Journal of

Electroanalytical
Chemistry
Journal of Electroanalytical Chemistry 591 (2006) 9398
www.elsevier.com/locate/jelechem

A copper bismuth lm electrode for adsorptive cathodic

stripping analysis of trace nickel using square wave voltammetry
Sophie Legeai *, Stephanie Bois, Olivier Vittori
Laboratoire dElectrochimie Analytique, Universite Claude Bernard Lyon I, 43 bd du 11 Novembre 1918, 69622 Villeurbanne Cedex, France
Received 10 October 2005; received in revised form 17 February 2006; accepted 22 March 2006
Available online 19 May 2006

Abstract
Stripping analysis has been widely recognised as a powerful tool for trace metal analysis. Its remarkable sensitivity is attributed to the
combination of a preconcentration step coupled with pulse measurements that generates an extremely favourable signal-to-background
ratio. Mercury-based electrodes have been traditionally employed for achieving high reproducibility and sensitivity of the stripping technique. However, because of the toxicity of mercury, new alternative electrode materials are highly desired, particularly for on-site monitoring. Recently, the use of bismuth thin lms, deposited in situ on platinum or glassy carbon substrates, has been proposed as a possible
alternative to mercury: bismuth is environmentally-friendly, with a low toxicity and a widespread pharmaceutical use. In this paper, the
preparation of bismuth lm electrodes by electrodeposition on a copper substrate and their adsorptive stripping performances are presented. Bismuth lm electrodes were prepared ex-situ by pulsed potential electrodeposition. The analytical performances of these electrodes for adsorptive cathodic stripping voltammetry of nickel were evaluated in nondeaerated solutions. Well-dened peaks along with
low background current were obtained in ammonia buer by square wave voltammetry, using dimethylglyoxime as complexing agent.
Linear calibration curves were obtained for Ni2+ concentrations ranging from 1 10 8 to 1 10 7 mol L 1 and from 1 10 7 to
1 10 6 mol L 1 with relative standard deviations of 5% (n = 15) at 1 10 7 mol L 1 level. The analytical methodology was successfully
applied to monitor Ni2+ content in industrial electrolytic baths, ground water and tap water. These results open the possibility to use
bismuth lm electrodes as an alternative to mercury-based electrodes for nickel analysis.
 2006 Elsevier B.V. All rights reserved.
Keywords: Adsorptive cathodic stripping voltammetry; Bismuth lm electrodes; Copper substrate; Nickel analysis; Natural waters; Industrial wastes

1. Introduction
Electrochemical stripping analysis has long been recognized as a powerful technique for trace heavy metal detection in environmental samples. Its remarkable sensitivity is
attributed to the combination of an eective preconcentration step with pulse measurement techniques that generates
an extremely favorable signal-to-background ratio. Two
electrode systems, the mercury lm electrode and the hanging mercury drop electrode, have been traditionally
employed for achieving high reproducibility and sensitivity
of the stripping technique [13]. The excellent perfor*

Corresponding author. Tel.: +33 4 72 43 14 13; fax: +33 4 72 44 84 79.

E-mail address: [email protected] (S. Legeai).

0022-0728/$ - see front matter  2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.jelechem.2006.03.054

mances of mercury-based electrodes are due to their unique

ability to preconcentrate target metals during the accumulation step [4]. Using anodic stripping voltammetry (ASV),
the metals are cathodically preconcentrated into the mercury electrode by amalgam formation and then stripped
out of the electrode by applying an anodic potential scan
(e.g. Cd, Pb, and Zn) [5]. In certain cases, where the metals
form intermetallic compounds, react irreversibly or cannot
form an amalgam (e.g. Co and Ni), the adsorptive cathodic
stripping protocol is employed. A complexing agent like
dimethylglyoxime (DMG) is added to the measuring solution, forming a complex with the target metal. This complex is then subsequently accumulated onto the electrode
surface and stripped o voltammetrically or chronopotentiometrically [69].

