Dehydration of glycerin to acrolein over H3PW12O40 heteropolyacid catalyst supported on silica-alumina

Article abstract of DOI:10.1016/j.molcata.2014.10.015

A series of tungstophosphoricacid (H₃PW₁₂O₄₀) catalysts supported on silica-alumina (SA-X (X = 0, 15, 30, 50, 70, 85, and 100)) with different SiO₂ content (X, mol%) were prepared, and they (H₃PW₁₂O₄₀/SA-X) were applied to the gas-phase dehydration of glycerin to acrolein. The effect of support composition on the physicochemical properties and catalytic activities of H₃PW₁₂O₄₀/SA-X catalysts was investigated. Total acidity and Br?nsted acidity of H₃PW₁₂O₄₀/SA-X catalysts determined from pyridine-adsorbed in situ FT-IR spectroscopy measurements showed volcano-shaped trends with respect to SiO₂ content. Among the catalysts, H₃PW₁₂O₄₀/SA-50 exhibited the largest total acidity while H₃PW₁₂O₄₀/SA-85 showed the largest Br?nsted acidity. The results of ²⁷Al NMR spectroscopy analyses of H₃PW₁₂O₄₀/SA-X (X = 15, 30, 50, 70, and 85) catalysts clearly demonstrated that Br?nsted acidity of the catalysts was proportional to the amount of Al^IV species. It was revealed that yield for acrolein increased with increasing Br?nsted acidity rather than total acidity of H₃PW₁₂O₄₀/SA-X catalysts. Among the catalysts tested, H₃PW₁₂O₄₀/SA-85 with the largest Br?nsted acidity showed the highest yield for acrolein. A synergistic effect between H₃PW₁₂O₄₀ and SA-85 for H₃PW₁₂O₄₀/SA-85 catalyst was also observed in the dehydration of glycerin to acrolein. Moreover, the catalytic performance of regenerated H₃PW₁₂O₄₀/SA-85 catalyst was almost recovered compared to that of fresh H₃PW₁₂O₄₀/SA-85 catalyst.

Full text of DOI:10.1016/j.molcata.2014.10.015

Journal of Molecular Catalysis A: Chemical 396 (2015) 282–289
Contents lists available at ScienceDirect
Journal of Molecular Catalysis A: Chemical
journal homepage: www.elsevier.com/locate/molcata
Dehydration of glycerin to acrolein over H PW O heteropolyacid
3
12 40
catalyst supported on silica-alumina
a
a
a
a
a
Tae Hun Kang , Jung Ho Choi , Yongju Bang , Jaekyeong Yoo , Ji Hwan Song ,
b
b
a,∗
Wangrae Joe , Jun Seon Choi , In Kyu Song
School of Chemical and Biological Engineering, Institute of Chemical Processes, Seoul National University, Shinlim-dong, Kwanak-ku,
Seoul 151-744, South Korea
a
b
LG Chem. Ltd., Moonji-dong, Yuseong-Gu, Daejeon 305-380, South Korea
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 7 August 2014
Received in revised form 2 October 2014
Accepted 15 October 2014
Available online 23 October 2014
A series of tungstophosphoricacid (H PW12O40) catalysts supported on silica-alumina (SA-X (X = 0, 15, 30,
3
5
0, 70, 85, and 100)) with different SiO2 content (X, mol%) were prepared, and they (H3PW12O40/SA-X)
were applied to the gas-phase dehydration of glycerin to acrolein. The effect of support composition on the
physicochemical properties and catalytic activities of H3PW12O40/SA-X catalysts was investigated. Total
acidity and Brønsted acidity of H3PW12O40/SA-X catalysts determined from pyridine-adsorbed in situ
FT-IR spectroscopy measurements showed volcano-shaped trends with respect to SiO2 content. Among
the catalysts, H3PW12O40/SA-50 exhibited the largest total acidity while H3PW12O40/SA-85 showed the
Keywords:
Heteropolyacid
Silica-alumina
Dehydration of glycerin
Acrolein
2
7
largest Brønsted acidity. The results of Al NMR spectroscopy analyses of H3PW12O40/SA-X (X = 15, 30,
5
0, 70, and 85) catalysts clearly demonstrated that Brønsted acidity of the catalysts was proportional
IV
to the amount of Al species. It was revealed that yield for acrolein increased with increasing Brønsted
acidity rather than total acidity of H3PW12O40/SA-X catalysts. Among the catalysts tested, H3PW12O40/SA-
Brønsted acidity
8
5 with the largest Brønsted acidity showed the highest yield for acrolein. A synergistic effect between
H3PW12O40 and SA-85 for H3PW12O40/SA-85 catalyst was also observed in the dehydration of glycerin
to acrolein. Moreover, the catalytic performance of regenerated H3PW12O40/SA-85 catalyst was almost
recovered compared to that of fresh H3PW12O40/SA-85 catalyst.
©
2014 Elsevier B.V. All rights reserved.
1
. Introduction
from catalyst deactivation and low catalytic activity due to coke
formation and considerable activation energy, respectively.
Heteropolyacids (HPAs) are early transition metal-oxygen anion
clusters that possess much stronger Brønsted acid than other acids
such as zeolites, metal oxides, and concentrated sulfuric acid [8,9].
