Content uploaded by Parisa Javidan
Author content
All content in this area was uploaded by Parisa Javidan on Apr 19, 2023
Content may be subject to copyright.
Content uploaded by Parisa Javidan
Author content
All content in this area was uploaded by Parisa Javidan on Nov 15, 2022
Content may be subject to copyright.
* Corresponding author.
1944-3994/1944-3986 © 2022 Desalination Publications. All rights reserved.
Desalination and Water Treatment
www.deswater.com
doi: 10.5004/dwt.2022.28563
264 (2022) 259–269
July
A tailored metal–organic framework applicable at natural pH
for the removal of 17α-ethinylestradiol from surface water
Parisa Javidan, Majid Baghdadi*, Ali Torabian, Behnoush Aminzadeh Goharrizi
School of Environment, College of Engineering, University of Tehran, 1417853111, Tehran, Iran,
Tel. +98 21 61113171, Fax: +98 21 66407719; emails: m.baghdadi@ut.ac.ir (M. Baghdadi) ORCID 0000-0001-9816-764X,
p.javidan71@ut.ac.ir (P. Javidan), atorabi@ut.ac.ir (A. Torabian), bamin@ut.ac.ir (B.A. Goharrizi)
Received 6 November 2021; Accepted 9 May 2022
abstract
The use of metal–organic frameworks as efficient adsorbents for micropollutants removal has
attracted significant interest recently, while the effect of linkers on their performance was not
examined extensively. The present paper was devoted to examining the removal of 17α-ethinyl-
estradiol (EE2), a synthetic hormone with high-potency estrogenic contamination, from surface
water by adsorption process using MIL-53(Fe) and NH2-MIL-53(Fe) as metal–organic frameworks
adsorbents constructed with different linkers. Batch experiments were designed using the Design-
Expert software by considering the effective parameters, including the initial concentration of EE2,
adsorbent dosage, and pH. The effects of total dissolved solids and temperature were investigated
as well. The adsorbent synthesis was done by the solvothermal method, and field-emission scan-
ning electron microscopy, X-ray diffraction, Fourier-transform infrared spectroscopy, Brunauer–
Emmett–Teller, and energy-dispersive X-ray spectroscopy analyses were performed to evaluate
the characterization of the prepared adsorbent. The results of characterizations showed the crys-
talline structure of NH2-MIL-53(Fe) with a high specific surface area of 743 m2/g. According to
the preliminary tests, NH2-MIL-53(Fe) was more efficient than MIL-53(Fe), especially in natural
waters with neutral pH. In the case of using 2-aminoterephthalic acid as a linker, in addition to Van
der Waals attraction, electrostatic interaction is involved in the adsorption of EE2 and improves
removal efficiency. The easy regeneration and reusability of NH2-MIL-53(Fe) for several cycles
make it suitable for EE2 removal in natural water resources.
Keywords: Endocrine-disrupting chemicals; 17α-ethinylestradiol; Metal–organic frameworks; Adsorp-
tion; Surface water
1. Introduction
Water quality is directly related to human health and
living organisms. Over the years, various pollutants,
including organic and various chemical contaminants in
water, have been identified, and their effects on human
health have been studied. In recent years, micropollutants
have attracted the attention of researchers because they
are far more harmful than some other pollutants [1,2].
Pharmaceuticals and personal care products (PPCPs), and
endocrine disruptors are groups of organic contaminants
that have been identified in waters around the world [3].
Meanwhile, endocrine disruptors have received more
attention from law enforcement agencies due to the possi-
ble high exposure to these substances and their dangerous
effects on human and wildlife health, although they exist at
low concentrations [1,4–6].
Among these compounds, 17α-ethinylestradiol (EE2),
a synthetic estrogen, is widely used in birth control pills
and drugs aimed at treating hormonal diseases. According
to several studies, EE2 is more resistant to biological deg-
radation than other similar natural hormones, thereby
P. Javidan et al. / Desalination and Water Treatment 264 (2022) 259–269260
having a much longer half-life than natural hormones [7].
This means that it does not tend to decompose in nature,
so living organisms are much more likely to be suffered
from exposure to this particular hormone. As a result,
EE2 is classified as a high-risk pollutant for humans and
fish [8–10]. Adverse effects of this endocrine disruptor
on humans include decreased sperm count, infertility,
increased breast disease, testicular, prostate, and cervical
cancer [11–13].
In general, various biological, physical, and chemi-
cal processes, such as the adsorption process, are used
to remove EE2 [14–17]. The research showed the ineffi-
ciency of the biological and some physical processes in
removing EE2; for example, Damkjaer et al. [8] examined
EE2 amounts in the influent and effluent of two stabiliza-
tion ponds of wastewater treatment systems in Morogoro,
Tanzania and indicated these amounts could pose a high
risk of interfering with the normal function of the endo-
crine systems of fish and wildlife, thereby causing an
adverse impact on their reproduction and population.
Mohagheghian et al. [18] studied EE2 and some other
hormones in raw sewage influent and final treated efflu-
ent of seven wastewater treatment plants in Tehran, Iran
in two seasons (summer–autumn), and the result showed
that the mean removal efficiency of EE2 was about 80%.
Also, Silva et al. [15] investigated membrane separation
to eliminate mentioned contaminant, establishing the best
result of 57% for EE2 removal. Furthermore, the formation
of byproducts, which are sometimes more dangerous than
the primary contaminant, has been reported in chemical
processes [19–21]. Moriyama et al. [22] investigated the
chlorination by-products of EE2, showing some of them
are more dangerous than the primary contaminant.
The adsorption process has attracted the attention of
researchers because of the possibility of regenerating and
tailoring the adsorbents with highly selective interactions
[23]. Kent and Tay [14] examined the adsorption and deg-
radation process in an aerobic granular sludge sequencing
batch reactor, concluding that the main removal mech-
anism for EE2 was adsorption with an average removal
efficiency of 77%.
Some of the common adsorbents, such as activated car-
bon and anthracite were used for adsorptive removal of EE2
in previous research and their adsorption capacities were
reported to be about 0.4 and 0.3 mg/g, respectively [24–26].
In this regard, the use of nanoparticles in the removal of
contaminants in water and wastewater has gained a lot of
attention due to the high adsorption capacity. Metal–organic
frameworks (MOFs) are nanostructures with unique fea-
tures such as very high surface area and porosity, uni-
form and ultra-regular pore size, a high degree of design
capability, easy synthesis conditions, and maintaining the
integrity of the flexible structure after adsorption or with-
drawal of guest molecules. MOFs are constructed of metal
clusters and organic linkers, which are connected via coor-
dination bonds and form a uniform and porous crystalline
structures [27–30]. Abazari et al. [31] and Azhar et al. [32]
investigated the efficiency of metal–organic frameworks
for antibiotics (amoxicillin, ampicillin, cloxacillin, and sul-
fachlorpyridazine) removal, resulting in approximately
90% removal efficiency.
