Evaluation of sorption and desorption

Talanta 57 (2002) 277–287
Evaluation of sorption and desorption characteristics of
cadmium, lead and zinc on Amberlite IRC-718
iminodiacetate chelating ion exchanger
Mo´nica E. Malla
a
,Mo´nica B. Alvarez
a
, Daniel A. Batistoni
b,c,
*
a
Department of Chemistry and Chemical Engineering, Uni6ersidad Nacional del Sur,
(8000)
Bahı´a Blanca,
Pro6incia de Buenos Aires, Argentina
b
Chemistry Unit, Constituyentes Atomic Center, Comisio´n Nacional de Energı´a Ato´mica, A6enida General Paz
1499
,
(1650)
San Martı´n, Pro6incia de Buenos Aires, Argentina
c
INQUIMAE, School of Sciences, Uni6ersidad de Buenos Aires, Buenos Aires, Argentina
Received 27 July 2001; received in revised form 19 December 2001; accepted 8 January 2002
Abstract
A chelating type ion exchange resin (Amberlite IRC-718), containing iminodiacetate groups as active sites, has been
characterized regarding the sorption and subsequent elution of Cd, Zn and Pb, aiming to metal preconcentration
from solution samples of different origins. The methodology developed is based on off-line operation employing mini
columns made of the sorbent. The eluted metals were determined by flame atomic absorption spectrometry. The effect
of column conditioning, influent pH and flow rate during the sorption step, and the nature of the acid medium

employed for desorption of the retained metals were investigated. Working (breakthrough) and total capacities were
measured under dynamic operating conditions and the results compared with those obtained with Chelex-100, a resin
extensively employed for analytical preconcentration. Structural information on the complexation of metals by the
chelating groups was obtained by Fourier Transform infrared spectrometry. The analytical response of the Amberlite
sorbent was assessed for the analysis of water samples and digestates of marine sediments. © 2002 Elsevier Science
B.V. All rights reserved.
Keywords
:
Ion exchange; Chelating resin; Atomic absorption; Cadmium; Lead; Zinc
www.elsevier.com/locate/talanta
1. Introduction
Solid organic and inorganic ion exchangers
constitute the basis of widely employed chemical
separation procedures, with applications ranging
from analytical and environmental chemistry re-
search to water purification, waste management
and material technologies (such as in the nuclear
and electroplating industries) [1– 4]. The principal
interest of their use in trace analytical chemistry
lies on the design of methods for separation,
preconcentration and, more recently, speciation of
metals and non-metals. In the first two cases,
* Corresponding author. Fax: +54-11-6772-7886.
E-mail address
:
(D.A. Batistoni).
0039-9140/02/$ - see front matter © 2002 Elsevier Science B.V. All rights reserved.
PII: S0039-9140(02)00034-6
M.E. Malla et al.
/

Talanta
57 (2002) 277 – 287
278
procedures are aimed to enhancing the sensitivity
of the method through analyte enrichment and
simultaneously to lessen the influence of sample
macrocomponents able to interfere with the sub-
sequent (off-line or on-line) measurements [5–7].
Among the numerous types of organic ion ex-
changers (anionic, cationic, weak base anionic),
chelating ion exchange resins have ionogenic
groups that can form coordination bonds with
many metals. Such donor groups are principally
constituted by oxygen, nitrogen, sulfur or a com-
bination of these elements in the same functional
group. A very comprehensive and detailed review
covering theoretical aspects and the principal
characteristics and methods of synthesis of this
type of resins has been published by Sahni and
Reedijk [8].
A great deal of materials based on chelating
groups bound to polymeric cross-linked chains
have been synthetized and characterized in con-
nection to their ability to selectively adsorb ele-
ments or groups of elements, particularly
transition and heavy metals [5–11]. The weakly
acidic nature of the chelating groups makes the
desorption of metals by the action of acids rela-
tively straightforward after removing alkaline and
alkaline earth ions, enabling the subsequent re-

