New nickel catalysts for the formation of filamentous carbon in the reaction of methane decomposition

Article abstract of DOI:10.1006/jcat.1999.2562

The catalyst synthesis method was based on impregnating a porous metal oxide with the precursor of a textural promoter, followed by reduction of the latter. The desired texture of NiO was provided by precalcining it at a certain temperature within 300°-900°C. The yield of carbon and mechanical strength of carbon particles growing on the catalyst during methane decomposition increased with nickel concentration in the catalyst, reaching the maxima at 90-96% nickel. The carbon yield was highest (375-384 g/g C/Ni) for catalyst particles of 10-40 nm average dia. The carbon yield from SiO₂ as textural promoter was higher than that from Al₂O₃, MgO, TiO₂, or ZrO₂.

Full text of DOI:10.1006/jcat.1999.2562

Journal of Catalysis 187, 77–84 (1999)
Article ID jcat.1999.2562, available online at http://www.idealibrary.com on
New Nickel Catalysts for the Formation of Filamentous Carbon
in the Reaction of Methane Decomposition
1
M. A. Ermakova, D. Yu. Ermakov, G. G. Kuvshinov, and L. M. Plyasova
Boreskov Institute of Catalysis, Siberian Branch of the Russian Academy of Sciences, Prospekt Ak. Lavrentieva 5, Novosibirsk 630090, Russia
Received January 22, 1999; revised April 7, 1999; accepted May 6, 1999
2
2). Most of the authors propose CFC as a starting material
A method for intentional synthesis of metal catalysts, specifi- for synthesis of new adsorbents, composites, and catalysts.
cally nickel catalysts for methane decomposition, was developed. Among the other applications, CFC is suggested for use as
The method was based on impregnation of a porous metal oxide
with the precursor of a textural promoter followed by reduction
coats for different materials to vary the properties of the
latter (textural parameters, physical properties, catalyst se-
of the latter. The desired texture of NiO was provided by precal-
lectivity, etc.). In the meantime, the applicability of CFC to
�
cining it at a certain temperature within the range 300 to 900 C.
some areas, such as manufacturing of electrodes, lining of
A family of nickel-based catalytic systems with various concentra-
rectors and heat exchangers, the metallurgy and chemical
industry, pigments and fillers for varnish and paint, elec-
trical and rubber industries, and manufacturing of various
lubricants, was not examined. The purity and low price of
tions of the active component, metal particle sizes, and stability to
deactivation in the course of synthesis of filamentous carbon was
prepared. The yield of carbon and mechanical strength of carbon
particles growing on the catalyst during methane decomposition
were found to increase with the concentration of nickel in the cata- a product are often of chief importance in these cases. We
lyst, to reach their maxima at 90–96% nickel. The highest yield of believe that CFC can be used in these industries. However,
carbon (375–384 g carbon per g nickel) was observed for the catalyst there are as yet unresolved problems in the field. Among
with particles of 10–40 nm average diameter. The effect of textural
promoters (SiO2, Al2O3, MgO, TiO2, and ZrO2) on the catalyst per-
formance was studied; the highest carbon yield was obtained with
these problems, which impede the wide application of CFC,
is one of manufacturing considerable quantities of this ma-
terial with reproducible properties.
Notice that little attention is paid in the literature to the
problem of the quantitative yield of CFC. Meanwhile, a
specific feature of the reaction is that the volume of an
initially loaded catalyst is many times increased. And, it
must be vigorously stirred during the process, which causes
problems even with the scaling of laboratory reactors, not
to mention industrial use. Some of approaches to solving
SiO2.
�c 1999 Academic Press
Key Words: nickel catalysts for methane decomposition; impreg-
nation of porous oxide; textural promoters; yield of filamentous
carbon per gram of nickel.
1
. INTRODUCTION
Catalytic filamentous carbon (CFC) is the subject of nu- this problem are discussed in Ref. (23). Another feature
merous current studies. A unique filamentous structure of of large-scale CFC production is the use of the cheapest
CFC makes it a promising material for various technologies hydrocarbon feedstock such as natural gas with methane
(
1–7). First, EM images were recorded for CFC as early as as the main component. Therefore, there arises a need for
in 1950 (8, 9). In the 1970s, this material was applied as a developing highly effective catalysts for decomposition of
new promising material. Since that time, the studies in the methane. Iron and nickelare the active componentsofthese
field have not ceased. Quite extensive literature is available catalysts. The nickel-based catalysts are active at lower tem-
now in the field of CFC. Filaments and nanotubes of various peratures and provide a higher yield of CFC per mass unit
types have been synthesized, and textural and adsorptive of the active component. For this reason, the nickel metal-
properties of this material, as well as kinetics and thermo- based catalysts seem the most attractive for the industrial
dynamics of CFC formation over some catalysts, have been use. Ni–Al O catalysts for methane decomposition pre-
2
3
studied (1, 9–18). A variety of hydrocarbons and catalytic pared by coprecipitating the components were thoroughly
systems for hydrocarbon decomposition have been used for investigated (24–26). The highest yield of carbon per cata-
synthesis of CFC of different structures and textures. The lyst mass unit was obtained using the catalyst comprising
pertinent reaction mechanisms have been suggested (19– 90% nickel, the initial nickel particles of the catalyst being
ca. 14 nm in average size. The yield was as high as 145 g
carbon per gram of catalyst. Introduction of copper, 3% ,
1
Corresponding author. E-mail: erm@sirius.catalysis.nsk.su.
