Peculiarities of nickel oxide structure transformation upon CO hydrogenation: II. Dynamics of phase transformation

Article abstract of DOI:10.1006/jcat.1996.0002

The effect of initial microstructure and morphology of the NiO precursor on morphological peculiarities and catalytic properties of the multiphase catalysts for the CO hydrogenation was studied using in situ X-ray powder diffraction and transmission electron microscopy. Under the CO hydrogenation conditions, the NiO sample with the most developed (100) plane was found to be transformed into the Ni/Ni₃C/NiO catalyst, giving CH₄ as a major product of the reaction. The NiO sample with the most developed (111) plane was transformed into the catalyst containing Ni₃C/NiO and producing equal amounts of CH₄ and CO₂. The catalytic properties of the samples were shown to be interrelated to both the composition of the multiphase catalyst and the morphology of each of the component. A step-by-step comparison between variations in the phase composition and the catalytic activity of the samples during a long-term treatment under the reaction conditions was carried out.

Full text of DOI:10.1006/jcat.1996.0002

JOURNAL OF CATALYSIS 158, 13–22 (1996)
ARTICLE NO. 0002
Peculiarities of Nickel Oxide Structure Transformation
upon CO Hydrogenation
II. Dynamics of Phase Transformation
O. S. Morozova,*,1 G. N. Kryukova,† D. P. Shashkin,* L. M. Plyasova,† and O. V. Krylov*
*Semenov Institute of Chemical Physics of Russian Academy of Sciences, ul. Kosygina 4, Moscow 117334, Russia; and
†Boreskov Institute of Catalysis of Siberian Branch of Russian Academy of Sciences, prospekt Lavrentieva 5, Novosibirsk 630090, Russia
Received January 24, 1994; revised August 15, 1995
the samples during a long-term treatment under the CO
The effect of initial microstructure and morphology of the hydrogenation conditions.
NiO precursor on morphological peculiarities and catalytic
properties of the multiphase catalysts for the CO hydrogenation
was studied using in situ X-ray powder diffraction and trans-
mission electron microscopy. Under the CO hydrogenation con-
ditions, the NiO sample with the most developed (100) plane
was found to be transformed into the Ni/Ni3C/NiO catalyst,
giving CH4 as a major product of the reaction. The NiO sample
with the most developed (111) plane was transformed into the
catalyst containing Ni3C/NiO and producing equal amounts of
CH4 and CO2. The catalytic properties of the samples were
shown to be interrelated to both the composition of the multi-
phase catalyst and the morphology of each of the component.
A step-by-step comparison between variations in the phase
composition and the catalytic activity of the samples during a
long-term treatment under the reaction conditions was carried
The aim of our study is to find the correlation between
the catalytic activity and the phase composition of Ni-
containing catalyst. The major interest involves the effect
of initial microstructure and morphology of NiO on the
morphological peculiarities and catalytic activity of the
multiphase catalyst. Transmission electron microscopy
(TEM) was used for studying the morphology of the
samples.
EXPERIMENTAL
Two samples of NiO, whose properties were described
in detail previously (2), were utilized as precursors. Table
1 summarizes the properties of these oxides.
All the experiments were carried out at atmospheric
pressure using a reaction mixture containing CO and H2
in the ratio of CO : H2 ϭ 1 : 1.5 with an overall flow rate
of 6 ml minϪ1 at the temperatures of 240–285ЊC. As a rule,
the CO conversion did not exceed 10–12%.
out.
1996 Academic Press, Inc.
INTRODUCTION
Oxides of transition metals are commonly used as the
precursors of the catalysts for the reaction of CO hydroge-
nation. Active catalysts containing the metal, carbide, and
oxide phases are formed upon the treatment in the reaction
mixture (1–5). The phase composition, as well as the cata-
lytic activity of the sample, change during the treatment
or under the reaction conditions. In the previous reports
(2–4), we demonstrated the effect of the initial microstruc-
ture and morphology of NiO and Ͱ-Fe2O3 on the regularity
of the solid-phase transformation and morphological pecu-
liarities of the newly formed catalysts. In this work with
the use of in situ X-ray diffraction (XRD) technique, we
carried out a step-by-step comparison between the varia-
Reactor
The flow reactor cell used in the simultaneous catalytic
and XRD tests was similar to that employed in the previous
studies (3, 7). The reactor was a 0.48-ml-vol stainless-steel
chamber housed in an electric furnace and combined on-
line with a gas chromatograph. Temperature was measured
by using a chromel–alumel thermocouple located inside
the reactor cell so that the temperature of the sample was
registered. The gradientless conditions of the experiment
were found before monitoring the catalytic and struc-
tural properties.
tions in the phase composition and the catalytic activity of Experimental Procedure
The experiments were carried out according to the fol-
lowing procedure:
1 To whom correspondence should be addressed.
13
0021-9517/96 $12.00
Copyright 1996 by Academic Press, Inc.
All rights of reproduction in any form reserved.

14
MOROZOVA ET AL.
