Structure of the active component and catalytic properties of catalysts prepared by the reduction of layered nickel aluminosilicates

Article abstract of DOI:10.1134/S002315840603013X

The reduction of Ni-Mg aluminosilicates with the amesite structure was studied using thermogravimetry, high-resolution electron microscopy, XPS, and XRD. It was found that the reduction with hydrogen at 920 K resulted in the formation of nickel particles coated with a difficult-to-reduce amorphous oxide shell. The reduced samples were incapable of chemisorbing oxygen; however, they exhibited a high adsorption capacity for hydrogen. The Ni⁰ core-oxide shell decorated particles were highly active in steam methane reforming and CO hydrogenation reactions. At the same time, they were inactive in the formation of graphite-like carbon in both methane decomposition and CO disproportionation. MAIK Nauka/Interperiodica , 2006.

Full text of DOI:10.1134/S002315840603013X

ISSN 0023-1584, Kinetics and Catalysis, 2006, Vol. 47, No. 3, pp. 412–422. © MAIK “Nauka /Interperiodica” (Russia), 2006.
Original Russian Text © A.A. Khasin, T.M. Yur’eva, V.V. Kaichev, V.I. Zaikovskii, M.P. Demeshkina, T.P. Minyukova, N.A. Baronskaya, V.I. Bukhtiyarov, V.N. Parmon, 2006,
published in Kinetika i Kataliz, 2006, Vol. 47, No. 3, pp. 420–430.
Structure of the Active Component
and Catalytic Properties of Catalysts Prepared
by the Reduction of Layered Nickel Aluminosilicates
A. A. Khasin, T. M. Yur’eva, V. V. Kaichev, V. I. Zaikovskii, M. P. Demeshkina,
T. P. Minyukova, N. A. Baronskaya, V. I. Bukhtiyarov, and V. N. Parmon
Boreskov Institute of Catalysis, Siberian Division, Russian Academy of Sciences, Novosibirsk, 630090 Russia
Received November 21, 2004
Abstract—The reduction of Ni–Mg aluminosilicates with the amesite structure was studied using thermo-
gravimetry, high-resolution electron microscopy, XPS, and XRD. It was found that the reduction with hydrogen
at 920 K resulted in the formation of nickel particles coated with a difficult-to-reduce amorphous oxide shell.
The reduced samples were incapable of chemisorbing oxygen; however, they exhibited a high adsorption capac-
ity for hydrogen. The Ni0 core–oxide shell decorated particles were highly active in steam methane reforming
and CO hydrogenation reactions. At the same time, they were inactive in the formation of graphite-like carbon
in both methane decomposition and CO disproportionation.
DOI: 10.1134/S002315840603013X
INTRODUCTION
EXPERIMENTAL
The nature of the support and the history of the sam-
ple have a great effect on the structure and properties of
catalysts containing Group VIII metal particles. Com-
plex oxide systems containing Group VIII metal cat-
ions, which are reduced upon thermal treatment in
hydrogen with the formation of metal particles, are
interesting materials. In these systems, the formation of
highly dispersed metal particles, which strongly inter-
act with the support and exhibit unique catalytic prop-
erties, would be expected. Thus, the reduction of lay-
ered nickel–magnesium aluminosilicates with the
structures of amesite and its polymorph (chlorite) is
accompanied by the formation of dispersed nickel par-
ticles immobilized on a layered aluminosilicate sub-
strate [1]. A special feature of these particles is their
low activity in the formation of graphite-like carbon in
methane and carbon monoxide atmospheres [2]. This
work was devoted of the structure and chemical compo-
sition of the surface and the catalytic properties of
metal nickel particles formed upon the reduction of
Ni−Mg−Al aluminosilicates with the amesite structure.
Thermogravimetry, high-resolution electron micros-
copy, X-ray photoelectron spectroscopy (XPS), X-ray
diffraction (XRD), and adsorption techniques for sur-
face characterization were used in this study. Catalytic
properties were studied in the reactions of steam meth-
ane reforming, CO hydrogenation, methane formation
and thermal decomposition, and CO disproportion-
ation.
Sample Preparation and Characterization
Three samples of the Ni–Mg–amesite system with
different nickel contents (hereafter, NAS1, NAS2, and
NAS3) were prepared. For comparison, a Ni/MgO cat-
alyst was also used. It was prepared by the coprecipita-
tion of Ni2+ and Mg2+ cations from nitrate salt solutions
with a mixture of NaOH and Na2CO3 solutions at pH
10.5 and T = 343 K. The precipitate was washed with
distilled water, dried under an IR lamp, and calcined at
725–775 K in a flow of argon. To prepare Ni–Mg–
amesite samples, SiO2 powder was mixed with an aque-
ous aluminum nitrate solution; an aqueous ammonia
solution was added to pH 8, and the mixture was stirred
for 30 min. The resulting gel was filtered and washed
with distilled water. Magnesium hydroxide was pre-
pared by the precipitation of Mg2+ cations from a mag-
nesium nitrate solution with an aqueous KOH solution
at pH 10.5–11.0 and 298–303 K. The precipitate was
filtered off and washed with distilled water. The freshly
prepared Mg(OH)2 and nickel acetate were added to a
silicon–aluminum-containing gel, and the resulting
paste was stirred for 1 h until a homogeneous state was
achieved. The resulting paste was steam heated at
575 K and 9 MPa for 4 h and then filtered off and
washed with distilled water. Table 1 summarizes the
composition of samples based on data obtained by
atomic absorption spectrometry (AAS) and tempera-
ture-programmed reduction and oxidation (TPR–TPO)
with the measurement of weight changes using a
412

STRUCTURE OF THE ACTIVE COMPONENT AND CATALYTIC PROPERTIES
413
Table 1. Composition of the materials examined according to AAS, TPR–TPO, and thermogravimetric data [1]
Cationic composition (relative atomic content)
Phase content, wt %
Sample
Ni
Mg
Al
Si
NiO
Ni0*
NAS1
NAS2
NAS3
Ni/MgO
0.30
0.54
1.10
1
1.90
1.46
1.00
1
1.6
2.0
1.8
–
1.2
1.0
1.1
–
7.7
13.8
26.1
65
6.1
11
21
24
*In reduced samples (determined from thermogravimetric data for the oxidation of the sample with atmospheric oxygen); the reduction
temperature was 825 and 925 K for Ni/MgO and NAS samples, respectively.
