Kinetic and spectroscopic characterization of cluster-derived supported Pt-Au catalysts

Journal of Catalysis 212, 125–135 (2002)

doi:10.1006/jcat.2002.3770

Kinetic and Spectroscopic Characterization of Cluster-Derived

Supported Pt–Au Catalysts

∗

∗^,1

Corina Mihut, Claude Descorme,† Daniel Duprez,† and Michael D. Amiridis

∗

Department of Chemical Engineering, College of Engineering and Information Technology, Swearingen Engineering

Center, University of South Carolina, Columbia, South Carolina 29208; and †Laboratoire de Catalyse en Chimie

Organique, UMR 6503 CNRS, Universite´ de Poitiers, 40 avenue du Recteur Pineau, 86022 Poitiers, France

Received February 5, 2002; revised August 5, 2002; accepted August 6, 2002

in stabilizing Au in a small cluster form when these con-

ventional preparation methods are used (2). Such segre-

gation problems can be resolved, at least in principle, by

the use of preformed bimetallic cluster precursors, fre-

quently stabilized by organic ligands. Carbonyl-ligated clus-

ters are preferred precursors for heterogeneous catalysts

because of the advantage of the easy removal of the car-

bonyl groups by thermal treatment (3, 4). However, there

are no known bimetallic Pt–Au clusters exclusively ligated

by carbonyl groups. In contrast, there are a few examples

of phosphine stabilized bimetallic Au-containing clusters

(i.e., [Pt(Au(PPh₃)₈](NO₃)₂, [Pt(PPh₃)(AuPPh₃)₆](NO₃)₂,

and [Pt(AuPPh₃)₂(PPh₃)₂](NO₃)₂) that have been immobi-

lized and thermally activated on solid supports (5, 6). These

clusters have high Au content, and the presence of phos-

phorous in the ligands represents a potential poison for

several reactions. Recently, Chandler et al. (7–10) reported

the preparation of a supported Pt–Au catalyst from a clus-

ter precursor that exclusively contains acetylide ligands and

they concluded that true bimetallic Pt–Au particles can be

formed via this synthetic route. These ligands are easily re-

movable at high temperatures.

Silica-supported bimetallic Pt–Au catalysts prepared via differ-

ent synthetic routes have been investigated in terms of their struc-

tural properties, adsorption of CO, and catalytic activity for the

selective catalytic reduction of NO by propylene, the oxidation of

propylene in the absence of NO, and the ¹⁶O/¹⁸O homoexchange re-

action. Catalysts prepared by incipient wetness impregnation from

individual Pt and Au precursors exhibited characteristics very sim-

ilar to those of monometallic Pt catalysts, indicating that in these

cases the presence of Au did not affect the catalytic performance of

Pt in any signiﬁcant way. This behavior is consistent with a model

in which the two metals remain segregated due to their miscibility

gap, and only Pt participates in the adsorption of CO and the re-

actions under consideration. In contrast, catalysts prepared from

t

≡

a Pt₂Au₄(C C Bu)₈organo–bimetallic cluster precursor exhibited

different behavior both in terms of CO adsorption and their cata-

lytic activity for the three reactions examined. The combination

of the kinetic, spectroscopic, and structural characterization data

suggests that in this case Pt and Au remain intimately mixed in the

form of Pt–Au bimetallic particles and that the presence of Au in

these particles modiﬁes the behavior of Pt.

ꢀc 2002 Elsevier Science (USA)

Key Words: platinum; gold; synthesis; organometallic clusters;

CO adsorption; NO_xreduction; propylene oxidation; homoex-

change reaction.

In this paper, we followed the same synthetic approach

and we compared the cluster-derived catalysts to catalysts

having the same nominal composition, which were pre-

pared by incipient wetness impregnation from hexachloro-

platinic and tetrachloroauric acids. The structure of these

catalysts was characterized by transmission electron mi-

croscopy (TEM), hydrogen chemisorption, and Fourier

transform infrared spectroscopy of adsorbed ¹²CO/¹³CO

mixtures. The chemical activity was probed by kinetic stud-

ies of the selective catalytic reduction of NO by propylene,

the oxidation of propylene in the absence of NO, and the

¹⁶O/¹⁸O homoexchange reaction.

1. INTRODUCTION

Supported bimetallic Pt–Au catalysts are primar-

ily synthesized by impregnation, co-impregnation, and

deposition-precipitation methods from individual Pt and

Au salts. Phase separation of the two metals and inho-

mogeneous composition of the resulting bimetallic parti-

cles are common problems with this type of catalyst due

to the large miscibility gap in the bulk phase between Pt

and Au (i.e., between 18 and 97 wt% Pt) (1). Further-

more, the low melting point of Au and, therefore, the in-

creased mobility of the Au atoms at high temperatures re-

sults in broad Au particle size distributions and difﬁculties

2. EXPERIMENTAL METHODS

2.1. Catalyst Preparation

¹To whom correspondence should be addressed. Fax: (803) 777-8265.

E-mail: amiridis@engr.sc.edu.

Monometallic Pt/SiO₂and Au/SiO₂catalysts with Pt

loadings of 1 or 0.15 wt% and Au loadings of 2 or 0.3% were

125

0021-9517/02 $35.00

c

ꢀ 2002 Elsevier Science (USA)

126

MIHUT ET AL.

prepared by incipient wetness impregnation of the SiO₂gen at different pressures was recorded as a function of

support (Davison Syloid 74; calcined overnight at 500^◦C the equilibrium pressure up to monolayer coverage. The

prior to impregnation; BET surface area 300 m²/g) with monolayer coverage was determined based on the “extrap-

aqueous solutions of hexachloroplatinic (H₂PtCl₆· xH₂O, olation to zero pressure” method, which accounts for and

Aldrich) and tetrachloroauric acids (HAuCl₄· 3H₂O, subtracts the amount of physisorbed hydrogen. Measure-

Aldrich). Additionally, bimetallic catalysts containing ments performed with the 2% Au sample and with the bare

1 wt% Pt–2 wt% Au or 0.15 wt% Pt–0.3 wt% Au were silica support indicated no H₂uptake in these cases. A sto-

prepared by co-impregnation (i.e., in this case both precur- ichiometry of one adsorbed hydrogen atom per Pt surface

sors were added in the same impregnation solution) using atom was assumed in the calculations.

the same support material and precursors. Following im-

pregnation, these catalysts were dried in vacuum for 2 h at

2.4. Transmission Fourier-Transform Infrared

Spectroscopy (FTIR) of Adsorbed CO

120^◦C and calcined in air for 5 h at 500^◦C.

Cluster-derived catalysts were prepared via a wet-

Transmission FTIR spectra were collected in the single-

beam mode with a resolution of 2 cm⁻¹using a Nicolet

740 FTIR spectrometer equipped with an MCT-B detector.

A 10-cm-long stainless steel IR cell, with NaCl windows

cooled by ﬂowing water, was used to collect in situ spectra.