TABLE 1
Description of Nickel Oxide Samples
Specific surface
1
Sample
NiO(I)
Preparation
Morphological peculiarity
area (m2 gϪ
)
Plasmochemical method (6)
Cubic microcrystals with (100) de-
veloped planes. Mean particle
size is equal to 40 nm, whereas
larger crystals (over 200 nm)
also exist.
20
NiO(II)
Air calcination of NiCO3
(Merk) at 700ЊC for 5 h
Platelets with (111) developed
planes and surface steps on lat-
eral sides. Mean particle size is
equal to 150 nm.
4.9
(i) The unreduced precursor NiO (ca. 0.4 g) was On the basis of special calibrations and balance calcula-
placed into the reactor cell and heated in a flow of He up tions, reactions [1] and [2] were separated. We showed
to the initial temperature of the run.
that ȁ79.3% of CH4 was produced by reaction [1].
(ii) At the given temperature, He was replaced by the
(iiii) When the process of CH4 formation was com-
reaction mixture. In order to ensure steady-state conditions pleted, the reaction mixture was again passed through the
in both the catalytic reaction and the structural transforma- reactor, and the experiment continued as was described
tions, the CO hydrogenation was performed during a pe- in (ii).
riod of about 90 or 300 min, as dictated by the sample and
To study the kinetics of the phase transformation, we
systematically recorded the XRD pattern containing the
characteristic peaks of NiO, Ni, and Ni3C (Figs. 1 and 4).
A scan speed of 1Њ/min was used. A Be(110) peak was
used as an internal standard. The phase composition of
the catalyst was verified by the JSPDC data files. The
relative concentrations of the different phases in the sam-
ple were calculated on the basis of integral intensities of
the characteristic diffraction peaks. The amount of Ni was
determined from the calibration curve obtained for the
mixture of Ni and NiO containing 0.9–100 mol% of the
metal.
temperature of the run. When the steady-state conditions
were achieved, the temperature was quickly (about 1 min)
raised by 15ЊC. Thereafter, the experiment continued at
the next temperature. A composition of the reaction mix-
ture on the outlet of reactor and the phase composition
of the sample were tested simultaneously every 10 min.
Each of the analyses took ȁ10 min.
(iii) The XRD measurements indicated some amount
of Ni3C accumulated in the sample during the catalytic
reaction (Figs. 1a and 4a). To estimate the content of Ni3C,
we used the reaction
TEM investigations of the catalyst were carried out on
the different stages of transformation. A JEM-100 CX
electron microscope with 0.3-nm resolution and accelerat-
ing potential of 100 kV was used. The samples were pre-
pared from ethanol suspension and replaced into the cop-
per grids covered by the amorphous carbon.
Ni3C ϩ 2H2 Ǟ CH4 ϩ 3Ni,
[1]
by which Ni3C was transformed into CH4 and Ni quantita-
tively. The reaction was carried out in the flow of mixture
containing 5% H2 in He. When the steady-state conditions
were achieved, this mixture replaced reaction mixture. Un-
der the treatment, the rate of CH4 formation increased,
passed through the maximum, and then decreased, as the
content of Ni3C decreased (Fig. 3). The disappearance of
the Ni3C peak in the XRD pattern (Figs. 1b and 4b) indi-
cated the completion of the reaction [1]. The integral inten-
sity of the Ni3C diffraction peak was correlated with the
RESULTS
NiO(I) Sample
The treatment of the initial NiO(I) in the reaction mix-
amount of Ni3C. The latter was calculated through the ture at 240ЊC resulted in the partial reduction to Ni after
content of CH4 being formed in the period when the peak ȁ80 min of the induction period. When the phase of Ni
disappeared. Some carbon being accumulated in the sam- was detected in XRD pattern, CH4 was found in the efflu-
ple could also be hydrogenated under the same conditions: ent gas. It took about 125 min to obtain the system con-
taining ȁ4 mol% Ni in NiO. The conversion of CO on
C ϩ 2H2 Ǟ CH4.
[2] the catalyst did not exceed 1%. The products of the CO

NiO STRUCTURE TRANSFORMATION, II
15
FIG. 2. Change in the integral intensities of Ni(220) and Ni3C(103)—
data from the on-line, time-resolved diffractograms in Fig. 1a—and varia-
tion in the rates of CH4 and CO2 formation obtained during the CO
hydrogenation at 255ЊC.
FIG. 1. In situ XRD spectra of NiO(I) obtained on FeKͰ radiation
during the CO hydrogenation at 255ЊC (a), during the treatment in a
mixture containing 5% H2 in He at 270ЊC (b), and during the CO hydroge-
nation at 270ЊC (c).
specific surface area of the total sample remained virtu-
ally unchanged.
