Journal of Molecular Liquids 268 (2018) 728–733

Contents lists available at ScienceDirect

Journal of Molecular Liquids

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

Predicting equilibrium time by adsorption kinetic equations and

modifying Langmuir isotherm by fractal-like approach
Qili Hu a, Ye Liu a, Chuanping Feng b, Zhenya Zhang a,⁎, Zhongfang Lei a, Kazuya Shimizu a
a
Graduate School of Life and Environmental Sciences, University of Tsukuba, 1-1-1 Tennodai, Tsukuba, Ibaraki 305-8572, Japan
b
School of Water Resources and Environment, China University of Geosciences (Beijing), Beijing 100083, China

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

Article history: In this work, the pseudo-first-order (PFO), pseudo-second-order (PSO) and Vermeulen equations were modified
Received 22 May 2018 by introducing an adjustable parameter i.e. the fractional surface coverage θ. The Langmuir kinetic equation was
Received in revised form 9 July 2018 modified by the fractal-like approach and the separation factor RH was proposed to express essential character-
Accepted 29 July 2018
istics of the fractal-like Langmuir isotherm. The kinetic and isotherm data obtained from nitrate adsorption on the
Available online 30 July 2018
PAN/AC composite were used to evaluate the validity of these models. Results indicated that the modified PSO
Keywords:
equation could accurately predict the equilibrium time at different initial nitrate concentrations (R2 N 0.998)
Adsorption and that the fractal-like Langmuir isotherm could better describe the equilibrium data at different temperatures
Equilibrium time (R2 N 0.994). The magnitude of parameter b reflected the affinity of the adsorbent for the adsorbates. The inno-
Kinetics vation and significance of the present study was that the modified kinetic and isotherm models could predict
Langmuir isotherm the equilibrium time and describe the heterogeneous surfaces, respectively. Therefore, this work is expected to
Fractal-like extend the application scope of the PFO, PSO and Vermeulen equations and Langmuir isotherm.
© 2018 Published by Elsevier B.V.

1. Introduction attempts to modify the PFO, PSO and Vermeulen equations to predict
the equilibrium time of nitrate adsorption on the PAN/AC composite.
Adsorption is an attractive and promising technique for the purifica- On the other hand, the Langmuir isotherm [8] is usually not suitable
tion of municipal and industrial wastewaters [1]. Adsorption efficiency to describe the isotherm data obtained from aqueous solution due to a
is related to not only the equilibrium features but also the kinetics of consequence of the fact that its underlying assumptions such as mono-
the adsorption process for specific adsorption systems [2]. Therefore, a layer coverage of the adsorbent surface and energetic homogeneity of
study of adsorption equilibrium and kinetics is essential to design in- the adsorption sites may not be fulfilled [9]. This is particularly true
dustrial reactors, reduce operating cost and gain insights into adsorp- for the most important adsorbent activated carbon. In adsorption stud-
tion processes [3]. ies, to enhance the affinity and selectivity of the adsorbent for solute
The pseudo-first-order (PFO) [4] and pseudo-second-order (PSO) molecules, some surface techniques such as protonation and chemical
[5] equations are the most widely used kinetic models that describe modification are adopted [10]. Thus, most of the modified adsorbents
the adsorption rate based on adsorption capacity. The Vermeulen equa- often have a heterogeneous surface with different types of adsorption
tion, derived as an approximate solution of a diffusion-based Boyd sites. The Langmuir isotherm can only describe monolayer adsorption
equation, can predict the experimental data within the whole adsorp- on the energetically homogeneous solid surface [11], which limits its
tion time [6]. Although adsorption mechanisms cannot be directly application scope. Haerifar and Azizian demonstrate that the adsorption
assigned by fitting the three models [7], they have received extensive kinetics on energetically heterogeneous solid surfaces can be described
attention due to their simple mathematical forms and good fitting per- by a fractal-like approach [12]. Herein, we make attempts to introduce
formance. Furthermore, the three equations can also provide some im- the fractal-like concept into the Langmuir kinetic equation and thus
portant parameters such as the rate constant, equilibrium adsorption propose the fractional surface coverage dependence of the adsorption
capacity and intraparticle diffusion coefficient. However, they do not di- rate constant.
rectly predict the equilibrium time. Taking the importance of the equi- The innovation of this work is that the modified PFO, PSO and
librium time into consideration in adsorption studies, we make Vermeulen equations can be directly used to predict the equilibrium
time and that the fractal-like Langmuir isotherm can describe the ad-
⁎ Corresponding author. sorption of the adsorbates on the heterogeneous surfaces. Besides, in-
E-mail address: [email protected] (Z. Zhang). troduction of the parameter h into the Langmuir isotherm contributes

