Highly efficient and robust Mg0.388Al2.408O4 catalyst for gas-phase decarbonylation of lactic acid to acetaldehyde

Article abstract of DOI:10.1016/j.jcat.2015.05.016

Abstract The process for decarbonylation of lactic acid into acetaldehyde over magnesium aluminum oxides was explored. Magnesium aluminum oxides were prepared with co-precipitation method by varying pH values, Mg/Al molar ratios and calcination temperatures. The as-prepared magnesium aluminum oxides were characterized by nitrogen adsorption-desorption, XRD, FT-IR, NH₃-TPD, CO₂-TPD and SEM, and were employed to catalyze the gas-phase decarbonylation of lactic acid to produce acetaldehyde. It is found that pH value is a crucial factor for the formation of magnesium aluminum oxides. At pH = 7-8, the obtained magnesium aluminum oxide is indexed to Mg_0.388Al_2.408O₄, while at pH > 8, it is ascribed to MgAl₂O₄ spinel. At low calcination temperature such as 550°C, Mg_0.388Al_2.408O₄ can be formed, and it enhances crystallinity with an increase of calcination temperature. However, as the calcination temperature exceeded 1200°C, the structure of Mg_0.388Al_2.408O₄ encountered a serious destruction. Comparative study on catalytic performance for Mg_0.388Al_2.408O₄ and MgAl₂O₄ spinel suggests that the former has more excellent performance than the latter. Besides mixtures including Mg_0.388Al_2.408O₄ and Al₂O₃, pure MgO and pure Al₂O₃ were also investigated on their catalytic performance. In the presence of Mg_0.388Al_2.408O₄, the stability experiment was performed at high LA LHSV such as 13.0 h^-1. Encouragingly, the decarbonylation reaction of lactic acid proceeded efficiently at around 500 h on stream, and acetaldehyde selectivity remained constant (ca. ～93%).

Full text of DOI:10.1016/j.jcat.2015.05.016

Journal of Catalysis 329 (2015) 206–217
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Highly efficient and robust Mg0.388Al2.408O catalyst for gas-phase
4
decarbonylation of lactic acid to acetaldehyde
Congming Tang a, , Zhanjie Zhai , Xinli Li , Liangwei Sun , Wei Bai
⇑
a
a
a
b
a
Chemical Synthesis and Pollution Control Key Laboratory of Sichuan Province, China West Normal University, Nanchong, Sichuan 637002, PR China
Chengdu Institute of Organic Chemistry, Chinese Academy of Sciences, Chengdu, Sichuan 610041, PR China
b
a r t i c l e i n f o
a b s t r a c t
Article history:
The process for decarbonylation of lactic acid into acetaldehyde over magnesium aluminum oxides was
explored. Magnesium aluminum oxides were prepared with co-precipitation method by varying pH val-
ues, Mg/Al molar ratios and calcination temperatures. The as-prepared magnesium aluminum oxides
Received 20 January 2015
Revised 13 April 2015
Accepted 20 May 2015
3 2
were characterized by nitrogen adsorption–desorption, XRD, FT-IR, NH -TPD, CO -TPD and SEM, and
were employed to catalyze the gas-phase decarbonylation of lactic acid to produce acetaldehyde. It is
found that pH value is a crucial factor for the formation of magnesium aluminum oxides. At pH = 7–8,
Keywords:
Decarbonylation
Lactic acid
Bulk chemicals
Magnesium aluminum oxide
Acetaldehyde
the obtained magnesium aluminum oxide is indexed to Mg0.388Al2.408
to MgAl spinel. At low calcination temperature such as 550 °C, Mg0.388Al2.408
it enhances crystallinity with an increase of calcination temperature. However, as the calcination temper-
ature exceeded 1200 °C, the structure of Mg0.388Al2.408 encountered a serious destruction. Comparative
study on catalytic performance for Mg0.388Al2.408 and MgAl spinel suggests that the former has
more excellent performance than the latter. Besides mixtures including Mg0.388Al2.408 and Al , pure
MgO and pure Al were also investigated on their catalytic performance. In the presence of
the stability experiment was performed at high LA LHSV such as 13.0 h .
4
O , while at pH > 8, it is ascribed
2
O
4
O
4
can be formed, and
O
4
O
4
2 4
O
O
4
2 3
O
2 3
O
ꢀ1
Mg0.388Al2.408
O
4
,
Encouragingly, the decarbonylation reaction of lactic acid proceeded efficiently at around 500 h on
stream, and acetaldehyde selectivity remained constant (ca. ꢁ93%).
Ó 2015 Elsevier Inc. All rights reserved.
1
. Introduction
preparation of carbonyl compounds such as acetic acid, acrylic
acid, and other
a,b-unsaturated acids via carbonylation reaction
In organic synthesis, acetaldehyde is a very important com-
of corresponding substrates [9–13]. Carbon monoxide is also used
to synthesize propionaldehyde and other aldehydes through
hydroformylation reaction of ethylene and other alkenes [14–17].
As a platform molecule, LA is utilized to produce many value added
chemicals such as acrylic acid [18–23], acetaldehyde [8],
2,3-pentanedione [24,25], propionic acid [26], pyruvic acid [27]
and polylactic acid [28]. Except for corn starch, rich and
inexpensive biomass materials such as cellulose [29], sugars [30],
and sorbitol [31] have also been used to produce LA. Thus
pound and has been utilized as a useful synthon for various impor-
tant chemicals such as peracetic acid, pentaerythritol, pyridine
bases, butyleneglycol, and chloral [1,2]. So far, acetaldehyde is
2 2
mainly produced via ethylene partial oxidation using PdCl –CuCl
catalyst. Due to the increasing depletion of petroleum reserves that
are used to produce ethylene via high temperature pyrolysis or cat-
alytic cracking, this route on ethylene partial oxidation will be
restricted in the near future. For this reason, it has become a heated
research that acetaldehyde is produced from the biomass. Several
references have recently reported that acetaldehyde is produced
via catalytic dehydrogenation of ethanol or partial oxidation of
ethanol [1,3–7]. Besides, acetaldehyde is also obtained through
decarbonylation of lactic acid (LA) accompanying with
co-product of carbon monoxide [8]. However, carbon monoxide
is viewed as an important carbonyl synthesis source for the
acetaldehyde produced from LA has displayed
a potential
perspective. However few researches on decarbonylation of LA to
acetaldehyde have been reported so far. Katryniok et al. [8]
reported silica supported heteropolyacids for the catalytic decar-
bonylation of LA to acetaldehyde, achieving 81–83% yield of
acetaldehyde. More recently, we have reported metal sulfates
and aluminate phosphate as catalysts for the decarbonylation of
LA to acetaldehyde [32,33]. However, to my knowledge, magne-
sium aluminum oxides have not been used to catalyze decarbony-
lation of LA to acetaldehyde so far.
⇑
Corresponding author.
E-mail address: tcmtang2001@163.com (C. Tang).
http://dx.doi.org/10.1016/j.jcat.2015.05.016
021-9517/Ó 2015 Elsevier Inc. All rights reserved.
0