The calculation being made on the basis of crystallo-
hydrogenation consisted of CH4 (30%) and CO2(70%); wa- graphic parameters of the unit cell, on average size of the
ter was not analyzed. particle, and specific surface area of the NiO(I) sample
Figure 1a shows the XRD pattern obtained at 255ЊC indicated that the phase transformations at 255ЊC touched
under the reaction conditions. The (103) characteristic the surface and subsurface layers of the oxide particles:
peak of Ni3C at ⍜ ϭ 52.17Њ was defined in the spectra after from a layer of 1 unit cell thick to that of 3 unit cells thick.
some induction period. The kinetic curves in Fig. 2 indicate
At 270ЊC, the carbidization process renewed and
that there were two pathways of NiO transformation: the reached the steady-state conditions in 60 min, when the
reduction and carbidization processes took place simulta- reduction process evolved into the bulk of the oxide parti-
neously until the formation of Ni3C was stabilized.
cles. At the end of the run, the reduction degree was found
During the solid-phase transformations, the rate of CO2 to be ȁ60 mol%. The calculations give the average depth
formation peaked in the first 10–12 min of the treatment of the NiO bulk conversion equal to 6 unit cells.
and then decreased. The rate of CH4 formation increased
The rate of the carbidization reaction may fall and restart
during the first 90 min of the run and then stabilized due to the broad variation in the size of the NiO(I) particles
8
8
1
2
(1.13 ϫ 10Ϫ –1.18 ϫ 10Ϫ mol secϪ mϪ ), as well as the (from 20 to 200 nm). The reduction of the big crystals is
formation of Ni3C. In Table 2 we relate directly the struc- slow, whereas a probability of Ni3C formation is high. The
tural and catalytic properties of the sample at 255ЊC. On reduction of the smaller microcrystals proceeds faster, so
the basis of XRD measurements, we estimated the size of the carbidization of these particles is suppressed because
crystallites and specific surface area for the phase of NiO, the Ni layers form on the surface.
Ni, and Ni3C. The average size of the crystallites and the
Following the solid-phase transformation, the rate of
specific surface area of NiO were found to be ȁ44 nm CH4 stabilized in 60–70 min and later fluctuated in the
1
8
8
1
2
and 20.6 m2 gϪ , respectively. The average size of the Ni range from 3.3 ϫ 10Ϫ to 3.6 ϫ 10Ϫ mol secϪ mϪ . The
crystallites and the specific surface area were found to be rate of CO2 formation decreased fivefold during the run.
1
ȁ40 nm and 20 m2 gϪ , respectively. For Ni3C, the average It seems to be the result of covering the whole particles
size of crystallites and the specific surface area were found of the catalyst by the metal, (100)Ni.
1
to be ȁ33 nm and 16.1 m2 gϪ , respectively. During the
To compare the catalytic properties of the Ni/Ni3C/NiO
experiment, these values varied insignificantly. Thus, the and Ni/NiO samples, we treated the first one in a mixture

16
MOROZOVA ET AL.
TABLE 2
Dynamics of NiO(I) Structural and Catalytic Properties on the First Cycle of Carbidization
1
2
Rate (mol secϪ mϪ
)
Phase
Selectivityb (%)
composition (mol%)
ta
Conversion
108
1010
109
T (ЊC)
(min)
CO (%)
CH4
C2H6
CO2
CH4
C2H6
CO2
Ni
Ni3C
NiO
255
30
60
90
3.5
3.9
4.3
4.9
70.7
74.5
79.9
85.6
4.6
6.0
5.9
6.6
24.4
19.6
14.2
7.9
0.77
1.1
1.13
1.18
3.3
4.4
4.2
4.9
3.3
2.9
2.1
7.0
12.6
17.0
24.0
—
93.0
83.3
75.4
68.3
4.0
7.4
7.7
120
1.11
270
30
60
90
10.0
9.7
10.3
10.0
82.5
85.4
87.3
87.7
6.2
7.0
5.8
8.1
11.3
7.5
6.8
3.15
3.2
3.1
11.9
13.2
9.7
4.3
2.8
2.4
29.4
34.3
37.2
40.5
13.6
17.8
17.7
18.3
57.0
47.9
45.0
41.0
120
4.2
3.5
16.5
1.65
a A time count from the beginning of the run.
b Product distribution is based on the content of CH4, C2H6, and CO2 detected continuously at every run.
containing 5% H2 in He to remove Ni3C. Figure 1b demon-
strates the changes in XRD pattern of the catalyst during
the treatment. The integral kinetic curves describing the
solid-phase reaction and the CH4 formation are shown in
Fig. 3. After the treatment, the catalyst consisted of ȁ70
mol% Ni in NiO. The estimation on the basis of the mass
When the newly formed Ni/NiO catalyst was treated
upon the reaction mixture, Ni3C was detected after 30 min
of the treatment at 270ЊC (Fig. 1c). The rate of carbidiza-
tion at 270ЊC was lower than that observed at 255ЊC for
the initial sample. According to TEM data, the layer of
metallic nickel (100) was formed on the surface of NiO
particles because of the orientation relationship [100]Ni/
[100]NiO (2). It seems that these layers suppressed partially
the carbidization process. The relationship between the
catalytic properties and the phase composition is exhibited
in Table 3. At 255ЊC, the conversion of CO was found to
be almost 2-fold higher, as compared to that for the Ni/
Ni3C/NiO catalyst (Table 2). The distribution of the prod-
ucts was also changed: the rate of C2H6 formation was
found to be almost 10-fold higher and the rate of CO2
formation was found to be almost 10-fold lower than these
observed on the initial sample (Table 2). At the same time,
the rates of CH4 formation were of the same order of
magnitude for the Ni/NiO and Ni/Ni3C/NiO catalysts.
