Study of the decomposition of supported nickel acetylacetonate by thermal techniques

Article abstract of DOI:10.1016/S0040-6031(02)00486-0

The decomposition of supported nickel acetylacetonate (AcacNi) was studied by using thermal techniques to evaluate the nature of the calcination products. Thermogravimetric techniques used in this work indicated transformations occurring during preparatio

Full text of DOI:10.1016/S0040-6031(02)00486-0

Thermochimica Acta 400 (2003) 101–107
Study of the decomposition of supported nickel
acetylacetonate by thermal techniques
a
a
a
a,b,∗
Nora N. Nichio , Mónica L. Casella , Esther N. Ponzi , Osmar A. Ferretti
a
Centro de Investigación y Desarrollo en Procesos Catal ´ı ticos (CINDECA), Universidad Nacional de La Plata,
calle 47 No. 257, CC59, 1900 La Plata, Argentina
Departamento de Ingenier ´ı a Qu ´ı mica, Facultad de Ingenier ´ı a, Universidad Nacional de La Plata,
b
calle 47 No. 257, CC59, 1900 La Plata, Argentina
Received 2 June 2002; received in revised form 20 September 2002; accepted 20 September 2002
Abstract
The decomposition of supported nickel acetylacetonate (AcacNi) was studied by using thermal techniques to evaluate the
nature of the calcination products. Thermogravimetric techniques used in this work indicated transformations occurring during
preparation and calcination steps. Results demonstrated the important influence of support modifications upon the catalyst
nature finally obtained.
©
2002 Elsevier Science B.V. All rights reserved.
Keywords: Nickel catalysts; Nickel acetylacetonate; Methane oxidation; Synthesis gas
1
. Introduction
[7–9]. Previous studies have demonstrated that the
use of nickel acetylacetonate (AcacNi) and alumina
supports enabled a catalyst of good activity and sta-
bility, specially concerning the deactivation due to
carbon deposition, compared to catalysts prepared
with nickel nitrate as precursor compound [10,11].
A large amount of information about a catalytic
system can be obtained by using thermogravimetric
techniques (working in a non-steady state and with an
appropriate temperature program) [12]. From thermal
data, it is possible to achieve information regarding
the composition and structure of different phases of a
given sample. Our principal objective is the study of
the decomposition of the nickel precursor compound
(AcacNi) during calcination to evaluate the nature of
phases formed through support modifications.
Activity and selectivity are well-known properties
of supported nickel catalysts in different natural gas
reforming reactions (steam reforming, reforming with
CO2, oxyreforming and mixed reforming processes)
for the obtention of synthesis gas [1,2]. However,
these catalysts suffer severe deactivation problems
due to carbon deposition, nickel sintering and phase
transformations [3–6]. Therefore, the development
of a new stable Ni-catalyst is of great interest for
the practical application of these reactions. The sup-
port nature, precursor compounds used to generate
the active phase, and pre-treatments have an impor-
tant role in the performance of the resulting catalyst
∗
Weight changes as a function of the temperature
registered through TGA technique are caused by the
formation and rupture of chemical bonds at high
Corresponding author. Tel.: +54-221-421-0711;
fax: +54-221-425-4277.
E-mail address: ferretti@quimica.unlp.edu.ar (O.A. Ferretti).
0
040-6031/03/$ – see front matter © 2002 Elsevier Science B.V. All rights reserved.
PII: S0040-6031(02)00486-0

1
02
N.N. Nichio et al. / Thermochimica Acta 400 (2003) 101–107
temperatures, by processes that cause volatile com-
pounds or by reaction products that modify the sample
weight. The exit flow analysis of the TGA equipment
is very useful to get information on the stoichiomet-
ric ratio of the thermally decomposed compounds.
The coupled analyzer, a mass spectrometer (MS),
makes it possible the determination of the effluent gas
composition.
reaction was carried out in a fixed bed reactor, at tem-
perature between 423 and 573 K, atmospheric pres-
sure and a feed of 4.5% IPA in He, with a flow of
3
−1
40 cm min
.
