Comparative study of vapour phase glycerol dehydration over different tungstated metal phosphate acid catalysts

Article abstract of DOI:10.1039/c9nj04484a

Tungstated metal phosphate (WO_x/MP; M = Al, Zr and Ti) solid acid catalysts were examined for vapor phase glycerol dehydration to acrolein and hydroxyacetone. The physico-chemical characteristics of the samples were explored using X-ray diffraction (XRD), N₂ sorption, Fourier-transform infrared spectroscopy (FT-IR), UV-visible diffuse reflectance spectroscopy (UV-DRS), and ex situ FT-IR investigation of the adsorbed pyridine and ammonia temperature programmed desorption (NH₃-TPD) analysis to gain an understanding of the impact of the structural and acidic nature of the catalysts upon changing the metal component. The catalytic activity and selectivity functionalities were interpreted in terms of the acidity and structural properties of the WO_x/MP catalysts. The outcomes demonstrated that WO_x/TiP was the most selective and stable catalyst during the catalytic dehydration of glycerol owing to the acceptable amount of surface acidity and the substantial number of pores available on its surface. A detailed study on WO_x/TiP was also performed to understand the interaction between the tungsten and the TiP support.

Full text of DOI:10.1039/c9nj04484a

NJC
View Article Online
View Journal | View Issue
PAPER
Comparative study of vapour phase glycerol
dehydration over different tungstated metal
Cite this: New J. Chem., 2019,
phosphate acid catalysts†
ab
43, 16860
bc
b
Srinivasarao Ginjupalli,
*
Putrakumar Balla,
Hussain Shaikb,
b
b
Nagaraju Nekkala, Bhanuchander Ponnala and Harishekar Mitta
Tungstated metal phosphate (WO
phase glycerol dehydration to acrolein and hydroxyacetone. The physico-chemical characteristics of
the samples were explored using X-ray diffraction (XRD), N sorption, Fourier-transform infrared
spectroscopy (FT-IR), UV-visible diffuse reflectance spectroscopy (UV-DRS), and ex situ FT-IR
investigation of the adsorbed pyridine and ammonia temperature programmed desorption (NH -TPD)
x
/MP; M = Al, Zr and Ti) solid acid catalysts were examined for vapor
2
3
analysis to gain an understanding of the impact of the structural and acidic nature of the catalysts upon
changing the metal component. The catalytic activity and selectivity functionalities were interpreted in
Received 3rd September 2019,
Accepted 28th September 2019
terms of the acidity and structural properties of the WO
WO /TiP was the most selective and stable catalyst during the catalytic dehydration of glycerol owing to
the acceptable amount of surface acidity and the substantial number of pores available on its surface.
x
/MP catalysts. The outcomes demonstrated that
x
DOI: 10.1039/c9nj04484a
A detailed study on WO
and the TiP support.
x
/TiP was also performed to understand the interaction between the tungsten
rsc.li/njc
Catalytic glycerol dehydration is expected to be a result of a
double dehydration reaction that proceeds via the formation of
Introduction
The utilization of renewable biomass resources is inevitable for hydroxyacetone, with 3-hydroxypropionaldehyde as an inter-
1
11
the continuous development of chemical and fluid energy. As mediate (Scheme 1). The most astonishing selectivity to
a substantial excess of glycerol is created during biodiesel acrolein reported to date is 65–90%, achieved by catalytic
generation from the transesterification of vegetable oils, further glycerol dehydration in either the liquid or gas phase. It was
consideration must be focussed on novel utilization of glycerol found that a Hammet acidity (H ) at intervals of À10 and À16
o
from the perspective of both environmental protection and for an acid catalyst is more appropriate for glycerol dehydration
economic benefit. Various catalytic processes have been attempted to acrolein compare to a reduced acidity with a H
o
of around À2
1
2
for the transformation of glycerol into value-added chemicals. and À6. However, such heavy acidic forces have a tendency to
These different procedures produce a number of important deactivate quickly because of an abundance of carbonaceous
1
3
substance intermediates, including 1,2-propanediol and 1,3-pro- species which hinders the active surface. The textural pro-
2
panediol (hydrogenolysis and fermentation), dihydroxyacetone, perties also play an essential part alongside the acidity of the
1
4
glyceraldehyde, glyceric acid, glycolic acid and hydroxypyruvic active site, especially considering an expansive number of
3
4
15
acid (oxidation), polyglycerols (etherification), acrolein micropores. The selectivity and deactivation may be firmly
5
–7
(
dehydration),
oxygenated fuel additives (acetalization and influenced by the diffusion limitations owing to the develop-
acylation), bio fuels (fermentation), and polyglycerols and ment of coke.
polyglycerol esters (polymerization).
