Nickel-calcium phosphate/hydroxyapatite catalysts for partial oxidation of methane to syngas: Characterization and activation

Article abstract of DOI:10.1016/j.jcat.2003.07.004

Nickel-calcium phosphate/hydroxyapatite catalysts showed high activity and selectivity in the partial oxidation of methane. The characteristics of these catalysts were examined. Very fine, needle-shaped nickel particles of a few nanometers were present in the sample after the reaction. The crystalline phases observed were calcium phosphate, calcium hydroxyapatite, NiO, and metallic Ni. As the Ni/PO₄ ratio increased, the amount of NiO increased in the fresh samples and the amount of metallic Ni increased in the used (after-reaction) catalysts with decreasing NiO. Most of the catalytic activity was due to the fine Ni metal particles. The amount of H₂ chemisorption was very small (the percentage exposed of Ni < 1%) while the amount of O₂ sorption was quite large. The suppression of H₂ chemisorption might be due to the covering of the Ni particles with the phosphate and/or hydroxyl groups or to a strong interaction with those groups.

Full text of DOI:10.1016/j.jcat.2003.07.004

Journal of Catalysis 221 (2004) 178–190
www.elsevier.com/locate/jcat
Nickel–calcium phosphate/hydroxyapatite catalysts for partial oxidation
of methane to syngas: characterization and activation
a
b
c
c
c
Jin Hyuk Jun, Tae-Jin Lee, Tae Hoon Lim, Suk-Woo Nam, Seong-Ahn Hong,
a,∗
and Ki June Yoon
a
Department of Chemical Engineering, Sungkyunkwan University, Suwon 440-746, South Korea
b
School of Chemical Engineering & Technology, Yeungnam University, Kyongsan 712-749, South Korea
c
Battery and Fuel Cell Research Center, Korea Institute of Science and Technology, Seoul 136-791, South Korea
Received 12 May 2003; revised 16 July 2003; accepted 21 July 2003
Abstract
Nickel–calcium phosphate/hydroxyapatite catalysts exhibited high activity and selectivity in the partial oxidation of methane (POM).
Characterizations have been performed in order to understand the characteristics of these catalysts. TEM showed that very fine, needle-
shaped nickel particles of a few nanometers were present in the sample after the reaction. TPR showed the presence of three nickel species
in the fresh (before-reaction) sample, confirming that the nickel in the calcium hydroxyapatite and phosphate structures comes out as fine
particles and is reduced during the reaction. The crystalline phases observed were calcium phosphate, calcium hydroxyapatite, NiO, and
metallic Ni. As the Ca/PO ratio increased, the calcium hydroxyapatite phase tended to increase over the calcium phosphate phase. As the
4
Ni/PO ratio increased, the amount of NiO increased in the fresh samples and the amount of metallic Ni also increased in the used (after-
4
reaction) catalysts with the decrease of the NiO. The catalytically active component is certainly the metallic nickel that is produced under
reducing environment during the reaction, but its surface is totally reoxidized at 673 K and loses the activity completely. The catalyst could
be activated by the reactants only, without the H pretreatment. The first activation could be achieved at 923 K or below, and the subsequent
2
activation at 723 K.

2003 Elsevier Inc. All rights reserved.
Keywords: Nickel; Partial oxidation of methane; Calcium hydroxyapatite; Calcium phosphate; Syngas
1
. Introduction
tion produces significant quantities of carbon dioxide. Com-
pared with the highly endothermic steam reforming, POM is
mildly exothermic and does not require high operating pres-
sures and hence is more energy efficient. The reaction can be
carried out with lower investment and will produce a smaller
Hydrogen is regarded as the ultimate clean power source
of the future. Its application to fuel cells is a typical example.
Catalytic partial oxidation of methane (POM) for synthe-
sis gas and hydrogen production has been an active sub-
ject in recent years because it is currently more economical
and feasible than other methods such as water electrolysis
and photodecomposition. The process arouses renewed in-
terest as a promising alternative to the steam reforming of
methane [1–14]. Although the steam reforming of methane
is the dominant commercialized process for syngas produc-
tion, it has some drawbacks relating to the energy and capital
costs since it is a highly endothermic reaction and is operated
under high pressure. In addition, the water–gas-shift reac-
amount of CO . Recently, POM is of particular interest in
2
fuel processing applications such as in fuel cells for startup
from cold, owing to its exothermic nature [3,8]. However, if
POM is to be applied to low-temperature fuel cells such as
polymer electrolyte membrane fuel cells, subsequent treat-
ments such as conversion of CO to CO2 by the water–gas-
shift reaction and removal of trace CO by the preferential
oxidation of CO are required because CO severely poisons
the electrode.
There have been numerous studies on POM over vari-
ous metal catalysts. Ni and noble metals such as Rh, Pt,
and Pd are active for POM, and these are usually dispersed
on oxide supports. Rh is known to be the most active for
*
Corresponding author.
E-mail address: kijyoon@skku.edu (K.J. Yoon).
0021-9517/$ – see front matter  2003 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2003.07.004

