Highly reactive magnetite covered with islands of carbon: Oxidation of N and S-containing compounds in a biphasic system

Article abstract of DOI:10.1016/j.apcata.2012.10.017

In this work a series of magnetite were prepared by impregnating of natural goethite (limonite) with glycerol before thermal treatment under N₂ flow at 300, 400 and 500 °C. The formation of carbon spots (island) over magnetite was found. The magnetite was found to be the main phase after treatment at 400 and 500 °C with high content of Fe²⁺, which is the active specie in a Fenton like system. Furthermore, iron oxide particles with magnetic and amphiphilic properties were formed, which are interesting for using as a catalyst in refineries. All materials presented high catalytic activities for oxidation of quinoline and dibenzothiophene in a biphasic reaction system (water/toluene).

Full text of DOI:10.1016/j.apcata.2012.10.017

Applied Catalysis A: General 450 (2013) 106–113
Contents lists available at SciVerse ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Highly reactive magnetite covered with islands of carbon: Oxidation
of N and S-containing compounds in a biphasic system
Iara R. Guimarãesa,∗, Amanda S. Girotoa, Wladmir F. de Souzab, Mário C. Guerreiroa
a Department of Chemistry, Federal University of Lavras, P.O. Box 3037, CEP 37200-000, Lavras, MG, Brazil
b CENPES-Petrobras, University City – Q7, Ilha do Fundão, CEP 21949-900, Rio de Janeiro, RJ, Brazil
a r t i c l e i n f o
a b s t r a c t
Article history:
In this work a series of magnetite were prepared by impregnating of natural goethite (limonite) with
glycerol before thermal treatment under N2 flow at 300, 400 and 500 ◦C. The formation of carbon spots
(island) over magnetite was found. The magnetite was found to be the main phase after treatment at 400
and 500 ◦C with high content of Fe2+, which is the active specie in a Fenton like system. Furthermore, iron
oxide particles with magnetic and amphiphilic properties were formed, which are interesting for using
as a catalyst in refineries. All materials presented high catalytic activities for oxidation of quinoline and
dibenzothiophene in a biphasic reaction system (water/toluene).
Received 3 August 2012
Received in revised form
25 September 2012
Accepted 6 October 2012
Available online 7 November 2012
Keywords:
Limonite
© 2012 Published by Elsevier B.V.
Glycerol
Amphiphilic catalyst
Oxidative process
1. Introduction
complementary to the HDS process because is difficult or very
costly to use the existing HDS technology to reduce the sulfur in
Sulfur compounds in fuel oils are known to have a negative
impact onto the environment because of SOx emissions from their
combustion exhausts. Consequently, removal of sulfur (present in
organic sulfur compounds like thiophenes, benzothiophenes (DBT)
and its derivatives, thiois, mercaptans and others), is becoming
a worldwide challenge due to more stringent regulations and in
order to prevent exhaust catalyst deactivation [1–3]. The oxidative
desulfurization process (ODS) has received much attention for deep
desulfurization of middle distillates because of two main advan-
tages relative to the hydrodesulfurization (HDS) process [4]. First,
the major advantage of the ODS over HDS process is that it can be
carried out in the liquid phase and under very mild conditions. Sec-
ond, the most refractory sulfur-containing compounds to the HDS
process [e.g., dibenzothiophene (DBT) and its alkylated derivatives]
show high reactivity toward the oxidation by ODS method. DBT
can be oxidized by the electrophilic addition reaction to sulfones
and sulfoxide. The chemical and physical properties of sulphones
are significantly different from those of hydrocarbon in fuel oil.
Therefore, they can easily be removed by such separation opera-
tions as distillation, solvent extraction [5,6]. The ODS process can be
the fuels to less than 50 mg L−1 [7].
In ODS process is important to evaluate the effect of the pres-
ence of other compounds in the oil fraction. In addition, it has been
observed that the N- and S-containing compounds coexist in many
kinds of fuel oil [8]. Denitrogenation is considerably more difficult
than desulfurization because the organonitrogen compounds are
much less reactive than the organosulfur compounds [9,10]. As a
part of this work we evaluate the effect of the N-containing com-
pounds on the oxidative desulfurization reactions, i.e., oxidation of
dibenzothiophene in the presence of quinoline.