8
,9
10
1
1
Metal phosphate (MP) catalysts are well known as selective
and active catalysts for acid-catalyzed dehydration, isomerisation,
5
–9
Friedel–Crafts reactions and ketalization of ketones with diols.
a
Basic Science and Humanities, Swarnandhra College of Engineering and Technology,
Narsapur, 534280, India. E-mail: gsrao260@gmail.com; Tel: +91-9848088968
Inorganic and Physical Chemistry, Catalysis division, Indian Institute of Chemical
Technology, Hyderabad-500007, India
Several research studies have been carried out to resolve the
physicochemical and acidic properties of metal phosphate
b
c
16
catalysts. Furthermore, owing to their textural and acidic
1
7,18
properties
these metal phosphates can be utilized as catalyst
State key Laboratory of Catalysis, Dalian Institute of Chemical Physics,
Chinese Academy of Sciences, 457 - Zhong Shan Road, Dalian-116023, China
supports for metal oxides. In the field of biomass transforma-
†
Electronic supplementary information (ESI) available. See DOI: 10.1039/c9nj04484a tion, different studies have been performed on metal phosphate
16860 | New J. Chem., 2019, 43, 16860--16869 This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2019

View Article Online
Paper
NJC
Scheme 1 A schematic diagram of catalytic glycerol dehydration to acrolein and hydroxyacetone over solid acid catalysts.
catalysts for the selective production of hydroxymethylfurfural Material and methods
(
HMF) from carbohydrates, glucose to levulinic acid, and acro-
Preparation of catalysts
6
lein from glycerol in both the aqueous and gas phase.
Aluminium, titanium, and zirconium phosphates were pre-
pared by hydrothermal synthesis as described previously in the
Catalysts containing tungsten oxides have been used for acid-
19
catalysed reactions since the turn of the century. Tungsten oxide-
based catalysts have found different applications in petroleum
refineries, petrochemical, pollution control industries and biomass
2
8–30
literature
The metal phosphate supports were prepared by using their
alkoxide precursors and H PO as the phosphorous source.
with a M/P (metal to phosphorous) ratio of 1.0.
1
9,20
21
22
3
4
value addition reactions.
Mariano et al., Nadji et al., Song-Hai
et al. and Sung et al. have reported that metal oxide (Al
ZrO , TiO ) supported tungsten oxide catalysts exhibit unique
23
24
After the hydrothermal process, the solvent was filtered off and
the hydrogel was washed with twice its volume of millipore
water. The resultant filtrate was oven-dried at 110 1C for 12 h
and calcined at 500 1C under static air for 4 h.
2 3
O ,
2
2
catalytic properties in the dehydration of glycerol to acrolein.
For several of those studies, the active component of the
tungsten oxide phase assumes a key part in their overall
catalytic performance. The active phase of the supported tungsten
The supported tungsten oxide (15 wt% WO
were then prepared by the simple impregnation technique by
2,25,26 the addition of an aqueous solution of ammonium metatung-
x
/MP) catalysts
2
oxide catalysts is also modified by the preparation method
state to the calcined MP supports. The resultant material was
finally dried at 110 1C for 12 h and calcined in a muffle furnace
at 400 1C for 5 h.
and doping with Si and Nb to give a catalyst with a long lifetime for
24,27
glycerol dehydration.
In our previous work, we reported a comprehensive study of
WO /ZrP as a catalyst for glycerol dehydration to acrolein under
different reactions as well as different catalyst synthesis parameters.
x
9
Characterisation procedures
In the present investigation, we report the catalytic activity and X-ray diffraction (XRD) analysis was performed using a D8
selectivity of metal phosphate supported tungsten oxide (WO /MP: Diffractometer (Bruker, Germany), utilizing Cu-Ka radiation
x
M = Al, Zr, Ti) catalysts throughout the vapour phase glycerol (1.54 Å) at 30 mA and 40 kV. The Brunauer–Emmett–Teller
dehydration to acrolein. Aluminium phosphate (AlP), zirco- (BET) surface area and the pore related studies for all of the
2
nium phosphate (ZrP) and titanium phosphate (TiP) supports catalysts were evaluated by utilising N adsorption isotherms at
2
were synthesized using a hydrothermal process and supported À196 1C using the multipoint BET technique taking 0.162 nm
tungsten oxide catalysts were prepared by a simple impregnation as the cross-sectional territory utilising an Autosorb-1 (Quanta
step and characterized using physicochemical techniques. The chrome instruments, USA). The GBC UV-Visible Cintra 10e
acidic strength and nature of these catalysts were determined spectrometer with a coordinating circle reflectance was used
3
using ammonia temperature programmed desorption (NH -TPD) for recording the UV-Vis diffuse reflectance spectra (UV-DRS).
and Py-Fourier-transform infrared spectroscopy (FT-IR) spectro- The spectra were all recorded at around room temperature and
scopic investigation respectively. The final results indicate that the information was changed to agree with the Kubelka–Munk
the acidic properties of catalysts can be related to the catalytic condition f (R) = (1 À Ra)2/2ra. The NH
3
-TPD investigations
performance and have an effect on various reaction parameters were performed on the AutoChem 2910 instrument. Before
and reactant functionalities in the glycerol dehydration reaction. beginning the TPD examination, the specimen was flushed
À1
The present study also reports the effects of tungsten loading on with high purity (99.99%) helium (50 ml min ) at 300 1C for
a TiP support and product distribution during the dehydration of 1 h. After pretreatment the specimen was immersed in a 10%
À1
glycerol to acrolein. Our results provide a basis for correlation of NH
3
adjusted He blend (75 ml min ) at 80 1C for 1 h and then
À1
the catalyst acidity to the variation of the tungsten oxide content the lines were flushed with He gas (30 ml min ) for 1 h at the
in glycerol dehydration.
3
same temperature to expel the physisorbed NH . The TPD
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2019 New J. Chem., 2019, 43, 16860--16869 | 16861

View Article Online
NJC
Paper
investigation was completed by heating from the surrounding
À1
temperature to 700 1C with a 10 1C min heating rate. The
amount of NH desorbed was calculated by utilising GRAMS/32
3
programming. The infrared (IR) spectra were recorded with an
IR (model: GC-FT-IR Nicolet 670) spectrometer using a KBr
plate technique at room temperature.
