This journal is c the Owner Societies 2011 Phys. Chem. Chem. Phys., 2011, 13, 4607–4613 4607

Cite this: Phys. Chem. Chem. Phys., 2011, 13, 4607–4613

In situ Raman and in situ XRD analysis of PdO reduction and Pd1

oxidation supported on c-Al2O3 catalyst under different atmospheres

Alexandre Baylet,*aPatrice Marecot,

aDaniel Duprez,

aPaola Castellazzi,

b

Gianpiero Groppiband Pio Forzatti

b

Received 27th July 2010, Accepted 14th December 2010

DOI: 10.1039/c0cp01331e

Reduction of Pd1 and decomposition of palladium oxide supported on g-alumina were studied at

atmospheric pressure under different atmospheres (H2, CH4, He) over a 4 wt% Pd/Al2O3 catalyst

(mean palladium particle size: 5 nm with 50% of small particles of size below 5 nm). During

temperature programmed tests (reduction, decomposition and oxidation) the crystal domain

behaviour of the PdO/Pd1 phase was evaluated by in situ Raman spectroscopy and in situ

XRD analysis. Under H2/N2, the reduction of small PdO particles (o5 nm) occurs at room

temperature, whereas reduction of larger particles (45 nm) starts at 100 1C and is achieved at

150 1C. Subsequent oxidation in O2/N2 leads to reoxidation of small crystal domain at ambient

temperature while oxidation of large particles starts at 300 1C. Under CH4/N2, the small particle

reduction occurs between 240 and 250 1C while large particle reduction is fast and occurs between

280 and 290 1C. Subsequent reoxidation of the catalyst reduced in CH4/N2 shows that small and

large particle oxidation of Pd1 starts also at 300 1C. Under He, no small particle decomposition is

observed probably due to strong interactions between particles and support whereas large particle

reduction occurs between 700 and 750 1C. After thermal decomposition under He, the oxidation

starts at 300 1C. Thus, the reduction phenomenon (small and large crystal domain) depends on

the nature of the reducing agent (H2, CH4, He). However, whatever the reduction or

decomposition treatment or the crystal domain, Pd1 oxidation starts at 300 1C and is completed

only at temperatures higher than 550 1C. Under lean conditions, with or without water, the

palladium consists of reduced sites of palladium (Pd1, Pdd+ with d o 2 or PdOx with x o 1)

randomly distributed on palladium particles.

1. Introduction

Palladium is widely used as automotive catalyst (in substitution

of or in addition to Pt and Rh) for the abatement of HC, CO

and NOx emissions.1 It is the most active precious metal for

methane combustion in excess of oxygen (lean condition). For

that reason, palladium catalysts are strongly recommended

for depollution of natural gas-powered vehicles and in catalytic

processes for energy production from natural gas. Aging of

these catalysts is often caused by decomposition of the active

phase into metallic palladium. Pd1 is less active than oxidized

palladium in CH4 oxidation. Moreover, metallic palladium

particles sinter more easily than PdO particles, resulting in

an irreversible loss of activity.2 These properties have been

extensively detailed in past reviews on catalytic combustion of

methane over Pd-based catalysts.3–5 Many studies have been

devoted to better describe the optimum state of PdOx in

reaction: either chemisorbed oxygen on Pd1 or a PdO skin

on a Pd metal core or bulk PdO;4 most of these studies

concluded that the two latter forms would be the most active

state of palladium supported catalysts.6 For this purpose

many investigations (experimental or theoretical) on the

stability of the palladium oxide particles during reduction

and palladium metal reconstruction during oxidation have

been done over various catalysts: (i) single Pd crystal; (ii)

polycrystalline Pd foil;7–12 (iii) and Pd supported catalysts13–17

during in situ or ex situ experiments. However, the experi-

mental conditions may change the oxidation state (ultra high

vacuum, 10�10 atm, is required for atomic emission spectrometry

(AES), X-ray photoelectron spectroscopy (XPS), secondary

ion mass spectrometry (SIMS) or low-energy electron diffrac-

tion (LEED) analysis) or modify palladium particle morpho-

logy upon cooling for ex situ analysis (X-ray diffraction (XRD)

or Raman spectroscopy). From these studies, reduction models

show a core-shell development of the catalyst surface, whereas

a LACCO, UMR CNRS 6503, Universite de Poitiers, 40,Avenue du Recteur Pineau, 86022 Poitiers, Cedex, France.E-mail: alexandre.baylet@univ-poitiers.fr;Fax: +33-(0)5-49-45-34-99; Tel: +33-(0)5-49-45-39-98

b Laboratory of Catalysis and Catalytic Processes,Dipartimento di Energia, Politecnico di Milano,Piazza Leonardo da Vinci 32, 20133 Milano, Italy

PCCP Dynamic Article Links

www.rsc.org/pccp PAPER

Dow

nloa

ded

by U

nive

rsity

of

Cal

ifor

nia

- Sa

n D

iego

on

07 M

arch

201

3Pu

blis

hed

on 3

1 Ja

nuar

y 20

11 o

n ht

tp://

pubs

.rsc

.org

| do

i:10.

