organic photovoltaic systems. 37 CHAPTER 2: ORGANIC PHOTOVOLTAICS Me-Ptcdi N N /N Zn N \ / CHi-N N-CH3 N\ -N N ZnPc N 0 N N-&u- -0 N -N N N N N N PTCBI CuPc Figure 2.8 Small molecules which are typically vacuum-evaporated: phthalocyanine), Me-Ptcdi ZnPc (zinc- (N,N'-dimethylperylene-3,4,9,1 0-dicarboximide), fullerene C60 , copper phthalocyanine buckminster perylenetetracarboxylic bisbenzimidazole (PTCBI) [7,15]. 38 (CuPc), and the 3,4,9,10- CHAPTER 2: ORGANIC PHOTOVOLTAICS N 0 s MDMO-PPV PFB P3HT 0 --- N N Me - +01. -o NC CN-MEH-PPV PCBM F8BT Figure 2.9 Solution processed conjugated polymers. Upper row: p-type donor polymers: MDMO-PPV (poly[2-methoxy-5-(3,7-dimethyloctyloxy)]- 1,4-phenylenevinylene), P3HT (poly(3-hexylthiophene-2,5-diyl), and PFB (poly(9,9'-dioctylfluorene-co-bis-NN'-(4butylphenyl)-bis-N,N'-phenyl-1,4-phenylenediamine). Lower row: n-type acceptor (poly-[2-methoxy-5-(2'-ethylhexyloxy)- 1,4-( 1CN-MEH-PPV polymers: cyanovinylene)-phenylene), PCBM (1-(3-methoxycarbonyl) propyl-1-phenyl[6,6]C61), and F8TB (poly(9,9'-dioctylfluoreneco-benzothiadiazole) [7]. 39 CHAPTER 2: ORGANIC PHOTOVOLTAICS 2.4 Organic Solar Cell Characterization Figure 2.10 describes the fundamental nature of photoexcitation in a single layer organic solar cell, where localized Frenkel (charge transfer) excitons [5,27] are generated upon the optical excitation in the active organic regions of the device. As mentioned earlier, tightly bound electron/hole pairs of excitons should be dissociated in order to contribute to photocurrents in organic solar cells. In general, this process is assisted by the built-in electric field in the device. typically low (~10 6 The electric field strength in organic electronic devices is V/cm) [28], which leads to generally low exciton dissociation efficiency (U7ED < 10 %). However, after the exciton dissociation, the freed electrons and holes are mostly collected at the respective electrodes due to the presence of the built-in electric potential, approaching the charge collection efficiency (7cc) of -100 %. If most of the photons were absorbed by the organic layers, i.e., thicker than the optical absorption length, the absorption efficiency (A) can also be assumed to be -100 % [27]. 'qA0..0 C~1 00% Figure 2.10 Schematic illustration of the photocarrier generation process upon the incident light in a single organic layer. The dips in the energy level diagram describe the binding energy of the exciton [5]. The photocarrier generation process in a more energetically favorable structure, i.e., donor/acceptor heterojunction configuration, is also described in Figure 2.11. For a given exciton binding energy (Eex) and a charge-transfer state energy (IPD - EAA), 40 CHAPTER 2: ORGANIC PHOTOVOLTAICS exciton dissociation will only occur when Eex > IPD - EAA, since the charge-transfer process is energetically favorable under this condition but not the other. Ee Eex IPD-EAA IPD-EAA I. don )r donor acceptor Li acceptor Eex > IPD-EAA Eex < IPD-EAA Figure 2.11 Schematic illustration of photocarrier generation requirements in donor/acceptor heterojunction organic layers for Eex > IPD - EAA (left) and for Eex < IPD - EAA (right) [5]. 2.4.1 Quantum Efficiency Photocurrent generated in solar cells is directly related to the number of free charge carriers that reach the respective electrodes after the generation of the exciton. In general, the efficiency of this process can be described by the quantum efficiency, particularly as an external quantum efficiency (EQE). EQE describes the efficacy of the free charge carriers that are collected at the electrodes per incident photon and is an end result of several cascaded processes: * Photon absorption that leads to the exciton generation: 77A * Exciton diffusion to the donor/acceptor interface: 7JED * Exciton dissociation at the donor/acceptor interface: * Collection of free charge carriers at the electrodes: 7cc 77CT Therefore, in organic photovoltaic devices EQE can be defined as the following: 41 CHAPTER 2: ORGANIC PHOTOVOLTAICS 71EQE(XV) = 1)A(X)TED71CT(V)71CC(V) (2.5) = IIA(X)1QE(V) where X is the wavelength of the incident light, V is the applied voltage, and iIQE is the internal quantum efficiency (IQE) defined as, _Q - 91IQE # of carrierscollected at electrodes # of photons ABSORBED in the device (2.6) which should be well distinguished from the EQE, _ JEQE -# # of carrierscollected at electrodes of photons INCIDENT on the device Here, each parameter is associated with the fundamental processes occurring in excitonic organic solar cells upon the incident light absorption and describes the following process (Figure 2.12): " 77A describes the absorption efficiency of photons in the photoactive regions, i.e., nA Photon absorption # of photons absorbed in the active area Total # of incident photons on the active region efficiency be determined can from the organic layer thicknesses, dielectric constants, and the optical field intensities within the active layer [29]. . ?7ED is the efficiency of excitons that diffuse to the donor/acceptor interface before the recombination occurs. Due to the generally shorter exciton diffusion length (~50 - 100 A) compared to the optical absorption length (~500 - 1000 A) [30-32], this process is usually the primary limiting step. * 77cT is the efficiency of the charge-transfer process of excitons to dissociate into free charge carriers leaving holes in the donor and electrons in the acceptor. 42 CHAPTER 2: ORGANIC PHOTOVOLTAICS Since the time scale of charge-transfer (CT) is much shorter than the other processes, a few hundred femto-seconds [33,34], 7CT approaches ~100 % in typical organic electronic devices. * cc is the efficiency of freed charge carrier collection at the respective electrodes, which is typically considered to be -100 % in most organic photovoltaic devices under short circuit conditions. c= # of carrierscollected at electrodes Tcc T otal # of carriersgenerated after exciton dissociation (2.9 (29 ED LCT an A TICC I anode \ ! I 66 cathode donor Figure 2.12 Schematic illustration of the photocurrent generation process upon the incident light entering the donor/acceptor heterojunction organic layers [5]. We note well that the incident photon to charge carrier efficiency (IPCE) mentioned earlier refers to EQE (external quantum efficiency) not to IQE (internal quantum efficiency). The IPCE can be measured as follows: IPCE = (. x \ ( 'sc) } \'PINJ (2.10) where h is the Planck constant, c is the speed of the light, e is the elementary charge, JsC is the photocurrent density, A is the wavelength of the incident photon, and PIN is the 43 CHAPTER 2: ORGANIC PHOTOVOLTAICS incident photon flux at wavelength 2. The IPCE is generally expressed in percent, and hence it can be simplified as follows: IPCE (%) = 124oxphotocurrent density [m] w(.1 photon f lux[- .] xwavelength [nm] 2.4.2 Equivalent Circuit Model and Key Parameters of OPV Cells The simplest description of a solar cell can be given from an electric circuit where the solar cell is connected to an arbitrary external load as shown in Figure 2.13. Upon the incidence of light, that is when the solar cell is operational, the voltage developed across the terminal when it is disconnected is defined as the open circuit voltage Voc and the current that flows through the terminal when it is connected is called the short circuit current Isc. For any other intermediate load resistance RL, a relation between the current (I) and the voltage (V) across the external load can be developed simply from Ohm's law such that V = IRL. Therefore, the current-voltage characteristic of a solar cell under illumination can be determined from the measured values of V and I, and the photocurrent generated at the short circuit condition is dependent on the incident light. Solar Cell OD Figure 2.13 Schematic description of a solar cell connected to an external load. 44 CHAPTER 2: ORGANIC PHOTOVOLTAICS In Figure 2.13, a potential drop is also developed across the solar cell which generates a current in the opposite direction to the photocurrent, thus reducing the total net current from the short circuit value. This reverse current generated under the dark condition in response to the voltage is referred to dark current, which is approximated as the current that flows across the device under an applied bias in the dark condition. This assumption is generally employed for typical solar cell characterizations in practice via the superposition approximation [35]. The overall net current is thus the superposition of the short circuit photocurrent and the reverse dark current as shown in Figure 2.14. Jsc Light current C CD C Dark current Bias voltage, V Figure 2.14 Current density-Voltage characteristics of an ideal diode under the dark and the light conditions. Ideally, the total net current can be obtained by shifting the dark current by a constant amount which is equal to the short circuit photocurrent [35]. Although the behavior is not ideal, most photovoltaic devices generally show rectifying diode characteristics under a dark environment, that is a higher current flow under forward bias than reverse bias, and hence more detailed current-voltage characteristics of a solar cell can be established based on an equivalent circuit model incorporating a diode and a current source with parasitic resistances which account for the non-ideality of the diode as described in Figure 2.15. 45 CHAPTER 2: ORGANIC PHOTOVOLTAICS Rs Jdark J(R) Jph t V Rp J Figure 2.15 Equivalent circuit model of a solar cell with parasitic series (R,) and parallel ( Rp ) resistances accounting for the non-ideality of a diode. Jdark and jph are, respectively, the dark current density and photocurrent density, and J(Rp) is the current density due to RP, describing current flowing through Rp [29]. As mentioned earlier, the basis of this model lies in the photocurrent generation from the cell upon the incident light, i.e., the current source Jph, which flows in the opposite direction to the dark current Jdark and is limited by the series (R,) and parallel (Rp) resistances. The series resistance arises from several resistances in the solar cell material and the interfaces, and is a limiting factor to the current flow at higher bias. The parallel (shunt) resistance describes any leakage of the current flow from the cell and affects the rectifying behavior of the diode (see Figure 2.16). C a) C.) 1.. 0 Rs increasing R, decreasing 7 I Bias Bias I Figure 2.16 Effects of the series (R,) and parallel (Rp) resistances to the current-voltage characteristics [35]. 46 CHAPTER 2: ORGANIC PHOTOVOLTAICS From the equivalent circuit model in Figure 2.15, the current density-voltage (J-V) characteristics can be derived by the generalized Shockley equation as follows [30,36-38], R+R s xp .2 q (VJRs)+V( * n: ideality factor of the diode * kB: Boltzmann constant * T: temperature * q: charge of the electron * V: voltage across the solar cell * Rs: series resistance * RP: parallel resistances * Jph: photocurrent " Js: reverse saturation dark current of the diode defined by, Is =JsoCXP 2B) nk( (2.13) * Eg = IPD - EAA: activation energy barrier at the donor/acceptor heterojunction " JsO: temperature independent prefactor The photocurrent Jph is related to the incident solar spectrum and the cell photoresponse, i.e., quantum efficiency (EQE) of the cell which is the probability that an incident light photon at wavelength A (or energy E) delivers one electron to the external load or circuit, and can be calculated by convolving the two parameters as: Jph(V) = q f S(A)EQE(1, V)d1I (2.14) or Jph(V) = q f -P(A)EQE(A, hc 47 V)d1I (2.15) CHAPTER 2: ORGANIC PHOTOVOLTAICS where S(A) is the incident solar photon flux density, the number of photons of wavelength (A) between (2) + dA, per unit time, area, and wavelength. P(A) is the spectral irradiance of the incident light, the incident power of the solar radiation per unit area and wavelength. The photon flux density is thus related to the spectral power density as: S(,) = -P(,a) hc = (2,16) E via (2.17) or, to convert the photon energy into electron-Volts (eV). for wavelength A in nm. E 1240 eV (A/nm) (2.18) Figure 2.17 demonstrates a typical quantum efficiency of a gallium arsenide (GaAs) solar cell compared with the solar spectral intensity. QE of gallium arsenide cell -- Solar photon flux density U, U) (3 0 M~ i 200 400 600 800 1000 1200 1400 1600 1800 Wavelength / nm Figure 2.17 Quantum efficiency of a GaAs solar cell compared with the solar photon flux density in arbitrary untis [35]. 48 CHAPTER 2: ORGANIC PHOTOVOLTAICS By rearranging Equation 2.12 under open circuit conditions at J= 0, the open circuit voltage can be derived as follows: V"OC == nkBT q (ph(VOC) ~In \ JS + - + (2.19) Js p) JSRy) assuming that the parallel shunt resistance is much higher compared to the series resistance, then Equation 2.19 can be further simplified as: S= in (JPhVoc) (2.20) + 1) which is the more commonly used relation. Voc saturates at a maximum value, VOax, when the quasi-Fermi levels of the donor and acceptor materials are pinned at the high currents. By taking account of the exciton binding energy, the maximum achievable intrinsic open circuit voltage can then be obtained as [29]: 2 Vmax 4 q 1TEOErrDA (2.21) where * - 0 : vacuum permittivity * Er: relative dielectric constant of the bulk organic material * rDA: initial separation distance of the optically generated hole/electron pair in the donor/acceptor layers, immediately following the charge transfer * Eg: activation energy barrier at the donor/acceptor heterojunction * EB: binding energy of the bound electron/hole pair following the charge transfer. The region of solar cell operation is in the range between V [0, Voc], where the cell delivers power. The device power density P = JV reaches a maximum value Pm which occurs at a point (Im, Vm), called the operating point of the cell. 49 The power CHAPTER 2: ORGANIC PHOTOVOLTAICS conversion efficiency irp of the cell is defined as the ratio of the maximum power density Pm to the incident light power density PO as such: TI = im 0 PO (2.22) Another important parameter often used is the fill factor (FF) which describes the degree of rectification from the diode curve. FF = ImVm (2.23) The efficiency can be rewritten using FF from Equation 2.22 as: 11 = JscVocFF (2.24) The key parameters derived above are illustrated in Figure 2.18 from the currentdensity curve. 50 CHAPTER 2: ORGANIC PHOTOVOLTAICS 0" 20 E 10 Vm _ 0.O 00 0 SC 0-20 0.0 0.3 0.6 0.9 Voltage (V) Figure 2.18 Key photovoltaic parameters in organic solar cells. In general, current-voltage behavior of a solar cell can be characterized into three distinctive regions as follows [39]: " First linear region at negative and low positive biases: Current flow is limited by the parallel shunt resistance * Exponential regime at intermediate positive biases: Current flow is characterized by the diode behavior " Second linear region at higher biases: Current flow is limited by the series resistance These features are schematically described in Figure 2.19 with corresponding circuit elements in the equivalent circuit model of a typical solar cell. Under open circuit condition, there is no current flowing through the device and the photocurrent recombines internally via R. and the diode. Due to the relatively large values of R., the current 51 CHAPTER 2: ORGANIC PHOTOVOLTAICS density is dominated by the exponential behavior of the diode in most cases, and the open circuit voltage is also primarily affected by the properties of the diode. 1000 100 10 E -0 E fiaht 1 0.1 D 0.01 n mi -1.25 -0.75 -0.25 0.25 0.75 1.25 VaPP 1.75 2.25 U[Vm Figure 2.19 Illustration of the current density-voltage corresponding equivalent circuit model [39]. 2.5 characteristics with a Solar Radiation The power density of the solar radiation outside the earth's atmosphere on a plane perpendicular to the direction is referred to as the solar constant, with a constant value of 1353 W/m 2 . On the other hand, terrestrial solar radiation varies in the intensity as well as the spectral distribution depending on the position of the earth and the sun. Solar radiation is also attenuated when it passes through the earth's atmosphere. Since the spectral distribution of the solar radiation depends on the attenuation, various solar spectra can be measured at different locations on the surface of the earth. Therefore, a terrestrial solar radiation standard has been developed and the current standard in PV research uses AM (Air Mass) 1.5 radiation as the standard spectral distribution. The optical air mass is defined as the ratio of an actual path length of the sunlight to the 52 CHAPTER 2: ORGANIC PHOTOVOLTAICS minimal distance when the sun is at the zenith. When the sun is at its zenith, the optical air mass is unity and the radiation is referred to as air mass one (AM 1.0) radiation. When the sun is at an angle 0 to the zenith, the Air Mass is defined as [35] Air Mass = (cos)-1 (2.25) AM 0 radiation is the extraterrestrial spectrum of solar radiation outside the earth's atmosphere and AM 1.5 corresponds to an angle 0 = 48.20 between the zenith and the position of the sun. The irradiance of AM 1.5 radiation is 827 W/m 2 . However, for simplicity, the value of 1000 W/m 2 is adapted as the standard. The solar radiation spectrum is also related to the air mass and Figure 2.20 shows the spectral power density of some commonly used air mass radiation spectra. 2.51 6000 K black body 2.0 1.5 AMO radiation AM1.5 radiation 1.0 0 0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 Wavelength [pm] Figure 2.20 Spectral power density of sunlight at different radiation spectra [35]. 53 2.0 CHAPTER 2: ORGANIC PHOTOVOLTAICS (This page intentionally left blank) 54 CHAPTER 3: GRAPHENE Chapter 3 3. Graphene 3.1 Background Monolayer graphene is a hexagonal arrangement of carbon atoms forming a one-atom thick planar sheet. This layer is the building block of graphite and carbon nanotubes (Figure 3.1) [40] and it has been widely studied by theorists since the middle of the last century [41,42]. The successful isolation of single- and few- layer graphene by the mechanical cleaving of HOPG [43] has opened up an explosion of research activities, and significant attention has been captured by graphene's outstanding properties, such as high electron and hole mobility (-up to 200,000 cm 2 V-I s-') [44,45], high current carrying capability (up to 3 x 108 A/cm 2)[46] and high mechanical robustness [47]. The electrical conductance of graphene is sensitive to the absorption or desorption of a single gas molecule [48]. Graphene has also been shown to have a uniformly high transparency in the visible and near infrared region of the solar spectrum and can be used as ultra-thin transparent electrodes [49]. Furthermore, unlike other nanostructured materials, the twodimensional structure of graphene makes integration into planar devices possible. Therefore, graphene sheets show great potential as another material option for future electronics applications. 55 CHAPTER 3: GRAPHENE YY Y (a)0 (b) -Y A- (c) Figure 3.1 Graphene is a 2D building block for other carbon materials of all other dimensionalities. It can be (a) wrapped up into OD buckyballs, (b) rolled into ID nanotubes, and (c) stacked into 3D graphite [40]. 3.1.1 Synthesis Methods Driven by the enormous research interest and need, several techniques were developed for graphene synthesis or fabrication. These include: mechanical cleavage of HOPG [43]; ultra-high vacuum (UHV) annealing of single-crystal SiC (0001) [44,50]; dispersing graphite into solution via various chemical routes [51,52] and deposition on substrates; single crystal metal-based epitaxial growth under UHV [53] or chemical vapor deposition methods [54]. However, these methods either have a significant limitation to the 56 CHAPTER 3: GRAPHENE substrate type and/or produce only graphene flakes with a random shape, size, thickness and location on the substrate. Figure 3.2 illustrates several images of graphene synthesized among the various routes from the aforementioned methods. (c)i Figure 3.2 Graphene synthesis from various routes [55]. (a) Graphene flakes on a SiO 2 wafer prepared by the scotch tape method. (b) Left: Graphene suspension prepared from microcrystals obtained by the ultrasound cleavage of graphite in chloroform. Right: The suspension printed on a flexible substrate. (c) CVD graphene grown on a Ni thin film and transferred on a Si wafer. (d) Graphene grown by UHV annealing of single-crystal SiC. For large scale electronic applications, it is necessary to obtain large area, single crystalline graphene films with controlled number of layers, ideally on any substrate. In our group (Nano Materials and Electronics Lab, NMElab), we have successfully synthesized continuous large area single- to few- layer graphene films using chemical vapor deposition on a commercially available copper foil substrate under low pressure (LPCVD) or atmospheric pressure (APCVD) conditions. The details will be explained in Chapter 4. 57 CHAPTER 3: GRAPHENE 3.2 Physical Properties of Graphene 3.2.1 Electrical Properties The graphene honeycomb lattice is composed of two equivalent carbon sub-lattices, A and B, as shown in Figure 3.3(a). Shown in Figure 3.3(b) is the first Brillouin zone, with high-symmetry points M, K, K', and 1. s, px and pv orbitals of carbon atoms form (i bonds with neighboring carbon atoms, and n electrons in the pz orbital form n (bonding) and n* (anti-bonding) bonds. This 2D crystal lattice structure is the primary contributor to the many unique electronic properties of graphene, such as high carrier mobility and an ambipolar field effect. (a) a (b) ky K %\ b2 B\ A BZ Figure 3.3 K' b1 M A k (a) Graphene lattice structure and (b) the first Brillouin zone of graphene [56,57]. Indeed, Kim et al. reported a hole carrier mobility of over 200,000 cm 2/N-s from a mechanically exfoliated, single layer suspended graphene (Figure 3.4(a-b)) [45]. The ambipolar charge transport effect is illustrated in Figure 3.4(c) along with the band structure of graphene [40]. Under a positive gate bias, the Fermi level rises above the 58 CHAPTER 3: GRAPHENE Dirac point with populated electrons in the conduction band, and under negative gate bias, the Fermi level drops below the Dirac point with more holes introduced into the valence band. This ambipolarity can open up novel device structures by dynamically controlling the doping levels entirely by gating, which is a completely different approach than silicon based semiconducting logic. (c) I, - (b) E 0T 11Y 10 -1 0 I _Uo -30 0 V1(V) 30 ...... ....... .. 60 1i' n (10"cm-) Figure 3.4 (a) Scanning electron microscopy (SEM) image of a suspended graphene sheet in a field effect transistor. (b) Field-effect measurements indicating a mobility greater than 200,000 cm 2/(V-s) [45]. (c) Schematic diagram of the graphene band structures for electrons and holes, and the corresponding ambipolar field effect [40]. Due to the symmetry of the crystal lattice of graphene consisting of two equivalent carbon sublattices, charge carriers in graphene are described by the Dirac equation, rather than the usual Schrddinger equation for nonrelativistic quantum particles that is commonly adopted in solid state physics: In solid state physics, electrons and holes are typically described by the Schr6dinger equation with their effective masses whereas electrons and holes in graphene are interrelated (linked or conjugated), often referred to 59 CHAPTER 3: GRAPHENE as charge-conjugation symmetry [58,59]. The Dirac equation describes relativistic quantum particles with half integer spin, such as electrons and the Dirac spectrum is featured by the presence of anti-particles. For Dirac particles with mass m, there exists an energy gap between the minimum electron energy (E0 = mc2 ) and the maximum positron energy (-E). Here, the energy is linearly dependent on the wavevector only if the electron energy is E >> E.. On the other hand, for massless Dirac fermions, an energy gap does not exist and the linear dispersion relation is valid for any energy [60,61]. The overlap of the electron subbands at the Brillouin zone formed by superposition of a symmetric and an anti-symmetric wave functions on the two carbon sublattices results in the famous cone-shaped energy spectrum close to the Dirac points, K and K'. As a result, quasiparticles in graphene have a linear dispersion relation, E = hkvF, similar to massless relativistic particles, such as photons. Here, the role of the speed of light is incorporated in the Fermi velocity via VF ~ c/300. The band structure is described in Figure 3.5, where the conduction and the valence bands touch at the K and K' points. The unique feature of the graphene band structure is that the electron energy dispersion is linearly dependent on the wave vector around the symmetry points [62] and the spectrum is similar to the Dirac spectrum for massless fermions. Figure 3.5 Schematic illustration of the graphene band structure [63]. 