Selective vapor-phase oxidation of o-xylene to phthalic anhydride over Co-Mn/H3PW12O40@TiO2 using molecular oxygen as a green oxidant

Article abstract of DOI:10.1002/aoc.4461

The oxidation of o-xylene to phthalic anhydride over Co-Mn/H₃PW₁₂O₄₀@TiO₂ was investigated. The experimental results demonstrated that the prepared catalyst effectively catalyzed the oxidation of o-xylene to phthalic anhydride. Also, the synergistic effect between three metals plays vital roles in this reaction. From a green chemistry point of view, this method is environmentally friendly due to carrying out the oxidation in a fixed-bed reactor under solvent-free condition and using molecular oxygen as a green and cheap oxidizing agent. The resulting solid catalysts were characterized by FT-IR, XRD, XPS, ICP-OES, FESEM, TEM, EDX, DR-UV spectroscopy, BET and thermogravimetric analysis. The oxidation of o-xylene yields four products: o-tolualdehyde, phthaldialdehyde, phthalide and finally phthalic anhydride as the main product. The reaction conditions for oxidation of o-xylene were optimized by varying the temperature, weight hourly space velocity and oxygen flow rate (contact time). The optimum weight percentage of phosphotungstic acid (HPW) and Co/Mn for phthalic anhydride production were 15?wt % and 2?wt%, respectively. The best Co/Mn ratio was found to be 10/1. Oxygen flow rate was very important on the phthalic anhydride formation. The optimum conditions for oxidation of o-xylene were T?=?370?°C, WHSV?=?0.5?h^?1 and oxygen flow rate?=?10?mL?min^?1. Under optimized conditions, a maximum of 88.2% conversion and 75.5% selectivity to phthalic anhydride was achieved with the fresh catalyst. Moreover, reusability of the catalyst was studied and catalytic activity remained unchanged after at least five cycles.

Full text of DOI:10.1002/aoc.4461

Received: 11 March 2018
DOI: 10.1002/aoc.4461
Revised: 8 May 2018
Accepted: 11 May 2018
F U L L P A P E R
Selective vapor‐phase oxidation of o‐xylene to phthalic
anhydride over Co‐Mn/H3PW12O40@TiO2 using molecular
oxygen as a green oxidant
Kianoosh Masoomi1 | Mehran Ghiaci1
| A.M. Botelho do Rego2
1 Department of Chemistry, Isfahan
The oxidation of o‐xylene to phthalic anhydride over Co‐Mn/
H3PW12O40@TiO2 was investigated. The experimental results demonstrated
that the prepared catalyst effectively catalyzed the oxidation of o‐xylene to
phthalic anhydride. Also, the synergistic effect between three metals plays vital
roles in this reaction. From a green chemistry point of view, this method is
environmentally friendly due to carrying out the oxidation in a fixed‐bed reac-
tor under solvent‐free condition and using molecular oxygen as a green and
cheap oxidizing agent. The resulting solid catalysts were characterized by FT‐
IR, XRD, XPS, ICP‐OES, FESEM, TEM, EDX, DR‐UV spectroscopy, BET and
thermogravimetric analysis. The oxidation of o‐xylene yields four products: o‐
tolualdehyde, phthaldialdehyde, phthalide and finally phthalic anhydride as
the main product. The reaction conditions for oxidation of o‐xylene were opti-
mized by varying the temperature, weight hourly space velocity and oxygen
flow rate (contact time). The optimum weight percentage of phosphotungstic
acid (HPW) and Co/Mn for phthalic anhydride production were 15 wt % and
2 wt%, respectively. The best Co/Mn ratio was found to be 10/1. Oxygen flow
rate was very important on the phthalic anhydride formation. The optimum
conditions for oxidation of o‐xylene were T = 370 °C, WHSV = 0.5 h−1 and
oxygen flow rate = 10 mL min−1. Under optimized conditions, a maximum
of 88.2% conversion and 75.5% selectivity to phthalic anhydride was achieved
with the fresh catalyst. Moreover, reusability of the catalyst was studied and
catalytic activity remained unchanged after at least five cycles.
University of Technology, Isfahan
8415683111, Iran
2 Centro de Química ‐ Física Molecular
and IN, IST, Universidade Técnica de
Lisboa, Av. Rovisco Pais, 1049‐001,
Lisbon, Portugal
Correspondence
Mehran Ghiaci, Department of Chemistry,
Isfahan University of Technology, Isfahan,
8415683111, Iran.
Email: mghiaci@cc.iut.ac.ir
Funding information
Isfahan University of Technology
KEYWORDS
o‐Xylene, phosphotungstic acid, phthalic anhydride, TiO2, vapor‐phase oxidation
1 | INTRODUCTION
consumption of about 4,000,000 t year−1 in recent decades,
phthalic anhydride is listed as one of the most important
Selective oxidation, specially using heterogeneous vapor‐
phase oxidation processes, plays a vital role in chemical
industry.[1] Particularly, the catalytic oxidation of o‐xylene
to phthalic anhydride is a process of great industrial
and commercial importance. With increasing global
chemical intermediates in chemical industry.[2] Phthalic
anhydride has been commercially manufactured by BASF
since 1872 by developing the naphthalene oxidation pro-
cess.[3] The primary use of phthalic anhydride as a chemi-
cal intermediate is in the production of plastics from vinyl
Appl Organometal Chem. 2018;e4461.
wileyonlinelibrary.com/journal/aoc
Copyright © 2018 John Wiley & Sons, Ltd.
1 of 13
https://doi.org/10.1002/aoc.4461

