Fresenius J Anal Chem (1996) 354 : 718 721 © Springer-Verlag 1996
LECTURE
Frank Hißner · Jürgen Mattusch · Gerhard Werner
Determination of metal ions by ion chromatography
with precolumn electrochemical preconcentration
Received: 17 July 1995 / Revised: 30 August 1995 / Accepted: 3 September 1995
Abstract The determination of heavy metals in concen- especially in samples with high concentrations of alkaline
trations less than 10-6 mol/L by ion chromatography with ions. The suppressed ion chromatography is not applica-
conductivity detection requires a preconcentration step. ble for the analysis of divalent heavy metal ions due to the
Therefore, a special electrochemical equipment and method formation of non-soluble metal hydroxides during the
was developed for the on-line preconcentration of the di- suppressor reaction. Therefore, the single column ion
valent metals Ni, Co, Zn and Cd and their subsequent ion chromatography has to be used. The determination of the
chromatographic determination. The loop of the injection metals in concentrations less than 10-6 mol/L requires a
valve of an ion chromatograph was replaced by an elec- preconcentration step.
trochemical flow-through-cell with a gold working elec- Several on-line preconcentration methods for IC using
trode, a platinum auxiliary electrode and a silver/silver preconcentration columns are described in [1 4]. Cham-
sulphate reference electrode. The preconcentration step baz et al. [1] described a method where the sample loop
consists of the deposition of the reduced metals on the was substituted by a cartridge containing ethylenediamine
electrode surface during a continuous pumping of the triacetate bonded chelating silica for the preconcentration
sample solution through the cell. After switching of the of metal ions. The metals retained were desorbed with ni-
mobile phase through the cell, the analytes are injected af- tric acid; therefore only spectrophotometric detection with
ter their reoxidation directly into the mobile phase. A new post-column derivatization was possible.
preconcentration step is simultaneously possible during First electrochemical preconcentration methods were
the actual chromatographic run. An effective separation of developed for atomic absorption spectroscopy. Lund et al.
the analytes from the matrix is also possible with the pro- [5] described a method using a platinum wire which was
posed system. A maximum of metal ion accumulation introduced into the flame of the spectrometer. For ET-
was obtained after 120 min in the galvanostatic mode on a AAS Volland et al. [6] developed a procedure where the
gold tube electrode. The detection limits for Co(II), graphite tube (as cathode) was inserted into a continuous
Ni(II), Zn(II) and Cd(II) were improved by a factor of 7.7, flow system. A platinum/iridium wire was used as counter
10.4, 11.2, 14.0, respectively, and were in the 0.1 µmol/L electrode in the tube.
concentration range with a RSD of 2 6%. The accumula- Electrochemical preconcentration techniques provide
tion of metal ions was disturbed in the presence of Cr(III). good analyte/matrix separation with low contamination
risks [7]. Furthermore, the electrochemical preconcentra-
tion is compatible with conductivity detection and no
Introduction
post-column derivatization is necessary.
Ion chromatography is an alternative method to atomic
spectroscopy for the determination of heavy metal ions,
Experimental
Chemicals and reagents. Analytical-reagent grade tartaric acid,
citric acid and ethylenediamine were obtained from Merck (Darm-
stadt, Germany).
F. Hißner · G. Werner ( )
The metal ion standard solutions were prepared by dissolution
Institute of Analytical Chemistry, University Leipzig,
of the solid nitrate salts (Merck Darmstadt, Germany).
Linnéstrasse 3, D-04 103 Leipzig, Germany
Samples and eluents were prepared in Milli-Q water.
J. Mattusch
Instrumentation. The chromatographic separations were per-
Institute of Analytical Chemistry,
formed by a HPLC-pump module 2200 (Bischoff, Leonberg, Ger-
UFZ  Centre for Environmental Research, Permoserstrasse 15,
many), a Rheodyne 9125 injection valve, a Nucleosil SA 10 µm
D-04 318 Leipzig, Germany
719
Table 1 Chromatographic data for the separation of metal ions
Metal ion Capacity Eff. plate Eff. plate
factor number N height H
[mm]
Nickel 5.9 591 0.212
Zinc 8.9 2947 0.042
Cobalt 10.8 2871 0.044
Cadmium 17.6 3615 0.035
Nucleosil® SA 10 µm (125 × 4), tartrate/citrate/ethylenediamine
eluent, pH = 4
Table 2 Calibration data for the separation of the metal ions with-
out preconcentration
Fig. 1 Scheme of the electrochemical preconcentration cell
Analyte Calibration graph k Srel [%]a LOD [µg/L]b
column (125 × 4 mm) (Macherey Nagel, Düren, Germany) and a
Nickel y = 0.06 + 0.81x 0.9998 4.4 76.9
conductivity detector (Metrohm, Herisau, Switzerland). A second
Zinc y = 0.12 + 0.66x 0.9997 4.2 111.8
HPLC-pump (Knauer, Berlin, Germany) was used to circulate con-
Cobalt y = 0.17 + 0.59x 0.9973 4.0 80.7
tinuously the analyte solution through the selfmade electrochemi-
Cadmium y = 0.05 + 0.38x 0.9938 10.1 301.2
cal preconcentration cell depicted in Fig. 1. The potentiostat/gal-
vanostat PS 4 (Fl Meinsberg, Meinsberg, Germany) was applied
a
For a concentration of 10 µmol/L with 5 replicates
to polarize the working electrode during the preconcentration
b
LOD = 3 × noise × peak width / b
or stripping mode. The preconcentration of the metal ions was
electrochemically possible in a three-electrode technique with a
Ag/Ag2SO4 reference electrode and a platinum auxiliary electrode.