60 CHAPTER 3: GRAPHENE The dispersion relation of the n electrons can be described by the tight-binding model considering the first nearest neighbor interactions only as [42,64], E±(kX, kX) = yO 1 + 4 cos X 02 cos - + 2 4cos2 LL 2 (3.1) where a = Vacc and acc is the carbon-carbon distance (1.42 A), and the transfer integral y0 is the matrix element between the 7r orbitals of neighboring carbon atoms with a magnitude of approximately 3 eV. The minus sign corresponds to the 7r band (bonding state), and the plus sign refers to the 7* band (anti-bonding state). In graphene, the Tc band is fully occupied and the m* band is empty. By Taylor expanding Equation 3.1 around symmetry points K and K', a linear dispersion relation can be obtained as, Ez(k) = kylk| where y = hvF = V3ay 0 /2, respect to the K point. VF is the Fermi group velocity and k is measured with This linear dispersion relation is obtained due to the crystal symmetry of graphene and the linear E(k) is the origin of many unique physical properties of graphene such as the half-integer quantum Hall effect and Berry's phase [40,60,63]. 3.2.2 Optical Properties Graphene can be visualized from the optical image contrast on a Si/Si0 the Si0 2 as an optical spacer, a result of an interference effect. 2 substrate using The origin of this phenomenon is attributed to the increased optical path length in the given geometry and the superior opacity of the graphene, which is dependent on the thickness of SiO 2 and the wavelength of the light. In the following, the optical contrast of a structure consisting of Si, SiO 2 , and a single layer of graphene is derived under normal light incidence from the air with a refractive index n, = 1 (See Figure 3.6). Herein, Si is assumed to be semi61 CHAPTER 3: GRAPHENE infinite with a complex refractive index n 3 (2) and the SiO 2 layer is at thickness d 2 and n 2 (A) with real part only [65,66]. Graphene is described by a thickness d, and a complex refractive index n1 (2). d, is assumed ~0.34 nm which is the extension length of the n orbitals out of the plane [67]. n 1 (A) ~ 2.6 - 1.3i is described by the refractive index of bulk graphite since the optical response of graphite under an electric field that is parallel to the in-plane of graphene is dominated by the in-plane electromagnetic response [66]. air: n0=1 graphene: n,=2.6-1.3i SiO 2: n 2 (A) d,=0.34nm d2 Figure 3.6 The geometry of graphene for optical analysis [68]. From the geometry in Figure 3.6, the intensity of the reflected light is expressed as [68,69], 1(n,) = (r 1 ei(1+2) + r 2 ei((1 +2) + r 3 e i(1+t+D2) + rlr 2 r 3 ei(41 (ei((1+42) + rir 2 e i(4D12) + rir 3 e-i((+(2) + r 2 r 3 ei(01 *2))Y 2)) x 2 (3.2) where r1 , r2 , and r3 are relative indices of refraction, given as, ri = r2 = no+nl ni-n 2 nj+n2 62 (3.3) (3.4) CHAPTER 3: GRAPHENE r3 = n2(3.5) n 2 +n and cI1 and 42 3 are phase shifts due to the changes in the optical path as, = P2 = r (3.6) A (3.7) rn 2 d 2 Finally, the optical contrast C is defined as the relative intensity of the reflected lights in the presence (n, # 1) and the absence (n, = no = 1) of the graphene layer, C =I(n 1 )-i(nj) (3.8) I(nl=1) Figure 3.7 shows the theoretical expectation of color contrast from a single layer graphene flake on a Si/SiO 2 substrate as a function of both the wavelength and the SiO 2 thickness provided by Equation 3.8. 0.15 700 0.10 800 0.05 500 0.00 F A 400 0 100 200 300 Si0 2 thickness (nm) Figure 3.7 Color contrast of a single layer graphene on Si/SiO 2 substrate as a function of SiO 2 thickness and wavelength [68]. 63 ........ .. .. ........... ................ ............... . CHAPTER 3: GRAPHENE Figure 3.8 illustrates optical micrograph images of graphite flakes with a varying number of graphene layers on a Si wafer with different thicknesses of SiO 2 layers and illuminated under varying wavelengths. For 300 nm SiO 2, graphene is clearly visible both under white light illumination, whereas graphene is not well observed from the 200 nm thick SiO 2 at white light illumination as shown in Figures 3.8(a) and (c), respectively. The top and bottom panels indicate the same samples but illuminated through various narrow-band filters. 41Onm 470nm 590nm 530nm (c) 200nm SiO2 (b) X= 470nm A= 710nm 650nm 530nm 560nm 590nm white light 650nm X =71Onm Figure 3.8 Optical images of graphite with varying thicknesses of graphene on Si wafers with different thicknesses of SiO 2 and under illumination at different wavelengths. The traces in (b) indicate contrast changes in a stepwise manner for 1 - 3 layers suggesting that optical contrast can be applied to identify the number of layers in the graphene on a given substrate [68]. The optical transmittance of a free standing monolayer graphene can be determined by using the Fresnel equations with a fixed universal optical conductance [70], 64 --.........::= . ... . ...... CHAPTER 3: GRAPHENE Go = e 2 /(4h) ~ 6.08 x 10-sfl-1 (3.9) which yields T = (1 + 0.5a)-2 1 - wa ~ 97.7 % (3.10) where a = e 2 /(4wEhc) = G0/(wEoc) ~ 1/137 is the fine structure constant [71]. The optical absorption of a single layer graphene can thus simply be derived as, A ~ 1 - T ~w a %2.3 % (3.11) over the visual spectrum, which is proportional to the number of layers (Figure 3.9). Since monolayer graphene can be considered as a two dimensional electron gas, it can be assumed that little perturbation from the neighboring graphene layers occur in a few-layer graphene. Therefore, a few-layer graphene sheet can be considered as optically equivalent to the superposition of monolayers of graphene films with almost no interaction between the layers. Figure 3.10 illustrates the Rayleigh (elastic light scattering) image of graphite on Si/SiO 2 which demonstrates that the monochromatic contrast increases with the number of graphene layers as a result of the superposition of single sheets of graphene [72]. 65 CHAPTER 3: GRAPHENE 0 0 C E 9f8 I60 6 25 0 50 distance (pm) Figure 3.9 Transmittance of a single and bilayer graphene suspended on a porous membrane. Optical absorption is measured as 2.3 % per each layer. The inset shows the sample platform with several apertures where graphene flakes are placed [71]. Figure 3.10 layers [72]. Rayleigh image of a graphite flake with a different number of graphene 66 CHAPTER 3: GRAPHENE Furthermore, the absorption spectrum of monolayer graphene is relatively flat over the visible-infrared (~450 - 2,500 nm) region with a sharp peak in the ultraviolet region (-270 nm), as a result of the exciton-shifted van Hove singularity in the graphene density of states [73]. In a few-layer graphene sample, other absorption features associated with inter-band transitions can be seen at both the lower and higher photon energies [74,75]. 3.3 Possible Application of Graphene in Opto-electronic Devices There are several possible applications of graphene in opto-electronic devices such as in photovoltaic solar cells, light-emitting diodes, touch screens, and photodetectors as illustrated in Figure 3.11: Graphene can be used as window electrodes in solar cells (inorganic, organic, or dye-sensitized) or as charge injecting electrodes in light-emitting diodes. With the wide absorption spectrum (from ultraviolet to terahertz range) of graphene, it can work as an efficient photodetector with a broad spectral response. Furthermore, the favorable mechanical robustness, flexibility, or chemical stability of graphene satisfies the criteria required for touch screens and flexible smart windows. 67 CHAPTER 3: GRAPHENE (a) (b) Trrm inpo olocLoht ugh (d) H subetrAft (e) OutN tot 0uAPuA (f) irefoti.ve c:7"07 LUqud--Walt K FW P*-n~ -m (j) wit Wpof Graphson.*as.d trampo. 06ctrodo c ya dispay (i) tro (h) (g) LePdymerliquid aystals Figure 3.11 Illustration of graphene based opto-electronic devices [76]. Schematic diagram of (a) inorganic, (b) organic, and (c) dye-sensitized solar cells. Schematics of (d) organic LED (light-emitting diode) and (e) photodetector. (f-g) Schematic of capacitive touch screen (f) and resistive graphene touch screen (g) [77]. (h-j) Graphene based smart windows. 3.4 Justification of Graphene as Transparent Conducting Electrodes in Organic Photovoltaics Applications As discussed earlier, transparent conducting electrodes (TCE) are widely used in optoelectronic devices such as solar cells or light-emitting diodes: metal oxides such as indium tin oxide (ITO) have been commonly used as window electrodes. Usually used as 68 CHAPTER 3: GRAPHENE thin films, these materials require low sheet resistance (Rsh) with high transparency (). Currently the dominant material used in the industry is indium tin oxide (ITO), which is commercially available at T - 85 % (at 550 nm) and Rsh - 20 Q/sq. However, these materials are not ideal options for organic photovoltaic applications due to several reasons: (1) non-uniform absorption across the visible to near infrared region [78]; (2) chemical instability: ITO has been known to inject oxygen [79] or indium [80] ions into the active layers of a device; (3) metal oxide electrodes easily fracture under large bending [81], not suitable for flexible solar cell applications; (4) limited availability of indium on the earth leading to increasing costs with time. As a result, there has been a constant search for alternative electrode materials with good stability, high transparency, and suitable conductivity. Several materials such as metal oxides (ZnO, TiO 2 ) [82], metallic nanowires [83], and carbon nanotubes have been studied as possible alternative materials to ITO. Nonetheless, with recent progress in graphene research, graphene has also shown better potential as a transparent conducting material. Rsh and T values of several materials are compared in Figure 3.12 [76] and high quality graphene electrodes (T-- 90 % (at 550 nm) and Rsh - 30 Q/sq) have successfully been demonstrated by Bae et al. [77]. 69 CHAPTER 3: GRAPHENE (a) 10080- 60C Graphene 40- ITO ZnO/Ag/ZnO 20- TiO/Ag/TiO 2 Arc discharge SWNTs - 200 400 800 600 Wavelength (nm) (b) 1,000 I n= 2 x 10 3cm2 V~) s * 0 100 on * ITO 10 * SWNTs n * Graphene CVD D 0.1 12 cm-2 = 2 x104 cm2 V~ s1 * Ag nanowire mesh 1 3.4 x 10 Graphene calculated 1 10 Thickness (nm) 70 100 1,000 CHAPTER 3: GRAPHENE (c) 100n=3.4x106c -2 =2 x10 4 cM2 V 1 s1 C 80 n = 130 cm -2 y4 = 2 X 10 3cm2 V -1- E i -ITO C I- 60- + WNTs -- Graphene CVD * Ag nanowire mesh o Graphene calculated 40 (d1 1 10 Sheet resistance (&/3) 100 IA0- / A 80 - -*- LPE EI RGO 60 * + PA Hs -&- CVD 3.4 x 1012 cm-2 #'= 2 40 10 S3--. -+- MC s-1 x 104 cm2 V-1 - 10 10 5 10 Graphene calculated 7,9 106 101 Sheet resistance (0/0) Figure 3.12 Properties of several transparent conducting electrodes [76]. (a) Transmittance of different materials. (b) Logarithmic plot of thickness dependence of sheet resistance. (c-d) Transmittance as a function of sheet resistance from different materials (c) and of graphene prepared by various methods (d). In (d), LPE, RGO, PAHs, and MC stand for liquid-phase exfoliation of pristine graphene, reduced graphene oxide, polyaromatic hydrocarbons, and micromechanical cleavage. 71 CHAPTER 3: GRAPHENE RsA and T can be considered as related to each other in the sense that both parameters can be evaluated through the response of electrons to either static (voltage) or Therefore, a relation between RsA and T for dynamic (optical) electric fields [84]. graphene sheets which can be a useful guideline for a first order design principle can be developed. The transmittance is related to the optical conductivity u0 p by [85], T = (i + Zo where Zo = 1/EOC 2 t (3.12) ~ 377 fl is the free space impedance, E, is the permittivity of free space, c is the speed of light, and t is the film thickness. Optical conductivity is further related to the Lambert-Beer absorption coefficient a as [86], 2a (3.13) P 0,PZo The sheet resistance can be controlled by the DC conductivity aDC as, Rsh = (oDct)- (3.14) From Equations 3.12-3.14, by eliminating t, we obtain for the transmittance, T = (1 + \2Rsh ZO OP) (3.15) CDC Note that Equation 3.15 is derived for generalized three-dimensional DC and optical conductivities. For a mono-layer of graphene, two-dimensional DC (static) and optical (dynamic) conductivities have quantized values of [70,71,87], 2D CYDC = 72 (3.16) (316 CHAPTER 3: GRAPHENE 2D (3.17) 4 Although graphene is a two-dimensional material, the ratio between the DC and optical conductivities can be said to be equivalent in both two- and three- dimensions as, C =D -2DC Q _4e 2 ez4 P crop 2 /1h /4h ~ 2.55 and thus Equation 3.18 can be generally applied to graphene thin film structures. (3.18) Note that Equation 3.18 sets an upper limit for a pristine graphene thin film structures and is valid only when the Fermi level sits at the Dirac point. Once the Fermi level is shifted from its neutral position, the two-dimensional DC conductivity of graphene will no longer be quantized and will be variable depending on the doping level [40]. The modified DC conductivity can be expressed as [87], g2D = npte (3.19) where n is the electron or hole carrier density, y is the carrier mobility, and e is the elementary charge of the electron. Therefore, the more realistic relation, where graphene films can be unintentionally doped in most situations, between the DC and the optical conductivity can be given as follows, DC GOP _ nye e 2 /4h 73 _ 4hnM e (3.20) CHAPTER 3: GRAPHENE (This page is intentionally left blank) 74 CHAPTER 4: Graphene Synthesis via Chemical Vapor Deposition (CVD) Chapter 4 4. Graphene Synthesis via Chemical Vapor Deposition (CVD) 4.1 Graphene Synthesis and Electrode Fabrication Continuous large area single- to few- layer graphene films have been successfully synthesized via the chemical vapor deposition (CVD) method on a commercially available copper foil substrate, under low pressure (LPCVD) or atmospheric pressure (APCVD) conditions [88-90]. After the growth, the graphene films can be transferred to various substrates, such as SiO 2 , glass, or polyethylene terephthalate (PET). LPCVD is used for the monolayer graphene growth and APCVD can be used for a few-layer graphene synthesis [90]. Figure 4.1(a) shows a schematic illustration of the CVD method used for the graphene synthesis and Figure 4.2 describes the detailed growth parameters for each condition. Graphene synthesis (LPCVD, APCVD). Copper foil (25 pm in thickness, ALFA AESAR) was used as a metal catalyst for both of these growth conditions. First, we describe the conditions used for LPCVD. The CVD chamber was evacuated to a base pressure of 30-50 mTorr. The system was then heated to a growth temperature of 1000 'C under hydrogen (H 2, 10 sccm) gas (~320 mTorr) and annealed for 30 minutes. Subsequently, methane (CH 4 , 20 sccm) gas was introduced (total pressure: -810 mTorr) and graphene growth was carried out for 30 minutes. The chamber was then cooled down at ~45 'C/min to room temperature. APCVD. The chamber was heated to 1000 'C under H 2 gas (170 secm) and annealed for 30 minutes. After annealing, H 2 was reduced to 30 seem and CH 4 (1 secm) and Ar (1000 scem) were additionally introduced followed 75 CHAPTER 4: Graphene Synthesis via Chemical Vapor Deposition (CVD) by 30 minutes of growth. After growth, the chamber was cooled at -100 'C/min to room temperature. A Synthesis CH 4 + H2 CH4 + H2 , - - - - ~ - - - - - - - - - - - - - - - - - -- 4, U I eoo"- Transfer Cu etchant -k I Figure 4.1 (a) Schematic diagram of the graphene synthesis and transfer process. The last part of the transfer procedure is repeated to produce multi-layer graphene stacks for LPCVD graphene. (b) Photographs of copper foils (before and after LPCVD graphene growth. The unit for the length scale is shown at the bottom and is in inch. 76 CHAPTER 4: Graphene Synthesis via Chemical Vapor Deposition (CVD) (a) 1000 0C H2 10 sccm + H2 10 sccMn CH 4 20 sccm a. E k 100*C 20 min 30 min 30 min 45*C/min Time (b) 1000 OC H2 30 sccm H2 170 sccm + CH 4 I sccm Ar 1000 sccm E 0 1000 C I: 20 min 30 min 30 min K 100*C/min Time Figure 4.2 Illustration of the graphene growth process at different stages: (a) LPCVD and (b) APCVD. Transfer of synthesized graphene to desired substrate and electrode fabrication. Transfer was carried out using poly(methyl methacrylate) (PMMA, 950 A9, Microchem). Graphene on one side of the foil was removed via reactive ion etching (RIE) with 02 gas (Plasma-Therm, 100 Watt at 7x 10- Torr). Cu was etched by a commercial etchant (CE100, Transene). Graphene films were then thoroughly rinsed with diluted hydrochloric 77 CHAPTER 4: Graphene Synthesis via Chemical Vapor Deposition (CVD) acid (10 %) and de-ionized (DI) water to remove residual iron ions from the Cu etchant. The PMMA layer was removed by annealing at 500 'C for 2 hours under H 2 (700 sccm) and Ar (400 sccm). Repeated transfers were performed for LPCVD conditions to yield multi-layer graphene films. The transferred graphene films were finally patterned via RIE (reactive ion etching) into the desired geometry to form transparent conducting electrodes ready for further processing. Figure 4.3 shows a photograph of the patterned multi-layer graphene films transferred onto a quartz substrate. (Note: More details about the graphene electrodes preparation processes applied in the organic photovoltaic solar cell fabrication will be further explained in Chapter 5). Figure 4.3 Photograph of a patterned graphene films on a quartz substrate (0.5 inch x 0.5 inch). 4.2 Growth Mechanism Graphene growth on several transition metal substrates, such as Co [91], Fe [92], Ru [53,93], Pt [91], Ni [88], or Ir [94,95], have already been demonstrated by means of thermal decomposition from hydrocarbons on catalytic metal surfaces or the surface segregation of carbon atoms from the bulk metal upon cooling from a carbon-metal solid solution. In general, these thermal processes are intimately related to the carbon solubility in the metal catalyst which dominates the growth mechanism for the graphene synthesis. The process of sp2 crystalline carbon formation from the transition metal 78 CHAPTER 4: Graphene Synthesis via Chemical Vapor Deposition (CVD) catalyst is affected by the degree of catalytic power of the metal, i.e., carbon affinity. A phase diagram provides useful information to predict the overall growth procedures. Figure 4.4 shows binary phase diagrams of some commonly used materials for the graphene synthesis. Atomic Percent Carbon a a 3W. 0 4* *.*u to IF V so OpO L 3W. L+graphite (Cgraphite) .U. e CL 13265 C I-a) 0.6 (M) (Ni)+graphite C Weight Percent Carbon Atomic Percent Carbon Ni b I 95 L+C(graphite) 1320 C 0 (ciCo) L. M 0.9 S09(aCo)+graphite 2.6 4-0 ) E 422C (ECo) Co Weight Percent Carbon 79 C CHAPTER 4: Graphene Synthesis via Chemical Vapor Deposition (CVD) Atomic Percent Carbon C10 C1S36 20 is 139 L L + C(graphite) ......- yFe + L 00 (yFe) yFe+ Fe3C (or graphite) 911 740 C 800 E 0.6 E aFe + Fe3C aFe - 450 0 Fe 1 2 3 4 5 6 Weight Percent Carbon C Atomic Percent Carbon a -0 s. so - L(C *a *. * . t uwce +a L+(CQ4. 1 + (C) e (C4 (Cu)+(C) E O1100 C b 1084.87 (C' we (Cu 4 8 12 16 C content (at. ppm) 0.0076 (CU)+ (C) 1 ............................ I.......,........... ..... Cu Weight Percent Carbon C Figure 4.4 Binary phase diagram of carbon and four different transition metals [96]: (a) C-Ni, (b) Co-C, (c) Fe-C, (d) Cu-C, showing different phase diagrams. 80 CHAPTER 4: Graphene Synthesis via Chemical Vapor Deposition (CVD) The principal mechanism of CVD graphene growth on the copper catalytic surface is also based on the thermal decomposition of carbon containing gas species (e.g. CH 4) at elevated temperatures (~1000 'C). The process is a surface limited reaction and not an out-diffusion process from the bulk. Carbon species diffuse onto the catalyst surface or slightly into the catalyst, which leads to the nucleation of graphene islands and forms a mono- or a few- layer graphene films depending on the synthesis conditions. Figure 4.5(a) describes the detailed growth processes under the typical gas flow consisting of hydrogen, argon, and methane:[90] Methane (1) diffuses into the catalytic copper surface, (2) adsorbs on the surface, and (3) decomposes into active carbon species. (4) Carbon atoms diffuse along the catalyst and form the graphene lattice. (5) Inactive species are desorbed from the surface and (6) diffuse/swept away by the bulk gas flow. H4 (a) Boundary layer CHA H* + C*1 (b) Figure 4.5 Surface 1000 c, CHSH2 (a) Growth kinetics in the CVD graphene synthesis under steady state gas flow of methane and hydrogen. (b) Simplified description of CVD synthesis of graphene on a copper substrate consisting three major steps: removing native oxide on the copper substrate by hydrogen, nucleation/growth of graphene flakes, and coalescence of the graphene flake into a continuous film [90,96]. 81 CHAPTER 4: Graphene Synthesis via Chemical Vapor Deposition (CVD) As shown in Figure 4.5(b), the graphene synthesis process can be qualitatively summarized: (1) A copper substrate coated with native oxide is reduced via annealing in a hydrogen atmosphere which also helps to increase the grain size of the copper and to remove most of the surface defects. (2) Nucleation of islands of graphene sites initiates the growth process (3) followed by an increase of the graphene domain size and the coalescence of domains into a continuous graphene sheet. Figure 4.6 shows the SEM images of graphene growth on a copper foil at different time steps, i.e., nucleation of graphene islands, growth of graphene grain domain size, and coalescence into the continuous film. In Figure 4.6(a), graphene flakes with different sizes are observed (white circle) and nucleation sites can be seen inside the flake, as shown as dark dots (blue circle) in Figure 4.6(a). Figure 4.6(b) shows an intermediate step where the size of graphene domain increased just before coalescing into a continuous film (orange circle). Finally, a continuous graphene film is formed as shown in Figure 4.6(c). 82 CHAPTER 4: Graphene Synthesis via Chemical Vapor Deposition (CVD) Figure 4.6 SEM images of graphene grown at different time stages: (a) Formation of nucleation sites, (b) Growth of graphene domains, (c) Coalescence into a continuous graphene sheet. In (c), dark areas indicated by the blue circle are region of a few layered graphene [96]. Compared with other existing techniques introduced earlier, the CVD method offers several advantages as follows: * Large scale fabrication: the area of the graphene film is determined by the initial copper substrate, which can be scaled up by using a large chamber size. 83 CHAPTER 4: Graphene Synthesis via Chemical Vapor Deposition (CVD) * High throughput: the process is very similar to thin film deposition techniques in semiconductor fabrication technology, where wafers can be stacked in a tube furnace allowing high throughput. * Low cost: copper catalytic substrates are commercially available at lower costs, much cheaper than single or multi crystalline substrates. * Ability to integrate with Si or other semiconductors: graphene films can be transferred to arbitrary substrates after synthesis. * Size and position control: patterned graphene structures can be directly synthesized and transferred to a target substrate using a pre-patterned substrate (Figure 4.7). NI Figure 4.7 Direct growth of graphene patterns from pre-patterned Ni structures (a) An optical image of a pre-patterned Ni film on Si/Si0 2 . CVD graphene is grown on the surface of the Ni pattern. (b) An optical image of graphene is transferred from the Ni surface in (a) to other substrate (Si/Si0 2 ) [88]. 