View PDF VersionPrevious ArticleNext Article
 
DOI: 10.1039/C4RA08349K (Review Article) RSC Adv., 2015, 5, 949-959

Catalytic oxidation of CO on metals involving an ionic process in the presence of H2O: the role of promoting materials

Ken-Ichi Tanaka*a, Hong Hea and Youzhu Yuanb
aResearch Center for Eco-Environmental Sciences, Chinese Academy of Sciences, Beijing 100085, China. E-mail: ezz01606@nifty.com
bState Key Laboratory of Physical Chemistry of Solid Surfaces and National Engineering Laboratory for Green Chemical Productions of Alcohols-Ethers-Esters, College of Chemistry and Chemical Engineering, Xiamen University, Xiamen 361005, China

Received 8th August 2014 , Accepted 20th November 2014

First published on 20th November 2014

Abstract

A new catalytic oxidation of CO involving an ionic process in the presence of H2O is proposed on a Pt-catalyst with specific promoting materials (co-catalysts). Oxidation of CO is very slow at room temperature on ordinary Pt-catalysts such as Pt/SiO2, Pt/Al2O3, Pt/TiO2, Pt/Graphite, and Pt/carbon nano-tube (CNT), and H2 or H2O have no effect on the reaction. However, in the presence of specific co-catalysts, the oxidation of CO is markedly enhanced by H2 or H2O, so that highly selective preferential oxidation (PROX) of CO in H2 is attained. The role that co-catalysts play in the oxidation of CO enhanced by H2 or H2O was clarified by the experiments with Pt supported on CNT and carbon nano-fiber (CNF) that had Ni–MgO and FeOx at their one terminal end, respectively. Oxidation of CO was markedly enhanced by H2 on the Pt/CNT and Pt/CNF, but no enhancement was observed on the Pt/CNT-p and Pt/CNF-p, where the CNT-p and CNF-p were purified by removing Ni–MgO and FeOx. Similar enhancement of the oxidation of CO by H2 or H2O was observed on FeOx/Pt/TiO2 and FeOx/Au/TiO2, although no enhancement was observed on the Pt/TiO2 and Au/TiO2 catalysts. The in situ DRIFT spectra of the FeOx/Pt/TiO2 catalyst (Fe[thin space (1/6-em)]:[thin space (1/6-em)]TiO2[thin space (1/6-em)]:[thin space (1/6-em)]Pt = ca. 100[thin space (1/6-em)]:[thin space (1/6-em)]100[thin space (1/6-em)]:[thin space (1/6-em)]1) during reaction in a flow of (CO + O2 + H2) suggested the rate-determining slow step was HCOO(a) + OH(a) → CO2 + H2O. The oxidation of CO enhanced by H2O/D2O and H2/D2 showed a common hydrogen isotope effect of rH/rD = 1.4–1.5. The highly selective oxidation of CO in H2 on the Pt/CNT and Pt/CNF catalysts strongly suggests efficient transport of ionic intermediates from Ni–MgO or FeOx to Pt over the hydrophobic CNT and CNF surface according to the local potential gradient. According to this mechanism, selectivity in the preferential oxidation of CO in H2 is defined by the turnover number of a H2O molecule forming CO2 during its residence time on the catalyst, which is essentially different from the selectivity based on competitive adsorption and/or reaction. The role of the H2O molecule is as a kind of messenger molecule or a molecular catalyst to promote the reaction on the surface expressed by the equation n(CO + 1/2O2) + H2O → nCO2 + H2O. In this mechanism, the selectivity is given by n/(n + 1). Curious phenomena previously observed in the PROX reaction of CO in H2 on various catalysts are well explained by the mechanism including an ionic process proposed in this paper.

1. Introduction

Catalytic oxidation of CO on metals, especially on Pt, has been studied as a model of catalysis involving the reaction of adsorbed CO(a) with adsorbed O(a) or activated lattice oxygen. However, this reaction that appears simple is not so simple, as will be demonstrated in this review. The adsorption of CO on metals has been widely studied, both experimentally and theoretically. For example, it was shown that the adsorption bond of CO on metals was influenced by local electron density on the metal surface as experimentally demonstrated by King et al.1 on a K+/Ni(100) film (∼200 nm thick) by detecting radiating heat accompanying the adsorption of 50 ns pulses of 1012 molecules of CO, which was in good agreement with the bonding model given by Blyholder.2 The lack of activity of Au surface for the oxidation of CO was well explained theoretically by the model, given that it has no ability to adsorb CO. In contrast to this accepted knowledge, the unusual catalytic activity of Au particles, dependent on the particle size and support oxide used, was found by Haruta et al.3,4 That is, inactive Au particles change to an extremely active catalyst for the oxidation of CO at a crucial particle size ca. 3 nm. Haruta hypothesized that the active perimeter of the particles played a role in their activity in the reaction of adsorbed CO.5 However, the activation of Au particles smaller than a critical size of ca. 3 nm is difficult to explain simply by the perimeter length. In relation to this phenomenon, Goodman hypothesized the activity to be by a quantum effect,6 but later, Chen and Goodman showed the importance of the stability of specific materials.7

On the other hand, it is well known that the lattice oxygen becomes labile at the perimeter on Rh/Ce0.5Zr0.5O2, Pt–Rh/CeO2, and Pt–Rh/CeO2–ZrO2.8,9 Bollinger and Vannice10 performed an interesting experiment by depositing TiO2 on highly pure inactive Au powder (10 μm), and observed the activation of the Au particles for the oxidation of CO at 313 K. They postulated the formation of active sites at the Au–TiO2 interface. Farrauto et al.11 found activation of a 5 wt% Pt/γ-Al2O3 catalyst for the selective oxidation of CO in H2 on addition of FeOx and they proposed that a change in the electronic properties of the Pt particles caused by FeOx was responsible for the activity rather than the activation of the lattice oxygen of FeOx, although they observed no change in the IR spectrum of the CO adsorbed on Pt. So far, the oxidation mechanism of CO on metals has been explained based on the reaction of adsorbed CO(a) with adsorbed O(a) or labile lattice oxygen. Adsorption of CO on 1 wt% Pt/TiO2 as well as on 1 wt% Au/TiO2 is markedly influenced by the FeOx deposited on them as shown in Fig. 1, and the oxidation of CO on FeOx/Pt/TiO2 and FeOx/Au/TiO2 is markedly enhanced by H2O and/or H2. However, the rate-determining step of the oxidation of CO enhanced by H2O on the FeOx/Pt/TiO2 catalyst is neither the adsorption nor the subsequent reaction of the adsorbed CO, as discussed in this review. This fact suggests that the bonding form of adsorbed CO(a) may have no direct effect on the activity, but the adsorption ability should be indispensable.


image file: c4ra08349k-f1.tif
Fig. 1 Bridge-bonded CO peak (1844–1792 cm−1) on 1 wt% Pt/TiO2 and 1 wt% Au/TiO2 is markedly increased on the FeOx/Pt/TiO2 and FeOx/Au/TiO2 surfaces at 40 °C. The peaks at 2172 and 2118 cm−1 are gas-phase CO.13,16

