Temperature-programmed reduction study of chromium

oxide supported on zirconia and lanthanazirconia

D.L. Hoang, H. Lieske
*
Institute of Applied Chemistry, Berlin Adlershof, Richard-Willstatter-Str. 12, D-12489 Berlin, Germany
Received 23 August 1999; received in revised form 12 October 1999; accepted 13 October 1999
Abstract
Temperature-programmed reduction (TPR) has been applied to study the reduction behavior of chromium oxide supported
on pure zirconia and on lanthana-containing zirconia. The results revealed inuences of support nature and chromium loading
on the state of the supported chromium oxide. With catalysts containing lanthana and only low amounts of chromium, the
decrease of the chromium oxidation state during TPR corresponded to the reduction of Cr
6
to Cr
3
. With lanthana-free
catalysts, lower changes of oxidation state have been found, hinting on thermal instability of higher valent chromium species
during thermal catalyst pretreatment. Chromium oxide supported on pure zirconia is reduced at temperatures between 340 and
3708C and is not quantitatively reoxidized by an oxygen treatment. Chromium oxide on lanthanazirconia is reduced at higher
temperatures between 370 and 4158C and is completely reoxidable at Cr loadings _4.0 wt.%. The results suggest that lanthana
strengthens the interaction between the zirconia surface and chromium species and, therefore, stabilizes a dispersed state of
chromium species. # 2000 Elsevier Science B.V. All rights reserved.
Keywords: Temperature-programmed reduction; Chromia; Chromiazirconia; Lanthana; Reoxidation
1. Introduction
Chromia supported on alumina, silica or titania is
known to catalyze a wide range of industrially impor-
tant reactions, e.g. oxidation reactions, polymeriza-
tions, hydrogenationdehydrogenation reactions,
aromatization of C
6
parafnes or DeNO
x
[14].
Recently also, chromia supported on pure or modied
zirconia has been investigated as catalysts, in parti-
cular, for the dehydrogenation and aromatization of
hydrocarbons [47]. Lanthana proved to be a bene-
cial component in zirconia supported chromium
oxide as a catalyst for the selective aromatization of
C
6
parafnes [7]. One reason of the obvious versa-
tility of supported chromium oxides as catalysts is
certainly the variability of the chromium oxidation
state and of the aggregation state of chromiumspecies,
with these properties being strongly dependent on Cr
loading, treatment conditions as well as on the nature
of the support [2,3]. Therefore, investigations of these
properties are desirable. Besides spectroscopic meth-
ods [2,3], temperature-programmed reduction (TPR)
has proved to be an useful technique for the study of
the redox behavior of supported catalyst components
[8]. TPR has already been applied for the study of
chromia catalysts, too. However, most of the pub-
lished TPR studies focus on unsupported chromia or
on chromia supported on alumina, silica or titania [8
13], and not yet on chromia supported on zirconia.
Thermochimica Acta 345 (2000) 9399
*
Corresponding author. Fax: 49-30-6392-4370.
E-mail address: [email protected] (H. Lieske).
0040-6031/00/$ see front matter # 2000 Elsevier Science B.V. All rights reserved.
PII: S0 0 4 0 - 6 0 3 1 ( 9 9 ) 0 0 3 8 5 - 8
With our paper, we investigated the redox behavior
of catalysts consisting of chromium oxide supported
on pure zirconia and on lanthana modied zirconia by
means of TPR, hoping that this method would deliver
newinsight into the state of chromiumspecies and into
the benecial role of the lanthana component in the
chromium oxide/lanthanazirconia aromatization cat-
alysts.
2. Experimental
2.1. Catalyst preparation
Catalyst samples were prepared by impregnating
zirconium hydroxide based supports, i.e. amorphous
zirconium hydroxide pure or doped with 7.0 wt.%
La
2
O
3
(MEL, UK) with an aqueous solution of
(NH
4
)
2
CrO
4
. The supports were dried in air at
1208C for 12 h and then immersed in solutions con-
taining the appropriate amounts of (NH
4
)
2
CrO
4
to
yield samples loaded with 0.5, 1.0, 4.0 and
8.0 wt.% Cr. By ammonia addition, the solution
was kept at a pH of 10. Under stirring, the excess
water was slowly evaporated at 50608C. The pro-
ducts obtained were dried at 1208C for 12 h and
calcined in air at 6008C for 4 h, resulting in chromium
oxide supported on zirconia (Z) (CrO
x
/ZrO
2
) and on
lanthanazirconia (LZ) (CrO
x
/La
2
O
3
ZrO
2
). The cat-
alyst samples were kept dry over zeolite till further
use.