HPAs have been widely employed as catalysts in various acid-
catalyzed reactions such as dehydration of glycerin [10], hydration
of alkenes [11], and dehydration of alcohols [12]. However, low
Acrolein has attracted much attention due to its applicability as a
chemical intermediate for the synthesis of acrylic acid, methionine,
and super-absorber polymers [1]. Acrolein has been convention-
ally produced by the partial oxidation process of petroleum-based
propylene over multi-component metal oxide catalysts [2]. How-
ever, gas-phase dehydration of bio-based glycerin over acid catalyst
has garnered recent attraction as a promising alternative route for
acrolein production due to several economical and environmental
benefits, with increasing concerns about fossil fuel depletion and
global warming problem [2–4]. Several solid acid catalysts such as
zeolites [5] and phosphates [6] have been investigated as catalysts
for gas-phase dehydration of glycerin. However, a previous work
2
surface area (<10 m /g) of HPAs is a main drawback to main-
tain the catalytic performance during a long-time reaction under
harsh environments. To overcome this problem, HPAs supported on
porous materials with high surface area such as silica, alumina, zir-
conia, and carbon were investigated as a catalyst for acid-catalyzed
reactions. Recently, it has been reported that tungstophosphori-
[
7] revealed that acid-catalyzed dehydration of glycerin suffered
cacid (H PW12O40) catalyst supported on alumina [13] or zirconia
3
[
14] showed an excellent catalytic activity in the dehydration of
glycerin.
In this work, silica-alumina was employed as a benign support
∗ Corresponding author. Tel.: +82 2 880 9227; fax: +82 2 889 7415.
E-mail address: inksong@snu.ac.kr (I.K. Song).
for H3PW12O40 in the dehydration of glycerin due to its outstand-
ing physicochemical properties such as high surface area, large
http://dx.doi.org/10.1016/j.molcata.2014.10.015
381-1169/© 2014 Elsevier B.V. All rights reserved.
1

T.H. Kang et al. / Journal of Molecular Catalysis A: Chemical 396 (2015) 282–289
283
◦
pore volume, and excellent acid property [15–17]. Gas-phase dehy-
to 50 C, pyridine vapor (20 ml) was pulsed into the reactor with a
stream of He (40 ml/min) until the acid sites were saturated with
pyridine. The reactor was then evacuated at 50 C for 1 h. Brønsted
dration of glycerin to acrolein was carried out over H PW12O40
catalysts supported on silica-alumina (SA-X (X = 0, 15, 30, 50, 70,
3
◦
8
5, and 100)) with different SiO2 content (X, mol%). The prepared
acidity and Lewis acidity were determined from IR spectra of
31
◦
catalysts were characterized by ICP-AES, FT-IR, XRD, P MAS NMR,
and nitrogen adsorption–desorption analyses. Pyridine-adsorbed
in situ FT-IR spectroscopy analyses were conducted to examine
pyridine-adsorbed catalysts at 50 C within spectral range of
−
1
1600–1400 cm
.
the acid properties of H PW12O40/SA-X catalysts. 27Al NMR mea-
3
2.3. Gas-phase dehydration of glycerin
surements were also conducted to elucidate the acid properties of
H PW12O40/SA-X (X = 15, 30, 50, 70, and 85) catalysts. The effect
Gas-phase dehydration of glycerin was carried out in a con-
tinuous flow fixed-bed reactor under atmospheric pressure. Each
catalyst (0.1 g) was charged into a tubular pyrex reactor. The reac-
tor was preheated with a stream of nitrogen carrier gas (30 ml/min)
3
of support composition on the catalytic performance in the dehy-
dration of glycerin was investigated. A correlation between acid
property and yield for acrolein was then established and discussed.
◦
at 275 C for 0.5 h to achieve steady-state operation. Glycerin feed
(
7 mol% aqueous solution, 0.042 mol/h) was sufficiently vaporized
2
. Experimental
by passing through a pre-heating zone, and it was continuously fed
into the reactor together with a stream of nitrogen (30 ml/min). The
reaction was carried out at 275 C for 10 h. The reaction products
2.1. Preparation of catalysts
◦
were analyzed using a gas chromatograph equipped with a flame
ionization detector (Younglin, YL6100 GC) and a capillary column
A series of silica-alumina supports with different SiO2 content
were prepared by a sol–gel method [10]. Tetraethyl orthosilicate
TEOS, Sigma–Aldrich) was dissolved in distilled water, and pH of
(
Agilent, DB-5MS, 60 m × 0.32 mm). 1,4-Dioxane was used as an
(
internal standard for quantitative calculation. Conversion of glyc-
erin, selectivity for product, and yield for product were calculated
according to the following equations.
the solution was adjusted to 2.0 with nitric acid. The TEOS solution
◦
was vigorously stirred at 40 C for 1 h. Aluminum nitrate nonahy-
drates (Al(NO ) ·9H O, Junsei Chem.) was separately dissolved in
3
3
2
distilled water, and the solution was then added into the TEOS solu-
Conversion of glycerin (%)
◦
tion. The mixed solution was stirred at 40 C for 1 h. Ammonium
mole of glycerin inlet − mole of glycerin outlet
hydroxide solution (NH OH, Sigma–Aldrich) was added into the
4
=
× 100
× 100
(1)
mole of glycerin inlet
mixed solution at a constant rate to form a hydrogel at pH 8.5.