C. Tang et al. / Journal of Catalysis 329 (2015) 206–217
207
ꢀ
1
Magnesium aluminum oxides have attracted great interest in
academy and industry due to its unique properties such as high
resistance to chemicals, good mechanical strength in wide range
of temperatures, low dielectric constant, excellent optical proper-
ties, low thermal expansion and good catalytic performances
the range of 500–4000 cm on a Nicolet 6700 spectrometer. The
morphologic features of the catalysts are examined by scanning
electron microscope (SEM, JSM-6510) (shown in Fig. S1). The
specific surface areas of catalysts are measured through nitrogen
adsorption at 77 K using Autosorb IQ instrument. Prior to adsorp-
tion, the samples were treated at 250 °C under vacuum for 6 h
and the specific surface area was calculated according to the Bru
nauer–Emmett–Teller (BET) method. Pore size of catalysts is calcu-
lated from desorption branch data on the Barrett–Joyner–Halenda
(BJH) model. Surface acid and base properties of the samples are
[
34–41]. Several preparation methods such as hydrothermal tech-
niques, sol–gel, spray plasma, cool drying, controlled hydrolysis,
co-precipitation and aerosol method have been developed to syn-
thesize magnesium aluminum oxides [36,41–44]. Among these
methods, co-precipitation is viewed as a simple method to synthe-
size magnesium aluminum oxides. Due to excellent properties,
magnesium aluminum oxides are widely used for optical engineer-
ing applications, electronic humidity sensors, integrated electronic
devices, aluminum electronic cells and adsorbents [43]. Besides, its
low acidity and thermal stability made magnesium aluminum
oxide using as an excellent catalyst or catalyst support for oxida-
estimated by NH
Quantachrome Instrument. The sample (ca. 50–60 mg) is purged
with dry Ar (50 mL/min, purity >99.999 %) at 500 °C for 1.0 h, fol-
lowed by reducing the furnace temperature to 80 °C, and switching
to a flow of 8 % NH /Ar or 10 % CO /Ar for 1 h to execute NH or
CO adsorption. Then, NH or CO adsorbed on the sample is des-
orbed in the range of 80–700 °C at a rate of 10 °C/min.
3
-TPD and CO
2
-TPD, respectively, on
a
v
t
v
t
3
v
t
2
3
2
3
2
2 3
tion of SO to SO [45], selective catalytic reduction of NO [46],
water–gas shift reaction [47], and propane dehydrogenation [48].
In this work, magnesium aluminum oxides with different struc-
tures were prepared, and used as a catalyst for decarbonylation of
LA into acetaldehyde. Effect of pH values, calcination temperatures
and Mg/Al molar ratios on formation of magnesium aluminum oxi-
des was investigated. Based on these, we further discussed the
relationship between preparation conditions for magnesium alu-
minum oxides and catalytic performances.
2
.4. Catalyst evaluation
The synthesis of acetaldehyde from LA over the catalysts is car-
ried out in a fixed-bed quartz reactor with an 4 mm inner diameter
operated at atmospheric pressure. The catalyst (ca. 0.3 g, 20–40
meshes) is placed in the middle of the reactor and quartz wool is
placed in both ends. Firstly, the catalyst is pretreated at the
2
required reaction temperature (ca. 380 °C) for 1.0 h under N with
2
. Experimental section
high purity (0.1 MPa, 1.0 mL/min). The feedstock (20 wt% solution
of LA) is then pumped into the reactor (LA aqueous solution flow
rate, 1.0 mL/h) and driven through the catalyst bed by nitrogen.
The contact time of reactant over the catalyst is around 0.5 s, and
the contact time is estimated according to Eq. (1) [32,49]. The liq-
uid products are condensed using ice-water bath and analyzed
off-line using a SP-6890 gas chromatograph with a FFAP capillary
column connected to a FID. Quantitative analysis of the products
is carried out by the internal standard method using n-butanol as
the internal standard material. GC–MS analyses of the samples
are performed using Agilent 5973N Mass Selective Detector attach-
ment. The reaction tail gas is analyzed using GC with a packed col-
umn of TDX-01 connected to TCD detector. The conversion of LA
and the selectivity toward acetaldehyde or other by-products are
calculated according to Eqs. (2) and (3).
2
.1. Materials
Lactic acid (analytic grade, 85–90 wt%) is obtained from
Chengdu Kelong Chemical Reagent Co. and is used for the decar-
bonylation reaction of LA without further purification.
Triple-distilled water is prepared in the laboratory and used to
dilute lactic acid for required concentration. Aluminum nitrate
(
Al(NO
3
)
3
ꢂ9H
2
O), magnesium nitrate (Mg(NO
3
)
2 2
ꢂ6H O), ammonia
solution (25–28 wt%), acetaldehyde, acrylic acid, propionic acid,
acetic acid, 2,3-pentanedione and n-butanol, together with hydro-
quinone are purchased from Sinopharm Chemical Reagent Co., Ltd.
Acrylic acid, propionic acid, acetic acid, 2,3-pentanedione and
acetaldehyde are used for gas chromatograph reference materials,
and n-butanol is utilized as an internal standard material.
Hydroquinone (0.3 wt%) is used as a polymerization inhibitor.
3600 ꢃ 273:15 ꢃ Vcat:
t
C
¼ 2
ð1Þ
2; 400 ꢃ ðn
LA þ nH O þ n
C
Þ ꢃ T
2
2.2. Preparation of catalysts
tC: contact time (s); Vcat.: catalyst volume (mL); nLA: the moles of
lactic acid passed per hour; n(H2O): the moles of water in lactic acid
According to previous reports [40,43], magnesium aluminum
aqueous solution feed passed per hour; n : the moles of carrier gas
C
oxides are prepared with a co-precipitation method. In a typical
experiment, 5.0 g Mg(NO and 14.6 g Al(NO ꢂ9H
ꢂ6H
Mg/Al molar ratio = 1:2) are fully dissolved in 100 mL distilled
passed per hour; T: reaction temperature (K).
3
)
2
2
O
3
)
3
2
O
n
0
ꢀ n
1
Conversion ð%Þ ¼
Selectivity ð%Þ ¼
ꢃ 100
ð2Þ
(
n
0
water under a stirring state for 1 h at room temperature. Next,
the resultant solution is adjusted to different pH values (ca.
pH = 7–8, 8–9, 10–11, >11) to form a white precipitate by dropwise
addition of ammonium hydroxide solution (25 wt%). The resulting
precipitate is filtered, completely rinsed with distilled water and
dried at 120 °C for around 5 h. Besides other magnesium aluminum
oxides with different Mg/Al molar ratios are also prepared with a
similar method. Prior to use, the catalyst is calcined at demanded
temperature in air for 6 h.
np
ꢀ n
ꢃ 100
ð3Þ
n
0
1
where n
0
is the molar quantity of LA fed into reactor, n
1
is the
molar quantity of LA in the effluent, and n
p
is the molar quantity
of lactic acid converted to acetaldehyde or other byproducts such
as propionic acid, acrylic acid, acetic acid, 2,3-pentanedione.
Area-specific catalytic rate is defined as previous references
reported [50,51], and is determined with Eqs. (4) and (5).
2.3. Catalyst characterization
LA Consumption rate
amount of LA consumed per hour in the reactor ðmmol=hÞ
¼
Powder X-ray diffraction measurement is conducted on a
Dmax/Ultima IV diffractometer operated at 40 kV and 20 mA with
Cu Ka radiation. The FTIR spectra of the catalysts are recorded in
2
surface area of catalyst in the reactor ðm Þ
ð4Þ