Activity tests
The vapor-phase catalytic dehydration of 20 wt% glycerol was
conducted at 300 1C, under normal pressure in a vertical fixed-
bed quartz reactor using 200 mg of the sample. A fluid arrange-
ment consisting of 20 wt% glycerol was charged into the reactor
À1
using a micro-syringe pump at a stream rate of 0.5 mL h
(WHSV-2.6/h). The resultant products were consolidated in
an ice–water trap and accumulated hourly for investigation
utilizing a gas chromatograph GC-2014 (Shimadzu) provided
with a DB-wax capillary column (Agilent-123-7033) and a flame
ionization detector (FID).
Results and discussion
Physico-chemical properties of the catalysts
x
The crystallinity of various metal phosphates (MPs) and WO /MP
catalysts with 15 wt% tungsten oxide loading were distinguished
using XRD patterns. The samples did not demonstrate any XRD
diffraction indicating that the tungsten oxide is available in the well
dispersed amorphous state. Our previous studies also indicated that
31
6
the prepared AlP and ZrP had amorphous characteristics.
The BET surface area, pore volume, and average pore diameter
x
for metal phosphates and WO /MP catalysts evaluated from the
related adsorption isotherms are displayed in Table 1 and the BJH
pore size distribution of prepared catalysts have shown in Fig. 1.
All of the examined catalysts display moderate BET surface areas,
2
À1
which are seen to shift in the range from 150 to 400 m g
.
Amongst all of the catalysts examined, the AlP based samples
demonstrated the smallest surface area and pore volumes
2
À1
À1
(
1
x
AlP: ca. 204 m g with 0.28 cc g total pore volume; WO /AlP:
2
À1
À1
51 m g with 0.19 cc g total pore volume), while the ZrP based Fig. 1 BJH pore size distribution of the MP and WO
x
/MP catalysts.
catalysts showed the most astounding surface areas and pore
2
À1
À1
volumes (ZrP: 400 m g with 0.36 cc g total pore volume;
2
À1
À1
WO /ZrP: 195 m g with 0.24 cc g total pore volume).
impregnated catalysts owing to the W–Oc–W vibration of the two
x
The spectra of all the prepared catalysts demonstrated IR interfacing WO
6
octahedra (see Fig. 2). It is probably significant
À1
À1
absorption bands within the 1000–1100 cm area and the peak that a band was observed at 640 cm for all of the impregnated
À1
in the 490–510 cm area is caused by the asymmetric stretching catalysts which can be ascribed to a M–O vibration. As this IR
3
2
and bending mode of the phosphate ion, respectively. An extra band was missing in metal phosphate (MP) catalysts, it may
À1
33
IR band was observed in the 800–830 cm area in all of the WO
x
originate from a M–O–W extending vibration. These outcomes
3
Table 1 BET surface area, pore size distribution data and TPD of NH of MP and W/MP catalysts
BET surface
area (m g
Total pore
volume (cc g
Average pore
diameter (nm)
Total NH
mmol g
3
liberation
3
Total NH liberation
mmol m
2
À1
À1
À1
À2
S. no
Sample name
)
)
B/L ratio
1
2
3
4
5
6
AlP
ZrP
TiP
15W/AlP
15W/ZrP
15W/TiP
204.0
400.3
248.0
151.1
195.2
159.1
0.28
0.36
0.31
0.19
0.24
0.21
58.0
31.6
84.1
85.3
34.5
93.3
1.64
3.30
3.51
2.27
3.95
4.60
8.04
8.24
14.1
15.0
20.2
28.9
0.2
0.4
1.2
2.1
2.5
3.2
16862 | New J. Chem., 2019, 43, 16860--16869 This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2019

View Article Online
Paper
NJC
x
Fig. 2 FT-IR profiles of MP and WO /MP catalysts.
demonstrate the presence of the phosphate framework in the
metal phosphate materials and WOx interactions with metal
phosphate supports.
The UV-Vis spectra of all of the WO
x
x
/MP catalysts are Fig. 3 UV DRS profiles of the MP and WO /MP catalysts.
displayed in Fig. 3, alongside the range of metal phosphate
supports. The diffuse reflectance UV-Visible spectra of the AlP
catalyst possess a band at around 265 nm that is characteristic
The UV range of the MP support is forcefully adjusted by the
WO species. UV-vis DRS was utilized to test the tungsten oxide
The spectrum exhibits a small band at 280 nm ascribed to species scattered on the MP surface. The impact of the size can
3
+
34
of the d–d transition of Al ions.
x
3
5
the phosphate counter anions in the framework. The assimilation be obtained from the change in the retention edge. As the
2À
4+
x
edge at 220 nm for the ZrP sample is due to the O –Zr charge WO is impregnated on the MP, the force of the charge transfer
exchange transitions, relating to the excitation of the electrons transition band is also expanded, it shows the WO
from the valence band (having an O 2p character) to the with the MP support.
x
association
conduction band (having a Zr 4d character). The ZrP sample
3
The NH -TPD method used to be employed to examine the
demonstrated a band at 300 nm, which may be due to the acidic strength of catalysts. Ammonia is a small molecule in
Zr(IV) cations associated with the phosphate counter anions in size and it is frequently used as a probe molecule for the TPD
1
7
the system. The TiP sample exhibits two IR bands in the method owing to its strong basic strength and stability at
10–220 and 300–310 nm wavelength areas. The primary band high temperatures. Catalytic dehydration is an acid catalysed
2
2À
could be because of the electronic move from the O 2p to the reaction and generally takes place on the active and stable acid
Ti 3d orbital. The former high-energy absorption edge could sites present on the catalyst surface. Thus, resolution of the
be caused by the Ti(IV) cations interacting with the phosphate surface acidity will probably be a key factor used to check the
4
+
counter anions in the framework.
dehydration activity and selectivity.
