Dehydration of lactic acid to acrylic acid over lanthanum phosphate catalysts: The role of Lewis acid sites

Article abstract of DOI:10.1039/c6cp04163a

Lanthanum phosphate (LaP) nano-rods were synthesized using n-butylamine as a shape-directing agent (SDA). The resulting catalysts were applied in the dehydration of lactic acid to acrylic acid. Aiming to understand the nature of the active sites, the chemical and physical properties of LaP materials were studied using a variety of characterization techniques. This study showed that the SDA not only affected the porosity of the LaP materials but also modified the acid-base properties. Clearly, the modification of the acid-base properties played a more critical role in determining the catalytic performance than porosity. An optimized catalytic performance was obtained on the LaP catalyst with a higher concentration of Lewis acid sites. Basic sites showed negative effects on the stability of the catalysts. Good stability was achieved when the catalyst was prepared using the appropriate SDA/La ratio.

Full text of DOI:10.1039/c6cp04163a

PCCP
View Article Online
View Journal
PAPER
Dehydration of lactic acid to acrylic acid over
lanthanum phosphate catalysts: the role of
Lewis acid sites†
Cite this:DOI: 10.1039/c6cp04163a
Zhen Guo, De Sheng Theng, Karen Yuanting Tang, Lili Zhang, Lin Huang,
Armando Borgna and Chuan Wang*
Lanthanum phosphate (LaP) nano-rods were synthesized using n-butylamine as a shape-directing agent
(SDA). The resulting catalysts were applied in the dehydration of lactic acid to acrylic acid. Aiming to
understand the nature of the active sites, the chemical and physical properties of LaP materials were
studied using a variety of characterization techniques. This study showed that the SDA not only affected
the porosity of the LaP materials but also modified the acid–base properties. Clearly, the modification
of the acid–base properties played a more critical role in determining the catalytic performance than
porosity. An optimized catalytic performance was obtained on the LaP catalyst with a higher concentration
of Lewis acid sites. Basic sites showed negative effects on the stability of the catalysts. Good stability was
achieved when the catalyst was prepared using the appropriate SDA/La ratio.
Received 22nd June 2016,
Accepted 3rd August 2016
DOI: 10.1039/c6cp04163a
www.rsc.org/pccp
biomass and is becoming increasingly available in commercial
1
. Introduction
1
,11
quantities at low cost.
The catalytic dehydration of LA to AA has been studied using
but also provide opportunities for the conventional chemical both homogeneous and heterogeneous catalysts. With Na HPO
industry. Replacing fossil fuels with renewable feedstock is one and/or NaOH as homogenous catalysts, a yield of AA as high as
Increasing environmental concerns not only raise challenges
2
4
12
of the best solutions to mitigate those concerns. In line with 58% was reported. But the drawbacks of homogeneous catalysts
this, transformation of biomass to value-added chemicals has are their unfeasible separation and recovery. Therefore, signifi-
attracted significant attention in recent years. As renewable cant efforts were devoted to the development of heterogeneous
feedstock, biomass derived materials, mainly cellulose and catalysts. The most frequently used heterogeneous catalysts are
6
,8,10,13,14
5,15
7,9,16,17
lignocellulose, have been converted to platform chemicals such modified zeolites,
sulfates
and phosphates.
10
as furfural, 5-hydroxymethylfurfural (HMF), levulinic acid, and Sodium Y zeolites (NaY) have been modified by alkali metals,
lactic acid, among others. These platform chemicals can be alkaline earth metals and rare earth metals to improve the
further upgraded to fuels, high value chemicals, monomers and selectivity towards AA. 67.9% selectivity toward AA was achieved
polymers. Thus, the future chemical industry can be released on potassium iodide modified NaY zeolites (KI/NaY). In
from the full dependence on fossil fuels. One good example is addition, the stability of NaY zeolites was improved by using
the production of acrylic acid (AA), one of the important Na HPO as a modifier. 77% selectivity toward AA was reported
1
–4
13
8
3
,4
10
6
2
4
18
commodity chemicals with an estimated annual production using alkali metal modified ZSM-5 zeolites. BaSO showed an
4
1
capacity of 4 200 000 metric tons. AA and its alkyl esters are of efficient activity for the dehydration of LA. 74% selectivity
1
5
great industrial importance in the production of polyacrylic towards AA was obtained under optimal conditions. CaSO
4
5
esters, super-absorbent materials, washing reagents, paints, and doped with phosphates gave 63.7% yield. Most recently, high
1,5
adhesives.
Currently AA is manufactured through two-step selectivity toward AA (87.0%) has been reported by Matsuura
1
19
oxidation of propylene, which is a nonrenewable petrochemical.
et al., using hydroxyapatite catalysts. The deactivation mecha-
Recently, tremendous efforts have been devoted to the production nism of zeolite based catalysts was studied by N ¨a fe et al. They
of AA from renewable bio-resources, e.g. catalytic dehydration suggested that the deactivation was caused by the deposition of
6
–10
20
of lactic acid (LA).
LA is obtained from the fermentation of species containing carboxyl groups on the catalysts.
The catalytic mechanism of LA dehydration has been extensively
studied. It is believed that the acid–base properties of the catalyst
play an essential role in the selectivity toward AA. It has been
Institution of Chemical and Engineering Sciences, 1 Pesek Road, Jurong Island,
Singapore 627833. E-mail: wang_chuan@ices.a-star.edu.sg
†
Electronic supplementary information (ESI) available. See DOI: 10.1039/c6cp04163a proposed that Brønsted acid sites catalyze the decarbonylation
This journal is ©the Owner Societies 2016
Phys. Chem. Chem. Phys.

