Fuel 210 (2017) 380–389

Contents lists available at ScienceDirect

Fuel
journal homepage: www.elsevier.com/locate/fuel

Full Length Article

Thermodinamic study of a magnetic molecular imprinted polymer for MARK

removal of nitrogenous pollutant from gasoline
Liz Nayibe Martínez Saavedraa, Bruno Eduardo Lobo Baetaa, Márcio César Pereirab,
⁎
Luiz Carlos Alves de Oliveirac, Adilson Candido da Silvaa,
a
Departamento de Química, Instituto de Ciências Exatas e Biológicas (ICEB), Universidade Federal de Ouro Preto (UFOP), 35400-000 Ouro Preto, Minas Gerais, Brazil
b
Instituto de Ciência, Engenharia e Tecnologia (ICET), Universidade Federal dos Vales do Jequitinhonha e Mucuri (UFVJM), Campus Mucuri, 39803-371 Teófilo Otoni,
Minas Gerais, Brazil
c
Departamento de Química, Universidade Federal de Minas Gerais (UFMG), 31270-901 Belo Horizonte, Minas Gerais, Brazil

A R T I C L E I N F O A B S T R A C T

Keywords: In this work, magnetic molecularly imprinted polymers (MMIP) and magnetic non-imprinted polymers (MNIP)
Quinoline were tested as selective adsorbents for quinoline in n-heptane. The synthesis was carried out using chloroform as
Gasoline a porogenic solvent and maghemite (γ-Fe2O3) as the magnetic component. The thermogravimetric data revealed
Adsorption that the MMIP and MNIP are formed by 73 and 81 wt% of the polymer, respectively. The powder X-ray dif-
Selectivity
fraction patterns of the MMIP and MNIP confirmed that the magnetic component in the polymers was maghe-
magnetic molecularly imprinted polymers
mite. The adsorption equilibrium data of quinoline displayed that the Sips models better fitted the isotherms.
(MMIP)
The maximum quinoline adsorption capacity (qmax) at 298 K was 25 and 14 mg g−1 for the MMIP and MNIP,
respectively. The kinetics of quinoline adsorption followed the Elovich model. The negative values observed for
the Gibbs free energy (ΔG°) suggest a spontaneous adsorption while the enthalpy values (ΔH°) indicate that the
quinoline was attached to the polymers by physisorption. Selectivity tests showed that the MMIP was about 6
times more selective than MNIP for quinoline adsorption. Also, the MMIP was efficient to remove quinoline from
a complex matrix of gasoline (10.35 mg/g). Because of its magnetic properties, the MMIP could be easily re-
covered and reused for at least five adsorption cycles without significant loss in the adsorption capacity.

1. Introduction well-known catalytic processes in the refining industries. However, the

Concerns at a global level with the increased contamination rate of
the air, water, and soil caused by the industrial developments have
imposed stricter control measures against the emission of pollutants.
Heterocyclic nitrogen and sulfur compounds have received particular
attention in the last years regarding the generation of toxic molecules
such as NOx and SOx during their combustion [1]. Therefore, the re-
moval of these compounds from complex matrixes such as those found
in fuels is a significant challenge.
In the last years, the stricter environmental standard has been im-
posed with the aim of decrease the level of these contaminants in the
oil. In Brazil, the level of sulfur present in the diesel was reduced to less
than 50 ppm while for countries of European Union it must be than
15 ppm [2–5].
Nowadays, the most used processes for the removal of heterocyclic
nitrogen and sulfur compounds are based on hydrotreatments as hy-
drodenitrification (HDN) and hydrodesulfurization (HDS), which are
nitrogen compounds as the quinoline in the oil may hamper the effi-
ciency of such process, since these compounds are known to cause the
poisoning of the catalyst. Thus, technologies such as HDN and HDS used
for the removal of benzothiophene (BTs), dibenzothiophene (DBTs) and
indoles have not been efficient [6]. It should be emphasized that only a
small concentration equivalent to 5 ppm of quinoline is enough for
poisoning the catalyst [2].
An alternative for increasing the efficiency of the hydrotreatments
consists of removing the sulfur and nitrogen compounds by adsorption
on the surface of nanomaterials. However, the use of universal ad-
sorbents as activated carbon has some drawbacks such as low reuse
capacity and low selectivity. To solve these drawbacks the design of
selective adsorbents based on molecularly imprinted polymers (MIP)
has been proposed [7,8].
The MIPs are prepared by the bulk copolymerization of functional
monomer and cross-linker in the presence of a template molecule. After
polymerization, the material needs to be mechanically crushed, which

⁎
Corresponding author.
E-mail address: adilsonufla@gmail.com (A.C. da Silva).