77
0021-9517/99 $30.00
Copyright �c 1999 by Academic Press
All rights of reproduction in any form reserved.

7
8
ERMAKOVA ET AL.
to the catalyst allowed the carbon yield to be improved to solution of magnesium nitrate as the supplier of magnesia.
2
40 g per gram of catalyst at some decrease in methane The impregnated precursors were dried at room temper-
�
conversion (26).
ature and reduced in hydrogen at 550 C for 1 h. Before
Deactivation of catalysts for hydrocarbon decomposition being unloaded from the reactor, prepared catalysts were
is a fundamental problem. The deactivation mechanisms passivated with ethanol in order to prevent pyrophoric phe-
are discussed elsewhere (27, 28). The deactivation is usu- nomena. The TEOS to be introduced into pores of nickel
ally accounted for by either crushing of the active particles oxide was pre-hydrolyzed with a substoichiometric amount
(
atomic erosion or fragmentary dispersing) in the course of of water with hydrogen chloride added as the hydrolysis
the reaction or encapsulation of the metal particle into car- catalyst. TEOS molecules were thus coupled together to
bon. The latter reason seems more plausible for methane form oligomer chains through siloxane bonds to allow for
decomposition. In this case the deactivation rate is affected the rates of bulk aggregation of SiO2 particles and of gel for-
by diffusion phenomena in the nickel–carbon system and is mation on storing to be many times decreased. The other
determined by the rate ratio of formation of carbon atoms solutions were not subjected to any pretreatment proce-
to their diffusion toward the deposition sites. Hence, there dure.
is some optimal size of active metal species (nickel particles
A sample of 90% Ni–10% Al2O3 catalyst used for com-
in this case) to provide the longest equilibrium between the parison was prepared by coprecipitation according to the
rate of formation of atomic carbon species and the rate method given elsewhere (26). The components were pre-
of carbon diffusion to the aggregation points. However, cipitated by NaOH from an aqueous solution of Ni(NO3)2 �
it is our opinion that distortions in the metal crystal lat- 6H2O and Al(NO3)3 � 9H2O taken in known proportions.
tice induced by some impurities (for example, intermetallic Samples were carefully washed with distilled water until no
compounds of the metal and textural promoter) may have NO3 ions were detected in the remaining solution and were
additional impact upon the deactivation rate.
The present work was aimed at comprehensive studies of perature was elevated at the rate of ca. 5 C min to 350 C.
the influence of concentration and dispersion of the active Samples were reduced in flowing pure hydrogen for 3 h at
calcined in flowing nitrogen for 3 h. The calcination tem-
�
� 1
�
�
component in the initial catalyst, as well as the nature of the temperature elevated to 550 C at the same rate.
a textural promoter, on the time of deactivation of nickel
The texture of samples was characterized by low-temp-
catalysts in the course of methane decomposition to pro- erature nitrogen adsorption at 77 K using an automated
duce CFC.
installation ASAP-2400.
The average sizes of nickel oxide and nickel metal parti-
cles were determined by the Sherrer radiographic method
2
. EXPERIMENTAL
(
30) based on the half-width of diffraction lines assigned
Nickel oxide used for preparing nickel catalysts was syn- to (200), (111) NiO and (200), (111) Ni. The XRD tech-
thesized by precipitating a nickel nitrate with aqueous am- nique was used for determining the phase composition of
monia followed by the drying and calcining of the obtained the samples.
�
�
The catalysts were tested in a vibrofluidized bed at 550 C
precipitate (nickel hydroxide) in a muffle furnace at 300 C
for 4 h. The texture of the produced nickel oxide was varied using a laboratory installation with a flow quartz reactor.
by calcining it at temperatures 300 to 900 C.
�
3
The operation region of the reactor was 20 cm in volume.
Concentrations of alkali and earth metals in NiO samples Methane consumption was always preset at 120 liter per
were determined using an ionization spectroscopic tech- gram of nickel per hour. Hydrogen concentration in the
nique with a BAIRD plasma spectrophotometer. A curve mixture was chromatographically measured using a column
for thermal decomposition of nickel hydroxide was plotted filled with NaX zeolite. The catalyst activity was determined
using a Q-1500 derivatograph.
from the hydrogen percentage in the reaction mixture. The
To prepare catalysts, the precursor of an active compo- reaction was stopped as soon as the outlet hydrogen con-
nent (nickel oxide in our case) was impregnated with the centration reached 5% . The amount of carbon deposited
solution of the precursor of a textural promoter taken in on the catalyst during the reaction time was determined by
the calculated amount (29). Hard-to-reduce oxides (HRO) weighing unloaded samples.
