Mesoporous MgO and Ni-MgO prepared by using carboxylic acids

Article abstract of DOI:10.1039/b307144h

MgO and NiO-MgO with large mesopores were prepared by using the corresponding nitrates and carboxylic acids. Their pore structures were characterized by N₂ adsorption, and reduced Ni-MgO samples were used in the liquid-phase hydrogenation of ketone. The mesopore size of MgO was controllable with the alkyl-chain length of the carboxylic acid in the range between 13 and 38 nm. The mesopores are located at the MgO interparticles. In the hydrogenation of 4-heptanone to 4-heptanol, the catalytic activity of the Ni-MgO, which had mesopores at 11 nm, prepared using dodecanoic acid was higher than that of a commercial Raney Ni with mesopores around 4 nm, while the Ni surface of the Ni-MgO was lower than that of a Raney Ni catalyst. At an optimum regulated size of mesopores, the Ni-MgO catalyst would show high catalytic activity satisfying both rapid mass transfer of the reactants and high dispersion of Ni metals on the catalyst surface.

Full text of DOI:10.1039/b307144h

View Article Online / Journal Homepage / Table of Contents for this issue
Mesoporous MgO and Ni–MgO prepared by using carboxylic acids
Shoichi Takenaka, Satoshi Sato,* Ryoji Takahashi and Toshiaki Sodesawa
Department of Materials Technology, Faculty of Engineering, Chiba University,
Yayoi, Inage, Chiba 263-8522, Japan. E-mail: satoshi@faculty.chiba-u.jp
Received 23rd June 2003, Accepted 4th September 2003
First published as an Advance Article on the web 25th September 2003
MgO and NiO–MgO with large mesopores were prepared by using the corresponding nitrates and
carboxylic acids. Their pore structures were characterized by N2 adsorption, and reduced Ni–MgO samples
were used in the liquid-phase hydrogenation of ketone. The mesopore size of MgO was controllable with
the alkyl-chain length of the carboxylic acid in the range between 13 and 38 nm. The mesopores are located
at the MgO interparticles. In the hydrogenation of 4-heptanone to 4-heptanol, the catalytic activity
of the Ni–MgO, which had mesopores at 11 nm, prepared using dodecanoic acid was higher than that
of a commercial Raney Ni with mesopores around 4 nm, while the Ni surface of the Ni–MgO was lower than
that of a Raney Ni catalyst. At an optimum regulated size of mesopores, the Ni–MgO catalyst would show high
catalytic activity satisfying both rapid mass transfer of the reactants and high dispersion of Ni metals on the
catalyst surface.
1. Introduction
2
Experimental
Porous nickel catalysts like Raney nickel and supported-
type of catalysts such as Ni/Al2O3 and Ni/MgO have
been used for the liquid-phase hydrogenation of organic
compounds in organic syntheses.1–13 The pore structure
of the Ni catalysts is also important in the liquid-phase
reaction, because the pore structure affects mass transfer
of reactants.1,4,7,8,13 In addition to Ni surface area, much
attention should be paid to the pore structure of the Ni
catalysts.
We have reported that mesoporous Ni–MgO catalysts are
prepared from a melt of nickel nitrate, magnesium nitrate
and citric acid.13–16 The Ni–MgO showed catalytic activity
being higher than that of a commercial Raney Ni catalyst in
the liquid-phase hydrogenation of ketones.15,16 The pore size
of the Ni–MgO prepared with citric acid is larger than that
of a Raney Ni catalyst. Diffusion coefficient in 10 nm meso-
pores is one-order larger than that at 4 nm.17 The large meso-
pores would be more efficient for liquid-phase hydrogenation
2.1 Preparation of MgO and Ni–MgO
All reagents supplied by Wako Pure Chemical Ltd. were used
without further purification. Mg(NO3)2ꢀ6H2O and Ni(NO3)2ꢀ
6H2O were used as MgO and Ni–MgO sources, and various
monofunctional carboxylic acids (CH3(CH2)nCOOH) with
different alkyl-chain length were used as organic additives.
The molar ratio (Y ) of CH3(CH2)nCOOH to the sum of Mg
and Ni was typically 2.5. The mixture of Mg(NO3)2ꢀ6H2O,
Ni(NO3)2ꢀ6H2O and CH3(CH2)nCOOH was heated at 100ꢁC
for 1 h and then heated at 170ꢁC for 1 h to prepare organic-
inorganic (CH3(CH2)nCOOH–Mg, CH3(CH2)nCOOH–Mg-
Ni) precursors. Mesoporous MgO and NiO–MgO were
obtained by calcination of the precursors at 400ꢁC for 4 h at
the heating rate of 50 K hꢂ1. The resulting NiO–MgO samples
were reduced under hydrogen flow at 500ꢁC for 2 h to prepare
Ni–MgO catalysts.
