Fluorine-modified Mg-Al mixed oxides: A solid base with variable basic sites and tunable basicity

Article abstract of DOI:10.1016/j.apcata.2010.01.023

The fluorine-modified Mg-Al mixed oxides were synthesized by thermal decomposition of the fluorine-containing Mg-Al hydrotalcites, and their physicochemical properties were characterized by ICP, TGA, XRD, FTIR, CO₂-TPD and N₂ adsorption/desorption techniques. It was found that weak basic sites were gradually transformed into moderate and strong basic sites during thermal decomposition, and thus weak basic sites (OH^- groups), moderate basic sites (Mg-O, Mg-F and Al-O pairs) and strong basic sites (coordinatively unsaturated F^- and O^2- ions) were all produced in the as-prepared samples. Furthermore, the basicity of moderate basic sites could be controlled by the change of fluorine content, which caused the change in the amounts of Mg-F pairs. In the synthesis of propylene glycol methyl ether from methanol and propylene oxide, the base-catalytic performance of the obtained samples was shown to be closely associated with their moderate basic sites.

Full text of DOI:10.1016/j.apcata.2010.01.023

Applied Catalysis A: General 377 (2010) 107–113
Contents lists available at ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Fluorine-modified Mg–Al mixed oxides: A solid base with variable basic sites and
tunable basicity
a
a,
b,
b
Gongde Wu , Xiaoli Wang *, Wei Wei *, Yuhan Sun
a
Department of Environment and Technology, Nanjing Institute of Technology, Nanjing 211167, PR China
b
State Key Laboratory of Coal Conversion, Institute of Coal Chemistry, Chinese Academy of Sciences, Taiyuan 030001, PR China
A R T I C L E I N F O
A B S T R A C T
Article history:
The fluorine-modified Mg–Al mixed oxides were synthesized by thermal decomposition of the fluorine-
containing Mg–Al hydrotalcites, and their physicochemical properties were characterized by ICP, TGA,
Received 24 July 2009
Received in revised form 28 December 2009
Accepted 13 January 2010
Available online 20 January 2010
2 2
XRD, FTIR, CO -TPD and N adsorption/desorption techniques. It was found that weak basic sites were
gradually transformed into moderate and strong basic sites during thermal decomposition, and thus
À
weak basic sites (OH groups), moderate basic sites (Mg–O, Mg–F and Al–O pairs) and strong basic sites
À
2À
(
coordinatively unsaturated F and O
ions) were all produced in the as-prepared samples.
Keywords:
Furthermore, the basicity of moderate basic sites could be controlled by the change of fluorine content,
which caused the change in the amounts of Mg–F pairs. In the synthesis of propylene glycol methyl ether
from methanol and propylene oxide, the base-catalytic performance of the obtained samples was shown
to be closely associated with their moderate basic sites.
Hydrotalcite
Mg–Al mixed oxides
Fluorine
Tunable basicity
Propylene oxide
Methanol
ß 2010 Elsevier B.V. All rights reserved.
1
. Introduction
In recent years the solid base catalysts have received extensive
the literature [18–20]. It was found that weak, moderate and
strong basic sites were all produced in Mg–Al mixed oxides, these
À
sites were attributed to OH groups, Mg–O and Al–O pairs, and
2À
attention in general and fine chemical manufactures, because they
are less corrosive, easy to separate from the reaction and easy to
reuse [1,2]. The materials previously studied as solid bases include
the alkaline earth oxides [3,4], alkali metals or alkali metal
compounds supported on metal oxides [5,6], zeolites [7,8],
oxynitrides [9,10], mesoporous silica functionalized with amino
groups [11,12], hydrotalcite-like compounds and their calcined
derivations [13–15]. Although so many basic materials were
reported, there were only 10 solid base-classified industrial
processes compared to 103 for solid acids [16,17]. One reason
for the restricted application is that the generation of basic sites is
still not clearly understood [1,2]. Thus, there is an urgent need to
carry out an in-depth study of solid base, such results will be of
significance for the design of highly effective solid base catalysts.
Among the aforementioned basic materials, the Mg–Al mixed
oxides obtained from hydrotalcites were considered as promising
catalysts and catalyst supports owing to their advantages of high
surface area and thermal stability, along with the tunable basicity.
Furthermore, their base generation had been clearly described in
coordinatively unsaturated O ions, respectively. Since Davis and
Derouane first reported the Mg–Al mixed oxides modified by
platinum in 1991 [21], many groups tried to modify Mg–Al mixed
oxides with different cations in attempts to improve their basicity.
B ˆı rjega et al. found that the basicity of the Mg–Al mixed oxides
modified by rare-earth elements was significantly enhanced in
comparison with that of the pure Mg–Al mixed oxides [22].
Dussault et al. successfully prepared the Mg–Al mixed oxides
2
+
2+
modified by Ni and/or Cu , and the resulting materials also
exhibited much higher basicity than the pure Mg–Al mixed oxides,
especially with the Ni/Mg/Al molar ratio of 1/1/1 [23]. However, to
the best of our knowledge, there have been few reports on the
anion-modified Mg–Al mixed oxides and their applications.
It was well known that the fluoride ion was a basic anion and
promoted many base-catalytic reactions [24–26]. Choudary and
Gao et al. reported that layered double hydroxide fluoride showed
high catalytic performance both in Michael addition and in
transesterification reactions [27,28]. Kemnitz prepared new
magnesium oxide fluoride with hydroxy groups, and the resulting
sample was also an effective catalyst in Michael addition [29,30].