2
08
C. Tang et al. / Journal of Catalysis 329 (2015) 206–217
AD formation rate
such as specific surface area, pore volume and pore size. The speci-
fic surface area of sample is very sensitive to calcination tempera-
ture. The specific surface area of sample drastically reduces with an
amount of AD formed per hour in the reactor ðmmol=hÞ
¼
2
surface area of catalyst in the reactor ðm Þ
2
increase of calcination temperature (ca. 224.1 m /g at 550 °C and
ð5Þ
2
3
.2 m /g at 1200 °C). Similar results on Mg–Al–O composites were
reported by previous references [41]. Contrary to the specific sur-
face area, pore diameter of sample rapidly increases with an
increase of calcination temperature. Table 3 shows physical prop-
erties of magnesium aluminum oxides with different Mg/Al molar
ratios together with pure MgO and Al O . It is clearly observed that
3
. Results and discussion
3.1. Characterization
2
3
all magnesium aluminum oxides have a higher specific surface
2
2
3.1.1. BET
area than pure samples MgO (16.4 m /g) or Al O (38.7 m /g).
2
3
As for heterogeneous catalyst, its surface area is an important
factor due to the catalytic reaction occurred on the surface
18,52–54]. In this section, effect of preparation parameters such
as pH values, calcination temperatures and Mg/Al molar ratios on
the specific surface, pore volume and pore distribution of the cat-
alysts was investigated with nitrogen adsorption–desorption at
This indicates that new species can form from a mixture including
Mg and Al precursors calcined at high temperature such as 1000 °C.
Apart from MgO (ca. 2.9 nm) and the magnesium aluminum oxide
with Mg/Al = 1:1 (ca. 24.2 nm), the pore sizes for other samples are
close to each other, distributing between 9 nm and 13.1 nm.
[
7
7 K using Autosorb IQ instrument, and the results were depicted
3.1.2. XRD and FT-IR
in Tables 1–3 and Figs. S2–S4. From the BET data given in
Table 1, the specific surface of magnesium aluminum oxides
decreases with an increase of pH values except for the sample pre-
Fig. 1A shows the XRD patterns of Mg–Al–O composites pre-
pared at different pH values. It is obvious that the magnesium alu-
minum oxide with pH = 7–8 displays the different characteristic
diffraction peaks compared to other samples. The sample with
pH = 7–8 well matches with the standard magnesium aluminum
2
pared at pH = 8–9. For example, the specific surface is 63.8 m /g at
2
pH = 7–8, while it decreases to 46.4 m /g at pH > 11. As for the pore
size of sample, it increases with an increase of pH values. For
instance, the pore diameter of the sample at pH = 7–8 is low, only
oxide (Mg0.388Al2.408O , PDF#48-0528); and it exhibits strong char-
acteristic diffraction peaks at 19.2°, 31.6°, 37.3°, 45.4°, 56.5°, 60.2°
and 66.3°, which can be indexed to (111), (220), (311), (400),
(422), (511) and (440) diffractions, respectively. Other samples
4
1
3.1 nm, but it increases to 34.4 nm at pH > 11. Table 2 shows
effect of calcination temperature on sample physical properties
2 4
match well with the standard spinel (MgAl O , PDF#21-1152);
and display strong characteristic diffraction peaks at 19.2°, 31.4°,
37.0°, 44.9°, 59.4° and 65.5°, which can be ascribed to (111),
Table 1
a
BET data of magnesium aluminum oxides at different pH values.
(
220), (311), (400), (511) and (440) diffractions, respectively.
2
3
Pore sizeb (nm)
This result suggests that pH values have an important influence
in the formation of magnesium aluminum oxides. Fig. 1B depicted
effect of calcination temperature on the formation of Mg–Al–O
pH value
S
BET (m /g)
Vol (cm /g)
7
8
1
–8
–9
0–11
63.8
44.4
51.3
46.4
0.32
0.40
0.40
0.34
13.1
24.2
24.3
34.4
composite (Mg0.388Al2.408
is relatively lower, only 550 °C, we can observe the formation of
. But it exhibits low characteristic diffraction peaks,
indicating a low crystallinity for Mg0.388Al2.408 . When the calci-
nation temperatures enhanced from 550 °C to 1000 °C, the charac-
teristic diffraction peaks for Mg0.388Al2.408 gradually increased,
4
O ). When the calcination temperature
>11
a
Mg0.388Al2.408O
4
Mg/Al molar ratio = 1:2, catalyst calcined at 1000 °C.
Calculated from desorption branch data on the Barrett–Joyner–Halenda (BJH)
b
O
4
model.
O
4
suggesting an increase of the crystallinity with an increase of cal-
cination temperatures. As the calcination temperatures further
increased (ca. 1200 °C), new diffraction peaks occurred, indicating
that the structure of Mg0.388Al2.408O encountered damage, and
4
some converted other new species. Fig. 1C shows XRD patterns
of magnesium aluminum oxides with Mg/Al molar ratios as well
Table 2
BET data of magnesium aluminum oxides at different calcination temperatures.a
2
3
Pore sizeb (nm)
Calcination temperature (°C)
S
BET (m /g)
Vol (cm /g)
550
750
900
224.1
114.0
84.8
63.8
3.2
0.33
0.35
0.38
0.32
0.40
4.7
8.1
13.1
13.1
61.2
as pure MgO and Al
exhibited different diffraction peaks compared to other samples,
and it matched well with the standard spinel (MgAl
PDF#21-1152). It is known that pH value in synthetic conditions
for MgAl spinel is around 9 [40–43]. However, the Mg–Al–O
composites were prepared at pH = 7–8, lower than that prepared
2 3
O . The Mg–Al–O composite with Mg/Al = 1
1
1
000
200
a
2 4
O ,
Mg/Al molar ratio = 1:2, pH, 7–8.
Calculated from desorption branch data on the Barrett–Joyner–Halenda (BJH)
b
model.
2 4
O
for MgAl
for Al(OH)
2
O
4
spinel. We also note that solubility product constant
Table 3
ꢀ33
(4.57 ꢃ 10 ) is far less than that of Mg(OH)
2
a
3
ꢀ
BET data of magnesium aluminum oxides at different Mg/Al molar ratios.
11
(
1.82 ꢃ 10 ). It happened that high concentration of Mg precur-
sor compensated low pH value in the process of co-precipitation
using ammonia water as a precipitant. Thus MgAl spinel was
2
3
Pore sizeb (nm)
Mg/Al
MgO
S
BET (m /g)
Vol (cm /g)
16.4
55.1
63.8
64.1
62.0
53.2
38.7
0.24
0.37
0.32
0.29
0.20
0.19
0.21
2.9
24.2
13.1
13.1
10.0
9.0
2 4
O
1
1
1
1
1
:1
:2
:3
:6
:8
obtained at Mg/Al = 1. At the same time, a small quantity of MgO
also existed in Mg–Al–O composite. When Mg/Al molar ratio
increased to 2, pure Mg0.388Al2.408
increase of Mg/Al molar ratios, Al
in the samples.
O
4
was formed. With further
2 3
O content gradually increased
Al
2
O
3
10.0
a
In order to obtain information on the functional groups in the
prepared samples, IR analyses were carried out as shown in
Fig. 2A–C. From IR spectra of samples shown in Fig. 2A, three
Calcination temperature: 1000 °C, pH, 7–8.
Calculated from desorption branch data on the Barrett–Joyner–Halenda (BJH)
b
model.

C. Tang et al. / Journal of Catalysis 329 (2015) 206–217
209
Fig. 1. XRD of magnesium aluminate composites (A) prepared at different pHs,
B) prepared at different calcination temperatures and (C) prepared with different
Mg/Al molar ratios. Conditions: (A), Mg/Al molar ratio = 1:2, catalyst calcined at
(
Fig. 2. FT-IR of magnesium aluminate composites (A) prepared at different pH
values, (B) prepared at different calcination temperatures and (C) prepared with
different Mg/Al molar ratios. Conditions: (A), Mg/Al molar ratio = 1:2, catalyst
calcined at 1000 °C; (B), pH, 7–8, Mg/Al molar ratio = 1:2; (C), pH, 7–8, catalyst
calcined at 1000 °C.
1
000 °C; (B), pH, 7–8, Mg/Al molar ratio = 1:2; (C), pH, 7–8, catalyst calcined at
000 °C.
1

2
10
C. Tang et al. / Journal of Catalysis 329 (2015) 206–217
Fig. 3. TPD profiles of magnesium aluminum oxides (A) NH
NH -TPD of magnesium aluminum oxides prepared at different calcination temperatures and (D) corresponding CO
prepared with different Mg/Al molar ratios together with pure MgO and Al and (F) corresponding CO -TPD.
3
-TPD of magnesium aluminum oxides prepared with different pH values and (B) corresponding CO
2
-TPD; (C)
3
2
-TPD; (E) NH -TPD of magnesium aluminum oxides
3
2
O
3
2
samples prepared at pH > 8 are almost identical. Absorption bands
spinel structure in the samples [42]. The sample prepared at
pH = 7–8 displays slightly different IR spectra in range of 500–
ꢀ1
ꢀ1
of around 3440 cm
adsorption H O, respectively [38,41,42]. Besides, the IR spectra of
these three samples exhibit two characteristic frequencies at ca.
and 1630 cm
are assigned to OH and
ꢀ
1
2
900 cm , suggesting that the structure of Mg–Al–O composite
) obtained at pH = 7–8 is different from other sam-
). This result accords with the aforementioned XRD
characterization. Fig. 2B shows the IR spectra of the samples
(Mg0.388Al2.408O
4
ꢀ1
ꢀ1
5
33 cm and 700 cm attributing to [AlO
6
] groups and the lattice
2 4
ples (MgAl O
vibration of Mg–O stretching, indicating an existence of MgAl
2 4
O

