Kinetics, adsorption and desorption of Cd(II) and Cu(II) on natural

Geoderma 319 (2018) 70–79
Contents lists available at ScienceDirect
Geoderma
journal homepage: www.elsevier.com/locate/geoderma
Kinetics, adsorption and desorption of Cd(II) and Cu(II) on natural
allophane: Effect of iron oxide coating
⁎
T
⁎⁎
Jorge Silva-Yumia,b, , Mauricio Escudeya,c, , Manuel Gacituaa,c, Carmen Pizarroa,c
a
Facultad de Química y Biología, Universidad de Santiago de Chile, Av. B. O'Higgins, 3363, Santiago, Chile
Facultad de Ciencias, Escuela Superior Politécnica de Chimborazo, Grupo GETNano, Panamericana Sur km 1 1/2, Riobamba, Ecuador
c
Centro para el Desarrollo de la Nanociencia y Nanotecnología, CEDENNA, Santiago, Chile
b
A R T I C L E I N F O
A B S T R A C T
Handling Editor: Edward A Nater
Volcanic soils are a potential source of abundant and low-cost adsorbent materials. Allophane and iron oxides
are the main inorganic constituents of Andisols, soils derived from volcanic-ashes abundant in the central and
southern continental Chilean territory. The present study aimed to elucidate the effect of free iron oxides on the
adsorption of Cd(II) and Cu(II) from aqueous media on natural allophane (NA). Allophane samples consisted of
natural allophane, obtained from a Chilean Andisol series, with (NA-FeOx) and without (NA) an iron oxide
coating. Material characterisation showed that NA-FeOx was formed by nanometric (100–500 nm) sized spheres
coated with an iron oxide layer that most likely consisted of a ferrihydrite-like mineral. On the other hand, NA
consisted of aggregates (with particles as big as 1.0 μm) mostly constituted by allophane. The specific surface
area of NA-FeOx and NA were ca. 150 and 17 m2 g− 1, respectively, and presented variable surface charge with
an isoelectric point of 10.3. and 5.4. From the kinetic studies, equilibrium adsorption was achieved within
60 min and was almost independent of the presence of the iron oxides. Among the kinetic models evaluated, the
pseudo-second order model presented a better correlation with the experimental data. The adsorption studies
revealed that NA has a greater adsorption capacity of both Cd+ 2 and Cu+ 2 compared to NA-FeOx. Langmuir and
Freundlich adsorption models fit the experimental data well, suggesting that the adsorption process is a combination of physical and chemical phenomena. The Freundlich constant (adsorption coefficient, KF) values obtained confirmed the superior affinity of NA active surface sites compared to NA-FeOx. Desorption studies demonstrated that the majority of adsorbed cations were retained by the adsorbent, indicating irreversible
chemical adsorption on specific active sites of the adsorbents.
Keywords:
Volcanic ash derived soils
Iron oxide coating
Metal adsorption
Adsorption-desorption modelling
1. Introduction
Pollution by metals is a growing problem due to industrial activities
and anthropogenic activities. The main sources of heavy metals are
mining and refining processes; the use of pigments, plastics, and Cd-Ni
batteries; the foundry and electroplating industry, as well as the intensive use of phosphate fertilizers and pesticides (Hashim et al., 2011;
Ihsanullah et al., 2016; Molina et al., 2009; Uddin, 2017). Once introduced into the environment, metals permeate air, soil, and water
compartments. As metals are not biodegradable, they may accumulate
in organisms and are toxic to the aquatic flora and fauna as well as to
human beings (Hegazi, 2013; Uddin, 2017). Remediation technologies
for the removal of these pollutants from aqueous environments include
chemical precipitation, electrochemical treatments, oxidation, reduction, coagulation, ion exchange, membrane filtration, bioremediation,
⁎
phytoremediation, bioadsorption, and adsorption (Hashim et al., 2011;
Hegazi, 2013; Hua et al., 2012; Ihsanullah et al., 2016; Mahar et al.,
2016; Mosa et al., 2016). Among these options, adsorption processes
have garnered much attention due to their operation feasibility and
simple design, among other attributes.
One of the challenges of adsorption is identifying low cost materials
with a high adsorption capacity. As a result, adsorbents obtained from
natural sources have gained much consideration (De Gisi et al., 2016).
In this context, soils in general and particularly Andisols (soils of volcanic origin) constitute a potential economic source of absorbent materials due to their natural allophane content (De Gisi et al., 2016;
Kaufhold et al., 2009; Yuan and Wada, 2012). Allophane is an aluminosilicate with predominant short-range organisation that presents
large specific surface area and high buffering capacity, among other
properties. Allophane is known to have variable surface charge, which
Corresponding author at: Facultad de Ciencias, Escuela Superior Politécnica de Chimborazo, Grupo GETNano, Panamericana Sur km 1 1/2, Riobamba, Ecuador.
Corresponding author at: Centro para el Desarrollo de la Nanociencia y Nanotecnología, CEDENNA, Santiago, Chile.
E-mail addresses: [email protected] (J. Silva-Yumi), [email protected] (M. Escudey), [email protected] (M. Gacitua), [email protected] (C. Pizarro).
⁎⁎
https://doi.org/10.1016/j.geoderma.2017.12.038
Received 29 June 2017; Received in revised form 2 November 2017; Accepted 31 December 2017
0016-7061/ © 2018 Elsevier B.V. All rights reserved.
Geoderma 319 (2018) 70–79
J. Silva-Yumi et al.
2.1.4. Removal of free iron oxides
Removal of free iron oxides was performed according to the procedure of Mehra and Jackson (Mehra and Jackson, 1960) but modified
using the dithionite citrate bicarbonate (DCB) mixture. A 50 mL sample
of the < 0.002 mm fraction was mixed with 100 mL of sodium citrate
(0.3 M) and 12.5 mL of 1 N sodium bicarbonate in a thermostatic water
bath set at < 80 °C. Following addition of 1 g of sodium dithionite, this
mixture was allowed to react during 15 min with occasional stirring.
The solid was purified using centrifuge and re-suspended in 50 mL of
double distilled water. Purification steps were repeated four times. Finally, to remove the organic anions (citrate) adsorbed on the solids, the
samples were washed with H2O2 (Escudey et al., 1986). The final
sample was considered to be natural allophane (NA).
may be positive or negative depending on the conditions of the medium
(mainly pH), facilitating interactions with positive or negative ions
through ion exchange or adsorption reactions (Bolan et al., 2013).
Although allophane is a main component of Andisols, other minerals are also present in these types of soil. Iron oxides are widespread
minerals in materials derived from volcanic-ashes, distributed mainly
as coatings of primary or secondary minerals or infiltrated in clay aggregates. Their presence alters the electric charge and modifies the
surface properties of allophane. Several reports point out that iron
oxide minerals may participate in the adsorption mechanisms of toxic
trace elements as As, Sb, Cd, Cu, Pb, V, etc. (Michalkova et al., 2014;
Uwamariya et al., 2016; Yan et al., 2016). Michalkova et al. (2014)
demonstrated that amendment of soils with different types of iron
oxides has little effect on the adsorption of Cd and Cu. Synthetic
magnetite/bentonite compounds have been shown to enhance adsorptive capacities for immobilisation of Pb, Cd and Cu in water samples (Yan et al., 2016). Uwamariya et al. (2016) established that the use
of iron oxide-coated sand has interesting Cd and Cu removal capacities.
