Chemical Engineering Journal 225 (2013) 535–546

Contents lists available at SciVerse ScienceDirect

Chemical Engineering Journal

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

Binary adsorption of heavy metals from aqueous solution onto natural

clays
E. Padilla-Ortega, R. Leyva-Ramos ⇑, J.V. Flores-Cano
Centro de Investigacion y Estudios de Posgrado, Facultad de Ciencias Quimicas, UASLP, Av. Dr. Manuel Nava No. 6, San Luis Potosi SLP 78210, Mexico

h i g h l i g h t s g r a p h i c a l a b s t r a c t

 Binary adsorption of metals on clays

was interpreted by multicomponent
isotherms.
 Bentonite capacity for adsorbing
Cd(II) was similar to that for
adsorbing Ni(II).
 Zn(II) presented strong antagonism to
the adsorption of Cd(II) on sepiolite.
 Adsorption of Zn(II) on sepiolite was
slightly affected by presence of Cd(II).
 Vermiculite adsorption capacity
toward Pb(II) was higher than that
toward Cu(II).

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

Article history: The binary adsorption of Cd(II)–Ni(II) on bentonite, Zn(II)–Cd(II) on sepiolite and Pb(II)–Cu(II) on vermic-
Received 2 February 2013 ulite was investigated in this work. The experimental binary adsorption data were interpreted using the
Received in revised form 2 April 2013 following multicomponent adsorption isotherms: non-modified, extended and modified Langmuir; non-
Accepted 3 April 2013
modified and modified Redlich–Peterson; extended Freundlich; and Sheindorf–Rebuhn–Sheintuch. The
Available online 11 April 2013
modified Redlich–Peterson multicomponent isotherm best fitted the adsorption data of Cd(II)–Ni(II) on
bentonite and Pb(II)–Cu(II) on vermiculite, while the modified Langmuir multicomponent isotherm best
Keywords:
fitted the data of Zn(II)–Cd(II) on sepiolite. The competitive adsorption of Cd(II)–Ni(II) on bentonite
Adsorption
Binary
revealed that both metals presented very similar antagonism toward the adsorption of the other metal.
Clays The single metal adsorption on sepiolite showed that the Cd(II) adsorption capacity of sepiolite was 1.3
Competitive times greater than that of Zn(II); however, in the binary adsorption, the Zn(II) presented a strong antag-
Heavy metal onism to the adsorption of Cd(II), but Cd(II) did not affected the adsorption of Zn(II). The single adsorption
Multicomponent isotherms isotherms of Cu(II) and Pb(II) on vermiculite revealed that the affinity of the vermiculite towards Pb(II)
was slightly higher than that of the Cu(II). However, in the competitive adsorption, the affinity of Pb(II)
towards the vermiculite was much higher than that of the Cu(II).
Ó 2013 Elsevier B.V. All rights reserved.

1. Introduction teries. Wastewater from these industries normally contains Pb(II),

The heavy metals have very diverse applications and are com-
monly found in wastewater from the mining and metal finishing
industry, refining and smelting of metals, and manufacture of bat-
⇑ Corresponding author. Tel.: +52 444 826 2440; fax: +52 444 826 2372.
E-mail address: rlr@uaslp.mx (R. Leyva-Ramos).
Cu(II), Cd(II), Ni(II) and Zn(II), and the presence of these metals in
wastewater is of great concern because some of these metals are
very toxic to the human health and the environment [1].
The adsorption is a separation process widely used for the re-
moval of heavy metals from wastewater due to its versatility and
easy operation. Among the most commonly used adsorbents are
activated carbon, biosorbents, natural zeolites, and natural clays.
The individual adsorption of heavy metals from water solution

1385-8947/$ - see front matter Ó 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.cej.2013.04.011
536 E. Padilla-Ortega et al. / Chemical Engineering Journal 225 (2013) 535–546

Nomenclature

ai Redlich–Peterson isotherm constant for metal i, L/g Nm number of experiment data.

aij competition coefficient of component i by component j ni Freundlich isotherm constant for metal i.
bi Redlich–Peterson isotherm constant for metal i, (L/ qi uptake of metal i adsorbed per mass of clay, meq/g.
meq)bi qi,F1 uptake of metal i adsorbed per mass of clay at first equi-
Ci concentration of metal i in aqueous solution at equilib- librium in the displacement experiment, meq/g.
rium, meq/L qi,F2 uptake of metal i adsorbed per mass of clay at second
Ci,exp experimental concentration of metal i, meq/L equilibrium in the displacement experiment, meq/g.
Ci,F final concentration of metal i in aqueous solution at qi,j,cal uptake of metal of i corresponding to data number j cal-
equilibrium, meq/L culated with the multicomponent adsorption isotherm,
Ci,F1 final concentration of metal i at the first equilibrium of meq/g
the displacement experiments, meq/L qi,j,exp experimental uptake of metal i corresponding to data
Ci,F2 final concentration of metal i at the second equilibrium number j, meq/g
of the displacement experiments, meq/L qm,i Langmuir isotherm constant for metal i, meq/g
Ci,0 initial concentration of metal i in aqueous solution, qmax constant of extended Langmuir multicomponent iso-
meq/L therm, meq/g.
%D average percentage deviation. V0 initial volume of solution, L
Ki Langmuir isotherm constant of metal i, L/meq xi, yi, zi extended Freundlich multicomponent isotherm con-
KE,i extended Langmuir multicomponent isotherm constant stants for component i.
for component i, L/meq
ki Freundlich isotherm constant for metal i, meq(11/n) Greek letters
L(1/n)/g. bi Redlich–Peterson isotherm constant for metal i.
m mass of clay, g gi Interaction factor for metal i.
N number of metals in solution.

on these adsorbents has been extensively studied; however, indus-
trial wastewater normally contains two or more metals. The pres-
ence of another metal in solution affects the adsorption capacity of
the single metal. Therefore, it is very important to study the mul-
ticomponent adsorption of heavy metals and to find out about
the selectivity or affinity of each metal for a given adsorbent and
to know if there is competition between the metal cations for the
same adsorption sites.
In the last 10 years, the multicomponent or competitive adsorp-
tion of metals from water solution has been investigated on vari-
ous adsorbents. The following competitive adsorption systems
have been studied: Pb(II)–Zn(II) on sulfured orange peel [2];
Pb(II)–Cu(II) on Sphaerotilus natans [3], on cow bone char [4], on
granular activated carbon [5] and on Rhizopus arrhizus [6]; Pb(II)–
Cd(II) on clinoptilolite [7], on green macroalga Caulerpa lentillifera
[8] and on calcium alginate [9]; and Pb(II)–Cd(II)–Zn(II) on Eichhor-
nia crassipes [10]. In these works, it was shown that the Pb(II) pre-
sented higher affinity towards these adsorbents than the other
metals since the presence of Cu(II), Cd(II) or Zn(II) in water solution
affected very slightly the mass of Pb(II) adsorbed on these
adsorbents.
The single adsorption of heavy metals on natural clays has been
extensively investigated since natural clays have a high cation ex-
change capacity, mechanical and chemical stability and low-cost,
and are very abundant [11–19]. The main disadvantages are that
the natural clays have very low or no capacity for adsorbing anio-
nic species [20], and the removal efficiency of clays for heavy met-
als is normally less than that of natural zeolites. The adsorption
mechanisms of metal cations on natural clays have been argued
and reviewed in various works [21].
The binary adsorption of metals on bentonite [22], beidellite
[23], kaolinite [24], montmorillonite [25] and smectite [26] have
been examined very superficially in recent works, but the binary
adsorption of metals on sepiolite and vermiculite has not been
investigated. Zhi-rong and Shao-qi [22] evaluated the adsorption
of Cu(II)–Ni(II) on a bentonite from Gaomiaozi (Mongolia, China)
and found that the capacity of the bentonite for adsorbing Cu(II)
decreased slightly by the presence of the Ni(II) and the same trend
happened in the other way around. The single and multicompo-
nent adsorption of Cd(II), Cu(II), Pb(II), and Zn(II) from aqueous
solution onto kaolinite was analyzed by Srivastava et al. [24] and
their results showed that the adsorption capacity increased in
the following orders: Cu(II) < Zn(II) < Pb(II) < Cd(II) in the single
metal systems, and Pb(II) < Cu(II) < Zn(II) < Cd(II) in the multicom-
ponent system. Srivastava et al. [24] did not interpret the adsorp-
tion equilibrium data with the typical multicomponent adsorption
isotherm models.
In general, the experimental binary adsorption equilibrium data
of metals on natural clays has been obtained for few metals and
clays systems, but no work has been reported regarding the mod-
eling of the adsorption equilibrium data using the multicomponent
adsorption isotherms.
The main objectives of this work were to study the binary
adsorption of Cd(II)–Ni(II) on bentonite, Zn(II)–Cd(II) on sepiolite
and Pb(II)–Cu(II) on vermiculite, and to model the adsorption equi-
librium data with the binary adsorption isotherms. Furthermore, it
was investigated the effect of the presence of Cd(II) on the adsorp-
tion of Ni(II) on bentonite, Zn(II) on the adsorption of Cd(II) on
sepiolite, and Pb(II) on the adsorption of Cu(II) on vermiculite
and vice versa.
2. Single and multicomponent adsorption equilibrium
isotherms
The adsorption equilibrium data for a single component can be
normally interpreted by the Langmuir, Freundlich and Redlich–
Peterson isotherms, which are represented mathematically as
follows:
1=ni
qi ¼ ki C i ð1Þ
qm;i K i C i
qi ¼ ð2Þ
1 þ K iCi
ai C i
qi ¼ b
ð3Þ
1 þ bi C i i
 E. Padilla-Ortega et al. / Chemical Engineering Journal 225 (2013) 535–546 537

