[Introduction]

Organic semiconductors are attracting much attention for their potential application for flexible, low-cost and printable devices. The carrier transport mechanism in organic materials, though, is not thoroughly understood while some report that polaron-band mechanism is the most probable in case of crystalline low-molecular-weight materials.

The most widely used method to elucidate electronic structure of solids is angle-resolved photoemission spectroscopy (ARPES). The ARPES gives us powerful insight into the carrier transport properties, including effective mass of carriers, inter-molecular transfer integrals, and charge-phonon coupling constant. Hence, band dispersion measurement is essential for studying carrier transport mechanism as well as establishing molecular design guidelines for novel materials. The ARPES study of organic semiconductors, however, is quite challenging and only performed in limited systems due to following reasons. First, conductivity of organic materials is very low without doping. Therefore, positive charges generated during photoemission accumulate on the surface, and make the spectrum shifted toward high binding energy direction. Second, since photoemission method is surface sensitive in general, the surface for ARPES studies should be clean, free from absorbates or oxidized molecules. In order to satisfy this requirement, ARPES studies are generally performed with samples grown in-situ. Finally, as it is an "Angle-resolved" method, the sample should be single crystals.

In this research, I will demonstrate two novel methods to prepare well-defined organic surfaces which are suitable for surface sensitive characterization methods like ARPES. The first approach is to grow epitaxial film of organic molecules on conductive substrates with 1 monolayer thickness and control in-plane orientation to satisfy angle resolution. The second one is to prepare organic single crystals ex-situ, introduce it into the vacuum chamber, clean the surface, and then neutralize accumulated charges using electron shower.

[In-plane orientation control of organic epitaxial films on conductive substrates]

The requirement for the substrate is as follows. First, it should be conductive substrate as mentioned above. Second, the lattice mismatch between organic crystal and substrate need to be small. Finally, since the ππ overlap is essential for high performance of organic semiconductors, the molecules should be in "standing" orientation with its molecular plane perpendicular to the surface. This requires rather inert surface without strong interaction with molecular πorbital. The epitaxial growth of pentacene on Au or Cu single crystals, for instance, often results in "lying" orientation, in which the molecular long axis is parallel to the surface.

As for pentacene, which is a benchmark material of organic semiconductors, epitaxial growth of "standing" molecules was reported by us on Bi terminated Si(111)-(√3×√3) reconstructed surface. The lattice constant of reconstructed surface is 6.65 A, which is close enough to those of pentacene (a=5.93 A, b=7.59 A). However, since the substrate has three-fold symmetry and pentacene crystals have triclinic lattice, there are 6 possible in-plane orientations for pentacene. This does not satisfy angle resolution. In order to reduce surface symmetry, we used vicinally cut substrates instead of flat ones. By bunching steps and producing periodic steps as high as 3nm, we successfully controlled in-plane orientation of pentacene, reducing 6 orientations to 2. The pentacene monolayer film obtained in this method had thin-film phase structure, and band dispersion was studied. In this research, I demonstrate that orientation control using bunched steps is also applicable to other organic materials.

Experimental

The substrate is Si(111) with miscut toward [11 2 ] direction. The sample was introduced into vacuum chamber (base pressure ~3x10(-10) Torr) equipped with Knudsen cell, pyrometer, and Reflective High Energy Electron Diffraction (RHEED). First, the sample was flashed at 1260 °C for 20 s in order to obtain clean 1 x 1 surface. As it was cooled down, the 1 x 1 turned into 7 x 7 reconstruction at 860 °C, and confirmed by RHEED. From 960 °C to 860 °C was cooled gradually, at the rate of 6 °C/min, which gives smooth surface and bunched steps with ~3nm height. All of processes above were performed under < 2 x 10(-9)Torrconditions. One monolayer bismuth was then deposited from Knudsen cell at room temperature, and 1 x 1 image was confirmed.