A heating element wrapped around the cell allowed spectra

collection at elevated temperatures. The cell temperature

was monitored by a thermocouple placed in close proxi-

mity to the catalyst sample. Reference spectra of the clean

surfaces in He, 5% H₂/He or 5% O₂/He mixtures were

collected at different temperatures as needed. Difference

spectra between the samples and the corresponding refer-

ence are shown in this paper. Catalyst samples were self-

supported wafers with a diameter of 12 mm and a thickness

of approximately 20 mg/cm².

Two sets of experiments, each with different activation

protocols, were conducted with each catalyst sample. In

the ﬁrst set (i.e., “oxidized” samples) the catalysts were

exposed for 2 h in a ﬂowing 5% O₂/He mixture at 300^◦C

(cluster-derived samples) or 400^◦C (all other samples). In

the second set (i.e., “reduced” samples) the catalysts were

exposed for 2 h in a ﬂowing 5% H₂/He mixture at 300^◦C.

Following these pretreatments the samples were cooled in

either O₂/He or H₂/He mixtures to 100^◦C, ﬂushed with

He, and cooled in He to room temperature. At room tem-

perature a ﬂowing CO/He mixture (CO concentrations of

3000 ppm and 3%) was introduced to the cell and spec-

tra were collected at different time intervals until steady

state was reached (usually after 20 min). Helium was then

purged through the cell and spectra were collected at dif-

ferent time intervals and elevated temperatures until the

entire amount of adsorbed CO was removed from the

surface.

impregnation technique by adsorption of the organometal-

t

≡

lic Pt₂Au₄(C C Bu)₈cluster precursor (11) from hexane

solution onto the SiO₂support (Aldrich Davisil SiO₂; cal-

cined overnight at 500^◦C prior to impregnation; BET sur-

face area 360 m²/g) following the same procedure outlined

in previous literature reports (7). The concentration of the

cluster solution was adjusted so that the ﬁnal metal loadings

of the bimetallic catalysts obtained were approximately the

same as those of the catalysts obtained by co-impregnation.

Following impregnation, these catalysts were dried in vac-

uum for 2 h at 60^◦C, oxidized at 300^◦C (temperature ramp

10^◦C/min) in a ﬂowing 5% O₂/He mixture for 2 h, and ﬁ-

nally reduced at 200^◦C (same temperature ramp) for 1 h in

a 5% H₂/He mixture. The different activation protocol was

chosen in this case to assure mild removal of the acetylide

ligands. The metal loadings of all catalysts used in this study

were conﬁrmed with elemental analysis via ICP (Galbraith

Laboratories).

2.2. Transmission Electron Microscopy Measurements

In preparation for TEM imaging approximately 15 mg

of each sample were ﬁnely crushed and dispersed in 3 ml

deionized water by ultrasoniﬁcation for 30 min. A drop

of this ﬁne dispersion was deposited on a Cu grid and

then dried at 40^◦C for 5–10 min until the excess water

was removed. Samples were then examined using a Hitachi

200-kV microscope. Digital images were recorded at mag-

niﬁcations of approximately 200,000. Particle size distri-

butions were obtained by measuring 200 particles for

each sample analyzed, using at least three different micro-

graphs.

2.3. Hydrogen Chemisorption Studies

Experiments were also conducted at room temperature

Pt dispersions were obtained via hydrogen chemisorp- with mixtures of ¹²CO and ¹³CO at a total ﬂow rate of

tion at room temperature (Autosorb AS-1 gas sorption sys- 5 ml/min and at different ratios (35% ¹²CO–65% ¹³CO,

tem). Prior to these measurements, the samples were re- 50% ¹²CO–50% ¹³CO, 65% ¹²CO–35% ¹³CO). The CO in

duced in situ in ﬂowing H₂at a temperature of 300^◦C for this case was diluted in 95 ml/min of He and was ﬂowed

3 h, followed by evacuation at the same temperature for over the catalyst pellet for 30 min to allow equilibration

2 h. The samples were then cooled to 40^◦C and exposed to of the surface. The gas-phase spectra of the ¹²CO–¹³CO

H₂at different pressures. Preselected amounts of H₂were mixtures were subtracted from the spectra shown in this

added to the sample and the volumetric uptake of hydro- paper.

CHARACTERIZATION OF Pt–Au CATALYSTS

127

2.5. Activity Measurements

ture was held at 200 cm³/min (1 atm, 25^◦C) with a corre-

sponding gas hourly space velocity (GHSV) of 120,000 cm³/

gm/hr. Measurements with samples containing 0.15 wt%

Pt and/or 0.3 wt% Au used 300 mg of catalyst and a total

ﬂow rate of 100 cm³/min, so that the overall space velocity

was decreased proportionally to the Pt loading and simi-

lar overall NO conversions were obtained with both sets of

samples. The percentage NO reduction was calculated from

the amounts of N₂and N₂O formed, while the selectivity to-

ward N₂and N₂O was deﬁned as the amount of N₂formed

divided by the total N₂and N₂O formed. The only reac-

tion products observed from the oxidation of C₃H₆were

CO₂and H₂O, and the carbon mass balance was closed

within 5%.

2.5.1. ¹⁶O/¹⁸O homoexchange reaction. Activity mea-

surements for the ¹⁶O/¹⁸O homoexchange reaction were

carried out on 20-mg samples in a batch recirculation reac-

tor described in detail elsewhere (12). Prior to reaction, the

samples were oxidized at 400^◦C (10^◦C/min) in pure ¹⁶O₂

for 15 min, outgassed under vacuum at 400^◦C for 15 min,

reduced at the same temperature for 15 min in pure H₂, and

ﬁnally outgassed again at 400^◦C under vacuum for 30 min.

Following this pretreatment the samples were cooled to the

reaction temperature and an equimolar mixture of ¹⁸O₂and

¹⁶O₂(P_total= 50 mbar) was introduced to the reactor. The

progress of the reaction was followed by mass spectrome-

try. Masses 36 (¹⁸O₂), 34 (¹⁸O¹⁶O), 32 (¹⁶O₂), 28 (N₂), and

18 (H₂O) were monitored continuously. (Masses 28 and

18 were monitored to assure that no leak occured into the

reactor.)