such as silica, alumina, titanium dioxide, zirconium oxide,
Attrition resistance of carbon granules was determined
and magnesia were used as textural promoters. An alcohol by the method described in Ref. (31). A sifted carbon frac-
solution of tetraethoxisilane (TEOS) was used as the sup- tion (2–1.25 mm) in the amount 0.2 g was loaded into a
plier of silica, an aluminium isopropoxide solution in xylene conical drum with its wide end of 65 mm diameter rotated
and an aqueous solution of aluminium nitrate as suppliers at the rate of 2 turns per s. A porcelain ball of 20 mm diam-
of alumina, a solution of titanium ethoxide in xylene as the eter and 10 g weight was placed into the drum to intensify
supplier oftitanium dioxide, an aqueoussolution of zirconyl the attrition process. The test took 6 h; then the sample
nitrate as the supplier of zirconium oxide, and an aqueous was unloaded and sifted using a sieve with 1.25 mm mesh.

NICKEL CATALYSTS FOR METHANE DECOMPOSITION
79
Attrition was taken equal to the percentage of granules
passed through the mesh.
3
. RESULTS AND DISCUSSION
3
.1. Precursors of Active Component
Concentrations of alkali metals and calcium (<0.03% for
K; ca. 0.001% for Na; ca. 0.005% for Ca) not higher than
their concentrations in distilled water were determined in
nickel hydroxide using ionization spectroscopy. That was
why we did not take into account the influence of these
metals on catalyst sintering.
X-ray amorphous nickel hydroxide with the specific sur-
face area of 256 m /g was produced using the above-
described method. During thermal decomposition of it,
2
FIG. 1. Activity (% H2) of coprecipitated nickel–alumina catalyst and
nickel–silica catalyst, prepared from NiO calcined at 450 C, vs reaction
time.
�
principal endothermic transformations were observed at
�
2
90 C, generally in agreement with the pertinent literature
to as carbon yield) goes through a maximum at the range 90
to 96% nickel, which is in agreement with the conclusions
made in Refs. (24–26). The deactivation time of the 90%
Ni–10% SiO2 catalyst, which shows the best performance
among the catalysts discussed in the present work, is about
three times that observed for the coprecipitated Ni–Al2O3
catalyst with the same concentration of nickel (Fig. 1). Ac-
cordingly, the carbon yield is also improved. The rate of
carbon deposition is ca. 10 g/g Ni h for the nickel catalysts
presented in Fig. 1. Then, the carbon yield under these con-
ditions may be calculated as
data (24, 32).
�
Decomposition of nickel hydroxide at 300 C produced
nickel oxide with specific surface area 144 m /g and average
particle size 3–4 nm (from XRD data).
2
3
.2. Influence of Nickel Concentration in the Catalyst
on Carbon Yield and Granule Strength
For nickel–alumina catalysts, the yield of carbon per cata-
lyst mass unit was shown (26) to increase with the concen-
tration of nickel in the catalyst. The highest carbon yield,
equal to 145 g per gram of catalyst, was reached at 90% Ni.
However, the next point was the 100% nickel catalyst com-
prising no textural promoter. A sharp decrease in capacity
to yield carbon, to lower than one tenth, was observed.
GC = td � 10,
where G is the carbon yield and t is the lifetime (h) of the
C
d
In the present work we tried to study the nickel–silica catalyst until the complete deactivation.
system in the same manner. The system was prepared by
Notice that in the vibrofluidized catalyst bed, CFC is nu-
a unique method (29). Unlike the method described in cleated and grows in the form of irregularly shaped gran-
Ref. (26), this one provided a wider range, 30 through 96% , ules. The granule strength is obviously dependent on the
of nickel concentrations. Table 1 shows that the yield of car- nickel loading in the catalyst. The granule strength is seen
bon per active component mass unit (hereinafter referred in Fig. 2 to increase with an increase in the nickel concentra-
tion. Probable reasons are as follows. The lower the nickel
TABLE 1
Influence of Nickel Concentration (%) in Ni–SiO2 Catalysts, Pre-
�
pared by Calcining NiO at 300 C, on the Surface Area (Sspecif cat,
2
m /g) of Reduced Catalysts, the Size of Metal Particles (DNi, nm),
the Carbon Yield (Gc, g CFC/g Ni), and the Specific Surface Area
2
of CFC (Sspecif CFC, m /g)
%
Ni
Sspecif cat
DNi
Gc
Sspecif CFC
3
0
0
0
5
0
6
583
360
210
120
71
7
9
9
10
11
23
65
120
220
183
204
300
238
10
124
133
146
130
100
106
—
5
7
8
9
9
49
6
FIG. 2. Effect of nickel concentration in the catalyst on attrition re-
sistance of carbon granules deposited on this catalyst.
