Journal of Environmental Chemical Engineering 4 (2016) 2671–2682

Contents lists available at ScienceDirect

Journal of Environmental Chemical Engineering

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

Thermodynamic parameters of cadmium adsorption onto orange peel

calculated from various methods: A comparison study
Hai Nguyen Trana,b,* , Sheng-Jie Youb,* , Huan-Ping Chaob,*
a
Department of Civil Engineering, Chung Yuan Christian University, Chungli 320, Taiwan
b
Department of Environmental Engineering, Chung Yuan Christian University, Chungli 320, Taiwan

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

Article history:
Received 3 March 2016 Thermodynamic adsorption investigation plays a key role in estimating adsorptive mechanisms (i.e.,
Received in revised form 6 May 2016 physical or chemical). Accuracy estimation of the thermodynamic parameters is directly dependent on
Accepted 7 May 2016 the equilibrium constant between two phases (KC; dimensionless). In this study, the thermodynamic
Available online 9 May 2016 parameters were calculated from the KC constants derived from the adsorption isotherm constants (i.e.,
Langmuir, Freundlich, and Henry) and partition coefficient, with and without dimensionality
Keywords: consideration. Results showed that the optimal selection of KC is strongly dependent on: the adsorption
Equilibrium constant characteristics (i.e., Henry, Freundlich, and Langmuir) where equilibrium data are actually located; and
Thermodynamic parameters
the correlation coefficient of the van’t Hoff equation. Both the Langmuir and Freundlich constants
Physical adsorption
(dimensionless) are appropriate to calculate the thermodynamic parameters. The Langmuir constants
Electrostatic attraction
Cadmium from its four linear forms can be applied to calculate the thermodynamic parameters without significant
Orange peel difference. The Cd(II) biosorption process onto the orange peel (OP) occurred spontaneously (
DG ), in
an exothermic nature (
DH ), and with increased randomness (+DS ). The biosorption process involved
physical adsorption with minimal activation energy (Ea) and adsorption energy (E). The biosorption
phenomenon reached fast equilibrium and reversibility. The negatively charged carboxylic groups
(–COO
) on the OP’s surface play an important role (approximately 90%) in Cd2+ biosorption through
electrostatic attractions (out-sphere complexes).
ã 2016 Elsevier Ltd. All rights reserved.

1. Introduction (DH ), and the entropy change (DS ). It is noticeable that the
estimation of DG , DH , and DS is directly affected by the
Isotherms, kinetics, and thermodynamics studies play an thermodynamic equilibrium constant (KC; dimensionless). In
important role in thoroughly understanding the adsorption adsorption thermodynamics, the equilibrium constant KC can
process of organic and inorganic compounds in water solutions. derive from the constants of various adsorption isotherms (i.e.,
In the three studies, the thermodynamic adsorption is essential in Langmuir, Freundlich, Brunauer-Emmett-Teller (BET), Henry,
consideration of the types and mechanisms of the adsorption Frumkin, Flory–Huggins, and Redlich–Petersone) or the partition
process under variations of the solution temperature. The coefficient. The different estimation approaches might lead to
thermodynamics can effectively exhibit adsorption mechanisms great variations in the thermodynamic parameters; therefore, the
through a series of the adsorptive equilibrium experiments under most appropriate approaches should be recommended.
different temperatures and various initial adsorbate concentra- Several researchers presented the thermodynamic parameters
tions as well as fixed optimal conditions (i.e., pH of the solution, the of adsorption phenomena by employing the equilibrium constants
adsorbent’s particle size, ionic strength, and solid/liquid ratio). derived from the constants of various adsorption isotherms or the
When the adsorption phenomenon reaches equilibrium, the partition coefficient. For example, Lyubchik et al. [1] compared the
thermodynamic parameters can be calculated. These parameters thermodynamic parameters of the adsorption process of the
comprise the Gibbs energy change (DG ), the enthalpy change metals from liquid phase on activated carbons, estimated from
the various equilibrium constants derived from the constants of
Henry, Langmuir, Freundlich, and BET isotherms. Roth et al. [2]
applied the constants of Langmuir isotherm, distribution (the
* Corresponding author.
Henry’s constant), and partition for computing the thermodynam-
E-mail addresses: [email protected], [email protected]
(H.N. Tran), [email protected] (S.-J. You), [email protected] (H.-P. Chao). ic parameters of cadmium accumulation process onto different

http://dx.doi.org/10.1016/j.jece.2016.05.009
2213-3437/ã 2016 Elsevier Ltd. All rights reserved.
2672 H.N. Tran et al. / Journal of Environmental Chemical Engineering 4 (2016) 2671–2682

granulometric soil fractions. In addition, the comparison of the the point of zero charge (pHPZC), are shown in Figs. S1 and S2,
thermodynamic parameters of ammonium removal from an respectively. The results of FTIR indicated that the OP particles
aqueous solution by using clinoptilolite was also investigated by possess abundant carboxylic (–COOH) and phenolic (–OH) groups
Tosun [3]. These parameters were calculated based on the on their surface. The morphology and superficial element
constants of the Langmuir and Tempkin models, and partition. composition of OP before and after adsorption were also examined
Unfortunately, the important role of dimensionality of the using scanning electron microscopy (SEM; JSM-7600F, JEOL) at
thermodynamic equilibrium constant KC is often neglected. As a 15 kV and energy dispersive X-ray spectroscopy (EDX), respective-
result, both sign and magnitude of the thermodynamic param- ly.
eters—DG , DH , and DS —usually indicate a confused represen-
tation. Noticeably, the Langmuir constant has been widely applied 2.2. Batch adsorption experiment
to compute of the thermodynamic parameters in the literature.
However, there is a limitation in the comparison of the All Cd solutions were diluted from stock solutions of Cd
thermodynamic parameters derived from four linearized forms (1,000 mg/L), which were prepared by dissolving cadmium nitrate
of the Langmuir model. Therefore, it is necessary to clarify the tetrahydrate (Cd(NO3)24H2O; Alfa Aesar) in DD water.
difference in estimation methods. Experimental procedure followed our previous research [4].
The objectives of this study are to (1) compare the thermody- Adsorption isotherm studies were conducted with five initial Cd2+
namic parameters calculated from various equilibrium constants concentrations (i.e., 50, 100, 200, 400 and 600 mg/L) and in fixed
(KC) that were derived from the isotherm constants of Langmuir, other environmental parameters (i.e., an initial pH value of
Freundlich, Henry, and the partition coefficient with and without 7.0  0.2 and a solid/liquid ratio of 0.1 g/50 mL). The flask
considerations of the KC dimensionality; and (2) consider the containing the Cd–OP mixture was placed on a stirrer with
difference among the thermodynamic parameters calculated from hotplate (Corning PC-420D, Fisher Scientific) for 24 h with a
the four linearized forms of the Langmuir model. In addition, rotating speed of 600 rpm under predetermined temperatures
dominate adsorption mechanisms were also discussed based on 20–50  C (Fig. 1). After agitating 24 h, the mixture of solid and
the thermodynamic parameters, adsorption rates, activation liquid was separated using glass fiber filters. The liquid was filtered
energy, adsorption energy, and reversibility. through a 0.45-mm syringe filter. The Cd concentration in the
solutions was determined using atomic absorption spectrometry
2. Materials and methods (Avanta/AAS, GBC). The amount of Cd uptake at equilibrium, qe
(mg/g), was calculated using the mass balance equation, as follows:
2.1. Biosorbent preparation
ðC o
C e ÞV
qe ¼ ð1Þ
m
Orange peel (OP) was collected from a local market near Chung
Yuan Christian University, Taiwan. The collected OP was then where C0 (mg/L) is the initial Cd(II) concentration in the solution, Ce
washed with water at least three times and with deionized (mg/L) is the Cd(II) concentration in the solution at equilibrium, m
distilled (DD) water to remove any adhering dirt and impurity. (g) is the dried mass of used OP, and V (L) is the volume of the Cd(II)
They were then placed in a drying oven at 85  C for 24 h. The dried solution.
OP was sieved into particles ranging from 0.106 to 0.25 mm, which
was used as a green biosorbent without any further treatment. The
basic characteristics of orange peel (i.e., textural, physicochemical,
ultimate, and proximate) are shown in Table 1. The surface
chemistry of OP identified by using Fourier transform infrared
spectroscopy (FITR; FT/IR-6600 JASCO) and the pH value of OP at

Table 1
Basic characteristics of pristine orange peel.