Several types of oxidants and different catalysts have been
examined for the ODS process. Oxidants used include hydrogen
peroxide [11–13], organic hydroperoxides (e.g., tertbutyl hydroper-
oxide) [14,15], air [16], and other that including the combination
of H2O2 with organic acid [17,18]. Most of the reported ODS sys-
tems involve use of the oil-insoluble oxidants, H2O2 or peroxides,
which results in a biphasic oil–aqueous solution system. Biphasic
systems limit the mass transfer through the biphasic interfaces in
the oxidation process, which leads to decrease the oxidation rate.
Iron oxides have shown potential application in ODS system,
particularly after surface modifications to generate active species
[7,19]. Limonite, an iron oxide source that is a waste from nickel
ore, is a mixture of minerals that displays the goethite phase (␣-
FeOOH – 75%) as the main constituent that presents interesting
catalytic properties. However, this material remains preferably in
hydrophilic phase causing a low reaction rate and low conversions
∗
Corresponding author. Tel.: +55 3829 1626; fax: +55 3829 1273.
E-mail addresses: iarinha04@yahoo.com.br (I.R. Guimarães),
mandinhasg@hotmail.com (A.S. Giroto), wladmir@petrobras.com.br
(W.F. de Souza), guerreiro@dqi.ufla.br (M.C. Guerreiro).
0926-860X/$ – see front matter © 2012 Published by Elsevier B.V.
http://dx.doi.org/10.1016/j.apcata.2012.10.017

I.R. Guimarães et al. / Applied Catalysis A: General 450 (2013) 106–113
107
Fig. 1. Schematic representation on the limonite–glycerol composites (a) and formation of the carbon islands and regions containing active Fe2+ on the catalyst (b).
[20]. For this reason, limonite and other material based on iron
oxide need to be modified to generate hydrophobic regions on the
catalyst surface. In this work, the limonite was treated with a glyc-
erol to create islands of carbon on the surface of iron oxide in order
to generate amphiphilic properties and also create regions rich in
Fe2+ [21].
In this context, the objective in the present study was to explore
a novel ODS material, which is more energy efficient, present lower
cost and it is environment friendly. This ODS method combines a
catalytic oxidation step in the presence of H2O2 at mild conditions
and an extraction liquid step using methanol solution. The potential
advantages of this novel ODS method are: (i) the process attempt
to use green chemical as an oxidant, which are readily available,
low cost, and suitable for desulfurization applications; (ii) natural
iron oxide (Fenton active phase – Fe2+ and a hydrophobic region)
impregnated with island carbon (hydrophobic regions), forming
an amphiphilic catalyst to act in a biphasic oil–aqueous-solution
system.
2.2. Catalyst characterization
Mössbauer spectroscopy experiments were carried out at room
temperature in a spectrometer model MA250 with a 57Co/Rh
source, using ␣-Fe foil as reference. The powder XRD data were
obtained in a Rigaku model Geigerflex using Co K␣ radiation scan-
ning from 10◦ to 80◦ 2ꢀ at a scan rate of 48 min−1. Natural limonite
and the sample with treatment were analyzed by FTIR using an
Excalibur FTS 3000 from Bio Rad. Samples were pressed into 7 mm
discs using 1–5% mixture of the sample and spectroscopic grade
KBr. The limonite and Gli materials were observed by a JEOL JSM
6400 scanning electron microscopy. Moreover, an EDS probe was
used to determining their elemental compositions. TPR experi-
ments were performed in a CHEMBET 3000 using 40 mg sample
and a flow rate of 80 mL min−1 of a mixture of H2 (5%) and N2
with heating rate of 10 ◦C min−1 from room temperature to 1000 ◦C.
Thermogravimetric analysis TGA (Shimadzu DTG-60AH) was made
in the following conditions: N2 atmosphere with a heating rate of
10 ◦C min−1, gas flow of 30 mL min−1 and initial mass about 3.0 mg.
The surface area was determined with the BET method using N2
adsorption/desorption at 77 K in an Autosorb 1 from Quantachrome
Instruments.