3. RESULTS

3.1. TEM Measurements and H₂Chemisorption

The homoexchange reaction rate (V_q), expressed in oxy-

gen atoms exchanged per hour per surface Pt site was cal-

culated as follows:

Number-average metal particle sizes, as well as the results

of the hydrogen uptake measurements for different fresh

and “used” (i.e., exposed to HC-SCR conditions) samples

are summarized in Table 1. Furthermore, metal particle size

distributions for the 1Pt and the 1Pt–2Au samples prepared

by impregnation, and the cluster-derived 0.5Pt₂Au₄sam-

ple are shown in Fig. 1. The results indicate similar metal

particle size averages for the monometallic 1Pt and the co-

impregnated 1Pt–2Au samples. In contrast, a smaller av-

erage metal particle size (3.4 versus 4.6 nm) and a more

narrow distribution are observed with the cluster-derived

0.5Pt₂Au₄sample. Evidence that mild sintering is taking

place as a result of exposure to HC-SCR reaction condi-

tions becomes apparent when the average metal particle

sizes of the used samples are compared to the correspond-

ing averages for the fresh samples (Table 1). The observed

changes, however, are similar for all three samples and indi-

cate that the cluster-derived 0.5Pt₂Au₄catalyst maintains a

smaller average metal particle size than the other two sam-

ples during exposure to temperatures up to 500^◦C in the

presence of O₂(i.e., HC-SCR conditions). Finally, Pt disper-

sions estimated based on hydrogen uptake measurements

(Table 1) are consistent with the TEM measurements, in-

dicating a higher uptake for the cluster-derived 0.5Pt₂Au₄

sample and mild sintering for the used samples.

ꢀ

ꢁ ꢀ

ꢁ

d P^o

2 · MW_Pt

V_rV_c

+

¹⁸O¹⁶

O

V_q=

·

[1]

m · x(wt) · D · R

T_rT_c

dt

where MW_Ptis the atomic weight of Pt (g/mol), m is the sam-

ple mass (g), x(wt) is the Pt loading of the sample, D is the Pt

dispersion, R is the ideal gas constant (8.314 J K⁻¹mol⁻¹),

V_rand V_care the reactor and loop volumes (cm³), T_r

and T_care the reactor and loop temperatures (K), and

d P₁^o₈

16

is the initial rate of formation of ¹⁸O¹⁶O.

O

dt

2.5.2. Selective catalytic reduction of NO by C₃H₆.

Steady-state activity measurements of the selective cata-

lytic reduction of NO by C₃H₆(HC-SCR) were conducted

in a quartz single-pass ﬁxed-bed ﬂow reactor. Prior to the

activity measurements all catalyst samples were oxidized

in situ at 500^◦C in a 5% O₂/He mixture for 2 h. Certiﬁed

mixtures of 1.00% NO in He, 1.03% C₃H₆in He, 9.80%

O₂in He, and a 99.999% He carrier gas, provided and cer-

tiﬁed by National Specialty Gases, were used to prepare

the reacting gas mixture. Typical gas concentrations used

were 1000 ppm_vNO, 1000 ppm_vC₃H₆, 1% O₂, and bal-

ance helium. Inlet and outlet gas streams were analyzed

by the use of a gas chromatograph (Hewlett-Packard 5890)

equipped with a thermal conductivity detector and three

chromatographic columns: Molecular Sieve 5A (for the

analysis of N₂), Porapak Q (for the analysis of N₂O and

CO₂), and Graphpac GC (for the analysis of C₃H₆) con-

nected in parallel. The reaction temperature was varied be-

tween 200 and 500^◦C and was monitored by a thermocouple

placed in the catalyst bed.

TABLE 1

Properties of Different Catalysts Used in this Study

Average metal

Hydrogen uptake

Loading (%)

particle size (nm)

(moles H₂/mole Pt)

Catalyst

Pt

Au

Fresh

Used

Fresh

Used

1Pt

1Pt–2Au

0.5Pt₂Au₄

1.04

0.79

0.77

—

1.81

1.54

4.7

4.6

3.4

5.3

5.6

4.0

0.158

0.153

0.223

0.108

0.101

0.2

Measurements with samples containing 1% Pt and/or

2% Au used 100 mg of the catalyst in the form of 60/80 mesh

particles. The total volumetric ﬂow rate of the reactant mix-

128

MIHUT ET AL.

45

40

35

30

25

20

15

10

5

ing in the high-frequency region can be observed in the

CO–Au peak of the same spectrum. These features are

more pronounced than the corresponding asymmetric tail-

ing observed in the spectra of the monometallic Pt/SiO₂

and the co-impregnated Pt–Au/SiO₂samples (Spectra 2c

and 2b) and suggest that the cluster-derived catalyst con-

tains a broader distribution of CO adsorption sites of differ-

ent adsorption strengths. Indeed, attempts to deconvolute

these peaks using different software packages, yielded re-

sults indicating that a number of smaller intensity peaks is

needed in order to ﬁt the asymmetric tailing shown in Fig. 2,

spectrum a.

Purging the CO-saturated 0.5Pt₂Au₄sample with an in-

ert atmosphere results in the complete removal of the Au-

bound CO (Fig. 3) after approximately 40 min. At the same

time, the peak at 2064 cm⁻¹assigned to Pt-bound CO de-

creases slightly in intensity at room temperature and is com-

pletely eliminated at temperatures around 150^◦C. These re-

sults are in agreement with results from previous studies

on supported Au and Pt–Au catalysts, where outgassing at

room temperature completely depletes the band associated

with CO bound to Au (15–17).

0

1

2

3

4

5

6

7

8

9

10

(nm)

Spectra of adsorbed CO on a “used” catalyst follow-

ing a reduction treatment at 300^◦C are shown in Fig. 4.

Even after exposure to reaction conditions the two distinc-

tive IR features (namely, the red shift in the CO–Pt peak

and the appearance of the CO–Au peak) are still present.

This result is consistent with the kinetic results, indicat-

ing no catalyst deactivation or change in the selectivity to-

ward N₂even after prolonged exposure of the 0.5Pt₂Au₄

catalyst to the reaction conditions. Similar to the fresh

FIG. 1. TEM histograms of the fresh co-impregnated and cluster-

derived catalysts ( , 0.5Pt₂Au₄fresh sample; ꢀ, 1Pt fresh sample; ꢁ,

1Pt–2Au fresh sample).

3.2. FTIR Spectroscopy of Adsorbed CO

FTIR studies of adsorbed CO were performed on both

reduced and oxidized samples, in an attempt to understand

the effect of the Au presence on the structure and prop-

erties of Pt. Spectra collected at room temperature on re-

duced catalyst samples following exposure to 3.3 kPa of

CO are shown in Fig. 2. A strong peak at 2070 cm⁻¹(Fig. 2,

spectrum b) is observed in the spectrum of the monometal-

lic 1Pt sample, in agreement with literature reports (13).

This peak was assigned to CO linearly adsorbed on fully

reduced Pt sites. The addition of Au via co-impregnation,

as in the 1Pt–2Au sample, did not affect the CO adsorption

(Fig. 2, spectrum c), consistent with previous literature re-

ports (7, 13, 14). Furthermore, no peak assignable to CO

adsorbed on Au was observed for the 2Au sample (Fig. 2,

spectrum d).