So far, the oxidation of CO in H2 (in the presence of H2O) has been explained by the competitive reaction of CO(a) and H(a) with adsorbed oxygen or active lattice oxygen, but Daniells et al.36 proposed a new reaction mechanism of CO via hydroxyl-carbonyl intermediates on an Au/Fe2O3 catalyst. Tanaka et al.12–16 found unusual catalytic activity with Pt/TiO2 and Au/TiO2 covered with a large amount of FeOx (ca. 100 wt% in Fe) for the oxidation of CO in the presence of H2, that is, H2O. So far no one has observed CO adsorbed on the Au catalyst, but an IR peak appearing at 1803 cm−1 on the FeOx/Au/TiO2, assignable to bridge-bonded CO, was observed when 1 wt% Au/TiO2 was covered with a large amount of FeOx as shown in Fig. 1.13,16 Interestingly, the bridge bonded CO was also markedly increased on the Pt/TiO2 catalyst after addition of a large amount of FeOx, but the peak at 1629 cm−1 derives from a foreign species not related to the oxidation of CO, as deduced from a DRIFT spectroscopic study illustrated in Fig. 5. As will be described in more detail, adsorption of CO is very rapid compared to other elementary steps during the oxidation of CO.

To put it simply, the low temperature oxidation of CO is enhanced by H2O on FeOx/Pt/TiO2, and is different from the oxidation with O2 without H2 on FeOx/Pt/TiO2 and in the presence of H2 on Pt/TiO2. In fact, the in situ IR spectroscopy suggested the rate-determining reaction of HCOO(a) + OH(a) → CO2 + H2O on the FeOx/Pt/TiO2 catalyst. Based on the dynamics of the in situ DRIFT spectroscopy, a mechanism involving an ionic process in which ions are formed on the FeOx and are rapidly transported onto the Pt is strongly suggested, and the rate determining reaction mentioned above takes place on the Pt.13,14,16 The role of promoting materials in this oxidation of CO was proved more conclusively on Pt/CNT (carbon nano-tubes having Ni–MgO at the terminal end) and on the Pt/CNF (carbon nano-fibres having FeOx at the terminal end). If Ni–MgO and FeOx were removed from CNT (CNT-p) and CNF (CNF-p) by purification procedure, the oxidation of CO enhanced by H2O on Pt stopped. An important fact is that the Pt particles are completely separated from Ni-MgOx or FeOx on the Pt/CNT and Pt/CNF catalysts, but the promoting materials and H2O are indispensable for this new catalytic oxidation of CO.

Fukuoka et al.17 reported curious activity of Pt-nano-rods (3 nm × several hundred nm) for the selective oxidation of CO in H2 depending on their location, either inside of SiO2-nano-tubes or on the external surface of SiO2-nano-tubes as shown in Fig. 10, which was explained by the existence of a specific electronic state in the Pt-nano-rods in the SiO2-nano-tubes for the oxidation of CO(a) with O(a). Yuan et al.18,19 observed the effect of the crystalline shape of the CeO2 support in Au/CeO2 catalysts on their activity and selectivity for the oxidation of CO in H2. These phenomena are difficult to understand based on the mechanism premising the reaction of adsorbed CO(a) with adsorbed O(a) or lattice oxygen without considering the indispensable contribution of H2O, that is, a new oxidation mechanism of CO via an ionic process may take place, as presented in this paper. If the oxidation of CO proceeds by a new mechanism, as proposed in this review, the definition of the selectivity for the preferential oxidation of CO in H2 should be different from that of the traditional mechanism based on the competitive reaction of CO and H2. A meaningful definition of the selectivity in the preferential oxidation of CO in H2 is given by the turn-over number “n” of one H2O molecule forming a CO2 molecule as described by the equation n(CO + 1/2O2) + H2O → nCO2 + H2O, in the residence time of the H2O molecule on the catalyst, that is, H2O is a kind of messenger molecule or a molecular catalyst on the surface.

Many investigators have studied the activation or passivation effects of oxide layers deposited on metals. A well-known phenomenon is the growth of a passive oxide layer on precious metals by reduction, which was named the SMSI (strong metal support interaction) effect by Tauster et al.20 In this case an inactive TiO2−x layer is formed in a reducing atmosphere over precious metals supported on TiO2. Using STM, Netzer et al.21 demonstrated the wetting growth of an oxide layer on a metal with a V2O3 layer on a Pd(111) surface. Formation of active oxide layers on metals was also observed. Discontinuous changes in the activity of an oxide layer depending on its thickness were first observed by Lambert et al.22 for CeO2 deposited on a Pt(111) surface. One monolayer of CeO2 on the Pt(111) surface is inactive for the adsorption of CO as well as for the oxidation of CO; however, deposition of more than two monolayers of CeO2 abruptly gives superior activity in the oxidation of CO. Similar activation depending on the number of layers was observed for FeOx deposited on a Pt(111) surface by Shaikhutdinov et al.,23 that is, the structural change of the FeOx layers causes a change in the reactivity of the lattice oxygen. It is worth noting that the lattice activation was observed without H2O, and it is extremely interesting to consider the effect of H2O on the activation of lattice oxygen in order to understand the new catalytic oxidation of CO enhanced by H2O.

