Active gold species on cerium oxide nanoshapes for methanol steam reforming and the water gas shift reactions

Abstract

We report that a small amount (<1 at%) of gold on ceria single crystals prepared as nanorods (10 ± 2.8 by 50–200 nm) of {110} and {100} crystal surfaces shows excellent catalytic activity in both the steam reforming of methanol (SRM) and the water gas shift (WGS) reactions at low temperatures (<250 °C). The ceria nanorods bind and stabilize gold as atoms and clusters (<1 nm, TEM invisible). On the other hand, gold nanoparticles (∼3 nm) are found on the {100} surfaces of ceria nanocubes. Very low rates of SRM and WGS were measured on the Au–ceria {100} cubes, while the rates on Au–ceria {110} rods were at least an order of magnitude higher. However, the apparent activation energies did not depend on the shape of ceria. Strong bonded Au_n–O–Ce species are the active sites and these are present only in negligible concentrations on the {100} surfaces. Thus, both reactions are structure-insensitive on Au–ceria. SRM proceeds through the methyl formate route. The Au–ceria {110} catalyst shows both high SRM activity and high selectivity to CO₂ at temperatures below 250 °C.

---

Broader context

Novel catalysts for the efficient conversion of fuels to hydrogen for fuel cell applications are under development. Among them a prominent place is held by new catalysts that are both active and stable for the low-temperature water-gas shift reaction to upgrade hydrogen. Another need is highly active and selective catalysts for the reforming of methanol, the easiest alcohol to reform at low temperatures. Achieving either of these two goals would be a breakthrough in the efforts to generate energy cleanly and with high efficiency. Here we demonstrate the feasibility of using trace amounts of gold on cerium oxide as an active and selective catalyst for both the steam reforming of methanol (SRM) and the water-gas shift (WGS) reactions at temperatures below 300 °C. Furthermore, we explore mechanistic issues of the SRM reaction activity and selectivity on this very promising catalyst system for fuel cell applications.

1. Introduction

Hydrogen holds great promise as a future fuel, especially when it is produced at low cost from water splitting, once the relevant technologies are fully developed. In the interim, hydrogen can be derived from renewable biomass fuel conversion, or biomass plus fossil (or waste) fuel mixtures, in ways that minimize the carbon footprint of the process. In the latter approach, catalytic processes, such as the steam reforming of high-hydrogen carrier fuels, such as methanol, and the upgrade of hydrogen streams by the water-gas shift (WGS) reaction are integral parts of hydrogen production. The choice of catalysts determines the efficiency and stability of these processes.^1,2

Since the application of Au/Fe₂O₃ to water gas shift reaction was reported by Andreeva,³ Au catalysts have been widely investigated for this reaction. In a previous work in our group, we were first to identify Au/CeO₂ as a promising low-temperature shift catalyst⁴ and further found that atomically dispersed gold in ceria, Au_n–O–Ce, is responsible for the WGS activity.^5,6 On the other hand, few reports currently exist on the catalytic applications of gold supported on oxides for methanol reactions. Bowker et al.⁷ found that Au/TiO₂ photocatalyzes the reforming of methanol in an aqueous solution. Chang et al.⁸ reported that Au/TiO₂ catalysts show relatively high activity for partial oxidation of methanol, and correlated the activity with the particle size of metallic gold. At the same time, the mechanism behind the methanol reforming reaction is still under debate, since the methanol decomposition/water gas shift route is sometimes proposed, while the methyl formate route is supported by experimental evidence under different sets of conditions.^9,10

Because the SRM reaction may involve the WGS route in its mechanism, and our previous work has demonstrated that gold on nanoscale ceria is an excellent WGS catalyst,^4–6,11,12 the exploration of the SRM properties of the same material was a reasonable extension of our investigation.¹³ An additional objective was to investigate potential shape effects of the ceria nanocrystals on the Au–ceria catalyst activity.

