Adsorption and Removal Reactions of (CH3)2S on Rh/Al2O3/NiAl(100): Structural and Spectroscopic Study

We have studied the adsorption reaction of (CH3)2S on Rh/Al2O3/NiAl(100) and sulfur removal reaction depending upon heat-treatment using X-ray Photoelectron Spectroscopy and Atomic Force Microscopy techniques. The results show a difference in the adsorption reaction between the as-deposited Rh/Al2O3/NiAl(100) and that heat-treated at 1000 K. (CH3)2S dissociates on the Rh/Al2O3/NiAl(100) as-deposited surface at 300 K, and the sulfur removal reaction occurs at 523 K induced by O2 dosing. In contrast, the heat-treated surface is quite inert against the dissociative reaction of (CH3)2S, because the Rh atoms dissolute into the NiAl(100) substrate through the Al2O3 layer by the high temperature treatment (1000 K). The AFM results also show the morphological changes in these systems, such as many stripe structures for the Al2O3/NiAl(100), nanoclustered surface for the Rh/Al2O3/NiAl(100) as-deposited surface, and nano-hole structures for the Rh/Al2O3/NiAl(100) heated at 1000 K. [DOI: 10.1380/ejssnt.2009.239]


I. INTRODUCTION
Platinum group metals supported on oxides, such as Al 2 O 3 , CeO 2 and ZrO 2 , are widely used as the typical catalysts in the purification process of the automobile exhaust gas. For example, Rh/Al 2 O 3 catalyst is noticeable in the reduction of NO x , and Pd/Al 2 O 3 is available in the oxidation of CO. Those performances are frequently degraded by some deactivation factors because of the use in the severe environment. Especially, the poisoning of catalyst surface and the aggregation of catalyst metals under high temperature are regarded as the main problems [1].
The sulfuric atmosphere as low as few ppm causes the sulfur adsorption on the catalyst surface, which disturbs the other purification reactions; the so-called "Sulfur-Poisoning" is one of the most important deactivation reaction for metal catalysts [2,3]. Therefore not only the resistance against poisoning but also the removal of adsorbed sulfur is required for those catalysts to apply into the purification of the automobile and industrial exhaust gases. In this regard, many works have investigated the interactions between the sulfur-containing molecules and the transition metal surfaces [4][5][6][7]. These studies show that the sulfur poisoning is mainly caused by the adsorption of the atomic sulfur generated by the dissociation of the chemical bonds around sulfur atom. On the other hand, some studies take notice of the removal reaction of sulfur atoms from platinum group metal surfaces [8]. For example, Dohmae reveals that the adsorbed sulfur atoms on Rh polycrystalline surface desorb as SO 2 at around 500 K under O 2 environment [2]. This work also com-pares with Pt and Pd surfaces, and shows the excellent performance of Rh for the sulfur removal reaction. In recent years, the collection of those poisoned metals is difficult because of the low cost performance. Thus, the easy recycle of the catalyst metals can be expected by means of the sulfur removal reaction.
A high temperature environment (∼1000 K) also gives damages to the catalyst metals supported on Al 2 O 3 . One of the problems is the migration of metal atoms under high temperature, and it leads to the aggregation of catalyst metals. The metal dissolution into Al 2 O 3 is the second problem. It was reported that a part of Rh atoms was dissolved into Al 2 O 3 support material by oxidation at 873 K [9]. Both of these two problems are related to the decrease in the reaction area.
In our previous studies, the adsorption system of (CH 3 ) 2 S (dimethyl sulfide: DMS) on Rh(100) single crystal surface have been revealed, where we have investigated the temperature dependent dissociative reaction and the sulfur removal reaction [8,[10][11][12]. However, these systems are too fundamental to apply into practical use. Therefore, we are required to obtain the more actual findings of the sulfur poisoning by clarifying the adsorption behavior of DMS on Rh/Al 2 O 3 system. The aim of this work is firstly to study the growth of Rh on Al 2 O 3 /NiAl(100) depending upon heat treatment, and secondary to reveal the adsorption reaction of DMS on Rh/Al 2 O 3 /NiAl(100) surface and the sulfur removal reaction by means of Xray Photoelectron Spectroscopy (XPS) and Atomic Force Microscopy (AFM) techniques.

II. EXPERIMENTAL
A commercially available NiAl(100) single crystal (10 mm φ diameter, 2 mm thickness, Surface Preparation Laboratory Inc.) was mechanically polished with 0.05 µm Al 2 O 3 to a mirror finish. The NiAl(100) crystal was cleaned by the alternating cycles of Ar + ion sputtering (3 keV, 2 µA, 30 min) to remove sulfur, carbon and oxygen impurities and annealing at 1100 K for 30 min by an electron bombardment in an ultrahigh vacuum (UHV) chamber. The base pressure was better than 4×10 −8 Pa. The cleanliness of NiAl(100) surface was verified by XPS measurements (S 2p, C 1s and O 1s).