Netzsch STA-409 instrument for thermogravimetric the elements was no higher than 10%. For an increase
analysis.
in the depth of analysis, layer-by-layer ion etching was
used. The ion etching of samples was performed with
the use of a VG AG-21 argon-ion gun with a beam
energy of 2 keV; the rate of etching was 1–2 nm/min at
a current of ~20 µA [1].
The specific surface area of nickel metal in reduced
samples was determined from the temperature-pro-
grammed desorption (TPD) of hydrogen in a flow of
argon (1.8 l/h NTP) on heating to 673 K at a rate of
10 K/min. Before desorption, the adsorption of hydro-
gen onto a 50- to 100-mg sample was performed at
373 K in a flow of pure hydrogen for 12 h. The average
surface area occupied by a nickel metal atom was taken
equal to 6.5 × 10–2 nm2 [3]. The adsorption capacity for
oxygen was studied by means of N2O decomposition at
325 K under conditions of pulsed N2O supply and chro-
matographic analysis of decomposition products. Anal-
ogous measurements at 625 K allowed us to perform
the titration of total nickel metal. The amount of Ni0
was also determined by treatment in a flow of air at
675 K and weight measurements with a Netzsch
STA-409 thermal analyzer.
The electron-microscopic study of samples was per-
formed using a JEOL JEM-2010 transmission electron
microscope (resolution, 1.4 Å; accelerating voltage,
200 kV). The samples were applied onto perforated car-
bon substrates from alcohol suspensions, which were
ultrasonically dispersed.
Catalytic Tests
The catalytic activity of samples in the reaction of
steam methane reforming was determined in a flow-cir-
culation fixed-bed reactor with the circulation of a gas
mixture at a circulation ratio of 100. The fed mixture
contained 33 vol % CH4, 65 vol % H2O, and 2 vol % N2
as an internal standard. The tests were performed at
atmospheric pressure over a temperature range from
823 to 1023 K. A 0.3-g catalyst sample with a particle
size of 0.25–0.50 mm was used.
Phase analysis was performed on Siemens D-500
and Bruker D-8 X-ray diffractometers with the use of a
CuKα-radiation source (β lines were removed using a
graphite monochromator).
The chemical composition of the subsurface layers
of catalysts was analyzed by XPS. The XPS spectra
were measured on a VG ESCALAB HP photoelectron
spectrometer using AlKα radiation (hν = 1486.6 eV).
Samples were loaded into the vacuum chamber of the
spectrometer in a flow of argon in order to prevent con-
tact with air. The scale of binding energies (Eb) was
precalibrated using the positions of Au4f7/2 (84.00 eV)
and Cu2p3/2 (932.67 eV) lines in the spectra of cleaned
gold and copper foils, respectively. The charging effect
that appeared in the course of photoemission was taken
into consideration using the internal standard tech-
The catalytic activity of samples in CO hydrogena-
tion (Fischer–Tropsch synthesis) was determined in a
fixed-bed tube reactor. The feed had the composition
H2/CO/N2 = 6 : 3 : 1 (by volume). The tests were per-
formed at 483 and 503 K. A 1-g catalyst sample with a
particle size of 0.25–0.50 mm was used. The samples
were reduced in hydrogen at 1025 K immediately
before the catalytic tests.
The catalytic activity in the reaction of methane
nique; the C1s line of carbon (Eb = 284.8 eV) was used decomposition was studied by thermogravimetry both
as an internal standard. The relative concentrations of under temperature-programmed conditions and at a
the elements in an analytical zone (the depth of analysis constant temperature. The procedure used for the deter-
was 2–4 nm) were determined from the integrated mination of catalytic activity in methane decomposi-
intensities of photoelectron lines normalized to the cor- tion was described elsewhere [2]. The CO dispropor-
responding atomic sensitivity factors [4]. The relative tionation reaction was also studied under temperature-
determination error for the relative concentrations of programmed conditions using the Netzsch STA-409
KINETICS AND CATALYSIS Vol. 47 No. 3 2006

414
KHASIN et al.
Table 2. Data on the particle size, degree of dispersion, and specific surface area of metal nickel in NAS samples and a model
Ni/MgO sample obtained using XRD, the thermal desorption of hydrogen, and the decomposition of N2O
NH ×105,
mol/g
NO × 105,
Sample Ni0 content,a wt % dNb i , nm
D
b, c, %
Dd, %
De, %
Oads/Hads SNi, m2/g
2
mol/g
Ni/MgO
NAS1
24
6.1
29
15
11
12
4.2
8.2
12
3.7
12
13
5.8
7.1
27
3.30
0.68
0.42
0.43
1.13
0.19
0.07
0.12
7.1b, d, e
2.9d
1.4
1.6
3.1
NAS2
11
21
10.9
9.5
12.7
7.3
9.4d
NAS3
10.2d
a
From data on the titration of samples reduced at 825 (Ni/MgO) or 925 K (NAS) with atmospheric oxygen (thermogravimetry).