2. Experimental

2.1 Sample preparation

Gold–ceria was synthesized in a two-step process as previously reported.¹² Different shapes of nanoscale ceria (nanorods and nanocubes) were prepared via a controlled hydrolysis method using Ce(NO₃)₃·6H₂O as the precursor and NaOH as the base, followed by a hydrothermal treatment. Gold then was introduced as calcined (400 °C, air) ceria shapes by deposition–precipitation (DP) at pH = 9 (base: (NH₄)₂CO₃). NaCN leaching was preformed with aqueous 2% NaCN/NaOH solution (pH = 12) at room temperature.^4,5

2.2 Physical and chemical characterization

Bulk composition analysis of the catalysts was conducted in an inductively coupled plasma optical emission spectrometer (ICP-OES, Leeman Labs Inc.). The BET surface area was determined by single-point N₂ adsorption/desorption cycles in a Micromeritics AutoChem II 2920. The as-calcined (400 °C, air) samples were pretreated in He at 300 °C for 0.5 h before tests.

Transmission electron microscopy (TEM) and high-resolution TEM (HRTEM) were performed on a JEOL 200cx and JEOL 2010 at 200 kV, respectively. The TEM samples were prepared by drying an ethanol suspension containing dispersed catalyst powders on carbon film-coated copper grids.

A Kratos AXIS Ultra Imaging X-ray Photoelectron Spectrometer (XPS) with a resolution of 0.1 eV was used to determine the atomic metal ratios of the surface region and the metal oxidation states of gold–ceria catalysts. Samples in powder form were pressed on a double-side adhesive copper tape for analysis. All measurements were carried out at room temperature without any sample pre-treatment. An Al Kα X-ray source was used in this work. All binding energies were calibrated with the C 1s peak at 285.0 eV.

2.3 Temperature programmed reactions

Temperature Programmed Surface Reaction (TPSR) tests in a CO + H₂O gas mixture were conducted with, 60 mg of as prepared samples, the gas mixture (1% CO + 3% H₂O) obtained by flowing 1% CO through a water bubbler with a flow rate of 70 mL min⁻¹.

In CH₃OH-Temperature Programmed Desorption (TPD) and (CH₃OH + H₂O)-TPSR, typically 60 mg samples were used. The catalysts were diluted with quartz powder at a weight ratio of 1/3. Samples were first pre-reduced in 20% H₂–He flowing at 70 mL min⁻¹. In CH₃OH-TPD, 16% CH₃OH was obtained by flowing pure helium through the methanol bubbler. Following completion of the adsorption, the samples were purged with pure helium to clean the residual gases before each test. In CH₃OH + H₂O-TPSR, the mixture of methanol and water was also generated by flowing pure helium through a liquid bubbler filled with methanol and water.

All temperature programmed reactions were carried out with the heating rate of 5 °C min⁻¹ from room temperature to the designed temperature. The outlet gas was analyzed online by a quadrupole residual gas analyzer (MKS model RS-1).

2.4 Activity tests

For methanol steam reforming, a mixture of methanol and water was injected into the flowing gas by a syringe pump and vaporized in the heated gas line before entering the reactor. The feed gas comprised a mixture of CH₃OH/H₂O/He with a molar ratio of 2/2.6/95.4 and a total gas flow rate of 70 mL min⁻¹. The feed and product gases were analyzed online by a residual gas analyzer (MKS model RS-1).

The WGS reaction tests were conducted at atmospheric pressure with a loading of ∼100 mg catalyst powders. The reaction gas mixture was 2% CO/10% H₂O/He and a total gas flow rate of 70 mL min⁻¹ was used. Measurements of reaction rates and activation energy were also carried out in product-free (2% CO/10% H₂O/He) gases at low conversions to avoid mass transfer effects. CO₂ and CO concentrations in the feed and product gas streams were analyzed by an HP-6890 gas chromatograph (GC) or an IR 703 non-dispersive infrared (NDIR).

3. Results and discussion

3.1 Catalyst characterization

The bulk metal compositions for both the parent and cyanide-leached catalysts were determined by ICP-AES. Table 1 shows that the attained Au concentration (0.9 at%) is very close to the designed value (1 at%) in each sample, which confirms the effective gold deposition on the ceria surface via the applied DP method. The BET surface areas of the Au–ceria samples were the same as those of the corresponding CeO₂ supports (see Table 1), showing that the gold-deposition had no effect on the textural properties of the oxide matrix.

