Download as pdf or txt
Download as pdf or txt
You are on page 1of 12

Journal of Molecular Liquids 293 (2019) 111376

Contents lists available at ScienceDirect

Journal of Molecular Liquids

journal homepage: www.elsevier.com/locate/molliq

Adsorption of tetracycline, ofloxacin and cephalexin antibiotics on boron


nitride nanosheets from aqueous solution
Raghubeer S. Bangari, Niraj Sinha ⁎
Department of Mechanical Engineering, Indian Institute of Technology Kanpur, Kanpur 208016, India

a r t i c l e i n f o a b s t r a c t

Article history: Due to their ability to treat infections, antibiotics are being used in large quantities for human, animals and plants.
Received 26 March 2019 Although they treat diseases well, they pose serious environmental risk as a result of non-consumption of a large
Received in revised form 27 June 2019 fraction in metabolism. These unmodified antibiotics have been found to pollute surface, ground and drinking
Accepted 15 July 2019
water. Against this backdrop, we report the synthesis of large surface area boron nitride nanosheets (BNNSS) and
Available online 16 July 2019
their application as adsorbent for removal of tetracycline (TC), ofloxacin (OFL) and cephalexin (CFX) from water.
Keywords:
The synthesized adsorbent has been characterized using field emission scanning electron microscopy (FESEM),
Boron nitride nanosheets high resolution transmission electron microscopy (HR-TEM), atomic force microscopy (AFM), X-ray photoelectron
Adsorption spectroscopy (XPS), zeta potential, X-ray powder diffraction (XRD), Fourier transform infrared spectroscopy (FTIR)
Tetracycline and surface area analysis. The largest specific surface area (SBET = 1801.9 m2 g−1) of the synthesized sheets was
Ofloxacin obtained by optimizing the mixing time. The effect of pH, dose and contact time has been investigated to obtain
Cephalexin the optimum parameters for maximum adsorption. The maximum adsorption capacities of BNNSs have been
found to be 346.66 mg g−1 (TC), 72.50 mg g−1 (OFL) and 225.0 mg g−1 (CFX). This superior adsorption behavior
demonstrates the potential of BNNSs for commercial applications for potable and waste water treatment.
© 2019 Elsevier B.V. All rights reserved.

1. Introduction depends on properties such as surface area, porosity and pore diameter
of the adsorbent. Due to high surface area, nanomaterials have received
Due to their ability to treat diseases caused by infections, antibiotics significant attention for remediation of water pollutants as well as other
are being widely used for controlling diseases of humans, animals and applications [11–15]. In one of the earlier works, Wu et al. demonstrated
plants [1–3]. However, a significant fraction of these antibiotics is not the potential of non-functionalized carbon nanotubes as a hydrogen
used in metabolism, and therefore, excreted to the environment. As a sulphide adsorbent in fluid [16–19]. Recently, two-dimensional
result, they have been found to pollute surface and ground water [4,5]. nanomaterials such as graphene and boron nitride nanosheets
These can adversely affect the human health by lowering of human im- (BNNSs) have received significant attention for water remediation.
munity. They also pose potential threat to ecological sustainability by af- BNNSs that are isoelectronic with graphene possess distinct properties
fecting lower organism such as algae [6]. Therefore, it is of utmost that include polarity, high oxidation resistance, high surface area, high
importance to remove the antibiotic residues from the wastewater corrosion resistance, high thermal conductivity and stability, and inert-
from source such as households, hospitals and pharmaceuticals facto- ness to most of the chemicals. As a result, they have been used in water
ries before discharging them to environment. The techniques that remediation applications such as removal of dyes, oils, organic solvents,
have been commonly used for removal of antibiotics include ozonation proteins and heavy metals from water [20–26]. For example, Chen et al.
[7], coagulation [8,9], photo-catalytic degradation [10], ion exchange, synthesized ultrathin functionalized boron nitride (FBN)-based filtra-
liquid membrane separation, reverse osmosis and adsorption. Among tion membranes that showed stability for wide range of solvents at
these techniques, adsorption is preferred due to simplicity in design, high temperatures and extreme pH conditions, demonstrating their po-
ease of fabrication, high efficiency, relatively lower cost and absence of tential for industrial wastewater treatment [24,25]. In another work,
high toxic by-product. For its large-scale practical applications, an ad- Wang et al. have synthesized boron carbon nitride to adsorb dyes for
sorbent should be physically and chemically stable, should be non- wastewater treatment [26].
toxic, should be recyclable and easily separable from the water solution In this work, we have synthesized BNNSs by bottom-up approach
in addition to having high adsorption capacity. The adsorption capacity and have tuned their surface area by controlling the mixing time. The
synthesized BNNSs have been characterized using characterization
⁎ Corresponding author. techniques such as field emission scanning electron microscopy
E-mail address: nsinha@iitk.ac.in (N. Sinha). (FESEM), high resolution transmission electron microscopy (HR-TEM),

https://doi.org/10.1016/j.molliq.2019.111376
0167-7322/© 2019 Elsevier B.V. All rights reserved.
2 R.S. Bangari, N. Sinha / Journal of Molecular Liquids 293 (2019) 111376

atomic force microscopy (AFM), X-ray photoelectron spectroscopy (37%) was acquired from Fisher Scientific, India. TC and OFL were pur-
(XPS), zeta potential, X-ray powder diffraction (XRD), surface area anal- chased from Alfa Aesar, while CFX was acquired from Tokyo Chemical In-
ysis and Fourier transform infrared spectroscopy (FTIR). Further, we dustry, Japan. All the chemicals were used as received. Ultra-pure
have explored their application as adsorbents for removal of three anti- deionized (DI) water (18.2 Mῼ-cm) from a Millipore Milli-Q water puri-
biotics, viz. tetracycline (TC), ofloxacin (OFL) and cephalexin (CFX) from fication system was used to prepare all the solutions.
water. The mechanism of adsorption has been discussed and the ob-
tained adsorption capacities of these three antibiotics have been com-
pared with other adsorbents to depict the potential of BNNSs as 2.2. Synthesis of boron nitride nanosheets
adsorbents for removal of antibiotics from water.
Urea and boric acid were used as the precursors for the synthesis of
2. Experimental section BNNSs. The precursors were mixed in a fixed molar ratio of 1:30. 40 ml
of water and methanol (1:1) were added into the mixture and solution
2.1. Materials was kept at 45 °C for mixing. After mixing the solution, the mixture was
transferred to a tubular furnace and it was heated to 900 °C for 2 h. The
Urea (≥ 99.0% purity) and boric acid (≥ 99.5% purity) were purchased entire reaction was carried out under the nitrogen environment. After
from Himedia and Sigma-Aldrich, respectively. Sodium hydroxide pellets the furnace cooled down to room temperature, a white powder of
(≥ 97% purity) were acquired from Finar Limited, India. Hydrochloric acid BNNSs was obtained.

