Nature of Ni-Al developed phases during thermal activation in relation to the preparation techniques. Part II: Reduction and catalytic properties

Article abstract of DOI:10.1016/s0040-6031(99)00181-1

The reducibility of different catalysts based on Ni (bulk and supported) used to obtain syngas is studied in this work. Taking as test reaction the steam reforming of methane, it was possible to prove that the stability of catalytic systems is closely rel

Full text of DOI:10.1016/s0040-6031(99)00181-1

Thermochimica Acta 336 (1999) 55±60
Nature of Ni±Al developed phases during thermal activation in relation to
the preparation techniques. Part II: reduction and catalytic properties
M.A. Marturanoa,b, E.F. Agliettib, O.A. Ferrettia,b,c,*
a
 Â
Centro de Investigacion y Desarrollo en Procesos Catalõticos, CINDECA, CONICET-UNLP,
Calle 47 N8 257, C.C. 59 (1900), La Plata (Bs. As.), Argentina
 Â
Centro de Tecnologõa de Recursos Minerales y Ceramica, CETMIC, CIC-CONICET-UNLP,
b
Cno. Centenario y 506, C.C. 49 (1897) M.B. Gonnet (Bs. As.), Argentina
  Â
Departamento de Ingenierõa Quõmica, Facultad de Ingenierõa, Universidad Nacional de La Plata,
c
Calle 47 N8 257, C.C. 59 (1900), La Plata (Bs. As.), Argentina
Received 17 February 1999; accepted 15 June 1999
Abstract
The reducibility of different catalysts based on Ni (bulk and supported) used to obtain syngas is studied in this work. Taking
as test reaction the steam reforming of methane, it was possible to prove that the stability of catalytic systems is closely related
with the reduction temperature determined by TGA±DTG. Higher dif®culty in reduction of metallic phase implies higher
interaction between Ni and Al oxides; this leads to higher stability of catalytic system exposed to severe thermal treatments.
# 1999 Published by Elsevier Science B.V. All rights reserved.
Keywords: TGA±DTG; H2 reduction; Ni catalyst; Methane steam reforming
1. Introduction
Catalysts based on Ni are widely used for steam
reforming, partial oxidation or CO2 reforming
[1±4].
In Part I it was studied the evolution of dif-
ferent systems (bulk and supported catalysts)
during calcination step by using differential thermal
analysis (DTA) and thermogravimetry analysis
(TGA±DTG), complemented with X-ray diffraction
(XRD).
In this work the H2 reducibility of calcined
phases in function of preparation method is studied
by TGA±DTG. The relation between nature and redu-
cibility of active phase (Ni) with catalytic properties
(stability) is analysed, taking as reaction test the
methane steam reforming.
Due to the great demand of hydrogen, both for
petroleum re®ning processes and petrochemical
industry, its production by reforming of methane
has increased in importance. While most syngas is
produced by steam reforming of methane, some other
routes as partial oxidation or CO2 reforming of
methane may be more attractive, depending on
factors such as H2/CO ratio, environmental restric-
tions, etc.
*Corresponding author. Tel.: +54-221-4210711; fax: 54-221-
4254277
E-mail address: ferretti@dalton.quimica.unlp.edu.ar (O.A. Ferretti)
0040-6031/99/$ ± see front matter # 1999 Published by Elsevier Science B.V. All rights reserved.
PII: S 0 0 4 0 - 6 0 3 1 ( 9 9 ) 0 0 1 8 1 - 1