1039

/C0C

P013

31E

View Article Online / Journal Homepage / Table of Contents for this issue

4608 Phys. Chem. Chem. Phys., 2011, 13, 4607–4613 This journal is c the Owner Societies 2011

the experiments give two different ways depending on the

reducing agent: (i) reduction with H2 occurs via a shrinking

core mechanism; whereas (ii) reduction with CH4 occurs in

a ‘‘cauliflower-like structure’’ implying roughening of the

surface and so an increase of palladium oxide surface area.18

The oxidation process generally occurs in two steps: (i) oxygen

diffusion into Pd metal followed by (ii) oxygen diffusion

into PdO once the bulk oxide layer was formed. The oxygen

uptake was linearly proportional to the square root of the

time of treatment, No = K(P,T)t1/2. Thus the dissolution of

oxygen atoms into Pd metal followed the Mott-Cabrera

model19 with diffusion coefficient (DO–Pd) of 10�16 cm2 s�1

at 327 1C and activation energy of 60–85 kJ mol�1.20 The

diffusion of oxygen through the bulk oxide layer again follows

the Mott–Cabrera parabolic diffusion law with diffusion

coefficient (DO–PdO) of 10�18 cm2 s�1 at 327 1C and activation

energy of 111–116 kJ mol�1.20

In this work, the reduction or decomposition of PdO under

H2, CH4, lean mixtures (CH4/O2 or CH4/O2/H2O) or He and

the oxidation of Pd1 under O2 have been studied during

temperature programmed experiments by in situ Raman spectro-

scopy and in situ XRD analysis. The XRD technique, which is

more sensitive to disorder in the cationic sublattice compared

to anionic sublattice and allows characterising only the largest

particles, has been used in conjunction with Raman spectro-

scopy, which is primarily sensitive to oxygen-cation vibrations

(coherence length of 1 nm) and provides both short as well as

long range ordering to study the phase evolution of palladium

under different atmospheres. The catalyst was a 4 wt% Pd

supported on g-Al2O3. The 4 wt% Pd loading provided high

signal sensitivity and intensity during XRD and Raman

detection of Pd1 and PdO crystallites.

2. Experimental

2.1 Catalyst preparation

Alumina powder supplied by Sasols was used as precious

metal support. The support was calcined at 950 1C for 4 h

before palladium impregnation. The final alumina powder had

a specific surface area of 100 m2 g�1. The alumina powder

was sieved to retain particles of sizes between 140–200 nm. The

g-Al2O3 support was impregnated with palladium nitrate

dissolved in water. The weight of the salt precursor was

calculated to obtain a theoretical metal loading of 4 wt%.

For this purpose, the Pd salt solution (Pd(NO3)2, Alfa Aesar,

density of 1.427, 14.67 wt% Pd) was adjusted to fill only the

g-Al2O3 porosity, (0.49 cm3 g�1 equivalent to a mean alumina

pore diameter of 12.3 nm with the assumption of cylindrical

pores: d = 2V S�1). The solution containing the palladium

precursor was then added dropwise. The solid was dried under

air at 110 1C for 2 h. The resulting powder was calcined under

air flow at 600 1C for 10 h (ramp = 5 1C min�1).

2.2 Chemical and physical characterization

Specific surface areas were estimated from N2 adsorption

at �196 1C (BET method), using a Micromeritics Tristar

apparatus. The experimental Pd loading was determined by

atomic absorption method on a Varian AA110 apparatus.

Palladium dispersion was obtained by H2 chemisorption.

Analysis has been carried out on a Micromeritics AutoChemII

instrument. After reduction of the sample under 2% H2 in He

at 500 1C for 1 h and purge under Ar (at the same temperature

for 2 h), H2 pulses were injected after cooling at 70 1C at

regular intervals. Dispersion was calculated using eqn (1),

supposing the chemisorption of one hydrogen atom per

palladium atom:

Dð%Þ ¼ 2� P� VH2�mcata

106 � R� T � nPdð1Þ

with, P, atmospheric pressure (in Pa); V, volume of adsorbed

H2 (mL); R, 8.314 J K�1 mol�1; T, 295 K; mcata, sample weight

(g); nPd, mole of palladium as obtained by ICP.

Palladium particle size (A) was determined using eqn (2),

supposing hemispherical particles:

daverage ¼6CaPM � 109

rDNavð2Þ

with Ca, the concentration of surface metal atoms, equal to

1.27 � 1019 atoms m�2, PM Pd atomic mass, r, Pd volumetric

mass, equal to 12.02 � 106 g m�3,22 D metal dispersion and

Nav Avogadro number.

