Adsorption of NO on promoted Ag/α-Al2O3 catalysts

Journal of Catalysis 217 (2003) 442–456

www.elsevier.com/locate/jcat

Adsorption of NO on promoted Ag/α-Al₂O₃catalysts

Jale Müslehiddinog˘lu and M. Albert Vannice ^∗

Department of Chemical Engineering, 107 Fenske Laboratory, The Pennsylvania State University, University Park, PA 16802-4400, USA

Received 26 November 2002; revised 21 February 2003; accepted 25 February 2003

Abstract

Nitric oxide was used to probe Ag/α-Al O epoxidation catalysts. NO dissociated on reduced silver catalysts to oxidize metallic silver

2

3

sites and desorb N O at 300 K. Uptake measurements and DRIFTS results showed that Ag–O sites were the principal surface species formed

2

after NO adsorption at 300 K on reduced, unpromoted Ag catalysts, as indicated by the formation of gas-phase N O as well as a value of

2

unity for the NO/O uptake ratio. The formation of nitrite species was time dependent and a weak band due to chelating nitrito was observed

after long exposure times for the unpromoted catalyst after reduction at 473 K. On the unpromoted, O-covered Ag surface, NO adsorption

−1

produced a band at 1396 cm which was assigned to a nitrate species. Bands at 1238 and 1345 cm , assigned to chelating nitrito and

nitrate species, respectively, were detected after NO adsorption on Cs-promoted Ag surfaces reduced at 473 K or covered with chemisorbed

−1

oxygen following reduction at 673 K. Only the 1238 cm band was observed after NO adsorption on the Cs-promoted catalyst following

reduction at 673 K, and this band was the only absorption feature after NO adsorption on a Ag surface promoted with both Cs and Cl,

regardless of the pretreatment. In the presence of Cs, all results indicated that the chemistry of oxygen at the Ag surface was different and

more oxygen existed on the surface. The presence of surface oxygen enhanced NO adsorption, and the presence of both Cl and Cs may

x

decrease the surface concentration of nucleophilic oxygen, which forms nitrate species. After O and NO coadsorption on unpromoted Ag,

2

the O desorption peak was shifted 100 K higher to 660 K. When NO was adsorbed in the presence of gas-phase O , bulk Ag nitrate was a

2

dominant species.

Keywords: NO; Silver; Epoxidation; DRIFTS; TPD; Cs; Cl

1. Introduction

The distribution of electrons in the NO molecule is simi-

lar to that in the CO molecule, with the main difference being

that NO has one unpaired electron occupying the π^∗orbital.

Thus even slight changes in electron density due to NO ad-

sorption result in a noticeable change in the frequency of NO

stretching vibrations. The thermodynamic properties of NO

(ꢀG^◦_{f, 298 K}= 20.7 kcal/mol) indicate the potential instabil-

ity of the molecule [5]. For most transition metals, it has

been reported that the adsorption behavior of NO and CO

is similar [6], whereas for silver it differs in a way that the

heat of adsorption, Q_ad, is 5 kcal/mol for CO compared to

25 kcal/mol for NO, as reported by Behm and Brundle [6].

Most of the recent studies have been devoted to the investiga-

tion of NO_xreduction on Ag dispersed on γ -alumina [7–19],

whereas previous surface science studies focused on NO_x

adsorption on Ag single crystals, foils, or powders [20–40].

Contradictory results exist; for example, Lambert and co-

workers have reported molecular NO adsorption at 300 K

on the Ag (111) or Ag (110) surface [20–22], whereas the

Ag (111) surface has been found to be unreactive with NO at

A better understanding of the adsorption behavior of oxy-

gen and the effects of promoters on silver catalysts is crucial

for gaining insight into epoxidation reaction mechanisms.

CO has recently been used as a probe molecule to elucidate

the oxidation states of Ag and the effects of Cs and Cl [1].

In the current study, NO was also used to probe the effect of

promoters like Cs and Cl on the adsorption behavior of NO

and O₂, as well as oxygen adsorption on unpromoted Ag sur-

faces. Previous studies have used an O₂chemisorption–H₂

titration technique to determine the surface area and disper-

sion of silver in supported Ag catalysts [2–4]. If NO disso-

ciatively adsorbs on Ag at 300 K and its adsorption does not

lead to the formation of reactive NO_xproducts so that the

process can be stoichiometrically described, then NO might

also be used to count Ag sites.

*

Corresponding author.

E-mail address: mavche@engr.psu.edu (M.A. Vannice).

0021-9517/03/$ – see front matter

doi:10.1016/S0021-9517(03)00084-8

J. Müslehiddinog˘lu, M.A. Vannice / Journal of Catalysis 217 (2003) 442–456

443

295 K by others [26,34,37,38]. Surface science studies have

Table 1

Composition and Ag crystallite size of Ag/α-Al O catalysts

2

3

reported that NO adsorption is structure sensitive, and tem-

perature and pressure also affect its adsorption behavior. To

date researchers have mainly focused on the interaction be-

tween NO and Ag at subambient temperatures, which has re-

sulted in the formation of adsorbed (NO)₂dimers and N₂O,

as shown by X-ray photoelectron spectroscopy (XPS) [6,34]

and infrared spectroscopy [32,40].

The present study examines NO adsorption on α-Al₂O₃-

supported Ag catalysts as well as pure α-Al₂O₃to probe

the chemical state of the silver surface regarding the in-

ﬂuence of pretreatment and promoters by utilizing in situ

diffuse reﬂectance infrared Fourier transform spectroscopy

(DRIFTS), temperature-programmed desorption (TPD), and

chemisorption measurements. Unlike dioxygen, NO adsorp-

tion can give gas-phase products such as N₂O that are de-

tectable by FTIR as well as by TPD. Studies were conducted

on unpromoted Ag/α-Al₂O₃as well as this catalyst pro-

moted with either CsNO₃(to study the effect of only Cs)

or CsCl (to study the combined effect of both Cs and Cl).

a

b

c

d

Catalyst

Silver

(%)

Cesium

(ppm)

Chlorine

(ppm)

d

(nm)

UNP-I

UNP-II

CsN (409)

CsCl (412)

CsN (1174)

CsCl (994)

15.1

14.9

15.5

14.4

10.8

13.1

0

409

412

1174

994

0

130

0

310

–

275

283

247

354

–

a

Analyzed by ICP-OES.

Analyzed by AAS.

Nominal content based on cesium content.

Particle size calculated based on SEM analysis.

b

c

d

I LT (HT) (reduced Ag surfaces)

1. Calcine 2 h at 523 (773) K in 20% O₂at a total ﬂow of

50 sccm;

2. Reduce 1 h at 473 (673) K in 20% H₂at a total ﬂow of

50 sccm;

3. Purge at 473 (673) K for 30 min under ﬂowing He;

4. Cool to adsorption temperature.

2. Experimental

After the pretreatment, 4% NO in He (or a mixture of 3%

NO and 30% O₂in He) for 30 min at a total ﬂow of 20 sccm

was introduced.

Details regarding the preparation of these supported cat-

alysts, which contain approximately 15% Ag, have been

given elsewhere [1]. An unpromoted catalyst, designated

UNP-I, was prepared using α-Al₂O₃(Norton Corp., SA

5562, 0.78 m²/g) and then calcined in a forced air oven at

493 K for 30 min. Four Cs-promoted catalysts were prepared

with either a nominal 400 ppm Cs loading, designated CsN

(409) and CsCl (412), or a nominal 1200 ppm Cs loading,

designated CsN (1175) and CsCl (994), where the preﬁxes

CsN and CsCl indicate either CsNO₃or CsCl as the respec-

tive precursor and the number in parentheses indicates the

analyzed Cs content in ppm. A sample of α-Al₂O₃impreg-

nated with 1000 ppm Cs using CsNO₃precursor was also

prepared and it is designated Cs (1000, no Ag). After im-

pregnation, the promoted catalysts were calcined at 598 K in

air for 30 min. It has been reported elsewhere that the aver-

age Ag particle size in UNP-I is somewhat smaller than that

in the promoted catalysts, so experiments were also carried

out on another unpromoted Ag/α-Al₂O₃catalyst (UNP-II)

which was calcined at 598 K for 30 min, as were the pro-

moted catalysts, to compare a more similar Ag crystallite

size [41]. Properties of these six catalysts are tabulated in Ta-

ble 1. Experiments were also performed with pure α-Al₂O₃.

Four different pretreatments using two sets of tempera-

ture were applied and the procedures are given below, where

LT refers to the lower set of temperatures and HT refers to

the higher set of temperatures shown in parentheses. After

a speciﬁed pretreatment, these catalysts were characterized

by NO and O₂chemisorption, DRIFTS, TPD, and X-ray dif-

fraction (XRD).