J.H. Jun et al. / Journal of Catalysis 221 (2004) 178–190
179
POM [3]. However, Ni is the most extensively studied cata-
lyst because it is effective as well as cheap. Ni is usually
supported on alumina and silica, but it is susceptible to the
formation of inactive nickel aluminate and silicate. Sinter-
ing of Ni and the supports and coke formation are other
problems. Several attempts have been reported to overcome
these problems. Choudhary et al. [3] have shown that, in
supported Ni catalysts prepared by depositing NiO on alu-
mina and silica precoated with MgO, chemical interactions
of NiO with the support can be prevented by providing a
stable protective layer of magnesium aluminate and silicate.
In addition, MgO stabilizes nickel by forming a NiO–MgO
solid solution. Hayakawa et al. [7] and Takehira et al. [9]
have reported that Ni supported on perovskite-type oxides
prepared by solid-phase crystallization, such as Ni0.2/ATiO3
phosphate/hydroxyapatite is a new type of catalyst for POM
that has not been investigated by other researchers, char-
acteristics of this material were investigated by several
characterization methods. The activation behavior of those
catalysts was also studied and compared with that of nickel-
strontium phosphate and oxide-supported nickel catalysts.
2. Experimental
2.1. Catalyst preparation
The catalysts were prepared from nearly saturated aque-
ous solutions of calcium nitrate, nickel nitrate, and dibasic
ammonium phosphate. The pH of the solutions was adjusted
to 10–11 by adding ammonia water, and predetermined
amounts of the solutions were mixed at room temperature
with vigorous stirring in 1 h and the water was evaporated to
get a thick paste. The mixture was dried at 383 K overnight
and finally calcined in air at 1073 K for 2 h to obtain the
catalyst. The solid catalyst was crushed and sieved, and the
particles of 40- to 80-mesh size were used.
(
A: Ca, Sr or Ba), shows good activity and selectivity as well
as very low coke formation. They explain that this may be
due to highly dispersed and stable Ni metal particles on the
perovskite, where the nickel species thermally come out dur-
ing the reaction from the cations homogeneously distributed
in an inert perovskite matrix as the precursor. Takehira [12]
has also reported that a Ni/Al–Mg oxide catalyst prepared
by solid-phase crystallization, via synthesis of a hydrotalcite
The mole ratio of Ca/PO4 was varied from 8.0/6 to
10.0/6 and the Ni/PO4 ratio from 0.5/6 to 3.0/6. The cata-
−
2
(
e.g., Ni0.5Mg2.5Al(OH)8(CO3 )0.5·2H2O) and its calcina-
∗
∗
tion, had very fine Ni particles and showed high activity,
selectivity, and stability. Such studies suggest that high dis-
persion of metal species and/or the use of alkaline earth met-
als may be beneficial to improve the catalytic performance
for the POM. Liu et al. [13,14] have recently reported that
Ni/Ce-ZrO2/θ-Al2O3 shows high activity and stability and
that a protective layer of Ce-ZrO2 suppresses the formation
of inactive NiAl2O4.
lysts were designated to be Ca NiP(a), where is 60 times
the Ca/PO4 mole ratio and a is 60 times the Ni/PO4 mole
ratio employed. For example, Ca95NiP(25) has the composi-
tion ratio of Ca9.5Ni2.5(PO4)6. The multiplication by 60 was
taken for convenience sake in order to express the composi-
tion by integers, since calcium phosphate (Ca3(PO4)2) and
calcium hydroxyapatite (Ca10(PO4)6(OH)2) have the least
common multiple of 6 between the subscript numbers of the
phosphate group. In addition, -f or -u was affixed when nec-
essary, where f and u denoting fresh (before-reaction) and
used (after-reaction) catalyst, respectively.
From the work in our laboratory, it has recently been re-
ported that another type of new catalyst, nickel-strontium
phosphate, exhibits high activity and selectivity in POM
[
15–17]. No refractory oxide support was used in this cat-
alyst. It was found that this catalyst could be activated by the
reacting gases only (methane + oxygen) at around 850 K.
Very fine nickel particles were observed and they are con-
sidered to come out from the strontium nickel phosphate and
hydroxyapatitestructure. Over this catalyst, methane conver-
sion and H2 and CO concentrations close to or sometimes in
excess of those predicted by the thermodynamic equilibrium
were observed. This kind of observation has been reported in
some earlier works [6,18]. An explanation for this is that the
actual catalyst temperature is significantly higher than the
temperature usually measured by a thermocouple since the
reaction is exothermic. Actually, it has been reported that hot
spots in the catalyst bed were detected by an optical pyrom-
eter [6]. On the other hand, it is considered that the presence
of Sr imparts some alkalinity to the catalyst and might be
beneficial for the suppression of coke formation.
2.2. Catalyst characterizations
The catalysts were characterized with X-ray diffraction
(XRD) using Ni-filtered Cu-Kα (M18xHF-SRA, Mao Sci-
ence) and high-resolution transmission electron microscopy
(HRTEM, JEM3011, JEOL). The temperature-programmed
reduction (TPR) was carried out by using 5% H2/Ar with
a ramp rate of 10 K/min for 100 mg of the fresh cata-
lyst. X-ray photoelectron spectroscopy (XPS) measurements
were performed using Mg-Kα radiation with the reference
of Au 4f7/2 at 84.00 eV (ESCA 2000). Ni particle sizes
were estimated from XRD line broadening analysis using
the Scherrer equation. Hydrogen chemisorption experiments
were performed with a volumetric adsorption apparatus (Mi-
cromeritics ASAP-2000) as well as by employing pulse
experiments consisting of repeated injections of a small
amount of H2 in flowing Ar and determination of effluent
H2 by TCD. The hydrogen chemisorption was carried out
at room temperature for the fresh samples after reduction
by flowing H2 at 723 K for 2 h and subsequent evacuation
In this work, calcium, in the same group as Sr, was
employed instead of strontium, and nickel–calcium phos-
phate/hydroxyapatite catalysts were investigated for POM
as an extension of the previous works. Since nickel–calcium

180
J.H. Jun et al. / Journal of Catalysis 221 (2004) 178–190
when using the volumetric apparatus or at 1023 K for 2 h
and subsequent flushing with Ar when employing the pulse
experiment. The amounts of O2 sorption at 923 K were also
measured by employing the pulse experiments for the fresh
samples reduced and flushed as described above.
2
.3. Performance test
The catalyst performance was tested by a conventional
method using an 8-mm i.d. quartz-tube flow reactor. The re-
action temperature was measured by a thermocouple directly
contacting the catalyst particles and controlled by an electric
furnace. Unless specified otherwise, the following standard
experimental conditions were employed. The catalyst charge
was 0.2 g. The partial pressures of methane and oxygen were
1
6.2 and 8.1 kPa (0.16 and 0.08 atm), respectively. Ar was
3
used as the diluent gas and the total flow rate was 100 cm
STP)/min. The product gas was analyzed by two gas chro-
(
matographs using Carboxen 1004 columns (Supelco); one
used Ar as the carrier gas and the other used He. The latter
was needed especially when the concentration of produced
CO2 was too low to detect by using the Ar carrier gas. The
first reaction experiment for each catalyst was carried out at
(a)
6
73 K by flowing the reaction gas mixture without employ-
ing the hydrogen pretreatment.
3
. Results and discussion
3
.1. TEM results
Fig. 1 shows the TEM pictures for Ca95NiP(25)-f and -u.
The used catalyst here is the sample that had experienced a
set of reaction experiments in which the reaction temperature
was changed from 673 to 1073 K and then down to 673 K
◦
with an interval of 50 . As shown later in Section 3.6, the
catalyst lost almost all the activity when the temperature was
lowered to 673 K.
In the fresh sample, round particles of 5–20 nm in size
were observed. These particles were identified as NiO by
energy-dispersiveX-ray spectroscopy (EDS) equipped in the
TEM apparatus, since the major elements in the particles
were found to be Ni and O when irradiated with about a
1
0-nm size beam. At the other part of the catalyst, the major
elements were Ca, P, and O, together with a small amount
of Ni. In the used sample, very fine particles that had not
been seen in the fresh sample were observed. Many of them
were in a needle shape, with the width of about 1 nm and
the length of several nanometers. Round particles of a few
nanometers in diameter were also observed. For these parti-
cles a major element was also found to be Ni by EDS. This
suggests that the Ni comes out of the matrix during the POM
reaction. Similar results had been observed in the previous
work on the POM reaction over a Ni-strontium phosphate
catalyst [16]. This phenomenon appears similar to the solid-
phase crystallization noted in Introduction [9]. Since the
(
b)
Fig. 1. TEM images of Ca95NiP(25) catalysts. (a) Fresh Ca95NiP(25),
(b) used Ca95NiP(25).
used catalyst had been finally exposed to an oxidizing en-
vironment of flowing CH4–O2 at 673 K for about 1 h where
practically no reforming reaction occurred, the nickel sur-
face would have been fully oxidized. In addition, the nickel
that had come out did not go back into the matrix but most
of it appeared to remain as fine particles. This suggestion of

