A single-walled carbon nanotube (SWNT) is a novel one-dimensional material possessing attractive electric, mechanical, and thermal properties. Driven by the potential applications of SWNTs, many methods have been proposed to synthesize SWNTs. Among these, alcohol catalytic chemical vapor deposition (ACCVD) can yield high-quality SWNTs at moderate temperatures. It is also the first method by which vertically aligned SWNT arrays were obtained. However, the incomplete understanding of the growth mechanism in this process, such as insufficient information about the catalyst status and position during growth, hinders the full control over the final product.

Isotope labeling is a powerful technique for identifying the reaction pathway in chemical reactions. By feeding of two types of isotope-labeled ethanol in sequence and therefore forming 12C-13C SWNT junctions, we confirmed the root growth mechanism of SWNTs synthesized by ACCVD. This result is consistent with TEM observation, where catalyst particles are found only in one end of an array. Clarification of the root growth model is critical in understanding the growth and catalyst deactivation mechanisms in ACCVD. The no-flow condition we introduced is also able to produce high-quality SWNT arrays effectively.

In root growth mode, the feedstock molecules have to diffuse through the thick CNT array, reach the substrate where catalysts are located, and then contribute to the CNT growth. In this bottom-up growth process, the diffusion resistance of the feedstock from the top to the root arises as an obstruction, and can act as a unique decelerating growth mechanism. Existence of a feedstock diffusion resistance means that concentration of the carbon source at the CNT root should be lower than the bulk concentration. We proposed a method of using a non-dimensional modulus to quantitatively evaluate the degree of feedstock diffusion resistance (no diffusion resistance regime, transient regime, and strong diffusion limit regime). Five of the most frequently used systems are also discussed. The results show that, for mm-scale SWNT arrays, the feedstock concentration at the root of the array is much lower than the bulk concentration, while for mm-scale MWNTs the decreasing growth can not be attributed to a diffusion limit, which agrees well with the currently available experimental results.

Recent investigations have also improved our understanding of SWNT growth behavior. We previously developed an in situ optical absorption measurement that allows for convenient real-time measurement of the film thickness. This technique revealed, for the first time, in situ growth kinetics of a VA-SWNT array. It provided sub-second resolution and much more direct information to the previously black-box approach to studying the growth process. We investigate the influence of various species on ACCVD synthesis of VA-SWNTs. A small amount of acetylene (approximately 1% partial pressure) was found to accelerate the growth rate by almost ten times, as revealed by a distinct change in the growth rate determined from in situ optical absorption. This accelerated growth, however, only occurred in the presence of ethanol, whereas pure acetylene at the same partial pressure resulted in negligible growth and quickly deactivated the catalyst. The dormant catalyst could be revived by reintroduction of ethanol, indicating that catalyst deactivation is divided into reversible and irreversible stages. Since the thermal decomposition of ethanol also yields some amount of acetylene, we calculated theoretically and also measure experimentally the concentration of different species. Contribution of such gases to the formation of SWNTs is quantified.

The diameter of a SWNT affects its mechanical, chemical and especially electronic or optical properties. SWNTs with smaller diameters (approximately 1 nm) have a larger band gap and are more mechanically stable than larger SWNTs (e.g. dia. >2 nm). Furthermore, they are also more easily measurable by spectroscopic methods such as resonance Raman spectroscopy and photoluminescence excitation. However, most vertically-aligned SWNT arrays produce SWNTs synthesized to date have the average diameter of ~ 2 nm, or even larger (~ 3 nm). The difficulty of diameter control lies not only in the synthesis but also the characterization. By extensive study on the CVD parameter and catalyst recipe, we show that the average diameter of vertically aligned SWNTs can be arbitrarily tuned between 1.4 and 2.5 nm. The average diameter was determined by optical absorption spectra as well as high-resolution TEM observations. We also show that the diameter of SWNTs along the growth direction is not uniform, as evidenced by carefully decomposed optical absorption spectra. TEM observations and Raman spectra obtained from different locations along the height of the array are consistent with the absorption results. These results provide insight into the SWNT growth mechanism and catalyst behavior on a flat substrate.

Localizing the growth of CNTs, therefore, has attracted much attention, since it is a critical step for the fabrication of on-substrate devices. The conventional way is to pattern catalyst by sputtering or evaporation of metal through a physical mask or a pre-exposed photoresist layer. This conventional MEMS technique is, although effective, normally complicated and expensive. By identifying the role of the surface wettability in the deposition of catalyst by dip-coating, we demonstrate a method realize site-controlled SWNT growth. Hydrophilic/hydrophobic patterns were fabricated on a silicon substrate by self-assembled monolayer (SAM) surface functionalizaiton. Catalyst is found deposited only in the regions where surface is highly hydrophilic. Therefore the growth is successfully localized. The all-liquid-based procedure allows us to bypass conventional complicated fabrications and yield controlled patterns consisting of high-quality as-grown SWNTs. More importantly, when electron beam was utilized to destroy the SAM, sub-10-nm resolution may be obtained by this method. The ability of using commercial scanning electron microscope also facilitate the fabrication of SWNT based devices.