2 of 13
MASOOMI ET AL.
chloride. Phthalate esters, which function as plasticizers,
are derived from phthalic anhydride. Major use of phthalic
anhydride has been in the production of polyester resins
and alkyl resins used in paints and lacquers, certain dyes
(anthraquinone, phthalein, rhodamine, phthalocyanine,
fluorescein and xanthene dyes), insect repellents, and
urethane polyester polyols.[1]
Limitations on the use of naphthalene (high price,
low availability and low purity) led the research toward
the discovery of cheaper processes and alternative raw
materials, such as o‐xylene, which is nowadays available
in adequate quantities from cracking plants and refiner-
ies. On the other hand, o‐xylene has several advantages
compared to naphthalene such as lower price, easier
transportation, high excess and availability, high purity
of products and long catalyst lifetime.[3,4]
In the last few decades, the liquid‐phase oxidation of
o‐xylene and p‐xylene has been done in acetic acid,[5,6]
but because acetic acid at high temperatures is very corro-
sive, scientists pay more attention to vapor‐phase oxida-
tion of o‐xylene. This process not only increases the
phthalic anhydride yield but also saves raw materials
and decreases emission of COx gases, which occurs as
the main by‐product in this type of reaction.[2] Lack of
solvent in vapor‐phase oxidation reactions has promoted
it to an environmentally friendly process. From the stand-
point of environmental friendliness, much attention has
been paid to the development of metal catalysts for the
selective oxidation using molecular oxygen as a green
oxidizing agent in vapor‐phase.[1,4]
Co‐Mn catalytic systems without bromide ion for oxida-
tion of o‐xylene. Furthermore, using heteropolyacids
(HPAs) in this catalytic system was the most effective
approach.
It is well known that heteropolyacids (HPAs) have
been largely employed as solid‐acid catalysts in homoge-
neous and heterogeneous reactions due to their unique
structural and chemical properties.[9] Moreover, the high
acidity, redox properties and pseudo‐liquid behavior
of these compounds make them attractive catalysts for
a large number of reactions. Phosphotungstic acid
(H3PW12O40, HPW) is the strongest HPA among the
Keggin series and is a well‐established catalyst for selec-
tive oxidation reactions and acid catalyzed reactions.[10]
Although HPAs in their acidic form or transition metal
ion substituted forms are versatile compounds, their main
disadvantages being high solubility in polar solvents and
low surface area (<10 m2/g). Therefore, in a homoge-
neous reaction the isolation of the products and the reuse
of the catalyst after reaction become difficult.[10] Several
studies were published on the heterogenization and
immobilization of HPAs on various supports and their
use in catalysis as green catalysts.[11–13] Accordingly,
this is an attractive strategy to improve the ease of
separation, increasing surface area, and catalyst reusabil-
ity. Since physically adsorbed HPW on the support
surface can leach out easily, encapsulation of these
catalysts onto the porous material might be the most
effective approach.
In the present study and in continuation of our works
on stabilization of heterogeneous catalysts, we have syn-
thesized and characterized a composite of Co‐Mn/HPW
encapsulated into a mesoporous anatase TiO2 network,
and used as a heterogeneous catalyst for catalyzing the
vapor‐phase oxidation reaction of o‐xylene to phthalic
anhydride in a fixed‐bed reactor. The synergistic effects
between three metals have been considered as the main
reason responsible for the high catalytic performance of
the prepared catalysts.
Recently, considerable efforts have been devoted to
the design of catalysts based on transition metal oxides,
like CoOx and MnOx which exhibit catalytic activity for
oxidation reactions as both a homo and heterogeneous
catalyst because of their oxidation–reduction properties
and ability to demonstrate several valences (Co2+, Co3+
;
Mn2+, Mn3+, Mn4+, Mn7+) with easy electron transfer
between these states.[7] The most efficient cobalt oxide
for oxidation reaction is Co3O4 with a spinel structure,
but at high temperatures the active phase Co3O4 trans-
forms into the less active CoO, leading to a decrease in
activity of the catalyst.[8] To solve this problem, some pro-
moters such as Br− have been used to increase stability
of the cobalt oxide catalyst. For example, the oxidation
of m‐xylene in liquid‐phase, over the Co‐Mn and Br ion
as promoter of nanoparticles immobilized on activated
carbon was investigated.[9] Also in our previous work,
the Co‐Mn‐Br nanoparticles immobilized on modified
bentonite was used for oxidation of p‐xylene in liquid‐
phase.[6] Unfortunately, bromide ion corrodes the
expensive reactors and also forms methyl bromide which
must be sequestered because it can deplete the ozone
layer.[9] In this work, we have attempted to develop other
2 | EXPERIMENTAL
2.1 | Materials and Instruments
All materials were used for synthesis of the catalysts and
oxidation reaction obtained commercially without further
purification: cobalt (II) acetate tetrahydrate, manganese
(II) acetate tetrahydrate, phosphotungstic acid hydrate
(H3PW12O40.xH2O) and polyvinylpyrrolidone were pur-
chased from Merck Company. Titanium (IV) isopropoxide
(Ti [OCH (CH3)2]4, 97%) purchased from Sigma Aldrich
Company. The o‐xylene used in the present investigation
was obtained from Titrachem Company. Molecular