The calcination or activation of the solids was
carried out in an oxidizing atmosphere, after the
impregnation step, and it was continuously mon-
itored by thermogravimetric analysis (Shimadzu
TGA-50 equipment), differential scanning calorime-
try (DSC) (Shimadzu DSC-50 equipment) and in a
parallel analysis by mass spectrometry (Shimadzu
GCMS-QP5050A). Ten milligram sample were used
for the thermal decomposition of impregnated mate-
2
. Experimental
2
.1. Catalyst preparation
3
−1
rials. The samples were treated with a 30 cm min
gas feed flow composed of a He/O2 mixture (4% O2),
Four supported Ni catalysts, with 2 wt.% Ni,
−
1
were prepared using nickel acetylacetonate (AcacNi)
Aldrich). Table 1 summarizes the studied catalytic
with a heating program of 5 K min , from room
temperature up to 393 K, remaining at this level for
30 min (drying stage), and then from 393 to 1073 K,
with a heating rate of 5 K min (stage of calcination).
The GC/MS equipment had a capillary column of
30 m×0.25 mm, with SPB-5TM (Supelco) as station-
ary phase, operated in an isothermal mode at 303 K.
The temperature-programmed reduction (TPR) tests
were carried out in a conventional dynamic equipment,
with a ratio H2/N2 in the feed flow of 1/9, and a heat-
(
systems and their corresponding denomination.
The modification of the ␣-Al2O3 support was
carried out by impregnation with Al(NO3)3·9H2O
−
1
(Merck), in such a way that one obtains 1 wt.% Al in
the solid. The organometallic catalysts were prepared
by impregnation of the corresponding supports, pre-
◦
viously calcined at 500 C for 2 h. AcacNi was used
dissolved in benzene, with a solution volume of five
times the pore volume of the support. The system was
−
1
ing rate of 10 K min from room temperature up to
1273 K.
◦
left in contact for 72 h at 60 C, and later on the solid
was filtered and washed.
3. Results and discussion
2
.2. Apparatus and measurements
The final properties of a catalyst depend on the
The determination of the textural properties of the
nature of the precursor-support interaction generated
during the preparation step. In this case, the acid–base
properties of the support play an important role. For
this reason, a characterization of textural properties
and superficial acidity of different studied supports
was carried out (Table 2).
The isopropylalcohol reaction (IPA) was used
for the characterization of the superficial acidity.
The selectivity towards different reaction products
changes if the superficial acid–base properties are
modified. Thus, a higher selectivity towards propene
and di-isopropylether (SP + D) evidences the ex-
istence of a relatively higher concentration of acid
sites [11]. Supports based on ␥-Al2O3 present a
higher contribution of acid sites than supports based
on ␣-Al2O3. This fact may influence the interaction
supports (superficial area and pore volume) was car-
ried out in a N2 adsorption equipment (Micromeritics
◦
Accursorb 2100E) at −196 C on a 200 mg sample,
◦
previously outgased at 200 C under vacuum for 2 h.