Sizes of coherent scattering domains evaluated from the integrated XRD line widths.
b
c
d
e
Under the assumption that the particles are cuboctahedral in shape, according to Barbier et al. [6].
From data on the thermal desorption of H assuming that θ = 1.
2
H
From data on the adsorption of O from N O assuming that θ = 2.
2
O
instrument for thermogravimetric analysis. The sample tion capacities of samples for hydrogen and oxygen,
weight was ~100 mg. The temperature was increased which were experimentally determined using the ther-
from 300 to 1100 K and then decreased at a rate of mal desorption of hydrogen and O adsorption upon
4 K/min.
N2O decomposition, respectively.
The degree of dispersion was estimated from the
particle size of Ni0 under the assumption that the metal
particles are cuboctahedral in shape [6]:
RESULTS AND DISCUSSION
Active Component Particle Size and Adsorption
Capacity of Catalysts for Hydrogen and Oxygen
D = Nsurf/Ntotal
,
Nsurf = 12 + 10m(m – 2),
Ntotal = 3.33(m – 1)3 + 5(m – 1)2 + 3.67(m – 1) + 1,
Table 2 summarizes data on the concentrations of
nickel metal and the degree of dispersion (D) of a nickel
metal phase in the test samples (NAS series) and a
model reference sample (Ni/MgO). The structure of the
samples was studied in detail previously [1, 5]. X-ray
diffraction data for the Ni/MgO sample indicated that
the sample before reduction was a solid solution of
nickel cations in the magnesium oxide structure. The
cubic unit cell parameter a = 4.198 Å was significantly
smaller than that in the pure MgO oxide (4.210 Å).
According to thermogravimetric and in situ XRD data,
the reduction of nickel cations occurred at temperatures
higher than 625 K and the degree of reduction essen-
tially depended on temperature. Thus, 65% nickel
remained in the Ni2+ state after reduction at 725 K. The
degree of reduction reached ~45% after reduction at
825 K; this was accompanied by an increase in the par-
ticle size of Ni0 from 8 to 29 nm (as estimated based on
XRD line broadening). According to XRD and IR-
spectroscopic data, NAS samples were nickel metal
particles localized on the surface of a layered Ni–Mg
aluminosilicate with the magnesium chlorite–vermicu-
lite structure. The degree of nickel reduction at a reduc-
tion temperature of 923 K was close to 100%.
1/3
total
d = 0.27N
.
Here, Nsurf is the number of nickel atoms on the surface
of a particle, Ntotal is the total number of nickel atoms in
a particle, m is the number of atoms at the edge of a cub-
octahedron, and d is the particle diameter (nm).
In the evaluation of degree of dispersion from data
on the thermal desorption of hydrogen, we assumed
that the coverage of nickel metal with hydrogen (θH)
was close to 1 (according to Christmann et al. [7], the
degree of saturation of a nickel metal surface is θH =
(0.8–0.9) 20%). In the evaluation of the specific sur-
face area of nickel metal, we assumed that an atom at
the surface of nickel metal occupied a surface area of
6.5 × 10–2 nm2 [3].
The degree of dispersion estimated from XRD and
hydrogen desorption data were consistent with each
other for all of the samples. It is likely that the average
particle size of nickel found from XRD data was some-
what overestimated because highly dispersed nickel,
whose contribution to diffraction line widths was insig-
nificant, was ignored. At the same time, highly dis-
persed nickel particles make a considerable contribu-
tion to the total surface area of the metal.
Table 2 summarizes data on the degree of dispersion
of nickel. These data were obtained based on the sizes
of coherent scattering domains evaluated from the inte-
In the Ni/MgO sample, the surface area of nickel
grated XRD line widths. Table 2 also gives the degree metal estimated from the adsorption capacity for oxy-
of dispersion of Ni0 evaluated from data on the adsorp- gen is also consistent with data on the thermal desorp-
KINETICS AND CATALYSIS Vol. 47 No. 3 2006

STRUCTURE OF THE ACTIVE COMPONENT AND CATALYTIC PROPERTIES
415
Table 3. Relative atomic concentrations of cations on the surface of sample NAS2, as found from XPS data
Treatment conditions
[Mg]/[Si]
[Ni]/[Si]
[Al]/[Si]
[O]/[Si]
Initial
1.13
0.60
0.42
0.54
0.52
0.54
0.50
0.20
0.06
0.15
0.06
0.10
1.7
1.8
1.5
1.8
1.6
1.5
7.8
5.5
4.2
5.0
4.9
5.0
775 K, Ar
925 K, H2
925 K, H2, etching with Ar+ for 5 min
925 K, H2/CO/H2O, 2 MPa
925 K, H2/CO/H2O, 2 MPa, etching with Ar+ for 5 min
tion of H2 and XRD data. The estimation was per- composition of sample NAS2 after various types of
treatment. Figure 1 shows the Ni2p XPS spectra of the
formed based on the assumption that the ratio of the
amount of adsorbed oxygen to the amount of surface
nickel metal atoms is θO = O/Nis = 2 (see publications
concerning oxygen adsorption on clean nickel metal
surfaces [8] and supported nickel-containing catalysts
[9, 10]). Note that data obtained by Brennan et al. [8]
refer to the partial pressures of O2 lower than 13 Pa,
whereas the sample examined in our study was exposed
to pulses of N2O at 0.1 MPa. In this case, the partial
pressure of oxygen in decomposition products at 300–
325 K was lower than 200 Pa (the sensitivity level of
our analysis for oxygen).
test samples.