generation of the resin with an appropriate
medium.
Applications of polymeric resins containing
iminodiacetate groups as active sites are well doc-
umented in the literature. Among the commer-
cially available products, Chelex-100 (Bio-Rad) is
one of the most well characterized regarding its
applications [3,12–15]. Less information exists,
however, on the behavior and practical analytical
aspects of similar resin types which may poten-
tially be employed for trace metal preconcentra-
tion [16–20].
In this paper we present an evaluation of the
sorption– desorption properties of Amberlite
IRC-718 chelating resin, in connection with its use
for the separation, matrix elimination and precon-
centration of Cd, Pb and Zn present at trace
levels in natural aquatic systems. This resin is
claimed to present a macroporous (macroreticu-
lar) structure that provides high resistance to os-
motic shock and short diffusional paths that may
result in improved operation kinetics. The
methodology is based on off-line operation by
employing a micro column filled with the sorbent
for deposition of the metals, followed by desorp-
tion– elution and subsequent measurement by
flame atomic absorption spectrometry (FAAS). In
establishing the dynamic operation conditions for
the studied material, some measurements were
also performed with Chelex-100 for comparative

purposes.
2. Experimental
2
.
1
. Chemicals
Deionized water (ultra pure type II) was em-
ployed throughout. The acids employed were high
purity HCl, HClO
4
and HNO
3
(Erbatron RSE,
Carlo Erba). Other chemicals were of analytical
reagent grade. Stock solutions (1.000 mg ml
−1
)of
Cd and Zn were prepared from Riedel-de Haen
Fixanal, and appropriately diluted to the required
concentrations. A Pb solution of similar concen-
tration was prepared in 1% (v/v) HNO
3
in water
from analytical reagent grade Pb(NO
3
)
2
. The con-
centration of this stock solution was verified by
titrimetry. Multielement calibrant solutions of the

metals were prepared in 1% (v/v) HNO
3
by daily
dilution of the corresponding stock solutions.
Ammonium acetate buffer solution (1 M) was
prepared by mixing appropriate amounts of am-
monia solution (28%) and glacial acetic acid, fol-
lowed by dilution with water (final pH 9.5).
Required pH adjustments were performed with
ammonia or HCl. Amberlite IRC-718 resin in the
sodium form, 16–50 mesh, (equivalent particle
size 300–900 mm) made by Rhom & Haas,
Philadelphia, PA, was obtained from Biosix SA
(Argentina). Chelex-100 resin (sodium form, 100–
200 mesh, equivalent particle size 75–150 mm) was
from Bio-Rad Laboratories (Richmond, CA).
2
.
2
. Apparatus
FAAS measurements were performed with a
Hitachi Z 6100 flame atomic absorption spec-
trometer equipped with single element hollow-
M.E. Malla et al.
/
Talanta
57 (2002) 277 – 287
279
cathode lamps. The instrument was operated at
maximum sensitivity with an air-acetylene flame.

Analytical wavelengths (nm) and instrumental de-
termination limits (mgml
−1
) were: Cd 228.8/0.01;
Pb 283.3/0.34; Zn 213.9/0.01. No background cor-
rection was required. Data reduction and calibra-
tion calculations were performed by employing
the standard software provided by the manufac-
turer. Measurements of pH were performed with
a conventional pH meter (glass electrode).
Infrared spectra of the Amberlite resin in the
protonated and contacted with the metal forms
were obtained with a Nicolet 510 P Fourier trans-
form infrared spectrometer (KBr pellets). Elemen-
tal analyses of the sodium and protonated forms
of the resin were performed with a Carlo Erba EA
1108 microanalyzer.
2
.
3
. Procedures
An amount equivalent to 2 ml of resin was
packed in a glass mini column (8 cm length, 0.5
cm internal diameter). The circulation of liquid
through the resin bed was driven by gravity.
Conversely, a peristaltic pump was employed to
deliver the influent solution at a fixed rate. Flow
rate was generally 1 ml min
−1
during the condi-

tioning step, 3 ml min
−1
for the deposition–re-
tention step and 2 ml min
−1
for analyte recovery
by elution. A washing step with pure buffer solu-
tion previous to metal elution, at the same flow
rate, was employed for elimination of concomi-
tant elements that could be partially retained by
the column. Column blanks were obtained by the
same procedure with no analyte added.
Recovery studies were performed on river a
seawater samples by spiking with amounts of the
analytes equivalent to those tested with synthetic
samples. Generally, between 200 and 500 ml sam-
ple volumes containing 25 mg of Cd, Pb and Zn,
with the addition of the appropriate volume of
acetate buffer (final pH: 9.0), were employed in
the deposition step. Desorption of metals was
achieved with HNO
3
and the eluted solutions
analyzed by FAAS.
For the analysis of sediment digestates, por-
tions of 0.50090.001 g of dry sediment were
digested in a PTFE container with 12 ml of a
(5:5:2) mixture of concentrated HF/HCl/HNO
3
,