94

S. Legeai et al. / Journal of Electroanalytical Chemistry 591 (2006) 9398

The growing needs for on-site monitoring of priority

metal pollutants have led to the development of ow stripping systems [1013] and more recently to the introduction
of hand-held metal analysers [14,15]. A more attractive
in situ approach is to immerse the stripping electrode
directly in the natural matrix (e.g. natural waters or industrial wastes). However, despite the excellent performance of
mercury electrodes, future regulations and occupational
health considerations may severely restrict or even ban
the use of mercury as an electrode material because of its
high toxicity. Indeed, mercury compounds that are characterized by a high toxicity, like mercuric ions (Hg2+) or
organic compounds (e.g., dimethyl mercury), can be
formed by immersion of metallic mercury (Hg0) in natural
waters in the presence of dissolved oxygen or by the mean
of bacterial activity. A few years ago, Yosypchuk et al. [16]
described the use of non-toxic electrodes of solid mercury amalgams (Ag, Cu, and Au) for stripping analysis.
However, despite these electrode materials can be considered as non-toxic for laboratory use, it is not the case
for environmental on-site monitoring. Indeed, mechanical
and/or bacterial corrosion can lead to mercury release from
amalgams, as mercuric ions or metallic mercury.
New alternative electrode materials are then highly
desired to develop environmentally friendly stripping
sensors suitable for on-site heavy metals monitoring.
Numerous solid electrode materials, including gold [17],
carbon-based [9] or iridium [18], have been tested, but
their overall performances have never approached those
of mercury. Recently, Wang et al. [1921], Hutton et al.
[22], Krolicka et al. [23] and Baldo et al. [24] introduced
the use of bismuth thin lms, in situ electrodeposited onto
glassy carbon or platinum substrates, for use in stripping
analysis of trace heavy metals. The behaviour of the bismuth lm electrode (BiFE) was shown to compare
favourably to that of mercury electrodes, with its attractive properties including high sensitivity, well-dened
stripping signals, good resolution of neighbouring peaks
(e.g. Cd, Pb, and Zn), large cathodic potential range,
and insensitivity to dissolved oxygen in contrast to mercury electrodes. This latter characteristic is an essential
property for on-site monitoring. In addition, bismuth is
a more environmentally friendly material with a low toxicity and is widely used in medicine and cosmetics. However, in situ procedure for BiFE preparation, consisting of
Bi3+ addition to the sample and simultaneous deposition
of the bismuth lm and target metals on the substrate,
is not suitable for on-site monitoring. In a previous study
[25], we demonstrated that economic bismuth lm electrodes (BiFE) prepared by electrodeposition onto a copper substrate oer an alternative to mercury electrodes
for anodic stripping analysis of trace heavy metals, such
as cadmium, lead and zinc. In this paper, the application
of BiFE to trace nickel analysis by the adsorptive stripping voltammetry (AdSV) technique is presented. BiFE
were prepared by electrodeposition of a thin bismuth lm
onto a copper substrate that presents several advantages

compared to platinum and glassy carbon substrates, previously tested in literature reports. Indeed, copper is a
metal far from expensive, easy to manufacture and lends
itself to the realization, by lithography, of microsensors
of various sizes and geometries. It would then be easy
to adapt these microsensors by electrochemical bismuth
deposits.
2. Experimental
2.1. Apparatus
Stripping analysis was performed using square wave voltammetry (SWV) with a POL 150 potentiostat linked to a
polarographic stand MDE 150 and monitored by the
Tracemaster 5 software (Radiometer Analytical S.A.,
France). A copper disk electrode (diameter 3.1 mm,
Aldrich, France) served as substrate for bismuth deposition. Bismuth lm electrodes were prepared by electrodeposition of a bismuth lm from a stirred external plating
solution, using a Voltalab 40 potentiostat monitored by
the Voltamaster 5.0 software (Radiometer Analytical
S.A., France). Reference and auxiliary electrodes were an
Ag/AgCl sat. KCl electrode and a platinum wire (Radiometer Analytical S.A., France), respectively.
A LEICA scanning electron microscope (model Stereoscan 440) was used to determine the grain size of bismuth
deposits.
2.2. Reagents
All solutions were prepared with distilled water. Bismuth(III)
nitrate
pentahydrate
(99.999%),
dimethylglyoxime (DMG, >99.0%) and ammonium chloride
(>99.5%) were purchased from Acros Organics (France).
Nickel nitrate hexahydrate (99.999%) was supplied by
Aldrich (France). Nitric acid (65%) was purchased from
Carlo Erba Reagenti RPE (France) and sulfamic acid
(>99%) was supplied by Chimie-Plus (France). Ammonia
solution (25%) was purchased from Laurylab (France),
and gelatin powder was supplied by Prolabo RP (France).
2.3. Measurement procedures
Stripping analysis of nickel was carried out in
0.01 mol L 1 ammonia buer (pH 9), containing
10 3 mol L 1 DMG as the complexing agent. Prior to
use, the BiFE was activated in the measurement solution
using cyclic voltammetry at a scan rate of 50 mV s 1
between 700 and 1300 mV vs Ag/AgCl (three cycles).
A preconcentration potential of 700 mV vs Ag/AgCl
was applied to the working electrode under stirring conditions. After an equilibration period of 15 s, the stripping
voltammogram was recorded in the quiescent solution by
applying a negative-going potential scan using square-wave
voltammetry (SWV) with a frequency of 10 Hz, a potential
step of 5 mV and a pulse amplitude of 50 mV.