In this research, two MOFs were synthesized using
iron(III) cations as the core coordinated adsorbent with
terephthalic acid and 2-aminoterephthalic acid as ligands.
It is expected that the type of the linker in the MOF struc-
ture could be effective in the interactions involved in the
adsorption process, especially at neutral pH. Using the
adsorption process as a high-efficient and innovative
approach for removing EE2 with leaving no trace of dan-
gerous by-products was investigated in the presence of
mentioned MOFs as adsorbents. Because the linkers sig-
nificantly impact the efficiency of adsorption by MOF, dif-
ferent linkers were explored to select the suitable one for
removing EE2 at natural pH. The applied adsorbents in
this study, especially NH2-MIL-53(Fe), are known as selec-
tive adsorbents with porous structure and high stability,
making them effective in EE2 elimination from aquatic
environments. The prepared adsorbent was identified by
characterization tests, including scanning electron micros-
copy, Fourier-transform infrared spectroscopy (FTIR), X-ray
diffraction (XRD), Brunauer–Emmett–Teller (BET), and
energy-dispersive X-ray spectroscopy (EDS) analysis. The
central composite design (CCD) and the response surface
methodology (RSM) were applied simultaneously for the
determination of effective parameters in batch adsorption
experiments, such as pH, adsorbent dosage, and initial
concentration of EE2. More examinations of the adsorption
process were accomplished by investigation of isotherm
and thermodynamic parameters. Finally, the regeneration
capability of the adsorbent was checked in several cycles.
2. Materials and methods
2.1. Materials
EE2 (C20H24O2) was purchased as a white, odorless solid
crystalline powder from the Iran Hormone Pharmaceutical
Company and was used as the contaminant for spiking in
samples. The terephthalic acid (C8H6O4), 2-aminotereph-
thalic acid (HO2C−C6H3NH2−CO2H), FeCl3·6H2O (99.5%),
N,N-dimethylformamide (DMF, 99.8%), and methanol (99%)
were used to synthesize the adsorbent. The pH adjustment
was performed using HCl and NaOH solutions (0.1 mol/L).
Potassium dihydrogen phosphate (KH2PO4), dichlorometh-
ane (CH2Cl2), and acetonitrile (C2H3N) were supplied for
the sample preparation step. All these mentioned materials
and MgSO4, NaHCO3, KNO3, and CaCl2 for the preparation
of total dissolved solids (TDS) solutions were purchased
from Merck (Darmstadt, Germany) except 2-aminotere-
phthalic acid, which was purchased from Sigma-Aldrich
(Taufkirchen, Germany), and all of them were used without
further purification.
2.2. Synthesis of the adsorbents
In this research, the solvothermal method was used to
synthesize MOFs [33,34]. In brief, NH2-MIL-53(Fe) was syn-
thesized using a solution containing FeCl3·6H2O, 2-amino-
terephthalic acid (NH2-BDC), and N,N-dimethylformamide
with a molar ratio of 1:1:280 (FeCl3·6H2O:NH2-BDC:DMF)
was prepared. For this purpose, 1 g (3.7 mmol) of
FeCl3·6H2O and 0.67 g (3.7 mmol) of 2-aminoterephthalic
261P. Javidan et al. / Desalination and Water Treatment 264 (2022) 259–269
acid were dissolved in 80 mL (1,033 mmol) of N,N-
dimethylformamide. In the next step, the prepared solution
was placed in an autoclave for 17 h at 145°C. After cooling
the solution at ambient temperature, the obtained MOF was
washed with N,N-dimethylformamide and methanol and
centrifuged at 5,000 rpm several times to remove residuals.
Finally, it was dried for 12 h at 70°C [35–37]. It should be
noted that for MIL-53(Fe) preparation, the type of linker
was only changed, and the materials were combined with
the same molar ratio.
2.3. Characterization of the adsorbent
A field-emission scanning electron microscope (FE-SEM,
MIRA3, TESCAN, Czech Republic) was used to deter-
mine the morphological characteristics of the adsorbent.
Fourier-transform infrared spectrum (FTIR Spectrum RX I,
PerkinElmer, USA) was applied for the identification of the
surface functional groups of the adsorbent in the range of
400–4,000 cm–1. Brunauer–Emmett–Teller (BET, BELSORP
mini II, Japan) was employed for evaluating the pore diam-
eters, pore volumes, and specific surface area of the synthe-
sized adsorbent. X-ray diffraction patterns were recorded
at 1° < 2θ < 80° to investigate the crystal structure of MOFs
using (XRD, X’Pert PRO MPD, PANalytical, Netherland).
Energy-dispersive X-ray spectroscopy (EDS, Oxford
Detector, England) was utilized to investigate the elemental
content of the adsorbent.
2.4. Experimental design and data analysis
Batch experiments were used for EE2 adsorption in
this research. Effective parameters and their ranges were
determined based on the pre-tests. The experiments were
designed by Design-Expert 10 (Stat-Ease Inc., Minneapolis,
United States) for the optimization of the adsorption process
based on the Box–Behnken Design and RSM. The removal
efficiency as the response depends on pH, adsorbent dos-
age (mg/L), and initial concentration of EE2 (µg/L), which
are labeled as A, B, and C, respectively (Table 1).
The total number of the designed experiments was 17
runs and the assessment of experimental errors was con-
trolled via 5 center points, as shown in Table 2. The ade-
quacy of the quadratic model and correlation between
variables and the significance of each one was evaluated
by the analysis of variance (ANOVA). The role of the three
main factors (pH, adsorbent dosage, and initial concen-
tration) on the removal of EE2 and their interactions were
presented by the 3D response surface plots. Due to the
presence of inorganic ions in real water samples, the adsor-
bent performance was studied in the presence of other ions
by designing and performing a series of experiments. For
this purpose, a solution containing inorganic salts with
a final TDS of 2,000 mg/L was prepared as the TDS stock
solution, and it was used to prepare the samples with dif-
ferent TDS levels. The saturated adsorbent was placed in
15 mL of methanol and shaken for 5 min to determine the
number of regeneration cycles for the adsorbent. Because of
the high solubility of EE2 in methanol, the adsorbed EE2
can be released in methanol, thereby resulting in recovered
adsorbent with free active sites. The adsorption/desorp-
tion cycle continued until a significant downward trend in
the adsorption capacity of the adsorbent was observed.