2. Experimental
2.1. Catalyst preparation and characterization
2.3. Catalytic tests
All chemicals were high purity grade and were used as pur-
chased. Limonite ore samples (45–46 wt% of Fe 75% of which are as
goethite form), from mines located in Goias State (Central Brazil),
were ground and sieved below 100 mesh (Tyler series), dried for 1 h
at 120 ◦C. This material was impregnated with a glycerol–acetone
solution (10%, v/v), in proportion of iron oxide 5 g to 100 mL of
glycerol solution. The material was maintained under stirring at
80 ◦C until total evaporation of the solvent. The limonite–glycerol
composite was thermal treated under N2 atmosphere at three tem-
peratures: 300 ◦C (Gli 300), 400 ◦C (Gli 400) and 500 ◦C (Gli 500),
rate 10 ◦C min−1, with flow of 50 mL min−1. The scheme showed in
Fig. 1 represent all steps used for the treatment to the formation of
the amphiphilic materials.
The oxidative reactions were performed were magnetically
stirred at 25 ◦C ( 2 ◦C). The catalytic activity of the materials was
tested on the oxidation of quinoline (50 mg L−1) using toluene
as a solvent. The catalytic tests at pH 4.0 (natural pH of the
H2O2//HCOOH solution) were carried out with a total volume of
10 mL (9.9 mL of quinoline solution and 0.1 mL of the equimolar
solution H2O2//HCOOH) and 10 mg of catalyst. Quinoline removal
from toluene phase was monitored by GC–MS. Polar compounds
was monitored by electrospray ESI-MS (Agilent-1100), aiming to
identify intermediates formed during this reaction. ESI-MS/MS
experiments in the positive ion mode were also performed. The
reaction samples were analyzed by infusion into the ESI source with

108
I.R. Guimarães et al. / Applied Catalysis A: General 450 (2013) 106–113
100
96
Gli 500
Gli 400
92
Gli 300
88
Natural limonite
Fig. 2. The quinoline and DBT chemical structures.
84
80
76
a syringe pump at a flow rate of 5 ␮L min−1. The spectral data so
obtained were averaged from 50 scans of 0.2 s each. Typical ESI-MS
conditions were as follows: heated capillary temperature, 150 ◦C;
dry gas (N2) at a flow rate of 5 L min−1; spray voltage, 4 kV; capillary
voltage, 25 V; tube lens offset voltage, 25 V.
0
100
200
300
400
500
600
Temperature/ºC
The catalytic properties of the amphiphilic material were tested
via oxidation of dibenzotiophene (DBT), a model compound of sul-
fur contaminant from petroleum. In a typical run, 9.9 mL of a DBT
solution (50 mg L−1, in toluene), 10 mg of the catalyst and 0.1 mL
of equimolar mixture H2O2//HCOOH as oxidant were added in the
reactor. The oil phase was analyzed by GC-QP 2010 equipped with a
HP-5 capillary column (30 m × 0.25 mm × 0.32 ␮m film thickness)
and MS detector QP 2010 Plus.
Fig. 4. TGA profiles for samples of natural limonite and Gli 300, Gli 400 and Gli 500,
obtained under N2 atmosphere.
vibrations, while the bands at lower wave numbers correspond to
Fe O and Fe
[23].
O H stretching vibrations or lattice vibration bands
It was also tested the catalytic capability of simultaneous oxi-
dation of quinoline and DBT. The mixture contained 100 mg L−1
of quinoline and 50 mg L−1 of DBT in toluene was used. A self-
extracting ability of the biphasic system was assessed with a
solution methanol and assessed by ESI-MS. Fig. 2 presents the
quinoline and DBT chemical structure.
The FTIR spectra of Gli 300, Gli 400 and Gli 500 materials showed
two main absorption peaks in the range of 370–800 cm−1, which are
characteristic peaks of magnetite [24]. For Fe3O4 phase, two major
bands at 590 and 408 cm−1 are ascribed to Fe O stretching modes
of tetrahedral and octahedral sites in magnetite, respectively [25].
Signals related to vibration of OH groups of bulk are no longer
present in the material after thermal treatment during the dehy-
dration and reduction of goethite to magnetite. The natural and
synthetic limonite samples were analyzed by thermogravimetric
analysis (TGA). The TGA profile of the samples are shown in Fig. 4.
A total mass loss of about 12.9%, 8.2%, 4.6% and 2.8% was
observed for natural limonite, Gli 300, Gli 400 and Gli 500, respec-
tively, and this loss is predominantly in the temperature below
300 ◦C. The TGA profile shows at low temperature very small weight
losses (in the temperature below 130 ◦C), since the thermopro-
cesses only comprise the removal of physically adsorbed water on
the surface of materials. The loss of mass in temperature above
300 ◦C is related to structural loss of water due to loss of hydroxyl
groups of goethite. Decrease in mass with the thermal treatment
(300–500 ◦C) is related with conversion of goethite into magnetite
phase (as shown in XRD analysis, Fig. 6) due to loss of the structural
OH. The reduction capability of the samples were analyzed by TPR.
