1 Introduction  1 介绍

Selective catalytic reduction is a promising method for removing NO, a pollutant emitted from diesel engines, lean-burn engines, and combustors. Numerous studies have investigated the selective catalytic reduction of NO using reductants such as NH3 and hydrocarbons [1–3]. On the other hand, H2 and CO have not been regarded as effective reductants under oxidizing conditions, because they are more easily oxidized by O2 than by NO. Concerning NO reduction with CO, it is not required to add newly the reductant, because the CO was included together with combustion exhaust gases. Thus, the development of a catalyst that could effectively reduce NO to N2 with CO would contribute to more extensive use of selective catalytic reduction of NO.
选择性催化还原是去除 NO 的一种很有前途的方法，NO 是柴油发动机、稀薄燃烧发动机和燃烧室排放的污染物。许多研究调查了使用 NH3 和碳氢化合物等还原剂选择性催化还原 NO [1–3]。另一方面，H2 和 CO 在氧化条件下不被认为是有效的还原剂，因为它们比 NO 更容易被 O2 氧化。关于用 CO 还原 NO，不需要添加新的还原剂，因为 CO 与燃烧废气一起包括在内。因此，开发一种可以用 CO 有效将 NO 还原为 N2 的催化剂将有助于更广泛地使用选择性催化还原 NO。

Recently, a number of researches have been carried out for the selective catalytic reduction of NO with CO in the presence of O2. The Ir- and Rh-based catalysts have been reported to be effective in this reaction and show the unique catalytic property [4–9]. Ogura et al. [4] found that NO was selectively reduced with CO on Ir/silicalite in an oxidizing atmosphere. They showed that the NO reduction was promoted by coexisting O2. Haneda et al. [8] reported that the coexistence of SO2 greatly promoted the reduction of NO by CO over Ir/SiO2 in the presence of O2, whereas no promotional effect of SO2 was observed for Rh/SiO2. Fujitani et al. [10] examined the adsorption and reactivity of SO2 using the Ir(111) and Rh(111) surfaces, and demonstrated that the atomic sulfur produced by SO2 dissociation effectively removed surface oxygen on Ir, whereas on Rh, atomic sulfur did not react with oxygen. We concluded that the effect of SO2 was to prevent the oxidation of the Ir metal surface. The surface states of the Ir and Rh catalysts would be influenced under real reaction conditions. However, the detail of the surface state is not fully understood. Thus, we have interested in the nature of catalysis such as the active site and the mechanism for the NO reduction with CO over Ir and Rh from fundamental point of view.
最近，已经进行了大量研究，用于在 O2 存在下用 CO 选择性催化还原 NO。据报道，Ir 和 Rh 基催化剂在该反应中有效，并显示出独特的催化性能 [4–9]。Ogura 等 [4] 发现，在氧化气氛中，IR/硅酸盐上的 CO 选择性还原了 NO。他们表明，共存的 O2 促进了 NO 的减少。Haneda等[8]报道，在O2存在下，SO2的共存极大地促进了CO对NO的还原，而不是Ir/SiO2，而SO2对Rh/SiO2没有观察到促进作用。Fujitani 等 [10] 使用 Ir（111） 和 Rh（111） 表面研究了 SO2 的吸附和反应性，并证明 SO2 解离产生的原子硫有效地去除了 Ir 上的表面氧，而在 Rh 上，原子硫不与氧反应。我们得出结论，SO2 的作用是防止 Ir 金属表面的氧化。在实际反应条件下，Ir 和 Rh 催化剂的表面态会受到影响。然而，表面状态的细节尚不完全清楚。因此，我们从基本面对催化的性质感兴趣，例如活性位点以及用 CO 超过 Ir 和 Rh 还原 NO 的机制。