Table 1 ðn1 Þþx1 ðn1 Þþx2

Chemical composition of bentonite (B), sepiolite (S) and vermiculite (V). k1 C 1 1 k2 C 2 2
q1 ¼ q2 ¼ ð9Þ
C x11 þ y1 C z21 C x22 þ y2 C z12
Clay Weight percentage (%)
SiO2 Al2O3 Na2O CaO K2O MgO Fe2O3 (g) Sheindorf–Rebuhn–Sheintuch isotherm (SRSI) [31]:
 
B 59.4 15.7 0.84 3.04 2.70 3.52 4.48 ! n1 1
S 53.3 6.92 5.55 3.99 1.72 17.3 2.09
X
N i

V 36.5 9.9 0.16 1.53 4.28 23.0 13.5

qi ¼ ki C i aij C j ð10Þ
j¼1

where aii = ajj = 1.

Table 2 The competition and interaction factors of the multicomponent
Cation exchange capacity (CEC) and textural properties of bentonite (B), sepiolite (S)
isotherms were estimated using the Micromath Scientist software,
and vermiculite (V).
which is based upon the optimization algorithm of Levenberg–
Clay CEC Surface area Pore volume Average pore Marquardt. The minimizing function was the following:
(meq/g) (m2/g) (cm3/g) diameter (nm)
B 1.06 41.7 0.063 5.9 X
2 X
Nm

S 0.44 194 0.30 6.2 Least squares ¼ ðqi;j;exp  ¼ qi;j;cal Þ2 ð11Þ

V 0.42 15.2 0.049 13.0 i¼1 j¼1

The average percentage deviation was estimated with the fol-

lowing equation:
The multicomponent adsorption isotherms are normally ob- "  #
Nm  
tained by extending the above single component adsorption iso- 1Xi¼2
1 X qi;j;exp  qi;j;cal 
%D ¼    100% ð12Þ
therms and can be classified accordingly to the single adsorption 2 i¼1 N m j¼1  qi;j;exp 
isotherm that originated the multicomponent isotherm. The fol-
lowing multicomponent adsorption isotherms were used in this
work:
3. Materials and experimental methods

(a) Non-modified Langmuir multicomponent isotherm (NLMI)

3.1. Natural clays
[9]:
The calcium bentonite employed in this work was obtained
qm;i K i C i from a location in Guadalcazar, San Luis Potosi, Mexico. The sepio-
qi ¼ P ð4Þ
1 þ Nj¼1 K j C j lite sample was supplied by the company SEPIOLSA, and it was col-
lected from a location near Madrid, Spain. The vermiculite sample
(b) Extended Langmuir multicomponent isotherm (ELMI) [27]: was provided by the company Virginia Vermiculite, situated in Vir-
qmax K E;i C i ginia, USA. The particle average diameter of the three clays was
qi ¼ P ð5Þ
1 þ Nj¼1 K E;j C j 0.375 mm.

(c) Modified Langmuir multicomponent isotherm (MLMI) with

3.2. Characterization of clays
an interaction factor gi [28–30]:
qm;i K i ðC i =gi Þ The chemical composition of the clays was determined follow-
qi ¼ PN ð6Þ
1þ j¼1 K j ðC j =gj Þ ing the method proposed by Rettig et al. [32] and using an atomic
absorption spectrometer, PerkinElmer, model AAnalyst. The cation
(d) Non-modified Redlich–Peterson multicomponent isotherm exchange capacity (CEC) was quantified with the method proposed
(NRPMI) [29,30]: by Ming and Dixon [33]. The chemical composition and CEC of the
ai C i clays are shown in Tables 1 and 2, respectively. The textural prop-
qi ¼ PN bj
ð7Þ
1þ erties (surface area, mean pore and pore volume) of the clays were
j¼1 bj C j
determined using the N2-BET method. The analysis was performed
(e) Modified Redlich–Peterson multicomponent isotherm in a surface area and porosimetry analyzer, Micromeritics, model
(MRPMI) with an interaction factor gi [29]: ASAP 2010, and the results are given in Table 2.
ai ðC i =gi Þ The crystalline species present in the natural clays were identi-
qi ¼ PN bj
ð8Þ fied by X-ray diffraction (XRD) analysis, and the XRD patterns were
1þ j¼1 bj ðC j =gj Þ
obtained with a diffractometer, Rigaku, DMAX 2000. The examina-
(f) Extended Freundlich multicomponent isotherm (EFMI) tion of the XRD patterns showed that the main crystalline species
[9,29]: present in the clay samples were Montmorillonite (main compo-

Table 3
Parameter values of the Langmuir, Freundlich and Redlich–Peterson isotherms for the single metal adsorption on bentonite (B), sepiolite (S) and vermiculite (V) at T = 25 °C.

Clay Metal pH Langmuir Freundlich Redlich–Peterson

qm (meq/g) K (L/meq) %D k (meq11/n L1/n/g) n %D a (L/meq) b (L/g)b b %D
B Ni 7 0.581 0.875 5.47 0.246 2.49 5.45 0.753 1.84 0.829 2.88
Cd 0.692 2.41 7.31 0.413 3.15 10.1 1.81 3.05 0.906 6.45
S Cd 7 0.403 5.71 12.3 0.279 3.44 18.6 4.86 14.8 0.836 5.00
Zn 0.265 6.65 2.21 0.247 3.11 1.44 570 2308 0.680 1.55
V Pb 4 0.579 0.984 11.2 0.255 3.18 7.71 1.52 4.48 0.797 7.84
Cu 0.513 0.686 6.32 0.205 2.45 3.05 0.768 2.66 0.737 2.99
538 E. Padilla-Ortega et al. / Chemical Engineering Journal 225 (2013) 535–546

nent of bentonite), sepiolite and vermiculite. Besides, the sepiolite 0.7

contained dolomite as an impurity.

Metal uptake on bentonite, meq/g

0.6
3.3. Determination of heavy metals in aqueous solutions
0.5

The concentrations of Cd(II), Cu(II), Ni(II), Pb(II) and Zn(II) in 0.4

water solutions were quantified by atomic absorption spectros-
copy. The absorbance of a sample was measured using a double- 0.3
beam atomic absorption spectrophotometer, Varian, model Spec-
trAA-20, and the concentration of a metal was calculated from a 0.2
calibration curve prepared with standard solutions of the heavy Cd(II)
0.1 Ni(II)
metal. All the heavy metal solutions were prepared from Cd(NO3)2
4H2O, Cu(NO3)22.5H2O, Ni(NO3)22.6H2O, Pb(NO3)2 and Zn(NO3)2 0.0
0 1 2 3 4 5 6
6H2O, respectively, and the solution pH was adjusted by mixing
proper volumes of 0.01 M HNO3 and NaOH solutions. Concentration of the metal at equilibrium, meq/L
(a)
3.4. Method for obtaining the adsorption equilibrium data
0.4
The experimental adsorption equilibrium data for the single and