II LT (HT) (O-covered Ag surfaces)

1. After applying I-LT (HT), introduce 10% O₂for 30 min

at 443 K at a total ﬂow of 20 sccm;

2. Purge at 443 K for 30 min under He ﬂow;

3. Cool to adsorption temperature.

After this pretreatment, 4% NO in He (20 sccm total ﬂow)

for 30 min was introduced.

NO and O₂chemisorption measurements were performed

in a glass gas adsorption system capable of a vacuum of

< 1 × 10⁻⁵Torr in the adsorption cell [42]. The system

was equipped with an Edwards Model EO2K oil diffusion

pump backed by a Precision mechanical pump, and the

vacuum was monitored by a Granville-Phillips Model 260

gauge controller. Isotherm measurements were obtained us-

ing a Mensor DPG II Model 15000 pressure transducer with

0.01% full-scale accuracy.

Ultrahigh purity (UHP) grade He and H₂(99.999%, MG

Ind.) were further puriﬁed by passing them through indi-

cating Oxytraps (Alltech Assoc.). All gases including O₂

(99.999%, MG Ind.) were passed through molecular sieve

traps, and NO (99%, MG Ind.) was used after passing it

through Ascarite and CaSO₄traps to remove CO₂and wa-

ter, respectively. ¹⁵NO (Cambridge Isotope Laboratories,

98%), which was introduced under static conditions, was

used without further puriﬁcation. Precise gas ﬂow rates were

maintained using mass ﬂow controllers (Tylan Corp., Model

FC-260). All DRIFT spectra were obtained at 300 K. Details

444

J. Müslehiddinog˘lu, M.A. Vannice / Journal of Catalysis 217 (2003) 442–456

of the diffuse reﬂectance FTIR system used in the present

study and the standard procedures employed to record spec-

tra have been described in detail elsewhere [1]. After loading

the sample into the DRIFTS reactor cell, the catalyst was

purged overnight under a He ﬂow. It was then subjected to an

in situ pretreatment and cooled to 300 K to obtain the ﬁrst in-

terferogram which was used as the background reference for

the subsequent spectra for that sample. Following this, 4%

NO (or a mixture of 3% NO and 30% O₂) was introduced at

300 K for 30 min at 20 cm³/min, at which time a second

interferogram was collected and, for some samples, addi-

tional interferograms were also recorded during this 30-min

exposure to NO. Then the sample was purged with He for

30 min and another interferogram was taken at that tempera-

ture (third spectrum). The gases were fed under atmospheric

pressure in all pretreatments and experiments. When ¹⁵NO

was used, 30 Torr was admitted to the cell under nonﬂow

conditions for 30 min. To determine the thermal stability

of adsorbed species, the sample in the cell was heated to a

certain temperature and then cooled to 300 K and a fourth

interferogram was then collected. All spectra are reported

as absorbance because the absorbance and Kubelka–Munk

spectra were typically similar, but the K-M function does

not display negative bands.

The temperature-programmed desorption experiments

were performed at 1 atm using a small ﬁxed bed of sam-

ple (100 mg, unless otherwise stated) placed in a 4 mm i.d.

quartz tube that was 30 cm long. A mixture of 4% NO in

He (MG Ind.) was used after passing through Ascarite and

CaSO₄traps to remove CO₂and water, respectively. After

applying the I-HT pretreatment, either a mixture of 3% NO

(as 4% NO in He, MG Ind.) and 30% O₂(UHP, MG Ind.)

in He or 10% O₂(UHP, MG Ind.) in He was introduced for

30 min followed by a 30-min purge. After that, TPD spectra

were obtained by heating the sample from room tempera-

ture to 973 K at a rate of 50 K/min under a ﬂow of 30 sccm

He. A small portion of the efﬂuent gas was routed via a

needle valve to a leak valve and continuously monitored by

an on-line, computer-controlled UTI quadrupole mass spec-

trometer for analysis.

Table 2

Adsorption of O at 443 K and NO at 300 K on Ag catalysts and α-Al O

2 3

2

Catalyst

Pretreatment

Gas uptakes (µmol/g

)

cat

a

O

NO

NO /O

(ad) (ad)

2

UNP-I

I-HT

II-HT

3.4 (1.0)

–

7.1

4.5

1.0

0.62

UNP-II

I-HT

II-HT

2.8 (0.93)

–

5.5

2.6

0.98

0.46

CsN (409)

CsCl (412)

CsN (1175)

CsCl (994)

Ag powder

I-HT

II-HT

2.5

2.0

2.2

0.82

0

6.6

3.6

1.32

0.72

I-HT

II-HT

4.5

2.0

1.1

0.50

I-HT

II-HT

8.1

2.7

2.0

0.67

I-HT

II-HT

3.7

1.8

0.84

0.41

I-HT

II-HT

1.0

0.65

0.61

0.40

α-Al O

I-HT

0.28

–

2

3

a

O

2

adsorption at 300 K in parentheses.

443 K. The O₂and NO uptakes are listed in Table 2, and

the corresponding adsorption isotherms are provided else-

where [43]. Oxygen uptakes after a low-temperature (LT)

pretreatment involving reduction at 473 K have been dis-

cussed elsewhere in detail [41]. With the unpromoted cat-

alysts, the irreversible “O” uptake was higher on UNP-I

than on UNP-II (6.8 µmol/g_catversus 5.6 µmol/g_cat, respec-

tively). The irreversible uptake of 5.0 µmol O/g_caton CsN

(409) was slightly lower than on its unpromoted counter-

part, UNP-II, while lower oxygen uptakes of 4.0, 4.0, and

4.4 µmol O/g_cat, were measured on CsCl (412), CsN (1175),

and CsCl (994), respectively. The O uptake on Ag powder

was 1.64 µmol/g, and no irreversible oxygen adsorption oc-

curred on α-Al₂O₃alone.

To examine the effect of the support, NO chemisorption

at 300 K was ﬁrst measured on pure α-Al₂O₃after a I-HT

pretreatment and the uptake was 0.28 µmol/g, which indi-

cates that no signiﬁcant nitric oxide adsorption occurred.

The NO uptake was 1.0 µmol/g on reduced Ag powder,

7.1 µmol/g_caton UNP-I, and 5.5 µmol/g_caton UNP-II. The

irreversible NO uptakes of 8.1 and 6.6 µmol/g_caton CsN

(1175) and CsN (409), respectively, are higher than that on

their unpromoted counterpart, UNP-II, whereas the NO up-

takes of 3.7 and 4.5 µmol/g_catfor CsCl (994) and CsCl

(412), respectively, were lower than those on the UNP and

CsN catalysts. The ratio of the NO uptake on the reduced Ag

surface at 300 K to the O uptake at 443 K was found to be

close to or slightly higher than unity for all catalysts except

the two CsN catalysts, and this is discussed later. After a II-

HT pretreatment, NO uptakes on the O-covered unpromoted

catalysts were 4.2 and 2.6 µmol/g_catfor UNP-I and UNP-

II, respectively, while they were 2.7 and 3.6 µmol/NO/g_cat

on CsN (1175) and CsN (409), respectively. The NO uptake

was 1.8 µmol/g_caton O-covered CsCl (994), 2.0 µmol/g_cat

XRD measurements were made using a Philips X’Pert

MPD X-ray diffractometer with Cu-K_αradiation which was

operated at 40 kV, 20 mA, and 0.2^◦/min with a 2^◦slit width.

3. Results

3.1. Chemisorption of O₂and NO on Ag catalysts

Each sample was subjected to a I-HT pretreatment and

then cooled to 443 K, at which temperature a monolayer

coverage of oxygen is achieved [2–4], and O₂uptakes were

measured. NO chemisorption at 300 K was studied on Ag

surfaces reduced at 673 K, and NO adsorption at 300 K on

O-covered Ag surfaces was examined following either a II-

HT pretreatment or an O₂chemisorption measurement at

J. Müslehiddinog˘lu, M.A. Vannice / Journal of Catalysis 217 (2003) 442–456

445

on O-covered CsCl (412), and 0.65 µmol/g_caton O-covered

Ag powder.

band observed after a 30-min exposure. NO adsorption on a

UNP-I catalyst after a I-LT pretreatment gave no signiﬁcant

IR features [43].