MASOOMI ET AL.
3 of 13
oxygen (99.99%) was obtained from Argon Company. All
reactions were monitored by gas chromatography (Agilent
GC‐14C, column RTX‐50 equipped with a flame ionization
(0.0909 g, 0.00036 mol) and manganese (II) acetate
tetrahydrate (0.00909 g, 0.000037 mol) were added to
the solution. The resulting solution was sonicated for
30 min. Then, titanium (IV) isopropoxide (4.278 g,
0.015 mol) was added dropwise during 5 hr with continu-
ous stirring. Finally, the resulting precipitate was dried
at room temperature and then calcined at 450 °C in air
for 3 hr.
detector and
a
PEG‐20
M
capillary column,
30 m × 0.25 mm × 0.25 mm. Temperature program was
carried out as: initial temperature = 70 °C, 0 min; final
temperature = 280 °C, 1 min; with two ramps heating
rate = 20 and 10 °C/min, temperature of injector = 260 °C,
temperature of detector = 280 °C). Also, all products were
identified by GC‐Mass (Agilent 5975C gas chromatograph
coupled with a mass spectrometer).
2.3.1 | o‐Xylene oxidation procedure and
apparatus
The vapor‐phase tests were performed using approxi-
mately 0.44 g of the catalyst (mesh of 20–40) mixed with
1.0 g inert quartz beads in grain size (2–4 mm) to achieve
better heat transfer in a fixed‐bed vertical down flow
tubular reactor (length: 36 cm; diameter: 1 cm). The
Pyrex reactor was inserted in an electric oven controlled
by a thermocouple and operating at atmospheric pres-
sure. With this arrangement, the reaction temperature
could be controlled with a precision of 1 °C. The feed
(o‐xylene) was introduced by an infusion pump, GENIE,
Kent Scientific Corporation, and oxygen as the carrier
gas, regulated by a mass‐flow meter. The catalyst was pre-
viously activated in situ for 4 h at 250 °C under a
10 ml min−1 O2 flow. In each experiment, the products
were collected by condensing them in a trap containing
salt‐ice, and analyzed off‐line using an Agilent gas chro-
matograph, equipped with a PEG‐20 M capillary column
and a FID detector, and all products were identified by
GC‐Mass. After the reaction, the reactor was cooled to
room temperature and on the basis of consumed o‐
xylene, the mass balances were calculated from the
weights of the reaction mixture after the reaction; in each
reaction there was between 10–20% weight lost because of
decomposition of the main product and intermediates.
2.2 | Characterization Techniques
The catalysts were characterized by FT‐IR, DR‐UV, ICP‐
OES, FESEM‐EDX, TEM, XPS, XRD, TGA and N2
adsorption–desorption techniques. Infrared spectra were
recorded in the range of 4000–400 cm−1on a Jasco 680‐
plus FT‐IR spectrophotometer using potassium bromide
(KBr) pellets. Diffuse reflectance spectra were recorded
in the range of 200–900 nm with a Jasco V‐670 UV–vis.
FESEM images were obtained using the TESCAN Mira‐
3, 15 kV instrument and transmission electron micros-
copy (TEM) were registered on a Philips CM120 transmis-
sion electron microscope with accelerator voltage of
100 kV. Thermogravimetric analysis (TGA) was done
under an argon atmosphere at heating rate of 10 °C/
min on BAHR‐STA/TGA‐503. The weight percentage of
cobalt, manganese and tungsten in the catalysts were
obtained by inductively coupled plasma‐optical emission
spectroscopy (ICP‐OES) on Varian Vista‐PRO CCD
Simultaneous ICP‐OES Inductive gekoppeltes. Nitrogen
adsorption–desorption isotherms were measured with a
PHS‐1020 (PHS China) at 77 K. The surface area was cal-
culated by the Brunauer–Emmett–Teller (BET) method
and pore size distribution was derived from the adsorp-
tion branches of the isotherms using the distribution
Barrett–Joyner–Halenda (dBJH) method. X‐ray diffrac-
tion (XRD) measurements were carried out using a
Philips powder diffractometer type PW 1373 goniometer
(Cu Kα = 1.5406 Å, Philips X'pert), over an angular range
from 5° to 50°, and finally X‐ray photoelectron spectros-
copy (XPS) were recorded with a Kratos XSAM 800
(Al Kα, 1486.6 eV).
3 | RESULTS AND DISCUSSION
3.1 | Characterization of the catalyst
3.1.1 | FT‐IR analysis
FT‐IR spectra of mesoporous HPW@TiO2, Co‐Mn@TiO2,
Co‐Mn/HPW@TiO2 and reused catalyst after 5th run
are shown in Figure 1. There are four characteristic
peaks in the range of 1100 cm−1 to 800 cm−1 that may
be ascribed to HPW Keggin structure. The band at
1079 cm−1 was attributed to the stretching frequency
of P–O in the central PO4 tetrahedron. The peak at
982 cm−1 was assigned to the terminal bonds for W=Od
in the exterior WO6 octahedron. Bands at 891 cm−1 and
804 cm−1 were attributed to the W–Ob–W and W–Oc–W
2.3 | Catalyst preparation
For the synthesis of the Co‐Mn/H3PW12O40@TiO2, tita-
nium (IV) isopropoxide was used as the precursor of
TiO2. In the first step, 1 g polyvinylpyrrolidone (PVP)
was dissolved in 100 ml deionized water under magnetic
stirring at room temperature. Then, phosphotungstic acid
(0.75 g, 0.0002 mol), cobalt (II) acetate tetrahydrate

4 of 13
MASOOMI ET AL.
matrix is directly related to the preparation method used
here for preventing aggregation of the HPW crystallites
but such matrix was not able to stop leaching of the
HPW from the matrix. Although during the hydrolysis
of titanium (IV) isopropoxide in the presence of
H3PW12O40, the Keggin units were entrapped by the tita-
nia network, the surface hydroxyl groups of Ti (≡Ti–OH)
form hydrogen‐bonding with the HPW which results
in composites such as (≡TiOH2+) (H2PW12O40−) (see
Figure S1 at supporting information).[19] In such compos-
ites, hydrogen bondings such as W=Ot … HO–Ti, W–Oc …
HO–Ti, and W–Oe … HO–Ti also formed between the
oxygen atoms of Keggin ion and the ≡Ti–OH groups of
titania network.[14]
3.1.2 | XRD studies
According to reference pattern in Figure 2a, the broad dif-
fraction peaks appearing in Figure 2b‐d at 2θ = 25.2°
(101), 37.9° (103) and 48.3° (200) just could be related to
deformed anatase structure and the reason for peak
broadening is conceivably due to the HPW, cobalt oxide,
and manganese oxide in the TiO2 matrix which prevented
formation a uniform anatase crystalline TiO2.[14] No dif-
fraction peaks could be assigned to HPW, cobalt oxide,
and manganese oxide which means that these compo-
nents form too small crystallites or are well dispersed in
the mesoporous anatase TiO2 framework rather than
existing as aggregates in the matrix.[20–22] In authoriza-
tion of this comment, results obtained from EDX and ele-
mental mapping analysis (Figure 6d inset) indicate high
dispersion of HPW and other components over the TiO2
network, which could explain why XRD of the catalysts
did not show any crystalline phases related to the phos-
photungstic acid and others.[17,18,23]
FIGURE 1 FT‐IR spectra of (a) HPW@TiO2, (b) Co‐Mn@TiO2
(c) Co‐Mn/HPW@TiO2 and (d) reused Co‐Mn/HPW@TiO2 after
5th run
bridges, respectively.[12] For HPW@TiO2 and Co‐Mn/
HPW@TiO2 catalysts (Figures 1a, 1c and 1d) these peaks
are located at 805, 889, 974 and 1079 cm−1, so that the red
shift observed in the peak corresponding to W=Od bond
probably suggests the existence of hydrogen bonds
between oxygen of the W=Od bond in the Keggin
oxoanion and the ≡Ti–OH groups of titania net-
work.[14,15] The IR spectrum of the sample without
HPW shows none of these characteristic peaks (Figure 1
b). In Figure 1d, one could see that the primary Keggin
structure of the HPW has not been destroyed and remains
intact in the TiO2 network after oxidation reaction.
Furthermore, the FT‐IR spectra indicate a broad band
around 660 and 570 cm−1 that can be ascribed to Co–O
vibrations in the cobalt oxide lattice[16] which are over-
lapped with Mn–O stretching at 613 and 509 cm−1 and
Ti–O stretching at 550 cm−1 [15,17]
However, because of
.
the low quantities of Co and Mn in these catalysts these
bands were not seen clearly as distinct peaks.[16] Also,
the bands at around 1630 and 3400 cm−1 can be assigned
to the deformation modes of OH bonds in TiO2 and
adsorbed H2O.[18]
FIGURE 2 XRD patterns of (a) reference pattern of anatase
TiO2, (b) HPW@TiO2, (c) Co‐Mn@TiO2, (d) Co‐Mn/HPW@TiO2
and (e) commercial HPW
A possible mechanism for the formation of well‐
dispersed HPW and Co and Mn in the mesoporous TiO2