Fig. 1. HRTEM images of (a) BNNSs and (b) SAED of BNNSs (c) Elements distribution (d) FESEM image of BNNS.
R.S. Bangari, N. Sinha / Journal of Molecular Liquids 293 (2019) 111376 3

2.3. Effect of mixing time 2.5.1. Kinetics of adsorption


The kinetic studies were conducted by adding 40 mg of adsorbent
To investigate the effect of mixing time on surface area, the mixture into the 100 ml of stock solutions of the three antibiotics. The pH was
of precursors and methanol were dried at 45 °C for 0, 12, 24, 50, and adjusted to the optimum values mentioned in Section 2.4. The stock so-
100 h in an oven (Mahindra Scientific, India). The mixtures were then lutions were subjected to continuous shaking inside a thermo-
used to synthesize BNNSs as per the procedure explained in controlled orbital shaker. At the end of equilibrium time, all the samples
Section 2.2. The BNNSs were labelled as BNNSs-0, BNNSs-12, BNNSs- were withdrawn and were filtered with 0.22 μm filter paper. The ad-
24, BNNSs-50, and BNNSs-100, respectively. sorption capacity at different interval of time was found by using the fol-
lowing equation.

2.4. Characterization and measurements


ðC o −C t Þ V
qt ¼ ð2Þ
The properties of synthesized samples were characterized by m
using techniques mentioned in Section 1. The morphologies of syn-
thesized BNNSs were obtained with the help of field emission scan- where qt and Ct are adsorption capacity and concentration at time t,
ning electron microscopy (FESEM, Carl Zeiss) and high resolution respectively.
transmission electron microscopy (HR-TEM, FEI Titan G2 60-300 mi- Further, the experimental data was fitted into pseudo first order
croscope). The specific surface area (SBET), pore volume and pore (PFO) and pseudo second order (PSO) kinetic models to determine
size distribution were calculated from Nitrogen gas adsorption – the adsorption kinetic parameters. The linear forms of PFO and PSO
desorption isotherm at −196 °C (Quanta chrome Autosorb iQ). The are expressed by Eqs. (3) and (4) as
structure of BNNSs was examined by XRD (Panalytical X'Pert Pow-
der). XRD pattern of the synthesized product was recorded by Cu
Kα radiation (λ = 1.506 Å) from 10 to 60°, at a scan rate of 2°/min. ln ðqe −qt Þ ¼ lnqe −k1 t ð3Þ
The surface functional groups were qualitatively measured by FTIR
(Perkin Elmer Spectrum two spectrometer) by recording the trans-
mission/adsorption spectra in the range of 4000–400 cm −1 . The
zeta potential values were determined by NanoBrook 90plus Zeta,
Brookheaven Instrument, USA. AFM measurements were performed
on MFP-3D, Oxford Intruments. Powder sample was mounted on
conductive carbon tape and analysed by X-ray photoelectron spec-
troscopy (PHI 5000 versa probe-ULVAC-PHI Inc.). The remaining
concentration of the three antibiotics was measured by UV–vis spec-
trophotometer (Carry 60, Agilent Technologies).

2.5. Adsorption studies

Adsorption of the three antibiotics (TC, OFL and CFX) was studied
with stock solutions of 50, 30 and 20 ppm, respectively. The stock so-
lutions were prepared by careful addition of accurate weighted
amounts of the respective pharmaceutical in 1000 ml of distilled
water. The adsorption experiments were carried out in 50 ml poly-
propylene tubes. The samples for adsorption studies were prepared
by adding 10 mg of adsorbent in 25 ml of TC, OFL, and CFX, respec-
tively. The pH study was conducted by adjusting the pH from 2 to
14 with the help of 0.1 M HCl and 0.1 M NaOH. The optimum pH
was found to be at 8 for TC and OFL, while it was 12 for CFX. All the
samples were kept inside a thermo-controlled orbital shaker
(Mahendra Scientific, India) at 25 °C and 200 rpm. At the end of equi-
librium, the samples were taken out and were filtered with 0.2 μm
filter paper (Whatman) and analysed by ultra-violet spectropho-
tometry. The remaining concentrations of all the pharmaceuticals
were found by the measurement of absorbance at λmax = 356 nm,
λmax = 286 nm and λmax = 260 for TC, OFL, and CFX, respectively
against a standard calibrated curve. The adsorption capacity at equi-
librium was determined by the following equation.

ðC o −C e ÞV
qe ¼ ð1Þ
m

where C0 (mg l−1) and Ce (mg l−1) are the initial and equilibrium
concentrations of the adsorbates. V (l) is the volume of the solution
and m (g) is the mass of the adsorbent. All the adsorption experi-
ments reported in this study were performed in triplicate and have
been reported in this study. Fig. 2. (a) XRD pattern and (b) FTIR spectra of BNNSs.
4 R.S. Bangari, N. Sinha / Journal of Molecular Liquids 293 (2019) 111376

Fig. 3. High resolution XPS spectra of (a) B 1 s (b), N1 s, and (c) O1s.