Finally, Suda and Makino (2016) established that iron oxides from soils
have a significant capacity to absorb pollutants such as arsenate, vanadate, and antimonate, among others. However, there is no clear
consensus if iron oxides, when found forming coatings over natural
aluminosilicate minerals, play a detrimental role in the adsorption capacities of cationic pollutants over a classical-cation adsorbent, such as
allophane. Another aluminosilicate, also present on volcanic-ashes derived soil, at much lower concentration than allophane, is imogolite,
which has been studied for copper and cadmium adsorption kinetics by
Arancibia-Miranda et al. (2015).
The objective of this research was to obtain natural allophane with
(NA-FeOx) and without (NA) iron oxides from a Southern Chilean
Andisol, and determine the influence of the natural free iron oxide
coatings on the adsorption of cationic pollutants Cd(II) and Cu(II) in
aqueous media.
2.2. Adsorbent characterisation
2.2.1. Total elemental analysis
The elemental composition of samples (Al, Si, and Fe content) was
determined by X-ray fluorescence by energy dispersion on a Shimadzu
EDX-720 apparatus equipped with a rhodium (Rh) X-ray tube at a
voltage of 5 to 50 kV and an X-ray Si detector.
2.2.2. X-ray diffraction
The allophane samples, freeze-dried and homogenised in an agate
mortar, were analysed by means of non-oriented powder diagrams.
Diffractograms were acquired scanning between 10 and 80° (2θ) using
Cu Kα (λ = 0.154 nm) radiation, 40 mA, and 40 kV and a nickel filter
on a Bruker model D8 Advance diffractometer, with a path of 0.02° and
a time of 1 s per path.
2.2.3. Scanning electron microscopy (SEM)
The SEM micrographs of NA-FeOx and NA samples were obtained
on a model EVO/MA10 (ZEISS) high vacuum scanning electron microscope.
2. Materials and methods
2.2.4. Fourier-transform infrared spectroscopy (FTIR)
FTIR spectra of the aluminosilicates were obtained on a Bruker
Tensor 27 spectrometer (Bruker). A 3 mg dry sample was compacted in
a spectral-grade KBr matrix. The spectra were scanned 32 times at a
resolution of 2 cm− 1.
2.1. Adsorbent preparation
2.1.1. Soil samples
Andisol samples were taken from 20 to 40 cm depth of Santa
Barbara (ashy, medial, mesic, Typic Dystrandept; (Mella and Kuhne,
1985)) uncultivated soil (36°47′ S 72° 8′ W). The series are dominated
by allophane (> 50% of total minerals) and have a high content of Fe
oxides. This profile was selected because allophane is known to be more
concentrated in the B horizon of soil (Mella and Kuhne, 1985). Basic
soil characterisation included organic carbon (Walkley-Black method,
(Allison, 1965)), pH (1:2.5 soil: solution ratio), cation exchange capacity, and surface charge.
2.2.5. Mössbauer spectroscopy
The NA-FeOx samples were weighed to provide a concentration
between 5 and 10 mg of total Fe cm− 2. The sample, ground in an agate
mortar, was placed on an approximately 3.4 cm diameter sample holder
(~ 5 cm2), and the spectrum was recorded at ambient temperature
(25 °C) at a speed rate of ± 10 mm s− 1 using a transmission spectrometer with a constant acceleration and a 20 mCi source of 57Co inserted
in a Rh matrix. The calibration was made with respect to an iron foil
sheet, α-Fe; the middle point of the hyperfine spectrum of the Fe was
defined as zero on the velocity axis.
2.1.2. Removal of organic matter
The organic matter was removed by an oxidative process with 30%
hydrogen peroxide (H2O2) as described by Kunze and Dixon (Kunze and
Dixon, 1965). The soil samples were suspended with double distilled
water and heated to 60 °C. Addition of 10 mL of 30% H2O2 was followed by digestion for 24 h. This process was repeated until the residual organic carbon content was minimal, monitored by the wet
oxidation method described by Kunze and Dixon (Kunze and Dixon,
1965).
2.2.6. Total specific surface area (TSSE)
The specific surface area was determined gravimetrically by applying the ethylene glycol monoethyl ether (EGME) technique, proposed for determining the surface area of soils and their fine fraction, as
well as of mineral silicates (Carter et al., 1965).
2.2.7. External specific surface area (ESSA), pore size distribution and pore
volume
A sample of approximately 0.3 g was first degassed. Then, the adsorption-desorption isotherms of N2 at − 195 °C (77 K) were acquired,
changing the relative pressure (P/Po) of the gas and recording the volume adsorbed on the solid's surface. Specific surface area was calculated from the amount of N2 adsorbed employing the Brunauer-EmmettTeller or BET equation (Brunauer et al., 1983). Pore size distribution
2.1.3. Extraction of the < 0.002 mm fraction
After removal of the organic matter, the required fraction was extracted from aqueous suspensions of the soil treated by a sedimentation
procedure based on Stokes' law. This sample corresponds to natural
allophane with iron oxides (NA-FeOx).
71
Geoderma 319 (2018) 70–79
J. Silva-Yumi et al.
model succeeds in giving a perfect description of the experimental data.
The pseudo-second-order kinetic model postulates that solute adsorption occurs on two available surface sites of the adsorbent, the
pseudo-second-order adsorption kinetics (Arancibia-Miranda et al.,
2015). The integrated form of the pseudo-second order kinetic rate
equation is:
(pore volume distribution vs. pore size) was determined from the N2
adsorption and desorption curves applying the Barrett-Joyner-Halenda/
Dollimore-Heal or BJH/DH method using relative pressures > 0.35
(Barrett et al., 1951). Pore volume is derived from the amount of vapour adsorbed at a relative pressure close to one (P/P0 ~0.95), assuming that the pores were filled with adsorbate in the liquid phase.
t
1
1
=
+
t
Ct
(Cm − cal )2k2
Cm − cal
2.2.8. Surface charge determination
The IEP values were determined by micro-electrophoresis.
Electrophoretic mobility was measured with a zeta meter (ZM-77) apparatus. Dilute dispersions (0.05 g L− 1) were prepared in 10− 3 M KCl
and pH was adjusted with 10− 2 M HCl or NaOH. The mobility was
averaged and the zeta potential calculated using the HelmholtzSmoluchowski equation (Hunter, 1983).
The observed rate constant of the pseudo-second-order-equation
(k2) is a complex function of the initial solute concentration. This linear
plot may be described as a chemisorption process.
The initial adsorption rate, h (mg g− 1 min− 1), can be calculated
from the pseudo-second-order equation, which can be obtained directly
from the intercept of the straight line according to the equation:
2.2.9. Particle size distribution
The Particle Size Distribution of samples was carried out in a
Malvern Zetasizer Nano ZS Dynamic Light Scattering (DLS) system
(Freud, 2007; de Kanter et al., 2016). The device uses classical singlescattering DLS methods to calculate particle sizes in a range from
0.3 nm to 10 μm with a 633 nm laser and detects the backscattered light
at 173°. The device features a temperature-controlled holder for a
cuvette.
h = k2 (Cm − cal )2
2.3.3. Adsorption isotherms
In 50 mL polypropylene centrifuge tubes, 20 mL of 10, 20, 40, 60,
80, 100 and 120 mg L− 1 Cd2 + or Cu2 + solutions in 0.05 M KNO3, as
well as 1.5 mL of NA-FeOx or NA, were added to each tube. The tubes
were stirred at 200 rpm in a linear shaker at ambient temperature for
8 h and subsequently centrifuged. The supernatants were filtered and
saved for later quantification of cadmium and copper. The experiments
were performed in duplicate at 25 °C and the pH was controlled at each
point, keeping a fixed value of 5.5 ± 0.2 for the adsorption studies
(Clark and McBride, 1984). The amount of metal absorbed at equilibrium, Cs, can be determined by mass balance according to:
2.3.1. Theoretical speciation of Cd(II) and Cu(II)
The chemical speciation, or determination of the form in which the
ions are found in solution, was carried out with the GEOCHEM PC
program, designed for making speciation calculations at equilibrium
(Parker et al., 1995). The parameters for the chemical speciation were
the concentration of each of the species, expressed as –log C, a constant
ionic force of 0.05 M NaNO3, and a pH range from 2 to 11.