However, the rates of CH4 formation on the Ni/NiO cata-
lyst were virtually constant from the beginning of the run.
When the temperature increased to 270ЊC, the following
changes in catalytic activity were observed: the conversion
of CO, as well as the rates of CH4 and CO2 formation,
increased and the rate of C2H6 formation lowered some-
what. The production of the Ni3C phase was marked by
the drastic increase of the rate of CO2 formation. It was
found to be almost 10-fold higher than that at 255ЊC in
the absence of Ni3C, while the rate of CH4 increased by
1.5–2 times. The catalytic properties of the samples listed
in Tables 2 and 3 became approximately the same at 270ЊC.
All the parameters observed became virtually equal after
the 120-min exposure under the reaction conditions in spite
of the different bulk composition. This result points to the
3
balance indicated that ȁ1 ϫ 10Ϫ g moleculeϪ1 of Ni3C and
4
1
3 ϫ 10Ϫ g atomϪ of carbon were formed in the catalyst
during the experiment.
FIG. 3. Change in the integral intensities of NiO(200), Ni3C(103),
and NiO(222)—data from the on-line, time-resolved diffractograms in
Fig. 1b—and variation in the rate of CH4 formation obtained during the
treatment in the mixture of 5% H2/He at 270ЊC.

NiO STRUCTURE TRANSFORMATION, II
17
TABLE 3
Dynamics of NiO(I) Structural and Catalytic Properties on the Second Cycle of Carbidization
1
2
Rate (mol secϪ mϪ
)
Phase
composition (mol%)
Selectivityb (%)
ta
Conversion
108
109
1010
T (ЊC)
(min)
CO (%)
CH4
C2H6
CO2
CH4
C2H6
CO2
Ni
Ni3C
NiO
255
30
60
90
8.1
7.4
6.8
7.9
79.8
82.6
78.7
73.4
18.3
16.4
20.5
25.9
1.9
1.0
0.4
0.6
1.9
1.8
1.5
1.8
2.2
1.8
2.0
3.2
4.5
2.3
0.8
1.6
75.0
75.3
76.4
—
—
—
25.0
24.7
23.5
120
270
30
60
90
11.9
11.2
11.0
11.4
85.6
85.9
86.5
86.0
10.2
10.5
9.9
4.2
3.6
3.6
4.1
3.4
2.8
3.2
3.3
2.0
1.6
2.0
1.8
16.7
10.9
14.2
14.6
76.3
76.3
75.2
74.6
Traces
1.4
2.3
23.6
22.3
22.5
21.6
120
9.9
3.8
a A time count from the beginning of the run.
b Product distribution is based on the content of CH4, C2H6, and CO2 detected continuously at every run.
fact that the same surface conditions can be reached from The (201) and (112) diffraction peaks of Ni3C were defined
both oxide and metal/oxide initial systems, as was reported after 140 min of treatment. However, these peaks periodi-
in (1) for the iron catalyst.
cally appeared and disappeared on the diffractograms
within 30–50 min from the beginning of the run. During
the induction period, the lattice parameters of NiO in-
creased. This is an indication of the insertion of carbon
NiO(II) Sample
The treatment of the initial NiO(II) sample under reac- into the lattice of oxide. Table 4 summarizes results of
tion conditions at 285ЊC resulted in nothing but Ni3C for- the catalytic and XRD experiments. The initial catalytic
mation, while the phase of Ni was not detected (Fig. 4a). activity was found within 20 min from the beginning of
FIG. 4. In situ XRD specta of NiO(II) obtained on CuKͰ radiation under the treatment at 285ЊC in the mixtures, as follows: CO/H2 (a), 5%
H2/He (b), and CO/H2 (c).

18
MOROZOVA ET AL.