2. Co-catalysts for the oxidation of CO enhanced by H2O

Catalytic oxidation of CO with O2 on Pt has been explained as a model of the reaction of adsorbed species depending on the coverage of CO(a) and O(a). In fact, all adsorbed CO(a) and O(a) on the Pt surface appear to contribute to the catalytic reaction as Golchet and White24 demonstrated on a Pt foil by changing the ratio of O2 to CO pressure at 180–212 °C, which is apparent uniformity by rapid diffusion of adsorbed species. It is also known that the oxidation of CO on Pt catalysts is very slow at temperatures lower than 80 °C because of the strong adsorption of CO, but the PROX reaction of CO in H2 is catalyzed on Pt with specific promoting materials even at room temperature. So far, the role of co-catalysts has been explained as weakening the adsorption of CO on Pt or improving reactivity of CO and/or O(a) on Pt or of lattice oxygen of the support or promoting materials. As pointed out in this paper, the role of co-catalysts is difficult to understand based on these traditional explanations. We confirmed that the role of promoting materials is to provide a new reaction path for the oxidation of CO in the presence of H2O by using CNT and CNF with promoting materials only at their one terminal end. Yoshida et al.25 were the first to successfully observe the growing process of multi-wall CNT from a nano-size Fe3C particle by heating in C2H2 by in situ TEM, and it was shown that the nano-size catalyst surface was covered with graphene first, and expelled a multi-wall CNT, so that the size of the multi-wall CNT was decided by the size of the Fe3C particle. As the left TEM image of Fig. 2(a) shows, the width of CNT prepared by the Ni–MgO catalyst is narrower than the particle remaining at the terminal end of the CNT. This fact may indicate that the size of the active Ni catalyst (probably nickel carbide) is smaller than the particle size of Ni–MgO remaining at the terminal end of CNT. The Pt-particles are clearly smaller than the width of the CNT, and they are completely separated from the Ni–MgO or FeOx localized at the terminal end of the CNT. However, the Ni–MgO or FeOx at the terminal end of CNT is indispensable for the catalytic oxidation of CO enhanced by H2O on Pt/CNT and Pt/CNF catalysts. As shown in Fig. 2(a), a highly selective oxidation of CO took place in excess H2 at room temperature on the Pt/CNT catalyst. In contrast, if the Ni–MgO was removed from the CNT by purification (CNT-p), low temperature oxidation of CO in H2 was no longer catalyzed on the Pt/CNT-p catalyst.12,26 This result proves that the Ni–MgO is indispensable for the selective oxidation of CO in H2, but it is not necessary for it to be in contact with the Pt particles. One should be reminded that the Pt/CNT and Pt/CNT-p showed almost equal low activity for the oxidation of CO with O2 in the absence of H2. Taking these results into account, the hydrogen isotope effect for the oxidation of CO by H2/D2 and H2O/D2O was studied, and it was confirmed that H2/D2 and H2O/D2O gave equal hydrogen isotope effects for the oxidation of CO. Accordingly, we can conclude that H2O is the molecule responsible for the PROX of CO in H2.
image file: c4ra08349k-f2.tif
Fig. 2 (a) TEM images of CNT, Pt/CNT, and Pt/CNT-p, and the activity and selectivity of Pt/CNT and Pt/CNT-p for the PROX reaction of CO in H2. Conversion of CO (solid) and O2 (open) attained in a flow of 100 mL min−1 of (CO/O2/H2 = 1.5/1.5/42.0 mL min−1) with a balance of N2 on 15 wt% Pt/CNT and 15 wt% Pt/CNT-p (0.8 g). (b) Conversion of CO attained in a flow of (CO/O2/H2/N2) = 3.0/1.5/20.0/78.5 (mL min−1) on 5 wt% Pt/CNF (VGCF-X) (image file: c4ra08349k-u1.tif) and 5 wt% Pt/CNT-p (image file: c4ra08349k-u2.tif), 5 wt% Pt/VGCF (image file: c4ra08349k-u3.tif), 5 wt% Pt/VGCF-S (image file: c4ra08349k-u4.tif). (c) Conversion of CO (solid) and O2 (open) attained in a flow of (CO/O2/H2/N2) = 3.0/1.5/20/75.5 (mL min−1) on 15 wt% Pt/Graphite (0.8 g) (image file: c4ra08349k-u5.tif, image file: c4ra08349k-u6.tif) and 5 wt% Pt/Graphite doped with 5 wt% of Ni–MgO (image file: c4ra08349k-u7.tif, image file: c4ra08349k-u8.tif). PROX reaction of CO on Pt/Graphite and Pt/(Graphite + Ni–MgO).12,26

As shown in Fig. 2(a), nearly 100% conversion of CO was attained at a flow of 100 mL min−1 of (CO + H2 + O2 + N2) (CO[thin space (1/6-em)]:[thin space (1/6-em)]O2[thin space (1/6-em)]:[thin space (1/6-em)]H2 = 1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]28) over 15 wt% Pt/CNT (the amount of Pt was equivalent to 0.8 g of 0.75 wt% Pt/TiO2).27 If the Ni–MgO was removed from the CNT by chemical purification (CNT-p), the Pt/CNT-p lost the activity for the oxidation of CO in H2 (H2O in practice) as shown in Fig. 2(a). The essential role of Ni–MgO as a co-catalyst for the oxidation of CO enhanced by H2 (H2O in practice) was confirmed by using graphite, that is, the oxidation of CO in H2 on Pt/Graphite and Pt/Graphite with Ni–MgO.27 As shown in Fig. 2(c), Pt/Graphite was not active for the oxidation of CO at temperatures lower than 120 °C, but Pt supported on Ni–MgO-doped graphite was highly active in the temperature range of 40–120 °C. From these results, we can confidently conclude that Ni–MgO and FeOx are indispensable materials in the oxidation of CO enhanced by H2, actually by H2O, but the Pt-particles and Ni–MgO and FeOx do not need to be in contact with one another on the surface. To realize the rapid oxidation of CO enhanced by H2O on Pt/CNT and Pt/CNF, the rapid transport of intermediate species is necessary between the Ni–MgO and Pt-particles over the nano-tubes. This transport mechanism on the surface is crucial to understanding the oxidation mechanism of CO promoted by H2O.

To more strictly confirm the contribution of the co-catalysts separated from the Pt, the reaction was performed using CNF (VGCF-X; ϕ15 nm × 4 μmL of Showa Denko Co.), which were prepared using a ferrocene catalyst in the gas phase. The FeOx remaining at the terminal end of the CNF was completely removed by heating the fibers at 2900 °C (CNF-p). Three lots of CNF, VGCF; ϕ150 nm × 8 μmL, VGCF-H; ϕ150 nm × 6 μmL, and VGCF-S; ϕ80 nm × 10 μmL, were used. As shown in Fig. 2(b), the Pt/CNF (VGCF-X) catalyst was very active for the PROX reaction of CO in H2, and nearly 100% selective oxidation of CO (CO/O2 = 2/1) was attained in a flow of CO (3.0 mL min−1) + O2 (1.5 mL min−1) + H2 (20.0 mL min−1) + N2 (78.5 mL min−1) at temperatures lower than 80 °C. As expected, the Pt-catalyst supported on the CNF-p having no FeOx (Pt/VGCF, Pt/VGCF-H and Pt/VGCF-S) had very poor catalytic activity for the oxidation of CO at temperatures lower than 100 °C. Although Ni–MgO and FeOx are separated from the Pt particles on the CNT and CNF, these promoting materials are indispensable in the catalytic oxidation of CO on Pt enhanced by H2O. Accordingly, the total mechanism for the oxidation of CO promoted by H2O is accomplished by the effective transport of the intermediates between Ni–MgO or FeOx and Pt particles.28

A similar promoting effect of H2 and H2O on the oxidation of CO was observed on Pt/TiO2 overlaid with a large amount of FeOx.13–16 As shown in Fig. 3(a), the oxidation of CO was not enhanced by H2O (or H2) on a 1 wt% Pt/TiO2 catalyst, but was markedly enhanced on the FeOx/Pt/TiO2 catalyst at 60 °C by H2 and H2O as shown in Fig. 3(b) and (c),13,15 where the FeOx/Pt/TiO2 was prepared by overlaying a large amount of FeOx (ca. 100 wt% in Fe) on the 1 wt% Pt/TiO2. If H2 or H2O were removed from the gas phase, the oxidation of CO was quickly suppressed as shown in Fig. 3(c).


image file: c4ra08349k-f3.tif
Fig. 3 (a) The oxidation of CO with O2 on a 1 wt% Pt/TiO2 catalyst is not enhanced by H2O at 60 °C. (b) The oxidation of CO in a flow of [CO (3 mL min−1) + O2 (1.5 mL mi−1) + N2 (95.5 mL min−1)] on a FeOx/Pt/TiO2 catalyst is markedly enhanced by H2 (15.0 mL min−1) at 60 °C. (c) Enhancement of the oxidation of CO by H2O on FeOx/Pt/TiO2 is reversible at 60 °C.13