https://doi.org/10.1016/j.molliq.2018.07.113
0167-7322/© 2018 Published by Elsevier B.V.
 Q. Hu et al. / Journal of Molecular Liquids 268 (2018) 728–733 729

to improving fitting performance. The objectives of this study are: (i) to 2.2. Modification of Langmuir isotherm
predict the equilibrium time by modifying the PFO, PSO and Vermeulen
equations; and (ii) to extend the application scope of the Langmuir iso- The Langmuir kinetic equation, which is commonly used to describe
therm through introducing the fractal-like concept into the Langmuir monolayer adsorption on the energetically homogeneous solid surfaces,
kinetic equation. The goodness of fit for these models is evaluated by can be given as [15]:
the coefficient of determination (R2) and chi-squared analysis (χ2).
dθ
¼ ka C ð1−θÞ−kd θ ð10Þ
2. Theoretical analysis dt

2.1. Kinetic models where θ is the fractional surface coverage that is a dimensionless con-
stant; ka (L mg−1 min−1) and kd (min−1) are the adsorption and de-
In this study, kinetic models used to predict the equilibrium time in- sorption rate constants, respectively; and C (mg L−1) is the
clude the PFO, PSO and Vermeulen equations, which are expressed as [6, concentration of solute at time t.
13, 14]: The nonlinear form of the Langmuir isotherm at equilibrium is
expressed as [16]:
qt ¼ qe ð1− expð−k1 t ÞÞ ð1Þ
qmax K L C e

 qe ¼ ð11Þ
1 1 þ KL Ce
qt ¼ qe 1− ð2Þ
1 þ qe k2 t
where qmax (mg g−1) is the maximum adsorption capacity; KL (L mg−1)
vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
! is the Langmuir constant (KL = ka/kd); Ce (mg L−1) is the concentration
u
u Dπ2 of solute at equilibrium.
t
qt ¼ qe 1− exp − 2 t ð3Þ
r In order to facilitate the following analyses, another form of Eq. (11)
is written as:
where qt (mg g−1) and qe (mg g−1) are the amounts of solute adsorbed qe
per unit mass of the adsorbent at time t and at equilibrium, respectively; Ce ¼ ð12Þ
K L ðqmax −qe Þ
k1 (min−1) and k2 (g mg−1 min−1) are the PFO and PSO rate constants;
D (cm2 min−1) is the intraparticle diffusion coefficient; r (cm) is the ra-
Inspired by the fractal-like concept [3, 12], we argue that solute mol-
dius of the spherical adsorbent particles; and t (min) is the adsorption
ecules can be freely adsorbed on the bare solid surface (homogeneous
time.
or heterogeneous) in the initial phase of adsorption. As time goes on,
It is important to note that k1, qek2 and Dπ2/r2 have identical dimen-
solute molecules adsorbed on the solid surface may affect adsorption
sions (i.e. reciprocal time, min−1). According to the above three equa-
paths of the incoming solute molecules and thus adsorption of the in-
tions, the relationships between the fractional surface coverage (θ =
coming solute molecules on the solid surface is suffered from the frac-
qt/qe) at time τe and the parameters k1, qek2 and Dπ2/r2 are defined as:
tional surface coverage. Consequently, the adsorption rate constant is
a function of the fractional surface coverage. In this study, we introduce
1
k1 ¼ − ∙ ln ð1−θÞ ð4Þ the fractal-like concept or the fractional surface coverage dependence of
τe
the adsorption rate constant to the Langmuir kinetic equation, which is
1 θ expressed as:
qe k2 ¼ ∙ ð5Þ
τe ð1−θÞ
ka ¼ ka;0 θ−h ð0 ≤h≤1Þ ð13Þ