The position of the Ni2p3/2 line at 855.4–856.2 eV
and the presence of pronounced satellites 6 eV higher
than the main photoemission lines clearly indicate that
most of the nickel occurred in the Ni2+ state on the sam-
Intensity
855.4 856.2
Ni 2p
At the same time, the capacity of NAS samples to
adsorb oxygen is very low. The decrease in the capacity
to adsorb oxygen can be explained by the partial deco-
ration of the surface of nickel metal particles with sili-
con oxide. Thus, for example, Delmon [11] supposed
that nickel metal particles prepared by the reduction of
a layered nickel silicate were partially immersed in a
SiO2 substrate. Khassin et al. [12] used high-resolution
electron microscopy in order to observe complete sur-
face coverage of Co0 particles with silicon oxide in
Co/SiO2 samples prepared by the reduction of a layered
cobalt silicate with the stevensite structure at 1025 K. It
is likely that the decoration of Ni0 particles in NAS
samples is 90–95% of the metal surface. However, to
explain data on hydrogen adsorption, one has to assume
that the oxide shell can activate and chemisorb hydro-
gen. Moreover, the chemisorption of hydrogen on the
surface oxide layer is reversible.
852.8
1
2
×3
×3
3
4
5
Chemical Surface Composition of Silicon-Containing
Catalysts Prepared by the Reduction of Ni–Mg
Amesites (According to XPS Data)
850 855 860 865 870 875 880 885
Binding energy, eV
Fig. 1. Ni 2p XPS spectra of samples NAS2: (1) initial,
(2) calcined in Ar at 925 K, (3) calcined in Ar at 925 K and
The problem of the decoration of metal nickel parti-
cles can be clarified based on data on chemical surface
compositions and high-resolution electron-microscopic
data. Table 3 summarizes data on the chemical surface
reduced in H at 925 K, (4) reduced at 925 K and treated in
2
a mixture of CO/H /H O = 1: 2 : 2 at 2 MPa and 925 K, and
2
2
+
(5) sample 4 after etching with Ar for 5 min.
KINETICS AND CATALYSIS Vol. 47 No. 3 2006

416
KHASIN et al.
50 nm
5 nm
Ä
B
C
[110]
[020]
(‡)
(b)
Fig. 2. Electron micrographs of sample NAS2 after calcination in flowing argon at 925 K followed by reduction in H at 925 K.
2
Inset: The result of the Fourier filtration of the image of particle B.
ple surfaces [13–18] both before and after sample cal-
cination in flowing argon. Moreover, unexpectedly,
nickel almost completely occurred in the Ni2+ state on
the sample surface even after reduction in flowing
hydrogen (Fig. 1, curve 3). Only a weak shoulder
observed at ~853 eV may be indicative of the presence
of nickel metal on the surface or in subsurface layers.
At the same time, a dramatic decrease in the Ni/Si
atomic ratio upon a reductive treatment can be ascribed
to the reduction of Ni2+ and the formation of metal par-
ticles (Table 3). The surface etching of a reduced sam-
ple with a beam of Ar+ for 5 min resulted in the disap-
pearance of lines due to an oxidized state of nickel and
in the appearance of an intense Ni 2p3/2 peak due to Ni0
at 852.8 eV [13–19]. The etching was accompanied by
a considerable increase in the surface concentration of
nickel (see the Ni/Si ratio in Table 3). Thus, it is
believed that, as expected, the reduced sample con-
tained nickel metal particles, however, decorated with a
Ni2+ oxide film or particles with a characteristic thick-
ness of at least 2 nm (the depth of XPS analysis is
~3−4 nm). Taking into account the rate of surface etch-
ing (1–2 nm/min), the average particle size of nickel
metal was greater than 10 nm. However, note that the
detection of nickel metal after surface etching with a
beam of Ar+ can also be a consequence of the reduction
of Ni2+ with a high-energy beam of argon cations [15].
The probability of this oxide reduction depends on the
M–O bond energy, and it is high in the case that the
standard Gibbs function of oxide formation is no higher
than 250 kJ/mol on a metal basis. Kim and Winograd
[15] demonstrated that the reduction of a NiO surface
with a beam of Ar+ is possible (∆fG0 ~ 217 kJ/mol),
whereas the reduction of a Ni(OH)2 surface (∆fG0 ~
445 kJ/mol) does not occur.
