Temperature programmed reduction studies of spillover effect in Pd impregnated metal oxide catalysts

Article abstract of DOI:10.1007/s10973-005-0438-0

Temperature programmed hydrogen reduction studies have been carried out for SnO₂ and Ce-Sn mixed oxides with and without Pd metal impregnation, to demonstrate the existence of spillover of hydrogen from Pd metal centers to support oxides. TPR pattern of SnO₂ showed a main peak at ～973 K indicating the bulk reduction of this sample. In Pd metal impregnated sample, the bulk reduction peak shifts to lower temperature (～923 K) due to the spillover of activated hydrogen from Pd metal to SnO₂ at relatively lower temperatures and its subsequent reaction with SnO₂. For Pd impregnated Ce-Sn mixed oxide samples also, a similar effect or an enhanced reduction was observed indicating the spillover effect of hydrogen. These results have been further confirmed from ¹¹⁹Sn Moessbauer spectroscopic measurements carried out for some representative samples of SnO₂ and Pd/SnO₂ heated in hydrogen flow up to a temperature of 473 K. The value of Sn ²⁺/(Sn⁴⁺+Sn²⁺) ratio was found to be significantly higher for Pd impregnated sample. Both these observations provide direct evidence for the existence of spillover effect of hydrogen taking place in the metal impregnated samples.

Full text of DOI:10.1007/s10973-005-0438-0

Journal of Thermal Analysis and Calorimetry, Vol. 78 (2004) 723–729
TEMPERATURE PROGRAMMED REDUCTION
STUDIES OF SPILLOVER EFFECT IN Pd
IMPREGNATED METAL OXIDE CATALYSTS
*
R. Sasikala and S. K. Kulshreshtha
NM and SC Division, Bhabha Atomic Research Centre, Trombay, Mumbai 400 085, India
Abstract
Temperature programmed hydrogen reduction studies have been carried out for SnO
oxides with and without Pd metal impregnation, to demonstrate the existence of spillover of hydrogen
from Pd metal centers to support oxides. TPR pattern of SnO showed a main peak at ~973 K indicating
the bulk reduction of this sample. In Pd metal impregnated sample, the bulk reduction peak shifts to
lower temperature (~923 K) due to the spillover of activated hydrogen from Pd metal to SnO at rela-
tively lower temperatures and its subsequent reaction with SnO . For Pd impregnated Ce–Sn mixed ox-
ide samples also, a similar effect or an enhanced reduction was observed indicating the spillover effect of
2
and Ce–Sn mixed
2
2
2
119
hydrogen. These results have been further confirmed from Sn Mössbauer spectroscopic measurements
carried out for some representative samples of SnO and Pd/SnO heated in hydrogen flow up to a tem-
2
2
2+
4+
2+
perature of 473 K. The value of Sn /(Sn +Sn ) ratio was found to be significantly higher for Pd im-
pregnated sample. Both these observations provide direct evidence for the existence of spillover effect of
hydrogen taking place in the metal impregnated samples.
Keywords: mixed oxide, Mössbauer spectra, Pd impregnated oxides, temperature programmed re-
duction, tin oxide
Introduction
Metal oxide or mixed metal oxide catalysts are widely employed for oxidation/selec-
tive oxidation of hydrocarbons and also in automobile threeway catalysis [1–4]. For
these reactions, low temperature reducibility of oxides is desired and it can often be
achieved by the metal impregnation of these oxides. In order to have an idea about the
catalytic oxidation activity, the reduction behaviour of these oxides has to be studied
and different thermal techniques have been employed for this purpose [5, 6]. Tempera-
ture programmed reduction (TPR) technique is widely employed to study the reduction
behaviour of oxides. The TPR profiles thus obtained can give an idea about the differ-
ent types of reducible species and the catalytic oxidation activity of these oxides [7, 8].
Pd/SnO is a well-studied oxidation catalyst and a synergistic effect has been re-
2
ported in this system for CO oxidation reaction [9]. The existence of synergistic ef-
*
Author for correspondence: E-mail: kulshres@magnum.barc.ernet.in
1
388–6150/2004/ $ 20.00
Akadémiai Kiadó, Budapest
©
2004 Akadémiai Kiadó, Budapest
Kluwer Academic Publishers, Dordrecht