J.H. Jun et al. / Journal of Catalysis 221 (2004) 178–190
181
the Ni coming out of the matrix will be further examined by
XRD and TPR below. One thing to note here is that in the
fresh and used samples there may exist very large particles,
larger than a 100 nm, but due to the limitation of TEM they
would be hard to distinguish from thick parts of the sample.
3
3
.2. XRD analysis
.2.1. The crystalline phases
The XRD patterns for representative samples are shown
in Figs. 2–4. The used samples here are those that had been
used for the POM reaction. Each used sample had experi-
enced a set of reaction experiments in which the first experi-
ment was carried out at 1023 K in order to activate the cata-
lyst (see Section 3.6), the next at 1073 K, and the subsequent
experiments were carried out by decreasing the temperature
◦
with an interval of 50 down to 673 K where the activity usu-
ally became near zero. At each temperature the reaction was
carried out for about 1 h.
The XRD patterns of some samples of Ca80NiP(a) cata-
lysts are shown in Fig. 2. In these catalysts the calcium hy-
droxyapatite phase (hexagonal Ca10(PO4)6(OH)2; this will
be called as the apatite hereafter) was not observed. This
is because the Ca/PO4 mole ratio, 8.0/6, is considerably
lower than 10/6 so that it is difficult for the apatite to be
formed. Although the Ca/PO4 ratio was smaller than 9/6,
the calcium phosphate phase (hexagonal β-Ca3(PO4)2; this
will be called as the phosphate hereafter) was well formed.
Fig. 3. X-ray diffractograms for Ca85NiP(5), Ca85NiP(15), and
Ca85NiP(25) (P, β-Ca (PO ) ; ", Ca (PO ) (OH) ; 1, NiO; 2, Ni).
3
4 2
10
4 6
2
Fig. 2. X-ray diffractograms for Ca80NiP(5), Ca80NiP(15), and
Fig. 4. X-ray diffractograms for Ca95NiP(5), Ca95NiP(15), and
Ca95NiP(25) (P, β-Ca (PO ) ; ", Ca (PO ) (OH) ; 1, NiO; 2, Ni).
Ca80NiP(25) (P, β-Ca (PO ) ; ", Ca (PO ) (OH) ; 1, NiO; 2, Ni).
3
4 2
10
4 6
2
3
4 2
10
4 6
2

182
J.H. Jun et al. / Journal of Catalysis 221 (2004) 178–190
This suggests that it is highly probable that a part of the
system in the 0.34 ꢀ x ꢀ 0.43 range, there exist the
Sr3(PO4)2 phase and a solid solution of Sr9Fe1.5(PO4)7 and
SrFe2(PO4)2. They also showed that Ni, Co, Cu, Zn, Mn, and
Cd could substitute the Sr2 . The Sr3−xNix(PO4)2 system
exists as a solid solution of Sr2Ni(PO4)2 and Sr9Ni1.5(PO4)7
in the 0.32 ꢀ x ꢀ 0.39 range. This tells us that the nickel can
nickel (as Ni2 ) may incorporate into the phosphate struc-
+
2
+
ture by taking the position of Ca [19]. In the catalyst with
a low content of Ni such as Ca80NiP(5), NiO phase was
hardly observed. This may be due to the small amount of the
nickel or some of the nickel taking the position of the cal-
cium. In the catalyst with a higher content of nickel such as
Ca80NiP(15), the NiO phase was present in the fresh sample
but it was barely seen. In the used Ca80NiP(15), the NiO and
metallic Ni phases were hardly seen. As the nickel content
increased further, the amount of the NiO phase in the fresh
catalyst increased. In Ca80NiP(30) having the highest nickel
content among the Ca80NiP(a) catalysts investigated (ca.
+
substitute or take the position of Ca2 in the phosphate and
the apatite structure. In addition, this may also be applicable
to the explanation of complicated constitution of the phos-
phate and apatite phases depending on the composition.
The XRD patterns of some samples of Ca95NiP(a) cat-
alysts are shown in Fig. 4. In these catalysts the apatite
phase was the dominant phase, which is different from the
results for Ca80NiP(a) and Ca85NiP(a) catalysts. This is
certainly due to the higher Ca/PO4 ratio, which is favor-
able for the formation of the apatite phase. In Ca95NiP(5),
the phosphate phase was not seen, and the NiO phase was
hardly observed because the Ni content was low. In the cata-
lysts with higher contents of Ni such as Ca95NiP(15) and
Ca95NiP(25), the phosphate phase was present in a signifi-
cant amount, although it was considerably smaller than the
apatite phase. Metallic Ni species was observed in the used
catalysts and its amount appeared to increase with the Ni
content. For Ca90NiP(a) catalysts (not shown for brevity’s
sake), the XRD patterns were somewhere in the midway of
Ca85NiP(a) and Ca95NiP(a) catalysts.
+
1
6.0 wt% Ni), the amount of NiO in the fresh sample was
quite large. In the used Ca80NiP(30), the NiO almost dis-
appeared but the metallic nickel was observed in almost the
same amount of NiO in the fresh sample. This tells us that
most of the NiO can be reduced to the metallic state during
the reaction. Moreover, a considerable portion of the metal-
lic nickel remained not reoxidized even if the used sample
had been exposed to flowing CH4–O2 at 673 K for 1 h.
Fig. 3 shows XRD patterns for some samples of
Ca85NiP(a) catalysts. For the catalyst with a low content
of Ni such as Ca85NiP(5), the phosphate was the dominant
phase and only a small amount of the apatite was observed.
Although the Ca/PO4 mole ratio employed, 8.5/6, is lower
than 10/6, the presence of the apatite phase indicates that a
part of the Ni may be incorporated into the apatite struc-
ture. The presence of NiO or metallic Ni phase was not
clearly seen and the presence of nickel phosphate was not
observed. In the catalyst with a higher content of Ni such as
Ca85NiP(15), in contrast to Ca80NiP(a) and Ca85NiP(5),
both the phosphate and the apatite phases were present in
comparable amounts. In the fresh Ca85NiP(15), the pres-
ence of a significant amount of NiO was observed. In the
used Ca85NiP(15), the presence of metallic Ni was observed
while the amount of NiO was considerably decreased when
compared with that in the fresh sample. In the catalyst with
a higher content of Ni such as Ca85NiP(25), the phosphate
became again the dominant phase while the apatite phase
was present in a very small amount. The amount of NiO in
the fresh sample was large but in the used catalyst it became
smaller with the appearance of the metallic Ni. Despite the
high content of Ni, especially when the (Ca + Ni)/PO4 ra-
tio is higher than 10/6, dominance of the phosphate phase
suggests that Ni species may be apt to get together by them-
selves rather than to incorporate into the apatite or phosphate
structure, thus to form large NiO particles. The NiO in the
fresh samples is reduced to the metallic Ni under the reac-
tion conditions that provide a reducing environment by the
presence of H2, CO, and CH4.
For Ca100NiP(a) catalysts, the apatite was the domi-
nant phase while the phosphate was present in very small
amounts (Fig. 5). Because the (Ca + Ni)/PO4 mole ratio
Belik et al. [19] have recently reported that the Sr–Fe(or
Ni)–PO4 system exists as a single phase, as a mixture of
multiple phases, or as a solid solution of 2–4 phases, de-
pending on the composition. For example, SrFe2(PO4)2 and
Sr9Fe1.5(PO4)7 form a single phase. For the Sr3−xFex(PO4)2
Fig. 5. X-ray diffractograms for Ca100NiP(5), Ca100NiP(15), and
Ca100NiP(25) (P, β-Ca (PO ) ; ", Ca (PO ) (OH) ; 1, NiO; 2, Ni).
3
4 2
10
4 6
2