3.2. DRIFTS results

After a HT pretreatment, bands were similar for either

UNP-I or UNP-II [43]; therefore, only results with UNP-I

are presented here. Fig. 2a shows DRIFT spectra after NO

adsorption following reduction at 673 K. As with reduced

UNP-II, gas-phase N₂O was formed after the introduction

of NO for 1 min, and bands at 1640 and 1161 cm⁻¹were

observed after 30 min. To better identify these adsorbed

species, an isotope study was conducted and ¹⁵NO was ad-

sorbed on UNP-I after the same pretreatment but continuous

ﬂow was not used. The spectra in Fig. 2b show bands at 2154

and 1265 cm⁻¹which are due to ¹⁵N¹⁵N and ¹⁵NO stretch-

ing modes, respectively, of gas-phase ¹⁵N₂O. The position of

the 1165 cm⁻¹band was the same after either ¹⁴NO or ¹⁵NO

adsorption, which indicates that this band does not involve a

“N” atom. Unlike ¹⁴NO, ¹⁵NO was not introduced contin-

uously; thus, the intensity of gas-phase ¹⁵N₂O increased as

Gas-phase spectra for ¹⁴NO and ¹⁵NO have been pro-

vided elsewhere which show bands at 1875 (1841) and 2224

(2154) cm⁻¹due to the NO stretching vibration of ¹⁴NO

(¹⁵NO) and the N≡N stretching of ¹⁴N₂O (¹⁵N₂O), respec-

tively [43], because the latter gases were present as an im-

purity in the gas cylinders. These spectra were subtracted

from the corresponding spectra taken after exposure of any

sample to ¹⁴NO (¹⁵NO); therefore, any signiﬁcant N₂O for-

mation during NO adsorption could be differentiated from

the impurity.

DRIFT spectra after exposure of α-Al₂O₃to NO at

300 K following a I-HT pretreatment are given elsewhere

and demonstrate no detectable NO adsorption [43]. In ad-

dition to some OH, Al–O, and gas-phase N₂O stretching

bands, a weak band at 1640 cm⁻¹was observed which is

attributed to a bending mode of chemisorbed water because,

as stated later in this section, this band position did not shift

after ¹⁵NO adsorption on supported Ag catalysts.

Fig. 1 illustrates DRIFT spectra of NO adsorbed on UNP-

II after a I-LT pretreatment. After exposure of this sample

to NO for 1 min, a band at 2222 cm⁻¹due to the N≡N

stretching mode of N₂O was observed after subtracting the

gas-phase NO spectrum, and after a 30-min exposure to NO,

a band appeared at 1640 cm⁻¹; however, the intensity of

these bands is low and it is difﬁcult to distinguish them from

the impurity level as explained under Discussion. The time

dependency of the spectra was also investigated with this

sample and spectra were taken after 30, 60, and 100 min

of exposure to NO. These spectra exhibited the same IR fea-

tures, as shown by the third and fourth spectra in Fig. 1, and

the intensity of the band at 1232 cm⁻¹was greater after a

100-min NO exposure compared to that of the 1213 cm⁻¹

15

Fig. 1. DRIFT spectra of NO adsorbed on UNP-II at 300 K after a I-LT

pretreatment.

Fig. 2. DRIFT spectra of (a) NO and (b) NO adsorbed on UNP-I at 300 K

after a I-HT pretreatment.

446

J. Müslehiddinog˘lu, M.A. Vannice / Journal of Catalysis 217 (2003) 442–456

14

Fig. 4. (a) DRIFT spectra of NO adsorbed at 300 K on CsN (1175) after

15

either a I-LT or a II-LT pretreatment. (b) DRIFT spectra of NO adsorbed

at 300 K on CsN (1175) after a I-LT pretreatment.

14

Fig. 3. (a) DRIFT spectra of NO adsorbed on UNP-I at 300 K after a

15

II-HT pretreatment; (b) DRIFT spectra of NO adsorbed on UNP-I at

300 K after a II-HT pretreatment; inset shows comparison of spectra taken

14

15

after a 30-min purge, i.e., NO (third spectrum in (a)) and NO (second

spectrum in (b)).

Fig. 4a illustrates DRIFT spectra taken after NO adsorp-

tion on CsN (1175) following either a I-LT or a II-LT pre-

treatment. As with the Cs/α-Al₂O₃sample, a band near

1660 cm⁻¹was again observed, which suggests that it is a

surface species due to Cs. In addition, a dominant band ap-

peared at 1238 cm⁻¹with a shoulder around 1345 cm⁻¹and

weak “negative” bands at 1573 and 1407 cm⁻¹occurred for

the reduced sample while a different, sharper negative peak

occurred with the O-covered sample. After a 30-min expo-

sure of CsN (1175) to ¹⁵NO following a I-LT pretreatment,

gas-phase ¹⁵N₂O and bands at 1213 and 1300 cm⁻¹were

observed as shown in Fig. 4b. Their red shift compared to

their counterparts at 1238 and 1345 cm⁻¹indicated that both

are due to a species containing “N.”

Figs. 5a and 5b show NO adsorption on CsN (1175) af-

ter either a I-HT or a II-HT pretreatment. After exposure to

30 Torr NO for 1 min, gas-phase N₂O was present to give a

band at 2223 cm⁻¹and a broad band at 1510 cm⁻¹appeared

as soon as NO was introduced. As exposure time increased,

this latter peak shifted to 1520 cm⁻¹and its intensity in-

exposure time increased, indicating that continuous forma-

tion of nitrous oxide occurs during exposure to NO.

Figs. 3a and 3b depict spectra obtained after either ¹⁴NO

or ¹⁵NO adsorption on UNP-I after a II-HT pretreatment.

In contrast to the reduced catalyst, gas-phase N₂O was

not observed over the O-covered sample. The band near

1396 cm⁻¹after ¹⁴NO adsorption (Fig. 3a) was red-shifted

to 1352 cm⁻¹after ¹⁵NO adsorption (Fig. 3b), which implies

that this band is due to a species containing nitrogen.

A comparison between NO and O₂adsorption on Cs-

promoted α-Al₂O₃after a I-LT pretreatment has been shown

elsewhere [43]. Except for some gas-phase N₂O formation

during exposure to NO, the weak IR features were similar

after exposure to either NO or O₂. Considering possible as-

signment of the band positions, it is proposed that the bands

at 1655, 1390, and 1362 cm⁻¹were due to a CO₂impurity

in both the O₂and NO cylinders, as noted later.

J. Müslehiddinog˘lu, M.A. Vannice / Journal of Catalysis 217 (2003) 442–456

447

15

Fig. 6. DRIFT spectra of NO adsorbed on CsN (1175) at 300 K after a

I-HT pretreatment.

stretching modes, respectively. The band at 1217 cm⁻¹is

the counterpart of the 1237 cm⁻¹band, suggesting that this

surface species contains N. In addition, a band at 1522 cm⁻¹

,

similar to that at 1520 cm⁻¹after ¹⁴NO adsorption, existed,

and the absence of a red shift suggests that this is not an

N_xO_yspecies, as discussed later.

NO was adsorbed on the CsCl (994) catalyst to examine

the combined effect of Cs and Cl. As with all other samples,

after a 1-min exposure to NO following reduction at 473 K,

gas-phase N₂O was present, as shown in Fig. 7. The band

at 1238 cm⁻¹showed a time-dependent increase in inten-

sity after ¹⁴NO adsorption and it was red-shifted after ¹⁵NO

adsorption to 1211 cm⁻¹, while bands between 1616 and

1626 cm⁻¹and 1352 and 1381 cm⁻¹occurred after either

¹⁴NO or ¹⁵NO adsorption. After a 30-min purge, the band

intensity at 1238 (or 1211) cm⁻¹remained the same. Af-

ter heating the sample to 473 K and cooling to 300 K to

take a spectrum, a band near 1500 cm⁻¹band was formed,

as shown in Figs. 7a and 7b. After a I-LT, a II-LT, or a I-

HT pretreatment, absorption features were similar to those in

Fig. 7 [43], so the latter two sets of spectra are not presented.

Figs. 8a and 8b display DRIFT spectra taken after coad-

sorption of NO and O₂(3 and 30%) on supported Ag cat-

alysts, α-Al₂O₃, Cs/α-Al₂O₃, and promoted Ag/α-Al₂O₃

after either a 30-min exposure or a 30-min purge following

the NO + O₂exposure. When a mixture of NO and O₂was

introduced, gas-phase NO₂(1616 cm⁻¹) and N₂O₄(1756

and 1271 cm⁻¹) were formed. After purging, all gas-phase

NO_xcompounds were removed while the bands between

1355 and 1393, 1326 and 1361, and 1433 and 1474 cm⁻¹re-

tained the same intensity for all samples, and residual peaks

at 1757 and 2346 cm⁻¹remained with the supported Ag cat-

alysts (Fig. 8b). Bands at 1584, 1414, 1355, and 1322 cm⁻¹

were generated on alumina, while a dominant band appeared

at 1379 cm⁻¹with Cs/α-Al₂O₃.