7
24
SASIKALA, KULSHRESHTHA: Pd IMPREGNATED METAL OXIDE CATALYSTS
fect has been explained on the basis of a spillover effect [10], which is the migration
of reactant species from the metal sites to support after getting activated at metal
sites. The interaction of the activated CO with SnO leads to its reduction at relatively
2
lower temperature. TPR profiles of such samples should exhibit a shifting of reduc-
tion peak to low temperature as compared to that of oxides without metal because of
the activation of reactant species on metal sites at relatively lower temperatures. In
the case of partially reducible oxides as support, an enhanced reduction should take
place due to spillover effect. With this aim, the TPR studies have been carried out for
SnO and Ce–Sn mixed oxides with and without metal impregnation using hydrogen.
2
1
19
In order to get supplementary evidence for spillover effect, the Sn Mössbauer
spectra of some samples were also recorded after reduction in hydrogen and the ex-
2
tent of reduction was evaluated by calculating the fraction of Sn in these samples.
+
Experimental
SnO was prepared by the precipitation of tin chloride solution using ammonia. Sn–Ce
2
mixed oxides ((1–x)SnO +xCeO , where x=0.3 and 0.5 referred to as C30S and C50S,
2
2
respectively) were prepared by coprecipitation method from the chloride and nitrate
solutions of Sn and Ce respectively using NH OH at pH ~9. The oven-dried precipi-
4
tates after washing thoroughly were calcined at 773 K in air for 2 h. Pd impregnation
on these oxides was done by wet impregnation method using PdCl solution (Pd=2%
by mass of sample). After impregnation, the sample was heated in hydrogen flow at
2
473 K for 2 h for reducing PdCl to Pd metal followed by heating in air at 773 K for 2 h.
These samples were characterized by powder XRD technique using Philips 1820
2
diffractometer with CuK radiation. TPR experiments were carried out from room tem-
a
perature to 1223 K using an indigenously fabricated instrument. 25 mg sample was
–1
placed in a quartz tube and heated at 15°C min in H +Ar stream (8% H by volume).
2
2
All samples were given an in situ pretreatment of heating in Ar flow at 573 K for 2 h
followed by cooling to room temperature before a TPR run. Change in H concentra-
tion in the effluent gas due to the reduction of samples, was monitored by a thermal
2
119
conductivity detector. Sn Mössbauer spectra of pure and metal impregnated samples
were recorded at room temperature after heating the samples under different condi-
119
tions. The source used was Ba SnO and all values of isomeric shift are reported with
3
respect to this source. The spectra were least square fitted using Lorenzian line shape to
2+
4+
estimate the relative fractions of Sn and Sn species in the catalyst samples.
Results and discussion
Powder X-ray diffraction patterns of SnO , {(1–x)SnO +xCeO } with and without Pd
2
2
2
metal impregnation and heated at 773 K in air for 2 h are shown in Fig. 1. SnO showed
2
broad Bragg reflections with average ‘d’ values corresponding to the tetragonal (rutile)
phase. CeO showed better crystallinity and the unit cell parameter ‘a’ was found to be
2
0.5402 nm. Mixed oxides existed in two phases as that of SnO and CeO . The
crystallinity of mixed oxides improved with CeO concentration as can be seen from
2 2
2
J. Therm. Anal. Cal., 78, 2004

SASIKALA, KULSHRESHTHA: Pd IMPREGNATED METAL OXIDE CATALYSTS
725
2 2
Fig. 1 XRD patterns of SnO , Ce–Sn mixed oxides and CeO with and without Pd
heated in air at 773 K for 2 h
their XRD patterns. These patterns were not affected much after Pd metal impregnation
and heat treatment. The peaks due to Pd metal were not seen in any of these patterns, as
its concentration was very small. All these patterns were indexed and the values of lat-
tice parameters were calculated for those samples, which showed reasonable
crystallinity and were found to be close to the values reported for individual SnO
and
2
CeO phases (ASTM card no. 41-1445 for SnO and 34-394 for CeO ). For C50S and
2
2
2
Pd/C50S the values for SnO phase were ‘a’=0.4770 and 0.4737 and ‘c’=0.3168 and
2
0.3185 nm, respectively. The CeO phase present in C50S and Pd/C50S had the values
of unit cell parameter (‘a’) 0.5378 and 0.5399 nm, respectively.
2
TPR profiles of pure SnO and Pd impregnated SnO heated in air at 773 K for
2
2
2
h are shown in Fig. 2. For pure SnO an intense peak centered at 973 K is present
2
besides some small peaks at ~548 and 673 K. The low intense peaks are assigned to
the reduction of surface oxide species and the intense peak present at 973 K to the
bulk reduction of SnO to Sn metal.
2
The low temperature peaks are assigned to surface reduction, as these peaks are
not seen for sintered SnO sample, which has been heated in air at 1073 K for 18 h [11].
2
The total volume of hydrogen consumed during TPR corresponds to the complete re-
duction of SnO to Sn metal. The presence of Sn metal in the reduced samples was also
2
established from the XRD patterns of the samples recorded after TPR. The TPR profile
of Pd/SnO showed an intense peak at 923 K corresponding to the bulk reduction be-
2
sides small peaks due to surface reduction. It can be noted that the bulk reduction peak
of SnO has shifted to lower temperature in Pd/SnO indicating that Pd is playing a role
2
2
J. Therm. Anal. Cal., 78, 2004