A well-ordered ultrathin θ-Al 2 O 3 layer was grown on exposing the cleaned NiAl(100) to O 2 (∼1000 L) at 500 K and subsequently annealing at 1000 K for 5 min [13]. Rhodium was deposited on the Al 2 O 3 thin layer in-situ at room temperature with Rh vapor from an ultra-pure Rh wire (99.9%) set in an electron beam evaporator. The Rh deposition rate and time were 0.025 ML/min and 40 min, and then the total coverage was about 1.0 ML. After deposition, the sample was heated at 1000 K for 5 min. Three types of samples were prepared in this work, which were Rh/Al 2 O 3 /NiAl(100) as Rh deposited, Rh/Al 2 O 3 /NiAl(100) heated at 1000 K, and Al 2 O 3 /NiAl(100) surface.
The research grade DMS was purified by means of a few cycles of freezing with liquid N 2 under high vacuum condition and melting at ambient temperature. DMS gas was admitted via the variable leak valve into the UHV chamber in this procedure, as the sample surface was cooled down to 90 K with liquid N 2 . The dosage was set at 0.3 L. The temperature of substrate (90-300 K) was controlled by resistive heating with a tungsten-filament, which was set behind the crystal. The sulfur removal reaction was promoted under O 2 environment (2.7×10 −5 Pa for 1h) at 423, 473 and 523 K.
The surface morphology was obtained by ex-situ AFM measurement with NanoScope III-a (Veeco Instruments) with tapping mode at Innovation Plaza Hiroshima JST. XPS measurements were recorded by use of the concentric hemispherical electron energy analyzer (PHOIBOS 100-5ch, SPECS) with MgKα X-ray (1253.6 eV). All binding energies were referenced to Ni 3p 3/2 peak position of NiAl at 853.0 eV [14]. The initial thickness of Al 2 O 3 layer was 4-5Å estimated from the signal ratio of I Al2p Al2O3 and I Al2p NiAl were the Al 2p peak intensities of Al 2 O 3 and NiAl as shown in Fig. 3 [15]. The coverages (ML) of Rh and sulfur were obtained with their XPS peak intensities in Rh 3d and S 2p regions (1ML corresponds to the area density of fcc Rh(100) surface atoms, 1.37×10 13 atoms/mm 2 ).

III. RESULTS AND DISCUSSIONS
A. S 2p XPS Figure 1 shows the temperature dependent S 2p XPS spectra for DMS on Rh(1.0 ML)/Al 2 O 3 /NiAl(100) asdeposited, Rh(1.0 ML)/Al 2 O 3 /NiAl(100) heated at 1000 K, and Al 2 O 3 /NiAl(100) surfaces. The initial DMS coverages are estimated at 0.34, 0.34 and 0.26 ML for each sample, respectively. Those amounts mean almost monolayer adsorption. There is no sulfur contamination on each initial surface. On the Al 2 O 3 /NiAl(100) surface, most of DMS adsorb on Al 2 O 3 layer without dissociation into CH 3 S − and atomic S at 90 K. It is found that the sulfur in DMS has an interaction with the substrate and is picked some electrons up by the Al 2 O 3 layer, because the main peak position located at 164.3 eV undergoes a higher energy shift from that of the multilayer DMS at 163.8 eV [10]. Almost all of the adsorbates desorb from the Al 2 O 3 /NiAl(100) surface at 200 K, indicating no catalytic performance for DMS dissociative reaction. Dissociation of DMS also hardly occurs on the Rh/Al 2 O 3 /NiAl(100) as-deposited surface at 90 K. The peak position shifts to the lower binding energy side from that of the Al 2 O 3 /NiAl(100) surface. This shift gives the interaction between DMS and Rh layer. More than half the molecules (0.18 ML) remain on the surface at 200 K without desorption. Some of those molecules undergo the scission in the S-C bonds. The dissociation is promoted as the temperature rises, and most of the S-C bonds are cleaved at 300 K. Thus the Rh over-layer possesses a catalytic performance for the dissociative reaction of DMS.
In the case of the Rh/Al 2 O 3 /NiAl(100) heated at 1000 K, the dissociative reaction and the desorption behavior are quite equivalent to that on the Al 2 O 3 /NiAl(100) surface meaning inert against the dissociation process of DMS.
