Role of Surface Carboxylates in the Gas Phase Ozone-Assisted Catalytic Oxidation of Toluene

Article abstract of DOI:10.1007/s10562-017-2143-0

Abstract: Ozone-assisted catalytic oxidation of toluene was conducted over MnO_x/γ-Al₂O₃ to identify and differentiate the role of reaction byproducts. It was found that not only alumina acted as a reservoir for toluene, but also it interacted effectively with toluene to create surface carboxylate intermediates. Surface carboxylates were essential for an effective oxidation process, and they did not directly cause catalyst deactivation. The presence of Mn sites was necessary for further oxidation of the surface carboxylates. At 90 °C, a stable catalytic activity with 95% conversion was achieved. However, at 25 °C, byproducts such as acetic acid and formic acid accumulated on the surface of the catalyst and decreased the catalyst activity. Graphical Abstract: [Figure not available: see fulltext.].

Full text of DOI:10.1007/s10562-017-2143-0

Catal Lett
DOI 10.1007/s10562-017-2143-0
Role of Surface Carboxylates in the Gas Phase Ozone-Assisted
Catalytic Oxidation of Toluene
Mostafa Aghbolaghy1 · Jafar Soltan1 · Ning Chen2,3
Received: 28 April 2017 / Accepted: 10 July 2017
© Springer Science+Business Media, LLC 2017
Abstract Ozone-assisted catalytic oxidation of toluene
was conducted over MnOx/γ-Al2O3 to identify and differ-
entiate the role of reaction byproducts. It was found that
not only alumina acted as a reservoir for toluene, but also
it interacted effectively with toluene to create surface car-
boxylate intermediates. Surface carboxylates were essential
for an effective oxidation process, and they did not directly
cause catalyst deactivation. The presence of Mn sites was
necessary for further oxidation of the surface carboxylates.
At 90°C, a stable catalytic activity with 95% conversion
was achieved. However, at 25°C, byproducts such as ace-
tic acid and formic acid accumulated on the surface of the
catalyst and decreased the catalyst activity.
Graphical Abstract
Keywords Mechanism · Intermediate · Ozonation ·
VOC · Deactivation · DRIFTS
Electronic supplementary material The online version of this
article (doi:10.1007/s10562-017-2143-0) contains supplementary
material, which is available to authorized users.
1 Introduction
Volatile organic compounds (VOCs) are important indoor
and outdoor air pollutants. VOCs can have adverse health
effects on eyes, skin, lungs and the central nervous system
[1]. A number of techniques have been used for removal
of VOCs from air. The most common methods are adsorp-
tion [2], photocatalytic oxidation [3], non-thermal plasma
[4], biological treatment [5], catalytic oxidation and ozone-
assisted catalytic oxidation (i.e. catalytic ozonation) [6].
Among these, catalytic ozonation has gained increasing
attention in the recent years due to its advantages over
* Jafar Soltan
j.soltan@usask.ca
1
Department of Chemical and Biological Engineering,
University of Saskatchewan, 57 Campus Drive, Saskatoon,
SK S7N 5A9, Canada
2
Canadian Light Source Inc., University of Saskatchewan, 101
Perimeter Road, Saskatoon, SK S7N 0X4, Canada
3
Department of Geological Sciences, University
of Saskatchewan, 114 Science Place, Saskatoon,
SK S7N 5E2, Canada
Vol.:(0123456789)
1 3