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2019 New J. Chem., 2019, 43, 16860--16869 | 16863

View Article Online
NJC
Paper
3 x
Fig. 4 NH -TPD profiles of the MP and WO /MP catalysts.
The apparent acidity was determined as the total acidity and to note that we have applied ex situ FT-IR pyridine adsorption to
disclosed as the desorbed NH (mmol) per gram of catalyst. determine the nature of the acidic sites and the proportion of the
Moreover, the acid density is given as the desorbed NH
3
3
Brønsted and Lewis acidic sites similar to their relative areas.
The Py-FT-IR spectra of the distinct metal phosphates are
2
(
3
mmol) per m of surface area in Table 1. From the NH -TPD
3
profile (Fig. 4), it is clear that the NH -desorption of the AlP shown in Fig. 5 and the B/L proportions of the catalysts
catalyst is related in the two temperature areas compared to the are provided in Table 1. All of the catalysts have a IR band at
À1
weak and weak-moderate acid strengths. The ZrP and TiP 1444 cm for the Lewis sites and the other IR band confirmed
À1
samples have indicated one sharp peak in the weak acidic area at values up to 1550 cm is ascribed to the Brønsted acidic
and one broad peak in the moderate acidic area.
The TPD profiles of the WO impregnated metal phosphate posses both Lewis and Brønsted acidic sites in altered amounts
catalysts show the desorbed peaks (50–130 1C) are related to owing to the attributes of the metallic phosphate catalyst.
weak acidity, the intensity is attenuating and the desorbed As can be seen from the results reported in Table 1, the B/L
peaks (150–500 1C) related to moderate acidity are increasing. acidity proportion of the WO /MP catalysts was found to be
The overall acidity values of WO /MP are also multiplied com-
pared to the bare MP supports. However, in the WO /MP catalysts
sites. It is also obvious from Fig. 5, that all of the catalysts
x
x
x
x
the NH is customarily desorbed between 150 and 500 1C, which
3
is indicative of the presence of medium to strong acid sites. The
acidity of the WO
P–OH and to the WO
that the quantity of acidity and the strength of the acidic sites
increase upon addition of the WO component to the MP
support. This may also be noticeable from the TPD profiles of
the MP and WO /MP catalysts, the total acidity of the catalysts is
reduced in the following order: WO /TiP 4 WO /ZrP 4 WO /
x
/MP catalysts can be credited to the surface
3
6
x
groups. The TPD evaluation indicates
x
x
x
x
x
AlP 4 TiP 4 ZrP 4 AlP.
The TPD technique cannot separate the Lewis and Brønsted
acid sites. As a consequence, the TPD experiment does not
provide any information on the type of acidic sites. Ex situ FT-IR
examination of pyridine adsorbed samples was utilized to
describe the nature of the surface acidic sites. Before the ex situ
FT-IR examination, the samples were saturated with pyridine
vapor at 200 1C for 2 h. Relying on the characteristics of bonding
fashioned via the acidic site with pyridine, the IR-spectra offers
À1
distinctive bands for the Brønsted (B: 1540–1548 cm ) and
À1
Lewis (L: 1445–1460 cm ) acid sites. Moreover, the band at
À1
1
490–1500 cm is typically ascribed to the superposition of the
3
7
x
signals for the Brønsted and Lewis (B + L) sites. It is important Fig. 5 Pyridine adsorbed FT-IR profiles for the MP and WO /MP catalysts.
16864 | New J. Chem., 2019, 43, 16860--16869 This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2019

View Article Online
Paper
NJC
higher than the pure MP supports. The FT-IR spectra of the Catalytic activity: vapour phase dehydration of glycerol
pyridine adsorbed samples reveal that the intensity of the IR bands
at 1498 cm (B + L) and 1540 cm (B) was amplified in the
following order: TiP 4 ZrP 4 AlP. It is fascinating to note that with
The results of the acid catalysed glycerol dehydration activity
over different MP and WO /MP catalysts are shown in Fig. 6 and
x
the distribution of the resultant products is reported in Table 2.
Fig. 6 illustrates the catalytic activity of the different MP and
À1
À1
the addition of tungsten oxide upon the MP support the FT-IR band
À1
intensity in the 1450 cm region, corresponding to the Lewis WO /MP catalysts towards glycerol conversion and the selectivity
x
acidity, declined slightly. The intensity of the IR bands at 1498 and of the different reaction products at a time on stream (TOS) of
À1
1
540 cm were also improved in the WO /MP catalysts. The Py-FT- 10 h. The TiP-based catalysts presented the best conversion of
x
IR spectra indicate that the quantity of Brønsted sites is improved glycerol with a prevalent selectivity towards acrolein among all of
upon impregnation of WO
x
ions into the MP support.
the prepared catalysts. The results can be characterized by means
Fig. 6 Evolution of the glycerol conversion and selectivity to acrolein as a function of time over various MP and WO
x
/MP catalysts. Reaction conditions:
À1
À1
0
.2 g catalyst diluted with 3.0 g of quartz. Aqueous glycerol (20% wt/wt) fed at the rate of 0.5 g h at a temperature of 320 1C, under a 7 mL h
N
2
atmosphere.
