MATTSON ET AL. VOL. 5 ��� NO. 12 ��� 9710���9717 ��� 2011 www.acsnano.org 9710 November 21, 2011 C 2011 American Chemical Society Evidence of Nanocrystalline Semiconducting Graphene Monoxide during Thermal Reduction of Graphene Oxide in Vacuum Eric C. Mattson,��� Haihui Pu,��� Shumao Cui,��� Marvin A. Schofield,��� Sonny Rhim,��� Ganhua Lu,��� Michael J. Nasse,���,�� Rodney S. Ruoff, ) Michael Weinert,��� Marija Gajdardziska-Josifovska,��� Junhong Chen,*,��� and Carol J. Hirschmugl*,��� ���Department of Physics and Laboratory for Surface Studies, University of Wisconsin Milwaukee, Milwaukee, Wisconsin 53211, United States, ���Department of Mechanical Engineering and Laboratory for Surface Studies, University of Wisconsin Milwaukee, Milwaukee, Wisconsin 53211, United States, ��Synchrotron Radiation Center, Stoughton, Wisconsin 53589, United States, and)Department of Mechanical Engineering and the Texas Materials Institute, University of Texas at Austin, Austin, Texas 78712, United States Gnologiesfor raphene has demonstrated great po- tential novel electronic tech- 1 3 however, many emerg- ing applications require atomically thin films with a substantial band gap. One route to- wardthemanipulation oftheelectronic prop- erties of graphene-based materials is through chemical modification. Chemically modified graphene or graphene oxide (G-O)4 8 has already found various applications such as supercapacitors,9 sensors,10,11 and flexible transparent conductive electrodes.12 However, G-O is a nonstoichiometric, partially disordered material13,14 with poor electronic properties. While studies have indicated that on average the hexagonal lattice constant of graphene is preserved in G-O,13 15 amorphous regions coexist with the crystalline carbon lattice, with a number of oxygen functional groups, such as hydroxyl, carboxyl, carbo- nyl, epoxide, and intercalated water,15 17 bonded with no long-range order.13 These disordered qualities of G-O make it an un- appealing candidate for application in modern electronics. Thus, G-O is most commonly modified by various chemical6 and thermal reduction treatments15 17 with the goal of removing oxygen functional groups from G-O and producing graphene. While the quality of the resulting materials can in general be improved via chemical vapor deposition (CVD),18 these treatments generally fail to form graphene and pro- duce yet another nonstoichiometric par- tially reduced GO where oxygen remains disordered. We demonstrate the formation of a stoi- chiometric carbon oxide by vacuum an- nealing of multilayered G-O, a method com- monly used to make thermally reduced graphene oxide (TRG-O). Studies of TRG-O, however, have primarily focused on reduc- tion of monolayer or few layer systems,15 17 with little attention paid to thermal reduc- tion of systems with a large number of layers. While previous reports17 have shown * Address correspondence to cjhirsch@uwm.edu, jhchen@uwm.edu. Received for review August 21, 2011 and accepted November 20, 2011. Published online 10.1021/nn203160n ABSTRACT As silicon-based electronics are reaching the nanosize limits of the semiconductor roadmap, carbon-based nanoelectronics has become a rapidly growing field, with great interest in tuning the properties of carbon-based materials. Chemical functionalization is a proposed route, but syntheses of graphene oxide (G-O) produce disordered, nonstoichiometric materials with poor electronic properties. We report synthesis of an ordered, stoichiometric, solid-state carbon oxide that has never been observed in nature and coexists with graphene. Formation of this material, graphene monoxide (GMO), is achieved by annealing multilayered G-O. Our results indicate that the resulting thermally reduced G-O (TRG-O) consists of a two-dimensional nanocrystalline phase segregation: unoxidized graphitic regions are separated from highly oxidized regions of GMO. GMO has a quasi-hexagonal unit cell, an unusually high 1:1 O:C ratio, and a calculated direct band gap of ���0.9 eV. KEYWORDS: graphene oxide . thermal reduction . in situ electron diffraction . infrared spectroscopy . density functional theory . nanocrystals . semiconductors ARTICLE