In contrast, some signiﬁcant differences can be noticed

in the spectra of the 0.5Pt₂Au₄catalyst (Fig. 2, spectrum a)

in agreement with similar results reported by Chandler

et al. (7) for the adsorption of CO on cluster-derived Pt–Au

catalysts. In particular, a red shift from 2070 to 2064 cm⁻¹

is observed in the position of the peak assigned to CO lin-

early bound to Pt, and a new peak is observed at 2117 cm⁻¹

assigned to CO bound to ﬁnely dispersed Au. Furthermore,

a weak shoulder and substantial asymmetric tailing in the

low-frequency region can be observed in the CO–Pt peak

FIG. 2. IR spectra of adsorbed CO at room temperature (30^◦C) on

reduced SiO₂-supported Pt and Pt–Au catalysts. Spectra were collected

immediately following the removal of gas-phase CO: (a) 0.5Pt₂Au₄,

of Fig. 2, spectrum a. Similarly, substantial asymmetric tail- (b) 1Pt, (c) 1Pt–2Au, (d) 2Au.

CHARACTERIZATION OF Pt–Au CATALYSTS

129

FIG. 5. IR spectra of adsorbed CO at RT (30^◦C) over SiO₂-supported

oxidized Pt and Pt–Au catalysts. Spectra were collected immediately fol-

lowing removal of the gas-phase CO: (a) 0.5Pt₂Au₄, (b) 1Pt, (c) 1Pt–2Au.

FIG. 3. IR spectra of adsorbed CO on reduced 0.5Pt₂Au₄catalyst.

Spectra were collected following exposure to 0.3 kPa CO for 1 h at room

temperature (RT) and He purge: (a) immediately following CO gas-phase

removal at RT, (b) for 5 min at RT, (c) for 70 min at RT, (d) for 230 min at

RT, (e) for 10 min at 50^◦C, (f) for 10 min at 75^◦C, (g) for 10 min at 100^◦C,

and (h) for 10 min at 150^◦C.

Similarresultstotheonesobtainedwiththereducedsam-

ples were also observed with samples pretreated in oxygen.

In this case, however, the CO stretching frequencies occur

consistently at higher wave numbers. This could be due to

either a compression of the CO layer caused by co-adsorbed

oxygen, which predictably leads to higher frequencies than

the ones obtained on reduced surfaces, or to an electronic

effect associated with a weakening of the surface CO bond

under these conditions (18). Spectra of the oxidized sam-

ples after exposure to CO at room temperature are shown

in Fig. 5. Once again, a red shift of the Pt–CO peak and a

new peak at 2123 cm⁻¹characteristic of Au-bound CO are

evident in the spectrum of the 0.5Pt₂Au₄sample (Fig. 5,

spectrum a). Furthermore, a new peak is identiﬁed in

Fig. 5 (spectrum a) at 1844 cm⁻¹and can be assigned to

bridge-bonded CO on Pt (19, 20).

cluster-derived sample, purging with He at room temper-

ature quickly removed the CO–Au peak, while an elevated

temperature was required for the elimination of the CO–Pt

peak.

FTIR studies of adsorbed CO were also conducted with

the prereduced 1Pt, 1Pt–2Au, and 0.5Pt₂Au₄samples at el-

evated temperatures. Results of these studies are shown in

Figs. 6 (for 1Pt) and 7 (for 0.5Pt₂Au₄). The results obtained

for the 1Pt–2Au are identical to those obtained with 1Pt

and have not been included for brevity. In all cases, when

the temperature was increased in the presence of gas-phase

CO while the surface was allowed to equilibrate, the vibra-

tion frequency of adsorbed CO was noticeably decreased.

It was also noticed that total CO desorption from Pt sites

in the cluster-derived 0.5Pt₂Au₄catalyst occurs at higher

temperatures than from the monometallic 1Pt sample. This

suggests that on the average the CO–Pt bond is stronger

in the cluster-derived catalyst (i.e., the fraction of lower

coordination Pt sites is higher in this sample). This result

is consistent with the substantial asymmetric tailing in the

FIG. 4. IR spectra of adsorbed CO on used reduced 0.5Pt₂Au₄cata-

lyst. Spectra were collected immediately following the removal of the gas-

phase CO: (a) 20 min at RT, (b) 25 min at RT, (c) 1 h at RT, (d) 20 min

at 50^◦C, (e) 20 min at 80^◦C, (f) 20 min at 125^◦C, (g) 20 min at 150^◦C,

(h) 20 min at 190^◦C.

130

MIHUT ET AL.

indicating minimal coverage effects in this case. The clear

difference in the magnitude of the shifts of the CO–Au and

CO–Pt bands is another indication that the two bands are

uncoupled and belong to different species. It is worth noting

that previous room-temperature studies of the adsorption

of CO on monometallic Au/SiO₂catalysts indicate a blue

shift with decreasing CO pressure (15, 21–23).

These results appear to suggest that the role of Au in the

cluster-derivedcatalystisbeyondthatofsimplybreakingup

large Pt ensembles, and that Au may be electronically modi-

fying the properties of Pt. However, due to the possibility

of islands of CO formation even at low surface coverages

(24), in which situation the dipole–dipole coupling is not

completely eliminated, this hypothesis was further veriﬁed

by collecting spectra of adsorbed ¹²CO/¹³CO mixtures at

different ratios between the two isotopes. The dilution of

one isotope in the other and the subsequent extrapolation

of the observed frequencies to zero coverage yields another

estimateofthesingletonfrequency. Thefrequencyof¹³CO–

Pt was used in these calculations, because of interference

between the frequency of ¹²CO–Pt and that of ¹³CO–Au (at

FIG. 6. FTIR spectra of adsorbed CO on reduced 1Pt collected in

the presence of gas-phase CO at different temperatures: (a) RT, (b) 70^◦C,

(c) 105^◦C, (d) 145^◦C, (e) 175^◦C, (f) 204^◦C, (g) 220^◦C, and (h) 245^◦C.

low-frequency region of the CO–Pt peak, which was dis- approximately 2064 cm⁻¹) in the 0.5Pt₂Au₄sample.

cussed in a previous paragraph. Furthermore, the results of

The IR spectra collected with different mixtures of ad-

Figs. 6 and 7 suggest that the observed red shift at room sorbed ¹²CO/¹³CO on pre-reduced 1Pt are shown in Fig. 8.

temperature (i.e., full coverage) between CO adsorbed on The intensities associated with the two bands do not nec-

Pt sites on 1Pt (or 1Pt–2Au) and 0.5Pt₂Au₄may be related essarily correspond to the ratio of the two isotopes in

to differences in the singleton frequency of the adsorbed the mixture, since a transfer of intensity from one band

CO on these samples, since at the corresponding highest to the other is possible, as was discussed previously by

temperatures (i.e., lowest CO coverage) this red shift is fur- Toolenaar et al. (25) for Pt–Cu alloys. For the same rea-

ther enhanced (from 2063 cm⁻¹for 1Pt and 1Pt–2Au to son, spectra collection at molar fractions above 4 : 1 or be-

2049 cm⁻¹for 0.5Pt₂Au₄). Finally, no signiﬁcant shift was low 1 : 4 was virtually impossible. The spectra were decon-

observed at elevated temperatures with the CO–Au band, voluted using curve-ﬁtting software packages, and peak

FIG. 7. FTIR spectra of adsorbed CO on reduced 0.5Pt₂Au₄col-

lected in the presence of gas-phase CO at different temperatures: (a) RT,

(b) 60^◦C, (c) 70^◦C, (d) 85^◦C, (e) 110^◦C, (f) 249^◦C, and (g) 275^◦C.