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2019 New J. Chem., 2019, 43, 16860--16869 | 16865

View Article Online
NJC
Paper
Table 2 Effect of reaction temperature on product distribution at TOS =
The WO
/MP catalysts exhibit a low selectivity to hydroxy-
x
1
x
0 h over various WO /MP catalysts
acetone compared to the metal phosphate catalysts, indicating
that the increase in WO on the MP supports strengthens the
acidity as well as the Brønsted acidity of the catalysts, leading to
consistent selectivity for acrolein. Additionally, these catalysts
Conversion of glycerol
and selectivity to reaction
products (%)
WO
x
/AlP
WO
x
/ZrP
WO
x
/TiP
x
300 320 340 300 320 340 300 320 340
Glycerol
Acrolein
Hydroxyacetone
Acetaldehyde
Others
81 92 98 88 96 100 96 100 100 possess combined Lewis and the Brønsted acid sites (B + L) that
51 60 59 54 62
60 74
80 70
are beneficial for steady acrolein selectivity during the catalytic
dehydration of glycerol. From the characterization results the
18 12
4
12
21 15
8
14
7
7
5
8
7
3
13
4
6
6
8
25 22 25 19 16
18 12
x
14 P–OH and W–OH groups on the surface of the WO /MP catalysts
were found to be responsible for the generation of the Brønsted
acid sites that are active for the selective dehydration reaction.
Others include allylic alcohol, acetone, acetic acid, hydrogenolysis
products.
4
3
The Py-FT-IR spectra suggests that the WO /MP catalysts have
x
distinct Brønsted sites and the combination of B + L acidic sites.
of the acidity of the catalysts, for this reason the NH
Py-FT-IR results indicate that the TIP based catalyst have a greater sites along with B + L acidic sites. This can apparently improve
acid strength as well as B/L ratio compared to the other metal the selectivity of acrolein in WO /MP compared to the pure
3
-TPD and However, the MP catalysts have a greater number of Lewis acidic
x
phosphate-based catalysts.
metal phosphate catalysts.
The impact of the reaction temperature (300–340 1C) on
The AlP and ZrP catalysts presented a conversion of 68 and
8
0% respectively after a TOS of 10 h, but the selectivity to glycerol dehydration was studied over different WO /MP catalysts,
x
acrolein was found to be lower when compared with the TiP and the outcomes are listed in Table 2. At 300 1C, the trans-
catalysts. The AlP and ZrP catalysts demonstrated a hydroxy- formation of the reactant by all of the catalysts was found to be
acetone selectivities of around 16% and 15%, respectively. reduced and the greatest transformation was achieved at the
Generally, Lewis acid sites are favourable for the production highest temperature. The excessive selectivity to acrolein
of hydroxyacetone and this can be clarified by the B/L acidic site accomplished at average temperatures might be explained by
ratio which is determined by the Py adsorbed FT-IR spectra the specific performance of the stronger adsorption process of
(
Table 1). The ZrP and TiP catalysts show a total acidity with a glycerol via the 21-OH group, which leads to the kinetically
À1
small difference (0.21 mmol g ), although the catalytic activity favoured consecutive steps. The greater amount of hydroxy-
of TiP is higher when compared with ZrP. This could be acetone found at a lower temperature is conceivable because of
explained by the nature of the acidic sites determined using the distinction in the initiation energies of the response path-
the Py-FT-IR spectra. The TiP catalyst (B/L: 1.2) possessed ways as shown in Scheme 1. Although the higher temperatures
a greater quantity of Brønsted acidity compared to the ZrP favour the decay of hydroxyacetone, subsequently the selectivity
(
2
B/L: 0.4) catalyst (Table 1). Although ZrO -based catalysts have of the acetic acid and the acetaldehyde increases with the
3
8
been stated as being neutral in nature, it has been discovered reaction temperature.
that both acidic and basic sites can be accommodated on the At 320 1C the WO
surface of ZrO supports. Nonetheless, ZrP catalysts have shown conversion and selectivity. There is an appreciable effect on the
magnificent acidic properties in various acid catalyzed reactions,
and can provide some basic sites on their surfaces. The presence (Table 2). However, at a greater temperature (340 1C) the acrolein
x
/MP catalyst demonstrated the highest
39
2
40
glycerol transformation observed with the rise in the temperature
41
of these weak basic sites enables better selectivity for allylic alcohol selectivity diminished owing to the increment in the development of
42
and hydroxyacetone.
the acetic acid and the acetaldehyde along with a trace amount of
The AlP catalyst exhibits 68% glycerol conversion with a 40% unidentified products. These outcomes apparently indicate that at a
selectivity towards acrolein during glycerol dehydration. The higher temperature (340 1C) the catalyst assists in C–C bond
TPD profile of AlP shows reduced acidity and also that there is a cleavage resulting in an oxidation reaction rather than glycerol
shift in the NH -desorbed peak towards a lower temperature dehydration. Acetaldehyde is the major side product formed owing
3
region in comparison with the other samples. Hydroxyacetone to C–C bond cleavage and its further oxidation leads to the
and acetaldehyde are obtained as the major by-products along development of acetic acid. From the perspective of these
with acetone, allylic alcohol, methanol, acetic acid, hydrogeno- discoveries, 320 1C was chosen as the ideal reaction temperature
lysis products and also a trace amount of unidentified products for screening of the catalysts for further examination.
during glycerol dehydration over pure MP catalysts.