Table 1 Physical properties and activity of ceria and Au–ceria

Sample	Au^a (at%)	S BET/m^2 gcat^−1	Au dispersion^c (%)	Rate^d/µmol gcat^−1 s^−1
				SRM	WGS
a Determined by ICP-AES. b Determined by ICP-AES after NaCN leaching at room temperature. c The dispersion of gold was estimated from the TEM data. 100% dispersion assumed for the leached samples (in brackets). d Numbers in brackets are the temperatures at which reaction rates were measured for the parent and leached samples. e Number in bracket is the specific surface area of CeO2 {110} surfaces, based on the nanorod structure.
Ceria (nanorods)	—	105 (53)^e	—	4.3 (400)	NA
Ceria (nanocubes)	—	27	—	1.2 (400)	NA
1% Au–ceria (Rod)	0.9 (0.5)^b	100	100 (100)	12.3 (250)	3.9 (175)
1% Au–ceria (Cube)	0.9 (0.03)^b	24	30 (100)	0.2 (250)	0.3 (175)

---

Ceria nanoparticles maintained their original crystal shapes after gold deposition and air calcination. The CeO₂ nanorods have a uniform width of 10.1 ± 2.8 nm with a length of 50–200 nm (Fig. 1a), while the CeO₂ nanocubes have a side-size of 29.5 ± 10.6 nm (Fig. 2a). Based on the corresponding HRTEM images in Fig. 1b and 2b, we have determined the growth direction for ceria nanorods as {110}, and the exposed crystal planes as: {110} + {100} on the rods and {100} on the cubes. Fig. 2a and b show small gold nanoparticles with an average size of ∼3 nm on the {100} surfaces of the ceria cubes. However, for the rod samples, no such gold nanocrystals were found. This indicates that the CeO₂ nanorod surfaces can disperse and stabilize gold atoms and sub-nm clusters (TEM invisible) which are not the case for the {100} surfaces of the nanocubes. By deduction, the {110} surfaces of the ceria nanorods are assumed to contain all the gold.

Fig. 1 TEM images of 1% Au–ceria (Rod): (a) fresh; (b) fresh, high-resolution; (c) after use in SRM (2% CH₃OH + 2.6% H₂O/He, 15 h on-stream, T = 300 °C); (d) after use in WGS (2% CO + 10% H₂O/He, 15 h on-stream; T = 350 °C).

Fig. 2 TEM images of 1% Au–ceria (Cube): (a) fresh; (b) fresh, high-resolution; (c) after use in SRM (2% CH₃OH + 2.6% H₂O/He, 15 h on-stream, T = 300 °C); (d) after use in WGS (2% CO + 10% H₂O/He, 15 h on-stream, T = 350 °C).

The morphologies of the used Au–ceria nanorods and nanocubes after the SRM and WGS reactions were also investigated by TEM. Fig. 2c and d show that the cube shape of ceria was maintained after reaction in CH₃OH/H₂O or CO/H₂O atmospheres. The sizes of both CeO₂ nanocubes and Au nanoparticles did not change. For the rod samples, it can be seen from Fig. 1c and d that the reaction conditions had an effect on the morphologies of the CeO₂ crystals. Some short nanorods, together with some polycrystals, were formed after either the SRM or the WGS reaction. However, gold particles were still absent as checked by TEM/HRTEM; gold atoms and clusters (<1 nm) were thus stable on the {110} surfaces even after 15 h on stream.

From the XPS (Axis Ultra, Al-Kα) data listed in Table 2, it can be seen that the surface Au concentrations (0.8–1.3 at%) are in good agreement with the bulk (0.9 at%), indicating that the gold species are present on the surface or the sub-surface layers of the ceria after the 400 °C calcination step. Specific elemental XPS spectra in Fig. 3a show a distinct binding-energy shift of the Au 4f peaks between the rod and the cube samples. After peak-deconvolution, we found that ionic gold (Au⁺ and Au³⁺) is the main species in the rod sample, while metallic gold (Au⁰) is dominant on the ceria nanocubes. We also investigated the used catalysts after 15 h on-stream in the SRM and WGS reactions. A mixture of mainly metallic gold and some ionic gold was found (Fig. 3b and c). Pre-reduction of the Au–ceria samples at 300 °C with 20% H₂/He was employed before the SRM tests, while the catalysts were used in the as calcined state in WGS. Detailed XPS peak analysis is shown in Table 2. Furthermore, according to the TEM/HRTEM results (Fig. 1 and 2), the zero valent gold was still well dispersed on the WGS- and SRM-used rod samples, since no gold particles bigger than 1 nm were identified.