2.3 In situ temperature programmed characterization

2.3.1 H2 adsorption at low temperature. The H2 adsorption

properties of the supported Pd catalyst at low temperature

were evaluated on a Micromeritics AutoChemII instrument.

Prior to H2 adsorption experiments, 150 mg of the catalysts

were pre-treated under reduction treatment. The gas composi-

tion was 5% H2 in Ar. The reaction temperature range was

between 0 and 80 1C with a flow-rate of 50 mL min�1. The

amount of H2 consumption during the H2 adsorption is

measured by a thermal conductivity detector (TCD).

2.3.2 CH4-TPR. Temperature programmed reduction

under a methane containing atmosphere was studied using a

fixed bed tubular quartz microreactor (I.D. = 7 mm) at

atmospheric pressure, placed within an electrically heated

furnace. The catalytic bed consisted of 60 mg of catalytic

powder (74–105 mm), diluted by 180 mg of quartz. CH4-TPR

was carried out in the following conditions: 0.5% CH4 in He,

150 mL min�1 at STP (standard temperature and pressure),

GHSV (gas hourly space velocity) = 150 000 mL g�1 h�1; the

temperature was ramped from room temperature up to 500 1C

at 15 1C min�1. Reactants and product compositions at the

outlet of the reactor were monitored by a mass spectrometer

with quadrupole detector (Balzers QMS 422).

2.3.3 In situ Raman spectroscopy. In situ Raman measure-

ments were carried out on a Horiba Jobin Yvon HR800UV

LabRam spectrometer using green light laser (514 nm).

For measurements, the laser was focused on sample aggregate

(o2 mm). The in situ Raman analysis was carried out at room

temperature on the fresh sample in order to check the PdO

peak and to find the best compromise between acquisition time

and peak intensity. The acquisition time was 20 s. The two

main PdO peaks are located at 445 and 640 cm�1. Mamede

et al.21 have carried out in situ Raman analysis experiments on

Dow

nloa

ded

by U

nive

rsity

of

Cal

ifor

nia

- Sa

n D

iego

on

07 M

arch

201

3Pu

blis

hed

on 3

1 Ja

nuar

y 20

11 o

n ht

tp://

pubs

.rsc

.org

| do

i:10.

1039

/C0C

P013

31E

View Article Online

This journal is c the Owner Societies 2011 Phys. Chem. Chem. Phys., 2011, 13, 4607–4613 4609

1 wt% Pd/g-Al2O3 using a spectrometer equipped with a

Nd :YAG laser (excitation line at 532 nm). They found a slight

shift at lower values, 427 and 626 cm�1. They attributed the

Raman lines to the Raman active Eg and B1g vibration modes of

PdO. The other weak and broader bands were attributed to a

resonance effect induced by the use of a laser excitation near the

excitation state of PdO. Bell et al.15 found the main PdO peak at

651 cm�1 on 10 wt% Pd/ZrO2 with the 514.5 nm wavelength of

the Ar ion laser and Arai et al.22 observed evolution of the main

PdO peak between 626 cm�1 (40 1C) and 633 cm�1 (600 1C) on

sputtered PdOx thin film with the 632.8 nm wavelength of the

He–Ne laser. A relevant study, carried out by Otto et al.,23 has

demonstrated that Raman spectroscopy of palladium oxide on

g-alumina was useful for a quantitative and non-destructive

analysis. They studied a wide range of palladium content, between

0.05 and 20 wt%. They showed that for palladium loading lower

than 0.2 wt%, some palladium may interact with the support and

thus may not contribute to the signal related to crystalline PdO.

Moreover, McBride et al. have shown that the analysis depth for

Raman measurements at 514.5 nm was about 10 nm24 and that

for palladium loading lower than 2.5 wt%, the Raman signal

should be proportional to the concentration. These results were

confirmed by Otto et al.23 Indeed for palladium content higher

than 2 wt%, there were no linear relations between Raman signal

and Pd loading. They assumed the effect of the presence of large

particles and wide variations in the particle size. To conclude, the

Raman technique is able to give useful information about the

smallest particles present in the catalyst.

The experimental procedure is described as following: (i) the

sample was pre-treated under 1%O2/N2 at 600 1C for 1 h with a

total flow rate of 50 mL min�1. Raman spectra were recorded

after cooling to 30 1C under the same atmosphere; (ii) after

purge under N2, the catalyst sample was treated under different

gaseous atmospheres with a total flow rate of 50 mL min�1,

either under reducing conditions (2% H2/N2, 2% CH4/N2), or

in lean conditions (2% CH4/5% O2/N2 or 2% CH4/5%

O2/H2O/N2) or in pure He; (iii) heating step at 10 1C min�1

was stopped at different temperature values for 10 min in order

to record Raman spectra under N2. For each spectrum acquisi-

tion, CH4 was removed from the feed because its presence

modified the signal intensity.25 In certain experiments, the

heating was stopped after 30 min to evaluate the kinetics of

the PdO/Pd1 transformation. Virtually no further transforma-

tion was observed between 10 and 30 min. The state of the

catalyst is essentially a function of the temperature: after 10 min

and beyond, it depends little on the time at which the sample is

maintained at a given temperature.