The reduction profiles of the materials are shown in Fig. 5.
3. Results and discussion
3.1. Catalyst characteristics
FTIR spectra of all materials are shown in Fig. 3. The FTIR
spectrum of reference sample natural limonite corresponds to ␣-
FeOOH. The resulting Fe O and O H bond lengths of the surface are
once more different than those of the bulk [22]. The IR bands at 3442
and 3186 cm−1 correspond to the stretching vibrations of adsorbed
H2O molecules and structural OH groups, respectively. Intensive
bands at 900 and 793 cm−1 are assigned to Fe
O H bending
All TPR profiles were similar due to the fact that all sam-
ples presents similar composition and goethite (major constituent
of limonite) is transformed into hematite by heating during TPR
measurements. A sharp peak centered at approximately 425 ◦C
was observed. A second H2 consumption peak, centered at 750 ◦C
was also observed. The first consumption peak accounts for the
reduction of Fe2O3 (hematite) to Fe3O4 (magnetite). The second
H2 consumption peak is related to the reduction of magnetite to
wustite (FeO) and metallic iron (Fe0). The samples thermal treat-
ment leads to the decrease of the amount of hydrogen consumed,
observed from the area under the peak. This fact is due to the
reduction of Fe3+ to magnetite. A stead area decrease of the peak at
425 ◦C is observed with increase of temperature of treatment and
it is related to Fe2+ sample enrichment. This process was observed
previously for other iron oxide materials [13,18,26].
In order to identify the crystalline phases the materials were
characterized by X-ray diffraction. The XRD patterns of all samples
are shown in Fig. 6.
Fig. 3. FTIR spectra of samples natural limonite and Gli 300, Gli 400 and Gli 500,
recorded at 20 ◦C.

I.R. Guimarães et al. / Applied Catalysis A: General 450 (2013) 106–113
109
S2
S1
Natural Limonite
298 K
Fig. 5. TPR profiles for natural limonite, Gli 300, Gli 400 and Gli 500; heating rate
Bhf
Gli 400
298 K
10 ◦C min−1, under flow rate of 20 mL min−1 of 5% H2 in N2.
-12
-9
-6
-3
0
3
6
9
12
Velocity/mm s-1
XRD patterns of natural limonite sample (Nat Lim) confirmed
the presence of ␣-FeOOH as main crystalline phase (JCPDS). The
XRD pattern of Gli 300 material in Fig. 6 confirms the forma-
tion of mixture of iron oxide including a reduced iron oxide
phase (magnetite–Fe3O4), showing that the deposition of car-
bon on the material surface also modifies the structure of the
iron oxide. Gli 400 and Gli 500 presented magnetite as main
phase. From here on we focused in natural limonite and Gli 400
materials due to the latter sample presented high activity in
preliminary catalytic tests and the former was used for compari-
son.
In order to obtain information about the structural properties
Mössbauer spectra of the samples were recorded (Fig. 7). From
XRD analyses three phases are expected in the samples: ␣-FeOOH,
Fe3O4, and ␣-Fe2O3.
Mössbauer spectra were obtained at 298 K for both materi-
als. Natural limonite presents a central doublet due to the small
particle size, notwithstanding the Mössbauer spectrum recorded
at 77 K showed a sextet with magnetic ordered. The limonite
Fig. 7. Transmission Mössbauer spectra of natural limonite and Gli 400 powder
measured at 298 K.
spectrum is fitted by two magnetically split subspectra, S1 and S2.
The super-paramagnetic behavior of goethites has been attributed
to the reduction of the magnetic coupling between the crystallites
and the high concentration of vacant sites [27]. From the hyper-
fine parameters of the sextet, obtained in fitting, 66% of iron is
present as ␣-FeOOH (Bhf = 37.8 T). The remaining doublet can be
assigned to dispersed superparamagnetic Fe3+. The Gli 400 mate-
rial shows a Mössbauer spectrum characteristic for magnetite,
Fe3O4, with a typical sextet observed for magnetic iron oxide pre-
senting tetrahedral [Fe3+
298 K.
The materials were also analyzed by SEM–EDS (Fig. 8). The natu-
ral limonite showed flat crystal faces and Gli 400 showed a different
texture, with carbon island on the surface.