M. Aghbolaghy et al.
catalytic oxidation. Compared to catalytic oxidation, the
reaction temperature for catalytic ozonation is significantly
lower, leading to energy saving. In addition, catalysts based
on transition metal oxides have high efficiencies in removal
of VOCs by catalytic ozonation, eliminating the need for
expensive noble metals that are commonly used in catalytic
oxidation of VOCs by oxygen [7].
with no or negligible catalyst deactivation [7, 18, 20]. Study
of catalytic ozonation of VOCs at a relatively high tempera-
ture, with stable catalytic activity, and comparison of the
reaction characteristics with those at the room temperature
reaction, with catalyst deactivation, can provide more clear
understanding of the reaction intermediates and catalyst
stability. It is important to investigate the nature and roles
of intermediates and byproducts that are formed on the
surface of the catalyst during room temperature operation.
This can help us develop a better understanding of the reac-
tion mechanism and catalyst stability.
Manganese oxides (MnOx) are the most active metal
oxides for gas phase VOC removal in the presence of ozone
[8]. Manganese oxides supported on different supports
have been used for catalytic ozonation of acetone [9], ben-
zene [10], toluene [11], formaldehyde [12], cyclohexane
[13] and chlorobenzene [14]. It has been shown that cata-
lytic ozonation can oxidize the listed VOCs at tempera-
tures below 100°C. Supported manganese oxide possesses
higher catalytic activity than unsupported manganese oxide
[15]. In addition, studies on catalytic ozonation of VOCs,
using manganese oxide supported on γ-alumina, silica, tita-
nia, and zirconia, found that VOC removal rates normalized
by surface area of the catalysts were comparable or slightly
lower than those with the γ-alumina supported catalysts
[16–18].
In this work, catalytic ozonation of toluene was
investigated at two temperatures of 25 and 90°C on
MnOx/γ-alumina catalyst. A combination of in-situ diffuse
reflectance infrared Fourier transform spectroscopy, chemi-
cal analysis of the reaction products by GC-MS analysis,
and a number of temperature programmed techniques were
used to investigate the reaction pathways by differentiating
the roles of reaction intermediates and byproducts that are
formed on the surface of the catalyst.
Catalytic ozonation of VOCs at room temperature is
advantageous. However, catalyst deactivation is a serious
impediment. There have been a number of studies aimed
at understanding the nature of catalyst deactivation and for-
mation of reaction byproducts in catalytic ozonation. Reed
et al. [9] used in-situ Raman spectroscopy to study cata-
lytic ozonation of acetone on 10 wt% MnOx/SiO2. It was
concluded that acetone was physisorbed molecularly on
the silica surface. It was suggested that the silica support
was a reservoir for the adsorbed acetone, then the adsorbed
acetone migrated to Mn site and reacted with atomic oxy-
gen species produced from ozone decomposition. Einaga
et al. used in-situ Fourier transform infrared (FTIR) spec-
troscopy to investigate room temperature catalytic ozo-
nation of benzene over 5 wt% Mn/USY zeolite [19] and
5 wt% MnO2/Al2O3 [16]. It was observed that two types of
intermediates were formed on the surface of the catalysts.
These intermediates were categorized as weakly bound and
strongly bound compounds. Zhao et al. [12] employed in-
situ FTIR spectroscopy to investigate room temperature
catalytic ozonation of formaldehyde on MnOx catalyst.
Bidentate and monodentate carbonate species were found
on the catalyst surface and both increased gradually with
time. These studies have limited their investigations to ozo-
nation operating temperatures in the range of 22–35°C,
at which severe catalyst deactivation occurs. In addition,
they have not identified a mechanism for production of the
observed compounds, and therefore, have not differentiated
between the roles of reaction intermediates and byproducts.
It is known that catalytic ozonation of VOCs at a rela-
tively high temperature (i.e. 90°C) shows stable activity
2 Materials and Methods
2.1 Experimental Setup
Figure 1 shows a schematic diagram of the experimen-
tal setup that was used for catalyst activity and gas phase
analyses. A gas cylinder with ppm level concentration
of toluene balanced with nitrogen (Praxair, accuracy of
2%) was used to supply toluene. An ozone generator
(AZCO Industries LTD, HTU-500S) was employed to
produce ozone from a high purity oxygen stream (Praxair,
99.993%). Nitrogen gas was supplied from a high purity
nitrogen cylinder (Praxair, 99.999%) for dilution and
purging purposes. Gas flows were controlled by mass
flow controllers (Brooks, SLA 5850, accuracy of 1%).
Total flow rate of 250 ml/min (at STP) was used for all
experiments. Gas streams were combined and passed
through a horizontal Pyrex tube, filled with glass beads,
to improve mixing before entering the reactor. A reac-
tion chamber (Harrick, HVC) was used as an atmospheric
pressure reactor to conduct all the reaction experiments.
The reaction chamber was equipped with a heater that
allowed operations up to 900 °C. Exhaust gas from the
reactor was analyzed by a long-path gas cell (PIKE, vol-
ume 0.1 l 2.4 m optical length, KBr window) coupled
with a Nicolet iS50 FTIR spectrometer equipped with a
Deuterated l-alanine doped triglycine sulfate (DLaTGS)
detector. Spectra were collected at a resolution of 4 cm−1
in the range of 4000–400 cm−1. Homogeneous reaction
between ozone and toluene was negligible in this system.
1 3

Role of Surface Carboxylates in the Gas Phase Ozone-Assisted Catalytic Oxidation of Toluene
Fig. 1 Schematic of the experimental setup for the catalyst activity measurements
An ozone analyzer (Teledyne API M454) was used
to measure concentration of ozone in the exhaust gas
stream.
For each experiment, 0.06 g of the catalyst was heated
for 1 h at 475°C under 140 ml/min oxygen and 110 ml/
min nitrogen. Then, the catalyst was cooled down under
250 ml/min nitrogen flow and its temperature was fixed at
the desired reaction temperature of 25 or 90°C. The cata-
lyst was saturated by toluene before introducing ozone to
the reaction chamber.
Another configuration (Figure S.1 in the supplementary
document) was used for in-situ DRIFTS (diffuse reflectance
infrared Fourier transform spectroscopy) analyses, in which
the long-path gas cell was replaced with a DRIFTS acces-
sory (Harrick, Praying Mantis) on the FTIR spectrometer.
Then, the reaction chamber, equipped with ZnSe windows,
was coupled with the DRIFTS accessory and the FTIR
spectrometer. To achieve higher sensitivity and faster scan-
ning, a narrow band mercuric cadmium telluride (MCT-A)
detector was used for in-situ DRIFTS operations. Spectra
were collected at a resolution of 4 cm−1 in the range of
3900–1300 cm−1 to avoid saturation of the MCT-A detec-
tor. Exhaust stream of the reaction chamber was analyzed
by the ozone analyzer. A gas cylinder with ppm level con-
centrations of carbon dioxide and carbon monoxide bal-
anced with nitrogen (Praxair, accuracy of 2%) was used
for calibration purposes. COx yield was calculated using
Eq. (1):
2.3 Catalyst Characterization
Manganese loading on the catalyst was verified by induc-
tively coupled plasma mass spectrometry (ICP-MS) using
a Nexion 300D (PerkinElmer) instrument. N2 adsorption by
ASAP 2020 (Micromeritics) instrument was used to deter-
mine pore volume and Brunauer–Emmett–Teller (BET)
surface area of the catalyst. Topography of the catalyst
was recorded by transmission electron microscopy (Philips
CM10). Prior to measurements, the catalyst was ultrasoni-
cally dispersed in ethanol to produce a suspension. A drop
of the suspension was deposited onto the holey carbon film
and was left to dry. Then, it was used for the transmission
electron microscopy (TEM).
(
[
]) (
[
])
COx yield = 100 × [CO] + CO2 ∕ 7 ∗ C7H8, reacted
X-ray absorption near edge structure (XANES) of Mn
K-edge was collected at hard X-ray micro analysis beam-
line of the Canadian Light Source (CLS). Catalyst was
diluted with boron nitride (BN), ground and pressed to thin
disks. The prepared disks were protected by Kapton tape.
Measurements were conducted in transmission mode using
straight ion chamber detectors filled with helium gas, a Si(1
1 1) monochromator crystal, and Rh mirrors. For the pre-
edge and XANES regions, scan step-sizes were 10 eV/step
and 0.25 eV/step, respectively. ATHENA software [21] was
used for data processing. The reference materials includ-
ing Mn2O3 (Sigma-Aldrich, 99%, bixibyite), MnO2 (Alfa
(1)
2.2 Catalyst Preparation
MnOx/γ-Al2O3 catalyst was prepared by dry impregnation
of powdered γ-alumina (powder size less than 0.208 mm)
by using manganese(II) nitrate tetrahydrate (Sigma Aldrich,
97%) solution as manganese precursor. Nominal mass of
Mn per mass of the catalyst was 10%. The impregnated
support was dried for 10 h at 100°C and calcined for 4 h at
500°C in air. Then, the catalyst was crushed and sieved to a
final powder size of less than 0.208 mm.
1 3