Below, the catalyst samples will be denominated by
their Cr content and support nature. Thus, 0.5CZ and
0.5CLZ indicate 0.5% Cr on Z and 0.5% Cr on LZ
support, respectively.
As reference substances, the chromium-free Z and
LZ supports were dried and calcined under the above
conditions, too.
2.2. Temperature-programmed reduction
The TPR experiments have been performed using a
homemade gas ow system, including a xed-bed
quartz reactor and a gas analyzer unit. In this system,
samples can be pretreated in several gas ows at
desired conditions. The catalyst samples, with a par-
ticle size of 0.30.8 mm and calcined at 6008C for 4 h,
were placed into the reactor and heated in an argon
ow at 3008C for 1 h before the TPR experiment.
After cooling the sample in an argon ow to 508C, the
rst TPR run, TPR1, was carried out in 5.17% hydro-
gen in argon ow at a heating rate of 10 K/min and a
gas ow rate of 15 ml/min. Optimum sample weights
had been estimated according to the equation by Monti
and Baiker [14]. TPR1 was nished by an isothermal
period of 0.5 h at 7508C. Hydrogen consumed was
monitored by a calibrated thermal conductivity detec-
tor (GOW-MAC Instruments). The samples, having
been reduced during TPR1, were reoxidized in situ by
20% oxygen in helium ow at 6008C for 1 h. After-
wards, a second TPR run, TPR2, was carried out under
the conditions of TPR1.
From the TPR signal intensities, changes of the
oxidation state (n m), dened by
Cr
n

n m
2
h i
H
2
=Cr
m
(n m)H

(1)
have been calculated.
3. Results
3.1. TPR1 and TPR2 of CZ catalysts
TPR1proles of the CZsamples are showninFig. 1a.
The TPR prole of the chromium-free support, Z, is
shown, too. It demonstrates a hydrogen uptake of
the zirconia support at 5748C. For the CZ catalysts,
two resolved TPR signals can be observed, the rst
signal at T
max1
= 3423698C and the second one at
Fig. 1. TPR proles of Cr/ZrO
2
(CZ) catalysts; (a) TPR1, (b)
TPR2.
94 D.L. Hoang, H. Lieske / Thermochimica Acta 345 (2000) 9399
T
max2
= 4735338C. The intensity of the rst peak
increases with increasing Cr loading, that of the
second peak remains almost unchanged. With CZ
catalysts, a clear dependence of T
max1
on the Cr
content could not be observed. The position of T
max2
is slightly shifted to lower temperatures with increas-
ing Cr content.Reoxidation of 1CZ, 4CZ and 8CZ was
followed by TPR2 experiments, the proles of which
are given in Fig. 1b. Apparently, the TPR2 proles do
not signicantly differ from the TPR1 proles and
show TPR signals at almost the same temperatures,
namely at T
max1
= 3443688C and at T
max2
= 469
5258C. As with TPR1, increasing Cr content only
increases the intensity of the rst TPR2 peak but
not of the second one. Also, T
max2
slightly shifts to
lower values at higher Cr contents. As a further
observation, the intensities of the rst TPR2 peaks
are signicantly lower than those of the rst TPR1
peaks. Fig. 2 depicts the dependence on the Cr content
of hydrogen consumed within the rst TPR1 and
TPR2 peaks. As already mentioned, increasing Cr
content only increases the intensity of the rst TPR
peaks, but leaves that of the second ones practically
unchanged. Moreover, with Cr contents between 0 and
4.0 wt.%, a linear correlation between Cr loading and
hydrogen consumption under the rst TPR peaks is
observed. The hydrogen consumptions are lower than
expected for (n m) = 3, which would correspond to
a total reduction of Cr
6
to Cr
3
. Furthermore, Fig. 2
demonstrates that the deviations of the intensities of
the rst peak from a total reduction of Cr
6
to Cr
3
increase with higher Cr loading. Moreover, the hydro-
gen consumption under the rst peak of TPR2 is
always lower than that of TPR1.
