Archives of

1. Introduction
Industrial waste management is one of the significant challenges in making the balance between industry and the environment. Dairy industry wastewater with high organics is considered as one of the most polluting kinds of wastewater. It is mainly caused by washing tanks, pipes, and facilities. Approximately 0.5%-5% of input raw milk is detected in the effluent of the dairy industry, making the chemical and biochemical oxygen demand (COD and BOD) of dairy wastewater extremely high. On the other hand, such effluents would stink in a short time and cause an intense odor [1,2]. The conventional treatments for dairy wastewater mainly include physical (flotation, flocculation, and coagulation) and biological (aerobic and anaerobic) methods which are time-consuming and costly [3]. Considering that the main goal of dairy effluent treatment is to remove organic compounds, alternative treatment techniques (e.g., adsorption) could be taken into account. Adsorption is still a promising method owing to its effortless implementation, low operating costs, and the fact that it does not need sophisticated equipment [4-6].
The selection of a suitable and efficient adsorbent is the infrastructure of adsorption processes. Traditionally, zeolite as an adsorbent has always been considered because of its individual properties such as a high surface area and a large number of cavities. Even today, zeolite-based adsorbents have extensive applications [7-9]. Magnesia (magnesium oxide, MgO) has also been utilized as an adsorbent thanks to its unique properties such as facile synthesis, environmental compatibility, accessibility, and non-toxicity [10-12].
Zeolite-based magnesia compounds are among the most popular adsorbents that, by a synergistic effect, could overcome most problems triggered by the removal of organic and inorganic pollutants in liquid or gaseous phases [13-15].
Despite all the above-mentioned descriptions, the separation of the used adsorbent is the weakness of all the adsorbents, including the magnesia-zeolite composition. The most prevalent methods include filtration and precipitation, which are always problematic, since
Synthesis of Zeolite-Supported Magnesia/Magnetite Nanoparticles, and its Efficient Application in Ultrasound-Assisted Adsorption of Dairy Wastewater: Optimization and Modeling of the Process by RSM and ANN
Ali MehrizadID
Department of Chemistry, Tabriz Branch, Islamic Azad University, Tabriz, Iran
*Corresponding Author: Ali Mehrizad, Email: mehrizad@iaut.ac.ir
Received: June 3, 2021, Accepted: November 7, 2021, ePublished: September 29, 2022
https://jhygiene.muq.ac.ir
10.34172/AHS.11.3.78.1
Vol. 11, No. 3, 2022, 175-188
Original Article
Mehrizad
176 Arch Hyg Sci. Volume 11, Number 3, 2022
adsorbent particles may block the filter or pass through
it. Moreover, secondary pollution may occur by the
released adsorbents. In recent years, the magnetic
separation process has been presented as an alternative
method for overcoming these problems and facilitating
the separation and recovery of the adsorbents [16,17]. In
this regard, magnetite (iron oxide, Fe3O4) has been widely
used due to its facile synthesis, low cost, physicochemical
stability, and biocompatibility [18-20].
Ultrasound irradiation is able to accelerate physicochemical
processes due to aquatic cavitation, thus it is
expected that the adsorption process by the ultrasoundassisted
method is appropriate for wastewater treatment.
In fact, the cavitation phenomenon involves the
formation, growth, and collapse of bubbles in the liquid
phase, which can facilitate the adsorbate-adsorbent mass
transfer [21,22].
The aim/novelty of this study is to synthesize and evaluate
the performance of a new ternary magnesia-magnetitezeolite
composite (MgO-Fe3O4-zeolite) as an adsorbent
in the adsorption of dairy wastewater. The ternary
composite was prepared, characterized, and applied as
an adsorbent. To save time and cost, ultrasound-assisted
adsorption experiments were designed, optimized, and
modeled by response surface methodology (RSM) and
artificial neural network (ANN). Eventually, adsorption
isotherms, kinetics, and thermodynamics were evaluated
under optimal conditions.
2. Materials and Methods
2.1. Materials
Iron (II) chloride tetrahydrate (FeCl2.4H2O), iron
(III) chloride hexahydrate (FeCl3.6H2O), sodium
hydroxide (NaOH), and magnesium nitrate hexahydrate
(Mg(NO3)2.6H2O) were pursued from Merck
Company, Darmstadt Germany. Further, zeolite type
4A (Na12[(SiO2)12(AlO2)12]27H2O) was obtained from
Afrazand Mineral Company, Iran.
The dairy wastewater was collected from the industrial
wastewater treatment plant located in Parsabad, Iran. The
wastewater sample was obtained following the separation
of suspended solids/fat materials from the homogenizer
tank. The sample was stored at 4°C (COD: 900 ± 50 mg L-1).
Distilled (DI) water was used in all experiments, and
the pH of the solutions was adjusted using 0.1 M HCl or
NaOH.
2.2. Adsorbent synthesis and characterization
The adsorbent synthesis was performed according to the
following hierarchy:
To synthesize Fe3O4, 100 mL of DI water was poured into
a 3-necked round-bottomed flask, and the stoichiometric
amounts of iron (II) chloride (5 mmol) and iron (III)
chloride (10 mmol) were added under an N2 atmosphere.
The mixture was stirred vigorously. Once the temperature
of the mixture reached 60ºC, 10 mL of sodium hydroxide
(5 M) was gradually added, and the obtained mixture
was stirred for another 30 minutes. After cooling the
mixture, the sediments were gathered with a magnet and
washed with DI water to become neutral. The obtained
product was dried overnight at room temperature. Fe3O4
nanoparticles were obtained after crushing [23].
The magnesia-magnetite (MgO-Fe3O4) nanoparticles
were prepared through a procedure that was already
reported by Mahmoudi and Behnajady [24]. The
stoichiometric amounts of magnesium nitrate (10 mmol)
and sodium hydroxide (20 mmol) were separately
dissolved in 20 mL of DI water. Thereafter, 0.2 g of Fe3O4
was added to the solution containing Mg2 + ions, and
while the mixture was being heavily stirred, the aqueous
solution of NaOH was added dropwise. Afterward, the
sediments were centrifuged, washed with ethanol and DI
water, and dried overnight at room temperature. MgOFe
3O4 was obtained following powder calcination for 2
hours at 300ºC. The pure MgO was synthesized by the use
of this method without Fe3O4.
To prepare MgO-Fe3O4-zeolite, a proper amount of
zeolite was dispersed in 100 mL DI water in an ultrasonic
bath. Subsequently, the washed sediments from the
previous step (MgO-Fe3O4) were added to the zeolite
suspension, and the resulting mixture was stirred on a
magnetic stirrer for 4 hours. Following centrifugation and
drying at room temperature, the specimen was calcined
for 2 hours at 300°C until obtaining the composite [25].
The structural, morphological, surface, and magnetic
features of the synthesized powders were extensively
characterized by different methods, including X-ray
diffraction (XRD, X’ Pert Pro, Panalytical), field
emission-scanning electron microscopy (FE-SEM,
ZEISS-Sigma VP), and transmission electron microscopy
(TEM, ZEISS, EM10C, 100 kV). The other methods
were energy dispersive X-ray spectroscopy (EDX,
Oxford Instruments), vibrating sample magnetometer
(VSM, Meghnatis Daghigh Kavir Co. LBKFB), and
N2 adsorption/desorption isotherms (BElSORP Mini-
Microtrac Bel Corp).