Cs =
1
Cs = KF (Ce )
(1)
where C0 and Ct are the metal's concentrations (mg L ) in the supernatant at times 0 and t, respectively, V (L) is the total volume of solution in the tubes, and m is the mass (g) of allophane used. Adsorption
kinetic data were fitted employing the pseudo-first and –second order
models. The pseudo-first-order kinetic model assumes that a solute ion
is adsorbed on a surface site of the adsorbent (Arancibia-Miranda et al.,
2015). The integrated form of the pseudo-first order kinetic rate
equation is:
log(Cm − Ct ) = log Cm − cal
(5)
nfads
(6)
where KF and nfads are adjustable parameters with 0 < 1/nfads < 1
(usually), Cs is the concentration of adsorbed species (mg g− 1), Ce is the
concentration in the supernatant solution (mg L− 1), KF is the empirical
Freundlich adsorption coefficient, and nfads is a linearity factor.
The Langmuir model assumes that the adsorption of the solute on a
solid takes place forming a monolayer; it considers ideal solutions
where the adsorption is independent of the coverage of the surface, the
sites on which the adsorption takes place are distributed uniformly on
the adsorbent surface, and the energy due to the interaction is equal at
all the interaction sites. The model also assumes a fixed number of
exchange sites. The Langmuir model is applicable to L-type isotherms,
characterised by a high affinity between adsorbate and adsorbent,
commonly implying a chemical adsorption with the formation of inner
sphere complexes (Giles and Smith, 1974). The Langmuir model is
described by:
−1
kt
− 1
2.303
(C0 − Ce )
V
m
where Ce is the concentration of the metal in the supernatant after
reaching equilibrium. Adsorption isotherm data were fitted using
Freundlich and Langmuir adsorption models (Evangelou, 1998). The
Freundlich model does not have a mechanistic interpretation; it only
represents an empirical approach to predict the species distribution
between a solid/solution phase, and is described by:
2.3.2. Adsorption kinetics
In 50 mL polypropylene centrifuge tubes, 20 mL of 120 ppm Cd2 +
or Cu2 + solution in 0.05 M NaNO3, and a volume of NA or NA-FeOx
suspensions, equal to 200 mg of aluminosilicate, were added. The tubes
were stirred (200 rpm) in a reciprocating shaker using time intervals of
0.16, 0.33, 0.5, 0.66, 1, 2, and 5 h. At the end of each time-interval the
tubes were centrifuged. The supernatant was filtered and stored for
later quantification of Cd and Cu. The experiments were performed in
duplicate at 25 °C and the pH was controlled at each point, keeping a
fixed value of 5.5 ± 0.2 for the kinetic studies (Clark and McBride,
1984). The results of the experiment were used to calculate the amount
of adsorbed metal as a function of the time, Ct, determined by mass
balance according to the equation:
(C0 − Ct )
V
m
(4)
where Cm-cal is the total adsorbed amount calculated by the model.
2.3. Adsorption studies
Ct =
(3)
Cs = Cm − cal
(2)
KL Ce
1 + KL Ce
(7)
where Cs and Ce have the same meaning as in the Freundlich model, Cm−1
) calculated by the
cal is the maximum adsorption capacity (mg g
model, and KL is the empirical affinity Langmuir coefficient (L mg− 1).
The removal percentages, R, were calculated using Eq. 8 and are
presented with respect to the initial concentration in solution of pollutant (C0):
where Ct is the adsorbed quantity at any time (t), Cm is the maximum
experimental adsorbed amount (obtained from the kinetic curve), k1 is
a combination of the adsorption and the desorption rate constants, and,
finally, Cm-cal corresponds to the maximum adsorbed amount, as determined by the application of the model. Cm-cal and Cm will agree if the
72
Geoderma 319 (2018) 70–79
J. Silva-Yumi et al.
R=
C 0 − Ce
× 100%
C0
Table 1
Elemental composition, specific surface area and pore volume of aluminosilicate samples.
(8)
2.3.4. Desorption isotherms
Once the adsorption process had ended, the centrifuged solids were
washed with some drops of double distilled water to dilute and remove
residual solution. Consecutively, 20 mL of 0.05 M NaNO3 free of Cd2 +
and Cu2 + were added, and the mixtures were stirred at 200 rpm in an
orbital shaker at 25.0 °C for 8 h. After centrifugation, the supernatant
was filtered and stored for quantification through atomic absorption
spectroscopy. Desorption process was performed in duplicate only for
one desorption cycle. Since the Freundlich model fitted better the adsorption isotherms data, the same model was considered for adjusting
desorption isotherms. The amount of retained metal (mg of the metal
ion per g of allophane) was calculated from the difference between the
adsorbed amount and the desorbed amount according to the following
equation:
Cr = Cs −
Cdes
V
m
Si
Al
Fe
Al/Si
TSSAa
ESSAa
Total pore volume
Mean pore diameter
a
−1
where Cr (mg g ) and Cs (mg g ) are the amounts retained and adsorbed, respectively; Cdes (mg L− 1) is the equilibrium concentration in
the desorption process; V (L) is the volume of solution used for the
desorption process; and m (g) is the mass of allophane used. Desorption
percentages, P, were calculated from the adsorption and desorption
data as in the following equation:
P=
Cdes
× 100%
Cs
(10)
−1
where Cdes (mg g ) is the amount of metal desorbed by NaNO3 and Cs
(mg g− 1) is the amount of metal adsorbed by the solid. An index of the
desorption hysteresis (hysteresis index or HI) based on Freundlich's
exponents is calculated as a percentage according to:
HI =
nfdes
× 100%
nfads
m2 g− 1
m2 g− 1
mL g− 1
nm
NA
34.3 ± 0.3
42.8 ± 0.5
19.4 ± 0.0
1.25
301 ± 23
151 ± 1
0.14
3.7
44.8 ± 0.4
45.6 ± 0.5
4.8 ± 0.0
1.02
237 ± 14
17 ± 0
0.04
10.5
TSSA: total specific surface area; ESSA: external specific surface area.
However, after removing iron oxides (NA), the spectrum showed strong
signals assigned to halloysite and other weaker signals representative of
quartz, cristobalite, and feldspar. Surface parameters (Table 1) revealed
that specific surface (whether TSSA or ESSA) was greater in the aluminosilicate samples after removal of the organic matter contents (NAFeOx) when compared to the samples after elimination of iron oxide
coatings (NA). This difference may account for the higher degree of
aggregation observed in NA (Fig. 1) compared to NA-FeOx, as large
aggregates (and clusters) provide lower surface area than small ones.
Moreover, iron oxides are renowned materials for their large surface
area. The shape of the adsorption and desorption isotherms of N2
(Supplementary material, Fig. S4) used for the ESSA calculation suggests the presence of mesopores (2–50 nm) at NA-FeOx and macropores
(> 50 nm) at the NA samples, respectively. Specifically, desorption
curves (Supplementary material, Fig. S5) were used to evaluate the pore
size distribution, resulting on a pore-size maximum of 50 and 40 nm for
NA-FeOx and NA, respectively. Expressing the porosity by means of the
total pore volume and the pore size distribution (Table 1), NA-FeOx has
greater porosity than NA, but with a smaller pore diameter. The morphological characterisation of the samples is shown by the scanning
electron microscopy (SEM) micrographs presented in Fig. 1.