The vapour phase dehydration of glycerol over the MP and
WO /MP catalysts at 320 1C with respect to time is shown in
The results shown in Fig. 6 obviously indicate that the WO
x
/
x
MP catalysts are observed to be more active than pure MP supports Table 2. The greatest transformation of glycerol was found at a
during the glycerol dehydration reaction. The NH -TPD and Py-FT-IR TOS of 10 h, thereafter a decreasing pattern was observed for all
3
analysis clearly explain the low activity resulting from glycerol of the metal phosphate and tungstated metal phosphate catalysts.
dehydration over the MP catalysts. The TPD profile of the pure The best catalytic performance was obtained over the TiP based
metal phosphates appears in the low-temperature region with a catalysts. The TiP catalyst showed a relatively stable glycerol
reduced amount of acidity and also has a low B/L ratio of acidic conversion of 85% after 30 h versus 90% during 10 h and an
x
sites compared to the WO /MP catalysts.
acrolein selectivity of 56% after 30 h versus 60% during a TOS of
16866 | New J. Chem., 2019, 43, 16860--16869 This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2019

View Article Online
Paper
0 h. The WO
NJC
1
x
/TiP catalyst also showed a long lifetime and a analysis (Table 3) was also performed to evaluate the acidic
stable glycerol conversion of 90% after 50 h versus 100% during properties of the spent WO
0 h and an acrolein selectivity of 70% after 50 h versus 70% values of the used catalysts have been reduced compared to those of
during a TOS of 10 h. the fresh WO /MP catalyst. Nevertheless, the used WO /TiP catalyst
x
/MP catalysts. The total ammonia uptake
1
x
x
For the most part, the conversion of glycerol and selectivity demonstrated a greater NH uptake value compared to the other
3
towards acrolein relied upon the overall acidity and on the spent catalysts. This indicates that the narrow pore structure of AlP
textural properties. The TiP based catalysts demonstrated a and ZrP were more affected by the rapid coke formation on the
pronounced catalytic lifetime during the gas phase dehydration active sites which are responsible for the dehydration reaction.
of glycerol owing to the appropriate amount of the surface
x
The TG-DTG profiles of the used WO /MP catalysts are given
acidic density and the large pores on the surface. A significant in Fig. 7 and this suggests that the weight loss of all of the catalysts
decrease in both the conversion of glycerol and the selectivity of was observed below 400 1C, this indicates the coke deposited on
acrolein in the case of the AlP and ZrP based catalysts is the surface of the catalyst is amorphous. About 9%, 18%, and 28%
attributed to the narrow pores. The selectivity and deactivation of weight loss occurred in the 50–400 1C regions after a TOS of
may strongly suffer from dispersion limitations because of coke 10 h for TiP, AlP and ZrP respectively. From these results it is
x
development on the restricted pores. These problems are not clear that the surface of the WO /ZrP catalyst is most affected by
present in the TiP based catalysts and might explain the better coke formation during the catalytic reaction compared to the
longevity of this catalyst during glycerol dehydration.
other catalysts.
The characterisation of the spent catalysts provided superior
CHNS elemental analysis was utilized to determine the atomic
clarification of the structural and acidic properties of the WOx/ H/C ratio in the coke over the used catalysts and the results
MP catalysts during the catalytic glycerol dehydration. NH -TPD (Table 3) are in good agreement with the TGA-DTG analysis. From
3
Table 3 it can be clearly observed that the H/C ratio of the used
WO
uptake and CHNS analysis ratio of the used WO
x
x
/TiP is higher than the other spent catalysts. The higher H/C
Table 3 BET surface area, pore volume, NH
3
/TiP catalyst after a TOS of 10 h indicates
data for the spent WO
x
/MP catalyst
that the coke deposits have a less dense and aromatized structure
over the active sites.
WO
x
/TiP
WO
x
/ZrP
x
WO /AlP
The results reported in Table 3 suggest that a decline in the
surface area and the total pore volumes were found in used
2
À1
BET surface area (m g
)
124.0
0.18
4.34
1.92
128.7
0.17
2.95
1.28
104.4
0.14
2.00
1.36
À1
BJH pore volume (cc g
Total NH uptake (mmol g
H/C (CHNS analysis)
)
À1
3
)
WO /MP catalysts compared to the fresh catalyst (Table 1). This
x
could be due to the deposition of the coke on the active surface
of the catalysts during the dehydration of glycerol, leading to
deactivation of the active component of the catalyst.
In the present study we also report the effect of tungsten
oxide loading on the best support, TiP, and our results provide
a basis for correlating the catalyst acidity by varying the
tungsten oxide content in glycerol dehydration.
The effect of WO loading on the TiP support during the vapour
x
phase dehydration of glycerol by various WO /TiP catalysts is
x
shown in Table 4. As can be seen from Table 4, the results clearly
suggest that tungstated titanium phosphate (WO
are found to be more active than the pure TiP support. The 10 wt%
and 15 wt% WO /TiP catalysts showed a higher glycerol conversion
than the other catalysts. However, the 15 wt% WO /TiP catalyst
/MP catalysts after a TOS of 10 h. demonstrated the highest selectivity to acrolein (80%).
x
/TiP) catalysts
x
x
Fig. 7 TG-DTG curve of the spent WO
x
Table 4 Product distribution results for glycerol dehydration over pure and tungstated TiP catalysts
Selectivity of reaction products (%)
Catalyst
Pure TiP
Con. of glycerol (%)
Acrolein
Hydroxyacetone
Acetaldehyde
Allyl alcohol
Acetone
Acetic acid
Others
90
95
100
100
96
60
67
71
80
76
72
14
10
7
5
6
5
7
9
8
10
10
5
3
3
1
1
2
4
3
2
1
2
2
2
2
2
2
2
3
10
8
6
3
3
5
1
1
2
2
W-TiP
0W-TiP
5W-TiP
0W-TiP
5W-TiP
94
v
3
À1
À1
Reaction conditions: 0.2 g catalyst; 20% wt/wt aqueous glycerol; glycerol feed rate 0.5 g h ; 10 mL min
2
N atmosphere; TOS 10 h; reaction
temperature 300 1C.