3.2. TPR1 and TPR2 of CLZ catalysts
Fig. 3a shows the TPR1 proles of the lanthana-
modied CLZ catalysts. As in the case of the CZ
catalysts, the CLZ catalysts consume hydrogen under
two TPR peaks and increasing Cr loading increases
the intensities of the respective rst peaks. Moreover,
their T
max1
shifts to lower values, in detail from 415 to
3688C. The second TPR1 peaks of 0.5CLZ and 1CLZ
samples are found at 556 and 5428C, respectively.
With 4CLZ and 8CLZ, their second TPR1 peaks are
further shifted to lower temperatures and deformed at
the same time, resulting in shoulders at 542 and
4718C. The TPR prole of the LZ support exhibits
an hydrogen consumption peak in the same tempera-
ture range, at 5548C.
The TPR2 proles of 1CZL, 4CLZ and 8CLZ are
shown in Fig. 3b. The TPR2 spectra do not essentially
differ from the TPR1 proles in Fig. 3a. In Fig. 4, the
amounts of hydrogen consumed under both the peaks
measured by TPR1 and TPR2 are plotted versus the Cr
loading. Analogously to the CZ catalysts, increasing
Cr content increases the intensity of the rst hydrogen
consumption peaks, but leaves that of the second ones
nearly unchanged. Also here, the amounts of hydrogen
consumed deviate to some degree from those ones
expected in case of a total reduction of Cr
6
to Cr
3
.
Unlike the CZ catalysts, the TPR2 results on CLZ
samples do not signicantly differ from TPR1, except
for the sample 8CLZ.
Fig. 2. Quantitative TPR results of CZ catalysts; (*) hydrogen
consumption calculated according to Eq. (1) with n = 6 and m = 3.
Fig. 3. TPR proles of Cr/La
2
O
3
ZrO
2
(CLZ) catalysts; (a) TPR1,
(b) TPR2.
D.L. Hoang, H. Lieske / Thermochimica Acta 345 (2000) 9399 95
Quantitative results of TPR1 and TPR2 on the CZ
and CLZ samples, consisting of T
max
values, amount
of hydrogen consumed under the rst and second peak
as well as (n m) values calculated fromthe rst peak
are summarized in Table 1.
The main observations to be taken from Figs. 14
and Table 1 are the following:
v The chromium-free supports, Z and LZ, cause
hydrogen uptake signals at 574 and 5548C, respec-
tively.
v Both the CZ and CLZ catalysts give two separate
TPR signals. With CZ, these signals appear at
T
max1
= 3403708C and T
max2
= 4705338C,
with CLZ at T
max1
= 3704158C and
T
max2
= 4905568C. The temperature maxima
shift more or less to lower temperatures as the
Cr loading is increased. Increasing Cr content
increases the intensity of the first TPR peak, but
does not significantly influence the intensity of the
second one.
v The oxidation state change (n m) reaches values
near 3 with catalysts of low chromium content and
in the presence of lanthana. With higher chromium
loading and in the absence of lanthana, consider-
able deviations of (n m) from the value 3 were
observed.
v With the lanthanum-free CZ catalysts, the hydro-
gen consumption during the second TPR run
TPR2, i.e. after reoxidation, has been generally
lower than the hydrogen consumption during the
Fig. 4. Quantitative TPR results of CLZ catalysts; (*) hydrogen
consumption calculated according to Eq. (1) with n = 6 and m = 3.
Table 1
Summarized TPR results on CZ and CLZ catalysts
Catalysts %Cr T
max1
(8C)
H
2
consumption,
1st peak (mmol/g)
n m
a
T
max2
(8C)
H
2
consumption,
2nd peak (mmol/g)
TPR1
Z 0 574 0.0605
CZ 0.5 350 0.1310 2.72 533 0.0628
1.0 342 0.2574 2.68 523 0.0594
4.0 369 0.9971 2.59 503 0.0758
8.0 350 1.2933 1.68 473 0.1263
LZ 0 554 0.0558
CLZ 0.5 415 0.1423 2.96 556 0.0532
1.0 410 0.2874 2.99 542 0.0621
4.0 390 0.9697 2.52 522 0.0691
8.0 368 1.8063 2.35 487 0.1546
TPR2
CZ 1.0 344 0.1796 1.87 525 0.0523
4.0 367 0.596 1.55 505 0.11
8.0 350 0.9353 1.22 469 0.1118
CLZ 1.0 420 0.2352 2.63 537 0.0641
4.0 403 0.9749 2.53 522 0.053
8.0 368 1.1353 1.48 500 0.1317
a
The (n m) values were calculated from hydrogen consumed under the rst TPR peak.