2.3.4 In situ XRD analysis. Contrary to Raman spectro-

scopy, XRD analysis allows characterizing only the largest

particles. It is commonly admitted that diffractograms of

nanoparticles deposited on alumina supports can give useful

information only when particle sizes are greater than 4–5 nm.

Diffractograms obtained for in situ experiments were obtained

on a Bruker AXS D8 advance powder diffractometer using

Cu Ka radiation (lKa = 0.15186 nm). Patterns were recorded

for 2y values between 251 and 851 in 0.0301 steps, with step

duration of 0.2 s (i.e. 6 min 40 s for the whole diffractogram).

The 2y values used in order to identify PdO and Pd1 peaks are

35, 40, 54 and 591 and 39, 46 and 821, respectively. The cell

support for XRD measurements was composed of Kanthal

with the main diffraction peaks at 44.480, 64.779 and 82.2841

(relative intensity: 100/20/50) with the following dimensions:

depth o 1 mm, width = 10 mm, length = 20 mm. The

experimental procedure is described as follows: (i) the sample

was pre-treated under 1% O2/N2 at 600 1C for 1 h with a total

flow rate of 50 mL min�1. After cooling down to 30 1C under

the same atmosphere, the XRD diffractogram was recorded;

(ii) after purge under N2, the sample was treated in the same

conditions as for Raman studies: reducing atmosphere

(2% H2/N2, 2% CH4/N2), lean conditions (2% CH4/5% O2/N2

or 2% CH4/5% O2/H2O/N2) or pure He; (iii) the heating step

at 10 1C min�1 was stopped at different temperature values for

10 min in order to record the XRD pattern. In this experiment,

CH4 was kept in the feed because there was no modification of

the signal intensity. As for the Raman spectra, some XRD

spectra were recorded after 30 min at a given temperature: no

detectable transformation was recorded for times longer than

10 min at the same temperature.

3. Results and discussion

3.1 Characterization

The Pd loading is close to the theoretical value, 3.94 wt%, and the

specific surface area is around 105 m2 g�1. The 4 wt% Pd sample

shows a dispersion of 22% (i.e. mean particle size of 5 nm). The

crystallite size estimated on the basis of the width at half-height of

PdO peaks was about 6 nm, in good agreement with results of H2

chemisorption experiments. TEM pictures reveal that particle size

was smaller than 10 nm and the palladium particle size evaluated

from the TEM picture is given in Fig. 1. More than 50% of the

particles have a particle size diameter lower than 5 nm and 30%

lower than 4 nm. More information on the catalyst characteriza-

tion, including XRD patterns and TEM pictures, can be found in

our previous work.26 It can be expected that Raman spectroscopy

should only give useful information on particles smaller than

4 nm while XRD will be more sensitive to bigger particles of

size 44–5 nm.

Fig. 1 Palladium particle size distribution from TEM picture.26

Dow

nloa

ded

by U

nive

rsity

of

Cal

ifor

nia

- Sa

n D

iego

on

07 M

arch

201

3Pu

blis

hed

on 3

1 Ja

nuar

y 20

11 o

n ht

tp://

pubs

.rsc

.org

| do

i:10.

1039

/C0C

P013

31E

View Article Online

4610 Phys. Chem. Chem. Phys., 2011, 13, 4607–4613 This journal is c the Owner Societies 2011

3.2 Reduction under H2/N2 followed by oxidation under

O2/N2

PdO reduction by H2/N2 and Pd1 oxidation were followed by

in situ Raman spectroscopy. The overall Raman spectrum is

given in Fig. 2A and the zoom of the main PdO band at

640 cm�1 is presented in Fig. 2B. For the sake of clarity, in the

next parts, only the zoom of the PdO band will be presented.

The PdO band located at 640 cm�1 is in the same frequency

range (650 cm�1) as the band observed by Bell et al. obtained

with the 514.5 nm wavelength of the Ar.18 At room tempera-

ture, both phenomenon, PdO reduction by H2 and Pd1 oxida-

tion by O2, occur easily. Raman spectroscopy shows that small

particles are reduced at ambient temperature but this technique

cannot give clear information on the biggest particles. In order

to clarify the behaviour of PdO particles bigger than 5 nm,

in situ XRD analysis was carried out. The overall diffractogram

obtained during reduction under H2/N2 is presented in Fig. 3.