Table 2 XPS Au 4f peak analyses for gold on ceria nanoshapes

Sample	Condition^a	XPS Au 4f peak^b
		Au^c (at%)	Au^0 (%)	Au^+ (%)	Au^3+ (%)
a Fresh, WGS, and SRM refer to sample as-prepared after air calcination at 400 °C, 15 hon-stream in WGS, and 15 h on-stream in SRM. b The binding energy was calibrated to the C 1s peak at 285.0 eV. c Determined from Au 4f and Ce 3d cores in XPS spectra.
1% Au–ceria (Rod)	Fresh	0.9	—	86	14
	WGS	0.8	87	10	3
	SRM	1.1	85	11	4
1% Au–ceria (Cube)	Fresh	1.1	77	17	6
	WGS	1.3	85	9	6
	SRM	1.3	86	10	4

---

Fig. 3 XPS spectra of Au_4f over Au–ceria: (a) fresh, (b) after use in SRM (2% CH₃OH–2.6% H₂O/He, 15 h on-stream, T = 300 °C); (c) after use in WGS (2% CO + 10% H₂O/He, 15 h on-stream, T = 350 °C).

3.2 Au–ceria interaction

In previous work, H₂-TPR was used to study the reducibility of surface oxygen on each Au–ceria catalyst.¹² Au–ceria nanorods showed a reduction peak around 94 °C, which is due to good interaction of the gold with the surface oxygen of ceria.^5,11 In the cube sample, however, no obvious reduction peak appeared up to 300 °C, indicating the lack of interaction between Au particles and CeO₂ in this sample.

The interaction of the ceria rods and Au–ceria (Rod) surfaces with methanol using CH₃OH-TPD was evaluated here, as shown in Fig. 4a and b, respectively. Hydrogen and carbon monoxide elute simultaneously on the ceria nanorods, apparently due to the decomposition of the methoxy group, which was found to form upon adsorption of methanol on ceria surfaces.¹⁴ On the other hand, weak hydrogen desorption took place from the {100} ceria cube faces.¹³ This suggests a different interaction of the {110} and {100} surfaces with methanol.

Fig. 4 CH₃OH-TPD over (a) ceria (Rod) and (b) 1% Au–ceria (Rod) (16% CH₃OH/He for TPD adsorption, flow rate: 70 mL min⁻¹, heating rate: 5 °C min⁻¹).

On the Au–ceria samples, no CO peak was observed; only CO₂ and H₂ were produced. If the intermediate was the methoxy species, same as on ceria, we should observe the simultaneous desorption of CO and H₂, instead of CO₂ and H₂. Thus, methanol decomposition on ceria and gold–ceria may follow different pathways.^15,16 Indeed, we found methyl formate intermediate on the Au–ceria, but not on ceria during CH₃OH-TPSR.¹³

3.3 Activity evaluation

In steady-state WGS light-off tests, the CO conversion over the Au–ceria rod sample approached the maximum value (∼90%) dictated by equilibrium at 250 °C, while the cube sample only showed <10% CO conversion at 350 °C.¹² The rates of CO₂ produced on these Au–ceria catalysts are shown in Fig. 5a and Table 1. The rate on the rods was 3.9 µmol g_cat⁻¹ s⁻¹ at 175 °C, while on the cubes only 0.3 µmol CO₂ g_cat⁻¹ s⁻¹ was produced at the same temperature. Similar rates were measured on the corresponding leached Au–ceria samples and the apparent activation energy was similar, 37–44 kJ mol⁻¹, which indicates that the residual gold in ceria species (0.5 at% Au for the rod and 0.03 at.% Au for the cube) are the active sites. We can then calculate turnover frequencies by properly scaling the measured reaction rates, assuming 100% gold dispersion in the leached samples. The TOFs thus calculated were similar for both samples, e.g. the TOF = ∼0.13 s ⁻¹ at 175 °C on Au–ceria nanorods or nanocubes. These results strongly suggest that the WGS reaction mechanism over the Au–ceria {110} and {100} surfaces is the same, i.e. the reaction is structure insensitive.

Fig. 5 Steady-state reaction rates for (a) WGS reaction and (b) SRM reaction measured in 2% CO–10% H₂O–He for WGS and 2% CH₃OH–2.6% H₂O–He for SRM.

