Selective Cu2adsorption and recovery from contaminated water using [602525]

Selective Cu2+adsorption and recovery from contaminated water using
mesoporous hybrid silica bio-adsorbents
Mihaela Mureseanua, Nicoleta Cioateraa, Ion Trandafira, Irina Georgescub, François Fajulac,
Anne Galarneauc,⇑
aFaculty of Chemistry, University of Craiova, 107I Calea Bucuresti, 200144 Craiova, Romania
bFaculty of Chemical Engineering and Environmental Protection, Technical University of Iasi, 71 D. Mangeron, Iasi, Romania
cInstitut Charles Gerhardt Montpellier, UMR 5253 CNRS/UM2/ENSCM/UM1, ENSCM, 8 rue de l’Ecole Normale, 34296 Montpellier Cedex 5, France
article info
Article history:
Received 17 February 2011Received in revised form 15 April 2011Accepted 20 April 2011
Available online 28 April 2011
Keywords:
Mesoporous silica
MetallothioneinCopper adsorptionHybrid materialBiocomplexantabstract
Metallothioneins (MTs) are low-molecular weight proteins (1–10 kDa), which are known to bind selec-
tively metal ions such as Zn or Cd in metal–thiolate clusters. The present work describes the preparation
of copper–metallothionein (Cu–MT) and its immobilization by covalent grafting on mesoporous silica for
the selective uptake and recovery of Cu2+from water. The mesoporous silica used (SiDav) features 10 nm
pore size suitable to accommodate Cu–MT (6 nm size) and 200 lm particle size adequate for flow pro-
cesses. For the covalent coupling, SiDav was first functionalized with aminopropyl (SiDav–NH 2) or glyci-
doxypropyl (SiDav–Gly) functions before to react with Cu–MT. After decomplexation of Cu, the resultingMT–SiDav–NH
2and MT–SiDav–Gly materials were used to adsorb Cu2+from aqueous solutions in the
presence of various competing cations. The adsorption capacity of the hybrid biocomplexant silica mate-
rials was studied in batch and in column for flow process. Starting from a solution containing 2 mM offour cations, the maximum adsorption capacity under flow (1 mL/min, pH 6) was obtained forMT–SiDav–NH
2with a high selectivity for Cu2+:C u2+(0.210 mmol g/C01)/C29Cd2+(0.009 mmol g/C01)>Z n2+
(0.005 mmol g/C01)>P b2+(0.003 mmol g/C01). Furthermore, adsorbed Cu2+ions were quantitatively recov-
ered by simply eluting the column with HCl. This column was also successfully used to preconcentrateCu
2+contained in different water samples as tap or mineral waters for an easier analysis of Cu traces.
/C2112011 Elsevier Inc. All rights reserved.
1. Introduction
The present work illustrates the feasibility of using selective
biological chelators anchored at the surface of mesoporous silica
supports as reusable materials for the selective removal and recov-
ery of metal ions from contaminated waters. This material can also
be used to preconcentrate traces of cations for an easier analysis of
their content in different water samples, such as tap water or min-
eral waters. While synthetic ligands or chelators are widely stud-
ied, the use of bio-adsorbents prepared from biomass of bacteria,
fungi and algae are less examined, but biocomplexants have sev-
eral advantages over synthetic chemical ligands such as high selec-
tivity to various ions depending on their tertiary structures and
relatively mild conditions for adsorption and desorption [1]. Pep-
tides and proteins could be efficient chelating ligands, due to their
amino acids residues that contain metal binding functional groups
and because they can be produced at low cost. Previously, we had
shown that a small protein, named pyoverdin ( Mw= 1.1 kDa), anatural Fe3+chelator from a Pseudomonas fluorescens strain, cova-
lently anchored through a glycidoxypropyl linker into mesoporous
silica materials demonstrated very high selectivity and sorption
capacity for Fe3+ions contained in an aqueous multimetallic solu-
tion. Amount as high as 1.33 mg of Fe3+per g of solid was adsorbed
selectively from a solution of 40 mg/L of each metal (Fe3+,C r3+,
Cu2+,N i2+,Z n2+,C o2+,P b2+) and Fe3+was totally recovered after elu-
tion [2]. To complete this research in the selective uptake and
recovery of metals contained in waters, our attention was focused
on another family of biological chelators named metallothioneins.
Metallotheoneins are small metal-binding proteins found in all liv-
ing organisms. They regulate the amounts of heavy metals in a cell.
Metallothionein (MT) is a family of cysteine-rich, low-molecular
weight proteins ( Mwranging from 0.5 to 14 kDa). MTs have the
capacity to bind both physiological (such as zinc, copper, selenium)
and xenobiotic (such as cadmium, mercury, silver, arsenic) heavy
metals through the thiol group of its cysteine residues, which repre-
sents nearly the 30% of its amino acidic residues [3]. Cysteine is a
sulfur-containing amino acid, hence the name ‘‘-thionein’’. When a
metal enters a cell, it can be picked up by thionein, which thus
becomes metallothionein. In humans, there are four main isoforms,
which production is dependent on the type of metal exposure
1387-1811/$ – see front matter /C2112011 Elsevier Inc. All rights reserved.
doi:10.1016/j.micromeso.2011.04.026⇑Corresponding author.
E-mail address: anne.galarneau@enscm.fr (A. Galarneau).Microporous and Mesoporous Materials 146 (2011) 141–150
Contents lists available at ScienceDirect
Microporous and Mesoporous Materials
journal homepage: www.elsevi er.com/locate/micromeso

(zinc, copper and selenium). Each MT molecule is able to capture
several cations, such as 7 atoms/molecule for Zn, Cd, Pb metal ions
and 18 atoms/molecule for Ag and Hg [4]. Their occurrence is re-
ported not only for humans, but also throughout the animal king-
dom, in plants and for several microorganisms, such as yeasts. MTs
are produced by yeasts in response to metal threats provoked by ele-
vated concentrations of some metals [5]. The yeast Saccharomyces
cerevisiae has been commonly used to investigate copper incorpora-
tion in eukaryotic cells [6]and to study the properties of Cu–MT
within the cellular environment. Cu–MT from yeast has been used
in this work and is characterized by 53-residue polypeptide
(5.6 kDa) containing 12 cysteines and 8 copper atoms/molecule
[7]. Furthermore the unique feature of MT is the competition be-
tween protons and metal ions for protein binding. Protons can re-
place metals from MT and thus metal ions are released from theprotein gradually with decreasing pH. Therefore, a simple addition
of acid allows the metal recovery. As examples, Zn in MT is com-
pletely released at a pH lower than 4 and the same situation occurs
for Cd at pH 2 [8]. Lower pH values for release are related to a higher
binding affinity of the metal to MT. After removal of metals, in the
presence of metal ions, the H
+–MT can refold to the correct configu-
ration when the pH is raised to neutral [9]. This characteristic makes
MT a useful tool for the development of bio-adsorbents for the up-
take and recovery of metals because the protein can be used
repeatedly.
Since the MT capacity to bind metals such as zinc and cadmium
has been demonstrated (see below Ref. [10]) with 7 metal ions/MT,
we have attempted to achieve the selective and recyclable uptake
of copper from water, by producing a Cu–MT from a culture med-
ium with a high copper concentration and by using this natural li-
gand bounded to a mesoporous silica support. Among porous silica
supports, MCM-41-like material demonstrated superior perfor-
mance for iron uptake and release using pyoverdin as biochelator
compared to classical silica-gel due to a lower density of surface
silanol groups [2]. In the present case however, silica-gel was pre-
ferred in order to fulfill two main conditions: a large pore diameter
(10 nm) to accommodate MT (5.6 kDa, /C245/C26 nm) (which is much
larger than pyoverdin) and uniform particles of size large enough
(200lm) to perform studies in column under flow process without
drop pressure. Indeed the existence of such uniform large particles
of MCM-41 (without any aggregation) featuring large pores are not
yet available, even by the pseudomorphic synthesis [11–13] . Cu–
MT was expressed from yeast in a concentrated solution of copper
to induce the formation of the isoform selective to copper, and the
resulting Cu–MT was immobilized on silica support. Methods of
physical or chemical immobilization of biomolecules on solid sur-
faces are well known [14]. For instance, MT has been immobilized
on supports by several methods including unspecific adsorption/chemisorption, covalent binding or physisorption [15]. However,
the supports were almost always polymeric resins or organic gels
with limited capacity for reuse [16]. Inorganic carriers are more
expensive than their organic commercial counterparts but have
the advantages of being stable and reusable, which in some cir-
cumstances can decrease the effective cost of the process. The
encapsulation of MT (from Sccizosaccharomyces pombe ) into a
sol–gel silica has been studied by Bahrami et al. [10] for Cd and
Zn removal from water separately. Their experiments demon-
strated that MT has a greater capacity for Cd and Zn than non-bio-
logical chelators such as polyethyleneimine or EDTA. Furthermore,
the use of MT compared to non-specific biomass for metal adsorp-
tion offers the advantages of high metal binding capacities and
selectivity.