View Article Online
Paper
and/or decarboxylation of LA, leading to the formation of into the mixture of NH
PCCP
4
H
2
PO
4
and the SDA with vigorous stirring
1
5,21
acetaldehyde and thus poor selectivity toward AA.
However, at room temperature. A white solid precipitated immediately, and
the role of Lewis acids and bases during this dehydration the slurry was kept stirring for 24 h. The precipitate was recovered
process is still under debate. The high performance of NaY by centrifugation, washed with deionized water copiously and then
zeolites modified with rare earth metals (La, Ce, Sm and Eu) was dried at 60 1C overnight. Finally, the solid was calcined at 550 1C
8
explained by the higher content of weak acid sites. Whereas the for 4 h. The mole ratio of SDA/La was varied from 3 to 6.
enhanced selectivity toward AA obtained on barium doped NaY Accordingly, LaP catalysts were denoted as LaP-x (x = 3, 4, 5, 6),
zeolites (Ba/NaY) was attributed to its high content of medium where x indicates the SDA/La molar ratio. For the sake of compar-
1
3
basic sites. In the KI/NaY-catalyzed dehydration of LA, it was ison, a LaP sample without the addition of the SDA was also
proposed that dehydration of LA occurs on acid–base pairs, in prepared, and was labelled as LaP-0. To study the porosity effect of
+
which K ions act as Lewis acid sites and the framework oxygen LaP materials on the catalytic dehydration of LA, a modified LaP-3
1
0
plays the role of basic sites. Similar theory was proposed by sample with a relatively low surface area was also prepared. LaP-3
1
8
Yuan et al. for the ZSM-5 based catalysts. In the study on ZSM-5 was modified by loading a small amount of amorphous LaP to
modified by NaOH and Na HPO , Zhang et al. suggested that minimize the specific surface area by blocking its pores. The
medium acid sites and medium basic sites were important for introduction of amorphous LaP was carried out by impregnating
achieving high selectivity toward AA as well as long term LaP-3 with La(NO and NH PO , followed by washing, drying
2
4
3
)
3
4
H
2
4
2
2
stability. However, studies on ion-exchanged b zeolites showed and calcination. The loading of amorphous LaP was 2 wt%. The
that weak acid and weak base sites were desired for the opti- modified LaP-3 was denoted as LaP/LaP-3.
2
3
mized catalysts. For sulfate-based catalysts, weak acid sites
were found to be important to achieve high selectivity toward
2
.2 Catalytic dehydration of LA to AA
1
5
AA. In phosphate catalyzed systems, Miller et al. indicated that The dehydration of LA was carried out using a vertical, down-flow,
the formation of metal lactate as an intermediate is a critical packed bed reactor. 0.2 to 0.5 g of LaP-x were loaded in the middle of
1
6,17
step during the catalytic transformation of LA to AA.
This the reactor. Before each reaction, the loaded catalyst was dried at
idea was further developed by Ghantani and co-workers, who 350 1C for 1 h under a flowing mixture of He/N 5%. Nitrogen
2
suggested that the formation of metal lactate was accompanied was used as the internal standard for the gas phase analysis using
by a proton transfer to the phosphate, leading to in situ genera- online gas chromatography (GC). The flow rate of the carrier gas
7
ꢁ1
tion of P–OH Brønsted acid sites. Therefore, the metal cation, (He/N 5%) was kept at 35 ml min . 20 wt% of an aqueous solution
2
acting as a Lewis acid site, stabilized the negatively charged of LA was fed and vaporized at the top of the reactor along with the
carboxylic group of LA, while the P–OH group was responsible carrier gas. Gas phase products, such as CO, CO and a fraction of
2
for the dehydration process.
acetaldehyde, were determined by an online GC (Shimadzu 2014),
Lanthanum phosphate (LaP) has been reported as a dehydra- whereas the other products were condensed in a stainless steel trap
24,25
tion catalyst.
Rod-shaped nanocrystalline LaP containing a high located at the outlet of the reactor. The liquid products were
density of Lewis acid sites was used as a catalyst for the acetal- analyzed by an offline GC (Agilent 6890). Fouling of spent catalysts
26
formation reaction. However, the catalytic capability of LaP for was estimated from the weight loss after calcination. The carbon
the dehydration of LA has not been explored yet. In this work, we content in spent catalysts was estimated based on the method
28
report on the catalytic performance of LaP catalysts in the selective suggested by Miller et al. The conversion of LA and the selectivity
dehydration of LA to AA. LaP materials with relatively high specific toward AA were calculated based on the following equations:
surface areas were synthesized using n-butylamine as a shape-
fed LA in mole ꢁ unreacted LA in mole
directing agent (SDA). The influence of the organic SDA on the
structural and surface properties of LaP materials is discussed.
Also, a correlation between the acid–base properties of the LaP
catalysts and their catalytic behaviour is reported.