] ]
tetra and octahedral [Fe3+, Fe2+ octa sites at
Elemental analysis of the materials revealing that the region
in (a) likely composed of Fe and O are the dominant elements
detected on the surface of the natural limonite in the proportions
corresponding to FeOOH phase. However, the Gli 400 EDS (Fig. 8d)
reveals that C is added to the surface of iron oxide after of impregna-
tion with glycerol and thermal treatment. This elemental analysis
showed the presence of a significant amount of C (approximately
49%), which generated a partial hydrophobic properties to Gli 400
catalyst [28].
3.2. Catalytic tests
The catalytic activity of the natural limonite and modified mate-
rial were evaluated with respect to decomposition of hydrogen
peroxide with the results presented in Fig. 9.
The most active catalyst for hydrogen peroxide decomposi-
tion is thermal treated material. This effect may be due to the
fact that the surface reactivity of the oxides depended on the
presence of reduced iron on the material surface, such as Fe2+
.
Firstly, Fe2+ plays an important role for the initiation of the Fen-
ton reaction, according to the classical Haber–Weiss mechanism
and therefore the enhancement of the rate of •OH production. As
iron(II) presents a rate constant for hydrogen peroxide decomposi-
Fig. 6. XRD patterns of obtained powder samples, recorded at 20 ◦C (Gt = ␣-FeOOH,
Mt = Fe3O4, and Hm = ␣-Fe2O3) Cu K␣.
tion of kFe = 76 mol L−1
S
−1 which is faster than the rate constant
2+

110
I.R. Guimarães et al. / Applied Catalysis A: General 450 (2013) 106–113
Fig. 8. SEM images for natural limonite and Gli 400 (a and c) and EDS analyses (b and d).
mixed Fe2+–Fe3+ species present in the material, reactions involv-
ing both Fe2+ and Fe3+ occur. Therefore, the presence of iron(II) in
the material enhances the production rate of HO• [30]:
observed for iron(III) species kFe = 0.01 mol L−1
S
−1 [29]. In a het-
3+
erogeneous Fenton type reaction, a series of reactions occurs on
the material surface. If only Fe3+ phase is present, Fe2+ is slowly
generated by the reduction reactions as shown in Eqs. (2)–(4), ini-
tiating OH radical generation as shown in Eq. (1). In the case of
Fe
+ H2O2 → Fe3su+rface + HO− + HO•
(1)
2+
surface
Fig. 9. Rate of hydrogen peroxide decomposition (a) and organic model compound oxidation (b) using iron oxide as catalyst: natural limonite and thermal treated materials.
Experimental conditions: T = 25 ◦C; 30 mg of catalyst; 3.0 mL H2O2 50%; 5.0 mL H2O (a) and T = 25 ◦C; 10 mg of catalyst; 0.1 mL H2O2 50%; 9.9 mL methylene blue 50 mg L−1
.

I.R. Guimarães et al. / Applied Catalysis A: General 450 (2013) 106–113
111
S
S
S
O
O
O
DBT
DBTO
DBTO2
Fig. 10. Chemical structures of dibenzothiophene and its oxidation derivatives
dibenzothiophene sulfoxide (DBTO) and dibenzothiophene sulfone (DBTO2).
3+
Fe
Fe
Fe
+ H2O2 → Fe3su+rface(H2O2)
(H2O2) → Fe2su+rface(O2H) + H2O
(O2H) → Fe2su+rface + HOO•
(2)
(3)
(4)
surface
3+
surface
2+
surface
For Gli 400 material, after 20 min of reaction, about 11.9 mL of
Fig. 11. Dibenzothiophene oxidesulfurization over Gli 400. Experimental
conditions: [H2O2 = 0.088 mol L−1]; [catalyst]: 1 mg mL−1
; [DBT in toluene]:
O2 were formed while the natural limonite formed only 1.5 mL.