Fig. 5. DRIFT spectra of NO adsorbed on CsN (1175) at 300 K: (a) after a

I-HT pretreatment; (b) after a II-HT pretreatment.

creased; however, after exposures larger than 17 min, the

intensity of this band decreased simultaneously as the band

intensity at 1237 cm⁻¹increased. In addition, a weak band

at 1142 cm⁻¹was observed after a 10-min NO exposure

which remained as a shoulder after a 30-min purge. Bands

near 1670 and 1380 cm⁻¹were observed as with adsorption

of NO on the Cs/α-Al₂O₃sample. After heating this sample

to 473 K and cooling it to 300 K, the bands at 1670, 1520,

and 1380 cm⁻¹disappeared and a band at 2153 cm⁻¹devel-

oped. The intensity of the 1237 cm⁻¹band decreased and,

after heating to 673 K, it completely disappeared, as shown

in the eighth spectrum in Fig. 5a. After this last heating cy-

cle, the intensity of the band at 2153 cm⁻¹increased and a

band at 1117 cm⁻¹appeared.

To help identify these surface species, experiments were

conducted with ¹⁵NO, and Fig. 6 illustrates the spectra taken

after ¹⁵NO adsorption on CsN (1175) following reduction at

673 K. After ¹⁵NO adsorption, gas-phase ¹⁵N₂O was present

to give bands at 2154 and 1265 cm⁻¹due to NN and NO

448

J. Müslehiddinog˘lu, M.A. Vannice / Journal of Catalysis 217 (2003) 442–456

Fig. 8. DRIFT spectra of NO adsorbed in the presence of O (23:230 Torr)

2

at 300 K after a I-HT pretreatment, unless otherwise stated: (a) after 30-min

14

15

Fig. 7. DRIFT spectra of (a) NO and (b) NO adsorbed on CsCl (994)

at 300 K after a I-LT pretreatment.

exposure; (b) after a 30-min purge; inset shows expanded scale for spectra

−1

of alumina and CsN (1000, no Ag) in the 1000–2500 cm spectral region.

3.3. Mass spectroscopy and TPD

680, and 750 K. O₂TPD was also conducted to probe the ef-

fect of Cs and to compare the adsorption behavior of oxygen

in the presence and absence of NO. Fig. 11 displays TPD re-

sults for UNP-II and CsN (1175) after exposure to 76 Torr

O₂at 443 K following a I-HT pretreatment, and respective

O₂desorption peaks were observed at 560 and 530 K, indi-

cating weaker oxygen adsorption on Cs-promoted Ag.

The efﬂuent gas from the TPD cell was analyzed by mass

spectroscopy during NO adsorption at 300 K. Fig. 9a clearly

shows that the introduction of NO to a reduced UNP-I cata-

lyst initially produced a signiﬁcant amount of N₂O, and sim-

ilar behavior occurred with reduced samples of the promoted

Ag catalysts [43]. In contrast, with O-covered Ag surfaces,

the introduction of NO at 300 K did not produce any N₂O

formation, as seen in Fig. 9b, and similar results were ob-

tained with the promoted Ag catalysts and alumina after a

II-HT pretreatment [43]. During a TPD run after exposure

of a catalyst to a NO/O₂mixture following a I-HT pretreat-

ment, one desorption maximum was observed for NO, O₂,

N₂O, and N₂around 620–660 K with all the Ag catalysts,

and Fig. 10a shows TPD results with UNP-I as a represen-

tative run. On pure α-Al₂O₃after a I-HT pretreatment, TPD

was performed after the introduction of a NO/O₂mixture, as

shown in Fig. 10b. No O₂desorption peak occurred, while

three weak desorption maxima for NO were observed at 465,

4. Discussion

Previous NO adsorption studies have been conducted

with Ag single crystals and Ag ﬁlms, mostly at subambi-

ent temperatures although a few studies have utilized am-

bient temperature. Lambert and co-workers conducted one

of the earliest surface science studies on NO adsorption on

Ag (111) and Ag (110) at 300 K using LEED, AES, and

TDS [20–22], and NO adsorption was found to be molecu-

lar and reversible with a sticking probability of 0.1, which is

very high in comparison with the sticking probability of O₂

J. Müslehiddinog˘lu, M.A. Vannice / Journal of Catalysis 217 (2003) 442–456

449

Fig. 9. (a) N O formation during initial NO adsorption on UNP-I at 300 K

2

after a I-HT pretreatment; (b) during initial NO adsorption on UNP-I at

Fig. 10. TPD after NO adsorption in the presence of O (a mixture of 3%

2

300 K after a II-HT pretreatment.

NO and 30% O ) at 300 K on (a) UNP-I; (b) α-Al O .

2

2 3

(∼ 10⁻⁴on Ag (110) and 10⁻⁶on Ag (111)) [44]. The NO

coverage was low and gave a comparatively high desorp-

tion activation energy of 24.7 kcal/mol. In contrast to these

results, Edamoto et al. reported that no adsorbed species ex-

isted at 300 K [26]. Similarly, Bao et al. examined the coad-

sorption of NO and O₂on a Ag (110) surface and found that

when only NO was introduced below 375 K, no evidence

for NO adsorption was obtained using Raman spectros-

copy [37]. XPS and UPS monitored a small amount of bulk

oxygen, but no signals relevant to surface oxygen and nitro-

gen were detected. Hadjiivanov studied NO_xadsorption on

Ag-ZSM-5 at 300 K using IR spectroscopy and, after intro-

duction of 10 Torr NO, a weak band at 1634 cm⁻¹occurred

which was stable toward evacuation and was assigned t₋o

bridged surface nitrate species [19]. The formation of NO₃

species requires adsorbed O atoms; thus, this study implied

NO dissociation on Ag because this was the only O source.

In most of the studies conducted at subambient temperatures,

the formation of NO₂and NO₃species was not observed but

N₂O and the (NO)₂dimer were detected [6,27,32]. However,

these different conclusions were reached based on differ-

ent coverages, temperatures, and techniques. Bartollucci and

Fig. 11. TPD after O adsorption at 443 K on unpromoted Ag/α-Al O

2 3

2

and Cs-promoted Ag/α-Al O following reduction at 673 K.

2

3

Otto used SERS to investigate the transformation of NO and

NO₂on silver ﬁlms at 30–40 K, and they found that after ex-

posure to 10 L NO, strong signals were detected originating

from NO₂⁻, N₂O, and atomic oxygen [30]. Although this

work was performed at subambient temperatures similar to

450

J. Müslehiddinog˘lu, M.A. Vannice / Journal of Catalysis 217 (2003) 442–456

the previous studies, Bartollucci and Otto did not observe

any signal due to NO, which they interpreted to mean that

NO adsorption on Ag is structure sensitive.

4.1. Band assignments

NO chemisorption can be divided into reactive and non-

reactive adsorption. During nonreactive adsorption, surface

mononitrosyl and dinitrosyl complexes are formed [45].

DRIFT spectra have not detected the formation of nitrosyl

complexes, ν(NO), on Ag, which have wavenumbers rang-

ing from 1900 to 1500 cm⁻¹. This is in accordance with the

study by Hadjiivanov which reported that NO adsorption on

Ag-ZSM-5 did not lead to formation of nitrosyl complexes

due to the weak electrophilicity of the univalent Ag⁺ions

and the absence of any π back-donation of electrons [19].

Most of the studies of NO on Ag have reported the for-

mation of an N₂O or NO dimer on the surface at subambient

temperatures. In our current study, gas-phase N₂O was ob-

served by DRIFTS and mass spectroscopy during NO ad-

sorption at 300 K on all the reduced Ag catalysts, which

necessitates NO dissociation on metallic silver. This result

agrees with that of Brown et al., who reported that N₂O

adsorbed on silver surfaces at low temperatures begins to

desorb at 120 K [32]. An isotopic study using ¹⁵NO was

conducted with unpromoted and Cs-promoted samples by

introducing ¹⁵NO to the reduced sample under static condi-

tions [43]. The intensity of the band due to gas-phase ¹⁵N₂O

increased as the exposure time increased, which indicates

that NO continued to oxidize metallic silver atoms during

the 30-min exposure period. IR bands due to gas-phase N₂O

were also detected after NO adsorption on HT-pretreated

alumina, although with the same sample no N₂O was de-

tected using mass spectroscopy. The NO uptake on α-Al₂O₃

was only 0.28 µmol/g, which implies that NO may disso-

ciate on alumina but to a much lesser extent. With all the

supported Ag samples, a band between 1620 and 1640 cm⁻¹

was observed. Although many NO_xspecies (NO₂, nitrites,

nitrates) absorb in this region, our studies of NO adsorption

on alumina and ¹⁵NO adsorption on supported Ag catalysts

have demonstrated that this band is due to a bending mode

of trace amounts of water chemisorbed on alumina [1].