High-Resolution Electron-Microscopic Data
on the Structure of Reduced Ni–Mg Amesite
Figure 2 shows the electron micrographs of sample
NAS2 after calcination in a flow of argon at 925 K fol-
lowed by reduction in H2 at 925 K. The sample thus
treated exhibited metal nickel particles distributed over
the surface of a layered aluminosilicate. The electron-
microscopic data provide support for our previous con-
clusions [1] that the structure of a layered aluminosili-
cate remained unaffected in the course of calcination
and reduction. In this case, the particle size of Ni0 var-
ied from 4 to 8 nm, which is somewhat smaller than
sizes estimated from XRD data and the thermal desorp-
tion of hydrogen. High-resolution microscopy (Fig. 2b)
demonstrates that well-crystallized Ni0 particles
(labeled B) are coated with an amorphous oxide layer
(A) ~2 nm in thickness and the layered aluminosilicate
support is well-crystallized (C). Interplanar distances in
the test materials (0.20 and 0.45 nm for particle B and
support C, respectively) are in good agreement with
data on Ni0 (distance between the [111] planes of the
face-centered structure of nickel metal) and chlorite–
vermiculite structures (distance between the [020] and
[110] planes, which form a pseudohexagonal structure,
of the triclinic structure in accordance with JCPDS file
number 34-0163). We failed to detect atomic ordering
in the oxide layer (A) of a particle shown in Fig. 2, as
well as in other oxide shells of this kind. Note that, in a
recent study of analogous nickel metal core–oxide shell
particles in finely dispersed nickel powder prepared by
pulsed laser ablation followed by forced oxidation,
Sakiyama et al. [19] found a layer of amorphous nickel
oxide that coated a single-crystalline metal core.
KINETICS AND CATALYSIS Vol. 47 No. 3 2006

STRUCTURE OF THE ACTIVE COMPONENT AND CATALYTIC PROPERTIES
417
Table 4. Activity of the samples examined in steam methane reforming
Catalyst
T = 820 K
T = 920 K
T = 1020 K
Ni/MgO
NAS1
NAS2
NAS3
0.13
0.09
–
0.10
0.13
–
1.07
1.45
–
0.25
0.20
–
0.32
0.53
–
3.53
5.72
–
0.36
0.31
0.31
0.36
1.23
1.72
0.68
0.64
13.3
19.0
88
90
–
7.44
7.0
0.11
0.05
0.55
0.24
0.19
2.02
75
Note: Total pressure of 0.1 MPa; molar composition of the vapor–gas mixture, CH /H O = 0.5; w is the apparent specific (per gram
SMR
4
2
0
of catalyst) rate of methane conversion; k is the specific (per square meter of Ni surface) rate constant of the reaction; TOF
is the turnover frequency of a single surface site (see the text).
SMR
Structure and Surface Composition of Reduced
uents with the metal core of the particle. It is also
Samples Prepared by the Reduction of Ni–Mg Amesite believed that, although we found the decoration of
metal particles in NAS samples, these samples exhibit
The above set of data obtained using XPS, electron
catalytic activity in reactions that require the activation
microscopy, and the adsorption of oxygen from N2O
of hydrogen and, probably, hydrocarbons.
suggests the occurrence of a thick oxide shell around
nickel metal particles. As follows from electron-micro-
scopic and XPS data, the thickness of this shell can be
Activity of Samples Prepared by the Reduction
~2 nm. Indeed, in this case, reduced NAS samples
of Ni–Mg Amesite in the Reactions of Steam Methane
should not chemisorb oxygen on the surface, which is
Reforming and CO Hydrogenation
saturated with oxygen.
The nature of the assumed oxide shell is of impor-
tance. This shell is amorphous; evidently, it cannot be
simply pure NiO, which was observed previously in the
case of nucleus–shell particles [19]. Indeed, the sample
reduction temperature used in our experiments was suf-
ficiently high (925 K) for the reduction of any NiOx
oxide film. Thus, it is believed that, in addition to Ni2+
cations, Si4+ and, probably, Al3+ impurity cations
occurred as the constituents of the amorphous oxide
shell. This hypothesis was indirectly supported by XPS
data obtained in the course of the layer-by-layer analy-
sis of NAS samples. The Mg/Si and Al/Si atomic ratios
increased after sample etching with an argon ion beam
(Table 3). As mentioned above, the possibility of deco-
rating metal particles prepared by the reduction of lay-
ered silicates was repeatedly considered in the literature
[11, 12]. Indirect evidence for the partial decoration of
Co0 nanoparticles, which were prepared by the reduc-
tion of Co–Al hydrotalcites, with AlOx oxide clusters
was reported previously [20, 21].
Table 4 summarizes the results of the catalytic tests of
samples prepared by the reduction of Ni–Mg amesite at
920 K in steam methane reforming at 825–1025 K, a
TOFSMR, s–1
Ea = 90 kJ/mol
1
2
3
4
5
101
Ea = 110 kJ/mol
100
Ea = 75 kJ/mol
0.13
0.12
0.14
0.15
0.16
(RT)–1 × 103, mol/J
At the same time, note that reduced NAS samples
exhibited unexpected properties. It is likely that these
samples did not have a free metal surface. Nevertheless,
these samples were capable of activating and reversibly
chemisorbing hydrogen. This can be indicative of the
strong interaction of Ni2+ cations as oxide shell constit-
Fig. 3. The temperature dependence of the turnover fre-
quency of an active site in the reaction of steam methane
reforming inArrhenius coordinates: (1) Ni/MgO, (2) NAS1,
(3) NAS2, (4) NAS3, and (5) published data [23].
KINETICS AND CATALYSIS Vol. 47 No. 3 2006

418
KHASIN et al.
Table 5. Comparison between the activities of the catalysts in CO hydrogenation
wFT,
wFT,
TOFFT × 10–3, s–1
α
TOFFT × 10–3, s–1
T = 503 K
α
mmol (g Cat)–1 h–1
mmol (g Cat)–1 h–1
Catalyst
Ea, kJ/mol
T = 483 K
Ni/MgO*
NAS2
–
1.60
1.05
1.20
0.50
0.60
0.54
–
2.70
2.80
3.30
0.38
0.44
0.40
70
100
100
0.9
1.1
2.4
3.1
NAS3
Note: 0.1 MPa; H /CO = 2 : 1; CO conversion of 10–20%.