2.5. Analytical methods
After performing the adsorption process, the samples
were concentrated by liquid–liquid extraction and filtered
with PTFE syringe filters (0.2 µm). The residual concentration
of EE2 was determined using high-performance liquid chro-
matography (HPLC, Agilent 1100, Santa Clara, United States),
supplemented with a C18 column (5 µm, 4.6 mm × 150 mm)
and a UV detector set at 205 nm, and an isocratic elution
mode with a fixed mobile phase composition. Acetonitrile/
buffer (40/60) was used as the mobile phase with a flow rate
of 1 mL/min [38,39]. The pHs were detected by a pH meter
(Behineh B-2000, Iran), and TDS was measured by TDS/EC
conductometer (Metrohm 691, Switzerland).
2.6. pHzpc and ash content determination
The general procedure of pHzpc determination was per-
formed. In brief, 0.1 mol/L KCl solution was prepared in a
volume of 250 mL and was divided equally into 8 flasks of
25 mL to investigate pHzpc. 0.1 mol/L NaOH or HCl solu-
tions were used to adjust the pH in the range of 3–10. In
addition, 0.1 g adsorbent was introduced to each flask. The
samples were shaken for 24 h, and after that, the final pH
of each flask was measured after separating adsorbents
from samples using a centrifuge to establish pHzpc.
The ash content of the adsorbent also was determined
by heating in the muffle furnace. 1 g of NH2-MIL-53(Fe)
was placed in a crucible and then heated at 700°C for 12 h
to determine the ash content of mentioned adsorbent,
similar to previous studies [40].
3. Results and discussion
3.1. Comparison of synthesized MOFs performance
for adsorptive removal of EE2
In order to determine the appropriate adsorbent for the
removal of EE2, the performance of the prepared MOFs
was studied at various pH values. As shown in Fig. 1a,
the performance of both synthesized adsorbents signifi-
cantly depends on the pH value. By increasing the pH, a
significant decrease in the performance of MIL-53(Fe) was
observed, thereby resulting in low removal efficiency in
the pH range of 7–8.5. However, the performance of NH2-
MIL-53(Fe) in this pH range was significantly higher than
Table 1
Experimental levels and ranges of effective parameters for EE2
adsorption
Factors Unit Levels
–1 0 +1
A: pH – 4.5 7 9
B: Adsorbent dosage mg/L 100 250 400
C: Initial concentration µg/L 50 275 500
P. Javidan et al. / Desalination and Water Treatment 264 (2022) 259–269262
that of MIL-53(Fe). There are two interactions between
MIL-53(Fe) and EE2, electrostatic interactions due to its
active amine groups and Van der Waals forces because
of aromatic rings, which significantly impact the higher
adsorption capacity of the adsorbent. Positively charged
amino functional groups incorporated in NH2-MIL-
53(Fe) can interact with the negatively charged functional
groups of contaminants and improve removal efficiency.
According to the pH of surface waters, NH2-MIL-53(Fe)
was selected as the suitable MOF for the removal of EE2
in this study. In addition to the higher adsorption capacity
of NH2-MIL-53(Fe) compared to MIL-53(Fe) at the natu-
ral pH (surface water), it also exhibits better performance
in most different tested pHs, making it a better choice
to apply in various environments.
Fig. 1b indicates the pHzpc of NH2-MIL-53(Fe) is equal
to 9.1, according to obtained data. Therefore the adsorbent
has a positive charge at neutral pH and a negative charge
at alkaline pHs. This finding is consistent with the data in
Fig. 1a, in which as the pH of the sample goes to alkaline
ranges, the adsorption of EE2 declines due to the weaken-
ing of electrostatic interaction between adsorbent and con-
taminant. The ash content of NH2-MIL-53(Fe) was also
determined to be 16% of the total adsorbent mass.
3.2. Characterization of adsorbent
The morphological aspects of the adsorbent were inves-
tigated using the FE-SEM analysis (Fig. 2a–c). As can be
seen, the prepared MOFs have regular, porous, and uni-
formly framed octagonal crystals with an average edge
length of 400 nm. The porous morphology of the adsorbent
verifies its high capacity to adsorb pollutants.
According to the information obtained from the corre-
sponding EDS spectrum (Fig. 2d), as expected, the elements
including Fe, O, N, and C in the regular structure of the final
adsorbent were the same as the initial chemicals used for
the preparation of the adsorbent. As shown in Fig. 2d, the
amount of major elements of C, N, and O (49.45%, 10.6%,
Fig. 1. (a) Comparison of the MIL-53(Fe) and NH2-MIL-53(Fe)
performance in different amounts of pH for EE2 removal and
(b) pHzpc determination of NH2-MIL-53(Fe).
Table 2
RSM design values
Run Factors Response
A: pH B: Adsorbent dosage (mg/L) C: Initial concentration (µg/L) Removal efficiency (%)
1 9.5 250 500 68.1
2 4.5 250 500 80.5
3 7.0 250 275 80.3
4 4.5 250 50 0.3
5 9.5 100 275 50.1
6 4.5 100 275 55.2
7 7.0 250 275 82.6
8 7.0 400 50 39.9
9 7.0 250 275 83.4
10 7.0 100 50 0.5
11 7.0 250 275 75.8
12 7.0 400 500 89.0
13 7.0 250 275 69.4
14 7.0 400 500 92.1
15 9.5 400 275 65.2
16 4.5 400 275 85.5
17 9.5 250 500 79.8
263P. Javidan et al. / Desalination and Water Treatment 264 (2022) 259–269
and 28.4%, respectively) is well-matched with that reported
in previous research [41]. In addition, 11.5% of the iron
in the final structure proves the presence of iron clus-
ters in the adsorbent.
The crystallographic structure and phase purity of
the synthesized sample was corroborated by the XRD pat-
tern presented in Fig. 2f. The adsorbent represents a spec-
ified crystal structure that authenticated FE-SEM results.
Furthermore, the main diffraction peaks are according to
the previously reported data in the literature. Meanwhile,
no other peaks were detected, demonstrating the pure phase
of NH2-MIL-53(Fe) [27,41,42].
In order to further identify different chemical bonds
and functional groups in the prepared adsorbent, an
FTIR analysis was conducted (Fig. 2e). At about 3,458 and
3,375 cm–1, the wide peaks respectively are related to the
Fig. 2. FE-SEM images of NH2-MIL-53(Fe) on different scales: (a) 1 µm, (b) 500 nm, (c) 200 nm, (d) corresponding EDS spectrum and
elemental composition of NH2-MIL-53(Fe), (e) FTIR, and (f) XRD spectra of NH2-MIL-53(Fe).