The sulfur-poisoned Rh/Al 2 O 3 /NiAl(100) as-deposited surface was heated under O 2 environment to remove the sulfur atoms. The main peak is located at 162.0 eV showing the adsorption of the atomic sulfur. One can find the broad peak at 166 eV to 169 eV, when the sample is annealed at 423 K and 473 K. This chemical state is assigned to SO x species [2]. The broad peak disappears at 523 K, and simultaneously the total coverage of sulfur decreases. Those changes indicate the oxidation of sulfur and the desorption of SO x from the Rh/Al 2 O 3 /NiAl(100) as-deposited surface. The desorption onset temperature is higher than that reported with polycrystalline Rh and Rh(100) surfaces, on which the sulfur atoms desorb at 423 K [2,8]. Therefore, the performance of Rh in the sulfur removal reaction is weakened on the Al 2 O 3 layer.
B. Rh 3d XPS Figure 2 shows the Rh 3d 5/2 peaks for the Rh/Al 2 O 3 /NiAl(100) systems before DMS adsorption. The peak shape and the FWHM were estimated using a Shirley background and a mixed Gaussian-Lorentzian function with an asymmetric parameter. Table I shows these intensities, peak positions and FWHM of the Rh 3d 5/2 peak. Metallic Rh state has a single peak at 307.2 eV, shown as the spectrum of Rh(111) bulk [16]. The Rh 3d 5/2 peak of Rh/Al 2 O 3 /NiAl(100) as-deposited becomes broader than that of Rh(111) bulk. The peak position also shifts from that of Rh(111) bulk to 0.25 eV higher binding energy side. Therefore, the spectrum for Rh/Al 2 O 3 /NiAl(100) as-deposited seems to contain other components at the higher binding energy side in addition to the peak for metallic Rh state. The Rh atoms in second layer or more shows their peak at the corresponding position with metallic Rh state at 307.2 eV, because the spectrum becomes similar shape to the Rh(111) bulk spectrum as Rh coverage increases. On the other hand, the additional components shift from the metallic Rh state to the higher energy side, indicating the oxidation or the interaction with the Al 2 O 3 layer. The oxidized Rh state is reported to represent the Rh 3d 5/2 peak at 308.0 eV or more; for Rh 2 O 3 at 308.35 eV and for RhOOH at 308.55 eV [16]. Thus, these components are confirmed the first layer Rh atoms on the surface which interact with Al 2 O 3 layer. These Rh atoms are squeezed some electrons by the Al 2 O 3 layer, which causes the Rh 3d 5/2 peak shift to the higher binding energy side. We can also find the reverse shift of the Al 2 O 3 peak in Al 2p and O 1s spectra shown in Figs. 3 and 4. These results indicate multilayer or nanocluster structure composed with Rh more than two layers.
In contrast, the peak shape of the Rh/Al 2 O 3 /NiAl(100) heated at 1000 K can be described with a single state at 307.50 eV. The state extremely stabilizes against various treatments, such as the DMS adsorption, the heating and the O 2 exposure. The integral intensity decays in comparison with that of the Rh as-deposited sample by 22%. Hence, two possible processes can engender the stability and the decrease in the Rh 3d 5/2 peak intensity; the morphology of the Rh over-layer changes from a flat structure to a cluster structure; and the Rh atom dissolves into the substrate. We confirm the dissolution of Rh into the substrate under high temperature treatments because of the low reactivity to DMS dissociation. Al 2p and O 1s XPS results also agree with that standpoint in next chapter. Rh 3d 5/2 XPS result reveals the significantly different environment between Rh atoms on the as-deposited surface and that on the heat-treated surface (1000 K).  interaction between the Al 2 O 3 layer and the deposited Rh atoms. The lower shifted peak position does not move any longer in the adsorption system of DMS on Rh/Al 2 O 3 /NiAl(100) as-deposited surface, indicating that those adsorption reaction is exclusive on the Rh surface. A dramatic change is observed in the signal ratio of I Al2p Al2O3 /I Al2p NiAl after O 2 treatment at 423K, 473 K and 523 K. Thereby, the oxide thickness increases significantly after heating under O 2 environment. The Al 2 O 3 thicknesses are calculated to be 8Å, 11Å and 13Å at 423 K, 473 K and 523 K, respectively (the pristine Al 2 O 3 thickness is about 5Å). As mentioned by Song et al., the pristine Al 2 O 3 layer thickness is stable even at 1070 K without over-layer [17]. This result indicates that Rh layer behaves as a catalyst for oxidation of aluminum by supplying reactive oxygen atoms formed by dissociation of oxygen on Rh surface. A similar catalytic mechanism was reported in Pd/Al 2 O 3 /NiAl(110) system [17,18]. A part of the reactive oxygen atoms also have an important role in the sulfur removal reaction as revealed in our previous study [8].