4.3 Characterization of CVD Graphene Surprisingly, this one atomic thick layer of carbon atoms from the CVD process can be relatively easily characterized from optical microscopy, atomic force microscopy, or Raman spectroscopy. Figure 4.8 shows typical characteristics of monolayer graphene synthesized from LPCVD characterized by optical (Figure 4.8(a-b)), atomic force microscopy (Figure 4.8(c)), and Raman analysis (Figure 4.8(d)). The wrinkle formation, 84 CHAPTER 4: Graphene Synthesis via Chemical Vapor Deposition (CVD) a key signature of the presence of graphene, is clearly visible from the optical and AFM images. The Raman spectrum also indicates a high G' to G peak intensity ratio, which is a main characteristic feature used to distinguish monolayer graphene from other sp2 carbon materials. 20.9. (c) (d) (C) C 1200 .0 S0Raman 1400 1600 2400 2600 2800 3000 shift (cm*) Figure 4.8 Monolayer graphene characterized by (a-b) optical microscopy, graphene transferred on quartz (a) and Si/SiO 2 (b), (c) AFM, graphene transferred on Si/SiO 2, and (d) Raman spectroscopy, graphene transferred on Si/SiO 2. The transmittance of our CVD monolayer graphene was measured to be ~97 % (at 550 nm), which is in good agreement with the theoretically determined value of ~ 2.3 % opacity. Furthermore, the linear dependence of the transmittance on the number of graphene layers can also be observed in our graphene as shown in Figure 4.9(a). The conductivity of graphene films also improved as the number of graphene layers increased 85 CHAPTER 4: Graphene Synthesis via Chemical Vapor Deposition (CVD) (Figure 4.9(b)). This result confirms that our Cu grown graphene layers are indeed monolayers and the multiple transfer steps are successful to maintain the integrity of the graphene layers. As shown in Figure 4.9(c), further analysis of Raman spectroscopy on the graphene films with increasing number of layers from 1 - 3 also indicates a decrease in the G' to G peak intensity ratio, consistent with the Raman signature of multi-layer graphene films. 86 CHAPTER 4: Graphene Synthesis via Chemical Vapor Deposition (CVD) (a) 100 95 1C 90 (b) no. E Cas. 85 1- E Soo 100 9 m5 94 80 C 93 49,00 192 75170 IL 21L 31 a 400 1 ~ - Ml 2 a 500 M.mhrff 390 3 han~a Is gp 300 ,av y 700 600 . I 800 2 Number of layers Wavelength (nm) 1.2 2t 2.s 0.S 2L 3 Number of Lavers IL U) C 1000 1500 2000 2500 3000 3600 Raman shift (cm- ) Figure 4.9 (a) Transmittance of graphene films of 1 to 3 layers. As-grown LPCVDsynthesized graphene films are mostly single layered and each additional layer contributes approximately 2.3% opacity over the indicated range of wavelengths. The inset indicates the transmittance at 550 nm as a function of the number of the graphene layers. (b) The sheet resistance of graphene films transferred to quartz substrates as a function of the number of graphene layers. (c) Raman spectra of graphene films (1 - 3 Layers) on quartz substrates with a laser excitation at 532 nm (2.33 eV). 87 CHAPTER 4: Graphene Synthesis via Chemical Vapor Deposition (CVD) The surface morphologies of graphene films synthesized under atmospheric pressure conditions are presented in Figure 4.10 compared with those synthesized under low pressure conditions. From the figures one can observe that the graphene sheets grown under atmospheric pressure condition do not form uniform and continuous films of graphene layers. It appears that there are discontinuous regions of a few layers of graphene on top of the mono-layer background (optical images in Figures 4.10 (a-b) and atomic force microscopic (AFM) image in Figure 4.10 (d)). A mono-layer graphene sheet from LPCVD is also shown in Figures 4.10 (c, e) for comparison. Figure 4.10 Optical (a-c, scale bar: 10 pm) and AFM (d-e, scale bar: 1p m) images of graphene transferred on to Si/SiO 2 (300 nm) substrates synthesized under different pressure conditions: (a,b,d) APCVD. (c,e) LPCVD images are shown here as a reference for APCVD. (a-b) APCVD graphene consists of non-uniformly distributed multilayer regions on top of a mono-layer background, which can be clearly identified from the image taken at the edge of the graphene as shown in (b). (b) The dotted area indicates a region of the graphene film that is broken due to the transfer process, which further confirms the existence of the mono-layer background. (d) AFM image which illustrates the non-uniformity of APCVD graphene. The rms roughness of APCVD graphene is 1.66 nm compared to 1.25 nm for LPCVD graphene in (e). 88 CHAPTER 4: Graphene Synthesis via Chemical Vapor Deposition (CVD) The average sheet resistance and optical transmittance values of APCVD graphene were ~450 a/sq and ~92 % (at 550 nm), respectively. From the transmittance values, it can be inferred that a few layers of graphene can be grown under the AP conditions whereas under the LP conditions the graphene growth process is self-limited and yields mostly mono-layers. For comparison, graphene films consisting of three layers fabricated from LPCVD condition have average sheet resistance and transmittance values of~250 Q/sq and ~92 % (at 550 nm), respectively. 89 CHAPTER 4: Graphene Synthesis via Chemical Vapor Deposition (CVD) (This page is intentionally left blank) 90 CHAPTER 5: GRAPHENE ELECTRODE AND ORGANIC SOLAR CELL FABRICATION Chapter 5 5. Graphene Electrode and Organic Solar Cell Fabrication 5.1 Graphene as Transparent Conducting Window Electrodes in OPV Cells: Requirements Among the many interesting properties of graphene, such as superior electron and hole mobility or high current carrying capability, its uniformly high transparency in the visible and near infrared region, with good electrical conductivity and mechanical robustness [47], place graphene as a promising candidate for an alternative to indium tin oxide (ITO) [49] as a transparent conducting electrode (TCE) in organic photovoltaic solar cells. In reality, several criteria, such as electrical conductivity, optical transmittance, or work function (WF), need to be considered for the successful integration of graphene sheets as TCEs in OPV devices. Recent reports [77,97] have already demonstrated that graphene has high transmittance with moderate conductivity with a desirable mechanical flexibility and robustness. The most critical factor, however, is the energy level alignment between the work function of graphene and the highest occupied molecular orbital (HOMO) of the electron donor material. Bae et al. [77] reported a WF value of 4.27 eV (electron volts) for a monolayer of graphene synthesized from LPCVD, which is lower than that for ITO (~4.5 eV and the value can be increased up to ~5.0 eV after oxygen (02) plasma treatment or UV-ozone treatments) [98,99]. This low value of WF for graphene is obviously not a good match for most of the commonly used electron donor materials, such as (tetraphenyldibenzoperiflanthene (DBP), HOMO = 5.5 eV) [100], copper phthalocyanine (CuPc) (HOMO = 5.2 eV) [100] or poly(3-hexylthiophene) (P3HT) (HOMO = 5.2 eV) [101], which makes the energy level alignment between the 91 CHAPTER 5: GRAPHENE ELECTRODE AND ORGANIC SOLAR CELL FABRICATION graphene electrode and the HOMO of adjacent organic materials (5.2 - 5.5 eV) unfavorable, inducing in a large interfacial energy barrier at the interface. For ITO electrodes in OPV solar cells, a thin layer of conducting polymer, (PEDOT:PSS), is commonly poly(3,4-ethylenedioxythiophene):poly(styrenesulfonate) inserted before the deposition of the electron donor material in order to favor an ohmiccontact at the junction. The PEDOT:PSS hole transporting layer (HTL) with a WF of 5.2 eV not only facilitates the injection or extraction of holes but is also known to help planarize the rough surface of the underlying ITO, which often becomes a possible source of local shorting through the ultra-thin active layers. improves the overall device performance [102,103]. Thus the PEDOT:PSS layer Therefore smooth and complete coverage of the PEDOT:PSS layer on the underlying electrode surface plays a crucial role in the general OPV device performance. On the other hand, application of the PEDOT:PSS layer onto the graphene surface has been challenging due to the fact that the graphene surface is hydrophobic but the PEDOT:PSS is in an aqueous solution. The ITO surface, typically prepared via sputter deposition, is also hydrophobic but it is almost always pretreated with 02 plasma or UV-ozone, which renders the hydrophobic surface into a hydrophilic one by introducing hydroxyl (OH) and carbonyl (C=O) groups [104], thus enabling the conformal coverage of PEDOT:PSS. However, these active oxygen species from the plasma inevitably disrupt the aromatic rings of the graphene and reduce the conductivity significantly. In the case of a single- or a few- layer graphene electrode, the graphene film can completely lose the electrical conductivity after such treatments. Therefore, the key factor to the success of graphene electrode based OPV devices lies in the proper coating of the hole transporting buffer layer on the graphene surface for efficient charge extraction or injection at the interface between the graphene electrode and adjacent organic layer. This modification of graphene with an appropriate HTL (hole transporting layer) can possibly be realized in two ways: (1) Modifying the graphene surface (ideally non-destructively) to become suitable for the deposition of PEDOT:PSS HTL or (2) developing alternative HTL deposition methods without the need for the 92 CHAPTER 5: GRAPHENE ELECTRODE AND ORGANIC SOLAR CELL FABRICATION surface modification of graphene electrodes. The proposed approaches in this thesis include the following and will be investigated in detail in Chapter 6: " Surface modification of graphene films via chemical doping: In earlier work [97], it was demonstrated that doping graphene films with AuCl 3 can enhance the conductivity of the graphene films without sacrificing the transparency. " Application of alternative hole transporting materials to the conventional solution processed PEDOT:PSS layer, such as transition metal oxides (e.g. MoO 3 , molybdenum oxide): Unlike PEDOT:PSS, which is acidic (pH ~A) and corrosive to underlying electrodes, typical metal oxides HTLs have the following characteristics: - Wide band-gap p-type semiconductor behavior (3.5 - 4.0 eV) - Transparent in the visible spectral region - Appropriate energy level alignment to allow ohmic contact to the donor material - Ambient chemical stability and inertness with respect to the adjacent organic layers * 5.2 Non-destructive interface engineering of graphene films by conducting polymers Graphene Electrode Fabrication Large area (5" x 1") graphene sheets were synthesized via low-pressure chemical vapor deposition (LPCVD) using a copper foil (25 pm in thickness) as a metal catalyst. The copper foil (5" x 1") was placed in the CVD furnace and the chamber was evacuated to a base pressure of 30-50 mTorr. The system was then heated to a graphene growth temperature of 1000 'C under flowing hydrogen gas (total pressure: ~375 mTorr, 10 sccm (standard cubic centime per minute)) which removes oxide layers and other contaminants on the copper foil, and then the copper was annealed for 30 minutes to initiate grain growth. Subsequently, methane gas was introduced (total pressure: ~810 93 CHAPTER 5: GRAPHENE ELECTRODE AND ORGANIC SOLAR CELL FABRICATION mTorr, 20 secm) and the graphene growth was carried out for 30 minutes. After the completion of the graphene synthesis, the methane gas was terminated and the chamber was cooled down at 45'C /min under the hydrogen gas (10 sccm). As-grown graphene films require transfer to target substrates and patterning in order to be utilized as electrodes in OPV devices. The films were transferred using the poly(methyl methacrylate) (PMMA, PMMA A9, Microchem.) method described elsewhere.[88] Before the removal of the copper foil, the graphene on one side of the foil was removed via reactive ion etching (RIE) with oxygen gas (100 Watt at 7x 10- Torr), since the graphene growth occurs on both sides of the copper foil. If not removed, graphene pieces from the opposite side would be adsorbed underneath the floating graphene films during the copper etching, causing possible problems for OPV device fabrication afterwards, such as local shorting. The copper foil was etched using a commercially available etchant (CE-100, Transene) and the graphene films were rinsed with diluted hydrochloric acid (10 %) followed by de-ionized (DI) water to remove residual iron ions left over from the copper etchant, thus preventing unintentional doping of graphene. Finally, the PMMA layer was removed via several routes which will be discussed in the following and their effects on device performance were investigated. Three layers of graphene sheets were prepared as electrodes through layer-by-layer transfers to obtain a robust film with reasonable conductivity (-250 Q/sq) and transmittance (~92 % at 550 nm) for use as electrodes. Transferred graphene films require further patterning processes in specific geometries to be incorporated as electrodes in the device. Conventional photolithography can be used to pattern the film, but the photoresist residues are hard to remove [105] and can block the charge carriers to reach the graphene electrode since the resist is not conducting. In this thesis work, chromium (Cr) was used as a patterning mask. This method simplifies the fabrication procedure and provides a cleaner graphene surface. Cr patterns were directly deposited through a shadow mask onto the transferred graphene sheets by electron beam or thermal evaporation, and the graphene films were patterned through RIE afterwards. Subsequently, the Cr was removed using a commercially available 94 CHAPTER 5: GRAPHENE ELECTRODE AND ORGANIC SOLAR CELL FABRICATION etchant (CR-7, Cyantek). The sheet resistance of the graphene electrode was slightly increased from ~250 a/sq to ~350-500 Q/sq after patterning due to the processing but the transmittance remained the same. 5.3 Organic Solar Cell Fabrication and Measurements After patterning the graphene electrodes, archetypal standard bi-layer heterojunction small molecular OPV cells were fabricated via thermal evaporation with the following structure (see Figure 5.1): Anode (graphene or ITO)/hole transporting layer/donor (CuPc or DBP)/acceptor (C 60)/BCP (exciton blocking layer)/Cathode (Ag or Al). Donor nor A nodle c Figure 5.1 Schematic illustration of a typical bi-layer heterojunction small molecular organic solar cell. Organic layers [DBP (tetraphenyldibenzoperiflanthene, Luminescence Technology Corp., >99 %), CuPc (copper phthalocyanine, Acros Organics, ca. 95 % purity), C60 (fullerene, 95 CHAPTER 5: GRAPHENE ELECTRODE AND ORGANIC SOLAR CELL FABRICATION Sigma Aldrich, 99.9 %), BCP (bathocuprine, Luminescence Technology Corp., >99 %)] and top cathode [Al (Alfa Aesar, 3.175 mm slug, 99.999 %), Ag (Alfa Aesar, 1-3 mm shot, 99.999 %), Mg (Alfa Aesar, -4 mesh, 99.98 %)] were thermally evaporated through shadow masks at a base pressure of -1 x 10-6 Torr at rates of 1.0 A/s and 1.5 A/s, respectively. CuPc and C60 was purified once via thermal gradient sublimation before use. DBP, BCP, Al, and Ag were used as received. Pre-patterned ITO (Thin Film Devices, 20 f2/sq, 88 %T at 550 nm) substrates were cleaned by solvents followed by 30 seconds of 02 plasma (100 W, Plasma Preen, Inc.). Patterned graphene substrates were cleaned by annealing at 500 'C for 30 minutes under H2 (700 sccm) and Ar (400 sccm) to remove possible organic residues on the graphene surface. The device area defined by the opening of the shadow mask was 1.21 mm2 The current-voltage measurements were recorded by a Keithley 6487 picoammeter in a nitrogen atmosphere. 100 mW cm-2 illumination was provided by 150 W xenon arc-lamp (Newport 96000) filtered by an AM 1.5G filter. Shown in Figure 5.2 are photograph images of complete graphene- (Figure 5.2(a)) and ITO- (Figure 5.2(b)) based archetypal small molecular bi-layer heterojunction (CuPc/C 6o) devices, with a testing fixture (Figure 5.2(c)) and a description of the electrical connections (Figure 5.2(d)). 96 CHAPTER 5: GRAPHENE ELECTRODE AND ORGANIC SOLAR CELL FABRICATION (a) (b) (c) (d) Figure 5.2 Description of complete (a) graphene- or (b) ITO- based of OPV devices with CuPc/C 6 0 active layers and (c) the testing fixture. (d) Schematic illustration of the electrical connections: Graphene or ITO serves as the anode and top finger electrodes serve as the cathode. 5.4 Issues Related to the PMMA Transfer Method In general, PMMA on the graphene surface can be removed by common solvents such as acetone or chloroform. However, solvent cleaning usually tears the graphene film and introduces discontinuities in the film, thus reducing the conductivity of the graphene electrodes significantly. This method also suffers from PMMA residues left on the graphene surface. For instance, Figure 5.3 demonstrates the effect of PMMA residues on the graphene electrode applied in a typical CuPc/C6 0 OPV cell. Figure 5.3(a) shows an 97 CHAPTER 5: GRAPHENE ELECTRODE AND ORGANIC SOLAR CELL FABRICATION AFM image of a graphene surface on a quartz substrate after removing PMMA by immersion in acetone for 2 hours. A significant amount of PMMA residues on the graphene surface is still observed after such a treatment. Figure 5.3(b) demonstrates the current density vs. voltage (J-V) characteristics of an OPV cell fabricated from a graphene electrode with large amounts of PMMA residues and this figure illustrates quite a poor diode behavior. In fact, all the devices made from these graphene electrodes showed either similar behavior as shown in Figure 5.3(b) or no photo-response at all. There are several factors that could contribute to this degradation in behavior. Since PMMA is insulating, the residues will prevent the charge carriers from reaching the graphene electrode; large chunks of PMMA (>100 nm) cause a rough surface which could be detrimental for the OPV device operation (since each layer in the device is only tens of nm in thickness). Last but most importantly, PMMA residues appear to worsen the wetting of PEDOT:PSS on a graphene surface, even further, causing device failures. 98 CHAPTER 5: GRAPHENE ELECTRODE AND ORGANIC SOLAR CELL FABRICATION 21l nm 0.0 5.0 pm - (b) 9 o""' E m * 6 ' ~ ' Dark Light ." .* . EU mm 0 -3 bE o . -3 o 0 - "* C -9 . 4-. I ii I U' , * 0 U",e "g - C2 0 U U - -6 0 S 0 -. U.-12 -151 -0. 4 0 a -0.2 20 10 0 vc 20, 0.0 0.3 Voltage (V) 0.9 a 0.2 0.0 0.4 0.6 0.8 Voltage [V] Figure 5.3 Characteristics of a graphene OPV device with significant amounts of PMMA residues on the graphene surface. (a) AFM image of PMMA residues on the graphene electrode. PMMA was treated by acetone for 2 hours. (b) J-V response of a device under simulated AM 1.5G illumination at 100 mW/cm 2, made from graphene/PEDOT:PSS/CuPc(40nm)/C 6 0 (40nm)/BCP(1 0nm)/Al(1 00nm), graphene electrodes covered with large amount PMMA residues, showing poor diode characteristics. The inset describes the ideal behavior of the solar cell operation as a reference. 99 CHAPTER 5: GRAPHENE ELECTRODE AND ORGANIC SOLAR CELL FABRICATION Therefore, we considered several approaches to remove PMMA on graphene surface: (1) Immersing in acetone for 24 hours, (2) Remove most of the PMMA with acetone vapor to minimize the tearing of the graphene by direct immersion in acetone solution, then soaking in acetone for 24 hours, (3) Acetone vapor, brief acetone dipping for 2 minutes, followed by 3 hours of annealing, (4) Annealing for 3 hours at 500 'C under the protecting gas mixtures of hydrogen (700 sccm) and argon (400 sccm). AFM images show that the residues are almost removed for soaking in acetone for 24 hours (Figure 5.4(a) and (b)). Additional annealing after acetone treatment greatly improved in removing PMMA residues (Figure 5.4(c)). Annealing alone for 3 hours at 500 'C under hydrogen and argon was also enough to remove most of the PMMA from the graphene surface while minimizing tearing (Figure 5.4(d)). 100 CHAPTER 5: GRAPHENE ELECTRODE AND ORGANIC SOLAR CELL FABRICATION 20.0 nm 20.0 nm 0.0 5.0 pm 20.0 nm 0.0 5.0 pm 20.0 nm 0.0 5.0 pm 0.0 5.0 pm Figure 5.4 Graphene surface imaged by AFM after removal of PMMA via different routes. (a) Method (1), Immersing in acetone for 24 hours; (b) Method (2), First treated by acetone vapor followed by 24 hours immersing in acetone; (c) Method (3), Acetone vapor, 2 min acetone immersion, and 3 hours of annealing; (d) Method (4), 3 hours of annealing. Method (3) provides the cleanest graphene surface. Raman spectroscopy was also performed to investigate the defect density and the quality of graphene films. The collected signal showed pronounced D peaks (1350 cm) for acetone assisted cleaned samples (method (1), (2)) as illustrated in Figure 5.5. Since method (1) and (2) takes 24 hours, not very suitable for fabricating a large number of 101 CHAPTER 5: GRAPHENE ELECTRODE AND ORGANIC SOLAR CELL FABRICATION devices and minimal differences in the device performances from methods (3) and (4) were observed, in this thesis work we have mostly used method (4). Representative device characteristics from methods (3) and (4) are described in Figure 5.6. More details about these devices will be explained in Chapter 6. -Anneal VaporAcetoneAnneal A VaporAcetone 12 DO 1400 2400 1600 2600 2800 3000 Raman shift (cm') Figure 5.5 Raman spectra of graphene on 300 nm SiO 2 substrates where PMMA was removed via various methods as described in Figure 5.4. Raman spectra were obtained with a laser excitation wavelength of 532 nm (2.33 eV). 102 CHAPTER 5: GRAPHENE ELECTRODE AND ORGANIC SOLAR CELL FABRICATION * G(Cr, 15nm)/MoO3 (20nm)/Method-4 dark O G(Cr, 15nm)/MoO3 (20nm)/Method-4 light 2 * " o C4E 0 G(Cr, 15nm)/MoO3 (2Onm)/Method-3 dark G(Cr, 15nm)/MoO3 (20nm)/Method-3 light 0 -2 a 10 -4 -6 -L_0/ U -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 Voltage [V] Figure 5.6 Current density vs. voltage characteristics of graphene OPV devices with the MoO 3 hole transporting layer fabricated via cleaning methods (3) and (4) under simulated AM 1.5G illumination at 100 mW/cm 2. Graphene electrodes are patterned with 15 nm thick Cr metal masks and the thickness of MoO 3 layer is 20 nm. PCEs (power conversion efficiency) from methods (3) and (4) are 0.69 ± 0.02 % and 0.71 ± 0.01 %, respectively. 103 CHAPTER 5: GRAPHENE ELECTRODE AND ORGANIC SOLAR CELL FABRICATION (This page intentionally left blank) 104 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS Chapter 6 6. CVD Graphene based Organic Solar Cells With current understanding and knowledge on the synthesis of graphene and its desirable features suitable for transparent conducting electrodes applications, in this chapter we investigate on applying graphene as a transparent conducting window electrode material for organic photovoltaic solar cells applications, ultimately on paper thin flexible substrates. As a starting point, we will focus on the organic solar cell structure consisting of a standard bi-layer heterojunction configuration using small molecules as the active layer materials. Ultimately, other configurations of OPV cells utilizing bulk- heterojunction polymers, quantum dots, or organic-inorganic hybrid structures with graphene electrodes will be investigated. 