96 D.L. Hoang, H. Lieske / Thermochimica Acta 345 (2000) 9399
foregoing TPR1 run. With the lanthana-containing
CLZ catalysts, except 8CLZ, the TPR2 hydrogen
consumptions do not significantly differ from
TPR1.
v With identical Cr contents, T
max1
of CLZ catalysts
is always higher than T
max1
of CZ catalysts.
4. Discussion
In literature, the redox behavior of supported chro-
mium has been found to depend on the Cr loading, on
the support nature as well as on treatment conditions
[2,3]. For zirconia supported chromia, the existence of
Cr
6
and Cr
5
as well as of some Cr
2
O
3
has been
detected after calcination at 5508008C by several
methods [26,1518]. The higher valent Cr species are
believed to be anchored by esterication with support
surface hydroxyl groups, resulting in real surface
species. Beyond a saturation coverage with such sur-
face species, i.e. beyond a certain chromium content,
which is determined by the nature of the support,
excess chromium species are converted to bulky
Cr
2
O
3
by calcination at high temperature. By treating
at 1206008C with a reducing agent, chromium oxi-
dation states of 6 or 5 have been reported to be
reduced mainly to Cr
3
[26,1518], but partly also to
Cr
4
[5] or to Cr
2
[4,15].
Our aim was to investigate the catalyst system
CrO
x
/La
2
O
3
ZrO
2
, which had proved to be an inter-
esting and very selective dehydrocyclization catalyst
[7], by means of TPR, looking for insight in the
oxidation and dispersion state of the chromium oxide
component as well as on the role of lanthana within
this catalyst system.
TPR spectra can give, by the T
max
positions of their
hydrogen consumption signals, qualitative informa-
tion on the reduction behavior of the sample under
investigation and, by the intensity of the signals,
quantitative information on the amount of reducible
species. In case of supported substances, the results
can allow to calculate the change of the oxidation state
of a supported species during its reduction. If TPR
experiments are included into the investigations with
samples reoxidized after a rst TPR run and experi-
ments with varying content of the supported species,
the results can give valuable information on changes
of the chemical state or of the dispersion state during
the reduction and can allow to differentiate between
hydrogen consumption of a reducible species or of the
support, respectively.
All the TPR1 proles of CZ and CLZ catalysts
shown in Figs. 1 and 3 exhibit two hydrogen con-
sumption stages appearing as two peaks. As shown in
Table 1 and Figs. 2 and 4, only the intensities of the
rst TPR peak, but not of the second one, depend on
the Cr loading. This should mean that only the rst
TPR peak can be assigned to the reduction of sup-
ported Cr species. Hence, one should exclude that the
second peak is due to a further reduction of Cr
3
formed under the rst peak into Cr
2
[9,19,20] or to a
separate reduction of another higher valent Cr species
at higher temperature [4,10,15]. The range of the
T
max2
temperatures is near the hydrogen uptake tem-
perature of the pure zirconia Z. Hence, with high
probability the second peak is to be assigned to a
hydrogen uptake of the zirconia support, already
observed with pure zirconia [2123]. The second peak
shifts towards lower T
max2
values with increasing
chromium content. This can perhaps be explained
by an activation of hydrogen by chromium oxide
species, facilitating the hydrogen uptake of the zirco-
nia supports.
With the catalysts containing only 0.5 or 1 wt.% Cr,
especially with the samples additionally containing
lanthana, the amounts of hydrogen consumed within
the rst TPR peak, which is assigned to chromium
reduction, correspond to (n m) values near 3. This is
most probably explained by the assumption that the
initial Cr oxidation state had been predominantly
Cr
6
, which is reduced to Cr
3
during TPR. At higher
chromium contents, the (n m) values of both the CZ
and CLZ catalysts are more or less lower than 3, i.e.
lower than expected for a total reduction of Cr
6
to
Cr
3
. This deviation could be a hint on an incomplete
reduction of Cr
6
[18]. More probably, however, this
can be explained by the assumption mentioned above
that an excess part of Cr
6
, which could not be
chemically stabilized on the support surface, is con-
verted into lower valent chromium already by the
preceding calcination at 5006008C, i.e. before TPR
[3,6,18].