MASOOMI ET AL.
5 of 13
3.1.3 | TGA studies
The thermal behavior of HPW@TiO2, Co‐Mn@TiO2 and
Co‐Mn/HPW@TiO2 were investigated by TGA analysis.
As shown in Figures 3a and 3c, there are two main
weight losses in TGA curves of the HPW@TiO2 and Co‐
Mn/HPW@TiO2 catalysts. The first mass loss of 1.8
and 2.6% (between 30 °C and 160 °C) may be due to
removal of surface adsorbed water from these catalysts,
respectively. The HPW@TiO2 and Co‐Mn/HPW@TiO2
catalysts were stable in the range of about 160–500 °C,
where they show a second weight loss (1.07 and 1.7%)
after 500 °C which could be attributed to the decomposi-
tion of the heteropolyacid and more condensation of the
metal oxides in the composites. In the Co‐Mn@TiO2 the
total weight loss (2.7 wt%) could be due to elimination
of adsorbed water and different types of condensations
that might occur in the metal oxides.[7]
FIGURE 4 DR‐UV spectra of (a) HPW@TiO2, (b) Co‐Mn@TiO2
and (c) Co‐Mn/HPW@TiO2
introduction of phosphotungstic acid in TiO2 framework
and changing the coordination environment of TiO2.[14]
The UV–Vis spectrum of the catalyst containing
cobalt and manganese is displayed in Figure 4b. The
absorption bands around 215, 420 and 720 nm were
observed for cobalt oxide. The band at around 215 nm
can be assigned to the oxygen to metal charge transfer
from O2− → Co3+. The characteristic bands related to
3.1.4 | DR‐UV studies
As illustrated in Figure 4, the DR‐UV spectrum of the
HPW@TiO2 sample has shown two absorption bands
near 245 and 315 nm with a well‐defined absorption edge
at 390 nm. These features agree very well with typical
anatase TiO2 spectrum reported elsewhere,[23] and were
related to O2− → Ti4+ charge transfer transitions. The
peaks at 260 nm and 310 nm were the typical absorption
bands of HPW anion.[24] The absorption band at 260 nm
could be attributed to the ligand to metal charge transfer
from O2− ion to a W6+ ion where leads to the following
process:
the d–d transition of T1g(F) → T1g(P) of Co3+ ions in
octahedral coordination and 4A2(F) → 4T1(P) of Co2+ ions
in tetrahedral coordination in the complex of Co3O4
(≡Co2+(Co3+)2O4) spinel were shown at around 400 and
700 nm, respectively, indicating Co3O4 should be the
dominant phase for cobalt oxide encapsulated in TiO2
network with most oxidative property.[7,27] Moreover,
the manganese oxide displayed an intense band at about
260–286 nm caused by the oxygen to metal charge transi-
tion in Mn3O4 (≡ Mn2+(Mn3+)2O4) complex. The weak
shoulders at around 370 and 470 nm could be assigned
4
4
Â
Ã
Â
Ã
6þ–O2−–W6þ → W5þ–O−–W6þ
5
5
5
5
W
to d–d transitions of the B1g → B2g and B1g → Eg of
Mn3+ in Mn2O3, and the weak peak at around 640 nm
The produced excited state, i.e. [W5+–O−–W6+] has a
strong oxidation ability for oxidation of the xylene.[25] In
addition to this, there is an absorption band at 310 nm
which is also characteristic of H3PW12O40.[26] The broad-
ening of the absorption bands could be due to the
6
corresponded to A1g
→
4T2g transition of Mn2+ in
MnO.[7,28] Therefore, a mixture of manganese oxides
(Mn3+ and Mn2+) was dispersed in the calcined Co‐
Mn@TiO2 catalyst.
3.1.5 | XPS studies
X‐ray photoelectron spectroscopy (XPS) is a powerful
technique used to investigate the chemical properties of
the species in HPW@TiO2, Co‐Mn@TiO2, Co‐Mn/
HPW@TiO2 and the reused Co‐Mn/HPW@TiO2 after
5th run. As shown in Figure 5, the studied XPS detailed
regions were Ti 2p, W 4d, Co 2p, Mn 2p and O 1 s. The
XPS of Ti 2p levels for the all samples are presented in
Figure 5a which show the doublet arising from spin‐orbit
splitting Ti 2p3/2 and Ti 2p1/2. The analysis of the Ti 2p3/2
level (more intense signal) shows the binding energy
value at 458.7 eV that is typical of Ti4+[29] whereas, in
FIGURE 3 TG curves of (a) HPW@TiO2, (b) Co‐Mn@TiO2 and
(c) Co‐Mn/HPW@TiO2

6 of 13
MASOOMI ET AL.
the reused Co‐Mn/HPW@TiO2 catalyst after 5th run, it
shifted to 459.1 eV. The XPS spectrum in Figure 5b shows
the most intense W 4d5/2 component of the W 4d doublet
at a binding energy of 247.5 eV which confirms the oxida-
tion state of +6 (W6+) for tungsten in the catalysts.
The spectrum in Figure 5c exhibits an intense doublet
due to the two spin–orbit components for Co 2p3/2
(780.8 eV) and Co 2p1/2 (795.5 eV). From the spin‐orbital
splitting value of the 2p level one could measure the oxi-
dation of the metal in the catalyst; the spin–orbit splitting
for Co3O4 is 15 eV, while it is 16 eV for CoO.[16] In the
prepared catalyst in this work the spin–orbit splitting of
14.7 eV is consistent with the Co3O4.[16,30,31] Clearly, Co
2p is detected, and quantified in the Co‐Mn@TiO2 and
Co‐Mn/HPW@TiO2 samples. On the contrary, in the
reused Co‐Mn/HPW@TiO2 catalyst after five consecutive
runs, its presence is ‘suspected’ but cannot be quantified
and it is probably duo to leaching of some of the cobalt
after five recycles which coincides with the result of ICP
studies (see section ICP Studies). Furthermore, Mn was
not detected by XPS after an accumulation of 50 sweeps
suggesting that the Mn surface concentration is lower
than the detection limit of the XPS analysis (Figure 5d)
and this result is consistent with the conclusion obtained
from the XRD analysis.
Finally, we must mention that the analysis of the O
1 s core levels (see Figure S2) reveals a peak at 530.2 eV
assigned to oxygen bound to metallic ions (mainly Ti4+
)
except in reused Co‐Mn/HPW@TiO2 after 5th run where
the region is dominated by higher binding energy compo-
nents (531.5 eV) corresponding to organic oxygen (bound
to carbon) and commonly to OH from hydroxyl ions.[29]
3.1.6 | FESEM, EDX and elemental map-
ping studies
The FESEM images of the HPW@TiO2, Co‐Mn@TiO2
and Co‐Mn/HPW@TiO2 samples in magnification of
500 nm are shown in Figure 6, and in other magnifica-
tions are demonstrated in Figures S3 and S4. As seen,
they have approximately similar surface morphology
and roughness. All of the samples are narrowly distrib-
uted and approximately having a uniform particle size.
These results indicate that the encapsulation method
developed in this work is highly efficient on increasing
the surface to volume ratio of the catalyst. To demon-
strate the presence of HPW, Co and Mn oxides in the
catalyst, elemental analysis was done using EDX and
the characteristic peaks were identified (Figure 6d).
Clearly, the analysis shows the presence of Ti, Co, Mn,
P, W, and O in the Co‐Mn/HPW@TiO2 catalyst. How-
ever, the XRD patterns do not show any peaks related
to cobalt oxide, manganese oxide and phosphotungstic
FIGURE 5 XPS patterns of (a) Ti 2p, (b) W 4d, (c) Co 2p and (d)
Mn 2p; For all samples: (I) HPW@TiO2, (II) Co‐Mn@TiO2 (III) Co‐
Mn/HPW@TiO2 and (IV) reused Co‐Mn/HPW@TiO2 after 5th run

