Reaction of ethanol over hydroxyapatite affected by Ca/P ratio of catalyst

Article abstract of DOI:10.1016/j.jcat.2008.08.005

The mineral hydroxyapatite [HAP; Ca₁₀(PO₄)₆(OH)₂] is the chief component of animal bones and teeth. It also is known to function as a catalyst with both acid and base sites, depending on the manner in which it is synthesized. We closely studied the reaction of ethanol over HAP using catalysts of different Ca/P molar ratios. These were prepared by controlling the pH of the solution during precipitation synthesis. We found that the distribution of acid sites and basic sites on the catalyst surface varied with the Ca/P ratio of HAP. The yields of ethylene, 1-butanol, and 1,3-butadiene were correlated with the ratio of acid sites and basic sites. We further found that yields of higher alcohols, such as 1-butanol, that are known as Guerbet alcohols and are characteristic products of ethanol over HAP, are functions of the probability of ethanol activation (α) on the catalyst surface.

Full text of DOI:10.1016/j.jcat.2008.08.005

Journal of Catalysis 259 (2008) 183–189
Contents lists available at ScienceDirect
Journal of Catalysis
www.elsevier.com/locate/jcat
Reaction of ethanol over hydroxyapatite affected by Ca/P ratio of catalyst
Takashi Tsuchida a,∗, Jun Kubo a, Tetsuya Yoshioka a, Shuji Sakuma a, Tatsuya Takeguchi b, Wataru Ueda b
a
Central Research Center, Sangi Co., Ltd., Fudoinno 2745-1, Kasukabe-shi, Saitama 344-0001, Japan
Catalysis Research Center, Hokkaido University, Kita 21 Nishi 10, Kita-ku, Sapporo 001-0021, Japan
b
a r t i c l e i n f o
a b s t r a c t
Article history:
The mineral hydroxyapatite [HAP; Ca10 (PO4)6(OH)2] is the chief component of animal bones and teeth. It
also is known to function as a catalyst with both acid and base sites, depending on the manner in which
it is synthesized. We closely studied the reaction of ethanol over HAP using catalysts of different Ca/P
molar ratios. These were prepared by controlling the pH of the solution during precipitation synthesis.
We found that the distribution of acid sites and basic sites on the catalyst surface varied with the Ca/P
ratio of HAP. The yields of ethylene, 1-butanol, and 1,3-butadiene were correlated with the ratio of acid
sites and basic sites. We further found that yields of higher alcohols, such as 1-butanol, that are known
as Guerbet alcohols and are characteristic products of ethanol over HAP, are functions of the probability
of ethanol activation (α) on the catalyst surface.
Received 18 June 2008
Revised 7 August 2008
Accepted 10 August 2008
Available online 11 September 2008
Keywords:
Hydroxyapatite
Acid basic catalyst
Ca/P ratio
Ethanol
1-Butanol
© 2008 Elsevier Inc. All rights reserved.
Guerbet alcohol
Guerbet reaction
Probability
Ethanol activation
1. Introduction
HAP also is known to have the character of both an acid catalyst
and a basic catalyst when its Ca/P ratio is between 1.50 and 1.67
Hydroxyapatite [HAP; Ca10(PO4)6(OH)2] is the main component
of animal bones and teeth. It is used widely in fields such as bone
regeneration, dental materials, fertilizers, and food supplements (as
a source of calcium). It also is used industrially in sensors, fluores-
cence materials, chromatography, and environmental phosphorus
recovery; its use in drug delivery systems has been well studied
[1–3]. As a catalyst, HAP has the unusual property of containing
both acid sites and basic sites in a single crystal lattice [4–13]. The
stoichiometric form of HAP is shown as Ca10(PO4)6(OH)2, and its
molar ratio is 1.67. The ionic radius of HAP’s component elements
allows a fair degree of transfer or loss of ions within its crystal
structure [14]. As a result, HAP is a highly nonstoichiometric cal-
cium phosphate compound with a Ca/P molar ratio ranging from
1.50 to 1.67. Joris and Amberg [15] postulated that HAP’s nonstoi-
chiometry is facilitated by the fact that the loss of Ca2+ ions and
[4–7]. However, there has been no quantitative report concerning
the reaction of ethanol over highly activated HAP acting as both an
acid catalyst and a basic catalyst. From a catalytic standpoint, it is
thought that active sites newly emerge on the surface of nonstoi-
chiometric HAP as it loses calcium, phosphoric acid, and hydroxyl
groups. We found that we could refine the reaction of ethanol over
HAP by using catalysts with different Ca/P molar ratios, which we
prepared by controlling the pH of the solution during precipitation
synthesis. Our results explain the peculiar distribution of reaction
products from ethanol over HAP catalysts.
2. Experimental
2.1. Catalyst preparation
+
resulting electrical imbalance are corrected by introduction of H
Four HAP catalysts were prepared by the precipitation method.
In each case, a solution containing 0.60 mol/l of Ca(NO3)2·4H2O
and a solution containing 0.40 mol/l of (NH4)2HPO4 [16,17] were
−
ions and depletion of OH ions, representing this by the formula
Ca10−Z (HPO4)Z (PO4)6−Z (OH)2−Z ; 0 < Z ꢀ 1. It is known that at a
Ca/P ratio of 1.50, highly crystalline HAP acts as an acid catalyst,
catalyzing chiefly ethylene synthesis from ethanol by dehydration;
however, at a Ca/P ratio of 1.67, it acts as a basic catalyst, catalyzing
chiefly acetaldehyde synthesis from ethanol by dehydrogenation.