t 1 t 3. Results and discussion


¼ þ ð4Þ
qt k2 q2e qe
3.1. Characterization of BNNSs
where k1 (min−1) is the pseudo-first order rate constant and k2 is the
pseudo-second order kinetic rate constant (g mg−1 min−1). The surface morphology of synthesized material was analysed by
HRTEM and FESEM. The HRTEM images in Fig. 1(a) clearly shows the
2.5.2. Adsorption isotherm study lattice fringes of BNNSs, which is found to be 3.33 nm. Selected area
Adsorption isotherm studies were conducted by adding 10 mg of electron diffraction (SAED) of BNNSs in Fig. 1(b) shows the diffuse
BNNSs in 25 ml of the three antibiotics solutions. The adsorption iso- ring pattern. Each ring from the inner most to outer most corre-
therm data were analysed by using Langmuir and Freundlich isotherm sponds to (002), (100) and (101) planes, respectively. Diffuse ring
models that are expressed by Eqs. (5) and (6), respectively. pattern also confirms the amorphous nature of BNNSs. Further, ele-
mental mapping (please refer to Fig. 1(c)) clearly shows the distribu-
ce 1 ce tion of each element present in BNNSs. FESEM image in Fig. 1
¼ þ ð5Þ
qe qm K L qm (d) depicts the layered structure of BNNSs. The structure of the as-
synthesized product was investigated by XRD. The XRD pattern in
1
ln qe ¼ ln K F þ lnce ð6Þ
n
Table 1
where KL (ml mg−1) is the Langmuir adsorption equilibrium constant
Elements atomic percentage.
related to free energy of adsorption, qm (mg g−1) denotes the maximum
adsorption capacity of the adsorbent and KF and n are constants that Elements Atomic %
measure the adsorption capacity and intensity, respectively. The Lang- B 42.0
muir model assumes that the adsorbate covers the adsorbent in the N 40.1
form of a monolayer, while the Freundlich model is based on the as- O 10.2
C 7.6
sumption that there is multilayer adsorption.
R.S. Bangari, N. Sinha / Journal of Molecular Liquids 293 (2019) 111376 5

Fig. 2(a) shows the characteristics peaks at 2θ =24.96° and 42.42°, 3.2. Effect of mixing time on specific surface area
which correspond to (002), (100) and (101) planes, respectively. It
closely matches with the hexagonal structure of boron nitride. Fig. 5(a) shows the N2 adsorption–desorption isotherm, which is in-
Peaks broadening in the XRD pattern show the nanometer size ac- dicative of the volume occupied by synthesized absorbents with differ-
cording to Scherrer equation and implies amorphous nature of ent mixing times. The shape of the optimum isotherm comes under the
BNNSs. Chemical bonds present in the synthesized material was type IV isotherm category as directed by IUPAC (Union of Pure and Ap-
investigated by FTIR spectra. In Fig. 2(b), the peak at ~800 cm−1 indi- plied Chemistry) [34] and is indicative of multilayer adsorption. The
cates in-plane stretching vibration of B\\N. This stretching vibration pore size distribution (PSD) is shown in Fig. 5(b). As the mixing time in-
is attributed to sp2 bonding in B\\N. The peak at ~1386 cm−1 corre- creases from 0 to 100 h, the specific surface area occupied by the adsor-
sponds to out of plane bending vibration of B-N-B. The peaks at bents increases initially. Afterwards, it starts to decrease upon achieving
2925 cm−1 and 3410 cm−1 are indicative of the bending vibration the optimum value. The optimum value of SBET was found to be
peaks of B-NH2 and B-OH, respectively. These bonds are formed 1801.9 m2 g−1 for BNNSs-24. From all the isotherms, it is clear that
due to presence of electron deficient boron and a lone pair of nitro- the hysteresis loop was shown only by BNNSs-12 and BNNSs-24, indi-
gen atom at the edges of hexagonal boron nitride. cating increase in proportion of mesopores. Further, Su et al. reported
XPS results of boron nitride nanosheets (BNNSs) are shown in that an increase in mesopores enhances the surface area and pore vol-
Fig. 3. The XPS results showed the presence of boron, nitrogen and ume [35]. The SBET, total volume (Vtotal), volume of mesopores
oxygen on the adsorbent. Table 1 shows the atomic percentage dis- (Vmeso) and average pore size (Davg) of all the samples are graphically
tribution of elements calculated from the XPS results. The presence represented in Fig. 6(a)–(d), respectively. From the figures, it clear
of oxygen was treated as impurity and might have come from the en- that BNNSs-24 has highest Vmeso and Vtotal. Therefore, BNNSs-24
vironment. The obtained percentage of oxygen is in line with the showed the highest specific surface area. This can be attributed to the
previous studies [27]. There was slight presence of carbon that can fact that the inter-connected porous structure in case of BNNSs-24
be attributed to the carbon base used for analysis. Further, core- was more effective in preventing the restacking of BNNSs. Similar obser-
level high resolution spectrums indicate the bonding between the el- vations have been made in case of graphene in literature [36]. The per-
ements. In Fig. 3(a), B1s spectrum shows the main peak at 190.5 eV, centage increase in SBET, average pore size and total pore volume of
which corresponds to B\\N bonds [28]. Another peak at 192 eV indi- BNNS-24 as compared to BNNS-0 are 53.3, 198, and 279, respectively.
cates the B\\O bond that is due to hydroxylation of boron [29–31].
N1s spectrum shows (please refer to Fig. 2(b)) two peaks. The 3.3. Effect of adsorption parameters
main peak at 398 eV confirms the presence of N\\B bonds [32],
while another small peak at 398.6 eV shows the presence the N\\H 3.3.1. Effect of pH
bonds [30]. The O1s core level shows strong peak at 532.5 eV and a The surface charge on the adsorbent varies with the pH of the solu-
small peak at 533 eV as shown in Fig. 3(c). These peaks indicate the tion as shown by variation of the zeta potential value in Fig. 7. The iso-
presence of O\\H and O\\B bonds, respectively [33]. All the bonds electric point (IEP), at which the adsorbent surface is neutral, can be
depicted by XPS analysis have been confirmed by FTIR spectrum. determined from Fig. 7. IEP was found at pH = 2.89. Therefore, it can
The AFM results are shown in Fig. 4. From the AFM analysis, the be inferred that the surface charge on BNNSs is positive and negative
thickness of BNNSs was found approximately as 1.4 to 2.0 nm that if the value of pH is less than or greater than the value of IEP, respec-
correspond to 4–6 layers. tively. Therefore, adsorption capacity is directly influenced by the pH

Fig. 4. AFM pictures of BNNSs inset image shows the thickness of BNNSs.
6 R.S. Bangari, N. Sinha / Journal of Molecular Liquids 293 (2019) 111376

Fig. 5. (a) N2 adsorption-desorption Isotherms and (b) pores size distributions of BNNSs samples mixed for different hour.