heating to near dryness. Then 2 ml of HClO
4
were
added, repeating the heating until white fumes,
and the residue was redissolved with concentrated
HNO
3
, diluting to 200 ml with water. The result-
ing solution, buffered to pH 8.5–9.0 was passed
through the column. Elution of the analytes was
performed, after a washing step, with 50 ml of 1
M HNO
3
, and the solution analyzed for Cd, Pb
and Zn by FAAS. To correct for fractional recov-
ery from the column the procedure was applied in
parallel to a similarly treated sample spiked with
25 mg of each metal.
For dynamic capacity measurements, appropri-
ately buffered solutions of each analyte at fixed
pH were continuously passed at a given flow rate
through the column. Successive fractions of 10 ml
of the effluent were collected and metal concentra-
tions determined by FAAS. Calculation of the
dynamic working and total capacities were per-
formed as described by Wang and Barnes [5].
For the FT-IR studies, portions of about 2 g of
the resin were contacted in batch during several
hours respectively with 25 ml of 50% HCl (v/v)
solution and with 200 ml of 1.0 g ml

−1
buffered
solutions of the individual analytes. The resulting
samples were air dried under an infrared lamp
before preparing KBr pellets for FT-IR
spectroscopy.
3. Results and discussion
3
.
1
. Purification of the resin
Preliminary tests demonstrated that erratic and
abnormally high blank levels were observed when
the resin was employed as received. In conse-
quence we employed a previous purification step
based on washing with ethanol followed by suc-
cessive portions of 5 M HNO
3
, water, 5 M HCl,
and water, stirring in each case during 15 min.
This procedure left the resin in protonated form.
No significant changes in the position and shape
of IR bands of the resin were observed by FT-IR
spectrometry in the 4000 –500 mm range when
compared with the original non-purified material
(also in protonated form).
M.E. Malla et al.
/
Talanta
57 (2002) 277 – 287

280
3
.
2
. Effect of column conditioning and influent
pH
Because the active iminodiacetate groups of the
Amberlite IRC-718 are weak acids, the degree of
protonation will critically affect the ability of the
resin to retain metal cations. This situation should
be similar with that observed for the Chelex-100
resin. In this resin protonation of the carboxylates
and the donor N atom are reported to be com-
plete at pH 2.21 [21]. A completely deprotonated
form is reached at pH=12.30. To evaluate the
effect of the influent pH on retention of the
analytes, a column 2-ml of IRC-718 resin was
conditioned by passing 10 ml of ammonium ace-
tate buffer. A volume of analyte solution at a
given pH containing the equivalent to 25 mgof
each element was passed through the column and
the amount of non-retained metal measured in the
percolated solution. The total amount of analyte
deposited was subsequently estimated after des-
orption by employing 50 ml of1MHNO
3
solu-
tion. Analyte recovery data, corresponding
respectively to the descending curve (non retained
metal) and ascending curves (eluted metal), are

graphically depicted in Fig. 1. A consistent trend
is observed, indicating that the retention may be
more favorable at pH higher than around 8.0.
These results differ from those reported for the
Chelex-100 resin, for which near quantitative re-
covery of similar metals is reported at pH as low
Fig. 2. Effect of HNO
3
concentration on desorption of metals.
Circles: Cd; filled squares: Pb; triangles: Zn.
as 5.0 [12 –15]. In consequence, further analyte
enrichment experiments with Amberlite IRC-718
were performed in the 8.5–9.0 pH range. Higher
pH values were avoided to prevent the precipita-
tion of metal hydroxides, particularly in the case
of Pb. However, as described in the corresponding
section, because resin capacity measurements re-
quired a relatively high concentration of this ele-
ment in the influent solution, capacity
measurements for Pb were carried out at pH 7.5.
3
.
3
. Acid elution of metals
Desorption of electrostatically bound metals is
expected to be achieved by proton exchange from
acidic solutions. After deposition of the metals
following the procedure described in the prece-
dent section, desorption was tested by employing
25 ml total volumes of HNO