◦
The hydrogel was further aged at 40 C for 1 h, and subsequently,
◦
it was filtered, washed, and dried at 100 C for 10 h. The result-
ing white solid was finally calcined at 550 C for 3 h. The prepared
Selectivity for product (%)
mole of product outlet
◦
silica-alumina supports were denoted as SA-X (X = 15, 30, 50, 70,
=
(2)
(3)
and 85), where X represented SiO content (mol%) in silica-alumina.
mole of glycerin inlet − mole of glycerin outlet
2
Pure alumina (SA-0) and silica (SA-100) supports were also pre-
pared by the similar methods described above. H PW12O40/SA-X
3
mole of product outlet
catalysts were prepared by an impregnation method using an aque-
Yield for product (%) =
× 100
mole of glycerin inlet
. Results and discussion
.1. Characterization of catalysts
Fig. 1 shows the FT-IR spectra of H PW12O40/SA-X (X = 0,
ous solution of tungstophosphoricacid (H PW12O40). Content of
3
H PW12O40 in all the supported catalysts was fixed at 10.0 wt%. The
3
3
◦
impregnated solid catalysts were dried at 80 C for 10 h, and then
◦
they were calcined at 300 C for 3 h. The prepared catalysts were
3
denoted as H PW12O40/SA-X (X = 0, 15, 30, 50, 70, 85, and 100).
3
3
2.2. Characterization of catalysts
15, 30, 50, 70, 85, and 100) catalysts. H3PW12O40/SA-0 exhib-
−
1
ited broad bands within the range of 1250–700 cm
, cor-
Infrared spectra of H PW12O40/SA-X catalysts were obtained
responding to the disordered metastable spinel structure of
3
using a Fourier transform infrared (FT-IR) spectrometer (Thermo
Scientific, Nicolet 6700) with a diffuse reflectance accessory. Chem-
ical compositions of the catalysts were determined by inductively
coupled plasma-atomic emission spectrometry (ICP-AES) analyses
(
Shimadzu, ICP-1000IV). Nitrogen adsorption–desorption experi-
ments (BEL Japan, BELSORP-mini II) were conducted to examine the
textural properties, and pore size distributions were determined by
BJH (Barret–Joyner–Hallender) method. X-ray diffractograms were
obtained (Rigaku, D-MAX2500-PC) using Cu-K␣ radiation operated
at 50 kV and 100 mA. Structural state of H PW12O40 was confirmed
3
31
by P magic-angle spinning nuclear magnetic resonance (MAS
NMR) spectroscopy analyses (Bruker, AVANCE 400 WB). Aluminum
coordination state was examined by a 400-MHz solid state 27Al
nuclear magnetic resonance (NMR) spectrometer (Bruker, AVANCE
4
00 WB) at frequencies of 104.3 MHz.
Acid properties of H PW12O40/SA-X catalysts were determined
3
by pyridine-adsorbed in situ FT-IR spectroscopy measurements.
Each catalyst (0.03 g) was pressed into a pellet with a radius of
◦
0
.65 cm. The pellet was preheated at 300 C for 1 h with a stream of
He (40 ml/min) in a tubular quartz reactor. After cooling the pellet
Fig. 1. FT-IR spectra of H3PW12O40/SA-X (X = 0, 15, 30, 50, 70, 85, and 100) catalysts.

2
84
T.H. Kang et al. / Journal of Molecular Catalysis A: Chemical 396 (2015) 282–289
Fig. 3. 31P MAS NMR spectra of H3PW12O40/SA-X (X = 0, 15, 50, 85, and 100) catalysts.
Fig. 2. XRD patterns of H3PW12O40/SA-X (X = 0, 15, 30, 50, 70, 85, and 100) catalysts.
increasing SiO2 content, in good agreement with the FT-IR results.
␥
-Al O [18]. H PW12O40/SA-100 showed broad bands at around
However, H PW12O40/SA-X catalysts did not show the character-
2
3
3
3
−
1
−1
1
250–1000 cm
(asymmetric stretching Si
O
Si), 800 cm
istic XRD peaks for H PW12O40. The above results indicated that
3
(
symmetric stretching Si
O
Si), and a weak shoulder band at
H PW12O4O was finely dispersed on the supports [20].
3
−1
31
around 974 cm (surface OH group), which were attributed to SiO2
framework [19,20]. FT-IR spectra of H PW12O40/SA-X (X = 15, 30,
5
P MAS NMR measurements were carried out to confirm the
structural state of H PW12O40. Fig. 3 shows the 31P MAS NMR spec-
3
3
31
0, 70, and 85) catalysts exhibited characteristics of both ␥-Al O
tra of H PW12O40/SA-X (X = 0, 15, 50, 85, and 100) catalysts.
P
2
3
1
3
−
and SiO . The intensity of IR bands at around 1250–1000 cm
MAS NMR spectra showed three peaks due to different structural
2
corresponding to asymmetric stretching vibrations of Si
O
Si
state of H PW12O40; hydrated H PW12O40 species (−15.5 ppm),
3
3
bonding increased with increasing SiO2 content. In addition, IR
bands due to asymmetric stretching vibrations of Si Si bonding
shifted to higher frequencies with increasing SiO2 content, which
were attributed to an increased cross-linking of Si Si network
H PW12O40 species interacting with surface acid sites of alumina
3
7
−
O
(−13.9 ppm), and lacunary species ([PW11O39
]
) or another frag-
mented forms of heteropolyanion (−11.6 ppm) [25]. H PW12O40
3
O
supported on SiO -rich silica-alumina supports exhibited a peak
2
[
21,22]. It has been reported that H PW12O40 exhibited four charac-
at −15.5 ppm, indicating that the primary structure of H PW12O40
3
3
−
1
−1
−1
teristic IR bands at 1080 cm (P O), 983 cm (W O), 890 cm
(
was retained as a hydrated form. Although H PW
O
12 40
supported
3
−1
W
Oc W), and 812 cm
(W Oe W) [23]. However, the char-
on Al O -rich silica-alumina supports showed partially fragment
2 3
acteristic IR bands corresponding to H PW12O40 were not clearly
species at −11.6 ppm due to strong interaction with alumina sur-
face, the primary structure of H PW was still maintained as
3
observed in the H PW12O40/SA-X catalysts, because the character-
O
12 40
3
3
istic IR bands of H PW12O40 were overlapped with large and broad
IR bands of supports [20].