Metal uptake on sepiolite, meq/g

binary metal systems were obtained in a batch adsorption mode.
The batch adsorber was a 50 mL conical vial. A volume of 40 mL 0.3
of a single or binary metal solution and 0.2 g of clay were poured
into the adsorber. The metal solution had a predetermined initial
concentration of one metal or two metals, and the solution was pre- 0.2
pared at a predetermined pH by mixing proper volumes of 0.01 M
HNO3 and NaOH solutions. The adsorber was placed in a thermo-
static bath, and the metal solution and clay were left in contact until 0.1
equilibrium was attained. The solution pH was measured frequently
Cd(II)
with a potentiometer, Termo Electron Corporation, model Orion 420
Zn(II)
A+, and was adjusted periodically by adding few drops of 0.01 M 0.0
HNO3 and NaOH solutions. The total volume of 0.01 M NaOH and 0.0 0.2 0.4 0.6 0.8 1.0
HNO3 solutions added to the adsorber was less than 0.3 mL. Conse- Concentration of the metal at equilibrium, meq/L
quently, the initial volume of the solution was assumed to remain (b)
constant during the adsorption. Preliminary runs showed that
7 days was enough to reach equilibrium. After 7 days a sample of 0.6
the solution was taken out, and the final concentration of each metal
Metal uptake on vermiculite, meq/g

was determined by the method previously described. The mass of 0.5

each metal adsorbed was estimated by performing a mass balance
of metal i that can be mathematically expressed as: 0.4

V 0 ðC i;0  C i;F Þ
qi ¼ ð13Þ 0.3
m

0.2
3.5. Procedure for performing the displacement experiments
0.1
The displacement experiment consisted of two stages. First, an Pb(II)
adsorption equilibrium experiment of metal i was carried out as Cu(II)
0.0
described above and the amount of metal i adsorbed was calcu- 0 2 4 6 8 10
lated using the Eq. (13). This stage was called as the first equilib- Concentration of the metal at equilibrium, meq/L
rium and the equilibrium conditions of metal i were denoted as (c)
qi,F1 and Ci,F1.
The procedure for the second stage was as follows. After Fig. 1. Single adsorption isotherms of (a) Cd(II) and Ni(II) on bentonite at pH = 7, (b)
approaching the first equilibrium, a sample of the adsorber solu- Cd(II) and Zn(II) on sepiolite at pH = 7 and (c) Cu(II) and Pb(II) on vermiculite at
pH = 4. The lines represent the Redlich–Peterson isotherm. T = 25 °C.
tion was taken out and an aliquot of a standard solution with a
known concentration of the metal j and the same pH as the adsor-
ber solution was poured into the adsorber solution. The volume of using Eq. (13) and the amount of metal i that was not displaced
the sample (1 mL) was the same as that of the aliquot to keep the or remained adsorbed was estimated by performing a mass bal-
volume of the adsorber solution constant (V = 40 mL). The initial ance of metal i:
concentration of the metal j in the adsorber solution was calculated
from the initial concentration of the standard solution of metal 2
V
and the dilution factor (1/40). The pH of the adsorber solution qi;F2 ¼ qi;F1 þ ðC i;F1  C i;F2 Þ ð14Þ
was kept constant as described earlier. The clay and binary metal m
solution were allowed to reach equilibrium, which was called sec- The first term in the right side of Eq. (14) represents the amount
ond equilibrium. Afterwards, the solution was sampled and the of metal i adsorbed in the first equilibrium, and the second term
concentrations of both metals at equilibrium were quantified as corresponds to the amount of metal i displaced or desorbed in
described before. The amount of metal j adsorbed was calculated the second equilibrium.
 E. Padilla-Ortega et al. / Chemical Engineering Journal 225 (2013) 535–546 539

Table 4
Metal selectivity and physicochemical property sequences.

Property Metal sequence

Bentonite Cd(II) > Ni(II) Sepiolite Cd(II) > Zn(II) Vermiculite Pb(II) > Cu(II)
Electronegativitya Ni(1.91) > Cd(1.69) Cd(1.69) > Zn(1.65) Pb(2.33) > Cu(1.9)
Ionic radius, Å Cd(0.97) > Ni(0.69) Cd(0.97) > Zn(0.74) Pb(1.19) > Cu(0.73)
Softnessb Cd(3.04) > Ni(2.82) Cd(3.04) > Zn(2.34) Pb(3.58) > Cu(2.89)
Hydrolysis constantc Ni(9.9) > Cd(10.1) Zn(9.0) > Cd(10.1) Pb(7.6)  Cu(7.5)
Ionization potential Cd(8.99) > Ni(7.64) Zn(9.39) > Cd(8.99) Cu(7.73) > Pb(7.42)
a
Pauling electronegativity.
b
Softness values from Misono et al. [39].
c
Hydrolysis constant, pK, is for the reaction: Me2+ + H2O = Me(OH)+ + H+.

Table 5
Parameters of the multicomponent isotherm models for the binary adsorption of Cd(II) and Ni(II) on bentonite, Cd(II) and Zn(II) on sepiolite, and Pb(II) and Cu(II) on vermiculite.

Model Single isotherm parameters Interaction factors Bentonite Sepiolite Vermiculite

Parameters %D Parameters %D Parameters %D
NLMI 4 0 49 87 20
MLMI 4 2 gCd = 1.85 11 gCd = 1.24 22 gPb = 1.05 10
gNi = 0.569 gZn = 0.088 gCu = 1.48
ELMI 0 3 qmax = 0.575 11 qmax = 0.274 24 qmax = 0.545 9
KE,Ni = 1.86 KE,Zn = 107 KE,Pb = 1.06
KE,Cd = 1.95 KE,Cd = 10.7 KE,Cu = 0.442
NRPMI 6 0 33 43 11
MRPMI 6 2 gCd = 1.42 10 gCd = 0.739 42 gPb = 0.828 8
gNi = 0.614 gZn = 0.584 gCu = 1.01
EFMI 4 6 xNi = 0.390 12 xZn = 0.105 24 xPb = 0.644 8
yNi = 0.152 yZn = 0.105 yPb = 0.137
zNi = 1.29 zZn = 1.13 zPb = 1.10
xCd = 0.546 xCd = 0.847 xCu = 0.462
yCd = 1.17 yCd = 5.02 yCu = 1.32
zCd = 0.672 zCd = 0.806 zCu = 0.621
SRSI 4 4 aCd,Cd = 1 15 aCd,Cd = 1 26 aPb,Pb = 1 12
aCd,Ni = 2.41 aCd,Zn = 12.8 aPb,Cu = 0.39
aNi,Ni = 1 aZn,Zn = 1 aCu,Cu = 1
aNi.Cd = 0.58 aZn,Cd = 0.026 aCu,Pb = 5.59

4. Results and discussion
4.1. Single metal adsorption isotherms on natural clays
The single metal and clay systems investigated in this work
were Cd(II) and Ni(II) on bentonite at pH = 7, Cd(II) and Zn(II) on
sepiolite at pH = 7, and Cu(II) and Pb(II) on vermiculite at pH = 4.
The solution pH was chosen as the pH where the clay presented
a reasonable adsorption capacity towards both metals, and no me-
tal precipitation occurred in the solution. In all the systems, the
temperature of the solution was 25 °C.
The Langmuir, Freundlich and Redlich–Peterson isotherms were
fitted to the experimental adsorption equilibrium data for a single
metal. The parameters or constants of these isotherms as well as
the average percentage deviations are shown in Table 3. The %D
ranged from 2.21% to 12.3%, 1.44% to 18.6% and 1.55% to 7.84%
for the Langmuir, Freundlich and Redlich–Peterson isotherms,
respectively. It was assumed that the isotherm model that best
fitted the adsorption equilibrium data was the model, which
presented the lowest %D. Accordingly to this criterion, the
Redlich–Peterson isotherm best fitted 4 out of the 6 single metal
clay systems shown in Table 3. Furthermore, the Freundlich model
provided slightly better interpretation of 2 of the 6 single metal
systems than the Redlich–Peterson Isotherm. Thus, it was consid-
ered that the Redlich–Peterson isotherm best fitted the single
adsorption equilibrium data. This result can be attributed to
the fact that the Redlich–Peterson model is a three-parameter
isotherm whereas the Freundlich and Langmuir models are two-
parameter isotherms.
The single metal adsorption equilibrium data and the Redlich–
Peterson isotherms of Cd(II) and Ni(II) on bentonite, Cd(II) and
Zn(II) on sepiolite, and Cu(II) and Pb(II) on vermiculite, are graphed
in Figs. 1a–c, correspondingly. As seen in Fig. 1a, the capacity of
bentonite for adsorbing Cd(II) is higher than that of Ni(II). For
example, the amount of Ni(II) and Cd(II) adsorbed on bentonite
was 0.44 and 0.64 meq/g, respectively, when the concentration of
the metal at equilibrium was 5 meq/L. The selectivity ratio, S(Cd/
Ni), can be calculated from the following equation:
qCdðIIÞ 0:64
SðCd=NiÞ ¼ ¼ ¼ 1:45 ð15Þ
qNiðIIÞ 0:44
Thus, the bentonite capacity towards Cd(II) was 1.45 times
greater than that towards Ni(II). The maximum capacity of an Ira-
nian bentonite for adsorbing Cd(II) was found to be 0.762 meq/g
(0.381 mmol/g) at pH = 6 [34], this capacity is higher than that re-
ported in this work. The differences in the adsorption capacities
can be attributed to the sources of the bentonite.
For a metal concentration at equilibrium of 0.8 meq/L, the
capacity of the sepiolite for adsorbing Cd(II) and Zn(II) was 0.33
and 0.22 meq/g, correspondingly, and the S(Cd/Zn) was approxi-
mately 1.5. This means that the Cd(II) capacity of the sepiolite
was 1.5 times greater than that of Zn(II). Alvarez-Ayuso and
García-Sánchez [35] found similar adsorption capacities for a
540 E. Padilla-Ortega et al. / Chemical Engineering Journal 225 (2013) 535–546