of the solution. Consequently, the effect of the pH was investigated by = 8.0. OFL exists in the solution in the anionic form when the pH is fur-
adjusting the pH of the solution from 2 to 14 with the help of 0.1 M ther increased, while the adsorbent surface remains negative. Therefore,
HCl and 0.1 M NaOH. The molecular structure of TC contains different electrostatic repulsion between the adsorbent and adsorbate increases.
functional group like amino, carboxyl, ketone, and phenols. Conse- Consequently, the adsorption capacity decreases rapidly. The adsorp-
quently, TC has three pKa values (pKa1 = 3.3, pKa2 = 7.7, pKa3 = 9.7). tion diminishes at pH = 12 due to strong electrostatic repulsion [40].
Therefore, it can exist in cationic, zwitterionic and anionic forms de- The maximum adsorption of OFL was found to be 42 mg g−1 at pH 8.
pending on the pH of the solution [37]. TC can exist as cation when Therefore, zwitterionic form in this case as well shows the highest
the pH is b3.3, as zwitterion when the pH is in between 3 and 8, and adsorption.
as anion for pH N8. Fig. 8(a) shows the variation of adsorption capacity So far as CFX is concerned, pka1 and pka2 values in this case are 2.56
of TC with pH. The adsorption capacity increases initially with an in- and 6.88, respectively. The adsorption capacity of CFX increases with the
crease in the pH and attains the maximum value (75 mg g−1) at pH 8 increase in pH (please refer to Fig. 8(c)). Negligible adsorption capacity
where it exists in zwitterionic form. This form contains both positive was found at low pH (pH = 2) since at this pH the adsorbent surface is
and negative charge centre on TC. Therefore, zwitterionic form can eas- protonated due to high H+ ions concentration. Consequently, the sur-
ily interact with the adsorbent by cation exchange as well as surface face charge on the absorbent is positive (please refer to Fig. 7). The
complexation reaction [38]. To support cation exchange and surface value zeta potential was found to be 8.46 mV. Moreover, CFX also exists
complexation reaction, the surface charge on adsorbent should be neg- in the form of cation at this pH. Therefore, the lowest adsorption capac-
ative. Since the zeta potential value at pH 8 was −22.59 mV (please ity can be attributed to the repulsion between the positive charge on the
refer to Fig. 7), it helped in enhancing the cation exchange surface com- adsorbent surface and positive charge on CFX (pka1 = 2.56) [41,42]. Al-
plexation reaction. OFL exists in cationic and anionic forms when the pH Khalisy et al. have also reported the low adsorption at low pH due to
is b6.08 and N8.28, respectively. Between 6.08 and 8.28, OFL exists in high competition between the hydrogen ions and CFX [43]. The maxi-
zwitterionic form [39,40]. Fig. 8(b) shows the effect of pH of the solution mum value of adsorption capacity was found to be 15 mg g−1 at
on the adsorption capacity. The cationic and zwitterionic form of OFL pH 12. At pH 12, more number of OH– are present in the solution,
adsorb on the negatively charged surface of BNNSs. Therefore, the elec- which activates the amino group present in the cephalexin. As a result,
trostatic attraction between the adsorbent and OFL dominates up to pH the adsorption of cephalexin increases due to amino affinitive behaviour
R.S. Bangari, N. Sinha / Journal of Molecular Liquids 293 (2019) 111376 7

Fig 6. Effect of mixing time on BNNSs (a) specific surface area and (b) total pore volume and (c) average pore size (d) mesopore volume.

of BNNSs. However, the adsorption capacity decreases with further in- sites [44]. The high concentration of hydroxyl ions (OH−1) forms
crease in pH (pH = 13 and pH = 14). The adsorption capacities were aqua-complexes. These aqua-complexes reduce the adsorption capacity
found to be 11 and 9 mg g−1 at pH = 13 and pH = 14, respectively. [45,46]. It was observed that TC showed higher adsorption capacity than
The decline in adsorption capacity might be due to the competition be- OFL. This might be due to presence of more number of aromatic rings in
tween the high concentration of OH−1 with the CFX ions for binding the chemical structure of TC than OFL in addition to absence of amino
group in OFL [47].

3.3.2. Effect of contact time


The kinetic behaviour of TC, OFL and CFX adsorptions on BNNSs were
investigated to find the adsorption capacity with time. Fig. 9 shows the
effect of contact time on the adsorption of pharmaceuticals capacity by
BNNSs. The trend shown by all three pharmaceutical can be divided in
three zones. In the initial zone the adsorption rate increases rapidly
with time, which can be attributed to presence of many active surface
sites for adsorption [44,45]. As the time proceeds, the adsorption rate
slows down gradually (second zone). This indicates increased utiliza-
tion/occupation of active sites on the adsorbent's surface. Finally, the ad-
sorption rate almost becomes constant with time and the equilibrium is
achieved (third/final zone). From Fig. 9, the time range of initial zone
shown by OFL, TC and CFX are 100,150, and 250 min, respectively. Fur-
ther, the second zone exists in range of 250–300, 100–250 and
150–400 min for CFX, OFL and TC, respectively. Finally, the equilibrium
is attained in the order as OFL N TC N CFX. The experimental adsorption
capacities of TC, OFL and CFX with time at the end of equilibrium were
found to be 75 mg g−1, 42.0 mg g−1 and 15.0 mg g−1, respectively.
The equilibrium adsorption data was analysed by PFO and PSO kinetic
Fig. 7. Zeta potential value under different pH. models. The kinetic constants K1 and K2 were determined by the slope
8 R.S. Bangari, N. Sinha / Journal of Molecular Liquids 293 (2019) 111376

Fig. 8. Effect of pH on the adsorption capacity of (a) TC, (b) OFL and (c) CFX.
Fig. 9. Effect of contact time on the adsorption capacity of (a) TC, (b) OFL and (c) CFX by
BNNSs.
of PFO and PSO, respectively. The correlation values of PSO are greater
than PFO kinetic model (please refer to Table 2). This indicates that an-
tibiotics adsorption is governed by PSO kinetic model. The correlation were found to be 3.06 × 10−4, 14.3 × 10−4 and 2.67 × 10−4, respec-
values (R2) of PSO kinetic model of TC, OFL and CFX are 0.99, 0.99 and tively. Based on K2 values, it is concluded that the rate of adsorption of
0.96, respectively. The PSO kinetic constant values of TC, OFL and CFX three antibiotics followed sequence of OFL N TC N CFX to attain the
R.S. Bangari, N. Sinha / Journal of Molecular Liquids 293 (2019) 111376 9

Table 2
Pseudo first -order and pseudo-second-order kinetic parameters.