56
M.A. Marturano et al. / Thermochimica Acta 336 (1999) 55±60
Table 1
Nomenclature of the catalytic systems
Start solution
Method
Bulk systems
Supported systems
a-Al2O3
a-Al2O3 ‡ Al
Ni nitrate
Heat treatment
Heat treatment
Precipitation
NiEV
NiAlEV
±
±
Ni nitrate ‡ Al nitrate
Ni nitrate ‡ urea
Ni nitrate ‡ Al nitrate ‡ urea
Ni nitrate
±
±
NiCOPU
±
±
Co-precipitation
Impregnation
NiAlCOPU
±
±
±
±
±
Ni/S
NiAl/S
NiDP/S
±
Ni nitrate ‡ Al nitrate
Ni nitrate ‡ urea
Co-impregnation
Deposition±precipitation
±
NiDP/S ‡ Al
2. Experimental methods
time t and the speci®c constant of initial reaction rate
(ꢀ ˆ kt/k0, supposing reaction of ®rst order).
Catalytic systems were prepared following the
methodology presented in Part I. Table 1 summarizes
prepared catalytic systems and their nomenclature.
The Ni reducibility was studied following the
weight change observed (TGA±DTG) when pre-
viously calcined samples (calcination temperature,
1173 K) were exposed to hydrogen ¯ow. Samples
were placed in alumina crucibles and a-Al2O3 was
used as a reference. The equipment was evacuated and
then ®lled with N2 to eliminate the oxygen present.
The experiments were performed in H2 atmosphere at
a ¯ow rate of 30 cm3 min 1, between 373 and 1273 K
with a heating rate of 10 K min 1. A Netzsch STA 409
equipment, model 414/2 was used.
3. Results and discussion
3.1. Reduction step
The reduction of supported metallic catalysts is
considered of great interest because of: (i) the reaction
temperature must be known to obtain the adequate
active phase and (ii) the reduction temperature may be
related with the nature of this phase and with the inter-
actions generated with the support and/or promoters.
Fig. 1 shows the TGA±DTG diagrams of the reduc-
tion step for the NiEV solid achieved by calcination of
nickel nitrate (Ni(NO3)2Á6H2O). In these conditions
the solid phase represents a NiO bulk [5], without
interaction with other components. The NiO reduction
(NiOsolid ‡ H2 ! Nisolid ‡ H2O) occurs in a narrow
temperature range (between 473 and 773 K), with a
peak temperature at 574 K, where the maximum rate
of weight loss by reduction is produced.
Fig. 2 shows the reduction diagram for NiCOPU
system, presenting a similar behaviour to the one of
the NiEV system. This indicates that Ni precipitation
by alkalinization via urea hydrolysis and subsequent
calcination leads to formation of nickel oxide bulk.
Fig. 3 shows TGA±DTG diagrams of NiAlEV
solid. The weight loss observed is developed within
a wider temperature range than for NiEV (between
473 and 1173 K) recording two temperature peaks.
The lower temperature peak is well de®ned at 603 K,
while the higher temperature peak is at about 893 K
corresponding to a wide not well de®ned zone.
Catalytic tests and stability measures of catalysts
were performed in a conventional microreactor of
®xed bed at atmospheric pressure with the following
program:
1. Reduction in N2/H2 (20/80) ¯ow during 1 h at
1023 K.
2. Reaction test in N2/H2/CH4/H2O (55/20/5/20) flow
at 873 K (180 cm3min 1, catalyst mass ˆ 0.05 g).
3. Deactivating treatment in N2/H2/H2O (58/21/21)
flow during 2 h at 1023 K.
4. Reaction test item 2.
5. Deactivating treatment, item 3, during 5 h.
6. Reaction test item 2.
7. Deactivating treatment, item 3, during 5 h.
The stability of catalytic systems was determined
through the evolution of activity coef®cient (ꢀ) with
deactivation time. This coef®cient is de®ned as the
ratio between the speci®c constant of reaction rate at

M.A. Marturano et al. / Thermochimica Acta 336 (1999) 55±60
57
Fig. 4. TGA±DTG diagrams for NiAlCOPU samples. Initial
Fig. 1. TGA±DTG diagrams for NiEV samples. Initial
mass ˆ 73 mg.
mass ˆ 71 mg.
The shifting of the reduction zone to higher tem-
peratures observed for NiAlEV and NiAlCOPU sys-
tems with respect to NiO is in agreement with
conclusions of Part I. The existence of a speci®c
interaction between Ni and Al, after calcination step,
leads to formation of mixed oxide structures strongly
interactive. The presence of two reduction zones may
be explained by the existence of two phases: (i) a
mixed oxide, NiAl2O4 type (the reduction temperature
of NiAl2O4 is higher than 1000 K [6,7]); (ii) a NiO
interacting with the previous phase (its temperature
peak is higher than the one of the NiO bulk). This
result is in agreement with the stoichiometric ratio
used Al/Ni ˆ 0.5, where the Ni-excess respect to
the amount required for spinel (NiAl2O4) is found
as NiO.
Fig. 2. TGA±DTG diagrams for NiCOPU samples. Initial
mass ˆ 72 mg.
Fig. 5 shows TGA±DTG diagrams for the sup-
ported Ni/S system. This catalyst presents a behaviour
similar to the one of NiO bulk although the tempera-
ture range (ÁT) is more prolonged (473±973 K) and
the peak temperature is at 613 K shifted in 40 K with
respect to the one of NiO. These results would indicate
Fig. 3. TGA±DTG diagrams for NiAlEV samples. Initial
mass ˆ 71 mg.
The TGA±DTG diagrams for the NiAlCOPU solid
(Fig. 4) show a similar behaviour to the one of the
NiAlEV system with a reduction shift towards higher
temperatures. The temperature range is between 673
and 1173 K. Two temperature peaks are observed at
893 and 1053 K.
Fig. 5. TGA±DTG diagrams for Ni/S samples. Initial
mass ˆ 300 mg.