The acid–base properties of the different supports were
determined by the isopropylalcohol (IPA) decomposi-
tion reaction, which is an indirect method [13]. This
Table 1
Nomenclature of the supports and catalysts studied
Support
Catalyst name
O
Ni
α
␣
-Al2O3 (T-708 Girdler) (␣)
O
␣-Al2O3 (T-708 Girdler) modified with
Ni
α+Al
1
wt.% Al (α + Al)
O
γ
O
γp
␥
␥
-Al2O3 (Cyanamid Ketjen) (γ)
-Al2O3 (Prepared in the laboratory) (γp)
Ni
Ni

N.N. Nichio et al. / Thermochimica Acta 400 (2003) 101–107
103
Table 2
cess. During the drying step (up to 393 K), the sam-
ple presented a mass loss of the order of 10.4% (Pw),
due to water loss. In the calcination step, three main
Textural and acid–base properties of the supports employed
Support BET surface
Pore volume IPA
SP + D
b
2
−1
3
−1
area (m g
)
(cm g
)
conversion (%)
peaks are observed: a minor one, P , at 477 K (that
1
a
(%)
represents a mass loss of 18.7%), a bigger one, P2, at
α
5.1
5.5
216
185
0.22
0.20
0.50
0.54
12
30
60
80
83
92
6
5
00 K with a shoulder, P3, at 633 K (together represent
1.8% of mass loss). The theoretical total mass loss
α + Al
γp
γ
would represent 71% if AcacNi decomposes form-
ing NiO. Results of this test gave a mass loss deter-
mined during the calcination step of 70.5% measured
as (P1 + P2 + P3); this result demonstrates that the
decomposition product of AcacNi is NiO.
100
100
a
Test conditions: 523 K, atmospheric pressure, feed composi-
3
−1
tion 4.5% IPA in He, feed flow 40 cm min
.
b
SP + D = selectivity to propene and diisopropylether, mea-
sured at a conversion of IPA of 40%.
Fig. 2 shows TGA diagrams corresponding to the
calcination step (>393 K) for all studied catalysts. The
decomposition pattern is strongly influenced by the
degree of the nickel precursor compound and the sup-
port during the preparation step. The modification of
the ␣-Al2O3 support with an aluminum oxide layer
O
support nature. The Ni catalyst presents a TGA dia-
α
(
α + Al) causes a certain increase in the total acidity
gram (Fig. 2a) quite similar to that of non-supported
AcacNi. Differences that can be pointed out are a shift
of all peaks to higher temperatures (approximately
measured by a higher activity and selectivity towards
di-isopropylether and propene (SP + D).
The evolution of pure AcacNi by TGA and DSC
was determined to obtain decomposition patterns dur-
ing the calcination step. The derivative of the weight
1
30 K) and a slightly bigger contribution of P peak.
This P peak appears at 516 K, representing a mass
1
loss of 33%, and the second peak, P , appears at 603 K
2
−1
change from TGA tests (DrTGA/mg min ) is rep-
resented in Fig. 1 for better visualization of weight
changes. Results obtained by DSC show that all de-
composition reactions are exothermic transformations,
except for the water loss occurring in the drying step
with a shoulder, P , at 683 K, representing a mass loss
3
(P + P ) of 67%.
2
3
O
α+Al
Fig. 2b shows the Ni
sample behavior. This
sample also presents three decomposition peaks (P at
1
502 K, P at 606 K and P at 683 K). In this case, peaks
2
3
O
(
T < 393 K), which is naturally an endothermic pro-
appear at temperatures similar to those of Ni catalyst,
α
Fig. 1. TGA–DSC curves for the decomposition of bulk AcacNi.

1
04
N.N. Nichio et al. / Thermochimica Acta 400 (2003) 101–107
O
α
O
α+Al
, (c) NiO , and (d) Ni .
O
Fig. 2. TGA curve for the decomposition of catalyst: (a) Ni , (b) Ni
γp γ
however, the first peak, P1, is noticeably bigger in
(endothermic) is more important than in the region
corresponding to P2 peak, where the combustion re-
actions (exothermic) are markedly predominant. The
decomposition at high temperatures would indicate a
stronger interaction of AcacNi with the support. Ac-
cording to these results, the growing order in the in-
the case of NiO , since it represents 70% of the
α+Al
O
α
total mass loss in the calcination step, while for Ni
catalyst, P1 peak represents a mass loss of 33%.
O
O
Results obtained with Ni and Ni catalysts differ
γp
γ
O
α
O
from those obtained for Ni , Ni
and AcacNi sam-
α+Al
O
O
teraction with the support would be Ni < Ni
<
α
α+Al
ples. Fig. 2c presents the TGA diagram corresponding
to Ni catalyst, where a single decomposition peak
O
O
O
γp
Ni < Ni . The P1 peak contribution (represent-
γp
γ
ing lesser exothermic processes) related to the sum
appears, with a maximum at approximately 613 K.