The images for both samples show nanometric-sized spheres, a
morphology characteristic of allophane. Moreover, the material's tendency to form aggregates with different shapes and sizes, as has been
described in the literature (Henmi and Wada, 1976), was also observed
even after subjecting the samples to ultrasound for dispersion. In the
case of NA-FeOx, aggregates were seen in the order of 100 to 500 nm,
with some clusters (adhesion of aggregates) reaching the order of micrometres. In contrast, a greater degree of aggregation was seen in NA
(larger aggregates) compared to those seen in NA-FeOx, exceeding
1.0 μm. The fact that NA forms larger aggregates than NA-FeOx may
affect the total and external specific surface area of the samples. FTIR
spectroscopy characterisation (Supplementary material, Fig. S2) produced characteristic transmittance bands for natural allophane. NAFeOx typically produces a signal at 930 cm− 1, corresponding to the
FeeO bond stretching (Mora et al., 1994), while NA produces a signal
between 770 and 400 cm− 1 representative of (HO)Si(OAl)3 allophanewalls vibrations (Levard et al., 2012). The most important finding from
the FTIR results was the disappearance of FeeO signals in NA samples,
which confirmed the removal of free iron oxides by the DCB treatment.
57
Fe Mössbauer spectroscopy characterisation of the NA-FeOx sample
(Supplementary material, Fig. S3) did not yield conclusive results with
respect to iron-oxide mineralogy due to the low resolution of the existing sextet. The presence of Fe3 + in octahedral of silicate structures
was clearly identified, but since the structural unit of all the iron oxides
is the octahedron, this observation does not provide further information
for species identification. If Mössbauer results are compared to those
reported in the literature for other Andisol series (Pizarro et al., 2000),
the oxides present may be assigned to a ferrihydrite-like mineral, a
stable iron-oxide known for its association with allophane (Bartoli
et al., 2007; Filimonova et al., 2016; Latrille et al., 2003). Taken together, the general characteristics from adsorbents (Table 1) indicate
(9)
−1
%
%
%
NA-FeOx
(11)
where nfdes and nfads are calculated from exponents of Freundlich's
equation in the desorption and adsorption, respectively.
3. Results
3.1. Characterisation
The Santa Bárbara soil series was chosen because allophane is
dominant in its mineralogy, comprising > 50% content, followed by
halloysite and vermiculite, with 1–5% (Mella and Kuhne, 1985). It is a
typical Andisol with acidic characteristics and a negative surface charge
(pHwater and pHKCl are 6.22 and 5.45, respectively), with an isoelectric
point (IEP) of 5.6 and 4.4% of organic carbon. Its texture is defined by
sand (9.6%), silt (65.9%), and clay (24.4%).
Removal of the organic matter was carried out by sequential
treatments with hydrogen peroxide and the respective sample without
organic matter was fractionated by the pipet method to separate
the < 0.002 mm part (the Particle Size Distribution by DLS analysis
showed that the average size of the fraction is 392 nm); this sample
corresponds to natural allophane without organic matter, with its iron
oxides deposited on the surface (NA-FeOx). Then, the destruction of the
iron oxides was carried out by the DCB method, obtaining natural allophane without organic matter or iron oxides (NA). The results of the
chemical composition by X-ray fluorescence of these materials, as well
as the specific surface area and pore volume, are presented in Table 1.
Reduction of the iron contents from 19.4% (NA-FeOx) to 4.8% (NA)
by the DCB method suggested successful removal of the free iron oxides.
Conversely, the XRD analysis for the NA-FeOx samples confirmed the
predominance of minerals with short-range organisation due to the
absence of intense diffraction signals (Supplementary material, Fig. S1).
73
Geoderma 319 (2018) 70–79
J. Silva-Yumi et al.
Fig. 1. SEM micrographs for natural allophane (NA) and natural allophane with iron oxides (NA-FeOx). Arrows indicate large and small aggregates found at NA and NA-FeOx micrographs, respectively.
determine the maximum pH at which these experiments could be performed in order to avoid confusing the adsorption phenomena with
precipitation of the metals as the hydroxides M(OH)2 (Pokrovsky et al.,
2012). The GEOCHEM-PC software is able to deduce the speciation of
Cd and Cu ions in solution as a function of pH. At an initial concentration of 120 ppm and an ionic strength of 0.05 M NaNO3, the
Cu2 + and Cd2 + metal ions started precipitating as M(OH)2 at pH values
of 5.6 and 8.5, respectively (Supplementary material, Fig. S6). Consequently, adsorption experiments were conducted at pH 5.5, consistent
with the typically acidic pH of soil derived from volcanic ashes (with
pH as low as 4.1, (Escudey et al., 2004)), which suffer additional
acidification processes due to agronomic exploitation (Hirzel et al.,
2011; Matsuyama et al., 2005).
Adsorption kinetics determined the rate at which the pollutants,
Cd+ 2 and Cu+ 2, were adsorbed over the aluminosilicate surfaces, NAFeOx and NA, as well as the time needed to reach equilibrium. A deeper
analysis can even allow the determination of the type of adsorption and
understand the reaction's route and mechanism (Worch, 2012). Fig. 3
shows the adsorption kinetics of both metals on the surfaces in study
and the fit of the adsorption models applied in each case.
The adsorbents were rapidly saturated, reaching a constant adsorption value after the first hour of the experiment, with most of this
saturation being achieved in < 60 min. This behaviour was not dependent on the adsorbed cation. The maximum amount of cation adsorbed was different between NA-FeOx and NA, as it was always higher
at the latter. Maximum amount of adsorbed Cd+ 2 on NA-FeOx and NA
were ca. 4.0 and 9.0 mg g− 1, respectively, while in the case of Cu+ 2
the values were approximately 6.5 and 10.0 mg g− 1. This behaviour
showed that the removal of iron oxides had a beneficial effect on the
cation adsorptive capacity of the allophane. Surface charge changes
were likely one of the factors involved in the observed effects.
Comparing the two pollutant cations, the results were similar, with a
slightly greater adsorption affinity of copper over cadmium. The fit of
the curves was carried out using pseudo-first- and pseudo-second-order
kinetic models, with their respective fitting parameters presented in
Table 2.
Considering the correlation coefficient (R2) as well as the value of
chi square (χ2), the model with the best fit corresponds to the pseudosecond-order model, suggesting that the adsorption of the cations occurs on the materials at two different sites. Thus, if the values of
pseudo-second order reaction rate constant (k2) are compared, it can be
concluded that the adsorption of the pollutants occurs about 3 to 4
times faster over NA than NA-FeOx. Therefore, the presence of the iron
oxide coating somehow deaccelerates the adsorption of the cationic
pollutants, a possible consequence of the different surface charge with
adsorbents. Adsorption kinetics on natural allophanes (with and
without FeOx coatings) can be advantageous compared to kinetics
that NA-FeOx possesses higher specific surface area/mass and bigger
pores than NA. Another important surface characterisation was determination of the variations at the isoelectric point due to destruction
of the OM and the removal of free iron oxides. Fig. 2 presents the
changes of the zeta potential vs. pH curves for each sample.
The acidic isoelectric point (IEP) of an Andisol has its origin in the
negative surface charge of allophane, intensified by the organic matter.