an interacted form with silica-alumina surface. Therefore, it can
be said that no characteristic diffraction peaks (Fig. 2) for any het-
3
Fig. 2 shows the XRD patterns of H PW12O40/SA-X (X = 0, 15,
eropolyacid species indicate high dispersion of H PW12O40 on the
3
3
3
0, 50, 70, 85, and 100) catalysts. H PW12O40/SA-0 showed XRD
silica-alumina supports.
Fig. 4 shows the nitrogen adsorption–desorption isotherms
and pore size distributions of H PW12O40/SA-X (X = 0, 15, 50,
3
◦
◦
peaks at 2ꢀ = 45 and 67 indicative of ␥-Al O3 structure [24].
H PW12O40/SA-100 showed a broad peak at 2ꢀ = 22 corresponding
2
◦
3
3
to amorphous SiO2 framework. XRD patterns of H PW12O40/SA-X
85, and 100) catalysts. H PW12O40/SA-X catalysts exhibited
3
3
(
X = 15, 30, 50, 70, and 85) catalysts showed both ␥-Al O and amor-
type-IV isotherm with H2 or H3 hysteresis loop. Nitrogen
adsorption–desorption isotherms and pore size distributions of
2
3
phous SiO diffraction characteristics with weak intensity. Intensity
2
of diffraction peak attributed to SiO2 framework increased with
H PW12O40/SA-X catalysts were similar to those of supports (not
3
Fig. 4. (a) Nitrogen adsorption–desorption isotherms and (b) pore size distributions of H3PW12O40/SA-X (X = 0, 15, 50, 85, and 100) catalysts.

T.H. Kang et al. / Journal of Molecular Catalysis A: Chemical 396 (2015) 282–289
285
Table 1
SiO2 content, surface area, pore volume, and average pore diameter of H3PW12O40/SA-X catalysts.
SiO2 content (mol%)a
Surface area (m /g)
2
b
Pore volume (cm /g)c
3
Average pore diameter (nm)d
Catalyst
H3PW12O40/SA-0
H3PW12O40/SA-15
H3PW12O40/SA-30
H3PW12O40/SA-50
H3PW12O40/SA-70
H3PW12O40/SA-85
H3PW12O40/SA-100
0
14
29
52
71
189.2
252.0
263.6
284.8
221.6
284.1
354.8
0.33
0.44
0.41
0.41
0.34
0.37
0.62
8.2
8.2
7.2
7.2
7.2
6.3
8.2
85
100
a
Determined by ICP-AES measurement.
Calculated by BET equation.
BJH desorption pore volume.
b
c
d
BJH desorption average pore diameter.
All the catalysts exhibited IR bands corresponding to Brønsted
−
1
−1
acid sites (1545 cm ) and Lewis acid sites (1450 cm ). IR spec-
tra at 1490 cm
−
1
corresponding to both Brønsted acid sites and
Lewis acid sites were overlapped. Brønsted acidity and Lewis
acidity of H PW12O40/SA-X catalysts were quantitatively mea-
3
sured according to the method reported in the literature [26]
as summarized in Table 2. Total acidity was calculated as the
sum of Brønsted acidity and Lewis acidity. Total acidity of the
catalysts exhibited a volcano-shaped trend with respect to SiO2
content. Among the catalysts tested, H PW12O40/SA-50 showed
3
the largest total acidity. Brønsted acidity of the catalysts also
showed a volcano-shaped trend with respect to SiO2 content. In
this case, however, H PW12O40/SA-85 showed the largest Brønsted
3
acidity. H PW12O40/SA-0 retained the smallest Brønsted acidity
3
and the largest Lewis acidity. In addition, H PW12O40/SA-100
3
showed very weak IR bands compared to the other catalysts
(
(
Fig. 5), resulting in quite small total and Brønsted acidities
Table 2).
Fig. 5. Pyridine-adsorbed in situ FT-IR spectra of H3PW12O40/SA-X (X = 0, 15, 30, 50,
0, 85, and 100) catalysts.
7
3
.3. Aluminum coordination state in the catalysts
shown here), indicating that the mesoporous structure of supports
was maintained even after the loading of H PW12O on the sup-
ports. Surface area, pore volume, and average pore diameter of
27Al NMR measurements were conducted to elucidate the dif-
3
40
ference in acid properties of H PW12O40/SA-X (X = 15, 30, 50,
70, and 85) catalysts. Fig. 6 shows the
H PW12O40/SA-X (X = 15, 30, 50, 70, and 85) catalysts. All the Al
NMR spectra exhibited three peaks because of different aluminum
coordination states; tetrahedral-coordinated aluminum species
3
H PW12O40/SA-X catalysts are summarized in Table 1. It was found
27
3
Al NMR spectra of
that surface area of H PW12O40/SA-X catalysts increased with
27
3
3
increasing SiO2 content. However, pore volume and average pore
diameter of H PW12O40/SA-X catalysts did not show any notice-
3
able trends with respect to SiO2 content. It was also observed that
the measured SiO2 content was well consistent with the designed
value.