Gli 300 (6.5 mL of O2) and Gli 500 (8.0 mL of O2) are much less
active than Gli 400, taking into consideration the activity of Gli 400
catalyst. Furthermore, the materials were evaluated in the oxida-
tion of methylene blue, used as model molecule. According to the
data presented in Fig. 9b, the material treated at 400 ◦C showed
the highest rates of oxidation due to the presence of reduced iron
sites combined of the accessible area of the material surface. The
reduction at 500 ◦C compresses the material and make the pro-
cess of diffusion at the interface of the catalytic material more
difficult. So the Gli 400 material was select for the next catalytic
tests.
Oxidation of sulfur atom in liquid phase followed by extrac-
tion of oxidized species can lead to desulfurization of diesel fuels.
DBT oxidation mainly after reaction with Gli 400 catalyst generated
DBTO (dibenzothiophene sulfoxide) and DBTO2 (dibenzothiophene
sulfone) as oxidation products (Fig. 10). The oxidation of the sulfur
compounds in the toluene medium with Gli 400 catalyst was first
carried without quinoline in a preliminary experiment aiming at
identifying the reaction products. Fig. 11 shows the kinetics of DBT
oxidation by the Gli 400//H2O2 system.
50 mg L−1
.
thermal treatment with glycerol) are playing an important role in
the catalyst activity. This catalyst, that was very active removing
almost 100% of the DBT, has a great potential to be used in the
removal S-compounds from fuel by an oxidative process, leading
to more polar compounds as DBTO and DBTO2 that can be removed
from fuel by partition process [31].
Both nitrogen-containing and sulfur-containing compounds
coexist in fossil fuels. Therefore it is essential to investigate
the oxidation of sulfur-containing compounds in the presence of
nitrogen-containing ones. It is known that nitrogen-containing
compounds have a negative effect in the oxidation of sulfur-
containing compounds [32]. The authors attributed the lower
activities to the poisoning of the active sites by nitrogen-containing
compounds. On the other hand, Ishihara et al. [33] reported that
nitrogen-containing compounds could be oxidized in the ODS oxi-
dation system. This result suggests that the decrease in ODS activity
can be related to the competition of the parallel oxidation of
nitrogen-containing compounds. To clarify the effects of nitrogen-
containing compounds in the DBT oxidation it was investigated the
behavior of the oxidation process using both model compounds
of nitrogen and sulfur containing molecules. The system used was
quinoline/DBT//H2O2. Fig. 12 shows the DBT conversion in the pres-
ence of quinoline.
The catalytic test for the oxidation of dibenzothiophene–DBT (a
model molecule for fuel contaminant) was carried out in toluene
and using the natural limonite and Gli 400 as a catalyst, the
later showed high catalytic activity. The Gli 400 presented higher
conversion (97%) after 180 min of reaction time (treatment with
natural limonite removed about 45% of DBT after 180 min of reac-
tion). This result clearly shows that the Fe2+ species (formed after
70
Quinoline
DBT
100
80
60
40
20
0
100
80
60
40
20
0
DBT
DBTO
100
2
60
80
50
40
60
S
N
30
40
20
20
10
S
S
0
O
0
O
0
30
60
90
120
150
180
0
30
60
90
Time/min
120
150
180
Time/min
Fig. 12. DBT conversion with reaction time in the oxidation over Gli 400 catalyst in the presence of quinoline. Experimental conditions: [H2O2 = 0.088 mol L−1]; [catalyst]:
1 mg mL−1; [DBT in toluene]: 50 mg L−1; [quinoline in toluene]: 50 mg L−1
.

112
2.2
I.R. Guimarães et al. / Applied Catalysis A: General 450 (2013) 106–113
DBT Quinoline
-
1.7
1.2
0.7
0.2
Standard
S
N
*
7.5
10.0
12.5
15.0
17.5
20.0
2.0
Gli400
120min
1.5
1.0
0.5
S
O
*
O
*
*
S
N
7.5
10.0
12.5
15.0
17.5
20.0
2.0
1.5
1.0
0.5
0.0
Gli400
180min
S
O
O
N
*
Fig. 14. Analysis by ESI-MS of aqueous fraction extracted from the oxidation of DBT
in the presence of quinoline.
7.5
10.0
12.5
15.0
17.5
20.0
tR/min
Fig. 13. GC–MS chromatograms of standard and oxidized quinoline and DBT. Exper-
imental conditions: [H2O2 = 0.088 mol L−1]; [catalyst]: 1 mg mL−1; [DBT in toluene]:
50 mg L−1; [quinoline in toluene]: 50 mg L−1. *Column bleeding.
give interesting characteristics to the iron oxide such as hydropho-
bicity increasing the catalytic activity. The scheme shown in Fig. 16
shows the biphasic reaction in the presence of natural limonite
(a) and Gli 400 (b), suggesting the ability of the thermal treated
material in remain in the interface of aqueous and apolar solvent.