TABLE 4
Dynamics of NiO (11) Structural and Catalytic Properties on the First and Second Cycles of Carbidization at 285؇C
1
2
Rate (mol secϪ mϪ
)
Phase
Selectivityb (%)
composition (mol%)
ta
Conversion
108
109
108
(min)
CO (%)
CH4
C2H6
CO2
CH4
C2H6
CO2
Ni
Ni3C
NiO
The first cycle
A
30
90
0.55
0.86
1.15
1.9
2.2
2.6
23.8
33.1
33.1
33.7
34.8
38.9
40.8
42.8
—
—
—
—
—
—
—
—
66.2
66.9
66.9
66.3
65.2
61.1
59.2
57.2
0.28
0.64
1.0
1.3
1.5
2.0
2.3
2.9
—
—
—
—
—
—
—
—
0.9
1.3
2.3
2.6
2.8
3.2
3.4
3.8
—
—
—
—
—
—
—
—
—
Traces
Traces
0.96
1.04
2.5
100
100
100
120
150
180
210
240
270
B
99.1
99.0
97.5
97.0
96.1
3.0
3.3
3.0
3.9
30
60
90
120
150
180
210
240
270
4.6
5.4
6.5
7.5
8.4
8.3
8.7
9.0
9.7
39.7
43.2
45.8
45.3
46.2
45.8
48.1
48.4
47.6
—
—
—
3.3
3.6
3.5
4.0
4.4
4.4
60.3
56.8
54.2
51.4
50.2
50.6
47.9
47.2
48.0
3.5
4.5
5.5
6.6
7.4
7.3
8.1
8.4
7.9
—
—
5.3
6.0
6.5
7.5
8.1
8.0
8.0
8.2
7.9
—
—
—
—
—
—
—
—
—
3.9
4.0
4.6
4.8
—
5.0
5.1
5.4
5.1
96.1
96.0
95.3
95.2
—
95.0
94.9
94.6
94.9
2.1
2.9
2.8
3.4
3.8
3.6
The Second Cycle
30
60
90
8.5
9.7
10.9
13.3
48.8
44.2
44.8
44.8
8.3
5.4
4.5
4.6
42.9
50.3
50.6
50.6
8.1
8.1
10.1
12.0
1.7
5.9
5.0
6.2
7.1
9.3
11.5
13.5
16.0
15.5
12.9
11.6
0.98
2.6
3.8
82.9
81.9
83.2
82.2
105
6.2
Note. A, the first day under the treatment. B, the continuation of the treatment after 20 h; the catalyst was left at 20ЊC on the He flow.
a A time count from the beginning of the run.
b Product distribution is based on the content of CH4, C2H6, and CO2 detected continuously at every run.
the run, when the phase of Ni3C was not observed. TEM According to XDR results, the system being prepared con-
data also evidenced for the absence of the Ni3C particles tained ȁ18 mol% of Ni. TEM data indicated the polycrys-
in the NiO(II) sample on this activation stage. The conver- talline structure of the metal.
sion of CO and the rate of CH4 formation were found to
As shown in Fig. 4c, the Ni/NiO sample easily trans-
be very low in the first ȁ120 min of the run that indicate formed into the multiphase catalyst Ni3C/Ni/NiO under
a small amount of the active sites on the surface. We have the reaction conditions. The catalytic and structural prop-
no data with which to discuss the nature of these sites. erties are summarized in Table 4. Figure 5 exhibits that a
Nevertheless, the surface steps of the oxide microcrystal direct relationship between the content of Ni3C and the
observed by TEM, inclusions of the active carbon, or the rates of CH4 and CO2 formation is retained. In the presence
small amount of the high dispersed Ni3C seems to be associ- of metallic Ni, the conversion of CO and the rates of CH4
ated with these. When the phase of Ni3C appeared, the and CO2 were found to be several times higher, as com-
conversion of CO, as well as the rates of CH4 and CO2, pared to those in the absence of the metal (Table 4). How-
increased. Figure 5 exhibits a direct relationship between ever, the difference between the properties observed was
the content of Ni3C and the parameters of the catalytic not so drastic, as in the case of the catalysts obtained from
process. After ȁ480 min, when the catalyst contained ȁ5.1 the sample NiO(I).
mol% of Ni3C, nearly steady-state conditions of both the
catalytic and topochemical processes were reached.
Both the CH4 and C2H6 selectivity on the catalyst ob-
tained from NiO(II) was found to be nearly twice as small
In the next stage of the experiment, we obtained a Ni/ as those on the catalyst obtained from NiO(I) because of
NiO system through the treatment of the Ni3C/NiO cata- the relatively high rate of CO2 formation. According to
lyst in the mixture of 5% H2 in He at 285ЊC (Fig. 4b). the data listed in Tables 2–4, this rate was found to be

NiO STRUCTURE TRANSFORMATION, II
19
and (ii) the solid-phase transformations and the catalytic
reaction are interdependent ones because of the participa-
tion of CO and H2 in both of the processes.
The effect of the initial microstructure of the NiO precur-
sor on the catalytic properties can be easily realized
through the interdependence between the topochemical
and catalytic reactions.
Previously, we reported that even for the polycrystalline
powder samples the mechanism of solid-phase transforma-
tion depends on the structural similarity between the crys-
tal lattices of the reagent and the product (2). Thus, due
to the structural similarity between cubic unit cells of NiO
and Ni, a layer of the metal, (100)Ni, was first of all formed
on the surface (100) of the NiO(I) sample under the treat-
ment in the reaction mixture. Dufour et al. (8–10) indicated
a leading role of atomic hydrogen in the reduction of the
(100) plane of NiO. The point defects and Ni atoms with
the low coordination number were found to be the centers
of H2 dissociation. These studies strongly suggest that the
growth of the metallic phase passes through nucleation in
the hexagonal structure near 250ЊC. At the higher tempera-
tures, Ni nucleated in its usual face-centered cubic struc-
ture. The interface between NiO and the newborn Ni phase
is constituted by the (111) planes, which correspond in
both lattices and placed lengthwise on the (111) plane
of NiO.