The reaction behaviors of NO or CO on the well-defined Ir and Rh single crystal surfaces have been studied by means of surface science techniques. For the NO reactivity on the Rh single crystal, the adsorption, dissociation and desorption properties have been extensively examined using various crystal planes such as (111) [11–16], (100) [17–21], (110) [22–26], and stepped [19, 27–29] surfaces. The adsorbed NO on the Rh surface dissociates completely upon heating at low NO coverage, whereas the dissociation percentage decreases with increasing NO coverage. The kinetics of NO dissociation was strongly dependent on the Rh crystal face [12, 18, 24]. Many studies have also investigated the adsorption and desorption of CO using various Rh single crystal surfaces [30–35]. The adsorption site, structure and energy of CO were shown to depend on the Rh surface structure.
已经通过表面科学技术研究了 NO 或 CO 在明确定义的 Ir 和 Rh 单晶表面上的反应行为。对于Rh单晶上的NO反应性，已经使用各种晶面进行了广泛的吸附、解离和解吸特性，如（111） [11-16]、（100） [17-21]、（110） [22-26]和阶梯状[19-27-29]表面。 在低 NO 覆盖率下加热时，Rh 表面吸附的 NO 完全解离，而解离百分比随着 NO 覆盖率的增加而降低。NO 解离的动力学强烈依赖于 Rh 晶面 [12， 18， 24]。许多研究还研究了使用各种Rh单晶表面对CO的吸附和解吸[30–35]。CO 的吸附位点、结构和能量与 Rh 表面结构有关。

For Ir single crystal, the reactivity of NO [36–38] or CO [39–44] has been reported. The NO adsorbed on Ir(111) decomposed to N2, where no formation of N2O was observed. The NO dissociation percentage was slightly higher than that on the Rh surface. The CO was found to adsorb molecularly in a (√3 × √3)R30° ordered structure on Ir(111) at room temperature. The influence of the Ir crystal face and the Ir step site on the NO dissociation activity and CO adsorption property was not entirely examined. For the NO reduction with CO, it is important to clarify the influence of coadsorbed CO on the adsorption and dissociation of NO. However, the adsorption site and reaction property on Ir and Rh surfaces under the coexistence of NO and CO were not fully understood. Therefore, we have investigated concerning the coadsorption of and reaction between NO and CO over the Ir and Rh single crystal surfaces. In this paper, we summarized our previous works concerning the adsorption and reactions of NO over the clean and CO-adsorbed Ir(111) and Rh(111) surfaces.

2 Experimental

Infrared reflection absorption spectroscopy (IRAS) and temperature-programmed desorption (TPD) experiments were carried out in an ultrahigh vacuum apparatus described in detail elsewhere [45, 46]. IRA spectra were recorded with a resolution of 4 cm−1 with 100 scans (30 s). TPD experiments were performed with a 1.0 K/s heating rate.

Ir(111) and Rh(111) single-crystal discs (8 mm diameter, 1 mm thickness, 99.999% purity) polished on one side only were used in this study. The crystal orientation was accurate to within 1°, and the surface roughness was <0.03 μm. The surfaces were cleaned by cycles of Ar+ sputtering and annealing in oxygen at 1,000 K and then by annealing at 1,200 K in a vacuum. The surface cleanliness of the samples was verified by Auger electron spectroscopy.

Adsorption experiments were carried out with 15NO (99.9% pure) and CO (99.999% pure) at (5 × 10−9) to (5 × 10−8) Torr (0–60 L; 1 L = 10−6 Torr s) and a sample temperature of 263–273 K. 15NO gas was used to distinguish between N2 and CO desorption peaks and between N2O and CO2 desorption peaks during TPD.