As such, the system's overall charge becomes more negative as the organic matter content increases. The Santa Barbara soil was determined
to have an IEP of 5.6. When the organic matter was removed (NAFeOx), the curve increased to 10.3 on IEP, in agreement with the destruction of the organic matter and the exposition of Fe-OH and Al-OH
active sites. It is also known that the iron oxide layer, in addition to the
terminal –OH groups of the aluminium oxides octahedral structures
that cover the allophane, contribute with positive charge to the surface
of the particle (Cornell and Schwertmann, 2003). Removal of the free
iron oxides with the DCB treatment decreased the IEP of the allophane
particles from 10.3 to 5.4, pointing out the negative contribution of SiOH active sites of allophane (Escudey et al., 1986; Escudey et al., 2004).
The differences in IEP and changes of surface charge with pH between
NA-FeOx and NA may be at least partially responsible for the different
adsorption capacities of cationic species seen; i.e., surfaces with higher
negative surface charge character should present higher adsorption
capacities towards cations like Cd2 + and Cu2 +.
3.2. Adsorption studies
Prior to performing the adsorption studies, it was necessary to
Fig. 2. Zeta potential vs pH for Santa Barbara soil and aluminosilicates samples.
74
Geoderma 319 (2018) 70–79
J. Silva-Yumi et al.
Fig. 3. Adsorption kinetic of Cd+ 2 and Cu2 + over NA-FeOx and NA with the respective kinetic model adjustments.
may be involved (Sullivan, 1977). On these terms Cu+ 2 is a harder
(acid) cation than Cd2 +, and therefore would present affinity to bond
with hard bases, like terminal –OH moieties at the allophane structure.
This may partially explain the preferences for copper over cadmium.
The isotherms revealed that adsorption of both pollutants is higher
over NA than NA-FeOx. The lower adsorption of cationic pollutants
over NA-FeOx compared to NA may be at least partially attributed to
differences in sign and magnitude of the adsorbent surface charge at
equilibrium pH (5.5), where iron oxide layer provides a positive-net
charge to the allophane surface, while natural allophane is almost at its
IEP. The interaction of cations with the surface would be different in
electrostatic terms. The same trends were observed when analysing the
removal percentages of pollutants from solution, R (Fig. 5), where the
highest R values observed were associated with the lowest development
of positive charge on the adsorbent surface and vice versa.
As concentration in solution increases, the percentage of pollutants
removed from the solution (R) decreases, limited by the adsorption
capacity of the adsorbents. Independent of the type of pollutant, the
decrease in R as function of the metal concentration in the solution was
more pronounced for NA-FeOx; this agrees well with the differences
observed for the maximum adsorption capacity for each adsorbent
(Fig. 4), confirming the effect of the iron oxide coating on the cation
adsorption capacity of allophane. Also Cu2 + tended to be more easily
removed by the same adsorbent compared to Cd2 +, a behaviour that
may also be attributed to the HSAB theory. In order to better characterise the systems, the Freundlich and Langmuir models were applied
to the adsorption isotherms (Fig. 4) and evaluated to determine which
model yielded a superior fit. Fitting parameters are presented on
Table 3.
As indicated by the low difference in correlations coefficient, both
the Freundlich and the Langmuir models fit the isotherms. This suggests
that adsorption is a combination of chemical and physical processes.
However, comparison of the R2 and χ2 values reveals that the fit of the
Freundlich model is superior to the Langmuir model. Additionally, the
Cm-cal values predicted by Langmuir model do not always agree with the
experimental Cm. The Freundlich model has been previously used for
Table 2
Kinetic models parameters.
NA-FeOx
NA
Cd2 +
Cu2 +
Cd2 +
Cu2 +
Pseudo 1st order model
Cm
mg g− 1
k1
min− 1
R2
χ2
3.89
2.70
0.85
0.32
6.35
0.21
0.99
0.02
9.27
0.25
0.99
0.01
10.11
0.37
0.99
0.32
Pseudo 2nd order model
Cm
mg g− 1
k2
g mg− 1 min− 1
h
mg g− 1 min− 1
R2
χ2
4.12
0.07
1.19
0.88
0.25
6.43
0.13
5.37
0.99
0.00
9.32
0.24
20.8
0.99
0.01
10.1
0.51
52.4
0.99
0.01
observed for same cations on synthetic imogolite were equilibrium was
reached after about 800 min (Arancibia-Miranda et al., 2015), while
about 60 min were required on allophanes samples studied on present
work.
The adsorption isotherms for each cation on both allophanes were
obtained by exposing the adsorbents to different concentrations of
pollutants until equilibrium was reached (Fig. 4).
As the graphs show, adsorption capacity increased rapidly at low
pollutant concentrations. The isotherms for Cd2 + correspond to L-type
isotherms, which tend to reach a maximum in the adsorption capacity.
In contrast, the isotherms for Cu2 + resemble H-type isotherms, indicating that adsorbent has more affinity for copper than cadmium.
Adsorption of Cu2 + was greater than that of Cd2 + per mass of adsorbent, regardless of allophane-type. As has been long proposed
(Sullivan, 1977), the “replacing-power” of cations is not easy to predict
since it depends on several system features including adsorbent nature,
concentration of metal in solution, type of ion adsorption originally on
exchange surface, temperature, and pH. Nevertheless, it has been proposed that the Pearson theory of Hard and Soft Acids and Bases (HSAB)
75
Geoderma 319 (2018) 70–79
J. Silva-Yumi et al.
Fig. 4. Adsorption isotherms of pollutants (points) and model fittings (lines) over natural allophane (NA) and natural allophane with iron oxides (NA-FeOx).
Table 3
Langmuir and Freundlich parameters for adsorption data.
NA-FeOx
Cd
Freundlich
nfads
KF
R2
χ2
Langmuir
Cm-cal
KL
R2
χ2
mg g− 1
mg g− 1
L mg− 1
2+
NA
Cu
2+
Cd2 +
Cu2 +
2.54
1.11
0.98
0.09
5.97
4.16
0.98
0.19
2.95
4.65
0.99
0.24
5.99
8.83
0.95
1.06
5.98
0.10
0.95
0.20
7.00
2.01
0.91
0.80
12.4
0.46
0.94
1.09
9.93
45.3
0.84
3.32
situation occurs for Cu2 +. On this case, KF values over NA-FeOx and NA
also implies that copper affinity is reduced by the iron oxide coating.
In order to fully characterise the system under study, the desorption
process must be analysed as well. The corresponding desorption isotherms were obtained by plotting the amount of metal retained on the
solid after the process of desorption as a function of the equilibrium
concentration of the metal in the solution. The desorption isotherms
calculated for each metal and adsorbent are presented in Fig. 6.
In each case, comparison of the magnitude between the respective
adsorption and desorption isotherm reveals that adsorbed species consistently exceed desorbed species. This observation is consistent with
the existence of an irreversible adsorption process over the adsorbents,
which can be explained in part by electrostatic considerations and additionally by a ligand exchange mechanism; combining physical and
chemical processes as previously suggested during model fitting analysis. This situation would be environmentally favourable, as it minimises the possibility of leaching which would return the retained
Fig. 5. Removal percentages for different adsorbent and pollutants, as a function of the
initial metal concentration in solution (C0).
fitting adsorption data from Cd2 + and Cu2 + on other types of aluminosilicate, such as on zeolite and bentonite samples from Anarak and
Firouzkoh mines in Iran (Hamidpour et al., 2010) and on kaolinite and
montmorillonite (Gupta and Bhattacharyya, 2008).