Orange peel
1. Textural parameters
BET surface area (m2/g) 19.01
Langmuir surface area (m2/g) 23.83
Total pore volume (cm3/g) 0.021

2. Physicochemical properties
Hardness (%) 39.7  1.15
Bulk density (g/cm3) 0.61  0.05
pH1:20 4.69  0.18
pHpzc 5.18  0.07

3. Ultimate analysis
C (wt%) 44.34
H (wt%) 5.91
N (wt%) 1.14
O (wt%)a 48.61

4. Proximate analysis
Moisture (wt%) 8.54  1.98
Total ash (wt%) 3.49  0.99
Volatile (wt%) 74.34  2.95
Fixed carbon (wt%)a 13.62  2.18
Fig. 1. Experimental set-up of adsorption process of Cd2+ ions onto orange peel.
a
Calculated by difference.
 H.N. Tran et al. / Journal of Environmental Chemical Engineering 4 (2016) 2671–2682 2673

The reversibility of Cd(II) adsorption on OP was studied with the the decreased number of active sites responsible for adsorption
desorption experiments by using (1) an electrolyte solution 1 M when the increased temperature.
KNO3 to inhibit the electrostatic adsorption between Cd2+ and the To identify the best-fit isotherm model, linear Chi-square values
OP’s surface and (2) a complexing agent 0.05 M EDTA (pH 6.0) to (x2) were examined along with linear regressions (R2). The
replace Cd2+ cations complexed by the OP’s acidic functional statistics of the Chi-squared test is basically the sum of the
groups. The experimental procedure followed the works published squared errors of the differences between experimental data and
elsewhere [5,6]. data obtained through calculations by using the models. Each
The chemicals used in this study were of analytical reagent squared difference is then divided by the corresponding data
grade. All experiments were conducted in triplicate, and blank obtained through calculations by using the models. If qe,iso values
samples without OP were also conducted together with adsorption obtained using a model are similar to the experimental values (qe),
experiments. The difference between the maximum monolayer the x2 value will be very small and close to zero. High x2 implies a
adsorption capacities calculated from four linear forms of the high bias between the experiment and model. Therefore, analyzing
Langmuir model at various temperatures was statistically analyzed the data set of the Chi-squared test possibly confirms the best-fit
using the Statgraphic Plus 3.0 statistical program. Differences were isotherm for the adsorption system. The mathematical statement
considered to be significant at p < 0.01 [4]. of Chi-squared test is possibly explained as below:

X ðqe
qe;iso Þ2
3. Results and discussion x2 ¼ ð2Þ
qe;iso
3.1. Adsorption isotherm
where qe (mg/g) is obtained from Eq. (1), qe,iso (mg/g) is the amount
of Cd uptake obtained from isotherms.
“Adsorption isotherm” is generally used to describe the
The Langmuir theoretical equation (1918) was developed by
relationship between concentration of adsorbate in solution
assuming the following: (1) a fixed number of accessible sites are
(liquid phase) and adsorbent (solid phase) at a constant tempera-
available on the adsorbent surface, and all sites have the same
ture under given conditions (i.e., pH and ionic strength, fixed mass
energy (the homogeneous surface); (2) adsorption is reversible; (3)
and particle size of adsorbent). Kumar [7] suggested that
once an adsorbate occupies a site, no further adsorption can occur
presenting a plot of qe versus Ce of a complete adsorption isotherm
on that site; and (4) no interaction occurs between adsorbed
in adsorption studies played a key role in identifying the region
species. The non-linearized form of Langmuir model is described in
(e.g., Henry, Freundlich, Langmuir, and BET) where the experi-
Eq. (3) and its four linearized forms are shown in Eqs. (4)–(7).
mental equilibrium data are actually located. The adsorption
isotherms of cadmium by OP at various temperatures are Q o max K L C e
qe ¼ ð3Þ
presented in Fig. 2. The region, where the experimental data of 1 þ K L Ce
adsorption equilibrium are located, might be the Langmuir region
Langmuir linearization (Type 1):
characterized by a saturation at high concentrations. Therefore, in
 
the this study, it is feasible to apply the Langmuir constant (KL) as Ce 1 1
¼ o Ce þ o ð4Þ
the thermodynamic equilibrium constant (KC) in estimating the qe Q max Q max K L
thermodynamic parameters. Moreover, Fig. 2 shows that the
temperature remarkably affected the metal-binding capacity. The Lineweaver-Burk linearization of Langmuir (Type 2):
 
amount of Cd2+ adsorbed by OP decreased with the increased 1 1 1 1
temperatures, suggesting that a lower temperature favored the Cd ¼ þ ð5Þ
qe Q max K L C e Q omax
o

removal from the aqueous solution. This tendency might be due to

Eadie-Hoffsie linearization of Langmuir (Type 3):
 

1 qe
qe ¼ þ Q omax ð6Þ
K L Ce
Scatchard linearization of Langmuir (Type 4):
qe
¼
K L qe þ Q omax K L ð7Þ
Ce
where Qomax (mg/g) is the maximum saturated monolayer
adsorption capacity of OP under the given conditions, KL (L/mg)
is the Langmuir constant related to the affinity between an
adsorbent and adsorbate, and Ce and qe are obtained from Eq. (1).
The Langmuir parameters at various temperatures calculated
from its four types of linearization are listed in Table 2. In a given
dried mass of OP, the concentration increase of Cd2+ ions can cause
the occupation of binding sites on the OP’s surface to reach a
saturation in the surface layer. After the saturation is reached, no
more binding sites are available on the OP’s surface and the
maximum equilibrium adsorptive capacity can be obtained (Fig. 2).
Obviously, the values of Qomax and KL are dependent on the
linearized forms of the Langmuir model. The different linear
equations can obtain slightly different values of the Langmuir
parameters. This is because the different weighting of isotherm
sections is attributed to the transformation of the variation.
Fig. 2. Equilibrium adsorption isotherms at various temperatures for Cd(II)–OP
However, this difference is negligible and statistically insignificant
system (mean values  standard deviation).
(p > 0.01). Therefore, the Langmuir constants (KL) from the four
2674 H.N. Tran et al. / Journal of Environmental Chemical Engineering 4 (2016) 2671–2682

Table 2
Freundlich and Langmuir parameters for Cd(II) adsorption onto orange peel at various temperatures.

