Introduction

Activation of C-H bonds of adsorbed alkanes through interaction with metal surfaces is a key process of dehydrogenation on heterogeneous catalysts. Although alkane molecules are chemically inert in nature, there is an electronic interaction between C-H σ* orbitals of adsorbed alkanes and metal d-states (C-H…metal interaction) [1].This interaction causes the softening of C-H bonds of adsorbed alkanes, which is evidenced by the significant red-shift of the C-H stretching vibrational mode (soft mode) [2]. I reported that the soft mode of cyclohexane adsorbed on the Rh(111) surface consists of multiple peaks that are broadened with increasing temperature (Figs. 1(a)-(d)) [3]. The observed broadening can be explained by the dephasing due to anharmonic coupling between the soft mode and thermally excited frustrated modes. On the other hand, the soft mode disappears completely on the hydrogen-saturated Rh(111) surface, indicating drastic weakening of the cyclohexane-substrate interaction by preadsorption of hydrogen (Fig. 1(e)).

In this thesis, I investigated interactions and two-dimensional (2D) superstructures of cyclohexane on the clean and hydrogen-preadsorbed Rh(111) surfaces using several experimental techniques to obtain further insight into the nature of the C-H…metal interaction, and the role of preadsorbed hydrogen.

Results and Discussion

A.2D superstructures and energy level alignment of cyclohexane on Rh(111)

Fig. 2(a) shows an STM image of cyclohexane adsorbed on the clean Rh(111) surface at cyclohexane coverage (θ) below 0.7 ML (1 ML=0.17 molecules/surface Rh atom). Adsorbed molecules form an ordered structure due to an attractive intermolecular interaction. In addition, molecular images show a periodic change in apparent height and shape (moire structure). Fig. 2(b) shows a line profile of the apparent height along the line in Fig. 2(a). The corrugation amplitude (height difference between the highest and lowest molecules) is about 0.2 Å. The observed superstructure can be assigned to a higher-order commensurate (HOC) (2√79×2√79) R17.0° structure. The intermolecular distance (dm) of this superstructure is 6.83Å. The variation in the apparent height results from the intermolecular difference in the adsorption distance. Thus, the multiple peaks of the soft mode originate from the variation in the adsorption distance. At θ≧0.7 ML, the structural phase transition from the moire structure to a densely packed structure (Fig. 3) was observed. The dense cyclohexane layer shown in Fig. 3(a) is an HOC (2√13×2√13)R13.9° superstructure (dm =6.47Å). A line profile (Fig. 3(b)) shows periodic distribution of lower protrusions (denoted by "O") which are assigned to molecules adsorbed on the on-top site.

An STM image of adsorbed cyclohexane on the hydrogen-saturated Rh(111) surface is shown in Fig. 4(a). There is no clear difference in the apparent height and the shape of each molecular image of adsorbed cyclohexane. In a line profile shown in Fig. 4(b), the variation in apparent height is less than 0.05Å, much smaller than those observed in the cyclohexane layers on the clean surface (Figs.2 and 3). The cyclohexane overlayer on the hydrogen-saturated surface is incommensurate with respect to the substrate, and the value of dm was estimated to be 6.44Å from LEED measurements. The results indicate that the adsorption state of cyclohexane on the hydrogen-saturated surface is insensitive to the adsorption site.

A C 1s XPS spectrum of 1 ML cyclohexane [(2√13×2√13)R13.9° structure] on the clean Rh(111) surface is shown in Fig. 5(a). On the clean Rh(111) surface, there are two C 1s components (284.04 eV and 283.69 eV). The peak at 283.69 eV is assigned to molecules on the on-top site. The intermolecular difference in the C 1s binding energy results from the site-specific energy level alignment of cyclohexane. The energy level alignment of cyclohexane on the Rh(111) surfaces is determined by the vacuum level (VL) shift and the final-state screening effects; both factors depend on the adsorption distance.

A C 1s spectrum of cyclohexane on the hydrogen-saturated Rh(111) surface (Fig. 5(b)) shows a significant peak shift to the lower binding energy by 0.6 eV compared to the main peak on the clean Rh(111) surface. This is a consequence of the longer cyclohexane-surface distance due to the Pauli repulsion between cyclohexane and preadsorbed hydrogen, which leads to the smaller VL shift and the insufficient screening on the hydrogen-saturated surface as schematically illustrated in Fig. 6.

B. Kinetic and geometric isotope effects in molecular adsorption of cyclohexane on Rh(111)

Fig. 7 shows the desorption energy of cyclohexane on the clean Rh(111) surface obtained from TPD measurements at θ < 0.7 ML. The desorption energies of C6D(12) are smaller than those of C6H(12), which is opposite to the normal kinetic isotope effect (inverse kinetic isotope effect).