Conversion of LA ¼
fed LA in mole
ꢂ 100%;
Amount of carbon in the product ðmoleÞ
Amount of carbon in fed LA ðmoleÞ
Yield of product ¼
2
. Experimental
ꢂ 100%;
2.1 Synthesis of lanthanum phosphate materials
Amountof carbonintheproductðmoleÞ
All chemicals were obtained from Sigma-Aldrich and used with-
out further purification. Lanthanum phosphate (LaP) materials
Selectivityof product¼
Amountof carboninconvertedLAðmoleÞ
ꢂ100%:
2
7
were prepared based on the method reported by Onoda et al. In
a typical synthesis process, solutions of ammonium dihydrogen
phosphate (NH H PO ) and lanthanum nitrate hexahydrate
4
2
4
2
.3 Characterization
physisorption isotherms were measured at ꢁ196 1C on a static
of the SDA in water. The SDA was added into the solution of volumetric instrument (Autosorb-6b, Quanta Chrome). The specific
(
La(NO ) ꢀ6H O) were prepared separately. A 50 wt% of ethanol
3
3
2
aqueous solution was used as a solvent due to the low solubility
N
2
NH
4
H
2
PO
4
. The dissolved La(NO
3
)
3
was then introduced dropwise surface area was estimated using the Brunauer–Emmett–Teller
Phys. Chem. Chem. Phys.
This journal is ©the Owner Societies 2016

View Article Online
PCCP
Paper
2
9
(
BET) method, and the pore size was calculated by the Barrett–
Joyner–Halenda (BJH) method, using the desorption branch of
3
0
the isotherms. Powder X-ray diffraction (XRD) measurements
were collected under Cu Ka radiation using a Bruker D8 X-ray
diffractometer over a 2y range of 20–801, employing a step size of
0.0171 and a step time of 1 s. Diffuse reflectance infrared Fourier
transform spectroscopy (DRIFTS) spectra were recorded using
Bio-Rad FT-IR 3000 MX equipped with a mercury cadmium
telluride (MCT) detector. Every spectrum was obtained based
ꢁ
1
on 128 scans with a resolution of 4 cm . Before each measure-
ment, the sample was pre-treated at 200 1C in helium for 1 h to
remove surface moisture and contaminants. Acid sites were
evaluated using temperature programmed desorption of NH3.
3
In a typical NH -TPD experiment, 0.2 g of sample was pre-treated
in helium flow at 400 1C for 2 h. Then, the temperature was
lowered to 100 1C, and the sample was treated with a flowing gas
Fig. 1 The XRD patterns of LaP-x samples prepared using different
mixture of He/NH
helium to remove physisorbed NH
then carried out from 100 to 450 1C using a heating rate of
3
5%. Afterwards, the sample was purged with SDA/La ratios.
3
. TPD measurements were
ꢁ
1
5
1C min . Transmission electron microscopy (TEM) images
were captured using a Tecnai F20 equipped with a field emission
electron gun and operated at 200 kV and a TEM instrument
operated at 200 kV. X-ray photoelectron spectroscopy (XPS)
spectra were recorded on a XPS spectrometer (Thermo VG
Scientific ESCALAB 250), using Al Ka (1486.6 eV). IR spectra
of adsorbed pyridine on the catalysts were measured on a
BIO-RAD spectrometer. Solid samples were pressed into a wafer
and loaded in an IR cell. The wafer was treated at 300 1C under
vacuum for 30 min. The background spectrum was recorded
under vacuum at room temperature (RT), before collecting
the spectra of samples. The wafer was then exposed to a vapor
of dried and purified pyridine at RT for 30 min, followed by
evacuation for 30 min at 22, 100, 200 and 300 1C, respectively.
All spectra were recorded under vacuum at RT.
Fig. 2 (a) N
2
physisorption isotherms and (b) pore size distributions
of LaP-x samples.
3
. Results and discussion
2
ꢁ1
3.1 Characterization of lanthanum phosphate (LaP)
area of LaP-3 is around 179.1 m g , this value being higher
than that reported for lanthanum phosphate materials prepared
using cetyl trimethylammonium bromide (CTAB) as an SDA
The crystallinity of LaP-x prepared with different amounts of
SDA was examined by powder XRD. As shown in Fig. 1, all LaP
materials show the typical XRD pattern of rhabdophane type
LaP. With increasing amounts of SDA, a new crystalline phase
becomes more and more noticeable, which can be attributed to
the formation of a new type of LaP mixed with a minor quantity
of lanthanum oxide (La
SDA concentration implies that the SDA may hinder the inter-
action of La and P–O during precipitation, resulting in the
segregation of crystalline phases and the formation of La
Generally, high surface area catalysts are preferred in hetero-
geneous catalysis, due to their increased density and enhanced Table 1 Texture properties of LaP-x catalysts
2
ꢁ1
24
(
123.0 m g ) using the same SDA/La molar ratio of LaP-3.
3
1
However, a further increase in the amount of SDA results in a
decrease of the specific surface area of the LaP materials. When a
SDA/La ratio = 3 was used during the synthesis, the pore size
distribution is very uniform and the average pore size is around
.8 nm (see Fig. 2b). However, increasing the amount of SDA leads
to a less uniform pore-size distribution, along with an increase in
3
1,32
2
O
3
).