Another Ni–MgO catalyst was prepared by using citric acid
(CA) according to the procedure described elsewhere.13,14 The
mixture of Mg(NO3)2ꢀ6H2O, Ni(NO3)2ꢀ6H2O and CA was
heated at 70ꢁC, then melted into homogenous solution. The
molten mixture was evacuated under the pressure of 0.7 kPa
in a rotary evaporator at 70ꢁC. The viscosity of the melt
increased gradually, and finally the mixture was solidified.
The solid was heated at 170ꢁC for 1 h and then calcined at
400ꢁC for 4 h at the heating rate of 50 K hꢂ1 to prepare a
NiO–MgO sample. The molar ratio of (Mg + Ni):CA was
1:1, and Ni content was 70 wt.%. The resulting NiO–MgO
sample was reduced under hydrogen flow at 500ꢁC for 2 h to
prepare a Ni–MgO catalyst (abbreviated as NM-CA).
A reference Raney Ni catalyst was prepared by leaching Al
from a commercial alloy (50 wt.% Ni, Wako Pure Chemical
Ltd.) with aqueous NaOH at 50ꢁC.21 The sample was washed
by water at 50ꢁC, ethanol and hexane every three times before
the catalytic reactions.
using solid catalysts.
We have recently prepared mesoporous TiO2
18,19
and
20
ZrO2 with controlled pore size by calcinating a complex of
the corresponding alkoxide and carboxylic acids. The porous
oxides showed narrow pore size distribution in the mesopore
range, and the pore size increased with increasing the
alkyl-chain length of carboxylic acid.
In the present work, first we prepare a pure MgO with
controlled pore size by using magnesium nitrate and car-
boxylic acids. The effects of alkyl-chain length of car-
boxylic acids on the structure of precursor and MgO
are investigated. Second we prepare mesoporous Ni–MgO
catalysts by using the carboxylic acids and the correspond-
ing nitrates. We investigate the catalytic activity for the
liquid-phase hydrogenation of
a ketone over the Ni–
MgO catalysts. The correlation between pore structure
and catalytic activity is also discussed by comparing the
structural and catalytic properties of the present Ni–
MgO catalysts with those of a commercial Raney Ni cat-
alyst and a reference Ni–MgO catalyst prepared with citric
acid.13–16
2.2 Characterization of MgO, Ni–MgO and their precursors
Fourier transform infrared (FT-IR) spectra of the precursors
were obtained on an FT/IR-350 spectrometer (JASCO).
4968
Phys. Chem. Chem. Phys., 2003, 5, 4968–4973
This journal is # The Owner Societies 2003
DOI: 10.1039/b307144h

View Article Online
X-ray diffraction (XRD) experiments were performed on an
M18XHF (Mac Science) using Cu-Ka radiation (l ¼ 0.154
nm) to investigate the structure of the precursors.
N2 adsorption–desorption isotherms of the calcined NiO–
MgO were obtained on an automated physisorption instru-
ment (Omnisorp 100CX, Coulter). Each sample was degassed
at 300ꢁC for 1 h before the measurement. Specific surface area
was calculated from the adsorption branch according to the
Brunauer–Emmett–Teller (BET) method.22 Pore size distribu-
tion was obtained by analyzing the desorption branch accord-
ing to the Dollimore–Heal method,23 which dose not require
any approximation for multi-layer desorption, as used in the
method of Barrett et al.24 Ink-bottle-shaped pores account
for hysteresis in the adsorption–desorption isotherm.25 Pore
size distribution obtained from the adsorption branch provides
information on the maximum size of pores, while that obtained
from the desorption shows the neck size of pores, which is
more important than the maximum size of pores from the
viewpoint of catalysis, where mass transfer of reactant is
restricted. Pore volume was calculated from the nitrogen
adsorption volume at 0.95 < P/P0 < 1.00.
Ni surface area of the reduced samples was calculated from
the amount of chemisorbed H2 at 0ꢁC in a conventional volu-
metric gas adsorption apparatus. Before the measurement,
each sample was reduced by H2 under an initial pressure of
39.5 kPa at 500ꢁC for 2 h, and then the sample was degassed
at 500ꢁC for 1 h to remove hydrogen adsorbed on the sample.