◦
fed slowly into a container and mixed by stirring for 24 h at 80 C.
The precipitate was then washed with deionized water, filtered,
◦
◦
dried at 140 C and calcined at 600 C for 2 h in atmosphere. Dif-
ferent levels of nonstoichiometry in the four HAP catalysts were
obtained by controlling the pH of the mixture of solutions at 7.0,
9.0, 10.0, and 10.5, using aqueous ammonia. CaO and MgO were
used as representative base catalysts for comparison; these were
Corresponding author. Fax: +81 48 752 0120.
E-mail address: takashi.tsuchida@sangi-j.co.jp (T. Tsuchida).
*
0021-9517/$ – see front matter © 2008 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2008.08.005

184
T. Tsuchida et al. / Journal of Catalysis 259 (2008) 183–189
synthesized from materials purchased from Wako Pure Chemical
Industries Inc. CaO was synthesized by dehydration of special-
◦
grade Ca(OH)2 at 600 C for 2 h in air and MgO by dehydration of
◦
Mg(OH)2 (0.07 μm) at 500 C for 30 min in air. β-TCP (Ca3(PO4)2)
(Taihei Chemical Industrial Co., Ltd.; Ca/P ratio, 1.50), was used as
a representative calcium phosphate catalyst.
2.2. Characterization
The catalysts were characterized as follows. Crystal structures
were specified by powder X-ray diffraction (XRD) using a model
M18XHF22-SRA diffractometer (Mac Science Co., Ltd., now Bruker
Japan Co., Ltd.; monochromatized CuKα radiation; tube voltage,
◦
◦
45 kV; tube current, 300 mA; slit width, DS 1.00 , SS 1.00 , RS
0.30 mm). The morphological appearance of the catalysts was ob-
served by scanning electron microscopy (SEM) using a Hitachi
S-4500 scanning electron microscope. Surface areas were mea-
◦
Fig. 1. XRD patterns of the four HAP catalyst powders (calcined at 600 C for 2 h in
sured with
a conventional BET nitrogen adsorption apparatus
atmosphere).
(Coulter SA3100). Bulk Ca/P ratios of HAP were determined by in-
ductively coupled plasma-atomic emission spectrometry (ICP-AES)
using a Seiko Instruments SPS1700R device. Surface Ca/P ratios
of HAP were determined from the binding energy (BE) of Ca 2p
and P 2p by X-ray photoelectron spectroscopy (XPS) using a JEOL
JPS-9010MC spectrometer with monochromatized MgKα radiation.
The densities of acid sites and basic sites on the catalysts were
measured by temperature-programmed desorption (TPD) using a
BEL-CAT (BEL Japan Inc.) equipped with a mass spectrometer. For
TPD measurement, 0.1 g of catalyst was used. This was dehydrated
◦
at 500 C for 30 min in a 50-ml/min helium flow, then cooled to
◦
250 C. Probe gas (5% NH3/He for acid sites and 5% CO2/He for ba-
sic sites) was then adsorbed for 30 min in a 50-ml/min flow. After
purging at the same temperature, the temperature was raised at a
◦
◦
rate of 10 C/min to 1050 C in a 30-ml/min helium flow. The den-
sities of acid sites and basic sites on the catalysts were determined
by mass spectrometry (Q-MS) using an Pfeiffer Vacuum OmniStar
◦
spectrometer with peaks of M/Z = 16 at around 320 C for acid
◦
sites and M/Z = 44 at around 330 C for basic sites.
2.3. Reaction methods
Ethanol conversion was carried out using 4.0 ml of each re-
spective catalyst (the four HAP catalysts, CaO, MgO, and β-TCP),
and the reaction products were analyzed. In each case, catalyst
in powder form was placed in a molding machine to form pel-
lets. The pellets were then lightly crushed to a particle size of
◦
Fig. 2. SEM images of the four HAP catalyst powders (calcined at 600 C for 2 h in
atmosphere).
14–26 mesh. In each experiment, 16.4 vol% of ethanol gas di-
−1
whereas those with high Ca/P ratios are more nodular and shorter
in length. We believe that this difference arises from the process
of HAP synthesis, because in precipitation solutions with a low pH,
tabular brushite (CaHPO4·2H2O) is precipitated as a precursor and
is gradually transformed into HAP [1–3]. Other physical properties
of the catalysts (BET surface area and density of acid and basic
sites) are summarized in Table 1, which also gives the Ca/P ratio
and precipitation suspension pH data for the four HAP catalysts.
The fact that the Ca/P ratio of HAP can be changed by control-
ling the pH of the precipitation solution during synthesis derives
from the morphological and electrical stability of phosphate ions
luted with helium was passed at 116 ml/min (GHSV: 2000 h
)
through a fixed-bed silica tubular reactor with interior diame-
ter of 5 mm at atmosphere. The reaction products were analyzed
by gas chromatography–mass spectrometry (GC-MS) using a Shi-
madzu GC-MS-QP5000 chromatograph in the range M/Z = 10–400
and by GC with a flame ionization detector (GC-FID) using a Shi-
madzu GC-14A chromatograph. In both cases, columns supplied
by J&W Scientific Corporation (liquid phase, DB-1; film thickness,
5.00 μm; column dimensions, 30 m × 0.323 mm) were used. The
performance of the reactions (ethanol conversion and product se-
lectivity) was calculated as in our previous work [16,17].