Clark summarized in detail the advantages and disadvantages of
the most popular fluoride ion sources, which comprised KF, CsF,
KF–(18-crown-6), TBAF, TEAF, BTMAF, fluoride resins, TBAF–silica
*
Corresponding authors. Tel.: +86 25 86118960; fax: +86 25 86118974.
E-mail addresses: wangxiaoli212@njit.edu.cn, wangxiaoli212@126.com
(X. Wang), weiwei@sxicc.ac.cn (W. Wei).
2 3 2 3
and KF/Al O [25]. Among them, the heterogeneous KF/Al O had
0926-860X/$ – see front matter ß 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.apcata.2010.01.023

1
08
G. Wu et al. / Applied Catalysis A: General 377 (2010) 107–113
Table 1
been widely used by organic chemists because of its inexpensive-
ness, efficient catalytic performance and easy post-treatment. But
the idea of its basic centers was remained controversial during the
past two decades [26]. Simultaneously, the support of alumina
with low and untunable basicity also limited the further
Composition and structure of samples.
a
2
Samples
Temp. (K)
F content (wt%)
S
BET (m /g)
F5.8/Mg(Al)O-573
573
673
723
873
1073
723
723
723
723
5.8
6.2
6.4
6.6
6.6
0
24
56
78
36
2
F6.2/Mg(Al)O-673
application of the heterogeneous KF/Al
2
O
3
. Accordingly, it is
F6.4/Mg(Al)O-723
F6.6/Mg(Al)O-873
F6.6/Mg(Al)O-1073
essential to investigate systematically the basic centers of the
fluorine-containing materials and to look for more excellent
supports for fluoride ions.
Mg(Al)O-723
226
85
F1.6/Mg(Al)O-723
F3.3/Mg(Al)O-723
F11.8/Mg(Al)O-723
1.6
3.3
11.8
Recently, we found that fluoride ion could be introduced into
Mg–Al mixed oxides by thermal decomposition of the fluorine-
containing Mg–Al hydrotalcites. The resulting fluorine-modified
Mg–Al mixed oxides showed much higher basicity than the pure
Mg–Al mixed oxides, which was ascribed to the appearance of Mg–
83
55
a
Decomposition temperature.
À
F pairs and coordinatively unsaturated F ions [31]. However, the
changing the concentration of NH
4
F solution (0.025, 0.05, 0.1,
process of base generation was not clear, and it was also difficult to
control the base of the fluorine-modified Mg–Al mixed oxides. In
the present work, we further investigated the base of the fluorine-
modified Mg–Al mixed oxides under different decomposition
temperatures. The results revealed that weak basic sites were
gradually transformed into moderate and strong basic sites during
thermal decomposition. Moreover, the amounts of moderate basic
sites could be controlled by the change of fluorine content. The
resulting solid base catalysts also showed high catalytic perfor-
mance in the synthesis of propylene glycol methyl ether from
methanol and propylene oxide (PO).
0.2 mol/L), other F /Mg(Al)O-723 (x = 1.6, 3.3, 6.4, 11.8) was also
synthesized with different fluorine contents (see Table 1).
x
2.2. Characterization
Elemental chemical analysis was performed using the induc-
tively coupled plasma-optical (ICP) emission spectroscopy (Perki-
nElmer ICP OPTIMA–3000). Thermogravimetric analysis (TGA) was
carried out on a Setaram TGA–92 thermal analyzer connected to a
PC via a TAC7/DX thermal controller. The sample was heated from
303 to 1073 K at 10 K/min under nitrogen. Nitrogen adsorption/
desorption isotherms were obtained at 77 K on a Micromeritics
ASAP-2000 instrument (Norcross, GA), using static adsorption
procedures. Samples were degassed at 573 K for a minimum of 5 h
2
. Experimental
À6
2.1. Catalyst preparation
under vacuum (10 Torr) prior to the measurement. Surface areas
were measured using the BET method. Powder X-ray diffraction
(XRD) experiments of samples were carried out on a Rigaku
2
À
(HT–CO
The Mg–Al hydrotalcite intercalated with CO
3
3
) was
synthesized using alkali-free procedures followed by controlled
hydrothermal treatment [32,33]. Typically, a 120 mL aqueous
Miniflex diffractometer using a Cu target with a Ni filter in a 2u
range of 5–708. The X-ray gun was operated at 50 kV and 30 mA,
using a scan speed rate of 0.28/min. The Fourier transform infrared
(FTIR) spectroscopy results of the samples were recorded on a
NICOLET NEXUS 470FT-IR spectrophotometer after the samples
were pressed into 13 mm discs with freshly dried KBr salt. The
solution A contained 0.09 mol of MgCl
2
Á6H
O, and a 120 mL aqueous solution B contained 0.16 mol of
CO . The two solutions were added dropwise with stirring to
2
O and 0.03 mol of
AlCl
3
4
Á3H
2
(NH
)
2
3
1
20 mL deionized water at room temperature. The addition was
performed over 1.0–1.5 h, and the pH value of the obtained mixed
solution was maintained close to 8.0 by the addition of appropriate
2 2
effluents of CO temperature-programmed desorption (CO -TPD)
were analyzed with a gas chromatograph that employed a thermal
conductivity detector. Catalysts (0.1 g, 40–60 mesh) were placed in
a quartz reactor bed. After 3 h pretreatment in Ar flow at the
heating temperatures of each sample, the catalysts were cooled to
4
amounts of NH OH (35% aqueous ammonia solution). The
resulting gel-like slurry was transferred to an autoclave and
hydrothermally treated at 373 K for 24 h. The as-prepared product
was collected after being filtered and washed with deionized water
until pH = 7. Then the precipitate was dried at 373 K in air for
approximately 12 h. The resulting hydrotalcite was heated in an
oven at 723 K for 8 h under a flowing stream of pure nitrogen. Then
the Mg–Al mixed oxide was obtained, it is denoted as Mg(Al)O-
room temperature. CO
valve till reaching saturation. Once the physically adsorbed CO
was purged off, the CO -TPD experiments were started from 303 K
2
was pulsed to the reactor using a six-way
2
2
to the heating temperatures of the samples with a heating rate of
10 K/min under Ar flow (50 mL/min).
7
23.