IV
(Al , 50–65 ppm), pentahedral-coordinated aluminum species
(Al , 30–35 ppm), and octahedral-coordinated aluminum species
V
VI
(Al , 0–7 ppm). The amount of aluminum species determined from
27
peak area of Al NMR spectra is summarized in Table 2. The amount
of Al species decreased with increasing SiO2 content due to the
VI
3.2. Acid properties of catalysts
increase of agglomeration of isolated Al framework during the for-
IV
Pyridine-adsorbed in situ FT-IR spectroscopy measurements
mation of Al
O
Al bonds. However, the amount of Al species
were conducted to investigate the acid properties of the cata-
lysts. Fig. 5 shows the pyridine-adsorbed in situ FT-IR spectra
increased with increasing SiO content because of the substitution
2
of Al3+ into Si sites at tetrahedral SiO2 framework structure [27].
4+
V
of H PW12O40/SA-X (X = 0, 15, 30, 50, 70, 85, and 100) catalysts.
The amount of Al species, which located at the interface between
3
Table 2
Brønsted acidity, Lewis acidity, total acidity, and amount of aluminum species of H3PW12O40/SA-X catalysts.
Catalysts
Brønsted acidity
mmol-pyridine/g-cat.)
Lewis acidity
(mmol-pyridine/g-cat.)
Total acidity
(mmol-pyridine/g-cat.)
Amount of aluminum
species (%)
a
(
AlIV
AlV
–
AlVI
H3PW12O40/SA-0
H3PW12O40/SA-15
H3PW12O40/SA-30
H3PW12O40/SA-50
H3PW12O40/SA-70
H3PW12O40/SA-85
H3PW12O40/SA-100
0.05
0.16
0.31
0.44
0.60
0.77
0.10
0.70
0.62
0.68
0.59
0.42
0.16
0.09
0.75
0.78
0.99
1.03
1.02
0.93
0.19
–
–
15.9
17.7
19.4
37.8
48.6
–
4.4
79.7
73.1
71.0
45.7
46.4
–
9.2
9.5
16.5
5.0
–
a
Determined by 27Al NMR spectroscopy.

2
86
T.H. Kang et al. / Journal of Molecular Catalysis A: Chemical 396 (2015) 282–289
Fig. 6. 27Al NMR spectra of H3PW12O40/SA-X (X = 15, 30, 50, 70, and 85) catalysts.
Fig. 8. A correlation between yield for acrolein after a 10 h-reaction over
H3PW12O40/SA-X (X = 0, 15, 30, 50, 70, 85, and 100) catalysts and Brønsted acidity
of the catalysts.
mixed silica-alumina and aluminum oxide phase [28], did not show
a noticeable trend with respect to SiO2 content.
It is interesting to note that Brønsted acidity of the catalysts
showed the same trend with the amount of Al species (Table 2).
It has been reported that Brønsted acid sites of silica-alumina are
attributed to tetrahedral-coordinated Al cations in Si–O–Si network
are summarized in Table 3. According to the literature [5], acrolein
was mainly produced during the reaction, while hydroxyacetone
and allyl alcohol were produced as by-products. It is known that
dehydration of glycerin follows elimination mechanism, and the
acid catalyst initiates the formation of primary and secondary car-
bocations [2]. Secondary carbocation is more stable than primary
carbocation due to the thermodynamic effect. Therefore, acrolein
is mainly produced through two-step dehydration via secondary
carbocation. Hydroxyacetone produced from primary carbocation
is the major by-product, while other compounds such as allyl alco-
hol, 1,2-propanediol, acetaldehyde, and acetic acid are the possible
minor by-products in the dehydration of glycerin [1,7].
IV
[
27,28]. Therefore, Brønsted acidity of the catalysts increased with
IV
increasing SiO2 content and with increasing the amount of Al
species. Among the catalysts, H PW12O40/SA-85 with the largest
3
IV
amount of Al species showed the largest Brønsted acidity. It
is believed that H PW12O40/SA-0 showed the smallest Brønsted
3
acidity and the largest Lewis acidity, because ␥-Al O3 retained
2
only Lewis acid sites [29]. Moreover, it can be inferred that neu-
tral silica support with partial surface silanol group [30] might be
responsible for small total and Brønsted acidities of H PW12O40/SA-
It was revealed that yield for acrolein increased in the
order of H PW12O40/SA-0 < H PW12O40/SA-100 < H PW12O40/SA-
3
1
00. Consequently, Brønsted acidity of the catalysts exhibited a
volcano-shaped trend with respect to SiO2 content (Table 2), and
H PW12O40/SA-85 showed the largest Brønsted acidity.
3
3
3
15 < H PW12O40/SA-30 < H PW12O40/SA-50 < H PW12O40/SA-
3
3
3
3
70 < H PW12O40/SA-85. H PW12O40/SA-0 and H PW12O40/SA-100
3
3
3
exhibited lower yield for acrolein than H PW12O40/SA-X (X = 15,
3
3.4. Catalytic performance in the gas-phase dehydration of
30, 50, 70, and 85) catalysts. The trend of yield for acrolein of
glycerin
H PW12O40/SA-X (X = 0, 15, 30, 50, 70, 85, and 100) was not
3
directly correlated with the trend of total acidity of the catalysts.