Dπ 2
1  
¼ − ∙ ln 1−θ2 ð6Þ where ka,0 (L mg−1 min−1) is the fractal-like adsorption rate constant, h
r 2 τe
is a constant parameter.
When adsorption process reaches equilibrium, substitution of
Substitution of Eq. (4) to Eq. (1), Eq. (5) to Eq. (2) and Eq. (6) to Eq. (13) into Eq. (10) leads to
Eq. (3) leads to


   ðhþ1Þ
qe
1 Ce ¼ ð14Þ
qt ¼ qe 1− exp ln ð1−θÞ t ð7Þ bqmax ðqmax −qe Þ
h
τe
0 1 where b (L mg−1) is the fractal-like isotherm constant (b = ka,0/kd). It is
B C
1 worthy to note that introducing the fractal-like concept into the Lang-
qt ¼ qe B
@1−  C A ð8Þ muir kinetic equation extends the application scope of the Langmuir
θ
1þ t isotherm. The fractal-like Langmuir isotherm can describe solute ad-
ð1−θÞτe
sorption on the heterogeneous surfaces according to the above analysis.
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

 ffi In addition, Eq. (14) includes the description of the equilibrium be-
1   
qt ¼ qe 1− exp ln 1−θ2 t ð9Þ tween the feed concentration C0 and the saturation level q0, which
τe may occur in the fixed-bed adsorption column. Dimensionless concen-
trations, Y = (qe/q0) and X = (Ce/C0), are introduced to Eq. (14),
One can readily see that the terms k1, qek2 and Dπ2/r2 are converted resulting in
to the equilibrium time (τe) under specific θ value and the mathematical
structure of the above equations has not changed through this transfor-
 −h 
mation. It should be emphasized that θ is an adjustable constant accord-
ðhþ1Þ C e 1 þ bC 0 qq0
ing to the actual requirements. Thus, selecting the appropriate θ value as qe max
¼
 −h  ð15Þ
the state of equilibrium will contribute to reducing the operating cost q0
C 0 1 þ bC e qqe
max
and optimizing the design of adsorption systems.
730 Q. Hu et al. / Journal of Molecular Liquids 268 (2018) 728–733

Herein, the favorable adsorption process can be also determined

when the right side of inequality 1/(1 + C0) b 1 in batch experiments.
Substitution of Eq. (16) to Eq. (15) leads to

X
Y ðhþ1Þ ¼ ð18Þ
RH þ ð1−RH ÞX

3. Materials and methods

3.1. Adsorbent preparation

The adsorbent used in this study was prepared by loading

polyaniline (PAN) on powdery activated carbon (AC). The PAN/AC com-
posite was synthesized by in-situ chemical oxidative polymerization in
the presence of HCl as the dopant. Specifically, 8.1 g of FeCl3·6H2O was
dissolved into a 250 mL of conical flask containing 150 mL of 0.1 M HCl
Fig. 1. Schematic diagram of the equilibrium time for different rate constants.
at ambient temperature, and then 1.5 g of AC and 1.5 mL of aniline in
turn were added to the mixed solution for 12 h under magnetic stirring.
The precipitate was filtrated using a vacuum filter and thoroughly
Similarly, according to the procedure proposed by Hall et al. [17], a washed with deionized water. The filter residue was dried in an oven
dimensionless constant RH (i.e. the separation factor or equilibrium pa- at 50 °C until the mass became constant.
rameter) can be used to reflect the essential characteristics of the
fractal-like Langmuir isotherm, which is defined as: 3.2. Kinetic and isotherm experiments
1
RH ¼  −h ð16Þ Kinetic experiments: 50 mL of nitrate solution (25, 50 and
1 þ bC 0 q0
qmax 75 mg L−1) was added to a series of conical flasks containing 0.5 g of ad-
sorbent, respectively, which were sealed and agitated in a thermostatic
shaker (80 rpm and 25 °C). A required amount of sample was taken
Although the saturation level q0 is expediently obtained for the influ-
from the conical flasks at preset time.
ent concentration C0 in the fix-bed experiment, q0 is difficult to be deter-
Isotherm experiments: 10 mL of nitrate solution (25–200 mg L−1)
mined in batch experiments. We note that:
were added to a series of conical flasks containing 0.1 g of adsorbent
and operated at 25, 35 and 45 °C according to the above conditions. A re-
1 quired amount of sample was taken from the conical flasks at
RH ≤ ð17Þ
1 þ bC 0 equilibrium.