2.3. Ultrasound-assisted adsorption experiments
Batch adsorption experiments were conducted on a
laboratory scale. Typically, 200 mL of the dairy effluent
with a specific pH and initial COD was poured into
an Erlenmeyer flask and sonicated after adding an
appropriate amount of the adsorbent. Sampling was
performed at a given time, and the adsorbent was seceded
from the solution by employing a strong magnet. The
supernatant was transferred into a vial, and COD was
quantified using HACH-DR/890 colorimeter equipped
with the HACH-DRB 200 thermal reactor as per the
instruction. The extent of the COD reduction (CODr
%) and the adsorption capacity (mgg-1) at time t (qt)
Arch Hyg Sci. Volume 11, Number 3, 2022 177
Synthesis of Zeolite-Supported Magnesia/Magnetite Nanoparticles, and its Efficient Application in Ultrasound
and equilibrium (qe) were calculated using Eqs. (1-3),
respectively:
0
0
(%) = ( - t ) ×100
r
C C
COD
C (1)
0 (
- ) = t ×
t
C C
q V
m (2)
0 (
- ) = e ×
e
C C
q V
m (3)
where C0, Ct, and Ce are the initial, time t, and the
equilibrium COD contents of the effluent (mgL-1),
respectively. Further, V and m denote the volume of
the solution (L) and the amount of the adsorbent (g),
respectively.
After performing preliminary trials and finding
the major influential parameters, the effects of four
factors at five different levels (Table 1) were surveyed
and optimized by the central composite design (CCD)
based on RSM (Design-Expert® 11 software). The ANN
modeling was also utilized to select the important
operational parameter and the reoptimization of the
adsorption process. The analysis of variance (ANOVA)
was performed to assess the accuracy and validity of the
models. Detailed explanations of RSM and ANN are
provided in our previous research work [26].
3. Results
3.1. Characterization of synthesized powders
The FE-SEM and TEM images were enacted to enlighten
the morphological properties of the prepared powders
(Figure 1). MgO and Fe3O4 spherical nanoparticles with
relatively uniform sizes (Figures 1a and b), as well as 4A
zeolite cubic particles (Figure 1c), were well formed based
on visual observation. Furthermore, we clearly observed
the widely distributed MgO aggregated with Fe3O4
particles (Figure 1d), along with the settled MgO-Fe3O4
on the porous surface of zeolite with abundant irregular
protrusions (Figure 1e).
In addition, the TEM image (Figure 1f) confirmed the
composite formation, in which the dark spots represented
the Fe3O4 nanoparticles and the other components
(MgO and zeolite) had light colors; the difference in
the components’ colors is due to their different levels of
electron permeability.
Figure 2a shows the XRD patterns of the prepared
powders. In the XRD pattern of MgO, five distinct peaks
appeared at the diffraction angles of 36.79, 42.83, 62.14,
74.44, and 78.46º, respectively. These angles correspond
to crystal planes with Miller indices (1 1 1), (2 0 0), (2
2 0), (3 1 1), and (2 2 2), which are fully matched with
the peaks of the pure cubic MgO index (JCPDS card no.
45-0946) [27]. For Fe3O4, the prominent peak at 35.71º
belongs to a crystal plane with Miller indices (3 1 1). The
other broad peaks at 2θ = 30.31º (2 2 0), 43.37º (4 0 0),
57.39º (5 1 1), and 62.86º (4 4 0), along with a noisy peak
at 74.02º (4 4 4), are in an acceptable agreement with the
standard magnetite diffraction pattern (JCPDS card No.
19-0629) [23]. The broad peaks suggest the formation
of fine particles with low crystallinity, while sharp peaks
indicate a highly crystalline structure [28]. Accordingly,
Table 1. Domain of the operational parameters and designed experiments,
along with experimental and predicted results
Parameters
Levels
-2 -1 0 + 1 + 2
X1: COD0 (mg L-1) 100 300 500 700 900
X2: Adsorbent
dosage (g L-1)
1 2 3 4 5
X3: pH 3 5 7 9 11
X4: Time (min) 5 10 15 20 25
Std.
Coded values of
parameters
CODr (%)
X1 X2 X3 X4 Exp. RSM Pred. ANN Pred.
1 -1 -1 -1 -1 53.11 53.49 53.58
2 1 -1 -1 -1 44.83 42.57 45.13
3 -1 1 -1 -1 63.82 62.38 64.51
4 1 1 -1 -1 51.92 50.91 52.37
5 -1 -1 1 -1 61.89 63.03 62.54
6 1 -1 1 -1 50.2 50.97 50.61
7 -1 1 1 -1 74.1 75.80 75.00
8 1 1 1 -1 63.21 63.19 63.89
9 -1 -1 -1 1 66.13 65.82 66.87
10 1 -1 -1 1 56.57 55.52 57.11
11 -1 1 -1 1 68.68 68.56 69.47
12 1 1 -1 1 59.17 57.70 59.76
13 -1 -1 1 1 74.08 75.75 74.98
14 1 -1 1 1 63.2 64.31 63.88
15 -1 1 1 1 80.43 82.36 81.46
16 1 1 1 1 70.09 70.37 70.91
17 -2 0 0 0 71.26 68.95 72.10
18 2 0 0 0 44.05 46.04 44.33
19 0 -2 0 0 57.32 56.76 57.88
20 0 2 0 0 71.47 71.71 72.31
21 0 0 -2 0 45.51 49.31 45.82
22 0 0 2 0 75.64 71.51 76.57
23 0 0 0 -2 58.25 58.79 58.82
24 0 0 0 2 79.17 78.31 80.17
25 0 0 0 0 60.89 60.47 61.52
26 0 0 0 0 57.93 60.47 58.50
27 0 0 0 0 63.8 60.47 64.49
28 0 0 0 0 61.28 60.47 61.92
29 0 0 0 0 59.43 60.47 60.03
30 0 0 0 0 59.51 60.47 60.11
Note. RSM: Response surface methodology; ANN: Artificial neural network;
COD: Chemical oxygen demand.
Mehrizad
178 Arch Hyg Sci. Volume 11, Number 3, 2022
Figure 1. FE-SEM images of MgO (a), Fe3O4 (b), 4A-zeolite (c), MgO-Fe3O4 (d), and MgO-Fe3O4-zeolite (e), TEM image of MgO-Fe3O4-zeolite (f). Note. FE-SEM:
Field emission scanning electron microscopy; MgO: Magnesium oxide iron oxide; Fe3O4: Iron oxide; TEM: Transmission electron microscopy
Figure 2. XRD patterns (a), Elemental mapping (b), VSM analysis (c), and N2 adsorption-desorption Plot (d) of prepared samples. Note. XRD: X-ray diffraction;
VSM: Vibrating sample magnetometer
Arch Hyg Sci. Volume 11, Number 3, 2022 179
Synthesis of Zeolite-Supported Magnesia/Magnetite Nanoparticles, and its Efficient Application in Ultrasound
the appearance of the magnetite XRD pattern represents
the formation of extremely fine particles with low
crystallinity, which was confirmed by the FE-SEM image
(Figure 1b). The XRD pattern of zeolite is in agreement
with the peaks of the 4A zeolite index (JCPDS card No.
39-0222) [29]. For MgO-Fe3O4, the peaks were obtained
through combining the peaks of the MgO and Fe3O4
specimens, indicating the formation of MgO-Fe3O4. As
observed in the MgO-Fe3O4-zeolite pattern, all the peaks
of the composite components (MgO, Fe3O4, and 4A
zeolite) were well observed, highlighting the successful
synthesis of the ternary composite without any impurity.
The composite’s EDX elemental mapping (Figure 2b)
exhibits the presence of Mg, O, Fe, Si, Al, and Na elements.
The elements were distributed in the composite texture.