MASOOMI ET AL.
7 of 13
FIGURE 6 FESEM images of (a) HPW@TiO2, (b) Co‐Mn@TiO2 and (c)Co‐Mn/HPW@TiO2 at magnification of 500 nm, (d) EDX and
elemental mapping images (inset) of Co‐Mn‐HPW@TiO2 catalyst
acid particles. Therefore, one might conclude that HPW,
cobalt oxide and manganese oxide are uniformly dis-
persed in the anatase matrix. According to the elemental
mapping analysis (Figure 6d inset), all of the elements
have been discovered with a homogeneous and uniform
distribution in the Co‐Mn/HPW@TiO2 catalyst.
particles in the HPW@TiO2, Co‐Mn@TiO2 and Co‐Mn/
HPW@TiO2 catalysts were found to be around 4, 7 and
5 nm in diameter, respectively (Size distribution histo-
gram was shown in Figure S5).
3.1.8 | N2 adsorption–desorption iso-
therms studies
3.1.7 | TEM studies
TEM images show the presence of overall and uniform
nanoparticles in the as‐prepared catalysts in different
magnifications (20, 60 and 100 nm) in Figure 7a–i. The
images also show that most of the nanoparticles have
spherical shape which are coherent together. We could
judge that the mesoporous structure is obtained by
the agglomeration of mono‐dispersed TiO2 particles; the
particles dispersion is obvious on the TiO2 matrix.
This dispersion of particles is due to the use of PVP in
preparation of the catalyst and by burning PVP during
calcination of the catalyst at 450 °C, leads to a uniform
dispersion of HPW, manganese and cobalt oxides in
the anatase TiO2 matrix. The average sizes of most
The specific surface area is an important representative of
a catalyst. The specific surface area, pore diameter and
pore volume of the catalysts were determined by BET
and dBJH methods. The specific surface area of the
HPW@TiO2, Co‐Mn@TiO2 and Co‐Mn/HPW@TiO2 cat-
alysts were 109, 122 and 90 m2/g, respectively (Table 1).
The porosities of these catalysts were investigated by
nitrogen adsorption–desorption techniques. According
to IUPAC definition the nitrogen adsorption–desorption
isotherms show type IV behavior for all samples with
steep hysteresis loop at high relative pressure, which is
typical for mesoporous materials (Figure 8a–c). As dem-
onstrated in Table 1, pore diameter (~2.6 nm) and pore

8 of 13
MASOOMI ET AL.
FIGURE 7 TEM images of (a‐c) HPW@TiO2, (d‐f) Co‐Mn@TiO2 and (g‐i) Co‐Mn/HPW@TiO2 at magnification of 100, 60 and 20 nm
TABLE 1 Textural properties of the prepared catalysts
Entry
Catalyst
SBET (m2/g)
Vpore (cm3/g)
Dpore (nm)
1
2
3
HPW@TiO2
109
122
90
0.199
0.359
0.194
2.6
2.6
2.6
Co‐Mn@TiO2
Co‐Mn/HPW@TiO2
volume (~0.1–0.3 cm3/g) of the catalysts were calculated
using the dBJH method (Figure 8a–c, inset).
3.1.9 | ICP studies
To determine the contents of W, Co and Mn in the Co‐
Mn/HPW@TiO2 catalyst before and after five consecutive
runs, ICP‐OES technique was applied and the data of
analysis were shown in Table 2. As observed, the fresh
catalyst had W (2.67 w%), Co (0.917 w%) and Mn (0.099
w%), which perfectly agrees with the theoretical calcula-
tions (Co/Mn ratio is 10/1). From the ICP data, we could
judge that the catalyst has lost 1.793 wt% of HPW,
FIGURE
8
N2 adsorption‐desorption isotherms of (a)
HPW@TiO2, (b) Co‐Mn@TiO2 and (c) Co‐Mn/HPW@TiO2 and
dBJH‐Plots pore size distribution of these samples (inset)

MASOOMI ET AL.
9 of 13
TABLE 2 ICP analysis results of fresh and reused Co‐Mn/HPW@TiO2 catalyst
Entry
Catalyst
Co (wt %)
Mn (wt %)
W (wt %)
1
2
Co‐Mn/HPW@TiO2 (Fresh)
0.917
0.698
0.099
0.069
2.67
Co‐Mn/HPW@TiO2 (After 5 runs)
0.877
0.219 wt% cobalt and 0.03 wt% manganese after 5 runs.
These results indicate that the activity of this catalyst
remained at a satisfactory state after at least five cycles.
to the instability of the heteropoly acid at high tempera-
tures which shows itself in the selectivity of the main
product; therefore, at 370 °C this catalyst exhibited a high
conversion and selectivity.
3.2 | Catalytic performance
3.2.1 | Effect of temperature
3.2.2 | Effect of WHSV
Weight hourly space velocity (WHSV) plays an important
role in this type of reactions. The influence of feed rate of
the reactant (o‐xylene) over Co‐Mn/HPW@TiO2 catalyst
was studied by using 2 ml o‐xylene at 370 °C. Hence,
the amount of catalyst in each run was determined by
the selected WHSV defined as follows:
The reaction temperature has a great impact on both con-
version and selectivity. As shown in Table 3 the effect of
temperature on this reaction was studied by changing
the temperature from 300 °C to 450 °C while the other
parameters remained unchanged. The conversion and
selectivity of phthalic anhydride increased when the tem-
perature reached to 370 °C. Beyond this temperature the
efficiency of Co‐Mn/HPW@TiO2 slightly decreased due
“Mass flow of o−xylene ðg=hÞ”
WHSV ¼
ð“Weight of catalyst used ðg”ÞÞ
TABLE 3 Effect of temperature on the oxidation of o‐xylene
Selectivity (%)
Entry
Temperature (°C)
Conversion (%)
1
2
3
4
5
6
7
8
300
330
350
370
390
410
430
450
0.2
22.4
64.6
88.2
56.1
54.2
51.2
36.4
88.7
10.3
7.6
11.2
44
0
0
28
17.5
48.8
75.5
52.3
45.1
43.0
3.4
20.6
4.2
22.9
18.1
24.5
27.3
29.4
16
2.1
4.2
18.8
19.1
15.4
14.4
8.2
7.3
66.1
Reaction conditions: o‐xylene: 2 ml, WHSV: 0.5 h−1, O2 flow rate: 10 ml min−1, amount of catalyst: 0.44 g selected catalyst (Co‐Mn/HPW@TiO2).
TABLE 4 Effect of WHSV on the oxidation of o‐xylene
Selectivity (%)
Entry
WHSV (h−1
)
Conversion (%)
1
2
3
4
0.25
0.5
1
78.3
88.2
38.1
22.5
5.5
2.1
8
10.7
4.2
23.3
18.1
22.7
22.7
60.3
75.5
42.8
35.7
26.4
29.3
1.5
12
Reaction conditions: o‐xylene: 2 ml, temperature: 370 °C, O2 flow rate: 10 ml min−1, amount of catalyst: 0.44 g selected catalyst (Co‐Mn/HPW@TiO2)