2
w
is the apparent specific rate of CO conversion in the Fischer–Tropsch synthesis; TOF is the turnover frequency of a surface
FT
FT
site; α is the parameter of theAnderson–Schulz–Flory product distribution by number of carbon atoms; E is the estimated activation
a
energy.
*Quoted from [24].
total pressure of 0.1 MPa, and the molar ratio atom), we used the specific surface areas of nickel
CH4/H2O = 0.5. The specific rate constants of the reac- metal particles calculated from data on the thermal des-
tion were calculated assuming a first-order reaction orption of hydrogen. Figure 3 shows the temperature
with respect to methane. For the evaluation of the turn- dependence of the TOF of an active site (more specifi-
over frequency (TOF) of a single catalytic site (nickel cally, a hydrogen adsorption site) in steam methane
200 nm
10 nm
(‡)
(b)
200 nm
40 nm
(c)
(d)
Fig. 4. Electron micrographs of (a, b) Ni/MgO and (c, d) NAS3 samples after methane decomposition at 790 K.
KINETICS AND CATALYSIS Vol. 47 No. 3 2006

STRUCTURE OF THE ACTIVE COMPONENT AND CATALYTIC PROPERTIES
419
40 nm
(b)
10 nm
(‡)
Fig. 5. Electron micrographs of carbon fibers formed in the course of methane decomposition (a) on NAS3 at 790 K and (b) on
NAS2 at 1070 K.
10 nm
10 nm
(‡)
(b)
Fig. 6. Electron micrographs of particles that are inactive in methane decomposition (a) on NAS3 after methane decomposition at
790 K and (b) on NAS2 after methane decomposition at 1070 K.
reforming (TOFSMR) on NAS catalysts and the NiMg decorated Ni0 particles exhibit catalytic activity in
sample in Arrhenius coordinates. For comparison, Fig. 3
also shows experimental data for a catalyst containing
25 wt % Ni supported on MgO promoted with 6% Al
(according to Rostrup-Nielsen [22, 23], sample A1). It
can be seen that the specific activity of nickel in NiMg
and NAS1 samples was close to published data on the
activity of catalyst A1. The estimated value of the
apparent activation energy of reaction on the Ni/MgO
catalyst was ~90 kJ/mol (according to Rostrup-Nielsen
[23], E‡ = 110 kJ/mol for catalyst A1). For the reduced
samples of Ni–Mg amesites, the apparent activation
energies were close to the Ea of the Ni/MgO sample.
steam methane reforming, and this activity is compara-
ble with the activity of nickel particles with clean sur-
faces.
Note that the catalytic properties of NAS samples in
the reaction of CO hydrogenation (Fischer–Tropsch
synthesis) at 483–503 K are little different from those
observed previously [24] in Ni/MgO catalysts. As a
result of the reaction, the formation of light saturated
hydrocarbons to heptane was observed. The product
composition was adequately described by the Ander-
son–Schulz–Flory distribution. Table 5 characterizes
the catalytic properties of NAS catalysts in this reac-
tion. The specific activities of NAS samples at 483 K
were somewhat lower than those of Ni/MgO catalysts.
A comparison between the values of TOFSMR
showed that the specific activity of the active compo-
nent in catalysts NAS2 and NAS3 was lower than the
corresponding value in the Ni/MgO sample by a factor
of only 1.5–2 over the test temperature range. Thus, The parameter α, which characterizes the selectivity of
KINETICS AND CATALYSIS Vol. 47 No. 3 2006

420
KHASIN et al.
oxide shell particles that exhibit high activity in the
reaction of steam methane reforming on catalysts pre-
pared by the reduction of Ni–Mg amesites.
1200
1100
1000
900
800
700
600
500
400
300
710
Temperature
990
Electron Microscopy of Samples after Performing
Methane Decomposition
DTG
signal
Previously [2], it was found that NAS samples were
less active than the Ni/MgO model catalyst in the reac-
tion of methane decomposition by a factor of 20–40.
Khasin and Kovalenko [2], who used in situ thermo-
gravimetry in a flow of methane, managed to find that,
for the reaction of methane decomposition in the exter-
nal kinetic region, the activation energy was equal to
190 5 kJ/mol for both NAS samples and the Ni/MgO
catalyst. However, the Ni–Mg amesite catalysts were
less active than Ni/MgO in this reaction. In this work,
we studied these samples after methane decomposition
using high-resolution electron microscopy. It can be
seen (Fig. 4) that the micrographs of Ni/MgO and NAS
samples carbonized in a flow of methane were essen-
tially different.
1
565
× 10
740
2
DTG
signal
600
620
100
200
Time, min
300
Fig. 7. The time dependence of the rate of carbon formation
in a flow of CO under temperature-programmed conditions.
Samples: (1) Ni/MgO and (2) NAS2.
The carbonized Ni/MgO sample consisted of a
dense ball of carbon fibers with the fishbone structure
(Figs. 4a, 4b). Only a few metal particles can be visu-
ally observed in the micrographs; all of these particles
were involved in the process of carbon formation. The
greatest particles were coated with a thick shell of car-
bon; the majority of particles of size smaller than 40 nm
were enclosed in a carbon fiber or arranged at its end;
this is consistent with many published data [26–29].