http://dx.doi.org/10.1016/j.fuel.2017.08.087
Received 18 April 2017; Received in revised form 20 August 2017; Accepted 25 August 2017
0016-2361/ © 2017 Elsevier Ltd. All rights reserved.
L.N.M. Saavedra et al.
results in particles with an irregular shape. Irregular particle size may
hamper its use in adsorption column. Alternatively, the MIPs synthe-
sized by the bulk method can be utilized in a continuous stirred–tank
reactor. However, by this process, the recovery of the MIP adsorbent
from the fluid is difficult to perform. Therefore, in this paper, we focus
on the synthesis of a magnetic molecularly imprinted polymer (MMIP).
The magnetic properties of these materials allow them to be easily re-
covered in magnetic separators [9–11]. After separation, the template
can be extracted and the MMIP adsorbents reused.
Extensive efforts have been made to synthesize and utilize these
materials as selective adsorbents. The few studies reported on the use of
MMIP adsorbents for removing pollutants derived from petroleum
[8,17], dyes [18], and antibiotics [19,20] have shown to be efficient.
Therefore, in this work, it was developed magnetic molecularly im-
printed polymers (MMIP) to be used as selective adsorbents for qui-
noline contaminated liquids. The kinetic and thermodynamic studies on
quinoline adsorption by MMIP in n-heptane and gasoline were de-
scribed in detail.
2. Materials and methods
2.1. Chemicals
Quinoline (98%), Dibenzothiophene (98%), Carbazole (≥95%),
methacrylic acid (MAA), ethylene glycol dimethacrylate (EGDMA), and
2, 2′-azobisisobutyronitrile (AIBN) were purchased from Sigma
Chemical Company, U.S.A. Toluene (Quimex), cyclohexane (Dinâmica),
methanol, NaOH (Synth), and n-Heptane (Neon Ltda) were acquired
from Brazilian Chemical Industries. Glacial acetic acid (Merck KGaA)
from Germany. All chemicals were used as received from the manu-
facturer without additional purification.
381
Fuel 210 (2017) 380–389
Fig. 1. Schematic process for the MMIP preparation.
2.2. Apparatus
Thermogravimetric analyses were performed on an SDT2960 from
room temperature to 1000 °C with a rate of 10 °C per minute under an
inert atmosphere of N2. FTIR spectra were scanned in a range of
400–4000 cm−1 using an ABB Bomem FTIR spectrometer equipped
with accessories of diffuse attenuated reflectance (DRIFT). The quan-
tification and identification of quinoline, dibenzothiophene, and car-
bazole were carried out on a spectrophotometer UV–VIS (Hewlett
Packard 8453) in the wavelength of 314 nm, and on a Gas
Chromatograph BID 2010 Plus SHIMADZU equipped with plasma de-
tector and a column SH-Rtx-5. SEM images were carried out on a JSM-
5500 scanning electron microscope. Surface area and porosity were
determined from adsorption experiments using an AutosorbIQ surface
area and pore size analyzer. X-ray diffraction analyses were performed
using a Panalytical Empyrean diffractometer equipped with a graphite
monochromator, nickel filter and CuKα radiation (λ = 1.54056 Ǻ,
40 kV, and 20 mA). For the analysis of the polymers, a scanning (2θ) of
10–80° and step of 0.5° min−1 was applied.
2.3. Preparation of γ-Fe2O3
γ-Fe2O3 nanoparticles were prepared by dissolving 5.56 g Fe
(SO4)2(NH4)2·6H2O in highly purified water. After this stage 100 mL of
NaOH (2 mol L−1) were added slowly drop by drop until the formation
of a green precipitate. After the formation of a precipitate, 5 mL of H2O2
30% (v/v) were immediately added with continuous stirring. The pre-
cipitate was washed several times with distilled water and dried in a
vacuum desiccator at room temperature.
L.N.M. Saavedra et al.
2.4. Preparation of MMIP and MNIP
The magnetic molecularly imprinted polymer was synthesized by
the method “Bulk”. The polymerization mixture was formed adding the
template (Quinoline) in 1 mmol (0.13 mL) and 3 mmol (0.26 mL) of
functional monomer (MAA) dissolved in 15 mL of chloroform in a
closed amber bottle. The mixture was kept under stirring for 24 h. After
the solubilization, 22.5 mmol (4.25 mL) of the cross-linking agent
(EGDMA) and 1 g of maghemite γ-Fe2O3 previously solubilized in 1 mL
of oleic acid were added to the reaction. This mixture was maintained
under ultrasonic stirring for 30 min. After this time 1 mL of the radical
initiator (AIBN) was added. The reaction medium was purged with the
N2 flow for 8 min to ensure an inert atmosphere. The polymerization
was carried out at 60 °C under constant stirring for 24 h. The template
was removed by Soxhlet extraction with a solution containing MeOH
and glacial acetic acid (90:10 v/v) at 80 °C for 48 h until no quinoline
was detected in the eluate. At the same time, the magnetic non-im-
printed polymer (MNIP) was prepared as a reference using the same
reaction mixture without the addition of the template and being treated
by the same procedure used for preparing the MMIP. A schematic re-
presentation of the process employed for the preparation of MMIP is
shown in Fig. 1.