The present work describes the immobilization of Cu–MT (from
baker yeast) into a mesoporous silica previously functionalized
with amino or glycidoxy groups to covalently bind MT to the mate-
rial and its use for the selective uptake of Cu
2+from a multimetallicaqueous solution in batch and in flow process and the recovery of
the Cu2+ions.
2. Experimental
2.1. Reagents and materials
Analytical reagent-grade chemicals were used to prepare the
solutions required for the biosynthesis and purification of MTs
(from baker yeast). Freshly prepared Milli Q (Millipore) ultra pure
water was used in all experiments. The support used was a com-
mercial silica obtained as a gift from Grace-Davison named Davicat
Si 1452 (named below SiDav) with 200 lm particle size, 390 m2/g
specific surface area, 1.1 mL/g pore volume, 10 nm pore diameter.
Before the covalent coupling of MT, this support was functional-ized with 3-aminopropyl-triethoxysilane (APTES, Fluka) or 3-glyci-
doxypropyl-trimethoxisilane (GPTMS, Fluka) accordingly to Ref.
[2]. Stock solutions (1000 ppm) of Cu(NO
3)2, Cd(NO 3)2, Zn(NO 3)2
and Pb(NO 3)2were prepared in distilled water. The pH was ad-
justed between 2 and 8 using diluted solutions of HCl or NaOH.
2.2. Apparatus
Electronic absorption spectra were recorded on a UVIKON XL
UV–vis spectrometer from Bio-Tek Instruments and diffuse reflec-
tance in the UV–vis range were obtained on a Lambda 14 spec-
trometer (Perkin-Elmer Inc., Shelton, USA) with an integrating
sphere (Labsphere, North Sutton, USA) for solid samples. The latter
were held in 0.05 nm thick cuvettes (100 QS, Helma, Mulheim,
Germany). Atomic adsorption spectroscopy (FAAS) measurements
were performed on a Spectra AA-220 Varian Spectrometer with
an air–acetylene flame. Thermogravimetric analysis was carriedout in a Netzsch TG 209C thermobalance. About 15 mg of solid
sample was loaded, and the air-flow used was 50 cm
3min/C01. The
heating rate was 20 /C176C min/C01and the final temperature was 850 /C176C.
2.3. Procedures
2.3.1. Production, isolation and characterization of the metallothionein
Cu–MT was excreted from baker yeast cells grown in a medium
containing 150 mg L/C01of Cu2+, according to the method of Weser
and Hartmann [17]. Baker yeast was cultivated in a medium com-
posed of 25 g yeast extract, 50 g gelatine hydrolisate, 175 g glu-
cose, 1.5 g NaCl, 1.5 g KH 2PO4and 375 mg L/C01CuSO 4in 3 L of
water. After 48 h, the cells were harvested by centrifugation at
3000 rpm for 15 min and were washed three times, each in 10
vol. of deionised water. The total Cu2+concentration in the yeast
was determined by FAAS. Washed cells (20 mg dry weight of cells)
were digested by adding 1 mL 6 N HNO 3in boiling water for
20 min. After this acid extraction, the samples were diluted to
5 mL with distilled water, centrifuged and the copper content
was measured. In order to isolate Cu–MT, the copper loaded
washed yeast cells were suspended in an equal volume of 20 mM
Tris/HCl buffer, at pH 7 and ruptured in a Polytron homogenizer
(0–4 /C176C). The homogenate, diluted two fold with 10 mM Tris/HCl
buffer (pH 7), was centrifuged at 10,000 rpm for 1 h. The total
supernatant solution was lyophilized. This powder was dissolved
in a minimum volume (<2 mL) of N 2-saturated 10 mM Tris/HCl,
at pH 7. Then Cu–MT complexes were purified by size exclusion
chromatography. The solution was applied to a Sephadex G-75 size
exclusion chromatography column (80 /C22.5 cm) with a 10 mM
Tris/HCl (pH 7) mobile phase in the presence of 0.1% 2-mercap-
toethanol to avoid uncontrolled oxidation of thiolate sulfur.
The fractions from the Mw10 K region were chromatographed
on Sephadex G-50 column (2.5 /C240 cm) equilibrated with142 M. Mureseanu et al. / Microporous and Mesoporous Materials 146 (2011) 141–150

N2-saturated 10 mM Tris/HCl (pH 7). The detection and localization
of Cu–MT during the preparation was controlled by copper analysis
by FAAS in the respective chromatographic fractions and UV elec-
tronic absorption spectra. The presence of copper within these
complexes was confirmed by FAAS. The H+–MT was prepared using
the chelating agent diethyldithiocarbamate (DTC) [18]. The pH of
the Cu–MT solution as isolated by G 50 gel filtration was adjusted
to pH 5.0 by using 3 M sodium acetate (final concentration of ace-
tate 0.1 M). One milligram of solid DTC/mL of solution was added
and the sample was incubated at room temperature for 1 h. The
colloidal Cu–DTC complex was removed by filtration through a
0.22lm filter. The slightly yellow filtrate was desalted on a Sepha-
dex G-25 column in 0.1% TFA and the fractions containing H+–MT
were freeze-dried.
2.3.2. Immobilization of Cu–MT in mesoporous silica by adsorption
The SiDav support (1 g) was activated at 180 /C176C under vacuum
for 1 h. A column was filled with the solid material and 14 mL of
a solution containing 437 mg Cu–MT in 10 mM Tris/HCl buffer
(pH 7) containing 0.1% of 2-mercaptoethanol was recirculated with
a peristaltic pump. The copper concentration in this solution was
determined by FAAS technique and the Cu–MT solution recircula-
tion was continued until the inlet copper concentration is the same
with the outlet concentration. The obtained adsorbent (Cu–MT–Si-
Dav) was washed several times with water until no copper was de-
tected in washing, and was then dried under vacuum at room
temperature and stored at /C05/C176C.
2.3.3. Immobilization of Cu–MT into amino- or glycidoxy-
functionalized mesoporous silica
Cu–MT was immobilized onto amino- or glycidoxy-functional-
ized silica by covalent binding. The silica support was firstly silan-
ized with APTES or GPTMS according to the previously described
procedure [2,19] . SiDav was freshly activated overnight at 180 /C176C
under vacuum (1 g), and APTES or GPTMS (1 mL) were mixed in
50 mL of dry toluene. After stirring the solution (reflux, 2 h), the re-
leased ethanol was distilled off and the mixture was kept under re-
flux for 90 min. The functionalized silica (referred as SiDav–NH 2
and SiDav–Gly, respectively) were filtered and washed with tolu-ene, ethanol and then diethyl ether. They were then submitted to
a continuous extraction run overnight in a Soxhlet apparatus using
diethyl ether/dichloromethane (v/v, 1/1) at 100 /C176C and dried over-
night at 130 /C176C.
The procedure of Cu–MT immobilization onto the amino-func-
tionalized support involves the mixing of 1 g SiDav–NH
2with
4 mL of 0.1 M sodium phosphate buffer (pH 7). One milliliter of
an aqueous 25 wt% glutaraldehyde solution was added leading to
a total concentration of 0.54 M of glutaraldehyde (or 2.5 mmol glu-taraldehyde/g SiDav–NH
2) and the mixture was stirred for 30 min
at room temperature. Then, the excess of glutaraldehyde was re-
moved during three cycles of centrifugation/washing with 10 mL
buffer solution each and the resulting solid was dispersed in
14 mL of 10 mM Tris/HCl buffer solution (pH 7) containing
460 mg lyophilized Cu–MT corresponding to 0.08 mmol Cu–MT
per g of SiDav–NH 2. The suspension was stirred at 25 /C176C for 24 h,
centrifuged at 3000 rpm for 10 min to remove the buffer, and
washed several times with water until no metallothionein was de-
tected by UV in the washing. The final solid, referred as Cu–MT–Si-
Dav–NH 2, was dried under vacuum and stored at /C05/C176C.