−
(i.e., H2PO4 , HPO24−, and PO43−) in relation to pH. Phosphate ions
3. Results and discussion
in solution exist in equilibrium, as shown in (1), and the solubil-
ity product indicates that as the pH of the solution rises, the most
highly ionized form (PO34−) becomes most prevalent [18]:
3.1. Characterization of catalysts
Fig. 1 shows XRD patterns for the four HAP catalysts. These
findings confirm that all four are composed of crystalline HAP and
no other substances. Fig. 2 shows SEM images of the four HAP
catalysts. Those with low Ca/P ratios appear rodlike or platelike,
K1
K2
K3
H3PO4 ꢁH2PO− ꢁHPO42− ꢁPO34−
,
4
−3
−8
−12
.
K1 = 7.50 × 10
,
K2 = 6.20 × 10
,
K3 = 1.70 × 10
(1)

T. Tsuchida et al. / Journal of Catalysis 259 (2008) 183–189
185
Thus, according to the chemical formula for nonstoichiometric
HAP, Ca10−Z (HPO4)Z (PO4)6−Z (OH)2−Z ; 0 < Z ꢀ 1, if we want to
synthesize HAP with a high Ca/P ratio (i.e., decrease the Z value),
then we need to reduce the number of HPO24− ions. In other words,
we need to raise the pH of the precipitation solution and increase
the number of PO34− ions.
The results of XPS analysis (Table 1) show that the surface Ca/P
ratios of HAP were smaller than the bulk Ca/P ratios. The same
tendency was seen in previous studies [19–22]. On the other hand,
the surface Ca/P ratio of β-TCP was same as its bulk Ca/P ratio. Be-
cause β-TCP is a mineral without nonstoichiometric composition
[2,3], we can conclude from the findings that a difference in sur-
face and bulk Ca/P ratios is a unique characteristic peculiar to HAP.
The right column of Table 1 gives the density of acid sites and
basic sites on the catalysts. For the four HAP catalysts, the higher
the Ca/P ratio, the lower the acid site density; however, acid sites
were still present on those catalysts with a high Ca/P ratio. In con-
trast, basic sites barely existed on the catalysts with a Ca/P ratio
ꢀ1.62; however, a high basic site density was observed on the cat-
alysts with a Ca/P ratio >1.62. No evidence of acidic properties
was observed in CaO and MgO, and only a slight amount was con-
◦
firmed in β-TCP in our TPD analysis. At around 320 C, the CaO
catalyst also showed no evidence of basic properties, although a
◦
larger CO2 peak than that for HAP was seen above 500 C. But the
MgO and β-TCP catalysts had basic site densities equal to about
one-third that of the HAP catalyst with a Ca/P ratio of 1.65 and al-
most the same as that of the HAP catalyst with a Ca/P ratio of 1.67.
We considered the relationship between the HAP catalysts’ Ca/P
ratio and their numbers of acid and basic sites in the following
manner. It is known that the acid/basic property of nonstoichio-
metric HAP is governed by its Ca2+ ions; that is, its basic property
derives from the presence of Ca2+ ions, and its acidic property de-
rives from the deficiency of Ca2+ ions [1–3]. Moreover, the HAP
crystal grows along its c-axis, so that the a-faces, which contain
phosphate groups related to acidity, are mainly exposed [23]. The
Ca/P ratio is lower at the surface of HAP particles than within their
bulk, as our XPS study revealed. This also suggests that HAP’s cal-
cium ions, considered related to basic properties, are not readily
expressed at the surface level. SEM showed that nonstoichiomet-
ric HAP with a Ca/P ratio of ꢀ1.62 was morphologically rodlike or
platelike. We concluded that in this case, the phosphate-bearing a-
faces were largely exposed and surface expression of calcium was
limited, resulting in a relatively high density of acid sites. In con-
trast, SEM showed that HAP with a Ca/P ratio of ꢂ1.65 (i.e., at or
close to stoichiometric composition) was more globular and com-
pact. We concluded that in this case, HAP’s calcium-bearing c-faces
were more highly exposed, resulting in the high basic site density
that was confirmed by our TPD analysis.
Table 1
Characteristic properties of catalysts
Catalyst pHa
Ca/P molar ratio
Bulkd Surfacee
7.0 1.59
9.0 1.62
Zb
BET
Densityc (μmol/m2) of
(m2/g)
Acid sites
Basic sites
HAP-1
HAP-2
HAP-3
HAP-4
CaOf
1.40
1.43
1.44
1.50
–
0.46
0.28
0.10
0.00
–
27.5
35.7
40.3
37.8
6.4
0.038
0.029
0.011
0.006
0.000
0.000
0.008
0.01
0.02
0.38
0.53
0.00
0.13
0.60
10.0 1.65
10.5 1.67
–
–
–
–
MgOg
β-TCP
–
–
–
–
166.0
1.2
1.50
1.50
Note. HAP: Hydroxyapatite, TCP: tricalcium phosphate (Ca3(PO4)2).
a
pH of precipitation suspension during catalyst synthesis.
Amount of Ca loss in the formula Ca10−Z (HPO4)Z (PO4)6−Z (OH)2−Z ; 0 < Z ꢀ 1
b
[15].
c
d
e
f
◦
3.2. Catalytic reaction
Data measured by TPD. Probe gas was adsorbed at 250 C.
Molar ratio determined with ICP.
Molar ratio determined from XPS.