2 3
The existence of La O at high
4
3
+
ꢁ
2 3
O .
accessibility of active sites. The porosity of LaP-x samples was
2
ꢁ1
3
ꢁ1
Catalysts Surface area (m g ) Pore volume (cm g ) Pore size (nm)
examined by using N physisorption. As shown in Fig. 2a, both
2
LaP-3
LaP-4
LaP-5
179.1
175.5
160.3
153.5
0.21
0.28
0.43
0.58
4.0
6.1
9.6
the isotherm shape and hysteresis loop of the LaP catalysts
change with the increasing amount of SDA, suggesting modifica-
tion of the pore shapes. As shown in Table 1, the specific surface LaP-6
12.8
This journal is ©the Owner Societies 2016
Phys. Chem. Chem. Phys.

View Article Online
Paper
PCCP
the average pore size (see Fig. 2b, LaP-4, LaP-5 and LaP-6). These the interaction between the SDA and LaP particles during the
results may also serve as evidence of the specific interaction growth of LaP nano-rods. The aggregation degree decreases with
between the SDA and LaP, which promotes the formation of the amount of SDA used, possibly because of a higher number of
LaP samples with a high specific surface area.
SDA molecules surrounding the surface of the LaP nano-rods
TEM was employed to further study the structure of the at higher concentrations of the SDA, hence hindering the
LaP-x samples. Fig. 3 shows that all LaP materials are in the form agglomeration of LaP nano-rods during precipitation. The SDA
of aggregated nano-rods. Generally, these nano-rods are 20 to may react with residual P–OH groups on the surface, giving rise
30 nm long and 5 to 10 nm wide. No well-defined meso- to the formation of ion pairs. The SDA may also be coordinated
pores are observed. Therefore, the mesopores detected by N with the surface lanthanum cations located at defect sites of the
2
physisorption measurement can be attributed to the formation crystals. Similar interactions between templates and calcium
of pores among the intercalated nano-rods. Based on the hydroxyapatite (CaHAP) nanoparticles have also been reported
3
3
characterization results reported above, the changes in pore by Tanaka et al. Mesopores, generated by the aggregation of
size distributions of the LaP materials could be explained by CaHAP nanoparticles, were also found after the removal of the
template. In line with this assumption, the surface properties of
these LaP materials could also be modulated by the amount of
SDA, resulting in changes of crystallinity and porosity. Usually,
the surface properties of the catalytic materials determine their
catalytic performance.
The surface composition of LaP-x was characterized by XPS.
The XPS spectra of La, P and O are shown in Fig. 4a–c, respectively.
By comparing the peaks assigned to La 3d5/2, a negative shift is
observed for the catalysts prepared using the SDA, which may
3
4
be due to the formation of La
binding energy, a change in the intensity ratio of the satellite
peak to the main peak (I /I ) is also observed. As shown in
Table 2, this I /I ratio gradually increases with the amount of
SDA. The satellite peak at higher binding energy is attributed to
2
O
3
.
Besides the change in
s
m
s
m
9
1
35
the O 2p–La 4f charge transfer ending in a 3d 4f final state.
The increase of the I /I ratio also indicates the formation
s
m
3
5
of La
2
O
3
.
P and O peaks do not show any observable shift.
Based on our XPS data and tabulated sensitivity factors, semi-
quantitative analyses were performed to study the chemical
composition on the surface of LaP-x samples. To minimize
systematic errors, atomic ratios of La/P and O/P instead of
absolute compositions are listed in Table 2.
Fig. 3 TEM images of LaP-x samples prepared using different SDA/La
ratios.
Fig. 4 XPS spectra of LaP-x: (a) La 3d5/2, (b) P 2p, (c) O 1s.
Phys. Chem. Chem. Phys.
This journal is ©the Owner Societies 2016

View Article Online
PCCP
Paper
Table 2
I
s
/I
m
ratio of La 3d5/2 and surface atomic ratio derived from XPS
3
Table 3 Density of acid sites derived from deconvolution of NH -TPD
measurements
peaks
Atomic ratio
Total acid
Weak acid
Medium acid
Strong acid
ꢁ
1
ꢁ1
ꢁ1
ꢁ1
Catalysts
(mmol g
)
(mmol g
)
(mmol g
)
(mmol g
)
Catalysts
s
I /I
m
La/P
O/P
LaP-3
LaP-4
LaP-5
LaP-6
670
603
444
375
328
277
244
236
235
211
147
113
107
115
53
LaP-3
LaP-4
LaP-5
LaP-6
0.79
0.81
0.83
0.86
0.52
0.56
0.66
0.71
2.25
2.46
2.73
2.87
26
To further understand the acid–base properties of the LaP-x
Both the La/P and O/P ratios increase with the amount of samples, DRIFTS was also used to characterize the surface
SDA used during the synthesis, which confirms the formation species. Before measurements, all samples were heated up to
of La
2 3 2
O on the surface of LaP nano-rods. These results are in 200 1C and purged with pure N for 2 h, in order to remove
good agreement with XRD results, suggesting the interaction of physically adsorbed water. In the region of OH group vibra-
ꢁ
1
the SDA and LaP particles. This specific interaction promotes tions, ranging from 2900 to 3800 cm (see Fig. 6a), the main
ꢁ1
the formation of La
LaP nano-rods, resulting in surface defects.