Fig. 2. The plot of qt versus t for nitrate adsorption on the PAN/AC composite under different concentrations: (a) Equilibrium time (θ(t) = 0.95); (b) Pseudo-first-order equation;
(c) Pseudo-second-order equation; (d) Vermeulen equation.
 Q. Hu et al. / Journal of Molecular Liquids 268 (2018) 728–733 731

Table 1 evaluated by the coefficient of determination (R2) [7] and chi-squared

Kinetic parameters obtained from the PFO, PSO and Vermeulen equations. value (χ2) [19], which can be expressed as:
C0 (mg L−1) PFO R2 χ2
X
n
qe (mg g−1) k1 (min−1) τe (min) ðycal −yÞ2
25 2.21 0.289 10.4 0.991 4.23 × 10−3 2
R ¼ i¼1
ð19Þ
50 4.09 0.233 12.8 0.994 9.68 × 10−3 X
n n 
X 2
75 5.43 0.220 13.6 0.992 2.30 × 10−2 ðycal −yÞ2 þ ycal −yexp
i¼1 i¼1
C0 (mg L−1) PSO R2 χ2
 2
qe (mg g−1) k2 (g mg−1 min−1) τe (min) yexp −ycal
X
n
2
25 2.30 0.283 29.2 0.999 6.67 × 10−5 χ ¼ ð20Þ
ycal
50 4.27 0.114 38.9 0.998 3.14 × 10−3 i¼1
75 5.69 0.078 42.8 0.999 3.61 × 10−3
where yexp (mg g−1) is the observed value; ycal (mg g−1) is the pre-
C0 (mg L−1) Vermeulen R2 χ2
dicted value;y (mg g−1) is the average value of yexp; and n is the number
qe (mg g−1) τe (min) of data points. A larger R2 or smaller χ2 value denotes good fitting
25 2.23 13.5 0.996 1.87 × 10−3 performance.
50 4.13 16.9 0.999 1.28 × 10−3
75 5.49 18.3 0.999 2.96 × 10−3 4. Results and discussion

4.1. Equilibrium time

All of the samples were filtrated using 0.45 μm of filter membrane. The schematic diagram of the equilibrium time for different rate
The concentration of nitrate (as NO−3 -N) in the filtrate was measured constants is depicted in Fig. 1. It is obvious that the increase in the frac-
using a UV-VIS spectrophotometer (UV–1800, Shimadzu, Japan) ac- tional surface coverage (θ(t) = qt/qe) is extremely fast in the initial
cording to the standard method (HJ/T 346–2007). stage, followed by a slow increase and finally the adsorption process ap-
proaches dynamic equilibrium. The fast initial adsorption rate mainly
3.3. Error analysis depends on the affinity between the adsorbent and the adsorbates.
The subsequent slower adsorption rate is obtained as a consequence
Considering that linearization of nonlinear equations in regression of the decrease in available adsorption sites. Due to the asymptotic
analyses may involve arbitrary data exclusion to improve linearity and form of the kinetic curves with time, the equilibrium time cannot be de-
thus create various errors [18]. In order to avoid these errors caused termined exactly. Considering that the breakthrough curve also has
by the linear regression, in this study, all experimental data are fitted analogous behavior in the fixed-bed column and that the time required
by the nonlinear curve fit using Origin 9 software (OriginLab Corp., to reach 0.95 of the ratio of the effluent to influent concentrations is de-
USA). The goodness of fit and the validity of various models are termined as the saturation time [9]. Similarly, the time at θ(t) = 0.95 is