T Freundlich parameters Langmuir parameters

(K)
KF 1/n R2 x2 Q max KL R2 x2
(mg/g)(L/mg)1/n (mg/g) (L/mg)
293 17.3 (0.42) 0.212 0.94 2.17 Type 1 59.5 (0.53) 0.076 (8.57) 0.99 6.57
303 16.3 (0.39) 0.211 0.95 1.64 56.5 (0.50) 0.070 (7.83) 0.99 4.54
313 15.0 (0.37) 0.215 0.94 1.80 53.8 (0.48) 0.063 (7.02) 0.99 3.72
323 14.9 (0.35) 0.205 0.95 1.08 51.6 (0.46) 0.056 (6.28) 0.99 4.90
293 Type 2 51.8 (0.46) 0.182 (20.37) 0.88 4.73
303 50.3 (0.45) 0.149 (16.80) 0.91 3.09
313 47.9 (0.43) 0.125 (14.04) 0.90 2.94
323 45.3 (0.40) 0.132 (14.84) 0.89 2.70
293 Type 3 54.3 (0.48) 0.162 (18.18) 0.76 3.94
303 52.0 (0.46) 0.135 (15.15) 0.80 2.67
313 49.7 (0.44) 0.113 (12.66) 0.80 2.49
323 46.9 (0.42) 0.119 (13.40) 0.79 2.32
293 Type 4 57.2 (0.51) 0.122 (13.74) 0.76 3.80
303 54.2 (0.48) 0.108 (12.18) 0.80 2.54
313 51.9 (0.46) 0.090 (10.12) 0.80 2.36
323 48.9 (0.44) 0.094 (10.60) 0.79 2.24
293 Standard deviation 3.1 (0.028) 0.046 (5.20) – –
303 2.5 (0.023) 0.034 (3.79) – –
313 2.4 (0.021) 0.027 (2.99) – –
323 2.4 (0.022) 0.031 (3.48) – –

In the Langmuir parameters, values in parentheses presenting the unit: mmol/g for Q max and L/mmol for KL; in the Freundlich parameters, values in parentheses presenting
the unit (mmol/g)(L/mmol)1/n for the KF.

types of linearization might be appropriate for calculation of the which confirms that the adsorption reaction was preferred at a low
thermodynamic parameters. Table 2 shows that the Qomax and KL temperature and exothermic in nature. According to high
values decreased with the temperature rise, which means that the correlation coefficients (R2) and low linear Chi-square values
Cd2+ adsorption phenomenon on OP is governed by an exothermic (x2), the Freundlich constant (KF) might be appropriate for
process. This phenomenon is one of the typical characteristics of calculating the thermodynamic parameters.
physical adsorption, and weak electrostatic forces between Besides the Langmuir and Freundlich, the Dubinin–Radushke-
negative charge sites, available on the OP’s surface and cadmium vich equation was also employed for the estimation of adsorption
cationic species in the solution, might be responsible for the energy (E). In essence, the E magnitudes can provide useful
adsorption behavior. information on adsorption mechanisms, such as physical adsorp-
The earliest empirical equation proposed by Freundlich (1932), tion (E < 8 KJ/mol), ion exchange (8 < E < 16 KJ/mol), and chemical
is useful for describing equilibrium data and adsorption character- adsorption (E > 16 KJ/mol) (Table S2). The Dubinin–Radushkevich
istics for a heterogeneous surface. Unlike the Langmuir, the equation and its isotherm parameters are described in Section S1
Freundlich equation approaches neither (arithmetic) linearity at and Table S1 (Supplementary data), respectively.
low adsorbate concentrations (high dilution) nor a limiting (fixed)
adsorptive capacity as concentration reaches any saturation. It 3.2. Fundamental thermodynamics of adsorption
means that the Freundlich isotherm does not have the Henry’s law
limitation and a saturation limitation. Therefore, infinite surface In the adsorption study, it is essential to establish the
coverage is predicted mathematically, which shows multilayer adsorption mechanisms (i.e., either chemical or physical). Physical
adsorption on the heterogeneous surface. The Freundlich equation adsorption (also known as physisorption) arises from relatively
and its linear form can be expressed as: weak interactions such as van der Waals force, while chemical
adsorption (or chemisorption) involves stronger chemical inter-
qe ¼ K F C 1=n
e ð8Þ actions (chemical bonding) with attendant transfer of electrons
between the adsorbent and adsorbate [8]. The key distinction
between physical and chemical adsorption is described in detail in
1 Table S2. Both the physical and chemical adsorption can be
log qe ¼ log C e þ log K F ð9Þ
n thoroughly recognized through the thermodynamic parameters,
where KF (mg/g)(L/mg)1/n is the Freundlich constant and 1/n including the Gibbs energy change (DG ), the enthalpy change
(dimensionless; 0 < n < 10) is a Freundlich intensity parameter, (DH ), and the entropy change (DS ).
which indicates the magnitude of the adsorption driving force or In general, the increased temperature often coincides with a
surface heterogeneity, and Ce and qe are obtained from Eq. (1). decreased adsorption capacity, which reflects the exothermic
When the n value is higher than 10, the adsorption isotherm is nature of the adsorption process (DH < 0). Noticeably, an
approaching a so-called rectangular isotherm (analogous to an exothermic adsorption process (
DH ), which releases energy in
irreversible isotherm). the form of heat to its surroundings, involves either physisorption
The parameters of the Freundlich equation can be obtained or chemisorption, or a mixture of both processes (comprehensive
from the intercept and slope of the straight line of the plot logqe adsorption). This exothermic phenomenon is because the total
versus log Ce. Table 2 depicts the Freundlich parameters at various energy absorbed in bond breaking is lower than the total energy
temperatures. The Freundlich constants (KF)—characterized as the released in bond making between Cd2+ and OP, causing the release
strength of adsorption—diminished the following order: 17.3 at of extra energy in the form of heat [9]. By contrast, an endothermic
293 K > 16.3 at 303 K > 15.0 at 313 K > 14.9 (mg/g)(L/mg)1/n at 323 K, process (+DH ), which adsorbs energy in the form of heat from its
 H.N. Tran et al. / Journal of Environmental Chemical Engineering 4 (2016) 2671–2682 2675

surroundings, is unequivocally attributable to chemisorption. This The Gibbs energy change is directly calculated from Eq. (10),
is due to the fact that the Cd2+ ions must displace more than one while the enthalpy change and the entropy change were
water molecule for their adsorption, which causes the endother- determined from the slope and intercept of a plot of lnKC against
micity of the adsorption process [9]. Therefore, to establish the 1/T (Eq. (12)). It is well-known that the equilibrium constant (KC)
plausible adsorption mechanisms, it is essential to further consider must be dimensionless. In fact, the common units of DG , the gas
the other thermodynamic parameters. constant, and temperature are J/mol, J/mol  K, and K; therefore,
The Gibbs energy change is a critical factor for determining the the equilibrium constant KC in Eq. (10) must be a parameter
degree of spontaneity of the adsorption process; a higher negative without any unit. This discussion is in accordance with the
value of DG indicates a more energetically favorable adsorption. assumption of the International Union of Pure and Applied
The adsorptive process will occur favourably and spontaneously at Chemistry. Clearly, inaccurate application of the equilibrium
a given temperature if DG exhibits a negative quantity; in constants will result in incorrect estimation of the thermodynamic
contrast, it will be non-feasible and non-spontaneous. Further- parameters, and vice versa.
more, the magnitude of DG also reflects the types of physical or
chemical adsorptions. Meanwhile, the magnitude and sign of DS 3.3. Thermodynamic parameters calculated from various constants
play a contributing role in reflecting whether the organization of
the adsorbate at the solid/solution interface during the adsorption 3.3.1. Equilibrium constant derived from the Langmuir constant
process becomes less random (DS < 0) or more random (DS > 0). The Langmuir equation initially derived from kinetic study and
The negative value of DS also implies that the adsorption subsequently from thermodynamic study. The derivation of the
phenomenon involves an associative mechanism, while the DS Langmuir equation from the thermodynamic aspect can be
positive corresponds to a dissociative mechanism. Moreover, the described as following by Crittendent et al. [10]. The relationship
positive value of DS also involves the increased degree of freedom between vacant surface sites on the surface of adsorbent
of Cd2+ ions in the solution. (Sv; mmol/m2) and adsorbate species in solution (A; mmol) and
The thermodynamic parameters can be computed according to adsorbate species bound to surface sites (SA; mmol/m2) can be
the thermodynamic laws through the following equations: described as the following reaction:

DGo ¼
RTlnK C ð10Þ SV þ A$SA

The relationship of DG to DH and DS is described as: On the basis of the Langmuir expression, it is assumed that the
reaction has a constant Gibbs energy change (DG ; J/mol) for all
DGo ¼ DHo
T DSo ð11Þ sites, so the thermodynamic equilibrium constant (KC; dimension-
The well-known van’t Hoff equation is obtained by substituting less) can be expressed as:
Eq. (10) into Eq. (11)
DG o
DGo ¼
RTlnK C ,K C ¼ e =RT
ð13Þ

DH 1 DS
o o
lnK C ¼ x þ ð12Þ Another Langmuir expression is that each site is assumed to be
R T R
capable of binding at most one adsorbate molecule (mono-layer).
where R is an universal gas constant (8.3144 J/mol  K), and T is According to the equilibrium condition, the thermodynamic
absolute temperature in Kelvin (K).

Table 3
Thermodynamic parameters calculated from the Langmuir constant.

Linear forms T (K) van’t Hoff equation KC DG DH DS

of Langmuir (KJ/mol) (KJ/mol) (J/mol)
Type 1 293 y = 986x + 9.71 R2 = 0.994 475,473
31.86
8.20 80.81
303 434,695
32.72
313 389,470
33.52
323 348,717
34.29

Type 2 293 y = 1,084x + 10.19 1,130,362
33.97
9.02 84.76

R2 = 0.803
303 932,275
34.65
313 779,091
35.32
323 823,582
36.60

Type 3 293 y = 1,051x + 10.20 1,009,091
33.70
8.73 84.81

R2 = 0.799
303 840,909
34.39
313 702,532
35.05
323 743,968
36.32

Type 4 293 y = 924x + 10.37 762,792
33.01
7.68 86.21

R2 = 0.821
303 675,768
33.83
313 561,494
34.47
323 588,078
35.68

Standard deviation 293 – – 0.94 0.60 2.32

303 – – 0.86
313 – – 0.80
323 – – 1.03

KC values calculated from Eq. (17): KC = 55.5  1000  KL.

2676 H.N. Tran et al. / Journal of Environmental Chemical Engineering 4 (2016) 2671–2682

equilibrium constant may be written as: easily obtained as a dimensionless parameter by multiplying KL by
55.5 and then 1000 (Eq. (17)). This finding was originally proposed
SA
DG o
=RT
KC ¼ ¼e ð14Þ by Milonji
c [11] and then developed by Zhou and Zhou [12] and Liu
SV ½A
[13]. The parameters of DG , DH , and DS can be calculated from
where [A] is the concentration of adsorbate A in solution at the following equations:
equilibrium (mg/L), and R and T are the gas constant and
K C ¼ 55:5x1; 000xK L ð17Þ
temperature described in Eq. (12), respectively.
The critical problem in Eq. (14) is due to two unknown
parameters (i.e., SV and [A]). However, this problem can be solved if
the total sites available or monolayer coverage (ST; mol/m2) are DGo ¼
RTlnð55:5x1; 000xK L Þ ð18Þ
fixed:
SA
DHo 1 DSo
ST ¼ SV þ SA ¼ þ SA , lnð55:5x1; 000xK L Þ ¼ x þ ð19Þ
K C ½A R T R
ð15Þ
K ½AST
SA ¼ C where the factor 55.5 is the number of moles of pure water per
1 þ K C ½A
liter (dividing 1000 g/L by 18 g/mol) and the term 55.5  1000
The unit of SA expressed as mmol/m2 might not be particularly  KL is dimensionless.
useful in mass balances; mass of adsorbate adsorbed per unit mass The thermodynamic parameters estimated from Eqs. (17)–(19)
of adsorbent (qe; mg/g) is much more useful. The term of mass are summarized in Table 3. Noticeably, the KC values from the four
loading (qe) can be given in mg/g (Eq. (1)) by multiplying both sides linear forms were significantly different. For example, the KC values
of Eq. (15) by the surface area (SBET; m2/g) of the adsorbent and ranged from 348,717 to 475,473 for Langmuir’s Type 1, Type 2
molecular weight of the adsorbate (Mw; g/mol). Eq. (15) will (823,582–1,130,362), Type 3 (743,968–1,009,091), and Type 4
become: (588,078–762,792) and tended to decrease when the temperatures
Q omax K L C e increased from 293 to 323 K. Nevertheless, the magnitudes of
K C ½AST SBET Mw
SA:SBET :Mw ¼ , qe ¼ ð16Þ thermodynamic parameters–DG , DH , and DS –calculated from
1 þ K C ½A 1 þ K L Ce
the linear forms were nearly the same with a small standard
where qe (mg/g) = SA  SBET  Mw is described as Eq. (1); deviation (SD; 2.3).
Ce (mg/L) = [A] is the concentration of Qomax (mg/g) In case of the KL unit expressed in liter per milligram, Milonji

c
= ST  SBET  Mw is the maximum monolayer adsorptive capacity [11,14] stated that the KL could be obtained as dimensionless by
of adsorbent when surface sites are saturated with adsorbate; and multiplying KL by 106 (Eq. (20)). However, Zhou and Zhou [12]
the relationship between KL and KC can be described in Eq. (20) or recommended that KC can be obtained as dimensionless
(23). by multiplying KL by the molecular weight of adsorbate (Cd;
When the adsorption study is conducted in an aqueous solution 112.41 g/mol), 1000, and then 55.5 (Eq. (23)). Therefore, it is
along with the unit of KL given liter per millimole, the KC could be necessary to discuss whether the thermodynamic parameters

Table 4
Thermodynamic parameters calculated from the Langmuir constant.

Linear forms Temperature (K) van’t Hoff equation KC DG (KJ/mol) DH (KJ/mol) DS (J/mol)
of Langmuir
Type 1 293 y = 994 + 7.86 R2 = 0.994 76,398
27.41
8.26 65.38
303 69,822
28.11
313 62,500
28.75
323 55,843
29.37

Type 2 293 y = 1,087x + 10.19 1,132,718
33.98
9.04 84.72

R2 = 0.802
303 929,972
34.64
313 779,844
35.32
323 823,635
36.60

Type 3 293 y = 1047x + 8.38 161,679
29.23
8.71 69.66

R2 = 0.799
303 134,773
29.77
313 112,670
30.29
323 119,302
31.41

Type 4 293 y = 924x + 8.54 122,300
28.55
7.68 71.01

R2 = 0.820
303 108,300
29.22
313 90,000
29.70
323 94,300
30.78

Standard deviation 293 – – 2.89 0.59 8.37

303 – – 2.89
313 – – 2.94
323 – – 3.16

KC values calculated from Eq. (20): KC = 106  KL.