For the covalent coupling of Cu–MT onto GPTMS-functionalized
silica, SiDav–Gly (1 g) was activated under vacuum at 130 /C176C for
1 h. The solid was dispersed in DMF (14 mL) before adding
460 mg of lyophilized Cu–MT (corresponding to 0.08 mmol Cu–
MT per g of SiDav–Gly) dissolved in the minimum amount of water
[2]. The mixture was heated at 60 /C176C under stirring, for 48 h. The
resulting material was collected by filtration and washed severaltimes with water. The final material, referred as Cu–MT–SiDav–
Gly was then dried under vacuum and stored at /C05/C176C.
Aliquots of the rinsing waters were analyzed by FAAS to quan-
tify the amount of complex that was not covalently coupled. In par-
allel to quantify the amount of Cu immobilised into the silica
support, a decomplexation of the entrapped Cu was performed
using the strong chelating agent DTC [20]. After the solubilization
of Cu–DTC complex in CCl 4, the amount of copper was quantified.
The total amounts of immobilized Cu–MT in the different silica
supports were evaluated from the TG/DTA curves and compared
with the results of the FAAS analysis.
2.3.4. Decomplexation of copper from Cu–MT immobilized in the silica
supports
In order to use the immobilized MT in adsorption studies, Cu2+
was decomplexed from the Cu–MT. For 1 g of Cu–MT–SiDav–NH 2
(or other hybrid material containing Cu–MT), 50 mg DTC in10 mL of a 0.1 M sodium acetate buffer (pH 5.0) were used. The so-
lid was then washed several times with CCl
4and ethanol and dried
at 80 /C176C.
2.3.5. Cations adsorption studies over MT–silica materials in batch
mode
The sorption of Cu2+,C d2+,Z n2+and Pb2+metal ions from aque-
ous solutions was investigated in batch mode at room tempera-
ture. All experiments were conducted in acid-washed PTFE
bottles by stirring (300 rpm) 0.1 g of dried hybrid materials with
20 mL aqueous solutions (5 g L/C01) with the desired initial concen-
tration of metal ions. The contents of the bottles were equilibrated
at room temperature for 2 h. After equilibration, samples for metal
analysis were taken after filtration through a 0.45 lm nitrocellu-
lose membrane. The samples where then diluted with distilled
water to be within 0.05–20 mg/L. The concentrations of Cu2+,
Cd2+,Z n2+and Pb2+were determined by flame atomic absorption
spectrometry (FAAS).
2.3.5.1. Influence of pH. The influence of the pH on the adsorption
capacity of Cu2+,C d2+,Z n2+,P b2+ions onto the hybrid biochelatant
mesoporous silica MT–SiDav–NH 2has been performed at room
temperature using separated solutions of 0.5 mmol L/C01of each cat-
ion. The pH in the range 2–8 was adjusted using diluted solutions
of HCl or NaOH.
2.3.5.2. Influence of ionic strength. The influence of NaCl concentra-
tion at pH 6 on the adsorption capacity of Cu2+,P b2+,C d2+and Zn2+
into MT–SiDav–NH 2has been performed using separated solutions
of 1 mmol L/C01of each metal ions.
2.3.5.3. Adsorption isotherms. Adsorption isotherms of Cu2+were
established for SiDav, SiDav–NH 2, MT–SiDav–NH 2materials at pH
6 by increasing Cu solution concentration from 0.1 to 5 mmol L/C01.
After adsorption, the concentration of remaining Cu was evaluated
by FAAS and corresponds to the equilibrium concentration, Ce
(mmol L/C01). The Cu adsorption capacity of materials, qe(mmol g/C01),
was plotted as a function of the equilibrium concentrations of Cu2+
ions, Ce(mmol L/C01).
2.3.5.4. Ions adsorption kinetics. Adsorption kinetics experiments
were carried out to determine the adsorption rates and the reac-
tion time to reach adsorption equilibrium. The effect of contact
time on the uptake of metal ions by the hybrid materials was
investigated at pH 6.0 by shaking 0.1 g of dried material in 20 mL
of solutions with initial concentration in metal ions of 1 mM for
each separated solution of cations. Samples were taken at different
time intervals and the residual concentration of metal was deter-
mined by FAAS.M. Mureseanu et al. / Microporous and Mesoporous Materials 146 (2011) 141–150 143

2.3.5.5. Competitive adsorption over MT–SiDav–NH 2.The competi-
tive adsorption of ions in batch mode was performed with MT–Si-
Dav–NH 2at pH 6. Selectivity coefficients were determined by
shaking 0.1 g of dried material in 20 mL of a multimetallic solution
containing Cu2+,C d2+,Z n2+and Pb2+ions with a concentration of
2 mM of each ion.
2.3.6. Cations adsorption studies over MT–silica materials in flow
through column
Continuous flow experiments were performed using plastic col-
umns with a length of 3 cm and a diameter of 1 cm. The hybrid
materials (0.5 g) were placed between two small sand beds and
were separated on top and bottom by two 45 lm nitrocellulose
membranes.
2.3.6.1. Breakthrough curves. The Cu2+solution having an initial
concentration of 0.8 mM (50 ppm) was allowed to flow downward
through the column at the desired flow rate between 1 and
3 mL min/C01by using a peristaltic pump. Samples were collected
from the outlet of the column at different time intervals and ana-
lyzed for Cu2+concentration. The operation of the column was
stopped when the initial Cu2+concentration matches the outlet
Cu2+concentration. The outlet Cu2+concentrations were plotted
versus volume to give the breakthrough curves. The outlet Cu2+
concentration was also plotted against flow rate.
2.3.6.2. Competitive adsorption over MT–SiDav–NH 2.The competi-
tive adsorption experiment of metal ions in continuous flow was
carried out at pH 6 by passing through a column filled with MT–Si-
Dav–NH 2100 mL of a multimetallic solution containing Cu2+,P b2+,
Cd2+and Zn2+ions with 2 mM concentration of each ion at a flow
rate of 1 mL min/C01.
2.3.6.3. Column regeneration and copper recovery. Column regenera-
tion and copper recovery was achieved by washing the column
with 1 M HCl solution until no Cu2+was detected in the effluents.
Thereafter, the column was carefully washed with water until neu-
tral pH.
2.3.6.4. Preconcentration effect of MT–SiDav–NH 2and quantification
of Cu2+traces. In order to determine the preconcentration factor a
solution with the initial concentration of 0.08 mM (5 ppm) Cu2+
was passed through the column of MT–SiDav–NH 2, with a flow rateof 1 mL min/C01and the outlet copper concentration was measured.
The effluent volume that allows the maximum dynamic loading
capacity was considered until no cooper was measured in the efflu-
ents. Thereafter the column was washed with water and the ad-
sorbed copper was eluted into the minimum volume of 1 M HCl
for which the metal ion recovery was the best. The preconcentra-
tion factor was calculated as the ratio between the volume of efflu-
ent used for the loading capacity and the volume of 1 M HCl
solution necessary for Cu2+recovery.
2.3.6.5. Copper determination in natural and mineral water sam-
ples. Tap water (pH 6.4, conductivity 150 lS/cm) and commercial
mineral water (pH 6.6, conductivity 460 lS/cm) were used without
any pretreatment. Samples of 300 mL of tap or mineral water
unspiked and spiked (with 10 lgL/C01Cu2+solution) were passed
through a column filled with MT–SiDav–NH 2at a flow rate of
1 mL min/C01. Then the column was washed with 20 mL deionized
water. The adsorbed copper was eluted with 6 mL of 1 M HCl and
copper content determined by FAAS.
3. Results and discussion
3.1. Properties of metallothionein and its copper chelate in solution
Copper–metallothionein (Cu–MT) was excreted from baker
yeast and purified. Two significant copper containing fractions
were identified by size exclusion chromatography. The more abun-
dant fraction is a low-molecular (LMW) component with an elution
time of 120–150 min. The second component is a high-molecular-
weight (HMW) complex, which eluted from the column in the void
volume (elution time /C2460 min). The UV–vis absorption spectrum
of the LMW component showed a resolved shoulder at 260 nmand a broad signal at approximately 340 nm while a single broad
absorption shoulder at about 270 nm was obtained from the he
HMW copper-containing component. In agreement with previous
studies only the LMW protein can be identified as Cu–MT [20].