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2019 New J. Chem., 2019, 43, 16860--16869 | 16867

View Article Online
NJC
Paper
The pure TiP support exhibited 90% glycerol conversion textural and acidic properties of the catalysts. The characterization
with 60% selectivity towards acrolein under a N feed flow. of used catalysts also explains the superior acidic and structural
2
The P-TiP support gave hydroxyacetone and acetaldehyde as the properties of the WO /TiP catalyst for selective acrolein production
x
major by-products along with 21% of other products including via the catalytic glycerol dehydration reaction.
allylic alcohol, acetic acid, acetone, methanol, hydrogenolysis
products and trace amounts of unidentified products. The
decrease in the dehydration activity can be explained by the
acidity of the P-TiP support determined using the NH -TPD
3
method (Fig. S3, ESI†). The TPD profile of P-TiP exhibited a
reduced amount of acidity and the TPD profile appeared at a
lower temperature region compared to other samples. The
Conflicts of interest
There are no conflicts to declare.
References
selectivity to hydroxyacetone was reduced for the WO /TiP
x
catalysts compared to the pure support suggesting that the
1 C. Daniel, L. W. Tyler, C. Stefan and J. Whitney, Green
Chem., 2014, 16, 384.
2 S. P. Samudrala, S. Kandasamy and S. Bhattacharya, Sci.
Rep., 2018, 8, 1.
3 D. Georgios, S. Stefan and T. Harun, ACS Catal., 2018, 87, 6301.
4 O. T. L. Caroline, L. R. Leticia, A. S. B. Marcos, D. F. Lindoval,
L. C. F. Orcid and E. H. Carla, Energy Fuels, 2017, 315, 5158.
5 G. Srinivasarao, N. P. Rajan, P. K. Vanama and K. V. R.
Chary, Chem. Technol. Biotechnol., 2014, 89, 1890.
6 G. Srinivasarao, N. P. Rajan, H. S. Mitta, A. Sarthani and
K. V. R. Chary, J. Mol. Catal. A: Chem., 2014, 395, 486.
7 G. Srinivasarao, S. K. Hussain and K. V. R. Chary, Mater.
Today: Proc., 2018, 5, 25773.
addition of WO
acetone and as a result the acrolein selectivity increases. However, at
lower loading values (5 wt% WO /TiP) by-products and other
products are formed in greater amounts because the TiP support
is not covered fully by the WO . The superior catalytic performance
x
depletes the active sites leading to hydroxy-
x
x
of these catalysts could imply that the Brønsted acid sites and
the combination of both the Lewis and the Brønsted acid sites
(Fig. S4, ESI†) are advantageous for stable acrolein production
from glycerol dehydration.
Characterization of the WO
x
/TiP catalysts suggests that the
active species corresponds to P–OH and W–OH species, which
generates Brønsted acid sites which are active and selective for
4
3
the dehydration reaction. Whereas a decrease in the catalyst
8 A. Sarthani, G. Srinivasarao and K. V. R. Chary, Appl.
Petrochem. Res., 2018, 8, 107.
activity at high WO (above 15 wt%) loading is attributed to
3
condensation of the W–OH groups and formation of three-
9 S. Gadamsetti, N. P. Rajan, G. Srinivasarao and K. V. R.
Chary, J. Mol. Catal. A: Chem., 2015, 410, 49.
4
4
dimensional WO clusters which reduce the acidity. The XRD
x
(
Fig. S1, ESI†) and UV-DRS (Fig. S2, ESI†) results also further 10 H. Mar ´ı a and P. F. Elena, Microb. Biotechnol., 2017, 10, 528.
support these findings as crystalline peaks of WO
above those of the 15 wt% WO /TiP.
x
x
appeared 11 N. R ´e mi, G. Nicolas, K. S. Abhishek, P. Florian, S. Daniel,
S. Cathleen, K. S. Sunil, H. Rainer and L. Christophe,
Catalysts, 2017, 7, 123; M. J. Antal, W. S. L. Mok, J. C. Roy
and A. T. Raissi, J. Anal. Appl. Pyrolysis, 1985, 8, 291.
1
2 S. Chai, H. Wang, Y. Liang and B. Xu, Green Chem., 2007,
Conclusions
9, 1130.
Pure MP and WO /MP solid acid catalysts have been employed 13 J. Lourenco, M. I. Macedo and A. Fernandes, Catal. Commun.,
x
for selective gas phase glycerol dehydration to acrolein. The
2012, 19, 105.