The self-made flow through cell consists of platinum and gold
cations and the stationary phase which prolong the reten-
tubes (Goodfellow, Cambridge) of 0.8 mm i.d. and 0.7 mm i.d., re-
tion times to about 40 min. Increasing the pH-value to 5.5,
spectively, 1 mm o.d. and a length of 20 mm each. The reference
the two metal ions can be separated in 12 min. Unfortu-
electrode was located between the working and the auxiliary elec-
nately, the separation efficiency of the other metal ions is
trode by using a silver sulphate coated silver wire in the Teflon
tube connecting the ends of the electrodes. The complete cell was low when using these conditions. Therefore, our investi-
protected by epoxy resin to get a suitable pressure resistance of ap-
gations were focused on the determination of metal ions
proximately 3 MPa. The volume of the cell was determined to be
whose chromatographic data are collected in Table 1.
26.7 µL.
To compare the results without and with preconcentra-
Chromatographic conditions. The mobile phase consists of 2 tion, calibrations of the 4 metal ions were carried out us-
mmol/L tartaric acid, 0.5 mmol/L citric acid and 3 mmol/L ethyl-
ing peak areas (Table 2). The limits of detection (LOD)
enediamine. The pH-value of 4 was adjusted with 1 mol/L nitric
are in the concentration level of 70 to 300 µg/L with suf-
acid. Prior to use the prepared eluent was degassed in an ultrasonic
ficiently good correlation coefficients of 0.994 or even
bath for 5 min under vacuum and filtered through a cellulose ac-
better for a linear dynamic range of two decades.
etate filter (Nalgene, Rochester, USA) with 0.45 µm pores. The
eluent is applicable approximately 10 days.
Precolumn preconcentration conditions. A supporting electrolyte
Preconcentration of the metal ions
solution consisting of 0.1 mol/L sodium sulphate and 1 mmol/L ni-
tric acid was preferred for the electrochemical preconcentration of
the metal ions on the gold working electrode. Sodium sulphate was For a further improvement of the sensitivity the injection
used to form the silver/silver sulphate reference electrode and ni-
loop of the injection valve was substituted by a self-made
tric acid to prevent the precipitation of metal hydroxides during
electrochemical flow-through cell to accumulate the metal
electrochemical preconcentration. The sample dissolved in sup-
ions by their electrochemical reduction to the metals and
porting electrolyte solution was continuously pumped during the
accumulation time with a flow rate of 0.8 mL/min. their desorption on gold working electrodes in the precon-
centration mode shown in Fig. 1. During the following in-
jection the metals will again be oxidized to form com-
Results and discussion plexes with components of the mobile phase for an appro-
priate chromatographic separation. The analytes can im-
Separation of nickel, zinc, cobalt, cadmium mediately be stripped by an abrupt change of the polarity
of the working electrode from the reduction to the oxida-
With the strong cation exchange column Nucleosil® SA tion mode. The resulting injection volume is very small.
10 µm as stationary phase and a mobile phase of pH 4 the The complete scheme of the enrichment and reoxidation
divalent cations nickel, zinc, cobalt and cadmium were of the metal ions from the working electrode is shown in
separated in approximately 20 min. Under these condi- Fig. 2. During the chromatographic run of the metal ions
tions it is not possible to separate other interesting ions just stripped from the working electrode the preconcentra-
like copper or lead in the same run. The reason is attrib- tion equipment can again be switched to the starting point
uted to the additional non-ionic interactions between these for an additional cycle.