Table 2 shows the selectivities (C wt%) of main reaction prod-
ucts from ethanol at 10% and 20% conversion. Only a trace of the
C1 product methane was formed on all catalysts evaluated, indicat-
ing that ethanol was not decomposed on them. The change in the
◦
At around 320 C, the CaO catalyst showed no evidence of basic properties al-
◦
though a larger CO2 peak than for HAP was measured at above 500 C.
g
◦
MgO was calcined at 500 C for 0.5 h in atmosphere, and other catalysts at
◦
600 C for 2 h in atmosphere.
Table 2
Main product selectivities at 10% and 20% conversion of ethanol on catalysts (contact time 1.78 s). Selectivity was calculated by Arrhenius plot using the experimental data
Catalyst
HAP-1
1.59
HAP-2
1.62
HAP-3
1.65
HAP-4
1.67
CaO
–
MgO
–
β-TCP
1.50
Ca/P molar ratio
Conversion (%)
10.0
42
371
20.0
84
387
10.0
29
320
20.0
58
350
10.0
20
275
20.0
40
296
10.0
19
272
20.0
38
298
10.0
85
397
20.0
170
421
10.0
7
352
20.0
14
385
10.0
367
405
20.0
735
436
Areal rate (μmol/(m2 s)) × 10
−3
◦
Reaction temp. ( C)
Selectivity (C wt%)
CH4
0.0
82.4
910
4.8
0.0
0.0
0.2
0.0
0.0
0.0
12.3
0.0
0.6
0.0
87.3
1929
3.4
0.0
0.0
0.2
0.0
0.0
0.0
8.8
0.0
1.3
0.0
7.7
77
0.0
16.7
332
4.9
0.0
13.8
3.2
0.6
0.4
5.6
2.2
39.2
941
1.7
1.9
0.2
0.0
0.9
16
0.0
1.0
36
1.7
0.0
1.1
0.5
0.1
0.3
5.1
0.4
68.8
72
6.5
5.0
1.0
0.5
0.0
0.5
16
0.0
0.6
37
0.2
55.1
–
0.3
57.0
–
0.1
18.2
–
0.1
20.8
–
0.1
15.7
7671
21.0
0.1
2.8
1.3
1.2
0.8
10.2
9.9
29.5
181
0.3
0.7
0.0
0.0
0.1
CH2=CH2
22.9
22,470
18.6
0.1
4.6
1.1
1.3
1.3
7.0
9.6
25.6
314
0.3
1.0
0.0
−3
(TOF) (s−1) × 10
CH3CHO
7.0
2.9
0.0
1.0
1.0
0.2
0.2
6.8
0.4
67.3
35
2.2
0.0
0.7
0.5
0.2
0.2
5.4
0.2
70.7
25
1.7
0.0
0.8
0.3
0.1
0.2
3.5
0.2
69.8
50
26.5
0.0
0.5
1.6
0.0
0.4
0.0
0.0
2.7
–
18.7
1.4
0.4
1.5
0.0
0.9
0.0
0.0
2.3
–
8.8
0.0
4.4
1.2
0.5
0.2
11.3
0.0
31.3
17
6.9
0.7
5.1
1.0
0.8
0.4
10.3
0.0
27.9
30
(CH3)2CO
1,3-butadiene
C4 olefins
0.0
4.9
1.8
=
C4 aldehydesa
0.7
0.4
12.6
2.2
49.6
595
2.8
1.6
Butyraldehyde
=
C4 OHsb
C2H5OC2H5
1-C4H9OH
−3
(TOF) (s−1) × 10
C2H5CH(C2H5)CH2OH
1-C6H13 OH
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
6.5
3.1
0.6
0.2
7.4
3.8
0.7
0.2
7.0
5.8
1.1
0.1
0.1
0.0
0.0
0.1
0.1
0.0
0.0
0.7
1.2
0.2
0.1
1.1
1.7
0.3
0.2
C4H9CH(C2H5)CH2OH
1-C8H17 OH
0.2
0.1
0.1
0.6
0.0
Aromatics
0.0
0.3
0.0
0.3
0.1
8.5
0.1
9.4
0.0
8.8
0.2
7.8
0.2
7.2
0.5
7.7
0.4
1.0
1.3
2.2
0.1
6.3
0.2
6.0
Others
12.4
16.3
20.7
20.6
a
Unsaturated C4 aldehydes.
Unsaturated C4 alcohols.
b

186
T. Tsuchida et al. / Journal of Catalysis 259 (2008) 183–189
(a)
(b)
(c)
(d)
Scheme 1. 1-Butanol formation mechanism in ethanol conversion reaction (a) activation of ethanol (aldehyde formation); (b) aldol condensation; (c) dehydration of aldol;
(d) hydrogenation of unsaturated aldehyde.
ethanol conversion rate had almost no effect on product selectiv-
ity for any of the catalysts, except for the fact that the selectivity
of ethylene increased and that of acetaldehyde decreased for all
catalysts with increasing ethanol conversion. At 20% ethanol con-
version, for instance, ethylene selectivity was 87.3% on HAP with
Ca/P = 1.59 (HAP-1), 57.0% on CaO, 22.9% on β-TCP, 20.8% on
MgO, 16.7% on HAP with Ca/P = 1.62 (HAP-2), 1.0% on HAP with
Ca/P = 1.65 (HAP-3), and 0.6% on HAP with Ca/P = 1.67 (HAP-4).