FT-IR and UV-DRS results showed the presence of the phos- 14 T. M. Neves, J. O. Fernandes, L. M. Liao, E. D. da Silva,
phate framework in the metal phosphate substrates and WO
interactions with metal phosphate supports. The reaction para-
x
C. A. da Rosa and V. B. Mortola, Microporous Mesoporous
Mater., 2019, 275, 244.
meters (temperature and time on stream) have been optimized to 15 B. Katryniok, S. Paul, V. Belli `e re-Baca, P. Rey and
accomplish the catalytic conversion of glycerol. The catalytic F. Dumeignil, Green Chem., 2010, 12, 2079.
activity and selectivity of the glycerol dehydration mainly depends 16 A. Clearfield and D. S. Thakur, Appl. Catal., 1986, 26, 1.
upon the density and nature of the surface acidic sites and the 17 K. N. Rao, A. Sridhar, A. F. Lee, S. J. Tavener, N. A. Young
long performance lifetime of the catalysts depends on the
porosity of the MP supports. The TiP based catalysts demon- 18 N. Pethan Rajan, G. Srinivsa Rao, V. P. Kumar and
strated the best performance, with a higher conversion and K. V. R. Chary, Catal. Sci. Technol., 2014, 4, 81.
acrolein selectivity after a TOS of 10 h. AlP and ZrP based 19 M. Scheithauer, R. K. Grasselli and H. Knozinger, Langmuir,
catalysts showed a poor performance, probably as a result of 1998, 14, 3019.
and K. Wilson, Green Chem., 2006, 8, 790.
diffusion and acid site deactivation problems. The TiP based 20 P. Bhaumik and P. L. Dhepe, ChemCatChem, 2016, 8, 1.
catalysts demonstrated a long lived activity during the catalytic 21 M. Massa, A. Andersson, E. Finocchio and G. Busca,
reaction owing to the appropriate amount of surface acidic
density and the large pores on the surface. The results indicate 22 L. Nadji, A. Masso, D. Delgado, R. Issaadi, E. Rodriguez-
J. Catal., 2013, 307, 170.
that glycerol conversion and selectivity to acrolein depends not
only on the reaction parameters, but can also be affected by the
Aguado, E. Rodriguez-Castellon and J. M. Lopez Nieto, RSC
Adv., 2018, 8, 13344.
16868 | New J. Chem., 2019, 43, 16860--16869 This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2019

View Article Online
Paper
NJC
23 S. H. Chai, B. Yan, L. Z. Tao, Y. Liang and B. Q. Xu, Catal. 34 Y. N. Tchenar, A. C. Braham and R. Bachir, Bull. Mater. Sci.,
Today, 2014, 234, 215.
2012, 35, 673.
24 K. H. Sung and S. Cheng, RSC Adv., 2017, 7, 41880.
35 G. A. Ozin and S. Ozkar, J. Phys. Chem., 1990, 94, 7556.
25 R. Znaiguia, L. Brandhorst, N. Christin, V. B. Baca, P. Rey, 36 T. Y. Kim, D. S. Park, Y. Choi, J. Baek, J. R. Park and J. Yi,
J. M. Millet and S. Loridant, Microporous Mesoporous Mater.,
014, 196, 97.
6 G. Srinivasa Rao, M. Sowmya, N. Pethan Rajan, B. Putra
Kumar and K. V. R. Chary, Appl. Surf. Sci., 2014, 309, 154.
7 P. L. Garbey, S. Loridant, V. Belliere-Baca, P. Rey and
J. M. M. Millet, Catal. Commun., 2011, 16, 170.
8 T. Lindblad, B. Rebenstorf, Z. G. Yan and S. L. T. Andersson,
Appl. Catal., 1994, 112, 187.
9 H. Fei, X. Zhou, H. Zhou, Z. Shen, P. Sun, Z. Yuan and
T. Chen, Microporous Mesoporous Mater., 2007, 100, 139.
0 T. Z. Ren, Z. Y. Yuan and B. L. Su, Chem. Commun., 2004, 2730.
J. Mater. Chem., 2012, 22, 10021.
37 T. Barzetti, E. Selli, D. Moscotti and L. Forni, J. Chem. Soc.,
Faraday Trans., 1996, 92, 1401.
38 J. Goscianska, M. Ziolek, E. Gibson and M. Daturi, Catal.
Today, 2010, 152, 33.
39 K. Pokrovski, K. T. Jung and A. T. Bell, Langmuir, 2001,
17, 4297.
40 G. Srinivasa Rao, N. Pethan Rajan, M. Hari Shekar, S. Ammaji
and K. V. R. Chary, J. Mol. Catal. A: Chem., 2014, 395, 486.
41 D. Spielbauer, G. A. H. Mekhemer, T. Riemer, M. I. Zaki and
H. Kno1zinger, J. Phys. Chem. B, 1997, 101, 4681.
2
2
2
2
2
3
3
1 K. V. R. Chary, G. Kishan, K. Ramesh, C. P. Kumar and 42 P. L. Garbey, S. Loridant, V. B. Baca, P. Rey and J. M. M. Millet,
G. Vidyasagar, Langmuir, 2003, 19, 4548. Catal. Commun., 2011, 16, 170.
2 A. Baiker, P. Dollenmeier, M. Glinski and A. Reller, Appl. 43 Y. T. Kim, K. D. Jung and E. D. Park, Appl. Catal., B, 2011,
Catal., 1987, 35, 351. 107, 177.
3 A. Dutta, A. K. Patra, S. Dutta, B. Saha and A. Bhaumik, 44 S. T. Wong, T. Li, S. Cheng, J. F. Lee and C. Y. Mou, J. Catal.,
J. Mater. Chem., 2012, 22, 14094. 2003, 215, 45.
3
3
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2019 New J. Chem., 2019, 43, 16860--16869 | 16869