 H.N. Tran et al. / Journal of Environmental Chemical Engineering 4 (2016) 2671–2682 2677

calculated from the different assumptions are significantly dissimi- recommendations of Milonji
c [11,14] and Zhou and Zhou [12]
lar. provide the same methods of calculation for the thermodynamic
parameters. Astonishingly, the results in Type 1, 3, and 4 contain
K C ¼ 106 K L ð20Þ some deviation; only Type 2 demonstrates identical results
(Tables 3–5).
Tables 3–5 show that the signs of DG , DH , and DS were in
DGo ¼
RTlnð106 K L Þ ð21Þ accordance with experimental data of the adsorption isotherms
(Fig. 2). The negative DH represented the exothermic nature of the
adsorption process, demonstrated by a decrease in the adsorption

DHo 1 DSo capacity (qe) and the equilibrium constant (KC) with higher–
lnð106 K L Þ ¼ x þ ð22Þ
R T R temperature conditions. The exothermic nature also implies the
emission of energy in form of heat into its surroundings during the
where the fact 106 is the solution density (assumption on the
adsorptive process, and the shift of equilibrium in the opposite
density of pure water being 1.0 g/mL) and the term 106  KL is direction of the reaction. Meanwhile, the negative values of DG at
dimensionless. all temperatures suggested that the adsorption phenomenon
K C ¼ 112:41x55:5x1; 000xK L ð23Þ occurred favourably and spontaneously. The increased negative
values of DG , along with the increased temperatures, confirmed
that the accumulation phenomenon of Cd2+ ions onto OP was more
DGo ¼
RTlnð112:41x55:5x1; 000xK L Þ ð24Þ favorable at a lower temperature. The positive value of DS is
corroborated by the increased randomness and disorder at the
solid/solution interface during the adsorption of Cd2+ ions onto OP.
Besides, the (+DS ) value of also indicated that the biosorption

DHo 1 DSo
lnð112:41x55:5x1; 000xK L Þ ¼ x þ ð25Þ process is entropy-governed rather than enthalpy-governed. It can
R T R
be concluded that difference among the thermodynamic param-
where the fact 55.5 is the number of moles of pure water per liter eters, which were computed from the equilibrium constants
and the term of 112.41  55.5  1000  KL is dimensionless. derived from Eqs. (17), (20), and (23), was negligible with a small
The thermodynamic parameters calculated from Eqs. (20)–(22) standard deviation.
and Eqs. (23)–(25) are depicted in Table 4 and Table 5, respectively. The low adsorption enthalpy (DH < 10 KJ/mol) can confirm
Clearly, the KC values were greatly dependent on both the that the biosorption process was physical adsorption (see Table S2)
Langmuir’s linear forms and the methods that demonstrated KC with a weak interaction between Cd(II) and the OP’s surface, which
as a dimensionless. Take Langmuir’s Type 1 as a typical example: can be plausibly explained by the pHPZC of OP. The electrical state of
the KC values in Table 4 ranged from 55,843 to 76,398, while these OP’s surface in solution is characterized by the point of zero charge
values ranged 348,393–476,631 in Table 5. This indicates that the (PZC). The pH value at which the net total (external and internal)
KC values were strongly affected by the estimation methods. surface charge of OP is zero is defined as pHPZC. The pHPZC of OP is
However, the thermodynamic parameters are less relative to the approximately 5.0, indicating that the total acid groups are more
linearized forms of the Langmuir model. As expected, both two dominate than basic groups on the OP’s surface (Fig. S2). When the

Table 5
Thermodynamic parameters calculated from the Langmuir constant.

DG (KJ/mol) DH (KJ/mol) DS (J/mol)


Linear forms T (K) van’t Hoff equation KC
of Langmuir
Type 1 293 y = 994x + 9.70 R2 = 0.994 476,631
31.87
8.26 80.61
303 435,605
32.73
313 389,922
33.52
323 348,393
34.29

Type 2 293 y = 1087x + 10.19 1,132,718
33.98
9.04 84.72

R2 = 0.802
303 929,972
34.64
313 779,844
35.32
323 823,635
36.60

Type 3 293 y = 1,047x + 8.38 1,008,675
33.69
8.71 84.88

R2 = 0.799
303 840,814
34.39
313 702,919
35.05
323 744,295
36.33

Type 4 293 y = 924x + 8.54 763,000
33.01
7.68 86.23

R2 = 0.820
303 675,657
33.83
313 561,488
34.47
323 588,315
35.69

Standard deviation 293 – – 0.94 0.59 2.43

303 – – 0.85
313 – – 0.80
323 – – 1.03

KC values calculated from Eq. (23): KC = 112.41  55.5  1000  KL.

2678 H.N. Tran et al. / Journal of Environmental Chemical Engineering 4 (2016) 2671–2682

pH of solution (pH  7.0; see in the Materials and Methods section)