3 Adsorption and Decomposition of NO on Ir(111)

Adsorption of NO on Ir(111) was investigated using IRAS. In the IRA spectra of adsorbed NO on Ir(111) exposed to 15NO gas at 0.2–45 L at 263 K (Fig. 1), the peaks due to NO adsorbed on hollow and atop sites were observed at 1,400–1,444 and 1,799–1,820 cm−1, respectively. Only the peak for NO adsorbed on atop sites was observed at the initial state of exposure to NO, which indicates that the NO was selectively adsorbed on atop sites at low NO coverage. The intensity of the peak for NO adsorbed on atop sites increased with increasing NO exposure up to 2 L, and then slightly decreased. The intensity of the peak for atop-NO was constant at exposures above 15 L. In contrast, NO began to adsorb on hollow sites at exposures above 2 L. The intensity of the hollow-NO peak increased with increasing NO exposure up to 15 L and then became saturated. As can be seen in Fig. 1, the peak intensity of the hollow-NO species was much smaller than that of the atop-NO species at saturation, although the hollow-NO coverage was twice that of the atop-NO species from the result of X-ray photoelectron spectroscopy (XPS) measurement [47]. Aizawa et al. [48] have reported normal-model frequencies and peak intensities of NO/Pt(111) at various coverages, and they found that the N–O stretching peak intensity of fcc-hollow species greatly decreased in the presence of coexisting atop species. This effect is not due to the well-known intensity-transfer effect derived from dynamic dipole-dipole coupling but to a change in the electronic state of the adsorption system caused by the coexistence of the atop species. The calculated data [48] are in good agreement with our experimental results.

Fig. 1
figure 1

IRA spectra for Ir(111) surface during 15NO exposure at 0.2–45 L and 263 K

Full size image

We investigated the thermal reactivity of 15NO on the Ir(111) surface using TPD. In the TPD spectra for 15NO and 15N2 after saturation of the surface with NO at 263 K (Fig. 2), we did not observe any desorption peaks for N2O and O2 on the Ir(111) surface, which indicates that atomic oxygen dissociated from NO remained on the surface at temperatures below 900 K. Two desorption peaks for 15NO were observed, at 393 and 455 K. In light of the XPS and electron energy loss spectroscopy results [47], which indicate that only atop-NO species existed on the Ir surface after annealing to 373 K, we assigned the peaks at 393 and 455 K to desorption of 15NO adsorbed on hollow and atop sites, respectively. Simultaneously with 15NO desorption, the 15N2 desorption peaks were observed at 471 and 574 K.

Fig. 2
figure 2

TPD spectra of 15NO and 15N2 for the Ir(111) surface after saturated adsorption of 15NO at 263 K

Full size image

The mechanism of the NO decomposition on Rh(111) was reported by Root et al. [11], who observed the two desorption peaks of N2 at 473 and 560 K at saturation coverage of NO. They suggested that the N2 desorption peak at 473 K could be attributed to a disproportionation reaction, NOa + Na → N2 g + Oa, and they attributed the peak at 560 K to recombination of Na. On the basis of their results, we assigned the 15N2 peaks we observed at 471 and 574 K to a disproportionation reaction between atop-NO and Na dissociated from hollow-NO and to the recombination of Na, respectively.

4 Adsorption and Reaction of NO on CO-preadsorbed Ir(111)

We used IRAS to obtain the details regarding the adsorption state of NO on Ir(111) with CO preadsorbed at a coverage below 0.33 (Fig. 3). In the IRA spectra of NO exposed at 273 K at 45 L on the Ir(111) surface preadsorbed at various CO coverages, the peak intensities of NO adsorbed on atop sites decreased with increasing CO coverage, and the peaks disappeared after the preadsorbed CO coverage reached approximately 0.27. In contrast, we observed no change in the peak for NO adsorbed on hollow sites, even when preadsorbed CO was present in saturation coverage. These results clearly demonstrate that preadsorbed CO, which selectively adsorbed on atop sites of Ir(111), inhibited adsorption of atop-NO, whereas the adsorption of NO on hollow sites was not affected by the presence of adsorbed CO.

Fig. 3
figure 3

IRA spectra after 15NO adsorption at 45 L at 273 K on the CO-preadsorbed Ir(111) surface at various coverages

Full size image

The thermal reaction between CO and 15NO on Ir(111) was examined by TPD. In the 15NO spectra (Fig. 4), the peaks that appeared at around 460 K decreased as the amount of preadsorbed CO increased, and the peaks disappeared at CO coverage of 0.33. However, the 15NO desorption peak at 400 K remained constant regardless of the coverage of preadsorbed CO. The results shown in Fig. 3 indicate that NO adsorbed on atop sites decreased with increasing CO precoverage, but hollow-NO species was present on Ir(111) at any coverage of preadsorbed CO. Thus, the desorption peaks at 400 and 460 K were due to desorption of NO adsorbed on hollow and atop sites, respectively.