1
00

8
0
ERMAKOVA ET AL.
concentration is in the initial catalyst, the longer is the dis-
tance between the carbon filaments growing on the active
species and the lower is their concentration per unit vol-
ume of the carbon granule. Thus, the density and, hence,
the strength characteristic of the carbon formed on high-
loaded nickelcatalystscannot be reached on the low-loaded
catalyst at the moment of the catalyst deactivation. Again,
the strength and density of carbon granules also depend on
the carbon yield provided by the catalyst. The higher the
carbon yield, the higher the density of it. For example, the
3
pore volume equals 0.27 cm /g in the carbon produced on
9
0% Ni–10% SiO2 (at the carbon yield of 300, see Table 1),
3
but it equals 0.34 cm /g in the carbon synthesized on 30%
Ni–70% SiO2 (at the carbon yield of 120). There is an ap-
parent difference between the strengths of these carbon
samples (Fig. 2).
One can also assume, based on the data available in
Table 1 for the given series of catalysts, that the carbon
yield is affected by the nickel particle size and reaches a
maximum at the average diameter equal to 10–20 nm. It
also should be notice that somewhat of a decrease in the
specific surface area of CFC is observed at the maximal
nickel concentration in the catalyst and the maximal car-
bon yield. This effect is, probably, the result of an increase
in the diameter of carbon filaments.
FIG. 3. NiO specific surface area as a function of calcination temper-
ature.
Thus, oxide systems comprising NiO particles of average
diameter from 3 to more than 100 nm were synthesized
(
Table 2). Stabilization of the obtained textural parameters
was achieved by impregnating the oxide systems with SiO2
in amounts of 4 to 10% . As was to be expected, reduction of
the catalysts by hydrogen demonstrated that NiO particles
to be reduced to the metal state are stabilized more reliably
at a higher proportion of SiO2. For example, when the oxide
particles are 4 nm in average size, the proportion of silica
3
.3. Influence of Average Size of Nickel Particles
on Carbon Yield
The influence of the size of active particles on the car- of 4% was, apparently, insufficient to prevent them from
bon yield was studied thoroughly using a series of catalytic sintering, and the diameter of the formed metal particles
systems with the same proportions of components but dif- was almost 7–8 times that of the precursor particles. At the
ferent average diameters of nickel particles. The catalysts same time, the sintering process resulted in no more than
were synthesized by a unique method. A nickel oxide sam- a fourfold increase in the diameter of the primary particles
ple was divided into eight portions, and each of them was in the presence of 10% silica. As the size of the primary
�
calcined at temperatures ranging from 300 to 900 C in or- NiO particles increased, the stabilizing effect was equal-
der to prepare precursors of the active component with a ized and became independent of the proportion of intro-
preset texture. Figure 3 shows specific surface area of NiO duced silica. As for the carbon yield, the highest capacity to
against the calcination temperature.
yield carbon was observed for the systems comprising 10%
TABLE 2
Influence of Temperature of Nickel Oxide Calcination on Textural Parameters of Ni Catalysts and
2
the Carbon Yield (Gc, g CFC/g Ni), Particle Size (DNiO, nm), Specific Surface Area (Sspecif, m /g), Pore
3
Volume (Vpore, cm /g), Average Pore Size (Dpore, nm)), and size of metal particles (DNi, nm)
9
6% Ni–4% SiO2
90% Ni–10% SiO2
Tcalcin.
NiO ( C)
Sspecif.
NiO
�
Vpore
Dpore
DNiO
DNi
GC
DNi
GC
3
4
4
5
6
7
8
9
00
00
50
00
00
00
00
00
144
90
56
29
17
7
0.57
0.40
0.28
0.15
0.05
0.03
0.013
—
17
18
20
20
20
20
20
—
3
12
19
35
40
23
20
20
25
32
38
—
—
238
10
10
17
25
30
40
—
—
300
238
300
224
200
107
375
375
340
370
384
2
65
88
3
0.7
1.5
0.5
>100
0.5

NICKEL CATALYSTS FOR METHANE DECOMPOSITION
81
silica and particles of 10 to 40 nm average size. The further
TABLE 3
increase in the average size of nickel particles resulted in a
sharp decrease in the amount of carbon deposited thereon.
Thus, silica introduced according to the above-described
procedure can be considered as an effective textural pro-
moter to provide synthesis of the catalytic system with the
required textural parameters. Therefore, the given method
can be used for preparing catalysts with preset parameters
and for other reactions.
Effect of the Nature of Textural Promoter on the Size of Metal
Particles (DNi, nm) in the Reduced 90% Nickel Catalysts and the
Carbon Yield (Gc, g CFC/g Ni)
T of NiO calcination
T of NiO calcination
�
�
300 C
450 C
Promoter
DNi
Gc
DNi
Gc
SiO2
Al2O3
Al2O3
ZrO2
TiO2
MgO
11
28
43
16
30
17
300
220
20
272
200
220
17
22
20
21
26
22
375
315
280
300
218
287
a
3
.4. Effect of Chemical Nature of the Textural Promoter
in Nickel Catalysts on the Carbon Yield
The effect of the chemical nature of textural promot-
ers introduced into a catalytic system on the efficiency of
a
Al2O3 was introduced into the nickel oxide matrix from the aqueous
nickel catalysts was studied with respect to the reaction solution of aluminum nitrate.
of methane decomposition to produce filamentous carbon.