As mentioned for Raman spectra, in the next parts, only the 2yrange of the PdO peak will be presented. The results show

that it is not possible to reduce large PdO particles under H2 at

room temperature. Higher temperatures are required to clearly

observe this reduction, which starts between 50 and 100 1C and

it is completed between 100 and 150 1C. Indeed the peak at

33.91 disappears completely. In the same time, Pd1 crystalli-

sation seems to be achieved at 150 1C. However, H2-TPR

measurements carried out on Pd catalysts show that PdO

reduction occurs at low temperature (o100 1C).27,29,30 The

shift in the reduction temperature between XRD and H2-TPR

measurements could be due to the difference between the

phenomenon detected. In the case of H2, the H2 consump-

tion reveals the reduction of PdO oxide whereas XRD

analysis reveals PdO phase transformation into Pd1 metallic

phase.

Reduction under H2 at ambient temperature was further

complicated by the potential formation of palladium hydride.

Some works have been carried out in order to study the

kinetics and mechanism of the dissolution of H atoms in

palladium supported catalysts28–30 or palladium single crystal.31–33

We carried out H2 adsorption experiment on pre-reduced

sample (H2 + PdO - Pd1 + H2O) in order to evaluate the

effect of the low ambient temperature on the reduction process.

Fig. 4 shows the evolution of H2 signal during two heating

[0–80 1C] and cooling [80–0 1C] cycles under H2. It can be

observed that H release occurs during heating at 68 1C and H

uptake occurs during cooling at 18 1C. Palladium crystallites

absorb H to form PdH (a or b) phase (yH2 + 2Pdb 2 2PdbHy).

The a phase contains less H atoms than the b phase. Both

phases of Pd metal have the fcc structure but with different

lattice parameters (0.39 and 0.40 nm, respectively) whereas in

the PdO structure, Pd atoms are in planar coordination with

four O atoms and the O atoms are in tetrahedral coordination

with four Pd atoms. The dominant phase of PdH depends on

H pressure and temperature.

After reduction treatment, Pd1 oxidation was carried out

under O2/N2 and followed by in situ XRD. PdO and Pd1 peak

evolutions located in the 30–421 range are shown in Fig. 5.

The intensity of the Pd1 reflection rapidly decreases above

300 1C, completely disappearing at 400 1C. On the other

hand, only a very broad PdO peak appears at 350 1C which

only gradually sharpens up to 550 1C, suggesting that reoxida-

tion occurs in the 300–400 1C temperature range, forming

highly disordered PdO, which only crystallizes at higher

temperatures.

3.3 Reduction under CH4/N2 followed by oxidation under

O2/N2

Small PdO particle reduction by CH4/N2 was followed by

in situ Raman spectroscopy. The zoom of the PdO band is

presented in Fig. 6A. Under CH4/N2, the reduction of small

PdO particle into Pd1 starts at 240 1C and is completed at

250 1C (i.e. relative Raman signal being equal to 0). This

reduction was followed by CH4-TPR as described in section

2.3.2. Methane consumption is given in Fig. 6B. The tempera-

ture of the reduction process is similar to those obtained by

Raman. Indeed, the reduction starts at 250 1C and is complete

at 310 1C. This temperature range difference could be attributed

to the size effect distribution. In the case of Raman, only small

palladium particles are analyzed, whereas in CH4-TPR,

methane consumption is due the overall palladium content

(i.e. every particle). In order to follow the large PdO particle

reduction, in situ XRD analysis was carried out. The results

presented in Fig. 7 shows that reduction is a fast process

occurring in a narrow temperature range between 280 and

290 1C with simultaneous Pd1 reconstruction. After the reduc-

tion treatment, Pd1 oxidation was carried out under O2/N2

and followed by in situ Raman and XRD analysis. Small

particles are not reoxidized even after oxidation treatment at

300 1C (Fig. 8A). It can be assumed that after reduction under

CH4 conditions, the surface is covered by C or unburned CH4

located in part on the metal. Pd1 particles could only be

reoxidized when these species are removed from the metal

surface. Moreover, Liu et al. showed, thanks to a density

functional theory study, that the C-forming reaction from

CH4 is structure sensitive and is easier on kink and step sites,

which are more abundant on small particles.34 However, large

Pd1 particle reoxidation follows the same process as after

reduction under H2/N2, i.e. slow gradual phenomenon starting

only at 300 1C and finishing at a temperature higher than

550 1C. XRD PdO and Pd1 peak evolution located in the

30–421 range is shown in Fig. 8B. The oxidation process after

CH4/N2 reduction may include three steps: (i) oxidation ofFig. 2 Overall Raman spectra (A) and zoom of the PdO band

(B) after reduction under 2% H2/He.

Dow

nloa

ded

by U

nive

rsity

of

Cal

ifor

nia

- Sa

n D

iego

on

07 M

arch

201

3Pu

blis

hed

on 3

1 Ja

nuar

y 20

11 o

n ht

tp://

pubs

.rsc

.org

| do

i:10.