sepiolite from Zaragoza, Spain. In the case of the vermiculite, the
uptake of Pb(II) and Cd(II) was 0.50 and 0.46 meq/g, respectively,
when the metal concentration at equilibrium was 8 meq/L. There-
fore, the affinity of vermiculite towards Pb(II) was slightly higher
than that of Cu(II) and the S(Pb/Cu) was close to 1.1. However, a
vermiculite from Brazil presented the metal affinity order of
Cu(II) > Pb(II) [36]. Again, this discrepancy in the metal affinity
order may be due to the different origins of the vermiculite.
The metal affinity sequence of clays is normally associated with
the metal physicochemical properties. The hydrated ionic radii of
the metals increased in the following order: Pb (4.01 Å) > Ni
(4.04 Å) > Cu (4.19 Å) > Cd (4.26 Å) > Zn (4.30 Å) [37]. The discrep-
ancies in the adsorption capacities of the clays towards the metals
cannot be ascribed to accessibility of the metal cations to the clays
pores because the average pore diameter of the bentonite, sepiolite
The interpretation of the multicomponent experimental data
was also performed with the following Freundlich-type isotherm
models: the extended Freundlich multicomponent isotherm (EFMI)
and Sheindorf–Rebuhn–Sheintuch isotherm (SRSI). Both isotherm
models used the single component Freundlich isotherm parame-
ters and various interaction and competition factors. Based upon
the values of %D, the EFMI fitted the experimental data slightly bet-
ter than the SRSI because the %D of EFMI was lower than that of the
SRSI for the three binary metal-clay systems. As an example, for the
Pb(II)–Cu(II)/vermiculite system, the %D of EFMI was 8%, whereas
0.7
qpred = 1.1 q exp

Metal uptake predicted, meq/g

0.6
and vermiculite (see Table 2) was at least 14, 15 and 31 times
greater than the hydrated ionic radii of the metals. 0.5
Auboiroux et al. [38] used the Hard Soft Acid–Base (HSAB) the-
0.4 qpred = q exp
ory to explain the metal affinity sequence in a Ca-Montmorillonite,
and argued that the metal affinity sequence was due to differences 0.3
on electronegativity and softness of metal cations. The metal affin- qpred = 0.9 q exp
ity and the physicochemical property sequences for each clay are 0.2
shown in Table 4, and the properties were reported by Kinraide
MRPMI
and Yermiyahu [37]. Examination of the sequences in Table 4 re- 0.1
Ni(II)
veals that the affinity of the metals on the clays is dependent upon Cd(II)
0.0
the softness and ionic radius of the metal cations, but no upon 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Pauling electronegativity, hydrolysis constant and ionization po- Experimental metal uptake, meq/g
tential. No further examination of the affinity sequences was con-
ducted since the metal affinity order for each clay was only
(a)
obtained for two metals. 0.5
qpred = 1.1 q exp
4.2. Multicomponent adsorption isotherms
Metal uptake predicted, meq/g

0.4
qpred = q exp
The values of the multicomponent adsorption isotherm con-
stants and the average percentage deviation are shown in Table 5. 0.3
The multicomponent adsorption equilibrium data were interpreted
with the three multicomponent adsorption models based on the
0.2 qpred = 0.9 q exp
Langmuir isotherm. The values of the % D shown in Table 5 indi-
cated that the non-modified Langmuir multicomponent isotherm
(NLMI) cannot fit the experimental data since the average percent- 0.1
MLMI
age deviation ranged from 20% to 87%. This result can be explained Zn(II)
recalling that this isotherm does not incorporate any parameter Cd(II)
0.0
accounting for the interactions between metal ions of different 0.0 0.1 0.2 0.3 0.4 0.5
species and competition among metal cations. On the other hand, Experimental metal uptake, meq/g
the modified Langmuir multicomponent isotherm (MLMI) and ex-
(b)
tended Langmuir multicomponent isotherm (ELMI) adjusted rea-
sonably well the experimental data since both isotherms
0.7
provided very similar %D, ranging from 9% to 24%. However, the
qpred = q exp
MLMI was considered as the best option of these Langmuir-type
Metal uptake predicted, meq/g

0.6
models because the MLMI uses the single metal Langmuir isotherm
parameters. All the parameters of the ELMI have to be estimated by 0.5
fitting the multicomponent adsorption equilibrium data, and no qpred = 1.1 q exp
0.4
information of the single metal Langmuir isotherm is required.
The non-modified (NRPMI) and modified Redlich–Peterson 0.3
multicomponent isotherms (MRPMI) were fitted to the experimen- qpred = 0.9 q exp
tal adsorption equilibrium data, and the %D of the MRPMI was al- 0.2
ways lower than that of the NRPMI (see Table 5). For example, the MRPMI
MRPMI presented a %D of 10% for the Cd(II)–Ni(II)/bentonite sys- 0.1
Cu(II)
tem, whereas the %D of NRPMI was 33%. This result was attributed Pb(II)
0.0
to the fact that the NRPMI comprised 6 parameters from the single 0.0 0.1 0.2 0.3 0.4 0.5 0.6
component R–P isotherms; meanwhile, the MRPM included 6 Experimental metal uptake, meq/g
parameters from the single component R–P isotherm plus 2 inter-
action factors related to the multicomponent interactions. In other
(c)
words, the MRPMI better fitted the data because the multicompo- Fig. 2. Comparison between the metal uptake predicted with the best model and
nent interactions were incorporated into this model. This is why the experimental metal uptake. (a) Cd(II)–Ni(II) on bentonite, (b) Cd(II)–Zn(II) on
the MRPMI model was chosen over the NRPMI model. sepiolite and (c) Cu(II)–Pb(II) on vermiculite.
 E. Padilla-Ortega et al. / Chemical Engineering Journal 225 (2013) 535–546 541

that of SRSI was 12%. For a binary system, 6 and 4 additional fitting
parameters related to multicomponent interactions are required
by the EFMI and SRSI models. In the case of SRSI, the number of
additional parameters is only 2 instead of 4, since the parameters
a11 and a22 are equal to 1. Thus, the SRSI was considered to better
adjust the experimental data than the EFMI.
Based upon the lowest value of %D, the best models were
MRPMI, MLMI and MRPMI for the binary adsorption of Cd(II)–Ni(II)
on bentonite, Zn(II)–Cd(II) on sepiolite and Pb(II)–Cu(II) on vermic-
ulite, respectively. The values of %D for the above models were 10%,
22% and 8%, respectively.
The interpretation of the binary adsorption data was also ana-
lyzed by comparing the amount of metal adsorbed predicted with
the best model, qpred, vs. the experimental amount of metal ad-
sorbed, qexp. The plots qpred vs. qexp for the three binary-metal sys-
tems are shown in Fig. 2a–c. To help analyze the fitting results, the
straight lines qpred = qexp, qpred = 1.1 qexp and qpred = 0.9 qexp were
also graphed in these Figures. These lines represent exact predic-
tion, overpredicted by 10% deviation and underpredicted by 10%
deviation. The results plotted in Fig. 2a–c clearly showed that the
0.7
Uptake of Cd(II) adsorbed, meq/g
best models predicted reasonably well the experimental data.
The MRPMI model overpredicted the uptake of Cd(II) on bentonite,
but underpredicted the uptake of Ni(II) on bentonite. The uptake of
Zn(II) on sepiolite was overpredicted by MLMI model whereas the
Cd(II) uptake on sepiolite was underpredicted.
4.3. Binary adsorption of Cd(II) and Ni(II) on bentonite
The binary adsorption equilibrium data of Cd(II)–Ni(II) on ben-
tonite and the prediction of MPRMI model are plotted in Fig. 3a and
b. The effect of the presence of Ni(II) on the uptake of Cd(II) can be
seen in Fig. 3a, and in the concentration of Ni(II) ranging from 0 to
2 meq/L, the uptake of Cd(II) diminished drastically while increas-
ing the concentration of Ni(II) at equilibrium. Then, the uptake of
Cd(II) diminished slightly for concentrations of Ni(II) greater than
2 meq/L. For example, at a concentration of Cd(II) at equilibrium
of 5 meq/L, the uptake of Cd(II) adsorbed was 0.60, 0.41, 0.33 and
0.30 meq/g, when the concentration of Ni(II) at equilibrium was
of 0, 2, 4 and 5 meq/L, respectively. The capacity of the bentonite
for adsorbing Cd(II) decreased 32%, 20% and 9% when the concen-