The (n m) values of the CLZ samples are gen-
erally higher than those of the CZ samples, see Table
1. This can most clearly be seen by comparing the
samples 8CZ and 8CLZ. In terms of the above model
D.L. Hoang, H. Lieske / Thermochimica Acta 345 (2000) 9399 97
of the co-existence of stabilized surface species and
thermally instable bulky excess species before the
reduction, this nding indicates that the surface of
the LZ support contains a higher number of the
stabilizing surface groups, leading to a higher capacity
for the more stable chromium surface species. The
shift of T
max1
to lower temperatures with increasing Cr
loading can be explained by means of this model, too.
Increasing chromium content leads beyond the satura-
tion limit to more bulky and non-stabilized chromium
species which are reduced at lower temperatures.
A feature distinguishing the TPR1 proles of the
CLZ catalysts from those of the CZ catalysts are the
T
max1
values. The values of the CLZ samples are
always higher than the corresponding values of the
CZ samples (with the same Cr loading). This suggests
that not only a higher number of Cr
6
species can be
stabilized on the lanthana containing support, but that
lanthana also strengthens the interaction between
higher valent chromium species and the support sur-
face. Such a stronger interaction means a stabilization
of Cr
6
species and results in higher reduction tem-
peratures. The more detailed mode of action of the
lanthana addition is not yet clear.
According to former works, Cr
3
species formed on
supports after reduction proved to be reoxidable to
Cr
6
to some extent [3,13,18]. This behavior may be
of interest, if, for example, an oxidative regeneration
of Cr based catalysts is necessary, as is usual with
hydrocarbon conversion reactions, or if such catalysts
shall be applied in deep oxidation reactions, which
require a steady oxidationreduction cycle of the
catalytically active element [24]. However, on differ-
ent supports different types of supported Cr can exist
[2,10], which differ in their degree of reoxidability.
Therefore, it was of interest to investigate this problem
also with our catalysts, taking the TPR2 experiments
as an appropriate tool.
With the TPR2 proles of the CZ catalysts, Fig. 1b,
Fig. 2 and Table 1, the intensities of the rst TPR peak
have been always lower than with the corresponding
TPR1 spectra. But, in case of the CLZ samples, Fig. 3
as well as Fig. 4 and Table 1, the intensities of the rst
TPR peak are almost identical with TPR2 and TPR1,
with the exception of the high chromium content
sample 8CLZ. This indicates that the reduction of
the lanthanum-containing CLZ catalysts preferably
results in Cr
3
species, the nature of which enables
a reoxidation. Having in mind that the TPR1 results on
the lanthanum-containing CLZ samples are best
explained by a lanthana induced stabilization of
Cr
6
species, which are surface dispersed on a mole-
cular level, the reoxidable Cr
3
should be of high
dispersion. This is in line with the picture that high
chromium content and low chromium-support inter-
action (CZ samples) result in bulky chromia already
during high temperature calcination. According to the
results by Zaki et al. for alumina supported chromia
[15], those chromium species, which can reversibly be
reduced and reoxidized, may be polychromates. The
anchoring of polychromate species on zirconia was
actually evidenced by UVVIS diffuse reectance
spectroscopy and Raman spectroscopy [2,3,17].
5. Conclusions
v The results of the TPR study on chromium oxide/
zirconia and chromium oxide/lanthanazirconia
can be explained in terms of a model, which
assumes the stabilization of Cr
6
species on the
surface of the zirconia support and, beyond a
saturation coverage, a co-existence of such surface
stabilized and therefore molecularly dispersed
Cr
6
with non-stabilized bulky Cr
6
species.