Gas-phase dehydration of glycerin to acrolein was carried out
However, the trend of yield for acrolein of H PW12O40/SA-X
3
over H PW12O40/SA-X (X = 0, 15, 30, 50, 70, 85, and 100) cata-
(X = 0, 15, 30, 50, 70, 85, and 100) was well consistent with
the trend of Brønsted acidity of the catalysts. Fig. 8 shows the
correlation between yield for acrolein after a 10 h-reaction over
3
lysts. Fig. 7 shows the catalytic performance of H PW12O40/SA-X
3
◦
catalysts with time on stream at 275 C. Conversion of glycerin,
selectivity for product, and yield for acrolein after a 10 h-reaction
H PW12O40/SA-X (X = 0, 15, 30, 50, 70, 85, and 100) catalysts
3
Fig. 7. (a) Conversion of glycerin and (b) yield for acrolein over H3PW12O40/SA-X (X = 0, 15, 30, 50, 70, 85, and 100) catalysts in the dehydration of glycerin to acrolein with
◦
time on stream at 275 C.

T.H. Kang et al. / Journal of Molecular Catalysis A: Chemical 396 (2015) 282–289
287
Table 3
Conversion of glycerin, selectivity for product, yield for acrolein, carbon balance, and amount of carbon deposition over H3PW12O40/SA-X catalysts after a 10 h-reaction.
Catalysts
Conversion of
glycerin (%)
Selectivity for product (%)
Yield for
acrolein (%)
Carbon
balance (%)
Amount of carbon
deposition (wt%)
Acrolein Hydroxyacetone Allyl
Acetaldehyde Acetic
acid
Othersa
alcohol
H3PW12O40/SA-0
H3PW12O40/SA-15
H3PW12O40/SA-30
H3PW12O40/SA-50
H3PW12O40/SA-70
H3PW12O40/SA-85
13.2
15.7
18.8
22.9
27.8
42.8
23.3
27.3
30.3
34.3
40.4
46.4
25.7
21.9
19.2
14.6
14.5
11.5
8.5
1.6
1.5
1.6
2.3
2.1
1.9
2.2
0.9
0.5
0.3
0.6
0.7
0.6
0.8
0.2
25.8
24.8
28.9
30.6
31.7
28.5
23.2
3.1
4.6
5.7
74.6
75.7
79.5
85.7
90.9
89.3
64.3
4.3
6.4
6.6
6.8
7.8
9.8
4.5
2.5
2.8
3.5
4.8
2.9
1.1
7.9
11.2
19.9
4.2
H3PW12O40/SA-100 16.4
13.2
a
Others include acetone, acrylic acid, and several unidentified products.
◦
and Brønsted acidity of the catalysts. It should be noted that
yield for acrolein increased with increasing Brønsted acidity of
the catalysts. This result was well consistent with the result of
previous works [31–33] in a sense that Brønsted acid sites strongly
affected the acrolein formation in the dehydration of glycerin. In
case of Brønsted acid sites, migration of proton from the catalyst
to internal oxygen of glycerin is facile due to the absence of
steric effect. Therefore, protonation of internal oxygen leads to
the formation of acrolein via secondary carbocations. However,
interaction of glycerin with Lewis acid sites can be affected by
steric constraints due to the electronic structure of Lewis acid sites,
and terminal oxygen of glycerin reacts with catalyst more easily
than internal oxygen of glycerin. As a result, hydroxyacetone is
mainly produced on the Lewis acid sites via primary carbocations.
stream at 275 C. It is noteworthy that conversion of glycerin and
yield for acrolein over H PW12O40/SA-85 were much higher than
3
those over H PW12O40 and SA-85. This means that H PW12O40
3
3
and SA-85 synergistically affected the catalytic performance of
H PW12O40/SA-85. Because of the complexity of the reaction mech-
3
anism of glycerin dehydration, deconvolution of the synergistic
effect was not a simple task. However, it is inferred that silica-
alumina with moderate Brønsted acidity and high surface area not
only provided additional active sites for dehydration of glycerin
but also led to high dispersion of HPA, resulting in outstanding cat-
alytic performance of H PW12O40/SA-85. It was also revealed that
Brønsted acidity of H PW12O40 (0.62 mmol-pyridine/g-cat.) was
larger than that of SA-85 (0.53 mmol-pyridine/g-cat.) (not shown
here). Although initial activity of SA-85 was higher than that of
3
3
Among the catalysts tested, H PW12O40/SA-85 with the largest
H PW12O40, yield for acrolein after a 10 h-reaction over SA-85
3
3
Brønsted acidity showed the highest yield for acrolein. Therefore,
it is concluded that Brønsted acidity of the catalysts is an important
factor determining the catalytic performance in the dehydration
of glycerin to acrolein.
(10.1%) was lower than that of H PW12O40 (12.6%) (Fig. 9), in good
3
agreement with the main conclusion of this work in a sense that
large Brønsted acidity of the catalyst was more favorable for dehy-
dration of glycerin to acrolein (Fig. 8).
Catalytic performance of H PW12O40/SA-X catalysts gradually
3
decreased with time on stream during the 10 h-reaction. The spent
3.5. Regeneration test of H PW12O40/SA-85
3
H PW12O40/SA-X catalysts were collected after the 10 h-reaction
3
for CHNS analyses. The amount of carbon deposition in the spent
catalysts determined by CHNS analyses is summarized in Table 3.
The result showed that severe coke formed during the reaction
blocked the active sites of the catalysts, resulting in catalyst deac-
tivation. Moreover, carbon balance was less than 100% due to the
carbon deposition derived from oligomerization of glycerin [34],
polymerization of acrolein [35], and formation of undetectable
Regeneration of deactivated catalyst is an important issue in the
catalytic dehydration of glycerin [1]. Therefore, the regeneration
test for H PW12O40/SA-85 was conducted in order to investi-
3
gate the stability and reusability of the catalyst. Fig. 10 shows
the temperature-programmed oxidation (TPO) profile of spent
H PW12O40/SA-85 catalyst used for 10 h-reaction. During the TPO
3
measurement, the desorbed CO2 was monitored using a GC-MSD
cyclic C -compounds [36].