The wet cells contained 1.1 mg copper/g cell (determined by
FAAS analysis) that correspond to 12.55 mg MT/g cell if 8 Cu per
MT has taken in the calculation. Indeed, the crystal structure of
yeast Cu–MT has been resolved recently (in 2005) by Calderone
et al. [21] and shows cluster of 8 copper atoms complexed through
Cys–S–Cu bonds ( Scheme 1 ). The Cu–MT structure shows the larg-
est known oligonuclear Cu thiolate cluster in biology, consisting of
O
OSiO
OO
+H2N MT-Cu
O
OSiO
OOH
N MT-CuHbO
OSiO
NH2
H2N MT-CuO O
O
OSiO
N
O
O
OSiO
N N MT-Cuc
H2N MT-Cua
Scheme 1. (a) Schematic representation of Cu–MT with its cluster of eight copper atoms adapted from Ref. [21] and Cu–MT immobilization by covalent binding onto (b)
GPTMS (SiDav–Gly) and (c) APTES (SiDav–NH 2) functionalized silica support.144 M. Mureseanu et al. / Microporous and Mesoporous Materials 146 (2011) 141–150

six trigonally and two diagonally coordinated Cu ions. Cu–MT
shows a different structure of the metal cluster in comparison to
other metallothionein, but also the main differences lie in the cys-
teine topology and in the conformation of some portions of the
backbone.
The H+–MT was prepared using the chelating agent diethyldi-
thiocarbamate (DTC) by the method proposed by Hunziker [18].
The amount of decomplexed copper corresponded to that deter-
mined for the Cu–MT samples by FAAS. All the copper complexed
by MT was depleted by DTC chelating agent. After extraction and
purification 500 mg of Cu–MT were recuperated from 118 g of
wet cells.
3.2. Metallothionein immobilized on mesoporous silica supports
Cu–MT was immobilized on mesoporous silica support (SiDav)
by adsorption and different covalent graftings. SiDav features
10 nm pore diameter, large enough to accommodate Cu–MT
(/C245–6 nm), and particle size of 200 lm to allow the study of
adsorption in flow process without pressure drop. For the covalent
coupling of Cu–MT, SiDav was first functionalized with 3-amino-
propyltriethoxysilane (SiDav–NH 2) or with 3-glycidoxypropyltri-
methoxysilane (SiDav–Gly) through a silanization reaction
(Scheme 1 )[2,19] .
The TGA profile of SiDav, SiDav–NH 2and SiDav–Gly are pre-
sented in Fig. 1 . The SiDav support exhibits a first mass loss
(1.42%) in the temperature range 20–120 /C176C, corresponding to
the physically adsorbed water. The second weight loss (2.77%) be-
tween 250 and 800 /C176C is attributed to water loss due to the con-
densation of the silanols. The amount of silanols is therefore
estimated at 3.08 mmol SiOH/g silica and (if there are all in sur-
face) to a density of 4.7 SiOH/nm2, which is a classical value for sil-
ica gel materials. In the case of functionalized silica materials, the
loss of physisorbed water (below 120 /C176C) was about 2%. Consider-
ing the TGA profiles, the degradation of the anchored organic moi-
eties occurs between 150 and 800 /C176C. The corresponding weight
losses were 8.2% and 11.7% for SiDav–NH 2and SiDav–Gly, respec-
tively. This corresponds to an anchorage capacity of 91.3 mg of
propylamine and 135 mg of propylglycidoxy per g of silica (cal-
cined material) for SiDav– and SiDav–Gly, respectively. TGA results
are in good accordance with the values found by elemental analy-
sis corresponding to 102 and 122 mg of grafting agent per g of sil-
ica for SiDav–NH 2and SiDav–Gly, respectively ( Table 1 ), though
chemical analysis slightly overestimates the amount of amino
groups whereas it underestimates the amount of glycidoxy. For
comparison purposes and self-consistency, the grafting densities
have been calculated from TGA results. For SiDav–NH 2, this leadsto a density of 1.60 mmol NH 2/g silica or 1.72 mmol NH 2/g native
silica (2.66 NH 2/nm2), if additional silica due to the grafting is ta-
ken into account in the calculation with the hypothesis of a biden-
tate grafting (the value becomes 2.60 NH 2/nm2for a tridentate
grafting hypothesis). For SiDav–Gly, TGA result indicates a density
of 1.18 mmol Gly/g silica or 1.27 mmol Gly/g native silica (1.96
Gly/nm2) with a bidentate grafting hypothesis (or 1.91 Gly/nm2
with a tridentate grafting hypothesis). The grafting density valuesare in agreement with previous results found for the grafting of or-
ganic functions on silica gel and MCM-41 with values ranging from
1 to 2.5 molecules/nm
2with an expected higher grafting density
for aminopropyl groups in comparison to less polar functionalchains [2,19] .
The functionalized silica materials were used thereafter for the
Cu–MT immobilization by direct covalent coupling for SiDav–Gly
and through the glutaraldehyde intermediate coupling agent for
SiDav–NH
2(Scheme 1 ) to obtain the hybrid biocomplexant silica
materials, named Cu–MT–SiDav–Gly and Cu–MT–SiDav–NH 2,
respectively. Cu–MT was immobilized rather than H+–MT in order
to preserve the conformation of the protein during the grafting, as
it was previously done for the Fe–pyoverdin grafting in mesopor-
ous silica materials [2], and to protect the functional groups in-
volved in the metallic ions complexation, as the –SH groups of
cysteine may oxidize into –S–S– bridges. The TGA profiles of all hy-
brid biocomplexant silica materials Cu–MT–SiDav, Cu–MT–SiDav–
NH 2and Cu–MT–SiDav–Gly are presented in Fig. 2 . The loss of the
organic moieties occurred between 150 and 800 /C176C and corre-
sponds to the loss of MT and the organic function used for the silica
grafting.
The amount of immobilized Cu–MT was determined both by
TGA and by copper mass balance by FAAS ( Table 2 ). The results ob-
tained by both methods stand in good agreement. Thus, for Cu–MT
adsorbed on SiDav (Cu–MT–SiDav) only 1% of the Cu–MT engaged
in the physical adsorption procedure was immobilized onto the
support, which corresponds to 0.001 mmol Cu–MT g/C01silica.
100 200 300 400 500 600 700 8008790939699
(4)
(3)(2)Weight loss (%)
Temperature ( °C)(1)
Fig. 1. Thermogravimetric analysis (TGA) of (1) mesoporous silica SiDav, and
functionalized SiDav with (2) 3-aminopropyltriethoxysilane (SiDav–NH 2) and (3) 3-
glycidoxypropyltrimethoxysilane (SiDav–Gly). Curve (4) is SiDav–NH 2cross-linked
with glutaraldehyde.Table 1Chemical composition of the functionalized organic entities (‘‘NH
2‘‘ and ‘‘Gly’’)
grafted on SiDav materials.
Material Elemental analysis Organic group (mg g/C01silica)
C (%) N (%) H (%)
SiDav–NH 2 6.30 2.46 1.45 101.9a99.7b92.9c
SiDav–Gly 7.62 0.02 1.28 122b135.8c
aDetermined from N elemental analysis.
bDetermined from C elemental analysis.
cDetermined from TGA.
100 200 300 400 500 600 700 8007580859095100
(3)(2)Weight loss (%)
Temperature ( °C)(1)
Fig. 2. Thermogravimetric analysis (TGA) of Cu–MT-immobilized silica supports:
(1) Cu–MT immobilized by adsorption in SiDav, (2) Cu–MT covalently grafted onSiDav–Gly and (3) Cu–MT covalently grafted on SiDav–NH
2.M. Mureseanu et al. / Microporous and Mesoporous Materials 146 (2011) 141–150 145

Moreover, the solid prepared by simple adsorption Cu–MT–SiDav
was deactivated after several metal ions adsorption/desorption cy-
cles due to the progressive leaching of the biocomplexant. For Cu–
MT grafted on SiDav–Gly, the difference of weight losses in therange 150–800 /C176C between the TGA analysis of Cu–MT–SiDav–
Gly (307 mg/g silica) and SiDav–Gly (135 mg/g silica) allows to cal-
culate an amount of 172 mg Cu–MT/g silica (calcined material),
corresponding to 0.034 mmol Cu–MT/g silica or 0.037 mmol Cu–
MT/g native silica (bidentate grafting hypothesis). This represents
a density of 0.057 Cu–MT/nm
2or 17.5 nm2of surface covered per
Cu–MT molecule, equivalent to an average diameter of 4.7 nm
per molecule, so to a densely packed Cu–MT monolayer at the sur-
face of the silica support. Indeed the size of a Cu–MT molecule in a
crystal is 5 /C26n m [21].