Kinetic model Parameters Tetracycline Ofloxacin Cephalexin

Pseudo –first-order qe,exp (mg g−1) 75.0 42.0 15.0


qe,(cal) (mg g−1) 42.59 21.13 19.43
k1(min−1) 0.0124 0.0145 0.0098
R2 0.954 0.857 0.951
Pseudo-second-order qe,cal (mg g−1) 79.37 43.47 20.66
k2 (g mg−1 min−1) 3.06E-04 1.43E-03 2.67E-04
R2 0.999 0.999 0.961

equilibrium. Therefore, OFL achieved the equilibrium in the least time


among all antibiotics. PSO also indicates that the adsorption rate limit-
ing step might be governed by chemisorption [48].

3.3.3. Effect of adsorbent dose


The effect of adsorbent dose was studied by increasing the dose from
0.2 g to 1.0 g of adsorbent per litre. Fig. 10 shows the effect of adsorbent
dose on the adsorption capacity and percentage removal of three antibi-
otics considered in this study. The percentage removal increases, while
the adsorption capacity decreases with an increase in the dose of the ad-
sorbent for all three antibiotics. From Fig. 10(a), it can be seen that the
adsorption capacity of TC decreased from 145 mg g−1 to 45 mgg−1
and the corresponding percentage removal increased from 55 to 82.5
when the dose increased from 0.2 to 1.0 g per litre. The same dose
rate studies were also carried out for OFL and CFX. In case of OFL, it
was found that the adsorption capacity decreased from 57.0 mg g−1 to
22.6 mg g−1 and the corresponding percentage removal increased
from 39 to 75 (please refer to Fig. 10(b)). Finally, as can be seen in
Fig. 10(c), the adsorption capacity for CFX antibiotics decreased from
20 mg g−1 to 9.5 mg g−1 and associated percentage removal increased
from 18.2 to 55. The high adsorption capacity at 0.2 g of adsorbent per
litre can be explained as follows: as the adsorbent dose increases, num-
ber of active sites also increases. Therefore, more number of adsorbate
molecules are adsorbed, resulting in enhanced removal of antibiotics.
The adsorption capacity curve is very steep for 0.2 g adsorbent per
litre. The steepness of the curve and the capacity decreases afterwards
with an increase in the dose. The decrease in adsorption capacity at a
particular dose can be attributed to the aggregation of adsorbent,
which reduces the number of exposed active sites.

3.3.4. Adsorption isotherm


Adsorption isotherm models have been used to describe the distri-
bution of adsorbed molecules on the adsorbents after reaching equilib-
rium state. The adsorption isotherm of the three antibiotics is shown in
Fig. 11. It can be seen in the figure that the adsorption capacity increases
with increase in the initial concentration. The experimental data of ad-
sorption isotherm was fitted into the linear form of Langmuir and
Freundlich isotherm models as stated earlier in Eqs. (5) and (6), respec-
tively. Adsorption parameters were calculated from the slopes and in-
tercepts of the linear equations and are shown in Table 3.
Experimental isotherm values of TC and OFL pharmaceuticals followed
the Langmuir model as evident from their corresponding correlation co-
efficient values (R2 = 0.9817 and 0.978, respectively). This shows that
TC and OFL form a monolayer on the surface of BNNSs. The maximum
adsorption capacities (qmax) of TC and OFL were experimentally found Fig. 10. Effect of adsorbent (BNNSs) dosage on the adsorption capacity and percentage
to be 346.66 and 72.5 mg g−1, respectively. Here it is important to men- removal of (a) TC, (b) OFL and (c) CFX.
tion that the BNNSs-24 synthesized and optimized in this study showed
1.2 times higher adsorption capacity of TC than the Porous BNNS (p- 0.996). Therefore, there was multilayer adsorption of CFX on BNNSs.
BNNS) reported previously in the literature (please refer to Table 4) The maximum adsorption capacity of CFX was found to be
[47]. It is to be noted that the SBET of BNNSs-24 is also approximately 225.0 mg g−1. When compared to other adsorbents and their adsorp-
1.2 times higher than SBET of p-BNNS. Therefore, it can be concluded tion capacities, the adsorption capacity of CFX found in this study is sig-
that the high adsorption capacity of TC with BNNSs-24 was due to nificantly higher.
high SBET of BNNSs-24. In case of CFX, the adsorption isotherm of the Table 4 compares different adsorbents and their adsorption capaci-
experimental values were found to be in close agreement with the ties against these three antibiotics. It is clear from the table that
Freundlich isotherm model with very high correlation value (R2 = BNNSs-24 performs better than graphene oxide, magnetically reduced
10 R.S. Bangari, N. Sinha / Journal of Molecular Liquids 293 (2019) 111376

graphene oxide, carbon nanotubes, activated carbon composite and p- Table 3


BNNS as adsorbent for TC. In terms of percentage, BNNSs-24 adsorption Langmuir and Freundlich adsorption isotherm parameters for TC, OFL and CFX.

capacity is higher by 10.75, 890.45, 246.66, 32.31 and 22.06, respectively Isotherm models Isotherm parameters Adsorbent used BNNSs
for these adsorbents. The studies related to OFL adsorption are very lim- Tetracycline Ofloxacin Cephalexin
ited in literature. The OFL adsorption capacity obtained in this study is
Langmuir qmax,exp (mg g−1) 346.667 72.50 225.0
q max,cal (mg g−1) 384.62 75.18 a

KL (L mg−1) 0.0200 0.0711


R2 0.9817 0.978
Freundlich kF (mg1−n Ln g−1) 48.211 19.286 1.905
n 2.913 3.93 1.100
R2 0.9872 0.965 0.996
a
Calculations are not reported due to low correlation.

approximately 2.5 times higher than the highest adsorption capacity


(30 mg g−1) reported previously. In case of CFX, our synthesized mate-
rial showed higher adsorption capacity than variants of activated car-
bon, bentonite, natural zeolite, and magnetite coated zeolite. The
adsorption capacity of BNNSs-24 is higher by 2000%, 1300% and 400%
for bentonite, natural zeolite and activated carbon, respectively. The su-
perior adsorption performance of BNNSs-24 for the three antibiotics can
be attributed to its large surface area, pore size and pore volume. The π-
π interactions between the aromatic rings of antibiotics and the ring
network of BNNSs-24 is considered to be responsible for adsorption
[47]. Another adsorption mechanism is related to the amino affinity be-
tween the antibiotics and BNNSs (that is present in case of TC and CFX).
Therefore, TC and CFX showed more adsorption capacity than OFL.
Among all the pharmaceuticals, TC showed the highest adsorption ca-
pacity due to the presence of more number of aromatic rings in its struc-
ture than the other two.