C. Tang et al. / Journal of Catalysis 329 (2015) 206–217
211
calcined at different temperatures. When the calcination tempera-
ture is below 1000 °C, IR spectra for theses samples are almost con-
stant except that the intensities for absorption bands are different
from each other. However as the calcination temperature further
increases from 1000 °C to 1200 °C, a drastical change takes place
in the IR spectra. This result is also in concert with that obtained
from XRD characterization in Fig. 1B. From IR spectra of the sam-
ples with different Mg/Al molar ratios shown in Fig. 2C, the weak
ꢀ1
absorption band at 819 cm for the sample with Mg/Al molar
ratio = 1 obviously disappears in comparison with the sample with
Mg/Al molar ratio = 2, indicating a difference in their molecular
structures.
3
.1.3. NH
Fig. 3A–F shows the NH
ples varying pH values, calcination temperatures and Mg/Al molar
ratios together with corresponding CO -TPD profiles. In
3
-TPD/CO
2
-TPD of Mg–Al–O composites
3
-TPD profiles of the as-prepared sam-
2
Fig. 3A and B, the sample with pH = 7–8 displayed a broad deso-
rption peak in the region of 150–300 °C in which were regarded
as weak-medium acidity (alkalinity), and was different in TPD
curves from others obtained at pH > 8. The peak centered at around
2
00 °C for other samples splits into two peaks, suggesting that
acidity (alkalinity) increased with an increase of pH values. From
the profiles shown in Fig. 3C and D, desorption peak at around
4
00 °C characterized medium-strong acidity (alkalinity) moved to
the region in low temperatures as calcination temperature
increased, indicating that acidity (alkalinity) for samples
decreased. It is noteworthy that the desorption peaks of the sample
almost disappeared when the calcination temperature further
increased to 1200 °C. These indicated that both acidity and alkali-
nity for the sample calcined at 1200 °C became the weakest.
Fig. 3E and F shows the TPD profiles for Mg-Al-O composites
with different ratios as well as pure Al
By comparing the NH -TPD curve of MgO with corresponding
CO -TPD curve, it is found that they exhibit double peaks at
50 °C and 400 °C, and the high temperature desorption peak for
2 3
O and MgO, respectively.
3
2
2
Fig. 4. Performance of magnesium aluminum oxides at different pHs by the time
the former is higher than the latter. This result suggests that acid-
ity–alkalinity balance does not exist on the surface of MgO.
course of LA conversion (A) and acetaldehyde selectivity (B). (a) Catalyst volume,
0
.38 mL, magnesium aluminum oxides prepared through Mg(NO
precursors at calcination temperature of 1000 °C, reaction temperature, 380 °C,
particle size: 20–40 meshes, carrier gas N : 1 mL/min, feed flow rate: 1 mL/h, LA
3 2 3 3
) and Al(NO ) as
Through further observation for CO
found that weak alkalinity sites are more abundant than strong
alkalinity sites on the surface of MgO. As for pure Al , a broad
2
-TPD curve for MgO, it is easily
2
feedstock: 20 wt% in water. (b) LA: lactic acid, AD: acetaldehyde, PA: propionic acid,
ACA: acetic acid, AA: acrylic acid, PD: 2,3-pentanedione.
2 3
O
and slight desorption peak at 450 °C occurs in Fig. 3E except for
the peak at 200 °C, suggesting that weak acidity sites as well as
medium strong acidity sites exist on the surface of pure Al
2 3
O .
dis-
Table 4
Effect of pH values.
Observing the CO -TPD curves shown in Fig. 3F, pure Al
2
2
O
3
a
plays only a desorption peak at 200 °C. All the Mg–Al–O compos-
ites with different Mg/Al molar ratios except for the sample with
Mg/Al = 1 exhibit a broad desorption peak in low temperature
region (150–300 °C), suggesting that weak and medium acidity
pH
LA conv.
%)
Sel. (%)b
Area-specific catalytic
ꢀ1
ꢀ2
(
rate (mmol h
PA ACA AA PD LA
consumption formation
m )
AD
AD
(
alkalinity) sites exist on the surface of the samples. In addition,
7
8
1
–8
–9
0–
100
79.1
11
91.4 2.6 2.2
90.2 3.1 1.9
78.7
2.6 0.9 878.1
2.7 1.1 963.6
88.6
802.6
869.2
3.8
the intensity of desorption peak is slightly different from each
other, indicating that the number of weak and medium acidity
(
alkalinity) sites is also different from each other.
1.1
859.4
761.5
2.0
4.3
>
11
67.9
83.4 6.4 2.5
6.1 1.5 819.8
683.7
3
3
.2. Catalyst evaluation
a
Catalyst, 0.38 mL, 0.28–0.29 g; magnesium aluminum oxides prepared through
Mg(NO and Al(NO as precursors at calcination temperature of 1000 °C; reac-
tion temperature, 380 °C; particle size, 20–40 meshes; carrier gas N , 1 mL/min;
2
feed flow rate: 1 mL/h, LA feedstock: 20 wt% in water; TOS, 4–6 h.
LA: lactic acid, AD: acetaldehyde, PA: propionic acid, ACA: acetic acid, AA:
acrylic acid, PD: 2,3-pentanedione.
.2.1. Effect of catalyst preparation means on the catalytic
)
3 2
3 3
)
performance
.2.1.1. pH values. Catalytic reactions for gas phase decarbonylation
3
b
of LA over Mg–Al–O composites were performed at 380 °C with LA
concentration (20 wt%) and feed flow rate (1 mL/h), and the results
were shown in Fig. 4A and B and Table 4. It is clearly seen from
Fig. 4A that LA conversion is drastically influenced by pH values.
For example, LA conversion decreases with an increase of pH val-
ues. Based on the effect of TOS (time on stream) on LA conversion
at different pH values, different catalysts display evidently
different stabilities. It is noted that the Mg–Al–O composite pre-
pared at pH = 7 offers an excellent stability. For example, LA was
almost fully converted on the whole time (TOS, 1–8 h). However

2
12
C. Tang et al. / Journal of Catalysis 329 (2015) 206–217
LA conversions over other Mg–Al–O composites prepared at pH > 8
drastically decreased as the time on stream lengthened. From the
results shown in Fig. 4B, similar experimental phenomena were
observed except for the Mg–Al–O composite prepared at pH = 8–
9
. AD selectivity decreased with an increase of pH values.
Furthermore, AD selectivity over all the Mg–Al–O composite cata-
lysts slightly fluctuated with a change in reaction time. In addition,
as an important method characterized catalytic performance,
area-specific catalytic rates were calculated at TOS = 4–6 h
(
depicted in Table 4). We can easily observe that area-specific cat-
alytic rate decreased with an increase of pH values except for the
Mg–Al–O composite prepared at pH = 7–8. It is noteworthy that
LA conversion over the Mg–Al–O composite catalyst prepared at
pH = 7–8 attains 100%. It is likely that some of the catalyst surface
was not utilized. However the area-specific catalytic rate for these
catalysts was determined based on the total surface of the catalyst.
Thus the catalyst prepared at pH = 7–8 displayed lower
area-specific catalytic rate ostensibly. In fact, it can display higher
area-specific catalytic rate (confirmed by the following
Section 3.2.2.3 Effect of LA LHSV). According to XRD and FTIR char-
acterizations shown in Figs. 1A and 2A, we found that the Mg–Al–O
composite obtained at pH = 7–8 belongs to Mg0.388Al2.408
4
O , while
Mg–Al–O composites obtained at other pH values ascribe to spinel
2
(
2 4 4
MgAl O ). Besides, Mg0.388Al2.408O (63.8 m /g) has bigger specific
2
surface area than other spinels (44.4–51.3 m /g). It is known that
as for heterogeneous catalysts, bigger specific surface area favors
to the catalytic reaction. Thus it is not astonishing that the highest
LA conversion over the Mg–Al–O composite at pH = 7–8 was
achieved. In addition, decarbonylation of LA to acetaldehyde cat-
alyzed by weak-medium acid has been recognized by previous
reports [32,33]. Through NH
composites with different pH values, the Mg–Al–O composite
) with pH = 7–8 displays a broad desorption peak
3
-TPD characterization for Mg–Al–O
(
Mg0.388Al2.408O
4
in the region of 150–300 °C in which were regarded as
weak-medium acidity, and is different in TPD curves from others
obtained at pH > 8. The peak centered at around 200 °C for other
samples splits into two peaks, suggesting that acidity increases
with an increase of pH values. With respect to the acidity
2 3
Fig. 5. Performance of magnesium aluminum oxides together with MgO and Al O
by the time course of LA conversion (A) and acetaldehyde selectivity (B).
4
of the catalysts, the Mg–Al–O composite (Mg0.388Al2.408O )
with pH = 7–8 has more appropriate acidity compared to
others for decarbonylation of LA to acetaldehyde. Considering
the specific surface area and acidity of catalysts, the Mg–Al–O
Table 5
Effect of Mg/Al molar ratio.
a
Sel. (%)b
Mg/
Al
LA conv.
(%)
Area-specific catalytic
composite (Mg0.388Al2.408
O
4
) obtained at pH = 7–8 is a potentially
ꢀ1
ꢀ2
rate (mmol h
ACA AA PD LA
consumption formation
m )
excellent catalyst for gas phase decarbonylation of LA to
acetaldehyde.
AD
PA
AD
MgO
12.9
80.7
100
89.4
74.3
82.1
70.3
69.9 13.6 8.9
6.0 1.3 725.8
507.3
712.2
802.6
658.5
515.2
674.8
787.5
3
.2.1.2. Mg/Al molar ratio. Fig. 5 and Table 5 show the catalytic per-
formance of Mg–Al–O composites with different Mg/Al molar
ratios as well as pure MgO and Al . From Fig. 5A, time on stream
1:1
89.9
91.4
90.3
87.7
89.2
91.2
3.3 2.2
2.6 2.2
2.3 3.4
3.7 3.5
2.5 4.1
2.7 2.7
3.5 1.0 792.2
2.6 0.9 878.1
2.7 1.0 729.3
3.6 1.2 587.4
3.0 1.0 756.5
2.3 0.9 863.5
1:2
1:3
2
O
3
1
1
:6
:8
had an important influence in LA conversion over some catalysts
such as Mg–Al–O composites with Mg/Al = 1:6 and 1:8, respec-
tively, pure Al O and MgO. For these catalysts except for pure
2 3
MgO, they displayed an excellent initial activity, while as time on
stream lengthened the activity (LA conversion) drastically
decreased. Interestingly, with an increase of Mg/Al molar ratios
the catalyst displayed better activity and stability. For example,
LA conversion maintained 100% on the whole reaction time (TOS:
2 3
Al O
a
Catalyst, 0.38 mL, MgO, 0.17 g, Al
prepared through Mg(NO and Al(NO
of 1000 °C, 0.30–0.32 g, pH, 7–8, particle size: 20–40 meshes, carrier gas N
mL/min, feed flow rate: 1 mL/h, LA feedstock: 20 wt% in water, reaction temper-
ature: 380 °C.
2
O
3
, 0.33 g, magnesium aluminum oxides
) as precursors at calcination temperature
)
3 2
3 3
2
:
1
b
LA: lactic acid, AD: acetaldehyde, PA: propionic acid, ACA: acetic acid, AA:
acrylic acid, PD: 2,3-pentanedione.
0
–8 h) over the catalyst with Mg/Al = 1:2. However with further
increase of Mg/Al molar ratios (ca. Mg/Al = 1:1), the catalyst
offered lower LA conversion (ca. 70–81%) although the catalyst
yet remained better stability. Unlike effect of time on stream on
LA conversion, it has a slight influence in acetaldehyde selectivity
ratios. The fastest area-specific catalytic rate (LA consumption:
ꢀ1
ꢀ2
ꢀ1
ꢀ2
878.1 mmol h
m
and AD formation: 802.6 mmol h
m )
(
shown in Fig. 5B). Pure MgO (AD sel.: 65–70%) displayed less
acetaldehyde selectivity compared to other catalysts (AD Sel.:
7–92%). From the data obtained at TOS = 6–8 h shown in
Table 5, area-specific catalytic rate changed with Mg/Al molar
was achieved over the Mg–Al–O composite with Mg/Al = 1:2. As
discussed above (shown in Fig. 1C), the Mg–Al–O composite
8
obtained at Mg/Al = 1:2 is attributed to Mg0.388Al2.408
4
O with high
purity. When the Mg/Al molar ratio (ca. 1:1) is higher than 1:2, the