10 of 13
MASOOMI ET AL.
TABLE 5 Effect of O2 flow rate on the oxidation of o‐xylene
Selectivity (%)
Entry
O2 flow rate (ml min−1
)
Conversion (%)
1
2
3
5
10
20
41.6
88.2
20.6
6.1
2.1
8.4
37.3
4.2
25.4
18.1
28.4
31
75.5
18.8
44.2
Reaction conditions: o‐xylene: 2 ml, temperature: 370 °C, WHSV: 0.5 h−1, amount of catalyst: 0.44 g of the selected catalyst (Co‐Mn/HPW@TiO2).
TABLE 6 Effect of HPW and Co‐Mn load
Selectivity (%)
Entry
Catalyst
Conversion (%)
1
2
3
4
5
6
7
8
9
Co‐Mn (1%)/HPW (15%)@TiO2
Co‐Mn (2%)/HPW (15%)@TiO2
Co‐Mn (4%)/HPW (15%)@TiO2
Co‐Mn (2%)/HPW (10%)@TiO2
Co‐Mn (2%)/HPW (25%)@TiO2
Co‐Mn (2%)/HPW (15%)@TiO2*
HPW (15%)@TiO2
30.3
88.2
66.7
36.4
58.7
51
10.8
2.1
3.1
9
23.5
4.2
28.6
18.1
24.2
33.1
28.7
24.3
29
36.9
75.5
61.3
35.3
53
11.3
22.4
14.7
25.4
28.1
31.7
‐
3.5
9.3
8.8
9.5
‐
41
43.3
48.1
0
34
Co‐Mn (2%)@TiO2
26.7
‐
32
Blank
‐
Reaction conditions: o‐xylene: 2 ml, temperature: 370 °C, WHSV: 0.5 h−1, O2 flow rate: 10 ml min−1, amount of catalyst: 0.44 g, Co/Mn ratio: 10/1 in all samples,
*the catalyst was prepared in the absence of PVP.
While the catalyst weight was kept at 0.44 g, the con-
tact time was varied by changing the feed rate, for exam-
ple, 125, 250, 500 and 1000 μl/h (i.e. 0.25–1.5 h−1). On the
basis of the experimental data, 0.5 h−1 was chosen as
selected WHSV for the oxidation reaction (Table 4). In
addition, the conversion and selectivity decreased by
increasing of WHSV from 0.5 to 1.5 h−1. According to def-
inition of WHSV, mass flow has a reverse relation with
weight of the catalyst, therefore increasing the mass flow
rate has a direct relation with decreasing the weight of
catalyst and decreasing the oxidative sites which is
consistent with the results presented in Table 4.
quartz tubular reactor and the catalyst do not have
enough time to oxidize o‐xylene. So, the optimum oxygen
flow rate for this oxidation reaction is 10 ml min−1and
lower than this flow rate, oxidation reaction faces to lack
3.2.3 | Effect of oxygen flow rate
The effect of oxygen flow rate on the o‐xylene conversion
was studied and it was measured with gas flow meter
(i.e., ml/min). According to the experimental results
shown in Table 5, the conversion decreased with an
increase of oxygen flow rate, because the oxygen passes
quickly through the catalyst in a fixed‐bed down flow
FIGURE 9 Reusability of the Co‐Mn/HPW@TiO2 catalyst over
oxidation of o‐xylene to phthalic anhydride