The magnetic properties of the prepared powders were
studied with VSM (Figure 2c). The hysteresis showed the
ferromagnetic behavior of binary and ternary composites
with a saturation magnetization (Ms) in the range of 6-8
emug-1, which was lower than the magnetite sample. This
might be due to the presence of fewer magnetic magnetite
particles within the composites. Based on the obtained
data, the Ms value of Fe3O4 is 52 emug-1, which is slightly
less than the reported bulk value of this compound (≈70
emug-1) [30]. Ms is highly dependent on the particle size
and decreases with the reduction in particle size. In this
study, the low Ms value of Fe3O4 could be attributed to
the small particle size, large surface-to-volume ratio, spin
canting effect on the surface, and incomplete particle
crystallization [31,32], which is supported by FE-SEM
and XRD analysis (Figures 1b and 2a). On the other
hand, the coercivity was reduced after compositing
the magnetic phase with two non-magnetic phases
(MgO and zeolite), and subsequently, an increase was
observed in the coercive force. The coercivity depends on
various factors, including the substance microstructure,
magnetic anisotropy, grain size, and shape. Placing the
particles of non-magnetic phases among the magnetite
particles affects their surface anisotropy and the way they
interact at the interface among the magnetic particles.
As a result, it affects the mechanism of electron spinning
and the displacement of magnetic domain walls. These
parameters lead to a change in the specimen’s coercivity
after compositing the magnetic phase with non-magnetic
phases.
Considering that surface properties are essential
parameters in adsorption processes, the textural features
of MgO-Fe3O4-zeolite were scrutinized using the N2
adsorption-desorption isotherm. Figure 2d depicts the
volume of adsorbed-desorbed nitrogen by the composite
as a relative pressure function (p/p0). As shown, the
isotherm has a concave-convex shape with a sharp knee
point in the low p/p0 that corresponds to the IUPAC
type IV notation with the H3 hysteresis loop. This also
proves the mesoporous nature of the MgO-Fe3O4-zeolite
[33]. Isotherm hysteresis indicates that the geometric
shape of the pores is cylindrical, and there is a relatively
strong relationship between the specimen surface and the
adsorbent [34]. Based on the inset of Figure 2d and the
Brunauer-Emmett-Teller method, the specific surface
area and the pore volume of the composite were 68.90
m2g-1 and 0.59 cm3g-1, respectively. Additionally, the
pore size distribution of the MgO-Fe3O4-zeolite was
determined by the analysis of adsorption-desorption
isotherms using the Barrett-Joyner-Halenda method.
The mean size of composite pores regarding adsorption
and desorption branches was 12.20 nm and 14.03 nm,
respectively, indicating the mesoporous nature of the
specimen, which is significant in adsorption processes.
3.2. Preliminary screening of adsorbents
The initial step in the investigation of the adsorption
process is to find the equilibrium time and adsorbent
capacity. For this purpose, the individual adsorption
performance of synthesized powders and their binary
and ternary composites were assessed by ultrasoundassisted
adsorption. Based on the results, the adsorption
rate was high for all the adsorbents at initial times,
and the equilibrium was achieved after a short time. It
is worth noting that parallel adsorption experiments
were conducted under magnetic stirring, and the
results revealed that the equilibriums were obtained
over longer times (Figure not shown). The hybrid
adsorption + ultrasound process could accelerate the
mass transfer between the adsorbate and adsorbent,
thereby reducing the equilibrium time.
Following the comparison of the CODr extent at the
equilibrium time, MgO-Fe3O4-zeolite was selected as
a susceptible adsorbent in the adsorption of the dairy
effluent with an equilibrium time of 25 minutes and used
in further experiments under sonication.
3.3. Preliminary trials of dairy wastewater adsorption
onto MgO-Fe3O4-zeolite
To find parameters affecting the adsorption process and
the range of each parameter, the effects of four critical
factors were studied, namely, the initial COD, adsorbent
dosage, pH, and contact time on CODr efficiency.
To investigate the impacts of the effluent’s initial
concentration, the samples with a specific concentration
(COD0: 100, 300, 500, 700, and 900 mg L-1) were prepared
by diluting the stock solution. After adding the adsorbent
at pH = 7 (natural pH of a dairy effluent), the experiments
were followed for 25 minutes. According to Figure 3a, as
the initial COD increased, CODr decreased efficiency. To
study the effects of the adsorbent dosage, the experiments
were performed using five different amounts of MgOFe
3O4-zeolite (1, 2, 3, 4, and 5 gL-1). Based on data in
Figure 3b, the percentage of CODr increased with the
increase in the dosage of the adsorbent. The effect of pH
Mehrizad
180 Arch Hyg Sci. Volume 11, Number 3, 2022
on the adsorption process is influenced by the adsorbate
solution pH and the adsorbent pHpzc, and the pHpzc of
MgO-Fe3O4-zeolite was primarily measured accordingly.
Next, dairy effluent samples with a pH of 3, 5, 7, 9, and
11 were prepared, and the experiments were followed.
According to Figure 3c, the extent of CODr increased
with the increase in the solution’s pH. The effects of the
contact time on the adsorption process also implied that
the adsorption efficiency increased over time (Figure 3d).
3.4. Assessment of parameter interaction and
optimization of the process by RSM
As mentioned in Section 2, RSM-based CCD was used
to design the experiments, investigate the parameter
interaction, and achieve optimal conditions. Table 1
presents the domain of the parameters and the designed
experiments, along with experimental and predicted
results.
After importing the experimental data into the software,
a quadratic polynomial equation, Eq. (4), was suggested
and assessed by ANOVA (Table 2).
1 2 3 4 1 2
2
1 3 1 4 2 3 2 4 3 4 1
2 2 2
2 3 4
(%) 60.47 - 5.73 3.74 5.55 4.88 - 0.14 -
0.28 0.15 0.97 -1.54 0.96 - 0.75
0.94 - 0.015 2.02
= + + +
+ + +
+ +
CODr X X X X X X
X X X X X X X X X X X
X X X
(4)
According to the ANOVA results, F (32.69) and P
( < 0.0001) values indicated the high adequacy of the
proposed model. The lack of fit F-value (1.63) implies
that the lack of fit is not considerable, and the model
Figure 3. Effect of initial COD (a), adsorbent dosage (b), pH (c), and time (d) on CODr efficiency. Note. COD: Chemical oxygen demand
Arch Hyg Sci. Volume 11, Number 3, 2022 181
Synthesis of Zeolite-Supported Magnesia/Magnetite Nanoparticles, and its Efficient Application in Ultrasound
covered the data well. Additionally, the correlation
coefficient (R2) value of the model regression was
0.9682, whose appropriate adaption with the adjusted R2
(0.9387) demonstrated a proper correlation between the
experimental and predicted results.
Based on data in Table 2, the “Prob > F” values of X1X2,
X2X3, and X2X4 were less than 0.0500, and these terms can
have a significant effect on CODr.
The 3D diagrams (response surfaces) of the parameter
interaction (X1X2, X2X3, and X2X4) were depicted after
confirming the accuracy and precision of the selected
model.
As shown in Figure 4a, the CODr efficiency increased
with an increase in the adsorbent dosage and a decrease
in the initial COD.
According to Figures 4b and 4c, the extent of CODr
increased with an increase in the solution’s pH and
contact time.
Considering that this experimental design mainly
aimed to achieve optimal conditions, the ultrasoundassisted
adsorption process was optimized by the
software (DX 11), and it was found that under optimal
conditions (COD0: 300 mgL-1, adsorbent dosage: 4 gL-1,
pH: 9, and contact time: 20 minutes), the maximum CODr
efficiency was 83.22%. Performing the experiments under
optimal conditions showed that practical efficiency was
approximately 80%, with the proximity of experimental
and predicted yields re-emphasizing the accuracy and
precision of the RSM proposed model.