FIG. 8. RT spectra of mixtures of ¹²CO/¹³CO adsorbed on pre-

reduced 1Pt: (a) 100% ¹³CO, (b) 35% ¹²CO/65% ¹³CO, (c) 50% ¹²CO/50%

¹³CO, (d) 65% ¹²CO/35% ¹³CO, and (e) 100% ¹²CO.

CHARACTERIZATION OF Pt–Au CATALYSTS

131

100

2025

2000

1975

80

60

40

20

0

20

40

60

80

100

(% ¹³CO)

0

200

250

300

350

400

450

500

FIG. 9. RT stretching frequency of Pt-bound ¹³CO adsorbed on dif-

ferent Pt and Pt–Au catalysts as a function of the molar fraction of ¹³CO

in the ¹²CO/¹³CO mixture: ꢂ, 1Pt; ꢃ, 1Pt–2Au; ꢄ, 0.5Pt₂Au₄.

Temperature (⁰C)

FIG. 10. NO reduction and C₃H₆oxidation vs. temperature over dif-

ferent SiO₂-supported Pt and Pt–Au catalysts: ꢅ, ꢃ, 1Pt; ꢁ, ꢀ, 1Pt–2Au;

ꢆ, ꢄ, 0.5Pt₂Au₄; ꢇ, ꢂ, 2Au. Open symbols: C₃H₆oxidation; closed sym-

bols: NO reduction (1000 ppm NO, 1000 ppm C₃H₆, 1% O₂, balance He,

positions of ¹³CO were thus extracted. Plots of the fre-

quency of adsorbed ¹³CO on Pt obtained from this de-

convolution procedure versus the ¹³CO mole fraction in

the ¹²CO/¹³CO mixture for the three catalysts examined

are shown in Fig. 9. Extrapolation to zero concentration of

¹³CO in the gas phase yields the ¹³CO singleton frequency

at full coverage, eliminating the dipole–dipole interactions

between neighboring adsorbate molecules. These frequen-

cies are shown in Table 2. Similar values were obtained for

1Pt and 1Pt–2Au (2008 and 2014 cm⁻¹, respectively), while

a lower singleton frequency (1990 cm⁻¹) was obtained for

the cluster-derived 0.5Pt₂Au₄sample, consistent with the

results of the studies conducted at elevated temperatures.

The combination of the two results suggests that in the case

of the cluster-derived catalyst the electronic properties of

Pt have been modiﬁed by the presence of Au.

GHSV = 120,000 ml h⁻¹

g

⁻¹).

In both cases, the NO reduction and propylene oxidation

curves closely match each other up to the temperature of

maximum NO reduction. At higher temperatures, the re-

duction of NO decreases, while the propylene oxidation

reaches 100% and remains constant at this level. The se-

lectivity toward N₂in both cases ranges from 45 to 55%,

consistent with what was previously observed with similar

catalysts (26, 27).

A signiﬁcantly different behavior was observed for the

cluster-derived catalyst (0.5Pt₂Au₄). In this case, a temper-

ature delay of approximately 150^◦C was observed in both

the NO reduction and propylene oxidation curves. In ad-

dition, a slight separation of the C₃H₆oxidation and NO

reduction curves was observed, with C₃H₆oxidation pre-

ceding the NO reduction by approximately 10^◦C. Further-

more, the selectivity toward N₂in this case was in the range

of 70 to 80%, which is signiﬁcantly higher than what was ob-

served with the previous samples. Finally, the monometallic

2Au catalyst prepared by impregnation showed no activity

toward NO reduction and only minor oxidation activity at

temperatures above 450^◦C under the conditions studied.

A similar behavior was also observed for the oxidation

of propylene in the absence of NO. Results of these studies

are shown in Fig. 11. Once again, the performance of 1Pt

and 1Pt–2Au was very similar with only a minor tempera-

ture delay of approximately 10^◦C observed in the light-off

curve of the 1Pt–2Au sample. A signiﬁcant shift of 150^◦C

toward higher temperatures was observed with the cluster-

derived catalyst, and the 2Au catalyst was only active at

3.3. Kinetic Results

3.3.1. Selective catalytic reduction of NO by propylene.

The NO reduction and propylene oxidation versus temper-

ature curves for catalysts containing approximately 1 wt%

Pt and for 2 wt% Au loadings are shown in Fig. 10. Identi-

cal results were obtained with the 1Pt and 1Pt–2Au samples.

TABLE 2

Singleton Frequencies Obtained from

¹²CO/¹³CO Experiments

Catalyst sample

Singleton frequency (cm⁻¹

)

1%Pt/SiO₂

0.8%Pt–1.8%Au/SiO₂

0.4%Pt₂Au₄/SiO₂

2008

2014

1990

132

MIHUT ET AL.

100

the absence of NO. Finally, we should point out that the

results of kinetic studies conducted with a monometallic

0.15Pt catalyst prepared using an organometallic tert-butyl

ligated Pt cluster (10) indicate that the catalytic behavior

of this material is identical to that of the monometallic 1Pt

sample.

80

60

40

20

0

Previous studies have led to the conclusion that the

oxidation of hydrocarbons over Pt is a structure-sensitive

reaction. Carballo and Wolf (29), for example, reported that

propylene oxidation over alumina-supported Pt proceeds

at a signiﬁcantly lower rate on smaller Pt particles. In that

study, differences of 10 nm in particle size (i.e., for 1 to

11 nm) resulted in a ten-fold increase in the turnover fre-

quency. Similar results were also reported by Garetto and

Apesteguia for the oxidation of cyclopentane, methane, and

benzene (30, 31). A similar Pt particle size effect has also

been reported for the selective reduction of NO by hydro-

carbons. Lower turnover frequencies over smaller Pt crys-

tallites have been observed, for example, for the reduction

of NO by either propylene (32, 33) or octane (34). Further-

more, previous attempts to promote the catalytic activity

of Pt for the selective catalytic reduction of NO by propy-

lene by the addition of Au appear at ﬁrst to have had con-

tradictory results. Bimetallic Pt–Au catalysts prepared by

incipient wetness impregnation from individual metal salts

(35) or by redox techniques (36) have similar activity with

the monometallic Pt only catalysts, in agreement with the

results of this study. In contrast, reports from the patent

150

200

250

300

350

400

450

500

Temperature (⁰C)

FIG. 11. C₃H₆oxidation in the absence of NO vs. temperature

over different SiO₂-supported Pt and Pt–Au catalysts: ꢅ, 1Pt; ꢁ, 1Pt–

2Au; ꢆ, 0.5Pt₂Au₄; ꢇ, 2Au (1000 ppm C₃H₆, 1% O₂, balance He,

GHSV = 120,000 ml h⁻¹

g

⁻¹).

temperatures above 400^◦C. These results are consistent

with our previous mechanistic conclusions regarding the

kinetic signiﬁcance of the hydrocarbon activation during

HC-SCR at temperatures below the temperature of max-

imum NO reduction (28). In fact, the light-off curves of

Fig. 11 indicate that the observed shift in the NO reduction

proﬁle for the cluster-derived catalyst under HC-SCR con-

ditions can be directly attributed to the lower propylene

oxidation activity of this catalyst. In contrast, the presence

of Au in the co-impregnated sample does not have such an

effect.