The fluorine-containing Mg–Al hydrotalcite (HT–F) was pre-
2.3. Catalytic test
pared using the ‘‘memory effect’’ according to the literature [28].
Typically, the as-prepared Mg(Al)O-723 (1.0 g) was treated with an
The synthesis of propylene glycol methyl ether from methanol
and PO was studied in a 100 mL stainless steel autoclave equipped
with a magnetic stirrer. In a typical experiment, 0.3 g of freshly
activated catalyst, 5.8 g of POand 16 g of methanolwereintroduced;
then the reaction was performed at 393 K for 15–900 min under
autogenous pressure. The products were analyzed using a gas
chromatograph with a capillary 30 m HP-5 column and a flame
ionization detector after centrifugal separation from the catalyst.
4 2
aqueous solution of NH F (0.1 mol/L, 100 mL) under N atmo-
sphere with stirring at room temperature for 48 h, and the pH
value of the solution was adjusted to 8.0 by the addition of
4
appropriate amounts of NH OH (35% aqueous ammonia solution).
To prevent CO from contaminating the aqueous solution,
2
deionized decarbonated (DD) water was used here. The resulting
product was filtered, washed extensively with the DD water and
2
dried at 373 K for 12 h under N atmosphere. The HT–F was thus
prepared. The obtained HT–F was further heated at the desired
temperatures in an oven for 8 h under a flowing stream of pure
nitrogen. Then the fluorine-modified Mg–Al mixed oxides were
3. Results and discussion
3.1. Thermal evolution and structure
obtained at different temperatures and denoted as F
T = 573, 673, 723, 873, 1073 K), where the subscript x and T stand
for fluorine content and heating temperature, respectively. By
x
/Mg(Al)O-T
(
The thermal stability of the as-prepared HT–F was examined by
TGA experiment. Two major weight loss peaks are observed in

G. Wu et al. / Applied Catalysis A: General 377 (2010) 107–113
109
obtained F
hysteresis loops (see Fig. 2a), characteristic of a mesoporous solid
38,39]. Simultaneously, the pore-size distribution gradually
x
/Mg(Al)O-T all displayed type IV isotherms with clear
[
narrowed (see Fig. 2b), and a significant improvement in BET
surface area was also found (see Table 1). This suggested that
porosity was enhanced with the rise of decomposition tempera-
ture. However, when the temperature was further increased from
x
723 to 1073 K, the hysteresis loop of F /Mg(Al)O-T gradually
weakened. At the same time, the pore-size distribution curve
broadened and was nearly a straight line at 1073 K. These changes
might be related to the collapse of porous structure, which was in
agreement with the significant decrease in BET surface area.
The XRD patterns of HT–CO
They both showed the pattern characteristic of hydrotalcites
28,32,40,41], indicating that HT–CO and HT–F were successfully
prepared. And the basal spacing of HT–F (9.02 A) was higher than
3
and HT–F are displayed in Fig. 3a.
[
3
Fig. 1. Thermogravimetric weight loss curve and derivative plot of the fluorine-
˚
containing hydrotalcite.
À
˚
3
(7.74 A). This suggested that F anions with less
that of HT–CO
electrostatic force took the place of CO
However, the 110 peaks of the two hydrotalcites were observed at
the same 2 value, indicating that HT–F and HT–CO still kept the
2
À
Fig. 1. The first peak was associated with the elimination of
physically adsorbed and interlayer water molecules, while the
second composite peak was due to the removal of interlayer anions
and the dehydroxylation of layer hydroxyl groups [19]. During the
first weight loss (340–470 K), the HT–F layered structure was not
destroyed, but the gallery spacing decreased. And the interaction
might take place between fluorine anions and the brucite-like
layers [34]. During the second weight loss (600–840 K), the
amorphous species were formed due to the complete collapses of
the HT–F layered structure [35–37]. Above 900 K, no obvious
weight loss was observed, but the crystallization of new phases
took place [35]. Thus, the phase changes of samples were closely
associated with the temperatures of their thermal decompositon.
3
anions in HT interlayer.
u
3
same Mg/Al molar ratio.
The XRD patterns of
x
F /Mg(Al)O-T obtained at different
decomposition temperatures are illustrated in Fig. 3b. At the
temperature of 573 K, HT characteristic peaks of 0 0 3, 1 1 0 and
1 1 3 were still present in the XRD pattern. Simultaneously, two
2
new phases of Mg–Al mixed oxides and MgF (PDF no. 41–1443)
were also observed. This indicated that HT–F was partially
decomposed and converted into the two new phases. For the
sample obtained at 673 K, the diffraction peaks due to HT phase
completely disappeared, and only two phases of Mg–Al mixed
The textural properties of F
decomposition temperatures are valued by N
tion measurement. When HT–F was heated from 573 to 723 K, the
x
/Mg(Al)O-T obtained at different
2
oxides and MgF were detected, suggesting that HT–F was
2
adsorption/desorp-
completely decomposed at this temperature. When the tempera-
ture was further increased from 673 to 723 K, the peak intensities
2 x x
Fig. 2. (a) N adsorption/desorption isotherms of F /Mg(Al)O-T obtained at different decomposition temperatures; (b) pore-size distributions of F /Mg(Al)O-T obtained at
different decomposition temperatures.
x 2
Fig. 3. XRD patterns: (a) the HT precursors; (b) F /Mg(Al)O-T obtained at different decomposition temperatures: (&) hydrotalcite, (&) MgF , (*) mixed oxides, (*) spinel.