Fig. 13 shows the GC–MS chromatograms of the
DBT–quinoline/toluene obtained before and after the oxida-
tion experiment. The peaks corresponding to DBT and quinoline
were identified by comparison of their retention time with refer-
ences compounds while the peak of its corresponding sulfone was
identified using a GC–MS and by comparison with an electronic
mass spectrum library.
The model mixture of quinoline–DBT presents both peaks of
N- and S-compounds (t = 0 min, standard sample), but after 180
or 240 min of reaction time we observed that both compounds
were oxidized considerably. The addition of DBT suppressed the
performance of the Gli 400//H2O2 in the oxidation of quinoline
in approximately 50%. In order to investigate the mechanism of
quinoline oxidation, the reaction products were extracted using
methanol and analyzed by ESI-MS (Fig. 14). It was detected
peaks with m/z = 201 related to hydroxylated DBT (DBT-OH)
and m/z = 146 and 178 related to hydroxylated quinoline. These
results confirm that the catalyst Gli 400 is very active to
transfer oxygen species as •OH groups to DBT and quinoline
molecules, which increase their polarity and permit the removal
from fuel by partition using a polar solvent [34,35]. We pro-
posed the structures for the oxidized compounds as shown in
Fig. 15.
N+
m/z =130
H
S
m/z =184
HO
H
N+
H
m/z =146
S+
HO
HO
m/z =201
H2-
m/z =178
(a)
OH
N+
H
HO
The difference in the activity of the materials may be related to
the nature iron oxides phases and the presence of carbon islands
present on the Gli 400. During the thermal treatment, glycerol
mixed with ␣-FeOOH reduces the iron oxide surface to produce
iron(II) species that are much more active in the generation of
hydroxyl radicals. Moreover, the formation of elemental carbon can
(b)
Fig. 15. Proposal structures based on signals found in ESI-MS analyses.

I.R. Guimarães et al. / Applied Catalysis A: General 450 (2013) 106–113
113
Fig. 16. Scheme of biphasic reaction in the presence of natural limonite (a) and Gli 400 sample (b).
4. Conclusions
[12] W.F. de Souza, I.R. Guimaraes, M.C. Guerreiro, L.C.A. Oliveira, Appl. Catal. A 360
(2009) 205–209.
[13] I.R. Guimaraes, A. Giroto, L.C.A. Oliveira, M.C. Guerreiro, D.Q. Lima, J.D. Fabris,
Appl. Catal. B 91 (2009) 581–586.
[14] A. Ishihara, D. Wang, F. Dumeignil, H. Amano, E.W. Qian, T. Kabe, Appl. Catal. A
279 (2005) 279–287.
[15] V.V.D.N. Prasad, K.-E. Jeong, H.J. Chae, C.U. Kim, S.Y. Jeong, Catal. Commun. 9
(2008) 1966–1969.
[16] J.T. Sampanthar, H. Xiao, J. Dou, T.Y. Nah, X. Rong, W.P. Kwan, Appl. Catal. B 63
(2006) 85–93.
[17] W. Ferraz, L.C.A. Oliveira, R. Dallago, L. Conceic¸ ãO, Catal. Commun. 8 (2007)
131–134.
[18] W.F. de Souza, I.R. Guimaraes, L.C.A. Oliveira, M.C. Guerreiro, A.L.N. Guarieiro,
K.T.G. Carvalho, J. Mol. Catal. A: Chem. 278 (2007) 145–151.
[19] A. Matsumura, T. Kondo, S. Sato, I. Saito, W.F. de Souza, Fuel 84 (2005)
411–416.
Glycerol was carbonized over the iron oxide surface at 400 ◦C
under N2 flow producing a hydrophobic carbon layer. Further-
more, the thermal decomposition of glycerol caused the formation
of reduced iron species with high capacity to decompose hydro-
gen peroxide. The material can be easily recovered using a magnet
device to be reused. The magnetic catalyst presented high catalytic
activity for oxidation of quinoline and DBT in organic medium
as showed by mass spectrometry analyses. We also observed an
amphiphilic property for the reduced materials presenting a great
potential for an ODS fuel treatment.