The amount of chemisorbed hydrogen was estimated by the
Langmuir adsorption equation for the hydrogen uptake. Ni
surface area of the samples was calculated by assuming that
chemisorption stoichiometry of H/Ni was 1, and that Ni
surface area was occupied by H atoms equaling 0.065 nm2.26
Stepwise temperature-programmed reduction (TPR) profiles
of the NiO–MgO samples were obtained by using a conven-
tional apparatus under the H2/N2 (containing 10 vol% H2)
flow rate of 10 cm3 minꢂ1 at the heating rate of 5 K minꢂ1
.
In the first step, the temperature was raised from 25ꢁC to
500ꢁC and kept at 500ꢁC for 2 h. In the second step, the tem-
perature was raised again to 900ꢁC. The reducibility of NiO–
MgO samples was estimated from the proportion of H2
consumption in the first step to the total H2 consumption.14
Fig. 1 IR spectra of the precursors heated at 170ꢁC. (A) The n in
CH3(CH2)nCOOH ¼ (a) 10, (b) 12, (c) 14, (d) 16 at CH3(CH2)n-
COOH/Mg (Y) ¼ 2.5. (B) CH3(CH2)16COOH/Mg ¼ (e) 0, (f) 0.5,
(g) 1.5, (h) 2.0, (i) 2.5.
Several absorption peaks relating to CH3(CH2)nCOOH are
observed in the precursors. The sharp peaks observed at
2850–2950 cmꢂ1 are assigned to a stretching vibration of C–
H in CH3(CH2)nCOOH. The broad peak observed around
3400 cmꢂ1 is assigned to a stretching vibration of O–H in
the mixture of CH3(CH2)nCOOH and Mg(NO3)2ꢀ6H2O. The
peak observed at 1700 cmꢂ1 is assigned to a stretching vibra-
2.3 Catalytic reaction
Liquid-phase hydrogenation of 4-heptanone to 4-heptanol was
performed over Ni catalysts. A mixed solution (10 cm3) of 4-
heptanone (9.6 mmol) and a solvent hexane was added to
the glass vessel with a Ni catalyst reduced under H2 flow con-
dition. Hydrogenation was performed in a stirred pressure
batch reactor (Taiatsu Techno Co., TVS-1) under the H2 pres-
sure of 1.1 MPa, the stirring speed of 1350 rpm, and the reac-
tion temperature of 0ꢁC. The reaction solution was analyzed
by FID-GC with a fused silica capillary column of TC-WAX
(30 m) at 30ꢁC.
The reaction rate of the liquid-phase hydrogenation was
calculated from the conversion of 4-heptanone to 4-heptanol.
The conversion data were fitted to the integral form of first-
order rate, kt ¼ ln[1/(1 ꢂ X)], where k is the first-order reaction
rate constant, t is process time and X is conversion of ketone.27
The rate constant, k, was divided by the catalyst weight and Ni
surface area to obtain the first-order rate constant per unit
weight of catalyst, kw , and per unit Ni surface area, ks ,
respectively.
=
tion of C O in CH3(CH2)nCOOH. Two absorption peaks
assigned to a bidentate carboxylate anion (RCOOꢂ) are
observed at 1400-1500 cmꢂ1 and 1500–1600 cmꢂ1 in the pre-
cursors.
Fig. 1B shows the IR spectra of the MgO precursors pre-
pared at different CH3(CH2)16COOH content. The molar ratio
(Y ) of CH3(CH2)nCOOH to Mg metal affects the coordination
of RCOOꢂ with Mg. At Y < 2 (Fig. 1e–g), the absorption
peaks assigned to a stretching vibration of N–O in Mg(NO3)2
were observed at 1390 and 1640 cmꢂ1, while the peaks disap-
pear at Y ꢃ 2. At Y ꢃ 2, two absorption peaks assigned to a
bidentate RCOOꢂ are observed at 1400–1500 cmꢂ1 and
1500–1600 cmꢂ1. The IR results of the precursor indicate that
carboxylic acid acts as a bidentate RCOOꢂ ligand and coordi-
nates with a Mg atom to form a 2:1 CH3(CH2)nCOOH–Mg
complex.
Fig. 2 shows the XRD patterns of the MgO precursors pre-
pared with different carboxylic acids. The diffraction peaks of
lamellar mesophase are observed in the precursors. The inter-
planar distance of lamellar mesophase in the precursors
sequentially increases with increasing n (Table 1). No diffraction
3. Results
3.1 MgO
Fig. 1A shows the IR spectra of the MgO precursors prepared
with CH3(CH2)nCOOH (n ¼ 10–16, Ni content: 0 mol%).