Fig. 3. Isotherm curves for nitrate adsorption on the PAN/AC composite at different temperature: (a) 25 °C; (b) 35 °C and (c) 45 °C, (d) Effect of the separation factor RH on the shape of
isotherm curves.
732 Q. Hu et al. / Journal of Molecular Liquids 268 (2018) 728–733

Table 2 time required to reach the equilibrium state is longer at higher concen-
Isotherm parameters obtained from the Langmuir and fractal-like Langmuir isotherms. tration. In this study, the equilibrium time predicted by the PSO
Temperature (°C) Langmuir R2 χ2 equation is 29.2, 38.9 and 42.8 min respectively at the initial concentra-
qmax (mg g−1) KL (L mg−1)
tions of 25, 50 and 75 mg L−1. This result agrees well with the experi-
mental values. Therefore, the PSO equation can more accurately
25 8.90 6.47 × 10−2 0.996 4.54
predict the equilibrium time for nitrate adsorption on the PAN/AC
35 8.61 5.98 × 10−2 0.993 9.43
45 8.26 5.62 × 10−2 0.993 9.96 composite.

Temperature (°C) Fractal-like Langmuir R2 χ2 4.2. Fractal-like Langmuir equation

qmax (mg g−1) b (L mg−1) h
The isotherm curves at different temperatures are depicted in
25 9.50 3.82 × 10−2 0.74 0.999 0.37
35 9.12 3.72 × 10−2 0.68 0.996 6.20 Fig. 3a–c. It is evident that the Langmuir isotherm curves, to some ex-
45 8.56 3.57 × 10−2 0.42 0.994 9.43 tent, deviate from the experimental values whereas the fitting curves
provided by the fractal-like Langmuir equation almost completely pass
through the data points. As shown in Table 2, the fractal-like Langmuir
therefore defined as the equilibrium time in this study. The rate con- equation has lower χ2 and higher R2 values at all temperatures. As a re-
stant is a time-scaling factor and thereby a larger rate constant results sult, the fractal-like Langmuir equation can better describe nitrate ad-
in a shorter time required to reach equilibrium state [7]. According to sorption on the PAN/AC composite. Table 3 summarizes the maximum
Eqs. (7), (8) and (9), the equilibrium time is calculated from the follow- adsorption capacity of nitrate (as NO− 3 -N) obtained from different acti-
ing formulae. vated carbon. Although these values (qmax) are attained under different

 experimental conditions, they can be useful in measuring the adsorp-
3:0 tion capacity of each adsorbent for nitrate. In this study, the PAN/AC
qt ¼ qe 1− exp − t ð21Þ
τe composite shows a moderate adsorption capacity (9.50 mg g−1) com-
0 1 pared with other adsorbents. Therefore, the PAN/AC composite has a
potential of practical application for nitrate removal. In addition, the
B 1 C
qt ¼ qe B
@1−
C ð22Þ Langmuir constant KL reduces with the increase in temperature, indicat-
19 A ing that the affinity of the PAN/AC composite for nitrate ions decreases
1þ t
τe [29].
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