2.5. Adsorption tests
In the adsorption studies, n-heptane was used as a solvent for si-
mulating the gasoline medium. The quinoline quantification was made
in triplicate and the adsorption capacities determined by Eq. (1).
(Ci−Cf ) v
qe =
m (1)
−1
where qe (mg g ) is the equilibrium adsorption capacity, Ci and Cf
(mg L−1) are the adsorbate concentrations before and after the ad-
sorption, respectively; v (L) is the volume of the solution, and m (g) is
the mass of adsorbent.
Adsorption isotherm studies were carried out in a Shaker Tecnal
(TE-421) with temperature control device at 288.15, 298.15, 308.15,
and 318.15 K. Quinoline solutions were prepared at different con-
centrations 4, 12, 20, 28, 36, 44, 52, 60, 68, 76 and 82 mg L−1 and in
each assay were added 10 mg of the polymeric materials in an amber
bottle with 5 mL of quinoline solutions previously prepared. The sam-
ples were kept under stirring at 180 rpm for 4 h. All the experiments of
isotherms adsorption were carried out in triplicate. The data were fitted
to nonlinear models of Langmuir, Freundlich, Sips, and Tempkin.
2.6. Kinetic studies
5 mL of 50 mg L−1 quinoline solution was placed in contact with
10 mg of MMIP and MNIP and kept under stirring at 180 rpm in a
Shaker Tecnal incubator (TE-421) at 288.15, 298.15, 308.15, and
318.15 K. Aliquots of 2 mL were taken at predetermined times (5, 15,
30, 60, 90, 120, 180, 240 and 300 min), centrifuged and analyzed. The
experimental data were fitted to nonlinear pseudo-second order,
Elovich, and intraparticle diffusion models.
2.7. Thermodynamic parameters
The thermodynamic of quinoline adsorption on MMIP and MNIP
was evaluated by determining the Gibbs free energy (ΔG°), enthalpy
(ΔH°), and entropy (ΔS°). The parameter ΔG° was calculated by the Eq.
(2);
ΔG° = −RTlnKp (2)
−1
where Kp (L g ) is the thermodynamic equilibrium constant, T(K) is
the absolute temperature, R (8.314 J mol−1 K−1) is the ideal gas con-
stant. The equilibrium thermodynamic constant was obtained from the
382
Fuel 210 (2017) 380–389
Langmuir equation [32–34]. The variation of Gibbs free energy also is
related to the change in entropy and enthalpy at a constant tempera-
ture, thus:
−ΔH ° ⎞ 1 ΔS °
lnKp = ⎛ ⎛ ⎞+
⎝ RT ⎠ ⎝ T ⎠ R (3)
The plot lnKp versus 1/T follows a straight line, and the slope and
intercept coefficient values allow to obtain the values of ΔH° and ΔS°,
respectively. The equation of Van't Hoff shows the dependence of the
adsorption equilibrium constant and the temperature [35].
2.8. Selectivity tests for the quinoline adsorption
Experiments involving competitive adsorption of quinoline with
other compounds by the MMIP were compared to the corresponding
MNIP to assess the molecular imprinting effect set up in the polymers. A
solution with carbazole (CB), dibenzothiophene (DBT), and quinoline
(QUI) of 50 mg L−1 in n-heptane was prepared. The experiment was
done by adding 10 mg of MMIP or MNIP in a bottle containing 5 mL of
the ternary solution. The flasks were subjected to constant stirring at
180 rpm and temperature of 298.15 K for 6 h. After the adsorption, the
supernatant was separated and analyzed as described in Section 2.2.
Through the parameters obtained in equations (4), (5), and (6) specific
coefficients were determined to evaluate the selectivity of the printed
materials. The distribution coefficient Kd (L g−1) for each compound
was calculated by the Eq. (4).
(C i−C e) V
Kd = x
Ce M (4)
Where Ci and Ce are the initial and equilibrium concentration
(mg L−1) of quinoline and other compounds (DBT and CB), V is the
solution volume (L), and M is the polymer mass (mg). Selectivity
coefficients for quinoline about the competitive compounds (CC) are
defined by Eqs. (5) and (6).
K d (quinoline)
KMMIP =
K d (CC) (5)
KMNIP =
K d (quinoline)
K d (CC) (6)
where Kd (quinoline) and Kd (CC) are distribution constant of quinoline
and competitor compounds, respectively. From Eqs. (5) and (6) the
relative selectivity coefficient K′ was defined as in Eq. (7).
kMMIP
k′ =
kMNIP (7)
2.9. Adsorption of quinoline in gasoline
The adsorption study of MMIP in gasoline was evaluated with the
addition of 80 mg L−1 of quinoline in a real matrix. The experiment was
carried out by adding 10 mg of the respective adsorbent in a 20 mL
amber bottle containing 10 mL of gasoline in which quinoline was
added. The flasks were kept under constant stirring at 180 rpm and
298.15 K for 6 h. After adsorption, the supernatant was separated and
analyzed as described in Section 2.2.
2.10. Recyclability studies
The recyclability tests of MMIP were performed under the same
conditions outlined in Section 2.8. After the adsorption, the adsorbent
was collected and separated from the solution. The adsorbent was
washed several times at 25 °C with a mixture of methanol and acetic
acid in the ratio 9:1 (v/v) until no quinoline was detected in the el-
uated. Five regeneration cycles were performed.
L.N.M. Saavedra et al. Fuel 210 (2017) 380–389