C. Al 2p and O 1s XPS
When the Rh as-deposited sample is heat-treated at 1000 K, the Al 2 O 3 binding energy returns to a position near that before the deposition of Rh. Therefore, the interaction between the Al 2 O 3 layer and the Rh atoms is lost by heat-treatment. We can consider two possible processes such as the morphological change of Rh layer and the dissolution of Rh into the substrate as shown in Chapter 3. 2. We conclude the dissolution of Rh into the NiAl(100) substrate under high temperature treatments, when DMS molecules adsorb on that surface. The S 2p XPS result shows the charge transfer from DMS to the Al 2 O 3 layer. Simultaneously, the shift to the lower binding energy side is observed in the DMS on Al 2 O 3 /NiAl(100) system as shown in Fig. 3. Similar shift has occurred on the Rh/Al 2 O 3 /NiAl(100) heated at 1000 K. Furthermore, the growth of Al 2 O 3 layer has not been promoted by O 2 exposure. Therefore, it indicates that the deposited Rh atoms do not present on the Al 2 O 3 layer but it dissolves into the NiAl(100) substrate through the Al 2 O 3 layer. This diffusion into the substrate is possible through the defects of the Al 2 O 3 layer [17,19]. The O 1s XPS results shown in Fig. 4 also represent the similar shift as the Al 2p XPS results, indicating that the Rh and the adsorbate have some interaction with not only aluminum atoms but also oxygen atoms in the Al 2 O 3 layer. Figure 5 shows the AFM images obtained from Al 2 O 3 /NiAl(100), Rh/Al 2 O 3 /NiAl(100) as-deposited and Rh/Al 2 O 3 /NiAl(100) heated at 1000 K. We can see a lot of stripe structures on the Al 2 O 3 /NiAl(100) surface growing preferentially along the crystal directions, consistent with the other reports [13,20,21]. These stripes disappear when the surface covered with Rh layer. The deposited Rh on Al 2 O 3 /NiAl(100) has a protruded structures with the height of 3-7Å. Thereby, it can be speculated that nanoclusters consisted of 1-3 Rh layer formed on the Al 2 O 3 /NiAl(100) surface because the diameter of a Rh atom is estimated to be 2.7Å. After the heattreatment at 1000 K, there is no protruded structure but a lot of nano-hole structures as shown in Fig. 5(c). The former XPS results show the absence of Rh atoms on the surface. Thus, it is supposed that this outmost surface is mainly composed with the Al 2 O 3 layer, which undergoes the structural change by the diffusion process of Rh into the NiAl(100) substrate through the Al 2 O 3 layer. In three types of surfaces, only the Rh/Al 2 O 3 /NiAl(100) asdeposited surface has the activity for the dissociative reaction of between DMS and O 2 , because the Rh atoms are exposed on the surface. The proposed mechanism for the interaction of Rh/Al 2 O 3 /NiAl(100) with DMS and O 2 at elevated temperatures is shown in Fig. 6. The process I shows the initial surface of Rh/Al 2 O 3 /NiAl(100). The adsorbed DMS dissociates on the Rh nanocluster surface. The Rh surface is covered with the sulfur atoms in process II. The dissociative adsorption of O 2 and the growth of the Al 2 O 3 layer occur at the temperature range of 423-473 K (process III). Finally, a part of the sulfur atoms desorbs from the surface as SO 2 in the process IV.

IV. CONCLUSION
We have investigated the adsorption reaction of DMS on Rh/Al 2 O 3 /NiAl(100) surface and the sulfur removal reaction related with heat-treatment with XPS and AFM techniques. Atomic sulfur remains on the Rh/Al 2 O 3 /NiAl(100) as-deposited surface at 300 K because of DMS dissociative reaction. The sulfur removal reaction is induced by O 2 gas at 523 K with the formation and desorption of SO 2 . Simultaneously, the growth in the thickness of Al 2 O 3 layer occurs since the Rh layer serves the catalytic effect for the oxidation of aluminum in the NiAl(100) substrate. On the other hand, the Rh/Al 2 O 3 /NiAl(100) heat-treated at 1000 K is quite inert against DMS dissociation. The reason for the inertness is confirmed the dissolution of Rh into the NiAl(100) substrate through the Al 2 O 3 layer occurring by the high temperature treatment. The AFM results show the morphological change in these systems; many stripe structures for the Al 2 O 3 /NiAl(100), nanoclustered surface that consists of Rh for the Rh/Al 2 O 3 /NiAl(100) as-deposited surface, and nano-hole structures formed with Al 2 O 3 layer for the Rh/Al 2 O 3 /NiAl(100) heated at 1000 K.