Uptake of Zn(II) adsorbed, meq/g

0.05 0.10 0.15 0.20 0.25 0.30
0.6
0.5
0.4
0.3
0.2
0.1

1.
Co n 0 0 . 8
0

5 .6
7 4 c en .5 0
Conc
6 5 4 3 q/ L t ra t 0 . 6 0 0 .4 0 q/L
entrat 3 2 , me io n 0 .3 , me
ion of 1 0 1
2
f C d(II) .
of C 4 0 . 2 0 .2 Z n(II)
Ni(II) o d (I I . 1 no f
, meq 0 tion ), m 0 0
0
ratio
/L oncentra eq/L Concent
C
(a) (a)
0.7

0.20 0.25 0.30

Uptake of Ni(II) adsorbed, meq/g

Uptake of Cd(II) adsorbed, meq/g

0.6
0.5
0.4

0.0 0.05 0.10 0.15

0.3
0.2
0.1
0

5 10 0. 8 1 .0
8 0 .8
4
3 6 eq/L Conc 0.6 0 .6 eq/L
Conc
entrat 4 II) , m entrat 0 .4 0 .4 Cd(I
I), m
f Ni (
2
ion of 1 2 o io n of Z 0.2 0 .2 o f
Cd(II 0 0 tion n(II), 0 0
ratio
n
), meq/L c entra meq/L cent
Con Con
(b) (b)
Fig. 3. Binary adsorption isotherms of Cd(II)–Ni(II) on bentonite. The surfaces are Fig. 4. Binary adsorption isotherms of Cd(II)–Zn(II) on sepiolite. The surfaces are
predicted with the MRPMI model. (a) Cd(II) uptake and (b) Ni(II) uptake at pH = 7 predicted with the MLMI model. (a) Zn(II) uptake and (b) Cd(II) uptake. pH = 7 and
and T = 25 °C. T = 25 °C.
542 E. Padilla-Ortega et al. / Chemical Engineering Journal 225 (2013) 535–546
0.30
Uptake of Zn(II) adsorbed, meq/g
0.25
0.20
0.15
0.10
Concentration of Cd(II) on equilibrium, meq/L
0.05 0 0.2 0.4
0.6 0.8
0.00
0.0 0.1 0.2 0.3 0.4 0.5 0.6
Concentration of Zn(II) at equilibrium, meq/L
(a)
0.4
Concentration of Zn(II) at equilibrium, meq/L
Uptake of Cd(II) adsorbed, meq/g
0 0.2 0.4 0.6 1 ), m
0.3
0.2
0.1
0.0
0.0 0.2 0.4 0.6
Concentration of Cd(II) at equilibrium, meq/L
(b)
Fig. 5. (a) Adsorption isotherms of Zn(II) at various isoconcentrations of Cd(II) and
(b) Cd(II) at various isoconcentrations of Zn(II) on sepiolite at pH = 7 and T = 25 °C.
The lines represent the MLMI model.
tration of Ni(II) at equilibrium was varied from 0 to 2, 2 to 4 and 4
to 5 meq/L, correspondingly.
The adsorption isotherm surface of Ni(II) on bentonite is shown
in Fig. 3b and the dependence of the uptake of Ni(II) on the concen-
tration of Cd(II) at equilibrium is depicted in this figure. The uptake
of Ni(II) diminished slightly due to the competition of the Cd(II),
and the reduction of the Ni(II) uptake was almost proportional to
the enhancement of the Cd(II) concentration. At a Ni(II) concentra-
tion at equilibrium of 5 meq/L, the uptake of Ni(II) adsorbed was of
0.53, 0.39, 0.32 and 0.29 meq/g for a concentration of Cd(II) at
equilibrium of 0, 2, 4 and 5 meq/L, respectively. The capacity of
the bentonite for adsorbing Ni(II) was diminished 26%, 18% and
9% by raising the concentration of Cd(II) at equilibrium from 0 to
2, 2 to 4 and 4 to 5 meq/L, correspondingly.
The single adsorption isotherms of Cd(II) and Ni(II) on bentonite at
pH = 7 (Fig. 1a) indicated that the S(Cd/Ni) of the bentonite was 1.45
at concentration of metal at equilibrium of 5 meq/g. However, the
competitive adsorption data revealed that the S(Cd/Ni) was 1.03
when the concentrations of both metals at equilibrium were
5 meq/L. In the competitive adsorption, the uptake of Cd(II) decreased
to a greater extent than that of Ni(II). Therefore, the Ni(II) exhibited
greater antagonism to Cd(II) during the competitive adsorption.
4.4. Binary adsorption of Cd(II) and Zn(II) on sepiolite
Fig. 4a and b show the competitive adsorption equilibrium data
of Zn(II) in the presence of Cd(II) and that of Cd(II) in the presence
CEC = 0.436 meq/g
0.5
Total uptake of metals, qT, meq/g
0.4
0.3
0.2
1
0.1
0.7 0.8 0.9
0
1.0
Con 0.8
cen 0.6 0.6
trat 0.4 eq/L
ion 0.4 I), m
of C 0.2 0.2 f Zn(I
io n o
d(II
0 0
n centrat
eq/ L Co
Fig. 6. Total uptake of metals adsorbed on sepiolite at pH = 7 and T = 25 °C. The
surface is predicted by the MLMI model.
of Zn(II) on sepiolite at pH = 7 and T = 25 °C, respectively. Also, the
adsorption isotherm surfaces of Cd(II) and Zn(II) adsorbed on sepi-
olite were drawn with the MLMI model. This model was used suc-
cessfully by Ho and McKay [40] and Leyva-Ramos et al. [41] to
0.8 1.0 represent the binary adsorption isotherms of Co(II) and Ni(II) on
peat, and Cd(II) and Zn(II) on activated carbon, respectively.
The adsorption isotherm surface of Zn(II) (Fig. 4a) shows that
the uptake of Zn(II) does not change by varying the concentration
of Cd(II) at equilibrium. For example, at a concentration of Zn(II) at
equilibrium of 0.8 meq/L, the uptake of Zn(II) adsorbed was
smoothly reduced from 0.26 to 0.24 meq/g when the concentration
of Cd(II) was raised from 0 to 0.8 meq/L.
As seen in Fig. 4b, the adsorption isotherm surface of Cd(II)
exhibits a very different behavior. The uptake of Cd(II) adsorbed
was substantially decreased by increasing the concentration of
Zn(II). For example, at an equilibrium concentration of Cd(II) of
0.8 meq/L, the amount of Cd(II) adsorbed was 0.32, 0.075, 0.042,
0.03 and 0.02 meq/g for a concentration of Zn(II) of 0, 0.2, 0.4, 0.8
and 1 meq/L, respectively. This means that the Cd(II) adsorption
capacity of sepiolite was reduced near 77% when the concentration
at equilibrium of Zn(II) was slightly changed from 0 to 0.2 meq/L.
The adsorption isotherms of Zn(II) at various isoconcentrations
of Cd(II), and the adsorption isotherms of Cd(II) at various isocon-
centrations of Zn(II) are shown in Fig. 5a and b, respectively. The
isotherms of Zn(II) or Cd(II) were predicted with the MLMI model
by keeping constant the concentration of the other metal. These
graphs are very useful for observing the decrease in the uptake
of a metal by the presence of the other metal. Fig. 5a shows that
the Zn(II) adsorption capacity of the sepiolite was slightly affected
by the presence of Cd(II); however, in the Fig. 5b, it can be clearly
observed that at a concentration of Cd(II) at equilibrium of
0.8 meq/L, the uptake of Cd(II) adsorbed diminished from 0.33 to
0.02 meq/g, increasing the concentration of Zn(II) from 0 to
0.8 meq/L. This means that the capacity of the sepiolite for adsorb-
ing Cd(II) was decreased 16.5 times due to the competition of
Zn(II).
At the concentrations at equilibrium of 0.8 meq/L, the S(Cd/Zn)
was 1.5 for the single adsorption isotherms of Cd(II) and Zn(II),
whereas the S(Cd/Zn) was 0.08 for the competitive adsorption.
Hence, the Zn(II) exhibited strong antagonism in the competitive
 E. Padilla-Ortega et al. / Chemical Engineering Journal 225 (2013) 535–546 543