In order to turn 1 x 1 Bi into √3×√3 reconstructed surface, the sample was annealed after Bi deposition. It is known that depending on the annealing temperature, Bi coverage changes in the same √3×√3 manner. When annealed at less than 400 °C, Bi coverage remains 1 ML and called β-phase. On the other hand, annealing at 450 °C and 590 °C lead to 0.66 and 1/3 ML, respectively. 1/3ML thickness Bi-Si is called a-phase. Since both a-phase and β-phase have √3×√3 reconstructed surface, it is difficult to assign them using RHEED patterns. Therefore, the coverage of Bi was estimated using ex-situ X-ray Photoemission Spectroscopy (XPS) by comparing intensity ratio of Bi 4f(7/2) and Si2p peaks.

Results and discussion

First, I report the difference in surface energy depending on surface reconstruction manners. Fig. 2 shows DPh-BTBT crystals grown on a-phase and β-phase Bi-Si(111)-(√3×√3). On a-phase, DPh-BTBT crystal grew in two-dimensional mode, resulting in monolayer films. When grown on β-phase, though, needle-like, three dimensional crystals were obtained. The same tendency was also observed in DNTT crystals. This result indicates that the surface energy of β-phase is smaller than that of a-phase.

Fig. 3 is the optical microscope image of DPh-BTBT needle-like crystal grown on vicinal substrate. When grown on flat Si(111) substrate, needle-like crystals grew in 6 directions. Even though the needle-like crystals are polycrystalline, molecular layers in the vicinity of substrates seem to be affected by lattice periodicity. In order to control in-plane orientation of the crystal, the DPh-BTBT was deposited onto vicinal Si(111) substrate terminated by Bi. It was found that even when DPh-BTBT form needle-like crystals, which are larger than steps on surfaces, in-plane orientation was controlled. The step height necessary to control orientation was found to be 4.4nm. Interestingly, the direction of controlled crystal growth was perpendicular to the step edges. It means that crystals growing over steps are selected. This result indicates that steps do not work as obstacles of crystal growth, but they stabilize crystal growth or nucleus formation in certain direction.

On the other hand, the orientation of monolayer films on a-phase Bi-Si required steps more than 10nm to be controlled. Figure 4 shows the Transverse shear microscopy (TSM) images of DPh-BTBT monolayer grown on stepped surfaces. The contrast in TSM image reflects the in-plane orientation of each grain. Figure 4 (b) shows clear contrast in and between grains while (d) shows no contrast. Moreover, Reflective High Energy Electron Diffraction (RHEED) analysis indicates that in-plane orientation was successfully limited down to at least two directions.

[Development of surface cleaning technique for organic single crystals]

The second approach is to clean the surface of ex-situ grown single crystals and neutralize positive charges. First, I will report the development of surface cleaning technique for organic crystals. Since many low-molecular-weight organic semiconductors have layered crystal structure, it was found that cleavage can be employed to obtain clean surface. However, cleaving crystal often require thick crystals. In our research, only crystals thicker than several hundred μm were cleaved successfully. These thick crystals, though, often have low conductivity and poor crystallinity. Therefore, we have to develop a novel surface cleaning technique suitable for crystals thinner than 1 μm.was found that cleavage can be employed to obtain clean surface. However, cleaving crystal often require thick crystals. In our research, only crystals thicker than several hundred μm were cleaved successfully. These thick crystals, though, often have low conductivity and poor crystallinity. Therefore, we have to develop a novel surface cleaning technique suitable for crystals thinner than 1 μm.

Experimental

The single crystal of tetracene was grown by Physical Vapor Transport (PVT) method, in which tetracene was sublimated inside quatz tube with two temperature zones and Ar gas flow. The tetracene single crystal, with typical dimension of 0.2×2 ×0.001mm3 was laminated onto Au (17nm) / Ti (3nm) / SiO2 substrate. The surface morphology was investigated using RHEED equipped with micro channel plate on the screen (MCP-RHEED).