With Cs-promoted samples, bands were observed be-

tween 1500 and 1480 cm⁻¹with either negative or positive

intensities. Negative bands appeared in this spectral region

with LT- and II-HT-pretreated samples, and this suggests that

certain species on the surface were removed during NO ex-

posure. With a I-HT-pretreated CsN (1175) catalyst, a band

developed at 1520 cm⁻¹during the initial adsorption of

NO but its intensity decreased as exposure time increased.

Similar behavior was observed for a 1472 cm⁻¹band with

I-HT-pretreated CsCl (994). Fig. 12 provides a comparison

between CO and NO adsorption on CsN (1175) and shows

that bands in the 1470–1530 cm⁻¹spectral region were evi-

dent after adsorption of either gas, which suggests that these

bands may be due to carbonate compounds. Similar features

Fig. 12. Comparison of DRIFT spectra taken for CsN (1175) at 300 K:

(A) after a 10-min exposure to 30 Torr NO after a I-HT pretreatment; (B) af-

ter a 30-min exposure to 76 Torr O after a I-HT pretreatment; (C) after a

2

30-min CO adsorption (76 Torr) following a II-LT pretreatment; (D) after a

30-min CO adsorption (76 Torr) following a II-HT pretreatment (all spectra

were taken after a 30-min purge).

were present after O₂adsorption ([43] and spectrum B in

Fig. 12). Although it might be considered that these bands

were due to molecularly adsorbed oxygen, as reported earlier

for metal oxides [46–49], the observation of similar bands

after CO adsorption on O-covered Ag catalysts provides a

strong argument that these IR features are due instead to car-

bonate formation. This proposal implies that both O₂and

NO cylinders have a CO₂impurity. Because the Ag sur-

face areas of the supported catalysts are small, even a low

concentration of impurities could have a signiﬁcant effect,

and our use of Ascarite to remove CO₂may not be com-

pletely efﬁcient. After the initial exposure of CsN (1175)

to NO, a band at 1520 cm⁻¹occurred, which was assigned

to carbonate formation on AgCs_xO_ysites [1]. As exposure

time increased, NO adsorption led to an exchange reaction

2−

with CO₃groups because NO competes for basic oxy-

gen anions on the surface and can displace CO₂as it is a

stronger Lewis acid, as shown in Fig. 5a. Zemlyanov and

Schlögl studied the interaction of the Ag (111) surface with

NO, O₂, and an NO/O₂mixture at 330 K using XPS, and

they reported that two peaks were observed which are char-

acteristic of carbonates and likely reﬂect that the oxygen

atoms associated with these carbonate species possess dif-

ferent charges [38]. In fact, the carbonate species is easily

produced by the following reaction:

CO₂+ O⁼_(ad)ꢀ CO₃_(ad)

.

=

This implies that in the current study CO₂(a Lewis acid) de-

sorbed from the surface and left basic oxygen anions on the

surface; therefore, gas-phase CO₂should be present during

loss of surface carbonate groups. However, CO₂formation

was not always detected in these experiments, which is pre-

sumably due to the ﬂow of the He carrier stream which

J. Müslehiddinog˘lu, M.A. Vannice / Journal of Catalysis 217 (2003) 442–456

451

Table 3

−

NO band assignments

2

Compound

Asymmetric

Symmetric

NO

Reference

[49]

2

N–O stretch

2

N–O stretch

2

deformation

−1

ν (cm

a

)

ν (cm

s

)

δ (cm

)

−

Free NO

1250

1335

–

2

Nitro complex

1470–1370

1335–1440

1340–1320

1350–1315

–

[49]

[59]

ν(N=O)

1485–1400

1450–1470

ν(NO)

1110–1050

1050–1065

–

[49]

Nitrito complex

Chelating nitro

–

[59]

[49]

1260–1390

1170–1210

840–860

–

Bridging nitrite

1220–1205

[45]

Inorganic compound

1343–1388

1330

1284–1328

1328–1375

1337–1429

1222–1260

1240

1176–1261

1180–1318

1114–1220

813–866

810

828–867

810–832

817–846

[52]

[53]

[54]

[55]

[56]

would continuously purge any desorbed CO₂. A displace-

ment reaction is apparent, but it is not clear why the CO₂

impurity is preferably adsorbed on the surface during the

initial NO exposure because NO is a stronger acid. These re-

sults have shown that the LT pretreatment at 473 K does not

completely remove carbonate compounds which were pre-

sumably formed during exposure of the catalyst to CO₂in

the ambient air. Carbonate compounds were also formed af-

ter a II-LT or a II-HT pretreatment, which included exposure

to O₂at 443 K for 30 min and therefore to CO₂as an impu-

rity in the oxygen cylinder.

Table 4

−

NO band assignments [Ref. from [49,57–61]]

3

Compound

Vibration

Band positions

−1

(cm

)

−

Free nitrite ion NO

Bidentate nitrate

ν

1380

3

(NO ,as)

2

ν

1200–1310

1500–1620

1003–1040

(NO ,as)

2

(N=O)

(NO ,s)

2

ν

1450–1570

1250–1330

970–1035

(NO ,as)

2

Unidentate nitrate

Bridging nitrate

(NO ,s)

2

(NO)

ν

1170–1300

1000–1030

1590–1660

(NO ,as)

This displacement reaction gave rise to a band around

1238 cm⁻¹for both Cs and CsCl-promoted catalysts, re-

gardless of the pretreatment, which could be ascribed to

NO₂⁻, NO₃⁻, NO⁻, or (NO)₂²⁻species. The possible pres-

ence of NO⁻species will ﬁrst be discussed. When NO is

adsorbed on a metal oxide such as Cr₂O₃and MgO, the

negatively charged nitrosyl ion, NO⁻, and the dimeric hy-

ponitrite ion, (NO)₂²⁻, are also observed in addition to ni-

trite and nitrate species [50,51]. The formation of a bond

between nitric oxide and any atom or group of atoms is ac_∗-

companied by an increase of the electron density in the π

orbitals of NO, which weakens the N–O bond, increases in

the N–O distance, and leads to a shift in ν(NO) to as low

as 1040 cm⁻¹[50]. However, when the sample was oxi-

dized, NO⁻species were not observed [50]. In our study,

the same band was obtained after NO adsorption on either

a clean Ag surface or an O-covered Ag surface; thus, the

1238 cm⁻¹band cannot be assigned to NO⁻. In the same

manner, (NO)₂²⁻can be observed only on a reduced surface;

hence, this band cannot be assigned to (NO)₂²⁻. A recent

review of the identiﬁcation of NO_xsurface species by IR

spectroscopy has pointed out the disagreement that exists in

2

(NO ,s)

2

(NO)

Inorganic compounds

ν

1470–1531

1253–1304

970–1055

(NO ,as)

2

(NO ,s)

2

(NO)

some guidelines were provided for the identiﬁcation of these

surface species [45]. There have been many studies in the

litera₋ture to differentiate among band positions assigned to

−

NO₂and NO₃groups. The early studies between 1957

and 1970 were performed with inorganic complexes such as

KNO₂, KNO₃, and NaNO₂[52–59] and they showed that

nitrite complexes and nitrate complexes gave similar bands.