For Cu–MT grafted on SiDav–NH 2, calculations are more com-
plex as the amount and the way of the glutaraldehyde is anchored
to the amino groups of the solid (1 or 2 bonds with two adjacent
amino groups) is unknown. In the preparation, there is enough glu-
taraldehyde to cross-link all amino groups of the surface. However
the time of reaction is very short (30 min at room temperature)
and several washings have been performed before to introduce
the solution of Cu–MT as well as after the cross-linkage with MT.
In a first hypothesis, we can assume that the amount of glutaralde-
hyde grafted corresponds to that necessary for the cross-linkage
with Cu–MT. The weight loss due to this linker (66 g/mol) would
be then negligible in comparison to the weight loss due to Cu–
MT (5000 g/mol); the other grafting species on the material surface
being the amino groups. In this case, the same calculation as above
for Cu–MT–SiDav–Gly can be performed by subtracting the weight
loss of organics of SiDav–NH 2(91.3 mg/g silica) to that of Cu–MT–
SiDav–NH 2(321 mg/g silica). A value of 230 mg Cu–MT/g silica is
obtained, which is in good agreement with the amount of Cu–MT
calculated from the amount of copper in the material determined
by FAAS (260 mg Cu–MT/g silica) ( Table 2 ). The MT amount de-
rived from TGA corresponds to 0.046 mmol Cu–MT/g silica or
0.049 mmol Cu–MT/g native silica (bidentate grafting hypothesis)
and represents a density of 0.076 Cu–MT/nm2or 13 nm2of surface
covered per Cu–MT molecule (equivalent to an average diameter of
4.1 nm per molecule), so to a higher amount of Cu–MT at the sur-
face of SiDav–NH 2in comparison to SiDav–Gly and a more densely
packed Cu–MT monolayer. However, if the intermediate material
SiDav–NH 2cross-linked with glutaraldehyde is isolated and dried,
TGA ( Fig. 1 ) shows a weight loss of 118.6 mg/g silica. By taking the
hypotheses that either (i) each glutaraldehyde is only cross-linked
with one amino group or (ii) the majority of glutaraldehyde is
cross-linked with two adjacent amino groups, the resulting
amounts of Cu–MT immobilized in SiDav–NH 2are 209 and
203 mg/g silica for each hypothesis, respectively. These values
are far from the results obtained by FAAS copper analysis
(260 mg Cu–MT/g silica). The result obtained using the TGA analy-
sis of the intermediate SiDav–NH 2–glutaraldehyde material could
arise from an excess of glutaraldehyde retained at the surface
due to the drying step, whereas some glutaraldehyde could have
been washed away during the entire preparation, which includes
several washings, of Cu–MT–SiDav–NH 2. Therefore we have chosenfor the discussion of the following sections to remain with the re-
sult of the first hypothesis (230 mg Cu–MT/g silica – 0.046 mmol
Cu–MT/g silica).
Copper decomplexation was achieved by using diethyldithio-
carbamate (DTC) as a chelating agent. All the copper complexed
by MT was depleted by DTC making the resulting H+–MT available
for a new complexation reaction.
3.3. Ions adsorption properties of the metallothionein immobilized into
mesoporous silica supports
In order to establish that the chelation of metals occurs thanks
to the presence of MT rather than on the silica matrix and/or on theamino groups, the removal of four metallic ions (Cu
2+,C d2+,Z n2+,
Pb2+) from aqueous solutions was investigated over the parent
mesoporous silica SiDav, the functionalized silica SiDav–NH 2, and
the hybrid biocomplexant materials MT–SiDav–NH 2and MT–Si-
Dav–Gly ( Table 3 ). MT covalently grafted on silica materials reveals
much higher affinity for copper in comparison to SiDav and SiDav–
NH 2as demonstrated below (Section 3.3.3 ). The two hybrid bio-
complexant materials MT–SiDav–NH 2and MT–SiDav–Gly showed
the same kind of behavior in the metal adsorption studies with a
somewhat higher capacity for MT–SiDav–NH 2due to a higher
amount of MT on this material ( Table 3 ). The amount of metal re-
tained per unit mass of silica was evaluated using the following
expression:
qe¼VðC0/C0CeȚ
Wð1Ț
where qeis the amount of metal ion adsorbed per g of silica
(mmol g/C01);Vthe volume of the aqueous phase (L); Wis the silica
weight of the material (g); C0and Ceare the concentrations of metal
ions in the initial aqueous phase and equilibrium concentrations
(mmol L/C01).
The adsorption of metal ions to the chelating agent MT immo-
bilized onto the mesoporous silica is affected by several factors.
These include the physicochemical parameters of solution such
as pH, ionic strength, initial metal concentration, contact time,
interaction with other ions, etc. The influence of these factors onthe adsorption process is discussed below. For the sake of clarity,
the only results described hereafter will be those obtained with
MT–SiDav–NH
2.
3.3.1. Influence of pH
The influence of the pH on the adsorption capacity of separated
ion (Cu2+,C d2+,Z n2+,P b2+) solutions onto the hybrid biochelatant
mesoporous silica is illustrated in Fig. 3 . This study has been
performed using separated solutions of 0.5 mM of each cation corre-
sponding to a potential amount for adsorption of 0.10 mmol ion g/C01
silica therefore to 1/4 of the expected maximum capacity of MT–Si-
Dav–NH 2for copper ions (0.046 mmol of MT g/C01silica with 8 Cu
atoms per MT). The adsorption capacity of ions increased with the
Table 3
Parameters for the Langmuir and Freundlich isotherm models for the adsorption of
Cu2+in the hybrid biocomplexant MT–SiDav–NH 2and MT–SiDav–Gly materials and
in the parent mesoporous silica SiDav.
Materials Langmuir isotherm parameters Freundlich isotherm
parameters
q0
(mmol g/C01
silica)b
(L mmol/C01)R2KF
(mmol1/C0n
g/C01Ln)nR2
MT–SiDav–NH 20.480 4.463 0.997 2.599 1.815 0.851
MT–SiDav–Gly 0.349 5.222 0.999 1.724 1.949 0.856
SiDav 0.041 1.932 0.994 41.932 1.532 0.945Table 2Amount of the Cu–MT complex immobilized on different mesoporous silica supportsper g of silica (calcined material).
Materials Cu–MT (mg g/C01silica) Cu–MT (mmol g/C01silica)
Cu–MT–SiDav 5.7a3.8b,c0.0013a,c0.0008b
Cu–MT–SiDav–NH 2 230a260b,c0.0460a,c0.0520b
Cu–MT–SiDav–Gly 172a181b,c0.0344a,c0.0362b
aDetermined by ATG.
bDetermined by FAAS.
cConsidering the molecular weight of metallothionein being Mw= 5 kDa.146 M. Mureseanu et al. / Microporous and Mesoporous Materials 146 (2011) 141–150

pH and presented a maximum around pH 6, this maximum of
adsorption is maintained until pH 8. Metallothionein being an intra-
cellular peptide, complexation of the metal ions occurs preferen-
tially at pH values close to the physiological one. In the following,
a pH of 6 was selected. The maximum of adsorption reached in this
experiment was 0.066 mmol Cu g/C01silica, corresponding to 66% of
copper removal. For the other cations, the adsorption capacity was
lower with 0.055, 0.023 and 0.006 mmol ion g/C01silica for Pb2+,
Cd2+,Z n2+, respectively.
3.3.2. Influence of ionic strength
The ionic strength of the solution also affects the metal ions
adsorption. Fig. 4 shows the influence of NaCl concentration on
the adsorption capacity of Cu2+,P b2+,C d2+and Zn2+into MT–Si-
Dav–NH 2. This study has been performed using separated solutions
with a concentration corresponding to a potential amount for
adsorption of 0.20 mmol ion g/C01silica therefore half of the ex-
pected maximum capacity of MT–SiDav–NH 2for copper ions
(0.046 mmol of MT g/C01silica with 8 Cu atoms per MT). For copper,
the maximum of adsorption reached in this experiment was
0.186 mmol Cu g/C01silica, corresponding to 93% of copper removal,
for a NaCl concentration between 0 and 40 mmol L/C01. For higher
NaCl concentration (>50 mmol L/C01), the amount of metal adsorbed
on MT–SiDav–NH 2materials drastically decreased (3-/4-fold),
indicating a conformational change of the tertiary structure of
the protein and the metals binding sites and therefore a loss of
the biocomplexant binding ability at high ionic strength.3.3.3. Adsorption isotherms
Adsorption isotherms of Cu2+were established by plotting the
adsorption capacities of MT–SiDav–NH 2,qe(mmol g/C01) as a function
of the equilibrium concentrations of Cu2+ions, Ce(mmol L/C01)(Fig. 5 ).