O
γ
of peaks (P2 + P3) would indicate the rupture ratio
Ni catalyst (Fig. 2d) presents its single decomposi-
and combustion reactions during the decomposition
tion peak at approximately 783 K, a temperature no-
ticeably higher than in the other samples.
All decomposition reactions occurring during the
calcination step are exothermic transformations ob-
O
samples, re-
α+Al
sults indicate that with the modified support α + Al,
the contribution of rupture reactions is more important
than when the ␣ support is used.
O
O
served by DSC. In the case of Ni and Ni
samples,
α
α+Al
In agreement with what has been proposed for
characterization results of different supports em-
ployed, the acid nature of ␥-Al2O3 supports generates
a stronger interaction with the precursor compound,
the second peak, P2, represents a more exothermic
process than the first peak, P1. This behavior could
be explained by the fact that in the region correspond-
ing to P1 peak the contribution of rupture reactions

N.N. Nichio et al. / Thermochimica Acta 400 (2003) 101–107
105
causing its decomposition at higher temperatures than
pure AcacNi or supported on ␣-Al2O3.
higher relative contribution of rupture reactions in the
region of P1 peak. With the interaction increase of
AcacNi with the support, it is possible to suppose that
a weakening of the molecule internal bonds could
be produced causing consequently an increase in the
number of ruptures and also a percentage increase of
light fragments during decomposition.
Additional GC/MS tests were performed on sam-
ples to analyze the composition (both qualitative and
quantitative) of gases evolved during their decomposi-
tion. Results obtained in these analyses together with
those obtained by TGA indicate that there are no or-
O
O
When studying the thermal decomposition of NiO
ganic fragments left on the surface in Ni and Ni
α
α+Al
γp
and NiO catalysts, the process is different to the
catalysts after reaching 723 K. Results obtained by
IR spectroscopy on the calcined samples were similar
with the subsequent disappearance of the wide band at
600 cm assigned to organic fragments of AcacNi
14].
For a better comprehension of results obtained by
γ
AcacNi, NiO and NiO
decomposition. The main
fragment is not the one corresponding to CO2 (m/z =
4) but to a fragment with m/z = 43 (Fig. 3) and
heaviest fragments do not appear at m/z = 85 and
α
α+Al
−1
1
4
[
1
00 but rather up to m/z = 58. These results are sim-
GC/MS, values of relative abundance (%) of main
fragments (m/z) are shown in Fig. 3 for each one of
the tested materials. With regard to the composition
of effluent gases produced during the calcination of
pure AcacNi, the main fragment corresponds to CO2
ilar to the previous ones: γp and γ supports presented
higher relative acidity than the a and α + Al sup-
ports, a result that has been correlated with a stronger
precursor-support interaction. These facts may be
responsible for a shift of the decomposition towards
higher temperatures, together with a more important
contribution of rupture reactions.
(
m/z = 44) in agreement with DSC results that show
the high decomposition exothermicity.
O
O
α+Al
If the composition of effluents for Ni and Ni
α
Reduction profiles were also determined by TPR for
all calcined catalysts (Fig. 4). As it is well documented
in numerous studies carried out on the nickel reduction
samples is analyzed, the most significant differences
when compared with AcacNi are the major abun-
dance of small fragments (m/z = 12) with smaller
CO2 amounts (m/z = 44). This is consistent with a
Fig. 3. Decomposition fragments detected by GC/MS analysis,
obtained during the calcination of bulk AcacNi and Ni catalysts.
Fig. 4. TPR profiles for Ni studied catalysts.

