Carbon has various allotropes including diamond and graphite, which exhibit quite different physical properties and chemical reactivity. For instance, graphite shows 1015 times larger electrical conductivity than diamond due to the existence of delocalized π electrons. Thus, it is expected that the electronic properties of carbon materials can be modified by controlling the ratio of sp2/sp3, and doping of atoms with different electronegativity into carbon is one route to realize sp2/sp3 control. Introduction of oxygen into graphitic carbon, consisting of highly and/or purely sp2 hybridized bonds is a promising pathway, because oxygen can react with carbon to form stable carbon-oxygen bonds even in the ambient atmosphere.

So far, oxidized graphene has been synthesized mainly by a wet process developed by Hummers and Offeman (the so called as Hummers method). However, the Hummers method has been done under strong oxidation conditions, so that various functional groups such as epoxy, carboxyl, hydroxyl and carboxyl groups are introduced in a non-uniform manner.

In this study, formation of a carbon-oxygen bond and/or oxygen-containing groups has been examined using two different oxidation methods. One is solid-phase oxidation on oxide substrates, in which oxygen ions supplied from substrate react with carbon. The other is a novel oxidizing technique based on a solid-gas reaction using a reactive oxygen gas species. For the latter process, two atomic oxygen sources were utilized and one of them yields mildly oxidized graphene, which has a smaller variety of oxygen-containing groups and less disordering of carbon atoms in the graphene hexagonal lattice.

Solid-phase oxidation of carbon thin films on oxide substrates by annealing at higher temperatures

was investigated in order to dope oxygen atoms from the substrate to the film. 10 nm-thick carbon films were deposited on refractory oxide substrates such as yttria stabilized zirconia (YSZ), magnesium oxide and sapphire using the pulsed laser deposition technique (λ = 266 nm) in a vacuum chamber, which was followed by ex-situ annealing at higher temperatures, ranging from 1500 to 2700 °C under high vacuum and argon atmosphere conditions, respectively.

As-deposited carbon films showed a significant contribution from sp3 according to the Raman spectrum (Fig. 1a). Successive post-annealing at lower temperatures below 1500 °C caused graphitization (conversion from sp3 to sp2), and nano-graphitic carbon films were formed (Fig. 1b). At higher annealing temperatures, a reaction occurs at the interface between the carbon film and oxide substrate, resulting in the exfoliation of the film from the substrate. This indicates that ionic oxygen continuously supplied from the substrate during the annealing is highly reactive and removes carbon atoms at the interface as gaseous CO/CO2.

Graphene grown on metal surface by chemical vapor deposition

High quality graphene films with well-regulated numbers of layers were fabricated using the chemical vapor deposition (CVD) technique. As is well known, thin graphene layers are grown by decomposition of hydrocarbon gas and/or segregation of carbon at the surfaces of transition metals. In this study, copper was chosen as a substrate to obtain mono-/bi-layer graphene at the surface. Hydrogen and methane gases were introduced to a quartz tube as cleaning and source gasses, respectively. The process temperature, pressure and time were typically 1000 °C, 20 Pa, and 1-10 min, respectively. Characterization of graphene was carried out with Raman spectroscopy, scanning electron microscopy (SEM) and scanning tunneling microscopy (STM).

A Raman spectrum (excitation wavelength λ = 514, 532 and 633 nm) (Fig. 2) of the obtained film showed a sharp G peak at 1587 cm-1 and 2D peak at 2681 cm-1. The full width at half maximum (FWHM) of the 2D peak is 48 cm-1, indicating that mono- and/or bilayer graphene was obtained. The value of the 2D peak is fairly wider than those previously reported, 30 cm-1. This broadening may be due to orbital hybridization between graphene and the underlying copper or surface plasmon resonance of copper, which would modify the band structure of the graphene film. Even in the film deposited for the longer time of 30 min, the Raman spectrum did not essentially change, implying that the number of layers is limited to two irrespective of the deposition time.