Similar carbon fibers can also be found in carbon-
ized samples NAS2 and NAS3 (see the survey micro-
graph of NAS3 in Fig. 4c, fibers are marked with
arrows); however, the fibers are few. Figure 5 exempli-
fies these fibers. At the same time, it can clearly be seen
that the vast majority of metal nickel particles was not
involved in the formation of catalytic fibrous carbon.
Moreover, high-resolution electron microscopy (Fig. 6)
demonstrated that these metal nickel particles did not
hydrogenaton, of the Anderson–Schulz–Flory distribu-
tion is consistent with the previously found dependence
of selectivity on active metal particle size [25].
A conceivable reason for this considerable catalytic
activity of fully decorated metal particles could be the
degradation of an oxide shell in the reaction atmo-
sphere. However, as mentioned above, XPS data for
sample NAS2 reduced and then treated under condi-
tions of steam methane reforming at 2 MPa suggest that
the surface composition of the sample in the reaction
atmosphere remained unchanged and Ni2+ cations
remained predominant on the sample surface. The deg-
radation of an oxide shell is even less probable under
conditions of CO hydrogenation, that is, at tempera-
tures 400 K lower than the temperature of the preceding have even trace graphite structures in the nearest envi-
ronment. Instead, Ni0 particles were coated with a layer
of an amorphous phase, which was similar to that
observed on reduced NAS samples (Fig. 2).
reductive treatment. Consequently, it is the Ni0 core–
Table 6. Experimental rates of carbon formation in the
course of CO disproportionation under temperature-pro-
grammed conditions
Consequently, decorated metal particles do not par-
ticipate in methane decomposition, and the low activity
of NAS samples in the reaction of carbon formation
from methane cannot be ascribed to Ni0 core–oxide
shell particles. It is likely that the formation of carbon
can be related to the presence of an amount of ordinary
undecorated Ni0 particles in reduced NAS samples.
wmax
,
w990 K,
wt %/min
Sample
Yield of C, wt %
wt %/min
Ni/MgO
1.15
0.30
0.46
0.33
0.750
0.025
0.110
0.035
170
14
NAS1
NAS2
NAS3
Catalytic Properties of Samples Prepared
by the Reduction of Ni–Mg Amesite in the Reaction
of CO Disproportionation
28
16
The NAS samples were also found inactive in the
reaction of CO disproportionation. Figure 7 compares
the dependence of the rate of reaction on time (temper-
Note: All of the data are given with reference to the weights
of reduced samples.
KINETICS AND CATALYSIS Vol. 47 No. 3 2006

STRUCTURE OF THE ACTIVE COMPONENT AND CATALYTIC PROPERTIES
421
ature) in the course of tests performed under tempera-
REFERENCES
ture-programmed conditions in an atmosphere of CO
for Ni/MgO and NAS2 samples. The rate profile of car-
bon formation exhibited two maximums at 700 and
1000 K, which were attributed previously to the forma-
tion of egg-shell graphite around metal particles and
carbon nanotubes, respectively [30].
It can be seen that sample NAS2 was less active and
rapidly deactivated in the course of reaction. Moreover,
this sample did not exhibit noticeable activity in the for-
mation of nanotubes. The two other NAS samples were
found even less active in the disproportionation of CO.
Table 6 summarizes the results of tests of the samples
in the disproportionation of CO. The activities of sam-
ples NAS1 and NAS3 in the region of nanotube forma-
tion (temperatures above 850 K) were lower than the
activity of Ni/MgO by one order of magnitude.
1. Khassin, A.A., Yurieva, T.M., Demeshkina, M.P., Kus-
tova, G.N., Itenberg, I.Sh., Kaichev, V.V., Plyasova, L.M.,
Anufrienko, V.F., Molina, I.Yu., Larina, T.V., Baron-
skaya, N.A., and Parmon, V.N., Phys. Chem. Chem.
Phys., 2003, vol. 5, no. 18, p. 4025.
2. Khasin, A.A. and Kovalenko, A.S., Dokl. Akad. Nauk,
2004, vol. 397, no. 6, p. 786.
3. Klemperer, D.F. and Stone, F.S., Proc. Roy. Soc., 1957,
vol. A243, p. 375.
4. Handbook of X-ray Photoelectron Spectroscopy, Wag-
ner, C.D., Riggs, W.M., Davis, L.E., Moulder, J.F., and
Muilenberg, G.E., Eds., Eden Prairie, Minn.: Perkin-
Elmer, 1978.
5. Plyasova, L.M., Kriger, T.A., Khasin, A.A., and Par-
mon, V.N., Dokl. Akad. Nauk, 2002, vol. 382, no. 4,
p. 505.
6. Barbier, A., Pereira, E.B., and Martin, G.A., Catal. Lett.,
1997, vol. 45, p. 221.
CONCLUSIONS
7. Christmann, K., Behm, R.J., Ertl, G., van Hove, M.A.,
and Weinberg, W.H., J. Chem. Phys., 1979, vol. 70,
p. 4168.
The reduction of Ni–Mg aluminosilicates with the
amesite structure with hydrogen at 920 K resulted in the
formation of metal nickel particles coated with a thick
amorphous oxide shell (2 nm in thickness). Because of
this, reduced NAS samples cannot chemisorb oxygen;
however, they exhibit high adsorption capacity for
hydrogen. Correspondingly, these samples exhibit cat-
alytic properties that are much different from those of
Ni/MgO.