FIG. 5. The dependence between the content of Ni3C and the rates
of CH4 and CO2 formation on the initial (I) and the additional (II)
carbidization stages.
The interesting feature of the study was the observation
that CO reacts with the (100) plane of NiO only in the
presence of the hydrogen traces. No reduction was de-
tected below 270ЊC in pure CO. Our studies (2) indicated
the key role of CO in the topochemical process: the dissoci-
ative adsorption of CO resulted in the insertion and accu-
mulation of carbon into the bulk of NiO accompanied
by the destruction of NiO(I) particles. As a result, the
multigraded NiO microcrystals were formed.
more 10 to 100 times higher on Sample II than that on
Sample I, whereas the rates of CH4 and C2H6 were of the
same order of magnitude for both of these samples. In
contrast to Sample I, the formation of CO2 was not sup-
pressed in the presence of metallic Ni on Sample II (the
second cycle of carbidization).
On the basis of the model for reduction of the (100)
NiO proposed in (8), our XRD and TEM data reported
in (2, 11, 12), and the aforementioned experimental results,
we formulated the scheme for the solid-phase transforma-
tion of NiO(I) initiated by the interaction with the reagents
DISCUSSION
The careful examination of the experimental results led
us to the following important conclusions: (i) the initial
microstructure and morphology of NiO define the structure
and the catalytic properties of the newly formed catalyst, (CO and H2) (Fig. 6). Through this scheme, we attempted
FIG. 6. A scheme for the solid-state transformation of NiO(I) under reaction conditions.

20
MOROZOVA ET AL.
to illustrate the initial stages of the reduction–carbidization the planes of the Ni3C hexagonal unit cell. Previously (2,
process and to correlate step-by-step the peculiarities of 12), we reported that the orientation similarity between
the topochemical and catalytic reactions.
the parent and carbide phases is the necessary condition
According to (11), the concentration of point defects in for the Ni3C formation. Close inspection of the Ni/NiO
cubic structure of the initial NiO(I) is high. Therefore, system indicates that the interface between Ni and NiO
after the H2 dissociation, the interaction of H2 and CO could be the only plane for the Ni3C formation. It is located
with the oxide could proceed through the stages, as follows: lengthwise on the (111) plane of NiO, corresponds with
the (111) plane of Ni, and has a crystallographic similarity
2H2 ϩ Olat Ǟ H2O ϩ Vlat
[3] with the [0001] base plane of Ni3C (Fig. 6). The formation
of Ni3C requires only the stoichiometric amount of carbon
atoms accumulated on the interface through the diffusion
CO ϩ Olat Ǟ CO2 ϩ Vlat
[4]
CO ϩ Vlat Ǟ Cintrinsic ϩ Olat
CO ϩ Ni Ǟ COadsNi
[5] of Cintrinsic (13). Indeed, the induction period of the carbidi-
zation reaction was actually observed (Fig. 2). When Ni3C
[6]
was defined in X-ray spectra for the first time, its content
COads ϩ COads Ǟ Cintrinsic ϩ CO2,
[7] was in good agreement with the amount of carbon being
accumulated in the catalyst during the experiment. The
where Olat and Vlat are the lattice oxygen and oxygen va- estimation was carried out assuming reaction [7] to be
cancy, respectively. The interaction of H2 with the surface the major contributor of carbon. Figure 6 shows that the
of NiO results in the formation of Ni atoms scattered on stepped surface of the NiO microcrystal is also suited for
the oxide. These atomic species and the oxygen vacancies the formation of Ni3C: according to TEM data, the steps
could be considered as adsorption sites for the CO mole- are placed lengthwise to the (111) plane of NiO and closely
cules (10). Reduction of the surface [3, 4], as well as Cintrinsic related to (111) plane of Ni.
formation [5, 7], also take place. Atomic carbon being
When the phase of Ni3C appeared, the rate of CH4
formed accumulates in the bulk of the oxide (2). It seems increased proportionally to its content until steady-state
to localize on the metal-oxide interface or ‘‘fall’’ into the conditions were achieved. On the basis of TEM data, we
crystal volume with the formation of carbon clusters (2, suppose that steady-state conditions correspond to the mi-
13–15). The fact that the NiO lattice parameters increase crocrystals of NiO(I) covered by the metal, (100)Ni. On
in the first stage of the treatment justifies our assumption. this kind of sample, the formation of Ni3C is suppressed
TEM also indicates the presence of the small carbon clus- because the (100) plane of Ni is devoid of an orientation
ters in the subsurface layers of the NiO(I) microcrystals similarity with the Ni3C planes, and the rate of CH4 forma-
after the treatment at 240ЊC. As was shown by TEM (2), tion stabilizes. On the basis of these data, it is conceivable
the destruction of the perfect microcrystals of NiO(I) into that two types of the centers for the CH4 formation are
the stepped ones took place under the action of carbon presented on the surface. Under steady-state conditions,
clusters. The adsorption and activation of the CO and H2 the major part of the centers seems to be located on the
molecules, as well as the reduction of NiO, seem to proceed (100)Ni surface because the rate of CH4 formation is inde-
faster on the fresh surface of the steps. The activated mole- pendent of the reduction degree. Otherwise, when both
cules take place in both the reduction of NiO and formation reduction and carbidization processes take place, the most
of CH4. Thus, the topochemical process is in competition active centers could be located on the interfaces between
with the catalytic one. When the treatment was carried out the newly formed phases. The unusual hexagonal nuclei
at 240ЊC, the conversion of CO was only a fraction of a of Ni were found to be formed on the interface (100)NiO/
percentage. As the phase of Ni appeared, methane was Ni (10). Under the non-steady-state conditions, these nu-
detected in the products of the reaction, but CO2 was found clei or other unstable but active centers located on the
to be the major one. Until Ni3C was defined, the rate of interface can be responsible for part of the CH4 formation.