(Agilent, MSD-6890N GC). The TPO profile of spent H PW12O40/SA-
85 catalyst showed two peaks at 429 C and 536 C, indicating that
two kinds of carbonaceous materials (soft coke and hard coke) were
6
3
◦
◦
Fig. 9 shows the conversion of glycerin and yield for acrolein
over H PW12O40/SA-85, H PW12O40, and SA-85 with time on
3
3
Fig. 9. (a) Conversion of glycerin and (b) yield for acrolein over H3PW12O40/SA-85, SA-85, and H3PW12O40 catalysts in the dehydration of glycerin to acrolein with time on
◦
stream at 275 C.

2
88
T.H. Kang et al. / Journal of Molecular Catalysis A: Chemical 396 (2015) 282–289
Fig. 12. 31P MAS NMR spectra of fresh, spent, and regenerated H3PW12O40/SA-85
catalysts.
Fig. 10. TPO profile of spent H3PW12O40/SA-85 catalyst used for 10 h-reaction.
form of H PW12O40. This result clearly indicates that the primary
3
structure of H PW12O40 in the H PW12O40/SA-85 catalyst was
3
3
maintained even after the catalytic reaction and regeneration pro-
cess. Therefore, it is concluded that H PW12O40/SA-85 served as a
3
stable, reusable, and highly active catalyst for the dehydration of
glycerin.
4. Conclusions
Gas-phase dehydration of glycerin to acrolein was carried out
over H PW12O40 catalysts supported on silica-alumina supports
3
(SA-X) with different SiO2 content (X, mol%). Total acidity and
Brønsted acidity of H PW12O40/SA-X catalysts determined from
3
pyridine-adsorbed in situ FT-IR spectroscopy showed volcano-
shaped trends with respect to SiO2 content. Among the catalysts,
H PW12O40/SA-50 and H PW12O40/SA-85 exhibited the largest
3
3
27
Fig. 11. Yield for acrolein over fresh H3PW12O40/SA-85 and regenerated
total acidity and Brønsted acidity, respectively. Al NMR spec-
H3PW12O40/SA-85 catalysts in the dehydration of glycerin to acrolein with time
on stream at 275 C.
troscopy analyses of H PW12O40/SA-X (X = 15, 30, 50, 70, and 85)
3
◦
catalysts revealed that Brønsted acidity of the catalysts increased
IV
with increasing the amount of Al species. It was found that the cat-
formed during the 10 h-reaction. Based on this result, the regen-
eration temperature was chosen to be 500 C. The regeneration of
alytic performance of H3PW12O40/SA-X (X = 0, 15, 30, 50, 70, 85, and
100) in the dehydration of glycerin was closely related to Brønsted
acidity of the catalysts rather than total acidity of the catalysts. The
catalytic performance of H3PW12O40/SA-X (X = 0, 15, 30, 50, 70, 85,
and 100) catalysts increased with increasing Brønsted acidity of
the catalysts. Among the catalysts tested, H3PW12O40/SA-85 with
the largest Brønsted acidity showed the highest yield for acrolein.
Thus, Brønsted acidity of the catalysts served as an important role
in the catalytic dehydration of glycerin to acrolein. A synergistic
effect between H3PW12O40 and SA-85 for H3PW12O40/SA-85 was
also observed in the dehydration of glycerin. Thus, H3PW12O40/SA-
85 served as a stable, reusable, and highly active catalyst for the
dehydration of glycerin.
◦
spent H PW12O40/SA-85 was done with a mixed stream of nitrogen
3
◦
(
30 ml/min) and oxygen (10 ml/min) at 500 C for 10 h.
Fig. 11 shows the yield for acrolein over fresh H PW12O40/SA-
3
8
5 and regenerated H PW12O40/SA-85 catalysts in the dehydration
3
◦
of glycerin with time on stream at 275 C. It should be noted
that the catalytic performance of regenerated H PW12O40/SA-85
3
catalyst was almost the same as that of fresh H PW12O40/SA-
3
8
5 catalyst during the 10 h-reaction. Furthermore, the amount
of carbon species on the regenerated H PW12O40/SA-85 catalyst
3
determined by CHNS analysis was less than 0.1%. This indicates
that H PW12O40/SA-85 catalyst was completely regenerated under
3
our regeneration conditions and that H PW12O40/SA-85 catalyst
3
showed excellent reusability in the dehydration of glycerin to
acrolein.
Acknowledgements
This work was supported by LG Chemical Corporation. This work
was also supported by Mid-career Researcher Program of National
Research Foundation of Korea (NRF) grant funded by the Korea
government (MEST) (No. 2012-R1A2A4A01001146).
3
.6. Structural stability of regenerated catalyst
31P MAS NMR measurements were carried out to investigate the
structural state of fresh, spent, and regenerated H PW12O40/SA-85
3
catalysts. Fig. 12 shows the 31P MAS NMR spectra of fresh, spent,
References
and regenerated H PW12O40/SA-85 catalysts. 31P MAS NMR analy-
3
ses revealed that fresh, spent, and regenerated H PW12O40/SA-85
catalysts exhibited a single peak at −15.5 ppm due to the hydrated
[1] B. Katryniok, S. Paul, M. Capron, F. Dumeignil, ChemSusChem 2 (2009) 719–730.