M. Aghbolaghy et al.
Aesar, 99.9%, pyrolusite), MnO (Alfa Aesar, 99%, man-
ganosite), and Mn3O4 (Sigma-Aldrich, 97%, hausmannite)
were also analyzed to provide reference points.
Figure 2 shows Mn K-edge XANES spectra of the
catalyst along with spectra of MnO, Mn3O4, Mn2O3 and
MnO2 as the pure reference materials. At energies higher
than 6550 eV, catalyst spectra were a mix of spectra of the
reference materials. On the other hand, at energies below
6550 eV, catalyst spectra followed the Mn2O3 spectra.
Quantification of the results by Linear Combination Fit-
ting (LCF) of Mn K-edge XANES data showed that Mn2O3
(82%) was the major manganese phase in the catalyst fol-
lowed by MnO2 (11%) and Mn3O4 (7%). Raman spectra of
the catalyst confirm that Mn2O3 was the major manganese
phase. Figure 3 shows Raman spectra of the catalyst and
the reference materials [8]. There were four distinct peaks
at 199, 312, 614, and 698 cm−1 on the spectra of the cat-
alyst, which indicate a significant presence of the Mn2O3
phase.
Raman spectroscopy was performed by Renishaw InVia
Raman microscope. A 514.5 nm Ar+laser (Spectra Phys-
ics) and a CCD detector were used. A 10 s detector expo-
sure time and 32–128 spectra accumulations were used for
the measurement and the instrument was operated in the
line focus mode [8].
Thermo-gravimetric analysis was used to analyze tem-
perature programmed weight loss of the spent catalyst. A
thermal analyzer (TGA Q500, TA Instruments) with stand-
ard furnace type was employed for this purpose. Samples
were heated at 20°C/min to 790°C under a pure nitrogen
flow.
Carbonaceous species accumulated on the catalyst were
extracted with dichloromethane and were analyzed by
injecting the extracted sample into a gas chromatograph
(Agilent, 7890A) coupled with mass spectrometer (Agilent,
5975C). An Agilent standard HP-5MS column was used in
the gas chromatograph.
Figure 4 depicts transmission electron microscopy
image of the catalyst. TEM image showed that the cata-
lyst consists of metal cluster sizes of about 20 nm. This is
confirmed by X-ray diffraction (XRD) results (Figure S.2)
that showed Mn2O3 and MnO2 in the catalyst had metal
cluster sizes of 23 and 17 nm, respectively. Mn3O4 was not
detected by XRD. This is due to detection limit of the XRD
analyzer. XRD analyzer can detect particles larger than
5 nm, indicating that metal clusters of Mn3O4 phase were
highly dispersed [22].
For temperature programmed oxidation (TPO), spent
catalyst was immediately purged with nitrogen for
10 min. Then, the catalyst was heated at a rate of 10°C/
min to 745°C under a 250 ml/min atmospheric flow of
oxygen–nitrogen (20–80 v%). Similarly, for temperature
programmed desorption (TPD), spent catalyst was imme-
diately purged with nitrogen for 10 min. Then, the sam-
ple was heated at 20°C/min to 870°C under a 250 ml/min
atmospheric flow of nitrogen.
3.2 Catalyst Activity
Catalytic ozonation of toluene was carried out at 25
and 90°C. The experiments were conducted over
MnOx/γ-Al2O3, pure γ-alumina, and pure Mn2O3. Toluene
and ozone conversion profiles along with carbon dioxide
and carbon monoxide concentrations in the exhaust stream
3 Results and Discussion
3.1 Catalyst Characterization
Table 1 presents manganese loading, BET surface area
and pore volume of the catalyst. ICP-MS analysis verified
that the manganese loading on the catalyst (9.8 wt%) was
close to its nominal 10 wt% loading. Also, surface area and
pore volume results showed that addition of manganese on
γ-alumina decreased the pore volume and surface area by
15 and 16%, respectively.
Table 1 Manganese loading, BET surface area and pore volume of
the catalyst
Catalyst
γ-alumina
MnOx/γ-
Alumina
Mn loading (wt %)
BET (m2/g)
9.8
183
0.51
219
0.60
Fig. 2 XANES spectra of Mn K-edge; (a) Mn2O3, (b) Mn3O4, (c)
MnO2, (d) MnO, (e) MnOx/γ-alumina
Pore volume (cm3/g)
1 3