Table 4
Comparison of maximum adsorption capacity of pharmaceuticals reported by using differ-
ent adsorbents.

Antibiotics Adsorbent pH Conc. Adsorption Ref.

(mg/L) Capacity
(mg g−1)

Tetracycline Grapheme oxide 3.6 333 313 [49]


Fe3O4 rGO composite 4.5–5.6 75 35 [50]
Multi-walled CNT 6 84 100 [48]
Single-walled CNT 5.7 84 340
MnFe2O4-activatedcarbon 5 222 262 [51]
composite
Porous BNNSs 7 120 284 [47]
(Ce)
Polygorsite 8.7 800 99 [52]
This study 8 290 346.66
(Ce)
Ofloxacin Chitosan/reed biochar 7.0 4–20 6.64 [39]
composite (CRBC)
β-CD-grafted hydroxyapatite – 2.0 30.0 [53]
(β-CD-g-HA)
This study 8 166 72.50
(Ce)
Cephalexin Modified activated carbon 7 – 137.02 [54]
namely KAC
Modified activated carbon 7 – 118.08
namely KCAC
Alligator activated carbon – – 45 [55]
Natural zeolite 7 – 16.10 [56]
Zeolite coated with 7 – 24.5
manganese oxide
AC – – 66.0 [57]
AC-Cu – – 78.0
AC-Fe – – 75.0
Activated carbon 6.1 – 17.361 [43]
Bentonite 6.1 – 10.384
Natural zeolite 7.0 – 16.100 [58]
Fe3O4 coated zeolite 7.0 – 24.900
This study 12 200 225
(Ce)
Fig. 11. Adsorption isotherm of (a) TC, (b) OFL and (c) CFX.
R.S. Bangari, N. Sinha / Journal of Molecular Liquids 293 (2019) 111376 11