1
06
N.N. Nichio et al. / Thermochimica Acta 400 (2003) 101–107
supported on alumina, three reduction peaks can be
found: the first, in the region of 600–700 K assigned
to the presence of bulk NiO or a nickel oxide having
a weak interaction with the support; the second, in
the region between 700 and 1000 K, corresponding to
nickel species and aluminum mixed oxides in strong
interaction with the support; and the last, in the region
of 1000–1300 K assigned to the existence of NiAl2O4
species, a crystalline spinel characterized by a strong
interaction of the nickel with the support [15,16].
Fig. 4 shows the effect of the support nature upon
TPR profiles. When comparing the catalyst supported
on α+Al with the catalyst supported on ␣, the peak of
maximum H2 consumption is shifted towards higher
temperatures in the first catalyst. In the case of NiαO
catalyst, reduction temperatures do not differ appre-
ciably from what is generally assigned to nickel with
Table 3
Catalytic performance in the partial oxidation of methane reaction
Catalyst
Methane
Stability
coefficient
aCH4
0.74
0.94
0
Particle size
increment
(%)
conversiona
XCH4 (%)
b
b
O
Ni
α
64
82
<5
<5
95
21
n.d.
n.d.
O
Ni
α+Al
O
Ni
γ
O
γp
Ni
0
a
Test conditions: conventional flow reactor, 773 K, atmo-
spheric pressure, feed composition N2/CH4/O2:11/2/1, feed flow:
3
−1
6
5 cm min , conversion measured at 2 h on stream.
b
Determined after 24 h on stream. aCH4 : ratio between the rate
of consumption of CH4 after 24 h on stream and the initial rate
value. Particle size increment (%): ((mean particle size after 24 h
on stream −mean particle size of the fresh catalyst)/mean particle
size of the fresh catalyst) × 100, measured by TEM. n.d.: non
determined.
relatively weak interaction with the support. On the
O
presents a resistance to sintering that it is noticeably
other hand, for Ni
be present in the region at which the reduction occurs,
probably precursors of the NiAl2O4 spinel. Niγp and
catalyst, mixed oxides could
α+Al
O
higher than in the case of Ni catalyst. These results
α
O
confirm what has been observed by TGA and TPR.
The existence of nickel species with a higher interac-
tion degree was demonstrated when the α+Al support
was used.
O
γ
Ni catalytic systems present reduction temperatures
above 1150 K. These reduction temperatures evidence
the formation of the NiAl2O4 spinel.
The sequence of reduction temperatures measured
by TPR, Ni Oα < Ni
O
< Ni < Niγ goes in the
O
O
α+Al
γp
4. Conclusions
same direction than that of the interaction degree de-
termined by TGA and GC/MS, and indicates that the
metal-support interaction degree achieved during the
preparation step is maintained after the activation step
of catalytic systems.
The interaction among nickel acetylacetonate and
different alumina based supports is strongly dependent
on the support surface acidity, with the sequence: γ >
γp > α + Al > α. Results of the IPA reaction on
the supports, together with those obtained by TGA
and CG/MS during the calcination stage of catalysts
indicate that the aluminum oxide layer deposited on
␣-Al O by means of an impregnation with aluminum
Results obtained when the four catalytic systems
were submitted to the partial oxidation of methane
reaction are presented in Table 3. Marked differences
among different catalysts are observed with respect to
their activity and catalytic stability. Catalysts based on
2
3
␥
-Al2O3 supports presented a low activity level when
nitrate results in a support having a nature different
from the one of transition aluminas here studied.
Following the decomposition of AcacNi by TGA
during the calcination step, there is a shift of peaks
towards higher temperatures with the increase of the
interaction degree between the nickel precursor and
the support. The presence of lighter fragments was
detected by GC/MS, surely due to a major contribu-
tion of rupture reactions. It was demonstrated that
above 723 K there are no organic fragments left on
the surface, in agreement with previously published
IR results.
compared with catalysts based on ␣-Al2O3. Taking
into account that the active phase under the operating
conditions of the oxyreforming reaction is the reduced
Ni [8], low activity levels observed for Ni and NiγO
O
γp
could be assigned to the low reducibility of nickel
species in these catalysts.