Phys. Chem. Chem. Phys., 2003, 5, 4968–4973
4969

View Article Online
Fig. 2 XRD patterns of the precursors heated at 170ꢁC. The n in
CH3(CH2)nCOOH ¼ (a) 10, (b) 12, (c) 14, (d) 16. Y ¼ 2.5.
Fig. 3 Pore size distribution of the MgO calcined at 400 ꢁC. The n in
CH3(CH2)nCOOH ¼ (a) 10, (b) 16. Y ¼ 2.5.
peak of lamellar mesophase was observed in the MgO
prepared by calcinating the CH3(CH2)nCOOH–Mg precursor
at 400ꢁC (patterns not shown).
in the first step and shows a broad one in the second step
(Fig. 6b, c). The reducibility of the NiO–MgO estimated from
the proportion of H2 consumption in the first step to the total
H2 consumption is listed in Table 2. The reducibility increases
with increasing Ni content.
Fig. 7 shows the Ni surface area and the rate constant (kw)
for the liquid-phase hydrogenation of 4-heptanone over the
Ni–MgO catalysts prepared at n ¼ 10. Ni surface area gradu-
ally increases with increasing Ni content maximizing at the Ni
content of 80 mol%, and it decreases steeply at the Ni content
of 90 mol%. The kw is small up to the Ni content of 50 mol%,
and it steeply increases showing a maximum at the Ni content
of 60 mol%.
Fig. 8 shows the pore size distribution of a Raney Ni (Raney
50), the NiO–MgO prepared with citric acid (NM-CA), the
NiO–MgO prepared at n ¼ 10 (NM-10) and the NiO–MgO
prepared at n ¼ 16 (NM-16). Raney 50 shows extremely nar-
row pore size distribution at 4 nm. NM-CA also shows narrow
pore size distribution at 4 nm, but the distribution is broader
than that of Raney Ni to have some mesopores whose size is
Fig. 3 shows typical adsorption–desorption isotherms of
MgO prepared with CH3(CH2)nCOOH. And Fig. 4 shows
the pore size distribution of the MgO prepared with
CH3(CH2)nCOOH. Each MgO shows narrow pore size distri-
bution in the mesopore range, and the pore size sequentially
increases from 13.2 to 37.5 nm with increasing n. Both the spe-
cific surface area and pore volume of the MgO prepared at dif-
ferent n are maximized at n ¼ 12 (Table 1). Table 1 lists the
crystallite size estimated from XRD peak broadening at
2y ¼ 42.8ꢁ of MgO (200) plane according to the Scherrer for-
mula.28 Little difference is observed in the crystallite size of the
MgO prepared with CH3(CH2)nCOOH.
3.2 NiO–MgO
Fig. 5 shows the pore size distribution of the NiO–MgO pre-
pared with CH3(CH2)10COOH. The pure MgO (Ni content:
0 mol%) shows narrow pore size distribution in the mesopore
range as described above (Fig. 4), and the pore structure is
maintained after the incorporation of Ni up to the Ni content
of 80 mol%. Both the specific surface area and pore volume of
the NiO–MgO decrease with increasing Ni content (Table 2).
Fig. 6 shows the stepwise TPR profiles of the NiO–MgO
prepared with CH3(CH2)10COOH.The pure MgO shows little
H2 consumption (Fig. 6a). The unsupported NiO (Ni content:
100 mol%) shows a reduction peak in the first step (the reduc-
tion temperature <500 ꢁC) and shows no peak in the second
step ( > 500 ꢁC) (Fig. 6d). The NiO–MgO shows some peaks
Table 1 Structural properties of MgO prepared with different
CH3(CH2)nCOOH
n
SAa /m2 gꢂ1
PVa /cm3 gꢂ1
PSb /nm
Dc /nm
d
d /nm
10
12
14
16
20
175
193
171
154
131
0.46
0.51
0.47
0.48
0.40
13.2
14.9
14.9
19.5
37.5
8.0
8.1
2.6
2.7
2.9
3.1
5.7
8.9
8.8
10.7
a
b
SA, specific surface area; PV, pore volume. Pore size: Peak in the
c
pore size distribution in Fig. 4. Crystallite size estimated using the
Scherrer formula from XRD peak broadening at 2y ¼ 42.8ꢁ (patterns
not shown). d Interplanar distance of lamellar mesophase in the MgO
precursors.