P. Javidan et al. / Desalination and Water Treatment 264 (2022) 259–269264
asymmetrical and symmetrical stretching vibrations of the
N–H bond, while the peak at 1,652 cm–1 presents bending
vibrations of the N–H bond, emphasizing the presence of
amino groups in the synthesized adsorbent. The two strong
peaks at 1,579 and 1,383 cm–1 are correlated to the asymmet-
ric and symmetric stretching vibrations of carboxyl groups
in aminoterephthalic acid. In addition, the peaks at 1,256
and 1,386 cm–1 also refer to the C–N stretching bond, which
is due to the presence of aromatic amines. Furthermore,
the peaks at 759 and 573 cm–1 can be dedicated to the C–H
bending vibrations of the aromatic ring and the Fe–O
vibration bond, respectively. It is worth noting that all
available peaks, as expected, were in full compliance with
those reported for the synthesized MOFs with iron cores,
establishing that the synthesis of adsorbent has been per-
formed successfully [27,41,43].
BET method is the most well-known procedure used to
assess the surface area of MOFs. According to BET results,
the specific surface area of the adsorbent was 743 m2/g,
which is a hundred times higher than that of some other
MOFs; for example, Yılmaz et al. [44] reported 23 m2/g for
the specific surface area of MIL-53(Fe). Also, the average
diameter of the adsorbent pores is 2.4 nm, which confirms
the microporous structure of the adsorbent and illustrates
its suitability for trapping the target micropollutant. Also,
the pore volume of the synthesized NH2-MIL-53(Fe) is
0.45 cm3/g, which is larger than the pore volume of acti-
vated carbon (0.215 cm3/g) or graphene nanosheets reported
(0.1–0.18 cm3/g) in previous research [45,46].
3.3. Effect of pH, adsorbent dosage, and initial
concentration of EE2
As mentioned earlier, A (pH), B (adsorbent dosage),
and C (initial concentration of contaminant) were consid-
ered the effective parameters, and the following equation,
which has been acquired from the Design-Expert soft-
ware, demonstrates the correlation between A, B, and C
and removal efficiency as the response obtained by the
software.
Removal
189 73 41 38 020056
0 010 001001
2
.. ..
...
ABC
AB AC BC
.. ..84 001001
222
ABC (1)
The ANOVA results are shown in Table 3. The p-val-
ues less than 0.05 indicate the significant effect of the cho-
sen factors on the removal performance. The effect of pH,
adsorbent dosage, and initial concentration of EE2, as well
as the interaction of pH and adsorbent dosage, were signif-
icant (p-value > 0.05). The p-value and F-value of the model
indicate that only 0.01% of the model matching may have
occurred randomly. Additionally, no significant lack of fit
confirmed the validity of the model. On the other hand,
R2-values are defined as correlation coefficients and describe
the consistency between the calculated and experimental
data and also the difference between the predicted deter-
mination coefficient (Pred. R2) and adjusted determina-
tion coefficient (Adj. R2) was about 0.2, which indicates the
model precision. Eventually, the amount of Adeq. precision
was much more than 4, which is an appropriate feature.
The predicted data obtained from Eq. (1) were plot-
ted against actual data (Fig. 3a). As can be seen, all results
are close to the experimental data with no significant devi-
ation, which proves an excellent correlation between the
proposed model and experimental results. Also, as can be
seen in Fig. 3b, the normal plot of residuals exhibit linear
behavior and constant distribution, confirming the accuracy
of the assumption of normality.
The perturbation plot is shown in Fig. 3c and displays
the effect of all parameters in one plot by shifting one
parameter in a specified range while keeping other factors
constant. The vertical axis is associated with the removal
efficiency, whereas the horizontal axis shows the 3 levels
of independent variables. A high slope for a parameter is a
Table 3
Analysis of variance (ANOVA) for the response surface quadratic model for EE2 removal
Factor Sum of square Degree of freedom Mean square F-value p-value
Model 15,933.50 9 1,770.39 164.69 <0.0001 Significant
A: pH 171.13 1 171.13 15.92 0.0072
B: Adsorbent dosage 820.13 1 820.13 76.29 0.0001
C: Initial concentration 10,512.50 1 10,512.50 977.91 <0.0001
AB 56.25 1 56.25 5.23 0.0622
AC 36.00 1 36.00 3.35 0.1170
BC 441.00 1 441.00 41.02 0.0007
A21,190.25 1 1,190.25 110.72 <0.0001
B22.25 1 2.25 0.21 0.6634
C22,704.00 1 2,704.00 251.53 <0.0001
Residual 64.50 6 10.75
Lack of fit 35.75 3 11.92 2.28 0.4310 Not significant
Pure error 28.75 3 9.58
Cor. total 15,998.00 15
R2 = 0.9960, Adjusted R2 = 0.9899, Predicted R2 = 0.9611, Adeq. Precision = 36.072
265P. Javidan et al. / Desalination and Water Treatment 264 (2022) 259–269
sign of high sensitivity of the result to this parameter, so the
initial concentration of EE2 is a much more critical factor
than other parameters. At first, the slope is sharp and when
the concentration increase, decreasing slope confirms that
its effect on removal efficiency declines. Also, the pH graph
indicates a positive slope in acidic regions and a negative
slope in alkaline regions. The maximum removal occurred
at neutral pH, thereby making the adsorbent application
practical in real water samples (pH of real water sample:
7.1). As can be clearly seen in Fig. 3d, at low pH and high
EE2 concentrations, the adsorption operation is performed
with much greater efficiency, but at higher pH, the effect
of the pollutant concentration is decreased. These results
are quite similar to previous reports and despite a large
number of MOFs whose active pH range is narrow, NH2-
MIL-53(Fe) can remove EE2 in a wide pH range and show
its best application in neutral pH [44,47]. Also, as indi-
cated in Fig. 3e, the high initial concentration of the con-
taminant and adsorbent dosage are much better conditions
for adsorbing EE2 using NH2-MIL-53(Fe) because more
adsorbent dosage provides more surface area and more
active sites leading to a more effective interactions with
additional EE2 and enhanced removal efficiency [44,47].
According to the three-dimensional graphs, it was found
that the best conditions for removing EE2 are neutral pH
and an adsorbent dosage of 400 mg/L. The validation of
the proposed model also was conducted in the point pre-
diction of Design-Expert, where it was established that the
mean removal of EE2 according to designed experiments
is 74%. The point prediction also showed the EE2 removal
efficiency is in the range of 70% to 79% (77%).