Two series of nickel catalysts were prepared by calcining
�
NiO at 300 and 450 C, respectively. The oxide calcined at
(c) Intermediate efficiency was observed with MgO and
�
the minimal temperature (300 C) was shown above to con- ZrO as textural promoters.
2
sist of the finest particles of NiO (ca. 3 nm). Hence, impreg-
�
When NiO calcined at 450 C was the catalyst precur-
sor, the sintering processes were confined to the inside of
the paternal particles of ca. 19 nm, whatever textural pro-
moter among those under consideration was used. Figure 4
demonstrates diffraction patterns recorded for 90% Ni
catalysts with various textural promoters and for oxide
nickel precursor of these catalysts. Broadening of the (111)
and (200) lines used by us to compute the average size of
nickel particles is seen to be approximately equal for all
the catalytic systems and to correspond to broadening of
the (111) line of nickel oxide. Nevertheless, the highest car-
bon yield was observed with the Ni–SiO2 catalysts. It seems
appropriate to explain this phenomenon in terms of the
presence or absence of intermetallic Ni–HRO compounds
in the system, which may effect the process of carbon diffu-
sion through the nickel particle. Notice that no unambigu-
ous solution of this problem is available in the literature.
The authors’ opinions are quite contradictory. For example,
the presence of chemical compounds which inhibit decom-
position of hydrocarbons is assumed in the Ni/SiO2 systems
nation of this oxide with various HRO can illustrate the
efficiency of these compounds as textural promoters pro-
tecting the metal particles against sintering in the course
of the reduction. This knowledge may be useful to de-
velop high-loaded metal catalysts comprising active parti-
cles of the minimal size. However, methane decomposition
needs catalytic systems with coarser particles. These parti-
cles are formed when NiO calcined at temperatures above
�
4
00 C is used as the catalyst precursor. SiO2, ZrO2, MgO,
TiO2, and Al2O3 were introduced into oxide matrices. Note
that alumina was introduced into the NiO pores from the
solutions of various compounds. The interaction between
alumina, prepared by decomposition of aluminium iso-
propoxide, and nickel oxide was unlikely to produce chem-
ical compounds, whereas such a potentiality was not ex-
cluded to occur during impregnation of nickel oxide with an
aqueous solution of aluminium nitrate. These phenomena
may affect the texture of the catalysts upon further treat-
ments. Later, the essential difference between the catalyst
prepared using aluminium isopropoxide and that prepared
from aluminium nitrate was found.
(
33–35). On the other hand, the authors of Refs. (36, 37) be-
lieve that intermetallic Ni–Si compounds cannot be formed
at temperatures below 1000 K, and Ni is only preserved
in the metal state. It is our opinion (which agrees with the
understanding of other authors (32)) that the interaction
between the components is minimal in the Ni–SiO2 system
Table 3 presents the data on the size of metal particles
and the carbon yield obtained over the corresponding cata-
lysts. What is seen for the catalyst prepared using nickel
�
oxide precursor calcined at 300 C is as follows:
(
a) The highest level of protecting metal particles against compared to that in the other systems, e.g., Ni–Al O , with
2 3
sintering during the reduction is achieved in the presence any method used for preparing the catalysts. This kind of in-
of SiO2. In this case, the particles formed are as little as 3 or teraction is even less probable in the catalysts prepared by
4
times coarser relative to the average size of the precursor the method proposed by us because the textural promoter
(
NiO) particles;
precursor is introduced into the calcined preformed NiO.
(
b) Al2O3 and TiO2 used as the textural promoter pro- At this stage any interaction of the components may hardly
vided the lowest level of protecting Ni particles against sin- occur when tetraethoxisilane is used as the precursor of
tering;
HRO. Again, the following stage of reduction in hydrogen

8
2
ERMAKOVA ET AL.
transition metals, resulted in remarkable improvement of
the carbon yield (26). Thus, the above discussion allows the
conclusion that the yield of carbon produced by methane
decomposition depends on the purity of nickel particles;
the lower is the concentration of distortions of the crystal
lattice induced by unreduced intermetallic compounds, the
higher is the yield.
If we return to the catalytic systems presented in Table 3,
the best performance of the Ni–SiO2 system may be also ac-
counted for by the lowest density of silica among the HRO
3
under study. The density is 2.65 cm /g for quartz, while it is
3
3
3
3
.96 cm /g for Al2O3, 3.58 cm /g for MgO, 4.3 cm /g for TiO2,
3
and as high as 5.68 cm /g for ZrO2. Therefore, at the same
weight ratio of the constituents (90% Ni and 10% HRO),
the Ni–SiO2 catalyst comprises27vol% silica, which ismuch
higher than the volume proportion of HRO in the other sys-
tems under consideration: 20 vol% for Al2O3, 22 vol% for
MgO, 19 vol% for TiO2, and 15 vol% for ZrO2. This fact is
of great importance because the catalytic system may be-
come rather mobile in the methane medium (38, 39). This
is the case in which the higher volume of the textural pro-
moter is to prevent more effectively the sintering of the
nickel particles at the earliest stage of the reaction.