MATTSON ET AL. VOL. 5 ��� NO. 12 ��� 9710���9717 ��� 2011 www.acsnano.org 9711 that the chemistry of intercalated water in multilayer G-O is significant, no investigations have reported the effect of vacuum annealing a large number of G-O layers to high temperature. We have found that the multilayer structure of G-O thin films with intercalated water results in a previously unobserved atomic struc- ture and morphology when annealed in vacuum: a two-dimensional phase segregation produces graphi- tic regions with little or no oxidation that coexist with oxidized regions with an unusually high oxygen con- tent. Our work, combining in situ selected area electron diffraction (SAED) studies with synchrotron-based in- frared microspectroscopy (IRMS) and density func- tional theory (DFT) calculations, reveals the atomic structure of the ordered oxidized regions coexisting with unoxidized graphene-like regions. From DFT modeling guided by experimentally derived structu- ral data, we determined that the resulting oxidized structure, Graphene Monoxide (GMO), consists of a quasi-hexagonal unit cell with two carbon atoms bridged by a double-epoxide pair. GMO formation occurs because the large number of layers effectively limits the reduction process, and conversion of the remaining oxygen functional groups to a double- epoxide-type configuration is energetically favorable. Furthermore, the resulting material demonstrates ap- pealing transport properties, which could potentially be tuned owing to GMO's direct gap of 0.9 eV. RESULTS AND DISCUSSION Figure 1a compares SAED patterns of a multilayer G-O film before (left) and during (right) in situ vacuum annealing, at 750 ��C. Before annealing, the primary features evident in Figure 1a are the diffraction rings (labeled as I and II) from spacings of 0.213 and 0.123 nm, respectively, corresponding to the {100}- and {110}- type reflections of graphene. A ring pattern is observed rather than a spot pattern due to the fact that the sample consists of a large number of randomly oriented layers (see Methods and Figure S1). Analysis of the relative intensities shows that the {100}-type reflec- tions produce a greater diffracted intensity than the {110} reflections, indicating that the layers are mono- layers with disordered stacking, as opposed to the few- layer Bernal-stacked graphite oxide.14 Moreover, SAED patterns (Figure S2, Table S1) recorded at higher scat- tering angles (i.e., smaller lattice spacing) clearly indi- cate weaker higher order rings also consistent with crystalline graphene. In addition, SAED patterns re- corded before annealing (Figure 1a, left) show two broad amorphous rings centered at about 0.27 and 0.52 �� 1 (0.370 and 0.185 nm in real space). The amorphous rings are attributable to the first- and second-order reflections from nearest-neighbor disor- dered species. While annealing the multilayer G-O film, two promi- nent rings (labeled as III and IV) develop corresponding to spacings of about 0.260 and 0.152 nm, respectively, while the graphene rings (I and II) remain largely unchanged (Figure 1a). Thus the SAED data demon- strate that a new crystalline phase develops upon an- nealing. A visualization of the complete temperature- dependent evolution of SAED patterns from the G-O film annealing (extracted from a real time movie see Supple- mentary movie) is shown in Figure 1b. Diffraction rings in a conventional SAED pattern (Figure 1a) appear as horizontal lines in the representation of Figure 1b. Figure 1. Evolution of Electron Diffraction patterns with temperature. (a) SAED patterns of G-O sample before (left) and after (right) vacuum reduction anneal. Overlay is the radial average of intensity showing peaks at ring positions I and II before annealing, and the addition of peaks III and IV after annealing. (b) Radially averaged profiles from SAED patterns as a function of temperature. Profiles displayed as overlay in top panel are equiva- lent to profiles along the corresponding dashed lines (arrowed) in middle panel of figure. Peaks III and IV evolve continuously from the broad background, most notably in the temperature range from 500 to 700 ��C. (c) Temperature-dependent evolution of unwrapped SAED patterns. As temperature is increased, new peaks appear with indicated spacings and grow in intensity. Peaks I and II are attributed to graphene regions of the sample, while peaks III and IV are attributed to regions of nanocrystalline graphene monoxide. ARTICLE