Fig. 2. (a) Thermogravimetric profiles and (b) FTIR spectra of MMIP and MNIP.

3. Results and discussion (163 m2 g−1), probably due to the higher content of maghemite
(79 m2 g−1) in MMIP than that in the MNIP. Moreover, the amount of
3.1. Characterization of the imprinted and non-imprinted polymers mesoporous of MMIP was significantly greater than that of MNIP. It
should be noted that the molecular imprinting caused an increase in the
3.1.1. TGA and FTIR analysis average pore diameter of MMIP (103.5 Å) compared to that of MNIP
Fig.2a presents the results of TGA for the MMIP and MNIP. It is (69.3 Å), which was due to the imprinting process and further removal
possible to observe an initial mass loss of 10% at 50 °C that may be of the template.
attributed the presence of volatile compounds trapped inside the Scanning electron micrographs of the MMIP and MNIP are shown in
polymer cavities [36]. The higher mass loss for MNIP (71%) and MMIP Fig. 3. It can be observed that both polymers exhibited an agglomerated
(63%) were observed in the range of 250 to 450 °C. For the polymers of pseudo-spherical particles, thus creating a rough surface that might
studied no loss of mass was observed after 450 °C. The TGA data re- minimize the diffusion effects.
vealed that the MMIP presented 27 wt% of γ-Fe2O3 in it structure while TEM images for the MMIP and MNIP are shown in Fig. 4. It can be
the MNIP had 19 wt%. This may occur due to the loss of maghemite observed that both polymers exhibited quasi-spherical particles of
during the synthesis since the same amount of iron oxide was added to maghemite (darker) dispersed in the organic matrix formed by the
both materials. It should be noted that the interaction between the polymerization step of MAA during the MMIP synthesis.
nitrogen of quinoline (Lewis base) and the γ-Fe2O3 (Lewis acid) occurs
only in MMIP due to the presence of the template (quinoline), which 3.1.3. X-ray diffraction of MMIP and MNIP
may favor the incorporation of maghemite in the polymer. The XRD patterns of MMIP and MNIP (Fig.5a) displayed diffraction
Fig. 2b shows that the FTIR spectra of MMIP and MNIP are similar, peaks typical of maghemite, which confirms the attachment of ma-
indicating that all the quinoline in MMIP was removed after the ex- ghemite to the polymer. No diffraction was observed for the polymers
traction stage. A wide band in the range of 3672–3170 cm−1 belonging indicating that they are amorphous. Also, it is possible to observe that
to OeH stretching vibration of functional monomers (MAA) was ob- the intensities of the diffraction peaks due to maghemite in the MMIP
served [37]. The band at 2960 cm−1 was attributed to CeH asymmetric were more intense than those found for the MNIP. These results cor-
stretching of eCH2 and eCH3 in the polymer chain and EGDMA. The roborate with the results achieved from TGA analysis (Fig.2a), which
band at 1480 cm−1 was assigned to eCOO stretching vibration of suggests a greater insertion of the maghemite to the MMIP.
MMA. Stretching vibrations at 1750 cm−1 correspond to eC]O, and
the signals at 1230 and 1125 cm−1 are characteristics of the cross-
linking agent (EGDMA) [8–38]. A broad band in the range of 3.2. Adsorption kinetics
625–560 cm−1 was attributed to the stretching vibration of the FeeO
bond, which suggests that the γ-Fe2O3 was successfully inserted into the Fig. 6a-c show the adsorption kinetics of quinoline by MMIP and
polymer network [18–38]. MNIP. It is possible to observe an increase in the removal of quinoline
with time. At the equilibrium, the maximum adsorption capacity (qe)
after 240 min and 308.15 K was 11.52 mg g−1 for MNIP and
3.1.2. Surface area, porosity, SEM, and TEM images of MMIP and MNIP 14.30 mg g−1 for MMIP. The higher value of qe for MMIP, when com-
Table 1 shows the results of nitrogen adsorption/desorption iso- pared to MNIP, can be attributed the higher affinity of quinoline by the
therms for the MMIP and MNIP. The value of specific area for the MMIP selective cavities, while the adsorption onto the MINP occurs possibly
(190 m2 g−1) was higher than that observed for the MNIP by non-selective interactions. The higher difference observed between
the MMIP and MNIP was achieved at 298.15 K. The interpretation of
Table 1 the experimental kinetic data was evaluated by pseudo-second-order,
Textural data obtained for MNIP and MMIP. non-linear Elovich, and intraparticle diffusion models.
The results obtained were summarized in Tables 2 and 3. The va-
Material Specific area Average pore diameter Average pore volume
lidity of models was evaluated by the correlation coefficient (R2), the
(m2 g−1) (Å) (cm3 g−1)
comparison between the values of Qe obtained experimentally and
Micro Meso Micro Meso Total those achieved by the models as well as means of root-mean-squared
error (RMSE). It was found that the Elovich and intraparticle diffusion
γ-Fe2O3 79 13.2 74.3 0.05 1.20 1.25
models well described the adsorption kinetic of the polymers compared
MNIP 163 17.5 69.3 0.06 1.11 1.18
MMIP 190 17.8 103.5 0.07 1.07 1.14 to the other models. The Elovich model indicates that the active sites of
the adsorbents are of heterogeneous nature and present different

383
L.N.M. Saavedra et al.
energies of activation [39]. The evaluation of the model parameters was
done considering the temperature of 308 K because it was the tem-
perature in which were observed the higher values of Qe for the MMIP
and MNIP. For the MMIP the value of the constant a related to the rate
of sorption was 77.56 mg g−1 min−1 while that to MNIP the value was
42.58 mg g−1 min−1. The value of constant b that is related to the
384
Fuel 210 (2017) 380–389
Fig. 3. SEM images of (a) MNIP and (b) MMIP.
Fig. 4. TEM images of (a, b) MMIP and (c, d) MNIP.
Fig. 5. XRD patterns for (a) maghemite and (b) MNIP e
MMIP.
surface coverage was higher to the MNIP (0.78 g mg−1) when com-
pared to the MMIP (0.69 g mg−1). This indicates that adsorption on
MNIP does not occur at specific sites [40]. As the constant b is related to
extend surface coverage, the increase of the concentration of quinoline
will decrease the available adsorption surface.
Fig.6c shows the plot of qt versus t1/2 indicating multilinearity with
L.N.M. Saavedra et al. Fuel 210 (2017) 380–389

three stages in the adsorption process. The first stage represents the
transport of the particle from boundary layer to the surface of the ad-
sorbent, second stage corresponds to the gradual adsorption in which
intraparticle diffusion is the rate-limiting step, and the third stage is the
equilibrium final, where the diffusion decreased in function of the low
concentration of quinoline in the solution and the low availability of
active adsorption sites [41,42]. When the intraparticle diffusion is the
rate-limiting stage during the adsorption, the plot passes through the
origin. If the plot does not pass through the origin, the linear coefficient
obtained from the intraparticle diffusion model can indicate the
thickness of boundary layer [50]. It is possible to observe in Fig.6c that
the thickness of the boundary layer for the MMIP was higher when
compared to MNIP, which indicate a larger number of binding sites on
the external surface of the adsorbent through the outer liquid film [50].

3.3. Adsorption isotherms

Fig. 7 shows the adsorption isotherms at different temperatures

obtained for the studied polymers. It is possible to observe an increase
of the adsorption in higher temperature. The adsorption isotherms were
adjusted with various models and the parameters obtained can be seen
in Table 4. Among the fitted models, the Sips showed the best fit to the
experimental data for all materials studied. The best fit of these models
can be rationalized by the lower values of RMSE, satisfactory values of
R2 and similarity of Qe experimental with those values predicted by the
model.
The Sips model bypasses some limitations present in the Freundlich
equation to growing concentration of the adsorbate. This model too is
named as Langmuir-Freundlich by being a combined form of the
Langmuir and Freundlich models. It is used to describe adsorption onto
heterogeneous surfaces. This agrees with the results found in the kinetic
study that showed the data to be fitted to the Elovich model.
At 298 K, it is possible to observe that the values of equilibrium
adsorption capacity (Qe) obtained from the Sips model were higher in
the MMIP (25.24 mg g−1) when compared with the MNIP
(14.01 mg g−1). Also, it was noted an increase in the difference be-
tween the MMIP and MNIP at a higher temperature, which confirms the
molecular imprinting on the surface of the adsorbent [43]. Differently,
state of the art on quinoline adsorption by MIP indicate a decreasing of
Qe with the increase of the temperature. This suggests that the ma-
ghemite nanoparticles in the MMIP change the adsorption mechanism
of quinoline. The n parameter calculated by SIPs model indicates the
heterogeneity of the system and can vary from 0 to 1. If n is 1, the
system is predominantly homogeneous matching the Langmuir model,
while that n < 1 indicates heterogeneity on the surface. All n values
(see Table 4) were < 1 suggesting that in the polymers there is a dis-
tribution of energetically heterogeneous sites [44].

