Continuous, Highly Flexible and Transparent. Graphene Films by Chemical Vapor Deposition for. Organic Photovoltaics

Transcription

1 Supporting Information for Continuous, Highly Flexible and Transparent Graphene Films by Chemical Vapor Deposition for Organic Photovoltaics Lewis Gomez De Arco 1,2, Yi Zhang 1,2, Cody W. Schlenker 2, Koungmin Ryu 1, Mark E. Thompson 2 and Chongwu Zhou 1 * 1 Department of Electrical Engineering, University of Southern California, Los Angeles, CA Department of Chemistry, University of Southern California, Los Angeles, CA * chongwuz@usc.edu Supporting Information Available: 1. X-Ray Diffraction Spectrum of High-Temperature Annealed Ni Film. 2. TEM, SAED and Raman of CVD Graphene. 3. Atomic Force Microscopy of CVD Graphene, ITO and SWNT Films. 4. CVD Graphene Organic Photovoltaic Cells on Rigid Substrates. 5. Model Parameter Extraction for CVD Graphene and ITO Based OPVs. 6. Conductance vs. Bending Cycles of Flexible CVD Graphene Films. 1. X-ray diffraction of high-temperature annealed Ni film. X-ray diffraction spectra were collected on the annealed Ni substrates over which graphene films are synthesized. To this end, elemental Ni was thermally evaporated onto precleaned Si/SiO 2 substrates up to a thickness of

2 ~1000 Å. Subsequently, Ni/Si/SiO 2 substrates were taken into a sealed hightemperature furnace and heated to 900 C under a hydrogen flow rate of 600 sccm. A Rigaku x-ray diffractometer equipped with a 12 kw rotating anode x-ray generator was employed to investigate the distribution of crystalline planes on the annealed polycrystalline Ni. Figure S1a shows an AFM (a) (b) Intensity (a.u.) Ni (111) Ni(200) Si(100) θ (degrees) Figure S1. (a) AFM image of a 300 nm Ni film deposited on Si/SiO2 substrate after high temperature annealing. (b) Typical X-ray diffraction spectrum of annealed Ni film. image of a typical annealed Ni surface. Irregular and faceted-shape surface are consistent with polycrystalline formation. X-ray diffraction spectrum shown in Figure S1b reveals the presence of (111) and (200) planes. Furthermore, it is clear that the surface of annealed Ni film is comprised predominantly by the (111) plane.

3 2. TEM, SAED and Raman of CVD Graphene. Free-standing PMMA/CVD graphene was obtained by etching the underlying polycrystalline Ni surface with aqueous nitric acid solution. Consecutively, acetone is used to dissolve the PMMA layer and allow the deposition of clean CVD graphene films on standard grids for TEM analysis. Figure S2a shows a low magnification TEM image of the deposited CVD graphene films, while the selected area diffraction pattern along the z- direction shows well ordered typical graphite lattice structure. Further spectroscopic evidence is obtained by micro- Raman (a) (b) Raman shift (cm -1 ) Figure S2. (a) Low magnification TEM image of CVD graphene films. Inset shows a selected area electron diffraction (SAED) pattern of the few-layer graphene film. (b) Raman spectrum of CVD graphene film. Figure S2b shows the Raman spectrum corresponding to the sample analyzed in Figure S2a. The observed G-band centered at 1581 cm -1 is characteristic of the C-C stretching in the sp 2 structure of graphene. Another band of interest, the G -

4 band that appears at 2697 cm -1, presents a fairly symmetric lineshape that suggests the films are comprised of few layers of graphene (1-5 layers). The very low D-band intensity at 1354 cm -1 with respect to the intensity of the G-band, indicates a low defect density in the transferred CVD graphene. 3. Atomic force microscopy of CVD Graphene, ITO and SWNT films. In general, CVD graphene obtained under the experimental conditions described here contains 1-5 graphene layers. To further test this point, we perform height profile measurements on the transferred films. Cross section analysis of AFM images taken on transferred CVD graphene on glass substrates allows a quantitative estimate of the film thickness. Figure S3a shows an AFM image of an opening in the graphene film with clear edges. Figure 3Sb shows the height profile along the straight line depicted in Figure 3Sa. A height step of ~1nm can be clearly observed between the substrate surface and the graphene film. Larger vertical distances can be found between the substrate surface and the CVD graphene edge step as well. However, due to irregularities on Ni surface, and as transferred graphene films may suffer from bending, folding and mechanical stress, they may not be lying fully extended and flattened on the receiving substrate. Thus, the lowest vertical distance within the profile edge steps can be regarded as a good estimate of the film thickness. In this case, a thickness of ~1 nm indicates the CVD graphene