C. Tang et al. / Journal of Catalysis 329 (2015) 206–217
213
Mg–Al–O composite obtained is attributed to MgAl
While the Mg/Al molar ratio is lower than 1:2, the mixture includ-
and Al is obtained. It is known from the dis-
cussion in Section 3.2.1.1 that Mg0.388Al2.408 has more excellent
activity due to appropriate acidity sites on its surface.
2
O
4
spinel.
high temperature favored the formation of Mg0.388Al2.408
4
O .
Furthermore Mg0.388Al2.408 was vied as active species. Thus with
O
4
ing Mg0.388Al2.408
O
4
2
O
3
an increase of calcination temperature, the catalytic performance
became better and better. Unfortunately, LA conversion drastically
decreased to 24.3% as calcinations temperature increased to
O
4
1
200 °C. But it is not surprising that under the calcination temper-
ature of 1200 °C the catalyst partly decomposed or transformed to
other species. In addition, similar to LA conversion, acetaldehyde
selectivity slightly fluctuated within 750–1000 °C. This can be
3
.2.1.3. Calcination temperature. Fig. 6 and Table 6 show effect of
calcination temperature on the performance of catalysts. LA con-
version increased from 75.9% to 100% as the calcination tempera-
ture increased from 550 °C to 750 °C. Subsequently, LA
conversion remained constant (100%) in the range between
4
explained using Mg0.388Al2.408O selectively catalyzed decarbony-
lation of LA into acetaldehyde.
7
50 °C and 1000 °C. But LA conversion drastically decreased with
3
3
.2.2. Reaction conditions
.2.2.1. Reaction temperature. According to the discussion on prepa-
further increase of calcination temperature. For example, 24.3%
of LA conversion was achieved at 1200 °C of calcination tempera-
ture. Although area-specific catalytic rate (LA consumption:
ration conditions of catalysts in the preceding sections, we
obtained the optimal preparation conditions for Mg0.388Al2.408
O
4
ꢀ1
ꢀ2
ꢀ1
ꢀ2
2
247.5 mmol h
m
and AD formation: 1613.7 mmol h
m )
(
pH, 7–8; calcination temperature, 1000 °C; Mg/Al molar ratio,
on the catalyst calcined at 1200 °C was the highest among all the
catalysts calcined at different temperatures (ca. 550–1200 °C), LA
conversion as well as acetaldehyde selectivity is the lowest due
1
:2.). Next, effect of reaction conditions on the decarbonylation
of LA would be investigated. As an important factor determining
the reaction rate and reaction selectivity, reaction temperature
was firstly discussed [55,56]. From effect of reaction temperature
on decarbonylation of LA shown in Fig. S5A, LA conversion
increased with an increase of reaction temperature from 320 °C
to 380 °C. For a fixed catalyst, area-specific catalytic rate also
increases with an increase of reaction temperature. For this reason,
LA conversion increased with an increase of reaction temperature.
However as the reaction temperature increased to 400 °C, LA con-
version rapidly reduced at 2–6 h on stream. Interestingly,
area-specific catalytic rate at 400 °C also decreased compared to
the data obtained at 360 °C or 380 °C, indicating part deactivation
of catalyst active sites. As for acetaldehyde selectivity (shown in
Table 7), it slightly decreased with an increase of reaction temper-
ature. Besides, from correlation between TOS and acetaldehyde
selectivity depicted in Fig. S5B, it is clearly seen that acetaldehyde
selectivity was slightly influenced by TOS.
2
to its low specific surface area (only 3.2 m /g far lower than others
2
(
63.8–224.1 m /g)). In order to better understand the effect of cal-
cination temperature on the catalytic performance, we can draw
support from the XRD characterization shown in Fig. 3B. With an
increase of calcination temperature, characteristic diffraction
4
peaks for Mg0.388Al2.408O also gradually increased, suggesting that
3
.2.2.2. LA concentration. Effect of LA concentration on the reaction
performance was also investigated over Mg0.388Al2.408 at 380 °C
and the results were given in Table 8. Notably, unlike the results
obtained over Al (SO catalyst [32], acetaldehyde selectivity
O
4
2
4 3
)
was hardly influenced, whereas LA conversion was influenced with
change of LA concentration. LA was completely converted when LA
concentration was lower than 20 wt%. But a residue of LA began to
occur as LA concentration was more than 30 wt%. As for most of
by-products such as propionic acid, acetic acid and acrylic acid,
the selectivity fluctuated with an increase of LA concentration.
But 2,3-pentanedione selectivity regularly enhanced with an
increase of LA concentration. It is known that formation of
2
,3-pentanedione is via a Claisen condensation reaction of two lac-
tate moieties followed by decarboxylation and dehydration steps
24,54]. Thus higher LA concentration favors its condensation reac-
tion to form 2,3-pentanedione. However, the selectivity of
,3-pentanedione was far lower compared to other by-products.
[
2
This suggested that apart from LA concentration factor, acid–base
properties play an important role for catalytic formation of
2
2
,3-pentanedione [24,25,52]. According to CO -TPD characteriza-
tion shown in Fig. 3B, only a desorption peak occurred at low tem-
perature, indicating that weak alkalinity existed on the catalyst
surface. Therefore, low selectivity of 2,3-pentanedione was
4
obtained on Mg0.388Al2.408O catalyst at 380 °C.
3.2.2.3. Liquid hourly space velocity (LHSV). LHSV is generally used
to evaluate the performance of heterogeneous catalyst [53,57,58].
Table 9 shows the influence of LA LHSV on reaction performance.
The reaction was conducted at 380 °C with LA flow rate changed
Fig. 6. Performance of magnesium aluminum oxides calcined at different temper-
atures by the time course of LA conversion (A) and acetaldehyde selectivity (B).
ꢀ1
from 0.5 to 10 mL/h (corresponding LHSV = 1.3–26.3 h ). For LA