3
.5. Principle of Catalyst Formation
In coprecipitated high-loaded metal catalysts, the active
phase in the catalyst grain may be in the form of individ-
ual clusters divided by a textural promoter or in the form
of an infinite cluster with the embedded textural promoter
species (24). Actually, these are extreme cases which coex-
ist. In other words, there occur both individual metal par-
ticles and interconnected groups of such particles in the
catalytic system at practically any phase ratio. As a result,
there must be a wide diameter distribution of active parti-
�
FIG. 4. Diffraction patterns of NiO calcined at 450 C and catalysts
prepared thereof.
at a high rate of temperature elevation (no less than cles. For decomposition of hydrocarbons, the stable reac-
�
2
0 C/min) does not allow these compounds to be formed. tion can be provided by particles that are neither too fine
Some of our catalyst samples comprising 85% Ni and 15% nor too coarse. The finest particles on which formation of
SiO2 were studied using mass-spectrometry of secondary CFC was observed were 2 to 4 nm in size (14, 40). However,
ions. This is a very sensitive technique, but we failed to it is hardly possible to say anything about any quantitative
detect any Ni–Si compound in the catalyst reduced in hy- yield of carbon. As for coarser particles, they become en-
�
drogen at 550 C. On the contrary, the paracrystallinity of closed into the solid carbon crust a in short time during
nickel was revealed in the Ni–Al2O3 catalyst for methane the reaction and, therefore, are deactivated (41, 42). One
decomposition (25) due to the penetration of the hard-to- can assume that a more uniform system is produced by the
reduce oxide into the crystal nickel lattice, the effect of the method we proposed that results in a more narrow range of
paracrystal nickel state being not related to the carbon yield diameter distribution of the metal particles. This conclusion
in this reaction. However, from the data by the same au- is based on the following speculations: Hydrolysis of nickel
thors (26), the average size of nickel particles falls into the ammoniate results in formation and existence for a while of
optimal range determined by us but the carbon yield is not nickel hydroxide sol with particles of close size, 3–4 nm. The
higher than 145 g per gram of catalyst. The coprecipitation particles are not enlarged in the course of the topochemi-
method used to prepare these catalysts is known to favor cal reaction Ni(OH)2 → NiO. The specific surface area de-
a strong interaction between the components and, in this creases due to an increase in the density of NiO formed
3
3
case, the formation of spinels which are not reduced even (8.91 cm /g) against the density of Ni(OH)2 (4.1 cm /g).
�
at a temperature as high as 550 C. Addition of a minor No textural promoter being in the system that time,
amount of copper, which is known to promote reduction of there is nothing to hinder the nickel oxide particles from

NICKEL CATALYSTS FOR METHANE DECOMPOSITION
83
metal-to-HRO percentage ratio. This is the fundamental
difference between the systems under consideration and
those obtained by coprecipitation. Theoretically, the only
parameters affecting the amount of the introduced textural
promoter are the concentration of the HRO precursor in
the impregnating solution and the pore volume of the ac-
tive component precursor. In practice, the uniform mode
of HRO distribution in the subsurface region of the active
phase will be inevitably disrupted as the HRO concentra-
tion increases above a certain level.
Essentially, the catalytic systems discussed above resem-
FIG. 5. Pore volume of NiO samples, calcined at various tempera- ble impregnated catalysts but bear an opposite sense. Both
tures, vs specific surface area.
cases imply impregnation of a preformed solid matrix with a
liquid phase and adsorption of the solute on the surface fol-
lowed by thermal decomposition of this compound. As for
our case, the solid matrix being considerably transformed
during the reduction, the process of formation of the cata-
lytic system is more complicated and needs further study.
further enlargement, which depends on thermodynamic
factorsonly, such astemperature and calcination time. Thus,
highly dispersed NiO particles (which are formed at the
�
minimal temperature, 300 C, of decomposition of nickel
Nevertheless, the stabilizing effect of the textural promoter
hydroxide) coalesce in uniform manner at the sites of con-
introduced in such a manner is rather high, which allows
tacts at elevating temperature and the system remains uni-
the coalescing of the metal particles to be avoided in many
form. The NiO pore volume is proportional to a decrease in
cases under severe reduction conditions, even though the
the specific surface area (Fig. 5), the average pore diameter
particles are in direct contact with one another.