6.1 Doped Graphene Electrodes for Organic Solar Cells In this section, graphene films grown by chemical vapor deposition (CVD) with controlled number of layers are demonstrated as transparent window electrodes in small molecules (CuPc/C 60 ) organic photovoltaic (OPV) cells. It is found that for devices with pristine graphene electrodes, the power conversion efficiency (PCE) is comparable to their counterparts with ITO electrodes. Nevertheless, the chances for failure in OPV devices with pristine graphene electrodes are much higher than the ones with ITO electrodes, due to the surface wetting challenges between the conventional PEDOT:PSS hole transporting layer and the graphene electrodes. Various alternative routes were investigated and it was found that AuCl 3 doping on a graphene surface can alter the graphene surface wetting properties such that a uniform coating of the hole transporting 105 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS layer can be achieved and device success rate can be increased. Furthermore, the doping both improves the conductivity and increases the work function of the graphene electrode, resulting in improvements in the overall performance of the OPV devices. 6.1.1 Aim of the Work In conventional OPV devices with ITO anodes, PEDOT:PSS has been shown as an effective hole transporting buffer layer facilitating efficient charge extractions or injections [102,103]. Although ITO-based OPV devices fabricated without PEDOT:PSS interlayer have been reported [106], we have found that graphene-based OPV devices do not demonstrate photovoltaic rectifying diode behavior without the PEDOT:PSS interfacial layer. Therefore, uniform coverage of PEDOT:PSS on a graphene surface plays a crucial role in the performance of graphene photovoltaic devices. However, the hydrophobic surface of pristine graphene makes uniform and conformal coating of PEDOT:PSS via aqueous solution very challenging. As a result, the success rate of devices with graphene electrodes is less than 5 % compared to devices with ITO electrodes in our studies, and obviously this rate will drop significantly when the device area increases. Thus the goal of this part of the thesis work aims at improving the interfacial wetting of the PEDOT layer with the graphene, while achieving as high a power conversion efficiency as possible. 6.1.2 Device Description Large area (~cm 2) graphene films, mostly single layered (~95 % of the growth area), were synthesized via low pressure chemical vapor deposition (LPCVD) using copper foil (25 pm in thickness) as a metal catalyst substrate. The as-grown graphene film on Cu was transferred to quartz substrates using poly(methyl methacrylate) (PMMA). Since the as grown graphene is mostly single layered, in principle, one can control the number of 106 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS graphene layers through multiple transfers which results in the overall improvements in the conductivity of the graphene electrode: for the OPV devices in this work, 3-layered graphene sheets were used as transparent electrodes. The sheet resistance (Rsh) of the graphene films on quartz substrates was varied from ca. 500 to 300 a/sq and the transmittance from ca. 97.1 to 91.2 % for 1 to 3 layered graphene sheets. The optical transparency of our graphene sheet agrees quite well with the measurements performed by Nair et al. [71], where each graphene layer was reported to have approximately 2.3 % opacity. This result confirms that our Cu grown graphene layers. are mostly monolayer and the multiple transfer steps are successful to maintain the integrity of each graphene layer. Figure 6.1 shows the schematic diagram of the OPV device structure considered in this section. Three generations of devices (~300 devices) were fabricated and investigated, with multiple batches of devices made in each generation for repeated testing to confirm the consistency of the behaviors. In every batch, the standard ITO/PEDOT:PSS/CuPc/C 6o/bathocuproine (BCP)/Ag (or Mg/Ag) device is used as a reference. These devices are made by spin-coating ~40 nm PEDOT:PSS onto commercially available ITO (~145 nm, Thin Film Devices) on borosilicate glass and annealed at 210 'C for 5 min in air. Afterwards, small molecule organic layers [CuPc (30 nm in generation I, 21nm in generations II and III)/C 60 (40nm in generations I, II, and III)/BCP (10 nm in generations I, 9.3 nm in generations II and III)] and cathode electrodes (Ag:Mg/Ag, 50 nm/50 nm in generation I, Ag (100 nm) only in generations II and III) are deposited by thermal evaporation at room temperature under high vacuum (~10-6 Torr). The active area as defined by the overlap of the anode and cathode is 2 0.0121cm . For the devices with graphene electrodes, first, 3-layer graphene films are transferred onto quartz substrates and lithography steps are used to achieve the same pattern as the ITO electrodes. Afterwards, either the same coatings of PEDOT:PSS are applied (for devices in generations II and III, and some in generation I), or an equivalent PEDOT layer is coated (for other devices in generation I). The following active layer and cathode deposition steps are carried out side by side with the ITO reference electrodes in 107 .... ... ... .... ............ . .. ... . .. ...... .......... ..................... ................ CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS each batch. The measurements were carried out under simulated AM 1.5G (100 mW/cm 2) solar illumination in a nitrogen-filled glove box. Schematic diagram of the organic solar cell structure: graphene or Figure 6.1 ITO/PEDOT:PSS (or other equivalent layer)/CuPc/C 60/BCP/Ag (or Mg/Ag). 6.1.3 Results and Discussions Figure 6.2 shows the current density-Voltage (J-V) characteristics of a representative set of devices, with the numerical values summarized in Table 6.1 (short-circuit current density (Jsc), open-circuit voltage (Voc), fill factor (FF), and power conversion efficiency (PCE)) from each type of device considered in the following. The conventional OPV devices typically incorporate PEDOT:PSS as the hole transporting layer. For devices fabricated with ITO electrodes, 02 plasma treatment is typically applied to improve the wettability of PEDOT:PSS on ITO. However, active oxygen species from such treatment disturb the aromatic rings of graphene and greatly reduces the conductivity. Therefore, in generation I, we tried different ways of 108 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS improving the wettability of PEDOT on graphene surface by modifying hydrophobic PEDOT solutions via two different routes: (1) Dissolving and mixing the PEDOT:PSS solution with organic solvents such as dimethyl-sulfoxide (DMSO), dimethyl-formamide (DMF), isopropyl alcohol, N-methylpryrrolidone (NMP) [107,108] or (2) Replacing the PSS derivative with poly(ethylene glycol) (PEG) doped with perchlorate (PC) or paratoluenesulfonate (PTS) [109]. Mixing with organic solvents is commonly used to improve the conductivity of the doped PEDOT thin film, but additionally it renders the PEDOT:PSS solution more hydrophobic, thus yielding better wetting on the graphene surface. Replacing PSS with PEG, which polymerizes the PEDOT with PEG, makes the PEDOT soluble in hydrophobic solvents such as nitromethane (CH 3NO 2 ), thus enabling wetting of PEDOT on hydrophobic graphene surfaces. PEDOT:PEG is also much less corrosive than PEDOT:PSS at the electrode interface since PSS itself is a very strong acid. 109 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS (a) 20 . IA. ITO dark IA. ITO illuminated IB.Pristine graphene (PEDOT:PSS)_dartk 16 18. E 12 - 8 E C) Pristine graphene (PEDOT:PSS)_jlluminated IC. Pristine graphene (PEDOT:PSS-DMSO)dark - IC. Pristine graphene (PEDOT:PSS-DMSO)illuminated ID. Pristine graphene (PEDOT:PEG(PC))_dark A ID. Pristine graphene (PEDOT:PEG(PC))jIluminated 4 0 - -4 _______________- -8 * - - -- U -12 -0.2 -0.4 0.2 0.0 0.4 0.6 0.8 Voltage (V) 10 (b) E IIA ITOdark " IIA. ITO illuminated 0 U 8 lIlA. ITO 6 C) 4 E 2 (02 plasma )_dark lIA ITO ( plasma )_iluminated 0 0 ___________________________________________ I -2 0 -4 0 -8 -10 p5ER ITO I -0.4 -0.2 0.0 0.2 Voltage (V) 110 0.4 0.6 0.8 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS 12 (C) IIB. Pristine graphen e dark U I11, Pristine graphen e_illuminated a 9 111B. Graphene (0, plasma)_dark I1B. Graphene M, plasma)_illuminated ItC. Doped graphene -dark IIC. Doped graphene _illuminated 111C. Doped Graphen e (02 plasma)_dark 111C. Doped Graphen e (0, plasma)_illuminated C4 E E 6 X * 3 + - / r 0 I- p00 -3 -6 M I - -f~UDM -9 .12 I Graphene I a a -4.4 0.6 0.4 0.2 0.0 -0.2 . 0 .8 Voltage (V) (d) 8 6 N E 0 l1WA10 (02Plasma)darkIlA, ITO (02 plasma) illuminated 1C Graphene (Doped)dark O IIC. Graphene (Doped)Lilluminated E 4 E 2 0 0*M C -2 C. -4 -6 TO vs. Graphene .__ __ -0.4 ___ -0.2 . _ . . 0.2 0.0 p 0.4 . p 0.6 0.8 Voltage (V) Figure 6.2 J-V characteristics of organic solar cells with different anodes under dark and simulated AMI.5G illumination at 100mW/cm 2 . Device performances of various types: (a) different PEDOT layer processing (b) ITO and (c) graphene electrodes are demonstrated. Shown in (d) is the comparison of performances of ITO with modified PEDOT:PSS by 02 plasma (IIA) and graphene doped with AuCl 3 (10 mM) (IIC). 111 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS In Table 6.1, under generation I, the results of two such examples are presented (device types IC and ID) together with the data from ITO references (IA), and graphene electrode devices with conventional PEDOT:PSS (IB) are also listed for comparison (their J-V characteristics are shown in Figure 6.2(a)). All of the newly synthesized PEDOT solutions yield uniform coating on hydrophobic graphene films. As a result, the improved interface (and increased conductivity in some cases) resulted in improvements in Jsc in some types of devices (see, device types IC and ID in Table 6.1 and Figure 6.2(a)). However, possible mismatch of the work function for the new compounds (or the same compound by different processing) caused an overall decrease in Voc and FF which offset the benefit of increase in Jsc. Furthermore, most of the modified PEDOTs with aforementioned methods were observed with pronounced leakage currents (i.e. reduced shunt resistance) (see Figure 6.2(a)), which is also detrimental to the overall device performance. 112 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS raton Device description VOC FF 3.10 0.48 0.43 0.63 2.82 0.46 0.44 0.57 5.31 0.22 0.28 0.33 ID. Pristine Graphene/PEDOT:PEG (PC) 40 4.00 .8 0.28 03 0.32 3 0.35 IIA. ITO/PEDOT:PSS 6.48 0.45 0.46 1.33 aphetEDOT:PSS 3.46 0.48 0.45 0.75 ap ene*/PEDOT:PSS 6.44 0.46 0.52 1.51 lIA. ITO/PEDOT:PSSt 6.88 0.46 0.56 1.77 pheeieDOT:PSSt 6.32 0.49 0.44 1.37 9.15 04 0.43 042 0.42 16 1.63 2 rain(mA/cm IA. ITO/PEDOT:PSS Graph PEDOT:PSS (VMF (%) IC. Pristine I Graphene/PEDOT:PSSDMSO" I III IIIC. Doped9.5 Graphene*/PEDOT:PSSt * PEODT:PSS-DMSO (3:1 vol. of PEDOT:PSS and DMSO) *Graphene chemically doped (p-type) with AuCl 3 in nitromethane (10mM) t PEDOT:PSS treated with 02 plasma prior to active layer deposition Note: Jsc (short circuit current), Voc (open circuit voltage), FF (fill factor), PCE (power conversion efficiency) Table 6.1 Summary of photovoltaic performance parameters of OPV devices from Figure 6.2. As a result, in generations II and III, only devices with conventional PEDOT:PSS layers via aqueous solution were investigated. As mentioned previously, although both the represented data from the ITO reference and pristine graphene electrodes are listed in Table 6.1, it is noted that the OPV devices with pristine graphene electrodes have a much higher failure rate than the ones with ITO electrodes. For devices in generations II and III, the thicknesses of the active layers and the cathode were further optimized, as a result, the PCE of the ITO references were noticeably enhanced themselves. A further 113 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS improvement of the ITO references (and also the OPVs with pristine graphene electrodes) in generation III as compared to the ones in generation II was achieved from an additional 02 plasma treatment (5 seconds) to the PEDOT:PSS which was carried out after spin coating the PEDOT:PSS layer (comparison between the J-V characteristics of these two types of devices is shown in Figure 6.2(b)). This treatment possibly modifies the surface electronic structure and thus improves the charge carrier collection at the anode [103]. When comparing the devices having pristine graphene electrodes with the ITO reference devices in generations II and III (in fact, also in generation I), it can be seen from Table 6.1 that the overall performance of the ones with graphene electrodes are still less than the ones with ITO electrodes. This is likely due to the still much higher Rsh of the graphene electrode (Rsh: ~300 fl/sq) which results in higher bulk series resistance for the solar cell than those from the ITO electrode (Rsh: ~20 Q/sq). Therefore, we tried to use chemical doping to reduce the graphene sheet resistance. It was found that the p-type chemical doping of the graphene films significantly improved the device performances. Figure 6.2(c) shows the J-V characteristics of devices with graphene electrodes in all three generations. Particularly, for devices in generation II, the doped graphene devices perform even better than the ITO references. Figure 6.2(d) compares the best performances of devices with ITO electrodes and graphene electrodes, and it can be seen that the graphene electrode based devices are quite comparable to the ones with ITO electrodes. The doping was carried out with AuCl 3 in nitromethane (10 mM) which is explained in detail in section 6.1.4 [97]. It was reported that AuCl 3 doping on graphene films resulted in up to 77 %decrease in Rsh with only 2 %decrease in transmittance. For devices in this work, AuCl 3 solution (filtered by 0.2 pm PTFE filters from VWR International) is spin-coated (at 2500 rpm for 60 seconds) on patterned graphene electrodes and annealed at 210 'C for 1 min to remove any residual nitromethane. The improved performances are likely due to the reduced sheet resistance of the graphene electrodes and the tuning of the work function as a result of the doping [110]. 114 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS More importantly, we have found that the doping solves the aforementioned wetting problems: After the doping, graphene electrodes become hydrophilic and a uniform coating of the PEDOT:PSS layer could always be achieved. This significantly enhances the device success rate for the OPV devices with doped graphene electrodes in generation 1I. At the same time, a consistent improvement in PCE is always observed for these doped graphene electrode based devices. However, for the devices in generation III, only ~10% of the time was there a clear improvement in the AuCl 3 doped graphene based devices as compared to pristine graphene based devices. One possible reason for the inconsistency could be the formation of various sizes of Au nano-particles on the graphene films due to doping (the particle size varies from 10 to 100 nm in diameter). Under exposure to 02 plasma, there could be slight etching of the PEDOT:PSS layer and the larger Au particles could be exposed and become a possible source of local shunts, which counteracts the benefits of doping effect on graphene films. Other doping agents such as SOC12 [111] or HNO 3 [77] have been used to p-type dope the graphene which do not cause particle deposition issues on the graphene surface; however, these agents have long term stability issues and thus were not considered in this work. For a future step, either different chemical dopants which do not have either of the problems (i.e., particle deposition or stability issue) or improved processing steps which can overcome these issues need to be identified. 6.1.4 Doping of Graphene with AuC13 AuCl 3 is a common doping agent often used in organic conducting polymers such as poly(3-alkylthiophene) or poly(3-octylthiophene) [112,113]. When AuCl 3 is dissolvent in solvents, different ionic conformations can occur depending on the amount of coordinating solvents. For instance, for a poor coordinating solvent such as nitromethane, the following reaction is expected to take place on the surface of graphene during the doping process [113]: 2Graphene + 2AuCI 3 -> 2Graphene+ + AuC12(Au') + AuC1-(Au"') 115 (6.1) CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS 3AuC1- <-+ 2Au 0 I +AuC14 + 2C- (6.2) In the presence of excess AuCl 3, Cl- coordinates as follows, Cl- + AuCl 3 -+ AuCl1 (6.3) For other solvents such as water, AuCl 3 forms a square planar complex of AuCl- and AuCl can be directly reduced as,[ 114] AuC- + 3e -+ Au 0 I +4CL- (6.4) p-type doping of graphene with AuCl 3 can occur either through reactions 6.1-6.3 or via 6.4 with reduction potentials of 1.4 V and 1.0 V, respectively. 6.1.5 Conclusions In summary, in this work we investigated LPCVD grown graphene electrodes as transparent electrodes in OPV cells. For devices with pristine graphene electrodes, the overall performance is comparable, but slightly inferior, to their counterparts with ITO electrodes, possibly due to the relatively higher sheet resistance which still needs to be improved. In addition, as the pristine graphene electrodes are hydrophobic and a uniform coating of the PEDOT:PSS layer cannot be achieved, the device success rates are considerably lower than the devices with ITO electrodes, and the device area has to be limited to guarantee certain number of successful devices. As a result, a significant amount of effort in this work was devoted to finding ways to improve the surface wetting between the graphene electrode and the PEDOT:PSS hole transporting layer. We found that AuCl 3 doping significantly improves the graphene OPV device performances, and also improves the surface wetting of the graphene electrodes with the PEDOT:PSS layer, thus improving the device success rate. However, the issues with cost and availability still are the limiting factors for this method. Gold is known to have less amount (ca. ~20 %), and thus higher cost (ca. ~10000 %) than Indium. 116 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS 6.2 Transition Metal Oxide Hole Transporting Layer In this section, organic photovoltaics (OPV) with graphene electrodes are constructed where the effect of graphene morphology, hole transporting layers (HTL), and counter electrodes are presented. Instead ethylenedioxythiophene):poly(styrenesulfonate) of the PEDOT:PSS conventional HTL, an poly(3,4alternative transition metal oxide HTL (molybdenum oxide (MoO 3)) is investigated to address the issue of surface immiscibility between graphene and PEDOT:PSS. The morphology of the graphene electrode and HTL wettability on the graphene surface are shown to play important roles in the successful integration of graphene films into the OPV devices. The effect of various cathodes on the device performance is also studied. These factors (i.e. suitable HTL, graphene surface morphology, and the choice of well matching counter electrodes) will provide a better understanding in utilizing graphene films as transparent conducting electrodes in future solar cell applications. 6.2.1 Aim of the Work As mentioned earlier, one of the major challenges in the integration of graphene in organic photovoltaics is the incompatibility between graphene and conventional PEDOT:PSS hole transport layer (HTL) which significantly increases device failure rate. When hydrophilic PEDOT:PSS is spin-coated onto graphene, it is difficult to achieve uniform and conformal coating due to the hydrophobic nature of the graphene surface. In Chapter 6.1, we showed that the compatibility of PEDOT:PSS with graphene films can be significantly improved by doping graphene films with AuCl 3 . Doping the graphene film also improves the overall power conversion efficiency (PCE) by increasing the conductivity of graphene. Nevertheless, the doping process introduces large Au particles (up to 100 nm in diameter) onto the graphene film, which can create shorting pathways through the device, thus possibly reducing the device yield. A planarizing buffer layer with good wettability is therefore necessary for more effective integration of graphene electrodes into organic photovoltaics. 117 ........ ... ....................... .............. . CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS Shrotriya et al. [115] and Godoy et al. [116] reported that in both polymer and small molecule based solar cells utilizing ITO anodes, a transition metal oxide buffer layer (Molybdenum trioxide, MoO 3) can be used in place of PEDOT:PSS as an efficient HTL. In this work, we propose MoO 3 as another type of HTL in OPV devices using graphene electrodes and confirm that MoO 3 can be successfully integrated between the graphene sheet and the subsequent organic layer as an alternative HTL. The use of a direct thermal evaporation of MoO 3 provides a HTL on the graphene surface with better wetting compared to hydrophilic PEDOT:PSS. We then further characterize the influence of the graphene surface morphology and counter electrodes (cathode) on the performance of graphene based organic solar cells. 6.2.2 Device Description A standard small molecular bi-layer heterojunction OPV structure, anode/PEDOT:PSS(40 nm)/CuPc(40 nm)/C 60(40 nm)/BCP(10 nm)/Ag(100 nm), was utilized in this work and was compared to devices made with MoO 3 replacing the PEDOT:PSS HTL. The device schematic and corresponding flat band energy levels are shown in Figure 6.3. HTL (b) CuPc C6o BCP -2.9 3.5 3.5 C3.7 Graphene 4.0-4.5 -"".. Ca 4.5 Mg ... . 4.2 Al 4.3 Ag - 5.3 Au .8 PED 5.2 6.2 Figure 6.3 6.5 (a) Device schematic of a standard solar cell considered in this work: Graphene/HTL/CuPc(40nm)/C 60(40nm)/BCP(1 Onm)/Ag(1 00nm); (b) Schematic diagram of flat band energy levels of each materials with various metal electrodes (in units of eV). 118 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS 6.2.3 Effect of Graphene Surface Morphology and the MoO3 HTL OPV devices employing a graphene anode and a transition metal oxide HTL were constructed, with a device structure of graphene/MoO 3/CuPc(40 nm)/C 6o(40 nm)/BCP(1 0 nm)/Ag(100 nm). The MoO 3 layer, with its wide band gap of 3 electron volts (eV), is relatively transparent, with a transmittance ranging from 85 - 95 % for 20 - 40 nm films at a wavelength of 550 nm (Figure 6.4(a)). AFM images of MoO 3 (10 nm) surfaces on bare quartz and graphene/quartz substrates are presented in Figures 6.4(b) and 6.4(c), illustrating the smooth surface profile of the MoO 3 layer (Figure 6.4(b), rms roughness of ~0.4 nm) and conformal coverage of thin layer of MoO 3 on the graphene surface (Figure 6.4(c)). Further SEM characterization of the MoO 3 layer deposited on graphene (Figures 6.4(d,e)) shows in fact there is still not 100 % wettability of MoO 3 on graphene (as indicated by the holes in Figures 6.4(d,e)). Nevertheless, this wetting behavior is much better than for the PEDOT case (further illustrated in Figure 6.5) and the device yield with MoO 3 HTL (43 ± 14.1 %) was significantly improved from the PEDOT HTL based devices, (3 ± 11.0 %). 119 .. .............. -- ........... -- - ___- . . ... ........ ........ CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS (a) IUU Onm 90 owffr 780 E 60 300 400 500 600 700 00 Wavelength [nm] Figure 6.4 (a) Ultraviolet-visible transmittance spectra of a MoO 3 layer on a quartz substrate with varying thicknesses (20, 30, 40 nm). The UV transmittance measured at 550 nm is 93.7 %, 89.6 %, and 86.2 % with increasing thicknesses. AFM images of (b) MoO 3 (10 nm) (c) MoO 3 (10 nm)/Graphene (3L) on quartz substrates. (d, e) Scanning electron micrographs (SEM) of MoO 3 (20 nm) on graphene/quartz. Bright (d and inset, with in-lens detector) and dark (e, without in-lens detector) spots indicate that the graphene openings not covered by the MoO 3 film. Scale bars are 1 pm for (b, c), 20 Am for (d, e) and 200 pm for the inset of (d). The height bars in (b, c) are 20 nm. (a) PEDOT:PSS MoO 3 (b) Graphene PEDOT:PSS )1% Graphene ;01 MoO3 Figure 6.5 Optical micrographs of graphene films on quartz substrates deposited with PEDOT:PSS (a) and MoO 3 (20nm, (b)). As shown in (a), spin-casting PEDOT:PSS results in only sporadic covering of PEDOT:PSS (dark dots) on the graphene surface. But the MoO 3 layer (b) is much more uniform. 