Fig. 3(a) is an STM image taken on the graphene film, clearly resolving both an atomic honeycomb lattice and Moire patterns with lateral size over 20 nm. From the observation of Moire patterns, it can be concluded that monolayer graphene was grown in this region. Such the hexagonal lattice and Moire patterns were reproducibly observed, indicating that monolayer graphene covers most of the films obtained here. SEM observations revealed that the graphene film is of polycrystalline form with grain sizes of sub-microns (Fig. 4). Thus, the density of grains can be estimated to be several per micrometer square. As a consequence, there are many grain boundaries, at which defects or edges, such as 5-/7-membered aromatic rings missing carbon atoms in the honeycomb lattice, might exist.

Atomic oxygen exposure to graphene for gas-phase oxidation

Gas-phase oxidation of graphene has been investigated in an ultra-high vacuum (UHV) chamber by irradiation of atomic oxygen to graphene grown on a copper surface. Two atomic oxygen sources were tried in an effort to confirm correlation in the dissociation ratio of oxygen species. The irradiation time was typically 5-15 min. During the irradiation of oxygen species, the graphene sample was not heated, in order to prevent from carbon etching. The Raman spectra of the oxidized graphene exhibited a sharp D peak centered at around 1350 cm-1 (Fig. 5). This D peak in the spectra was relatively narrower than that of unreduced graphene oxide prepared using the Hummers method (FWHM: typically around 100 cm-1) although there was little change in the G and 2D peaks.

Infrared spectroscopic and X-ray photoelectron spectroscopic studies revealed that this oxidation method allowed us to introduce oxygen atoms including both ether and carbonyl groups onto graphene. Notably, STM images of the presently prepared oxidized graphene showed a clear sp2-based honeycomb lattice similar to those of pristine graphene before oxidization. Even after repeating the irradiation process, there was no significant difference in nanoscale topographic images. By combining all the results discussed above, it is likely that the oxidation occurs mainly at the grain boundary regions, because carbon atoms near edges or defects at the boundaries seem to be more reactive than those in the honeycomb lattice.

Summary

In this work, solid-phase and gas-phase oxidation of graphitic carbon thin films including graphene layers were investigated, followed by synthesis of these thin films. Post-annealing of graphite on transition metal oxide substrates exhibited the removal of carbon located at the film/substrate interface as gaseous CO/CO2. Irradiation of oxygen to the graphene surface was found to be a moderate oxidation process and formation of carbon-oxygen bond was confirmed. The oxidized graphene films showed the shaper D peak in Raman spectroscopy, suggesting that they contain a smaller variety of the oxygen-containing groups or defects compared with those prepared using the Hummers method.

Fig. 1 Raman spectra from the sample (a) before and (b) after post-annealing (λexc = 532 nm). (c) A topographic image of the carbon film after post-annealing taken with AFM (20×20×1 μm).

Fig. 2 A Raman spectrum of as-grown pristine CVD-grown graphene (λ = 514 nm). FWHM(2D) = 48 cm-1. The dotted curves at bottom are curve-fitted peaks of D, G and 2D peaks. The intensity (area) ratio I(G)/I(D) is 13.4.

Fig. 3 Images of monolayer graphene grown on copper substrate taken by UHV-STM at 4.7 K: (a) topographic image and (b) auto-correlated image of (a). (Vbias = - 0.2 V, Itunneling = 2 nA.)

Fig. 4 An SEM image of as-grown graphene. Sub-micrometer bright regions correspond to sub-monolayer graphene islands.

Fig. 5 Raman spectra of graphene after atomic oxygen exposed ((a, b, c) three spectra from three different points within the same sample) and (d) before exposed. (λexc = 532 nm, laser spot size: 1 μm) (FWHM(2D) = 37 cm-1)

Fig. 6 Reflection infrared Fourier-transform spectrum of graphene oxidized by microwave-assisted atomic oxygen.

Fig. 7 Core-level X-ray photoelectron spectra of (a) C 1s and (b) O 1s. Blue and red curves correspond to those for graphene before and after oxidation. All spectra were collected at the same photon flux.