本論文は"Controlled Growth of Vertically Aligned Single-Walled Carbon Nanotubes for Devices (デバイス応用に向けた垂直配向単層CNTの合成制御)"と題し,ナノテクノロジーの中心的素材として注目を集めている単層カーボンナノチューブ(single-walled carbon nanotubes, SWNTs)のデバイス応用に向けて,基板と垂直に配向したSWNTs膜のCVD合成制御を試みたものである.13C同位体エタノールを用いたCVD合成実験や理論による垂直配向膜の合成メカニズム解明,エタノールの熱分解生成物であるアセチレンによる合成速度増大の発見,触媒金属による直径分布の制御などの基礎的な知見を元に自己組織化膜(SAM膜)を用いたパターニングの提案と実証を行ったものであり,論文は全6章よりなっている.

第1章は,"Introduction(序論)"であり,CNTなどの炭素の同位体の幾何学構造,SWNTの幾何学構造と電子物性,光学物性,CVD合成技術,アルコールを炭素源とするCVD法(ACCVD法) ,垂直配向膜のCVD合成および応用について議論し,論文全体の流れを述べている.

第2章は,"Root growth mechanism and diffusion limit(根元成長機構と拡散律速)"である.垂直配向単層CNT膜のCVD中のレーザー吸収リアルタイム測定による膜厚のモニタリングおよびCVDの途中から炭素源のエタノールを13C同位体エタノールに換える同位体実験を行い,合成された13C垂直配向CNT膜の位置を顕微ラマン分光で特定することで,触媒金属が合成中つねに基板上にある根元成長機構であることを明らかとした.さらに,根元成長であると炭素源となるエタノール分子は成長しつつある垂直配向膜中を拡散後に触媒金属に達する必要がある.この拡散抵抗の程度について,簡単な一次元拡散モデルによって検討し,垂直配向単層CNTの場合には膜厚が数mmになってはじめて拡散抵抗律速となることを明らかとした.

第3章は,"Growth acceleration and ethanol decomposition(合成速度増大とエタノールの熱分解)"である.炭素源であるエタノールの気相における熱分解反応によってエチレン,水や微量のアセチレンが生成する過程をFT-IRによる実測と化学反応モデル計算によって明らかとした.また,微量のアセチレンをエタノールに加えることで単層CNT膜のCVD合成速度を飛躍的に増大させられることを明らかとした.これらの実験結果をもとに,炭素源のエタノールの一部はアセチレンまで熱分解したのちに単層CNT合成に至ることを明らかとした.特に,低流量のCVDにおいては,この熱分解を経たCNT合成過程が優勢となり合成速度が増大する一方CNTの品質は劣化することを明らかとした.

第4章は,"Diameter Control of VA-SWNTs(垂直配向単層CNTの直径制御)"である.光吸収分光やラマン分光によって,垂直配向単層CNTの直径分布が成長に伴って若干変化し,合成の後半では直径が大きくなる傾向を明らかとした.直径分布に対するエタノール圧力やCVD温度の影響は比較的小さいが,デップコート法における仕込みの触媒金属濃度を変化させることで,直径分布を大きく変化させることができることを明らかとしている.

第5章は,"High-precision patterned growth(高解像度パターン合成)"であり,様々なデバイス応用に向けて,パターン合成法を議論している.最初に,従来から報告のある方法を発展させ,シリコン酸化膜のパターンをMEMS技術で合成し,単層CNTの垂直配向膜が酸化膜部分から合成されることを示した.さらに,デップコート法による触媒担持特性が基板の濡れ性に強く影響されることに着目し,疎水性の自己組織化膜(SAM膜)を基板上に合成し,その一部をUV照射あるいは電子線照射で除去し,その部分にデップコート法で触媒を担持して,単層CNTを合成する方法を提案し実証した.SAM膜を用いることで極めて高解像度のパターン合成が実現し,将来のデバイス応用の可能性を示した.

第6章は,"Conclusion and prospect(結論と今後の展開)"であり,上記の研究結果をまとめたものである.

以上を要するに,本論文では垂直配向単層CNT膜の合成機構の解明を進め,微量分子添加による合成速度増大や直径分布の制御を実現するとともに,自己組織化膜(SAM膜)を用いたパターニングの提案と実証を行ったものである.本論文はSWNTsの合成機構とCVD合成の制御に関する新たな知見を与えており,分子熱工学の発展に寄与するものであると考えられる.

よって本論文は博士(工学)の学位請求論文として合格と認められる。