Results and Discussion

Typical terrace width of tetracene crystals were around several μm, and molecularly flat surface was obtained. The MCP-RHEED image of the surface, though, indicate that the top-most surface do not have crystalline order. Only when the incident angle of electrons was more than 8 °, which means that proving depth was large, streaky patterns were obtained.

First, I show the effect of vacuum annealing on the disordered top-most surface of tetracene crystals. Fig. 6 show the surface of tetracene single crystal annealed in the vacuum of ~10(-8) Torr. It was found that vacuum annealing make the surface roughening, probably due to random molecule desorption from the surface. The MCP-RHEED image of vacuum annealed sample was spotty (Fig. 6 (a)), indicating that the surface hasnm size bumps and holes.

On the other hand, when annealed under atmospheric pressure argon, layer-by-layer sublimation was achieved. The schematic image of experimental setup is shown in Fig. 7 (e). It was observed that the tetracene molecules sublimated from teracces with layer-by-layer mode, without random desorption and roughening. Moreover, the MCP-RHEED image of tetracene surface after blowing hot Ar gas was significantly improved and streaky image was recovered even when incidence angle was sufficiently low. This result suggests that contamination and disordered layer at top-most surface was successfully removed.

[Conclusion]

In this research, I have demonstrated two novel methods to prepare well-defined samples suitable for surface sensitive characterization methods like ARPES of organic semiconductors. For the first approach, in which singly-oriented epitaxial film of organic molecules were grown on conductive substrate, it was found that the coverage of Bi greatly affects surface energy. The role of steps in orientation control was also studied, and it was assumed that periodic steps stabilize nucleus formation. In the second approach, the surface cleaning technique for organic crystals was developed. These methods will be widely used to prepare clean and highly crystalline organic surface in-situ, and promote researches on fundamental transport mechanism in organic functionalized materials.

Fig. 1 The molecular structure of (a) pentacene , (b) tetracene, and (c) 2,7 -Diphenyl[1]benzothieno[3,2-b][1]benzot hiophene (DPh-BTBT).

Fig. 2 (a) Schematic illustration of the Bi/Si(111)-(√3×√3) surface structures. (b) DPh-BTBT monolayer film grown on a-phase (1/3 ML). The cross section profile is shown in (d). (c) DPh-BTBT grown on β-phase (1 ML). Unlike a-phase, needle-like crystals were obtained.

Fig. 3 The optical microscope images of (a) DPh-BTBT crystals grown of flat β-phase Bi-Si(111) surface, and (b) vicinally cut β-phase Bi-Si(111).

Fig. 4 The AFM and TSM images of DPh-BTBT films grown on (a, b) 4nm steps and (c, d) 10nm steps. The grains in the same contrast is indicated in (b) and (d).

Fig. 5 MCP-RHEED patterns of tetracene single crystal grown ex-situ. The incidence angle of electrons is ~8°.

Fig. 6 (a) MCP-RHEED pattern and (b) AFM image of tetracene single crystal after vacuum annealing.

Fig. 7 (a-c) The AFM images of tetracene single crystal annealed in atmospheric pressure. (d) The schematic image of hot Ar annealing. (e) The MCP-RHEED pattern after annealing in atmospheric pressure Ar for 5 minutes. Streaky pattern was recovered.

有機半導体の基礎物性に関しては、未だ論争が続いている点が多く残っており、角度分解光電子分光(ARPES)による電子状態の解明が望まれている。ARPESなどの分光手法においては、結晶配向、表面清浄性などの点で、構造が良く規定された試料を準備することが必要である。本研究では、エピタキシャル薄膜の配向制御とex-situで成長させた単結晶の表面清浄化という2つのアプローチから、角度分解測定に十分な単一配向性と、表面敏感な測定に適した清浄な表面を持つ有機結晶表面を作製する手法を開発している。