These results strongly suggest that the role of H2O is not the activation of the Pt or FeOx, but provides a new catalytic oxidation reaction of CO(a) as discussed in detail below. It is worth noting that the Pt/TiO2 and FeOx/Pt/TiO2 used here have almost equal activity for the oxidation of CO with O2 at 60 °C in the absence of H2 or H2O, that is, the H2O molecule enhances the oxidation of CO only on the FeOx/Pt/TiO2 surface but not on the Pt/TiO2, although their activity for the oxidation of CO in the absence of H2 or H2O is almost equal at temperatures higher than 100 °C. Therefore, the role of H2O molecules in reaction (ii) in eqn (1) is like a messenger molecule or a molecular catalyst.

 
image file: c4ra08349k-t1.tif(1)

Since Langmuir's work in 1921,29 the catalytic oxidation of CO with O2 on Pt catalysts has been explained by the mechanism of reaction (i), that is, the reaction of adsorbed CO(a) with adsorbed O(a) or labile lattice oxygen. As mentioned above, Golchet and White24 showed that all adsorbed CO(a) and O(a) on the Pt-foil during reaction depended on the ratio of P(O2)/P(CO), and all adsorbed species contribute to the oxidation reaction at 180 °C and 212 °C, but steady state adsorption on the Pt surface during catalysis depended on the ratio of P(O2)/P(CO). One should be reminded that this result does not mean homogeneous activity of the Pt surface and/or homogeneous reactivity of the adsorbed CO(a) and O(a) species, because rapid surface diffusion of adsorbed species smears the heterogeneity of the Pt surface and the reactivity of the adsorbed species.

When the oxidation of CO and H2 proceeds independently on Pt surface as premised in the L–H mechanism, the two species neither enhance nor inhibit oxidation of the other, rather the reaction of CO and H2 is controlled by the competitive mechanism of the adsorption and reaction of CO(a) and H(a). Therefore, if the activity of a Pt catalyst is changed by some additives or support materials, this should be explained as the result of a change in CO(a) or O(a) caused by the change of Pt or by the formation of active lattice oxygen. However, the role of the support or promoting materials in catalysis is not so simple as is rationalized by the reactivity of adsorbed species or lattice oxygen. To illustrate this point, here we take up the results of Korotkikh and Farrauto et al.11 in experiments on the Pt/γ-Al2O3 and FeOx/Pt/γ-Al2O3 catalysts. The activity of Pt/γ-Al2O3 for the oxidation of CO in H2 was markedly improved by the addition of FeOx, but the IR spectrum of CO adsorbed on FeOx/Pt/γ-Al2O3 showed no difference compared to that of CO on Pt/γ-Al2O3. They explained the activation of the FeOx/Pt/γ-Al2O3 catalyst as a synergy effect of Pt and FeOx, that is, as they premised the reaction of adsorbed CO with adsorbed oxygen or lattice oxygen, and the difference of the activity between Pt/γ-Al2O3 and FeOx/Pt/γ-Al2O3 was explained by the term synergy effect, but the physical or chemical meaning of the synergy effect was not made clear.

In contrast to the CO adsorbed on Pt/γ-Al2O3 and FeOx/Pt/γ-Al2O3, the CO adsorbed on FeOx/Pt/TiO2 was entirely different from that on the Pt/TiO2. That is, linearly bonded CO(a) becomes predominant on the Pt/TiO2 catalyst but bridge-bonded CO(a) (1809 cm−1) is the main adsorbed species on the FeOx/Pt/TiO2 catalyst. Furthermore, the oxidation of CO was markedly enhanced by H2 on the FeOx/Pt/TiO2 catalyst as shown in Fig. 3(b) and (c), but the reaction on the Pt/TiO2 catalyst was not enhanced by H2 as shown in Fig. 3(a).13,14 It should be emphasized that the bonding of CO is different on FeOx/Pt/TiO2 and Pt/TiO2, but we could not realize the difference in apparent reactivity for the two forms of CO, bridge-bonded CO and linearly bonded CO, in the catalytic oxidation of CO enhanced by H2 or H2O on FeOx/Pt/TiO2. This fact convinces us that the rate-determining step of the oxidation of CO enhanced by H2O on the FeOx/Pt/TiO2 catalyst is not the oxidation of adsorbed CO(a) with adsorbed O(a) or lattice oxygen. As argued in this paper, the oxidation of CO enhanced by H2O, reaction (ii), is a new oxidation reaction via intermediates [Xi(a)] in eqn (1). In fact, the oxidation of CO enhanced by H2/D2 showed an equal hydrogen isotope effect of rH/rD = 1.4–1.5 on FeOx/Pt/TiO2 and Pt/CNT as well as on Au/CeO2 and FeOx/Au/TiO2 catalysts, which is equal to the isotope effect when enhanced by H2O/D2O,14,18,30 which indicates that the H2O molecule is responsible for the enhancement in either case. To understand the oxidation mechanism of CO promoted by H2O, it is essential to deduce the rate-determining slow step and to identify the intermediates [Xi], as discussed in the following section.

3. Intermediates and rate-determining slow step

When a catalytic reaction proceeds at a steady-state, chemical potentials and the amounts of the intermediates can be illustrated by a series-connected water tank model shown in Fig. 4 as illustrated by Tamaru.31
image file: c4ra08349k-f4.tif
Fig. 4 A series-connected water tank model showing the chemical potential of intermediates (ξ1) and (ξ2) in a steady-state reaction of (A) to (B).31 The model shows a case when step (iii) is the rate-determining step.

In situ DRIFT spectra attained on a FeOx/Pt/TiO2 catalyst in a flow of (CO + O2 + N2) and (CO + H2 + O2 + N2) at 60 °C are shown in Fig. 5(a) and (b) in blue. CO(a) peaks of almost equal intensity were observed in a flow of (CO + O2 + N2) and (CO + O2 + H2 + N2) on FeOx/Pt/TiO2, although the reaction rates were quite different, as shown in Fig. 3(a) and (b). When the CO was removed from a flow of (CO + O2 + H2 + N2), the IR peaks for CO(a) were very quickly erased as shown in Fig. 5(b), but the decrease in the CO(a) was very slow in the absence of H2 as shown in Fig. 5(a). As it is known that the desorption of CO(a) on Pt is very slow at room temperature, the rapid decrease of the CO(a) is caused by its rapid reaction in the presence of H2, that is, in the presence of H2O. Therefore, the adsorption of CO, step (i), is very rapid, and the adsorption of CO on the FeOx/Pt/TiO2 catalyst may be nearly a dynamic saturation in (CO + O2) or (CO + O2 + H2). If step (ii) is the rate-determining slow step, (ξ1), that is, CO(a) is detectable but (ξ2) may be lower than the detectable level in steady-state reaction. If step (iii) is the rate-determining slow step, both (ξ1) and (ξ2) may be detectable on a steady-state catalyst surface. Therefore, our goal is to detect intermediate (ξ2) on a steady-state catalyst during the oxidation reaction of CO, as will be discussed below.


image file: c4ra08349k-f5.tif
Fig. 5 Time-resolved in situ DRIFT spectra of CO(a) attained on a FeOx/Pt/TiO2 catalyst at 60 °C. (a) CO was removed from (CO + O2 + N2); (b) CO was removed from (CO + H2 + O2 + N2). A series of spectra were recorded at time intervals: (1) is immediately after the removal of CO, and (2)–(6) are 5, 10, 15, 20, 30, and 45 min after CO removal.28,30

In situ DRIFT spectra include every species existing on the catalyst surface. Therefore, the spectrum responsible for the reaction intermediates must be extracted from the spectrum. One promising way to do this is to extract the part of the spectrum that changes on abrupt removal of CO from the reactant gas. In this procedure, the spectrum attained on a steady-state catalyst is subtracted as a back-ground, and the spectra that change with time during the reaction are extracted. In this way, the change in the spectrum with time caused by removing CO from the reactant gas was extracted. If step (iii) in eqn (2) is the rate-determining slow step, the intermediates of (ξ1) and (ξ2) are nearly in dynamic equilibrium in the steady-state reaction. As mentioned above, the oxidation of CO enhanced by H2 or H2O exhibits an equal hydrogen isotope effect of rH/rD = 1.4–1.5 for H2/D2 and H2O/D2O, which suggests that H2O is responsible for the oxidation of CO enhanced by H2.28,30 The isotope effect is caused by either the kinetic isotope effect given by the rate constant (kH > kD) or the thermodynamic pre-equilibration of the intermediates (concentration) proceeding with the rate-determining step (kH/kD).