1039

/C0C

P013

31E

View Article Online

This journal is c the Owner Societies 2011 Phys. Chem. Chem. Phys., 2011, 13, 4607–4613 4611

residual C at the surface; (ii) O� and O2� migration in the Pd1

bulk; (iii) reconstruction of PdO crystals.

3.4 Thermal decomposition under He followed by oxidation

under O2/N2

PdO decomposition under He and Pd1 oxidation by O2/N2 was

only followed by in situ XRD analysis. Indeed, Raman results

are not shown here due to the experimental limitation at

700 1C. However, at this temperature, surface PdO was still

present. The zoom on the PdO peak during He decomposition

and oxidation treatment are presented in Fig. 9A and Fig. 9B,

Fig. 3 Overall XRD diffractogram after reduction under 2% H2/He.

Fig. 4 H2-TPR at low temperature [0–80 1C].

Fig. 5 XRD diffractograms focused on PdO and Pd1 peaks after

oxidation under 1% O2/N2.

Fig. 6 Raman spectra focused on PdO band (A) after reduction

under 2% CH4/N2 and CH4-TPR; (B) under 0.5% CH4/He.

Fig. 7 XRD diffractogram focused on PdO and Pd1 peaks after

reduction under 2% CH4/N2.

Fig. 8 Raman spectra focused on PdO band after reduction under

2% CH4/N2 followed by oxidation under 1% O2/N2 (A) and XRD

diffractogram focused on PdO and Pd1 peaks after oxidation under

1% O2/N2 (B).

Dow

nloa

ded

by U

nive

rsity

of

Cal

ifor

nia

- Sa

n D

iego

on

07 M

arch

201

3Pu

blis

hed

on 3

1 Ja

nuar

y 20

11 o

n ht

tp://

pubs

.rsc

.org

| do

i:10.

1039

/C0C

P013

31E

View Article Online

4612 Phys. Chem. Chem. Phys., 2011, 13, 4607–4613 This journal is c the Owner Societies 2011

respectively. Decomposition of large PdO particles starts at

750 1C and is a gradual phenomenon. Total decomposition of

PdO is completed at 800 1C. This decomposition process can

be due to (i) release of O from the surface toward the gas

atmosphere and (ii) O� or O2� anion migration from bulk to

the surface. This second step allows maintaining constant

oxidation of the surface and would explain the Raman results.

Large Pd1 particle reoxidation under O2/N2 follows the same

slow process as after reduction under H2/N2 or CH4/N2. The

results of Pd1 oxidation under O2/N2 followed by in situ

Raman corroborate our previous results on 1% Pd/y-Al2O3

catalyst during temperature programmed desorption of O2.27

It was found that PdO decomposition into metallic palladium

started at 790 1C and was completed at 920 1C, whereas Pd1

was found to reoxidize during cooling at a temperature lower

than 550 1C (i.e. beginning of oxygen consumption) and is

completed at 400 1C.

3.5 Lean conditions

As already mentioned, palladium is used in many oxidation

processes (gas turbine, NGV catalytic converter). In order to

obtain information on the surface and bulk behaviour of the

4 wt% Pd/g-Al2O3 catalyst during CH4 oxidation, with or

without water, in situ Raman spectroscopy and in situ XRD

analysis were used.

3.5.1 Lean condition without water: CH4/O2/N2. The

evolution of the PdO Raman band and PdO diffraction peak

during temperature combustion test under lean conditions

without water (2% CH4/5% O2/N2) are shown in Fig. 10A

and Fig. 10B, respectively. In dry reaction medium, between

200 and 300 1C, the Raman PdO peak area decreases but its

surface seems to be stable beyond 350 1C. It can be assumed

that the PdO particles are modified during CH4 combustion.

However, it is difficult to conclude if the PdO surface is

composed of a mixture of PdO and Pd1 phases or of an under-

stoichiometric PdOx phase (x o 1). Moreover, until 600 1C,

there is no reduction of large PdO particles (Fig. 10B) and thus

CH4 does not modify bulk PdO.