720
Fig. 2 Working cycle of preconcentration
Table 3 Calibration data with galvanostatic preconcentration
(500 µA, 30 min)
Metal ion Calibration graph k LOD µg/La
Nickel y = 2.15 + 2.71x 0.9844 29.0
Zinc y = 1.03 + 2.81x 0.9975 26.2
Fig. 3 Dependence of peak area on preconcentration time
Cobalt y = 0.79 + 2.05x 0.9985 23.4
Cadmium y = 0.71 + 1.59x 0.9935 71.3
a
LOD = 3 × noise × peak width / b
In the galvanostatic, three-electrode mode based on a
constant current of 500 µA, higher preconcentration effi-
ciencies were obtained for Ni(II), Co(II), Cd(II) and
Zn(II) (see Table 3). All metal ions are reduced to the met-
als in acidic media in a two electron step. In the case of
the hexaquo complexes of cobalt and nickel, the experi-
mentally found reduction potential is about 600 mV more
negative than the standard potentials [8, 9]. On the metal
layers of cobalt and nickel the electrochemically evoluted
Fig. 4 Chromatogram of separation without and after 1 and 2 h of
hydrogen is adsorbed. This fact causes a bad electrode re-
preconcentration
sponse of thicker nickel and cobalt layers due to their brit-
tleness after hydrogen adsorption. The dependence of
peak areas on preconcentration time shows a linear corre-
lation in the case of Zn(II) and Cd(II), but a non-linear be- The accumulation of the metal ions on the working
haviour for Ni(II) and Co(II) due to instable layers by hy- electrode or the chromatographic separation was already
drogen formation (Fig. 3).
interfered with by Mg(II), Mn(II) and Cr(III) with con-
The optimized preconcentration conditions are: gal- centrations similar to those of the analytes. After precon-
vanostatic mode at 500 µA, supporting electrolyte solu- centration of the metal ions, magnesium gave a very in-
tion 0.1 mol/L K2SO4, 1 mmol/L HNO3, gold working
tensive peak, too. This behaviour can only be explained
electrode, flow rate 0.8 mL/min, preconcentration time up
by a precipitation of magnesium with electrochemically
to 120 min.
generated hydroxide directly on the electrode surface. An
If lower detection limits are required, longer precon- electrochemical reduction of Mg(II) to the metal is not
centration times can be used especially for the determina- possible under the given conditions. On the other hand,
tion of zinc and cadmium. Figure 4 illustrates the compar- chromium (III) blocked the reduction of the other analyte
ison of chromatograms obtained with different preconcen- ions by reoxidation of the metals with electrochemically
tration times and the chromatogram using the electro- generated chromium (II).
chemical flow-through cell as a normal injection loop
Using this method for the determination of Zn(II),
with 10-6 mol/L of each metal ion. The peak area is ap- Cd(II), Ni(II) and Co(II) in sodium sulphate (pure), 82.9
proximately 10 times higher with a preconcentration time
µg Zn (0.0058%) and 50.4 µg Co (0.0036%) were found
of 120 min.
after 30 min of preconcentration. To eliminate the inter-
721
the relative standard deviation was determined to be 4%.
All calibration curves were linear in the examined range
from 1 to 15 µmol/l. The enrichment of the metals is in-
fluenced by Cr(III) which decreases the precipitation of
the other heavy metals and by Mn(II) (10-fold) which in-
terferes with the cadmium peak.
This method can be applied to the determination of Ni,
Zn, Co, and Cd in ammonium sulphate fertilizer and sea-
water samples. Shorter preconcentration times and higher
accumulation rates should be possible with a coulometric
device [10]. Another interesting point of view could be
the possibility to reduce the metal ions from different
complexes to achieve a speciation analysis.
References
1. Chambaz D, Edder P, Haerdi W (1991) J Chromatogr 541 :
Fig. 5 Determination of Zn and Co by standard addition
443 452
2. Boomer DW, Powell MJ, Hipfner J (1990) Talanta 37 : 127
134
fering effects of other components the standard addition
3. Tapparo A, Bombi GG (1990) Anal Chim Acta 238 : 279 284
4. Fong WL et al (1991) Spectrosc Lett 24 : 931 941
technique is recommended (Fig. 5).
5. Lund W, Thomassen Y, Doule P (1977) Anal Chim Acta 93 :
53 56
6. Volland G, Tschöpel P, Tölg G (1977) Anal Chim Acta 90 :
Conclusion
15 19
7. Manthey M et al (1988) International Laboratory Nov 1988 :
20 26
The preconcentration procedure described was optimized
8. Proszt J (1967) Polarographie. Akademiai Kiado, Budapest
with respect to time, voltage/current and electrode design.
9. Krjukova TA, Sinjakova SI (1964) Polarographische Analyse.
After 120 min of preconcentration in galvanostatic mode
Deutscher Verlag für Grundstoffindustrie, Leipzig
on a platinum tube electrode, the detection limit was im- 10. Beinrohr E, Németh M, Tschöpel P, Tölg G (1992) Fresenius J
Anal Chem 344 : 93 99
proved 10-fold. The average reproducibility in terms of