0.7
Second Equilibrium of Pb(II) for [Cu(II)] 0 (meq/L):

Uptake of Pb(II) adsorbed, meq/g

0.6 1.3 3.2 6.3 12.5
0.6

Single Adsorption
0.5 Isotherm of Pb(II)
Uptake of Cu(II) adsorbed, meq/g
0.5

0.4
0.4

0.3

0.2
0.3

0.1
0.2

First equilibrium of Pb(II)

0.0
0 1 2 3 4 5 6 7
0.1

Concentration of Pb(II) at equilibrium, meq/L

(a)
0

12
10 12
Co n 10
cen 8 6 8 0.7
trat 6 eq/L
ion
of P 4 2 II), m Second Equilibrium of Cu(II) for [Pb(II)]0 (meq/L):

Uptake of Cu(II) adsorbed, meq/g

4 C u (
b(II 2 n of 0.6 1.3 3.2 6.3 12.5
), m 0 0 ratio
eq/L o n cent
C 0.5
Single Adsorption
Isotherm of Cu(II)

(a) 0.4

0.3

0.2
0.6

0.1
Uptake of Pb(II) adsorbed, meq/g
0.5

0.0
0 2 4 6 8 10 12
Concentration of Cu(II) at equilibrium, meq/L
0.4

(b)
0.3

Fig. 8. Displacement of Pb(II) adsorbed on vermiculite by adsorption of Cu(II) at

pH = 4 and T = 25 °C. (a) Uptake of Pb(II) and (b) uptake of Cd(II).
0.2

4.5. Binary adsorption of Cu(II) and Pb(II) on vermiculite

0.1

At pH = 4 and T = 25 °C, the experimental competitive adsorp-

0

12
10 12 tion data of Cu(II) and Pb(II) on vermiculite as well as the adsorp-
Con 10
cen 8 8 tion surface predicted with the MRPMI model are illustrated in
trat /L
ion 6 6 meq Fig. 7a and b.
of C 4 2 4 Pb (II),
u(II 2 n of The effect of the presence of Pb(II) on the capacity of the ver-
), m 0 0 ratio
eq/L on cent miculite for adsorbing Cu(II) is graphed in Fig. 7a. The competitive
C
adsorption surface of Cu(II) shows that the presence of Pb(II) re-
(b) duced drastically the uptake of Cu(II) adsorbed on the vermiculite.
A very large reduction of the uptake of Cu(II) adsorbed can be
Fig. 7. Binary adsorption isotherms of Cu(II)–Pb(II) on vermiculite. The surfaces are
noted when the concentration of Pb(II) at equilibrium was varied
predicted with the MRPMI model. (a) Cu(II) uptake and (b) Pb(II) uptake. pH = 4 and
T = 25 °C. in the range from 0 to 3 meq/L. The decrease of the uptake of Cu(II)
was less pronounced for a Pb(II) concentration above 3 meq/L. For
example, at a concentration of Cu(II) at equilibrium of 8 meq/L, the
adsorption of Cd(II), whereas the Cd(II) did not affect the compet- uptake of Cu(II) was 0.46, 0.23, 0.17 and 0.15 meq/g when the con-
itive adsorption of Zn(II) or did not present antagonism. centration of Pb(II) at equilibrium was 0, 3, 6 and 8 meq/L. These
The total metal uptake on the sepiolite is calculated by adding results reveal that the vermiculite capacity for adsorbing Cu(II) de-
the amount of Cd(II) adsorbed and the amount of Zn(II) adsorbed creased 50% and 12% by raising the concentration of Pb(II) from 0
on sepiolite. The total metal uptake is represented as qT = qCd(II) + to 3 meq/L and from 6 to 8 meq/L, correspondingly.
qZn(II). The adsorption surface of qT is shown in Fig. 6, and this The binary adsorption surface of Pb(II) in the presence of Cu(II)
adsorption surface was graphed using the MLMI models of Cd(II) is depicted in Fig. 7b and the uptake of Pb(II) adsorbed on vermic-
and Zn(II). It can be observed that the adsorption surface levels ulite was very slightly affected by the competition of Cu(II). At a
off at qT = 0.27 meq/g. Hence, both metals are competing for the concentration of Pb(II) of 8 meq/L, the amount of Pb(II) adsorbed
same cationic sites of the sepiolite. Besides, qT = 0.27 meq/g is was 0.52, 0.43, 0.38 and 0.37 meq/g while the concentration of
shorter than the CEC capacity of the sepiolite (CEC = 0.44 meq/g). Cu(II) was 0, 3, 6 and 8 meq/L, respectively. These results revealed
This result can be attributed to the fact that some cationic sites that the capacity of vermiculite for adsorbing Pb(II) was reduced
of the clays are not available or accessible to both Cd2+ and Zn2+ 17% and 3% by varying the concentration of Cu(II) from 0 to 3,
ions. and 6 to 8 meq/L, respectively.
544 E. Padilla-Ortega et al. / Chemical Engineering Journal 225 (2013) 535–546

0.7 0.7
Second Equilibrium of Cu(II) with [Pb(II)] 0 (meq/L): Cu(II)

Metal uptake predicted with MRPMI, meq/g

qpred = 1.1 qexp
Uptake of Cu(II) adsorbed, meq/g

1.3 3.2 6.3 12.5 Pb(II)

0.6
0.6
First Equilibrium of Cu(II)
0.5 qpred = qexp
Single Adsorption 0.5
Isotherm of Cu(II)
0.4
0.4
0.3
qpred = 0.9 qexp
0.3
0.2

0.1 0.2

0.0 0.1
0 2 4 6 8 10 12
Concentration of Cu(II) at equilibrium, meq/L
0.0
(a) 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Experimental metal uptake, meq/g
0.7
Second Equilibrium of Pb(II) with [Cu(II)]0 (meq/L):
(a)
Uptake of Pb(II) adsorbed, meq/g