Fig. 4
figure 4

TPD spectra of 15NO (a), 15N2 (b), CO (c), and CO2 (d) after the saturated adsorption of 15NO at 273 K on the CO-preadsorbed Ir(111) surface at various CO coverages

Full size image

Similar changes in the TPD spectra were observed for the desorption of 15N2. The peaks at 470 K decreased with increasing CO precoverage, whereas the peaks at 570 K remained constant regardless of CO coverage. Because atop-NO species were reduced by the presence of CO, the 15N2 peak due to the disproportionation reaction between atop-NO and Na dissociated from hollow-NO disappeared with increasing CO exposure. This result clearly indicates that the 15N2 peaks at 470 and 570 K were due to a disproportionation reaction between atop-NO and Na dissociated from hollow-NO and to the recombination of Na, respectively. A CO desorption peak was observed at around 600 K, and this peak increased with increasing CO exposure. Subsequent to N2 desorption, the CO2 formation peak was observed at 574 K, a temperature that was lower than the CO desorption temperature (600 K). Thus, CO2 was produced by the reaction of adsorbed CO and atomic O produced by dissociation of NO adsorbed in the hollow sites.

5 Adsorption and Decomposition of NO on Rh(111)

We investigated NO adsorption on the Rh(111) surface using IRAS to clarify the adsorption behavior of NO (Fig. 5). A peak was observed at 1,477 cm−1 for 0.2 L of NO exposure. An increase in the peak intensity and a peak shift to higher frequency were seen as the NO exposure was increased to 4 L, at which point the peak had shifted to 1,606 cm−1. At exposures above 5 L, the peak intensity gradually decreased with increasing NO exposure; and at 5 L, a new peak appeared at 1,813 cm−1, and the peak increased in intensity with increasing NO exposure up to 12–16 L. A third peak was observed at 1,495 cm−1 above 16 L. The intensities of the three NO adsorption peaks at 1,495, 1,606, and 1,813 cm−1 became constant above 48 L. From the literature data [16, 49], the peaks at 1495, 1606, and 1,813 cm−1 were assigned to NO adsorbed on hcp, fcc, and atop sites, respectively. In the NO adsorption process, the fcc-NO peak shifted continuously from 1,477 to 1,606 cm−1 with increasing NO exposure between 0.2 and 4 L; note that the peak shift is due to a change in the intermolecular interactions that accompany the increased NO coverage.

Fig. 5
figure 5

IRA spectra for the Rh(111) surface during 15NO exposure at 0.2–60 L and 273 K

Full size image

We have also examined the NO adsorption on the Rh(111) surface using XPS. NO coverage increased with increasing NO exposure, and saturation coverage of ΘNO = 0.71 was obtained at about 50 L [50], which was in good agreement with the exposure (48 L) at which saturation NO adsorption was measured by IRAS (Fig. 5). On the basis of the IRAS and XPS [50] results, we summarize the NO adsorption process on the Rh(111) surface as follows. NO adsorbed initially on the fcc sites. After the fcc-NO coverage reached 0.23 at 4 L of NO exposure, NO started to adsorb on the atop sites. At 12 L, the atop-NO coverage was estimated to be 0.23 because the total NO coverage was 0.46. At 16 L, the NO began to adsorb on the hcp sites, and then the NO adsorption reached saturation at 48 L (ΘNO = 0.71), at which point the hcp-NO coverage was estimated to be 0.26. NO coverage for each type of the sites—fcc, atop, and hcp—was almost the same (0.23–0.26) for the saturated adsorption surface, a result that was consistent with that for the (2 × 2)-3NO structure reported as a saturated adsorption structure of NO on Rh(111) [13]. NO has been reported to adsorb on a Pt(111) surface at the fcc, atop, and hcp sites in that order [51, 52]. Thus, the NO adsorption process on Rh(111) was similar to that on Pt(111) but was significantly different from that on Ir(111) shown in Fig. 1.