2
14
C. Tang et al. / Journal of Catalysis 329 (2015) 206–217
Table 6
a
Effect of calcination temperature.
Calcination temperature (°C)
LA conv. (%)
Sel. (%)b
Area-specific catalytic rate
ꢀ
1
ꢀ2
(
mmol h
m )
AD
PA
5.0
3.0
2.4
2.6
12.7
ACA
AA
PD
LA consumption
AD formation
5
7
9
000
200
50
50
00
75.9
100
100
100
24.3
83.5
88.4
91.0
91.4
71.8
5.3
4.7
3.3
2.2
4.7
4.2
2.7
2.2
2.6
6.5
1.7
1.0
0.9
0.9
4.0
166.0
443.9
616.6
878.1
2247.5
138.6
392.4
561.1
802.6
1613.7
1
1
a
Mg0.38Al2.4
O
4
catalyst, 0.38 mL, Mg/Al molar ratio = 1:2, particle size: 20–40 meshes, carrier gas N
2
: 1 mL/min, feed flow rate: 1 mL/h, LA feedstock: 20 wt% in water,
reaction temperature: 380 °C, TOS: 6–8 h.
b
LA: lactic acid, AD: acetaldehyde, PA: propionic acid, ACA: acetic acid, AA: acrylic acid, PD: 2,3-pentanedione.
Table 7
Effect of reaction temperature.a
Reaction temperature (°C)
LA conv. (%)
Sel. (%)b
AD
Area-specific catalytic rate
ꢀ
1
ꢀ2
(
mmol h
m )
PA
ACA
AA
PD
LA consumption
AD formation
3
3
3
3
4
20
40
60
80
00
54.7
82.3
91.9
100
93.7
93.5
92.0
91.4
90.1
1.2
1.4
1.9
2.6
2.8
1.9
1.5
1.8
2.2
2.5
2.6
2.5
3.0
2.6
2.0
0.3
0.9
1.0
0.9
0.9
437.4
670.1
776.5
878.1
716.5
409.9
626.6
714.4
802.6
645.6
80.0
a
4 2
Mg0.38Al2.4O catalyst, 0.38 mL, 0.28 g, Mg/Al molar ratio = 1:2, pH = 7–8, catalyst calcination temperature, 1000 °C, particle size: 20–40 meshes, carrier gas N : 1 mL/min,
LA feedstock: 20 wt% in water, feed flow rate: 1 mL/h, LA feedstock: 20 wt% in water, TOS: 6–8 h.
b
LA: lactic acid, AD: acetaldehyde, PA: propionic acid, ACA: acetic acid, AA: acrylic acid, PD: 2,3-pentanedione.
Table 8
Effect of LA concentration.a
LA concentration (wt%)
LA conv. (%)
Sel. (%)b
AD
Area-specific catalytic rate
ꢀ
1
ꢀ2
(
mmol h
m )
PA
ACA
AA
PD
LA consumption
AD formation
10
15
20
30
40
100
100
100
91.8
90.5
91.2
91.5
91.7
91.2
91.7
2.2
1.8
2.5
2.3
1.6
4.0
3.5
2.6
3.2
2.9
1.5
2.2
2.0
2.0
2.2
0.6
0.6
0.8
0.9
1.0
476.4
685.5
878.1
1081.9
1286.6
434.5
627.2
805.2
986.7
1179.8
a
Mg0.38Al2.4
O
4
catalyst, 0.38 mL, 0.28–0.30 g, Mg/Al molar ratio = 1:2, particle size: 20–40 meshes, carrier gas N
2
: 1 mL/min, feed flow rate: 1 mL/h, reaction temperature:
3
80 °C, TOS: 1–2 h.
b
LA: lactic acid, AD: acetaldehyde, PA: propionic acid, ACA: acetic acid, AA: acrylic acid, PD: 2,3-pentanedione.
Table 9
Effect of LA LHSV.a
LA solution LHSV (h 1
ꢀ
)
LA conv. (%)
Sel. (%)b
AD
Area-specific catalytic rate
ꢀ
1
ꢀ2
(
mmol h
m )
PA
ACA
AA
PD
LA consumption
AD formation
1
2
3
5
8
3.0
0.8
6.3
.3
.6
.9
.2
.4
100
100
100
100
100
97.2
88.5
82.9
89.0
91.7
92.1
91.9
93.2
94.0
95.0
95.1
4.0
2.5
1.9
1.8
1.5
1.3
0.9
0.9
3.2
2.6
2.3
2.2
1.8
1.3
1.3
1.4
2.1
2.0
2.7
3.2
2.1
2.4
2.0
2.0
1.4
0.9
0.6
0.6
0.9
0.7
0.4
0.2
228.5
878.1
203.4
805.2
1919.3
3317.2
7463.8
19785.9
45294.6
74249.9
1767.6
3048.5
6956.3
18598.7
43029.9
70611.6
1
2
2
a
4 2
Mg0.38Al2.4O catalyst, 0.38 mL, 0.29–0.30 g, Mg/Al molar ratio = 1:2, particle size: 20–40 meshes, carrier gas N : 1 mL/min, LA feedstock: 20 wt% in water, reaction
temperature: 380 °C, TOS: 1–2 h.
b
LA: lactic acid, AD: acetaldehyde, PA: propionic acid, ACA: acetic acid, AA: acrylic acid, PD: 2,3-pentanedione.
conversion, it almost remained at 100% when LA LHSV was lower
formation of acetaldehyde, indicating that decarbonylation reac-
tion of LA is faster than other side reactions. Area-specific catalytic
rate quickly increased with an increase of LA solution LHSV. For
ꢀ1
ꢀ1
ꢀ1
than 8.4 h . As LA LHSV increased from 8.4 h to 26.3 h , LA
conversion gradually decreased from 100% to 82.9%.
Acetaldehyde selectivity, unlike LA conversion, increased slowly
with an increase of LA LHSV. It is known that LA contact time on
ꢀ1
ꢀ2
example, LA consumption rate was 228.5 mmol h
m
at LA
m
ꢀ1
ꢀ1
ꢀ2
solution LHSV = 1.3 h while it attained to 74249.9 mmol h
ꢀ
1
the surface of Mg0.388Al2.408
O
4
catalyst shortens with the enhance-
at LA solution LHSV = 26.3 h . For that reason, LA conversion
ꢀ1
ment of LA LHSV. Enhancement of LA LHSV favored the selective
(82.9%) was very high at high LHSV of 26.3 h
.