3
and HCl solutions
of increasing molarity. Results are depicted in
Figs. 2 and 3 as % recovery vs. acid molar con-
centration. In the case of HNO
3
similar recoveries
were obtained with acid concentrations between 1
and 2 M. However, recoveries are lower than
100%, particularly for Cd and Pb, suggesting that
the employed volume may not be suited for com-
plete elution of the analytes. Elution of Pb seems
to be weakly influenced by the molarity of HCl,
but the maximum recovery is only about 70%.
Furthermore, in the case of Zn and Cd the mea-
Fig. 1. Effect of influent solution pH on deposition of metals.
Circles: Cd; filled squares: Pb; triangles: Zn.
M.E. Malla et al.
/
Talanta
57 (2002) 277 – 287
281
sured recovery decreases with increasing acid mo-
larity. Formation of stable complexes of the ana-
lytes in the presence of chloride ions could explain
these results. At the very low pH values corre-
sponding to higher chloride concentrations, the
coordinating resin groups become fully proto-
nated and the resin behaves as a weak anion
exchanger regarding the negatively charged com-

plexes (i.e. MCl
3
−
, MCl
4
−
,… with M=metal).
Consequently these species will tend to be
strongly bound, lowering the efficiency of the HCl
eluent for metal desorption. The observed behav-
ior of the sorbent concurs qualitatively with that
reported for chloride metal complexes in an anion
exchange resin. The tabulated log D
max
values [22]
for Cd, Pb and Zn are respectively 3.5 (2 M HCl),
1.5 (1 M HCl) and 3.2 (2 M HCl), corresponding
to the trend observed in the elution curves: Pb
tends to be the most easily lost, while Cd seems to
be the strongly retained. Additionally, a mecha-
nism of this sort has also been invoked by
Hashemi and Olin [23] for a FIA-ICP-AES system
based on preconcentration to explain the low
exchange rate of Cd retained in Chelex-100 and in
an iminodiacetate based sorbent (Novarose
®
)
when relatively high HCl concentrations are em-
ployed for desorption. Similarly, Knezˇevic´ et al.
[24] attributed the non-quantitative recovery of

Pb, Zn and Cr from Chelex-100 to the formation
of neutral and negatively charged complexes of
the metals with chloride and acetate ions.
Fig. 4. Elution curves for different concentrations of HNO
3
in
the eluent. Squares: 0.1 M; circles: 1 M; triangles: 2 M. (a) Cd;
(b) Pb; (c) Zn.
Fig. 3. Effect of HCl concentration on desorption of metals.
Circles: Cd; filled squares: Pb; triangles: Zn.
Preconcentration factors for a given volume of
the sample solution passed through the column
depend upon the original sample volume and the
volume of acid solution required to quantitatively
elute the metal sorpted onto the resin. We ob-
tained elution curves for the studied metals em-
ploying different HNO
3
eluent concentrations by
measuring the amount of analytes in successive 10
ml fractions of the percolated solution collected
after previous deposition of 25 mg of each metal.
Results are shown in Fig. 4(a)–(c). It is observed
that with 1 M HNO
3
, the curves leveled off at an
eluted volume of 50 ml. Assuming a maximum
M.E. Malla et al.
/
Talanta

57 (2002) 277 – 287
282
volume of 500 ml of sample originally passed
through the column, an enrichment factor of ten
times could be achievable. However, were the
recoveries not quantitative, this could result in a
degradation of the enrichment factor. Conse-
quently recovery factors should be taken into
consideration for more accurate estimation of the
analyte concentrations.
3
.
4
. Dynamic capacity measurements
One parameter that describes the operational
characteristics of a ion exchanger is the capacity,
resulting from the effective number of functional
active groups per unit of mass of the material.
The theoretical value depends upon the nature of
the material and the form of the resin. When the
column operation mode is employed, the opera-
tional capacity is usually lower than the available
capacity, and depends on several experimental
factors such as flow rate, temperature, particle
size and concentration of the feeding solution.
Besides, the ‘breakthrough’ of solution from the
column defines a working capacity, which is lower
than the total capacity [5]. The defined working
capacity corresponds to the maximum amount of
analyte that is retained with minimum leakage of

the element from the influent solution. The vol-
ume of solution percolated from the breakthrough
point to the point of leveling of the loading curve
for a given solution flow rate also depends upon
the kinetics of exchange.
We obtained saturation curves by circulating
analyte containing solutions at a predetermined
pH and collecting successive volumes of 10 ml of
effluent. The concentration (C
i
) of metal in each
fraction was determined by FAAS, and the ratio
of each concentration to the concentration of the
influent feeding solution (C
0
) was plotted vs. the
effluent volume.
The behavior of Amberlite IRC-718 resin is
depicted in Figs. 5 and 6. Analogous plots, ob-
tained with a column of similar dimensions con-
taining Chelex-100 are presented, for comparative
purposes, in Fig. 7. The shapes of the dynamic
capacity curves were found to depend upon sev-
eral experimental parameters. Fig. 5 graphically
depicts the variation observed in the case of Zn
Fig. 5. Cd and Zn dynamic capacity curves for Amberlite
IRC-718 at different effluent flow rates. Metal initial concen-
tration (C
0
): 600 mgml