5 can be as thin as bilayer graphene. In addition, domains showing a large height profile may correspond to mechanical distortions, such in-plane film compression or film folding, or to multilayer graphene stacks with more than 5 layers. (a) (b) 1.0 nm Figure S3. (a) AFM image of a transferred CVD graphene film onto glass substrate. (b) Cross section measurement of the height of the CVD graphene. Typical thickness exhibited by the transferred films is found within the range 1-3 nm. One important aspect in the use of CVD graphene as transparent conductive electrode in photovoltaic devices is its surface smoothness. Figure S4 shows the AFM images from which r.m.s. surface roughness values were obtained for CVD graphene, ITO and SWNTs on different substrates. Surprisingly, CVD graphene has the ability to form continuous films over large areas while maintaining an outstanding smoothness; such smoothness is substantially higher than that of a

6 percolating nanotube network and similar to that of ITO on glass and PET substrates. RMS: 0.9 nm (a) RMS: 1.3 nm (d) 0.5 μm 0.5 μm RMS: 0.7 nm (b) RMS: 1.1 nm (e) 0.5 μm 0.5 μm RMS: 8.4 nm (c) 0.5 μm Figure S4. AFM images depicting the r.m.s. surface roughness of typical CVD graphene (a), ITO (b), and SWNT (c) films on glass substrates. used as electrodes AFM image of a transferred CVD graphene film onto glass substrate. (d) and (e) show AFM images and r.m.s. surface roughness of CVD graphene and ITO films on PET flexible substrates. All images have the same scale in the z-direction. 4. CVD Graphene Organic Photovoltaic Cells on Rigid Substrates We fabricated solar cells on CVD graphene films transferred on glass (R Sheet : 1.2 kω/sq, T: 82% at 550 nm) and glass substrates coated with ITO (Thickness: 150 nm, R Sheet : 20 Ω/sq, T: 84% at 550 nm). J(V) characteristics of the fabricated devices are plotted in semi-log (Fig. S5a) and linear (Fig. S5b) scale for both

7 devices. Again, CVD graphene solar cells on rigid transparent substrate showed distinct diode behavior with little leakage current at reverse bias in the dark, while exhibiting high dark current under forward bias. Under illumination, the ITO device gives J sc of 5.41 ma/cm 2, V oc of 0.47 V, FF of 0.54 and power conversion efficiency (η) of 1.39%. On the other hand, the CVD graphene device exhibited J sc of 3.45 ma/cm 2, V oc of 0.47 V, FF of 0.47 and (η) of 0.75%. a b Current density (A/cm 2 ) Current density (ma/cm 2 ) Graphene Dark Graphene Light ITO Dark ITO Light ITO Dark ITO Light Graphene Dark Graphene Light Voltage (V) Voltage (V) Figure S5. Logarithmic (a) and linear (b) current density vs voltage characteristics of CVD graphene and ITO photovoltaic cells on glass under 100 mw/cm 2 AM1.5 spectral illumination. Structure of the devices is given by [CVD graphene / PEDOT / CuPc / C 60 / BCP / Al] and [ITO / CuPc / C 60 / BCP / Al] for CVD graphene and ITO OPVs, respectively.

8 Comparison of these devices shows that even though conversion efficiency of the CVD graphene device is lower, the overall performance of the CVD graphene photovoltaic cell is competitive, with FF comparable to that of the control ITO device. Higher transparency of the ITO film may lead to a higher exciton generation rate, which in turn is reflected in higher J sc values. However, smoothness and thickness of the graphene film may favor charge injection and transport. The disparate power conversion efficiency observed between the two cells can be attributed to the higher sheet resistance and lower transparency of the graphene electrode in the G-OPV. Table S1 summarizes representative solar cell performance parameters measured for the CVD graphene and ITO cells as well as reduced GO devices reported in the literature. Clearly, CVD graphene in all cases, on rigid or flexible substrates, compares favorably against reduced GO as transparent anode in OPV cells. Anode J sc (ma/cm 2 ) V oc (V) FF η (%) CVD Graphene ITO Red. GO (Wu et al.) Red. GO (Wang et al.) Table S1. Comparison of performance details of OPV cells built on glass substrates. The structure of the devices is given by [CVD graphene/pedot/cupc/c60/bcp/al] and [ITO/CuPc/C60/BCP/Al] for CVD graphene and ITO OPVs, respectively.