J.H. Jun et al. / Journal of Catalysis 221 (2004) 178–190
183
Fig. 6. Relative peak intensities of NiO with respect to the sum of calcium
phosphate and apatite peak intensities against the Ni content and Ca/PO
Fig. 7. Relative peak intensities of Ni with respect to the sum of calcium
phosphate and apatite peak intensities against the Ni content and Ca/PO
4
4
ratio for fresh catalysts (the most intense peak for each phase is chosen for
comparison).
ratio for used catalysts (the most intense peak for each phase is chosen for
comparison).
employed is considerably higher than 10/6, the apatite phase
will be much more favorable to form. As the Ni content
increased, the amount of NiO again increased in the fresh
catalysts and the amount of metallic nickel also increased in
the used catalysts with the decrease of the NiO.
for the relative amount of metallic Ni in the used samples
(
Fig. 7). However, except for Ca90NiP(a) catalysts, the max-
imum amount of the Ni appeared at the Ni/PO4 ratio of
.5/6 for the samples with a fixed ratio of Ca/PO4. The
2
amount of Ni in the samples with the Ni/PO4 ratio of 3.0/6
became smaller, and this is considered due to that a signif-
icant portion of the NiO, probably that in the core of the
large particles, could not be reduced. Actually, the presence
of considerable amounts of unreduced NiO was observed in
those used samples.
3
.2.2. Relative amount of Ni and NiO depending on the Ni
content and Ca/PO4 ratio
From the XRD results, the relative amounts of NiO and
Ni depending on the nickel content and the Ca/PO4 ratio
were determined. In the investigated catalysts, the phosphate
and apatite phases were present in a mixed state and the ra-
tio of the two phases varied depending on the composition.
Nevertheless, the amounts of the phosphate and the apatite
in a fresh sample appeared to be nearly the same as those
in the corresponding used sample. Therefore, the sum of the
main peak intensities of the apatite (211) plane (IAP) and
the phosphate (0.2.10 + 217) plane (ICP) was used as the
reference for convenience sake since the difference in the
Ca/PO4 ratios in these two phases was small. The ratio of
NiO (211) peak intensity (INiO) to the reference and the
ratio of Ni (111) peak intensity multiplied by 0.81 to the
reference were obtained in order to determine the relative
amounts. The factor of 0.81 means the intensity ratio of NiO
Fig. 8 shows the relative amount of (Ni + NiO) in the
used samples. One thing to note here is that the amount
of (Ni + NiO) in the used sample was considerably larger
compared with the amount of NiO in the fresh sample. This
strongly supports the suggestion that in the fresh sample
a significant portion of the nickel is taking the position of
2
+
Ca in the phosphate and apatite structures but it comes
out of the structures to form metallic nickel particles under
reducing environment during the reaction, resulting in more
moles of Ni present on the catalyst surface compared with
the fresh sample. Another thing to note here is that not all
the NiO and metallic Ni particles in the used sample may
have been observed by XRD. That is, although a portion of
the nickel that had come out of the phosphate and apatite
structures might agglomerate and form big particles, a sig-
nificant portion of the nickel remained as very fine particles
smaller than a few nanometers, as observed by HRTEM, and
these particles would not have been detected by XRD due to
its limitation [20].
(
211) peak to Ni (111) peak (or the molar sensitivity ratio of
NiO to Ni being 1/1.23 = 0.81) obtained by calibration for
a mixture of the same moles of NiO and Ni.
Fig. 6 shows the relative amount of NiO in the fresh sam-
ples. As the nickel content increased, the amount of the NiO
increased. As the Ca/PO4 ratio increased, the amount of
the NiO tended to increase. Similar results were obtained