1
10
G. Wu et al. / Applied Catalysis A: General 377 (2010) 107–113
3.2. Surface base
The CO
obtained at different decomposition temperatures are given in
Fig. 5a. It was found that they all showed three CO desorption
peaks, which could be assigned to weak basic sites (around 410–
20 K), moderate basic sites (around 455–465 K) and strong basic
2 2 x
-TPD profiles of Mg(Al)O-723, MgF and F /Mg(Al)O-T
2
4
sites (around 520–540 K). For Mg(Al)O-723, weak basic sites were
À
ascribed to OH groups, moderate basic sites were related to Mg–O
and Al–O pairs, and strong basic sites were associated with
2
À
coordinatively unsaturated O ions [19,20]. For MgF
2
, its weak,
À
moderate and strong basic sites could be assigned to the OH
groups, Mg–F pairs and coordinatively unsaturated F ions,
À
respectively [31]. XRD characterization showed that the diffraction
peaks assigned to MgF
in the patterns of F /Mg(Al)O-T. Thus, the F
contain the above six basic centers that originated from Mg(Al)O-
2
and Mg–Al mixed oxides were both present
x
x
/Mg(Al)O-T should
7
23 and MgF
2
. According to the CO
2
desorption temperatures of
2 x
Fig. 4. FTIR spectra of MgF and F /Mg(Al)O-T obtained at different decomposition
À
temperatures.
the six basic centers, the OH groups as well as the Mg–O, Al–O and
Mg–F pairs were reasonably classified into weak basic sites and
moderate basic sites, respectively. The remaining two basic centers
of the coordinatively unsaturated F and O
assigned to strong basic sites. Prescott and Verziu reported that F
assigned to the above two new phases gradually increased,
indicating an increase in their crystallinity. Noticeably, at the
À
2À
ions could be
À
temperature of 873 K, another new phase of MgAl
no. 21–1152) appeared, which led to the collapse of porous
structure and the decrease in BET surface area of F /Mg(Al)O-T.
Interestingly, it was found that MgF phase completely disap-
2
O
4
spinel (PDF
2
À
À
and O ions were similar ‘‘basic’’ anions, and that F ion was a
2
À
x
weaker basic center than O ions [6,29]. So, the above assignment
is reasonable, and the broad peak around 520–540 K might
originate from the overlapping of the two peaks of the
coordinatively unsaturated F and O ions. Accordingly, there
were six kinds of basic centers: OH groups (weak basic sites), Mg–
2
peared at 1073 K. Moreover, the peak intensities of Mg–Al mixed
oxides sharply increased, and their peak positions shifted to lower
À
2À
À
2
u
values. These changes all suggested that MgF
lattice of Mg–Al mixed oxides to form a solid solution.
The FTIR spectra of MgF and F /Mg(Al)O-T obtained at different
2
dissolved into the
O, Mg–F and Al–O pairs (moderate basic sites), coordinatively
À
2À
2
x
unsaturated F and O ions (strong basic sites) present in F
Mg(Al)O-T. In order to calculate the amounts of the three kinds of
basic sites, we divided the CO -TPD profile of F5.8/Mg(Al)O-573
x
/
decomposition temperatures are shown in Fig. 4. An intensive
À1
sharp peak at about 454 cm was observed for MgF
2
, this peak
dimer [42,43]. Such a
characteristic band was also found without any band shift for F
Mg(Al)O-T obtained at 573 K (F5.8/Mg(Al)O-573) [43], confirming
that the new species of MgF had formed in F /Mg(Al)O-T.
2
was attributed to the ring stretch of a Mg
2
F
4
into three peaks (see Fig. 5a), and the amounts of the three kinds of
basic sites were calculated from the peak area. Similarly, the
basicity of the other samples was also carefully calculated, and the
results are listed in Table 2. One could see that the total basicity of
samples increased with the increasing of decomposition tempera-
ture and reached the maximum at 723 K. However, as the
decomposition temperature was further increased, the total
basicity sharply decreased and reached the minimum at 1073 K.
This could be related to the gradual collapse of porous structure
x
/
2
x
However, this characteristic band gradually shifted to high-
wavenumber region with the rise of decomposition temperature,
À1
and the position of the characteristic band increased to 487 cm
for the sample obtained at 1073 K. This could be attributed to the
2
variable geometry of MgF at different decomposition tempera-
tures [43–45], such a change was consistent with the XRD results
that the phases of samples varied from 573 to 1073 K. In addition,
the intensive peaks at 3435 and 1636 cm could be attributed to
2 4
due to the presence of the new MgAl O phase at higher
temperature, which led to a sharp decrease in the amounts of
accessible basic sites.
À1
the
y
OH stretching and
d
OH bending vibrations of hydroxyl groups,
Noticeably, the relative amounts of the three kinds of basic sites
were also varied with the decomposition temperature (see
Table 2). From 573 to 723 K, the relative amounts of weak and
moderate basic sites gradually decreased, whereas strong basic
À1
and the broad band around 1400 cm was ascribed to the
vibration of CO
contamination during measurement.
y
3
2
À
3
.