 The effects of the separation factor RH on the isotherm curves are il-
2:3 lustrated in Fig. 3d according to Eq. (18). It is obvious that Y(h+1) identi-
qt ¼ qe 1− exp − t ð23Þ
τe cally equals to unity for different X when RH = 0 (b → ∞), suggesting
that this type of adsorption can be attributed to an irreversible process.
where τe (min) represents the equilibrium time. The shape of isotherm curves is convex (0 b RH b 1), showing a favorable
The plot of qt versus t for nitrate adsorption on the PAN/AC compos- adsorption. It should be noted that isotherm curves are closer to vertical
ite is illustrated in Fig. 2. It is observed that the kinetic curves share sim- axis when the RH value is smaller, indicating a more favorable adsorp-
ilar trend at different initial nitrate concentrations and that the rate of tion. The value of Y(h+1) increases linearly with the increase in X when
nitrate removal is very fast (Fig. 2a). The high removal rate indicates RH = 1. The isotherm curve is concave biased toward horizontal axis
that the surface of the adsorbent has a high density of active sites for ni- when RH N 1, indicating the adsorption is unfavorable. These results
trate adsorption [20]. In this study, the amount of nitrate uptake at equi- are consistent with the separation factor of Langmuir eq. [17]. As a re-
librium is 2.27, 4.20 and 5.52 mg g−1 at the initial concentrations of 25, sult, the separation factor RH can be used to describe the feasibility of ad-
50 and 75 mg L−1, respectively. Moreover, when the time at θ(t) = 0.95 sorption process and characterize the degree to which the adsorption
is used as the equilibrium state, the corresponding equilibrium time ob- reaction proceeds. It should be emphasized that these results apply
tained are 26.2, 37.3 and 39.1 min, respectively. One can readily see that only to cases where the adsorption equilibrium follows the fractal-like
the fitting curves provided by the PFO, PSO and Vermeulen equations Langmuir equation. In this study, all of the RH values at different temper-
can well describe the kinetic data (Fig. 2b–d). atures and initial concentrations are between 0.10 and 0.54, showing fa-
It can be seen from Table 1 that the three equations have good fitting vorable adsorption for nitrate removal by the PAN/AC composite.
performance with high coefficient of determination (R2 N 0.99) and low The effect of the parameter b on the isotherm curves is depicted in
chi-square value (χ2 b 0.01) and that the experimental values are Fig. 4. It is observed that the isotherm curves are biased toward the hor-
mostly consistent with the predicted values with respect to the amount izontal axis with the increase in b and that the larger b value suggests a
of nitrate uptake. The magnitude of the rate constant is usually depen- more favorable adsorption. Therefore, analogous to the Langmuir con-
dent on the applied initial concentration and the rate constant decreases stant KL [30], the parameter b can represent the affinity of adsorption
with the increase in initial concentration [21], which is in agreement sites for solute molecules and a measure of how strong solute molecule
with the finding of this study. That is, the rate constants k1 and k2 de- is attracted onto the adsorbent surface. Besides, the parameter b reduces
crease with the increase in the initial concentration. Consequently, the with the increase in temperature, indicating that the decrease in

Table 3
Comparison of nitrate maximum adsorption capacity (as NO−
3 -N) obtained in this work with previously reported activated carbons.

Adsorbents qmax (mg g−1) Dosage (g L−1) Concentration (mg L−1) Contact time (h) Temperature (°C) Rotate speed (rpm) References

Origin GAC 14.25 2.5 25.1–376 4 23 ± 2 150 [22]

F400-DA 9.30 2 100–800 Overnight 25 100 [23]
PC-PU 6.30 10 1.4–11.2 24 25 300 [24]
Ox-9OG 2.94 2 – 24 25 100 [25]
Zn(II)/GAC 10.26 10 5–200 2 25 ± 2 – [26]
PJAC 10.99 1 1–100 1 30 100 [27]
CAG 9.10 4 – 1.5 25 100 [28]
PAN/AC 9.50 10 25–200 2 25 80 This work
 Q. Hu et al. / Journal of Molecular Liquids 268 (2018) 728–733 733