184
J.H. Jun et al. / Journal of Catalysis 221 (2004) 178–190
Fig. 8. Relative peak intensities of Ni + NiO with respect to the sum of
calcium phosphate and apatite peak intensities against the Ni content and
Ca/PO ratio for used catalysts (the most intense peak for each phase is
4
chosen for comparison).
Fig. 9. TPR patterns of (a) Ca85NiP(15) and (b) Ca95NiP(25).
3
.3. TPR results
fore, it would be more accurate to consider that the apatite
phase actually consists of calcium hydroxyapatite and cal-
cium nickel hydroxyapatite. The phosphate phase would be
the same case. The OH–Ni–O–PO3– structure in the apatite
is considered to be chemically less stable than the –PO3–O–
Ni–O–PO3– structure in the phosphate when comparing the
OH–Ni– and –PO3–O–Ni– structures, and hence the sec-
ond peak is attributable to the Ni in the former structure.
The amount of the free NiO is much larger than the sum
of the amounts of the Ni in the apatite and phosphate. The
amount of the Ni in the phosphate is larger than that of the
apatite, which is consistent with the argument that the nickel
in the phosphate structure is more stable. In other words, in-
corporation of nickel into the phosphate structure is more
favorable than into the apatite structure.
Fig. 9 shows the TPR results for Ca95NiP(25) and
Ca85NiP(15). For the Ca95NiP(25), hydrogen consumption
started at around 600 K and there appeared at least three
peaks. The first peak at around 690 K is considered due to re-
duction of free NiO, according to earlier works [14,21–23].
The second and third peaks at around 715 and 780 K are at-
tributable to nickel species that are less easily reducible than
NiO. One might think that the second peak is attributable to
very fine NiO particles that have a stronger interaction with
the substrate (apatite and phosphate) than the larger NiO
particles. However, since the presence of very fine particles
of a few nanometers, especially of the needle shape, in the
fresh sample was not observed by HRTEM, this is not con-
sidered to be the case. Instead, they may be assigned to the
nickel ions substituting Ca2+ in the apatite and phosphate
For the Ca85NiP(15), two peaks, the first and the third,
were clearly observed but the presence of the second peak
was not clear. The first peak also appeared at around 690 K
again due to the reduction of free NiO. The other peak at
around 820 K is also assigned to the nickel ions incorporated
into the phosphate structure. In this sample, the nickel ions
incorporated into the apatite structure were not clearly ob-
served. This again means that nickel would be more difficult
to incorporate into the apatite structure than the phosphate. It
would be worthwhile to note that the reduction temperature
of the nickel that has a strong interaction with θ-Al2O3 has
been reported to be 913 K from a TPR study [23], which is
considerably higher than the temperature of the above third
peak.
structures. Although presence of calcium nickel hydroxyap-
2
+
atite and calcium nickel phosphate, in which a part of Ca
was substituted with Ni2 , could not be identified by XRD,
this does not necessarily rule out their presence if their XRD
patterns are almost the same as the calcium hydroxyapatite
and the calcium phosphate, respectively (note: there appear
no diffraction files for calcium nickel hydroxyapatite and
calcium nickel phosphate in JCPDS and no other reports
for these phases have not yet been found in the literature).
The presence of strontium nickel phosphate in the previ-
ous works [17,19] strongly supports the possible presence of
the calcium nickel hydroxyapatite and calcium nickel phos-
phate because the substituting chemistry would be the same
+
(
Ca and Ni have the same oxidation number of +2). There-

J.H. Jun et al. / Journal of Catalysis 221 (2004) 178–190
185
Table 1
N physisorption and H chemisorption on fresh catalysts
2
2
Ca95NiP(10)
16.8
Ca95NiP(15)
14.2
Ca95NiP(20)
18.2
Ca95NiP(25)
8.6
Ca95NiP(30)
15.6
BET surface area
of catalyst (m /g)
2
Pore volume
3
0.03
4.59
0.05
5.44
0.05
3.78
0.21
0.02
2.70
0.08
17.0
0.66
(
cm /g)
a
H chemisorption
2
(µmol/g)
Exposed of Ni (%)
0.48
Ca100NiP(5)
1.21
0.38
Ca100NiP(15)
1.64
0.12
Ca100NiP(25)
4.26
b
H chemisorption
2
(µmol/g)
Exposed of Ni (%)
0.24
0.12
0.18
a
With volumetric adsorption apparatus.
b
With H pulse experiment.
2
3
.4. Hydrogen chemisorption and oxygen sorption
Hydrogen chemisorption was performed on reduced
Ca95NiP(a) with Micromeritics ASAP-2000 and on reduced
Ca100NiP(a) with H2 pulse chemisorption, as described in
Section 2.2 under Experimental. Table 1 shows the hydrogen
chemisorption results. The amounts of H2 chemisorption
measured by the volumetric adsorption apparatus ranged
from 2.7 to 17.0 µmol/gcat and the percentages exposed of
Ni were below 0.7%. These are quite small amounts, al-
most in the error range of the measurement. Therefore, it
was hard to tell any trend with the Ni content. However,
when the pulse chemisorption was employed, the amount
of H2 chemisorption tended to increase with the Ni content,
although the amounts were very small.
Fig. 10 shows the oxygen pulse sorption results for
Ca95NiP(25) and Ca100NiP(a). Sorption means here the
chemisorption at the surface plus penetration into or oxida-
tion of the bulk nickel. Before oxygen sorption the samples
were reduced in hydrogen at 1023 K for 2 h. The molar
ratio of total nickel to the total sorbed oxygen atom for
Ca100NiP(5) was only 0.19, while that for the other catalysts
was similar to each other, ranging from 0.59 to 0.69. The sig-
nificantly lower ratio for the catalyst with low nickel content
is considered due to the reason that most of the nickel would
be present in the phosphate and apatite structures and be
more stable (more difficult to be reduced) or that the nickel
present deep inside the large crystallites would not have the
opportunity to contact the H2 and migrate to the surface.
On the contrary, for the catalyst with high nickel content,
a significant portion of the nickel in the fresh sample is al-
ready present as NiO, probably at the surface, and it would
be more easily reducible and sorb oxygen. Moreover, the
nickel in the phosphate and apatite structures might be more
easily reducible due to the higher extent of the substitution
of Ca. The nickel unreduced by the hydrogen treatment may
be again due to that present deep inside the large crystal-
lites of NiO, calcium nickel phosphate, and calcium nickel
hydroxyapatite. Nevertheless, the portion of reduced nickel
was very much higher compared with the percentage ex-
Fig. 10. The O
and Ca100NiP(a) reduced at 1023 K.
/Ni
at 923 K vs pulse number over Ca95NiP(25)
sorbed
total
posed of metallic Ni (less than 0.7%) determined by the H2
chemisorption.
3
.5. XPS analyses
Fig. 11 presents XPS spectra in the Ni 2p region for
Ca85NiP(30). Since the samples had been exposed to an
oxidizing environment, the presence of metallic nickel (the
binding energy (b.e.) =∼ 852 eV) at the surface could not
be expected. The peaks at the b.e. of ∼ 857 and 874 eV are
for Ni2 2p3/2 and 2p1/2. The peak at the b.e. of ∼ 862 eV
is the shakeup satellite peak [6]. The b.e. of Ni 2p3/2 in NiO
is reported to be ∼ 854 eV, and that in Ni(OH)2, NiAl2O3,
NiSO4, or NiSiO3 to be 856–857 eV [6,24]. Therefore,
the peak at the b.e. of ∼ 857 eV is considered to come
+
from Ni2 in the phosphate (P–O–Ni–O–P) or the apatite
+