3.5.2 Lean condition in presence of water: CH4/O2/H2O/N2.

In our previous study,26 the methane conversion on 1%, 2%

and 4 wt% Pd/g-Al2O3 was followed during temperature

programmed combustion in the presence of water (2% CH4/5%

O2/2% H2O/N2) and transient experiment (successive Pd

oxidation/reduction cycles). For the 2% and 4% catalysts

with similar Pd particle sizes, 5 nm, determined by H2

chemisorption, Pd catalyst reactivation and PdO reformation

both occurred gradually with comparable time scales. These

results suggested a role of a mixed PdO/Pd1 phase in temporarily

enhancing CH4 combustion activity, but no definite conclu-

sions were given on the issue of whether complete or partial

re-oxidation of Pd was required to achieve maximum CH4

combustion activity. In this complementary work (Fig. 11A),

in situ Raman spectroscopy reveals that small PdO particles

are partially reduced during lean conditions in the presence of

water (2% CH4/5% O2/2% H2O/N2) whereas large PdO

particles are not modified (Fig. 11B). The state of palladium

in reaction is intermediary between PdO and Pd1 but, like for

dry condition, it is not possible to conclude on the nature of

the active site: either a PdO thin film covering a metallic

palladium core PdO/Pd1 phase or a palladium suboxide

(PdOx, x o 1). Specchia et al. have studied the surface

chemistry and surface reactivity of palladium particles over

ceria-zirconia support.35 The authors assumed that the improved

catalytic activity of the fresh catalyst for CH4 combustion at

low temperature is due to its constitution, partly oxidized very

small Pd metal particles and dispersed Pd oxide species. The

progressive oxidation of highly dispersed small Pd particles to

PdOx as well as the coalescence of dispersed Pd oxide species

result in fully oxidized PdOx particles, at least at the surface,

which are less active (in the low temperature range) with

respect to the active species of the fresh catalyst. Moreover,

it is also possible to take into consideration reduction

assistance between particles of different sizes. For instance,

Martin and Duprez36 showed that big particles of rhodium on

alumina reduced before smaller ones. However, it should be

necessary to know the degree of intimacy of the biggest and

the smallest particles and the mean distance between them. At

this stage, it is difficult to conclude definitively.

4. Conclusion

The reactivity study of the PdO and Pd1 particles of the 4 wt%

Pd/Al2O3 catalysts under different atmospheres, characterized

by in situ XRD and Raman analysis, highlighted the following

behaviours:

(1) The reduction behaviour depends on the reducing agent:

H2 can reduce PdO particles at much lower temperatures

(100–150 1C) than methane (280–290 1C) while PdO decomposi-

tion into Pd1 in neutral gas (He) occurs only above 700 1C.

(2) Whatever the reduction or decomposition treatment,

re-oxidation of large Pd1 particles is a gradual phenomenon,

starting at 300 1C and being completed at temperatures higher

than 550 1C. The oxidation process may occur in three steps:

(i) oxidation of the C species of the surface (if CH4 is the

reducer) and of the surface Pd1; (ii) O migration in the Pd1

bulk (DO–Pd1) and (iii) PdO reconstruction (DO–PdO).

(3) Finally, during experiments in lean conditions, with or

without water, the PdO surface is partially reduced whereas

bulk PdO is not modified.

Fig. 9 XRD diffractogram focused on PdO and Pd1 peaks during

thermal decomposition under He (A) and during oxidation under

1% O2/N2 (B).

Dow

nloa

ded

by U

nive

rsity

of

Cal

ifor

nia

- Sa

n D

iego

on

07 M

arch

201

3Pu

blis

hed

on 3

1 Ja

nuar

y 20

11 o

n ht

tp://

pubs

.rsc

.org

| do

i:10.

1039

/C0C

P013

31E

View Article Online

This journal is c the Owner Societies 2011 Phys. Chem. Chem. Phys., 2011, 13, 4607–4613 4613

Acknowledgements

The French ‘‘Agence De l’Environnement et de la Maıtrise de

l’Energie’’ (ADEME) is acknowledged for financially supporting

this work.

References

1 R. J. Farrauto, M. C. Hobson, T. Kennelly and E. M. Waterman,Appl. Catal., A, 1992, 81, 227–237.

2 J. Chen and E. Ruckenstein, J. Catal., 1981, 69, 254–273.3 D. Ciuparu, M. R. Lyubovsky, E. Altamn, L. Pfefferle andA. Datye, Catal. Rev. Sci. Eng., 2002, 44, 593–647.

4 P. Gelin and M. Primet, Appl. Catal., B, 2002, 39, 1–37.5 T. V. Choudhary, S. Banerjee and V. R. Choudhary, Appl. Catal.,A, 2002, 234, 1–23.

6 J. G. McCarty, Catal. Today, 1995, 26, 283–283.7 G. Ketteler, D. F. Ogletree, H. Bluhm, H. Liu, E. L. D. Hebenstreitand M. Salmeron, J. Am. Chem. Soc., 2005, 127, 18269–18273.

8 E. H. Voogt, A. J. M. Mens, O. L. J. Gijzeman and J. W. Geus,Surf. Sci., 1997, 373, 210–220.

9 D. Zemlyanov, B. Aszalos-Kiss, E. Kleimenov, D. Teschner,S. Zafeiratos, M. Haevecker, A. Knop-Gericke, R. Schlogl,H. Gabasch, W. Unterberger, K. Hayek and B. Klotzer, Surf.Sci., 2006, 600, 983–994.