Similar experiments were also conducted over a series

of catalysts with a nominal Pt loading of 0.15 wt%. As

explained in the experimental section, the experiments in

these cases were conducted at a proportionally lower space

velocity(20,000mlh⁻¹g⁻¹)tofacilitatecomparisonsamong

different catalyst samples. A small temperature delay (i.e.,

20–25^◦C) in both the NO reduction and C₃H₆oxidation

curves was observed with the co-impregnated 0.15Pt–0.3Au

sample compared to the monometallic 0.15Pt sample at

temperatures below the temperature of the maximum NO

reduction (Fig. 12). The selectivity toward N₂was approx-

imately 50%, and the magnitude of the maximum NO re-

duction was comparable in both cases. Once again, the

0.075Pt₂Au₄cluster-derived catalyst exhibited a signiﬁcant

shift toward higher temperatures in both the NO reduc-

tion and the C₃H₆oxidation proﬁles. As with the 0.5Pt₂Au₄

cluster-derived sample, the selectivity toward N₂was also

in the range of 70 to 80% in this case. A similar temper-

100

80

60

40

20

0

200

250

300

350

400

450

500

Temperature (⁰C)

FIG. 12. NO reduction and C₃H₆oxidation vs. temperature over dif-

ferent SiO₂-supported Pt and Pt–Au catalysts: ꢅ, ꢃ, 0.15Pt; ꢁ, ꢀ, 0.15Pt–

0.3Au; ꢆ, ꢄ, 0.075Pt₂Au₄; open symbols: C₃H₆oxidation; closed sym-

bols: NO reduction (1000 ppm NO, 1000 ppm C₃H₆, 1% O₂, balance He,

ature shift was observed with the propylene oxidation in GHSV = 20,000 ml h⁻¹

g

⁻¹).

CHARACTERIZATION OF Pt–Au CATALYSTS

133

It was previously shown that the ¹⁶O/¹⁸O homoexchange

reaction over supported Pt catalysts is a structure-sensitive

reaction favored over larger Pt particles or ensembles (38–

40). In light of this previous work, our results suggest that

in the case of the cluster-derived 0.5Pt₂Au₄catalyst the Pt

ensembles are smaller than in the case of the 1Pt or the

1Pt–2Au samples prepared by impregnation. The 1Pt and

1Pt–2Au samples appear to have similar Pt particle or en-

semble sizes. From this point of view, the results from the

¹⁶O/¹⁸O exchange reaction are in complete agreement with

the results of the selective catalytic reduction of NO by

propylene described in the previous section.

57

56

55

54

53

52

51

50

4. DISCUSSION

1.25

1.35

1.45

1.55

1.65

1.75

1.85

1.95

1000/T (K^-1)

Kinetic results obtained for the selective catalytic reduc-

tion of NO by propylene, the oxidation of propylene in

the absence of NO, and the ¹⁶O/¹⁸O homoexchange reac-

tion indicate signiﬁcant differences in the behavior of the

various SiO₂-supported Pt and Pt–Au catalysts examined

in this study. In particular, the cluster-derived 0.5Pt₂Au₄

catalyst shows a signiﬁcantly lower activity for all three re-

actions, which is manifested in the case of the HC-SCR

and the propylene oxidation as a shift of the conversion-

temperature curves toward higher temperatures. In addi-

tion, a higher selectivity for N₂is observed with this cata-

lyst during HC-SCR. In contrast, the 1Pt–2Au catalyst pre-

pared by co-impregnation from individual precursors of Pt

and Au exhibits behavior which is almost identical to that

of the monometallic 1Pt catalyst. Corroborated with the

fact that the monometallic 2Au catalyst prepared by im-

pregnation is inactive for all three reactions, these results

demonstrate that while Pt and Au do not affect each other

in the 1Pt–2Au system, the presence of Au has a strong ef-

fect on the catalytic behavior of Pt in the cluster-derived

0.5Pt₂Au₄catalyst.

During the presentation of the results it was indicated

that both the oxidation of hydrocarbons and the ¹⁶O/¹⁸O

homoexchange reactions are known to be structure sensi-

tive over supported platinum catalysts. In fact, the origin

of the structure sensitivity is the same in both cases and is

associated with the higher rate of the activation of molec-

ular oxygen over larger Pt ensembles. Our previous work

in the HC-SCR area has concluded that the activation of

propylene by molecular oxygen is the rate-determining step

of this reaction at temperatures below the temperature of

maximum NO conversion (28, 41). Consequently, it is not

surprising for HC-SCR to exhibit the same type of struc-

ture sensitivity as well. In fact, there are published litera-

ture reports that have experimentally demonstrated such

a structured sensitivity effect over supported Pt catalysts

with different Pt particle sizes (33, 34). Hence, in light of

these reports, it is reasonable to attempt to attribute the

differences in the kinetic results obtained in this study to

FIG. 13. Arrhenius plot for the ¹⁶O/¹⁸O homoexchange reaction: ꢄ,

1Pt; ꢃ, 1Pt–2Au; ꢀ, 0.5Pt₂Au₄.

literature indicate a promoting effect when the Au was

added by ion exchange or deposition methods (37).

In light of these reports, the results obtained in our study

suggest the possibility of similar Pt particle or ensemble (in

the case of the bimetallic catalysts) sizes for the monometal-

lic Pt/SiO₂and the co-impregnated Pt–Au/SiO₂samples,

and a signiﬁcantly smaller Pt particle or ensemble size for

the cluster-derived Pt₂Au₄/SiO₂catalysts.