3.4. Effect of TDS on removal efficiency of EE2
The drinking water in different areas has different val-
ues of TDS (up to 1,200 mg/L), which may affect the adsor-
bent behavior in the adsorption process, so the effect of
TDS on the adsorption process was investigated. Therefore,
the samples with the optimal conditions mentioned ear-
lier were prepared, and their TDS concentrations were
adjusted at 0, 250, 500, and 1,000 mg/L using a TDS stock
solution with a composition presented in Table 4. After sep-
arating the adsorbent and sample preparation, the concen-
tration of EE2 remaining in the solution was determined by
HPLC. It should be considered that in the absence of TDS,
the removal efficiency of EE2 was about 90%. As shown
in Fig. 4, the removal efficiency increased with increasing
TDS, which can result from the salting-out effect, thereby
decreasing the water solubility of EE2, which is consistent
with the results observed in previous studies [48]. Existing
Fig. 3. (a) Predicted against actual response values, (b) normal plot of residuals for the model, (c) deviation from the reference point,
and 3D response surface plots indicating the interactive impact of (d) pH and EE2 concentration and (e) dosage of adsorbent and EE2
concentration.
P. Javidan et al. / Desalination and Water Treatment 264 (2022) 259–269266
aromatic rings in the proposed adsorbent cause the hydro-
phobic interaction between adsorbent and contaminant and
result in adsorbing EE2. The higher TDS of a sample can
decline EE2 solubility, enhancing the hydrophobic interac-
tion between adsorbent and adsorbate and subsequently
increasing EE2 removal efficiency. The findings indicate that
synthesized adsorbent has great potential to remove the
EE2 from real samples that usually contain high TDS.
3.5. Adsorption isotherms
The adsorption isotherm indicates the relationship
between the equilibrium concentration of adsorbate on the
surface of the adsorbent and in the solution. In the present
study, the closest isotherms to the trend of experimental
data were obtained from batch experiments at the optimum
level of effective parameters. The examined models were
Langmuir and Freundlich, and their equations and the values
of the model parameters in addition to the calculated errors,
are shown in Table 5. According to the OLS (ordinary least
squares) values at 298 K, which were calculated from Eq. (1),
the Langmuir model is less consistent with the experimental
data, thereby estimating inaccurate qm. On the other hand,
the Freundlich model has less OLS and is more consistent
with the predicted results, which is according to the reported
results [49]. The amount of Langmuir constant (b) indicates
the adsorption constant, which is related to the interaction
between adsorbent and adsorbate. In the Freundlich model,
more nF expresses the more strong interaction between the
adsorbent and the pollutant. In this research, the value of
1/nF is equal to 1.22, which indicates favorable adsorption.
The adsorption isotherm at the temperature of 298 and 323 K
was investigated and the corresponding parameters are pre-
sented. It is evident that the adsorption capacity is increased
at a higher temperature. However, according to OLS, the
experiments at 298 K are more consistent with the calculated
data from both models. Further discussions on the impact
of temperature on adsorption experiments are provided in
the thermodynamic section. As it is obvious, the Freundlich
model exhibited less OLS at both 298 and 323 K and is more
consistent to describe the experiment.
OL
Scal
qq
i
n
exp
2
1
(2)
3.6. Adsorption thermodynamic
Another parameter affecting the rate of adsorption is
temperature. In thermodynamic studies, the effect of tem-
perature on the adsorption process is examined, result-
ing in thermodynamic parameters [50]. Since temperature
is one of the influential factors in the removal process of
EE2, its effect was investigated at 275, 283, 298, and 323 K.
The experiments were performed under optimal condi-
tions. Fig. 5 shows the plot of lnKc vs. 1/T through which
the thermodynamic parameters were calculated. Table 6
indicates the values of standard entropy change (ΔS° (J/
mol K)), Gibbs free energy change (ΔG° (kJ/mol)), and
standard enthalpy change (ΔH° (kJ/mol)), which were cal-
culated according to Eqs. (2) and (3) using the slope and
y-intercept of the diagram in Fig. 5.
GHTS
(3)
ln
KS
R
H
RT
c
(4)
where R and Kc are related to the universal gas constant
(8.314 J/mol K) and the thermodynamic equilibrium con-
stant of adsorption, respectively. Negative values of ΔG°
Fig. 4. TDS effect on the removal efficiency of EE2.
Table 4
Composition of TDS stock solution (2,000 mg/L)
Cation mg/L Anion mg/L
Ca2+ 235.8 SO4
2– 394.2
Na+158.4 Cl–422.4
K+106.4 NO3
–169
Mg2+ 98.6 HCO3
–420
Table 5
Parameters and values of isotherm models for EE2 adsorption
Isotherm Equation Value of parameter (298 K) Value of parameter (323 K) Calculated error
Freundlich
ln
ln lnqK
n
C
ef e
1
1/nF = 1.22 1/nF = 01.03 OLS = 0.0030 (298 K)
Kf (mg/g)(L/mg)1/n = 13.7 Kf (mg/g)(L/mg)1/n = 15.4 OLS = 0.0034 (323 K)
Langmuir
C
qbq
C
q
e
em
e
m
1
qm (mg/g) = 5,536 qm (mg/g) = 6,741 OLS = 0.0214 (298 K)
b (L/mg) = 0.0016 b (L/mg) = 0.0019 OLS = 0.0247 (323 K)
pH: 7, adsorbent dosage: 0.4 g/L, time: 30 min
267P. Javidan et al. / Desalination and Water Treatment 264 (2022) 259–269
are an indication of the spontaneity of the adsorption, and
the positive value of ΔS° shows that randomness at the
solid–liquid interface in the adsorption process increased
and also confirms that the reaction is entropy-driven.
Positive values of ΔH° demonstrate that the process is nat-
urally endothermic, which proves that the adsorption effi-
ciency increases by increasing temperature.
3.7. Regeneration of NH2-MIL-53(Fe) characteristics
The reusability of the adsorbent after each adsorp-
tion operation under optimal conditions was evaluated.
The regeneration tests for NH2-MIL-53(Fe) were carried
out in optimum experimental conditions (adsorption: pH:
7, adsorbent dosage: 0.4 g/L, time: 5 min, EE2 concentra-
tion: 500 µg/L, temperature: 298 K; washing: methanol:
15 mL, time: 20 min). The adsorbent was separated from
the sample and placed in methanol solution, which is a
suitable solvent for dissolving EE2, to release the adsorbed
EE2 into the methanol. The adsorption/desorption cycle
continued until a significant downward trend in adsorp-
tion capacity was observed. As shown in Fig. 6, the high
removal efficiency was observed in the first stage, but the
efficiency decreased slightly in each cycle. The decrease in
efficiency can be attributed to the binding of the pollutant
at the adsorption sites, occupating active sites, as well as the
loss of adsorbent amounts at the repetition of each cycle.