Time-resolved DRIFT spectra obtained by removing CO from a steady flow of (CO + H2 + O2) are shown in Fig. 6, which reflect the decrease of intermediates from the FeOx/Pt/TiO2 catalyst due to the reaction, that is, the spectra correspond to the decrease of (ξ1) and (ξ2) in eqn (2) by the reaction. Large negative peaks at 2172 and 2119 cm−1 are due to gas phase CO decreased by the removal. At the same time, linearly-bonded CO(a) (2065, 2067 cm−1) and bridge-bonded CO(a) (1836 cm−1), ξ1, are rapidly decreased by changing to (ξ2) on the FeOx/Pt/TiO2 catalyst by the reaction. That is, ξ1 CO(a) is very rapidly changed to (ξ2) in eqn (2) because the desorption of CO is very slow. Intermediates (ξ2) are also decreased with time by the reaction, which is observed as the growth of large negative peaks at 1522, 1354 and 1296 cm−1 and a large broad band from 3000–3800 cm−1 centred at ca. 3400 cm−1 in Fig. 8.28,30 Large negative peaks at 1522 and 1354 cm−1 with a shoulder at 1296 cm−1 are assignable as HCOO−(a), and the broad band from 3000–3800 cm−1 centred at 3400 cm−1 is assigned as the highly reactive OH(a). These two are the key intermediates (ξ2) in eqn (2). The most remarkable features are the broad band centred at 3400 cm−1 assigned as OH, and the rapid growth of large negative peaks with time.


image file: c4ra08349k-f6.tif
Fig. 6 Time-resolved in situ DRIFT attained by removing CO from a flow of (CO + H2 + O2 + N2) at 60 °C. The spectra were obtained at the 30 s, 1 min, 5 min, 10 min, 20 min, 30 min, and 40 min. after the removal of CO.28,30

It is worth noting that the large broad band assigned as OH was rapidly decreased from the surface by the reaction, but the characteristic scissors mode of adsorbed H2O at 1650 cm−1 was not observed. This fact indicates that the amount of H2O on the catalyst changed little on removal of gas phase CO, although H2O formation should be increased by removing CO from (CO + H2 + O2). The lack of an increase or decrease in the scissors mode of adsorbed H2O (1650 cm−1) indicates that neither the increase nor the decrease of adsorbed H2O occurs on removal of gas phase CO. This fact indicated saturation of the catalyst surface with H2O(a) in the steady-state reaction of (CO + H2 + O2). Taking these results into account, the broad band centred at 3400 cm−1 is not adsorbed H2O(a) but the reactive OH(a) species, which undergo reaction with HCOO(a) at the rate-determining step as described by step (iii) in eqn (2).13,26 In fact, the HCOO(a) and OH(a) peaks at 1522 and 3355 cm−1 (negative growth of the peaks) rapidly decreased at the same rate, as shown in Fig. 7(b).


image file: c4ra08349k-f7.tif
Fig. 7 (a) Change of in situ DRIFT spectra of the FeOx/Pt/TiO2 catalyst with time when CO was removed from the gas phase in the presence of H2 (blue) and in the absence of H2 (red). (b) Decrease of HCOO(a) (1522 cm−1) and OH (at 3355 cm−1) (solid lines) with time attained by removing gas phase CO in the presence of H2. Broken line indicates slow decrease of bicarbonate peak (1632 cm−1) with time.

If the reaction of HCOO(a) with OH(a) on the Pt surface is the rate-determining slow step as deduced, the concept for the selectivity of preferential oxidation of CO in H2 should be different from the traditional selectivity based on the competitive reaction of adsorbed CO(a) and H(a) with O(a). As was proved above, even if Pt particles are not in contact with the Ni–MgO or FeOx on the CNT or CNF, the oxidation of CO expressed by reaction (ii) in eqn (1) occurs on the Pt, where the role of H2O is like a molecular catalyst working on the surface. Taking this mechanism into account, the selectivity for the preferential oxidation of CO in H2 is given by the turn-over number “n” of H2O molecules, that is, a meaningful selectivity is the formation of “n” CO2 molecules by one H2O molecule in a steady-state reaction. According to this definition, if one H2O molecule produces 10 molecules of CO2 during its stay on the catalyst surface, the selectivity for oxidation of CO in H2 (H2O is formed by the oxidation of H2) becomes 91% (10/11). An interesting question is why the oxidation of CO is enhanced by H2O on hydrophobic Pt/CNT and Pt/CNF surfaces. Surface diffusion or surface transport of adsorbed species is essential in order to explain various phenomena taking place on the surfaces including catalysis, but its embodiment is quite difficult. One interesting example is the reduction of pale yellow WO3 powder with H2 in the presence of Pt/Al2O3 as shown by Boudart et al.32 They showed that well-evacuated WO3 did not change its colour to blue in H2 even when mixed with Pt/Al2O3 powder. However, if a small amount of H2O or alcohols was added to the H2, pale yellow WO3 was readily changed to blue HxWO3 by H2 at 50 °C, and the efficiency of the promoting molecules depends on their proton affinity. Taking these results into account, the role of H2O was explained by an ionic process on the surface, that is, adsorbed H(a) on Pt is ionized in the presence of H2O, H(a) + H2O → H3O+ + e, and H3O+ diffuses over the surface and reacts on the WO3 with an electron transported through the bulk to form blue HxWO3, which is described as WO3 + H3O+ + e → HxWO3 + H2O. In this case, H2O is an indispensable molecule for forming the H3O+ ion on the Pt surface, and the H3O+ ion is transported to WO3. This process is quite interesting in considering the role of Ni–MgO and FeOx on the Pt/CNT and Pt/CNF catalysts.