Role of Surface Carboxylates in the Gas Phase Ozone-Assisted Catalytic Oxidation of Toluene
are presented in Fig. 5. Standard error for toluene and
ozone conversion was within 3%. Also, Standard errors
for carbon dioxide and carbon monoxide concentrations
were within 7 and 3 ppm, respectively.
For catalytic ozonation over MnOx/γ-alumina at 90°C, a
stable reaction was observed with a high conversion of 95%
for both toluene and ozone. However, at 25°C operation,
a decline in toluene and ozone conversions was observed.
After 150 min of the reaction at 25°C, toluene and ozone
conversions dropped to 71 and 35%, respectively. At 90°C,
CO and CO2 concentrations increased gradually with reac-
tion time and reached constant concentrations of 167 and
413 ppm, respectively. On the other hand, at 25°C reac-
tion, CO and CO2 concentrations increased initially and
reached maximum values of 58 (at 105 min) and 178 ppm
(at 55 min), respectively, followed by a decline in CO2 con-
centration. After 150 min of reaction, COx yields for the 25
and 90°C reactions were 35 and 87% with CO2/CO ratios
of 2.38 and 2.47, respectively.
For catalytic ozonation over pure γ-alumina, rapid cata-
lyst deactivation was observed at both 25 and 90°C. How-
ever, toluene and ozone conversions were relatively higher
at 90°C compared to those at the 25°C reaction. This
observation emphasizes the important role of the Mn sites
in the catalytic activity of the MnOx/γ-alumina catalyst.
Moreover, this observation suggests that γ-alumina is not
an inert support and it can oxidize toluene in the presence
of ozone, although the reaction is not stable and a rapid cat-
alyst deactivation is observed.
Fig. 3 Raman spectra; (a) MnOx/γ-alumina, (b) Mn3O4, (c) Mn2O3,
(d) MnO2, (e) MnO
For catalytic ozonation over pure Mn2O3, a decline in
toluene and ozone conversions was observed at 25°C oper-
ation. However, at 90°C a stable reaction with relatively
low conversions of toluene (37%) and ozone (46%) was
observed. CO and CO2 concentrations for reactions over
pure γ-alumina and Mn2O3 were significantly lower than
those for reactions over MnOx/γ-Al2O3 catalyst. Unlike
pure Mn2O3, manganese oxides in the MnOx/γ-alumina are
dispersed on the large surface area of γ-alumina. Also, it
was indicated that γ-alumina is not inert and interacts with
toluene. These factors help the MnOx/γ-alumina catalyst
to achieve a stable reaction at 90°C with high toluene and
ozone conversions.
To better understand the effect of temperature on cata-
lyst activity and byproduct formation, in-situ DRIFTS
studies were conducted. Figure 6 depict in-situ DRIFTS
spectra of the catalytic ozonation of toluene at different
reaction times. Figure 6a, b show in-situ DRIFTS spectra
of the catalytic ozonation of toluene over MnOx/γ-Al2O3
catalyst at 25 and 90°C, respectively. Toluene adsorption
on the catalyst did not change spectra of the fresh catalyst,
due to low concentration of toluene in the inlet stream.
Once ozone was introduced into the reactor, a number
of peaks appeared. The peaks at around 1429, 1740, and
Fig. 4 Transmission electron microscopy image of MnOx/γ-alumina
catalyst
1 3

M. Aghbolaghy et al.
Fig. 5 Catalytic ozonation of toluene at 25 and 90°C, [O3]=1150 ppm, [toluene]=100 ppm; a Conversions of toluene and ozone, b CO and
CO2 concentrations in the exhaust stream; (1) MnOx/γ-Al2O3, (2) γ-Al2O3, (3) Mn2O3
1 3

Role of Surface Carboxylates in the Gas Phase Ozone-Assisted Catalytic Oxidation of Toluene
Fig. 6 In-situ DRIFTS spectra of catalytic ozonation of toluene, [O3]=1150 ppm, [toluene]=100 ppm; a MnOx/γ-Al2O3 at 25°C, b
MnOx/γ-Al2O3 at 90°C, c γ-Al2O3 at 25°C, d Mn2O3 at 25°C
1603 cm−1 were the main bands corresponding to C–H
asymmetric deformation vibration, C=O stretching, and
overlap of aromatic ring stretching and COO− stretching
of surface carboxylates, respectively. Moreover, there was
a broad band from 2400 to 3750 cm−1 corresponding to the
overlap of OH stretching of alcohols, carboxylic acids and
water. Complete list of significant bands and correspond-
ing functional groups are summarized in Table 2 [16, 22,
23]. The in-situ spectra suggest that catalytic ozonation
started with a similar mechanism at both 25 and 90°C. This
caused appearance and initial increase of similar bands at
both temperatures. However, the bands corresponding to
the byproduct formation (i.e. 1740, and the broad band at
2400–3750 cm−1) were stronger at 25°C, which is in agree-
ment with the deactivation pattern of the catalyst at 25°C.
In-situ DRIFTS spectra for catalytic ozonation of tolu-
ene using pure γ-alumina at 25°C are depicted in Fig. 6c.
The significant bands were at around 1325–1393, 2896,
and 1600 cm−1 corresponding to COO− stretching of the
surface carboxylates, saturated C–H stretching, and overlap
of aromatic ring stretching and another COO− stretching,
respectively. In addition, there was a broad band from 2500
to 3750 cm−1 corresponding to the overlap of OH stretching
of alcohols, carboxylic acids and water. The most important
difference between the spectra obtained from MnOx/γ- alu-
mina and γ-alumina was the presence of a very strong band
at around 1740 cm−1 (corresponding to C=O stretching) on
the spectra of the MnOx/γ-alumina catalyst. The same band
appeared only as a shoulder on the spectra of the γ-alumina
catalyst. This implies that the presence of manganese oxide
1 3