[4] S. Lagergren, About the theory of so-called adsorption of soluble substances, K. Sven.
Vetenskapsakad. Handl. 24 (1898) 1–39.
[5] G. Blanchard, M. Maunaye, G. Martin, Removal of heavy metals from waters by
means of natural zeolites, Water Res. (1984) 1501–1507.
[6] T. Vermeulen, Theory for irreversible and constant-pattern solid diffusion, Ind. Eng.
Chem. 45 (1953) 1664–1670.
[7] K.L. Tan, B.H. Hameed, Insight into the adsorption kinetics models for the removal of
contaminants from aqueous solutions, J. Taiwan Inst. Chem. Eng. 74 (2017) 25–48.
[8] I. Langmuir, The adsorption of gases on plane surfaces of glass, mica and platinum, J.
Am. Chem. Soc. 40 (1918) 1361–1403.
[9] E. Worch, Adsorption Technology in Water Treatment: Fundamentals, Processes and
Modelling, Walter de Gruyter, Berlin/Boston, 2012.
[10] P. Loganathan, S. Vigneswaran, J. Kandasamy, Enhanced removal of nitrate from
water using surface modification of adsorbents – a review, J. Environ. Manag. 131
(2013) 363–374.
[11] D.M. Ruthven, Principles of Adsorption and Adsorption Processes, John Wiley &
Sons, New York, 1984.
[12] M. Haerifar, S. Azizian, Fractal-like kinetics for adsorption on heterogeneous solid
surfaces, J. Phys Chem. C 118 (2014) 1129–1134.
[13] Q. Hu, N. Chen, C. Feng, J. Zhang, W. Hu, L. Lv, Kinetic studies of nitrate removal from
aqueous solution using granular chitosan-Fe(III) complex, Water Sci. Technol. 73
(2015) 1211–1220.
Fig. 4. The effect of the parameter on b the isotherm curve (qmax = 9.50 mg g−1, h = 0.74). [14] P. Ganesan, R. Kamaraj, G. Sozhan, S. Vasudevan, Oxidized multiwalled carbon nano-
tubes as adsorbent for the removal of manganese from aqueous solution, Environ.
Sci. Pollut. Res. 20 (2013) 987–996.
temperature will facilitate adsorption, which is consistent with the fact [15] S. Azizian, Kinetic models of sorption: a theoretical analysis, J. Colloid Interface Sci.
276 (2004) 47–52.
that nitrate adsorption on the PAN/AC composite has a higher adsorp- [16] M. Naushad, S. Vasudevan, G. Sharma, A. Kumar, Z.A. ALOthman, Adsorption kinet-
tion capacity at low temperature. Understanding the effect of the pa- ics, isotherms, and thermodynamic studies for Hg2+ adsorption from aqueous me-
rameter b on the isotherm curve sheds light on the affinity of dium using alizarin red-S-loaded amberlite IRA-400 resin, Desalin. Water Treat. 57
(2016) 18551–18559.
adsorbent for solute molecules under different experimental conditions. [17] K.R. Hall, L.C. Eagleton, A. Acrivos, T. Vermeulen, Pore- and solid-diffusion kinetics in
fixed-bed adsorption under constant-pattern conditions, Ind. Eng. Chem. Fundam. 5
5. Conclusions (1966) 212–223.
[18] S. Wongcharee, V. Aravinthan, L. Erdei, W. Sanongraj, Use of macadamia nut shell
residues as magnetic nanosorbents, Int. Biodeterior. Biodegrad. 124 (2017)
The equilibrium time under specific conditions could be easily pre- 276–287.
dicted by the modified PFO, PSO and Vermeulen equations by adjusting [19] R. Kamaraj, A. Pandiarajan, M.R. Gandhi, A. Shibayama, S. Vasudevan, Eco–friendly
the θ value. In this study, the equilibrium time obtained from the PSO and easily prepared graphenenanosheets for safe drinking water: Removal of
chlorophenoxyacetic acid herbicides, ChemistrySelect 2 (2017) 342–355.
equation was 26.2, 37.3 and 39.1 min at the initial concentrations of [20] N. Li, R. Bai, C. Liu, Enhanced and selective adsorption of mercury ions on chitosan