Diethylether selectivity was 8.8%, and the combined selectivity of
ethylene and diethylether was 96.1% on HAP-1; thus, this catalyst
is considered a typical solid acid catalyst. In contrast, acetaldehyde
selectivity at 20% ethanol conversion was 18.7% on CaO, 18.6% on
β-TCP, 6.9% on MgO, 4.9% on HAP-2, 3.4% on HAP-1, and 1.7% on
HAP-3 and HAP-4. We had expected to find much higher selectiv-
ity to acetaldehyde on CaO, which is a typical basic catalyst. We
suggest that this relatively low selectivity to acetaldehyde on CaO
under the reaction conditions used resulted from the fact that de-
hydration of ethanol occurred predominantly at the high reaction
Turnover frequency (TOF) for ethylene synthesis and 1-butanol
synthesis is also given in Table 2. We supposed that ethanol was
converted to ethylene on the catalysts’ acid sites and to 1-butanol
on their basic sites. We defined TOF as the number of ethanol
molecules converted per active site per second and calculated it
by using the values for density of acid sites and basic sites on the
catalysts given in Table 1. Comparing the TOF results shows that
ethylene synthesis (i.e., dehydration of ethanol) was more predom-
inant on HAP without basic sites than on HAP with both acid sites
and basic sites. In contrast, HAP with basic sites had higher TOF
values than MgO in 1-butanol synthesis, suggesting that HAP cat-
alysts with basic sites synthesize 1-butanol more effectively than
MgO.
We can explain the difference in 1-butanol synthesis between
MgO and HAP as follows. A reaction that synthesizes higher alco-
hols from lower alcohols (including ethanol) is known as a Guerbet
reaction [24–32]. In our proposed mechanism, Guerbet alcohols are
synthesized via the aldol condensation of two aldehydes generated
by the dehydrogenation of alcohol. Our proposed mechanism for
1-butanol synthesis from ethanol over HAP is shown in Scheme 1.
By dissociative adsorption of ethanol in the vapor phase, an ethox-
◦
temperature of 421 C. Selectivity to 1-butanol was 69.8% on HAP-
4, 68.8% on HAP-3, 39.2% on HAP-2, 27.9% on MgO, 25.6% on β-TCP,
2.3% on CaO, and undetectable on HAP-1.

T. Tsuchida et al. / Journal of Catalysis 259 (2008) 183–189
187
Scheme 2. Dehydrogenation mechanism from ethanol on MgO catalyst.
ide intermediate is adsorbed on Lewis acid sites, and proton-like
hydrogen is adsorbed on Brönsted basic sites. The ethoxide is then
dissociated into an aldehyde intermediate and hydride-like hydro-
gen (a). Next, one of two neighboring aldehyde intermediates is
decomposed to an enolate (carbanion intermediate), which reacts
with the other aldehyde intermediate to form aldol (aldol conden-
sation) (b). Unsaturated aldehyde is then generated by dehydration
of the aldol (c), and, finally, 1-butanol is synthesized via hydro-
genation of the unsaturated aldehyde (hydride reduction) (d), tak-
ing up hydrogen generated by dissociative adsorption during steps
(a) and (b).
Our results indicate that the nature of hydrogen adsorption on
HAP and MgO differs slightly. As shown in Table 1, on TPD analy-
◦
sis, a peak was observed for MgO at around 320 C, similar to the
peaks observed for HAP-3 and HAP-4. This suggests that like those
catalysts, MgO has basic sites. But MgO showed much higher se-
lectivity to acetaldehyde and unsaturated alcohols than any of the
HAP catalysts (see Table 2). Our hypothesis to explain this find-
ing is illustrated in Scheme 2. In the case of MgO (Mg-O distance,
0.210 nm), ethanol was dissociatively adsorbed to ethoxide on Mg
(Lewis acid sites) and to proton-like hydrogen on ionically bound
O (hard basic sites) near these acid sites. This ethoxide then dis-
sociated into aldehyde and hydride-like hydrogen. But because of
the short distance between them, the released hydrogen entities,
rather than being trapped on the catalyst surface, migrated to form
H2 molecules. These H2 molecules desorbed into the gas phase
with a certain probability. In contrast, in the case of HAP (Ca–O
distance, 0.239 and 0.240 nm), we postulate that ethanol was dis-
sociatively adsorbed to ethoxide on Ca (Lewis acid sites) and to
proton-like hydrogen on covalently bound O in phosphate groups
(soft basic sites) near the acid sites. Dissociative adsorption there
resulted in the formation of stable hydrogen phosphate groups. The
ethoxide then dissociated into aldehyde and hydride-like hydro-
gen. We postulated that in the case of the HAP catalyst, there was
sufficient distance between the hydrogen entities released during
dissociative adsorption for them to become trapped on the catalyst
surface without migrating to form molecular hydrogen in the gas
phase. As a result, yields of Guerbet alcohols (which are saturated
alcohols) were higher on HAP than on MgO.
Fig. 3. Selectivity of main products at 50% conversion of ethanol: (Q) total Guerbet
alcohols, (") 1,3-butadiene, (2) ethylene. Reaction conditions: catalyst, 4 ml; GHSV,
−1
2000 h
.
ethanol conversion over HAP catalysts with high Ca/P ratios. This
tendency reflects the relationship between the basic site density
and Ca/P ratio of the HAP catalysts, also shown in Table 1. We
conclude that HAP is a unique catalyst that synthesizes mainly
Guerbet alcohols at high Ca/P ratios. In contrast, we found that
CaO, a well-known base catalyst, synthesizes mainly ketones and
ethylene at 50% ethanol conversion (data not shown). Both CaO
and HAP have Ca–O bonds, but their reaction products differ. As
seen in Table 1, CaO has neither acid nor basic properties at around
◦
320 C, as HAP does, but exhibits basic properties only at higher
temperature. These data indicate that the reaction of ethanol on
HAP catalysts differs greatly from that on CaO catalysts.