Fig. 8 shows the work function change as a function of cyclohexane coverage on the clean Rh(111) surface measured by UPS. The work function change by adsorption of C6D12 molecules is smaller than that by C6H(12) adsorption at any cyclohexane coverage. This indicates that the C6D12 molecules are slightly more distant from the surface than the C6H(12) molecules because the work function change is mainly controlled by the adsorption distance [1].

Fig. 9 illustrates adsorption potential energy curves that are derived from the observed results of TPD and UPS at θ= 0.3 ML. The adsorption potential is different between C6H(12) and C6D12. Note that no isotope effect was observed on the hydrogen-saturated Rh(111) surface where the C-H…metal interaction is drastically weakened. This indicates that the C-H…metal interaction plays a key role in the observed isotope effects. Based on the similarity between the C-H…metal interaction and the typical hydrogen bond, the origin of the isotopic difference in the adsorption potential should be attributed to the quantum nature of hydrogen in the C-H…metal bond.

The isotope effects on the cyclohexane-substrate interaction also influence the 2D superstructures. Fig. 10 shows LEED patterns of C6H(12) and C6D12 adsorbed on the clean Rh(111) surface (θ= 0.3 ML). The diffraction patterns are clearly different between two isotopes. The dm in the C6D12 superstructure (6.73Å) is smaller than that in the C6H(12) superstructure (6.83Å). This is a lateral geometric isotope effect (GIE) inthe superstructure of adsorbed cyclohexane observed at θ< 0.7 ML.

The intermolecular distance in cyclohexane superstructure is determined by the balance between the attractive van der Waals interaction and the repulsive interaction between the interfacial dipoles. The lateral GIE in the cyclohexane superstructure can be attributed to the isotopic difference in the repulsion between the interfacial dipoles. The interfacial dipole by C6D12 adsorption is smaller than that by C6H(12) adsorption, which is evidenced by the smaller work function change by C6D12 adsorption. Thus, the repulsive interaction between interfacial dipoles is weaker for C6D12 than C6H(12). As a result, C6D12 molecules can be more densely packed in the 2D superstructure.

Conclusions

The higher-order commensurate superstructures of adsorbed cyclohexane on the clean Rh(111) surface are controlled by the delicate balance between adsorbate-adsorbate and adsorbate-substrate interactions. There is a variation in the adsorption distance which is responsible for the multiple peaks of the soft mode, and the site-specific energy level alignment of adsorbed cyclohexane. The preadsorption of hydrogen atoms weakens the cyclohexane-substrate interaction. This affects both the superstructure and the energy level alignment of cyclohexane.

Cyclohexane adsorbed on the clean Rh(111) surface shows the "inverse kinetic" and "geometric" isotope effects in the adsorption energy and the 2D superstructure. These isotope effects originate from the quantum nature of hydrogen in the C-H…metal interaction as in the case of the typical hydrogen bond.

FIG.1. (a) - (d) IRAS spectra of cyclohexane on Rh(111) as a function of temperature. (e) An IRAS spectrum of cyclohexane on H-saturated Rh(111) at 90 K.

FIG.2. (a) An STM image of cyclohexane on the Rh(111) surface at θ< 0.7 ML (Vs = 0.31 V, It = 0.13 nA, Ts = 86 K). (b) A line profile along the line in (a).

FIG.3. (a) An STM image of cyclohexane on the Rh(111) surface at θ≧ 0.7 ML (Vs = 0.53 V, It = 0.19 nA, Ts = 95 K). (b) A line profile along the line in (a).

FIG.4. (a) An STM image of cyclohexane on the H-saturated Rh(111) surface (Vs = 0.63 V, It = 0.09 nA, Ts = 97 K). (b) A line profile along the line in (a).

FIG.5. C 1s spectra of monolayer cyclohexane on (a) clean and (b) H-saturated Rh(111) surfaces (hv = 380 eV, Ts = 90 K).

FIG.6. Schematic illustrations of the effects of the vacuum level (VL) shift and the final-state screening on the energy level alignment of cyclohexane.

FIG.7. Cyclohexane desorption energy as a function of coverage obtained from TPD measurements.

FIG.8. Work function change by cyclohexane adsorption as a function of cyclohexane coverage measured by UPS at 90 K.

FIG.9. Schematic adsorption potentials of C6H(12) and C6D12 (θ= 0.3 ML).

FIG.10. LEED patterns of (a)C6H(12) and(b)C6D12 on the clean Rh(111) surface at 90 K.