3.8. Comparison of NH2-MIL-53(Fe) and other adsorbents
Compared to other common adsorbents, NH2-MIL-
53(Fe) has many advantages for EE2 removal. The synthesis
process is straightforward and without using complicated
instruments. The advantages include higher adsorption
capacity, which is compared with other adsorbents in
Table 7. This superior performance of the NH2-MIL-53(Fe)
could be because of regular and porous crystals with a
high specific surface area. Furthermore, amine functional
groups in the structure of NH2-MIL-53(Fe) play an essen-
tial role in the adsorption of EE2. It can be deduced by con-
sidering operational conditions presented in Table 7 that
most experiments are conducted at natural pHs and room
temperature to increase the adsorbent applicability in real
works. Still, the NH2-MIL-53(Fe) indicates a higher adsorp-
tion capacity toward EE2 removal than other adsorbents
reported in other studies.
4. Conclusion
In this study, for the first time, the performance of
MIL-53(Fe) and NH2-MIL-53(Fe) was investigated for the
removal of EE2 as a resistant micropollutant. Compared to
MIL-53(Fe), there is an electrostatic interaction advantage
in addition to the Van der Waals forces in NH2-MIL-53(Fe),
which has positively charged amine functional groups, and
the EE2 with the negatively charged phenolic groups. The
Fig. 5. Thermodynamic plot for the adsorption of EE2 (pH: 7,
adsorbent dosage: 0.4 g/L, time: 30 min, EE2 concentration:
500 µg/L).
Table 6
Values of ΔS°, ΔG°, and ΔH°
Temperature (K) ΔG° (kJ/mol) ΔS° (J/mol K) ΔH° (kJ/mol)
275 –2.05
250.6 67.08
283 –4.02
298 –7.2
323 –14.2
Fig. 6. Removal efficiency after each cycle of desorption
experiments.
Table 7
Comparison of NH2-MIL-53(Fe) and other common adsorbents
for EE2 adsorption
Adsorbent T (K) pH qmax (mg/g) Reference
Powdered activated
carbon
– – 1.163 [24]
Granular activated
carbon
– – 0.436 [24]
un-anthracite 298 7 0.314 [25]
4 K anthracite 298 7 0.299 [25]
Multi-walled carbon
nanotube
298 6 0.472 [51]
Magnetic graphene
oxide
308 7 0.378 [52]
MIL-53(Fe) 298 7 1.26 This study
NH2-MIL-53(Fe) 298 7 1.850 This study
P. Javidan et al. / Desalination and Water Treatment 264 (2022) 259–269268
characterization analysis established that the porosity of the
selected adsorbent and its specific surface area was high,
indicating the formation of MOF crystals and suitable porous
spaces for adsorbing EE2. In spite of some other MOFs,
which have a narrow active pH range, NH2-MIL-53(Fe) is
an efficient adsorbent in a wide pH range and its best appli-
cation is in neutral pH, making it appropriate for natural
water treatment even in the presence of high amounts of
TDS. According to thermodynamic studies, the adsorption
process is endothermic, entropy-driven, and spontaneous.
The best isotherm to interpret the adsorption process was
Freundlich. Eventually, desorption experiments proved
that NH2-MIL-53(Fe) is a stable adsorbent and can be used
in different adsorption cycles. Because of the different func-
tional groups in the structure of MOFs, it can be predicted
that they can be used for the adsorption of various endocrine
disruptors and other perilous micropollutants.
Acknowledgments
Special thanks to the Nanotechnology Research Center
of the School of Environment, College of Engineering,
University of Tehran, Tehran, Iran, for supporting this
research.
Declarations
The authors have no relevant financial or non-financial
interests to disclose.
References
[1] M.O. Barbosa, N.F.F. Moreira, A.R. Ribeiro, M.F.R. Pereira,
A.M.T. Silva, Occurrence and removal of organic micropollu-
tants: an overview of the watch list of EU Decision 2015/495,
Water Res., 94 (2016) 257–279.
[2] K.O. K’oreje, F.J. Kandie, L. Vergeynst, M.A. Abira,
H. Van Langenhove, M. Okoth, K. Demeestere, Occurrence,
fate and removal of pharmaceuticals, personal care products
and pesticides in wastewater stabilization ponds and
receiving rivers in the Nzoia Basin, Kenya, Sci. Total Environ.,
637–638 (2018) 336–348.
[3] R. Tröger, P. Klöckner, L. Ahrens, K. Wiberg, Micropollutants
in drinking water from source to tap - method development
and application of a multiresidue screening method, Sci. Total
Environ., 627 (2018) 1404–1432.
[4] C.X. Huo, P. Hickey, EDC demonstration programme in the
UK – Anglian water’s approach, Environ. Technol., 28 (2007)
731–741.
[5] J.P. Laurenson, R.A. Bloom, S. Page, N. Sadrieh, Ethinyl
estradiol and other human pharmaceutical estrogens in the
aquatic environment: a review of recent risk assessment data,
AAPS J., 16 (2014) 299–310.
[6] V. Chander, B. Sharma, V. Negi, R.S. Aswal, P. Singh, R. Singh,
R. Dobhal, Pharmaceutical compounds in drinking water,
J. Xenobiot., 6 (2016) 5774, doi: 10.4081/xeno.2016.5774.
[7] Y. Feng, Z. Zhang, P. Gao, H. Su, Y. Yu, N. Ren, Adsorption
behavior of EE2 (17α-ethinylestradiol) onto the inactivated
sewage sludge: kinetics, thermodynamics and influence factors,
J. Hazard. Mater., 175 (2010) 970–976.
[8] K. Damkjaer, J.J. Weisser, S.C. Msigala, R. Mdegela, B. Styrishave,
Occurrence, removal and risk assessment of steroid hormones
in two wastewater stabilization pond systems in Morogoro,
Tanzania, Chemosphere, 212 (2018) 1142–1154.
[9] P. Schröder, B. Helmreich, B. Škrbić, M. Carballa, M. Papa,
C. Pastore, Z. Emre, A. Oehmen, A. Langenhoff, M. Molinos,
J. Dvarioniene, C. Huber, K.P. Tsagarakis, E. Martinez-Lopez,
S. Meric Pagano, C. Vogelsang, G. Mascolo, Status of hormones
and painkillers in wastewater effluents across several European
states–considerations for the EU watch list concerning
estradiols and diclofenac, Environ. Sci. Pollut. Res., 23 (2016)
12835–12866.