CH4 formation increased proportionally to the increase Until the interface Ni/Ni3C/NiO is accessible to the re-
of the metal content. Conversion of CO also increased, agents, the rate of CH4 formation is proportional to the
whereas the rate of CO2 decreased after peaking. This Ni3C content, which is placed on this interface.
effect seems to be associated with the newborn active Ni-
The solid-phase transformation of the NiO(II) sample
containing centers giving the portion of activated H2 mole- under the reaction conditions was described in the previous
cules sufficient for both of the processes: the continued report (2). Because of the crystallographic similarity be-
reduction and the catalytic hydrogenation.
tween the (111) plane of the initial NiO(II) and the [0001]
At 255ЊC, the phase of Ni3C was obtained in the Ni/ base plane of hexagonal crystal lattice Ni3C, only the phase
NiO catalyst. In general, the formation of Ni3C on the of Ni3C was formed. According to TEM results, forming
(100) planes of NiO or Ni is almost impossible due to lack Ni3C microcrystals were placed on the surface of the oxide,
of orientation similarity between these planes and any of as similar-sized platelets. Then, the Ni3C platelets lose con-

NiO STRUCTURE TRANSFORMATION, II
21
tact with the ‘‘parent’’ NiO particle followed by the forma- of Ni3C was restarted immediately under the reaction con-
ditions. In the presence of the metal, the rate of CH4
increased in the beginning of the run, as compared to the
rate of CO2 (Fig. 5). However, when Ni3C appeared, the
ratio between the rates of CO2 and CH4 formation was
found to be as before.
Thus, there is some evidence that CO activation may
take place on the surfaces of the oxide, carbide, and metal,
as well as on carbon inclusions. The carbon inclusions seem
to be used for Ni3C and CO2 formation. Therefore, the
interface between Ni3C and NiO was highly profitable
for CO2.
The surface of the metal, especially (100) Ni, or the Ni
clusters located on the surfaces of oxide or carbide phases
seem to be suitable for H2 activation. According to the data
obtained for Sample II, the CO2 formation was competitive
with the CO hydrogenation when the centers for the H2
activation were deficient. The catalyst with a high content
of the metal, especially (100) Ni, was better suited for
the methane formation. In none of the special prolonged
experiments with the fine nickel powder as a catalyst (aver-
age size of the particles of 35 nm) (20), the deposition of
coke or Ni3C was observed. This phenomenon is associated
with the high concentration of H atoms on the surface and
in the bulk of the catalyst. The concentration could be
enough for the hydrogenation of different forms of CO
adsorbed hindering the carbon and Ni3C production.
tion of the new Ni3C microcrystals.
In spite of the simplicity of the structural transformation
of NiO(II), the process began at a higher temperature and
after a longer induction period as compared to Sample I.
This seems to be associated with two factors: (i) the diffi-
culties in removing oxygen from the (111) plane of NiO
and (ii) the accumulation of the definite amount of Cintrinsic
in the subsurface layer of the oxide microcrystal.
According to TEM results (11), concentration of point
defects in this sample is low; therefore activation of H2
and CO takes place only on the surface steps of the NiO(II)
microcrystals. The concentration of such centers does not
seem to be high. Thus, the removal of oxygen from the
surface of oxide, as well as the accumulation of carbon, is a
slow process. The most likely speculation about the catalyst
state is that the partially reduced surface of the (111) plane
of NiO is promoted with carbon. According to (15, 16),
the rate of dissociative adsorption of H2 on this kind of
surface is low, as compared to that of Sample I. However,
the carbon clusters with a negative charge may be consid-
ered the additional centers of CO adsorption (17).