C. Tang et al. / Journal of Catalysis 329 (2015) 206–217
215
3.2.3. Correlation between the acid–base property and catalytic
performance of Mg–Al–O composites
It has been known that the catalytic decarbonylation rate of LA
into acetaldehyde over the calcium hydroxyapatite would be
dependent on the catalyst surface acid–base property [50,59].
Indeed, from the results shown in Fig. 3A and B and Fig. 7, surface
acid–base property of catalysts changes with pH values. It is noted
that surface acidity and basicity depicted in Fig. 7 are determined
only in the range of 150–300 °C of desorption temperatures, which
favors decarbonylation reaction of LA [33]. An attempt is then
made to correlate the rates of LA consumption and product AD for-
mation with the surface acidity and basicity of catalysts prepared
at different pH values, and the results were shown in Figs. 8 and
9
. It is clear that LA consumption rate increased with increasing
acid density apart from acid density of 1.2
mol/m2 (correspond-
ing pH = 7–8), but decreased with the surface base density. It
l
ꢀ1
ꢀ2
seems that LA consumption rate (878.1 mmol h
m
) obtained
2
at 1.2
Actually, LA consumption rate (74249.9 mmol h
sponding LA conversion = 82.9% shown in Table 9) obtained at
lmol/m (corresponding pH = 7–8) is lower than others.
ꢀ1
ꢀ2
m
, corre-
Fig. 8. Dependences of the area-specific catalytic rate of LA consumption on the
surface acidity and basicity of the magnesium aluminum oxides prepared with
different pH values.
2
1 lmol/m is two orders of magnitude faster than others
.2
ꢀ1
ꢀ2
(
819.8–963.6 mmol h
m
, corresponding LA conversion = 67.9–
7
9.1% shown in Table 4). Similarly, AD formation rate also
increased with increasing surface acid density but decreased with
surface base density. The observations on the increase with surface
acid density and decrease with base density of the area-specific
catalytic rates for LA consumption and AD formation agreed well
with the previous reports [33,50,51].
3.2.4. Catalyst stability
Long-term stability is a very important characteristic for a
heterogeneous catalyst [55,56,60–62]. The catalytic stability of
was investigated at 380 °C and LA feed flow rate
Mg0.388Al2.408O
4
ꢀ
1
of 5 mL/h (corresponding LA LHSV = 13.0 h ), and the results were
depicted in Fig. 10. LA conversion slowly gradually decreased with
an increase of time on stream. For example, LA conversion reduced
with only 25% (from 95% to 70%) within 400 h on stream. When the
time on stream lengthened to 500 h, LA conversion still remained
above 58%. More importantly, acetaldehyde selectivity almost
remained constant (>93%) during the whole time on stream. To
our great delight, so far this is the best result concerning LA con-
version, acetaldehyde selectivity and LA LHSV. It is noted that for
Fig. 9. Dependences of the area-specific catalytic rate of AD formation on the
surface acidity and basicity of the magnesium aluminum oxides prepared with
different pH values.
Mg0.388Al2.408
O
4
, LA LHSV is at least 5 times than previous reports
[
32,33]. Long-term stability as well as high selectivity of acetalde-
hyde is also related to catalyst acid–base balance. As mentioned
Fig. 10. Durability of Mg0.388Al2.408
Mg/Al molar ratio = 1:2, particle size: 20–40 meshes, carrier gas N
feedstock: 20 wt% in water, feed flow rate: 5 mL/h, reaction temperature: 380 C.
O
4
catalyst. Catalyst volume, 0.38 mL, 0.30 g,
Fig. 7. Surface acidity and basicity of magnesium aluminum oxides prepared with
different pH values.
2
: 1 mL/min, LA

2
16
C. Tang et al. / Journal of Catalysis 329 (2015) 206–217
above, for Mg0.388Al2.408
O
4
catalyst predominant weak-medium
the mechanism for decarbonylation of LA to acetaldehyde is rarely
investigated so far. Recently, Ghantani et al. [59] have proposed the
decarbonylation reaction mechanism of LA to form acetaldehyde
over the calcium hydroxyapatite. More recently, based on alu-
minum sulfate utilized as a catalyst we have also discussed the
mechanism on decarbonylation of LA to acetaldehyde [32].
Similarly, based on the experimental results achieved over
Mg0.388Al2.408O catalyst, we proposed the possible reaction mech-
4
anism on formation of acetaldehyde from LA as followed in
Scheme 1. Firstly, dissociative adsorption of LA occurs on
acidity as well as an excellent acid–base balance plays an impor-
tant role in durability and acetaldehyde selectivity. It is known that
strong acidic sites have a strong catalysis for decomposition of C–C
bond, resulting in deposition of carbon or formation of coke on the
catalyst surface [51,60,63]. Thus the catalyst with more strong
acidic sites rapidly deactivates in the process of the catalytic decar-
bonylation reaction of LA due to covering the active sites of the cat-
alyst surface by the formed carbon or cokes. But for
4
Mg0.388Al2.408O catalyst, it contains weak-medium acidity and
remains an excellent acid–base balance. Furthermore, it lacks
strong acidic sites on the catalyst surface. Therefore
4
Mg0.388Al2.408O catalyst, where the hydroxyl group in LA molecule
adsorbs on the Mg site, forming a C–O–Mg bond and the hydroxyl
proton gets abstracted by aluminate oxygen forming Al–OH.
Subsequently, Al–OH further reacts with the carboxylic –OH group
to form aluminate ester. In the end, this aluminate ester decom-
poses to produce acetaldehyde and carbon monoxide. In addition,
the possible mechanisms for the formation of acrylic acid and
Mg0.388Al2.408
O
4
catalyst offered an excellent performance includ-
ing high selectivity and excellent stability.
3.3. Reaction mechanism
2
,3-pentanedione have also been proposed, and shown in
Unlike the mechanism on dehydration of LA into acrylic acid
which has been discussed by many previous reports [50,51,59],
Scheme 1 and Scheme 2 in Supporting Information.
4
Scheme 1. Proposed mechanism for acetaldehyde formation from LA over Mg0.388Al2.408O catalyst.