In regards to the Freundlich parameters, the nfads gave similar values for the adsorption of each metallic cation independent of the adsorbent. For instance, nfads values for Cd2 + were similar on NA-FeOx
and NA, which was also the case for Cu2 +. Therefore, adsorption isotherms fitted with similar nfads values can be compared in terms of their
empirical Freundlich adsorption coefficient or KF. For Cd2 +, comparing
the KF value over NA-FeOx and NA indicates that iron oxide coating
reduces about 4 times the affinity of allophane for cadmium. The same
76
Geoderma 319 (2018) 70–79
J. Silva-Yumi et al.
Fig. 6. Adsorption and desorption isotherms (points) and their corresponding fitting by the Freundlich model (lines).
adsorption. Considering the relationship between R and P (comparing
Figs. 5 and 7), the most intense adsorption process corresponds to the
less intense desorption, when irreversible (chemical) adsorption processes predominate. Chemical adsorption of Cd2 + on a different aluminosilicate, montmorillonite from Gumma (Japan), has been also described (Kumar-Saha et al., 2003) with P values near 40%. As in the
adsorption experiments, the Freundlich model was also used for fitting
the desorption isotherms, for which parameters are presented in
Table 4.
Similar to the trends observed for Freundlich model's adjustment of
adsorption isotherms, the values of nfdes (Freundlich linearity factor for
desorption) were similar for each pollutant independent of the adsorbent. Therefore, KF (Freundlich coefficient) obtained for the desorption isotherms may be compared for each pollutant. The values of KF
for Cd+ 2 on NA was almost 4 times higher than on NA-FeOx, meaning
that the interaction of cadmium over the aluminosilicate was about 4
times stronger if compared to the iron oxides coated surface. In the case
of Cu2 +, KF values on NA practically doubled over that of NA-FeOx;
thus, the interaction of Cu2 + over the iron oxide coated allophane was
weaker compared to the bare allophane surface. These comparisons
between KF values for the desorption also explain the behaviour observed for the adsorption kinetics experiments (faster on NA than NA-
pollutants to the environment (water bodies, etc.). To further elucidate
the nature of the interaction between pollutants and adsorbents, the
corresponding desorption percentages, P, were analysed.
Fig. 7 demonstrates that the desorption processes were incomplete
in all cases. For the respective cation, the desorption percentages from
NA-FeOx were higher than those observed for NA. For both aluminosilicates Cd2 + desorption was higher than Cu2 + desorption. The mostly
irreversible character of adsorption can be attributed to a specific interaction in high-energy sites in the solids and would correspond to a
superficial ligand exchange reaction. This observation is in agreement
with the pseudo-second-order model that best represented the experimental values and is usually associated with a process of chemical
Table 4
Freundlich model desorption parameters.
NA-FeOx
Freundlich
nfdes
KF
R2
χ2
Fig. 7. Desorption percentages for adsorbents and pollutants, as a function of the initial
metal concentration in solution (C0).
77
mg g− 1
NA
Cd2 +
Cu2 +
Cd2 +
Cu2 +
1.96
0.63
0.95
0.08
5.55
3.62
0.85
0.79
2.08
2.43
1.00
0.03
4.16
7.40
0.77
4.19
Geoderma 319 (2018) 70–79
J. Silva-Yumi et al.
by the adsorbent (over allophane retention values were around 60% at
5.0 pH). Therefore, the results herein taken together with the literature
demonstrate that several types of aluminosilicate exhibit irreversible
chemical adsorption of Cd2 + and Cu2 +, showing NA-FeOx and NA the
highest irreversibility within the minerals considered. From an environmental point of view the high irreversible adsorption is an important outcome, because reduces the likelihood that the adsorbed
metallic cations return to solution phase putting them back to the environment.
Table 5
Freundlich exponent and hysteresis index for adsorption and desorption processes of
pollutant over adsorbents.
Pollutant
Parameters
NA-FeOx
NA
Cd2 +
nfads
nfdes
HI
nfads
nfdes
HI
2.54
1.96
77.2 (−)
5.97
5.55
93.0 (−)
2.95
2.08
70.5 (−)
5.99
4.16
69.4 (−)
Cu2 +
4. Conclusions
FeOx, see Fig. 3) and the desorption percentages (for each cation,
desorption is always greater when studied over NA-FeOx, see Fig. 7).
More specifically, a relationship can be established between adsorption and desorption through hysteresis, a phenomenon in which the
adsorption and desorption isotherms do not coincide. Various mechanisms have been suggested to explain hysteresis, including chemical
precipitation, migration, incorporation of the adsorbate in the solid's
matrix, and deformation of the micropores (Hamidpour et al., 2010).
Table 5 presents the calculated values of the hysteresis index for Cd2 +
and Cu2 + on each of the adsorbents.
As expressed by Hamidpour et al. (Hamidpour et al., 2010), HI
values lower than 100 (nfads > nfdes) correspond to negative hysteresis
(−) and represent a scenario in which it is difficult for adsorbed species
to be desorbed from adsorbent; HI = 100 (nfads = nfdes) means the
absence of hysteresis; HI higher than 100 (nfads < nfdes) corresponds to
positive hysteresis (+), which indicates a feasible desorption process of
the analyte from the adsorbent. Thus, the hysteresis index “quantifies”
the reversible character of the adsorption process. In allophane systems
studied, negative hysteresis was observed (HI –), which is consistent
with the assumption that chemical adsorption processes likely dominate
over electrostatic interactions in all systems under study. HI values
were higher for NA-FeOx than NA, again pointing to the effect of iron
oxide coating at the aluminosilicate surfaces. Iron oxides weaken the
interaction between the cationic pollutants and the adsorbent (in
agreement with the kinetics assays, Fig. 3, and removal percentages or
R, Fig. 5). Thus, NA is a better material than NA-FeOX in terms of adsorption speed and pollutant removal percent.
For the adsorption outcomes, experimental maximum adsorption
capacity over NA was 11.3 and 12.3 mg g− 1 for Cd2 + and Cu2 +, respectively, comparable to reports from other natural aluminosilicatebased adsorbents. For instance, recent reports have established for
Cd+ 2 maximum adsorption capacity values of 6.3, 11.0, 11.2 and
22.17 mg g− 1 on montmorillonite, kaolinite, bentonite and smectite
(Gupta and Bhattacharyya, 2008; Jitniyom et al., 2012; Yavuz et al.,
2003), while for Cu2 +, over the same adsorbent, maximum adsorption
values found were 17.9, 10.8, 17.8 and 42.43 mg g− 1, respectively
(Jitniyom et al., 2012; Sdiri et al., 2011; Yavuz et al., 2003; Zhi-rong
and Shao-qi, 2010). Surface modification of the adsorbents has been
shown to yield significant enhancement of adsorptive capacities (Erdem
et al., 2009; Mhamdi et al., 2014), but was not a scope for the current
study. Therefore, adsorption capacities determined for allophane samples are comparable to those found in the literature for other natural
aluminosilicate-based cation adsorbents. On the case of hysteresis, negative hysteresis for Cd2 + on similar adsorbents was found in literature, including montmorillonite and bentonite samples by Kumar-Saha
et al. (2003) and for zeolite by Hamidpour et al. (Hamidpour et al.,
2010).