[10] K. Rehberger, E. Wernicke von Siebenthal, C. Bailey, P. Bregy,
M. Fasel, E.L. Herzog, S. Neumann, H. Schmidt-Posthaus,
H. Segner, Long-term exposure to low 17α-ethinylestradiol
(EE2) concentrations disrupts both the reproductive and the
immune system of juvenile rainbow trout, Oncorhynchus mykiss,
Environ. Int., 142 (2020) 105836, doi: 10.1016/j.envint.2020.
105836.
[11] M. Adeel, X. Song, Y. Wang, D. Francis, Y. Yang, Environmental
impact of estrogens on human, animal and plant life: a critical
review, Environ. Int., 99 (2017) 107–119.
[12] S. Lecomte, D. Habauzit, T.D. Charlier, F. Pakdel, Emerging
estrogenic pollutants in the aquatic environment and breast
cancer, Genes (Basel), 8 (2017) 229, doi: 10.3390/genes8090229.
[13] IARC, Combined Estrogen–Progestogen Contraceptives.
IARC Monographs on the Evaluation of Carcinogenic Risks
to Humans, International Agency for Research on Cancer,
150 cours Albert Thomas, 69372 Lyon Cedex 08, France,
2005 (2012) 100A:283–317.
[14] J. Kent, J.H. Tay, Treatment of 17α-ethinylestradiol,
4-nonylphenol, and carbamazepine in wastewater using an
aerobic granular sludge sequencing batch reactor, Sci. Total
Environ., 652 (2019) 1270–1278.
[15] L.L.S. Silva, J.C.S. Sales, J.C. Campos, D.M. Bila, F.V. Fonseca,
Advanced oxidative processes and membrane separation for
micropollutant removal from biotreated domestic wastewater,
Environ. Sci. Pollut. Res., 24 (2017) 6329–6338.
[16] V. Kumar, D. Avisar, L. Prasanna V, Y. Betzalel, H. Mamane,
Rapid visible-light degradation of EE2 and its estrogenicity in
hospital wastewater by crystalline promoted g-C3N4, J. Hazard.
Mater., 398 (2020) 122880, doi: 10.1016/j.jhazmat.2020.122880.
[17] X. Ma, C. Zhang, J. Deng, Y. Song, Q. Li, Y. Guo, C. Li,
Simultaneous degradation of estrone, 17β-estradiol and
17α-ethinyl estradiol in an aqueous UV/H2O2 system, Int. J.
Environ. Res. Public Health, 12 (2015) 12016–12029.
[18] A. Mohagheghian, R. Nabizadeh, A. Mesdghinia, N. Rastkari,
A.H. Mahvi, M. Alimohammadi, M. Yunesian, R. Ahmadkhaniha,
S. Nazmara, Distribution of estrogenic steroids in municipal
wastewater treatment plants in Tehran, Iran, J. Environ. Health
Sci. Eng., 12 (2014) 97, doi: 10.1186/2052-336X-12-97.
[19] R.O. Pereira, C. Postigo, M.L. de Alda, L.A. Daniel, D. Barceló,
Removal of estrogens through water disinfection processes and
formation of by-products, Chemosphere, 82 (2011) 789–799.
[20] D. Bila, A.F. Montalvão, D. de A. Azevedo, M. Dezotti,
Estrogenic activity removal of 17β-estradiol by ozonation and
identification of by-products, Chemosphere, 69 (2007) 736–746.
[21] C. Postigo, S.D. Richardson, Transformation of pharmaceuticals
during oxidation/disinfection processes in drinking water
treatment, J. Hazard. Mater., 279 (2014) 461–475.
[22] K. Moriyama, H. Matsufuji, M. Chino, M. Takeda, Identification
and behavior of reaction products formed by chlorination of
ethynylestradiol, Chemosphere, 55 (2004) 839–847.
[23] S. Hemidouche, A. Assoumani, L. Favier, A.I. Simion,
C.G. Grigoras, D. Wolbert, L. Gavrila, Removal of Some
Endocrine Disruptors via Adsorption on Activated Carbon,
2017 E-Health and Bioengineering Conference (EHB), IEEE,
Sinaia, Romania, 2017, pp. 410–413.
[24] J. Hartmann, R. Beyer, S. Harm, Effective removal of estrogens
from drinking water and wastewater by adsorption technology,
Environ. Process., 1 (2014) 87–94.
[25] J. He, J. Guo, Q. Zhou, J. Yang, F. Fang, Y. Huang, Analysis of
17α-ethinylestradiol and bisphenol A adsorption on anthracite
surfaces by site energy distribution, Chemosphere, 216 (2019)
59–68.
[26] L. Wang, L. Liu, Z. Zhang, B. Zhao, J. Li, B. Dong, N. Liu,
17α-Ethinylestradiol removal from water by magnetic ion
exchange resin, Chin. J. Chem. Eng., 26 (2018) 864–869.
[27] J. Bedia, V. Muelas-Ramos, M. Peñas-Garzón, A. Gómez-
Avilés, J.J. Rodríguez, C. Belver, A review on the synthesis
269P. Javidan et al. / Desalination and Water Treatment 264 (2022) 259–269
and characterization of metal–organic frameworks for
photocatalytic water purification, Catalysts, 9 (2019) 52,
doi: 10.3390/catal9010052.
[28] N.B. Singh, G. Nagpal, S. Agrawal, Rachna, Water purification
by using adsorbents: a review, Environ. Technol. Innovation,
11 (2018) 187–240.
[29] E.M. Dias, C. Petit, Towards the use of metal–organic
frameworks for water reuse: a review of the recent advances in
the field of organic pollutants removal and degradation and the
next steps in the field, J. Mater. Chem. A, 3 (2015) 22484–22506.
[30] C. Petit, Present and future of MOF research in the field of
adsorption and molecular separation, Curr. Opin. Chem. Eng.,
20 (2018) 132–142.
[31] R. Abazari, A.R. Mahjoub, J. Shariati, Synthesis of a
nanostructured pillar MOF with high adsorption capacity
towards antibiotics pollutants from aqueous solution, J. Hazard.
Mater., 366 (2019) 439–451.
[32] M.R. Azhar, H.R. Abid, H. Sun, V. Periasamy, M.O. Tadé,
S. Wang, One-pot synthesis of binary metal–organic frameworks
(HKUST-1 and UiO-66) for enhanced adsorptive removal of
water contaminants, J. Colloid Interface Sci., 490 (2017) 685–694.