Ni and NiO
O
catalysts present a good activity
α
α+Al
level but with marked differences with respect to their
catalytic stability. Taking the increase of Ni mean par-
ticle size (measured after 24 h of reaction) as a pa-
rameter of the catalyst sintering, the NiO
catalyst
α+Al

N.N. Nichio et al. / Thermochimica Acta 400 (2003) 101–107
107
The metal-support interaction degree observed dur-
References
ing preparation of catalysts is maintained after the
activation step. TPR results clearly show that the se-
[1] J.R. Rostrup-Nielsen, in: J.R. Anderson, M. Boudart (Eds.),
Catalysis: Science and Technology, vol. 5, Springer, Berlin,
O
O
quence of reduction temperatures is Niγ > Ni
>
γp
1984, p. 1.
O
O
Ni
> Ni . Reduction temperatures indicate that
α
α+Al
[2] M.A. Peña, J.P. Gómez, J.L. Garc ´ı a Fierro, Appl. Catal. A
144 (1996) 7.
[3] I. Alstrup, B.S. Clausen, C. Olsen, R.H.H. Smits, J.R.
Rostrup-Nielsen, Stud. Surf. Sci. Catal. 119 (1998) 5.
after calcination steps, nickel interacts with the sup-
O
γ
port and forms mixed oxides in the cases of Ni ,
Ni and NiO
O
supports.
γp
α+Al
[
[
4] J.R. Rostrup-Nielsen, I. Alstrup, Catal. Today 53 (1999) 311.
5] D. Trimm, Catal. Rev. Sci. Eng. 15 (1977) 155.
The reaction test employed to determine the cat-
alytic properties of studied materials demonstrates that
in Ni and Ni systems the strong metal-support in-
[6] C. Bartholomew, Catal. Rev. Sci. Eng. 24 (1982) 67.
O
γ
O
γp
[
[
7] V.A. Tsipouriari, X.E. Verykios, J. Catal. 187 (1999) 85.
8] N.N. Nichio, M.L. Casella, O.A. Ferretti, M.G. González,
C. Nicot, B. Moraweck, R. Frety, Catal. Lett. 42 (1996)
teraction offers a high resistance of the nickel reduc-
tion to a metallic state producing low activity catalysts
under the reaction conditions studied. On the other
65.
[9] S. Tang, L. Ji, J. Li, H.C. Zeng, K.L. Tan, K. Li, J. Catal.
hand, NiO
and Niα systems present good activ-
O
194 (2000) 424.
10] S. Wang, G.Q. Lu G, Appl. Catal. B 19 (1998) 267.
[11] N. Nichio, M. Casella, E. Ponzi, O. Ferretti, Stud. Surf. Sci.
Catal. 119 (1998) 723.
12] W. Wendlandt, Thermal Methods of Analysis, 3rd ed., Wiley,
α+Al
[
ity levels in the oxyreforming reaction. Mixed oxide
phases developed in the case of Ni
tant metal-support interaction generate a more stable
system concerning deactivation by sintering.
O
with an impor-
α+Al
[
[
[
[
New York, 1986.
13] A. Gervasini, J. Fenyvesi, A. Auroux, Catal. Lett. 43 (1997)
219.
14] J.R. Van Veen, P.C. Of Jong-Versloot, G.M. Van Kessel, F.J.
Fels, Thermochim. Acta 152 (1989) 359.
15] A. Jones, B.D. McNicol, Temperature Programmed Reduction
for Solid Materials Characterization, Chem. Ind. 24, Marcel
Dekker, Inc., New York, 1986.
Acknowledgements
This work has been supported by CONICET (Con-
sejo Nacional de Investigaciones Cient ´ı ficas y Técni-
cas), Argentina.
[
16] R. Molina, G. Poncelet, J. Catal. 173 (1998) 257.