1,3-Butadiene was not synthesized at all on HAP-1, which had a
low Ca/P ratio of 1.59, and was synthesized to only a small degree
on HAP-3 and HAP-4, which had the highest Ca/P ratios. Maxi-
mum selectivity for 1,3-butadiene was in fact observed over HAP-2,
which had a Ca/P ratio of 1.62, the level at which a relative balance
of acid and basic sites was seen on the HAP catalyst (Table 1). The
reaction in which 1,3-butadiene is synthesized from ethanol on an
acid-base catalyst is known as the Lebedev reaction [33–37]. We
postulated that synthesis of 1,3-butadiene from ethanol over HAP
with a Ca/P ratio of ꢂ1.62 was due to a Lebedev reaction, because
over that Ca/P range, HAP showed both acid and basic properties,
albeit to a decreasing degree. Other products not shown in Fig. 3
were olefins, aldehydes, dienes, and aromatics. At higher conver-
sion of ethanol (i.e., higher reaction temperature), selectivity to
these products increased, resulting in a multicomponent compo-
sition that resembled gasoline [17].
Although β-TCP, like HAP, has PO34− ions, it cannot take a non-
stoichiometric form, as HAP can, and cannot trap hydrogen as
HPO24−. Thus, we concluded that hydrogen migrated more easily
to form molecular hydrogen, and acetaldehyde was more readily
synthesized, on the β-TCP catalyst surface than on HAP.
Fig. 3 shows the relationship between the Ca/P ratio of HAP
catalysts and selectivity to ethylene, 1,3-butadiene, and Guerbet
alcohols (total C4, C6, C8, and C10 alcohols) at 50% ethanol con-
version. Ethylene was the main product when the Ca/P ratio of
the catalysts was low, as we reported previously [16]. But ethylene
selectivity fell to about 5% when HAP catalysts with higher Ca/P
ratios were used. This tendency closely reflected the relationship
between the acid site density and Ca/P ratio of the HAP catalysts
given in Table 1. On the other hand, Guerbet alcohols were not
synthesized at all over the HAP catalyst with the lowest Ca/P ra-
tio and were synthesized only to a small degree over HAP with a
Ca/P ratio of 1.62. But Guerbet alcohols were the main products of
3.3. Scheme of Guerbet alcohol synthesis and distribution
As indicated above, our results demonstrate that the Guerbet
reaction plays a major role in ethanol conversion over HAP cat-
alysts. We attempted to explain the scheme of Guerbet alcohol

188
T. Tsuchida et al. / Journal of Catalysis 259 (2008) 183–189
(a)
(b)
Scheme 3. General formula for the synthesis of Guerbet alcohols from normal alcohols (m,n ꢁ 1 and natural numbers). Note: The above represents the standard case, in
which a branched Guerbet alcohol is formed. When one of the normal alcohols is an enolate derived from ethanol, where m = 1 as shown below, the Guerbet alcohol
synthesized is a normal alcohol.
Table 3
Synthesis of C2n Guerbet alcohol from normal alcohols during ethanol conversion
over HAP, showing type of alcohol synthesized, and generation probability
Carbon number of
Guerbet alcohol
Combination of
normal alcoholsa
Type of Guerbet
alcohol
Probability
4 (n = 2)
6 (n = 3)
(2, 2)
→
normal-
α2
α3
(2, 4)
(4, 2)
→
branched-
normal-
→
8 (n = 4)
(2, 6)
(4, 4)
(6, 2)
→
→
→
→
→
→
→
branched-
branched-
normal-
α4
10 (n = 5)
(2, 8)
(4, 6)
(6, 4)
(8, 2)
branched-
branched-
branched-
normal-
α5
Fig. 4. Yields of Guerbet alcohols by carbon number at 20% conversion of ethanol
on β-TCP, MgO, CaO, and HAP catalysts with Ca/P molar ratios of 1.62, 1.65 and 1.67.
The right column shows the probability of ethanol activation (α) on each catalyst.
Contact time, 1.78 s.
Note. α: probability of ethanol activation. n: number of ethanol molecules activated.
a
Combinations in parentheses show carbon numbers for the respective normal
number 2) reacts with an enolate sourced from 1-butanol (car-
bon number 4) to synthesize a branched C6 Guerbet alcohol, and
(4, 2) indicates that an aldehyde sourced from 1-butanol (carbon
number 4) reacts with an enolate sourced from ethanol (carbon
number 2) to synthesize a normal C6 Guerbet alcohol. In each
group of Guerbet alcohols that share the same carbon number,
a normal Guerbet alcohol is generated only when the enolate is
derived directly from ethanol and has a carbon number of 2. As
the table shows, the ratio of normal to branched Guerbet alco-
hol generated in each group decreases as the carbon number of
the group increases. The probability of generating Guerbet alco-
hols, both normal and branched, with a carbon number of 2n thus
can be expressed as αn, where α is the probability of ethanol acti-
vation. Summing up the foregoing relations, the total yields of C2n
normal and branched Guerbet alcohols can be expressed by the
following general formula:
alcohols (aldehyde on the left, and enolate on the right) that combine to form Guer-
bet alcohols. Note that when the enolate’s carbon number is 2, a normal Guerbet
alcohol is formed.
synthesis over HAP by means of a new concept—namely, the prob-
ability of ethanol activation (α) on the catalyst surface. Scheme 3
shows the general formula for synthesis of Guerbet alcohols from
two lower alcohols. Generally, Guerbet alcohols synthesized from
normal alcohols are branched alcohols. There is an interesting ex-
ception to this, however. When one of the normal alcohols is an
enolate derived from ethanol, whose carbon number is 2, then
m = 1 in the formula shown in Scheme 3, and the Guerbet alco-
hol synthesized (C2n+2H4n+5OH) is a normal alcohol whose carbon
number is two carbons higher than that of the original normal al-
cohol (C2nH4n+1OH). On the basis of the above, we propose the
following two-part hypothesis for synthesis of Guerbet alcohols
from ethanol over HAP when ethanol conversion is low:
Total yields of Guerbet alcohols
ꢁ
ꢂ
ꢃ
ꢄ
∞
ꢀ
1
1
(1) Some molecules of ethanol are activated and react. (We de-
fined the probability of this activation as α.)
