Logo: University of Southern California


Graphene synthesis, Transfer and Applications

Graphene is probably the best theoretically studied allotropic form of carbon. It consists of a 2-D hexagonal arrangement of carbon atoms, with a quasi-linear dispersion relation, for which the carrier effective mass is very low. As a consequence, it has a predicted mobility at room temperatures of the order of 106 cm2/V•s and an experimentally measured mobility of 15,000 cm2/V•s. The high mobility of this material opens the possibility of ballistic transport at submicrometer scales. However, the advance of graphene-based nanoelectronics has been hampered due to the difficulty in producing single- or few-layer graphene (FLG) over large areas. Several methods have been used to obtain single-layer graphene or few-layer graphene but those methods inherently limit realistic applications of graphene in electronics due to high cost, impossibility of being scaled to wafer-size dimensions or by producing graphene with large amount of defects. In our group, we investigate synthesis approaches that can provide high-quality graphene at large scale. In addition we perform complete characterization and evaluation of its electronic and mechanical properties with the ultimate goal of fabricating functional devices that meet realistic applications. Furthermore, we demonstrated the application of CVD graphene in photovoltaic cells. Graphene has the potential to replace ITO photovoltaic cells due to its high flexibility and transparency.

CVD graphene synthesis and device appications

Figure 1. (a) Schematic of full-wafer scale deposition of graphene layers on polycrystalline nickel by CVD. (b) E-beam evaporated nickel film (100 nm) on a 4” Si/SiO2 wafer. (c) AFM image of Ni film before (top) and after (bottom) thermal annealing. (d) AFM image of Graphene layers deposited on polycrystalline Ni. Raman spectra on the right were taken on locations marked as A and B. (e) G’ band Lorentzian fits of the Raman spectra showed in (d) from monolayer (A) and bilayer (B) graphene.

We have reported the implementation of a simple, scalable and cost efficient method to prepare few-layer graphene films by the use of chemical vapour deposition (CVD) on polycrystalline Ni on wafer scale areas by using methane as carbon feedstock. We found that the number of graphene layers deposited on polycrystalline Ni can be greatly reduced to one, two and few layers by using a low-molecular-weight hydrocarbon molecules as carbon source. CH4 is easily pyrolized on polycrystalline nickel films at high temperatures to produce highly ordered domains with monolayer, bilayer and few-layer graphene. Etching of Ni allowed graphene layers to be deposited on Si/SiO2¬ wafers where back-gated field effect transistors could be fabricated at waferscale.

Figure 2. Wafer scale fabrication of few-layer graphene devices. (a) photo-image of a 4 inch wafer after graphene layers deposition on polycrystalline Ni. Inset of Figure 1(a) shows an optical micrograph of deposited few-layer graphene on Si/SiO2. (b) Complete 4 inch wafer with back-gated few-layer graphene devices. (c) IDS-VDS measurements and (d) IDS-VG curve of one of the devices shown on (b).

ISD-VSD characteristics depicted in Figure 2c shows that the drain current increases with the increase of negative source-drain voltage across the channel, indicating a weak p-type behavior in the films. Figure 2d shows the transfer characteristics of the FET devices fabricated from the synthesized FLG. Most devices were highly conductive and exhibited a weak modulation of the drain current by the gate bias, which is consistent with a 2D semimetal. A single graphene layer is a zero-gap semiconductor, but interlayer interactions bring in a semimetal behavior in FLG. Therefore, the transfer characteristics observed in Figure 2d can be attributed to a screened gating effect due to irregularities of the film and the presence of more than two graphene layers.


Comparison of Graphene Growth on Single-Crystalline and Polycrystalline Ni by Chemical Vapor Deposition

Recently, chemical vapor deposition (CVD) has raised its popularity in the synthesis of graphene as a scalable and cost effective approach. Polycrystalline Ni has been shown to be a good substrate for graphene synthesis by CVD, but the percentage of monolayer or bilayer graphene is limited by the grain size of crystalline Ni obtained after thermal annealing of Ni thin film.

We have reported the influence of the concentration of Ni interface boundaries on the formation of multilayer graphene domains. Synthesis of graphene by CVD on the (111) face of single crystal Ni favors the formation of highly uniform monolayer/bilayer graphene on the Ni surface, and simultaneously hinders the formation of multilayer graphene domains. Our results are understood on the basis of the diffusion-segregation model for carbon precipitation on Ni surface, where the uniform and grain-boundary-free surface of Ni (111) single crystal provides a smooth surface for uniform graphene formation. In contrast, the rough surface of polycrystalline Ni with abundant grain boundaries facilitates the formation of multilayer graphene.