is higher than 5.18, the surface charge of OP becomes negative due !ð1
1=nÞ
to the ionization or dissociation of acidic oxygen surface groups 106
DGo ¼
RTlnðK F r ð27Þ
(i.e., carboxylic and phenolic). Therefore, it is favorable to adsorb r
Cd2+ cations in the solution through the weak force of electrostatic
attractions. Furthermore, the acidic oxygen surface groups on the
OP’s surface, which can be recognized by FTIR (Fig. S1), can also ! !
106
DHo 1 DSo
interact with Cd2+ ions to form surface complexes (non-electro- ln K F r ð1
1=nÞ
¼ x þ ð28Þ
r R T R
static attractions). The distinction between electrostatic and non-
electrostatic attractions can be identified through further dis- where r is the density of pure water (assumption on 1.0 g/mL).
cussions in the desorption study. Table 6 illustrates the thermodynamic parameters which were
Several possible explanations exist for the effect of temperature computed from the equilibrium constant obtained from Eq. (26).
on the adsorption capacity (the temperature increase that resulted Generally, the signs and magnitudes of DG , DH , and DS
in the decrease of Cd2+ retention capacity by OP). Firstly, the Cd2+ calculated from KF (Table 6) were consistent with those from KL
ions tended to escape from the solid phase to the bulk phase with (Tables 3–5). The correlation coefficients (R2) of the Langmuir
an increased temperature of the solution [15,16]. This also model were higher than those of the Freundlich, but the
demonstrated the increasing desorption tendency of Cd2+ ions Chi-squares (x2) of the Freundlich model were lower than the
from the interface to solution with increasing temperatures. Langmuir (Table 2), which indicated that the biosorption
Secondly, the decrease of electrostatic force resulted from the characteristics of Cd2+ on OP were also well described by
agitation increase of the dissolved chemical species, which Freundlich model. Moreover, the high correlation coefficient of
decreased the interactions between positively charged Cd(II) ions the van’t Hoff equation suggested the appropriate application of
and negatively charged sites on the OP’s surface [17]. Thirdly, the the Freundlich constant KF in estimating the thermodynamic
solubility of adsorbate species increased alongsidehigher–temper- parameters. Similar to the Langmuir’s model, the equilibrium
ature values, which contributed to stronger interaction forces constants KC obtained from Eq. (26) are inversely proportional to
between the adsorbate and solvent than between adsorbate and temperatures, which is evidence of exothermic phenomenon in
adsorbent [9]. Lastly, the characteristic adsorption energy (E) adsorption.
decreased from 0.42 to 0.32 KJ/mol when temperatures increased It is noted that the unit of KF is reliant on the used units of
223–293 K (Table S1). liquid-phase concentration (C) and solid-phase concentration (q).
Table S3 records the incorrect thermodynamic parameters, The units of C (mg/L or mmol/L) and q (mg/g or mmol/g) are the
which were directly calculated from the Langmuir constants most frequently used to demonstrate adsorption from water
without dimensionality considerations. The results demonstrate solutions. The different units of KF can be reciprocally converted by
that the signs and magnitudes of the thermodynamic parameters using Eq. (29).
are contradictory. For example, in the Langmuir’s Type 1, the DG
values, directly calculated from equation KC = KL (L/mg), are K F ¼ K
F ðMw Þ1
1=n ð29Þ
positive (6.27–7.75 KJ/mol), while the corresponding values from 1/n
where the unit of Freundlich constant KF is (mg/g)(mg/L) along
equation KC = KL (L/mmol) are negative (from
5.24 to
4.94 KJ/
with C (mg/L) and q (mg/g), the unit of K*F is (mmol/g)(mmol/L)1/n
mol). Similarly, the negative magnitude of entropy change (
DS )
along with C (mmol/L) and q (mmol/g), and Mw is the molecular
obtained from KC = KL (L/mg) was overwhelmingly higher than that
weight of the adsorbate. The exponent 1/n is not affected by the
from KC = KL (L/mmol), with
49.48 and
10.03 J/mol, respectively.
different units of liquid-phase concentration (C) and solid/phase
concentration (q) (see in Table 2).
3.3.2. Equilibrium constant derived from the Freundlich constant
The inaccurate thermodynamic parameters which were direct-
The Freundlich equation is consistent with the thermodynam-
ly calculated from the Freundlich constant KF without the
ics of heterogeneous adsorption [10]. The Freundlich constant KF
consideration of dimensionality are listed in Table S4.
can be obtained as a dimensions (Eq. (26)) [18].
!
106 ð1
1=nÞ 3.3.3. Equilibrium constant derived from partition constant
KC ¼ KF r ð26Þ The thermodynamic equilibrium constant (or the thermody-
r
namic partition coefficient) with changes in temperatures was
originally published by Biggar et al. [19]. The equilibrium constant

Table 6
Thermodynamic parameters calculated from the Freundlich constant and partition constant.

T (K) van’t Hoff equation KC DG (KJ/mol) DH (KJ/mol) DS (J/mol)
Freundlich constant (KF)
293 y = 492x + 8.07 R2 = 0.939 17,258
23.78
4.10 67.10
303 16,338
24.45
313 15,028
25.04
323 14,935
25.82

Partition constants (Kp)

293 y = 191x + 0.77 4.18
3.49
1.59 6.44
303 R2 = 0.529 4.14
3.58
313 3.86
3.52
323 4.00
3.73

In case of Ce(mg/L), Cs(mg/L), andqe (mg/g); and KC values calculated from Eq. (26) for the Freundlich constant.
 H.N. Tran et al. / Journal of Environmental Chemical Engineering 4 (2016) 2671–2682 2679

can be defined as follows: coefficient (Kd) was defined as the following equation:
as g Cs Kd ¼
qe
ð32Þ
Kp ¼ ¼ s ð30Þ
ae g e C e Ce

where as is the activity of the adsorbed Cd(II) onto OP; ae is the In this method, the Kd values are obtained by plotting ln(qe/Ce)
activity of Cd(II) in the solution at equilibrium; g s is the activity against Ce and extrapolating to zero Ce. After that, a straight line fits
coefficient of the adsorbed Cd(II) onto OP; g e is the activity into the data with a high correlation coefficient value (R2), and then
coefficient of Cd(II) in the solution at equilibrium; Cs is the its intersection with the vertical axis provides the value of Kd.
concentration of Cd(II) adsorbed onto OP at equilibrium (mg/L); Fundamentally, Eq. (32) might be derived from the Freundlich and
and Ce is the concentration of Cd(II) in the solution at equilibrium Langmuir equation. In the situation of n = 1, the Freundlich
(mg/L). The Cs is defined by the mass balance of adsorbate (Cd2+ equation (Eq. (8)) will become the Henry’s equation as a linearized
ions) that disappeared from the solutions, which should in the fact isotherm (Eq. (33)), and the Freundlich constant (KF) will be equal
appear on the adsorbent (orange peel). to the Henry constant (KH). In the situation of the very low
When the concentration of Cd2+ in the solution approaches zero concentrations of absorbate (KL and Ce << 1), the Langmuir
which results in Cs ! 0 and Ce ! 0, the activity of coefficient g equation (Eq. (3)) will become the Henry’s equation (Eq. (33)),
approaches unity at low concentrations. Eq. (30) can be written and the Henry constant will become KH = QomaxKL. The Henry
into: constant (KH) is also referred to as the distribution coefficient (Kd)

C s as qe ¼ K H C e ð33Þ
Lim ¼ ¼ Kp ð31Þ
Cs!0C e ae The method mentioned above has been widely applied to
The Kp values can be obtained by plotting a straight line of ln(Cs/Ce) calculation of the thermodynamic parameters in the literature.
against Cs and extrapolating to Cs to zero. The straight line fits to Unfortunately, the unit of Kd in Eq. (32) is liter per gram due to Ce in
data with a high correlation coefficient value (R2), and its milligram per liter and qe in milligram per gram. Therefore, it is
intersection with the vertical axis provides the value of Kp. In impossible to directly employ Kp as the equilibrium constant for
this case, the partition coefficient was in unison with the the calculation of DG , DH , and DS . This unit problem has been
equilibrium constant. The DG value was directly calculated from discussed by several researchers [11,21,22]. Canzano et al. [21]
Eq. (10), while the values of DH and DS were determined through suggested that the partition coefficient Kd (L/g) can convert into KC
the slope and intercept of Eq. (12), respectively. (as dimestionless) by multiplying Kd to a factor of 1000 as
Table 6 indicates the thermodynamic parameters calculated originally proposed by Milonji
c [11,14]. The parameters of DG ,
from the partition coefficient. It is worth nothing that the values of DH , and DS can be estimated from the following equations:
KC, DG , DH , and DS were not dependent on the units (mg/L or
DGo ¼
RTlnð1; 000K d Þ ð34Þ
mmol/L) of Cd2+ concentrations (detail data not shown). Although
the signs of DG , DH , and DS calculated from the partition and
the Langmuir constants are the same, the partition constant might

DH o 1 DS o
not be appropriate to calculate the thermodynamic parameters, lnð1000K d Þ ¼ x þ ð35Þ
R T R
due to the low correlation coefficient of the van’t Hoff equation
(R2 = 0.53). According to the United States Environmental Protec- Table 7 indicates the thermodynamic parameters estimated
tion Agency, the partition coefficient is only appropriate for from the distribution coefficient. Clearly, the signs of DG , DH , and
calculation of the thermodynamic parameters in a situation of low DS estimated from the distribution coefficient and the Langmuir
initial concentrations of absorbate. An analogous conclusion was constant are the same. However, the unit of Kd (L/g) is completely
also reported in the literature [13]. They highlighted that the different from KL (L/mg), and the adsorption equilibrium data do
partition coefficient might be only equal to the thermodynamic not fit the Henry’s adsorption isotherm (see Fig. 2). As a result,
equilibrium constant when adsorbed concentrations in the application of the distribution coefficient was not appropriate for
solutions are very low. In this situation, it can be used for calculating the thermodynamic parameters in this study; a low
calculation of the thermodynamic parameters. correlation coefficient of the van’t Hoff equation (R2 = 0.42) is
evidence of this inappropriateness.
3.3.4. Equilibrium constant derived from distribution coefficient Table S5 shows that the incorrect thermodynamic parameters
In 1987, Khan and Singh [20] proposed the method used to were directly calculated from the distribution coefficient without
calculate thermodynamic parameters (Eq. (32)). This method was the dimensionality consideration of the equilibrium constant.
developed from the method of Biggar et al. [19]. The distribution To thoroughly understand the adsorption mechanism as well as
to distinguish between electrostatic and non-electrostatic

Table 7
Thermodynamic parameters calculated from the distribution coefficient or Henry’s constant.