These peaks all derived from the air
Fig. 5. CO
2 2 x x
-TPD profiles: (a) Mg(Al)O-723, MgF and F /Mg(Al)O-T obtained at different decomposition temperatures (the broken lines were assigned to F /Mg(Al)O-573), (b)
F
x
/Mg(Al)O-723 with different fluorine contents.

G. Wu et al. / Applied Catalysis A: General 377 (2010) 107–113
111
Table 2
a
Basicity and basic-catalytic performance of samples .
b
c
d
d
À1
)
Samples
CO
2
uptake (
m
mol/g)
W.
Selectivity (%)
Initial Rate
mmol/g min)
TOF (s
(
Total
M.
S.
PPM
SPM
e
F
F
F
F
F
5.8/Mg(Al)O-573
6.2/Mg(Al)O-673
6.4/Mg(Al)O-723
6.6/Mg(Al)O-873
6.6/Mg(Al)O-1073
62.9
110.0
128.1
43.7
6.5 (10.3)
9.3 (8.5)
9.7 (7.6)
5.1 (11.7)
2.2 (13.5)
5.6
27.7 (44.1)
46.2 (41.9)
39.8 (31.0)
19.1 (43.6)
7.1 (43.6)
21.8
28.7 (45.6)
54.5 (49.6)
78.6 (61.4)
19.5 (44.7)
7.0 (42.9)
53.5
85.2
87.0
86.6
83.8
70.4
79.9
81.9
84.6
83.2
74.8
14.8
13.0
13.4
16.2
29.6
20.1
18.1
15.4
16.8
25.2
0.87
1.51
1.31
0.76
0.63
0.27
0.59
0.88
0.98
0.06
0.52
0.54
0.55
0.66
1.47
0.21
0.34
0.44
0.51
2.0
16.3
Mg(Al)O-723
80.9
F1.6/Mg(Al)O-723
F3.3/Mg(Al)O-723
F11.8/Mg(Al)O-723
112.7
120.4
115.1
25.0
5.8
29.2
77.7
8.4
33.4
78.6
7.5
31.8
75.8
MgF
2
2.2
0.5
22.3
a
Reaction conditions: catalyst, 0.3 g; methanol, 16.0 g; PO, 5.8 g; temperature, 393 K.
W.: weak basic sites; M.: moderate basic sites; S.: strong basic sites.
Reaction time, 600 min.
b
c
d
2
Values are given over the first 15 min. Turnover frequency bases on the CO adsorption of the moderate basic sites, given in mol PO converted per mol of moderate basic
sites in the catalyst per second.
e
Relative amounts of the basic sites (%).
sites increased (F5.8/Mg(Al)O-573 to
F
6.4/Mg(Al)O-723). This
defects’’ formed in the lattice of Mg–Al mixed oxides as the
samples became richer in fluorine [18]. The latter was related to
that the increasing number of Mg–F pairs, which was caused by the
indicated that weak and moderate basic sites might be partially
transformed into strong basic sites with the rise of decomposition
temperature. It is well known that hydrotalcites only show weak
basic sites that originate from interlayer OH groups [13,14,46,47].
During the thermal decomposition of HT–F, the dehydroxylation
took place, and thus Mg–F, Mg–O and Al–O pairs were produced in
À
increasing number of F ions in HT–F interlayer. Then the amounts
À
of moderate basic sites increased significantly [28,40,48]. Howev-
er, the basicity of strong basic sites was almost independent of the
fluorine content at 723 K. As mentioned above, the HT phase
disappeared at 673 K, and fluoride ions were transformed entirely
x
the resulting F /Mg(Al)O-T. Then the three pairs could partially
À
break up, which led to the generation of coordinatively unsaturat-
into coordinatively unsaturated F ions. So the strong basic sites
À
2À
2À À
ed ions (F or O ). Simultaneously, a small quantity of fluoride
ions, which intercalated into the interlayer of HT–F precursors,
might also be transferred to the layers and then directly form
(O
or F ) in F /Mg(Al)O-T, which were generated in the
x
temperature domain from 673 to 723 K, were mainly related to
the cleavage of moderate basic sites (Mg–F, Mg–O and Al–O pairs).
Owing to the different bond strengths of the three pairs, Mg–O
pairs (363.2 Æ 12.6 kJ/mol) should be easy to break up in comparison
with Mg–F and Al–O pairs (461.9 Æ 5.0 and 511.0 Æ 3.0 kJ/mol) at
723 K [49]. Results indicated that the generation of strong basic sites
at 723 K showed principally be ascribed to the cleavage of the Mg–O
À
coordinatively unsaturated F ions [29,30,34]. The XRD results
reveal that the layer structure of HT–F has been completely
destroyed at 673 K, indicating that the interlayer fluoride ions have
been already transformed into coordinatively unsaturated F ions
À
at 673 K. Thus, above the decomposition temperature of 673 K,
strong basic sites only come from the cleavage of moderate basic
sites (Mg–F, Mg–O and Al–O pairs). The process of basic site
generation is illustrated in Scheme 1. Noticeably, the relative
amounts of the three kinds of basic sites almost no longer changed
from 873 to 1073 K (F6.6/Mg(Al)O-873 and F6.6/Mg(Al)O-1073),
indicating that the transformation of weak basic sites into
moderate and strong basic sites had finished. Consequently, the
decomposition temperature played an important role in the
generation of different basic sites.
pairs. For the as-prepared F
(Mg–O and Mg–F) only occupies very small part of the entire Mg
content in F /Mg(Al)O-723, and their Mg–O pair amount will not be
affected by the increasing of Mg–F pairs. So the basicity of their strong
basic sites was nearly identical in the all F /Mg(Al)O-723. In addition,
x
/Mg(Al)O-723, the Mg content in pairs
x
x
when the fluorine content was further increased to 11.8%, the basicity
of F11.8/Mg(Al)O-723 was found to decrease. This might be due to the
2
formation of more MgF , which decreased remarkably the BET surface
area and reduced sharply the amounts of accessible basic sites.