The comparison of the adsorption capacity of MT–SiDav–NH 2
(0.046 mmol MT/g silica) with the parent mesoporous silica materi-als SiDav (3.08 mmol SiOH/g silica) and SiDav–NH
2(1.60 mmol
NH 2/g silica) has been performed. The maximum Cu2+adsorption
capacity of MT–SiDav–NH 2amounted to 0.45 mmol g/C01silica for
an equilibrium concentration of 3.5 mmol L/C01corresponding to a to-
tal occupancy of the adsorption sites by copper ions, whereas SiDav
and SiDav–NH 2adsorbed only 0.03 and 0.26 mmol g/C01silica, respec-
tively. The hybrid biocomplexant silica material MT–SiDav–NH 2
exhibited an enhanced adsorption performance for Cu2+evidencing
the higher affinity of Cu2+for MT molecule in comparison to NH 2and
OH groups. As MT is able to complex 8 Cu ions per molecule, the
loading value for MT–SiDav–NH 2corresponds to a total occupancy
of the adsorption sites by copper ions (0.37 mmol Cu/g silica) with
some additional copper (0.08 mmol Cu/g silica) presumably linked
on amino and/or silanol groups. It is to notice that a similar behavior
is occurring for MT–SiDav–Gly (0.034 mmol MT/g silica) which
complexes 0.35 mmol Cu/g silica ( Table 3 ), corresponding to a total
occupancy of the adsorption sites by copper ions (0.27 mmol g/C01sil-
ica). There is therefore also some additional copper (0.08 mmol Cu/g
silica) presumably linked to OH groups. Moreover, the isotherm of
MT–SiDav–NH 2shows a sharp initial slope, indicating that the
material acts as highly efficient adsorbent even at low metal
concentration.
The Langmuir and Freundlich isotherm models were applied to
the experimental data in order to get a better insight into the sorp-
tion mechanism. The Langmuir adsorption isotherm describes a
homogeneous surface, assuming that all the adsorption sites have
the same activity and that the adsorption at a site does not affect
adsorption at an adjacent site. Another assumption is that the
adsorption occurs through the same mechanism and only a mono-layer is formed at the maximum adsorption. The Langmuir equa-
tion is expressed in the following equation:
Ce
qe¼Ce
q0ț1
q0bð2Ț
where qeis the amount of solute adsorbed on the surface of the
material (mmol g/C01),Ceis the equilibrium concentration of the sol-
ute (mmol L/C01),q0is the maximum surface density at monolayer
coverage and bis the Langmuir adsorption constant (L mmol/C01)
related to the adsorption energy. Plots of Ce/qeversus Ceallow cal-
culation of q0and b. The Freundlich isotherm describes the equilib-
rium on heterogeneous surfaces and does not assume monolayer012340.00.10.20.30.40.5
MT-SiDav-NH2
SiDav-NH2Cu adsorbed, qe (mmol g-1 material)
Ce (Cu mmol/L)SiDav
Fig. 5. Adsorption isotherms at room temperature of Cu2+on MT–SiDav–NH 2,
SiDav–NH 2and SiDav for an adsorbent content of 5.0 g L/C01, at pH 6 and an
equilibrium time of 24 h.0123456789 1 00.000.010.020.030.040.050.060.070.08
Cu
Pb
Cd
ZnMetal adsorbed, qe (mmol g-1 material)
pH
Fig. 3. Influence of pH on metal ions removal using MT–SiDav–NH 2as adsorbent
(5 g L/C01) in batch experiments performed with separated aqueous metal solutions
with initial concentrations of 0.5 mM for Cu2+,C d2+,Z n2+,P b2+for each metal ions.
0 20 40 60 80 100 120 140 1600.000.020.040.060.080.100.120.140.160.180.20Metal adsorbed, qe (mmol g-1 material)
NaCl concentration (mM) Cu
Pb
Cd
Zn
Fig. 4. Influence of NaCl concentration on metal ions removal at pH 6 using MT–
SiDav–NH 2as adsorbent (5 g L/C01) in batch experiments performed with separated
aqueous metal solutions with initial concentrations of 1 mM for Cu2+,C d2+,Z n2+,
Pb2+for each metal ions.M. Mureseanu et al. / Microporous and Mesoporous Materials 146 (2011) 141–150 147

capacity. Such a model has been considered as it has been suggested
that the hybrid adsorbent features a more heterogeneous micro-
structure than an unmodified one due to the coating by organosilica
[22]. Moreover, the adsorption mechanisms for Cu2+may involve
not only the metal ion chelation by the thiol complexing functions
of cysteine groups of the metallothionein, but as well a surface com-
plexation by amino or Si-OH sites remainining uncovered. The sorp-
tion data representation according to the Freundlich equation takes
the form:
qe¼KFC1=nð3Ț
A linear form of the Freundlich equation is
lnqe¼lnKFț1=nlnCe ð4Ț
where KFis the Freundlich constant (mmol g/C01), which indicates the
sorption capacity and represents the strength of the absorbtive
bond and nis the heterogeneity factor, which represents the bond
distribution. According to Eq. (4)the plot of the ln qeversus ln Ce
gives a straight line and KFand nvalues can be calculated from
the intercept and slope of this straight line.
The Langmuir and Freundlich parameters for the sorption of
Cu2+onto hybrid biocomplexant silica materials MT–SiDav–NH 2
and MT–SiDav–Gly comparatively with parent mesoporous silicaSiDav are listed in Table 3 . It appears that the Langmuir isotherm
model fits better the experimental data than the Freundlich model
when the linearity coefficient R
2values are compared, particularly
for the MT–SiDav–NH 2and MT–SiDav–Gly materials with
R2> 0.997 for the Langmuir model. It can be then concluded that
the metal adsorption on the hybrid biocomplexant silica material
occurs most probably on a homogeneous surface and that Cu2+
adsorption on metallothionein immobilized onto mesoporous sil-ica proceeds via a monolayer formation through a complexation
mechanism.
3.3.4. Ions adsorption kinetics over MT–SiDav–NH
2
This study has been performed using separated solutions of
each cations with a concentration corresponding to a potential
amount for adsorption of 0.20 mmol ion g/C01silica, therefore the
half of the expected maximum capacity of MT–SiDav–NH 2for cop-
per ions (0.05 mmol of MT g/C01silica with 8 Cu atoms per MT).
Fig. 6 shows the specific adsorption capacity of MT–SiDav–NH 2
as a function of time. The hybrid biocomplexant silica material
reached the saturation capacity after one hour. The highest affinity
was for Cu2+and followed the sequence: Cu2+>P b2+>C d2+>Z n2+,
with cations uptake values of 0.18, 0.16, 0.06, 0.02 mmol g/C01silica,
respectively.
Adsorption rates have been analyzed by using two common
semi-empirical kinetic models, which are based on adsorptionequilibrium capacity: the pseudo-first-order and pseudo-second-
order models, proposed by Lagergreen [23] and Ho and McKay
[24], respectively. The pseudo-first-order equation relates the
adsorption rate to the amount of metal adsorbed at time tas
dqt
dt¼k1ðqe/C0qtȚð 5Ț
where qeand qtare, respectively, the adsorbed amounts of metal at
equilibrium and time t, expressed as mmol g/C01;k1is the pseudo-
first-order kinetic constant, expressed as min/C01. By integration
and rearrangement the linear form is obtained:
lnðqe/C0qtȚ¼lnqe/C0k1t ð6Ț
The pseudo-second-order equation may be written in the form:
dqt
dt¼k2ðqe/C0qtȚ2ð7Ț
where k2is the pseudo-second-order kinetic constant, expressed as
g mmol/C01min/C01. By integration of the differential equation, the lin-
ear form is obtained:
t
qt¼1
k2q2
ețt
qeð8Ț
h¼k2q2
e ð9Ț
where his the initial adsorption rate and is expressed as
mmol g/C01min/C01.
The linear forms (Eqs. (6) and (9)) are commonly used to check
the validity of these models and to obtain the model parameters
when the corresponding linear plot is adequate. From the applied
models, the pseudo-second-order model showed the best correla-
tion with the experimental data ( Table 4 ).