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。この章ではまず、有機半導体の歴史と、伝導機構に関する研究を概観し、現在論争の的となっている格子振動と電子状態のカップリングの取り扱い方を議論している。また、このような議論には、ARPES測定が決定的な情報を与えうることを指摘している。つづいて、有機半導体のARPES測定の例を紹介し、その多くがπ電子の重なり合いを持たない"寝た"分子配向試料に関するものであること、ならびに結晶構造や得られたバンド分散の妥当性に関して議論があることを指摘している。一方で、シリコン表面の周期的なステップ構造を利用したペンタセンエピタキシャル膜の面内配向制御について紹介し、この手法を他の分子に発展させ、配向制御の機構を解明することを目的の1つに据えている。

第2章は、実験手法と、各種手法の原理の説明である。本研究で用いた超高真空装置の構成と機器を解説し、評価手法である水晶振動子膜厚計(QCM)、反射高速電子線回折(RHEED)、Transverse Shear Microscopy(TSM)、微小角入射X線回折(GIXD)、光電子分光法の原理と、そこから得られる情報について詳説している。

第3章は2,7-Diphenyl[1]benzothieno[3,2-b][1]benzothiophene (DPh-BTBT)エピタキシャル単分子膜の、面内配向制御に関して述べている。まず、背景として、目的物質であるDPh-BTBTとその誘導体の物性と、有機-無機ヘテロエピタキシャル成長、Si(111)面の(7×7)再構成とステップバンチング、Bi終端化面の(√3×√3)再構成について説明している。さらに、実験手順として、シリコン基板傾斜面の研磨法や表面清浄化法について述べている。次に、様々な基板の表面エネルギーを考察したうえで、DPh-BTBT薄膜成長に適した基板の選定を行っている。各種基板、(i) Bi(0001) / Si(111)、(ii) Sb終端化Si(111)-(√3×√3)、(iii) 1 ML Sb / Bi(0001)、Bi終端化Si(111)-(√3×√3)再構成面の(iv) α-phase (1/3 ML Bi), (v) β-phase (1 ML Bi), (vi) mixed phase(0.66 ML Bi)における成長モードを比較したうえで、α-phase (1/3 ML Bi)のBi終端化Si(111)-(√3×√3)再構成表面が、高い表面エネルギーを持つDPh-BTBT結晶の薄膜成長に適した表面であることを報告している。

さらに、面内配向制御を行った結果についても議論している。ペンタセン薄膜は、2°微傾斜したSi(111)面上の高さ3nmのステップで配向が制御できたのに対し、DPh-BTBTの場合には、4°微傾斜面の高さ10nmの周期的ステップ構造が配向制御に必要であると述べている。RHEEDとGIXDからエピタキシーを決定し、DPh-BTBTの格子定数を初めて決定している。また、TSM像からは、平坦基板で6方向あった結晶成長方向が、2方向以下まで制限できていることを示している。選ばれた配向を比較・検討し、有機膜の表面エネルギー異方性を考慮したうえで、配向制御のメカニズムに関するモデルを提唱している。さらに、得られた疑似単一配向DPh-BTBT単分子膜のARPES測定を行い、バンド構造の報告を行っている。

第4章は、テトラセン単結晶表面の表面清浄化法の開発について述べている。まず、ex-situで成長させたテトラセンの単結晶表面は結晶性が悪く、酸化された分子を多く含むことを、RHEED及びXPS測定から確認している。この酸化層を除去するための手法として、通常用いられる真空中アニールと、今回開発した熱アルゴン気流下でのアニールについて検討している。分子の昇華とアニール後の表面の平坦性を比較し、真空中アニールでは熱による表面のラフニングが起きるのに対し、熱アルゴン気流によるアニールでは1層ずつ表面層が除去できると報告している。この違いの原因を、昇華の際の分子クラスターのサイズと、脱離位置の違いによるものと結論付けている。

第5章は結論と総括である。

以上のように、本研究では、ステップ構造による単一配向結晶の成長を達成するとともに、ファンデルワールス界面におけるステップによる配向制御メカニズムの提唱、ならびに単結晶表面の清浄化法の確立を通して、ARPESに適した試料表面の作製法を提案している。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。