3.5. Finding the most important parameter affecting the
adsorption process by the ANN
The ANN modeling was utilized to select the
important operational parameter and reoptimization
of the adsorption process. To this end, the values of the
experiments designed by RSM, along with experimental
CODr (%) results (Table 1) were entered to the ANN
toolbox of Matlab R2020b software as input and
output data. Data were processed using the Levenberg-
Marquardt back-propagation algorithm [26].
Diverse networks with hidden layer neurons in the
range of 2-10 were tested to attain supreme topology
based on the least mean square error (MSE) and
regression (R) values. Overall, 20 iterations of each
topology were performed to resolve errors (Figure not
shown). It was found that the epoch 2 had the optimum
validation performance with the MSE value of 0.0011,
and the structure of the nine neurons of the hidden layer
was the best configuration for process modeling.
In addition, the correlation between outputs and
targets was evaluated using the regression (R) values of the
diagrams illustrated in Figure 5. For all plots, the R-values
close to one meant a close relationship between outputs
and target values. This is quite clear from the comparison
of actual and ANN-predicted results in Table 1.
After model approval, the weight matrix was
Table 2. ANOVA results of the quadratic polynomial model
Source
Sum of
squares
DF
Mean
square
F-value
P value
Prob > F
Model 2649.83 14 189.27 32.69 < 0.0001
X1 787.42 1 787.42 136.01 < 0.0001
X2 335.33 1 335.33 57.92 < 0.0001
X3 739.59 1 739.59 127.75 < 0.0001
X4 571.45 1 571.45 98.70 < 0.0001
X1 X2 0.31 1 0.31 0.054 0.0199
X1 X3 1.29 1 1.29 0.22 0.6432
X1 X4 0.38 1 0.38 0.066 0.8009
X2 X3 15.04 1 15.04 2.60 0.0279
X2 X4 37.91 1 37.91 6.55 0.0218
X3 X4 0.15 1 0.15 0.025 0.8758
X1² 15.24 1 15.24 2.63 0.1256
X2² 24.22 1 24.22 4.18 0.0588
X3²
6.431E-
003
1
6.431E-
003
1.111E-
003
0.9739
X4² 111.75 1 111.75 19.30 0.0005
Residual 86.84 15 5.79
Lack of fit 66.47 10 6.65 1.63 0.3071
R2 = 0.9683 Adjusted R2 = 0.9387
Note. ANOVA: Analysis of variance; DF: Degree of freedom; R2: Coefficient
of determination.
Figure 4. Response Surfaces of Interaction of Adsorption Operating Parameters on CODr Efficiency. Note. COD: Chemical oxygen demand. The values of the
other two parameters in each graph have been considered equal to the central point of their ranges
Mehrizad
182 Arch Hyg Sci. Volume 11, Number 3, 2022
acquired, and Garson’s equation [26] based on the
connection weights (Table 3) was employed to assess
the relative significance of the operating factors on
CODr (%); the results are illustrated as a radar chart in
Figure 5. Undoubtedly, the amount of the adsorbent is
the vital operational factor with a 38.30% effect on CODr
efficiency.
3.6. Adsorbent stability and reusability assessment
The reusability of the adsorbent should be considered
due to economic aspects. For this purpose, the used
adsorbent was collected with a magnet, and it was dried
and reused following the desorption process (washing
with 0.5 M HNO3 and DI water). The results of frequent
experiments demonstrated that the adsorbent could be
reused after multiple times with a slight reduction in
efficiency. These findings confirmed the durability and
reusability of the adsorbent and implicated the economic
aspect of its application.
3.7. Adsorption isotherms
Designing an optimal adsorption system requires a better
Figure 5. Regression plots for the evaluation of the ANN Model validity, along with relative importance of operational parameters on CODr (%). Note. ANN:
Artificial neural network; COD: Chemical oxygen demand.
Arch Hyg Sci. Volume 11, Number 3, 2022 183
Synthesis of Zeolite-Supported Magnesia/Magnetite Nanoparticles, and its Efficient Application in Ultrasound
understanding of the adsorbent-adsorbate interaction.
Adsorption isotherms are equilibrium data used to
describe the adsorbent-adsorbate interaction at a constant
temperature. Isotherms also indicate the adsorption
capacity of an adsorbent.
The isothermal studies of dairy effluent adsorption by
MgO-Fe3O4-zeolite were conducted by varying the initial
COD (100-900 mgL-1) at 25 °C and the optimum specified
conditions. The adsorption data were fitted with the
Langmuir, Freundlich, Temkin, Redlich-Peterson, and
Dubinin-Radushkevich isotherm models; the equations
of these models and the values of calculated parameters
are summarized in Table 4. The correlation coefficient
(R2) and chi-squared (χ2), Eqs. 5 and 6, respectively, were
employed to judge the fit of the model with experimental
data.
2
2 1 ,exp ,
2
1 ,exp ,exp
( )
1-
( )
=
=
−
=
−
Σ
Σ
n
e e cal in
i e e
q q
R
q q
(5)
2
2 ,exp ,
1
,
( )
=
−
= Σn e e cal
i
e cal
q q
q
χ
(6)
where qe,exp and qe,cal are the experimental and calculated
equilibrium capacity (mgg-1), respectively. e,exp q
represents the average of equilibrium adsorption capacity
obtained by experimental data.
The Langmuir model describes a homogeneous
and monolayer adsorption of the adsorbate with the
same energy on the active sites of the adsorbent. The
Freundlich model is based on the heterogeneous and
multilayer adsorption of the adsorbate over the adsorbent
surface. The Temkin model supposes that the adsorption
heat resulting from adsorbent-adsorbate chemical
interactions decreases linearly with coverage. The model
of Redlich-Peterson is applied in both homogeneous
and heterogeneous adsorption processes, including the
privilege of both Langmuir and Freundlich models. The
Dubinin-Radushkevich model is often used to predict
the physical or chemical nature of adsorption where the
magnitude of the mean adsorption energy (ε) < 8, ≈ 8-16,
or > 20 kJmol-1 indicates the physical adsorption, ion
exchange, or chemisorption of the process [35,36].
Polymath software (version 6.10) was employed to draw
nonlinear graphs and compute isotherm constants. The
results (Figure 6a and Table 4) implied that the Temkin
model fitted well with the experimental data (because of
the highest R2 and the lowest χ2).
Table 3. Weight matrix
Neuron
IWj,i
* LWj,i
**
X1 X2 X3 X4 CODr (%)
1 1.625 1.437 0.702 -0.445 0.231
2 -0.105 -1.449 1.533 -0.198 0.643
3 0.308 -1.468 0.412 1.808 0.340
4 -0.313 2.004 -1.019 0.228 -0.400
5 -1.415 1.763 0.162 0.035 1.274
6 0.418 0.363 -0.776 -2.246 -0.495
7 0.334 1.491 -1.198 -0.978 0.485
8 -1.760 1.566 -0.434 -0.368 -0.368
9 -2.337 0.783 -0.038 0.688 -0.684
Note. IW: Input weight matrix; LW: LAyer weight.
* Weight to the jth neuron of the hidden layer from the ith input parameter.