120 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS During the MoO 3 evaporation, achieving conformal coverage on the wrinkles in the graphene sheet is critical since these "peaks" can serve as potential shunt pathways. For the stacked multi-layers of graphene, both the density of these wrinkles and the overall peak heights increase. Figure 6.6 shows a much denser concentration of wrinkles on three layered graphene films compared to a single layer graphene sheet. The surfaces of the stacked layers are usually much rougher than those for the single layer sheet with a root-mean-square (RMS) roughness more than three-fold higher: 2.0 nm for 3 layers and 0.6 nm for 1 layer. The corresponding cross-sectional profiles of dotted regions are also shown below the respective AFM images. 30.0 nm 05 .. 20.0 nm 0.0 s.0 pm 4011 2 30!! S 05 0. Figur 6. nm' FIiae ofth grpeemrpooy [fill ifeetubrofgahe Figure 6.6 AFM images of the graphene morphology of different number of graphene layers: (a) 3 layers and (b) 1 layer. Cross sectional profiles of dotted regions are shown below. The RMS surface roughness is 0.6 nm and 2.0 nm for the 1 layer and the 3 layers graphene films, respectively. 121 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS The surface homogeneity can be improved by controlling the thickness of the Cr mask during the graphene patterning process. Various thicknesses of Cr masks were considered in this work (10 - 40 nm). In our experiments, we have found when the Cr thickness is below 15 nm, it will not form a continuous film on graphene and thus cannot be used as a mask: As shown in Figure 6.7, cracks in the graphene sheets are observed after removing the Cr. Figure 6.7 Optical micrograph of graphene films on the quartz substrate after the film is patterned with 10 nm of Cr. The vertical scratch mark is intentionally created to distinguish the graphene region from the quartz substrate. Many cracked regions in the graphene film are observed which can lower the conductivity of graphene electrodes considerably. Since E-beam deposition of Cr is directional, it can be anticipated that the wrinkles that are higher than the Cr thickness should be exposed. Thicker Cr depositions will preserve the extruded wrinkles of graphene better than the thinner Cr, and these wrinkles could be the possible shunting pathways. During the RIE of the graphene patterns, the sharp peaks not fully covered by the thin Cr are expected to be smoothened out, thus reducing the 122 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS probability of shunting pathways. As shown in Figure 6.8(a), devices made from thinner (15 nm) Cr masks typically show better performance than devices with higher shunt resistance: 0.85 ± 0.02 % (PCE) for 15 nm Cr, and 0.50 ± 0.03 % (PCE) and 0.36 ± 0.02 % (PCE) for 25 nm and 40 nm Cr. This observation confirms the importance of surface morphology of graphene sheets in the OPV application. Figure 6.8(b) presents J-V characteristics of graphene devices (patterned with 15 rn Cr) fabricated with varying thicknesses of MoO 3 layers (20 - 40 nm) and Table 6.2 summarizes the key photovoltaic parameters (short circuit current density (Jsc), open circuit voltage (Voc), fill factor (FF), and power conversion efficiency (PCE)). As the MoO 3 thickness increases, the device becomes more resistive and the performance decreases, given the insulating nature of the MoO 3 layer. The PCEs are 0.71 0.01 % and 0.75 ± 0.01 % for 20 and 30 nM MoO 3 and the PCE then decreases to 0.31 0.01 % for 40 nm of MoO 3. The efficiency of the PEDOT:PSS reference cell was 0.85 0.03 %. MoO 3 based devices perform at -86 ± 2 % of the PEDOT:PSS reference cell. A similar relation was observed with devices using ITO electrodes, where these devices with MoO 3 HTLs performed ~88 ± 6 % of those with PEDOT:PSS HTLs. These numbers are not optimized but are rather intended to suggest a guideline for the graphene-metal oxide based OPV structure. A certain thickness of HTL is required to guarantee complete coverage on the graphene surface; however, too thick of a MoO 3 layer degrades the device performance as a result of the increased bulk series resistance. Since the overall device performance is not optimized (e.g. the thicknesses of the CuPc/C6 0 active layers), at this stage, no further work was carried out for the optimization of the MoO 3 HTL thickness. Nonetheless, the device yield with MoO 3 (~43 %) HTLs was much improved from the PEDOT:PSS (~3 %) case due to the improved wetting of HTL. 123 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS (a) 4 2 .5j E 0 E 0) -2 Oi' -4 C 1.. L. Cr15nm) Cr(15nm) 0 Cr(25nm) Cr(25nm) A Cr(40nm) Cr(40nm) a -6 0 -8 -10 - -0.4 a -0.2 0.0 PEDOT PEDOT PEDOT PEDOT PEDOT PEDOT dark light dark light dark light a I a 0.2 0.4 0.6 0.8 Voltage [V] (b) E 4 - A 2 0 0 -21I 0% 0 I. - I - I - I E) 0 U G/PEDOT G/MoO3(20 Im) G/MoO3(30 Im) G/MoO3(40r m) -4 -6 -8 -0.4 -0.2 I 0.2 0.0 0.4 0.6 0.8 Voltage [V] Figure 6.8 Current density vs. voltage characteristics of graphene and ITO OPV devices with PEDOT:PSS and MoO 3 hole transporting layers under simulated AM 1.5G illumination at 100 mW/cm 2 : (a) graphene electrodes patterned with 15, 25, and 40 nm thick Cr metal masks with PEDOT:PSS HTL. Using thinner Cr mask helps to reduce shunting pathways as confirmed by the increased shunt resistance and better diode behavior; (b) Devices using graphene electrode with varying MoO 3 HTL thicknesses (20 - 40 nm) under light. Graphene was patterned with thinner 15 nm Cr to ensure a smoother surface. 124 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS Anode HTL (mA/cm 2) C (V) FF Graphene PEDOT:PSS 4.43 0.53 0.36 0.85 Graphene MoO 3 (20nm) 3.93 0.49 0.37 0.71 Graphene MoO 3 (30nm) 3.41 0.54 0.40 0.75 Graphene MoO3 (4Onm) 2i47 0.50 0.25 0.31 % Table 6.2 Summary of photovoltaic parameters with PEDOT:PSS and MoO 3 HTLs for devices from Figure 6.8(b). 6.2.4 Effect of Oxygen Plasma on MoO3 HTL The primary role of a MoO 3 layer is to improve the charge transfer from the CuPc to the anode by reducing the effective energy barrier between the anode and the highest occupied molecular orbital (HOMO) of CuPc, and a work function of 5.2-5.3 eV has been commonly reported [117-119]. However, Kroger et al. [120] reported deep-lying electronic states of MoO 3 where the HOMO and LUMO (lowest unoccupied molecular orbital) levels are shifted down -4.3 eV with a work function of -6.7 eV. In contrast to previously published interpretations of MoO 3 induced enhancement of hole injection, they argue that hole injection occurs via electron extraction from the HOMO of the donor through the MoO 3 conduction band. Irfan et al. [121] also observed a similarly high work function of -6.75 eV for MoO 3 and attributed the previously reported lower values of-5.3 eV to air or oxygen exposure of the MoO 3 surface during the measurement. We thus investigate the effect of oxygen exposure on MoO 3 and the resulting device performance. After evaporation of MoO 3 onto the graphene surface, the graphene/MoO 3 electrode was removed from the evaporation chamber and briefly exposed to 02 plasma (8 seconds). Interestingly, the plasma treatment on MoO 3 did not result in any significant difference in the device performance for both ITO electrode and graphene electrode patterned with a thinner (15 nm) Cr mask, as shown in Figure 6.9(a) and 6.9(b). We suspect this could be due to the possible contamination of our evaporation chamber with 125 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS moisture, pinning the work function of MoO 3 with a lower value, although the exact reason is still under investigation. On the other hand, for graphene electrodes patterned with thicker (25 nm) Cr which may have more potential shorting pathways, the plasma treatment seemed to smoothen out rough regions of the graphene/MoO 3 surface. As a result, the overall PV performance was improved as shown in Figure 6.9(c). For 20 nm of MoO 3, the PCE increased from 0.42 ± 0.01 % to 1.03 ± 0.01 % on average (~145 % improvement). The key parameters of previous devices are summarized in Tables 6.3 and 6.4. 126 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS (a )2 I I G(Cr, 15nm)/MoO3 (20nm) dark E G(Cr, 15nm)/MoO3 (20nm) light * G(Cr, 15nm)/MoO3 (20nm)/O2 plasma dark o G(Cr, 15nm)/MoO3 (20nm)/02 plasma light =5 I -* n E WD * 0 on C]f 0 -4 -0.2 .4 0.2 0.0 0.4 0.6 0.8 Voltage [V] -- c (b) 4 o A E ITO/PEDOT o ITO/MoO3(20nm) ITO/MoO3(20nm)/02 plasma c 2 " 0- ' 0 Lo 0 .4 -0.2 0.2 0.0 Voltage [V] 127 0.4 0.6 0.8 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS - (C) 6 -U 0 E23 U U I G(Cr, 25nmy PEDOT G(Cr, 25nm)I AoO3(20nm) G(Cr, 25nm)lf Mo03(20nmYO2 plasma 0 E -. .~9 L.-12 -151 - -0. 4 -0.2 1 0.0 i 0.2 i 0.4 i 0.6 0.8 Voltage [V] Figure 6.9 Current density vs. voltage characteristics of graphene and ITO OPV devices with MoO 3 hole transporting layers with and without 02 plasma under simulated AM 1.5G illumination at 100 mW/cm 2 : (a) Graphene anode patterned by thinner (15 nm) Cr mask with MoO 3 (20 nm) HTL; (b) ITO anode with MoO 3 (20nm) along with PEDOT:PSS reference under light. For both (a) and (b), the effect of 02 plasma on the anode/MoO 3 surface appear to be minimal; (c) Graphene anode patterned by thicker (25 nm) Cr mask with 20 nm of MoO 3 HTL. It is observed that the use of 02 plasma on these rougher surfaces help to improve the device performance by planarizing the surface. 128 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS Anode HTL Jsc 2 (mA/cm') ITO ITO ITO Graphene-1 Graphene-2 Graphene-3 PEDOT 4.41 MoO 3 (2Onm) 3.92 4.24 3.93 MoO 3 (2Onm) + 02 plasma Mo 3 1 2Onm) FF 0.524 0.47 -0.481 0.491 0.48 0.58 0.56 0.37 MoO 3 (2Onm) 3.94 0.49 MoO 3 (2Onm) MoO 3 (20nm) 0.48 Graphene-1 MoO 3 (20nm)+ 02 plasma 3.55 3.99 4.10 4.34 Graphene-2 Graphene-3 Graphene-4 MoO 3 (20nm) + 02 plasma 4.44 MoO3 (20nm) + 02 plasma MoO 3 (20nm)± 02 plasma 4.45 4.31 Graphene-5 MoO 3 (20nm) + 02 plasma 4.11 Graphene-4 Graphene-5 Vocl (V) MoO 3 (2Onm) - 1 PCE (%) 1.11 1.07 1.14 0.71 0.37 0.37 0.71 0.47 0.46 0.42 0.37 0.70 0.36 0.69 0.43 0.79 0.41 0.42 0.41 0.43 0.42 0.78 0.79 0.41 0.73 0.411 0.42 0.70 0.62 Table 6.3 Summary of photovoltaic parameters (Figure 6.9) with 20nm MoO 3 HTL with or without air/oxygen exposure. Graphene electrodes are patterned with 15 nm Cr. (V) FF PEDOT JsC 2 (mA/cm2) 4.24 0.45 0.26 0.50 Graphene-1 MoO 3 (20nm) 4.50 0.36 0.26 0.42 Graphene-2 MoO 3 (20mn) 4.45 0.35 0.26 0.40 Graphene-3 MoO 3 (20nm) 4.47 0.35 0.26 0.40 Graphene-4 MoO 3 (20nm) 4.45 0.35 0.26 0.40 Graphene-5 Graphene-1 MoO 3 (20nm) 0.40 3 0.35 0.47 0.26 Mo 4.48 8.33 0.26 1.03 Graphene-2 Mo 3 8.10 -F 0.47 0.26 0.99 Graphene-3 MoO 3 (20nm)+0 plasma 8.28 0.48 0.26 1.02 Graphene-4 Graphene-5 MoO (2Onm) +02 plasma 8.24 0.48 0.26 1.02 MoO 3 (2Onm) +02 plasma 8.20 0.47 0.26 1.01 Anode HTL Graphene (2Onm)+ 02 plasma (20nm) + 02 plasma 2 j7 Table 6.4 Summary of photovoltaic parameters (Figure 6.9) of graphene devices having 20nm MoO 3 with or without air/oxygen exposure. Graphene electrodes are patterned with 25 nm Cr. 129 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS Figure 6.10(b) illustrates the external quantum efficiency (EQE) of graphene devices presented in Figure 6.9(c) with 20 nm of MoO 3, which describes the primary light absorption of CuPc and C60 from 400 - 800 nm. One can clearly see an improvement in EQE with 02 plasma treatment. 18 __ 15 U 12 U E* Cy 6 M 3 C - ITOiMoO 3 (20nm) * ITOIMoO 3 (20nm) + 0 plasma (8sec) - 0 500 400 600 700 800 Wavelength [nm] I * 15 r 0 12 - I - - (b) U 9 E 64 3 -+- x wi + 0 400 GrapheneMoO, (20nm) GrapheneMoO 3 (20nm) 500 + O plasma (sa*C) 600 700 Wavelength [nm] 130 800 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS Is 15 12 C 9 6 0U 3 -i-Graphene -- ITO 0 400 600 500 700 800 Wavelength [nm] Figure 6.10 The external quantum efficiency of devices with MoO 3 (20nm, with or without 02 plasma) HTL where light absorption primarily occurs in CuPc and C60 : (a) with ITO electrodes; (b) with graphene electrodes patterned with 25 nm Cr; (c) Comparison between ITO and graphene. For the relatively rough graphene electrode, air/oxygen exposure significantly increased the quantum efficiency. There was less improvement observed for thicker MoO 3 since a thicker MoO 3 layer results in fewer possible shunt paths and thus has less beneficial effect from the 02 plasma. For 30 nm of MoO 3, the PCE increased from 0.73 ± 0.05 % to 0.96 ± 0.03 % (~32 % improvement) and for 40 nm of MoO 3, the PCE increased from 0.73 ± 0.01 % to 0.97 ± 0.01 % (~33 % improvement) on average. The J-V characteristics are presented in Figure 6.11. 131 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS (a) * 6 I I I I I . *G(Cr, 25nmYMoO3(3Onm)a oG(Cr, 25nm)/MoO3(3Onm)102 plasma IN a I Di DU 0 Cr" E 3 E 0 IN 0 a 0 (I _UU 00 -9 0 GO 13 DU ID~ AU 12 -0 .4 -0.2 0.2 0.0 0.4 0.6 0.8 a 0.6 0.8 Voltage [V] (b) in., 6 - 3 E 0 G(Cr, 25nm)/MoO3(40nm) uG(Cr, 25nm)/MoO3(40nm)/02 plasma * -1 - E -31-6 C I~5 -9 - -0 -12 -15 I -0. 4 a -0.2 0 0.2 0.0 a 0.4 Voltage [V] Figure 6.11 Current density vs. voltage characteristics of graphene devices with MoO 3 hole transporting layers with and without 02 plasma under simulated AM 1.5G illumination at 100 mW/cm 2: (a) with 30 nm of MoO 3. (b) with 40 nm of MoO 3. Graphene anodes are patterned with thicker (25 nm) Cr mask. Although the FFs are slightly improved for both cases, the effect of oxygen is not as obvious as shown in the 20 nm of MoO 3 case from Figure 6.9(c). 132 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS 6.2.5 Effect of Cathode Work Function The built-in potential in OPV, one of the key parameters represented by the Voc, is mainly determined by the energy level difference between the HOMO of the donor and the LUMO of the acceptor. On the other hand, the choice of metal contact is very important for facilitating the extraction of charge carriers: Metals forming ohmic contact were shown to enhance the PCE over the Schottky type contact, because injection barriers at the Schottky contacts can lead to accumulation of charge carriers causing diffusion current which must be balanced by drift current at open circuit [122]. Previously, various studies were conducted on how metal cathodes affect the overall OPV device performance with ITO anodes [122-125]. Herein, we performed similar studies on how metal cathodes with varying work functions affect graphene anode based solar cells. The metals considered in this work are those with moderate work functions compared with graphene, i.e. magnesium (Mg, -3.7 eV), aluminum (Al, ~4.2 eV), and silver (Ag ~4.4 eV), and those with much lower and higher values of work functions than graphene, i.e. calcium (Ca, ~2.9 eV) and gold (Au, ~5.3 eV). Work function values are referred to the values found in the literature.[126] The work function of graphene generally ranges from 4.0 - 4.5 eV depending on the synthesis conditions. Solar cells with varying cathodes were fabricated and tested under the same conditions. Figure 6.12 presents J-V responses from various metal cathodes and Table 6.5 summarizes the key PV parameters. Similar to the previous works with ITO anodes, devices made from Ag, Al, and Mg top electrodes showed moderate behavior with similar performances: PCEs of 0.37 ± 0.01 %, 0.47 ± 0.01 %, and 0.56 ± 0.02 %, respectively. Au and Ca based devices, on the other hand, showed considerably reduced efficiency where PCEs of both devices are only 0.12 ± 0.03 %. The high work function of Au likely forms a Schottky type contact and limits charge extraction, in which case holes are injected back to the active layer which causes recombination of charge carriers, thus lowering the device performance [123,124,127]. Ca with its low work function should favor an ohmic contact to the adjacent organic layer. However, Ca cathode devices showed significantly reduced Jsc and FF, with PCE values of only 0.12 ± 0.03 %. 133 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS Eo et. al. argued that Ca with a work function lower than the LUMO of C60 forms a Schottky type contact as well as having a non-spontaneous electron extraction process which lowers the device performance [122,123]. Nonetheless, the exact reason for the observed behavior is still not clear at this point. These observations highlight the importance of appropriate choices of counter electrodes for successful integration of graphene electrodes in organic solar cells. (a )2 E 0 -2 44A CA4 * 0 *' * 0 .~AA .00 A -6 4 0 0 0 0 -8 0 .0. 4 - -0.2 0.0 Mg/Al Ca/Al Au a a 0.2 0.4 0.6 Voltage [V] (b) 6 ISC(M VOC(V PCE 0.6 I 0.5 4 E 2 E 0.4 00 C) _ 3 0.3 2 0.2 I ,AI. A . a 11 a Ca (2.9) Mg (3.7) Al (4.2) Ag (4.3) AU (5.3) (OV) Figure 6.12 (a) J-V characteristics of graphene devices with varying top metal electrodes under light: graphene/PEDOT:PSS/CuPc(40nm)/C 60(40nm)/BCP(1 Onm)/metals(1 00nm). (b) Jsc, Voc, and PCE of devices with different cathodes. Al was coated over Mg and Ca electrodes to prevent oxidization. 134 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS Anode HTL Cathode Work function (eV) (mA/cm) JCathode CM2 V FF PCE P Graphene PEDOT Ca/Al 2.9 0.99 0.47 0.25 0.12 Graphene PEDO Mg/Al 3.7 5.01 0 48 0.23 0.56 Graphene [piEDO[T Al 4.2 4.42 0.42 0.26 0.47 Graphene PEDOT Ag 4.3 3.85 0.39 0.25 0.37 Graphene PEDOT Au 5.3 2.64 0.18 0.26 0.12 Mg/Al 3.7 6.42 0.41 0.28 0.72 Graphene Table 6.5 3 (20nm) Summary of photovoltaic parameters of various metal cathodes devices in Figure 6.12. 6.2.6 Conclusions In summary, OPV devices using small molecules as active materials and a transition metal oxide (MoO 3) as HTL were fabricated with graphene anodes. By utilizing the thermally evaporated MoO 3 HTL, the wetting of HTL on a graphene surface can be improved compared to the conventional PEDOT:PSS HTL. The effects of the surface morphology of graphene, MoO 3 thickness, and air/oxygen exposure on a MoO 3/graphene surface on the OPV performance were investigated. Finally, we demonstrated how different metal cathodes with varying work functions affect the performance of solar cells constructed with graphene anodes. The intent of this work is to provide a better understanding of graphene-based organic solar cell fabrication, facilitating the progress towards the realization of graphene integration even beyond OPVs into areas such as OLEDs and flexible displays. 6.3 Vapor Printed PEDOT Hole Transporting Layer For the successful integration of graphene as a transparent conducting electrode in organic solar cells, proper energy level alignment at the interface between the graphene 135 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS and the adjacent organic layer is critical. The role of a hole transporting layer (HTL) thus becomes more significant due to the generally lower work function of graphene compared to ITO. A commonly used HTL material with ITO anodes is poly(3, 4- ethylenedioxythiophene) (PEDOT) with poly(styrenesulfone) (PSS) as the solid state dopant. However, graphene's hydrophobic surface renders uniform coverage of In the section, we PEDOT:PSS (aqueous solution) by spin-casting very challenging. introduce a novel, yet simple, vapor printing method for creating patterned HTL PEDOT layers directly onto the graphene surface. Vapor printing represents the implementation of shadow masking in combination with oxidative chemical vapor deposition (oCVD). The oCVD method was developed for the formation of blanket (i.e., unpatterened) layers of pure PEDOT (i.e. no PSS) with a systematically variable work function. In the unmasked regions, vapor printing produces complete, uniform, smooth layers, of pure PEDOT over graphene. Graphene electrodes were synthesized under low pressure chemical vapor deposition (LPCVD) using a copper catalyst. The use of another electron donor material tetraphenyldibenzoperiflanthene (DBP) instead of copper phthalocyanine (CuPc) in the organic solar cells also improved the power conversion efficiency. With the vapor printed HTL, the devices using graphene electrodes yields comparable performances to the ITO reference devices (qp, LPCVD= 3.01 %, and ip, ITO = 3.20 %). 6.3.1 Aim of the Work In this section, we introduce a novel, yet simple, HTL fabricated material by vapor printing of PEDOT [128] directly onto the unmodified graphene surface. In a single step, vapor printing combines: (i) the synthesis of conducting polymer chains from vapor- phase (3, 4 ethylenedioxythiophene) (EDOT) monomer, (ii) thin film formation of the PEDOT HTL, and (iii) patterning by in-situ shadow masking. Vapor printing is derived from the oxidative chemical vapor deposition (oCVD) (steps (i) and (ii) only). The oCVD blanket (i.e., unpatterned) PEDOT layers readily integrate with a wide range of substrates, because it is a dry process and the substrate temperature is mild (-120 C), the difficulties with film dewetting and substrate degradation by solvents or high temperatures can be completely avoided.[128] In addition, the WF of the oCVD PEDOT 136 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS layer can be tuned by controlling the doping level of Cl- ions [129]. In this work, we have found that the oCVD process is also compatible with graphene substrates. The oCVD polymer layer is formed by directly exposing the substrate to the vaporized monomer EDOT and an oxidizing agent (in this case, FeCl 3) under controlled reactor conditions. The relatively mild deposition conditions (low temperature 120 'C, moderate pressure ~10 mTorr, and no use of solvents) allow for PEDOT to be deposited without damaging or delaminating the graphene electrode. Furthermore, as vapor printing is generally substrate independent, requiring no substrate-specific optimization of the oCVD process or substrate pretreatment, thus it allows the direct printing of the PEDOT onto our graphene substrates. The graphene based solar cells fabricated with vapor printed PEDOT HTLs in this work achieve -94 % of the performance of their ITO counterparts without any additional treatment to the graphene sheets such as chemical doping. 6.3.2 Device Description Graphene films were synthesized under low pressure chemical vapor deposition (LPCVD) on Cu foils (25 pm in thickness and 99.8 % purity). This yields monolayer graphene on the Cu and afterwards, graphene anodes were prepared through layer-bylayer transfers by stacking 3 mono-layers of graphene sheets. The average sheet resistance (Rsh) and transmittance values of the graphene electrodes are ~300 Q/sq and -92 % (at 550 nm). After patterning the graphene electrodes, PEDOT (PEDOT:PSS or vapor printed PEDOT) and organic layers were subsequently deposited followed by the top capping electrode via thermal evaporation. illustrated in Figure 6.13(a). graphene)/HTL The PEDOT deposition process is The final device structure was: anode (ITO or (PEDOT:PSS or vapor printed PEDOT)/DBP/C6 0 (fullerene)/BCP (bathocuproine)/Al (aluminum). The complete solar cell structure is schematically shown in Figure 6.13(b). 137 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS (a) PEDOT:PSS spin-casting Vapor print PEDOT via c teper*twmv-croed stWg (20-1 Shadow-masked graphene Growing PEDOT HTL oxidant vapor (b) (~~ + - ~ qNq I.- eq.I I \ - - - - - I I I I I V 0, \ 0 Figure 6.13 Schematics outlining the fabrication process of PEDOT HTLs, and OPV devices. (a) PEDOT:PSS spin-coating vs. vapor printing of PEDOT deposition. The spin-casting layer covers the graphene and the surrounding quartz substrate while the vapor printed patterns align to produce PEDOT only on the graphene electrodes. (b) Graphene/ITO anode OPV structure: Graphene(or ITO)/PEDOT/DBP/C 60/BCP/Al. 138 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS 6.3.3 Vapor Printed PEDOT The oCVD reactor configuration elsewhere.