−

Tables 3 and 4 list assignments to NO₂and NO₃species

based on inorganic compounds and metal oxides [45,49,52–

61]. In the current study, nitrate species were observed be-

tween 1300 and 1475 cm⁻¹after coadsorption of NO and O₂

on all supported Ag catalysts, as discussed later; therefore,

the band at 1238 cm⁻¹was not assigned to a nitrate com-

pound. When the N–O bonds in the NO₂group are equiv-

alent, a single band around 1200–1220 cm⁻¹is observed

because both symmetric and antisymmetric stretching vi-

−

the assignment of IR bands to different NO_xspecies, and

452

J. Müslehiddinog˘lu, M.A. Vannice / Journal of Catalysis 217 (2003) 442–456

brations, ν(NO₂), have nearly the same frequency and are

not resolved [49]; consequently, the 1238 cm⁻¹band may

be assigned to ν(NO₂) of a chelating nitrito species. The

experimental isotope shift with ¹⁵NO was between 20 and

26 cm⁻¹for this band with I-HT-pretreated CsN (1175) and

CsCl (994) samples. The chelating nitrito species preserve,

in general, the C_2νsymmetry [44]. Pinchas and Laulicht

have reported that for harmonic, nonlinear, symmetrically la-

beled Y–X–Y molecules, the following relation holds [62],

Table 5

Infrared band assignments for NO adsorbed on α-Al O -supported Ag at

300 K

2

3

Observed

Assignment

Based on

Reference

wavenumber

−1

(cm

)

−

15

1160–1165 Superoxide (O

1232–1238 Chelating nitrito

1396–1420 Surface nitrate

)

NO and O studies

2

[46–49]

[49,50]

2

NO

Coadsorption of NO and [45,49]

15

O

and NO studies

2

ꢀ

ꢁ₂

ꢀ

ꢁ

1317–1474 Bulk Ag nitrate

XRD, IR spectrum of

bulk Ag nitrate

[64]

ν₃ⁱ/ν₃= m_Xm_Ym_Xⁱ+ 2mⁱ_Ysin²α

(after coadsorption

of NO and O )

2

ꢀ

ꢁ

/mⁱ_Xmⁱ_Ym_X+ 2m_Ysin²α ,

(1)

where ν₃is the normal asymmetric stretching frequency of

the Y–X–Y molecule, m_Xand m_Yare the atomic weights

of its X and Y atoms, α is half the YXY angle, and the

superscript i denotes quantities belonging to the respective

labeled molecule. Assuming that the NO₂⁻oscillator is har-

monic and the ONO angle is 134^◦[63], a wavenumber of

1211 cm⁻¹is calculated for ¹⁵NO [43], which agrees well

Bands at 2153 and 1117 cm⁻¹were observed after NO

adsorption at 300 K on CsN (1175) reduced at 673 K fol-

lowed by heating to 673 K. The band at 2153 cm⁻¹can be

assigned to NO⁺, while the band at 1117 cm⁻¹can be at-

tributed to the O–O stretching mode of O₂⁻, shifted presum-

ably due to the Cs on the surface. Because the intensity of

these bands increased simultaneously with a decrease of the

1237 cm⁻¹band intensity, it is presumed that these species

formed as the chelating nitrito species decomposed on the

surface. Infrared band assignments for NO adsorbed on Ag

at 300 K are given in Table 5.

with the experimental values between 1217 and 1211 cm⁻¹

.

A band between 1330 and 1396 cm⁻¹was observed with

unpromoted Ag, Cs, and CsCl-promoted Ag catalysts, and

the Cs/Al₂O₃sample. The bands at 1360 and 1390 cm⁻¹

for the Cs/Al₂O₃sample may be due to cesium carbon-

ate species because similar bands were observed after CO

adsorption on this sample [1]. In addition, after ¹⁴NO and

¹⁵NO adsorption on CsCl (994), no change in position oc-

curred for the 1381 cm⁻¹band, which implies that this band

involves no N atoms and therefore may be due to a COO⁻

stretching mode of a carbonate compound. The band in this

region with CsN (1175) after either a LT or a II-HT pre-

treatment may have contributions from both carbonate and

nitrate species. After ¹⁵NO adsorption on CsN (1175) fol-

lowing reduction at 473 K, the band at 1345 cm⁻¹shifted

to 1300 cm⁻¹, which supports the idea that this band is due

to NO_xspecies. A similar study was conducted with ¹⁵NO

and UNP-I, and after ¹⁵NO adsorption on the O-covered

4.2. Coadsorption of NO and O₂at 300 K

Vibrational spectroscopy techniques have been used more

extensively for NO₂adsorption on Ag, and Table 6 sum-

marizes the IR absorption features of NO_xspeci₋es. These

studies have indicated that the formation of NO₃species

occurred regardless of the temperature range or the form of

Ag. NO adsorption on a reduced Ag surface in the presence

of oxygen produced gas-phase NO₂giving an intense peak

at 1616 cm⁻¹. A temperature-dependent equilibrium exists

between NO₂and its gas-phase dimer, N₂O₄[63], i.e.,

NO + O₂ꢁ NO₂, K_{1(300 K)}= 1.54 × 10⁶,

(2)

(3)

1

catalyst, the band at 1396 cm⁻¹red-shifted to 1352 cm⁻¹

.

2

Coadsorption of NO and O₂was also useful in assigning

the bands observed in this spectral region. The introduction

of a mixture of NO and O₂created a band between 1355

and 1385 cm⁻¹with all samples, which was assigned to ni-

trate species, as discussed later. Therefore, we conclude that

after NO adsorption on O-covered unpromoted Ag and Cs-

promoted Ag catalysts or Cs-promoted catalysts which were

given an LT pretreatment, nitrate species were formed on

the surface. As noted previously, exposure of the CsCl (994)

catalyst to NO created no nitrate band, which suggests that

surface Cl may block adsorption sites.

2NO₂ꢁ N₂O₄, K_{2(300 K)}= 6.66,

with N₂O₄giving bands at 1750 and 1270 cm⁻¹. Both reac-

tions are reversible and thermodynamic equilibrium shifts to

the right at lower temperatures [63]. After ﬂowing this (3%

NO + 30% O₂) mixture over the Ag catalysts for 30 min at

300 K, no NO peak was observed because the excess oxy-

gen allowed the conversion of essentially all the NO to NO₂.

Polzonetti et al. investigated the adsorption and reaction of

NO₂on Ag (111) at temperatures of 25, 90, and 300 K and

found that adsorption is entirely dissociative at 300 K, giv-

ing adsorbed NO and O species which strongly interact, in

addition to some NO₋desorption [28,29]. Outka et al. ob-

served NO₂and NO₃species when NO₂adsorbed on Ag

(110), with part of the chemisorbed NO₂giving₋a band at

1165 cm⁻¹and desorbing at 270 K, while NO₃was the

most stable species [25]. Zemlyanov and co-workers investi-

gated the reaction between Ag foil and a mixture of NO+O₂

With the reduced, unpromoted Ag catalysts, gas-phase

N₂O formation was an indication of NO adsorption and dis-

sociation. In addition, an 1165 cm⁻¹band occurred after NO

adsorption on the UNP-I catalyst reduced at 673 K, and this

−

band can be assigned to the O–O stretching mode of O₂

because after ¹⁵NO adsorption on this catalyst, no red-shift

was observed for this band.

J. Müslehiddinog˘lu, M.A. Vannice / Journal of Catalysis 217 (2003) 442–456

453

Table 6

Reported frequencies of adsorbed NO species observed on Ag

x

a

−1

)

Compound

NO

Form of Ag

Ag powder

Form of NO

System

SERS

Temperature (K)

Vibration

ν (cm

Reference

[23]

x

NO, NO

2

ν

a

1285

γ

815

ONO

−

NO

ν

1045

1360

1605

3

NO

ν

S

a

ν

−

NO

Ag Powder

Ag (110)

NO /N O

SERS

EELS

293–353

95–355

819

1040

[24]

[25]

2

3

2

4

ONO

ν

NO

1930

1745

795

2

ν

a

N O

2

4

ONO

ν

1390

1555 (weak)

S

−

NO

ν

3

Ag ﬁlms

NO, NO

SERS

30–40

1876

1263

806, 840

1279

2222

1037

705

1341

1267

1382

1718

810

[30]

2

ν

a

−

NO

γ

2

ONO

ν

NO

NN

NO

N O

2

ν

γ

ONO

ν

−

NO

3

a

ν , antiphase

ν , inphase

S

ν

a

N O

2

γ

NO

4

NO

(NO)

Ag (111)

NO

RAIRS

86

1940

1860–1880

1860–1880, 1590, 1285, 774

1590, 1265, 1045, 761

1717–1766, 1305, 774

[31]

[19]

2

N O

2

3

NO

N O

2

3

4

+

NO

N O

Ag-ZSM-5

NO and O

IR

RT

2140

1880, 1590

1748

2

3

N O

2

4

Chemisorbed NO

1676

2

Physisorbed NO

1607

2

NO , bridged

1630

3

NO , bidentate

3

1576

a

SERS, surface enhanced Raman spectroscopy; EELS, electron energy loss spectroscopy; IR, infrared spectroscopy.

at 300 K using XPS, ISS, and SEM to identify a complicated

sandwich structure consisting of superimposed AgNO₃and

Ag₂O layers at the surface of metallic Ag [35,36,38,39].