Figure 3. Schematic diagrams of graphene growth mechanism on Ni (111) (a) and polycrystalline Ni surface (b). (c). Optical image of graphene/ Ni (111) surface after the CVD process. The inset is a three dimensional schematic diagram of a single graphene layer on Ni (111) surface. (d). Optical image of graphene/ polycrystalline Ni surface after the CVD process. The inset is a three dimensional schematic diagram of graphene layers on polycrystalline Ni surface. Multiple layers formed from the grain boundaries.

Micro-Raman surface mapping reveals that the area percentages of monolayer/bilayer graphene are 91.4% for the Ni (111) substrate and 72.8% for the polycrystalline Ni substrate under comparable CVD conditions.

Figure 4. (a). Maps of IG’/IG of 780 spectra collected on a 60*50 μm2 area on the Ni (111) surface and (b) 750 spectra collected on a 60*50 μm2 area on the polycrystalline Ni surface. Corresponding optical images to Ni (111) Raman map and polycrystalline Ni Raman map (c and d). (e) AFM image of graphene film transferred to SiO2/Si substrate from Ni (111). (d) Height analysis of the thickness of graphene film.


Large-Area Graphene Films by Chemical Vapor Deposition for Highly Flexible Organic Photovoltaics


In our recent work we explored the implementation of large area, nearly defect-free and highly smooth few-layer graphene films, synthesized from CVD, as the anode material in flexible and rigid OPV cells (Fig. 5a). The use of CVD graphene (CVD-G) is attractive because other graphene films, which are formed by stacked micron-size flakes, suffer from flake-to-flake contact resistance and high roughness. In contrast, grain boundaries of CVD-G films have the advantage of being formed in situ during synthesis; such a process is expected to minimize contact resistance between neighboring graphene domains and may result in smoother films with better conducting properties.

Figure 5. a, Schematic representation of the energy level alignment (upper) and construction of the heterojunction organic solar cell fabricated with graphene as anodic electrode: CVD-G / PEDOT / CuPc / C60 / BCP / Al. b, Schematic of the CVD-G transfer process onto transparent substrates. Photographs showing highly transparent graphene films transferred onto glass and PET are shown in c and d, respectively. e, Transmission spectra for CVD-G, ITO and SWNT films on glass. f, AFM images of the surface of CVD-G, ITO and SWNT films on glass. The scale bar in z-direction is 50 nm for all images. g, Transmission spectra of CVD-G with different sheet resistance (RSheet). h, Comparison of RSheet vs. light transmittance at 550 nm for CVD-G and reduced GO films reported in the literature.

Solar cells made with CVD-G exhibited performance that compares to ITO devices and surpasses that of ITO under bending conditions, with power conversion efficiencies of 1.18% when fabricated on flexible substrates.

Figure 6. Current density vs voltage characteristics of CVD-G (a) or ITO (b) photovoltaic cells under 100 mW/cm2 AM1.5 spectral illumination for different bending angles. Insets show the experimental set up employed in the experiments. c, Fill factor dependence of the bending angle for CVD-G and ITO devices. d, SEM images showing the surface structure of CVD-G (up) and ITO (down) photovoltaic cells after being subjected to the bending angles described in a and b.

Related Publications:

 1. “Comparison of Graphene Growth on Single-Crystalline and Polycrystalline Ni by Chemical Vapor Deposition”
Y. Zhang, L. Gomez, F. Ishikawa, A. Madaria, K. Ryu, C. Wang, A, Badmaev, and C. Zhou
Journal of Physical Chemistry Letters, 1, 3101–3107 (2010) (PDF)

2. “Continuous, Highly Flexible, and Transparent Graphene Films by Chemical Vapor Deposition for Organic Photovoltaics”,
L. Gomez De Arco, Y. Zhang, C. W. Schlenker, K. Ryu, M.E. Thompson, C. Zhou
ACS Nano, 4, 2865–2873 (2010). (PDF)

3. “Synthesis, Transfer, and Devices of Single- and Few-Layer Graphene by Chemical Vapor Deposition”,
L. Gomez, Y. Zhang, A. Kumar, and C. Zhou
IEEE Transactions on Nanotechnology, 8, 135 (2009). (PDF)