M. Aghbolaghy et al.
Table 2 Assignment of FT-IR
bands for catalytic ozonation of
toluene
Significant band (cm−1
)
Assigned functional groups
1308–1393
1429
1453
Antisymmetric and symmetric COO− stretching of carboxylates
C–H asymmetric deformation vibration
Aromatic ring stretching
1498
1568
1603
1720–1780
2350
Aromatic ring stretching
Antisymmetric COO− stretching of carboxylates
Overlap of aromatic ring stretching and COO− stretching of carboxylates
C=O stretching
Adsorbed carbon dioxide
2893–2944
3036–3072
2400–3750
Saturated C–H stretching
Unsaturated C–H and aromatic C–H stretching
OH stretching of alcohols, carboxylic acids and water
Fig. 7 Weight loss during thermo-gravimetric analysis for fresh and
spent MnOx/γ-Al2O3 catalysts used in catalytic ozonation of toluene
at 25 and 90°C
Fig. 8 In-situ DRIFTS spectra during TPD analysis of spent
MnOx/γ-Al2O3 catalyst used in catalytic ozonation of toluene at 25°C
is essential to effectively oxidize carbonaceous materials on
the surface of the catalyst.
the spent MnOx/γ-alumina catalysts at a rate of 20°C/min
to 790°C, the spent catalysts of 25 and 90°C reactions lost
22.9 and 5.9% of their weights, respectively. This indicates
that higher amount of byproduct was deposited on the cata-
lyst at the 25°C operation.
In-situ DRIFTS spectra for catalytic ozonation of tolu-
ene using pure Mn2O3 at 25°C are depicted in Fig. 6d.
Interestingly, the only significant band was at around
1760 cm−1 corresponding to C=O stretching. Also, surface
carboxylates, with a significant band at around 1600 cm−1,
were not formed in the absence of γ-alumina. Therefore,
both γ-alumina and manganese oxide play important but
different roles in catalytic ozonation of toluene over the
MnOx/γ-alumina catalyst.
Figure 8 depicts in-situ DRIFTS spectra at different
temperatures of 325, 525 and 870°C during temperature
programmed desorption (TPD) of the catalyst used in the
25°C catalytic ozonation. TPD was carried out under inert
N2 flow and the spent catalyst was heated up to 870°C at a
rate of 20°C/min. By heating the spent catalyst to 325°C,
intensities of all bands decreased to some extent. Further
increase in temperature to 525°C, reduced intensities of the
band at 1740 cm−1 (C=O stretching) and the broad band
at 2400–3750 cm−1 (OH stretching). On the other hand, it
increased intensities of the bands at 1429 (C–H asymmetric
deformation vibration) and 1600 cm−1 (overlap of aromatic
3.3 Temperature Programmed Analysis
Thermo-gravimetric analysis was conducted under inert
N2 environment. Figure 7 shows weight loss profile during
thermo-gravimetric analysis of the catalyst. After heating
1 3

Role of Surface Carboxylates in the Gas Phase Ozone-Assisted Catalytic Oxidation of Toluene
Table 3 Compounds detected in the exhaust gas during TPD
Sample
Temperature
range (°C)
Detected compounds
25°C reaction
25–325
Acetic acid, formic acid,
water vapor, CO, CO2
25°C reaction
25°C reaction
90°C reaction
90°C reaction
325–525
525–870
25–525
Water vapor, CO, CO2
Benzene, methane, CO, CO2
Water vapor, CO, CO2
525–870
Benzene, methane, CO, CO2
ring stretching and COO− stretching of surface carboxy-
lates). At 525°C, there were still two other bands at 2930
and 3060 cm−1 corresponding to saturated and unsaturated
C–H stretching, respectively. Further increase in tempera-
ture to 870°C, removed the bands at 1740 and 2930 cm−1
and reduced intensity of the bands at 1429 and 1600 cm−1.
Therefore, even at 870°C, the bands at 1429, 1530–1600,
and 3060 cm−1 remained on the spectra.
Fig. 9 In-situ DRIFTS spectra during TPO analysis of the spent
MnOx/γ-Al2O3 catalyst used in catalytic ozonation of toluene at 25°C
Evolved gases during TPD were analyzed and Table 3
presents a summary of the detected compounds at each
temperature range. CO and CO2 were found in the exhaust
gas during TPD of the spent catalyst at almost all TPD
temperatures (25–870°C) and their concentrations varied
by temperature (Figure S.3). The presence of significant
amounts of CO and CO2 in a wide range of temperature and
variation of their concentrations with temperature indicates
that the detected COx not only originated from desorption
of isolated CO and CO2, but also they evolved from decom-
position of larger carbonaceous compounds that were
deposited on the surface of the catalyst.
catalyst used in the 25°C catalytic ozonation. By increasing
temperature to 745°C almost all bands corresponding to
carbonaceous materials disappeared. This is different from
TPD (under N2) results that even at 870°C, some bands at
1429, 1530–1600, and 3060 cm−1 remained on the spectra.
Figure 10 shows the variations in CO and CO2 concen-
trations during TPO analysis. For the catalyst used in the
25°C catalytic ozonation (Fig. 10a), maximum CO gen-
eration was at 346°C with a concentration of 511 ppm.
Also, maximum CO2 concentration was 1861 ppm that
was obtained at 394°C. Most of the CO2 was generated
in the temperature range of 275–525°C. This is an agree-
ment with the DRIFTS spectra during TPO (Fig. 9) that
increasing temperature to 525°C decreased the intensity of
the band at 1740 cm−1 (C=O stretching) significantly. Ben-
zene and methane gases were not found during TPO analy-
sis. This indicates that the surface carboxylates, associated
with benzene and methane, were oxidized to CO and CO2.
Similar results were obtained for the catalyst used in the
90°C catalytic ozonation (Fig. 10b), however, CO and CO2
concentrations were relatively lower. In this case, maxi-
mum CO concentration was 95 ppm that was obtained at
292°C, and CO2 concentration reached a maximum value
of 333 ppm at 357°C. For the catalyst used in the 25°C
reaction, total amount of the evolved COx during TPO was
4.6 times higher than that of the 90°C reaction (Table 4),
indicating that higher amount of carbonaceous materials
accumulated on the catalyst in the 25°C reaction.
During TPD of the catalyst used for the 25°C catalytic
ozonation, acetic acid, formic acid and water were des-
orbed by heating the spent catalyst to 325°C. Benzene and
methane were detected during TPD from 525 to 870°C.
This is the same temperature range at which the bands
corresponding to saturated C–H stretching, aromatic ring
stretching and COO− stretching of surface carboxylates
were decreased/removed as well (Fig. 8). This implies
that benzene and methane evolved from decomposition of
larger surface carboxylate compounds. Surface carboxy-
lates have been found in catalytic ozonation of VOCs using
SiO2-based catalysts as well [18]. During TPD of the cata-
lyst used for the 90°C catalytic ozonation, benzene and
methane were detected at 525–870°C temperature range.
This indicates that the surface carboxylates, associated with
benzene and methane, were present at the 90°C catalytic
ozonation as well.
Temperature programmed oxidation (TPO) of the spent
catalysts was conducted under atmospheric flow of oxy-
gen–nitrogen (20–80 v%) and heating rate of 10°C/min.
Figure 9 depicts in-situ DRIFTS spectra during TPO of the
Table 4 shows the quantitative data for the overall car-
bon balance in the process. For the 25°C operation, about
15.03 mg carbon entered the reactor during the saturation
and reaction time. On the other hand, 13.76 mg carbon was
1 3