9 5. Model Parameter Extraction for CVD Graphene and ITO Based OPVs. The Shockley equation is given by: J V JR = J s 1 V JR s s exp ph nv + (S1) t Rp J where R s,, J s, J ph, n, and V t are the lumped series resistance, lumped parallel resistance, reverse-bias saturation current-density, photocurrent-density, diode ideality factor, and thermal voltage respectively for a single diode circuit model. The transcendental nature of Eq. S1 limits optimization of the model parameters. Recently, it has been shown 1-5 that an explicit solution may be obtained using the Lambert-W function 6. Rearranging Eq. S1 gives J + (J ph + J s ) V + R s exp R s V J nv t + R s = J s exp + R s nv t V + R. s (S2) After multiply both sides of Eq S1 by R s /nv t exp(α/β), where α = (J ph + J s )R s and β = nv t (R s + ). we define the quantities w R s nv t J + (J ph + J s ) V + R s (S3) and x R s nv t R p J s exp V + (J + J )R ph s s. + R s nv t + R s (S4) Since w=w 0 (x) and W(x)e W(x) =x(v), substituting and solving for J gives J = nv J t W s R s 0 R s nv t (R s + ) exp V + (J ph + J s )R s nv t (R s + ) (J ph + J s ) V, (S5) (R s + )

10 which expresses the measured current-density dependence on applied voltage in terms of the model parameters for a single diode equivalent circuit model. The ProductLog[z] implementation of the Lambert-W function in Mathematica 5.1 was used to perform nonlinear regression according to Eq. S5 to obtain the model parameters for CVD graphene and ITO based OPV devices on PET substrates. 6. Conductance vs. Bending Cycles of Flexible CVD Graphene Films. Further evaluation of the flexibility of the synthesized CVD graphene films was conducted by transferring CVD graphene films to transparent and flexible polyethylene therephthalate (PET) substrates. Conductance (norm.) Bending cycles Figure S6. Change in conductance vs. number of bending cycles for a CVD graphene film on PET substrate. The conductance values have been normalized with respect to the conductance exhibited by the film before bending.

11 After CVD graphene transfer to PET substrates, Palladium (Pd) electrodes were deposited by e-beam evaporation through a shadow mask. Two terminal measurements were used to monitor the change in conductance of the flexible CVD graphene film with the number of bending cycles ( ). Figure S6 shows a plot of the normalized conductance vs. the number of bending cycles. It is seen that after 100 bending cycles, the conductance of the CVD graphene film decayed by only 7.9%. Thus, in other words, up to 92.1% of the conductance was retained. This result is outstandingly good for a flexible, conducting film with potential use as transparent electrode in OPV cells; especially when considering that the ITO film on PET lost electrical conductance after 1 bending cycle. It is important to note that cracks and fissures were developed on the Pd electrodes as a result of bending. Therefore, it is possible that the observed drop in conductance with bending for the CVD graphene film is in part due to the discontinuity of the metallic electrodes; and the decrease in conductance upon 100 bending cycles for the pristine CVD graphene film may be even lower than 7.9%. References 1. Ortiz-Conde, A., Sanchez, F. J. G. & Muci, J. Exact analytical solutions of the forward non-ideal diode equation with series and shunt parasitic resistances. Solid-State Electronics 44, (2000). 2. Jain, A., Sharma, S. & Kapoor, A. Solar cell array parameters using Lambert W-function. Solar Energy Materials and Solar Cells 90, (2006).

12 3. Jain, A. & Kapoor, A. A new method to determine the diode ideality factor of real solar cell using Lambert W-function. Solar Energy Materials and Solar Cells 85, (2005). 4. Jain, A. & Kapoor, A. Exact analytical solutions of the parameters of real solar cells using Lambert W-function. Solar Energy Materials and Solar Cells 81, (2004). 5. Banwell, T. C. & Jayakumar, A. Exact analytical solution for current flow through diode with series resistance. Electronics Letters 36, (2000). 6. Hayes, B. Why W? American Scientist 93, (2005).