(2) Only normal alcohols react to synthesize Guerbet alcohols.
=
Cαn + 1 −
Cαn
,
where C: constant.
n − 1
n − 1
n=2
Here the first and second terms show the yield of normal Guerbet
alcohols and branched Guerbet alcohols, respectively.
Table 3 shows the combinations of normal alcohols that synthe-
size C2n Guerbet alcohols, the type of Guerbet alcohols synthesized
(normal or branched), and the probability of such Guerbet alco-
hols being generated according to the foregoing hypothesis. The
combination of normal alcohols (the carbon numbers of which
are shown in left–right juxtaposition in Table 3) is explained as
follows. Taking the synthesis of C6 Guerbet alcohols as an exam-
ple, (2, 4) indicates that an aldehyde sourced from ethanol (carbon
Fig. 4 illustrates the distribution of Guerbet alcohols synthe-
sized during ethanol conversion over CaO, MgO, β-TCP, and three
kinds of HAP catalyst at a conversion rate of 20%, and shows the
relationship between yield and carbon number for these alcohols.
The lowest line shows data obtained using a CaO catalyst, which
synthesized only a small yield of Guerbet alcohols. We believe that
selectivity to Guerbet alcohols over CaO was low because other

T. Tsuchida et al. / Journal of Catalysis 259 (2008) 183–189
189
synthesis of Guerbet alcohols from ethanol on HAP catalysts, we
compared the catalysts’ acid and basic properties with those of tra-
ditional base catalysts CaO and MgO. Based on our findings, we can
state the following conclusions:
1. Distribution of higher alcohols from ethanol over HAP by the
Guerbet reaction can be expressed with simple probability.
2. Yields of Guerbet alcohols over HAP can be quantitatively ex-
pressed as a function of the probability of activation (α) of the
raw material, ethanol.
3. α values differ for different catalysts and are influenced by the
degree of parallel reactions on the catalyst.
4. α values are correlated with the catalysts’ basic site density.
Fig. 5. Yields of Guerbet alcohols by carbon number at 8% yield of 1-butanol (all
lines cross here) on MgO and HAP catalysts with Ca/P molar ratios of 1.62, 1.65
and 1.67. The right column shows the probability of ethanol activation (α) on each
catalyst. Contact time, 1.78 s.
Acknowledgments
We thank the New Energy and Industrial Technology Devel-
opment Organization (NEDO) of Japan, which provided financial
assistance for this research. We also thank Dr. Roslyn Hayman for
advice.
parallel reactions occurring on the catalyst (e.g., dehydration, dehy-
drogenation, and the Lebedev reaction) took precedence over the
Guerbet reaction. The exponent of the slopes shown in the fig-
ure was α, and the smaller the slope, the larger the α value and
the greater the yield of Guerbet alcohols. Although the α value
for each catalyst was different, the linearity of the slopes indicated
that our hypothesis satisfied all of the experimental data obtained
for synthesis of Guerbet alcohols on these catalysts. In other words,
it supported our hypothesis that at the level of ethanol conversion
used in our study, the yield of Guerbet alcohols is regulated by the
degree of activation of surface ethanol.
References
[1] H. Aoki, Medical Applications of Hydroxyapatite, Ishiyaku EuroAmerica, Inc.
Tokyo, St. Louis, 1994, p. 1.
[2] J.C. Elliott, Structure and Chemistry of the Apatites and Other Calcium Or-
thophosphates, Elsevier Science B.V., Amsterdam, 1994, p. 1.
[3] T. Kanazawa, Inorganic Phosphate Materials, Kodansha, Tokyo, 1989, p. 15.
[4] L.C. Kibby, K.W. Hall, J. Catal. 31 (1973) 65.
As shown in Table 1, the density of basic sites on catalysts HAP-
2, -3, -4 and MgO was 0.024, 0.382, 0.534, and 0.133 μmol/m2,
respectively. As shown in Fig. 4, the probability of ethanol activa-
tion (α) for the same four catalysts was 0.082, 0.151, 0.162 and
0.134. This data confirmed that for catalysts HAP-2, -3, -4, and
MgO, the probability of ethanol activation over each catalyst in-
creased with the catalyst’s basic site density at the same ethanol
conversion rate. In addition, based on the data in Fig. 4, we postu-
late that the catalysts’ α values reflect the difference in the num-
ber of active sites contributing to the Guerbet reaction for each
catalyst at the same ethanol conversion rate. To confirm the char-
acter of the active sites that synthesize only Guerbet alcohols, we
therefore compared the distribution of Guerbet alcohols synthe-
sized from ethanol over HAP at the point in each reaction where
the yield of 1-butanol was 8%; these data are shown in Fig. 5. (In
the case of CaO and β-TCP, data were excluded, because the yield
of 1-butanol did not reach 8%.) We postulate that the high α value
of MgO (0.160) can be explained as follows. An aldehyde deriv-
ing from 1-butanol participates in the reaction as a regular basic
intermediate, causing the density of 1-butanol on the catalyst to
decrease and the ratio of higher Guerbet alcohols to increase, re-
sulting in a high estimated α value. HAP-3 and HAP-4 had roughly
the same α value (0.113 and 0.118, respectively), indicating their
virtually identical ability to catalyze the synthesis of Guerbet al-
cohols from ethanol. The α value of HAP-2 was lower, at 0.078.