[2] E.H. Xie, Q.L. Zhang, K.T. Chuang, Appl. Catal. A: Gen. 220 (2001) 215–221.
3

T.H. Kang et al. / Journal of Molecular Catalysis A: Chemical 396 (2015) 282–289
289
[
3] Y.-C. Bak, J.-H. Choi, S.-B. Kim, D.-W. Kang, Korean J. Chem. Eng. 13 (1996)
42–245.
4] Z.-M. Wang, J.-S. Lee, J.-Y. Park, C.-Z. Wu, Z.-H. Yuan, Korean J. Chem. Eng. 24
2007) 1027–1030.
[21] M.V. Landau, S.P. Varkey, M. Herskowitz, O. Regev, S. Pevzner, T. Sen, Z. Luz,
Microporous Mesoporous Mater. 33 (1999) 149–163.
[22] E.I. Kamitsos, A.P. Patsis, Phys. Rev. B 48 (1993) 12499–12505.
[23] C. Rocchiccioli-deltcheff, M. Fournier, R. Franck, R. Thouvenot, Inorg. Chem. 22
(1983) 207–216.
[24] A. Boumaza, L. Favaro, J. Lédion, G. Sattonnay, J.B. Brubach, P. Berthet, A.M.
Huntz, P. Roy, R. Tétot, J. Solid State Chem. 182 (2009) 1171–1176.
[25] E. Caliman, J.A. Dias, S.C.L. Dias, A.G.S. Prado, Catal. Today 107–108 (2005)
816–825.
2
[
(
[
[
[
[
[
5] C.-J. Jia, Y. Liu, W. Schmidt, A.-H. Lu, F. Sch u˝ th, J. Catal. 269 (2010) 71–79.
6] A. Ulgen, W. Hoelderich, Catal. Lett. 131 (2009) 122–128.
7] L. Liu, X.P. Ye, J.J. Bozell, ChemSusChem 5 (2012) 1162–1180.
8] I.V. Kozhevnikov, Chem. Rev. 98 (1998) 171–198.
9] T. Okuhara, N. Mizuno, M. Misono, Adv. Catal. 41 (1996) 113–252.
[
10] L. Shen, H. Yin, A. Wang, Y. Feng, Y. Shen, Z. Wu, T. Jiang, Chem. Eng. J. 180
2012) 277–283.
11] A.V. Inanov, E. Zausa, Y. Ben Taârit, N. Essayem, Appl. Catal. A: Gen. 256 (2003)
25–242.
[26] C.A. Emeis, J. Catal. 141 (1993) 347–354.
[27] R. Takahashi, S. Sato, T. Sodesawa, M. Shizukuishi, K. Morofuji, K. Ogura, J. Non-
Cryst. Solids 351 (2005) 826–832.
[28] E.J.M. Hensen, D.G. Poduval, P.C.M.M. Magusin, A.E. Coumans, J.A.R. Veen, J.
Catal. 269 (2010) 201–218.
[29] S.M. Riseman, F.E. Massoth, G.M. Dhar, E.M. Eyring, J. Phys. Chem. 86 (1982)
1760–1763.
[30] B.A. Morrow, I.A. Cody, J. Phys. Chem. 80 (1976) 1995–1998.
[31] A. Alhanash, E.F. Kozhevnokiva, I.V. Kozhevnikov, Appl. Catal. A: Gen. 378
(2010) 11–18.
[32] E. Tsukuda, S. Sato, R. Takahashi, T. Sodesawa, Catal. Commun. 8 (2007)
1349–1353.
(
[
2
[
[
12] D. Varisli, T. Dogu, G. Dogu, Chem. Eng. Sci. 62 (2007) 5349–5352.
13] A. Zsigmond, P. Bata, M. Fekete, F. Notheisz, J. Environ. Protect. 1 (2010)
2
01–205.
14] S.-H. Chai, H.-P. Wang, Y. Liang, B.-Q. Xu, Appl. Catal. A: Gen. 353 (2009)
13–222.
[
2
[
[
15] D. Leckel, Energy Fuels 23 (2009) 2370–2375.
16] K. Okada, T. Tomita, Y. Kameshima, A. Yasumori, T. Yano, K.J.D. MacKenzie,
Microporous Mesoporous Mater. 37 (2000) 355–364.
[
17] J. Lee, S. Hwang, J.G. Seo, U.G. Hong, J.C. Jung, I.K. Song, J. Ind. Eng. Chem. 17
[33] S.R. Ginjupalli, S. Mugawar, P. Rajan N., P.K. Balla, V.R.C. Komandur, Appl. Surf.
Sci. 309 (2014) 153–159.
(
2011) 310–315 (2007) 1027–1030.
[
[
[
18] G. Busca, V. Lorenzelli, G. Ramis, R.J. Willey, Langmuir 9 (1993) 1492–1499.
19] R. Al-Oweini, H. El-Rassy, J. Mol. Struct. 919 (2009) 140–145.
20] W. Kuang, A. Rives, M. Fournier, R. Hubaut, Appl. Catal. A: Gen. 250 (2003)
[34] W. Suprun, M. Lutecki, T. Haber, H. Papp, J. Mol. Catal. A: Chem. 309 (2009)
71–78.
[35] J. Barrault, J.-M. Clacens, Y. Pouilloux, Top. Catal. 27 (2004) 137–142.
[36] R.C. Schulz, Angew. Chem. Int. Ed. 3 (1964) 416–423.
2
21–229.