186
J.H. Jun et al. / Journal of Catalysis 221 (2004) 178–190
Fig. 12. CH conversion with repeated increasing (", first; !, second) and
4
decreasing (Q, first; P, second) temperature over Ca95NiP(25).
Fig. 11. XPS spectra for Ni 2p region in Ca85NiP(30).
(
HO–Ni–O–P) structure [16]. This is consistent with the
performance(see Section 3.6) despite the low H2 chemisorp-
tion. This is not well understood at present, but this might
be another kind of strong-metal-support interaction (SMSI).
A more detailed study may be needed in the future.
XRD result that the phosphate and the apatite phases were
present in fresh Ca85NiP(a) catalysts. That is, a part of
nickel was incorporated into the phosphate and the apatite
structure by taking the position of Ca2
+
.
Although the presence of large amounts of NiO and
metallic Ni was observed by XRD, the presence of NiO at
the surface was not clearly seen by XPS and its amount ap-
peared to be small. No significant differences in the XPS
spectra were observed between the fresh and used samples
even though many fine particles were formed in the used
sample. The relatively sharp XRD peak for NiO in the fresh
sample represents that most of the NiO particles are large,
larger than 100 nm when estimated by the Scherrer equa-
tion [20]. As a consequence, the amount of Ni atoms at the
surface of large NiO particles could be quite smaller than the
amount of Ni atoms at the surface of calcium nickel phos-
phate and calcium nickel hydroxyapatite. This could be a
3.6. Catalyst activation and performance
The POM reaction was carried out by sequentially in-
creasing and decreasing the temperature with 50 intervals,
◦
and the results over Ca95NiP(25) are shown in Figs. 12–14.
When the temperature was first increased, the activity of
the catalyst was very low below 723 K. Up to 873 K, the
methane conversion was lower than 25% and CO2 was prac-
tically the sole carbon product; the H2 yield was only about
0.1%. At 923 K, the methane conversion was about 25%
with no CO formation during initial 30 min of time on-
stream, but after 1 h the conversion abruptly rose to 69%
with CO selectivity of 75%, H2 yield of 62%, and H2/CO
ratio of 2.8. This indicates that the Ni is reduced to the metal-
lic state at this temperature. Above this, the conversion, CO
selectivity, and H2 yield increased gradually with the tem-
perature, and at 1073 K the conversion of 93%, the CO
selectivity of 96%, and the H2 yield of 90% were obtained.
These values are very close to the equilibrium values.
While the temperature was decreased afterward, repro-
ducible results were obtained down to 923 K. Below this,
however, the conversion decreased gradually down to 723 K
and did not follow the path of the first increasing sequence.
At 723 K, the conversion was 35% and the CO selectivity
was 9%. At 673 K, the activity dropped to near zero, but
it took about an hour to lose the activity completely. When
the temperature was increased and decreased again, the re-
sults reproducibly followed the path of the first decreasing
possible explanation for the small amount of Ni2 with a
b.e. of ∼ 854 eV compared with that of ∼ 857 eV in the fresh
sample. For the used sample, the reason for the small amount
of Ni2 with a b.e. of ∼ 854 eV is hard to understand when
considering the fact that many fine particles were formed.
However, when the small H2 chemisorption is taken into ac-
count, it is suggested that the fine particles might be covered
with, or decorated with, or strongly interact with the phos-
phate and hydroxyl groups, resulting in a relatively large
amount of Ni2 with a b.e. of ∼ 857 eV and, at the same
time, suppression of H2 chemisorption. Since the amount of
H2 chemisorption was very small on these catalysts, the H2
chemisorption method may not be helpful to a study for the
change of active metal surface due to sintering or coking.
It is notable that the catalysts exhibited excellent catalytic
+
+
+

J.H. Jun et al. / Journal of Catalysis 221 (2004) 178–190
187
Fig. 15. CH conversion with repeated increasing (", first; !, second) and
decreasing (Q, first; P, second) temperature over Ca90NiP(20).
Fig. 13. CO selectivity with repeated increasing (", first; !, second) and
decreasing (Q, first; P, second) temperature over Ca95NiP(25).
4
it takes about 1 h at 923 K for the catalyst to exert its full
potential. At 923 K during the first temperature increase, the
Ni in the phosphate and apatite structures is considered to
come out and reduce to the metallic Ni state, as is evidenced
by the TPR results. Once a part of the Ni is reduced, CO
and H2 concentrations become higher and reduction of the
Ni can be accelerated. While the temperature is decreased, it
is believed that the Ni is progressively oxidized. Eventually
at a certain low temperature, 673 K, the surface of Ni parti-
cles is all oxidized and exhibits no POM activity, but the Ni
in the core of the particles still remains as the metallic state,
as evidenced by the XRD data. When the temperature was
increased afterward, the activity can be restored more easily
at 723 K since the Ni surface was covered only with a thin
layer of oxidized nickel that is more readily reducible. This
activation temperature of 723 K is a little higher than the
reduction temperature of NiO under hydrogen environment,
6
80 K.
Similar results were observed for other catalysts (Figs.
5–16). During the first-increasing sequence, Ca85NiP(15)
1
Fig. 14. H yield with repeated increasing (", first; !, second) and decreas-
ing (Q, first; P, second) temperature over Ca95NiP(25).
exhibited the sudden increase of the activity at 723 K while
Ca90NiP(20) did at 873 K. Nevertheless, they all lost the
POM activity at 673 K. For Ca85NiP(15), the first activation
temperature was so low that the hysteresis was exception-
ally not observed between the temperature increasing and
decreasing sequences.
The first activation temperature of Ca85NiP(25) was
973 K (Fig. 17). This temperature is higher than that of
Ca85NiP(15) with the same Ca/PO4 ratio. This may be
partly explained by the TPR and XRD results. In the XRD
results, two samples had similar amounts of the NiO phase.
However, while similar amounts of the apatite and phos-
phate phases were present in Ca85NiP(15), the phosphate
2
sequence. At 723 K, the activity was readily restored after-
ward.
The above results can be explained as follows. In the fresh
catalyst, all the Ni is present as the oxidized state, and there-
fore only the completely oxidized products are produced at
low temperatures. While the temperature was first increased,
it needs higher temperature and more time for the Ni to be re-
duced when compared with the case of H2 being used as the
reducing agent, because the CH4–O2 mixture is a weaker re-
ducing agent than H2. This is well explained by the fact that