The parameter that influence the efficiency of the biocomplex-
ant material at the short contact times associated with dynamic
adsorption systems is the initial sorption rate, h. Furthermore, this
parameter becomes particularly important for selective sorption
since the retention of the metal ion can be modulated by changing
the contact time with the solid. The hvalues for the four ions gave
the following sequence: hCu2+>hPb2+>hCd2+>hZn2+.
3.3.5. Competitive adsorption over MT–SiDav–NH 2in batch mode
The competitive adsorption experiments by the MT–SiDav–NH 2
hybrid adsorbent were first carried out in a batch reactor from amultimetallic solution containing all the four investigated metal
ions (Cu
2+,P b2+,C d2+,Z n2+). This study has been performed using
cation concentrations corresponding to a potential amount for
adsorption of 0.40 mmol ion g/C01silica, therefore the total capacity
for copper ions in MT–SiDav–NH 2(0.05 mmol of MT g/C01silica with
8 Cu atoms per MT). The selectivity coefficients, ki, of the Cu2+in
the presence of the other metal (Me2+) ions were calculated
according to the following equation:
ki¼qeCu2ț
qeMe2țð10Ț
The results obtained in batch are presented in the first two columns
ofTable 5 .
0 1 02 03 04 05 06 07 00.000.020.040.060.080.100.120.140.160.180.20Metal adsorbed, qe (mmol g-1 material)
Time (min) Cu
Pb
Cd
Zn
Fig. 6. Adsorption capacity at pH 6 and room temperature of MT–SiDav–NH 2
material (5 g L/C01) for separated metal ion solutions (1 mM for each metal ion) as a
function of time in batch reactor.Table 4Kinetic parameters for the different metals adsorbed on the hybrid biocomplexant
MT–SiDav–NH
2materials calculated using pseudo-second-order model.
Metal ion solutions h(mmol g/C01min/C01) k2(g mmol/C01min/C01) R2
Cu2+0.0160 0.245 0.9926
Pb2+0.0076 0.137 0.9608
Cd2+0.0027 0.256 0.9345
Zn2+0.0021 0.356 0.9994148 M. Mureseanu et al. / Microporous and Mesoporous Materials 146 (2011) 141–150

The Cu2+adsorption capacity (0.34 mmol g/C01) of MT–SiDav–
NH 2from a multimetallic solution is very high and reaches 85%
of the total adsorption capacity of the bio-adsorbent with a very
low adsorption for the other metal ions (>0.01 mmol g/C01). The
selectivity coefficients of Cu2+in the presence of other divalent me-
tal ions, ki, are in the following order: kiPb2+/C29kiZn2+>kiCd2+
showing the very high selectivity of MT–SiDav–NH 2toward the
Cu2+ions. The adsorption behavior of Pb ions is noteworthy. In
the case of the adsorption from a monometallic solution, the
amount of Pb2+adsorbed (0.14 mmol g/C01) was close to the one of
Cu2+ions (0.18 mmol g/C01) (see Fig. 6 ). By contrast, Pb2+ions are
barely adsorbed (0.001 mmol g/C01) in MT–SiDav–NH 2from a com-
petitive multimetallic solution. This behavior is most probably
due to the greater ionic radius of Pb2+with respect to the other
metallic ions in the solution, which may limit its diffusion towards
the active adsorption site.
3.3.6. Adsorption over MT–SiDav–NH 2in flow processes
Continuous metal ions removal was investigated with MT–Si-
Dav–NH 2introduced in a packed-bed column passing aqueous
metallic solutions at controlled flow rates.
3.3.6.1. Breakthrough curves over MT–SiDav–NH 2in flow pro-
cesses. Continuous Cu2+removal was investigated with MT–Si-
Dav–NH 2in the packed-bed column from an aqueous solution
containing 0.8 mM metal at different flow rates from 1 to
3 mL min/C01(Fig. 7 ). As comparison, in batch mode (with an infinite
contact time), the effective copper adsorption capacity was
0.18 mmol g/C01silica (8.51 mg Cu/g material) for this metal concen-
tration, which will correspond to the adsorption of the copper con-
tained in 84 mL of solution for 0.5 g of bio-adsorbent as used for
the experiment reported in Fig.7 . Effluent Cu2+concentration levels
were plotted as a function of passed volume in the form of break-
through curves ( Fig. 7 ) for the different flow rates. Breaking throughdescribes the increase in effluent concentration of a metal ion due to
the reduced capacity of the adsorbent for the metal ion. The equilib-
rium loading will determine the maximum sharpness of break-
through that is, how efficiently the ion will be separated from
other components. The sharpness of breakthrough curves is indica-
tive as well as the thickness of the adsorption zone indicating that
the bed is loaded more completely. In the case of the adsorption of
Cu2+on MT–SiDav–NH 2in the packed-bed column, the break-
through curve is very sharp, which is the evidence of a complete
loaded bed and a high efficiency of this adsorbent for copper
removal. The breakthrough points stand at 20, 40 and 55 mL of solu-
tion for flow rates of 3, 2 and 1 mL min/C01, respectively, correspond-
ing to adsorption capacities of 0.04, 0.09 and 0.12 mmol g/C01silica,
therefore to 23%, 48% and 66% of the adsorption capacity obtained
in batch mode. Lower flow rate increased the adsorption capacityof the column. After the first breakthrough, the column adsorbed
additional ions on a large volume zone before to reach the maximum
of adsorption obtained when the concentration of the solution is
equal to the initial concentration ( C/C
0= 1). For all flow rates the
maximum dynamic loading capacity of 0.18 mmol g/C01silica
(8.5 mg Cu g/C01bio-adsorbent) was attained after passing 84 mL of
solution ( Fig. 7 ), showing a very high level of copper adsorption cor-
responding to 100% of the batch mode (for the same initial concen-
tration) and to 37% of the maximum achievable loading capacity of
the bio-adsorbent.
3.3.6.2. Competitive adsorption over MT–SiDav–NH 2in flow pro-
cesses. The competitive adsorption experiment of metal ions in
continuous flow was carried out by passing through the column
filled with MT–SiDav–NH 2a multimetallic solution containing
Cu2+,P b2+,C d2+and Zn2+ions with 2 mM concentration of each
ion at a flow rate of 1 mL min/C01. As comparison, in batch mode
(with an infinite contact time), the effective copper adsorption
capacity was 0.34 mmol g/C01silica for this metal concentration ( Ta-
ble 5 ). In flow process for this flow rate, a capacity for copper
adsorption of 0.21 mmol g/C01silica was reached corresponding to
62% of the adsorption capacity of the batch mode and to 52% of total
adsorption capacity of the bio-adsorbents. A very high selectivity
for copper in flow is proved by the low adsorption capacity of the
other metals, below 0.009 mmol g/C01. The most significant differ-
ence between batch and continuous flow experiments was ob-
served for the Pb2+ions (selectivity coefficient of 38 in flow
compared to 340 in batch mode). Most probably, the equilibrium
displacement for the more voluminous lead ions by the other metal
ions with a greater affinity for the adsorbent could not be attained
due to the shorter contact time than that achieved in batch mode.
For the other ions, the values of selectivity coefficients were in
the same range than those obtained in batch mode ( Table 5 ).
3.3.6.3. Column regeneration and copper recovery. When the effluent
concentration passes the breakpoint values, the feed of the column
is discontinued and the column should be regenerated. The column
containing 0.5 g of bio-adsorbent was regenerated by circulatingTable 5
Adsorption selectivity of MT–SiDav–NH 2for Cu2+in a multimetallic ions solution in
batch and in continuous flow processes.
Metal ions Batch mode Continuous flow mode
qbatch
e
Adsorption
capacity(mmol g
silica/C01)ki
Selectivity
coefficient ofCu
2+qcolumb
e
Dynamic
adsorptioncapacity(mmol g
silica/C01)ki
Selectivity
coefficient ofCu
2+
Cu2+0.340 1 0.210 1
Cd2+0.011 31 0.009 24
Zn2+0.008 42 0.003 68
Pb2+0.001 340 0.005 38
0 20 40 60 80 100 1200.00.20.40.60.81.01.2C/C0
V (mL) 3 mL/min
2 mL/min
1 mL/min
Fig. 7. Influence of flow rate on Cu2+adsorption onto MT–SiDav–NH 2. Column
experiment, adsorbent amount: 0.5 g, initial Cu2+concentration: 0.8 mM.Table 6Determination of Cu
2+in water samples over MT–SiDav–NH 2bio-adsorbent.