MASOOMI ET AL.
11 of 13
of the oxidizing agent which has led to decrease in
conversion.
study we decided to start with the ratio of Co/Mn = 10/
1, and optimize the loading of HPW and Co‐Mn oxides.
The Co‐Mn/HPW@TiO2 catalyst with different HPW,
cobalt, and manganese oxides loads were compared for
the o‐xylene oxidation at 370 °C, WHSV = 0.5 h−1, O2
flow rate = 10 ml min−1. Table 6 shows the o‐xylene oxi-
dation for five catalysts. The change of HPW and cobalt
and manganese oxides affected on conversion and selec-
tivity. The maximum o‐xylene conversion and phthalic
anhydride selectivity appeared on the catalyst with
3.2.4 | Effect of HPW and Co/Mn load
The efficiency of the Co‐Mn/HPW@TiO2 catalyst can be
improved by optimizing the HPW and manganese and
cobalt oxides loads. Our experience[6] had shown that
the Co/Mn ratio could influence the catalytic activity of
the catalyst in oxidation of o‐xylene. Therefore, in this
TABLE 7 Catalytic activities of Co‐Mn/HPW@TiO2 and some previously reported catalysts in the o‐xylene oxidation
Entry
1
Catalyst
Temperature (°C)
Conversion (%)
Phthalic anhydride selectivity (%)
Ref.
[32]
a
V2O5/TiO2
380–400
300
350
387
460
435
410
440
400
450
280
160
249
240
220
250
370
370
370
100
60–70
90.2
42
70–75
80–90
87.3
34
b
c
[1]
2
V‐Mn‐MCM‐41
V‐Mo‐MCM‐41
[33]
[2]
3
d
4
AgMoVO6
e
[34]
[35]
[36]
[36]
[36]
[37]
[38]
[39]
[40]
[41]
[42]
[43]
5
V2O5‐P2O5/TiO2
99.8
76
72.3
57
f
6
V2O5/TiO2–ZrO2
g
7
V/ZrO2
99.2
13.2
96.2
100
100
100
90
35.2
0.9
43.5
50–60
50
g
8
Sb/ZrO2
g
9
Sb‐V/ZrO2
h
10
11
12
13
14
15
16
17
18
19
V2O5/ZrO2
Vanadium‐PTGi
j
Pd/Al2O3
‐
k
Pd/Co3O4
‐
l
Mn‐CeOx
100
100
90
‐
m
α‐MnO2
‐
n
Au‐Pd/3DOM Mn2O3
‐
o
HPW@TiO2
43.3
48.1
88.2
34
This work
This work
This work
o
Co‐Mn@TiO2
32
o
Co‐Mn/HPW@TiO2
75.5
Reaction conditions:
ao‐xylene: 0.54 mol%, amount of catalyst: 0.5 g.
bamount of catalyst: 0.3 g, WHSV: 0.5 h−1, air flow rate: 0.02 mol h−1, V/Mn ratio = 50:50.
camount of catalyst: 0.2 g, feed flow rate ratio: 0.016 cc, o‐xylene: 200 cc, oxygen gas per minute, contact time (W/F0): 24.41 h, Si/(V + Mo) ratio = 46.
dfeed composition = o‐xylene: 1 vol %, O2: 20 vol%, H2O: 5 vol %, GHSV: 3000 h−1
eamount of catalyst: 2100 g, V2O5 (5.9 wt%), P2O5 (0.25 wt%), and TiO2 (93.8 wt.%), flow rate: 4.5 Nm3 h−1
famount of catalyst: 0.3–0.5 g, feed flow rate: 1.0–2.0 ml h−1, O2 flow rate: 80–90 ml min−1
.
.
.
gamount of catalyst: 0.06 g, (W/F) was 42 g.s.L−1, The molar concentrations of o‐xylene and oxygen in the feed were 0.8 and 20.8% respectively (78.4% nitrogen).
hTemperature: 310, 330, 350, 370, 390 and 410 °C and different WHSV: 1500, 2000 and 2500 h−1, initial concentration of o‐xylene − Cxylene = 42 g/nm3.
iPorous titania glass (PTG) has been loaded with vanadia.
jo‐xylene: 100 ppm, flow rate: 100 ml min−1 (20 vol% O2:80 vol% N2), GHSV: 50,000 h−1
kamount of catalyst: 0.1 g, o‐xylene: 150 ppm, flow rate: 100 ml min−1(20 vol% O2:80 vol% N2), GHSV: 60,000 h−1
lamount of catalyst: 0.25 g, o‐xylene: 700 ppm, flow rate: 50 ml min−1(20 vol% O2:80 vol% N2), GHSV: 8,000 h−1
mo‐xylene: 700 ppm, flow rate: 50 ml min−1 (20 vol% O2:80 vol% N2), GHSV: 8,000 h−1
namount of catalyst: 0.05 g, o‐xylene: 1000 ppm, flow rate: 16.6 to 166.0 ml min−1 (20 vol% O2:80 vol% N2), GHSV: 9,100–191,000 h−1
oo‐xylene: 2 ml, amount of catalyst: 0.44 g, temperature: 370 °C, WHSV: 0.5 h−1, O2 flow rate: 10 ml min−1
.
.
.
.
.
.

12 of 13
MASOOMI ET AL.
15 wt% HPW, and 2 wt% cobalt and manganese oxides.
Beyond this load the conversion and selectivity decreased
probably because of aggregation of the effective catalysts.
It should be mentioned that when the optimized catalyst
was prepared in the absence of PVP in solution the con-
version decreased appreciably (Table 6, entry 6); naturally
one would expect that PVP causes more dispersion of the
catalyst components in the TiO2 matrix. In addition, it
should be mentioned that Co‐Mn/HPW@TiO2 catalyst
was very superior to the HPW@TiO2 and Co‐Mn@TiO2
catalysts due to the synergistic effects between cobalt,
manganese, and tungsten (Table 6, entries 2, 7 and 8).
The support without HPW, cobalt and manganese oxides
showed no activity.
main product. (4) The conversion of o‐xylene to phthalic
anhydride increased by enhancing the temperature, and
the highest selectivity to main product was obtained at
a reaction temperature of 370 °C. (5) The catalyst was
reasonably active even after five consecutive runs.
ACKNOWLEDGEMENT
We are grateful for the financial support from the
Research Council of Isfahan University of Technology
(IUT), Iran.
ORCID
Mehran Ghiaci http://orcid.org/0000-0002-0686-7778
3.2.5 | Reusability of the catalyst
REFERENCES
The stability of the Co‐Mn/HPW@TiO2 catalyst was stud-
ied in the oxidation of o‐xylene by applying the catalyst in
consecutive runs. By separating the catalyst after each
run, it was washed with ethyl acetate and dried in an
oven for 1 h at 80 °C. Then, the recovered catalyst was
charged into the reactor, and preheated at 250 °C for
4 h in a flow of oxygen (10 ml min−1) at atmospheric
pressure. According to the results shown in Figure 9, both
conversion and selectivity remained at a satisfactory level
after five runs. As mentioned in section ICP studies, the
catalyst after five consecutive runs lost a percentage of
its components and in this respect the activity of the cat-
alyst decreased slightly. These results indicate that the
catalyst is reusable, stable and has strong oxidation ability
even after five runs.
[1] C. Mahendiran, T. Maiyalagan, P. Vijayan, C. Suresh, K.
Shanthi, Kinet. Mech. Cat. 2012, 105, 469.
[2] A. Karpov, C. Deissler, C. K. Dobner, H. Hibst, G. Cox, N.
Brem, S. A. Schunk, R. E. Dinnebier, F. Rosowski,
ChemCatChem 2010, 2, 1562.
[3] M. Gimeno, J. Gascón, C. Téllez, J. Herguido, M. Menéndez,
Chem. Eng. Process. 2008, 47, 1844.
[4] A. Rahimi, S. Hamidi, CPPM. 2011, 6.
[5] R. P. Chaplin, M. S. Wainwright, F. Yay, H. A. Ali, J. Mol.
Catal. 1987, 39, 359.
[6] M. Ghiaci, M. Mostajeran, A. Gil, Ind. Eng. Chem. Res. 2012, 51,
15821.
[7] W. hua Yu, C. hui Zhou, D. shen Tong, T. ning Xu, J. Mol,
Catal. A: Chem. 2012, 365, 194.
The data in Table 7, compare the results obtained over
Co‐Mn/HPW@TiO2 catalyst and other catalytic systems
such as V2O5/TiO2,[32] V‐Mn‐MCM‐41,[1] V‐Mo‐MCM‐
[8] B. Solsona, T. E. Davies, T. Garcia, I. Vázquez, A. Dejoz, S. H.
Taylor, Appl. Catal. Environ. 2008, 84, 176.
41,[33] V2O5/TiO2–ZrO2,[35] Pd/Al2O3,[39] Sb‐V/ZrO2,[36]
[9] H.‐f. Lv, S.‐q. Wu, N. Liu, X.‐l. Long, W.‐k. Yuan, Chem. Eng. J.
2011, 172, 1045.
[37]
and V2O5/ZrO2
which have been reported for vapor‐
[10] L. A. Sofia, A. Krishnan, M. Sankar, N. Kala Raj, P.
Manikandan, P. Rajamohanan, T. Ajithkumar, J. Phys. Chem.
C 2009, 113, 21114.
phase oxidation of o‐xylene to phthalic anhydride in a
fixed‐bed reactor. Clearly, the catalyst used in the present
study might be one of the best catalysts for producing
phthalic anhydride.
[11] N. A. Khan, S. H. Jhung, Fuel Process. Technol. 2012, 100, 49.
[12] F. Nezampour, M. Ghiaci, H. Farrokhpour, Appl. Catal. Gen.
2017, 543, 104.
4 | CONCLUSION
[13] N. Sammah, M. Ghiaci, Ind. Eng. Chem. Res. 2017, 56, 10597.
[14] Y. Yang, Q. Wu, Y. Guo, C. Hu, E. Wang, J. Mol. Catal. A:
Chem. 2005, 225, 203.
The following conclusions can be highlighted: (1) the
Co‐Mn/HPW@TiO2 catalyst was found to be the most
efficient catalyst among the investigated catalysts for oxi-
dation of o‐xylene to phthalic anhydride. (2) The presence
of PVP in aqueous solution when preparing the catalyst
lead to higher activity for the oxidation reaction. (3) The
loading of HPW, cobalt and manganese oxides play an
important role in the conversion and selectivity of the
[15] D. Huang, Y. Wang, L. Yang, G. Luo, Microporous Mesoporous
Mater. 2006, 96, 301.
[16] J. Taghavimoghaddam, G. P. Knowles, A. L. Chaffee, J. Mol.
Catal. A: Chem. 2012, 358, 79.
[17] Z.‐Y. Tian, P. Mountapmbeme Kouotou, N. Bahlawane, P. H.
Tchoua Ngamou, J. Phys. Chem. C 2013, 117, 6218.