[130] and general process procedure are described The oCVD PEDOT HTLs were all deposited under the same reaction conditions. The reactor pressure was held at ~10 mTorr and the substrate temperature was maintained at 120 'C. The monomer 3,4-ethylenedioxythiphene (Sigma Aldrich, 97 %), EDOT, was used as purchased. The EDOT was heated to 140 'C and introduced into the reactor at a flow rate of ~5 sccm. Iron (III) chloride (Sigma Aldrich, 99.99 %) was evaporated from a heated crucible between 130-160 'C. Different thicknesses were achieved by varying the time of reaction (1, 2, 4, and 8 minutes respectively to get HTL thicknesses of 2, 7, 15, and 40 nm). (PEDOT:PSS (Clevios TM P VP Al 4083) was filtered (0.45 pm), spin-coated at 4000 rpm for 60 seconds, and annealed at 210 'C for 5 minutes in air). The PEDOT was patterned using a pre-cut metal shadow mask with the same dimensions as the graphene electrode. The mask was visually aligned such that the PEDOT was deposited directly on top of the graphene. Shown in Figure 6.14(a) are sheet resistance (Rsh) and transmittance values of vapor printed PEDOT with varying thicknesses. The thinner PEDOT layers (2, 7, and 15 nm) have higher transmittance values (generally >90 %) that are presumably better for HTL layers, because losses in optical absorption through the transparent electrode could contribute to a decrease in device performance. On the other hand, the sheet resistance decreases with increasing thickness, with an abrupt change from 2 nm to the thicker layers (7, 15, and 40 nm) due to the amount of charge pathways increases (as shown in Figure 6.14(b) and Table 6.6). Having a lower sheet resistance HTL results in better charge transfer to the graphene electrode; however, the thicker the HTL, the further the charge must travel through the layer to reach the graphene electrode and the greater the transmission losses due to absorption. Nevertheless, in our experiments, we have found that the thicknesses in the range of 7-40 nm all give reasonable performances. 139 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS 100 . - - - 90 - 2 nm, Rsh - 100,000 ohms/sq -- 7 nm, Rsh - 3000 ohms/sq 15 rm, Rsh - 900 ohms/sq nm, Rsh - 500 ohms/sq 80 '--40 E 70 60 300 400 600 600 700 800 Wavelength [nmJ (b)oo 100000 * 80000 - '90 85 0 8_-J 80 60000 40000 E C 76 20000 70 0 65 2 nm 7 nm 15 nm 40 nm PEDOT:PSS oCVD PEDOT Thickness (not to scale) Figure 6.14 (a) Transmittance data for the oCVD PEDOT HTL layers, measured using ultraviolet-visible spectroscopy (UV-Vis) over wavelengths from 350-800 nm. The oCVD PEDOT layers decrease in transmittance and sheet resistance with increasing thickness. The three thinnest PEDOT layers (2, 7, and 15 nm) have high transmittance values (>90% over a majority of the range), which are preferred for HTL layers. (b) Sheet resistance values for each thickness and the transmittance at 550 nm. The oCVD PEDOT sheet resistance was measured using a 4-point probe (taking the average of 10 measurements). With increasing oCVD PEDOT thickness, there are more pathways for charge transfer, so the sheet resistance (Rsh) decreases. Rsh decreases dramatically from the thinnest (2 nm) to thicker PEDOT layers (7, 15, and 40 nm). Transmittance and RsA values of PEDOT:PSS are also shown for comparison. 140 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS oCVD PEODT (2 nm) oCVD PEODT (7 nm) Transmittance at 550 nm (%T) 98.0 [964 Rh(/q 100,000 3,000 oCVD PEODT (15 umn) 94.1 900 oCVD PEODT (40 nm) 67.4 500 PEDOT:PSS (40nm) 97.1 2,000 Table 6.6 Optical transmittance (%T) and sheet resistance (Rh) of oCVD PEDOT with varying thicknesses described in Figure 6.14. PEDOT:PSS values are also shown for comparison. 6.3.4 Vapor Printed PEDOT vs PEDOT:PSS Figure 6.15 shows the optical and SEM images of the graphene/quartz substrate after spin-coating PEDOT:PSS (left images, Figures 6.15(a-c)) in contrast with the vapor printed PEDOT (15 nm) (right images, Figures 6.15(d-f)). The optical image in Figure 6.15(a) shows that most of the spin-casted PEDOT:PSS dewetts over the graphene electrode as well as the adjacent bare quartz (Figures 6.15(b, c) shows more details). The dewetting of the PEDOT:PSS is clearly observed over the entire substrate, which signifies that we do not have good coverage of the graphene surface. In contrast, vapor printing provides a well-defined PEDOT region (light blue) (Figure 6.15(d)). Furthermore, the scanning electron microscopy (SEM) image in higher magnification shows that the coating of vapor printed PEDOT on graphene is uniform in finer detail (Figure 6.15(f)). This confirms the understanding that since oCVD is a dry process, the dewetting problem is avoided and the PEDOT can form a uniform film on the graphene. 141 .......... . . ... .... ........ ........... ........ ...... ........ CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS U)JQudz I GraphWW Rim of tic PEDOT PSS Qua Gaphen DOT DfpVW of PEDOTPSS oCVD of VD PEDOT PEDOT PSS / .. (a) Graphene (d) 0.5 I Graphene PE Figure 6.15 Comparing HTL coverage on quartz/graphene substrate. (a-c) Spin-coated PEDOT:PSS on quartz/graphene substrate, (d-f) oCVD PEDOT coating on quartz/graphene substrate: (a) schematic illustration of PEDOT:PSS spun-coated on a quartz substrate with graphene electrode. Most of the PEDOT:PSS layer is dewetted from the substrate with dark macroscopic defects visible to the naked eye. (b-c) Optical micrographs (at different magnifications) of the spin-cast PEDOT:PSS on the graphene surface illustrating the poor wettability of PEDOT:PSS on the graphene. In contrast, (d) is the schematic illustration of CVD PEDOT coated via vapor deposition on quartz/graphene substrate, where a uniform coating and patterning via shadow masking is achieved. The left side in (d) has the oCVD PEDOT coating whereas the right side is (e) Optical micrograph and (f) SEM image of oCVD PEDOT on shadow masked. graphene showing uniform coverage. 142 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS 6.3.5 Work Function Furthermore, the WF of vapor printed PEDOT on the graphene (3-layers) was evaluated by the Kelvin probe method. The measured value averaged over several regions was ~5.1 eV which was similar to the commonly reported WF value of PEDOT:PSS (~5.2 eV). This observation indicates that the injection/extraction of holes from the HOMO of electron donor now becomes energetically favorable compared to the interface of graphene only. 6.3.6 Results and Discussions Small molecule organic solar cells with graphene anodes were fabricated with device structures mentioned earlier. Figure 6.16(a) displays the current density-voltage (J-V) measurements of devices with various configurations using 3-layer graphene anodes: graphene with spin-coated PEDOT:PSS and vapor printed PEDOT (15 nm) HTLs along with ITO reference. Due to the poor wetting of PEDOT:PSS, graphene device with PEDOT:PSS typically shows the leaky behavior (not a diode behavior but rather like a linear resistor), and with poor photo-response (much smaller V, and J). On the other hand, the J-V responses from the devices having vapor printed PEDOT HTLs (with different thicknesses shown in Figures 6.16(b) and (c) for graphene and ITO electrodes) shows good diode behavior, and performance (Jsc(short-circuit current density) = 5.69 0.17 mA cm 2 , Voc (open-circuit voltage) = 0.88 + 0.01 V, FF (fill factor) = 0.60 ± 0.01, and q. (power conversion efficiency, PCE) = 3.01 ± 0.05%) is comparable to the ITO reference device with PEDOT:PSS (Jsc=5.14 ± 0.12 mA cm 2 , Voc = 0.92 ± 0.01 V, FF 0.68 ± 0.01, and , = 3.20 ± 0.05%). 143 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS (a) 4" -ITO/PEDOT:PSS --Graphene(LP)/PEDOT:PSS --- Graphe e LP)/vapor printed PEDOT 2 E < E 0 -2 C -4 -6 -c.5 0.0 0.5 1.( Voltage [V] (b) E 2 Vapor printed PEDOT 0 - m Vapor printed PEDOT - 75nm -Vapor printed PEDOT - 4nm E -2 g0 C a -4 -6 -0.5 Graphene anodes 0.0 0.5 Voltage [V] 144 1.0 . ... .......... ..... .. .... .. ...... ................... CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS (C) ) E . . 2 2 -PEDOT:PSS 0 -- Vapor printed PEDOT - 7nm Vapor printed PEDOT - 15nm -Vapor printed PEDOT - 40nm 4- ITO anodes -6 -0.5 0.6 0.0 1.0 Voltage [V] (d) 3.5 6.5[eV] GrapheneIPEDOTIDBP/Cr ,/ Graphene/PEDOTCuPc/C,./ IAI /Al Figure 6.16 J-V characteristics of representative graphene (3-layer, LPCVD)/ITO OPV devices (Graphene, ITO/PEDOT:PSS (40nm), vapor printed PEDOT (7-40nm)/DBP, 25nm/C 60 , 40nm/BCP, 7.5nm/Al, 100nm) under simulated AM 1.5G illumination at 100 mW/cm 2 . (a) Graphene devices with PEDOT:PSS and vapor printed PEDOT (15nm) HTL, compared with ITO/PEDOT:PSS reference device. (b) Graphene anode based cells with varying thicknesses of vapor printed PEDOT (7, 15, 40nm). (c) ITO anode devices with varying vapor printed PEDOT thicknesses (7, 15, 40nm) and a PEDOT:PSS reference. (d) Flat-band energy level diagram of the complete OPV device structure comparing DBP and CuPc electron donors. 145 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS With more devices fabricated using the vapor printed PEDOT on graphene electrodes, we have found that about 30% among the working devices show the (close to) ideal diode JV responses similar to the one presented in Figures 6.16(a-c), and with the other working devices non-ideal behaviors have been observed (both for graphene and ITO electrodes), as shown in Figure 6.17. Nevertheless, these non-ideal performances are still considerably better than the performance obtained with devices having spin-coated PEDOT:PSS on graphene. This ideal vs. non-ideal behavior does not appear to be related to the thickness of the vapor printed PEDOT on graphene, as can be seen in both Figures 6.16(b-c) and Figure 6.17, for thicknesses between 7-40 nm, the devices have displayed both behaviors. And even within the devices that showed ideal behaviors, there appear to be no direct correlation between the PCE value and the PEDOT thickness. At present the non-ideal behavior appears to be a process-related issue and should be further investigated in the future. 146 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS (a) 2 - '" E E 0 - GrapheneNapor printed PEDOT - 7nm GrapheneNapor printed PEDOT - 15nm GrapheneNapor printed PEDOT - 40nm 4-1,o -2 -4 -6 -8 m 0.5 0.0 -0.5 1.0 Voltage [V] (b) 2 -ITONapor ITO/Vapor E E" i 0 -ITO/Vapor printed PEDOT - 7nm printed PEDOT - 15nm printed PEDOT - 40nm -2 -4 -6 -8 -0.5 0.0 0.5 I .0 Voltage [V] Figure 6.17 J-V characteristics of representative graphene (3-layer, LPCVD)/ITO solar cell devices with non-ideal diode characteristics under simulated AM 1.5G illumination at 100 mW/cm 2 : Graphene, ITO/vapor printed PEDOT (7-40nm)/DBP, 25nm/C 60, 40nm/BCP, 7.5nm/Al, 100nm). (a) Graphene anode solar cells. (b) ITO anode solar cells. Corresponding key photovoltaic parameters of each device are summarized in Table 6.7. 147 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS Anode HTL Jsc (mA/cm 2) (V) ITO oCVD PEDOT - 7nm 4.93 ITO oCVD PEDOT5lnm ITO FF PCE (%) 0.73 0.32 1.16 5.10 0.74 0.35 1.34 oCVD PEDOT - 40nm 5.28 0.78 0.31 1.29 Graphene oCVD PEDOT - 7nm 3.95 0.83 0.29 0.97 Graphene oCVD PEDOT- l5nm 5.30 0.75 0.30 1.20 Graphene oCVD PEDOT - 40nm 5.38 0.78 0.29 1.22 Table 6.7 Summary of key photovoltaic parameters of graphene/ITO devices from Figure 6.17. Apart from the successful interface engineering by the vapor printed oCVD PEDOT on graphene, another reason for the enhanced performance can be attributed to the increased Voc observed in the cells compared to devices fabricated using CuPc as electron donor materials, one of the most widely used electron donor materials in the small molecules based OPV structure. As shown from the energy level diagram in Figure 6.16(d), the maximum Voc achievable from DBP/C6 0 pair is ~1.0 V, ~0.3 V higher than CuPc/C 6 0 . Therefore, improvements in Voc mostly originate from the deep-lying HOMO level of DBP compared to that of CuPc. 6.3.7 Organic Solar cells from APCVD Graphene Graphene synthesized from the Cu catalyst under LP condition succeeded in producing high quality mono-layer sheets, which could not be achieved using nickel (Ni) catalyst [88,131]. However, due to the self-limiting process of the LPCVD, achieving multi- layers of graphene from the Cu under LPCVD has been difficult [131]. In practice, a mono-layer of graphene sheet can be hardly used as an electrode due to defects induced from the processing issues such as transfer or patterning, as well as the generally lower conductivity compared to stacked multi-layers layers [132]). (Rsh decreases as a function of additional On the other hand, transferring multiple steps to obtain multi-layers add 148 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS complexity and cost to the fabrication process. Therefore, in this work, we also carried out few-layer graphene synthesis under atmospheric CVD (APCVD) condition using Cu foils, where only one-step transfer is needed for graphene electrodes. Figures 6.18(a, c) show the optical and atomic force microscope (AFM) images of the APCVD grown graphene, which has a non-uniform film thickness, and the optical and AFM images of LPCVD graphene are also presented in Figures 6.18(b, d) for comparison. The APCVD graphene layers have average sheet resistance and transmittance values of ~450 a/sq and ~92 % (at 550 nm), respectively. Even though the thicknesses of the APCVD grown graphene layers are non-uniform, we have verified that the vapor printing of oCVD PEDOT onto these graphene layers is as successful as the LPCVD graphene layers. This is consistent with the general observation that oCVD PEDOT deposition is substrate independent and it coats uniformly on the substrate. 149 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS I Figure 6.18 Optical (a-b, scale bar: 10 pm) and AFM (c-d, scale bar: 1p m, height bar: 20nm) images of graphene transferred on the Si0 2 (300 nm) substrates synthesized under different pressure conditions: (a, c) APCVD. (b, d) LPCVD images are shown for comparison. (a) APCVD graphene consists of non-uniformly distributed multilayer regions on top of the mono-layer background. (c) AFM image further illustrates the nonuniformity of APCVD graphene. The rms roughness of APCVD graphene is 1.66 nm compared to 1.17 nm for LPCVD graphene in (d). Solar cells fabricated with graphene anodes prepared under APCVD conditions give performance close to the devices made via LPCVD conditions ( LPCVD 3.01 %). p. APCVD = 2.49 %and ,, Figure 6.19(a) shows the J-V characteristics of graphene (APCVD) 150 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS -2 device with vapor printed PEDOT (Jsc=5.89 ± 0.03 mA cm2, Voc = 0.89 ± 0.03 V, FF 0.48 ± 0.01, and q. = 2.49 0.06 %) and ITO reference device with PEDOT:PSS (Jsc =5.14 ± 0.12 mA cm-2, Voc 0.92 ± 0.01 V, FF = 0.68 ± 0.01, and q, = 3.20 ± 0.05 %). Figure 6.19(b) compares the best device performance using graphene electrodes from different synthesis conditions (LPCVD vs. APCVD), illustrating performances (APCVD graphene performs ~83 % of LPCVD graphene). 151 comparable CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS (a) A"' 2 - E ITOIPEDOT:PSS Graphen (AP)Ivapor printed PEDOT Opp 0 E C -2 s i i/ -4 -6 0.0 -0.5 0.5 1.0 Voltage [V] (b) Eq 2 E E -- LP Graphene AP Grap ene 0 -2 -4 -6 Vapor printed PEDOT _____ -0.5 _____ ___ 0.5 0.0 1.0 Voltage [V] Figure 6.19 (a) J-V characteristics of representative graphene (APCVD) OPV devices (Graphene/vapor printed PEDOT, 15nm/DBP, 25nm/C 60, 40nm/BCP, 7.5nm/Al, 1 00nm) along with ITO/PEDOT:PSS reference device under simulated AM 1.5G illumination at 100 mW/cm 2. (b) Comparison of graphene-based device performances, where graphene electrodes are prepared under either LPCVD or APCVD conditions. 152 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS 6.3.8 Conclusions In summary, we introduced a novel, yet simple, method for vapor printing PEDOT onto the graphene surface, which yields well defined patterns using in situ shadow masking. The oCVD process, which is the foundation for vapor printing, results in smooth, complete coverage of PEDOT on the graphene electrode. In contrast, spin-casting PEDOT:PSS from an aqueous solution does not coat the graphene surface as well due to graphene's low surface free energy [133]. The oCVD process works well on both LPCVD grown graphene (with uniform thicknesses) and APCVD grown graphene (with non-uniform thicknesses). Furthermore, the use of small molecular electron donor material DBP combined with the vapor printed PEDOT HTL, yields more efficient graphene based devices with performances comparable to those of ITO reference devices. The results here represent a further step forward in the investigation of using graphene as an alternative transparent conducting electrode material for the replacement of ITO and open up opportunities in other applications as well, such as organic light-emitting diodes (OLEDs). 6.4 Non-destructive Interface Engineering of Graphene for Universal Applications in OPV and OLEDs Among the interesting properties of graphene, its uniformly high transparency in the visible and near infrared regions along with moderate conductivity and mechanical robustness, find a particular interest in optoelectronic applications. In this section, graphene is proposed as a promising candidate for an alternative material to indium tin oxide (ITO) as a transparent conducting electrode (TCE) in organic electronics such as organic photovoltaic (OPV) solar cells or organic light-emitting diodes (OLEDs). 153 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS 6.4.1 Aim of the Work Recently, many works [132,134-136] have shown the possible application of graphene as a TCE material in OPV as an ITO alternative. However, the performance is still less efficient than ITO based OPV devices and the graphene films usually require rigorous treatments (e.g. chemical doping) in order to function as appropriate electrodes, which frequently suffer from long-term stability issues [77,97,111]. The device yield is also quite low in general and often is not discussed very well as well as the reproducibility of the devices, which are in fact quite important factors for considerations in the real-life applications. All of these works have demonstrated that graphene electrodes satisfy most of the criteria required for TCE, such as electrical conductivity or optical transmittance except one critically important factor, the work function (WF). As reported by Bae et al. [77], the WF of monolayer graphene layer synthesized from low-pressure chemical vapor deposition (LPCVD) is ~4.27 eV (electron volts), which is much lower than ITO (~4.5 eV and is usually modified to ~5.0 eV after oxygen (02) plasma or UV-ozone treatments) [98,99]. This low value of WF makes the energy level alignment between the graphene electrode and the highest occupied molecular orbital (HOMO) of most adjacent organic materials (5.2 - 5.5 eV) [100,101] unfavorable, resulting in a large interfacial energy barrier. In this work, we explain current limitations of graphene in the TCE application and propose a novel yet simple solution via surface-engineering of the graphene using commercially available polymers. We demonstrate that this method can be universally applicable in both OPV and OLED devices with high yields and reproducibility. 6.4.2 Non-destructive Surface Modification of Graphene In order to reconcile the aforementioned challenge, the WF of graphene was modified via surface-engineering achieved by the insertion of a thin polymeric layer between the "unmodified" graphene surface and the PEDOT:PSS HTL (Bi-layer HTL structure, BLHTL). We use commercially available poly(3,4-ethylenedioxythiophene)-block- poly(ethylene glycol) (PEDOT:PEG) doped with perchlorate (PC) in nitromethane (CH 3NO 2 ) (1 wt. %, Sigma-Aldrich) as the buffer layer which is spin-casted. 154 The CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS molecular structures of each PEDOT are illustrated in Figure 6.20(a), and Figure 6.20(b) shows the transmittance of the spin-cast films on quartz substrates. The sheet resistance of PEDOT:PEG (PC) and PEDOT:PSS typically ranged 2.08 ± 0.3 Me/sq and 100 ± 5 kQ/sq at thicknesses of ~45 and ~20 nm, respectively. Replacing PSS with PEG, which polymerizes the PEDOT with PEG, makes the PEDOT soluble in highly polar solvents and enables efficient wetting of PEDOT on the hydrophobic graphene surfaces, as well as other organic substrates, which has been difficult to achieve from PEDOT:PSS. 155 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS (a) - , .\0 U) 0 CL0 O 0 SO3H SOiH 0 0 SO 3H O s S03- O ssS 0 0 S03H O 0 SO3H SO3H ~0 S03 0 0 SO 3H SO3H 0.. 0 O - \R'C 1010 \_/O O\_} 100 90 0* 80 70 -- PEDOT:PEG (PC) PEDOT:PSS PEDOT:PEG (PC)/PEDOT:PSS 60 50 300 - - 400 500 I . 600 I 700 800 Wavelength [nm] Figure 6.20 (a) Molecular structures of PEDOT:PEG (PC) and PEDOT:PSS. (b) UVVis (ultraviolet-visible spectroscopy) spectra of each polymeric layer and the bi-layer spun-cast on quartz substrates. Both layers were spun at 4000 rpm resulting in the film thicknesses of -45 and 25 nm, respectively. Varying the spin speed (1500 - 5000 rpm) of PEDOT:PEG (PC) did not have a significant variation on the transmittance (%T, < 5 0.3 MQ/sq), %) and film thicknesses (< 10 %) as well as the sheet resistance (2.08 possibly due to the originally large particle sizes of the polymer (~ 600 nm in suspension). Interestingly, there was a slight increase in the %T particularly in the lower wavelength region, presumably caused by the interference at the interface. The inset shows macroscopic images of each polymer spun-cast on quartz substrates. 156 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS The surface morphologies of both materials are also compared. The scanning electron microscopy (SEM) images clearly illustrate that PEDOT:PEG (PC) after spun-cast completely covers the graphene surface whereas a large irregular area of uncoated graphene region is produced by the dewetting of PEDOT:PSS (Figures 6.21(b-c)). The dewetted polymer sometimes forms a ring around the edge of the defect, which undergoes a fingering instability to produce droplets (Figure 6.22). Although the PEDOT:PEG (PC) surface is relatively rough compared to that of the PEDOT:PSS, it can be smoothed out by the subsequent coating of the PEDOT:PSS (Figure 6.21(d)), which can be further confirmed by the atomic force microscopy (AFM) images as shown in Figures 6.21 (e-g). Further SEM images describing the surface morphologies of the spun-cast films on the bare quartz substrates are shown in Figure 6.23. 157 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS 400.0 rin 0.0 nm S5 pm 4 3 1 158 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS 20.0 nm 0.0 nrn 5pm 4 33 22 1m 4 Figure 6.21 3 Characterizations of the bi-layer hole transporting layer structure. (a-d) SEM micrographs of graphene (a), graphene/PEDOT:PEG(PC) (b), graphene/PEDOT:PSS (c), graphene/PEDOT:PEG(PC)/PEDOT:PSS (d) all on quartz substrates. Dewetted PEDOT:PSS on the graphene surface is clearly observed whereas the graphene surface is completely coated by the PEDOT:PEG (PC). micrographs illustrating the surface morphologies of polymeric (e-g) AFM layers on quartz substrates: (e) PEDOT:PEG(PC), (f) PEDOT:PSS, (g) PEDOT:PEG(PC)/PEDOT:PSS. The rms (root mean square) roughness of PEDT:PEG (PC) layer has been reduced from 36.4 nm to 27.0 nm upon the deposition of PEDOT:PSS (rms: 0.83 nm). Both SEM and AFM images confirms that the rather rough surface of the PEDOT:PEG (PC) is smoothened by the PEDOT:PSS layer. 159 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS Figure 6.22 SEM image of PEDOT:PSS on graphene/quartz substrates around the edge of the defect region. 160 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS Figure 6.23 SEM micrographs of PEDOT:PEG(PC) PEDOT:PEG(PC)/PEDOT:PSS (c) on quartz substrates. 161 (a), PEDOT:PSS (b), CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS In addition, the PEDOT:PEG (PC) buffer layer is much less corrosive to the underlying electrode than PEDOT:PSS: Jong et al. [137] reported that PEDOT:PSS, with a pH value between 1-2 due to PSS being a strong acid, etches indium out of the ITO anode which leads to the diffusion of indium atoms into the active layers that can possibly degrade the lifetime and long term stability of the OPV devices. 6.4.3 Results and Discussions: OPV PEDOT:PEG (PC) alone, however, is not a suitable HTL for the graphene and adjacent donor material due to its relatively lower WF (~4.33 eV), which can be confirmed by the non-rectifying diode characteristics as shown in Figure 6.24(a). A graphene device with a PEDOT:PSS HTL alone (i.e. graphene with poor wetting of a proper HTL) resulting in an almost resistor-like device behavior is also shown to emphasize the important role of the energy level matching buffer layer in the TCE application of graphene (Figure 6.24(b)). 