From the dynamics of the DRIFT spectroscopy, the reaction of HCOO(a) with OH(a) was deduced as the rate-determining slow step of the oxidation of CO enhanced by H2O on the FeOx/Pt/TiO2 catalyst. One question regards the mechanism for the rapid formation of the HCOO intermediate on the FeOx/Pt/TiO2 catalyst. The formation of HCOOH by the reaction of CO with Ca(OH)2 is known as a long-established industrial process. Therefore, a question is where the HCOO− is formed, either on the FeOx or Pt surface by the reaction of CO with OH− ion. One important result was obtained in Fig. 5(b), that is, the CO(a) adsorbed on Pt of the FeOx/Pt/TiO2 was very quickly erased by the reaction in the presence of H2. This fact strongly suggests rapid transport of OH− ion from FeOx to Pt surface, and reacts with adsorbed CO on the Pt to form HCOO(a) + electron (e). The released electron reacts with H3O+ to form highly reactive OH(a) on the Pt, and the reaction of OH(a) with HCOO(a) is preponderant over the ordinary reaction of CO(a) with O(a) on Pt catalysts at low temperature. An overall reaction mechanism is schematically described in Fig. 8.


image file: c4ra08349k-f8.tif
Fig. 8 Oxidation mechanism of CO enhanced by H2O via ionic process. The upper illustration shows potential gradient between Pt and FeOx made by the population of H+ and OH− ions, and OH− ion transport from FeOx to Pt and HCOO− is formed by the reaction with CO. The lower part describes the steady reaction mechanism on the dual functional catalyst with FeOx and Pt. Selectivity for the preferential oxidation of CO in H2 is given by n/(n + 1), where “n“ is the number of CO2 molecules produced by repeated contribution of one H2O molecule staying on the catalyst.

Local potential gradient on the dual functional catalyst and the oxidation of CO on it enhanced by H2O

It is evident that the oxidation of CO taking place in the presence of H2O on FeOx/Pt/TiO2, Pt/CNT, and Pt/CNF catalysts is a new oxidation reaction of CO via ionic intermediates such as HCOO−. To accomplish this mechanism, effective transport of ionic intermediates from FeOx to the Pt surface over the support surface is indispensable. The local potential gradient formed on the surface with heterogeneous local charge between FeOx and the Pt particles may be the driving force for efficient transport of ions such as OH− and H3O+. As will be discussed below, the ratio of OH−/H+ ions on Ni–MgO or FeOx and Pt is the origin of the local potential gradient between the FeOx or Ni–MgO and Pt, keeping overall charge neutrality on the support. If we consider hydrated metal ions in aqueous solution, the ionization equilibrium is given by the balance of proton activity of the hydrated H2O molecules of the metal ion and aqueous solution as described by eqn (3).
 
M(H2O)nα+ + H2O ⇄ M(H2O)n−1(OH)(α−1)+ + H3O+ (3)

It has been considered that the pKa of metal ions with different valence (α) cannot be compared on a universal scale. However, a universal equation (i) of eqn (5) was derived for metal ions in aqueous solution by defining “the electronegativity of metal ions (χα)” given by eqn (4).33

The electronegativity of metal ions (χα) is defined by eqn (4), as briefly explained next. An original concept of the electronegativity was given empirically for the elements to understand the bond strength between hetero-atoms, A2 + B2 → 2A–B, by Pauling. A similar concept was derived theoretically by Mulliken in an equation χ0 = (I + A)/2, where “I” is the ionization energy of a neutral M0 to M+1 ion and “A” is the electron affinity of a neutral element defined as the formation energy of M−1 ion from M0. If the ionization energy of metal ions can be expressed by a quadratic equation with respect to charge, Mulliken's equation (I + A)/2 is the tangent of the ionization energy curve (a kind of chemical potential of electron) of neutral elements at charge zero as pointed out by Iczkowski et al.34 This idea was extended to metal ions (Mα+) by Tanaka et al.33 as the concept of electronegativity of metal ions (χi) according to eqn (4).

 
image file: c4ra08349k-t2.tif(4)
where χo is the electronegativity of neutral metal atoms expressed by the Pauling electronegativity scale. As shown in Fig. 9(a) and (b), the proton activity of hydrated metal ions (Mα+) is well expressed by the electronegativity of metal ions (χα) in formula (i) of eqn (5).
 
image file: c4ra08349k-t3.tif(5)

If a metal oxide is immersed in aqueous solution, the hydrated oxide is ionized to create proton activity in equilibrium with that of the aqueous solution, which is described by the equation [Oxide(H2O)n + H2O ↔ Oxide(H2O)n−1(OH)− + H3O+]. As a result, the hydrated oxide has charge depending on the proton activity of the hydrated surface with respect to that of the aqueous solution. When the proton activity (pH) of the solution becomes equal to that of the hydrated oxide surface, the charge over the hydrated oxide becomes neutral (zero charge), which is referred to as the zero point of charge (ZPC). Therefore, the proton activity of the hydrated oxide surfaces is expressed by the ZPC of the oxides, and it was shown that the ZPC of the oxide surface can be expressed by a linear relation with respect to the electronegativity of the metal ions (χα) as shown in Fig. 9(b), which is expressed by equation (ii) in eqn (5).


image file: c4ra08349k-f9.tif
Fig. 9 (a) Ionization constant (pKa) of aqueous metal ions in water at 25 °C depending on the electronegativity of +α valence ion (χα). (b) The pH of water providing ZPC of oxide surfaces plotted against the electronegativity of ions.33,35

As was deduced, the FeOx/Pt/TiO2 surface during the oxidation of CO in the presence of H2 is saturated with H2O, because the in situ IR spectrum showed neither an increase nor a decrease of the scissors peak of adsorbed H2O when the CO in gas phase was removed. If the FeOx surface is saturated with H2O molecules during reaction, an ionization equilibrium may be established on the catalyst surface. Taking account of the proton activity of the hydrated FeOx and Pt surface, excess negative charge on FeOx creates an equilibrium distribution of ions, OH−/H+ > 1. To maintain the overall neutrality of the surface, the distribution of ions creates a local potential gradient, δ+ and δ−, as shown schematically in Fig. 8. The local potential gradient between the FeOx and Pt is responsible for the rapid and efficient transport of OH− ions from FeOx onto Pt.

Unusually selective oxidation of CO in H2 has been reported on several catalysts, but it is difficult to rationalize by the traditional mechanism based on the reaction of adsorbed CO(a) with adsorbed O(a) or lattice oxygen. If we consider the ionic process such as deduced in this section, previously observed unusual activity and selectivity in the oxidation of CO in H2 are well rationalized as discussed in the next section.