Fig. 4 Pore size distribution of the MgO calcined at 400 ꢁC. The n in
CH3(CH2)nCOOH ¼ (S) 10, (N) 14, (L) 16, (c) 20. Y ¼ 2.5.
4970
Phys. Chem. Chem. Phys., 2003, 5, 4968–4973

View Article Online
Fig. 5 Pore size distribution of the NiO–MgO prepared with
CH3(CH2)10COOH. The Ni content (mol%) ¼ (S) 0, (N) 50, (L) 80,
(c) 100.
Fig. 6 Stepwise TPR profiles of the NiO–MgO prepared with
CH3(CH2)10COOH for (a-d) and citric acid (CA) for (e). The Ni
content (mol%) ¼ (a) 0, (b) 60, (c) 80, (d) 100, (e) 61.6 (70 wt.%).
> 4 nm. Both NM-10 and NM-16 show narrow pore size
distribution at > 10 nm.
mesophase ordered in the precursors. However, no ordered
mesophase was observed in the calcined TiO2 and ZrO2 . The
metal oxides consisted of the aggregates of crystallites, and
the mesopores were located at the intercrystallites. In contrast,
SnCl2 and CH3(CH2)nCOOH did not form complexes like the
mesophase, and the pore size of the calcined SnO2 was con-
trolled by only altering the calcination temperature of the
precursors.29 Controllability of the pore size of the resulted
metal oxides is summarized in Table 4.
Table 3 summarizes the structural and catalytic properties of
a Raney Ni and the Ni–MgO catalysts. Raney 50 shows the
largest Ni surface area among these samples. The Ni surface
area of NM-CA is larger than that of NM-10 and NM-16. In
the hydrogenation of 4-heptanone, the rate constant, ks , over
NM-10 is the highest of all. The ksvalues of NM-CA and
NM-16 are much larger than that of Raney.
In the present MgO precursor, a similar complex between
CH3(CH2)nCOOH and Mg was observed. Two absorption
peaks assigned to RCOOꢂ are observed in the IR spectra of
precursors (Fig. 1A). Two CH3(CH2)nCOOH molecules are
coordinated with an Mg atom to form a 2 : 1 CH3(CH2)n-
COOH–Mg complex (Fig. 1B). The XRD results (Fig. 2) also
indicate that the CH3(CH2)nCOOH–Mg complexes form
lamellar mesophase in the precursor.
4. Discussion
4.1 Effects of alkyl-chain length of carboxylic acid on the
structure of MgO and its precursor
In our previous work, we have reported the preparation of
20
and ZrO2 prepared with the corres-
18,19
mesoporous TiO2
No lamellar mesophase was observed in MgO calcined at
400 ꢁC. It is probable that the MgO consists of the aggregates
of crystallites and the mesopores are located at the intercrystal-
lites. The pore size of the calcined MgO increases with increas-
ing n (Fig. 4). Since no significant difference, except for n ¼ 20,
is observed in the crystallite size of the MgO prepared at
ponding metal alkoxides and carboxylic acid. Two strong
absorption peaks assigned to a carboxylate anion (RCOOꢂ)
were observed at 1400–1500 cmꢂ1 and 1500–1600 cmꢂ1 in
the IR spectra of the precursors, and we concluded that
CH3(CH2)nCOOH acted as a bidentate carboxylate ligand
and was coordinated with Ti and Zr to form 2 : 2 and 2 : 1
CH3(CH2)nCOOH-metal complexes, respectively. XRD
experiments also indicated that the complex formed
Table 2 Structural and catalytic properties of Ni–MgO prepared with
CH3(CH2)10COOH.
Ni content SAa /
PVa /
PSb / Red.c Ni SAd / kw
/
e
(mol%)
m2 gꢂ1 cm3 gꢂ1 nm
(%)
m2 gꢂ1
minꢂ1 gꢂ1
0
30
175
133
105
102
103
88
0.46
0.33
0.27
0.3
13
12
15
12
11
15
12
—
—
0
—
—
0
—
—
0.01
0.02
0.05
0.29
0.26
0.13
0.03
0.01
40
50
6.8
11.9
17.2
21.4
25.2
14
40
60
70
0.3
54
59
0.26
0.23
0.1
80
90
72
35
71
93
100
17
0.09
100
—
a
b
SA, specific surface area; PV, pore volume. Pore size: Peak in the
pore size distribution in Fig. 5. c Reducibility estimated from the TPR
d
data in Fig. 6. Ni surface area calculated from the amount of H2
e
adsorbed on the reduced sample. First-order rate constant per unit
weight of catalyst in the hydrogenation of 4-heptanone.