In our study, bands due to nitrate were obtained (Fig. 8),

and the TPD results were consistent with the DRIFT spec-

tra (Fig. 10a). The nitrate species formed was very stable

and decomposed to give NO, O₂, N₂O, and N₂peaks be-

tween 635 and 660 K for all samples. To gain additional

information about the form of nitrate, an XRD analysis was

conducted with UNP-II, and XRD patterns are shown in

Fig. 13 for both fresh UNP-II and UNP-II after exposure to a

mixture of NO and O₂at 300 K. For both samples, the prin-

cipal peaks for Ag and α-Al₂O₃were observed; however, for

the latter, additional AgNO₃peaks were also present, which

shows that bulk Ag nitrate formation is signiﬁcant. In fact,

the IR bands obtained after coadsorption of NO and O₂on

unpromoted Ag catalysts, shown in Fig. 8, are very similar

to an IR spectrum for AgNO₃[64]. It is difﬁcult to see the

effect of Cs and Cl because prominent IR features due to Ag

nitrate dominated for all supported Ag catalysts. The inset in

Fig. 8b shows a spectrum of alumina after the introduction

of a mixture of NO and O₂, and it reveals bands between

1585 and 1300 cm⁻¹which may be associated with nitrite

species because no TPD for O₂peak was detected. Recently,

Bao et al. studied coadsorption of NO and O₂on a Ag (110)

surface using XPS, UPS, and Raman spectroscopy, and they

found that the existence of oxygen enhances the adsorption

of NO_xspecies [37]. When they introduced₋a 1:10 mixture

of NO + O₂at 375 K, both NO₂and NO₃species were

formed, and Bao et al. reported that NO acted as a promoter

and led to cleavage of oxygen bond and that a surface NO–

O₂complex decomposed at 625 K, a result very similar to

ours. In fact, our results show that O₂desorption peaks were

observed around 530–560 K when only O₂was adsorbed

(Fig. 11), whereas the O₂desorption peak occurred around

630–660 K after coadsorption of NO and O₂(Fig. 10a),

454

J. Müslehiddinog˘lu, M.A. Vannice / Journal of Catalysis 217 (2003) 442–456

Fig. 13. XRD patterns: (A) UNP-II exposed to a NO/O mixture (22:220 Torr) at 300 K; (B) fresh UNP-II.

2

which indicates that in the presence of NO, a different form

of surface oxygen species exists which has a stronger inter-

action with the Ag surface.

4.3. NO adsorption at 300 K

Fig. 14 shows DRIFT spectra after NO adsorption on

supported Ag catalysts and Cs/alumina at 300 K follow-

ing a I-LT pretreatment. No signiﬁcant IR features occurred

with the unpromoted catalysts; however, the intensity of the

1232 cm⁻¹band due to chelating nitr₋ito increase₋s slightly

as exposure time increased. Both NO₂and NO₃species

were observed on the Cs-promoted Ag catalyst, as indi-

cated by the respective ₋band at 1238 and the shoulder at

1345 cm⁻¹, while NO₂was the only signiﬁcant N-con-

taining surface species with CsCl (994). It is possible that Cl

blocks sites for NO dissociation in the latter catalyst; there-

Fig. 14. DRIFT spectra of NO adsorbed at 300 K for a 30-min (unless oth-

erwise stated) at 300 K after a I-LT pretreatment.

−

fore, surface oxygen required for the formation of NO₃

species was not present.

With samples reduced at 673 K, no nitrate species were

observed after a 30-min NO exposure (Fig. 15). For the un-

promoted Ag catalysts, on₋ly a band near 1165 cm⁻¹due to

the O–O stretching of O₂was observed, which suggests

that NO dissociation occurred on the Ag catalyst; however,

no oxidation product was observed and only gas-phase N₂O

was detected. NO and O₂uptake measurements showed that

an NO/O_adratio close to unity was obtained for the unpro-

moted catalysts; consequently, it is proposed that NO disso-

ciates on the reduced Ag catalysts to form principally Ag–O

sites, i.e.,

Ag_s+ 2NO_(g)→ Ag_s–O + N₂O_(g)

.

This reaction could occur either by dissociative NO adsorp-

tion followed by N₂O formation or by the initial formation

of (NO)₂²⁻dimers followed by rapid decomposition to give

N₂O [6,32]. However, in the current study, there was no

spectral evidence for the formation of (NO₂)²⁻groups fol-

lowing NO adsorption. After N₂O desorption, the remaining

Fig. 15. DRIFT spectra of NO adsorbed at 300 K after a I-HT pretreatment

(all spectra were taken after a 30-min purge).

J. Müslehiddinog˘lu, M.A. Vannice / Journal of Catalysis 217 (2003) 442–456

455

Ag_s–O sites are similar to the sites produced after O₂ad-

sorption at 443 K.

as indicated by a band at 1396 cm⁻¹. The intensity of the

band due to chelating nitrito increased considerably for the

Cs-promoted Ag catalysts. This was not due to NO adsorp-

tion on Cs sites, which was demonstrated by experiments

with a Cs/Al₂O₃sample, but rather was related to a change

in chemical behavior of the surface oxygen associated with

Cs. After reduction at 473 K of CsN (1175), which con-

tained no Cl, and the formation of an oxygen monolayer,

nitrate species were formed during NO adsorption. In con-

trast, exposure of this catalyst to NO after reduction at 473 K

produced both nitrate and nitrite species, but only the latter

species were observed after NO exposure following reduc-

tion at 673 K. This was attributed to redistribution of Cs,

due to the higher temperature in the latter case, which de-

creased the nucleophilic oxygen on the surface associated

with the formation of nitrate species. Surface Cl blocked Ag

sites and prevented nitrate formation on CsCl (994); how-

ever, similar to CsN (1175), a chelating nitrito species was

formed on this catalyst regardless of the pretreatment. The

presence of adsorbed oxygen enhanced NO_xinteraction with

the Ag surface and, after O₂and NO coadsorption, this ad-

sorbed oxygen was more strongly bound on the Ag surface.

Bao et al. found that the existence of surface oxygen en-

hances the adsorption of NO_x[37]; therefore, this could

−

explain the enhanced formation of NO₂on all the Cs-

promoted Ag catalysts and the formation of nitrate on the

LT- and II-HT-pretreated CsN (1175) samples; i.e., the pres-

ence of Cs alters the adsorption behavior of NO and the

oxygen chemistry on the surface. These results imply that

more surface oxygen should exist on the CsN (1175) cata-

lyst, which is consistent with previous results showing that

the reversible O₂uptake increased by about 30% although

the irreversible uptake did decrease [41]. In addition, as

shown in Fig. 11, the desorption peak for O₂was 30 K lower

for CsN (1175) than for the unpromoted Ag catalyst, which

shows that O₂is more weakly adsorbed on CsN (1175) than

on UNP-II. Consequently, it is hypothesized that Cs facili-

tates₋NO dissociation to oxidi₋ze metallic Ag sites and form

NO₂(1238 cm⁻¹) and NO₃(1345 cm⁻¹) species. Lam-

bert and co-workers have reported that, in the presence of

Na, NO was more strongly adsorbed and underwent surface

dissociation to yield adsorbed O and N atoms whose sub-

sequent reactions led to the formation of N₂, N₂O, and O₂

as gaseous products, whereas nondissociative NO adsorp-

tion occurred on a clean Ag (111) surface [22]. This was

attributed to Na being an electron donor, thus increasing the

bond strength of adsorbed NO [21,22]. The DRIFTS results

here are consistent with the chemisorption studies. With the

reduced Cs-promoted Ag catalysts, the NO uptake increased,

the CsN (1175) catalyst had a NO_ad/O_adratio of 2.0, and

no nitrate species were formed. These results are consistent

with the presence of composite AgCs_xO_ysites [1]. It has

been discussed previously that after a HT-pretreatment, re-

distribution of Cs may occur at 673 K to form islands of Ag

atoms separated by Cs atoms [1], which might inhibit the

formation of nitrate species, and Goncharova et al. have re-

ported that Cs can decrease the concentration of nucleophilic

oxygen on a Ag surface [65].

Acknowledgments

Partial support of this study by Eastman Chemical Com-

pany is gratefully noted. One of us (J.M.) thanks the Turkish

Scientiﬁc and Research Council for a NATO Science Fel-

lowship.

References

[1] J. Müslehiddinog˘lu, M.A. Vannice, J. Catal. 213 (2003) 305.

[2] D.E. Strohmayer, G.L. Geoffroy, M.A. Vannice, Appl. Catal. 7 (1983)

189.

[3] S.R. Seyedmonir, D.E. Strohmayer, G.L. Geoffroy, M.A. Vannice,

J. Catal. 87 (1984) 424.

[4] S.R. Seyedmonir, D.E. Strohmayer, G.L. Geoffroy, M.A. Vannice,

J. Catal. 93 (1985) 288.

[5] C.L. Yaws, Chemical Properties Handbook, McGraw–Hill, New York,

1999.