酸素原子が導入されたカーボン系薄膜の構造についてはこれまで諸説あり、現在も議論が続いている。本論文では、カーボン薄膜の表面・界面における酸素原子との反応、および反応の結果得られる薄膜の構造について議論している。これまで主に用いられてきた溶液プロセスによる合成法とは異なり、固体平坦表面上で酸素原子を導入する手法について検討している。

本論文は以下の5章より構成されている。

第1章は序論であり、本論文の背景および目的が述べられている。本研究で用いられた気相・固相酸化手法の優位性を、既存の手法との対比を通じて論じている。既存の手法で得られた酸化グラフェンでは、酸素含有官能基が多様であるため、カーボンのハニカム構造が乱れ、キャリア移動度が相対的に低く留まっていると述べている。このキャリア移動度の低さが、電子デバイス・光エレクトロニクスへの応用上の障害であると指摘している。以上のような背景のもとで、グラフェンシート構造を保ったまま酸素を導入することを目的の一つとした上で、新たなグラファイト状カーボン薄膜の酸化法を探究するとしている。

第2章は、類似物質の研究事例の比較と分析手法の説明である。カーボンナノチューブ、グラフェン、酸化グラフェンなどの関連物質について、これまで明らかになっている基本的な性質をまとめるとともに、これらの物質に酸素を導入する研究について例示している。また、本研究で用いた分析手法について、ラマン分光法、X線光電子分光法(XPS)、走査型プローブ顕微鏡の原理とそれらから得られる情報について説明している。

第3章では、グラファイト状カーボン薄膜の固相酸化について報告している。冒頭で、本手法の新規性について、過去の研究事例と比較しながら述べている。実験手順として、高融点金属酸化物単結晶基板上に厚さ10 nmのカーボン薄膜を真空蒸着し、1500-2700 °Cの高温領域での加熱を行なっている。加熱処理により、カーボンのナノスケールでの結晶化が進むと同時に、1 μmを超える巨大な凹凸構造が生じることを見出している。電子線衝撃および近赤外レーザー照射の2種類の加熱法により固有の影響がないことを確認し、上記結果が高温加熱によるものであると述べている。凹凸構造が生成するメカニズムについて、幾つかのモデルを比較検討した上で、高温加熱顕微鏡での観察結果と併せ、界面での反応による気体成分発生により凹凸構造が生成したと結論している。また、ラマン分光測定の解析から見積もられるsp2ドメインサイズが類似の加熱実験のものと比べて小さいことから、グラファイト結晶化が界面での酸化反応により抑制されたと推論している。

第4章では、グラフェン薄膜の気相酸化について述べられている。前半部分では、銅多結晶箔上へのメタンの化学的気相蒸着(CVD)法によりグラフェン薄膜を合成し、ラマン分光および走査型トンネル電子顕微鏡観察により単層グラフェンを得たことを確認している。試料非加熱(室温)でのグラフェンシートへの酸素導入については反応性の高い酸素原子が有効であるとし、熱フィラメントおよびマイクロ波プラズマにより解離した原子状酸素を用いている(熱解離法およびマイクロ波解離法)。これら2種類の解離法での解離率およびプロセス圧力の違いについて述べた上で、それぞれの手法を実験により比較検討している。XPSおよびラマン分光により、炭素-酸素間の結合、グラフェンシートの結晶性についてそれぞれ評価を行なっている。XPS C 1sスペクトルのピーク分離とC 1s/O 1sスペクトル強度比から、熱解離法では酸素が導入されない一方で、マイクロ波プラズマ酸化法により酸素原子が導入されることを明らかにしている。他方、両者のラマンスペクトルではいずれも欠陥由来の強いピークが確認され、グラフェンシートに欠陥が導入されていると推論している。ピーク分離したXPS C 1sスペクトルとフーリエ変換赤外分光スペクトルから、C-O-C(エーテルおよびエポキシ基)ならびにC=O(カルボニル基)が同量程度生成されたと結論している。この結果は、5d金属上のグラフェンの酸化とは異なっているが、その原因は、下地基板原子種の違いによるグラフェン-基板相互作用の大きさの違いによるものと説明している。

第5章は本論文で示された結論と総括である。

以上のように、本研究は、気相および固相酸素原子との反応によりグラファイト状カーボン薄膜に酸素導入を試みた上で、複数の分光法および表面分析法を用いて、酸化後の薄膜構造を明らかにしたものである。これらの研究は理学の発展に大きく寄与する成果であり、博士(理学)に値する。なお本論文は複数の研究者との共同研究であるが、論文提出者が主体となって行ったものであり、論文提出者の寄与は十分であると判断する。

したがって、博士(理学)の学位を授与できると認める。