3.4. Thermodynamics parameters

Fig. 6. (a) Pseudo-second order and Elovich plots for (a) MNIP and (b) MMIP (c) in-
traparticle diffusion plots.
The variation of Gibbs free energy (ΔG°), enthalpy (ΔH°), and en-
tropy (ΔS°) are the real indicators for the application of the adsorption
process. The obtained parameters from the equation of ΔG° and Van’t
Hoff are given in Table 5.
The negative values of ΔG° indicated spontaneity of the adsorption
process [45]. The values decrease with the increase of the temperature
indicating higher spontaneity at a higher temperature. The ΔG° values
close to −20 kJ mol−1 suggest the presence of electrostatic interactions
between the charged molecules with the surface, indicating a physical
adsorption [46,47]. The ΔH° values obtained were less than
80 kJ mol−1, which suggest a physisorption process. The positive ΔS°
values show an increase in the spontaneity on the solid–liquid surface
and reversibility in the adsorption process [48]. Also, the higher TΔS°
values than those of ΔH° suggest that the adsorption process is en-
tropically governed. The release of the solvation solvent (n-heptane)
from the target molecule (quinoline) and of the sites on the surface of

385
L.N.M. Saavedra et al. Fuel 210 (2017) 380–389

Table 2
Kinetic parameters obtained for MNIP and MMIP.

Polymers MNIP MMIP

Model Parameters Temperature

288 K 298 K 308 K 318 K 288 K 298 K 308 K 318 K

−1
qe,exp (mg g ) 8.40 9.34 11.52 10.36 10.09 12.85 14.3 12.51
PseudoSecond Order qe,calc (mg g−1) 8.14 8.87 11.2 10.2 9.51 12.04 13.31 12.103
K2 (min−1) 9.90E−3 0.14 0.02 0.11 0.07 0.04 0.02 0.055
X2 0.16 0.09 0.15 0.04 0.33 0.74 0.77 0.134
Adj. R2 0.83 0.98 0.95 0.84 0.51 0.60 0.81 0.845
R2 0.85 0.98 0.96 0.86 0.54 0.64 0.83 0.864
RMSE 0.13 0.1 0.12 0.06 0.17 0.23 0.23 0.096

Elovich a (mg g−1 min−1) 485.646 4.48E+8 42.58 1.30E+9 7.20E+4 1.90E+3 77.56 2.30E+5
b (g mg−1) 1.432 2.917 0.78 2.65 1.75 1.06 0.69 1.47
X2 0.032 0.016 0.09 0.03 0.05 0.09 0.09 0.02
Adj. R2 0.964 0.997 0.97 0.89 0.92 0.95 0.98 0.98
R2 0.969 0.998 0.97 0.9 0.98 0.96 0.98 0.98
RMSE 0.056 0.041 0.08 0.04 0.06 0.08 0.07 0.03

Table 3 the adsorbent may contribute to an increase in the disorder degree of

Intraparticle diffusion parameters for MNIP and MMIP. the system, which justifies the higher values of ΔS°. The ΔS° values for
the MNIP were higher when compared to those obtained to MMIP. A
Polymers Intraparticle diffusion models
possible explanation for this behavior is the absence of specific binding
K1p K2p K3p C1 C2 C3 R1 2 R22 R32 sites on the surface, which increase the disorder degree and contributes
to higher values of ΔH°. Therefore, the system requires more energy to
(g mg−1 min−1/2) (mg g−1) release the surface.
MNIP 0.09 0.22 0.13 7.66 8.82 5.85 0.97 0.44 0.80
MMIP 0.26 0.27 0.12 8.65 8.86 10.87 0.99 1.00 0.82 3.5. Selectivity for the quinoline adsorption

In Fig. 8 it is possible to note a higher capacity of quinoline ad-

sorption onto MMIP (12.0 mg g−1) when compared with the values
obtained for competitive compounds as DBT (3.5 mg g−1) and

Fig. 7. Langmuir, Freundlich, Temkin, and

Sips plots at different temperatures for MNIP
and MMIP.

386
L.N.M. Saavedra et al. Fuel 210 (2017) 380–389

Table 4
Parameters obtained from the quinoline adsorption isotherms MMIP and MNIP.

Polymers MNIP MMIP

Model Parameters Temperature

288 K 298 K 308 K 318 K 288 K 298 K 308 K 318 K

−1
qexp (mg g ) 6.96 10.87 12.33 12.14 14.16 23.3 22.43 21.86
Freundlich Kf (mg g−1)(L g−1)1/n 0.59 0.68 0.68 0.97 0.93 1.58 1.66 1.88
1/n (L mg−1) 0.63 0.70 0.74 0.64 0.71 0.78 0.74 0.70
X2 0.33 0.53 0.65 1.36 1.31 3.97 2.03 4.03
Adj. R2 0.93 0.95 0.95 0.90 0.93 0.92 0.95 0.91
R2 0.96 0.98 0.98 0.95 0.97 0.96 0.98 0.96
RMSE 0.58 0.73 0.81 1.17 1.15 1.99 1.42 2.01