[6] R.J. Behm, C.R. Brundle, J. Vac. Sci. Technol. 2 (1984) 1040.

[7] T. Miyadera, Appl. Catal. B: Environ. 2 (1993) 199.

[8] K. Masuda, K. Tsujimura, K. Shinoda, T. Kato, Appl. Catal. B: Envi-

ron. 8 (1996) 33.

[9] K.A. Bethke, H.H. Kung, J. Catal. 172 (1997) 93.

[10] T. Miyadera, Appl. Catal. B: Environ. 13 (1997) 157.

[11] T.E. Hoost, R.J. Kudla, K.M. Collins, M.S. Chattha, Appl. Catal. B:

Environ. 13 (1997) 59.

[12] S. Kameoka, T. Chaﬁk, Y. Ukisu, T. Miyadera, Catal. Lett. 55 (1998)

211.

5. Summary

NO adsorption was examined on supported and promoted

Ag catalysts to gain insight into the nature of adsorption

sites and the effect of Cs and Cl promoters. At 300 K NO

dissociates on reduced silver catalysts by oxidizing metal-

lic silver surface atoms and desorbing N₂O. Chemisorption

results and DRIFT spectra indicated that Ag–O sites were

formed on unpromoted catalysts during NO adsorption after

reduction at either 473 or 673 K. The concentration of nitrite

species was time dependent and a weak band at 1232 cm⁻¹

due to a chelating nitrito species was observed for an unpro-

moted Ag catalyst following reduction at 473 K and longer

exposures times. After reduction at 673 K, chemisorption

of O₂to give a monolayer coverage, and exposure to NO,

nitrate species were formed on the unpromoted catalysts,

[13] H. Jen, Catal. Today 42 (1998) 37.

[14] D.B. Akolekar, S.K. Bhargava, J. Mol. Catal. A 157 (2000) 199.

[15] S. Kameoka, Y. Ukisu, T. Miyadera, Phys. Chem. Chem. Phys. 2

(2000) 367.

[16] V. Zuzaniuk, F.C. Meunier, J.R.H. Ross, J. Catal. 202 (2001) 340.

[17] I.H. Son, M.C. Kim, H.L. Koh, K.L. Kim, Catal. Lett. 75 (2001) 191.

[18] A. Keshavaraja, X. She, M. Flytzani-Stephonopoulos, Appl. Catal. B:

Environ. 27 (2000) L1.

[19] K.I. Hadjiivanov, Micropor. Mesopor. Mater. 24 (1998) 41.

456

J. Müslehiddinog˘lu, M.A. Vannice / Journal of Catalysis 217 (2003) 442–456

[20] R.A. Marbrow, R.M. Lambert, Surf. Sci. 61 (1976) 317.

[21] R.A. Marbrow, R.M. Lambert, Surf. Sci. 61 (1976) 329.

[22] P.J. Goddard, J. West, R.M. Lambert, Surf. Sci. 71 (1978) 447.

[23] K.U. Von Raben, P.B. Dorain, T.T. Chen, R.K. Chang, Chem. Phys.

Lett. 95 (1983) 269.

[24] P.B. Dorain, J.E. Boggio, J. Chem. Phys. 84 (1986) 135.

[25] D.A. Outka, R.J. Madix, G.B. Fisher, C. Dimaggio, Surf. Sci. 179

(1987) 1.

[43] J. Müslehiddinog˘lu, PhD thesis, The Pennsylvania State University,

2002.

[44] F. Besenbacher, J.K. Nørskov, Prog. Surf. Sci. 44 (1993) 5.

[45] K.I. Hadjiivanov, Catal. Rev. Sci. Eng. 42 (2000) 71.

[46] A.A. Davydov, M.P. Komarova, V.F. Anufrienko, N.G. Maksimov,

Kinet. Catal. 14 (1973) 1342.

[47] A.A. Tsyganenko, T.A. Rodionova, V.N. Filimonov, React. Kinet.

Catal. Lett. 11 (1979) 113.

[26] K. Edamoto, S. Maehama, E. Miyazaki, T. Miyahara, H. Kato, Surf.

Sci. 204 (1988) L739.

[48] A.A. Davydov, Kinet. Catal. 20 (1979) 1245.

[49] A.A. Davydov, Infrared Spectroscopy of Adsorbed Species on the Sur-

face of Transition Metal Oxides, Wiley, New York, 1984.

[50] A.A. Davydov, Yu.A. Lokhov, Yu.M. Shchekochikhin, Kinet. Catal.

19 (1978) 532.

[51] M.J.D. Low, R.T. Yang, J. Catal. 34 (1974) 3479.

[52] M.H. Brooker, D.E. Irish, Canad. J. Chem. 46 (1968) 229.

[53] M.H. Brooker, D.E. Irish, Inorg. Chem. 8 (1969) 219.

[54] D.M.L. Goodgame, M.A. Hitchman, Inorg. Chem. 4 (1965) 721.

[55] L. El-Sayed, R.O. Ragsdale, Inorg. Chem. 6 (1967) 1640.

[56] D.M.L. Goodgame, M.A. Hitchman, Inorg. Chem. 3 (1964) 1389.

[57] F.A. Cotton, M.L. Goodgame, R.H. Soderberg, Inorg. Chem. 2 (1963)

1162.

[58] N.F. Curtis, Y.M. Curtis, Inorg. Chem. 4 (1965) 804.

[59] K. Nakamoto, Infrared and Raman Spectra of Inorganic Coordination

Compounds, 4th ed., Wiley, New York, 1986.

[60] J.W. London, A.T. Bell, J. Catal. 31 (1973) 32.

[61] C.C. Addison, N. Logan, S.C. Wallwork, C.D. Garner, Q. Rev. 25

(1971) 289.

[27] S.K. So, R. Franchy, W. Ho, J. Chem. Phys. 91 (1989) 5701.

[28] G. Polzonetti, P. Alnot, C.R. Brundle, Surf. Sci. 238 (1990) 226.

[29] G. Polzonetti, P. Alnot, C.R. Brundle, Surf. Sci. 238 (1990) 237.

[30] F. Bartollucci, A. Otto, Surf. Rev. Lett. 1 (1994) 561.

[31] W.A. Brown, P. Gardner, D.A. King, Surf. Sci. 330 (1995) 41.

[32] W.A. Brown, P. Gardner, D.A. King, J. Chem. Phys. 99 (1995) 7065.

[33] S.R. Bare, K. Grifﬁths, W.N. Lennard, H.T. Tang, Surf. Sci. 342 (1995)

185.

[34] A.F. Carley, P.R. Davies, M.W. Roberts, A.K. Santra, K.K. Thomas,

Surf. Sci. 406 (1998) L587.

[35] D.Yu. Zemlyanov, A. Nagy, R. Schlögl, Appl. Surf. Sci. 133 (1998)

171.

[36] D.Y. Zemlyanov, A. Hornung, G. Weinberg, U. Wild, R. Schlögl,

Langmuir 14 (1998) 3242.

[37] X. Bao, U. Wild, M. Muhler, B. Pettinger, R. Schlögl, G. Ertl, Surf.

Sci. 425 (1999) 224.

[38] D.Y. Zemlyanov, R. Schlögl, Surf. Sci. 470 (2000) L20.

[39] D.Y. Zemlyanov, G. Weinberg, U. Wild, R. Schlögl, Catal. Lett. 64

(2000) 113.

[40] W.A. Brown, P. Gardner, M.P. Jigato, D.A. King, J. Chem. Phys. 102

(1995) 7277.

[62] S. Pinchas, I. Laulicht, Infrared Spectroscopy of Labelled Compounds,

Academic Press, London, 1971.

[63] J. Laane, J.R. Ohlsen, Prog. Inorg. Chem. 27 (1980) 465.

[64] R.J. Keller, The Sigma Library of FT-IR Spectra, 1st ed., 1986.

[65] S.N. Goncharova, E.A. Paukshtis, B.S. Bal’zhinimanev, Appl. Catal.

A: Gen. 126 (1995) 67.

[41] M.V. Badani, M.A. Vannice, Appl. Catal. A 204 (2000) 129.

[42] M.B. Palmer, M.A. Vannice, J. Chem. Technol. Biotech. 30 (1980)

205.

Article Doi

Adsorption of NO on promoted Ag/α-Al2O3 catalysts

DOI: 10.1016/S0021-9517(03)00084-8

Source and publish data:

Authors:

Article abstract of DOI:10.1016/S0021-9517(03)00084-8

Full text of DOI:10.1016/S0021-9517(03)00084-8

Products guided by the article

Relevant to this article

Hot Product