58
M.A. Marturano et al. / Thermochimica Acta 336 (1999) 55±60
Fig. 6. TGA±DTG diagrams for NiAl/S samples. Initial
Fig. 8. TGA±DTG diagrams for NiDP/S ‡ Al samples. Initial
mass ˆ 300 mg.
mass ˆ 300 mg.
that the Ni phase interacts with the support, although
the interaction would be slightly weak [8,9].
in a wide temperature range (ÁT ˆ 673±1173 K).
These diagrams are similar to the ones obtained for
the NiAlCOPU solid bulk, in agreement with results of
Part I, where it was stated that active phases, obtained
on modi®ed supports (NiDP/S ‡ Al), are similar with
the ones achieved by co-precipitation (NiAlCOPU). In
both solids, the interaction between Ni and Al surely
occurs. With respect to NiAl/S systems (co-impreg-
nation), the shifting of the reduction temperature
towards higher values would indicate a more ef®cient
interaction between Ni and Al, for NiDP/S ‡ Al
catalyst.
Through weight loss measured (Áw), it is possible
to estimate the reduction degree reached. In all cases
this value is compatible with a reduction degree of
NiO near 100%. Table 2 shows the coincidence exist-
ing between the weight change achieved by TGA
(Áwexp) and the theoretical one estimated for total
reduction from NiO composition determined by
atomic absorption (Áwtheor).
For the case of NiAl/S co-impregnated systems
(Fig. 6), the behaviour is similar to the one of the
NiAl bulk solids (ÁT ˆ 573±1173 K; two tempera-
ture peaks at 793 and 923 K). The peak at 923 K is
assigned to mixed oxide phase of NiAl2O4 type. The
peak at 793 K may be interpreted as the one corre-
sponding to the reduction of a NiO phase presenting a
high interaction degree with the mentioned mixed
oxide or with the support.
The systems prepared for deposition±precipitation
showed different behaviours. Two reduction peaks
appear in Fig. 7 for the catalyst prepared on a-
Al2O3 without modi®cation (NiDP/S); the ®rst peak
at the same temperature of NiO bulk (573 K), indi-
cative of a phase without interaction with the support,
while the second peak around 860 K corresponds to a
wide reduction zone (ÁT ˆ 573±1173 K). This last
peak evidences that even in the case of support without
modi®cation, a part of deposited Ni by this method
interacts with the support.
3.2. Stability tests
TGA±DTG diagrams for the NiDP/S ‡ Al system
are shown in Fig. 8. In this case, the reduction occurs
It is important to take into account the information
from the use of these thermal activation techniques to
characterize solids and predict their catalytic beha-
Table 2
Weight change determined by TGA and calculated by atomic
absorption
Catalyst
Ni (%)
Áwexp (mg)
Áwth (mg)
Ni/S
NiAl/S
8.1
5.9
2.6
1.3
6.4
4.5
1.8
1.1
6.6
4.8
2.1
1.1
NiDP/S
NiDP/S ‡ Al
Fig. 7. TGA±DTG diagrams for NiDP/S samples. Initial
mass ˆ 300 mg.