2
O
3
and possibly impairs the crystallinity of peak is observed at 3510 cm along with two shoulders that
can be detected on each side for all samples. To examine the
The acid–base properties of the catalysts play a critical role in differences among these signals, each broad absorption band
the catalytic dehydration of LA to AA. NH -TPD was employed was deconvoluted into three peaks. The small peak centered at
3
ꢁ
1
to evaluate the acidity of LaP-x materials. As shown in Fig. 5a, 3680 cm can be assigned to OH groups bonded to phosphorous
broad NH desorption bands can be observed for all LaP-x atoms (i.e., P–OH groups, see red dashed lines), while the main
samples between 100 and 400 1C, which evidences the presence peak centered at 3500 cm is due to OH ions in the lattice (blue
3
ꢁ1
ꢁ
2
6
33
ꢁ1
of acidic sites on all phosphate samples. The TPD signal was dashed lines). The broad band centered at 3280 cm (yellow
calibrated and the density of acidic sites was calculated based on dashed lines) can be assigned to strongly adsorbed water which
3
9
the area of the peaks. The acid density decreases monotonically can only be efficiently removed above 400 1C.
with the increasing amount of SDA, with a significant drop in
ꢁ1
The absorption at 3680 cm is negligible for all LaP-x samples,
acidity when the SDA/La ratio is larger than 4. These broad suggesting a very low amount of P–OH groups on surfaces. The
bands are further deconvoluted into three peaks in order to intensity of this absorption signal decreases with the increase of
study the contribution of acid sites with different strengths. The the SDA. This signal of P–OH groups almost disappears in the
concentration of acid sites with different strengths, derived from spectrum of LaP-6. The absence of P–OH groups on LaP-5 and
the deconvolution of the TPD curves, is reported in Table 3. By LaP-6 could be attributed to the interaction between the SDA and
plotting the relative density of acid sites versus the SDA/La ratio LaP surfaces, which promotes the formation of La
2 3
O , and thus
(
Fig. 5b), it is clear that all types of acid sites diminish with the P–OH groups may be covered by La . This assumption is
2 3
O
increase in the SDA, but the strong acid sites decline much supported by the XPS results, which indicate the formation of
faster, followed by the acid sites with medium strength, while La O on the surfaces when the concentration of SDA increases.
2
3
the decrease of weak acid sites is the slowest. However, the type In addition, the formation of La O is also evidenced by DRIFTS
2
3
ꢁ
1
of acid sites (Brønsted or Lewis acid sites) cannot be determined measurements. Absorption bands between 1400 and 1600 cm
,
3
6–38
from the NH
3
-TPD results.
assignable to v3 carbonate modes, are found on LaP catalysts
3
Fig. 5 (a) NH -TPD of LaP-x samples and density of acid sites; (b) relative
density of acid sites with different strength versus the SDA/La ratio used
during synthesis.
Fig. 6 DRIFTS spectra of LaP-x prepared with different SDA/La ratios.
This journal is ©the Owner Societies 2016
Phys. Chem. Chem. Phys.

View Article Online
Paper
PCCP
4
0
prepared using a SDA. The intensity of this peak increases with
the addition of the SDA. These carbonate species may result
from the adsorption of CO from the atmosphere at strong basic
2
4
1
sites of LaP-x, namely the La O deposited on the surfaces.
2
3
It is important to note that while acid sites are detected by
NH -TPD, DRIFTS measurements suggest a very low content of
3
Brønsted acidic sites (P–OH) on catalyst surface. Therefore, it
can be postulated that the acid sites on LaP prepared using SDA
(sample LP-3 to LP-6) are mainly Lewis acid sites. To confirm
this, FTIR measurements of adsorbed pyridine were carried
out to examine the nature of the acid sites of these samples.
ꢁ
1
The clear peak centred around 1443 cm can be assigned to
4
2
the adsorbed pyridine on Lewis acids. While the signal at
ꢁ
1
1
484 cm could be attributed to both Lewis and Brønsted
4
3
acids, the signal corresponding to the existence of only
Brønsted acidic sites (B1540 cm ) is absent for all samples.
ꢁ1
Fig. 7 FTIR spectra of adsorbed pyridine after evacuation at 100 1C on
LaP-3 and LaP-4 show strong signals of Lewis acids, while the LaP-x samples.
peak of Lewis acid sites becomes very weak for LaP-5. No
adsorbed pyridine can be detected on LaP-6 after evacuation at
1
00 1C. The pyridine adsorption results may also be explained by This could possibly be due to the smaller molecular size of NH
3
the interaction of the SDA and phosphate during the synthesis of compared to that of pyridine. As shown in Fig. 5b, the decreasing
these catalysts. It can be postulated that these acid sites are rate of the amount of acid sites with different strengths is in the
3
+
36
envisaged as La cations of LaPO
4
or La
2 3
O .
When a SDA/La following order: strong 4 medium 4 weak. Possibly because
3
+
ratio of 3 was used, the effect of the SDA on the crystallization of the strong Lewis acidic sites, i.e., La ions located at defects, are
LaPO
4
is small; thus almost all La ions are balanced with more unstable and easily converted to La O .