3.3.2. ¹⁶O/¹⁸O exchange reaction. Arrhenius plots (uti-

lizing TOFs) for the ¹⁶O/¹⁸O exchange reaction over 1Pt,

1Pt–2Au, and 0.5Pt₂Au₄catalysts are shown in Fig. 13. The

results indicate that the cluster-derived 0.5Pt₂Au₄sample

is less active for the homoexchange reaction than the 1Pt

or the 1Pt–2Au catalysts. The 1Pt and 1Pt–2Au samples ex-

hibited similar activities. Finally, the 2Au sample was found

to be completely inactive for this reaction, consistent with

the idea that oxygen cannot be activated by gold. Appar-

ent activation energies for the ¹⁶O/¹⁸O exchange reaction

were calculated for the different samples from the slopes

of the respective curves and the results are summarized in

Table 3. A signiﬁcantly higher apparent activation energy

was observed over the cluster-derived sample.

TABLE 3

Apparent Activation Energy for the

¹⁶O/¹⁸O Homoexchange Reaction over

Different SiO₂-Supported Pt and Pt–Au

Samples

Catalyst sample

E_a(kJ/mol)

1Pt

1Pt–2Au

0.5Pt₂Au₄

90

122

148

134

MIHUT ET AL.

differences in Pt ensemble sizes and, in particular, to a cern between dipole–dipole coupling shifts and shifts due

smaller Pt ensemble size in the case of the cluster-derived to changes in the electronic state of the surface metal site.

0.5Pt₂Au₄catalyst.

Such estimates were obtained by two different approaches:

Indeed, the results of the TEM and hydrogen chemisorp- surface equilibration with gas-phase CO at elevated tem-

tion measurements suggest that the cluster-derived cata- peratures (i.e., equilibrated low surface coverages of CO)

lyst has a slightly smaller metal particle size (on the and use of ¹²CO/¹³CO mixtures at full coverage. Although

order of 1 nm) and a higher Pt dispersion than either the at ﬁrst it appears that both approaches provide the same

monometallic 1Pt or the co-impregnated 1Pt–2Au systems. result, in reality two different values of the singleton fre-

However, given the magnitude of the structure sensitivity quencycanbeobtained. Thehigh-temperatureexperiments

effects reported previously for the reactions examined (33, reveal the frequency of a single CO molecule adsorbed on

34, 39, 40), it becomes clear that these differences in metal a practically empty surface, while the experiments utilizing

particle size are not sufﬁcient to justify the observed dif- mixtures of the two isotopes reveal the frequency of an un-

ferences in catalytic behavior. As mentioned in the previ- coupled ¹²CO (or ¹³CO) molecule adsorbed on a surface

ous paragraph, the kinetic results suggest that the active saturated with ¹³CO (or ¹²CO) molecules.

Pt ensemble size in the cluster-derived 0.5Pt₂Au₄catalyst

The important point from the perspective of this paper

is signiﬁcantly smaller than the active Pt ensemble size in is that both approaches demonstrated differences of about

the monometallic 1Pt and the co-impregnated 1Pt–2Au sys- 15–20 cm⁻¹in the singleton frequency of adsorbed CO

tems. Since identical kinetic results were obtained with the on the cluster-derived 0.5Pt₂Au₄catalyst, compared to the

monometallic Pt catalysts irrespective of the precursor used monometallic 1Pt catalyst or the 1Pt–2Au catalyst prepared

(salt-derived or cluster-derived) as was discussed in sec- by co-impregnation from individual precursors of Pt and

tion 3.3.1, the smaller Pt ensemble sizes in the former case Au. These differences are a clear indication that in the

can be attributed to the presence of Au. In turn, this can cluster-derived catalyst not only the Pt ensemble size has

be interpreted as the ﬁrst indication of the formation of been affected, but also the electronic properties of Pt have

Pt–Au bimetallic particles in the 0.5Pt₂Au₄catalyst. The been modiﬁed by the presence of Au. Hence, the combi-

presence of Au in the surface of such particles is expected nation of kinetic and spectroscopic evidence leads to the

to “break down” large Pt ensembles and result in an av- conclusion that in the cluster-derived catalyst Pt and Au are

erage Pt ensemble size, which is signiﬁcantly smaller than intimately mixed in the form of bimetallic Pt–Au particles.

the average metal particle size. In addition, it is possible These particles are much larger than the original six metal

that the presence of Au in such bimetallic particles also re- atom clusters and, therefore, agglomeration is taking place

sults in a modiﬁcation of the electronic, and hence catalytic, duringtheligand-removalprocess. However, itappearsthat

properties of Pt.

during this agglomeration no signiﬁcant segregation of Pt

Further evidence for the formation of Pt–Au bimetallic and Au takes place.

particles in the 0.5Pt₂Au₄catalyst can be found in the re-

The formation of bimetallic particles is greatly dependent

sults of the CO adsorption experiments conducted in this on the precursors used. In contrast to the cluster-derived

study. The FTIR spectra of this catalyst collected at room catalyst, the 1Pt–2Au catalyst prepared by co-impregnation

temperature and full CO coverage indicate the presence from individual precursors of Pt and Au exhibits both cata-

of a new Au-bound weakly adsorbed surface CO species lytic and spectroscopic properties that are very similar to

and a red shift in the frequency of the Pt-bound CO. Pre- those of the monometallic 1Pt sample. This material ap-

vious literature reports indicate that only ﬁnely dispersed pears to follow a model in which Pt and Au are segregated,

Au crystallites lower than 5 nm in diameter can adsorb which would be expected based on the miscibility gap be-

CO. Therefore, the FTIR results suggest that Au is ﬁnely tween the two metals. Furthermore, Au in this material ap-

dispersed in this sample. The red shift observed in the pears to be in the form of larger crystallites since the FTIR

same FTIR spectra in the frequency of the Pt-bound CO studies did not reveal any CO adsorption. In fact, it is rea-

can be associated with either a geometric or an electronic sonable to assume that the majority of Au is in the form

effect. In the ﬁrst case, the geometric (“dilution”) effect is of the larger particles observed in the tail end of the metal

the result of Au breaking up the large Pt ensembles, thus particle size distribution of this catalyst (Fig. 1).

creating a weaker lateral interaction between adsorbed CO

Finally, we should point out that although the use of the

molecules (i.e., decreased dipole–dipole coupling in the ad- organometallicPt–Auclusterprecursorsappearstoprevent

layer). In the second case, electronic donation from one the segregation of the two metals into separate particles,

metal to the other can increase the back donation in the metal mobility within these particles and the subsequent

2π^∗antibonding orbitals of adsorbed CO molecules, and surface-enrichment effect cannot be ruled out. Thermody-

consequently cause the observed red shift. The estimation namic arguments indicate that the surface of the bimetallic

of the singleton frequency (i.e., the frequency of a single particles is expected to be enriched with the element having

CO molecule adsorbed on one metal site) can help to dis- the lowest surface tension. In the Pt–Au pair, Au exhibits

CHARACTERIZATION OF Pt–Au CATALYSTS

8. Chandler, B. D., and Pignolet, L. H., Catal. Today 65, 39 (2001).

135

the lowest surface tension and therefore a Au enrichment

of the surface is possible. However, a lower hydrogen up-

take would be expected if the surface was enriched in Au,

and no such trend was observed in the data of Table 1. This

may be due to the nanoscale dimension of the bimetallic

particles involved in this study. At this scale, metal behav-

ior may differ signiﬁcantly from “bulk” behavior observed

with larger particles.