Fig. 7 (a) Ni surface area and (b) first-order rate constant (kw) for
the hydrogenation of 4-heptanone over Ni–MgO catalysts prepared
at n ¼ 10.
Phys. Chem. Chem. Phys., 2003, 5, 4968–4973
4971

View Article Online
4.3 Diffusion of reactants
In liquid-phase hydrogenation, hydrogen transfer from the gas
phase to liquid phase depends on stirring speed, H2 pressure,
reaction temperature and so forth.4,6–8,11,15,16 Woerde et al.
claimed that the reaction rate was limited by the pore diffusion
of dissolved H2 .2 We have investigated the effects of stirring
speed on the conversion of 4-heptanone in the liquid-phase
hydrogenation over Ni–MgO catalysts.15 The conversion
gradually increased with increasing stirring speed up to 1200
rpm, above which it was constant. This indicated that gas–
liquid diffusion of hydrogen was removed by the stirring speed
above 1200 rpm. We also investigated the effects of H2pressure
on the initial rate in the hydrogenation of 4-heptanone over
Ni–MgO catalysts.16 The initial rate increased with increasing
H2pressure up to 0.8 MPa, above which it was constant. In the
present system, therefore, the liquid-phase hydrogenation of
ketones was performed under the H2 pressure of 1.1 MPa,
the stirring speed of 1350 rpm and the reaction temperature
of 0 ꢁC.
In heterogeneous reactions over porous catalysts, either sur-
face reaction or pore diffusion of reactants could be the major
process that determines the catalytic activity.1,25,31,32 In the
liquid-phase reactions, pore diffusion is often competitive with
surface reaction. Nitta et al. confirmed that pore diffusion of
reactants in the pore size of < 10 nm was the rate-determining
step in the liquid-phase hydrogenation of methyl acetoacetate
over supported Ni catalysts.3 Since the mass transfer of
H2molecules is generally faster than that of organic reactants
in liquid-phase reactions,33 more attention should be paid to
the effects of pore diffusion of reactants.
Fig. 8 Pore size distribution of (S) Raney, (N) NM-CA, (L) NM-10,
(c) NM-16.
different n (Table 1), the pore size control of the MgO is appar-
ently caused by the difference in the aggregation density of the
MgO crystallites. As n is longer, the crystallites of the MgO
aggregate more loosely to increase pore size. We consider that
the formation of lamellar mesophase in the precursor affects
the aggregation behavior of the MgO crystallites to control the
pore size. The alkyl-chain length, n, is the only one parameter
that can control the mesopore size of MgO (Table 4) in the
same way as ZrO2 .20
Pore diffusion is severely restricted in the mesopore range:17
pore diffusion coefficient of solute at 4 nm is 1/10 as small as
that at 10 nm, which is less than 1/10 of that of free diffusion
of solvent. The pore diffusion coefficient is affected by the
kinds of solvent, and directly affects the reaction rate in small
mesopores.
4.2 Effects of Ni content on the structure of NiO–MgO
Parmaliana et al. have investigated the effects of Ni content on
the reduction of NiO in Ni–MgO catalysts by TPR.30 The TPR
experiments revealed the presence of several forms of NiO.
NiO reduced at lower temperature was located on the surface
of MgO where NiO had little interaction with MgO, and NiO
reduced at higher temperature was located in the MgO lattice
to form NiO–MgO solid solution. The reducibility of the NiO–
MgO increased with increasing Ni content. It was suggested
that strong interaction between NiO and MgO raised at low
Ni content.
In the present NiO–MgO system, several NiO forms are con-
firmed by TPR (Fig. 6). The peaks observed in the first step is
attributed to the reduction of NiO located on the surface of the
MgO to be reduced easily at the temperature <500 ꢁC, and the
peak in the second step can be assigned to the reduction of
NiO located inside the MgO to form NiO–MgO solid solution.
The results obtained in TPR experiments agree with those
reported previously.30 In the TPR profiles (Fig. 6), both
NM-10 and NM-CA show some peaks in the first step and
show a broad peak in the second step. Both the reducibility
and Ni surface area of NM-CA are higher than those of
NM-10 (Table 3).