2 3
, and very few P–OH groups are formed. Thus, the Lewis
acidity of LaP-3 is mainly attributed to the La ions in LaPO4 a catalyst with a higher concentration of Brønsted acids was
3
+
3ꢁ
PO
4
In order to get more insight into the effect of Brønsted acids,
3
+
crystals. However, the broad band of NH -TPD suggests the also prepared without the addition of the SDA. This sample was
3
existence of a spectrum of acid sites with different strengths denoted as LaP-0. This was inspired by the DRFTS results,
on the surface of the LaP samples (see Fig. 5a). This may be where the signal of P–OH groups is inversely proportional to
attributed to the heterogeneity of the LaP crystallinity. For the amount of SDA used. As shown in Fig. S1 in the ESI,† the
3
+
example, the charge of a La ion on a well-defined LaP crystal DRIFTS signal of P–OH groups on LaP-0 is significantly larger
3
ꢁ
is balanced by PO
4
. Therefore, the Lewis acidity of these than that of LaP-3, suggesting the presence of a higher density
3
+
3+
charge-balanced La ions is weak. While La ions residing at of Brønsted acids. The distinct acidity of LaP-0 may arise from
3
6,44
the defects of LaP crystals may exhibit strong Lewis acidity.
the low pH value (B1) during the synthesis of this sample.
When the SDA/La ratio is higher than 4, the crystallinity
on the surface of LaP nano-rods may be disturbed by the
3.2 Dehydration of lactic acid (LA) to acrylic acid (AA)
additional amount of SDA via electrostatic interactions. This The catalytic performance of LaP-x samples was tested in the
3
+
may lead to the segregation of a small amount of La cations, dehydration of LA to AA. Table 4 summarizes the catalytic results.
which eventually leads to the deposition of La on the surface Among these catalysts, LaP-3 shows the highest selectivity toward
of LaP nano-rods after calcination. The formation of La is AA, while the highest conversion is obtained on LaP-4. Further
supported by the XPS and DRIFTS results. These La species increasing the amount of SDA during synthesis leads to poorer
could cover the Lewis acid sites on the surface of the samples catalytic results and serious fouling.
2 3
O
2 3
O
2 3
O
when high SDA/La ratios are used, leading to a decrease in the
For the sake of comparison, a sample with higher Brønsted
density of acid sites (see Fig. 5a). The presence of SDA may also acidity was also tested (LP-0, see the ESI,† Table S1). A very high
promote the formation of defects on the surface of LaP-4, and conversion is obtained on LaP-0, while the selectivity to AA is
thus more strong acid sites on LaP-4 are formed, which may about 10% lower than that of LaP-3. As Ghantani et al.
7
explain the slight increase in strong acid sites (see Fig. 5b) suggested, Brønsted acids (P–OH groups on phosphates) could
when the SDA/La ratios are increased from 3 to 4. The increase facilitate decarbonylation, resulting in the formation of acet-
3
+
in defects may also improve the accessibility of La cations. aldehyde, which undermines the selectivity towards AA.
As a result, the signal of pyridine adsorbed onto Lewis acids in The catalytic performance is usually determined by both
LaP-4 is higher than that of LaP-3 (see Fig. 7). A further increase physical and chemical properties. In the LaP system, the different
of the amount of SDA results in the deposition of excess La O behaviours on the dehydration of LA could be attributed to the
which covers the acid sites and hinders the adsorption of structural and surface properties. From N physisorption analysis
pyridine. NH -TPD of LaP-5 and LaP-6 shows a desorption peak (Table 1), LaP-3 exhibits the highest surface area, along with the
2
3
2
3
which indicates that NH
3
are adsorbed on LaP-5 and LaP-6. highest selectivity toward AA (Table 4). To understand the effect
Phys. Chem. Chem. Phys.
This journal is ©the Owner Societies 2016

View Article Online
PCCP
Paper
Table 4 Catalytic results of the dehydration of lactic acid to acrylic acid using LaP-x catalysts
Selectivity (%)
a
b
c
Catalysts
Conversion (%)
Acrylic acid
Acetaldehyde
2,3-Pentanedione
Propanoic acid
Others
Fouling
Carbon balance
LaP-3
LaP-4
LaP-5
LaP-6
67.4
72.0
60.4
53.1
63.8
49.6
46.9
27.5
7.9
21.2
25.3
24.2
26.0
25.7
0.2
0.2
0.4
0.4
0.1
3.1
2.2
6.9
7.4
1.4
20.8
14.7
22.8
27.7
18.0
2.8
4.7
9.2
12.9
1.7
97.7
94.0
91.0
82.3
89.6
LaP/LaP-3
42.7
a
All results are based on 5 measurements of time on stream (TOS), reaction temperature: 350 1C, LA concentration: 20 wt%, catalyst loading: 0.2 g.
Others include acetic acid, ethanol, hydroxyacetone and CO. Fouling is estimated from the weight loss after calcination of spent catalysts.
b
c
of surface area a LaP-3 sample with a relatively low surface area
was prepared. The porosity of LaP-3 was undermined by loading a
small amount of amorphous LaP (2 wt%) on the catalyst using an
impregnation method. This sample was labelled as LaP/LaP-3.
Based on N
the surface area (from 179.1 to 109.4 m
2
sorption measurements, a sharp decrease in both
2
ꢁ1
g
) and the pore
3
ꢁ1
volume (from 0.21 to 0.08 m g ) is observed upon impregna-
tion of LaP-3 with additional LaP. However, the significant
change in porosity only has a very limited influence on the
catalytic performance, as shown in Table 4. This indicates
that the porosity of the catalyst is not a determining factor
for LaP-based catalysts.
Hence, the effect of acid and basic sites on the dehydration
of LA over LaP was evaluated. Based on the characterization of
the acid–base properties and the catalytic results, it is clear that
Lewis acid sites are crucial to obtain a high yield of AA. The
positive effect of Lewis acids has also been reported in the
KI/NaY-catalyzed dehydration of LA. The enhancement of
selectivity was related to the weakened Lewis acidity due to
Fig. 8 Correlation between the density of strong Lewis acid sites and the
catalytic performance.