−1
; pH: 9.0. Circles: Cd (1 ml min
−1
);
squares: Zn (2 ml min
−1
); filled squares: Zn (0.8 ml min
−1
).
for the same concentration of element in the
influent at two different elution ratios (0.8 and 2
ml min
−1
). Although the leveling of the curves at
Fig. 6. Pb dynamic capacity curves for Amberlite IRC-718 for
two different influent concentrations and flow rates, at pH 7.5.
Circles: 2 ml min
−1
, C
0
: 250 mgml
−1
; filled squares: 3
ml min
−1
, C
0
: 125 mgml
−1
.

M.E. Malla et al.
/
Talanta
57 (2002) 277 – 287
283
Fig. 7. Dynamic capacity curves for Chelex-100. Effluent flow
rate: 2 ml min
−1
, pH: 5.6. Filled circles: Pb (375 mgml
−1
);
squares: Cd (600 mgml
−1
); filled squares: Cd (350 mgml
−1
);
triangles: Zn (400 mgml
−1
).
(7.5), indicating that weak active sites may be
involved in the deposition of Pb at higher flow
rates and lower influent pH.
Results are qualitatively similar for Chelex-100,
as presented in Fig. 7. In the case of the plotted
curves the operating pH was 5.5 and flow rates
were maintained at a constant value of 2-ml min
−1
for variable concentrations of the metals.
Calculated capacities and distribution coeffi-
cients (K

D
) for both sorbents are presented in
Table 1. Assuming that equilibrium is attained
when the effluent volume reaches the curve
plateau, K
D
values were calculated as:
K
D=(mmol of element/g of resin)
/(mmol of element/ml of solution)
The approximate location of the initial points of
the plateaus at C/C
0
=1 were estimated, when
necessary, by extrapolation of the ascending region
of the capacity curves.
In general, total dynamic capacities are similar
or slightly higher for Chelex-100. This could be
attributed, in part, to the smaller particle size of
that resin. As already mentioned the Zn capacity
for Amberlite IRC-718 is particularly dependent
on the solution flow rate, pointing to significant
kinetics effects. Such effects are not absent in this
type of resins. The effectiveness of flow rates as low
as 0.2 ml min
−1
for the retention of metals by
Chelex-100 resin has been reported by Paulson
[26].
The differences in capacities observed among the

metals retained by the sorbent materials tested
C/C
0
=1 is not reached, the working capacity
results noticeably higher at the lower flow rate,
suggesting that strong retention sites in the resin
are more favorably involved in deposition when
the interaction sorbent–solution is longer [25]. An
acceptable breakthrough point is observed for Cd
at a flow rate of 1 ml min
−1
. The slope of the
ascending portion of the curve also suggests a
higher exchange rate. Similar trends in the shape of
the loading curves were obtained for Pb with
different combinations of metal concentration in
the influent and flow rates (Fig. 6). The break-
through points are not well defined at the pH tested
Table 1
Estimated dynamic capacities
Sorbent Working capacity (mmol g
−1
) Total capacity (mmol g
−1
) K
D
(ml g
−1
)Element
Cd

a
1981.06Amberlite IRC-718 0.65
0.096 80Pb
b
0.01
– 0.048 80Pb
c
Zn 0.10 1.04 113
Zn
d
0.71 1.77 193
1.000.67 321CdChelex-100
61Pb 0.081 0.11
Zn 2081.270.99
Influent flow rate: 2 ml min
−1
(otherwise indicated);
a
1mlmin
−1
;
b
2mlmin
−1
, pH: 7.5;
c
3mlmin
−1
, pH: 7.5;
d