4. Unusual effect of H2O molecule on catalytic oxidation of CO

Fukuoka et al.17 reported curious activity of Pt-nano-rods (3 nm × several hundred nm) in the oxidation of CO in H2 depending on their location, either inside of SiO2-nano-tubes or on the external surface of SiO2-nano-tubes. As shown in Fig. 10, ca. 100% conversion of CO (CO/O2 = 1) was attained at ca. 80 °C by flowing a mixture of (CO[thin space (1/6-em)]:[thin space (1/6-em)]O2[thin space (1/6-em)]:[thin space (1/6-em)]H2[thin space (1/6-em)]:[thin space (1/6-em)]N2) (1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]93[thin space (1/6-em)]:[thin space (1/6-em)]5 vol.%) at 1200 h−1 GHSV through the Pt-nano-rods prepared inside of SiO2-nano-tubes. In contrast, the Pt-nano-particles supported on the external surface of SiO2-nano-tubes showed very low activity at temperatures lower than 100 °C, which was almost equal to the activity of ordinary Pt/SiO2 and Pt/Al2O3 catalysts as shown in Fig. 10. This unusual catalytic activity of the Pt-nano-rods inside the SiO2-nano-tubes in the oxidation of CO was explained by formation of a specific state of Pt inside the SiO2-nano-tubes, that is, the Pt undergoes specific activation for the oxidation of adsorbed CO(a) with O(a), but they did not consider the effect of H2O. As the Pt nano-rod only exists inside the SiO2-nano-tubes, H2O and CO2 are only formed inside of the SiO2-nano-tubes, and the molecules produced inside SiO2-nano-tubes are effused out through the 3 nm exit of the SiO2-nano-tubes. The effusion rate is known to be in inverse proportion to the square root of the molecular weight. In effusion of free CO2 and H2O molecules through a pin-hole, the ratio is H2O[thin space (1/6-em)]:[thin space (1/6-em)]CO[thin space (1/6-em)]:[thin space (1/6-em)]O2[thin space (1/6-em)]:[thin space (1/6-em)]CO2 = 1[thin space (1/6-em)]:[thin space (1/6-em)]3.0[thin space (1/6-em)]:[thin space (1/6-em)]3.7[thin space (1/6-em)]:[thin space (1/6-em)]4.0[thin space (1/6-em)]:[thin space (1/6-em)]4.7. Accordingly, the gas phase inside the SiO2-nano-tubes should be considerably different from that outside the SiO2 nano-tubes. In particular, H2O molecules will stay for a far longer time on the inner surface of the SiO2-nano-tubes by wetting the hydrophilic SiO2 surface and clustering with hydrogen bonds. In contrast, in the case of Pt supported on the outer surface of SiO2 nano-tubes, H2O is formed on the outer surface. In a steady flow of reactant gas, the H2O pressure is far lower than that inside of the Pt/SiO2 nano-tubes. Therefore, if the oxidation of CO is enhanced by H2O on Pt/SiO2, it will prevail over ordinary oxidation of CO(a) with O(a) on the Pt held inside SiO2 nano-tubes at low temperature. That is, the observed result is explainable without assuming a curious state of Pt. If we measure an in situ IR spectrum during the reaction of CO in H2, we could find the difference of the oxidation mechanism of CO on Pt-nano-rod on the outer surface of SiO2-nano-tubes and on the inside SiO2-nano-tubes.
image file: c4ra08349k-f10.tif
Fig. 10 Model of nano-tube silica and conversion of CO in the PROX reaction in excess H2. Conversion of CO on Pt-rods supported within SiO2-nano-tubes (image file: c4ra08349k-u9.tif) and on the external surface of SiO2-nano-tubes (image file: c4ra08349k-u10.tif) in a flow of (CO[thin space (1/6-em)]:[thin space (1/6-em)]O2[thin space (1/6-em)]:[thin space (1/6-em)]H2[thin space (1/6-em)]:[thin space (1/6-em)]N2 = 1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]93[thin space (1/6-em)]:[thin space (1/6-em)]5 vol.%) at GHSV = 1200 h−1 over 100 mg catalyst (5 wt% Pt). Reference activities of Pt/SiO2 (image file: c4ra08349k-u11.tif) and Pt/Al2O3 (+) are indicated with a broken line.17

When the catalyst activity of metals is changed by the support oxide or additives, the activity is explained by the change of active area on metals or active lattice oxygen, but the same reaction mechanism is usually premised. A typical example is the SMSI effect. However, in the case of FeOx/Pt/TiO2, if a Pt/TiO2 catalyst was overlaid with a large amount of FeOx, the activity for the oxidation of CO in H2 was markedly improved by the addition of FeOx as shown in Fig. 3, and in the case of Pt/CNT and Pt/CNT-p, the activity is markedly improved in the presence of Ni–MgO separated from the Pt on CNT. These results strongly indicate the new functional role of additives.

To understand these phenomena, the ionic processes were deduced in the catalytic oxidation of CO enhanced by H2O on the Pt-catalysts with specific promoting materials and on FeOx/Au/TiO2 catalyst.16 Tanaka et al.16 showed that large Au particles supported on TiO2 are inactive for the oxidation of CO, but they change to being an active catalyst for the oxidation of CO in the presence of H2O after being loaded with a large amount of FeOx. Bollinger and Vannice10 also reported an interesting activation of highly pure Au powder (10 μm) for the oxidation of CO at 40 °C by adding TiO2, and Daniells et al.36 suggested the formation of hydroxyl-carbonyl intermediates on an Au/Fe2O3 catalyst.

Recently, Yuan et al.18,19 found an interesting phenomenon whereby the performance of Au/CeO2 catalysts depended on the crystalline shapes of the CeO2, that is, the selectivity for the oxidation of CO in H2 depended on the crystalline shapes in the order polyhedral ≈ rods > cubes, although O2 consumption in (H2 + CO + O2) showed little influence by the CeO2 morphology. They found differences in the ratio of Au+ and Au3+ ions depending on the CeO2 crystal shapes such that (Au+ + Au3+)/(Au0 + Au+ + Au3+) = 25.3[thin space (1/6-em)]:[thin space (1/6-em)]46.0[thin space (1/6-em)]:[thin space (1/6-em)]61.5 (%) for (cube)[thin space (1/6-em)]:[thin space (1/6-em)](polyhedron)[thin space (1/6-em)]:[thin space (1/6-em)](rod). It is known that the selectivity for the oxidation of CO depends on the (Au+[thin space (1/6-em)]:[thin space (1/6-em)]Au3+) ratio but that it has little effect on the total consumption of O2 (activity). It is a noteworthy fact that the oxidation of CO enhanced by H2/D2 showed an equal hydrogen isotope effect of rH/rD = 1.44 on the three Au/CeO2 catalysts with different crystalline morphologies of CeO2.19 These results suggest that the oxidation of CO on the Au/CeO2 catalyst may proceed via an ionic process, and the reaction is given by reaction (ii) of eqn (1), and that the selectivity is given by the turn-over number of H2O molecules on the Au/CeO2 surface, which takes on a larger “n” value on CeO2 surfaces with larger (Au+ + Au3+)/(Au0 + Au+ + Au3+) values.

Lastly, it should be mentioned that adsorbed HCOO(a) species are very often observed on various catalysts in the oxidation reaction of organic molecules as well as in the water gas shift reaction, but this does not mean similar catalysis, because a catalytic reaction is accomplished by combining all elementary processes required for the catalysis on the surface, that is, similarity of one intermediate does not mean the similarity of catalysis. The Horiuti–Polanyi mechanism is a typical example of confusion, in which the isomerization and the hydrogenation of olefins were premised as occurring by the same catalysis via common alkyl intermediates. However, it was proved that the active sites promoting the hydrogenation of olefins were clearly different from the sites for the isomerization of olefins, although both reactions proceed via alkyl intermediates.37 From this view point, the water gas shift reaction is undoubtedly different from the irreversible oxidation of CO with O2 enhanced by H2O, reaction (ii) in eqn (1), even if HCOO(a) intermediates are observed in both reactions.

5. Conclusions

(1) A new catalytic oxidation of CO enhanced by H2O via an ionic process: oxidation of CO with O2 on Pt and Au with specific promoting materials is markedly enhanced by H2O, which involves ionic processes, and the rate-determining slow step is the reaction of HCOO(a) + OH → CO2 + H2O on Pt. The role of H2O is as a kind of molecular catalyst messenger molecule on the surface as expressed by reaction (ii) in eqn (1): CO + 1/2O2 + H2O → CO2 + H2O.