[20] C. Jiang, J. Wang, S. Wang, H. Guan, X. Wang, M. Huo, Appl. Catal. B 106 (2011)
343–349.
Acknowledgments
[21] F. Magalhães, M.C. Pereira, J.D. Fabris, S.E.C. Bottrel, M.T.C. Sansiviero, A. Amaya,
N. Tancredi, R.M. Lago, J. Hazard. Mater. 165 (2009) 1016–1022.
[22] J.-F. Boily, A.R. Felmy, Geochim. Cosmochim. Acta 72 (2008) 3338–3357.
[23] S. Krehula, S. Music, J. Mol. Struct. 976 (2010) 61–68.
[24] X. Lianga, S. Zhu, Y. Zhong, J. Zhu, P. Yuan, H. Hea, J. Zhang, Appl. Catal. B 97
(2010) 151–159.
The authors thank the financial support of Fapemig, CNPq,
CAPES, and Petrobras-Brazil.
References
[25] G. Gnanaprakash, J. Philip, T. Jayakumar, B. Raj, J. Phys. Chem. B 111 (2007)
7978–7986.
[26] A.L. de Abreu, I.R. Guimarães, A.S. Anastácio, M.C. Guerreiro, J. Mol. Catal. A:
Chem. 356 (2012) 128–136.
[1] F. Li, C. Kou, Z. Sun, Y. Hao, R. Liu, D. Zhao, J. Hazard. Mater. 205–206 (2012)
164–170.
[27] L.C.A. Oliveira, J.D. Fabris, R.M. Lago, Stud. Surf. Sci. Catal. 130 (2000) 2165–2170.
[28] W.F. Souza, I.R. Guimaraes, D.Q. Lima, C.L.T. Silva, L.C.A. Oliveira, Energy Fuels
23 (2009) 4426–4430.
[29] E.G. Garrido-Ramírez, B.K.G. Theng, M.L. Mora, Appl. Clay Sci. 47 (2010)
182–192.
[30] X. Xue, K. Hanna, N. Deng, J. Hazard. Mater. 166 (2009) 407–414.
[31] W.F. Souza, M.C. Pereira, L.C.A. Oliveira, Fuel 96 (2012) 604–607.
[32] L.C. Caero, F. Jorge, A. Navarro, A. Gutierrez-Alejandre, Catal. Today 116 (2006)
562–568.
[33] A. Ishihara, D. Wang, F. Dumeignil, H. Amano, E.W. Qian, T. Kabe, Appl. Catal. A
279 (2005) 279.
[2] F.A. Duarte, P.A. Mello, C.A. Bizzi, M.A.G. Nunes, E.M. Moreira, M.S. Alencar, H.N.
Motta, V.L. Dressler, É.M.M. Flores, Fuel 90 (2011) 2158–2164.
[3] A. Giuseppe, M. Crucianelli, F. De Angelis, C. Crestini, R. Saladino, Appl. Catal. B
89 (2009) 239–245.
[4] X. Cui, D. Yao, H. Li, J. Yang, D. Hu, J. Hazard. Mater. 205–206 (2012) 17–23.
[5] M.A. Sobati, A.M. Dehkordi, M. Shahrokhi, Fuel Process. Technol. 91 (2010)
1386–1394.
[6] S. Liu, B. Wang, B. Cui, L. Sun, Fuel 87 (2008) 422–428.
[7] X. Ma, A. Zhou, C. Song, Catal. Today 123 (2007) 276–284.
[8] J. Zhang, A. Wang, X. Li, X. Ma, J. Catal. 279 (2011) 269–275.
[9] Y. Jia, G. Li, G. Ning, C. Jin, Catal. Today 140 (2009) 192–196.
[10] P. Zeuthen, K.G. Knudsen, D.D. Whitehurst, Catal. Today 65 (2001) 307–314.
[11] J.M. Campos-Martin, M.C. Capel-Sanchez, J.L.G. Fierro, Green Chem. 6 (2004)
557–562.
[34] Z. Juan, Z. Dishun, Y. Liyan, L. Yongbo, Chem. Eng. J. 156 (2010) 528–531.
[35] H. Lü, J. Gao, Z. Jiang, F. Jing, Y. Yang, G. Wang, C. Li, J. Catal. 239 (2006)
369–375.