本論文は、Rh(111)清浄表面および水素修飾Rh(111)表面におけるシクロヘキサンの吸着状態、二次元超構造、エネルギー準位アラインメント、脱離における速度論的同位体効果および吸着構造における幾何学的同位体効果について、赤外反射吸収分光(IRAS)、昇温脱離分光(TDS)、スポットプロファイル分析型低速電子回折(SPA-LEED)、真空紫外光電子分光(UPS)、放射光による高分解能軟X線光電子分光(HR-XPS)、走査型トンネル顕微鏡(STM)を用いた実験的研究について報告している。

第1章は本研究の背景と目的、第2章は試料および実験装置と実験方法、第3章はRh(111)表面における吸着シクロヘキサンの2次元超構造、第4章はRh(111)表面における吸着シクロヘキサンのエネルギー準位接続、第5章はRh(111)表面に吸着したシクロヘキサンで観測される速度論的および幾何学的同位体効果、第6章は結論が記述されている。以下にやや詳しく述べる。

第1章では、金属単結晶表面とアルカン分子の相互作用について先行研究を概観し、本論文の研究目的について述べている。

第2章では、Rh(111)単結晶表面の清浄化とシクロヘキサンの吸着方法について述べた後、本研究に用いたIRAS、TPD、SPA-LEED、UPS、HR-XPSなどの実験装置と実験方法について説明している。

第3章では、Rh(111)清浄表面と水素修飾Rh(111)表面に吸着したシクロヘキサンの2次元超構造について、IRAS、SPA-LEEDおよびSTMを用いて詳細に研究した。水素でRh(111)表面を修飾すると、水素吸着量によって金属表面とシクロヘキサンの相互作用を制御する、すなわち、系統的に弱めることができる。SPA-LEEDおよびSTMの結果から逆格子空間および実空間における定量的な情報を得ることができた。様々な2次元超構造を示すことが観測され、水素吸着量とシクロヘキサン吸着量を関数とする相図が得られた。吸着シクロヘキサンと基板との相互作用、吸着シクロヘキサン分子間の相互作用のバランスにより多様な超構造が形成されたと考えられ、各超構造に対するモデルを提案した。それらに基づき、吸着シクロヘキサンのCH伸縮振動で観測されていたソフトニングと今まで謎であった多重ピークの原因について説明することができた。

第4章では、Rh(111)表面に吸着したシクロヘキサン分子の電子状態をHR-XPSとUPSを用いて調べた。 高分解能C1s光電子スペクトルおよび価電子帯光電子スペクトルでは、シクロヘキサンに複数の吸着状態があることが観測された。分子軌道と金属基板の間のエネルギー準位アラインメントを、吸着分子による界面双極子および近接した吸着分子と金属基板によるスクリーニング効果で定量的に説明することに成功した。

第5章では、Rh(111)清浄表面に吸着したシクロヘキサン(C6H(12))と重水素化したシクロヘキサン(C6D(12))の脱離キネティクスと吸着超構造について詳細に研究した。その結果、脱離の活性化エネルギー(吸着エネルギー)は、C6H12のほうがC6D(12)より大きいことがわかった。さらに仕事関数を調べると、C6H(12)のほうがC6D(12)より変化が大きいこと、すなわち界面双極子はC6H(12)のほうがより大きく、相互作用がより強いことがわかった。これらの結果は、C6H(12)の吸着ポテンシャルエネルギー曲線がC6D(12)の場合より深いことを示唆している。2次元超構造をSPA-LEEDで調べてみると、C6H(12)分子間距離がC6D(12)の場合よりも長いことが観測された。これは、C6H(12)がより大きな界面ダイポールを持つので、分子間の双極子-双極子相互作用(斥力)が原因であると結論づけた。これらは、速度論的および幾何学的同位体効果が吸着分子系において初めて観測された事例であり、たいへん意義深い。さらにCHと金属との水素結合的な相互作用に量子的な効果が含まれていることを議論した。

第6章は結語であり、Rh(111)表面とシクロヘキサン分子の相互作用についてまとめている。

以上のように、 小板谷貴典氏は、Rh(111)表面とシクロヘキサンの相互作用について様々な実験手段を駆使して詳細な研究を行い、新たな知見を得ることに成功した。

なお、本論文の第3章は、向井孝三、吉本真也、吉信淳との共同研究、第4章は、清水皇、向井孝三、吉本真也、吉信淳との共同研究、第5章は、向井孝三、吉本真也、吉信淳との共同研究であるが、論文提出者が主体となって、実験の遂行、分析および検証を行ったもので、論文提出者の寄与が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。