MASOOMI ET AL.
13 of 13
[18] R. Pochamoni, A. Narani, M. Varkolu, M. D. Gudimella, S. S. P.
Potharaju, D. R. Burri, S. R. R. Kamaraju, J. Chem. Sci. 2015,
127, 701.
[34] V. Nikolov, A. Anastasov, D. Elenkov, G. Ganev, N. Dimitrov,
Chem. Eng. Process. 1988, 24, 157.
[35] K. N. Rao, P. Venkataswamy, P. Bharali, H. P. Ha, B. M. Reddy,
Res. Chem. Intermed. 2012, 38, 733.
[19] S. Ullah, J. J. S. Acuña, A. A. Pasa, S. A. Bilmes, M. E. Vela, G.
Benitez, U. P. Rodrigues‐Filho, Appl. Surf. Sci. 2013, 277, 111.
[36] C. Pieck, S. Del Val, M. L. Granados, M. Banares, J. Fierro,
Catal. Lett. 2003, 89, 27.
[20] H. Hamadi, M. Kooti, M. Afshari, Z. Ghiasifar, N. Adibpour,
J. Mol. Catal. A: Chem. 2013, 373, 25.
[37] L. Makedonski, Bulgarian Chem. Commun. 2009, 41, 303.
[21] P. Ferreira, I. Fonseca, A. Ramos, J. Vital, J. Castanheiro, Cat.
Com. 2011, 12, 573.
[38] C. Y. Hong, R. P. Cooney, R. F. Howe, Appl.Catal. 1991, 75,
237.
[39] S. Huang, C. Zhang, H. He, J. Environ. Sci. 2013, 25, 1206.
[22] M. Kooti, M. Afshari, Mater. Res. Bull. 2012, 47, 3473.
[40] Y. Wang, C. Zhang, Y. Yu, R. Yue, H. He, Catal. Today 2015,
242, 294.
[23] J. M. Gallardo‐Amores, T. Armaroli, G. Ramis, E. Finocchio, G.
Busca, Appl. Catal. Environ. 1999, 22, 249.
[41] Y. Wu, Y. Zhang, M. Liu, Z. Ma, Catal. Today 2010, 153, 170.
[24] X.‐M. Yan, G.‐S. Su, L. Xiong, J. Fuel, Chem. Tech. 2009, 37,
318.
[42] Y. Wu, Y. Lu, C. Song, Z. Ma, S. Xing, Y. Gao, Catal. Today
2013, 201, 32.
[25] C. Jiang, Y. Guo, C. Hu, C. Wang, D. Li, Mater. Res. Bull. 2004,
39, 251.
[43] S. Xie, Y. Liu, J. Deng, X. Zhao, J. Yang, K. Zhang, Z. Han, H.
Arandiyan, H. Dai, Appl. Catal. Environ. 2017, 206, 221.
[26] G. R. Rao, T. Rajkumar, J. Colloid Interface Sci. 2008, 324, 134.
[27] A. Katsoulidis, D. Petrakis, G. Armatas, P. Trikalitis, P.
Pomonis, Microporous Mesoporous Mater. 2006, 92, 71.
SUPPORTING INFORMATION
[28] Q. Tang, S. Hu, Y. Chen, Z. Guo, Y. Hu, Y. Chen, Y. Yang,
Additional supporting information may be found online
in the Supporting Information section at the end of the
article.
MicroporousMesoporous Mater. 2010, 132, 501.
[29] L. P. Rivoira, M. L. Martínez, H. Falcón, A. R. Beltramone, J.
M. Campos‐Martin, J. L. Fierro, P. Tartaj, ACS Omega. 2017,
2, 2351.
[30] J. Janas, T. Machej, J. Gurgul, R. P. Socha, M. Che, S. Dzwigaj,
How to cite this article: Masoomi K, Ghiaci M,
Botelho do Rego AM. Selective vapor‐phase
oxidation of o‐xylene to phthalic anhydride over
Co‐Mn/H3PW12O40@TiO2 using molecular oxygen
as a green oxidant. Appl Organometal Chem. 2018;
e4461. https://doi.org/10.1002/aoc.4461
Appl. Catal. Environ. 2007, 75, 239.
[31] Q. Tang, Q. Zhang, H. Wu, Y. Wang, J. Catal. 2005, 230, 384.
[32] M. S. Wainwright, N. R. Foster, Cat. Rev. ‐ Sci. Eng. 1979, 19,
211.
[33] M. Selvaraj, T. Lee, Microporous Mesoporous Mater. 2005, 85,
39.