4. Conclusions [18] X.C. Wu, W.J. Zhang, R. Sammynaiken, Q.H. Meng, Q.Q. Yang, E. Zhan, Q. Liu, W.
Yang, R. Wang, Non-functionalized carbon nanotube binding with hemoglobin, Col-
loids Surf. B: Biointerfaces 65 (1) (2008) 146–149.
A highly efficient adsorbent, BNNSs, having high surface area has [19] Wu, X. C.; Eric, Z.; Wu, L.; Wu, Y. H.; Zhang, W. J. On further understanding of inter-
been synthesized and characterized in the present study. It was found action of pristine carbon nanotubes with hemoglobin, serum and H2S. J. Nanomater.
Mol. Nanotechnol. 2013, 2 (2), 2–4.
that the surface area was dependent on the mixing time of the precur- [20] W. Lei, D. Portehault, D. Liu, S. Qin, Y. Chen, Porous boron nitride nanosheets for ef-
sors and methanol prior to synthesis. The highest specific surface area fective water cleaning, Nat. Commun. 4 (2013) 1777.
(SBET = 1801.9 m2 g−1) was obtained at 24 h mixing time. The synthe- [21] Y. Zhang, C. Chan, Z. Li, J. Ma, Q. Meng, X. Cheng, J. Fan, Lipid extraction by boron ni-
tride nanosheets from liquid-ordered and liquid-disordered nanodomains, Nano-
sized nanosheets were tested for removal of TC, OFL and CFX. Maximum scale 10 (2018) 14073–14081.
adsorption could be obtained at pH values of 8 (TC and OFL) and 12 [22] J. Li, X. Xiao, X. Xu, J. Lin, Y. Huang, Y. Xue, P. Jin, J. Zou, C. Tang, Activated boron ni-
(CFX), respectively. Studies showed high adsorption capacities of tride as an effective adsorbent for metal ions and organic pollutants, Sci. Rep. 3
346.66 mg g−1 (TC), 72.50 mg g−1 (OFL) and 225.0 mg g−1 (CFX) that (2013) 3208.
[23] Bangari, R. S.; Singh, A.; Namsani, S.; Singh, J. K.; Sinha, N. Magnetite coated boron
are comparable to other adsorbents reported in literature. The adsorp- nitride nanosheets for the removal of arsenic (V) from water. ACS Appl. Mater. In-
tions of TC and OFL were based on Langmuir isotherm, implying mono- terfaces 2019, 11 (V), 19017–19028.
layer adsorption. On the other hand, CFX adsorption was based on [24] Chen, C.; Wang, J.; Liu, D.; Yang, C.; Liu, Y.; Ruoff, R. S.; Lei, W. Functionalized boron
nitride membranes with ultrafast solvent transport performance for molecular sep-
Freundlich isotherm model, and therefore, indicated multilayer adsorp- aration. Nat. Commun. 2018, 9 (1), 1902.
tion. The three antibiotics investigated in this study obeyed PSO kinetics [25] C. Chen, D. Liu, J. Wang, L. Wang, J. Sun, W. Lei, Functionalized boron nitride mem-
strongly. BNNSs were able to adsorb these antibiotics significantly due branes with multipurpose and super-stable semi-permeability in solvents, J. Mater.
Chem. A 6 (42) (2018) 21104–21109.
to similarity in their ring structure, which generates π-π interaction [26] J. Wang, J. Hao, D. Liu, S. Qin, C. Chen, C. Yang, Y. Liu, T. Yang, Y. Fan, Y. Chen, W. Lei,
and amino affinity of BNNSs in some cases. These results demonstrate Flower stamen-like porous boron carbon nitride nanoscrolls for water cleaning,
considerable potential of BNNSs as an adsorbent for remediation of Nanoscale 9 (28) (2017) 9787–9791.
[27] S. Marchesini, A. Regoutz, D. Payne, C. Petit, Tunable porous boron nitride: investi-
pharmaceuticals from contaminated water in addition to the existing gating its formation and its application for gas adsorption, Microporous Mesoporous
adsorbents for such applications. Mater. 243 (2017) 154–163.
[28] V. Cholet, L. Vandenbulcke, J.P. Rouan, P. Baillif, R. Erre, Characterization of boron
nitride films deposited from BCl3-NH3-H2 mixtures in chemical vapour infiltra-
tion conditions, J. Mater. Sci. 29 (6) (1994) 1417–1435.
Acknowledgements
[29] Q. Liu, C. Chen, M. Du, Y. Wu, C. Ren, K. Ding, M. Song, C. Huang, Porous hexag-
onal boron nitride sheets: effect of hydroxyl and secondary amino groups on
This research was partially supported by grant given to Niraj Sinha photocatalytic hydrogen evolution, ACS Appl. Nano Mater. 1 (9) (2018)
4566–4575.
by IIT Kanpur (IITK/ME/2013370).
[30] L. Ci, L. Song, C. Jin, D. Jariwala, D. Wu, Y. Li, A. Srivastava, Z.F. Wang, K. Storr, L.
Balicas, F. Liu, P.M. Ajayan, Atomic layers of hybridized boron nitride and graphene
References domains, Nat. Mater. 9 (5) (2010) 430–435.
[31] D. Liu, W. Lei, S. Qin, Y. Chen, Template-free synthesis of functional 3D BN architec-
[1] L.S. Lee, Sorption of three tetracyclines by several soils: assessing the role of pH and ture for removal of dyes from water, Sci. Rep. 4 (2014) 1–5.
cation exchange, Environ. Sci. Technol. 39 (19) (2005) 7452–7459. [32] G. Ciofani, G.G. Genchi, I. Liakos, A. Athanassiou, D. Dinucci, F. Chiellini, V. Mattoli, A
[2] A. Sapkota, A.R. Sapkota, M. Kucharski, J. Burke, S. Mckenzie, P. Walker, R. Lawrence, simple approach to covalent functionalization of boron nitride nanotubes, J. Colloid
Aquaculture practices and potential human health risks: current knowledge and fu- Interface Sci. 374 (1) (2012) 308–314.
ture priorities, Environ. Int. 34 (8) (2008) 1215–1226. [33] A.S. Nazarov, V.N. Demin, E.D. Grayfer, A.I. Bulavchenko, A.T. Arymbaeva, H.J. Shin, J.Y.
[3] K. Kümmerer, Antibiotics in the aquatic environment – a review – part I, Choi, V.E. Fedorov, Functionalization and dispersion of hexagonal boron nitride (H-
Chemosphere 75 (4) (2009) 417–434. BN) nanosheets treated with inorganic reagents, Chem. Asian J. 7 (3) (2012) 554–560.
[4] T. Heberer, Occurrence, fate, and removal of pharmaceutical residues in the aquatic [34] K. Sing, D. Everett, R. Haul, L. Moscou, R. Pierotti, J. Rouquerol, T. Siemieniewska,
environment: a review of recent research data, Toxicol. Lett. 131 (1–2) (2002) 5–17. Reporting physisorption data for gas/solid systems with special reference to the
[5] B. Halling-Sorensen, S.N. Nielsen, P.F. Lanzky, F. Ingerslev, H.C.H. Lutzheft, S.E. determination of surface area and porosity, Pure Appl. Chem. 57 (4) (1985)
Jorgensen, Occurrence, fate and effects of pharmaceutical substances in the 603–619.
environment-a review, Chemosphere 36 (2) (1998) 357–393. [35] H. Su, Z. Ye, N. Hmidi, High-performance Iron oxide–graphene oxide nanocomposite
[6] P. Grenni, V. Ancona, A. Barra, Ecological effects of antibiotics on natural ecosystems: adsorbents for arsenic removal, Colloids Surf. A Physicochem. Eng. Asp. 522 (2017)
a review, Microchem. J. 136 (2018) 25–39. 161–172.
[7] J. Gomes, R. Costa, R.M. Quinta-ferreira, R.C. Martins, Application of ozonation for [36] Z. Wang, Z. Tang, Z. Han, S. Shen, B. Zhao, J. Yang, Effect of drying conditions on the
pharmaceuticals and personal care products removal from water, Sci. Total Environ. structure of three-dimensional N-doped graphene and its electrochemical perfor-
586 (2017) 265–283. mance, RSC Adv. 5 (26) (2015) 19838–19843.
[8] P. Taylor, W. Yang, Y. Wu, L. Zhang, J. Jiang, L. Feng, Removal of five selected pharma- [37] C. Gu, K.G. Karthikeyan, Sorption of the antibiotic tetracycline to humic-mineral
ceuticals by coagulation in the presence of dissolved humic acids and kaolin, complexes, J. Environ. Qual. 37 (2008) 704–711.
Desalin. Water Treat. 54 (2015) 1134–1140. [38] L. Ji, W. Chen, L. Duan, D. Supporting Information Zhu, Mechanism for strong
[9] C. Sheng, A.G.A. Nnanna, Y. Liu, J.D. Vargo, Removal of trace pharmaceuticals from adsorption of tetracycline to carbon nanotubes: a comparative study using
water using coagulation and powdered activated carbon as pretreatment to ultra fil- activated carbon and graphite as adsorbents, Environ. Sci. Technol. 43 (7)
tration membrane system, Sci. Total Environ. 550 (2016) 1075–1083. (2009) 2322–2327.
[10] E. Hapeshi, A. Achilleos, M.I. Vasquez, C. Michael, N.P. Xekoukoulotakis, D. [39] C. Zhu, Y. Lang, B. Liu, H. Zhao, Ofloxacin adsorption on chitosan/biochar composite:
Mantzavinos, D. Kassinos, Drugs degrading photocatalytically: kinetics and mecha- kinetics, isotherms, and effects of solution chemistry, Polycycl. Aromat. Compd. 39
nisms of ofloxacin and atenolol removal on titania suspensions, Water Res. 44 (6) (3) (2019) 287–297.
(2010) 1737–1746. [40] K.W. Goyne, J. Chorover, J.D. Kubicki, A.R. Zimmerman, S.L. Brantley, Sorption of the
[11] G. Ghasemzadeh, M. Momenpour, F. Omidi, M.R. Hosseini, M. Ahani, A. Barzegari, antibiotic ofloxacin to mesoporous and nonporous alumina and silica, J. Colloid In-
Applications of nanomaterials in water treatment and environmental remediation, terface Sci. 283 (1) (2005) 160–170.
Front. Environ. Sci. Eng. 8 (4) (2014) 471–482. [41] M.R. Samarghandi, T.J. Al-Musawi, A. Mohseni-Bandpi, M. Zarrabi, Adsorption of
[12] N. Sinha, J. Ma, J.T.W. Yeow, Carbon nanotube-based sensors, J. Nanosci. cephalexin from aqueous solution using natural zeolite and zeolite coated with
Nanotechnol. 6 (3) (2006) 573–590. manganese oxide nanoparticles, J. Mol. Liq. 211 (2015) 431–441.
[13] Y. Xia, Nanomaterials at work in biomedical research, Nat. Mater. 7 (10) (2008) [42] H.R. Pouretedal, N. Sadegh, Effective removal of amoxicillin, cephalexin, tetra-
758–760. cycline and penicillin G from aqueous solutions using activated carbon nano-
[14] Y. Sun, N. Liu, Y. Cui, Promises and challenges of nanomaterials for lithium-based re- particles prepared from vine wood, J. Water Process Eng. 1 (2014) 64–73.
chargeable batteries, Nat. Energy 1 (7) (2016) 1–12. [43] R.S. Al-khalisy, A.M.A. Al-haidary, A.H. Al-dujaili, R.S. Al-khalisy, A.M.A. Al-haidary,
[15] Y. Jia, H. Zhang, S. Yang, Z. Xi, T. Tang, R. Yin, W. Zhang, Electrospun PLGA membrane A.H. Al-dujaili, Aqueous phase adsorption of cephalexin onto bentonite and acti-
incorporated with andrographolide-loaded mesoporous silica nanoparticles for vated carbon, Sep. Sci. Technol. 45 (2010) 1286–1294.
sustained antibacterial wound dressing, Nanomedicine 13 (22) (2018) 2881–2899. [44] S.V. Mohan, J. Karthikeyan, Removal of lignin and tannin colour from aqueous solu-
[16] X.C. Wu, W.J. Zhang, D.Q. Wu, R. Sammynaiken, R. Wang, Q. Yang, Using carbon tion by adsorption onto activated charcoal, Environ. Pollut. 97 (1–2) (1997)
nanotubes to absorb low-concentration hydrogen sulfide in fluid, IEEE Trans. 183–187.
Nanobioscience 5 (3) (2006) 204–209. [45] S. Venkata Mohan, S.V. Ramanaiah, B. Rajkumar, P.N. Sarma, Biosorption of fluoride
[17] X.C. Wu, W.J. Zhang, R. Sammynaiken, Q.H. Meng, D.Q. Wu, Q. Yang, W. Yang, E.M. from aqueous phase onto algal Spirogyra IO1 and evaluation of adsorption kinetics,
Zhang, R. Wang, Measurement of low concentration and Nano-quantity hydrogen Bioresour. Technol. 98 (5) (2007) 1006–1011.
sulfide in sera using Unfunctionalized carbon nanotubes, Meas. Sci. Technol. 20 [46] S.V. Mohan, J. Karthikeyan, Removal of diazo dye from aqueous phase by algae Spi-
(10) (2009). rogyra species, Toxicol. Environ. Chem. 74 (3–4) (2000) 147–154.
12 R.S. Bangari, N. Sinha / Journal of Molecular Liquids 293 (2019) 111376