162 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS (a) 10 5 Eq E - Graphene PEDOT:PEG (PC)_dark - Graphene PEDOT: PEG (PC)Jlluminated 0 -5 E 4 -10 -15 -20 _0.4 -0.2 0.0 0.2 0.4 0.2 0.4 Voltage [V] (b) 10 - E E Graphene PEDOT:PSSdark Graphene PEDOT:PSS illuminated 5 0 7 -51 4P. -10 CU -16 -20 a 4 Ui 0.0 -0.2 Voltage [V] Figure 6.24 J-V measurements of devices with graphene electrodes using either one of the polymer layers. (a) Due to the inappropriate alignment of the energy level from the PEDOT:PEG (PC), non-rectifying diode behavior is observed, even though complete coverage of the buffer layer was achieved. (b) PEDOT:PSS HTL with its matching WF at the interface still resulted in an almost resistor-like device character from the inadequate wetting on the graphene surface. 163 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS Instead, PEDOT:PEG (PC) is used as an intermediate layer for the subsequent PEDOT:PSS layer which significantly improved the wetting efficiency of PEDOT:PSS on the graphene/PEDOT:PEG (PC) composite: PEDOT:PSS layer with its high WF aligns the energy level with the HOMO of the donor material and PEDOT:PEG (PC) layer enables the wetting between the graphene and the PEDOT:PSS HTL. This novel BLHTL (bi-layer hole transporting layer) structure from simple spin-casting in air allows for PEDOT to be deposited on the graphene surface without introducing any defect sites (i.e., 02 plasma or UV-ozone treatment) or requiring any unstable modifications on the graphene electrode [132,138] and thus can be applicable to even a single layer graphene sheet if necessary. Our BLHTL process is substrate independent (i.e., no substrate- specific optimization or substrate pretreatment was required to achieve PEDOT coverage onto the graphene) and readily integrates with a wide range of substrates by eliminating the wettability issues that arise when using PEDOT:PSS in aqueous solution. Consequently, this process allows for complete coverage of the graphene electrode with a smooth coverage of HTL without the film dewetting that is critical for the successful graphene-based OPV operation. By incorporating the BLHTL structure, small molecules OPV solar cells based on the "unmodified" (i.e. no doping or chemical treatment) graphene electrode were demonstrated with comparable performances to the ITO reference devices. The device structure as well as the cross-sectional transmission electron microscopy (TEM) image of the complete device is shown in Figure 6.25(a-b) along with the corresponding flat-band energy levels of each material (Figure 6.25(c)). The energy dispersive X-ray spectroscopy (EDS) shown on the right panel in Figure 6.25(b) indicates each materials in the device corresponding to the cross-section TEM image. Graphene films were synthesized under the LPCVD conditions and the graphene electrodes were prepared through the layer-by-layer transfer [134] by stacking 3 mono-layers of graphene sheets: The average sheet resistance (Rsh) and transmittance values were ~300 f2/sq and -92 % (at 550 nm), respectively. The final device structure was anode graphene (or ITO)/BLHTL (PEDOT:PEG (PC)-PEDOT:PSS)/tetraphenyldibenzoperiflanthene (DBP, 164 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS 25nm)/fullerene (C60 , 40nm)/bathocuproine (BCP, 8.5 nm)/aluminum (Al, 100nm) (see Figure 6.25(a)). 165 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS (a) I / I 4 / -I \~ f~\/ -o C60 -c - . DBP PEDOTPSS PEDOT:PEG(PC) 7 I ,--i - Graphene __ _ Quartz (SiO I 166 _ _ ......... ...... .... ........ .... ......... ...... ... ......... CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS (c) I I 3.5 eV I A~ C) 2 E - 2 e-I - 0- Al I A -1.2 eV "WI (d) 4.2 OV I 1.2eV - - 6.5 eV ITO-dark ITOilluminated Graphene-dark Grapheneilluminated 4 - -2 0 -4 -6 -0.5 0.5 0.0 Voltage [V] 167 1.0 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS (e) BLHTL dark BLHTL illuminated -ITO PEDOT:PSS dark -ITO~PEDO:PSS illuminated - 2 E 0 E C~ -ITO -ITO -2 -- - - - ---- -___ -4 -6 -0.5 0.0 0.5 1.0 Voltage [V] Figure 6.25 Descriptions of typical organic solar cells with a graphene electrode and the device performance. (a) Schematic diagram outlining the graphene anode OPV architecture: graphene/PEDOT:PEG(PC), 45nm/PEDOT:PSS, 25nm/DBP, 25nm/C 60 , 40nm/BCP, 8.5nm/Al, 100nm. (b) Cross-sectional TEM image (left panel) of the complete device described in (a) with EDS elemental line scan overlaid onto a schematic of the device architecture (right panel). Solid lines indicate interfaces identified using TEM, dashed lines indicate expected location of interfaces not quite resolvable by TEM or EDS but partly resolvable from the difference in the color contrast. (c) Flat-band energy level diagram of the complete structure. (d) Current density vs. voltage (J-V) characteristics of a representative graphene OPV device compared with an ITO reference cell under simulated AM 1.5G illumination at 100 mW/cm 2 illustrating comparable performances. (e) J-V characteristics of ITO-based devices with different configurations of HTL, i.e. PEDOT:PSS alone and PEDOT:PEG (PC)/PEDOT:PSS, showing similar performances. Figure 6.25(d) illustrates the current density-voltage (J-V) measurements of the solar cells described above. reference. ITO-based device with the same device structure is shown as a J-V response using the BLHTL structure indicates well-rectifying diode characteristics which perform similarly to the ITO control device: q, (power conversion efficiency, PCE) of graphene and ITO based devices are 2.91 ± 0.13 % and 3.19 ± 0.10 168 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS %, respectively. We also observed that the type of HTL, i.e., single or bi-layer HTL, has a negligible effect on ITO based devices (Figure 6.25(e)), and thus have adopted the BLHTL structure for ITO electrodes as well for the sake of consistency. The key photovoltaic parameters of the representative devices (Jsc (short-circuit current density), Voc (open-circuit voltage), and FF (fill factor)) are summarized in Table 6.8. Compared to devices fabricated with either PEDOT:PEG (PC) or PEDOT:PSS HTL alone (see Figure 6.24), significant improvements in the photo-response from the BLHTL configuration is observed as a result of the reduced interfacial energy barrier between the graphene electrode and the HOMO of the adjacent donor layer enabled by the smooth and complete coverage of the HTL. Anode HTL ITO PEDOT:PSS Graphene mACm 2) PEDTPSC PEDOT:EGPC Soi Voc (V) PCE (% FF 5.47 0.18 0.91 5.43 0.22 0.95 ± 0.04 0.01 10.65 0.01 3.19 0.10 0.57 ± 0.03 2.91 0.13 Table 6.8 Key photovoltaic parameters for devices described in Figure 6.25(d). Our finding suggests that the proposed BLHTL structure serves as a desirable hole transporting buffer layer for the integration with graphene-based OPV devices with several important features. First and most importantly, the performance of OPV solar cells utilizing graphene TCEs approaches close to that of the ITO reference devices (- 91 % of the PCE of the ITO control device). Remarkably, this enhanced performance was achieved on "unmodified" or "undoped" graphene electrodes consisting of only three stacked layers suggesting more room for improvements. The non-destructive nature as well as the substrate-independent process provides more favorable conditions to the graphene without causing any damage or delamination. Other important features regarding the proposed BLHTL configuration lie in the reproducibility and the device yield which are essential factors for the real-life applications but have hardly been discussed in graphene based OPVs works so far. Here, we emphasize that our structure is highly reproducible with desirable yields (- 85 ± 10 %, out of approximately 200 cells), 169 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS which is almost as good as ITO reference devices (~ 92 ± 6 %). The long-term stability and lifetime testing will also be conducted in the future. Using small molecules as the active materials, the efficiency of the graphene device in this work is higher than the most bulk-heterojunction (BHJ), P3HT:PCBM (poly(3-hexylthiophene):[6,6]-phenyl-C 6 1- butyric acid methyl ester), based devices [101,136]. Therefore, with the further optimization and utilization of BHJ structures, highly efficient graphene OPV devices can be expected. 6.4.4 Results and Discussions: OLED We also demonstrated the universal applicability of the BLHTL structure to graphene electrodes by fabricating organic light-emitting diodes, where graphene is serving as a charge injecting anode. An archetypal bilayer small-molecule OLED architecture [139] was investigated, consisting of a transparent anode coated with hole injecting polymer layers of PEDOT:PEG (PC) (45 nm) and PEDOT:PSS (25 nm), followed by a 50 nm thick N,N'-bis(3-methylphenyl)-N,N'-bis(phenyl) -9,9-spirobifluorene (spiro-TPD) HTL, a 50 nm thick tris(8-hydroxyquinoline) (Alq 3) electron transporting layer (ETL), and a 50 nm thick Ag/Mg alloy cathode with a 70 nm thick Ag protective overlayer (Figure 6.26(a)). The anodes used were: (Device A - control) ITO-coated glass; and (Device B) graphene deposited on quartz substrates. Active layers were grown simultaneously for Devices A and B. 170 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS (a) N -4. / \ / (b) \ / 8000 6000 1 4000 F (U 2000 1 w 01 300 400 500 600 700 800 Wavelength (nm) 171 900 1000 1100 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS (c) 10, 0(4 ITO o Graphene 0 E 101 I 100 r m~5 r S10 100 10~ r 0 102c2 10 r 10 - + M ~0-1 V+1 10 - 1' 0 100 Voltage (V) (d)10 0 o ITO o Graphene 10- U w 102 E -2 Gr C h. w 1; -4 in I10-4 10 10-2 10-1 10 101 Current density (A/cm2 Figure 6.26 Electroluminescence performance of organic light-emitting diodes using graphene anodes. (a) Schematic of the device structure: graphene/PEDOT:PEG(PC), 45 nm/PEDOT:PSS, 25 nm/spiro-TPD, 50 nm/Alq 3, 50 nm/Ag/Mg, 50 nm/Ag, 70 nm. (b) Electroluminescence (EL) spectrum and photograph of the graphene-based OLED at 4.5 V (8.69 mA/cm2 ) applied bias, exhibiting uniform green emission characteristic of Alq 3 . (c) Current density (J,circles) and luminance (L, squares) versus applied forward bias (V) for OLEDs based on graphene (red) and ITO (black). In both cases two conduction regimes are observed, each described by J oc Vm'": Ohmic (m = 0) or space-chargelimited conduction (m = 1) below EL turn-on at 2.4 V; and trap-limited conduction 172 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS (m > 2) at higher voltages. (d) External quantum efficiencies (EQEs) of OLEDs using graphene (red) and ITO (black) anodes, as a function of J. For the graphene devices, a maximum EQE of ~0.27 % and a luminance of ~77 cd/m 2 were reproducibly reached at 2 ~5 V (17.6 mA/cm2), beyond which failure was typically observed. Nevertheless, the similarity in J, L and EQE OLED performance when ITO is replaced with graphene attests to the potential of this approach. L and EQE are calculated based on emission from the front faces of the OLEDs, as described in Chapter 6.4.5. Under applied bias both devices exhibit uniform green electroluminescence (EL) characteristic of Alq 3 emission, as illustrated by the EL spectrum (centered at ~530 nm) and photograph for Device B in Figure 6.26(b). The uniformity of EL confirms that the graphene electrodes form complete films. Next, the forward-bias current-voltage- luminance (J-V-L) characteristics of the ITO- and graphene-based OLEDs are compared in Figure 6.26(c). In both cases, the J-V behavior is described by J oc V' (solid lines), a well-characterized signature of bulk-limited conduction in the presence of traps in this OLED architecture [139]. The observation of bulk-limited (versus contact-limited) currents, together with the functional similarity of the J-V curves for Devices A and B, suggests that graphene affects hole injection with comparable efficiency to ITO in these devices. This is corroborated by the very similar L-V results obtained with the two electrodes, with turn-on voltages of~2.4 V. The EQEs (external quantum efficiency) of the two OLEDs, shown in Figure 6.26(d) as a function of J, are also closely matched. Nevertheless, the EQE of the control, Device A, peaks at ~ 0.43 %, slightly lower than that obtained in an identical OLED without PEDOT:PEG (PC) (~ 0.7 %). We note, however, that our graphene-based device has not yet been optimized; alternative materials exhibiting better wetting on graphene than hydrophilic PEDOT:PSS and capable of effectively mediating hole injection from graphene into HTLs have since been found to yield improved OLED performance and are currently under investigation. Moreover, whilst Device B operated reproducibly (with ~80% yields) at biases of up to 2 ~5 V (J 17.6 mA/cm , EQE ~0.27 %), failure was typically observed at higher voltages. A recent report also showed a graphene-based OLED operating at up to ~5 V [140]. The failure we observe is most likely a result of the field-induced breakdown of the very thin graphene film and can presumably be remedied through the use of thicker graphene 173 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS anodes, although the effect of the increased thicknesses of the graphene films need to be carefully monitored. 6.4.5 OLED Device characterization Organic active layers (spiro-TPD (Luminescence Technology Corp., >99 %), Alq3 (Tokyo Chemical Industry Co., LTD., >98.0 %), and top cathode (Ag (Alfa Aesar, 1-3 mm shot, 99.999 %), Mg (Alfa Aesar, -4 mesh, 99.98 %)) were thermally evaporated under the same environment as the OPV. Alq3 was purified once before use via thermal gradient sublimation and the other materials were used as received. Measurements were performed in a nitrogen-filled glovebox. Current-voltage characteristics of our OLEDs were recorded using a computer-controlled Keithley 2636A current/voltage source meter. Simultaneously, EL emission power from the front faces of the devices was measured using a calibrated Newport 818-UV silicon photodetector with an 'under filled' active area [141], and recorded with a computer-interfaced Newport Multi-Function Optical Meter 1835-C. EQEs were then calculated as described in Ref. [141] EL spectra of biased devices were taken with a fiber-coupled Ocean Optics spectrometer, enabling luminance values to also be computed. EL was photographed with an unmodified digital camera. 6.4.6 Conclusions In this work, we have developed a novel yet simple method to non-destructively modify the graphene surface using a bi-layer HTL structure, and showed its general applications to both OPV and OLED devices. This method ensures smooth and complete coverage of the HTL which is an essential factor for the successful integration of graphene TCEs into the opto-electronic devices. We have demonstrated that graphene based devices utilizing this innovative BLHTL architecture show comparable performances to the conventional ITO based devices without any further treatments to the pristine graphene sheets. This 174 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS similarity suggests that graphene serves as a viable replacement/alternative for ITO in various opto-electronic applications. With desirable device yields, reproducibility, and efficacy, we anticipate that graphene based OPV or OLED devices have become one step closer toward realization for real-life applications. 175 CHAPTER 6: CVD GRAPHENE BASED ORGANIC SOLAR CELLS (This page is intentionally left blank) 176 CHAPTER 7: CONCLUSIONS Chapter 7 7. Conclusions Graphene, a hexagonal arrangement of carbon atoms forming a one-atom thick planar sheet, possesses several unique and outstanding electrical, mechanical, optical, and chemical properties. This two-dimensional building block for carbon materials of all other dimensionalities (OD buckyballs, 1D nanotubes, and 3D graphite) has been widely studied by theorists since the middle of the last century; however, it wasn't until the first demonstration of the successful isolation of single- and few- layer graphene by the mechanical cleaving of highly ordered pyrolytic graphite (HOPG) when graphene began to draw worldwide attention. Furthermore, continuous and scalable large area synthesis by chemical vapor deposition (CVD) has led to a significant increase in studies of numerous research areas. Among the many interesting properties of graphene, its uniformly high transparency in the visible and near infrared region, with good electrical conductivity and mechanical robustness, place graphene as a promising candidate for an alternative to indium tin oxide (ITO) as a transparent conducting electrode (TCE). For the successful integration of graphene films as transparent window electrodes in practical organic photovoltaics, several criteria, such as electrical conductivity, optical transmittance, and work function, must be considered. Recently, several works have already demonstrated that graphene possesses high transmittance with moderate conductivity and is also mechanically robust, thus suitable for TCE applications. The most critical factor, however, is the energy level alignment between the work function of graphene and the highest occupied molecular orbital (HOMO) of the electron donor material. The work function of graphene, especially for a monolayer of graphene synthesized from LPCVD, is much lower than that of ITO. This low value for graphene is obviously not a good match for most of the electron donor materials considered in 177 CHAPTER 7: CONCLUSIONS OPV devices, and can induce a large energy barrier at the interface between the graphene electrode and the adjacent organic layer. In this thesis work, we investigated the possible application of graphene as a substitute for ITO anodes in organic photovoltaic solar cells and explained current limitations of graphene in the TCE application and proposed several possible solutions. By applying the technology already developed in OPV and OLED from Professor Bulovid's group, we could investigate how graphene can be utilized as transparent electrodes in those devices. Therefore, we want to thank contributions from Jill Rowehl, Patrick Brown and Trisha Andrew for studying OPV devices and Geoffrey Supran for investigating OLED devices. We also note contributions from Rachel Howden and Miles Barr from Professor Gleason's group for oCVD PEDOT processes (Rachel) and introducing the novel electron donor material DBP (Miles). Here, graphene sheets synthesized from chemical vapor deposition (CVD) with a controlled number of layers were demonstrated as transparent window electrodes in organic photovoltaic devices. It was found that for devices with pristine graphene electrodes, the power conversion efficiency is comparable to their counterparts with ITO electrodes. Nevertheless, the chances for failure in devices with pristine graphene electrodes were much higher than the ones with ITO electrodes, due to the surface wetting challenge between the conventional hole transporting layer PEDOT:PSS and the graphene electrodes. Therefore, the key factor to the success of graphene electrode based OPV devices lies in the proper wetting of the hole transporting buffer layer on the graphene surface for efficient charge extraction or injection at the graphene electrode and organic interface. Various alternative routes were investigated in this work, such as modifying the graphene surface (ideally non-destructively) to become suitable for HTL deposition or developing alternative HTL deposition methods without the need for surface modification. We first showed that gold (III) chloride (AuCl 3) doping on graphene can alter the graphene surface wetting properties such that a uniform coating of the PEDOT:PSS HTL 178 CHAPTER 7: CONCLUSIONS can be achieved and device success rate can be improved. Furthermore, the doping both increased the conductivity and the work function of the graphene electrode, resulting in improved overall device performance. Nevertheless, the doping process introduces large Au particles (up to 100 nm in diameter) onto the graphene film, which can create shorting pathways through the device in many cases. This method is also less favorable due to the high cost of AuCl 3 dopant. Therefore, more stable and robust planarizing buffer layer with good wettability on the graphene surface is necessary, and an alternative transition metal oxide buffer layer (molybdenum trioxide, MoO 3), instead of PEDOT:PSS, was investigated to address the issue of surface immiscibility between graphene and PEDOT:PSS. In this work, we used MoO 3 as the HTL in OPV devices using graphene electrodes, and carried out further investigations in terms of the effect of graphene surface morphology on the OPV performance and the choice of the counter electrode (cathode). We confirmed that MoO 3 can be successfully integrated between the graphene sheet and the subsequent organic layer as an alternative HTL. The use of a direct thermal evaporation of MoO 3 gives a HTL on the graphene surface with better wetting compared to hydrophilic PEDOT:PSS. However, the device performance was not as efficient as the ITO control device with a MoO 3 layer alone and still required the use of PEDOT:PSS on top of the MoO 3 interfacial layer, which allowed better wetting of PEDOT:PSS on the MoO 3-coated graphene. From the next part of the thesis work, we presented a process where PEDOT is directly deposited onto the graphene surface via oxidative chemical vapor deposition (oCVD), patterned via in situ shadow masking. In this process, the polymer layer is formed by directly exposing the substrate to vaporized monomer (EDOT) and an oxidizing agent (FeCl 3 ) under controlled reactor conditions. oCVD is a dry process, which eliminates the wettability issues that arise when using solution-processed PEDOT:PSS, and allows for conformal coverage of the PEDOT on the graphene substrate. The relatively mild deposition conditions (low temperature, moderate pressure, and no use of solvents) allowed for PEDOT to be deposited without damaging or delaminating the graphene 179 CHAPTER 7: CONCLUSIONS electrode. We then demonstrated that graphene based solar cells, fabricated with oCVD PEDOT HTLs, achieve -90% of the performance of their ITO counterparts without any additional treatment to the graphene sheets such as chemical doping. We also showed that graphene films synthesized under atmospheric pressure chemical vapor deposition (APCVD) conditions, using a common copper foil, can be well adopted as transparent electrodes in OPV devices with oCVD PEDOT HTLs. The APCVD films demonstrate comparable performances to graphene films synthesized from LPCVD conditions and we explored possible benefits over the LPCVD process. In the last part of the thesis, we introduced a novel yet simple approach to apply HTL onto the graphene surface by non-destructive surface modification of graphene using commercially available conducting polymers. Here, the work function of the graphene was modified via surface-engineering achieved by the insertion of a thin polymeric layer, PEDOT:PEG (PC), between the "unmodified" graphene surface and the PEDOT:PSS. PEDOT:PEG (PC) is used as an intermediate layer for the subsequent PEDOT:PSS layer which significantly improved the PEDOT:PSS wetting on the graphene/PEDOT:PEG (PC) composite: PEDOT:PSS with its high work function aligns the energy level with the HOMO of the donor material and PEDOT:PEG (PC) layer enables the wetting at the graphene and PEDOT:PSS interface. With this innovative bi-layer hole transporting layer (BLHTL) structure, we demonstrated that the performance of graphene based OPV devices approached similar values to that of the conventional ITO based devices, without requiring any additional treatments to the pristine graphene electrodes. We also showed the general applicability of this BLHTL structure by fabricating organic light-emitting diodes, where graphene served as a charge injecting anode and observed comparable performance to the ITO counterparts. From the above several studies that we have investigated in this thesis work, we anticipate that graphene indeed serves as a promising replacement or alternative for ITO in various opto-electronic applications. With desirable device yields, reproducibility, and efficiency, we expect that graphene based OPV or OLED devices have moved forward to a practical realization of industrial applications. 