being practically unchanged (Table 2). At first glance, such
a phenomenon seems improbable, while numerous litera-
4
. CONCLUSIONS
ture data, e.g., those cited in Refs. (43–47), argue the same
to be observed upon thermal shrinkage of silica. Iler (47)
attributes this phenomenon to the formation, as a result
of sintering, of local regions evenly distributed through the
bulk of silica, which are free of pores, or the pores are not
The following conclusions can be made based on the re-
sults obtained:
(1) A unique method was developed for intentional syn-
exposed to adsorbate molecules. The further elevation of thesis of metal catalysts, specifically nickel catalysts for
the calcination temperature results in formation of new re- methane decomposition. The method allows the catalytic
gions, and the like. Such a NiO system seems to resemble systems to be synthesized in a wide phase range (active
silica, which is known to be a uniform system. The system
can be stabilized at a certain stage by impregnating it with
an HRO precursor. As a result, HRO molecules are ad-
sorbed on energetically favorable fragments of the oxide
surface (Fig. 6), viz. bridges between the particles to pre-
vent them from further coalescing. The stabilizing role of
these compounds is very significant, so the average size of
nickel oxide particles does not change upon calcining the
�
system at a temperature as high as 800 C, whereas the par-
ticles are about 5 times increased in the absence of HRO
at the same calcination temperature. The stabilizing fea-
ture of HRO during the reduction is more pronounced, the
smaller is its particle size relative to the size of particles
to be stabilized. That is why the complete stabilization of
the finest NiO particles cannot be achieved due to the close
sizes of HRO and NiO particles. Nevertheless, sintering can
be avoided in the most cases (see Tables 2 and 3).
Thus, one can assume that the prepared catalytic system
is built-up by rather uniform metal particles joining one
another to form an infinite cluster. HRO particles are al-
located among the pores between them and behave as the
textural promoter. Such a structure is independent of the matrix.
FIG. 6. Hard-to-reduce oxide arrangement in the structure of NiO

8
4
ERMAKOVA ET AL.
component to HRO ratio) and, on the other hand, active 15. Kuvshinov, G. G., Mogilnykh, Yu. I., and Kuvshinov, D. G., Catal.
Today 42, 357–360 (1998).
6. Kuvshinov, G. G., Mogilnykh, Yu. I., Kuvshinov, D. G., Zaikovskii,
V. I., and Avdeeva, L. B., Carbon 36, 87–97 (1998).
7. Mogilnikh, Yu. I., Kuvshinov, G. G., and Lebedev, M. Yu., “Confer-
ence Eurocarbon ’98, Extended Abstracts and Programme, Strasburg,
France, July 5–9, 1998,” Vol. 1, pp. 445–446, 1998.
particles of the required size to be obtained at the same
phase ratio.
1
1
(
2) The carbon yield and the strength of granules de-
posited on the catalyst upon methane decomposition were
shown to increase with an increase in the nickel concentra-
tion in the catalyst to reach a maximum in the range 90 to
1
1
2
2
2
2
2
8. Rodriguez, N. M., Chambers, A., and Baker, R. T. K., Langmuir 11,
3862–3866 (1995).
9
6% Ni.
3) The maximal carbon yield was shown to be attained
9. Chambers, A., Nemes, T., Rodriguez, N. M., and Baker, R. T. K.,
J. Phys. Chem. 102, 2251–2258 (1998).
0. Sacco, A., Thacker, P., Jr., Chang, T. N., and Chiang, A. T. S., J. Catal.
85, 224–236 (1984).
1. Baker, R. T. K., Barber, M. A., Harris, P. S., Feates, F. S., and Waite,
R. J., J. Catal. 26, 51–62 (1972).
2. Buyanov, R. A., Afanasjev, A. D., and Chesnokov, V. V., Kinet. Katal.
(
with the catalysts comprising a textural promoter in the
amount of 10% and active particles of 10 to 40 nm average
diameter.
(
4) Comparative studies of the stabilizing effect of vari-
oustexturalpromoters(SiO2, Al2O3, MgO, TiO2, and ZrO2)
demonstrated that the highest carbon yield (375 g) was
obtained with SiO2. Not far worse results were obtained
with the other textural promoters if the NiO precursor was
treated at the optimal temperature. It should be empha-
sized that the method used for preparing these catalytic
systems provides a minor chemical interaction between the
constituents. We think that this feature favors the maximal
carbon yield.
18, 839 (1977).
3. Kuvshinov, G. G., Zavarukhin, S. G., Mogilnikh, Yu. I., and Kuvshinov,
D. G., Khim. Prom. 5, 300–305 (1998).
4. Goncharova, O. V., Avdeeva, L. B., Fenelonov, V. B., Plyasova, L. M.,
Malakhov, V. V., Litvac, G. S., and Vlasov, A. A., Kinet. Katal. 36(2),
293–298 (1995).
2
2
5. Shaikhutdinov, Sh. K., Avdeeva, L. B., Goncharova, O. V., Kochubey,
D. I., Novgorodov, B. N., and Plyasova, L. M., Appl. Catal. A 126,
125–139 (1995).
6. Avdeeva, L. B., Goncharova, O. V., Kochubey, D. I., Novgorodov,
B. N., Plyasova, L. M., and Shaikhutdinov, Sh. K., Appl. Catal. A 141,
1
17–129 (1996).
7. Kuijpers, E. G. M., Tjepkema, R. B., and Geus, J. W., J. Mol. Catal. 25,
41–251 (1984).
8. Tibbetts, G. G., Devour, M. G., and Rodda, E. J., Carbon 25(3), 367–
75 (1987).
ACKNOWLEDGMENTS
2
2
2
3
2
The authors are grateful to Professor V. B. Fenelonov for the adsorption
studies.