** Weight to the output layer from the jth neuron of the hidden layer.
Table 4. Isotherm and kinetics parameters along with their regression analysis
and error functions
Model Equation Parameter
Isotherm
Langmuir
1
m L e
e
L e
q K C
q
K C
=
+
qm = 333.33 mg g-1
KL = 3.28 × 10-3 L mg-1
R2 = 0.930
χ2 = 0.051
Freundlich
1
n
qe = KF Ce
n = 1.40
KF = 2.74 (mg g-1)(mg-1)1/n
R2 = 0.942
χ2 = 0.058
Temkin e ( T e )
q RT ln K C
Q
=
Δ
ΔQ = 54.48 J mol-1
KT = 0.07 L g-1
R2 = 0.991
χ2 = 0.017
Redlich-
Peterson
1
e
e
e
AC
q
BCβ =
+
A = 1.008 (L mg-1)-g
B = 1 L mg-1
g = 0.47
R2 = 0.836
χ2 = 0.281
Dubinin-
Radushkevich qe = qm exp(−βε 2 )
Β = 10-4 mol2 kJ-2
qm = 109.07 mg g-1
R2 = 0.724
χ2 = 0.279
Kinetics
Lagergren’s
pseudo 1st order
(1 k1t )
t e q = q − e−
k1 = 0.15 min-1
qe = 65.77 mg g-1
R2 = 0.641
χ2 = 0.064
Ho’s pseudo 2nd
order
2
2
2 1
e
t
e
k q t
q
k q t
=
+
k2 = 9.97 × 10-4 g mg-1 min-1
qe = 62.34 mg g-1
R2 = 0.952
χ2 = 0.008
Elovich ( ) 1 ln t q= αβ t β
α = 18.48 mg g-1 min-1
β = 0.05 g mg-1
R2 = 0.945
χ2 = 0.012
Intraparticle
diffusion
0.5
t id q = k t +C
kid = 12.89 mg g-1 min-0.5
C = 0.40 mg g-1
R2 = 0.971
Mehrizad
184 Arch Hyg Sci. Volume 11, Number 3, 2022
3.8. Adsorption kinetics
One of the most important factors in designing an
adsorption system is to predict the adsorption process
rate. The adsorption kinetics represents the trend of
changes in adsorption capacity over time. The estimation
of this trend provides valuable insight for determining
the type of the adsorption process mechanism in a given
system. Hence, four well-known kinetics models (i.e.,
the Lagergren’s pseudo 1st order, Ho’s pseudo 2nd order,
Elovich, and intraparticle diffusion models) were adopted
to study the adsorption kinetics over a period time of
0-25 minutes under optimal conditions. Table 4 presents
the equation of the kinetics models, along with the data
extracted from these models.
The 1st order Lagergren model is based on adsorbent
capacity and is applicable once the adsorption occurs
using a diffusion mechanism through a boundary layer;
meanwhile, the 2nd order Ho equation shows that chemical
adsorption is a dominant and controlling mechanism in
the adsorption process [37,38]. The Elovich model is an
appropriate model for describing chemisorption processes
and is mainly used in heterogeneous adsorption [39]. The
Weber-Morris intraparticle diffusion model is utilized to
elucidate the diffusion mechanism of adsorption systems.
In this model, the intercept (C) indicates the thickness of
the boundary layer, where zero implies that the adsorption
process is only controlled by intraparticle diffusion [40].
Similar to the previous stage, Polymath software
was employed to draw nonlinear graphs and compute
rate constants. According to the obtained results
(Figure 6b and Table 4), the adsorption kinetics could be
better fitted by the pseudo 2nd order and Elovich models.
The comparison of the qe calculated from the Ho’s model
(62.34 mgg-1) with experimental qe (60.48 mgg-1) revealed
a good correlation between them. Furthermore, the value
of k2qe (6.62 × 10-2 min-1) indicated that adsorption onto
the ternary composite of MgO-Fe3O4-zeolite was a swift
process. In addition, the non-zero value of the intercept in
the intraparticle diffusion model (C = 0.4 mgg-1) showed
that this model was not the only rate-controlling stage,
and the adsorption of dairy effluent by MgO-Fe3O4-zeolite
followed the pseudo 2nd order + Elovich + intraparticle
diffusion kinetics models.
3.9. Adsorption thermodynamics
To evaluate the thermodynamics of the process, the
adsorption of the dairy effluent onto MgO-Fe3O4-zeolite
was studied under optimal conditions at three different
temperatures. The results demonstrated that CODr
efficiency increased with an increase in the temperature.
Thermodynamics parameters (ΔGº, ΔHº, and ΔSº) were
computed using Eqs. (7) and (8):
Δ =− C Gο RTlnK ( = e
C
e
q
K
C ) (7)
( 1 ) Δ Δ
=− + C
lnK H S
R T R
ο ο
(8)
where KC, qe (mgg-1), and Ce (mgL-1) are the
equilibrium constant, equilibrium adsorption capacity,
and equilibrium COD extent at different temperatures,
respectively [41].
Taking into account Eq. (7) and substituting equilibrium
constants at 298, 308, and 318 K, the ΔGº values of -101.90,
-514.02, and -939.58 Jmol-1 were obtained, implying that
the process was spontaneous.
Considering Eq. (8) and plotting the lnKC vs. 1/T
(Figure not shown), the values of ΔH° (12.37 kJmol-1) and
ΔS° (41.89 JK-1mol-1) were achieved from the slope and
intercept, respectively.
4. Discussion
The FE-SEM and TEM images (Figure 1) confirmed the
composite formation through the distribution of MgO
and Fe3O4 spherical nanoparticles on 4A zeolite cubic
particles. The XRD patterns of MgO, Fe3O4, and 4A
zeolite (Figure 2a) appeared at their respective diffraction
angles, highlighting the success of their preparation.
For MgO-Fe3O4-zeolite, characteristic diffraction peaks
corresponding to MgO, Fe3O4, and 4A zeolite justified the
creation of the ternary composite without impurities. The
EDX-elemental mapping analysis (Figure 2b) confirmed
the presence of Mg, O, Fe, Si, Al, and Na elements in
the composite texture. Moreover, the VSM analysis
(Figure 2c) demonstrated the ferromagnetic behavior
of the ternary composite. The N2 adsorption-desorption
results (Figure 2d) corroborated the construction of the
composite with a mesoporous structure.
Figure 6. Comparison of adsorption isotherm models (a) and adsorption
kinetics models (b) with experimental data
Arch Hyg Sci. Volume 11, Number 3, 2022 185
Synthesis of Zeolite-Supported Magnesia/Magnetite Nanoparticles, and its Efficient Application in Ultrasound
The experiments to find the equilibrium time revealed
that the equilibrium was achieved after a short time. In the
adsorption process by MgO and Fe3O4, the most significant
changes occurred during the first 5 minutes, after which
the CODr efficiency remained almost constant. The
results also demonstrated that the adsorption onto MgOFe
3O4 and MgO-Fe3O4-zeolite was impressive in the first
25 minutes. The number of active sites in the adsorbent
was extremely high in the initial stages, but the adsorption
sites became saturated over time, and the adsorption rate
was controlled by transferring from external to internal
sites [42]. Based on the results, there was a substantial
difference in CODr efficiency by individual adsorbents
compared with their binary and ternary composites. A
low CODr extent by single adsorbents could be ascribed
to limited surface sites, the high hydrophilicity of MgO,
and the agglomeration possibility of Fe3O4 nanoparticles
in aqueous solutions [19,43].