Under non-steady-state conditions, the major portion of
CO2 was produced by reactions [4] and [7]. It seems that
Ni3C, part of CO2, and CH4 were formed from Cintrinsic
as follows:
,
SUMMARY
Ni3C
Ni
The initial microstructure and morphology of the poly-
crystalline NiO precursor define the phase composition
and the catalytic properties of the multiphase catalysts
obtained for the CO hydrogenation. Under the treatment
in the reaction mixture, the precursor with the (100) most
developed plane was transformed into the Ni/Ni3C/NiO
catalyst containing the cubic microcrystals of NiO covered
by the layer of (100)Ni and polycrystalline Ni3C. The major
product of reaction was found to be CH4. Under similar
conditions, the precursor with the (111) most developed
plane was transformed into the Ni3C/NiO catalyst involv-
ing the NiO platelets covered by ‘‘monocrystals’’ of Ni3C
and separate particles of Ni3C. On this catalyst, CH4 and
CO2 were formed to the same extent. The catalytic proper-
ties were found to be interrelated with the phase composi-
tion of the sample, as well as with the morphology of each
of the phases.
Cintrinsic
H2O
H2
CH4
CH4 + CO2
(18, 19)
As for the water–gas shift reaction, as a contributor of
CO2, the Ni-containing catalyst does not seem to be good
for it.
We recall that both the rates of CO2 and CH4 formation
increased, as the content of Ni3C increased. Under the
non-steady-state conditions, the rate of CO2 exceeded the
rate of CH4. Under steady-state conditions, these rates
were found to be almost equal. The CH4 formation took
place with the formal stoichiometry, as follows:
2CO ϩ 2H2 Ǟ CH4 ϩ CO2.
[8]
ACKNOWLEDGMENT
When the Ni3C/NiO catalyst was transformed into the
Ni/NiO catalyst, the polycrystalline metal did not form a
continuous layer on the surface of NiO(II) due to the
absence of the structural similarity between the (111) plane
of NiO and the cubic unit cell of Ni. Thus, the surface of
oxide was still accessible for the reagents, and formation
The authors are thankful to Prof. A.Ya. Rozovskii (Moscow Institute
of Petrochemical Synthesis) for valuable comments and discussion.
REFERENCES
1. Reymond, J. P., Meriandeau, P., and Teichner, S. J., J. Catal. 75,
39 (1982).

22
MOROZOVA ET AL.
2. Morozova, O. S., Ziborov, A. V., Kryukova, G. N., and Plyasova,
L. M., J. Catal. 144, 50 (1993).
10. Boudriss, A., and Dufour, L.-C., in ‘‘Non-stoichiometric Compounds,
Surfaces, Grain Boundaries, and Structural Defects’’ (J. Nowotny
and W. Weppner, Eds.), p. 311. Kluwer Academic, London/New
York 1989.
11. Kryukova, G. N., Ph.D. thesis, Moscow, 1989.
12. Morozova, O. S., Ziborov, A. V., Kryukova, G. N., and Plyasova,
L. M., J. Solid State Chem. 101, 353 (1992).
13. Darling, G. R., Pendry, J. B., and Joyner, R. W., Surf. Sci. 221,
69 (1989).
14. Zonnevylle, M. C., Geerlings, J. J., and van Santen, R. A., Surf. Sci.
240, 253 (1990).
3. Morozova, O. S., Maksimov, Yu. V., Shashkin, D. P., Shirjaev, P. A.,
Matveev, V. V., Zhorin, V. A., Krylov, O. V., and Kryukova, G. N.,
Appl. Catal. 78, 227 (1991).
4. Morozova, O. S., Parshina, T. M., Sacharov, M. M., Maksimov, Yu.
V., and Krylov, O. V., in ‘‘Proceedings, 4th All Union Conf. on the
Mechanism of Catalytic Reactions, Moscow,’’ Vol. I, pp. 375–379,
1986. [In Russian]
5. Heon Jung, and Thomson, W. J., J. Catal. 139, 375 (1993).
6. Parkhomenko, V., and Tsybulev, P. N., in ‘‘Proceedings, 8th Interna-
tional Symposium on Plasmochemistry, Tokyo,’’ Vol. 4, pp. 2105–
2110, 1987.
7. Shirjaev, P. A., Kushnerev, M. Ya, Shashkin, D. P., and Krylov,
O. V., Kinet. Katal. 23, 1280 (1982).
15. Benziger, J. B., and Madix, R. J., Surf. Sci. 115, 279 (1982).
16. Allen, W. M., Jones, W. E., and Pacey, P. D., Surf. Sci. 220, 193
(1989).
17. Papageno, L., Conti, M., Caputi, L. S., et al., Surf. Sci. 219, L565 (1989).
18. Fan, C.-Z., and Teng, X.-B., J. Catal. (China) 11, 51 (1990).
19. Somorjai, G. A., Catal. Rev. 29, 181 (1983).
20. Petrov, Yu. I., in ‘‘Clusters and Small Molecules,’’ p. 11. Nauka,
Moscow, 1988. [In Russian]
8. Dufour, L.-C., Floquet, N., and De Rosa, B, in ‘‘Reactivity of Solids,’’
Mat. Sci. Monograph 28 A, p. 47. Elsevier, Amsterdam, 1985.
9. Dufour, L.-C., and Perdereau, M., in ‘‘Surface and Near-Surface
Chemistry of Oxide Materials’’ (J. Nowotny and L.-C. Dufour, Eds.),
p. 577. Elsevier, Amsterdam, 1988.