0.8 ml min
−1
.
M.E. Malla et al.
/
Talanta
57 (2002) 277 – 287
284
may evidence a negative steric effect on coordina-
tion with the iminodiacetate groups. The effective
ionic radius of Pb(II) is 119 pm, compared to 95
pm for Cd(II) and 74 pm for Zn(II) [27]. Stability
of the chelate is expected to be less favorable for
ions of larger size. Formation of strong chelate
bonds for metals with smaller ionic radii may
explain the values obtained for the working and
total capacities in the case of Chelex-100. The
same qualitative correlation regarding ionic size
was found for Cd and Pb in the IRC-718 resin,
but the measured working capacity for Zn is lower
than anticipated from the above considerations.
3
.
5
. FT-IR absorption spectra
In order to further characterize the active sites
responsible of the binding of metals in the Amber-
lite IRC-718 resin, we obtained infrared spectra
under different conditions of saturation. Sepa-
rated 2 g portions of the sorbent were equilibrated

for 72 h in batch (stirring occasionally) with 200
ml of solutions of the metals of concentration 1.0
mg ml
−1
, buffered with ammonium acetate at pH
9.0 for Cd and Zn, and at pH 7.5 for Pb. Simi-
larly, the resin in the protonated form was ob-
tained by contacting it with a 50% (v/v) solution
of HCl. In all cases the sorbent was separated by
filtration, washed with water and air dried under
IR lamp for several hours before preparing the
potassium bromide pellets for IR spectrometric
analysis.
The bands recorded in the 4000–500 cm
−1
wavenumber range are compared in Fig. 8. The
spectra for the forms protonated and saturated
with Pb are noticeably similar. The absorption
features near 1730, 1220 and 1396 cm
−1
corre-
spond respectively the two first to carbonyl
stretching and the third to OH bending [28],
indicating the presence of protonated carboxylic
groups. A band of significant intensity at about
1100 cm
−1
(tertiary amine) [21] is conspicuously
absent, indicating that the nitrogen atom in the
imino group is still protonated at the working pH

[8], and suggesting a lower involvement of the
group in the chelation of Pb. This may further
explain the relatively low resin capacity observed
for this metal.
Fig. 8. IR absorption bands of Amberlite IRC-718 in the
forms protonated and contacted with metals. (a) Cd; (b) H
+
;
(c) Pb; (d) Zn.
The spectra produced by the resin contacted
with Cd and Zn at a higher pH differ from that
recorded from the protonated form. The absorp-
tion bands due to carboxylic acid are not ob-
served, being in turn substituted by strong bands
at 1600 and 1400 cm
−1
, attributable to the pres-
ence of carboxylate anions. These groups present
two strongly coupled band, arising the more in-
tense one from the asymmetric stretching in the
1550– 1650 cm
−1
region and the weaker one from
the symmetric stretching near 1400 cm
−1
[28]. In
addition, the band at about 1100 cm
−1
is clearly
observed for the Cd and Zn saturated resin,

pointing to the presence of deprotonated nitrogen.
In consequence coordination with the metals
through the nitrogen atom of the imino group is
favored, allowing the resin to behave as a triden-
tate ligand. This behavior, generally accepted for
M.E. Malla et al.
/
Talanta
57 (2002) 277 – 287
285
Table 2
Chemical analysis of Amberlite IRC-718
Resin form C (%) H (%) N (%)
7.61Sodium 2.9334.62
6.28Proton 5.4363.29
for the sodium form, indicating that, apart from
the H attached to the N and the carboxylates of
the imino group, a significant amount of excess H
is present. Although the oxygen content was not
measured, the excess H may be attributable to the
presence of several water molecules associated to
the resin in the sodium form.
3
.
7
. Analyte reco6ery studies
In order to evaluate the response of Amberlite
IRC-718 resin in real analytical situations such as
those in which an enrichment step, prior to the
determination by FAAS is involved, we carried

out recovery studies on tap, stream and sea water
samples. Variable volumes were spiked with
known amounts of the analytes and passed
through the column at optimized conditions of
pH and flow rate. Subsequent elution of the
metals was carried out with 50 ml of1MHNO
3
.
Obtained results, expressed as % recoveries of the
added amounts of analytes, are presented in Table
3. An overall consideration of the figures without
taking into account the different origin of the
samples indicates that near quantitative recovery
(between the limits of experimental error), is ob-
served for about 40% of the data. In addition,
recoveries of 90% or higher were obtained for
about 70% of the samples tested. About 25% of
the measurements gave recoveries lower than
80%, suggesting that a recovery factor should be
this type of chelating resin, may justify the rela-
tively higher capacity observed for Cd and Zn. It
is worth mentioning, however, that the results for
Pb do not allow to rule out the possibility of
formation of weaker 1:2 type metal–ligand associ-
ations with the resin groups, as proposed by
several authors for Chelex-100 [15,20]. Further
experiments will be required to univocally clarify
this situation for the Amberlite IRC-718 resin,
which if confirmed, would be an additional expla-
nation of the lack of effectiveness of the sorbent