10 H. Gabasch, W. Unterberger, K. Hayek, B. Klotzer,E. Kleimenov, D. Teschner, S. Zafeiratos, M. Haevecker,A. Knop-Gericke, R. Schlogl, J. Han, F. H. Ribeiro, B. Aszalos-Kiss,T. Curtin and D. Zemlyanov, Surf. Sci., 2006, 600, 2980–2989.

11 J. Han, D. Y. Zemlyanov and F. H. Ribeiro, Surf. Sci., 2006, 600,2730–2744.

12 G. Zhu, J. Han, D. Y. Zemlyanov and F. H. Ribeiro, J. Am. Chem.Soc., 2004, 126, 9896–9897.

13 R. J. Farrauto, J. K. Lampert, M. C. Hobson andE. M. Waterman, Appl. Catal., B, 1995, 6, 263–270.

14 Y. S. Ho, C. B. Wang and C. T. Yeh, J. Mol. Catal. A: Chem.,1996, 112, 287–294.

15 S. Su, J. N. Cartens and A. Bell, J. Catal., 1998, 176, 125–135.16 D. Roth, P. Gelin, A. Kaddouri, E. Garbowski, M. Primet and

E. Tena, Catal. Today, 2006, 122, 134–138.17 K. Fujimoto, F. H. Ribeiro, M. Avalos-Borja and E. Iglesia,

J. Catal., 1998, 179, 431–442.18 A. K. Datye, J. Bravo, T. R. Nelson, P. Atanasova, M. Lyubovsky

and L. Pfefferle, Appl. Catal., A, 2000, 198, 179–196.19 A. S. Khanna, Introduction to High Temperature Oxidation and

Corrosion, ASM International, 2002.20 J. Han, D. Y. Zemlyanov and F. H. Ribeiro, Surf. Sci., 2006, 600,

2752–2761.21 A. S. Mamede, G. Leclerq, E. Payen, P. Granger and J. Grimblot,

J. Mol. Struct., 2003, 651–653, 353–364.22 T. Arai, T. Shima, T. Nakano and J. Tominaga, Thin Solid Films,

2007, 515, 4774–4777.23 K. Otto, C. P. Hubbard, W. H. Weber and G. W. Graham, Appl.

Catal., B, 1992, 1, 317–327.24 J. R. McBride, K. C. Hass and W. H. Weber, Phys. Rev.

B: Condens. Matter, 1991, 44, 5016–5028.25 M. Boulova, A. Gaskov and G. Lucazeau, Sens. Actuators, B,

2001, 81, 99–106.26 P. Castellazzi, G. Groppi, P. Forzatti, A. Baylet, P. Marecot and

D. Duprez, Catal. Today, 2010, 155, 18–26.27 A. Baylet, S. Royer, C. Labrugere, H. Valencia, P. Marecot,

J. M. Tatibouet and D. Duprez, Phys. Chem. Chem. Phys., 2008,10, 5983–5992.

28 D. Wang, J. D. Clewley, T. B. Flanagan, R. Balasubramaniam andK. L. Shanahan, Acta Mater., 2002, 50, 259–275.

29 C. Neyertz, M. A. Volpe and C. Gigola, Catal. Today, 2000, 57,255–260.

30 C. Amorim and M. A. Keane, J. Colloid Interface Sci., 2008, 322,196–208.

31 F. Leardini, J. F. Fernadez, J. Bodega and C. Sanchez, J. Phys.Chem. Solids, 2008, 69, 116–127.

32 C. W. Chou, T. P. Perng and C. T. Yeh, J. Phys. Chem. B, 2001,105, 9113–9117.

33 T. Kuji, Y. Matsumura, H. Uchida and T. Aizawa, J. AlloysCompd., 2002, 330–332, 718–722.

34 Z. P. Liu and P. Hu, J. Am. Chem. Soc., 2003, 125, 1958–1967.35 S. Specchia, E. Finocchio, G. Busca, P. Palmisano and

V. Specchia, J. Catal., 2009, 263, 134–145.36 D. Martin and D. Duprez, Appl. Catal., A, 1995, 131,

297–307.

Fig. 10 Raman spectra (A) and XRD diffractogram (B) focused on

PdO and Pd1 peaks during lean condition reaction under 2% CH4/5%

O2/N2.

Fig. 11 Raman spectra (A) and XRD diffractogram (B) focused on

PdO and Pd1 peaks during lean condition reaction under 2% CH4/5%

O2/2% H2O/N2.

Dow

nloa

ded

by U

nive

rsity

of

Cal

ifor

nia

- Sa

n D

iego

on

07 M

arch

201

3Pu

blis

hed

on 3

1 Ja

nuar

y 20

11 o

n ht

tp://

pubs

.rsc

.org

| do

i:10.

1039

/C0C

P013

31E

View Article Online