7
26
SASIKALA, KULSHRESHTHA: Pd IMPREGNATED METAL OXIDE CATALYSTS
2 2
Fig. 2 TPR profiles of SnO and Pd/SnO
Fig. 3 a – TPR profiles C30S and C50S, b – Pd/C30S and Pd/C50S
in the reduction process. The TPR profiles of Ce–Sn mixed oxides heated in air at
773 K for 2 h and Pd impregnated mixed oxides are shown in Fig. 3.
C30S showed two main peaks centered around 543 and 881 K corresponding to
the surface and bulk reduction of SnO . Contribution to the reduction peaks from
2
CeO is very little as the reducibility of CeO is very poor as compared to SnO ,
2
2
2
which has been demonstrated in our earlier work [11]. The enhanced reduction seen
in mixed oxides as compared to pure SnO at the low temperature region
2
(
around 543 K) is attributed to a synergistic effect existing in these mixed ox-
J. Therm. Anal. Cal., 78, 2004

SASIKALA, KULSHRESHTHA: Pd IMPREGNATED METAL OXIDE CATALYSTS
727
ides [11]. The shoulder peak seen around 973 K corresponds to the reduction of CeO2
whose reduction is incomplete in the temperature region studied [8, 12]. Similarly,
two main peaks are seen for C50S also around 563 and 855 K. The peak seen around
9
93 K has become more pronounced due to the increased concentration of CeO in
2
this composition. The TPR profiles of Pd impregnated samples clearly show the
shifting of reduction peaks to lower temperatures, as is clear form Fig. 3. The low
temperature peaks seen up to 673 K for Pd/C30S and Pd/C50S can be attributed to
the reduction of surface oxygen species. The bulk reduction peak of both these mixed
oxides also shift to lower temperature, though the effect is less for Pd/C50S. It can
also be seen that for Pd/C50S, the width of the bulk reduction peak is more compared
to C50S, indicating that the bulk reduction of CeO is taking place at lower tempera-
2
ture due to the presence of Pd metal in these samples.
The decreased shift seen in the high temperature reduction peak of Pd/C50S as
compared to C50S can be due to a metal support interaction occurring in this sample.
It has been reported that [13] during high temperature reduction of Pd/CeO , CeO2
2
reduces to form CeO and these suboxides can encapsulate Pd metal which retards
x
further reduction process of CeO . As the concentration of ceria is more in Pd/C50S,
2
the effect becomes more pronounced in it and hence a decreased shift is observed in
this sample as compared to that observed for Pd/C30S and C30S.
The XRD patterns of Pd/SnO and Pd/CeO after TPR are shown in Fig. 4. It can
2
2
be seen that SnO has been completely reduced to Sn metal. The pattern has been in-
2
dexed and the lattice parameters have been calculated (‘a’=0.5582 and ‘c’=0.3182 nm)
which are very close to the value reported for Sn metal (ASTM card no. 4-673). In case
of Pd/CeO , the reduction is partial and the reduced sample retains its CeO structure. It
2
2
2 2
Fig. 4 a – XRD patterns of Pd/SnO and b – PdCeO after TPR
J. Therm. Anal. Cal., 78, 2004

7
28
SASIKALA, KULSHRESHTHA: Pd IMPREGNATED METAL OXIDE CATALYSTS
is well known that ceria can retain its fluorite structure with a non-stoichiometric com-
position, CeO2–x, where 0£ x£0.5 [14].
The shifting of the bulk reduction peak to low temperature cannot be due to sur-
face effects arising from the difference in the crystallite size of oxides and Pd impreg-
nated oxides as is evident from the width of the XRD peaks of these samples shown
in Fig. 1. It can be seen from this figure that the line widths are more for oxide sam-
ples, which do not contain Pd metal indicating smaller crystallite size for them. Even
then, low temperature reduction is seen for metal impregnated samples, ruling out the
role of crystallite size effect during reduction.
The role of spilt over hydrogen from metal centers to oxide support has been
emphasized during hydrogenation reaction by a number of investigators [15] and
methods for detecting spilled hydrogen has been reviewed by Sermon et al. [16]. The
spilled hydrogen may reside on the surface of the support or diffuse into the bulk of
the support and a variety of interactions with the support can take place like remain-
ing adsorbed on the support, the partial reduction of the support or reaction with the
support forming a new compound like hydrogen bronze. When the spilt over hydro-
gen reacts with the support leading to its reduction, the reduction of the support
should take place at lower temperatures or at a faster rate in the vicinity of metal cen-
ters due to the migration of activated hydrogen after dissociation at the metal centers.
The shifting of reduction peaks to lower temperatures is observed in the TPR profiles
of Pd metal impregnated samples indicating a spillover effect of hydrogen during re-
duction. Hence, temperature programmed reduction technique can be used to study
the spillover phenomenon in metal impregnated oxides.
1
19
Figure 5 shows Sn Mössbauer spectra of SnO and Pd/SnO after heating in
2 2
hydrogen at 473 K for 2 h. Spectrum of untreated SnO is also included for compari-
2
son. The spectrum of SnO shows an unresolved quadrupole doublet characterized by
2
–
isomeric shift d»0.0 mm s and quadrupole splitting DEq»0.75 mm s . The spectra
1
–1
2
of the reduced samples show peaks corresponding to Sn whose isomeric shift is
+
1
19
2 2
Fig. 5 Sn Mössbauer spectra of SnO and Pd/SnO after heat by hydrogen for 2 h
J. Therm. Anal. Cal., 78, 2004