M. Aghbolaghy et al.
13.68 mg carbon was detected in the exhaust stream result-
ing in a 97% overall carbon balance.
3.4 Carbonaceous Deposits on the Spent Catalysts
The carbonaceous species accumulated on the spent
MnOx/γ-alumina catalyst were extracted with dichlo-
romethane and were analyzed with GC-MS. For the cata-
lyst used at 25°C, a number of organic compounds were
found such as formic acid, acetic acid, acetol, methylgly-
oxal, formyl acetate, acetic anhydride, acetoxyacetic acid,
maleic anhydride, and isopropyl methyl ketone. These
compounds are listed in order of increasing the number of
carbon atoms in their structures. In another attempt, the
catalyst that was used in the catalytic ozonation at 25°C,
was heated to 325°C under N2 flow. Then it was washed
with dichloromethane and the extract was analyzed with
GC-MS. The results did not show any extracted byproduct.
This indicates that all of the mentioned byproducts were
desorbed from the surface of the catalyst by heating the
spent catalyst to 325°C. For the catalyst used at 90°C, none
of the mentioned compounds were found by GC-MS analy-
sis in the extract.
3.5 Effect of the Surface Carboxylates
According to the thermo-gravimetric analysis results
(Fig. 7) for the catalyst used in 25°C operation, almost 70%
of the observed weight loss occurred by heating the catalyst
to 325°C. TPD results (Table 3) showed that heating up to
325°C was enough to desorb formic acid and acetic acid
from the surface of the catalyst. Also, other byproducts that
were detected by GC-MS analysis, were not found after
heating the catalyst to 325°C. Therefore, mainly surface
carboxylates remained on the surface of the catalyst after
heating it to 325°C.
Fig. 10 Variation of CO and CO2 concentrations during TPO analy-
sis of the spent MnOx/γ-Al2O3 catalyst used in catalytic ozonation of
toluene at a 25°C, b 90°C
Table 4 Breakdown of carbon distribution and overall carbon bal-
ance (values are in mg)
The catalyst that was used for catalytic ozonation at
25°C, was regenerated by heating to 325°C under N2 flow,
then the regenerated catalyst was used for a second cata-
lytic ozonation of toluene. The reaction conditions were
the same as the first catalytic ozonation. Figure 11 shows
toluene and ozone conversion profiles along with produced
CO and CO2 concentrations for the first and second cata-
lytic ozonations. Up to 75 min from the beginning of the
reaction, toluene conversion results for the first and second
catalytic ozonations were almost identical. Then, the regen-
erated catalyst showed more deactivation. After 150 min,
toluene conversions for the first and second catalytic ozo-
nations dropped to 71 and 56%, respectively. On the other
hand, ozone conversion for the second catalytic ozonation,
dropped significantly after 10 min from the beginning of
the reaction, followed by a gradual decline. After 150 min,
25°C reaction 90°C reaction
Total carbon ina
15.03
7.62
6.14
14.17
12.17
1.51
Total carbon out before TPOb
Total carbon evolved during TPOc
Total carbon out
13.76
13.68
aDuring reaction and catalyst saturation with toluene
bAs COx and unreacted toluene
cAs evolved COx during TPO
detected in the exhaust of the reactor as unreacted toluene,
produced COx during reaction, and COx during TPO. This
resulted in a 92% overall carbon balance. For the 90°C
operation, about 14.17 mg carbon entered to the system and
1 3