T KC = Kd  1000 KC = Kd  112.41  55.5

(K)
van’t Hoff KC DG (KJ/mol) DH (KJ/mol) DS (J/mol) van’t Hoff KC DG (KJ/mol) DH (KJ/mol) DS (J/mol)
equation equation
293 y = 234x + 7.36 R2 = 0.429 3,497
19.89
1.95 61.17 y = 234x + 9.19 21,814
24.35
1.95 76.39
303 3,459
20.54 R2 = 0.429 21,580
25.15
313 3,139
20.96 19,585
25.73
323 3,331
21.79 20,778
26.71

In case of Ce (mg/L) and qe(mg/g); KC = Kd  1000 [11,14,21], and KC = Kd  12.4  55.5 [12].
2680 H.N. Tran et al. / Journal of Environmental Chemical Engineering 4 (2016) 2671–2682

Fig. 3. Kinetic of Cd2+ adsorption on orange peel at various temperatures (mean

Fig. 4. Desorption of Cd2+ from orange peel by KNO3 and EDTA (mean values 
values  standard deviation).
standard deviation; digitals in parenthesis indicate values of the desorption rate).

attractions, the further experiments regarding kinetics and

desorption were conducted. 90%) was overwhelmingly higher than that by EDTA (approxi-
mately 10%), which confirmed that the electrostatic attraction was
3.4. Kinetic and reversible adsorption study the main underlying mechanism of Cd2+ biosorption onto OP. This
conclusion is consistent with the other studies. Zhong et al. [5]
Fig. 3 shows that the cadmium removal rate increased rapidly investigated the adsorption and desorption of Cd(II) in the tropical
over the initial 30 min; after that, the removal rate decreased soils during pedogenesis in basalt. The authors concluded that the
slightly until the point of equilibrium was reached at approxi- high desorption rate (87.5–92.6%) by using 1 M KNO3 confirmed
mately 90 min. The adsorption equilibrium can be achieved rapidly that Cd(II) was adsorbed onto OP mainly through electrostatic
due to the requirement of minimal activation energy. The amount attractions, and the adsorption highly depends on surface charge
of Cd uptake at equilibrium calculated using the pseudo-second- density of the soils. Furthermore, Xu and Zhao [6] used EDTA to
order equation (qe,cal), matched well with that obtained from the replace cadmium complexed with biochars to evaluate the
experiment (qe,exp); and correlation coefficients (R2) of the pseudo- complexing ability of the biochars with Cd2+; they concluded that
second-order (0.988–0.991) reaction were higher than those of the both electrostatic and non-electrostatic mechanisms contributed
pseudo-first-order (0.865–0.936) reaction (Table S6). Therefore, to the enhanced adsorption of Cd(II) on the biochars. The
the kinetic data were adequately described by the pseudo-second biosorption process can attain fast equilibrium and be easily
order reaction, and the rate constants of the second-order reaction reversible because the required energies in physisorption are
(K2) were applied for the estimation of activation energy. The small.
first- and second-order equations and their parameters are The analysis of results of texture (Table 1) and superficial
presented in Section S2 and Table S6 (Supplementary data), element composition (Fig. 5) shows that the cation exchange with
respectively. light metal ions (i.e., Ca2+, Na+, and K+) and pore filling did not
In the adsorption process, activation energy (Ea; KJ/mol) is involve the adsorptive mechanism of Cd(II) on OP. The presence of
defined as the minimum energy that must be overcome by the cadmium element on the OP’s surface (Fig. 5b) confirmed the
adsorbate molecules, and it can be estimated through the effectively adsorptive results of Cd2+ ions from the aqueous
Arrhenius equation: solution.
  The primary adsorption mechanisms of Cd2+ onto the OP’s
Ea 1
lnK 2 ¼
þ lnA ð36Þ surface can be summarized as follows:
R T

where K2 is the rate constant of the second-order reactions, A is the (1) Electrostatic attraction (out-sphere complexation; contribu-
pre-exponential factor (frequently factor), and R and T are the gas tion of approximately 90%): when pH of the solution (7.0) is
constant and temperature, respectively. higher than pHPZC of the orange peel (5.18), the deprotonation
The activation energy and pre-exponential factor obtained from reactions of the carboxylic and hydroxyl groups available on
the slope and intercept of Eq. (36) (Fig. S3) were
11.9 KJ/mol and the OP’s surface occurred as Eqs. (37)–(38), and the electro-
6.905  10
5, respectively. The negative value of Ea indicated the static attractions can occur between negative-charged sties on
exothermic nature of the adsorption process [9] and the negligible the OP’s surface and Cd2+ cations in the solution as Eqs.
value of Ea confirmed again the physical adsorption with the (39)–(40). Furthermore, outer-sphere surface complexes,
relatively weak attraction forces. which are considered in the triple-layer model, contain at
The results of the reversibility study are shown in Fig. 4. The least one water molecule between the adsorbing ion and the
amount of Cd2+ ions desorbed by KNO3 and EDTA from the OP’s surface functional group [23] (Eqs. (41)–(42)).
surface resulted from the relative contribution of the electrostatic
and non-electrostatic adsorption mechanisms, respectively. Fig. 4 (37)
shows that the desorption rate of Cd(II) by KNO3 (approximately
 H.N. Tran et al. / Journal of Environmental Chemical Engineering 4 (2016) 2671–2682 2681

Fig. 5. Scanning electron microscopy (SEM) and X-ray spectroscopy (EDX) spectra of before (a) 2000 magnification and after adsorption cadmium (b) 5000 magnification.

(38) can be found in Section 3 and the surface complexation triple-layer

model is presented in Fig. S4 (Supplementary data).