+
10
the presence of K . A similar conclusion was also drawn for
1
3
3+
LA dehydration over Ba/NaY. In the present case, the La
cations of LaPO
between the catalytic performance and the amount of Lewis an adjacent P–OH group or another LA.
while the P–OH group catalyzes the dehydration of the adsorbed
4
act as Lewis acids. Correlations can be found LA to AA. Finally, the AA is desorbed by accepting a proton from
acid sites with different strengths. A good linear correlation
From the viewpoint of application, stability is an important
between the catalytic performance and the amount of strong attribute in industrial applications. A long-term catalytic dehydra-
acid sites is shown in Fig. 8, indicating that these acid sites play tion test of LA was carried out using the most promising catalyst,
a key role in the selective dehydration of LA to AA.
LaP-3. These results are plotted in Fig. 9. A stabilization period
Basic sites lead to by-products and fouling. The carbonac- of roughly 5 h is required to achieve steady-state conditions.
eous species deposited on the catalysts can block the active This induction period may cause the adsorption/saturation of
sites, therefore inhibiting the dehydration process. The colour the catalyst using LA and AA. Similar observations have also
1
7
of the catalysts changed from white to light brown, suggesting been reported by others. As shown in Fig. 9, the conversion
that the deactivating carbon deposit consists not only of decreases with the TOS due to fouling. However, the decrease in
carbon-containing species but also other components, possibly conversion becomes slower and the conversion level seems to
2
8
oxygenated species. Thus the carbon balance becomes poor be stable at around 70%, leading to a pseudo-steady state at the
for LaP-6 which contains a higher density of strong base sites. end of the catalytic test. It is interesting to note that the
3
+
As summarized in Scheme 1a, Lewis acid sites, mainly La
selectivity increases gradually from 45% to 52% and seems to
cations of LaPO , play a critical role in the dehydration of LA reach a pseudo-steady state level after 28 h. This is probably due
4
over LaP-x, while strong base sites cause carbon deposition on to the preferential deposition of carbon species on the active
the catalysts. The catalytic mechanism may be similar to that sites that are responsible for the side reactions such as the basic
7
6
proposed for the dehydration of LA over CaHAP. As shown in sites which promote the formation of polylactate. Overall, the
3
+
Scheme 1b, the carboxylic group of LA interacts with La and yield of AA remains around 40%, after the induction period. The
transfers a proton to the phosphate, leading to the in situ long-term stability of LaP-3 can be ascribed to the low density
3
+
formation of a P–OH group. Therefore, La stabilizes the of basic sites, preventing the fast formation of carbonaceous
carboxylic group of LA avoiding the production of acetaldehyde, deposits on the catalyst.
This journal is ©the Owner Societies 2016
Phys. Chem. Chem. Phys.

View Article Online
Paper
PCCP
Acknowledgements
Funding from the ‘‘Cell Factory: Biomass to Chemicals’’
programme of the Science & Engineering Research Council
(SERC), Agency for Science, Technology and Research (A*STAR),
is gratefully acknowledged.
References
1
Y. Fan, C. Zhou and X. Zhu, Catal. Rev.: Sci. Eng., 2009, 51,
93–324.
2
2
D. M. Alonso, S. G. Wettstein and J. A. Dumesic, Chem. Soc.
Rev., 2012, 41, 8075–8098.
3
4
P. Gallezot, Chem. Soc. Rev., 2012, 41, 1538–1558.
M. Besson, P. Gallezot and C. Pinel, Chem. Rev., 2013, 114,
1
827–1870.
J. Zhang, J. Lin and P. Cen, Can. J. Chem. Eng., 2008, 86,
047–1053.
J. Zhang, Y. Zhao, M. Pan, X. Feng, W. Ji and C.-T. Au,
ACS Catal., 2011, 1, 32–41.
V. C. Ghantani, S. T. Lomate, M. K. Dongare and
S. B. Umbarkar, Green Chem., 2013, 15, 1211–1217.
H. Wang, D. Yu, P. Sun, J. Yan, Y. Wang and H. Huang,
Catal. Commun., 2008, 9, 1799–1803.
C. Tang, J. Peng, G. Fan, X. Li, X. Pu and W. Bai, Catal.
Commun., 2014, 43, 231–234.
5
6
7
8
9
1
Scheme 1 (a) Reaction pathways on different acid–base sites of LaP-x
catalysts during the dehydration of LA and (b) proposed mechanism.
1
1
1
1
1
1
1
1
1
1
0 P. Sun, D. Yu, Z. Tang, H. Li and H. Huang, Ind. Eng. Chem.
Res., 2010, 49, 9082–9087.
1 L. Ye, M. S. B. Hudari, Z. Li and J. C. Wu, Biochem. Eng. J.,
2
014, 83, 16–21.
2 P. J. McCrackin and C. T. Lira, Ind. Eng. Chem. Res., 1993, 32,
608–2613.
3 J. Yan, D. Yu, P. Sun and H. Huang, Chin. J. Catal., 2011, 32,
05–411.
4 J. Yan, D. Yu, H. Li, P. Sun and H. Huang, J. Rare Earths,
010, 28, 803–806.
5 J. Peng, X. Li, C. Tang and W. Bai, Green Chem., 2014, 16,
08–111.