for retention of Pb.
3
.
6
. Elemental composition of the resin
The analysis of the elemental composition of
the resin for the content of C, N and H in both
the original sodium form and in the protonated
form (the last prepared as previously described for
FT-IR analysis) was carried out, and the results
presented in Table 2. It was found that the C/N
ratio has a constant value of about 12. However
the C/H and N/H ratios are substantially lower
Table 3
Analyses of spiked water samples
% Recovery (9 SD)
a
Water sample Sample volume (ml) Concentration of metal added (mgml
−1
)
type
Cd Pb Zn
79.29 6.7200 96.49 1.6 104.0 9 3.60.125Tap
101.29 7.1500 99.09 6.20.05 96.29 2.2
91.59 5.783.59 1.063.09 2.4Stream 0.05500
87.09 7.4 97.59 5.0Sea c 1 200 0.125 65.59 4.2
200 0.125Sea c 2 81.29 4.5 63.69 5.1 99.19 8.6
84.09 2.895.09 7.186.59 3.50.062400
500 0.05 91.09 2.8 94.09 5.7 84.09 5.7
a

Average of three determinations.
M.E. Malla et al.
/
Talanta
57 (2002) 277 – 287
286
Table 4
Analyses of sediment digestates
Pb ZnSample Cd
Found ReportedReported Found Reported Found
0.489 0.01MURST-ISS-A1 21.09 2.10.549 0.02 20.89 0.2 51.99 3.2 54.89 0.2
2.99 0.2 79.393.3 88.896.43.19 0.2 1079 4PG-1 1229 3
1.059 0.01 23.29 0.9 25.09 1.8 507 9 15PG-2 4629 341.169 0.03
Values in mgg
−1
(9 SD).
employed in particular cases to reach more accu-
rate analytical estimations.
In addition to the above described experiments,
we assessed the applicability of the Amberlite
resin to the preconcentration of metals in solu-
tions arising from the acid digestion of marine
sediments. These samples are usually complex and
contain relatively large amounts of alkaline and
alkaline earth concomitants, as well as other sili-
cate components. The materials employed in our
study include a Certified Standard Reference Ma-
terial (MURST-ISS-A1, Antarctic bottom sedi-
ment), and two surface sediment samples
prepared in our laboratory. The mineralogical

and chemical composition of the latter regarding
the elements of interest have been reported else-
where [29]. Spiked samples were employed to
account for the partial recovery of the analytes
after the preconcentration step. A comparison of
certified (or reported) and obtained concentration
values is presented in Table 4.
4. Conclusions
The results reported in the present study
demonstrate the applicability of the chelating
resin Amberlite IRC-718 for off-line enrichment
of trace metals from relatively complex water and
sediment samples, prior to the spectrometric de-
termination by FAAS. If extreme enrichment fac-
tors are not required, the sorbent compares
favorably with the widely employed Chelex-100
resin. Apart from a significantly low recovery rate
of Pb in one of the sea water samples that may be
ascribed to analyte losses during operation, the
larger departures from 100% are observed in gen-
eral for Cd. Acceptable recoveries were obtained
for Pb and Zn, but the efficiency of retention for
Zn seems to be affected by the original sample
volume: the recovery decreases with the influent
sample volume. It is worth mentioning that the
metals considered are prone to strong complexa-
tion by organic species frequently present in natu-
ral water systems. Because the complexes are in
many cases more stable that the associations of
the metal with the iminodiacetate groups of the

resin, particularly in the case of Cd, the deposi-
tion may be seriously impaired. Also, saturation
of the active groups with weakly adsorbed alka-
line and alkaline earth metals due to a mass effect
could lessen the retention efficiency of trace
metals. Acceptable agreement between known and
found concentration values was achieved in the
analysis of sediment digestates that involves a
preconcentration step, providing that recovery
factors derived from the analysis of analyte spiked
samples are employed to account for the frac-
tional recovery of metals from the column. The
studied sorbent may show also utility for on-line
concentration of metals prior to their determina-
tion by atomic spectrometric methods.
Acknowledgements
The authors are indebted to Myriam Crespo
(CERZUS, CONICET) for her collaboration in
performing the atomic absorption analyses, to
Mireille Perec (INQUIMAE) for obtaining and
helping in the interpretation of the FT-IR spectra,
and to Marı´a dos Santos Afonso (INQUIMAE)
for performing the resin microanalyses. This work
was carried out as part of CNEA-CAC Projects
M.E. Malla et al.
/
Talanta
57 (2002) 277 – 287
287
95-Q-02-01 and 02– 03. Financial support was