The stabilized Cr
6
species are reduced at rela-
tively high temperature to dispersed Cr
3
species,
which are reoxidable. A part of the non-stabilized
species can already thermally decompose during
calcination, resulting in bulky and non-reoxidable
chromia. The rest of them is reduced at lower
temperatures resulting in Cr
3
species, which are
not quantitatively reoxidable.
v Lanthana increases the capacity of the support for
the formation of surface stabilized Cr
6
, leading to
a larger portion of highly dispersed Cr
3
during
reduction. This ability of lanthana might be one of
the reasons for its beneficial role in chromium
oxide/lanthanazirconia aromatization catalysts.
Acknowledgements
This work was supported by the Berlin Senat
Department of Science, Research and Culture (Project
No. 89 11 30 001) as well as by the Federal Ministry of
98 D.L. Hoang, H. Lieske / Thermochimica Acta 345 (2000) 9399
Education and Research of the Federal Republic of
Germany (Project No. 03C30120). The authors are
grateful to MEL Chemicals, Manchester, UK, for
supplying them with zirconia materials.
References
[1] Ch.P. Poole Jr., D.S. McIver, Adv. Catal. 17 (1967) 223.
[2] B.M. Weckhuysen, I.E. Wachs, R.A. Schoonheydt, Chem.
Rev. 96 (1996) 3327.
[3] B.M. Weckhuysen, R.A. Schoonheydt, Catal. Today 51
(1999) 223.
[4] V. Indovina, Catal. Today 41 (1998) 95.
[5] K. Arata, M. Hino, H. Matsuhashi, Appl. Catal. 100 (1993)
19.
[6] J.R. Sohn, S.G. Ryu, H.W. Kim, J. Mol. Catal. 135 (1998) 99.
[7] H. Lieske, D.L. Hoang, German patent DE-OS 196 12 000.
[8] N.W. Hurst, S.J. Gentry, A. Jones, B.D. McNicol, Catal. Rev.
24 (1982) 23.
[9] W. Grunert, W. Saffert, R. Feldhaus, K. Anders, J. Catal. 99
(1986) 149.
[10] M.I. Zaki, N.E. Fouad, G.C. Bond, S.F. Tahir, Thermochim.
Acta 285 (1996) 176.
[11] J. Engweiler, J. Nickl, A. Baiker, K. Kohler, C.W. Schlapfer,
A. von Zelewsky, J. Catal. 145 (1994) 141.
[12] L.I. Ilieva, D.H. Andreeva, Thermochim. Acta 265 (1995)
223.
[13] M.I. Zaki, M.A. Hassan, N.E. Fouad, Appl. Catal. A: General
171 (1998) 315.
[14] D.A.M. Monti, A. Baiker, J. Catal. 83 (1983) 323.
[15] A. Cimino, D. Cordischi, S. de Rossi, G Ferraris, D. Gazzoli,
V. Indovina, M. Occhiuzzi, M. Valigi, J. Catal. 127 (1991)
761.
[16] T. Yamaguchi, M. Tan-no, K. Tanabe, in: G. Poncelet, P.A.
Jacobs, P. Grange, B. Delmon (Eds.), Preparation of Catalysts
V, Elsevier, Amsterdam, 1991, pp. 567574.
[17] D.S. Kim, I.E. Wachs, J. Catal. 141 (1993) 166.
[18] J.R. Sohn, S.G. Ryu, Langmuir 9 (1993) 126.
[19] A. Hakuli, M.E. Harlin, L.B. Backman, A.O.I. Krause, J.
Catal. 184 (1999) 349.
[20] B.M. Weckhuysen, R.A. Schoonheydt, J.-M. Jehng, I.E.
Wachs, S.J. Cho, R. Ryoo, S. Kijistra, E. Poels, J. Chem. Soc.
Faraday Trans. 91(18) (1995) 3245.
[21] C. Dall
/
Agno, A. Gervasini, F. Morazzoni, F. Pina, G. Strukul,
L. Zanderighi, J. Catal. 96 (1985) 106.
[22] W.K. Jozwiak, React. Kinet. Catal. Lett. 30(2) (1986) 245.
[23] D.L. Hoang, H. Lieske, Catal. Lett. 27 (1994) 33.
[24] R. Prasad, L.A. Kenedy, A.E. Ruckenstein, Catal. Rev.-Sci.
Eng. 26(1) (1984) 1.
D.L. Hoang, H. Lieske / Thermochimica Acta 345 (2000) 9399 99