4.4 Correlation between pore structure and catalytic activity
of Ni–MgO catalysts
The catalytic activity of Raney 50, which has the highest Ni
surface area and the smallest pore size among these samples,
is much lower than those of Ni–MgO samples in the hydroge-
nation of 4-heptanone (Table 3). It appears that the reaction
rate is determined by the pore diffusion rather than the surface
reaction in the hydrogenation of 4-heptanone.
The catalytic activity of NM-10 is the highest, while the Ni
surface area of NM-10 is lower than that of NM-CA (Table 3)
and the pore size of NM-10 is larger than that of NM-CA (Fig.
8). The large mesopores at 10 nm in NM-10 are probably effec-
tive in rapid mass transfer of 4-heptanone to show high cata-
lytic activity. The surface reaction and the pore diffusion are
probably competitive in the hydrogenation of 4-heptanone
over NM-CA. In other words, the large mesopores of NM-
10 appear to be effective in rapid mass transfer of 4-heptanone
to compensate for the lower Ni surface area.
Table 3 Structural and catalytic properties of Raney Ni and Ni–MgO catalysts.
Sample
Ni content/wt.%
SA/m2 gꢂ1
PV/cm3 gꢂ1
PSa /nm
Red. b (%)
Ni SA/m2 gꢂ1
ksc /minꢂ1 mꢂ2
Raney Ni
NM-CAd
NM-10e
NM-16f
100
70
69g
69g
176
115
103
69
0.12
0.13
0.3
4
4
—
76
54
—
39
23
17
7
0.002
0.012
0.017
0.007
11
15
0.27
a
c
Pore size: Peak in the pore size distribution in Fig. 8. b Reducibility estimated from the TPR data in Fig. 6. First-order rate constant per unit
d
Ni surface area in the hydrogenation of 4-heptanone. Ni–MgO catalyst prepared with citric acid (CA). Ni–MgO catalyst prepared with
e
f
g
CH3(CH2)10COOH at Y ¼ 2.5. Ni–MgO catalyst prepared with CH3(CH2)16COOH at Y ¼ 2.5. 60 mol%.
4972 Phys. Chem. Chem. Phys., 2003, 5, 4968–4973

View Article Online
Table 4 Controllablity of mesopore size of oxides prepared by using
carboxylic acid
2
3
H. M. Woerde, L. J. Bostelaar, A. Hoek and W. M. H. Sachtler,
J. Catal., 1982, 76, 316.
Y. Nitta, O. Yamanishi, F. Sekine, T. Imanaka and S. Teranishi,
J. Catal., 1983, 79, 475.
Controllable pore size/nm
4
5
6
Y. Nitta, T. Imanaka and S. Teranishi, J. Catal., 1983, 80, 31.
M. A. Keane and G. Webb, J. Mol. Catal., 1992, 73, 91.
F. Notheisz, A. Zsigmond, M. Bartok, Z. Szegletes and G. V.
Smith, Appl. Catal. A, 1994, 120, 105.
U. K. Singh, R. N. Landau, Y. Sun, C. LeBlond, D. G.
Blachmond, S. K. Tanielyan and R. L. Augustine, J. Catal.,
1995, 154, 91.
Y. Sun, J. Wang, C. LeBlond, R. N. Landau and D. G.
Blachmond, J. Catal., 1996, 161, 759.
S. Xie, H. Li, H. Li and J.-F. Deng, Appl. Catal. A, 1999,
189, 45.
Precursor
Sample complex
a
n
Y
Tc
Ref.
18,19
20
TiO2
ZrO2
SnO2
MgO
2 : 2 (lamellar) 6.5–22
2 : 1 (hexagonal) 4.8–7.5
4.8–14
ꢄ
7
ꢄ
ꢄ
ꢄ
ꢄ
29
—
2 : 1 (lamellar)
ꢄ
6.5–26
ꢄ
13–38
This work
8
9
a
Coordination molar ratio of carboxylic acid to metal in the car-
boxylic acid-metal complex, and the structure of the complex is shown
in the parenthesis. n: CH3(CH2)nCOOH, Y ¼ CH3(CH2)nCOOH/
metal, Tc : calcination temperature. ꢄ, uncontrollable.
10 A. Louloudi and N. Papayannakos, Appl. Catal. A, 2000,
204, 167.
11 C. V. Rode, M. J. Vaidya, R. Jaganathan and R. V. Chaudhari,
Chem. Eng. Sci., 2001, 56, 1299.
In contrast, NM-16 shows the largest pore size at 15 nm
(Fig. 8). The catalytic activity, however, is lower than that of
NM-10 (Table 3). Thus, much larger mesopores could have
disadvantage to the reactants to approach the active sites.