188
J.H. Jun et al. / Journal of Catalysis 221 (2004) 178–190
influential. Other factors, such as the nickel content, may be
operative. In this case, as the nickel content becomes higher,
it appears that the first activation temperature gets higher.
Sometimes when the temperature decreased, the activity
did not drop abruptly at 673 K but sustained considerably
high down to about 600 K or below, as shown in Fig. 17.
However, this was not reproducible but the temperature of
sudden activity drop appeared irregular. This phenomenon
is considered due to the exothermic nature of the oxidation
reactions, resulting in a behavior of an autothermal reaction
or presence of hot spots as reported by Dissanayake et al. [6].
A further study may be needed to better understand this.
The hysteresis of the activity and selectivity observed
in the first cycle of temperature increasing–decreasing se-
quence has also been observed similarly in other cata-
lysts such as oxide-supported Ni catalysts [5,6,11] and Ni-
strontium phosphate catalysts [16]. For these catalysts, the
hysteresis was repeatedly observed in subsequent cycles. Al-
though some experimental conditions in those works and in
this work are different, especially as to the space velocity
and partial pressures of methane and oxygen in the feed
Fig. 16. CH conversion with repeated increasing (", first; !, second) and
decreasing (Q, first; P, second) temperature over Ca85NiP(15).
4
(
Table 2), the common results are as follows: at the lower
branch, where the methane conversion does not exceed 25%
the maximum conversion when only the complete oxidation
(
occurs with a CH4/O2 feed ratio of 2), CO2 and H2O are the
sole products; at the temperature-decreasing branch, the O2
conversion was virtually 100%. The differences are the tem-
peratures at which the sudden increase and decrease in the
methane conversion occur. Below 673 K, all the catalysts
lose the POM activity anyway.
Dissanayake et al. [5] as well as van Looij and Geus [11]
explained about the hysteresis as follows. At the tempera-
ture-increasing branch, the nickel is initially present as an
oxidized state, yielding completely oxidized products, but at
a certain high temperature it is reduced rapidly by H2, CO,
or CH4 to the metallic state, giving partially oxidized prod-
ucts. Upon decreasing the temperature, the metallic nickel in
gradually oxidized, and eventually the metallic nickel is all
reoxidized by the O2 at a certain low temperature. The main
point is that metallic nickel is the active and effective com-
ponent for the selective partial oxidation of methane. The
same explanation may be applicable to the catalysts in this
work.
Fig. 17. CH conversion with repeated increasing (", first; !, second) and
decreasing (Q, first; P, second) temperature over Ca85NiP(15).
4
Comparison of the catalysts in this work and the other
catalysts is summarized in Table 2. For the catalysts in this
work, however, the hysteresis was not observed any more
in the subsequent cycles, and this is a salient difference of
the Ni–calcium phosphate/hydroxyapatite catalysts from the
other Ni catalysts noted above, besides the activation tem-
perature being relatively lower. This suggests that the Ni–
calcium phosphate/hydroxyapatite catalysts are more easily
reducible than the other Ni catalysts. This may be attributed
to the absence of the oxide support or to a weaker metal–
support interaction compared with the oxide support. Actu-
ally, even after the used catalyst was exposed to ambient air
phase was dominant in Ca85NiP(25). As noted in the TPR
results, NiO was the most easily reducible, the Ni incorpo-
rated in the apatite was the next, and that in the phosphate
was the most difficult to reduce. Therefore, the nickel of
Ca85NiP(15) could be reduced at a lower temperature than
that of Ca85NiP(25). The first activation temperature of
Ca95NiP(25), 923 K, is lower than that of Ca85NiP(25)
but higher than that of Ca85NiP(15). This suggests that the
ratio of the apatite to the phosphate does not seem to be
determinative for the first activation temperature but partly

J.H. Jun et al. / Journal of Catalysis 221 (2004) 178–190
189
Table 2
Comparison of activation temperature and reaction conditions employed with increasing–decreasing temperature sequences for various Ni catalysts
3
Catalyst
Catalyst charge
P
(atm)
P
(atm)
O2
Total flow rate (cm /min)
First activation temp. (K)
Reference
CH4
Ca85NiP(15)
Ca90NiP(20)
Ca95NiP(25)
0.2 g
0.2 g
0.2 g
0.16
0.16
0.16
0.08
0.08
0.08
100
100
100
723
873
923
973
973
This work
This work
This work
0.5 g
Ca85NiP(25)
0.2 g
0.16
0.16
0.08
0.08
100
100
This work
[16]
Sr Ni (PO )
4 6
0.2 g
1023
973
873
9.0 1.0
0
1
.5 g
.0 g
2
5 wt% NiO/SiO
50 mg
50 mg
0.01
0.064
0.005
0.036
100
50
1050
1023
[11]
[5]
2
25 wt% NiO/Al O –
2
3
TiO –CaO
2
Fig. 18. Proposed scheme of changes of Ni state in POM reaction.
overnight, it was found that it could be reactivated at 723 K
under the standard reaction conditions.
hydroxyapatite, and NiO; the NiO particles probably lay on
the surface (Fig. 18A). As the reaction temperature increases
and reaches a certain temperature, the NiO is reduced from
the outside (Fig. 18B,a) or through the cracks in the crys-
tallite (Fig. 18B,b) under reducing environment that is pro-
vided by the consumption of O2 and the production of H2
and CO. The nickel in the phosphate and apatite structures
gradually comes out and forms fine particles and is reduced,
which is evidenced by the TPR, TEM, and activation experi-
3
.7. Proposal for changes of the nickel state
Fig. 18 shows the process for changes of the nickel state
based on the above discussion. Before the reaction, the cata-
lyst exists in a mixed state of calcium phosphate, calcium
hydroxyapatite, calcium nickel phosphate, calcium nickel