Analysing the desorption process, the irreversible retention of Cd2 +
ranges from 55.2 to 83.5% and for Cu2 + from 79.0 to 95.0% for NAFeOx and NA, respectively. Itami and Yanai (Itami and Yanai, 2006)
studied adsorption/desorption of Cd2 + and Cu2 + on a series of aluminosilicates from mineral deposits, including montmorillonite, sericite
and allophane; the authors observed in all cases that desorption process
was not complete and a major part of the metallic cations was retained
Adsorbent materials were prepared from a low-cost natural source
(Andisol) to study the removal of Cd2 + and Cu2 + from aqueous solutions: Allophane coated by its original iron-oxide layer (NA-FeOx) and
bare natural allophane (NA). At working pH (5.5) NA was practically at
its IEP (5.4) while NA-FeOx (10.3) likely displayed positive surface
charge, partially explaining the differences in adsorption capacities
observed between the adsorbents. From the kinetic studies, it was determined that the adsorption was fast (1 h), almost unaffected by the
presence of the iron oxides and was fitted by the pseudo-second order
model. The adsorption isotherm studies showed that NA had a greater
adsorption capacity of both Cd+ 2 and Cu+ 2 compared to NA-FeOx.
Langmuir and Freundlich adsorption models fit the experimental data
well, indicating the adsorption process is likely a combination of physical and chemical phenomena. From Freundlich model parameters
analysis it was established that Cd2 + and Cu2 + were adsorbed about 4
and 2 times more strongly over NA surface compared to NA-FeOx, respectively. Desorption studies demonstrated that most of the adsorbed
cations were retained by the adsorbent, suggesting irreversible chemical adsorption on specific active sites on the adsorbents. Hysteresis
index (HI) confirmed this irreversible behaviour which is advantageous
in environmental pollutant removal applications.
Supplementary data to this article can be found online at https://
doi.org/10.1016/j.geoderma.2017.12.038.
Acknowledgements
This work was supported by the Secretaría de Educación Superior,
Ciencia, Tecnología e Innovación (SENESCYT) of Ecuador through the
Programa de Estudios de Cuarto Nivel de Formación Académica en el
Exterior, Convocatoria Abierta 2011, the Center for the Development of
Nanoscience and Nanotechnology (grant number CEDENNA, FB0807),
and Proyectos Basales (grant number USA1555) USACH.
References
Allison, L.E., 1965. Organic Carbon. In: Black, C.A., Evans, D.D., White, J.L., Ensminger,
L.E., Clark, F.E. (Eds.), Methods of Soil Analysis. Part 2. Chemical and
Microbiological Properties. Agronomy Monograph 9. American Society of Agronomy,
Madison, pp. 1367–1378.
Arancibia-Miranda, N., Silva-Yumi, J., Escudey, M., 2015. Effect of cations in the background electrolyte on the adsorption kinetics of copper and cadmium and the isoelectric point of imogolite. J. Hazard. Mater. 299, 675–684.
Barrett, E.P., Joyner, L.G., Halenda, P.P., 1951. The determination of pore volume and
area distribution in porous substances. I. Computation from nitrogen isotherms. J.
Am. Chem. Soc. 73, 373–380.
Bartoli, F., Poulenard, J., Schouller, B., 2007. Influence of allophane and organic matter
contents on surface properties of Andosols. Eur. J. Soil Sci. 58, 450–464.
Bolan, N., Mahimairaja, S., Kunhikrishnan, A., Naidu, R., 2013. Sorption–bioavailability
nexus of arsenic and cadmium in variable-charge soils. J. Hazard. Mater. 261,
725–732.
Brunauer, S., Emmett, P.H., Teller, E., 1983. Adsorption of gases in multimolecular layers.
J. Am. Chem. Soc. 60, 309–319.
Carter, D., Heilman, M., Gonzalez, C., 1965. Ethylene glycol monoethyl ether for determining surface area of silicate minerals. Soil Sci. 100, 356–361.
Clark, C.J., McBride, M.B., 1984. Chemisorption of Cu(II) and Co(II) on allophane and
imogolite. Clay Clay Miner. 32, 300–310.
Cornell, R.M., Schwertmann, U., 2003. The Iron Oxides Structure, Properties, Reactions,
Occurrences and Uses, 2nd ed. Wiley-VCH, Weinheim.
De Gisi, S., Lofrano, G., Grassi, M., Notarnicola, M., 2016. Characteristics and adsorption
78
Geoderma 319 (2018) 70–79
J. Silva-Yumi et al.
organic matter in the B horizon of an Andosol. Eur. J. Soil Sci. 54, 357–364.
Levard, C., Doelsch, E., Basile-Doelsch, I., Abidin, Z., Miche, H., Masion, A., Rose, J.,
2012. Structure and distribution of allophanes, imogolite and proto-imogolite in
volcanic soils. Geoderma 183–184, 100–108.
Mahar, A., Wang, P., Ali, A., Awasthi, M., Lahori, A., Wang, Q., Li, R., Zhang, Z., 2016.
Challenges and opportunities in the phytoremediation of heavy metals contaminated
soils: a review. Ecotoxicol. Environ. Saf. 126, 111–121.
Matsuyama, N., Saigusa, M., Sakaiya, E., Tamakawa, K., Oyamada, Z., Kudo, K., 2005.
Acidification and soil productivity of allophanic andosols affected by heavy application of fertilizers. Soil Sci. Plant Nutr. 51, 117–123.
Mehra, O., Jackson, L., 1960. Iron oxide removal from soils and clays by a dithionitecitrate system buffered with sodium bicarbonate. Clay Clay Miner. 7, 317–327.
Mella, A., Kuhne, A., 1985. Sistemática y descripción de las familias, asociaciones y series
de suelos derivados de materiales piroclásticos de la zona central sur de Chile. In:
Tosso, J. (Ed.), Suelos volcánicos de Chile. INIA, Santiago, pp. 635–636.
Mhamdi, M., Elaloui, E., Trabelsi-Ayadi, M., 2014. Kinetics of cadmium adsorption by
smectite of Oued Tfal (Gafsa Basin). Desalin. Water Treat. 52, 4245–4256.
Michalkova, Z., Komarek, M., Sillerova, H., Della Puppa, L., Joussein, E., Bordas, F.,
Vanek, A., Vanek, O., Ettler, V., 2014. Evaluating the potential of three Fe- and Mn(nano)oxides for the stabilisation of Cd, Cu and Pb in contaminated soils. J. Environ.
Manag. 146, 226–234.
Molina, M., Aburto, F., Calderón, R., Cazanga, M., Escudey, M., 2009. Trace element
composition of selected fertilizers used in Chile: phosphorus fertilizers as a source of
long-term soil contamination. Soil Sediment Contam. 18, 497–511.
Mora, M., Escudey, M., Galindo, G., 1994. Síntesis y caracterización de suelos alofánicos.
Bol. Soc. Chil. Quim. 39, 237–243.
Mosa, K.A., Saadoun, I., Kumar, K., Helmy, M., 2016. Potential biotechnological strategies
for the cleanup of heavy metals and metalloid. Front. Plant Sci. 7, 1–14.
Parker, D., Norvell, W., Chaney, R., 1995. GEOCHEM-PC - a chemical speciation program
for IBM and compatible personal computers. In: Loeppert, R.H., Schwab, A.P.,
Goldberg, S. (Eds.), Chemical Equilibrium and Reaction Models. Soil Science Society
of America, Madison, pp. 253–269.
Pizarro, C., Furet, N., Venegas, R., Fabris, J., Escudey, M., 2000. Some cautions on the
interpretation of Mossbauer spectra in mineralogical studies of volcanic soils. Bol.
Soc. Chil. Quim. 45, 243–250.
Pokrovsky, O.S., Probst, A., Leviel, E., Liao, B., 2012. Interactions between cadmium and
lead with acidic soils: experimental evidence of similar adsorption patterns for a wide
range of metal concentrations and the implications of metal migration. J. Hazard.
Mater. 199-200, 358–366.