The thermal reactivity for NO adsorbed on each type of the sites over Rh(111) was examined using TPD (Fig. 6). The desorption peak of fcc-NO was observed at 490 K for the Rh(111) surfaces with NO adsorbed only on the fcc sites. On the other hand, the desorption peaks were observed at 395 and 490 K for the surface at ΘNO = 0.37, where NO adsorbed also on the atop sites, indicating that the peak at 395 K could be attributed to desorption of atop-NO. Between ΘNO = 0.61 and 0.71, the area of the desorption peak at 490 K increased with adsorption of NO on the hcp sites, which clearly indicated that fcc-NO and hcp-NO desorbed at the same temperature. Thus, the stability of NO adsorption on the atop sites over Rh(111) was lower than the stabilities on the fcc and hcp sites.

Fig. 6
figure 6

TPD spectra of 15NO (a) and 15N2 (b) for the Rh(111) surface-adsorbed 15NO at various coverages and 273 K

Full size image

For N2 desorption, peaks at 500 and 680 K were observed for the surfaces at all NO coverages, and both peak areas increased with increasing NO coverage. Previous papers have reported that the N2 desorption at lower temperature is due to a disproportionation reaction between NO and Na, and the desorption at higher temperature is due to a recombination of Na [20, 53]. Kao et al. [13] reported that no dissociation was observed for NO adsorbed on the atop sites on Rh(111), indicating that the Na was produced from fcc-NO and hcp-NO. Furthermore, atop-NO is absent on the Rh(111) surface in the temperature range of N2 formation. Therefore, the N2 desorption at 500 K is due to the disproportionation reaction of fcc-NO and hcp-NO with Na from NO on fcc and hcp sites. N2 desorption at 680 K is attributed to recombination of Na from fcc-NO and hcp-NO.

6 Adsorption and Reaction of NO on CO-preadsorbed Rh(111)

Coadsorption of NO and CO on the Rh(111) surface was examined with IRAS (Fig. 7). The peaks due to atop-NO and fcc-NO were observed at 1,797 and 1,543 cm−1, respectively, for the Rh(111) surface-preadsorbed atop-CO at ΘCO = 0.15. NO adsorption on hcp sites was entirely inhibited by the presence of CO on atop sites. Only the fcc-NO peak was observed at 1,537 cm−1 for the surface-preadsorbed CO on atop sites at ΘCO = 0.30, at which coverage the intensity of the fcc-NO peak was the same as that observed for the CO-preadsorbed Rh(111) surface at ΘCO = 0.15. This result indicates that the nearly saturated preadsorption of CO on atop sites completely inhibited the NO adsorption on atop sites but did not affect the adsorption of NO on fcc sites. The intensity of the fcc-NO peak decreased as that of the hollow-CO peak increased above ΘCO = 0.36. No fcc-NO peak was observed at ΘCO = 0.60. CO preadsorption on hollow sites inhibited completely adsorption of NO on fcc sites. In contrast, the adsorption of NO on hollow sites over Ir(111) was not inhibited by the preadsorbed CO, because the CO adsorbed only on the atop sites (Fig. 3). Therefore, it was suggested that the Ir surface, which was not affected for NO dissociation by coadsorption of CO, shows a high catalytic activity for NO + CO reaction compared with Rh surface.

Fig. 7
figure 7

IRA spectra after 15NO adsorption at 60 L and 273 K on the CO-preadsorbed Rh(111) surface at various CO coverages

Full size image

The TPD spectra after saturated adsorption of 15NO at 273 K on the CO-preadsorbed Rh(111) surface at various CO coverages were obtained (Fig. 8). NO desorption peaks due to atop-NO and fcc-NO were observed at 395 and 490 K, respectively, for the CO-preadsorbed Rh(111) surface at ΘCO = 0.15. No influence on the NO desorption temperature was observed in the presence of adsorbed CO. A desorption peak for fcc-NO was observed for the surfaces at ΘCO = 0.30 and 0.36, whereas no NO desorption peak was observed at ΘCO = 0.50, even in the presence of fcc-NO. At ΘCO = 0.60, NO desorption was not observed, because of complete inhibition of NO adsorption.