0.6 1.3 3.2 6.3 12.5

0.6

Metal uptake predicted with MRPMI, meq/g

0.5 qpred = 1.1 qexp
0.5
0.4

0.3 0.4 qpred = qexp

0.2
0.3
0.1

0.0 0.2 qpred = 0.9 qexp

0 2 4 6
Concentration of Pb(II) at equilibrium, meq/L
(b)
Fig. 9. Displacement of Cu(II) adsorbed on vermiculite by adsorption of Pb(II) at
pH = 4 and T = 25 °C. (a) Uptake of Cu(II) and (b) uptake of Pb(II).
The results of the competitive adsorption of Pb(II)–Cu(II) onto
vermiculite have demonstrated that the Pb(II) ions presented a
higher affinity for the cationic sites of the vermiculite than the
Cu(II) ions. For metal concentrations at equilibrium of 8 meq/L,
the single adsorption isotherms indicated that the S(Pb/Cu) of ver-
miculite was 1.1 (see Fig. 1c); however, the S(Pb/Cu) was approxi-
mately 2.5 for the binary adsorption. In other words, the
vermiculite was much more selective towards Pb(II) than to Cu(II)
in the competitive adsorption. The Pb(II) ions presented strong
antagonism against the adsorption of Cu(II), whereas the Cu(II)
ions exhibited light antagonism against the adsorption of Pb(II).
This behavior cannot be predicted from the single metal adsorption
isotherms.
4.6. Displacement of Cu(II) and Pb(II) adsorbed on the vermiculite
The displacement of a metal adsorbed on vermiculite was per-
formed by adsorbing a single metal on vermiculite, and attaining
adsorption equilibrium. Subsequently, the metal adsorbed on the
vermiculite was displaced from the adsorption sites by adding an-
other metal into the adsorber solution. The displacement or
desorption of the metal adsorbed was attributed to the competi-
tion by the other metal for the adsorption sites.
4.6.1. Displacement of the Pb(II) adsorbed on the vermiculite
The single adsorption equilibrium data and isotherm of Pb(II)
on vermiculite is plotted in Fig. 8a. The data and isotherm are de-
noted as the first equilibrium since in the first stage of the displace-
8 10 12
0.1
Cu(II)
Pb(II)
0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6
Experimental metal uptake, meq/g
(b)
Fig. 10. Comparison between the metal uptake predicted with the MRPMI model
and the experimental metal uptake. (a) Displacement of Pb(II) adsorbed on
vermiculite by Cu(II), and (b) displacement of Cd(II) adsorbed on vermiculite by
Pb(II).
ment experiments only Pb(II) was present in the solution. The
competitive adsorption occurred when Cu(II) was added to the
solution of Pb(II) at equilibrium. The Pb(II) adsorbed on the vermic-
ulite was displaced from the adsorption sites so that Cu(II) can be
adsorbed on the cationic sites. The experimental displacement
adsorption data is denoted as the second equilibrium and the up-
take of Pb(II) at the second equilibrium was diminished by the
competition of the Cu(II). As shown in Fig. 8a, the amount of Pb(II)
adsorbed at the second equilibrium was dependent on the initial
concentration of Cu(II) because greater competition of Cu(II) oc-
curred by increasing the initial concentration of Cu(II) in the solu-
tion. At an initial concentration of Cu(II) of 1.3 meq/L, the uptake of
Pb(II) at the second equilibrium was very similar to that at the first
equilibrium. Hence, the Pb(II) adsorbed on the vermiculite was not
displaced by Cu(II) for an initial concentration of Cu(II) below
1.3 meq/L of Cu(II). However, the uptake of Pb(II) was reduced con-
siderably when the initial concentration of Cu(II) was 12.5 meq/L.
For example, at a concentration of Pb(II) at equilibrium of 2 meq/
L, the uptake of Pb(II) was almost halved when the initial concen-
tration of Cu(II) was 12.5 meq/L.
 E. Padilla-Ortega et al. / Chemical Engineering Journal 225 (2013) 535–546
On the other hand, the adsorption equilibrium data of Cu(II) at
the second equilibrium is graphed in Fig. 8b. No data was reported
at the first equilibrium since no Cu(II) was present in the first stage
of the displacement experiments. As a reference, the single adsorp-
tion isotherm of Cu(II) on vermiculite is also shown in Fig. 8b. The
uptake of Cu(II) at the second equilibrium was always below the
single adsorption isotherm and as expected, the uptake of Cu(II)
was diminished by raising the initial concentration of Pb(II) in
the first equilibrium.
4.6.2. Displacement of the Cu(II) adsorbed on the vermiculite
In Fig. 9a are plotted the experimental single adsorption equi-
librium data of Cu(II) as well as the single adsorption isotherm of
Cu(II). This information is from the first stage or first adsorption
equilibrium of Cu(II) on the vermiculite. In the second stage, the
uptake of Cu(II) adsorbed was decreased because the Cu(II) was
displaced by the Pb(II) added to the solution. The uptake of Cu(II)
at the second equilibrium was dependent on the initial concentra-
tion of Pb(II), and was always shorter than the uptake of single
Cu(II). A comparison between Figs. 8a and 9a shows that a larger
amount of Cu(II) adsorbed is displaced by the Pb(II) than the other
way around. Moreover, the amount of Pb(II) adsorbed on vermicu-
lite at the second equilibrium is exhibited in Fig. 9b and the uptake
of Pb(II) at the second equilibrium decreased slightly increasing
the initial concentration of Cu(II). The Pb(II) displaced the Cu(II) ad-
sorbed on the vermiculite at low and high initial concentrations of
Pb(II).
4.6.3. Prediction of the displacement equilibrium
The metal adsorbed in the first equilibrium was displaced by
the other metal, and subsequently, both metals adsorbed on the
vermiculite and attained the binary adsorption equilibrium de-
noted as second or displacement equilibrium. It is expected that
the binary adsorption or displacement equilibrium can be pre-
dicted with the MRPMI model. The comparison between the metal
uptake predicted with the MRPMI model, and the experimental
metal uptakes in the second equilibrium are depicted in Fig. 10a
and b.
As seen in Fig. 10a, the uptake of Cu(II) and Pb(II) adsorbed on
vermiculite is predicted properly well with the MRPPI model when
the Pb(II) was displaced by the Cu(II). In the case of the Cu(II) dis-
placed by Pb(II), the MRPMI model predicted suitably well the up-
take of Cu(II) adsorbed, but the MRPMI model underpredicted the
uptake of Pb(II) adsorbed. Thus, the Pb(II) uptake predicted with
the MRPMI model was lower than the experimental Pb(II) uptake,
indicating that the Pb(II) adsorbed on sites occupied by Cu(II) as
well as empty sites.
5. Conclusions
The binary adsorption equilibrium data of heavy metals can be
predicted reasonably well by the multicomponent adsorption iso-
therm models reported in the literature. The binary adsorption
data of Cd(II)–Ni(II) on bentonite, Cd(II)–Zn(II) on sepiolite and
Cu(II)–Pb(II) on vermiculite were best interpreted with the MRPMI,
MLMI and MRPMI models, respectively.
In the competitive adsorption of Cd(II) and Ni(II) on bentonite,
both metals presented very similar antagonism against the adsorp-
tion of the other metal. In the single metal adsorption, the affinity
of Cd(II) for the bentonite was greater than that of Ni(II); however,
in the competitive adsorption, the affinities of both metals towards
the bentonite were very similar.
The single metal adsorption on sepiolite showed that the capac-
ity of sepiolite for adsorbing Cd(II) was 1.3 times greater than that
of Zn(II), whereas the competitive adsorption of Cd(II) and Zn(II) on
545
sepiolite presented a very different behavior. In the binary adsorp-
tion, it was noted that Zn(II) presented a strong antagonism against
the adsorption of Cd(II); however, Cd(II) did not affected the
adsorption of Zn(II). The uptake of Cd(II) in the presence of Zn(II)
was reduced up to 16.5 times due to the competition of the Zn(II)
for the same cationic sites of the sepiolite.
The single component adsorption isotherms of Cu(II) and Pb(II)
on vermiculite revealed that the selectivity of the vermiculite to-
wards Pb(II) was slightly higher than that of the Cu(II). However,
in the competitive adsorption, the affinity of Pb(II) for the vermic-
ulite was much higher than that of the Cu(II). The Pb(II) presented