Sdiri, A., Higashi, T., Hatta, T., Jamoussi, F., Tase, N., 2011. Evaluating the adsorptive
capacity of montmorillonitic and calcareous clays on the removal of several heavy
metals in aqueous systems. Chem. Eng. J. 172, 37–46.
Suda, A., Makino, T., 2016. Functional effects of manganese and iron oxides on the dynamics of trace elements in soils with a special focus on arsenic and cadmium: a
review. Geoderma 270, 68–75.
Sullivan, P.J., 1977. The principle of hard and soft acids and bases as applied to exchangeable cation selectivity in soils. Soil Sci. 124, 117–121.
Uddin, M., 2017. A review on the adsorption of heavy metals by clay minerals, with
special focus on the past decade. Chem. Eng. J. 308, 438–462.
Uwamariya, V., Petrusevski, B., Lens, P.N.L., Amy, G.L., 2016. Effect of pH and calcium
on the adsorptive removal of cadmium and copper by iron oxide-coated sand and
granular ferric hydroxide. J. Environ. Eng. 142 (C4015015), 1–9.
Worch, E., 2012. Adsorption Technology in Water Treatment. Hubert and Co, GmbH and
Co, Gottigen.
Yan, L., Li, S., Yu, H., Shan, R., Du, B., Liu, R., 2016. Facile solvothermal synthesis of
Fe3O4/bentonite for efficient removal of heavy metals from aqueous solution. Powder
Technol. 301, 632–640.
Yavuz, O., Altunkaynak, Y., Guzel, F., 2003. Removal of nickel, cobalt and manganese
from aqueous solution by kaolinite. Water Res. 37, 948–952.
Yuan, G., Wada, S., 2012. Allophane and imogolite nanoparticles in soil and their environmental applications. In: Barnard, A.S., Guo, H. (Eds.), Nature's Nanostructures.
Pan Stanford Publishing Pte. Ltd., Singapore, pp. 493–553.
Zhi-rong, L., Shao-qi, Z., 2010. Adsorption of copper and nickel on Na-bentonite. Process.
Saf. Environ. Prot. 88, 62–66.
capacities of low-cost sorbents for wastewater treatment: a review. Sustain. Mater.
Technol. 9, 10–40.
Erdem, B., Ozcan, A., Gok, O., Ozcan, A.S., 2009. Immobilisation of 2,2-dipyridyl onto
bentonite and its adsorption behaviour of copper(II) ions. J. Hazard. Mater. 163,
418–426.
Escudey, M., Galindo, G., Ervin, J., 1986. Effect of iron oxide dissolution treatment on the
isoelectric point of allophanic soils. Clay Clay Miner. 34, 108–110.
Escudey, M., Forster, J.E., Galindo, G., 2004. Relevance of organic matter in some chemical and physical characteristics of ash-derived soils. Commun. Soil Sci. Plant Anal.
35, 781–797.
Evangelou, V.P., 1998. Environmental Soil and Water Chemistry Principles and
Applications. John Wiley & Sons Inc., Toronto.
Filimonova, S., Kaufhold, S., Wagner, F., Hausler, W., Kogel-Knabner, I., 2016. The role of
allophane nano-structure and Fe oxide speciation for hosting soil organic matter in an
allophanic Andosol. Geochim. Cosmochim. Acta 180, 284–302.
Freud, P.J., 2007. Nanoparticle Sizing - Dynamic Light Scattering Analysis in the
Frequency Spectrum Mode. (Application Report: SL-AN-14 Revision E). Retrieved
from Microtrac Inc. website: https://www.microtrac.com/MTWP/wp-content/
uploads/2012/10/Microtrac-Application-Notes-Nanoparticle-Sizing-Dynamic-LightScattering-Analysis-in-Frequency-Spectrum-Mode.pdf.
Giles, C., Smith, D., 1974. A general treatment and classification of the solute adsorption
isotherm. I. Theoretical. J. Colloid Interface Sci. 47, 755–765.
Gupta, S.S., Bhattacharyya, K.G., 2008. Immobilisation of Pb(II), Cd(II) and Ni(II) ions on
kaolinite and montmorillonite surfaces from aqueous medium. J. Environ. Manag. 87,
46–58.
Hamidpour, M., Kalbasi, M., Afyuni, M., Shariatmadari, H., Holm, P., Christian, H.,
Hansen, B., 2010. Sorption hysteresis of Cd (II) and Pb (II) on natural zeolite and
bentonite. J. Hazard. Mater. 181, 686–691.
Hashim, M., Mukhopadhyay, S., Narayan, J., Sengupta, B., 2011. Remediation technologies for heavy metal contaminated groundwater. J. Environ. Manag. 92, 2355–2388.
Hegazi, H., 2013. Removal of heavy metals from wastewater using agricultural and industrial wastes as adsorbents. HBRC. J. 9, 276–282.
Henmi, T., Wada, K., 1976. Morphology and composition of allophane. Am. Mineral. 61,
379–390.
Hirzel, J., Undurraga, P., Gonzalez, J., 2011. Chemical properties of volcanic soil affected
by seven-year rotations. Chil. J. Agr. Res. 71, 304–312.
Hua, M., Zhang, S., Pan, B., Zhang, W., Lv, L., Zhang, Q., 2012. Heavy metal removal from
water/wastewater by nanosized metal oxides: a review. J. Hazard. Mater. 211–212,
317–331.
Hunter, R.J., 1983. Zeta Potential in Colloid Science. Principles and Applications.
Academic Press, London.
Ihsanullah, A.A., Al-Amer, A.M., Laoui, T., Al-Marri, M.J., Nasser, M.S., Khraisheh, M.,
Atieh, M.A., 2016. Heavy metal removal from aqueous solution by advanced carbon
nanotubes: critical review of adsorption applications. Sep. Purif. Technol. 157,
141–161.
Itami, K., Yanai, J., 2006. Sorption and desorption properties of Cadmium and Copper on
soil clays in relation to charge characteristics. Soil Sci. Plant Nutr. 52, 5–12.
Jitniyom, K., Suddhiprakarn, A., Kheoruenromne, I., 2012. In: Adsorption of lead, zinc,
copper and cadmium of smectite. Proceedings of the 50th Kasetsart University
Annual Conference, 2012. 2. Kasetsart University, Thailand, pp. 136–143.
de Kanter, M., Meyer-Kirschner, J., Viell, J., Mitsos, A., Kather, M., Pich, A., Janzen, C.,
2016. Enabling the measurement of particle sizes in stirred colloidal suspensions by
embedding dynamic light scattering into an automated probe head. Measurement 80,
92–98.
Kaufhold, S., Kaufhold, A., Jahn, R., Brito, S., Dohrmann, R., Hoffmann, R., Gliemann, H.,
Weidler, P., Frechen, M., 2009. A new massive deposit of allophane raw material in
Ecuador. Clay Clay Miner. 57, 72–81.
Kumar-Saha, U., Iwasaki, K., Sakurai, K., 2003. Desorption behavior of Cd, Zn and Pb
sorbed on hydroxyaluminum- and hydroxyaluminosilicate-montmorillonite complexes. Clay Clay Miner. 51, 481–492.
Kunze, G., Dixon, J., 1965. Pretreatment for mineralogical analysis. In: Campbell, G.S.,
Jackson, R.D., Mortland, M.M., Nielsen, D.R., Klute, A. (Eds.), Methods of Soil
Analysis Part 1 - Physical and Mineralogical Methods. American Society of
Agronomy, Madison, pp. 568–577.
Latrille, C., Denaiz, L., Lamy, I., 2003. Interaction of copper and zinc with allophane and
79