9. Chandler, B. D., Schabel, A. B., and Pignolet, L. H., J. Phys. Chem. B

105, 149 (2001).

10. Chandler, B. D., Ph.D. dissertation, University of Minnesota,

Minneapolis, 1999.

11. Espinet, P., Fornies, J., Martinez, F., Tomas, M., Lalinde, E., Moreno,

M. T., Ruiz, A., and Welch, A. J., J. Chem. Soc., Dalton Trans. 791

(1990).

12. Martin, D., and Duprez, D., J. Phys. Chem. 100, 9429 (1996).

13. Balakrishnan, K., Sachdev, A., and Schwank, J., J. Catal. 121, 441

(1990).

14. Schwank, J., Balakrishnan, K., Sachdev, A., Volter, J., Ryczkowski,

J., Figueras, F., Grunert, W., Menon, P. G., Clarke, J. K. A., Coenen, J.,

and Paffett, M., Stud. Surf. Sci. Catal. 75, 905 (1993).

15. Yates, D. J. C., J. Colloid Interface Sci. 29, 194 (1969).

16. Boccuzzi, F., Guglielminotti, E., Pinna, F., and Strukul, G., Surf. Sci.

377, 728 (1997).

17. Liu, H., Kozlov, A. I., Kozlova, A. P., Shido, T., Asakura, K., and

Iwasawa, Y., J. Catal. 185, 252 (1999).

18. Conrad, H., Ertl, G., and Kuppers, J., Surf. Sci. 76, 323 (1978).

19. Delacruz, C., and Sheppard, N., Spectrochim. Acta Part A 50, 271

(1994).

20. Podkolzin, S. G., Shen, J., de Pablo, J. J., and Dumesic, J. A., J. Phys.

Chem. B 104, 4169 (2000).

21. France, J., and Hollins, P., J. Electron Spectrosc. Relat. Phenom. 64/65,

251 (1993).

22. Schubert, M. M., Kahlich, M. J., Gasteiger, H. A., and Behm, R. J.,

J. Power Sources 84, 175 (1999).

23. Boccuzzi, F., Chiorino, A., Tsubota, S., and Haruta, M., J. Phys. Chem.

100, 3625 (1996).

5. CONCLUSIONS

The use of an acetylide-ligated organometallic Pt–Au

cluster as a precursor has led to the synthesis of supported

Pt–Au catalysts containing bimetallic Pt–Au particles. In

contrast, a conventional synthesis approach utilizing incip-

ient wetness impregnation of individual Pt and Au precur-

sors results in catalysts containing segregated Pt and Au

catalysts. Catalysts prepared using the cluster route exhibi-

ted signiﬁcantly different catalytic behavior for a series of

test reactions, in addition to different adsorption proper-

ties for CO. These differences are the result of the presence

of ﬁnely dispersed Au, smaller Pt ensembles, and modiﬁca-

tion of the electronic properties of Pt by Au in the Pt–Au

bimetallic particles. These particles appear to be fairly sta-

ble and retain their properties and composition even after

prolonged exposure to HC-SCR conditions (i.e., oxidative

atmosphere and temperatures up to 500^◦C).

24. Crossley, A., and King, D. A., Surf. Sci. 95, 131 (1980).

25. Toolenaar, F. J. C., Stoop, F., and Ponec, V., J. Catal. 82, 1 (1983).

26. Roberts, K. L., and Amiridis, M. D., Ind. Eng. Chem. Res. 36, 3528

(1997).

27. Captain, D. K., Roberts, K. L., and Amiridis, M. D., Catal. Today 42,

93 (1998).

ACKNOWLEDGMENTS

28. Captain, D. K., Mihut, C., Dumesic, J. A., and Amiridis, M. D., Catal.

Lett., submitted for publication.

29. Carballo, L. M., and Wolf, E. E., J. Catal. 53, 366 (1978).

30. Garetto, T. F., and Apesteguia, C. R., Catal. Today 62, 189

(2000).

The authors express their gratitude to the U.S. Department of Energy

(DE-FG02-96ER14666) for ﬁnancial support of this work. M. D. A. also

acknowledges support from an international supplement from the Na-

tional Science Foundation (CTS-9624433). Finally, Bert D. Chandler is

acknowledged for the preparation of the organometallic clusters used as

precursors and for his useful suggestions during this work.

31. Garetto, T. F., and Apesteguia, C. R., Appl. Catal. B 32, 83

(2000).

32. Lee, J.-H., and Kung, H. H., Catal. Lett. 51, 1 (1998).

33. Denton, P., Giroir-Fendler, A., Praliaud, H., and Primet, M., J. Catal.

189, 410 (2000).

REFERENCES

1. Bouwman, R., and Sachtler, W. H. M., J. Catal. 19, 127 (1970).

2. Sachdev, A., and Schwank, J., J. Catal. 120, 353 (1989).

3. Gates, B. C., Chem. Rev. 95, 511 (1995).

4. Psaro, R., and Recchia, S., Catal. Today 41, 139 (1998).

5. Graf, I. V. G., Bacon, J. W., Curley, M. E., Ito, L. N., and Pignolet,

L. H., Inorg. Chem. 35, 689 (1996).

34. Burch, R., Fornasiero, P., and Southward, B. W. L., Chem. Commun.

625 (1998).

35. Burch, R., and Watling, T. C., Appl. Catal. B 11, 207 (1997).

36. Demichelli, M. C., Hoang, L. C., Menezo, J. C., Barbier, J., and

Pinabiau-Carlier, M., Appl. Catal. A 97, L11 (1993).

37. Nakatsuji, T., EP 0 602 602 A1 (1993).

6. Chandler, B. D., Rubinstein, L. I., and Pignolet, L. H., J. Mol. Catal.

133, 267 (1998).

7. Chandler, B. D., Schabel, A. B., Blanford, F., and Pignolet, L. H.,

J. Catal. 187, 367 (1999).

38. Taha, R., and Duprez, D., J. Chim. Phys. 92, 1506 (1995).

39. Duprez, D., Stud. Surf. Sci. Catal. 112, 13 (1997).

40. Descorme, C., and Duprez, D., Appl. Catal. A 202, 231 (2000).

41. Captain, D. K., and Amiridis, M. D., J. Catal. 194, 222 (2000).

Article Doi

Kinetic and spectroscopic characterization of cluster-derived supported Pt-Au catalysts

DOI: 10.1006/jcat.2002.3770

Source and publish data:

Authors:

Article abstract of DOI:10.1006/jcat.2002.3770

Full text of DOI:10.1006/jcat.2002.3770

Products guided by the article

Relevant to this article

Hot Product