provided by Agencia Nacional de Promocio´nCi-
entı´fica y Tecnolo´gica (Project PICT-06-00000-
0354) and SGCYT, Universidad Nacional del Sur
(Projects 24M052 and 068).
References
[1] B.L. Karger, L.R. Snyder, C. Howath, An Introduction
to Separation Science, Wiley, New York, 1973.
[2] F.G. Helfferich, Ion Exchange, McGraw-Hill, New York,
1962.
[3] T.M. Florence, G.E. Batley, Talanta 23 (1976) 179.
[4] J.B. Brower, R.L. Ryan, M. Pazirandeh, Environ. Sci.
Technol. 31 (1997) 2910.
[5] X. Wang, R.M. Barnes, J. Anal. Atom. Spectrom. 4
(1989) 509.
[6] B. Noresson, P. Hashemi, A
,
. Olin, Talanta 46 (1998)
1051.
[7] B. Prusiz, W. Zyrnicki, Fresenius J. Anal. Chem. 363
(1999) 110.
[8] S.K. Sahni, J. Reedijk, Coord. Chem. Rev. 59 (1984) 1.
[9] H. Kumagai, Y. Inoue, T. Yokoyama, T.M. Suzuki, T.
Suzuki, Anal. Chem. 70 (1998) 4070.
[10] P. Hashemi, B. Noresson, A
,
. Olin, Talanta 49 (1999) 825.
[11] H. Kumagai, T. Yokoyama, T.M. Suzuki, T. Suzuki,
Analyst 124 (1999) 1595.
[12] J.P. Riley, D. Taylor, Anal. Chim. Acta 40 (1968) 479.
[13] R.E. Sturgeon, S.S. Berman, A. Desaulniers, D.S. Russell,

Talanta 27 (1980) 85.
[14] W.W. Van Berkel, A. Overbosch, G. Feenstra, F.J.M.J.
Maessen, J. Anal. Atom. Spectrom. 3 (1988) 249.
[15] M. Pesavento, R. Biesuz, M. Gallorini, A. Profumo,
Anal. Chem. 65 (1993) 2522.
[16] D. Muraiev, A. Gonzalo, M. Valiente, Anal. Chem. 67
(1995) 3028.
[17] A.K. Sengupta, Y. Zhu, D. Hauze, Environ. Sci. Technol.
25 (1991) 481.
[18] D. Voutsa, C. Samara, K. Fityanos, Th. Kountzis, Frese-
nius J. Anal. Chem. 330 (1998) 596.
[19] Ch. Park, K. Cha, Talanta 46 (1998) 1515.
[20] A. Yuchi, T. Sato, Y. Morimoto, H. Mizuno, H. Wada,
Anal. Chem. 69 (1997) 2491.
[21] G. Schmuckler, Talanta 12 (1965) 281.
[22] H.P. Walton, Principles and Methods of Chemical Analy-
sis, second ed., Prentice Hall, New York, 1964, p. 155.
[23] P. Hashemi, A
,
. Olin, Talanta 44 (1997) 1037.
[24] S. Knezˇevic´, R. Milacˇicˇ, M. Veber, Fresenius J. Anal.
Chem. 362 (1998) 162.
[25] M. Howard, H.A. Jurbergs, J.A. Holcombe, J. Anal.
Atom. Spectrom. 14 (1999) 1209.
[26] A. Paulson, Anal. Chem. 58 (1986) 183.
[27] N.N. Greenwood, A. Earnshaw, Chemistry of the Ele-
ments, second ed., Butterworth-Heinemann, Oxford,
1997.
[28] R.M. Sylverstein, G. Clayton Bassler, T.C. Morrill, Spec-
trometric Identification of Organic Compounds, fifth ed.,

Wiley, New York, 1991.
[29] M.B. Alvarez, M.E Malla, D.A. Batistoni, Fresenius J.
Anal. Chem. 369 (2001) 81.