The RSM-based CCD was applied to design the
adsorption experiments. The response surfaces of the
parameter interaction were illustrated after confirming
the accuracy and precision of the selected model
(Figure 4). Increasing the amount of the adsorbent
increased the number of adsorption sites, thereby
augmenting the number of adsorbed micropollutant
molecules. On the other hand, as the initial COD of the
dairy effluent increased, the competition for accessing the
active sites increased, and the entire adsorbent surface
was exposed to contaminant molecules. Moreover,
the electrostatic repulsion between the adsorbed and
soluble micropollutants led to a CODr reduction at
high concentrations [44,45]. Other researchers reported
similar results for the adsorption of organics in dairy
wastewater on modified nano-montmorillonite [6]. The
leading reason for increased efficiency by the increment
in the solution pH and adsorbent dosage was electrostatic
interactions between micropollutant molecules and
the surface of the MgO-Fe3O4-zeolite. In fact, the dairy
effluent pH is almost neutral or slightly alkaline, but it
may become relatively acidic given the presence of lactic
acid generated by lactose [1-3,46,47]. Accordingly, the
deprotonation of dairy effluent compounds is possible
in alkaline media. Therefore, a strong electrostatic
attraction could be expected between the positive surface
of the adsorbent (pHpzc = 11) and the deprotonated
organic species by the increase in the solution pH. The
increase in CODr over time represents the fact that the
probability of the adsorbent-adsorbate contact increases
over a longer duration, finally reaching its maximum
value at equilibrium.
Based on the ANN modeling results, the relative
significance of the operating factors was as adsorbent
dosage > initial COD > pH > sonication time.
The adsorption of dairy wastewater by the composite
was consistent with the Temkin isotherm model. It could
be speculated that the adsorption of dairy wastewater
by MgO-Fe3O4-zeolite was mainly due to chemical
interactions. The positive value of ΔQ indicates that the
adsorption of the dairy effluent by MgO-Fe3O4-zeolite is
an endothermic process supported by thermodynamics
results (ΔH° > 0).
The kinetics data were well described by both Ho’s
pseudo 2nd order and Elovich models. Hence, the
adsorption of the dairy effluent onto MgO-Fe3O4-zeolite
was most probably due to chemisorption, which is
supported by isothermal results.
Increasing CODr efficiency with the increase in the
temperature could be attributed to the increase in the
kinetic energy of micropollutant molecules with the rise
in temperature, as well as the increase in their collisions
with the adsorbent surface [48].
Given the positive values of enthalpy and entropy, the
process was endothermic with an increase in disorder at
the adsorbent-adsorbate interface.
To further evaluate the adsorption capacity of the
adsorbent prepared in this study, the adsorption
capacities of MgO-Fe3O4-zeolite and other adsorbents
prepared in the literature to dairy wastewater are listed
in Table 5 [49-54]. Based on the data, MgO-Fe3O4-zeolite
exhibits excellent adsorption performance and has great
competitiveness in this research field, indicating that it is
a potential adsorbent for dairy wastewater treatment. In
addition, MgO-Fe3O4-zeolite has magnetic behavior and
thus has the advantage of being recyclable.
4.1. Possible adsorption mechanism
Generally, various factors are involved in the accosting of
the adsorbate to the adsorbent surface, including physical,
chemical, and electrostatic interactions. According to the
aforementioned results, the coulomb’s attractive forces,
along with chemical adsorption were the main reasons
for dairy effluent adsorption by MgO-Fe3O4-zeolite.
Given that the obtained pHpzc for the composite was
11, the composite surface had a negative and positive
electric charge at pH > 11 and pH < 11, respectively.
Most compounds in dairy effluents are organics. It
is more likely that they are deprotonated in alkaline
Table 5. Comparison of different adsorbents for the adsorption of dairy
wastewater
Adsorbent Contact time CODr (%) Reference
Copper oxide nanoparticles 120 min 77 [49]
Carbon nanotubes 15 h 50 [50]
Modified dried activated sludge 120 min 80 [51]
Soil 9 h 46 [52]
Graphene oxide 30 min 80 [53]
Chitosan 240 min 90 [54]
MgO-Fe3O4-zeolite 20 min 83 This study
Note. COD: Chemical oxygen demand; Magnesium oxide iron oxide; Fe3O4:
Iron oxide.
Mehrizad
186 Arch Hyg Sci. Volume 11, Number 3, 2022
media [46,47]. Accordingly, it could be expected that by
increasing the pH of the solution, a strong electrostatic
attraction occurs between the adsorbent positive surface
and deprotonated organic species. On the other hand,
the nature of the chemisorption of the process was also
partially revealed by the results obtained from isothermal
and kinetics studies. In this regard, the role of hydrogen
bonding and Van der Waals interactions must be taken
into consideration [55,56].
5. Conclusion
In the current research, the MgO-Fe3O4-zeolite
nanocomposite was prepared as a novel adsorbent for
adsorbing the organic compounds of the dairy effluent
by ultrasound-assisted adsorption. It was concluded
that the synergistic effects of MgO, Fe3O4, and zeolite
could facilitate the adsorption of organics in the effluent
and increase the efficiency of the adsorption process.
The CCD based on RSM was exploited in the planning
stage of the experimental approach, and it was found
that the maximum CODr efficiency was 83.22% under
optimal conditions. The evaluation of the adsorption
isotherm and kinetics represented that the process was
well fitted by the Temkin isotherm model and Ho’s
2nd order kinetics model, indicating that the process
belonged to chemisorption. The thermodynamics studies
showed that the adsorption of the dairy effluent by the
magnetic nanocomposite was endothermic (ΔH° > 0) and
spontaneous (ΔG° < 0).
The significant results of this work emphasize the
applicability of the prepared magnetically recyclable
adsorbent toward sequestering different micropollutants
in practical engineering.
Acknowledgments
The author would like to thank Tabriz Branch, Islamic Azad
University for providing facilities and technical support.
Conflict of Interests
The author declares that he has no known competing financial
interests or personal relationships that could have appeared to
influence the work reported in this paper.
References
1. Amini M, Younesi H, Zinatizadeh Lorestani AA, Najafpour
G. Determination of optimum conditions for dairy
wastewater treatment in UAASB reactor for removal
of nutrients. Bioresour Technol. 2013;145:71-9. doi:
10.1016/j.biortech.2013.01.111.
2. Daneshvar E, Zarrinmehr MJ, Koutra E, Kornaros M, Farhadian
O, Bhatnagar A. Sequential cultivation of microalgae in raw
and recycled dairy wastewater: microalgal growth, wastewater
treatment and biochemical composition. Bioresour Technol.
2019;273:556-64. doi: 10.1016/j.biortech.2018.11.059.
3. Muniz GL, Silva T, Borges AC. Assessment and optimization
of the use of a novel natural coagulant (Guazuma
ulmifolia) for dairy wastewater treatment. Sci Total Environ.
2020;744:140864. doi: 10.1016/j.scitotenv.2020.140864.
4. Li Z, Zhang Y, Cao Y, Zhou S, Han Y, Yue L. Characterization of
Sn-MOF and its adsorption application for Acid Red 3R. Mater
Lett. 2021;282:128647. doi: 10.1016/j.matlet.2020.128647.
5. Liu WB, Cui GN, Wang H, Zhang DM, Wu RX, Li L, et
al. Efficient and selective adsorption of dye in aqueous
environment employing a functional Zn(Ⅱ)-based metal
organic framework. J Solid State Chem. 2020;292:121740.
doi: 10.1016/j.jssc.2020.121740.
6. Zhang W, Zhang L, Jiang X, Liu X, Li Y, Zhang Y. Enhanced
adsorption removal of aflatoxin B1, zearalenone and
deoxynivalenol from dairy cow rumen fluid by modified
nano-montmorillonite and evaluation of its mechanism.
Anim Feed Sci Technol. 2020;259:114366. doi:
10.1016/j.anifeedsci.2019.114366.
7. Hong M, Yu L, Wang Y, Zhang J, Chen Z, Dong L, et al.
Heavy metal adsorption with zeolites: the role of hierarchical
pore architecture. Chem Eng J. 2019;359:363-72. doi:
10.1016/j.cej.2018.11.087.