Role of Surface Carboxylates in the Gas Phase Ozone-Assisted Catalytic Oxidation of Toluene
Fig. 11 Catalytic ozonation of toluene at 25°C by using fresh and regenerated MnOx/γ-Al2O3 catalysts, [O3]=1150 ppm, and [tolu-
ene]=100 ppm; a conversions of toluene and ozone, b CO and CO2 concentrations in the exhaust stream
ozone conversions were 35 and 17% for the first and second
catalytic ozonations, respectively.
on the surface of the catalyst before the appearance of for-
mic acid and acetic acid. To investigate this possibility, a
one-minute catalytic ozonation reaction was conducted by
using fresh MnOx/γ-alumina at 25°C. Figure 12 depicts in-
situ DRIFTS spectra of the 1 min reaction. The significant
bands on the DRIFTS spectra are listed in Table 2. It can be
seen that the surface carboxylases were formed on the sur-
face of the catalyst from the early moments of the reaction.
Therefore, these compounds were present from the early
moments of the reaction, however, they did not reduce the
toluene and ozone conversions (See Fig. 11). The surface
carboxylates were observed during the catalytic ozonation
At the beginning, CO and CO2 concentrations for the
second catalytic ozonation were higher than those of the
first catalytic ozonation. However, their concentrations
deceased more rapidly and after 150 min CO and CO2 con-
centrations for the second catalytic ozonation were lower
than those of the first reaction.
Identical toluene conversions in the first 75 min of the
both reactions suggest that the presence of the surface car-
boxylates did not directly prevent toluene conversion. In
addition, high ozone conversion at the first 10 min of the
second catalytic ozonation indicates that the active sites of
the catalyst (i.e. Mn sites) were regenerated by heating the
spent catalyst up to 325°C and desorbing byproducts such
as formic acid and acetic acid. Also, higher CO and CO2
concentrations at the beginning of the second catalytic ozo-
nation are attributed to two factors. First, the active sites of
the catalyst were regenerated, and improved oxidation of
the toluene. Second, the presence of the regenerated active
sites and better ozone decomposition helped with further
oxidation of the previously deposited surface carboxylates.
It seems that improved oxidation of the surface carboxy-
lates generated more byproducts such as formic acid and
acetic acid, which subsequently reduced ozone decompo-
sition and COx formation. It has been reported [24] that
increase in ozone concentration enhances transformation
from species with COO− group (representing carboxylates)
to species containing C=O and C–O functional groups such
as carboxylic acids.
If byproducts such as formic acid and acetic acid are
mainly produced from oxidation of the surface carboxylate
intermediates, then these intermediates should be formed
Fig. 12 In-situ DRIFTS spectra of the one-minute catalytic ozona-
tion of toluene at 25°C using MnOx/γ-Al2O3
1 3

M. Aghbolaghy et al.
at 90°C as well, while a stable catalytic ozonation was
observed. The surface carboxylates were found in the cata-
lytic ozonation with pure γ-alumina as well (Fig. 6c), while
they were not found in the catalytic ozonation with pure
Mn2O3 (Fig. 6d). Therefore, it is plausible that the surface
carboxylases resided mainly on the γ-alumina part of the
MnOx/γ-alumina catalyst, and therefore they did not reduce
the catalytic activity. This conclusion is different from pre-
vious studies [16, 18, 19, 24] that have assumed all accu-
mulated materials are detrimental to the catalytic activity,
without differentiating between their roles in the reaction.
As discussed earlier, formic acid and acetic acid were
found during TPD of the catalyst used for catalytic ozona-
tion at 25°C (see Table 3). In addition to formic acid and
acetic acid, byproducts such as acetoxyacetic acid, formyl
acetate, acetol, methylglyoxal, acetic anhydride, maleic
anhydride, and isopropyl methyl ketone were detected in
the extract of the spent catalyst (see Sect. 3.4). It is believed
that these byproducts, which were found only in the 25°C
operation, accumulated on the surface of the catalyst and
caused catalyst deactivation.
Fig. 13 Breakthrough curves of toluene adsorption on
MnOx/γ-Al2O3, γ-Al2O3, and Mn2O3 at 25°C and [toluene]=100 ppm
toluene with metal oxides have shown [25–28] that toluene
interacts with the surface of metal oxides via abstraction of H
atoms from the methyl group. Therefore, a possible pathway
for catalytic ozonation of toluene on the MnOx/γ-alumina
catalyst is as follows:
3.6 Reaction Mechanism
It has been reported [9, 13] that ozone decomposition on
the surface of the catalyst generates highly reactive atomi-
cally adsorbed oxygen species (▪O) that rapidly oxidize
adsorbed VOCs:
C6H5−CH3 + ▴ ⇌ ▴C6H5−CH3
(5)
▴C6H5−CH3 + ▴ → ▴C6H5−CH2 + ▴H
(6)
▴C6H5−CH2 + 2▴O → ▴C6H5−COO + 2▴H
O3 + ▪ → O2 + ▪O
(7)
(2)
▴
where represents an alumina site. Equation (5) shows
▪O + O3 → O2 + ▪O2
(3)
adsorption of toluene on the first alumina site. Equation (6)
is based on dissociation of one H atom from the methyl
group of toluene [25–28]. Then, the unstable benzyl species
▪O2 → O2 + ▪
(4)
▪
where represents a surface metal site. As discussed ear-
▴
( C6H5−CH2) reacts with oxygen species (Eq. 7) to pro-
lier, ozone decomposition occurs on both γ-alumina and
▴
duce adsorbed benzoate ( C6H5−COO), which is a stable
▪
Mn sites of the MnOx/γ-alumina catalyst. Therefore, can
surface carboxylate intermediate. Equation (7) may include
multiple steps as described in the supplementary document
be either an alumina or Mn site.
Figure 13 shows breakthrough curves of toluene adsorp-
tion on MnOx/γ-alumina, pure γ-alumina, and pure Mn2O3.
Adsorption breakthrough times for MnOx/γ-alumina, and
pure γ-alumina were 6 and 8 min, respectively. However,
breakthrough time for pure Mn2O3 was very short (<1 min).
Therefore, toluene is mainly adsorbed on γ-alumina sites of
the MnOx/γ-alumina catalyst.
▴
(Section S.4). The adsorbed protons ( H) can react with the
oxygen species and produce water (Eqs. 8, 9, 10) [28, 29]:
▴H + ▴O ⇌ ▴OH
(8)
(9)
2▴OH ⇌ ▴H2O + ▴O
▴H2O ⇌ ▴ + H2O
(10)
As discussed in Sect. 3.2, alumina not only acts as a reser-
voir for toluene, but also it interacts effectively with toluene,
in the presence of ozone, to create surface carboxylate inter-
mediates. The presence of highly reactive atomic oxygen spe-
cies, resulted from ozone decomposition, creates an oxidative
environment. Many quick reactions can occur in this oxida-
tive environment, which makes it difficult to identify unsta-
ble reaction intermediates. However, studies on interaction of
Pure γ-alumina cannot further oxidize surface carboxy-
lates (see Sect. 3.2) and the presence of Mn sites is necessary
for further oxidation of the surface carboxylates (Eqs. 11, 12):
▴C6H5−COO + ∗ ⇌ ∗C6H5−COO
(11)
∗C6H5−COO + n ∗O → products
(12)
1 3