Langmuir Qmax (mg g−1) 12.23 23.67 28.98 22.19 32.65 40.75 32.73 45.41
b (L mg−1) 0.02 0.02 0.02 0.03 0.02 0.02 0.02 0.03
X2 0.14 0.27 0.37 0.78 0.83 2.70 1.03 2.36
Adj. R2 0.97 0.97 0.97 0.94 0.96 0.95 0.98 0.95
R2 0.98 0.99 0.98 0.97 0.98 0.98 0.98 0.97
RMSE 0.37 0.52 0.60 0.88 0.91 1.64 1.01 1.535

Sips Qmax (mg g−1) 7.96 14.01 14.53 12.51 15.99 25.24 27.05 23.7
b (L mg−1) 0.05 0.04 0.04 0.06 0.05 0.08 0.07 0.08
n 0.56 0.63 0.54 0.43 0.46 0.44 0.58 0.47
X2 0.02 0.11 0.06 0.12 0.21 0.20 0.09 0.24
Adj. R2 0.99 0.98 0.99 0.99 0.99 0.99 0.99 0.99
R2 0.99 0.99 0.99 0.99 0.99 0.99 0.99 0.99
RMSE 0.15 0.33 0.24 0.35 0.46 0.45 0.30 0.50

Temkin bT (kJ mol−1) 0.92 0.63 0.57 0.57 0.5 0.24 0.30 0.30
AT (L mg−1) 0.25 0.26 0.25 0.26 0.34 0.30 0.34 0.31
X2 0.09 0.68 0.96 0.57 2.63 1.65 1.60 0.96
Adj. R2 0.98 0.94 0.93 0.96 0.87 0.97 0.96 0.98
R2 0.99 0.97 0.97 0.98 0.94 0.98 0.98 0.99
RMSE 0.31 0.82 0.98 0.75 1.62 1.28 1.26 0.98

Table 5
Thermodynamics parameters for the adsorption of quinoline onto MMIP and MNIP.

Polymers ΔG° (kJ mol−1) TΔS°(kJ mol−1) ΔH°(kJ mol−1)

Temperature (K)

288,15 298,15 308,15 318,15

MNIP −18.28 −19.35 −20.50 −22.02 36.64 17.21

MMIP −18.26 −19.31 −20.37 −21.57 32.79 13.46

Table 6
Selectivity parameters for the competitive adsorption (quinoline/DBT/carbazole) on the
MMIP and MNIP.

Polymers Selectivity parameters Molecules

Quinoline DBT Carbazole

−1
MNIP qexp (mg g ) 2.50 0.21 5.04
Kd (L g−1) 0.05 0.0042 0.13
k – 13.13 0.44

MMIP qexp (mg g−1) 12.01 3.18 4.95

Kd (L g−1) 0.46 0.07 0.12
k – 6.57 3.8
k′ – 0.50 8.50

carbazole, which favors the binding of this compound onto imprinted

Fig. 8. Selectivity study for adsorption of quinoline onto MMIP and MNIP.
carbazole (4.6 mg g−1). Also, it is possible to observe that the carbazole
molecule have higher affinity to the MMIP when compared with the
DBT molecule. It may be due to the presence of nitrogen in the
cavities of MMIP. The sulfur of the DBT molecule difficult the binding
onto imprinted cavities. For the MNIP is observed a higher capacity of
adsorption for carbazole when compared with the quinoline. This is
possible due to the absence of specific binding sites stable on the surface
of the material.
As can be seen in Table 6, MMIP had higher selective coefficients for
quinoline/DBT and quinoline/carbazole systems than that for the