190
J.H. Jun et al. / Journal of Catalysis 221 (2004) 178–190
ments. Upon decreasing the temperature, the catalyst surface
is gradually reoxidized. Finally, at 673 K or below, the metal-
lic nickel surface is completely covered with a thin layer of
oxygen, NiO (Fig. 18C), but the nickel that has come out
does not go back to the phosphate or the apatite structure.
When the temperature is increased again, it returns to the B
state. This cycle can be repeated while the temperature is de-
creased and increased.
References
[1] C.T. Au, H.Y. Wang, J. Catal. 167 (1997) 337.
[2] Y. Chen, C. Hu, M. Gong, Y. Chen, A. Tian, Stud. Surf. Sci. Catal. 130
(
2000) 2543.
3] V.R. Choudhary, B.S. Uphade, A.S. Mamman, J. Catal. 172 (1997)
81.
[
2
[
[
4] A.M. Diskin, R.M. Ormerod, Stud. Surf. Sci. Catal. 130 (2000) 3519.
5] D. Dissanayake, M.P. Rosynek, K.C.C. Kharas, J.H. Lunsford, J. Ca-
tal. 132 (1991) 117.
[6] D. Dissanayake, M.P. Rosynek, J.H. Lunsford, J. Phys. Chem. 97
(1993) 3644.
4
. Conclusions
[
7] T. Hayakawa, H. Harihara, A.G. Anderson, K. Suzuki, H. Yasuda, T.
Tsunoda, S. Hamakawa, A.P.E. York, Y.S. Yoon, Appl. Catal. A 149
Nickel–calcium phosphate/hydroxyapatite, a new type of
(1997) 391.
catalyst with no oxide support, exhibited excellent perfor-
mance for the partial oxidation of methane, yielding values
close to the equilibrium. This material, if activated once,
could be activated at 723 K after total reoxidation of the
nickel surface without employing the H2 pretreatment but
with CH4–O2 only. This may be due to the absence of the
oxide support and thin layers of NiO at the particle surface
formed during the reoxidation. At least three nickel species
were present in the fresh sample and the crystalline phases
were calcium phosphate, calcium hydroxyapatite, calcium
nickel phosphate, calcium nickel hydroxyapatite, and NiO.
As the Ca/PO4 ratio increased, the amount of the apatite
phase became larger. As the Ni/PO4 ratio increase, the
amount of NiO in the fresh catalyst became larger. In the
used catalysts, the metallic Ni was observed with significant
decrease of the NiO. TEM showed that many fine nickel par-
ticles of a few nanometers in needle and round shapes that
had not been observed in the fresh sample were present in
the used sample. These are believed to be formed from the
nickel that has come out of the calcium nickel phosphate and
calcium nickel hydroxyapatite under reducing environment
during the reaction. Therefore, most of the catalytic activity
was attributed to the fine Ni metal particles. The amount of
H2 chemisorption was very small (the percentage exposed of
Ni < 1%) while the amount of O2 sorption was quite large
[
8] S. Liu, G. Xiong, H. Dong, W. Yang, S. Sheng, W. Chu, Z. Yu, Stud.
Surf. Sci. Catal. 130 (2000) 3567.
[9] K. Takehira, T. Shishido, M. Kondo, R. Furukawa, E. Tanabe, K. Ito,
S. Hamakawa, T. Hayakawa, Stud. Surf. Sci. Catal. 130 (2000) 3525.
10] P.M. Torniainen, X. Chu, L.D. Schmidt, J. Catal. 146 (1994) 1.
11] F. van Looij, J.W. Geus, J. Catal. 168 (1997) 154.
12] K. Takehira, from AISTI, Sinotech, Information on New Technology,
http://www.mri.co.kr (Oct. 2002).
[
[
[
[
13] Z.-W. Liu, K.-W. Jun, H.-S. Roh, S.-E. Park, Y.-S. Oh, Korean J.
Chem. Eng. 19 (2002) 735.
[14] Z.-W. Liu, H.-S. Roh, K.-W. Jun, S.-E. Park, T.-Y. Song, Korean J.
Chem. Eng. 19 (2002) 742.
15] S.-H. Lee, K.J. Yoon, Korean J. Chem. Eng. 18 (2001) 228.
16] S.J. Lee, J.H. Jun, S.-H. Lee, K.J. Yoon, T.H. Lim, S.-W. Nam, S.-A.
Hong, Appl. Catal. A 230 (2002) 61.
17] J.H. Jun, S.J. Lee, S.-H. Lee, T.-J. Lee, S.J. Kong, T.H. Lim, S.-W.
Nam, S.-A. Hong, K.J. Yoon, Korean J. Chem. Eng. 20 (2003), in
press.
[
[
[
[18] V.R. Choudhary, A.M. Rajput, B. Prabhakar, J. Catal. 139 (1993) 326.
[
19] A.A. Belik, B.I. Lazoryak, K.V. Pokholok, T.P. Terekhina, I.A.
Leonidov, E.B. Mitberg, V.V. Karelina, D.G. Kellerman, J. Solid State
Chem. 162 (2001) 113.
20] B.D. Cullity, in: Elements of X-Ray Diffraction, 2nd ed., Addison–
Wesley, Menlo Park, CA, 1978, pp. 284–285.
[21] W.-S. Dong, K.-W. Jun, H.-S. Roh, Z.-W. Liu, S.-E. Park, Catal. Lett.
78 (2002) 215.
22] H.-S. Roh, K.-W. Jun, W.-S. Dong, S.-E. Park, Y.-S. Baek, Catal. Lett.
4 (2001) 31.
23] H.-S. Roh, K.-W. Jun, W.-S. Dong, S.-E. Park, Y.-I. Joe, Chem. Lett.
2001) 666.
[
[
[
[
7
(
Osorb/Nitotal ∼ 0.6). The suppression of H2 chemisorption
(
might be attributed to the covering of the nickel particles
with the phosphate and/or hydroxyl groups or to a strong in-
teraction with those groups.
24] D. Briggs, M.P. Seah (Eds.), Practical Surface Analysis: Auger and
X-Ray Photoelectron Spectroscopy, Vol. 1, 2nd ed., Wiley, New York,
1990, App.5.