[33] R. Liang, L. Shen, F. Jing, N. Qin, L. Wu, Preparation of MIL-
53(Fe)-reduced graphene oxide nanocomposites by a simple
self-assembly strategy for increasing interfacial contact: efficient
visible-light photocatalysts, ACS Appl. Mater. Interfaces,
7 (2015) 9507–9515.
[34] Q. Sun, M. Liu, K. Li, Y. Zuo, Y. Han, J. Wang, C. Song,
G. Zhang, X. Guo, Facile synthesis of Fe-containing metal–
organic frameworks as highly efficient catalysts for degra dation
of phenol at neutral pH and ambient temperature, Cryst. Eng.
Comm, 17 (2015) 7160–7168.
[35] X. Liu, Y. Zhou, J. Zhang, L. Tang, L. Luo, G. Zeng, Iron
containing metal–organic frameworks: structure, synthesis, and
applications in environmental remediation, ACS Appl. Mater.
Interfaces, 9 (2017) 20255–20275.
[36] N.M. Mahmoodi, J. Abdi, Nanoporous metal–organic
framework (MOF-199): synthesis, characterization and
photocatalytic degradation of Basic Blue 41, Microchem. J.,
144 (2019) 436–442.
[37] M.A. Ghasemzadeh, B. Mirhosseini-Eshkevari, M.H. Abdollahi-
Basir, MIL-53(Fe) metal–organic frameworks (MOFs) as an
efficient and reusable catalyst for the one-pot four-component
synthesis of pyrano[2,3-c]-pyrazoles, Appl. Organomet. Chem.,
33 (2019) e4679, doi: 10.1002/aoc.4679.
[38] P. Gautam, T. Purvis, Method development and validation of
stability indicating RP-HPLC method for the determination
of female hormones in hormone concentrates creams, Pharm.
Anal. Chem., 3 (2017) 1000120, doi: 10.4172/2471-2698.1000120.
[39] B.M. Peake, R. Braund, A.Y.C. Tong, L.A. Tremblay, Detection
and Presence of Pharmaceuticals in the Environment, In: The
Life-Cycle of Pharmaceuticals in the Environment, Woodhead
Publishing which is an imprint of Elsevier, 80 High Street,
Sawston, Cambridge, CB22 3HJ, UK, 2016, pp. 77–107.
[40] C.M. Navarathna, N.B. Dewage, A.G. Karunanayake, E.L. Farmer,
F. Perez, E.B. Hassan, T.E. Mlsna, C.U. Pittman Jr., Rhodamine
B adsorptive removal and photocatalytic degradation on
MIL-53-Fe MOF/magnetic magnetite/biochar composites,
J. Inorg. Organomet. Polym. Mater., 30 (2020) 214–229.
[41] Z. Zhang, X. Li, B. Liu, Q. Zhao, G. Chen, Hexagonal
microspindle of NH2-MIL-101(Fe) metal–organic frameworks
with visible-light-induced photocatalytic activity for the
degradation of toluene, RSC Adv., 6 (2016) 4289–4295.
[42] Q. Xie, Y. Li, Z. Lv, H. Zhou, X. Yang, J. Chen, H. Guo, Effective
adsorption and removal of phosphate from aqueous solutions
and eutrophic water by Fe-based MOFs of MIL-101, Sci. Rep.,
7 (2017) 3316, doi: 10.1038/s41598-017-03526-x.
[43] U.B. Simsek, C. Geçgel, M. Turabik, Different Iron Based
Metal–Organic Frameworks Synthesis, Characterization and
Using for Amoxicillin Removal from Aqueous Solutions, 2nd
International Mediterranean Science and Engineering Congress
(IMSEC 2017), Adana, 2017, pp. 1–7.
[44] E. Yılmaz, E. Sert, F.S. Atalay, Synthesis, characterization
of a metal–organic framework: MIL-53(Fe) and adsorption
mechanisms of methyl red onto MIL-53(Fe), J. Taiwan Inst.
Chem. Eng., 65 (2016) 323–330.
[45] S.G. Herawan, M.S. Hadi, Md. R. Ayob, A. Putra, Characterization
of activated carbons from oil-palm shell by CO2 activation with
no holding carbonization temperature, Sci. World J., 2013 (2013)
624865, doi: 10.1155/2013/624865.
[46] L. Jiang, Y. Liu, G. Zeng, S. Liu, X. Hu, L. Zhou, X. Tan,
N. Liu, M. Li, J. Wen, Adsorption of estrogen contaminants
(17β-estradiol and 17α-ethynylestradiol) by graphene
nanosheets from water: effects of graphene characteristics and
solution chemistry, Chem. Eng. J., 339 (2018) 296–302.
[47] B.M. Jun, H.S. Hwang, J. Heo, J. Han, M. Jang, J. Sohn, C.M. Park,
Y. Yoon, Removal of selected endocrine-disrupting compounds
using Al-based metal–organic framework: performance and
mechanism of competitive adsorption, J. Ind. Eng. Chem.,
79 (2019) 345–352.
[48] Z. Luo, H. Li, Y. Yang, H. Lin, Z. Yang, Adsorption of
17α-ethinylestradiol from aqueous solution onto a reduced
graphene oxide-magnetic composite, J. Taiwan Inst. Chem.
Eng., 80 (2017) 797–804.
[49] L. Mita, M. Forte, A. Rossi, C. Adamo, S. Rossi, D.G. Mita,
M. Guida, M. Portaccio, T. Godievargova, I. Yavour, M. Samir,
M. Eldin, Removal of 17-α ethinylestradiol from water systems
by adsorption on polyacrylonitrile beads: isotherm and
kinetics studies, Ann. Environ. Sci. Toxicol., 2 (2017) 48–58.
[50] S. Rahman, A. Arami-Niya, X. Yang, G. Xiao, G. (Kevin) Li,
E.F. May, Temperature dependence of adsorption hysteresis in
flexible metal–organic frameworks, Commun. Chem., 3 (2020)
186, doi: 10.1038/s42004-020-00429-3.
[51] L.A. Al-Khateeb, A.Y. Obaid, N.A. Asiri, M. Abdel Salam,
Adsorption behavior of estrogenic compounds on carbon
nanotubes from aqueous solutions: kinetic and thermodynamic
studies, J. Ind. Eng. Chem., 20 (2014) 916–924.
[52] X. Wang, Z. Liu, Z. Ying, M. Huo, W. Yang, Adsorption of trace
estrogens in ultrapure and wastewater treatment plant effluent
by magnetic graphene oxide, Int. J. Environ. Res. Public Health,
15 (2018) 1454, doi: 10.3390/ijerph15071454.