387
L.N.M. Saavedra et al.
Fig. 9. (a) Adsorption of quinoline onto MMIP in gasoline and (b) Recyclability of the MMIP after five adsorption cycles.
MNIP. The values of Kd for the quinoline were always greater than
those observed for MNIP. The k values > 1 for quinoline/DBT and
quinoline/carbazole suggest that the adsorption occurs preferentially
for quinoline. Therefore, the results show that the synthetic conditions
for MMIP generate an adsorbent with an excellent selectivity for qui-
noline even in the presence of similar molecules.
3.6. Adsorption of quinoline in real matrix and recyclability study
In Fig. 9a it is possible to observe that the maximum capacity ad-
sorption of quinoline from gasoline was 10.35 mg g−1. This result was
similar to that seen in the selectivity study and indicates that the MMIP
maintains its selective properties even in a complex sample containing
multi-compounds such as gasoline.
This result was similar to that observed in the selectivity study and
indicates that the MMIP maintains its selective properties even in a
complex sample containing multi-compounds such as gasoline. In ad-
dition, as can be seen in Fig.9b the MMIP showed an excellent re-
generation capacity after successive cycles of use (5 cycles). After the
fifth cycle, the regeneration calculated about the adsorption capacity in
the first cycle was maintained in 98%. Thus, MMIP becomes an at-
tractive adsorbent to be used in the removal of quinoline in a real
process, since it is easy to recovery due to its magnetic properties and
regeneration capacity.
4. Conclusion
A magnetic molecularly imprinted polymer (MMIP) was efficient to
promote the selective removal of quinoline from gasoline. The kinetic
and thermodynamic results indicated that adsorption occurs on an en-
ergetically heterogeneous surface and physisorption governs the pro-
cess. Also, it was found that the MMIP had a higher selectivity for
quinoline in the presence of other compounds in both simple and
complex samples as the gasoline. The MMIP appears as a potential
adsorbent to be used for the removal of compounds as quinoline, which
is responsible for passivating the catalysts during the hydrotreatments
(HDN and HDS) in the petroleum refineries.
Acknowledgments
The authors would like to thank the (CNPq), (CAPES), FAPEMIG,
and the Universidade Federal de Ouro Preto for the financial support.
To the Microanalysis Laboratory of the Universidade Federal de Ouro
Preto, a member of the Microscopy and Microanalysis Network of
Minas Gerais State/Brazil/FAPEMIG, for the mineral chemistry ana-
lyses. The authors would like to acknowledge the Nucleus of
388
Fuel 210 (2017) 380–389
Microscopy and Microanalysis at the Universidade Federal de Viçosa,
(http://www.nmm.ufv.br/) for providing the equipment and technical
support for experiments involving electron microscopy.
References
[1] Kim JH, Ma X, Zhou A, Song C. Ultra-deep desulfurization and denitrogenation of
diesel fuel by selective adsorption over three different adsorbents: a study on ad-
sorptive selectivity and mechanism. Catal Today 2006;111:74–83.
[2] Liu D, Gui J, Sun Z. Adsorption structures of heterocyclic nitrogen compounds over
Cu(I)Y zeolite: a first principle study on mechanism of the denitrogenation and the
effect of nitrogen compounds on adsorptive desulfurization. J Mol Catal A: Chem
2008;291:17–21.
[3] ANP, Resolução No 30, 29.6.2015 da Agência Nacional do Petróleo, Gás Natural e
Biocombustíveis (ANP), 30 (2015) 1–4.
[4] Wen J, Han X, Lin H, Zheng Y, Chu W. A critical study on the adsorption of het-
erocyclic sulfur and nitrogen compounds by activated carbon: equilibrium, kinetics
and thermodynamics. Chem Eng J 2010;164:29–36.
[5] Song C. An overview of new approaches to deep desulfurization for ultra-clean
gasoline, diesel fuel and jet fuel. Catal Today 2003;86:211–63. http://dx.doi.org/
10.1016/S0920-5861(03)00412-7.
[6] Liu K, Ng FTT. Effect of the nitrogen heterocyclic compounds on hydro-
desulfurization using in situ hydrogen and a dispersed Mo catalyst. Catal Today
2010;149:28–34.
[7] Lucci P, Núñez O, Galceran MT. Solid-phase extraction using molecularly imprinted
polymer for selective extraction of natural and synthetic estrogens from aqueous
samples. J Chromatogr A 2011;1218:4828–33.
[8] Huang W, Li H, Xu W, Zhou W, Zhou Z, Yang W. Selective Adsorption of
Dibenzothiophene using Magnetic Molecularly Imprinted Polymers. Adsorpt Sci
Technol 2012;30:331–44.
[9] Li Y, Li Y, Yin XF, Yin XF, Chen FR, Chen FR, et al. Synthesis of magnetic mole-
cularly imprinted polymer nanowires using a nanoporous alumina template.
Macromolecules 2006;39:4497–9.
[10] Rocha TS, Nasciminto ES, Silva AC, Oliveira HS, Marsalha EG, Oliveira LCA, et al.
Enhanced photocatalytic hydrogen generation from water by Ni(OH)2 loaded on
Ni-doped δ-FeOOH nanoparticles obtained by one-step synthesis. RSC Adv
2013:20308–14.
[11] Oliveira LCA, Zaera F, Lee I, Lima DQ, Ramalho TC, Silva AC, et al. Nb-doped
hematites for decomposition of isopropanol: evidence of surface reactivity by in situ
CO adsorption. Appl Catal A Gen 2009;368:7–21.
[12] Cao Y, Xu W, Wu X, Li Y, Li H, Huang W. Synthesis of a molecularly imprinted
polymer on silica-gel surfaces for the selective adsorption of indole from fuel oil.
Adsorpt Sci Technol 2013;31:489–502.
[13] Liu W, Liu X, Yang Y, Zhang Y, Xu B. Selective removal of benzothiophene and
dibenzothiophene from gasoline using double-template molecularly imprinted
polymers on the surface of carbon microspheres. Fuel 2014;117:184–90.
[14] Hu T, Zhang Y, Zheng L, Fan G. Molecular recognition and adsorption performance
of benzothiophene imprinted polymer on silica gel surface. J Fuel Chem Technol
2010;38:722–9.
[15] Castro B, Whitcombe MJ, Vulfson EN, Vazquez-Duhalt R, Bárzana E. Molecular
imprinting for the selective adsorption of organosulphur compounds present in
fuels. Anal Chim Acta 2001;435:83–90.
[16] Yu P, Sun Q, Li J, Tan Z, Yan Y, Li C. Selective removal of 2,4-dichlorophenol in
aqueous solutions by nanoparticles modified yeast using miniemulsion imprinting
polymerization. J Environ Chem Eng 2015;3:797–806.
[17] Niu D, Zhou Z, Yang W, Li Y, Xia L, Jiang B, et al. Preparation and characterization