We conclude that for HAP-2, activated ethanol was taken up in a
Lebedev reaction (diene synthesis, as indicated in Table 2), so that
synthesis of higher Guerbet alcohols on HAP-2 was restrained.
[5] J.S. Joris, H.C. Amberg, J. Phys. Chem. 75 (1971) 3167.
[6] F. Nozaki, T. Sodesawa, Hyoumen (in Japanese) 23 (1985) 399.
[7] H. Monma, Syokubai (in Japanese) 27 (1985) 237.
[8] S.A.J. Bett, K.W. Hall, J. Catal. 10 (1968) 105.
[9] L.C. Kibby, K.W. Hall, J. Catal. 29 (1973) 144.
[10] H. F-Milhofer, V. Hlady, F.S. Baker, R.A. Beebe, N.W. Wikholm, J.S. Kittelberger,
J. Colloid Interface Sci. 70 (1979) 1.
[11] N.S. Resende, M. Nele, V.M.M. Salim, Thermochim. Acta 451 (2006) 16.
[12] N. Elazarifi, A. Ezzamarty, J. Leglise, L.C. Menorval, C. Moreau, Appl. Catal. A
Gen. 267 (2004) 235.
[13] A. Venugopal, M.S. Scurrell, Appl. Catal. A Gen. 245 (2003) 137.
[14] R.E. Kreidler, A.F. Hummel, Am. Mineral. 55 (1970) 170.
[15] J.S. Joris, H.C. Amberg, J. Phys. Chem. 75 (1971) 3167.
[16] T. Tsuchida, S. Sakuma, T. Takeguchi, W. Ueda, Ind. Eng. Chem. Res. 45 (2006)
8634.
[17] T. Tsuchida, T. Yoshioka, S. Sakuma, T. Takeguchi, W. Ueda, Ind. Eng. Chem.
Res. 47 (2008) 1443.
[18] A.K. Lynn, W. Bonfield, Acc. Chem. Res. 38 (2005) 202.
[19] S. Sugiyama, T. Miyamoto, H. Hayashi, J.B. Moffat, J. Mol. Catal. A Chem. 135
(1998) 199.
[20] H.B. Lu, C.T. Camphell, D.J. Graham, B.D. Ratner, Anal. Chem. 72 (2000) 2886.
[21] K. Kieswetter, T.W. Bauer, S.A. Brown, F.V. Lente, K. Merritt, Biomaterials 15
(1994) 183.
[22] M. Yoshinari, Y. Ohtsuka, T. Derand, Biomaterials 15 (1994) 529.
[23] H. Aoki, Hyoumenkagaku (in Japanese) 10 (1989) 14.
[24] J.P. Wibaut, US Patent 1 910 582 (1933), to Naamlooze Vennootschap de Bataaf-
sche Petroleum Maatschappij.
[25] R.T. Clark, C.C. Tex, US Patent 3 972 952 (1976), to Celanese Corporation.
[26] W. Ueda, T. Kuwabara, T. Ohshida, Y. Morikawa, J. Chem. Soc. Chem. Commun.
(1990) 1558.
[27] W. Ueda, T. Ohshida, T. Kuwabara, Y. Morikawa, Catal. Lett. 12 (1992) 97.
[28] M. Xu, M.J.L. Gines, A.-M. Hilmen, B.L. Stephens, E. Iglesia, J. Catal. 171 (1997)
130.
[29] M.L.J. Gines, E. Iglesia, J. Catal. 176 (1998) 155.
4. Conclusion
[30] J.I.D. Cosimo, C.R. Apesteguia, M.J.L. Gines, E. Iglesia, J. Catal. 190 (2000) 261.
[31] C. Carlini, M. Marchionna, M. Noviello, R.M.A. Galletti, G. Sbrana, F. Basile, A.
Vaccari, J. Mol. Catal. A Chem. 232 (2005) 13.
[32] V.K. Diez, C.R. Apesteguia, J.I.D. Cosimo, J. Catal. 240 (2006) 235.
[33] B.B. Corson, E.H. Jones, E.C. Welling, A.J. Hinckley, E.E. Stahly, Ind. Eng.
Chem. 42 (1950) 359.
[34] K.S. Bhattacharyya, D.N. Ganguly, J. Appl. Chem. 12 (1962) 97.
[35] K.S. Bhattacharyya, D.N. Ganguly, J. Appl. Chem. 12 (1962) 105.
[36] R. Ohnishi, T. Akimoto, K. Tanabe, J. Chem. Soc. Chem. Commun. (1985) 1613.
[37] Y. Kitayama, M. Satoh, T. Kodama, Catal. Lett. 36 (1996) 95.
Our findings indicate that HAP catalysts, without the addition
of noble metals, transition metals, or halogens, such as chlorine,
can be used for the highly selective synthesis of valuable com-
pounds, such as 1-butanol and 1,3-butadiene, from ethanol merely
by controlling the catalyst’s Ca/P molar ratio. We found that prod-
uct selectivity bore a strong correlation to the acid and basic prop-
erties of the HAP catalysts. In seeking to explain the characteristic