The difference in the (per gram catalyst) rates is due to the very different amounts of active gold stabilized on the different ceria nanoshapes. Thus, the observed shape effect is “indirect”, and can be understood in terms of the active phase being the fully dispersed gold on ceria.

Let us now consider the activity of the Au–ceria catalyst for the SRM reaction. Reduction of the samples with 20% H₂/He at 300 °C preceded the activity tests because the oxidized (as calcined) samples showed lower activity. Steady-state data were collected after 1.5 h testing at each temperature. Both temperature ramping up and down were followed. Addition of gold was found to improve the activity of ceria dramatically. The 1% Au–ceria (Rod) catalyst shows full methanol conversion at ∼300 °C.¹³ Reaction rates were much lower on the 1% Au–ceria (Cube) than on the 1% Au–ceria (Rod), although the apparent activation energies were the same, ∼110 kJ mol⁻¹, as shown in Fig. 5b. Similar arguments to those made above for the structure insensitivity of the WGS reaction can be made also for the SRM reaction on Au–ceria, because the same turnover frequencies (e.g. TOF = ∼0.42 s⁻¹ at 250 °C) were calculated for both ceria nanoshapes.

3.3.1 Temperature Programmed Surface Reaction (TPSR). (CO + H₂O)-TPSR was carried out to investigate the WGS reaction onset and product evolution as the temperature was ramped up. Fig. 6a shows that CO₂ and H₂ are produced together as the CO/H₂O signals begin to drop. The increase of the water signal before it began to drop is due to desorption of the surface water.

Fig. 6 (a) CO + H₂O-TPSR over 1% Au–ceria (Rod) (1% CO + 3% H₂O/He, heating rate = 5 °C min⁻¹); (b) CH₃OH + H₂O-TPSR over 1% Au–ceria (Rod); (c) CH₃OH + H₂O-TPSR over 1% Au–ceria (Cube) (gas composition: 16% CH₃OH + 3% H₂O/He, heating rate = 5 °C min⁻¹).

TPSR was carried out also with a mixture of methanol and water over the Au–ceria samples. As shown in Fig. 6b, during the temperature ramp over the 1% Au–ceria (Rod), hydrogen and carbon dioxide were formed beginning at ∼175 °C (temp. I). This agrees with our activity test results that showed that hydrogen and carbon dioxide were produced in the low-temperature region.¹³ Carbon monoxide was observed starting at ∼275 °C (temp. II, Fig. 6b). In the temperature range from 275 °C to 335 °C (temp. III), the formation of CO took place. There are two possible sources of carbon monoxide: one is the decomposition of methoxy adsorbed on ceria and the other is the reverse water gas shift reaction. Either possibility can account for the reduced CO₂ selectivity above 275 °C. In the temperature range between 335 °C and 395 °C (temp. IV), CO and water decrease sharply but CO₂ and H₂ increase. Since methanol conversion is complete at 325 °C, this manifests a sharp increase in the WGS reaction. Evidently, the presence of methanol inhibits the water gas shift reaction. At even higher temperatures (>395 °C, Fig. 6b) the reverse water gas shift reaction dominates.

While in CH₃OH-TPSR methyl formate species were found,¹³ when H₂O was used along with methanol as shown in Fig. 6b, no methyl formate was identified in the products. There was also no methane present. Trace amounts of formic acid were found, however. This suggests that methyl formate hydrolysis took place:^9,13

HCOOCH₃ + H₂O → HCOOH + CH₃OH

The formic acid is then decomposed to produce CO₂ and H₂.

As shown in Fig. 6c, similar product distribution was observed on the 1% Au–ceria (Cube) except that the corresponding elution temperatures shifted to higher values.

From the similarities of Fig. 6b and c, and the rate measurements of Fig. 5, we can argue that the SRM reaction on Au–ceria is structure-insensitive. The rates on the Au–ceria (Cube) are much lower due to the lack of active sites (Au_n–O–Ce) but the few sites that remain after leaching (0.03 at% Au on the {100} surfaces) have the same intrinsic activity as the gold clusters on the {110} surfaces of the ceria nanorods.