Sample Spiked ( lgL/C01) Found ( lgL/C01)aRecovery (%)
Deionized water 0 n.d. –
10 10.11 ± 0.84 101
Tap water 0 6.54 ± 0.60 –
10 16.45 ± 0.24 99
Mineral water 0 0.93 ± 0.65 –
10 10.88 ± 1.08 95
aAverage of three determination ± standard deviation, n.d. = not detectable.M. Mureseanu et al. / Microporous and Mesoporous Materials 146 (2011) 141–150 149

6 mL of 1 M HCl solution and 98% of the adsorbed copper was
recovered.
3.3.6.4. Preconcentration effect of MT–SiDav–NH 2bio-adsorbent and
quantification of Cu2+traces. The very high selectivity of Cu–MT and
the easy recovery of the adsorbed copper by using small amount of
acidic solution prompted us to explore the possibility of using the
bio-adsorbent for traces analysis. Estimation of the preconcentra-
tion factor of the column was made by using a diluted solution
of copper with a concentration of 0.08 mM (5 ppm) Cu2+which
was passed through the column containing 0.5 g MT–SiDav–NH 2
adsorbent. The maximum dynamic loading capacity was foundequal to 0.13 mmol g
/C01material (8.26 mg Cu g/C01material, equiva-
lent to that found in the breakthrough experiments) and was at-
tained after passing 850 mL solution with a quantitative copperrecovery (98.66%). Based on these data a preconcentration factor
of 142 was obtained. Such a material is therefore suitable for metal
ion detection and quantification from diluted solutions.
To demonstrate the promise of the MT–SiDav–NH
2material for
traces analysis, samples of 300 mL of dionized water, tap water and
mineral water were used for Cu2+preconcentration. After elution
with 6 mL of 1 M HCl the recovered copper was determined by
FAAS. The data obtained for the native and spiked (with added cop-
per) water samples are presented in Table 6 . The results of tripli-
cate analysis of each water sample showed that the copper
recovery by MT–SiDav–NH 2bio-adsorbent was quantitative (case
of spiked samples) and that the material can indeed be useful for
traces analysis (case of plain tap and mineral waters).
4. Conclusion
The anchoring of copper thionein grown from baker yeast to the
surface of a mesoporous silica support leads to a highly selective
bio-adsorbent for the removal of Cu2+from a multimetallic water
solution and its recovery as a pure component. The covalent immo-
bilization of the biochelator to the silica surface can be readily
achieved by using efficient spacers as glycidoxypropyl or amino-
propyl and allows multiple recycling and operation under continu-
ous flow without loss of performance. Related recent works
describing the selective removal of Cd2+and Zn2+by bio-adsorbents
consisting of immobilized metallothionein from S. pombe [10] and
of Fe3+by bio-adsorbents consisting of immobilized pyoverdin
from P. fluorescens [2]on mesoporous silica matrices from multi-
metallic contaminated waters show that this type of hybrid bioma-
terials could be implemented in biotechnologies processes, not
only for the decontamination of polluted waters but for the
recovery of pure metals as well. Processes in flow can be easily
developed for all of these mentioned bio-adsorbents. Furthermore,
thanks to the high preconcentration factor achieved, analytical de-
vices for the measurement of the copper content of spiked tapwater and mineral waters have been already successfully evalu-
ated. Other metallothioneins, selective for other metals (Se,
Ag, …) could be prepared as well and complete the spectrum of
bio-adsorbents prone to be immobilize. The high selectivity of the
biocomplexants can authorize the design of processes combining
several connected columns specific for a single metal for its uptake
and recovery.
Acknowledgment
I.G. wishes to acknowledge EURODOC ‘‘Doctoral Scholarships
for research performance at European level’’ project for support.
References
[1] D.L. Gutnick, H. Bach, Appl. Microbiol. Biotechnol. 54 (2000) 451–460.
[2] G. Renard, M. Mureseanu, A. Galarneau, D.A. Lerner, D. Brunel, New J. Chem. 29
(2005) 912–918.
[3] A. Sigel, H. Sigel, R.K.O. Sigel (Eds.), Metallothioneins and Related Chelators.
Metal Ions in Life Sciences, 5, RSC Publishing, Cambridge, 2009, ISBN 978-1-
84755-899-2.
[4] (a) A. Krezel, W. Maret, J. Am. Chem. Soc. 129 (2007) 10911;
(b) W.G. He, D.Y. Chu, J.Y. Yang, D.F. Yao, M.C. Shao, Chin. Chem. Lett. 10 (1999)
87–90.
[5] J.S. Turner, N.J. Robinson, J. Ind. Microbiol. 14 (1995) 119–125.
[6] A. Dancis, D.S. Haile, D.S. Yuan, R.D. Klausner, J. Biol. Chem. 269 (1994) 25660.
[7] S.S. Narula, D.R. Winge, I.M. Armitage, Biochemistry 32 (1993) 6773–6787.
[8] K.B. Nielson, D.R. Winge, J. Biol. Chem. 259 (1984) 4941–4946.[9] J.H. Kagi, Y. Kojima, EXS 52 (1987) 25–61.
[10] S. Bahrami, A. Bassi, E. Yanful, J. Sol-Gel Sci. Technol. 42 (2007) 119–126.
[11] T. Martin, A. Galarneau, F. Di Renzo, F. Fajula, D. Plee, Angew. Chem. Int. Ed. 41
(2002) 2590–2592.
[12] B. Lefevre, A. Galarneau, J. Iapichella, C. Pettito, F. Di Renzo, F. Fajula, Z.
Bayram-Hahn, R. Skudas, K. Unger, Chem. Mater. 17 (2005) 601–607.
[13] A. Galarneau, J. Iapichella, K. Bonhomme, F. Di Renzo, P. Kooyman, O. Terazaki,
F. Fajula, Adv. Funct. Mater. 16 (2006) 1657–1667.
[14] (a) O. Zaborsky, Immobilized Enzymes, CRC Press, Cleveland, 1973;
(b) W. Hartmeier, Immobilized Biocatalysis – An Introduction, Springer,
Berlin, 1988;
(c) M. Hartmann, D. Jung, J. Mater. Chem. 20 (2010) 844–857.
[15] (a) E. Casero, L. Vázquez, J. Martin-Benito, M.A. Morcillo, E. Lorenzo, F.
Pariente, Langmuir 18 (2002) 5909–5920;
(b) P. Zhang, C.J. Allen, Biol. Trace Elem. Res. 50 (1995) 135–148;
(c) P. Corbisier, D. Vanderlelie, B. Borremans, A. Provoost, V. Delorenzo, N.L.
Brown, J.R. Lyoyd, Anal. Chim. Acta 387 (1999) 235–244;(d) A. Maquieira, H.A. Elmahadi, R. Puchades, Anal. Chem. 66 (1994) 1462–
1467.
[16] G. Das, K.A. Prabbu, Enzyme Microb. Technol. 12 (1990) 625–630.
[17] U. Weser, H.-J. Hartmann, Methods Enzymol. 205 (1991) 274–278.
[18] P.E. Hunziker, Methods Enzymol. 205 (1991) 451–452.
[19] T. Martin, A. Galarneau, D. Brunel, V. Izard, H. Hulea, A.C. Blanc, S. Abramson, F.
Di Renzo, F. Fajula, Stud. Surf. Sci. Catal. 135 (2001) 4621–4628.
[20] (a) A. Presta, A.R. Green, A. Zelazowsk, M.J. Stillman, Eur. J. Biochem. 227
(1995) 226–240;
(b) A. Presta, M.J. Stillman, J. Inorg. Biochem. 66 (1997) 231–240.
[21] V. Calderone, B. Dolderer, H.J. Hartmann, H. Echner, C. Luchinat, C. Del Bianco,
S. Mangani, U. Weser, Proc. Natl. Acad. Sci. USA 102 (1) (2005) 51–56.
[22] J.M. Hao, M.-J. Han, X.G. Meng, J. Hazard. Mater. 167 (2009) 1215–1221.
[23] S. Lagergren, Kugliga Svenska Vetenskapsakademius, Handlingar 24 (1898)
1–39.
[24] Y.S. Ho, G. McKay, Adsorpt. Sci. Technol. 16 (1998) 243–255.150 M. Mureseanu et al. / Microporous and Mesoporous Materials 146 (2011) 141–150