180 CHAPTER 7: CONCLUSIONS (This page is intentionally left blank) 181 PUBLICATION Publications 1. K. K. Kim, R. R. Alfonso, Y. S. Shi, H. Park, L. Li, Y. H. Lee, and J. Kong, "Enhancing the conductivity of transparent graphene films via doping", Nanotechnology, 21, 285205 (2010) 2. H. Park, J. Rowehl, K. Kim, V. Bulovid, and J. Kong, "Doped graphene electrodes for organic solar cells", Nanotechnology, 21, 505204 (2010) 3. H. Park, P. R. Brown, V. Bulovid, and J. Kong, "Graphene as transparent conducting electrodes in organic photovoltaic: studies in graphene morphology, hole transporting layers, and counter electrodes", Nano letters, 12, 133-140 (2011) 4. H. Park,* R. M. Howden,* M. C. Barr, V. Bulovid, K. Gleason and J Kong, "Graphene Organic ethylenedioxythiophene) Solar Cells with Vapor Printed Poly (3,4- Hole Transporting Layers", ACS nano in revision (2012). *These authors contributed equally to this work. 5. H. Park, G. J. Supran, M. Smith, S. Grade'ak, V. Bulovid, and J. Kong, "Interface Engineering of Graphene Sheets for Applications in Organic Photovoltaics and Organic Light Emitting Diodes", in preparation (2012) 6. H. Park, S. Jang, M. Smith, S. Grade'ak, V. Bulovid, and J. Kong, "Toward Efficient Graphene Cathode-based Organic Solar Cells via Non-destructive Interface Engineering", in preparation (2012) 7. H. Park, V. Bulovid, and J. Kong, "Application of solvent modified PEDOT:PSS layer to graphene electrodes in organic solar cells", in preparation (2012) 8. H. Park,* S. Jang,* J. Cheng, V. Bulovid, S. Gradebak, and J. Kong, "Efficient Hybrid Organic-Inorganic Solar Cells using P3HT-ZnO nanowires on Interface Engineered Graphene Cathode", in contributed equally to this work. 182 preparation (2012). *These authors PUBLICATION (This page is intentionally left blank) 183 REFERENCES Bibliography [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] http://www.nrel.gov/analysis/pdfs/46025.pdf. Ginley D, Green MA. MRS Bulletin. 2008; 33: 355. Arango AC, Oertel DC, Xu YF, Bawendi MG, Bulovic V. Heterojunction Photovoltaics Using Printed Colloidal Quantum Dots as a Photosensitive Layer. Nano Lett 2009; 9: 860-863. Tang CW. 2-Layer Organic Photovoltaic Cell. Appl Phys Lett 1986; 48: 183-185. Peumans P, Yakimov A, Forrest SR. Small molecular weight organic thin-film photodetectors and solar cells. JApplPhys 2003; 93: 3693-3723. Nelson J. Organic photovoltaic films. Current Opinion in Solid State & Materials Science 2002; 6: 87-95. Hoppe H, Sariciftci NS. Organic solar cells: An overview. J Mater Res 2004; 19: 1924-1945. Wohrle D, Meissner D. Organic Solar-Cells. Adv Mater 1991; 3: 129-138. Handbook of Conducting Polymers V, edited by T.A. Skotheim (Marcel Dekker, Inc., New York, 1986). Brabec CJ, Sariciftci NS, Hummelen JC. Plastic solar cells. Advanced Functional Materials 2001; 11: 15-26. Winder C, Sariciftci NS. Low bandgap polymers for photon harvesting in bulk heterojunction solar cells. Journalof MaterialsChemistry 2004; 14: 1077-1086. Malliaras G, Friend R. An organic electronics primer. Physics Today 2005; May: 53. Dimitrakopoulos CD, Mascaro DJ. Organic thin-film transistors: A review of recent advances. IBM JRes Dev 2001; 45: 11-27. Gregg BA, Hanna MC. Comparing organic to inorganic photovoltaic cells: Theory, experiment, and simulation. JAppl Phys 2003; 93: 3605-3614. Kietzke T. Recent Advances in Organic Solar Cells. Advances in OptoElectronics 2007: 40285. Chen SG, Stradins P, Gregg BA. Doping highly ordered organic semiconductors: Experimental results and fits to a self-consistent model of excitonic processes, doping, and transport. JPhys Chem B 2005; 109: 13451-13460. Gregg BA, Chen SG, Cormier RA. Coulomb forces and doping in organic semiconductors. Chem Mat 2004; 16: 4586-4599. Gledhill SE, Scott B, Gregg BA. Organic and nano-structured composite photovoltaics: An overview. JMaterRes 2005; 20: 3167-3179. http://en.wikipedia.org/wiki/Solarcell. Hwang J, Wan A, Kahn A. Energetics of metal-organic interfaces: New experiments and assessment of the field. Mater Sci Eng R-Rep 2009; 64: 1-31. Hill IG, Kahn A, Soos ZG, Pascal RA. Charge-separation energy in films of piconjugated organic molecules. ChemicalPhysics Letters 2000; 327: 181-188. Knupfer M, Peisert H, Schwieger T. Band-gap and correlation effects in the organic semiconductor Alq(3). PhysicalReview B 2002; 65. 184 REFERENCES [23] Alvarado SF, Seidler PF, Lidzey DG, Bradley DDC. Direct determination of the exciton binding energy of conjugated polymers using a scanning tunneling microscope. Physical Review Letters 1998; 81: 1082-1085. [24] Knupfer M, Fink J, Zojer E, Leising G, Scherf U, Mullen K. Localized and delocalized singlet excitons in ladder-type poly(paraphenylene). Physical Review B 1998; 57: R4202-R4205. [25] Barth S, Bassler H. Intrinsic photoconduction in PPV-type conjugated polymers. Physical Review Letters 1997; 79: 4445-4448. [26] Bernede JC. Organic photovoltaic cells: History, principle and techniques. J Chil Chem Soc 2008; 53: 1549-1564. [27] Pope M. Electronicprocesses in organic crystals. Oxford: Oxford University Press; 1982. [28] Kersting R, Lemmer U, Deussen M, Bakker HJ, Mahrt RF, Kurz H et al. Ultrafast Field-Induced Dissociation of Excitons in Conjugated Polymers. PhysicalReview Letters 1994; 73: 1440-1443. [29] [30] Rand BP, Burk DP, Forrest SR. Offset energies at organic semiconductor heterojunctions and their influence on the open-circuit voltage of thin-film solar cells. PhysicalReview B 2007; 75. Pettersson LAA, Roman LS, Inganas 0. Modeling photocurrent action spectra of photovoltaic devices based on organic thin films. JAppl Phys 1999; 86: 487-496. [31] [32] Bulovic V, Forrest SR. Study of localized and extended excitons in 3,4,9,10perylenetetracarboxylic dianhydride (PTCDA) .2. Photocurrent response at low electric fields. Chem Phys 1996; 210: 13-25. Halls JJM, Pichler K, Friend RH, Moratti SC, Holmes AB. Exciton diffusion and dissociation in a poly(p-phenylenevinylene)/C-60 heterojunction photovoltaic cell. Appl Phys Lett 1996; 68: 3120-3122. [33] van Hal PA, Janssen RAJ, Lanzani G, Cerullo G, Zavelani-Rossi M, De Silvestri S. Full temporal resolution of the two-step photoinduced energy-electron transfer in a fullerene-oligothiophene-fullerene triad using sub-10 fs pump-probe spectroscopy. Chemical Physics Letters 2001; 345: 33-38. [34] Zerza G, Brabec CJ, Cerullo G, De Silvestri S, Sariciftci NS. Ultrafast charge transfer in conjugated polymer-fullerene composites. Synthetic Metals 2001; 119: 637-638. [35] [36] [37] Nelson J. The physcis of solar cells. London: Imperial College Press; 2003. Bube RH, Fahrenbruch AL. Advances in Electronicsand Electron Physics. New York: Academic Press; 1981. Fahrenbruch AL, Aranovich J. Solar Energy Conversion. New York: Springer- Verlag; 1979. [38] Nelson J, Kirkpatrick J, Ravirajan P. Factors limiting the efficiency of molecular photovoltaic devices. PhysicalReview B 2004; 69. [39] Waldauf C, Scharber MC, Schilinsky P, Hauch JA, Brabec CJ. Physics of organic [40] bulk heterojunction devices for photovoltaic applications. JAppl Phys 2006; 99. Geim AK, Novoselov KS. The rise of graphene. Nature Materials 2007; 6: 183- 191. [41] McClure JW. Band Structure Of Graphite and Dehaas-Vanalphen Effect. Physical Review 1957; 108: 612-618. 185 REFERENCES [42] [43] Wallace PR. The Band Theory of Graphite. Physical Review 1947; 71: 622-634. Novoselov KS, Jiang D, Schedin F, Booth TJ, Khotkevich VV, Morozov SV et al. Two-dimensional atomic crystals. Proceedingsof the NationalAcademy of Sciences of the United States ofAmerica 2005; 102: 10451-10453. [44] Berger C, Song ZM, Li XB, Wu XS, Brown N, Naud C et al. Electronic confinement and coherence in patterned epitaxial graphene. Science 2006; 312: [45] 1191-1196. Bolotin KI, Sikes KJ, Jiang Z, Klima M, Fudenberg G, Hone J et al. Ultrahigh electron mobility in suspended graphene. Solid State Communications 2008; 146: 351-355. [46] Moser J, Barreiro A, Bachtold A. Current-induced cleaning of graphene. Appl [47] Phys Lett 2007; 91: 163513. Bunch JS, van der Zande AM, Verbridge SS, Frank IW, Tanenbaum DM, Parpia JM et al. Electromechanical resonators from graphene sheets. Science 2007; 315: [48] 490-493. Schedin F, Geim AK, Morozov SV, Hill EW, Blake P, Katsnelson MI et al. Detection of individual gas molecules adsorbed on graphene. Nature Materials 2007; 6: 652-655. [49] [50] Eda G, Fanchini G, Chhowalla M. Large-area ultrathin films of reduced graphene oxide as a transparent and flexible electronic material. Nature Nanotechnology 2008; 3: 270-274. Berger C, Song ZM, Li TB, Li XB, Ogbazghi AY, Feng R et al. Ultrathin epitaxial graphite: 2D electron gas properties and a route toward graphene-based [51] nanoelectronics. JPhys Chem B 2004; 108: 19912-19916. Li XL, Wang XR, Zhang L, Lee SW, Dai HJ. Chemically derived, ultrasmooth graphene nanoribbon semiconductors. Science 2008; 319: 1229-1232. [52] Stankovich S, Dikin DA, Piner RD, Kohlhaas KA, Kleinhammes A, Jia Y et al. Synthesis of graphene-based nanosheets via chemical reduction of exfoliated graphite oxide. Carbon 2007; 45: 1558-1565. [53] Sutter PW, Flege JI, Sutter EA. Epitaxial graphene on ruthenium. Nature Materials 2008; 7: 406-411. [54] Somani PR, Somani SP, Umeno M. Planer nano-graphenes from camphor by [55] [56] CVD. Chemical Physics Letters 2006; 430: 56-59. Geim AK. Graphene: Status and Prospects. Science 2009; 324: 1530-1534. Rao CNR, Sood AK, Subrahmanyam KS, Govindaraj A. Graphene: The New Two-Dimensional Nanomaterial. Angewandte Chemie-InternationalEdition 2009; 48: 7752-7777. [57] [58] Saito R, Dresselhaus G, Dresselhaus MS. Physical properties of carbon nanotubes. ImperialCollege Press 1998; London. Semenoff GW. Condensed-Matter Simulation of a Three-Dimensional Anomaly. PhysicalReview Letters 1984; 53: 2449-2452. [59] Haldane FDM. Model for a Quantum Hall Effect without Landau Levels: Condensed-Matter Realization of the "Parity Anomaly". PhysicalReview Letters [60] Katsnelson MI. Graphene: carbon in two dimensions. Mater Today 2007; 10: 20- 1988; 61: 2015-2018. 27. 186 REFERENCES [61] Morozov SV, Novoselov KS, Geim AK. Electron transport in graphene. Phys Usp 2008; 51: 744-748. [62] Slonczewski JC, Weiss PR. Band Structure of Graphite. PhysicalReview 1958; 109: 272-279. [63] Katsnelson MI, Novoselov KS, Geim AK. Chiral tunnelling and the Klein paradox in graphene. Nat Phys 2006; 2: 620-625. [64] Castro Neto AH, Guinea F, Peres NMR, Novoselov KS, Geim AK. The electronic properties of graphene. Rev Mod Phys 2009; 81: 109-162. [65] Henrie J, Kellis S, Schultz SM, Hawkins A. Electronic color charts for dielectric films on silicon. Opt Express 2004; 12: 1464-1469. [66] Palik ED. Handbook of Optical Constants of Solids. New York: Academic Press; [67] Pauling L. The Nature of the Chemical Bond. Ithaca: Cornell University Press; [68] 1960. Blake P, Hill EW, Neto AHC, Novoselov KS, Jiang D, Yang R et al. Making 1991. [69] [70] graphene visible. Appl Phys Lett 2007; 91. Anders H. Thin Films in Optics. London: Focal Press; 1967. Kuzmenko AB, van Heumen E, Carbone F, van der Marel D. Universal optical conductance of graphite. PhysicalReview Letters 2008; 100. [71] Nair RR, Blake P, Grigorenko AN, Novoselov KS, Booth TJ, Stauber T et al. Fine structure constant defines visual transparency of graphene. Science 2008; [72] 320: 1308-1308. Casiraghi C, Hartschuh A, Lidorikis E, Qian H, Harutyunyan H, Gokus T et al. Rayleigh imaging of graphene and graphene layers. Nano Lett 2007; 7: 27112717. [74] Li GH, Luican A, dos Santos J, Neto AHC, Reina A, Kong J et al. Observation of Van Hove singularities in twisted graphene layers. Nat Phys 2010; 6: 109-113. Mak KF, Shan J, Heinz TF. Electronic Structure of Few-Layer Graphene: Experimental Demonstration of Strong Dependence on Stacking Sequence. PhysicalReview Letters 2010; 104. [75] Wang F, Zhang YB, Tian CS, Girit C, Zettl A, Crommie M et al. Gate-variable [76] optical transitions in graphene. Science 2008; 320: 206-209. Bonaccorso F, Sun Z, Hasan T, Ferrari AC. Graphene photonics and optoelectronics. Nat Photonics 2010; 4: 611-622. [77] Bae S, Kim H, Lee Y, Xu XF, Park JS, Zheng Y et al. Roll-to-roll production of [73] 30-inch graphene films for transparent electrodes. Nature Nanotechnology 2010; [78] [79] [80] 5: 574-578. Phillips JM, Kwo J, Thomas GA, Carter SA, Cava RJ, Hou SY et al. Transparent Conducting Thin-Films of GaInO 3. Appl Phys Lett 1994; 65: 115-117. Scott JC, Kaufman JH, Brock PJ, DiPietro R, Salem J, Goitia JA. Degradation and failure of MEH-PPV light-emitting diodes. JAppl Phys 1996; 79: 2745-2751. Schlatmann AR, Floet DW, Hilberer A, Garten F, Smulders PJM, Klapwijk TM et al. Indium contamination from the indium-tin-oxide electrode in polymer light- emitting diodes. Appl Phys Lett 1996; 69: 1764-1766. 187 REFERENCES [81] [82] [83] [84] [85] [86] [87] [88] [89] [90] [91] [92] [93] [94] [95] [96] [97] [98] [99] Hur SH, Park 00, Rogers JA. Extreme bendability of single-walled carbon nanotube networks transferred from high-temperature growth substrates to plastic and their use in thin-film transistors. Appl Phys Lett 2005; 86. Granqvist CG. Transparent conductors as solar energy materials: A panoramic review. Solar Energy Materials and Solar Cells 2007; 91: 1529-1598. De S, Higgins TM, Lyons PE, Doherty EM, Nirmalraj PN, Blau WJ et al. Silver Nanowire Networks as Flexible, Transparent, Conducting Films: Extremely High DC to Optical Conductivity Ratios. ACS Nano 2009; 3: 1767-1774. De S, Coleman JN. Are There Fundamental Limitations on the Sheet Resistance and Transmittance of Thin Graphene Films? A CS Nano 2010; 4: 2713-2720. Dressel M, Gruner G. Electrodynamics of Solids: OpticalPropertiesofElectrons in Matter. Cambridge: Cambridge University Press; 2002. Dan B, Irvin GC, Pasquali M. Continuous and Scalable Fabrication of Transparent Conducting Carbon Nanotube Films. ACS Nano 2009; 3: 835-843. Novoselov KS, Geim AK, Morozov SV, Jiang D, Katsnelson MI, Grigorieva IV et al. Two-dimensional gas of massless Dirac fermions in graphene. Nature 2005; 438: 197-200. Reina A, Jia XT, Ho J, Nezich D, Son HB, Bulovic V et al. Large Area, FewLayer Graphene Films on Arbitrary Substrates by Chemical Vapor Deposition. Nano Lett 2009; 9: 30-35. Reina A, Son HB, Jiao LY, Fan B, Dresselhaus MS, Liu ZF et al. Transferring and Identification of Single- and Few-Layer Graphene on Arbitrary Substrates. JournalofPhysical Chemistry C 2008; 112: 17741-17744. Bhaviripudi S, Jia XT, Dresselhaus MS, Kong J. Role of Kinetic Factors in Chemical Vapor Deposition Synthesis of Uniform Large Area Graphene Using Copper Catalyst. Nano Lett 2010; 10: 4128-4133. Hamilton JC, Blakely JM. Carbon Segregation to Single-Crystal Surfaces of Pt, Pd and Co. SurfSci 1980; 91: 199-217. An H, Lee WJ, Jung J. Graphene synthesis on Fe foil using thermal CVD. Curr Appl Phys 2011; 11: S81-S85. Himpsel FJ, Christmann K, Heimann P, Eastman DE, Feibelman PJ. Adsorbate Band Dispersions for C on Ru(000 1). SurfSci 1982; 115: LI 59-LI 64. Kholin NA, Rutkov EV, Tontegode AY. The Nature of the Adsorption Bond Between Graphite Islands and Iridium Surface. Surf Sci 1984; 139: 155-172. Coraux J, N'Diaye AT, Busse C, Michely T. Structural coherency of graphene on Ir(l 11). Nano Lett 2008; 8: 565-570. Mattevi C, Kim H, Chhowalla M. A review of chemical vapour deposition of graphene on copper. JournalofMaterials Chemistry 2011; 21: 3324-3334. Kim KK, Reina A, Shi YM, Park H, Li U, Lee YH et al. Enhancing the conductivity of transparent graphene films via doping. Nanotechnology 2010; 21. Park Y, Choong V, Gao Y, Hsieh BR, Tang CW. Work function of indium tin oxide transparent conductor measured by photoelectron spectroscopy. Appl Phys Lett 1996; 68: 2699-2701. Milliron DJ, Hill IG, Shen C, Kahn A, Schwartz J. Surface oxidation activates indium tin oxide for hole injection. JAppl Phys 2000; 87: 572-576. 188 REFERENCES [100] Fujishima D, Kanno H, Kinoshita T, Maruyama E, Tanaka M, Shirakawa M et al. Organic thin-film solar cell employing a novel electron-donor material. Solar Energy Materials and Solar Cells 2009; 93: 1029-1032. [101] Wang Y, Chen XH, Zhong YL, Zhu FR, Loh KP. Large area, continuous, fewlayered graphene as anodes in organic photovoltaic devices. Appl Phys Lett 2009; 95: 063302. [102] Gunes S, Neugebauer H, Sariciftci NS. Conjugated polymer-based organic solar cells. ChemicalReviews 2007; 107: 1324-1338. [103] [104] Peumans P, Forrest SR. Very-high-efficiency double-heterostructure copper phthalocyanine/C-60 photovoltaic cells. Appl Phys Lett 2001; 79: 126-128. Murakami TN, Fukushima Y, Hirano Y, Tokuoka Y, Takahashi M, Kawashima N. Surface modification of polystyrene and poly(methyl methacrylate) by active oxygen treatment. Colloids and Surfaces B-Biointerfaces 2003; 29: 171-179. [105] Ishigami M, Chen JH, Cullen WG, Fuhrer MS, Williams ED. Atomic structure of graphene on Si02. Nano Lett 2007; 7: 1643-1648. [106] Xue JG, Uchida S, Rand BP, Forrest SR. 4.2% efficient organic photovoltaic cells with low series resistances. Appl Phys Lett 2004; 84: 3013-3015. Jonsson SKM, Birgerson J, Crispin X, Greczynski G, Osikowicz W, van der Gon AWD et al. The effects of solvents on the morphology and sheet resistance in [107] poly (3,4-ethylenedioxythiophene)-polystyrenesulfonic acid (PEDOT-PSS) films. Synthetic Metals 2003; 139: 1-10. [108] Kim JY, Jung JH, Lee DE, Joo J. Enhancement of electrical conductivity of poly(3,4-ethylenedioxythiophene)/poly(4-styrenesulfonate) by a change of solvents. Synthetic Metals 2002; 126: 311-316. [109] Sapp S, Luebben S, Losovyj YB, Jeppson P, Schulz DL, Caruso AN. Work function and implications of doped poly(3,4-ethylenedioxythiophene)-co- [110] poly(ethylene glycol). Appl Phys Lett 2006; 88: 152107. Shi YM, Kim KK, Reina A, Hofmann M, Li LJ, Kong J. Work Function Engineering of Graphene Electrode via Chemical Doping. A CS Nano 2010; 4: [111] 2689-2694. Dettlaff-Weglikowska U, Skakalova V, Graupner R, Jhang SH, Kim BH, Lee HJ et al. Effect of SOC12 treatment on electrical and mechanical properties of singlewall carbon nanotube networks. JAm Chem Soc 2005; 127: 5125-5131. [112] Ciprelli JL, Clarisse C, Delabouglise D. Enhanced stability of conducting poly(3octylthiophene) thin films using organic nitrosyl compounds. Synthetic Metals [113] 1995; 74: 217-222. Abdou MSA, Holdcroft S. Oxidation of pi-conjugated polymers with gold [115] trichloride - enhanced stability of the electronically conducting state and electroless deposition of Au(O). Synthetic Metals 1993; 60: 93-96. Choi HC, Shim M, Bangsaruntip S, Dai HJ. Spontaneous reduction of metal ions on the sidewalls of carbon nanotubes. JAm Chem Soc 2002; 124: 9058-9059. Shrotriya V, Li G, Yao Y, Chu CW, Yang Y. Transition metal oxides as the [116] buffer layer for polymer photovoltaic cells. Appl Phys Lett 2006; 88: 3. Godoy A, Cattin L, Toumi L, Diaz FR, del Valle MA, Soto GM et al. Effects of [114] the buffer layer inserted between the transparent conductive oxide anode and the 189 REFERENCES organic electron donor. Solar Energy Materialsand Solar Cells 2010; 94: 648- 654. [117] You H, Dai YF, Zhang ZQ, Ma DG. Improved performances of organic lightemitting diodes with metal oxide as anode buffer. JAppl Phys 2007; 101: 3. [118] Reynolds KJ, Barker JA, Greenham NC, Friend RH, Frey GL. Inorganic solutionprocessed hole-injecting and electron-blocking layers in polymer light-emitting [119] diodes. JAppl Phys 2002; 92: 7556-7563. Chu CW, Li SH, Chen CW, Shrotriya V, Yang Y. High-performance organic thin-film transistors with metal oxide/metal bilayer electrode. Appl Phys Lett [120] 2005; 87: 3. Kroger M, Hamwi S, Meyer J, Riedl T, Kowalsky W, Kahn A. Role of the deeplying electronic states of MoO3 in the enhancement of hole-injection in organic [121] thin films. Appl Phys Lett 2009; 95: 3. Irfan, Ding HJ, Gao YL, Small C, Kim DY, Subbiah J et al. Energy level evolution of air and oxygen exposed molybdenum trioxide films. Appl Phys Lett 2010; 96: 3. [122] Brabec CJ, Cravino A, Meissner D, Sariciftci NS, Fromherz T, Rispens MT et al. Origin of the open circuit voltage of plastic solar cells. Advanced Functional Materials 2001; 11: 374-380. [123] Eo YS, Rhee HW, Chin BD, Yu JW. Influence of metal cathode for organic [124] [125] photovoltaic device performance. Synthetic Metals 2009; 159: 1910-1913. Mihailetchi VD, Koster LJA, Blom PWM. Effect of metal electrodes on the performance of polymer: fullerene bulk heterojunction solar cells. Appl Phys Lett 2004; 85: 970-972. Mihailetchi VD, Blom PWM, Hummelen JC, Rispens MT. Cathode dependence of the open-circuit voltage of polymer: fullerene bulk heterojunction solar cells. J Appl Phys 2003; 94: 6849-6854. [126] Michaelson HB. The work function of the elements and its periodicity. JAppl [127] Phys 1977; 48: 4729-4733. Frohne H, Shaheen SE, Brabec CJ, Muller DC, Sariciftci NS, Meerholz K. Influence of the anodic work function on the performance of organic solar cells. [128] Chemphyschem 2002; 3: 795-+. Barr MC, Rowehl JA, Lunt RR, Xu J, Wang A, Boyce CM et al. Direct Monolithic Integration of Organic Photovoltaic Circuits on Unmodified Paper. Adv Mater 2011; 23: 3500-3505. [129] Im SG, Gleason KK, Olivetti EA. Doping level and work function control in oxidative chemical vapor deposited poly (3,4-ethylenedioxythiophene). Appl Phys Lett 2007; 90: 152112. [130] [131] [132] Lock JP, Im SG, Gleason KK. Oxidative chemical vapor deposition of electrically conducting poly(3,4-ethylenedioxythiophene) films. Macromolecules 2006; 39: 5326-5329. Li XS, Cai WW, An JH, Kim S, Nah J, Yang DX et al. Large-Area Synthesis of High-Quality and Uniform Graphene Films on Copper Foils. Science 2009; 324: 1312-1314. Park H, Rowehl JA, Kim KK, Bulovic V, Kong J. Doped graphene electrodes for organic solar cells. Nanotechnology 2010; 21: 6. 190 REFERENCES [133] Wang SR, Zhang Y, Abidi N, Cabrales L. Wettability and Surface Free Energy of Graphene Films. Langmuir 2009; 25: 11078-11081. [134] Park H, Brown PR, Buloyic V, Kong J. Graphene As Transparent Conducting Electrodes in Organic Photovoltaics: Studies in Graphene Morphology, Hole Transporting Layers, and Counter Electrodes. Nano Lett 2012; 12: 133-140. [135] De Arco LG, Zhang Y, Schlenker CW, Ryu K, Thompson ME, Zhou CW. Continuous, Highly Flexible, and Transparent Graphene Films by Chemical [136] Vapor Deposition for Organic Photovoltaics. A CS Nano 2010; 4: 2865-2873. Wang YWY, Tong SW, Xu XF, Ozyilmaz B, Loh KP. Interface Engineering of Layer-by-Layer Stacked Graphene Anodes for High-Performance Organic Solar Cells. Adv Mater 2011; 23: 1514-1518. [137] [138] de Jong MP, van Ijzendoom LJ, de Voigt MJA. Stability of the interface between indium-tin-oxide and poly(3,4-ethylenedioxythiophene)/poly(styrenesulfonate) in polymer light-emitting diodes. Appl Phys Lett 2000; 77: 2255-2257. Choe M, Lee BH, Jo G, Park J, Park W, Lee S et al. Efficient bulk-heterojunction photovoltaic cells with transparent multi-layer graphene electrodes. Organic Electronics 2011; 11: 1864-1869. [139] Burrows PE, Shen Z, Bulovic V, McCarty DM, Forrest SR, Cronin JA et al. Relationship between electroluminescence and current transport in organic heterojunction light-emitting devices. JAppl Phys 1996; 79: 7991-8006. [140] Han T, Lee Y, Choi M, Woo S, Bae S, Hong B et al. Extremely efficient flexible organic light-emitting diodes with modified graphene anode. Nat Photonics 2012; 6: 105-110. [141] Forrest SR, Bradley DDC, Thompson ME. Measuring the efficiency of organic light-emitting devices. Adv Mater 2003; 15: 1043-1048. [142] "During this investigation we found another work [ref 31] reported the use of MoO3 as interfacial layer in OPV with graphene electrode. Nevertheless, in [ref 31] a combination of MoO3 with PEDOT is used." 191