Furthermore, the influence of fluorine content on the base of F
Mg(Al)O-723 was also investigated. The CO -TPD profiles of F
x
x
/
/
2
3.3. Base-catalytic performance
Mg(Al)O-723 are illustrated in Fig. 5b. It was found that the basicity
of weak and moderate basic sites both reached the maximum at
the fluorine content of 6.4%. The former was ascribed to ‘‘many
Propylene glycol methyl ether is an important solvent in the
field of printing inks, dyes and agricultural chemicals, and it is
Scheme 1. The possible generation process of basic sites during thermal decomposition (W.: weak basic sites; M.: moderate basic sites; S.: strong basic sites).

1
12
G. Wu et al. / Applied Catalysis A: General 377 (2010) 107–113
x 2 x
Fig. 6. Influence of reaction time on the PO conversion: (a) F /Mg(Al)O-T obtained at different decomposition temperatures, (b) MgF and F /Mg(Al)O-723 with different
fluorine contents.
expected to function as a safe substitute for toxic ethylene glycol
ether [3]. Among its several synthesis methods, the propylene
oxide method is the most convenient and industrially feasible [50].
And the mildly toxic 1-methoxy-2-propanol (PPM) is the principal
product with low selectivity to 2-methoxy-1-propanol (SPM)
under the base condition [3]. Thus, in the present work, the
sites, while there is a linear relation between PO conversion and
the amount of moderate basic sites (see Fig. 7). Thus, the moderate
basic sites might play an important role in this reaction.
Furthermore, the relation of PO initial rates and the amount of
moderate basic sites was further investigated. PO initial rates were
calculated from Fig. 6 and all listed in Table 2. It was found that F6.2
/
catalytic performance of the as-prepared F
in the synthesis of propylene glycol methyl ether from methanol
and PO.
Fig. 6a and b depicts the relation between reaction time and PO
conversion over all the samples. It can be found PO conversion
significantly increased with reaction time until 600 min. This
indicated that the reaction had almost reached equilibrium at
x
/Mg(Al)O-T was tested
Mg(Al)O-673 with the maximum amounts of moderate basic sites
exhibited the maximum initial rate. This also suggested that the
moderate basic sites might be active sites in this base-catalytic
reaction. In order to further understand the role of the moderate
x
basic sites, we also investigated the catalytic performance of F /
Mg(Al)O-723 (x = 0, 1.6, 3.3, 6.4, 11.8). The results displayed that
PO initial rate was gradually increased with the fluorine content
and reached the maximum at the fluorine content of 6.4%.
Obviously, such a tendency was good in agreement with that for
the varied amounts of moderate basic sites, which further
indicated that the moderate basic sites were catalytic active sites
in this base-catalytic reaction.
6
00 min, and the optimum PO conversion had reached 96.1% with
PPM selectivity of 87.0% over F6.2/Mg(Al)O-673 as catalyst (see
Table 2). Generally speaking, the catalytic performance of solid
bases was not only related to their basic sites, but also correlated
x
with their structure. For the F /Mg(Al)O-T obtained by thermal
decomposition of HT-F at different temperature, both their basic
sites and their structure were varied with the decomposition
temperature. So there would be no obvious relation between their
Considering that the moderate basic sites were composed of
Mg–O, Mg–F and Al–O pairs, we then investigated the differences
in their activities. First of all, MgF
among the catalysts. This suggested that Mg-F pairs are the most
active among three pairs. Secondly, F6.6/Mg(Al)O-873 and F6.6
Mg(Al)O-1073 with less moderate basic sites (19.1 and 7.1 mol/
g) showed higher PO initial rates in comparison with Mg(Al)O-723
(21.8 mol/g). This should be ascribed to their possesing Mg–F
pairs, which were more active than Mg–O and Al–O pairs in this
reaction. And the analogous results were also found between F5.8
Mg(Al)O-573 and F1.6/Mg(Al)O-723. F5.8/Mg(Al)O-573 with the
relatively less moderate basic sites (27.7 mol/g) showed higher
PO initial rate than F1.6/Mg(Al)O-723 with more moderate basic
sites (29.2 mol/g), which was also ascribed to the appearance of
2
shows the highest TOF value
x
catalytic performance and their basic sites. However, for the F /
Mg(Al)O-T with different fluorine content obtained at 723 K, their
catalytic performance could be closely associated with their basic
sites owing to their similar structure. The relation between PO
conversion and the amounts of three basic sites was investigated in
detail. And the result showed that there is no obvious correlation
between PO conversion and the amount of weak or strong basic
/
m
m
/
m
m
more Mg–F pairs. In addition, Mg–O pairs were more efficient than
Al–O pairs in base-catalytic reactions due to their higher basic
strength [31]. As a result, the catalytic activities of the three pairs in
this base-catalytic reaction decreased in the following sequence:
Mg–F > Mg–O > Al–O. In order to further understand the propor-
tion of three pairs among moderate basic sites. The TOF values of
moderate basic sites are illustrated in Table 2. And it can be seen
that the TOF values are different, suggesting that not a single pair
but the three pairs acted as an active site together in this reaction.