25, 50 and 75 mg L−1, respectively, which agreed well with the experi- beads grafted with polyacrylamide via surface-initiated atom transfer radical poly-
mental values. The fractal-like Langmuir isotherm could well describe merization, Langmuir 21 (2005) 11780–11787.
[21] W. Plazinski, W. Rudzinski, A. Plazinska, Theoretical models of sorption kinetics in-
adsorption of nitrate on the PAN/AC composite. The magnitude of the cluding a surface reaction mechanism: a review, Adv. Colloid Interf. Sci. 152 (2009)
separation factor RH reflected the feasibility of adsorption process and 2–13.
characterized the degree to which the adsorption reaction proceeded. [22] D.W. Cho, C.M. Chon, Y. Kim, B.H. Jeon, F.W. Schwartz, E.S. Lee, H. Song, Adsorption
of nitrate and Cr(VI) by cationic polymer-modified granular activated carbon, Chem.
In this study, all of the RH values ranged from 0.10 to 0.54, showing fa- Eng. J. 175 (2011) 298–305.
vorable adsorption for adsorption of nitrate on the PAN/AC composite. [23] T. Iida, Y. Amano, M. Machida, F. Imazeki, Effect of surface property of activated car-
The larger the fractal-like Langmuir constant b value represented the bon on adsorption of nitrate ion, Chem. Pharm. Bull. 61 (2013) 1173–1177.
[24] G.V. Nunell, M.E. Fernandez, P.R. Bonelli, A.L. Cukierman, Nitrate uptake improve-
better affinity of adsorption sites.
ment by modified activated carbons developed from two species of pine cones, J.
Colloid Interface Sci. 440 (2015) 102–108.
Acknowledgments [25] K. Ota, Y. Amano, M. Aikawa, M. Machida, Removal of nitrate ions from water by ac-
tivated carbons (ACs)-Influence of surface chemistry of ACs and coexisting chloride
and sulfate ions, Appl. Surf. Sci. 276 (2013) 838–842.
The authors sincerely express their appreciation to Dr. Nan Zhang [26] A. Bhatnagar, M. Ji, Y.H. Choi, W. Jung, S.H. Lee, S.J. Kim, G. Lee, H. Suk, H.S. Kim, B.
(University of Tsukuba). The first author acknowledges the support of Min, S.H. Kim, B.H. Jeon, J.W. Kang, Removal of nitrate from water by adsorption
the China Scholarship Council (No. 201606400063). onto zinc chloride treated activated carbon, Sep. Sci. Technol. 43 (2008) 886–907.
[27] S.V. Manjunath, M. Kumar, Evaluation of single-component and multi-component
adsorption of metronidazole, phosphate and nitrate on activated carbon from
References Prosopıs julıflora, Chem. Eng. J. 346 (2018) 525–534.
[28] R.M. Paixão, I.M. Reck, R. Bergamasco, M.F. Vieira, A.M.S. Vieira, Activated carbon of
[1] A. Bhatnagar, M. Sillanpää, A review of emerging adsorbents for nitrate removal Babassu coconut impregnated with copper nanoparticles by green synthesis for the
from water, Chem. Eng. J. 168 (2011) 493–504. removal of nitrate in aqueous solution, Environ. Technol. 39 (2018) 1994–2003.
[2] W. Rudzinski, W. Plazinski, Studies of the kinetics of solute adsorption at solid/solu- [29] P. Chowdhury, C. Bikkina, S. Gumma, Gas adsorption properties of the chromium-
tion interfaces: on the possibility of distinguishing between the diffusional and the based metal organic framework MIL-101, J. Phys. Chem. C 113 (2009) 6616–6621.
surface reaction kinetic models by studying the pseudo-first-order kinetics, J. Phys. [30] D.D. Do, Adsorption Analysis: Equilibria and Kinetics, Imperial college press, London,
Chem. C 111 (2007) 15100–15110. 1998.
[3] M. Haerifar, S. Azizian, Fractal-like adsorption kinetics at the solid/solution interface,
J. Phys. Chem. C 116 (2012) 13111–13119.