(39) 4. Conclusions

After considering the problems regarding the dimensionality of

(40) the equilibrium constant, it can be concluded that:

1. The values of the equilibrium constant (KC) were strongly

(41) dependent on the applied constants, such as Langmuir,
Freundlich, Henry, and partition.
2. The thermodynamic parameters—DG , DH , and DS —calculat-
(42) ed from the constants of Langmuir, Freundlich, Henry and
partition had the same sign. However, the magnitude of these
parameters varied greatly according to the applied constants.
3. The adsorption equilibrium data play a vital role in determining
(2) Non-electrostatic attraction (inner-sphere complexation): this the constants (i.e., Langmuir, Freundlich, Henry, and partition)
attraction plays a minor role in the adsorption process that are most appropriate to correctly calculate the thermody-
(approximately 10%) and is described as Eqs. (43)–(44). namic parameters. Furthermore, the correlation coefficient (R2)
of van’t Hoff equation is also another determining factor in
(43) selecting the optimal derivation for KC. In this study, the
thermodynamic parameters of the biosorption process of
cadmium onto orange peel were appropriately calculated based
(44) on both the Langmuir and Freundlich constant.
4. The four linear forms of the Langmuir model were properly
applied to calculate the thermodynamic parameters, demon-
The pKa values of the carboxylic (2.0–4.0) and phenolic strating some deviation in Type 1, 3, and 4; only Type 2
(8.0–9.0) groups also play an essential part in properly explaining demonstrated identical results.
the OP’s surface charge. Both the carboxylic and phenolic groups 5. The biosorption process of Cd2+ onto orange peel was
can be dissociated and predominantly negative-charged spontaneous (
DG ) and exothermic (
DH ) in nature.
(Eqs. (37)–(38)) when pH of the solution is higher than their Moreover, the (+DS ) value indicated an increased randomness
pKa values. pH of the solution in this study is nearly 7.0; therefore, at the solid-solution interface during the adsorption interaction.
the carboxylic groups are directly accountable for the capture of Cd
(II) onto the OP’s surface. More detailed discussions on the The Cd2+ adsorption process from the aqueous solution onto
distinction between electrostatic and non-electrostatic attractions orange peel involves physical adsorption. This is evident not only
2682 H.N. Tran et al. / Journal of Environmental Chemical Engineering 4 (2016) 2671–2682

by low activation energy (Ea), adsorption energy (E), and enthalpy [6] R.-K. Xu, A.-Z. Zhao, Effect of biochars on adsorption of Cu(II), Pb(II) and Cd(II)
change (
DH ), but also by fast-reached equilibrium and by three variable charge soils from southern China, Environ. Sci. Pollut. Res. 20
(12) (2013) 8491–8501.
reversibility. The electrostatic attractions between negatively [7] K.V. Kumar, Comments on Adsorption of acid dye onto organobentonite, J.
charged sites on the OP’s surface and Cd2+ cations in the solution Hazard. Mater. 137 (1) (2006) 638–639.
plays a major role (90%) in the adsorption mechanisms, [8] G. Hill Jr, T.W. Root, Elements of Heterogeneous Catalysis. Introduction to
Chemical Engineering Kinetics and Reactor Design, Second edition, John Wiley
compared to the minor role played non-electrostatic attractions & Sons, 1977, pp. 152–188.
(10%). The carboxylic groups are mainly accountable for Cd2+ [9] S. Papita, C. Shamik, Insight into adsorption thermodynamics, Thermody-
adsorption on OP through electrostatic attractions. namic, INTECH Open Access Publisher, 2011, pp. 349–364.
[10] J.C. Crittenden, H. Montgomery Watson, Adsorption, MWH’s Water Treatment
Principles and Design, Third edition, John Wiley & Sons, 2012, pp. 1136–1150.
Acknowledgements [11] S.K. Milonji
c, A consideration of the correct calculation of thermodynamic
parameters of adsorption, J. Serb. Chem. Soc. 72 (12) (2007) 1363–1367.
[12] X. Zhou, X. Zhou, The unit problem in the thermodynamic calculation of
This current work was financially supported by Chung Yuan
adsorption using the Langmuir equation, Chem. Eng. Commun. 201 (11) (2014)
Christian University (CYCU) in Taiwan. The first author would like 1459–1467.
to thank CYCU for the Distinguished International Graduate [13] Y. Liu, Is the free energy change of adsorption correctly calculated? J. Chem.
Students (DIGS) scholarship to pursue his doctoral studies. Thanks Eng. Data 54 (7) (2009) 1981–1985.
[14] S.K. Milonji
c, Comments on removal of uranium (VI) from aqueous solution by
are also due to Michael Riches for assistance with English editing. adsorption of hematite, by X. Shuibo, Z. Chun, Z. Xinghuo, Y. Jing, Z. Xiaojian, W.
Jingsong, J. Environ. Radioact. 100 (10) (2009) 921–922.
Appendix A. Supplementary data [15] S. Karaca, et al., Kinetic modeling of liquid-phase adsorption of phosphate on
dolomite, J. Colloid Interface Sci. 277 (2) (2004) 257–263.
[16] Ş. Kubilay, et al., Removal of Cu(II): Zn(II) and Co(II) ions from aqueous sol-
Supplementary data associated with this article can be found, in utions by adsorption onto natural bentonite, Adsorption 13 (1) (2007) 41–51.
the online version, at http://dx.doi.org/10.1016/j.jece.2016.05.009. [17] M. Antunes, et al., Removal of diclofenac sodium from aqueous solution by
Isabel grape bagasse, Chem. Eng. J. 192 (2012) 114–121.
[18] P.S. Ghosal, A.K. Gupta, An insight into thermodynamics of adsorptive removal
References of fluoride by calcined Ca-Al-(NO3) layered double hydroxide, RSC Adv. 5 (128)
(2015) 105889–105900.
[1] S. Lyubchik, A. Lyubchik, et al., Comparison of the thermodynamic parameters [19] J.W. Biggar, M.W. Cheung, Adsorption of picloram (4-amino-3,5,6-tri-
estimation for the adsorption process of the metals from liquid phase on chloropicolinic acid) on panoche, ephrata, and palouse soils: a thermodynamic
activated carbons, Thermodynamics-Interaction Studies-Solids, Liquids and approach to the adsorption mechanism1, Soil Sci. Soc. Am. J. 37 (6) (1973).
Gases, INTECH Open Access Publisher, 2011, pp. 95–112. [20] A.A. Khan, R.P. Singh, Adsorption thermodynamics of carbofuran on Sn(IV)
[2] E. Roth, V. Mancier, B. Fabre, Adsorption of cadmium on different granulo- arsenosilicate in H+, Na+ and Ca2+ forms, Colloids Surf. 24 (1) (1987) 33–42.
metric soil fractions: influence of organic matter and temperature, Geoderma [21] S. Canzano, et al., Comment on Removal of anionic dye Congo red from
189 (2012) 133–143. aqueous solution by raw pine and acid-treated pine cone powder as adsor-
_ Tosun, Ammonium removal from aqueous solutions by clinoptilolite: de-
[3] I. bent: equilibrium, thermodynamic, kinetics, mechanism and process design,
termination of isotherm and thermodynamic parameters and comparison of Water Res. 46 (13) (2012) 4314–4315.
kinetics by the double exponential model and conventional kinetic models, [22] C.S. Patrickios, E.N. Yamasaki, A correction to the calculation of the Gibbs free
Int. J. Environ. Res. Public Health 9 (3) (2012) 970–984. energy of adsorption for biomolecules in ion-exchange systems, Biophys.
[4] H.N. Tran, S.-J. You, H.-P. Chao, Effect of pyrolysis temperatures and times on Chem. 69 (2–3) (1997) 219–220.
the adsorption of cadmium onto orange peel derived biochar, Waste Manage. [23] S. Goldberg, Surface complexation modeling, Reference Module in Earth
Res. 34 (2) (2016) 129–138. Systems and Environmental Sciences, Elsevier, 2013, pp. 1–14.
[5] K. Zhong, et al., Adsorption and desorption of Cu(II) and Cd(II) in the tropical
soils during pedogenesis in the basalt from Hainan, China, Carbonates
Evaporites 25 (1) (2010) 27–34.