stronger antagonism against the simultaneous adsorption of Cd(II),
but not in the other way around.
The displacement experiments of Cu(II) and Pb(II) adsorbed on
vermiculite by the other metal showed that both metals can be
desorbed by the presence of the other metal. Furthermore, the
amount of Cu(II) displaced by the adsorption of Pb(II) is greater
than the amount of Pb(II) displaced by the adsorption of Cu(II).
The amount of Cu(II) and Pb(II) displaced was up to 88% and 58%,
respectively.
References
[1] F. Fu, Q. Wang, Removal of heavy metal ions from wastewaters: a review, J.
Environ. Manage. 92 (2011) 407–418.
[2] S. Liang, X. Guo, Q. Tian, Adsorption of Pb2+ and Zn2+ from aqueous solutions by
sulfured orange peel, Desalination 275 (2011) 212–216.
[3] F. Pagnanelli, A. Esposito, F. Vegliò, Multi-metallic modeling for biosorption of
binary systems, Water Res. 36 (2002) 4095–4105.
[4] J.C. Moreno-Piraján, R. Gómez-Cruza, V.S. García-Cuello, L. Giraldo, Binary
system Cu(II)/Pb(II) adsorption on activated carbon obtained by pyrolysis of
cow bone study, J. Anal. Appl. Pyrol. 89 (2010) 122–128.
[5] M. Machida, M. Aikawa, H. Tatsumoto, Prediction of simultaneous adsorption
of Cu(II) and Pb(II) onto activated carbon by conventional Langmuir type
equations, J. Hazard. Mater. B120 (2005) 271–275.
[6] M. Alimohamadi, G. Abolhamd, A. Keshtkar, Pb(II) and Cu(II) biosorption on
Rhizopus arrhizus modeling mono- and multi-component systems, Miner. Eng.
18 (2005) 1325–1330.
[7] M.S. Berber-Mendoza, R. Leyva-Ramos, J. Mendoza-Barron, R.M. Guerrero-
Coronado, Competitive exchange of lead(II) and cadmium(II) from aqueous
solution on clinoptilolite, Studies Surf. Sci. Catal. 142 (2002) 1849–1856.
[8] R. Apiratikul, P. Pavasant, Sorption isotherm model for binary component
sorption of copper, cadmium, and lead ions using dried green macroalga,
Caulerpa lentillifera, Chem. Eng. J. 119 (2006) 135–145.
[9] S.K. Papageorgiou, F.K. Katsaros, E.P. Kouvelos, N.K. Kanellopoulos, Prediction
of binary adsorption isotherms of Cu2+, Cd2+ and Pb2+ on calcium alginate
beads from single adsorption data, J. Hazard. Mater. 162 (2009) 1347–1354.
[10] C. Mahamadi, T. Nharingo, Competitive adsorption of Pb2+, Cd2+ and Zn2+ ions
onto Eichhornia crassipes in binary and ternary systems, Biores. Technol. 101
(2010) 859–864.
[11] F. Barbier, G. Duc, M. Petit-Ramel, Adsorption of lead and cadmium ions from
aqueous solution to the montmorillonite/water interface, Colloid. Surf A
Physicochem. Eng. Aspects 166 (2000) 153–159.
[12] A. Chakir, J. Bessiere, K.E.L. Kacemi, B. Marouf, A comparative study of the
removal of trivalent chromium from aqueous solution by bentonite and
expanded perlite, J. Hazard. Mater. B95 (2002) 29–46.
[13] A.S. Khan, R.-U. Rehman, M.A. Khan, Adsorption of chromium (III), chromium
(VI) and silver (I) on bentonite, Waste Manage. 15 (4) (1995) 271–282.
[14] S. Kubilay, R. Gürkan, A. Savran, T. Sahan, Removal of Cu(II), Zn(II) and Co(II)
ions from aqueous solutions by adsorption onto natural bentonite, Adsorption
13 (2007) 41–51.
[15] D. Ozdes, C. Duran, H.B. Senturk, Adsorptive removal of Cd(II) and Pb(II) ions
from aqueous solutions by using Turkish illitic clay, J. Environ. Manage. 92
(2011) 3082–3090.
[16] G. Purna Chandra Rao, S. Satyaveni, A. Ramesh, K. Seshaiah, K.S.N. Murthy, N.V.
Choudary, Choudary, Sorption of cadmium and zinc from aqueous solutions by
zeolite 4A, zeolite 13X and bentonite, J. Environ. Manage. 81 (2006) 265–272.
[17] S. Sen Gupta, K.G. Bhattacharyya, Immobilization of Pb(II), Cd(II) and Ni(II) ions
on kaolinite and montmorillonite surfaces from aqueous medium, J. Environ.
Manage. 87 (2008) 46–58.
[18] S.S. Tahir, R. Naseem, Removal of Cr(III) from tannery wastewater by
adsorption onto bentonite clay, Sep. Purif. Technol. 53 (2007) 312–321.
[19] M. Ulmanu, E. Marañón, Y. Fernández, L. Castrillón, I. Anger, D. Dumitriu,
Removal of copper and cadmium ions from diluted aqueous solutions by low
cost and waste material adsorbents, Water, Air and Soil Pollut. 142 (2003)
357–373.
[20] B.S. Krishna, D.S.R. Murty, B.S. Jai Prakash, Thermodynamics of chromium(VI)
anionic species sorption onto surfactant-modified montmorillonite clay, J.
Colloid Interf. Sci. 229 (2000) 230–236.
546 E. Padilla-Ortega et al. / Chemical Engineering Journal 225 (2013) 535–546
[21] K.G. Bhattacharyya, S. Sen Gupta, Adsorption of a few heavy metals on natural
and modified kaolinite and montmorillonite: a review, Adv. Colloid Interf. Sci.
140 (2008) 114–131.
[22] L. Zhi-rong, Z. Shao-qi, Adsorption of copper and nickel on Na-bentonite,
Process Safety Environ. Protec. 88 (2010) 62–66.
[23] Ö. Etci, N. Bektasß M.S. Öncel, Single and binary adsorption of lead and
cadmium ions from aqueous solution using the clay mineral beidellite,
Environ. Earth Sci. 61 (2010) 231–240.
[24] P. Srivastava, B. Singh, M. Angove, Competitive adsorption behavior of heavy
metals on kaolinite, J. Colloid. Interf. Sci. 290 (2005) 28–38.
[25] A. Sdiri, T. Higashi, R. Chaabouni, F. Jamoussi, Competitive removal of heavy
metals from aqueous solutions by montmorillonitic and calcareous clays,
Water Air Soil Pollut. 223 (2012) 1191–1204.
[26] M. Eloussaief, W. Hamza, N. Kallel, M. Benzina, Wastewaters decontamination:
mechanisms of Pb(II), Zn(II), and Cd(II) competitive adsorption on Tunisian
smectite in single and multi-solute systems, Environ. Prog. Sustain. Energy,
doi: 10.1002/ep.11609.
[27] R.T. Yang, Gas Separation by Adsorption Processes, Butterworth, Boston, 1987.
[28] V.C. Srivastava, I.D. Mall, I.M. Mishra, Equilibrium modeling of single and
binary adsorption of cadmium and nickel onto bagasse fly ash, Chem. Eng. J.
117 (2006) 79–91.
[29] V.C. Srivastava, I.D. Mall, I.M. Mishra, Modeling individual and competitive
adsorption of cadmium (II) and zinc (II) metal ions from aqueous solution onto
bagasse fly ash, Sep. Sci. Technol. 41 (2006) 2685–2710.
[30] V.C. Srivastava, I.D. Mall, I.M. Mishra, Removal of cadmium (II) and zinc(II)
metal ions from binary aqueous solution by rice husk ash, Colloid. Surf.
Physicochem. Eng. Aspects 312 (2008) 171–184.
[31] C. Sheindorf, M. Rebhun, M. Sheintuch, A Freundlich-type multicomponent
isotherm, J. Colloid Interf. Sci. 79 (1) (1981) 136–142.
[32] S.L. Rettig, J.W. Marinenko, H.N. Khoury, B.F. Jones, Comparison of rapid
methods for chemical analyses of milligram samples of ultrafine clays, Clays
Clay Miner. 31 (1983) 440–446.
[33] D.W. Ming, J.B. Dixon, Quantitative determination of clinoptilolite in soils by a
cation-exchange capacity method, Clays Clay Miner. 35 (1987) 463–468.
[34] H. Faghihian, M. Nejati-Yazdinejad, A comparative study of the sorption of
Cd(II) and Pb(II) ions from aqueous solution by local bentonite and
clinoptilolite, Adsorp. Sci. Technol. 27 (2009) 107–115.
[35] E. Alvarez-Ayuso, A. Garcia-Sanchez, Sepiolite as a feasible soil additive for the
immobilization of cadmium and zinc, Sci. Total Environ. 305 (2003) 1–12.
[36] M.G. da Fonseca, M.M. de Oliveira, L.N.H. Arakaki, J.G.P. Espinola, C. Airoldi,
Natural vermiculite as an exchanger support for heavy cations in aqueous
solution, J. Colloid Interf. Sci. 285 (2005) 50–55.
[37] T.B. Kinraide, U. Yermiyahu, A scale of metal ion binding strengths correlating
with ionic charge, Pauling electronegativity, toxicity, and other physiological
effects, J. Inorg. Biochem. 101 (2007) 1201–1213.
[38] M. Auboiroux, F. Melou, F. Bergaya, J.C. Touray, Hard and soft acid-base model
applied to bivalent cation selectivity on a 2:1 clay mineral, Clays Clay Miner.
46 (1998) 546–555.
[39] M. Misono, E. Ochiai, Y. Saito, Y. Yoneda, A new dual parameter scale for the
strength of Lewis acids and bases with the evaluation of their softness, J. Inorg.
Nucl. Chem. 29 (1967) 2685–2691.
[40] Y.S. Ho, G. McKay, Competitive sorption of copper and nickel ions from
aqueous solution using peat, Adsorption 5 (1999) 409–417.
[41] R. Leyva-Ramos, L.A. Bernal-Jacome, L. Fuentes-Rubio, R.M. Guerrero-
Coronado, Competitive adsorption of Cd(II) and Zn(II) from aqueous solution
onto activated carbon, Sep. Sci. Technol. 36 (16) (2001) 3673–3687.