x
For F /Mg(Al)O-T obtained at different temperatures, their TOF
values always increased with the increasing of decomposition
temperature and reached the maximum at 1073 K. This was
related to the increasing proportion of Mg–F pairs during thermal
decomposition. When the decomposition temperature increased
À
from 573 to 673 K, interlayer F ions were grafted onto layers to
Fig. 7. The correlation between PO conversion and the amounts of moderate basic
form more Mg–F pairs. With the increasing of decomposition
temperature from 673 to 723 K, Mg–O pairs had priority for
sites in F
x
/Mg(Al)O-723 with different fluorine contents. Reaction conditions:
catalyst, 0.3 g; methanol, 16.0 g; PO, 5.8 g; temperature, 393 K; time, 600 min.

G. Wu et al. / Applied Catalysis A: General 377 (2010) 107–113
113
[
[
7] Z.-H. Fu, Y. Ono, Catal. Lett. 21 (1993) 43–47.
8] M.D. Romero, G. Ovejero, A. Rodr ı´ guez, J.M. G o´ mez, I. A´ gueda, Ind. Eng. Chem. Res.
breaking up compared with Mg–F and Al–O pairs, so the
proportion of Mg–F pairs increased among moderate basic sites.
This could afford the increase in the TOF value of moderate basic
sites. As the decomposition temperature further increased from
43 (2004) 8194–8199.
[9] P.W. Lednor, Catal. Today 15 (1992) 243–261.
10] M.J. Climent, A. Corma, V. Forn e´ s, R.M. Martin-Aranda, F. Rey, J. Catal. 134 (1992)
[
[
58–65.
7
23 to 1073 K, the proportion of Mg–F pairs among moderate basic
sites also increased due to the appearance of MgAl spinel which
reduced the amounts of Mg–O and Al–O pairs. So the TOF value
increased and reached the maximum at 1073 K. For the F
Mg(Al)O-723 (x = 0, 1.6, 3.3, 6.4, 11.8), the TOF value of moderate
basic sites increased from 0.21 (Mg(Al)O-723) to 0.55 (F6.4
11] D.J. Macquarrie, D.B. Jackson, Chem. Commun. (1997) 1781–1782.
[12] X. Zhang, N. Zhao, W. Wei, Y. Sun, Catal. Today 115 (2006) 102–106.
[13] B.F. Sels, D.E. De, P.A. Vos, Jacobs, Catal. Rev. 43 (2001) 443–488.
14] F. Cavani, F. Trifiro, A. Vaccari, Catal. Today 11 (1991) 173–301.
15] A. Corma, V. Forn e´ s, F. Rey, J. Catal. 148 (1994) 205–212.
16] H.C. Greenwell, P.J. Holliman, W. Jones, B.V. Velasco, Catal. Today 114 (2006) 397–
402.
17] K. Tanade, W.F. H o¨ lderich, Appl. Catal. A: Gen. 181 (1999) 399–434.
18] R.J. Chiment a˜ o, S. Abell o´ , F. Medina, J. Llorca, J.E. Sueiras, Y. Cesteros, P. Salagre, J.
Catal. 252 (2007) 249–257.
2
O
4
[
[
[
x
/
/
[
[
Mg(Al)O-723), which associated with the increasing amounts of
Mg–F pairs. While the fluorine contents increased from 6.4 to
1
1.8%, the TOF value of moderate basic sites decreased contrarily.
[19] J.I. Di Cosimo, V.K. D ı´ ez, M. Xu, E. Iglesia, C.R. Apestegu ı´ a, J. Catal. 178 (1998) 499–
10.
5
This might be due to more MgF formation, which blocked
2
[20] H.A. Prescott, Z.J. Li, E. Kemnitz, A. Trunschke, J. Deutsch, H. Lieske, A. Auroux, J.
channels and decreased the amounts of accessible Mg–F pairs.
Therefore, the catalytic activity was also influenced by the relative
proportion of the three pairs among the moderate basic sites.
Catal. 234 (2005) 124–135.
[21] R.J. Davis, E.G. Derouane, Nature 349 (1991) 313–315.
[
22] R. B ˆı rjega, O.D. Pavel, G. Costentin, M. Che, E. Angelescu, Appl. Catal. A: Gen. 288
2005) 185–193.
(
[23] L. Dussault, J.C. Dupin, H. Martinez, E. Dumitriu, A. Auroux, C. Guimon, Surf.
Interface Anal. 38 (2006) 234–237.
4
. Conclusions
[
[
[
24] T. Ando, J. Yamawaki, Chem. Lett. (1979) 45–46.
25] J.H. Clark, Chem. Rev. 80 (1980) 429–452.
26] H. Kabashima, H. Tsuji, S. Nakata, Y. Tanaka, H. Hattori, Appl. Catal. A 194/195
(2000) 227–240.
The fluorine-modified Mg–Al mixed oxides could be prepared
by thermal decomposition of the fluorine-containing Mg–Al
hydrotalcite. During thermal decomposition, six kinds of basic
centers: OH groups (weak basic sites), Mg–O, Mg–F and Al–O
[27] L. Gao, B. Xu, G. Xiao, J. Lv, Energy Fuels 22 (2008) 3531–3535.