Fig. 8
figure 8

TPD spectra of 15NO (a), 15N2 (b), CO (c), and CO2 (d) after the saturated adsorption of 15NO at 273 K on the CO-preadsorbed Rh(111) surface at various CO coverages

Full size image

In the N2 desorption spectra, peaks due to the disproportionation reaction and Na recombination were observed at 500 and 680 K, respectively, on the CO-preadsorbed Rh(111) surfaces at ΘCO = 0.15, 0.30, and 0.36, as well as on a clean surface. However, only a desorption peak due to the recombination of Na was observed for the surface at ΘCO = 0.50, which indicates that all of adsorbed NO dissociated on this surface because NO desorption was not observed.

In the CO spectra, a desorption peak for atop-CO was observed at 420 K for ΘCO = 0.15. A broad desorption peak for atop-CO at 460 K was observed at ΘCO = 0.30. The desorption temperature of atop-CO came close to that for the surface with adsorbed CO alone [53], indicating that the repulsive interaction became small with increasing CO coverage. On the other hand, a desorption peak was detected at 470 K at ΘCO = 0.36 and 0.50, indicating that both atop-CO and hollow-CO desorbed at the same temperature. Thus, the desorption temperature of atop-CO was affected by fcc-NO, but that of hollow-CO was not affected. Two CO desorption peaks were observed at 460 and 530 K for the surface at ΘCO = 0.60, on which no NO adsorbed.

In the CO2 spectra, the desorption peak was observed at 460 K for the CO-preadsorbed Rh(111) surface at ΘCO = 0.15–0.50, indicating that the adsorbed CO reacted with atomic oxygen (Oa) that was produced by dissociation of fcc-NO. The CO2 desorption peak was significantly large at ΘCO = 0.50 compared with the one at ΘCO = 0.15–0.36, although the amount of N2 formation is small; that is, Oa coverage was small. This result indicates that the CO2 production increased remarkably in the presence of hollow-CO. It may be considered that the Oa from dissociated fcc-NO reacted effectively with hollow-CO rather than with atop-CO. At ΘCO = 0.60, no CO2 desorption was observed, because of absence of Oa on this surface. It was thus suggested that the NO reduction with CO on the Rh catalyst surface proceeded via the direct decomposition of NO.

7 Conclusions

  1. 1)

    Two NO adsorption states of fcc-hollow sites and atop sites were observed for the Ir(111) surface. NO adsorbed preferentially on atop sites at low NO coverage and then began to adsorb on hollow sites as NO coverage increased. The dissociated Na desorbed as N2 at the two temperatures, by recombination of Na at 574 K and by a disproportionation reaction between atop-NO and Na at 471 K.

  2. 2)

    Preadsorbed CO selectively adsorbed on atop sites of Ir(111). The preadsorbed CO inhibited adsorption of atop-NO, whereas the adsorption and dissociation of NO on hollow sites were not affected by the presence of adsorbed CO. TPD results showed the formation of CO2 subsequent to N2 desorption, which indicates that the adsorbed CO reacted with Oa dissociated from hollow-NO.

  3. 3)

    NO adsorbed on Rh(111) on the fcc, atop, and hcp sites in that order. The adsorbed NO desorbed at 395 and 490 K, and the desorption peaks at these temperatures were attributed to desorption of atop-NO and fcc-NO + hcp-NO, respectively. Peaks for N2 desorption by disproportionation and recombination were observed at 500 and 680 K, respectively, and these peaks were related to fcc-NO and hcp-NO.

  4. 4)

    NO adsorption on the hcp sites over Rh(111) was entirely inhibited by the preadsorption of CO on atop sites, and NO adsorption on atop and fcc sites was inhibited by the CO preadsorbed on each type of the sites, which indicates that the NO and CO competitively adsorbed on Rh(111).