2
4
2
1
6 G. C. Gunter, R. Craciun, M. S. Tam, J. E. Jackson and
D. J. Miller, J. Catal., 1996, 164, 207–219.
7 G. C. Gunter, D. J. Miller and J. E. Jackson, J. Catal., 1994,
Fig. 9 Stability test of LaP-3 during the dehydration of LA. Reaction
temperature: 350 1C, catalyst loading: 0.5 g, LA concentration: 20 wt%.
148, 252–260.
8 C. Yuan, H. Liu, Z. Zhang, H. Lu, Q. Zhu and Y. Chen, Chin.
J. Catal., 2015, 36, 1861–1866.
9 Y. Matsuura, A. Onda and K. Yanagisawa, Catal. Commun.,
2014, 48, 5–10.
4
. Conclusions
In the present study, LaP-x samples with a high specific surface
area were prepared using n-butylamine as a SDA. Both meso- 20 G. N ¨a fe, M. A. L ´o pez-Mart ´ı nez, M. Dyballa, M. Hunger,
porous structural and surface properties are affected by varying Y. Traa, T. Hirth and E. Klemm, J. Catal., 2015, 329, 413–424.
the amount of SDA. The porosity of the LaP catalysts shows a 21 B. Katryniok, S. Paul and F. Dumeignil, Green Chem., 2010,
very limited influence on the dehydration of LA to AA. Instead, 12, 1910–1913.
the acid–base properties of the LaP catalysts significantly affect 22 X. Zhang, L. Lin, T. Zhang, H. Liu and X. Zhang, Chem. Eng.
the catalytic performance. A combination of characterization J., 2016, 284, 934–941.
techniques reveals that Lewis acids on LaP materials, mainly 23 B. Yan, A. Mahmood, Y. Liang and B.-Q. Xu, Catal. Today,
3
+
La cations, play an essential role in the catalytic dehydration
2016, 269, 65–73.
of LA to acrylic acid, while basic sites are responsible for 24 K. Ramesh, Y. Goh, C. G. Gwie, C. Jie, T. White and
catalyst deactivation.
A. Borgna, J. Porous Mater., 2012, 19, 423–431.
Phys. Chem. Chem. Phys.
This journal is ©the Owner Societies 2016

View Article Online
PCCP
Paper
2
5 H. Onoda, H. Nariai, A. Moriwaki, H. Maki and I. Motooka, 35 S. Jørgensen, J. A. Horst, O. Dyrlie, Y. Larring, H. Ræder and
J. Mater. Chem., 2002, 12, 1754–1760. T. Norby, Surf. Interface Anal., 2002, 34, 306–310.
6 K. Rajesh, P. Shajesh, O. Seidel, P. Mukundan and 36 K. Mtalsi, T. Jei, M. Montes and S. Tayane, J. Chem. Technol.
2
K. G. K. Warrier, Adv. Funct. Mater., 2007, 17, 1682–1690.
7 H. Onoda and K. Asai, Ceramics, 2012, 56, 102–106.
8 D. C. Wadley, M. S. Tam, P. B. Kokitkar, J. E. Jackson and
D. J. Miller, J. Catal., 1997, 165, 162–171.
Biotechnol., 2001, 76, 128–138.
37 G. Wu, A. Wang, X. Li, Y. Wang and Y. Hu, Chin. J. Catal.,
2010, 31, 1139–1144.
38 N. Fripiat, M.-A. Centeno and P. Grange, Chem. Mater., 1999,
11, 1434–1445.
39 T. Ishikawa, M. Wakamura and S. Kondo, Langmuir, 1989, 5,
140–144.
40 L. Borum and O. C. Wilson Jr, Biomaterials, 2003, 24,
3681–3688.
41 F. Khosrow-poura, M. Aghazadeh, B. Sabourb and S. Dalvandb,
Ceram. Interfaces, 2013, 39, 9491–9498.
42 M. C. Kung and H. H. Kung, Catal. Rev., 1985, 27,
425–460.
2
2
2
3
3
3
3
3
9 S. Brunauer, P. H. Emmett and E. Teller, J. Am. Chem. Soc.,
1938, 60, 309–319.
0 E. P. Barrett, L. G. Joyner and P. P. Halenda, J. Am. Chem.
Soc., 1951, 73, 373–380.
1 K. Ramesh, J. E. Zheng, E. G. Y. Ling, Y.-F. Han and
A. Borgna, J. Phys. Chem. C, 2009, 113, 16530–16537.
2 H. Bojari, A. Malekzadeh, M. Ghiasi, A. Gholizadeh, R. Azargohar
and A. Kumar Dalai, Cryst. Res. Technol., 2013, 48, 355–362.
3 H. Tanaka, A. Yasukawa, K. Kandori and T. Ishikawa,
Langmuir, 1997, 13, 821–826.
43 M. R. Basila, T. R. Kantner and K. H. Rhee, J. Phys. Chem.,
1964, 68, 3197–3207.
4 M. F. Sunding, K. Hadidi, S. Diplas, O. M. Løvvik,
T. E. Norby and A. E. Gunnæs, J. Electron Spectrosc. Relat. 44 V. R. Choudhary, S. A. R. Mulla and B. S. Uphade, Fuel, 1999,
Phenom., 2011, 184, 399–409. 78, 427–437.
This journal is ©the Owner Societies 2016
Phys. Chem. Chem. Phys.