8. Piri F, Mollahosseini A, khadir A, Milani Hosseini
M. Enhanced adsorption of dyes on microwaveassisted
synthesized magnetic zeolite-hydroxyapatite
nanocomposite. J Environ Chem Eng. 2019;7(5):103338. doi:
10.1016/j.jece.2019.103338.
9. Jiang N, Shang R, Heijman SGJ, Rietveld LC. Adsorption of
triclosan, trichlorophenol and phenol by high-silica zeolites:
adsorption efficiencies and mechanisms. Sep Purif Technol.
2020;235:116152. doi: 10.1016/j.seppur.2019.116152.
10. Mohamed RM, Shawky A, Mkhalid IA. Facile synthesis of
MgO and Ni-MgO nanostructures with enhanced adsorption
of methyl blue dye. J Phys Chem Solids. 2017;101:50-7. doi:
10.1016/j.jpcs.2016.10.009.
11. Cui W, Li P, Wang Z, Zheng S, Zhang Y. Adsorption study of
selenium ions from aqueous solutions using MgO nanosheets
synthesized by ultrasonic method. J Hazard Mater.
2018;341:268-76. doi: 10.1016/j.jhazmat.2017.07.073.
12. Zheng X, Zhou Y, Liu X, Fu X, Peng H, Lv S. Enhanced
adsorption capacity of MgO/N-doped active carbon
derived from sugarcane bagasse. Bioresour Technol.
2020;297:122413. doi: 10.1016/j.biortech.2019.122413.
13. Bekhti H, Bouchafaa H, Melouki R, Travert A, Boucheffa
Y. Adsorption of CO2 over MgO−Impregnated NaYzeolites
and modeling study. Microporous Mesoporous Mater.
2020;294:109866. doi: 10.1016/j.micromeso.2019.109866.
14. Guo J. Adsorption characteristics and mechanisms of high-levels
of ammonium from swine wastewater using natural and MgO
modified zeolites. Desalin Water Treat. 2016;57(12):5452-
63. doi: 10.1080/19443994.2014.1003334.
15. Gao F, Wang S, Chen G, Duan J, Dong J, Wang W. A
facile approach to the fabrication of MgO@Y composite
for CO2 capture. Adsorption. 2020;26(5):701-9. doi:
10.1007/s10450-019-00147-w.
16. Sivashankar R, Sathya AB, Vasantharaj K, Sivasubramanian
V. Magnetic composite an environmental super adsorbent
for dye sequestration–a review. Environ Nanotechnol Monit
Manag. 2014;1-2:36-49. doi: 10.1016/j.enmm.2014.06.001.
17. Kefeni KK, Mamba BB, Msagati TAM. Application of spinel
ferrite nanoparticles in water and wastewater treatment:
a review. Sep Purif Technol. 2017;188:399-422. doi:
10.1016/j.seppur.2017.07.015.
18. Salem ANM, Ahmed MA, El-Shahat MF. Selective adsorption
of amaranth dye on Fe3O4/MgO nanoparticles. J Mol Liq.
2016;219:780-8. doi: 10.1016/j.molliq.2016.03.084.
19. Liu X, Tian J, Li Y, Sun N, Mi S, Xie Y, et al. Enhanced
dyes adsorption from wastewater via Fe3O4 nanoparticles
functionalized activated carbon. J Hazard Mater.
2019;373:397-407. doi: 10.1016/j.jhazmat.2019.03.103.
20. Zhao L, Lv W, Hou J, Li Y, Duan J, Ai S. Synthesis of magnetically
Arch Hyg Sci. Volume 11, Number 3, 2022 187
Synthesis of Zeolite-Supported Magnesia/Magnetite Nanoparticles, and its Efficient Application in Ultrasound
recyclable g-C3N4/Fe3O4/ZIF-8 nanocomposites for
excellent adsorption of malachite green. Microchem J.
2020;152:104425. doi: 10.1016/j.microc.2019.104425.
21. Baziar M, Azari A, Karimaei M, Gupta VK, Agarwal S,
Sharafi K, et al. MWCNT-Fe3O4 as a superior adsorbent
for microcystins LR removal: investigation on the magnetic
adsorption separation, artificial neural network modeling,
and genetic algorithm optimization. J Mol Liq. 2017;241:102-
13. doi: 10.1016/j.molliq.2017.06.014.
22. Dashamiri S, Ghaedi M, Dashtian K, Rahimi MR, Goudarzi
A, Jannesar R. Ultrasonic enhancement of the simultaneous
removal of quaternary toxic organic dyes by CuO
nanoparticles loaded on activated carbon: central composite
design, kinetic and isotherm study. Ultrason Sonochem.
2016;31:546-57. doi: 10.1016/j.ultsonch.2016.02.008.
23. Rajput S, Pittman CU Jr, Mohan D. Magnetic magnetite (Fe3O4)
nanoparticle synthesis and applications for lead (Pb2 + ) and
chromium (Cr6 + ) removal from water. J Colloid Interface Sci.
2016;468:334-46. doi: 10.1016/j.jcis.2015.12.008.
24. Mahmoudi E, Behnajady MA. Synthesis of Fe3O4@NiO
core-shell nanocomposite by the precipitation method
and investigation of Cr(VI) adsorption efficiency. Colloids
Surf A Physicochem Eng Asp. 2018;538:287-96. doi:
10.1016/j.colsurfa.2017.11.020.
25. Ataei A, Mehrizad A, Zare K. Photocatalytic degradation of
cefazoline antibiotic using zeolite-supported CdS/CaFe2O4
Z-scheme photocatalyst: optimization and modeling of
process by RSM and ANN. J Mol Liq. 2021;328:115476. doi:
10.1016/j.molliq.2021.115476.
26. Attarikhasraghi N, Zare K, Mehrizad A, Modirshahla N,
Behnajady MA. Zeolite 4A supported CdS/g-C3N4 type-
II heterojunction: a novel visible-light-active ternary
nanocomposite for potential photocatalytic degradation
of cefoperazone. J Mol Liq. 2021;342:117479. doi:
10.1016/j.molliq.2021.117479.
27. Wang B, Xiong X, Ren H, Huang Z. Preparation of MgO
nanocrystals and catalytic mechanism on phenol ozonation.
RSC Adv. 2017;7(69):43464-73. doi: 10.1039/c7ra07553g.
28. Mahdavi M, Bin Ahmad M, Haron MJ, Gharayebi Y,
Shameli K, Nadi B. Fabrication and characterization of
SiO2/(3-aminopropyl) triethoxysilane-coated magnetite
nanoparticles for lead (II) removal from aqueous solution. J
Inorg Organomet Polym Mater. 2013;23(3):599-607. doi:
10.1007/s10904-013-9820-2.
29. Koh PY, Yan J, Teja A. Precipitation and growth of magnesium
hydroxide nanopetals on zeolite 4A surfaces. J Cryst Growth.
2011;331(1):56-63. doi: 10.1016/j.jcrysgro.2011.07.006.
30. Roca AG, Marco JF, del Puerto Morales M, Serna CJ.
Effect of nature and particle size on properties of uniform
magnetite and maghemite nanoparticles. J Phys Chem C.
2007;111(50):18577-84. doi: 10.1021/jp075133m.
31. Wang T, Zhang L, Wang H, Yang W, Fu Y, Zhou W, et al.
Controllable synthesis of hierarchical porous Fe3O4 particles
mediated by poly(diallyldimethylammonium chloride)
and their application in arsenic removal. ACS Appl Mater
Interfaces. 2013;5(23):12449-59. doi: 10.1021/am403533v.