[47] D. Liu, W. Lei, S. Qin, K.D. Klika, Y. Chen, Superior adsorption of pharmaceutical mol- [54] M.J. Ahmed, S.K. Theydan, Adsorption of cephalexin onto activated carbons from
ecules by highly porous BN nanosheets, Phys. Chem. Chem. Phys. 18 (1) (2016) Albizia Lebbeck seed pods by microwave-induced KOH and K2CO3 activations,
84–88. Chem. Eng. J. 211–212 (2012) 200–207.
[48] L. Ji, W. Chen, L. Duan, D. Zhu, Mechanisms for strong adsorption of tetracycline to [55] M. Miao, Q. Liu, L. Shu, Z. Wang, Y. Liu, Removal of cephalexin from effluent by acti-
carbon nanotubes: a comparative study using activated carbon and graphite as ad- vated carbon prepared from alligator weed: kinetics, isotherms, and thermody-
sorbents, Environ. Sci. Technol. 43 (7) (2009) 2322–2327. namic analyses, Process. Saf. Environ. Prot. 104 (2016) 481–489.
[49] Y. Gao, Y. Li, L. Zhang, H. Huang, J. Hu, S. Mazhar, X. Su, Adsorption and removal of [56] M. Reza, T.J. Al-musawi, A. Mohseni-bandpi, M. Zarrabi, Adsorption of cephalexin
tetracycline antibiotics from aqueous solution by graphene oxide, J. Colloid Interface from aqueous solution using natural zeolite and zeolite coated with manganese
Sci. 368 (1) (2012) 540–546. oxide nanoparticles, J. Mol. Liq. 211 (2015) 431–441.
[50] Y. Zhang, B. Chen, L. Zhang, J. Huang, F. Chen, Z. Yang, J. Yao, Z. Zhang, Controlled as- [57] H. Liu, W. Liu, J. Zhang, C. Zhang, L. Ren, Y. Li, Removal of cephalexin from aqueous
sembly of Fe3O4 magnetic nanoparticles on graphene oxide, Nanoscale 3 (4) (2011) solutions by original and Cu(II)/Fe(III) impregnated activated carbons developed
1446. from lotus stalks kinetics and equilibrium studies, J. Hazard. Mater. 185 (2–3)
[51] L. Shao, Z. Ren, G. Zhang, L. Chen, Facile synthesis, characterization of a MnFe2O4/ac- (2011) 1528–1535.
tivated carbon magnetic composite and its effectiveness in tetracycline removal, [58] A. Mohseni-bandpi, T.J. Al-musawi, E. Ghahramani, M. Zarrabi, S. Mohebi, S.
Mater. Chem. Phys. 135 (1) (2012) 16–24. Abdollahi, Improvement of zeolite adsorption capacity for cephalexin by coating
[52] P. Chang, Z. Li, T. Yu, S. Munkhbayer, Sorptive removal of tetracycline from water by with magnetic Fe3O4 nanoparticles, J. Mol. Liq. 218 (2016) 615–624.
palygorskite, J. Hazard. Mater. 165 (2009) 148–155.
[53] W. Tang, J. Zhao, B. Sha, H. Liu, Adsorption and drug release based on β-
cyclodextrin-grafted hydroxyapatite composite, J. Appl. Polym. Sci. 127 (4) (2013)
2803–2808.

You might also like