3.3.2 Mechanistic considerations. To further investigate the mechanism of methanol steam reforming over Au–ceria, carbon monoxide was added to the feed mixture of methanol and water. 5% CO was introduced into the reaction gas mixture (2.0% MeOH, 2.6% water, balanced with pure helium) at 225 °C under steady-state operation.¹³ The conversion of methanol remained the same, supporting the TPSR results of Fig. 6b and c that the water-gas shift reaction is not involved in the SRM reaction pathway on gold catalysts.

Methanol is known to adsorb weakly and molecularly on a clean gold surface^15,17 However, on oxygen-activated gold (111) surfaces, reconstruction and formation of small gold islands take place and on these, stronger adsorption of methanol has been reported.^15,17,18 In the present work, ceria can serve as the source of active oxygen to the small gold clusters stabilized on its surface. Based on previous studies on gold single crystals,^15,19 formaldehyde is stable on gold surfaces and can easily continue to react with methoxy or other sources of oxygen to form methyl formate groups. From CH₃OH-TPSR tests,¹³ we have found that methyl formate and hydrogen appear together, and H₂ and CO₂ are the dominant products.^13,20

On the basis of the above, we suggest a cooperative mechanism for Au–ceria in steam reforming of methanol: reduced ceria adsorbs methanol as methoxy which combines with the formaldehyde adsorbed on gold.¹³ The methyl formate pathway is then followed, as depicted in the graphical abstract. At the same time, gold facilitates the reduction of ceria, which in turn can adsorb water, as has been reported for the Rh/ceria(111) system.²¹ Adsorbed water helps to hydrolyze methyl formate to formic acid, which is then decomposed to CO₂ and H₂. In summary, the following steps are proposed for steam reforming of methanol over Au–ceria:

2CH₃OH → HCOOCH₃ + 2H₂

HCOOCH₃ + H₂O → HCOOH + CH₃OH

HCOOH → CO₂ + H₂

The other major finding of this work is that different surfaces of ceria bind and stabilize gold clusters and atoms differently. Thus, very different amounts of atomically dispersed gold are present on the {110} and {100} surfaces of the ceria nanoshapes. This in turn controls the reaction rates for both the SRM and WGS reactions. It is documented that ceria planes have different bonding characteristics, namely the very stable and neutral {111} surface, the less stable, slightly puckered {110} surface, and the higher energy {100} surface.²² According to the model proposed by Wang and Feng,²³ the surface density of atoms in the corresponding planes follows {111} > {100} > {110}. The energy of formation of oxygen vacancies follows the same order on these surfaces. Thus, the {110} surfaces have a higher number density of oxygen vacancies. For CeO₂ {100} polar plane, which is composed of positively charged Ce layers and negatively charged O layers, this structure is not stable upon heating and will relax to a low-energy surface.²⁴

3.4 Stability tests

In cyclic heating to 225 °C/cooling to RT in the WGS and SRM reaction mixtures over the Au–ceria (Rod) lasting a total of 12 h, no catalyst deactivation was found. We further examined the stability of Au–ceria (Rod) in shutdown/restart operation. As shown in Fig. 7, after 2 h on-stream at 225 °C under SRM conditions, shutdown to RT for 2 h was used twice, followed by reheating to 225 °C. The catalyst did not suffer any activity loss due to these treatments, Fig. 7. After 12 h on-stream, the Au–ceria (Rod) catalyst was oxidized using temperature-programmed oxidation to check for any carbon deposition. No evidence of carbon deposition was found. Hence these materials show promise for practical applications, e.g. in fuel cell systems requiring frequent shutdowns during typical operation.

Fig. 7 Stability of 1% Au–ceria (Rod) in cyclic start up/shutdown SRM tests (gas composition: 2% CH₃OH/2.6% H₂O/bal. He, GHSV: 42 000 h⁻¹).

4. Conclusions

We have found that a small amount (<1 at%) of fully dispersed gold on ceria nanorods exhibits good catalytic activity and selectivity for the low-temperature (<250 °C) methanol steam reforming reaction. The same material is also an excellent WGS catalyst. Fully dispersed gold (<1 nm, TEM invisible) on the {110} surfaces of ceria nanorods catalyzes the SRM reaction. Gold nanoparticles (∼3 nm) on the {100} surface of ceria nanocubes are inactive for both the SRM and WGS reactions. Interestingly, the WGS pathway is not involved in the SRM reaction on Au–ceria; methanol dehydrogenation, methyl formate hydrolysis and formic acid decomposition are the steps producing CO₂ and H₂. While both reactions are catalyzed by Au–ceria at the same low temperatures, we have found that methanol is the preferred adsorbate over carbon monoxide on gold.