C. Tang et al. / Journal of Catalysis 329 (2015) 206–217
217
[
[
16] V. Toan Van, H. Kosslick, A. Schulz, J. Harloff, E. Paetzold, J. Radnik, U. Kragl, G.
Fulda, C. Janiak, T. Nguyen Dinh, Micropor. Mesopor. Mater. 177 (2013) 135.
17] T.T. Adint, G.W. Wong, C.R. Landis, J. Org. Chem. 78 (2013) 4231.
4
. Conclusions
Magnesium aluminum oxides prepared with co-precipitation
[18] J.F. Zhang, Y.L. Zhao, X.Z. Feng, M. Pan, J. Zhao, W.J. Ji, C.T. Au, Catal. Sci.
Technol. 4 (2014) 1376.
[
[
[
method were firstly used to catalyze the decarbonylation of LA to
acetaldehyde. Formation of magnesium aluminum oxides is very
sensitive to pH values. Low pH values are favorable to the forma-
19] J.S. Peng, X.L. Li, C.M. Tang, W. Bai, Green Chem. 16 (2014) 108.
20] J. Yan, D.H. Yu, H. Li, P. Sun, H. Huang, J. Rare Earth 28 (2010) 803.
21] P. Sun, D.H. Yu, Z.C. Tang, H. Li, H. Huang, Ind. Eng. Chem. Res. 49 (2010) 9082.
tion of Mg0.388Al2.408
O
4
while high pH values are favorable to the
spinel. The crystallinity of Mg0.388Al2.408
increases with an increase of calcination temperature in range of
50–1000 °C. Among magnesium aluminum oxides,
[22] T.M. Aida, A. Ikarashi, Y. Saito, M. Watanabe, R.L. Smith, K. Arai, J. Supercrit.
Fluid 50 (2009) 257.
formation of MgAl
2
O
4
O
4
[
[
23] J.F. Zhang, Y.L. Zhao, M. Pan, X.Z. Feng, W.J. Ji, C.T. Au, ACS Catal. 1 (2011) 32.
24] D.C. Wadley, M.S. Tam, P.B. Kokitkar, J.E. Jackson, D.J. Miller, J. Catal. 165
(1997) 162.
5
[
[
[
[
[
[
25] G.C. Gunter, R.H. Langford, J.E. Jackson, D.J. Miller, Ind. Eng. Chem. Res. 34
Mg0.388Al2.408
Mg0.388Al2.408
O
4
offers the best catalytic performance. For
catalyst, LA conversion increases drastically with
(
1995) 974.
O
4
26] T.J. Korstanje, H. Kleijn, J.T.B.H. Jastrzebski, R.J.M.K. Gebbink, Green Chem. 15
(2013) 982.
27] S. Lomate, T. Bonnotte, S. Paul, F. Dumeignil, B. Katryniok, J. Mol. Catal. A-
Chem. 377 (2013) 123.
28] M. Dusselier, P. Van Wouwe, A. Dewaele, E. Makshina, B.F. Sels, Energy
Environ. Sci. 6 (2013) 1415.
29] Y.L. Wang, W.P. Deng, B.J. Wang, Q.H. Zhang, X.Y. Wan, Z.C. Tang, Y. Wang, C.
Zhu, Z.X. Cao, G.C. Wang, H.L. Wan, Nat. Commun. 4 (2013) 2141.
30] D. Esposito, M. Antonietti, Chemsuschem 6 (2013) 989.
an increase of reaction temperature while acetaldehyde selectivity
decreases slowly. Effect of LA LHSV on reaction performance sug-
gests that acetaldehyde selectivity increases with an increase of
LA LHSV. Decarbonylation reaction of LA over Mg0.388Al2.408
alyst is not sensitive to LA concentrations. Under the optimal reac-
tion conditions, the stability of Mg0.388Al2.408 catalyst was
evaluated at LA LHSV = 13.0 h . Encouragingly, LA conversion
reduced with only 25% (from 95% to 70%) within 400 h on stream.
More importantly, acetaldehyde selectivity almost remained con-
stant (>93%) during the whole time on stream.
4
O cat-
O
4
ꢀ1
[31] C.A. Ramirez-Lopez, J.R. Ochoa-Gomez, S. Gil-Rio, O. Gomez-Jimenez-
Aberasturi, J. Torrecilla-Soria, J. Chem. Technol. Biotechnol. 86 (2011) 867.
[
32] Z.J. Zhai, X.L. Li, C.M. Tang, J.S. Peng, N. Jiang, W. Bai, H.J. Gao, Y.W. Liao, Ind.
Eng. Chem. Res. 53 (2014) 10318.
[
33] C.M. Tang, J.S. Peng, X.L. Li, Z.J. Zhai, W. Bai, N. Jiang, H.J. Gao, Y.W. Liao, Green
Chem. 17 (2015) 1159.
[
[
34] H. Klym, A. Ingram, I. Hadzaman, O. Shpotyukde, Ceram. Int. 40 (2014) 8561.
35] L. Lallemant, J. Petit, S. Lalanne, S. Landais, S. Trombert, L. Vernhet, R.
Viroulaud, J. Eur. Ceram. Soc. 34 (2014) 791.
Acknowledgments
[
[
36] P. Orosco, L. Barbosa, M.D. Ruiz, Mater. Res. Bull. 59 (2014) 337.
37] F.F. Canova, A.S. Foster, M.K. Rasmussen, K. Meinander, F. Besenbacher, J.V.
Lauritsen, Nanotechnology 23 (2012) 1.
This work was supported by Scientific Research Fund of
Chemical Synthesis and Pollution Control Key Laboratory of
Sichuan Province with project number of CSPC-2014-3-1,
Scientific Research Fund of China West Normal University with
project number of 12B019 and Scientific Research Fund of
Sichuan Provincial Educational Department with project number
of 14ZA0128.
[38] Z. Mosayebi, M. Rezaei, N. Hadian, F.Z. Kordshuli, F. Meshkani, Mater. Res. Bull.
7 (2012) 2154.
[
[
[41] G.D.B. Nuernberg, E.L. Foletto, L.F.D. Probst, C.E.M. Campos, N.L.V. Carreno,
M.A. Moreira, Chem. Eng. J. 193 (2012) 211.
[
[
[
[
[
4
39] F.S. Al-Hazmi, W.E. Mahmoud, J. Eur. Ceram. Soc. 34 (2014) 3047.
40] B. Ismail, S.T. Hussain, S. Akram, Chem. Eng. J. 219 (2013) 395.
42] M.Y. Nassar, I.S. Ahmed, I. Samir, Spectrochim. Acta A 131 (2014) 329.
43] I. Ganesh, Int. Mater. Rev. 58 (2013) 63.
44] X.L. Chen, X.Q. Cao, B.L. Zou, J. Gong, C. Sun, J. Eur. Ceram. Soc. 35 (2015) 227.
45] J.A. Wang, C.L. Li, Appl. Surf. Sci. 161 (2000) 406.
46] C. Kienle, C. Schinzer, J. Lentmaier, O. Schaal, S. KemmlerSack, Mater. Chem.
Phys. 49 (1997) 211.
47] N.P. Hua, H.T. Wang, Y.K. Du, M. Shen, P. Yang, Catal. Commun. 6 (2005) 491.
48] R.J. Rennard, J. Freel, J. Catal. 98 (1986) 235.
49] J.F. Zhang, J.P. Lin, P.L. Cen, Can. J. Chem. Eng. 86 (2008) 1047.
50] B. Yan, L.Z. Tao, Y. Liang, B.Q. Xu, ACS Catal. 4 (2014) 1931.
51] B. Yan, L.Z. Tao, Y. Liang, B.Q. Xu, Chemsuschem 7 (2014) 1568.
52] J.F. Zhang, X.Z. Feng, Y.L. Zhao, W.J. Ji, C.T. Au, J. Ind. Eng. Chem. 20 (2014) 1353.
53] C.M. Tang, J.S. Peng, X.L. Li, Z.J. Zhai, N. Jiang, W. Bai, H.J. Gao, Y.W. Liao, RSC
Adv. 4 (2014) 28875.
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.jcat.2015.05.016.
[
[
[
[
[
[
[
References
[
[
1] P. Liu, E.J.M. Hensen, J. Am. Chem. Soc. 135 (2013) 14032.
2] C. Caro, K. Thirunavukkarasu, M. Anilkumar, N.R. Shiju, G. Rothenberg, Adv.
Synth. Catal. 354 (2012) 1327.
[
[
54] C.M. Tang, J.S. Peng, G.C. Fan, X.L. Li, X.L. Pu, W. Bai, Catal. Commun. 43 (2014)
[
[
3] I. Abdullahi, T.J. Davis, D.M. Yun, J.E. Herrera, Appl. Catal. A-Gen. 469 (2014) 8.
4] J.C. Bauer, G.M. Veith, L.F. Allard, Y. Oyola, S.H. Overbury, S. Dai, ACS Catal. 2
231.
55] M. Behrens, F. Studt, I. Kasatkin, S. Kuhl, M. Havecker, F. Abild-Pedersen, S.
Zander, F. Girgsdies, P. Kurr, B.L. Kniep, M. Tovar, R.W. Fischer, J.K. Norskov, R.
Schlogl, Science 336 (2012) 893.
56] M.S. Holm, S. Saravanamurugan, E. Taarning, Science 328 (2010) 602.
57] C. Li, B. Wang, Q.Q. Zhu, T.W. Tan, Appl. Catal. A-Gen. 487 (2014) 219.
58] B. Wang, C. Li, Q.Q. Zhu, T.W. Tan, RSC Adv. 4 (2014) 45679.
59] V.C. Ghantani, S.T. Lomate, M.K. Dongare, S.B. Umbarkar, Green Chem. 15
(
2012) 2537.
[
5] J.J. Murcia, M.C. Hidalgo, J.A. Navio, V. Vaiano, P. Ciambelli, D. Sannino, Catal.
Today 196 (2012) 101.
[
[
[
[
[
[
[
[
6] Y. Nakamura, T. Murayama, W. Ueda, Chemcatchem 6 (2014) 741.
7] Y. Guan, E.J.M. Hensen, J. Catal. 305 (2013) 135.
8] B. Katryniok, S. Paul, F. Dumeignil, Green Chem. 12 (2010) 1910.
9] C.M. Tang, Y. Zeng, X.G. Yang, Y.C. Lei, G.Y. Wang, J. Mol. Catal. A-Chem. 314
(
2013) 1211.
(
2009) 15.
[
[
60] J.M. Sun, K.K. Zhu, F. Gao, C.M. Wang, J. Liu, C.H.F. Peden, Y. Wang, J. Am. Chem.
Soc. 133 (2011) 11096.
[
[
[
10] C.M. Tang, Y. Zeng, P. Cao, X.G. Yang, G.Y. Wang, Catal. Lett. 129 (2009) 189.
11] A. Brennfuhrer, H. Neumann, M. Beller, Chemcatchem 1 (2009) 28.
12] A. Scrivanti, V. Beghetto, M. Zanato, U. Matteoli, J. Mol. Catal. A-Chem. 160
61] L. Deiana, Y. Jiang, C. Palo-Nieto, S. Afewerki, C.A. Incerti-Pradillos, O. Verho,
C.W. Tai, E.V. Johnston, A. Cordova, Angew. Chem. Int. Ed. 53 (2014) 3447.
62] X.J. Jin, K. Yamaguchi, N. Mizuno, Angew. Chem. Int. Ed. 53 (2014) 455.
63] A. Vjunov, M.Y. Hu, J. Feng, D.M. Camaioni, D.H. Mei, J.Z. Hu, C. Zhao, J.A.
Lercher, Angew. Chem. Int. Ed. 53 (2014) 479.
(
2000) 331.
[
[
[
[
[
13] C.M. Tang, X.L. Li, G.Y. Wang, Korean J. Chem. Eng. 29 (2012) 1700.
14] K. Xu, X. Zheng, Z. Wang, X. Zhang, Chem.-Eur. J. 20 (2014) 4357.
15] D.G. Hanna, S. Shylesh, P.A. Parada, A.T. Bell, J. Catal. 311 (2014) 52.