SASIKALA, KULSHRESHTHA: Pd IMPREGNATED METAL OXIDE CATALYSTS
729
–
d»3.0 mm s and DEq»1.9 mm s . The fraction of Sn /(Sn +Sn ) calculated for
1
–1
2+
4+
2+
SnO and Pd/SnO are 0.13 and 0.23 respectively. From this result it is clear that un-
2
2
2
der identical conditions of hydrogen treatment the fraction of Sn species formed for
+
metal impregnated sample is more than that of pure SnO . This provides a direct evi-
2
dence for the spillover of activated hydrogen from metal sites to SnO . This observa-
2
tion is consistent with the TPR results of these samples shown in Fig. 2. The low tem-
perature portion of the TPR patterns of SnO and Pd/SnO clearly shows that the on-
set of reduction takes place at lower temperature in the case of Pd/SnO as compared
2
2
2
to SnO . However, a comparison of the quantity of SnO , which has undergone re-
2
2
duction by both these techniques, cannot be made due to the fact that TPR is done in a
dynamic mode whereas the Mössbauer spectra were recorded after the reduction of
these samples for 2 h at 473 K.
Conclusions
TPR studies of tin oxide and mixed oxides of Ce and Sn with and without Pd metal
give a clear evidence for the existence of spillover of hydrogen in Pd impregnated
samples during reduction as the reduction peaks shift to lower temperatures in these
1
19
samples as compared to oxides not having Pd metal. Sn Mössbauer results further
2
confirm this effect from the increased concentration of Sn present in Pd/SnO as
+
2
compared to pure SnO2.
References
1
2
3
4
5
6
7
8
9
V. D. Sokolovski, Catal. Rev.- Sci. Eng., 32 (1990) 1.
M. C. K. Franca, R. A. da Silva San Gil and J. G. Eon, Catal. Today, 78 (2003) 105.
T. Feng and J. M. Vohs, J. Catal., 221 (2004) 619.
A. Trovarelli, Catal. Rev.- Sci. Eng., 38 (1996) 439.
M. Pospisil, V. Cuba and D. Polakova, J. Therm. Anal. Cal., 75 (2004) 35.
M. M. Pineda, S. Castillo, M. Asomoza and R. Gomez, J. Therm. Anal. Cal., 73 (2003) 341.
N. W. Hurst, S. J. Gentry and A. Jones, Catal. Rev.- Sci. Eng., 24 (1982) 233.
R. Sasikala, S. Varma, N. M. Gupta and S. K. Kulshreshtha, J. Mat. Sci. Lett., 20 (2001) 1131.
D. P. Schreyer, B. T. Upchurch, J. D. van Norman, K. G. Brown and J. Schreyer, J. Catal.,
1
22 (1990) 193.
th
0 C. Bond, M. J. Fuller and L. R. Molly, in Proc. 6 Intern. Cong. Catalysis, 1977 (Eds G. C. Bond,
1
P. B. Well and F. C. Tomkins), 1 (1977) 366.
1
1
1
1
1 R. Sasikala, N. M. Gupta and S. K. Kulshreshtha, Catal. Lett., 71 (2001) 69.
2 H. C. Yao and Y. F. Yao, J. Catal., 86 (1984) 254.
3 M. D. Mitchell and M. A. Vannice, Ind. Eng. Chem. Fundam., 23 (1984) 88.
4 V. Perrichon, A. Laachir, G. Bergeret, R. Frety and L. Tournayan, J. Chem. Soc. Faraday Trans.,
90 (1994) 773.
1
1
5 S. J. Teichner, Appl. Catal., 62 (1990) 1.
6 A. Sermon and G. C. Bond, Catal. Rev., 8 (1974) 212.
J. Therm. Anal. Cal., 78, 2004