Role of Surface Carboxylates in the Gas Phase Ozone-Assisted Catalytic Oxidation of Toluene
where * represents a surface Mn site. At 90°C, the surface
carboxylates are quickly oxidized to carbon dioxide and
carbon monoxide. At 25°C, in addition to carbon dioxide
and carbon monoxide, byproducts such as acetic acid and
formic acid are produced and reduce the catalyst activity.
References
1. Guieysse B, Hort C, Platel V, Munoz R, Ondarts M, Revah S
(2008) Biotechnol Adv 26:398–410
2. Li M-S, Wu SC, Shih Y-H (2016) J Hazard Mater 315:35–41
3. Huang Y, Ho SSH, Lu Y, Niu R, Xu L, Cao J, Lee S (2016) Mol-
ecules 21:56
4. Rodrigues A, Tatibouet J-M, Fourre E (2016) Plasma Chem
Plasma Process 36:901–915
4 Conclusions
5. Mohamed EF, Awad G, Andriantsiferana C, El-Diwany AI
(2016) Environ Technol 37:1197–1207
Catalytic ozonation of toluene at 25 and 90°C was con-
ducted over MnOx/γ-alumina catalyst. Distinguishing the
role of surface carboxylates from that of byproducts such
as acetic acid and formic acid helped to develop a better
understanding of the reaction mechanism and catalyst sta-
bility. In the presence of ozone, alumina interacts with
toluene to produce surface carboxylate intermediates.
These intermediates are further oxidized on the manga-
nese sites. At 25°C, a decline in toluene and ozone conver-
sions is observed. It is believed that at 25°C, byproducts
such as acetic acid and formic acid accumulated on the
surface of the MnOx/γ-alumina catalyst and decreased the
catalyst activity. However, at 90°C, these byproducts are
quickly oxidized to carbon dioxide and carbon monoxide.
Therefore, a stable catalytic activity is observed at 90°C.
The catalyst deactivation at 25°C is reversible, as a high
catalytic activity is restored by heating the spent catalyst
to 325°C and desorbing the byproducts such as acetic acid
and formic acid.
6. Naydenov A, Mehandjiev D (1993) Appl Catal A Gen 97:17–22
7. Rezaei E, Soltan J, Chen N, Lin J (2013) Chem Eng J
214:219–228
8. Rezaei E, Soltan J (2012) Chem Eng J 198–199:482–490
9. Reed C, Xi Y, Oyama ST (2005) J Catal 235:378–392
10. Huang H, Ye X, Huang W, Chen J, Xu Y, Wu M, Shao Q, Peng
Z, Ou G, Shi J, Feng X, Feng Q, Huang H, Hu P, Leung DYC
(2015) Chem Eng J 264:24–31
11. Jung S-C, Park Y-K, Jung HY, Kang U Il, Nah JW, Kim SC
(2016) Res Chem Intermed 42:185–199
12. Zhao D-Z, Shi C, Li X-S, Zhu A-M, Jang BW-L (2012) J Hazard
Mater 239:362–369
13. Einaga H, Futamura S (2004) React Kinet Catal Lett 81:121–128
14. Jin D, Ren Z, Ma Z, Liu F, Yang H (2015) Rsc Adv
5:15103–15109
15. Einaga H, Maeda N, Teraoka
142:406–413
Y (2013) Appl Catal B
16. Einaga H, Futamura S (2004) J Catal 227:304–312
17. Einaga H, Futamura S (2005) Appl Catal B 60:49–55
18. Einaga H, Ogata A (2009) J Hazard Mater 164:1236–1241
19. Einaga H, Teraoka Y, Ogata A (2013) J Catal 305:227–237
20. Wang HC, Liang HS, Chang MB (2011) J Hazard Mater
186:1781–1787
21. Ravel B, Newville M (2005) J Synchrotron Radiat 12:537–541
22. Rezaei E, Soltan J, Chen N (2013) Appl Catal B Environ
136–137:239–247
Acknowledgements Authors would like to thank the University of
Saskatchewan and the Natural Sciences and Engineering Research
Council of Canada (NSERC) for their financial support of this
research. XANES experiments were performed at the Canadian Light
Source, which is supported by the Canada Foundation for Innovation,
NSERC, the University of Saskatchewan, the Government of Sas-
katchewan, Western Economic Diversification Canada, the National
Research Council Canada, and the Canadian Institutes of Health
Research.
23. Coates J (2000) In: Meyers RA (ed) Encyclopedia of analytical
chemistry. John Wiley & Sons Ltd., Chichester, pp 10815–10837
24. Liping L, Jianguo Z, Lixian Y, Mingli F, Junliang W, Bichun H,
Daiqi Y (2011) Chinese J Catal 32:904–916
25. Andersson SLT (1986) J Catal 98:138–149
26. Busca G, Cavani F, Trifirò F (1987) J Catal 106:471–482
27. Irigoyen B, Juan A, Larrondo S, Amadeo N (2003) Surf Sci
523:252–266
28. Menon U, Galvita VV, Marin GB (2011) J Catal 283:1–9
29. Rezaei E, Soltan J (2014) Appl Catal B Environ 148–149:70–79
1 3