8. Brennan, D., Hayward, D.O., and Trapnell, B.M.W.,
Proc. Roy. Soc. London, Ser. A, 1960, vol. 256, p. 81.
9. Criado, J.M., React. Kinet. Catal. Lett., 1978, vol. 8,
p. 143.
10. Smith, J.S., Thrower, P.A., and Vannice, M.A., J. Catal.,
1981, vol. 68, p. 270.
Thus, the Ni0 core–oxide shell particles are highly
active in steam methane reforming. In this case, the
specific activity of a single hydrogen adsorption site is
no less than half of the observed activity of nickel metal
in the Ni/MgO catalyst. The specific activity of the Ni0
core–oxide shell particles in the hydrogenation of CO is
little different from the specific activity of nickel metal
in the Ni/MgO catalytic system.
11. Delmon, B., J. Mol. Catal., 1990, vol. 59, p. 179.
12. Khassin, A.A., Yurieva, T.M., Kustova, G.N., Plya-
sova, L.M., Itenberg, I.Sh., Demeshkina, M.P., Cher-
mashentseva, G.K., Anufrienko, V.F., Zaikovskii, V.I.,
Larina, T.V., Molina, I.Yu., and Parmon, V.N., J. Mol.
Catal. A: Chem., 2001, vol. 168, p. 209.
13. McIntyre, N.S. and Cook, M.G., Anal. Chem., 1975,
vol. 47, p. 2208.
At the same time, the Ni0 core–oxide shell particles
are completely inactive in the reactions of graphite-like
carbon formation in both methane decomposition and
CO disproportionation. The trace activity of NAS cata-
lysts observed in these processes can be attributed to
the impurities of undecorated Ni0. Based on the exper-
imental data, we can conclude that metal nickel parti-
cles decorated with an oxide shell exhibit high selectiv-
ity in steam methane reforming but do not exhibit activ-
ity in the formation of graphite-like phases.
14. McIntyre, N.S., Rummery, T.E., Cook, M.G., and Owen, D.,
J. Electrochem. Soc., 1976, vol. 123, p. 1164.
15. Kim, K.S. and Winograd, N., Surf. Sci., 1974, vol. 43,
p. 625.
16. Shavloy, R.B., Reucroft, P.J., and Davis, B.H., J. Catal.,
1979, vol. 56, p. 336.
17. Li, C.P., Proctor, A., and Hercules, D.M., Appl. Spec-
trosc., 1984, vol. 38, p. 880.
18. Carley, A.F., Jackson, S.D., O’Shea, J.N., and Ro-
berts, M.W., Surf. Sci., 1999, vol. 440, p. L868.
19. Sakiyama, K., Koga, K., Seto, T., Hirasawa, M., and
ACKNOWLEDGMENTS
Orii, T., J. Phys. Chem. B, 2004, vol. 108, p. 523.
We are grateful to Dr. Sci. (Chem.) L.M. Plyasova
and Cand. Sci. (Chem.) I.Sh. Itenberg for their assis-
tance and helpful discussions and toA.S. Kovalenko for
participation in the experiments. This work was sup-
ported by the Siberian Division of the Russian Acad-
emy of Sciences (a grant for young scientists). A.A.
Khasin acknowledges the support of the Russian Sci-
ence Support Foundation.
20. Khassin, A.A., Yurieva, T.M., Kaichev, V.V., Bukhti-
yarov, V.I., Budneva, A.A., Paukshtis, E.A., and Par-
mon, V.N., J. Mol. Catal. A: Chem., 2001, vol. 175,
p. 189.
21. Khassin, A.A., Anufrienko, V.F., Ikorskii, V.N., Plya-
sova, L.M., Kustova, G.N., Larina, T.V., Molina, I.Yu., and
Parmon, V.N., Phys. Chem. Chem. Phys., 2002, vol. 4,
p. 4236.
KINETICS AND CATALYSIS Vol. 47 No. 3 2006

422
KHASIN et al.
22. Rostrup-Nielsen, J.R., Catal. Sci. Technol., 1984, vol. 5, 27. Baker, R.T.K. and Harris, P.S., Chemistry and Physics of
p. 1.
Carbon, Walker, P.L. and Thrower, A., Eds., New York:
Marcel Dekker, 1978, vol. 14, p. 83.
23. Rostrup-Nielsen, J.R., J. Catal., 1973, vol. 31, p. 173.
28. Benissad, F., Gadelle, P., Coulon, M., and Bonnetain, L.,
24. Khasin, A.A., Yur’eva, T.M., and Parmon, V.N., Dokl.
Akad. Nauk, 1999, vol. 367, no. 3, p. 367.
Carbon, 1988, vol. 26, p. 425.
29. Nolan, P.E., Schabel, M.J., Lynch, D.C., and Cutler, A.H.,
25. Khasin,A.A. and Parmon, V.N., Dokl. Akad. Nauk, 1999,
Carbon, 1995, vol. 33, p. 79.
vol. 368, no. 4, p. 503.
30. Khassin, A.A., Yurieva, T.M., Zaikovskii, V.I., and Par-
mon, V.N., React. Kinet. Catal. Lett., 1998, vol. 64,
p. 63.
26. Avdeeva, L.B., Kochubey, D.I., and Shaikhutdinov, Sh.K.,
Appl. Catal., A, 1999, vol. 177, p. 43.
KINETICS AND CATALYSIS Vol. 47 No. 3 2006