Overall, the Au–ceria catalyst shows good activity and stability for both the WGS and SRM reactions. Because of its high CO₂ selectivity in SRM below 250 °C, this catalyst merits further evaluation for practical applications to future fuel processing and fuel cell systems.

Acknowledgements

The financial support of this work by the DOE/BES-Hydrogen Fuel Initiative program (#DE-FG02-05ER15730) is gratefully acknowledged. We thank the Air Force Office of Scientific Research (R. S.) for partial support of this work. We also thank Micromeritics for generously supplying the Autochem II 2920 as part of its Instrument Grant Program.

References

1. D. R. Palo, R. A. Dagle and J. D. Holladay, Chem. Rev., 2007, 107, 3992.
2. R. M. Navarro, M. A. Peña and J. L. G. Fierro, Chem. Rev., 2007, 107, 3952.
3. D. Andreeva, V. Idakiev, T. Tabakova and A. Andreev, J. Catal., 1996, 158, 354.
4. Q. Fu, A. Weber and M. Flytzani-Stephanopoulos, Catal. Lett., 2001, 77, 87.
5. Q. Fu, H. Saltsburg and M. Flytzani-Stephanopoulos, Science, 2003, 301, 935.
6. W. Deng, A. I. Frenkel, R. Si and M. Flytzani-Stephanopoulos, J. Phys. Chem. C, 2008, 112, 12834.
7. M. Bowker, L. Millard, J. Greaves, D. James and J. Soares, Gold Bull., 2004, 37, 170.
8. F. W. Chang, H. Y. Yu, L. S. Roselin and H. C. Yang, Appl. Catal., A, 2005, 290, 138.
9. C. J. Jiang, D. L. Trimm, M. S. Wainwright and N. W. Cant, Appl. Catal., A, 1993, 97, 145.
10. B. A. Peppley, J. C. Amphlett, L. M. Kearns and R. F. Mann, Appl. Catal., A, 1999, 179, 31.
11. Q. Fu, W. Deng, H. Saltsburg and M. Flytzani-Stephanopoulos, Appl. Catal., B, 2005, 56, 57.
12. R. Si and M. Flytzani-Stephanopoulos, Angew. Chem., Int. Ed., 2008, 47, 2884.
13. N. Yi, R. Si, H. Saltsburg and M. Flytzani-Stephanopoulos, Appl. Catal., B, 2010, 95, 87.
14. R. M. Ferrizz, G. S. Wong, T. Egami and J. M. Vohs, Langmuir, 2001, 17, 2464.
15. D. A. Outka and R. J. Madix, J. Am. Chem. Soc., 1987, 109, 1708.
16. A. Gazsi, T. Bánsági and F. Solymosi, Catal. Lett., 2009, 131, 33.
17. B. J. Xu, X. Y. Liu, J. Haubrich, R. J. Madix and C. M. Friend, Angew. Chem., Int. Ed., 2009, 48, 4206.
18. J. L. Gong, D. W. Flaherty, R. A. Ojifinni, J. M. White and C. B. Mullins, J. Phys. Chem. C, 2008, 112, 5501.
19. D. A. Outka and R. J. Madix, Surf. Sci., 1987, 179, 361.
20. N. Yi, PhD Dissertation, Tufts University, Medford, MA, USA, in progress.
21. L. Kundakovic, D. R. Mullins and S. H. Overbury, Surf. Sci., 2000, 457, 51.
22. M. Baudin, M. Wójcik and K. Hermansson, Surf. Sci., 2000, 468, 51.
23. Z. L. Wang and X. D. Feng, J. Phys. Chem. B, 2003, 107, 13563.
24. M. Baudin, M. Wójcik and K. Hermansson, Surf. Sci., 2000, 468, 51.

---

Footnote

† This article was submitted as part of a Themed Issue on fuels of the future. Other papers on this topic can be found in issue 3 of vol. 3 (2010). This issue can be found from the Energy & Environmental Science, homepage http://www.rsc.org/publishing/journals/ee/.

---