There would be an appropriate mesopore size having high
probability to contact reactants with the catalyst surface.
12 D. Tichit, R. Durand, A. Rolland, B. Coq, J. Lopez and P.
Marion, J. Catal., 2002, 211, 511.
13 T. Nakayama, K. Yamashiro, S. Sato and F. Nozaki, Appl. Catal.
A, 1997, 151, 437.
14 T. Nakayama, N. Ichikuni, S. Sato and F. Nozaki, Appl. Catal. A,
1997, 158, 185.
15 S. Sato, R. Takahashi, T. Sodesawa, F. Nozaki, X.-Z. Jin,
S. Suzuki and T. Nakayama, J. Catal., 2000, 191, 261.
16 H. Tsai, S. Sato, R. Takahashi, T. Sodesawa and S. Takenaka,
Phys. Chem. Chem. Phys., 2002, 4, 3537.
5. Conclusion
17 R. Takahashi, S. Sato, T. Sodesawa and T. Ikeda, Phys. Chem.
Chem. Phys., 2003, 5, 2476.
18 R. Takahashi, S. Takenaka, S. Sato, T. Sodesawa, K. Ogura and
K. Nakanishi, J. Chem. Soc. Faraday Trans., 1998, 94, 3161.
19 S. Takenaka, R. Takahashi, S. Sato and T. Sodesawa, J. Sol–Gel
Sci. Technol., 2000, 19, 711.
20 S. Takenaka, R. Takahashi, S. Sato, T. Sodesawa and T.
Uematsu, J. Ceram. Soc. Jpn, 2003, 111, 16.
21 J. Freel, W. J. M. Pieters and R. B. Anderson, J. Catal., 1996,
14, 247.
22 S. Brunauer, P. H. Emmett and E. Teller, J. Am. Chem. Soc.,
1938, 60, 309.
23 D. Dollimore and G. R. Heal, J. Appl. Chem., 1964, 14, 109.
24 E. P. Barrett, L. G. Joyner and P. P. Halenda, J. Am. Chem. Soc.,
1951, 73, 373.
25 J. W. McBan, J. Am. Chem. Soc., 1935, 57, 699.
26 C. H. Bartholomew and R. J. Farrauto, J. Catal., 1976,
45, 41.
27 J. M. Smith, Chemical Engineering Kinetics, McGraw-Hill,
New York, 3rd edn.,1981, p. 62.
We prepared MgO and NiO–MgO with large mesopores by
using the corresponding nitrates and carboxylic acids. The
resulting MgO has mesopores controlled in the range between
13 and 38 nm. The MgO consists of the aggregates of crystal-
lites and the mesopores are located at the intercrystallites. It is
advantageous that we can use a nitrate instead of using
expensive metal alkoxides in this process.
In the liquid-phase hydrogenation of 4-heptanone to 4-hep-
tanol, the catalytic activity of the Ni–MgO, which had meso-
pores at 11 nm, prepared using dodecanoic acid was higher
than those of a commercial Raney Ni with mesopores at 4
nm and of another Ni–MgO with mesopores around 16 nm,
which was prepared using stearic acid. Although the Raney
Ni catalyst had the highest Ni surface among the Ni catalysts
we tested, it did not work efficiently. The pore size of the Ni–
MgO catalyst was larger than that of a Raney Ni catalyst. We
consider that the large mesopores around 10 nm are effective in
rapid mass transfer of the reactants to show high catalytic
activity.
28 P. Scherrer, Gottinger Nachrichten, 1918, 2, 98.
29 S. Takenaka, R. Takahashi, S. Sato, T. Sodesawa, F. Matsumoto
and S. Yoshida, Microporous Mesoporous Mater., 2003, 59, 123.
30 A. Parmaliana, F. Arena, F. Frusteri and N. Giordano, J. Chem.
Soc. Faraday Trans., 1990, 86, 2663.
31 R. J. Madon and M. Boudart, Ind. Eng. Chem. Fundam., 1982,
21, 438.
32 R. I. Masel, Chemical Kinetics and Catalysis, John Wiley & Sons,
New York, 2001, p. 738.
33 M. Boudart and W.-C. Cheng, J. Catal., 1987, 106, 134.
References
1
P. N. Rylander, Catalytic Hydrogenation in Organic Syntheses,
Academic Press, New York, 1979, p. 1.
Phys. Chem. Chem. Phys., 2003, 5, 4968–4973
4973