32. Petcharoen K, Sirivat A. Synthesis and characterization of
magnetite nanoparticles via the chemical co-precipitation
method. Mater Sci Eng B. 2012;177(5):421-7. doi:
10.1016/j.mseb.2012.01.003.
33. Rouquerol J, Rouquerol F, Llewellyn P, Maurin G, Sing
KS. Adsorption by Powders and Porous Solids: Principles,
Methodology and Applications. Academic Press; 2013. doi:
10.1016/b978-0-12-598920-6.x5000-3.
34. Thommes M, Kaneko K, Neimark AV, Olivier JP, Rodriguez-
Reinoso F, Rouquerol J, et al. Physisorption of gases, with
special reference to the evaluation of surface area and pore
size distribution (IUPAC Technical Report). Pure Appl Chem.
2015;87(9-10):1051-69. doi: 10.1515/pac-2014-1117.
35. Gimbert F, Morin-Crini N, Renault F, Badot PM, Crini G.
Adsorption isotherm models for dye removal by cationized
starch-based material in a single component system:
error analysis. J Hazard Mater. 2008;157(1):34-46. doi:
10.1016/j.jhazmat.2007.12.072.
36. Azari A, Mahmoudian MH, Hazrati Niari M, Eş I, Dehganifard
E, Kiani A, et al. Rapid and efficient ultrasonic assisted
adsorption of diethyl phthalate onto FeIIFe2IIIO4@GO:
ANN-GA and RSM-DF modeling, isotherm, kinetic and
mechanism study. Microchem J. 2019;150:104144. doi:
10.1016/j.microc.2019.104144.
37. Yin L, Lin Y, Jia L. Graphene oxide functionalized magnetic
nanoparticles as adsorbents for removal of phthalate
esters. Microchim Acta. 2014;181(9):957-65. doi:
10.1007/s00604-014-1187-8.
38. Ho YS, McKay G. Pseudo-second order model for sorption
processes. Process Biochem. 1999;34(5):451-65. doi:
10.1016/s0032-9592(98)00112-5.
39. Xiao G, Wang Y, Xu S, Li P, Yang C, Jin Y, et al. Superior
adsorption performance of graphitic carbon nitride
nanosheets for both cationic and anionic heavy metals from
wastewater. Chin J Chem Eng. 2019;27(2):305-13. doi:
10.1016/j.cjche.2018.09.028.
40. Hasanzadeh V, Rahmanian O, Heidari M. Cefixime
adsorption onto activated carbon prepared by dry
thermochemical activation of date fruit residues. Microchem
J. 2020;152:104261. doi: 10.1016/j.microc.2019.104261.
41. Mehrizad A. Adsorption studies of some phenol derivatives
onto Ag-cuttlebone nanobiocomposite: modeling of process
by response surface methodology. Res Chem Intermed.
2017;43(7):4295-310. doi: 10.1007/s11164-017-2874-y.
42. Mehrizad A, Gharbani P. Decontamination of 4-chloro-2-
nitrophenol from aqueous solution by graphene adsorption:
equilibrium, kinetic, and thermodynamic studies. Pol J Environ
Stud. 2014;23(6):2111-6. doi: 10.15244/pjoes/26779.
43. Priyadarshini B, Patra T, Sahoo TR. An efficient and
comparative adsorption of Congo red and Trypan blue
dyes on MgO nanoparticles: kinetics, thermodynamics and
isotherm studies. J Magnes Alloy. 2021;9(2):478-88. doi:
10.1016/j.jma.2020.09.004.
44. Chowdhury S, Mishra R, Saha P, Kushwaha P. Adsorption
thermodynamics, kinetics and isosteric heat of
adsorption of malachite green onto chemically modified
rice husk. Desalination. 2011;265(1):159-68. doi:
10.1016/j.desal.2010.07.047.
45. Behnajady MA, Yavari S, Modirshahla N. Investigation on
adsorption capacity of TiO2-P25 nanoparticles in the removal
of a mono-azo dye from aqueous solution: a comprehensive
isotherm analysis. Chem Ind Chem Eng Q. 2014;20(1):97-
107. doi: 10.2298/ciceq120610105b.
46. Dela Justina M, Rodrigues Bagnolin Muniz B, Mattge Bröring
M, Costa VJ, Skoronski E. Using vegetable tannin and
polyaluminium chloride as coagulants for dairy wastewater
treatment: a comparative study. J Water Process Eng.
2018;25:173-81. doi: 10.1016/j.jwpe.2018.08.001.
47. Murcia JJ, Hernández-Laverde M, Rojas H, Muñoz E, Navío
JA, Hidalgo MC. Study of the effectiveness of the flocculationphotocatalysis
in the treatment of wastewater coming
from dairy industries. J Photochem Photobiol A Chem.
2018;358:256-64. doi: 10.1016/j.jphotochem.2018.03.034.
48. Mehrizad A, Zare K, Aghaie H, Dastmalchi S. Removal of
4-chloro-2-nitrophenol occurring in drug and pesticide waste
by adsorption onto nano-titanium dioxide. Int J Environ Sci
Mehrizad
188 Arch Hyg Sci. Volume 11, Number 3, 2022
Technol. 2012;9(2):355-60. doi: 10.1007/s13762-012-0038-6.
49. Al-Ananzeh NM. Treatment of wastewater from a dairy plant
by adsorption using synthesized copper oxide nanoparticles:
kinetics and isotherms modeling optimization. Water Sci
Technol. 2021;83(7):1591-604. doi: 10.2166/wst.2021.089.
50. Moradi O, Maleki MS. Removal of COD from dairy
wastewater by MWCNTs: adsorption isotherm modeling.
Fuller Nanotub Carbon Nanostructures. 2013;21(10):836-48.
doi: 10.1080/1536383x.2011.613547.
51. Bazrafshan E, Kord Mostafapour F, Alizadeh M, Farzadkia
M. Dairy wastewater treatment by chemical coagulation
and adsorption on modified dried activated sludge: a pilotplant
study. Desalin Water Treat. 2016;57(18):8183-93. doi:
10.1080/19443994.2015.1018331.
52. Al-Jabari M, Iqefan N, Zahdeh N, Dweik H. Adsorption of
organic pollutants from dairy wastewater on soil: pollution
problem and control. Int J Glob Environ Issues. 2017;16(1-
3):149-61. doi: 10.1504/ijgenvi.2017.083425.
53. Falahati F, Baghdadi M, Aminzadeh B. Treatment of
dairy wastewater by graphene oxide nanoadsorbent
and sludge separation, using In Situ Sludge Magnetic
Impregnation (ISSMI). Pollution. 2018;4(1):29-41. doi:
10.22059/poll.2017.233196.276.
54. Eulmi A, Hazourli S, Abrane R, Bendaia M, Aitbara A, Touahria
S, et al. Evaluation of electrocoagulation and activated
carbon adsorption techniques used separately or coupled to
treat wastewater from industrial dairy. Int J Chem React Eng.
2019;17(12):20180229. doi: 10.1515/ijcre-2018-0229.
55. Catherine HN, Ou MH, Manu B, Shih YH. Adsorption
mechanism of emerging and conventional phenolic compounds
on graphene oxide nanoflakes in water. Sci Total Environ.
2018;635:629-38. doi: 10.1016/j.scitotenv.2018.03.389.
56. Suo F, You X, Ma Y, Li Y. Rapid removal of triazine
pesticides by P doped biochar and the adsorption
mechanism. Chemosphere. 2019;235:918-25. doi:
10.1016/j.chemosphere.2019.06.158