3
9. Ermakova, M. A., Ermakov, D. Yu., Kuvshinov, G. G., and Plyasova,
L. M., Kinet. Katal. 5, 796 (1998).
0. Guinier, A., “Theorie et Technique de la Radiocristallographie.”
Dunod, Paris, 1956.
31. Kuvshinov, G. G., Mogilnykh, Yu. I., Kuvshinov, D. G., Ermakov,
D. Yu., Ermakova, M. A., Salanov, A. N., and Rudina, N. A., in
“Symposium on Microscopic Studies of Coal and Carbon, August 22–
27, 1998,” American Chemical Society Division of Fuel Chemistry
Preprints of Symposia, Vol. 43, No. 4, pp. 946–950. Am. Chem. Soc.,
Washington, DC, 1998.
REFERENCES
1
2
3
. Baker, R. T. K., Carbon 27(3), 315–327 (1989).
. Rodriguez, N. M., J. Mater. Res. 8(12), 3233–3250 (1993).
. Kuvshinov, G. G., Avdeeva, L. B., Goncharova, O. V., and Mogilnykh,
Yu. I., in “Extended Abstracts, ACHEMA’94 International Meeting
on Chemical Engineering and Biotechnology, Thermal Process Engi-
neering, Frankfurt on Main, 5–11 June 1994,” paper 9.6., 1994.
. Shaikhutdinov, Sh. K., Zaikovskii, V. I., and Avdeeva, L. B., Appl.
Catal. A 148, 123–133 (1996).
. Fenelonov, V. B., Avdeeva, L. B., Zheivot, V. J., Okkel, L. G.,
Goncharova, O. V., and Pimneva, L. G., Kinet. Katal. 34, 545 (1993).
. Kim, M. S., Rodriguez, N. M., and Baker, R. T. K., in “Synthesis and
Properties of Advanced Catalytic Materials” (E. Iglesia, P. W. Lednor,
4
5
6
32. Noskova, C. P., Ph.D. thesis, Novosibirsk, Boreskov Institute of
Catalysis, 1975.
33. Zhao, Y. X., Bovers, C. W., and Spain, J. L., Carbon 26, 291 (1988).
34. Kepinnski, L., Carbon 30, 949 (1992).
35. Nakayama, T., Arai, V., and Nishiyama, Y., J. Catal. 87, 108 (1984).
D. A. Nagaki, and L. T. Thompson, Eds.), MRS Symp. Proceedings, 36. Weissman, J. G., Ko, E. J., and Wynblatt, P., J. Catal. 125, 9 (1990).
Vol. 368, p. 99, 1994. 37. Lamber, R., Jeager, N., and Schulz-Ekloff, G., Surf. Sci. 227, 268 (1990).
. Molchanov, V. V., Chesnokov, V. V., Buyanov, R. A., and Zaizeva, 38. Rostrup-Nielsen, J. R., J. Catal. 85, 31–43 (1984).
7
8
9
N. A., Kinet. Katal. 39(3), 407–415 (1998).
39. Krivoruchko, O. P., and Zaikovskii, V. I., Kinet. Katal. 39(4), 607–617
. Watson, J. H. L., Leger, A. E., and Oulett, J., Phys. Colloid Chem. 54,
(1998).
9
69 (1950).
. Watson, J. H. L., Vanpu, M., and Lind, S. S., Phys. Colloid Chem. 54,
91 (1950).
40. Speck, J. S., Endo, M., and Dvesselhaus, M. S., J. Cryst. Growth 94(4),
834 (1989).
41. Audier, M., Bowen, P., and Jones, W., J. Cryst. Growth 63, 125 (1983).
3
1
1
1
0. Kim, M. S., Rodriguez, N. M., and Baker, R. T. K., J. Catal. 134, 253– 42. Guinot, J., Audier, M., Coulon, M., and Bonnetain, L., Carbon 19, 95
68 (1992). (1981).
1. Snoeck, J. W., Froment, G. F., and Fowles, M., J. Catal. 169, 240–249 43. Teichner, S. J., Nicolaon, G. A., Vicarini, M. A., and Gardes, G. E. E.,
1997). Adv. Colloid Interface Sci. 5, 245 (1976).
2. Shoeck, J. W., Froment, G. F., and Fowlest, M., J. Catal. 169, 250–262 44. Van Nordstrand, R. A., Kreger, V. E., and Ries, H. E., Jr., J. Phys.
2
(
(
1997).
Colloid Chem. 55, 641 (1951).
1
1
3. Alstrup, J., J. Catal. 109, 241–251 (1988).
45. Bastick, J., Bull. Soc. Chim. Fr. 20, 437 (1953).
4. Zaikovskii, V. I., Chesnokov, V. V., and Buyanov, R. A., Appl. Catal. 46. Kohlschutter, H. W., and Kampf, G., Z. Anorg. Chem. 292, 298 (1957).
8, 41–52 (1988). 47. Iler, R. K., “The Chemistry of Silica.” Wiley, New York, 1979.
3