of magnetic molecularly imprinted polymers for selective recognition of 3-methy-
lindole. Appl Polymer 2013:2859–66. http://dx.doi.org/10.1002/app.39511.
L.N.M. Saavedra et al.
[18] Lin Z, Zhang H, Peng A, Lin Y, Li L, Huang Z. Application of magnetic molecularly
imprinted polymers in the detection of malachite green in fish samples. Food Chem
2016;200:32–7.
[19] Huang M, Zhou T, Wu X, Mao J. Adsorption and degradation of norfloxacin by a
novel molecular imprinting magnetic Fenton-like catalyst. Chinese J Chem Eng
2015;23:1698–704.
[20] Chen L, Liu J, Zeng Q, Wang H, Yu A, Zhang H, et al. Preparation of magnetic
molecularly imprinted polymer for the separation of tetracycline antibiotics from
egg and tissue samples. J Chromatogr A 2009;1216:3710–9.
[21] Ho YS, Mckay G. The kinetics of sorption of divalent metal ions onto sphagnum
moss peat. Water Res 2000;34:735–42.
[22] Barkat M, Nibou D, Chegrouche S, Mellah A. Kinetics and thermodynamics studies
of chromium(VI) ions adsorption onto activated carbon from aqueous solutions.
Chem Eng Process Process Intensif 2009;48:38–47.
[23] Juang R, Chen M. Application of the Elovich equation to the kinetics of metal
sorption with solvent-impregnated resins. Ind Eng Chem Res 1997;36:813–20.
[24] Song G, Zhu X, Chen R, Liao Q, Ding Y-D, Chen L. An investigation of CO2 ad-
sorption kinetics on porous magnesium oxide. Chem Eng J 2016;283:175–83.
[25] Sekar M, Sakthi V, Rengaraj S. Kinetics and equilibrium adsorption study of lead(II)
onto activated carbon prepared from coconut shell. J Colloid Interface Sci
2004;279:307–13. http://dx.doi.org/10.1016/j.jcis.2004.06.042.
[26] Agarry S, Ogunleye O. Chemically treated kola nut pod as low-cost natural ad-
sorbent for the removal of 2,4-dinitrophenol from synthetic wastewater: batch
equilibrium, kinetic, and thermodynamic modelling studies. Turkish J Eng Environ
Sci 2014;38:11–40.
[27] F. Cabrera Capitán, Estudio de la adsorción isoterma de aniones ortofosfato por
óxidos metálicos existentes en la naturaleza, (1975) 54.
[28] Jeppu GP, Clement TP. A modified Langmuir-Freundlich isotherm model for si-
mulating pH-dependent adsorption effects. J Contam Hydrol 2012;129–130:46–53.
[29] Ho YS, Porter JF, McKay G. Equilibrium isotherm studies for the sorption of diva-
lent metal ions onto peat: copper, nickel and lead single component systems. Water
Air Soil Pollut 2002;141:1–33.
[30] Sánchez RA, Chamorro AF. Estudio de la adsorción de plomo en suelos de la región
minera en el distrito de buenos aires en el departamento del cauca, Colombia. Rev.
Ciencias Univalle. 2012;16:145–60.
[31] Chen C. Evaluation of equilibrium sorption isotherm equations. Open Chem Eng J
2013;7:24–44.
[32] Milonjić SK. A consideration of the correct calculation of thermodynamic para-
meters of adsorption. J Serbian Chem Soc 2007;72:1363–7.
[33] Zhou X, Zhou XIN. The unit problem in the thermodynamic calculation of ad-
sorption using the Langmuir equation. J Chem Eng Commun 2014;201:1459–67.
http://dx.doi.org/10.1080/00986445.2013.818541.
[34] Anastopoulos I, Kyzas GZ. Are the thermodynamic parameters correctly estimated
in liquid-phase adsorption phenomena? J Mol Liq 2016;218:174–85. http://dx.doi.
389
Fuel 210 (2017) 380–389
org/10.1016/j.molliq.2016.02.059.
[35] Denderz N, Lehotay J, Čižmárik J, Cibulková Z, Šimon P. Thermodynamic study of
molecularly imprinted polymer used as the stationary phase in high performance
liquid chromatography. J Chromatogr A 2012;1235:77–83.
[36] Niu D, Zhou Z, Yang W, Li Y, Xia L, Jiang B, et al. Preparation and characterization
of magnetic molecularly imprinted polymers for selective recognition of 3-methy-
lindole. J Appl Polym Sci 2013;130:2859–66.
[37] Lee SH, Doong RA. Adsorption and selective recognition of 17β-estradiol by mo-
lecularly imprinted polymers. J Polym Res 2012;19.
[38] Fayazi M, Taher MA, Afzali D, Mostafavi A. Preparation of molecularly imprinted
polymer coated magnetic multi-walled carbon nanotubes for selective removal of
dibenzothiophene. Mater Sci Semicond Process 2015;40:501–7.
[39] Inyang HI, Onwawoma A, Bae S. Soil & Tillage research the Elovich equation as a
predictor of lead and cadmium sorption rates on contaminant barrier minerals. Soil
Tillage Res 2016;155:124–32. http://dx.doi.org/10.1016/j.still.2015.07.013.
[40] Cheung CW, Porter JF, Mckay G. Sorption kinetic analysis for the removal of cad-
mium ions from effluents using bone char. Water Res 2001;35:605–12.
[41] Guimarães Gusmão KA, Alves Gurgel LV, Sacramento Melo TM, Gil LF. Application
of succinylated sugarcane bagasse as adsorbent to remove methylene blue and
gentian violet from aqueous solutions – Kinetic and equilibrium studies. Dye
Pigment 2012;92:967–74.
[42] Wang L, Zhang J, Zhao R, Li Y, Li C, Zhang C. Adsorption of Pb(II) on activated
carbon prepared from Polygonum orientale Linn.: kinetics, isotherms, pH, and ionic
strength studies. Bioresour Technol 2010;101:5808–14.
[43] Guo W, Hu W, Pan J, Zhou H, Guan W, Wang X, et al. Selective adsorption and
separation of BPA from aqueous solution using novel molecularly imprinted poly-
mers based on kaolinite/Fe3 O4 composites. Chem Eng J 2011;171:603–11. http://
dx.doi.org/10.1016/j.cej.2011.04.036.
[44] Moreno-Piraján JC, Gómez-Cruz R, García-Cuello VS, Giraldo L. Binary system Cu
(II)/Pb(II) adsorption on activated carbon obtained by pyrolysis of cow bone study.
J Anal Appl Pyrolysis 2010;89:122–8.
[45] Liu K, Zhu B, Feng Q, Wang Q, Duan T, Ou L, et al. Adsorption of Cu(II) ions from
aqueous solutions on modified chrysotile: thermodynamic and kinetic studies. Appl
Clay Sci 2013;80–81:38–45.
[46] Mahmoodi NM, Sadeghi U, Maleki A, Hayati B, Najafi F. Synthesis of cationic
polymeric adsorbent and dye removal isotherm, kinetic and thermodynamic. J Ind
Eng Chem 2014;20:2745–53.
[47] Bayramoglu G, Arica MY, Liman G, Celikbicak O, Salih B. Removal of bisphenol A
from aqueous medium using molecularly surface imprinted microbeads.
Chemosphere 2016;150:275–84.
[48] Ali RM, Hamad HA, Hussein MM, Malash GF. Potential of using green adsorbent of
heavy metal removal from aqueous solutions: adsorption kinetics, isotherm, ther-
modynamic, mechanism and economic analysis. Ecol Eng 2016;91:317–32.