[28] B.M. Choudary, M.L. Kantam, V. Neeraja, K.K. Rao, F. Figueras, L. Delmotte, Green
Chem. 3 (2001) 257–260.
À
À
2À
pairs (moderate basic sites), coordinatively unsaturated F and O
ions (strong basic sites) were all present in the resulting samples of
/Mg(Al)O-T. Moreover, the proportion of these basic sites was
[
29] H.A. Prescott, Z. Li, E. Kemnitz, J. Deutsch, H. Lieske, J. Mater. Chem. 15 (2005)
4616–4628.
[
[
30] S. Wuttke, G. Scholz, S. R u¨ diger, E. Kemnitz, J. Mater. Chem. 17 (2007) 4980–4988.
31] G. Wu, X. Wang, B. Chen, J. Li, N. Zhao, W. Wei, Y. Sun, Appl. Catal. A: Gen. 329
F
x
varied with decomposition temperature, and the basicity of
moderate basic sites could be controlled by the change of fluorine
content. In the synthesis of propylene glycol methyl ether from
methanol and PO, it was the moderate basic sites (Mg–F, Mg–O and
Al–O pairs) that acted as active sites, and the catalytic activities of
the three pairs decreased in the following order: Mg–F > Mg–
O > Al–O. When the reaction was carried out with a molar ratio of
methanol to PO of 5:1 at 393 K for 600 min, the PO conversion
(
2007) 106–111.
[32] D.G. Cantrell, L.J. Gillie, A.F. Lee, K. Wilson, Appl. Catal. A: Gen. 287 (2005) 183–
90.
33] Z.P. Xu, G.S. Stevenson, C.Q. Lu, G.Q. Lu, P.F. Bartlett, P.P. Gray, J. Am. Chem. Soc. 12
2006) 36–37.
[34] T.S. Stanimirova, I. Vergilor, G. Kirov, N. Petrova, J. Mater. Sci. 34 (1999) 4153–
161.
35] F.M. Labajos, V. Rives, P. Malet, M.A. Ulibarri, M.A. Centeno, Inorg. Chem. 35
1996) 1154–1160.
[36] F.M. Labajos, M.D. Sastre-Prieto, R. Trujillano, V. Rives, J. Mater. Chem. 9 (1999)
033–1039.
37] M.J. Holgado, F.M. Labajos, M.J.S. Montero, V. Rives, Mater. Res. Bull. 38 (2003)
879–1891.
1
[
(
4
[
(
1
could reach 94.4% with 86.6% selectivity towards PPM over F6.2
Mg(Al)O-673.
/
[
[
1
38] K.S.W. Sing, D.H. Everett, R.A.W. Haul, L. Moscou, R.A. Pierotti, J. Rouquerol, T.
Sieminiewska, Pure Appl. Chem. 57 (1985) 603–619.
39] G. Leofanti, M. Padovan, G. Tozzola, B. Venturelli, Catal. Today 41 (1998) 207–219.
40] H. Zhou, G.L. Zhuo, X.Z. Jiang, J. Mol. Catal. A: Chem. 248 (2006) 26–31.
Acknowledgements
[
[
The authors acknowledge the financial supports from National
Science Technology Foundation of China (No. 2006BAC2A08) and
the NJIT Scientific Research Foundation for Introduce Talents (KXJ
[41] M. Bellotto, B. Rebours, O. Clause, J. Lyneh, D. Bazin, J. Elkaim, J. Phys. Chem. 100
(1996) 8527–8534.
[
[
42] M.L. Lesiecki, J.W. Nibler, J. Chem. Phys. 64 (1976) 871–884.
43] Inorganics IR Grating Spectra, Sadtler Research Laboratories, Inc., Philadelphia,
PA 19104, 1974, Y799K.
0
8033 and KXJ 08034).
[44] A. Snelson, J. Phys. Chem. 70 (1966) 3208–3217.
References
[45] E.F. Hayes, A.K.Q. Siu, D.W. Kisker, J. Phys. Chem. 70 (1966) 4587–4588.
[46] F. Winter, X. Xia, B.P.C. Hereijgers, J.H. Bitter, A.J. van Dillen, M. Muhler, K.P. de
Jong, J. Phys. Chem. B 110 (2006) 9211–9218.
[
[
[
1] H. Hattori, Chem. Rev. 95 (1995) 537–558.
2] Y. Ono, J. Catal. 216 (2003) 406–415.
3] W. Zhang, H. Wang, Q. Li, Q. Dong, N. Zhao, W. Wei, Y. Sun, Appl. Catal. A: Gen. 294
[47] F. Winter, V. Koot, A.J. van Dillen, J.W. Geus, K.P. de Jong, J. Catal. 236 (2005) 91–
100.
[
[
48] C.V.A. Duck, J.M. Miller, J.H. Clark, J. Mol. Catal. A 62 (1990) 233–242.
49] D.R. Lide (Ed.), CRC Handbook of Chemistry and Physics, Internet Version 2005,
CRC Press, Boca Raton, FL, 2005, , <http://www.hbcpnetbase.com>, p. 52, 55, 56
(2005) 188–196.
[
[
[
4] H. Matsuhashi, Top. Catal. 52 (2009) 828–833.
5] B. Veldurthy, J.-M. Clacens, F. Figueras, J. Catal. 229 (2005) 237–242.
6] M. Verziu, M. Florea, S. Simon, V. Simon, P. Filip, V.I. Parvulescu, C. Hardacre, J.
Catal. 263 (2009) 56–66.
(in Section 9).
[50] H.C. Chitwood, J. Am. Chem. Soc. 68 (1946) 680–683.