(2) Function of co-catalysts: the co-catalysts are indispensable for the oxidation of CO on Pt promoted by H2O, but Pt and promoting materials do not need to be in contact with one another. It was proved by the Pt supported on CNT and CNT-p, where the CNT has Ni–MgO at its terminal end but the CNT-p has no Ni–MgO. The Pt/CNT-p had no activity for the oxidation of CO enhanced by H2O, but the Pt/CNT was extremely active in the presence of H2O although the Pt and Ni–MgO are separated one from another on the CNT.

(3) A new concept for the selectivity of preferential oxidation of CO in H2: selectivity for the preferential oxidation of CO in H2 is given by n/(n + 1), where “n” is the number of CO2 molecules produced by the repeated contribution of one H2O molecule according to the equation CO + 1/2O2 + H2O → [Xi(a)] → CO2 + H2O, where H2O is produced by the oxidation of H2. The “n” is the turn-over number of one H2O molecule to nCO2 molecules before it leaves the catalyst.

Acknowledgements

We acknowledge the financial support from the MOST of China (2011CBA00508) and the National Natural Science Found for Creative Research Groups of China (51221892), and thank Mrs Qingcai Feng of the Research Center for Eco-Environmental Sciences of the Chinese Academy of Sciences for her help in accomplishing this review. K.-I. Tanaka acknowledges Mr Mitsushi Umino, the President of Astech Co., for his support of the research.

Notes and references

  1. N. Al-Sarraf, J. T. Stuckless and D. A. King, Nature, 1992, 360, 243 CrossRef CAS.
  2. G. Blyholder, J. Phys. Chem., 1964, 68, 2772 CrossRef CAS.
  3. M. Haruta, N. Yamada, T. Kobayashi and S. Iijima, J. Catal., 1989, 115, 301 CrossRef CAS.
  4. M. Date, M. Okumura, S. Tsubota and M. Haruta, Angew. Chem., Int. Ed., 2004, 43, 2129 CrossRef CAS PubMed.
  5. M. Haruta, Catal. Today, 1997, 36, 153 CrossRef CAS.
  6. M. Valden, X. Lai and D. W. Goodman, Science, 1998, 281, 1647 CrossRef CAS.
  7. M. S. Chen and D. W. Goodman, Science, 2004, 306, 252 CrossRef CAS PubMed.
  8. G. Vlaic, P. Fornasiero, S. Geremia, J. Kaspar and M. Grazianj, J. Catal., 1997, 168, 386 CrossRef CAS.
  9. Y. Zhang, S. Anderson and M. Muhammed, Appl. Catal., B, 1995, 6, 325 CrossRef CAS.
  10. M. A. Bollinger and M. A. Vannice, Appl. Catal., B, 1996, 8, 417 CAS.
  11. O. Korotkikh and R. Farrauto, Catal. Today, 2000, 62, 249 CrossRef CAS.
  12. K.-I. Tanaka, M. Shou, H. B. Zhang, Y. Z. Yuan, T. Hagiwara, A. Fukuoka, J. Nakamura and D. L. Lu, Catal. Lett., 2008, 126, 89 CrossRef CAS.
  13. K.-I. Tanaka, M. Shou, H. He and X. Y. Shi, Catal. Lett., 2006, 110, 185 CrossRef CAS.
  14. K.-I. Tanaka, M. Shou, H. He, X. Y. Shi and X. L. Zhang, J. Phys. Chem. C, 2009, 113, 12427 CAS.
  15. M. Shou, K.-I. Tanaka, K. Yoshioka, Y. Moro-oka and S. Nagano, Catal. Today, 2004, 90, 255 CrossRef CAS PubMed.
  16. M. Shou, H. Takekawa, D. Y. Ju, T. Hagiwara, D. L. Lu and K.-I. Tanaka, Catal. Lett., 2006, 108, 119 CrossRef CAS PubMed.
  17. A. Fukuoka and M. Ichikawa, Top. Catal., 2006, 40, 103 CrossRef CAS PubMed.
  18. G. Q. Yi, Z. N. Xu, G. C. Guo, K.-I. Tanaka and Y. Z. Yuan, Chem. Phys. Lett., 2009, 479, 128 CrossRef CAS PubMed.
  19. G. Q. Yi, H. W. Yang, B. D. Li, H. Q. Lin, K.-I. Tanaka and Y. Z. Yuan, Catal. Today, 2010, 157, 83 CrossRef CAS PubMed.
  20. S. J. Tauster, S. C. Fung and R. L. Garten, J. Am. Chem. Soc., 1978, 100, 170 CrossRef CAS.
  21. S. Surnev, J. Schoiswohl, G. Kresse, M. G. Ramsey and F. P. Netzer, Phys. Rev. Lett., 2002, 89, 246101 CrossRef CAS.
  22. C. Hardacre, R. M. Ormerod and R. M. Lambert, J. Phys. Chem., 1994, 98, 10901 CrossRef CAS.
  23. Y.-N. Sun, Z.-H. Qin, M. Lewandowski, E. Carrasco, M. Sterrer, S. Shaikhutdinov and H.-J. Freund, J. Catal., 2009, 266, 359 CrossRef CAS PubMed.
  24. A. Golchet and J. M. White, J. Catal., 1978, 53, 266 CrossRef CAS.
  25. H. Yoshida, S. Takeda, T. Uchiyama, H. Kohno and Y. Honnma, Nano Lett., 2008, 8, 2082 CrossRef CAS PubMed.
  26. K.-I. Tanaka, M. Shou and Y. Z. Yuan, J. Phys. Chem. C, 2010, 114, 16917 CAS.
  27. K.-I. Tanaka, H. He, M. Shou and X. Y. Shi, Catal. Today, 2011, 175, 467 CrossRef CAS PubMed.
  28. X. Y. Shi, K.-I. Tanaka, H. He, M. Shou, W. Xu and X. Zhang, Catal. Lett., 2008, 120, 210 CrossRef CAS.
  29. I. Langmuir, Trans. Faraday Soc., 1921, 17, 607 RSC.
  30. M. Show and K.-I. Tanaka, Catal. Lett., 2006, 111, 115 CrossRef PubMed.
  31. K. Tamaru, Dynamic Heterogeneous Catalysis, Acad. Press Inc., London, 1978 Search PubMed.
  32. R. B. Levy and M. Boudart, J. Catal., 1974, 32, 364 CrossRef.
  33. K.-I. Tanaka and A. Ozaki, J. Catal., 1976, 8, 1 CrossRef.
  34. R. P. Iczkowski and J. L. Magrave, J. Am. Chem. Soc., 1962, 83, 3547 CrossRef.
  35. K.-I. Tanaka, Bull. Chem. Soc. Jpn., 1973, 46, 3887 CrossRef CAS.
  36. S. T. Daniells, M. Makkee and J. A. Moulijn, Catal. Lett., 2005, 100, 39 CrossRef CAS.
  37. K.-I. Tanaka, Adv. Catal., 1985, 33, 99 CAS.
This journal is © The Royal Society of Chemistry 2015