M.A. Marturano et al. / Thermochimica Acta 336 (1999) 55±60
59
viour. These catalysts present good levels of initial
activity in steam, CO2 or air reforming of methane to
synthesis gas [10±12]. One of the main problems is the
deactivation by sintering, mechanism favoured by
high temperatures and by the presence of steam.
For this reason, steam reforming of methane is a very
good stability test.
With the conditions used, NiEV and NiCOPU sys-
tems (pure NiO) are completely deactivated after 7 h
deactivating treatment. Instead, NiAlEV and NiAl-
COPU systems are much more stable and after 7 h
deactivating treatment, they maintain approximately
more than 80% initial activity (ꢁꢀ0.80). In these last
systems, due to the existence of structures where Ni
and Al are intimately related, aluminium oxide phases
disperse crystallites of the metallic nickel (active
phase) acting as barriers inhibiting Ni sintering. This
effect favours notably the thermal stability of the
system. Although intermediate stages depend on pre-
paration method used (hydrotalcites, nitrates, see Part
I), solids obtained at the end of the calcination stage
present properties and tendencies qualitatively simi-
lar: aluminate phases (XRD), comparable reduction
zones (TGA±DTG), stability with respect to sintering.
Figs. 9 and 10 show catalytic results obtained for
supported catalysts prepared by impregnation and
deposition±precipitation, respectively. For both pre-
paration methods, the Al presence (co-impregnation
or deposition±precipitation on modi®ed supports) is
essential to generate solids with higher stability. As it
was stated for bulk catalysts, the Al presence generates
barriers to the sintering mechanism of nickel phase,
which improves notably the stability of these catalytic
Fig. 10. Activity coefficient evolution (ꢀ) vs. deactivation time (t)
for NiDP/S and NiDP/S ‡ Al catalysts.
systems. The greatest dif®culty in the reduction of
nickel in NiDP/(S ‡ Al) solid with respect to NiAl/S
(see Figs. 6 and 8) may be explained by the higher
interaction between Ni and Al in the system prepared
by deposition±precipitation. This fact is in agreement
with the higher relative stability observed in curves
represented in Figs. 9 and 10. This interpretation is
also valid in the supported systems on a-Al2O3 with-
out modi®cation since NiDP/S catalyst is more stable
than Ni/S catalyst and this result is coherent with
TGA±DTG diagrams shown in Figs. 5 and 7.
4. Conclusions
The use of TGA±DTG permits to study the H2
reducibility of different catalysts based on Ni (bulk
and supported), as a function of the different prepara-
tion techniques used.
Those preparations that lead to a strong interaction
between Ni and Al (co-precipitation, co-impregnation,
deposition±precipitation), generate Ni phases more
dif®cult to be reduced. This is produced by the for-
mation of mixed oxides (spinel type non-stoichio-
metric) when a Al/Ni ˆ 0.5 atomic ratio is used.
The catalytic stability of these systems, in the
methane steam reforming process, is related to the
catalyst reducibility. When the Ni reduction is more
dif®cult to be performed, there exists higher interac-
tion between Ni and Al and higher thermal stability.
It was possible to ®x Ni via alkalinization with the
urea on a-Al2O3 supports modi®ed by an aluminium
oxide layer. This method led to the formation of Ni
Fig. 9. Activity coefficient evolution (ꢀ) vs. deactivation time (t)
for Ni/S and NiAl/S catalysts.

60
M.A. Marturano et al. / Thermochimica Acta 336 (1999) 55±60
[2] C.H. Bartholomew, Chem. Eng. (1982) 96.
[3] M.A. PenÄa, J.P. Gomez, G.L.G. Fierro, Appl. Catal. A 144
(1996) 7.
supported phases more dif®cult to be reduced and with
higher thermal stability.
[4] N. Nichio, M. Casella, O. Ferretti, M. Gonzalez, C. Nicot, B.
Moraweck, R. Frety, Catal. Lett. 42 (1996) 65.
[5] A. Jones, B.D. McNicol, Temperature-programmed reduction
for solid materials characterization, Chem. Ind. 24 (1986).
[6] C.H. Bartholomew, R.J. Farrauto, J. Catal. 45 (1976) 41.
[7] O.A. Ferretti, H. Mendiondo, J. Cendra, N. Firpo, R. Brizzi,
Procc. XII Symp. Ib. Am. Catal. 3 (1990) 62.
[8] J.T. Richardson, M.V. Twigg, Appl. Catal. A 167 (1998) 57.
[9] J.R.H. Ross, M.L.F. Steel, A. Zein-Isfahani, J. Catal. 52
(1978) 280.
Acknowledgements
This work was sponsoned by the Consejo Nacional
Â
Â
de Investigaciones Cientõ®cas y Tecnicas (CONI-
CET), Argentina. We thank Lic. S. Conconi for the
performing of the thermal analysis.
[10] M.A. Marturano, Ph.D. Thesis, Univ. Nac. La Plata,
Argentina, 1998.
References
[11] N.N. Nichio, M.L. Casella, E.N. Ponzi, O.A. Ferretti, Stud.
Surf. Sci. Catal. 119 (1998) 723.
[1] J.R. Rostrup Nielsen, Steam Reforming Catalyst, Danish
Technical Press, 1975.
[12] N.N. Nichio, M.L. Casella, O.A. Ferretti, React. Kin. Catal.
Lett. 66 (1999) 27.