S. Legeai et al. / Journal of Electroanalytical Chemistry 591 (2006) 9398

All measurements
temperature.

were

carried

out

at

room

2.4. Preparation of BiFE

The optimisation of BiFE preparation using a copper
substrate was described in a previous report [25]. The procedure established in previous work is described below.
Before use, the copper disk electrode, which served as the
substrate electrode for bismuth lm electrodeposition,
was mechanically polished with wet SiC abrasive disks
(ESCIL, France), starting with P 600 and working down
by steps to P 2400. The electrode was then rinsed thoroughly with distilled water.
The electrodeposition of the bismuth lm onto the copper substrate was then realised by chronoamperometry
using a pulsed potential procedure. The principle is as follows: a potential of 1000 mV vs Ag/AgCl is applied for
0.1 s, then it is suddenly decreased down to 50 mV vs
Ag/AgCl for 0.02 s. This process has to be repeated 1000
times. This pulsed procedure coupled with the use of gelatine, a surface-active compound inhibiting dihydrogen evolution during electrodeposition in acidic media, leads to a
slow and regular growth of small electrodeposited particles.
Indeed, it is well-known that pulse plating improves the
morphology and properties of electrodeposits due to its
positive eects on mass transport, electrode kinetics and
nucleation of growth centres [26,27]. Pulsed electrodeposition yields a ner-grained and more homogeneous appearance of the deposit because a higher instantaneous
current density is possible during deposition by using pulse
plating instead of potentiostatic plating. This results in an
increased nucleation rate leading to the formation of ner
grains [28]. Moreover, applying short anodic pulses
between cathodic pulses allows preferential redissolution
of dendritic excrescences that might have grown during
the cathodic pulses [2932]. Figs. 1a and b show the scan-

95

ning electron microscope (SEM) images of bismuth electrodeposits on the copper substrate, using potentiostatic and
pulsed electrodeposition, respectively. It appears obviously
that deposits realised under potentiostatic conditions are
characterized by a larger grain size (14 lm) than using
pulsed procedure (0.10.5 lm).
The pulse procedure lead to a greater electrode eective
surface and consequently to higher measured currents, as
observed experimentally during trace nickel analysis.
3. Results and discussion
3.1. Evaluation of BiFE suitability for nickel analysis
Preliminary experiments were run in ammonia buer
0.1 mol L 1 electrolyte and Fig. 2a shows typical AdSV
responses of ammonia buer, DMG and Ni(DMG)2 complex. These responses were compared to those obtained
using the hanging mercury drop electrode (HMDE)
(Fig. 2b).
The stripping voltammograms obtained using the
HMDE are in agreement with what was published in previous literature for similar experimental conditions (pH
9, ammonia buer, DMG as complexing agent) [7]. A
well-dened stripping peak is obtained at 950 mV vs
Ag/AgCl during the cathodic scan following the
Ni(DMG)2 complex adsorption onto the HMDE. Similar
voltammograms are obtained using the BiFE, although a
higher background current is observed. This high capacitive current can be explained by the roughness of the bismuth deposit compared to the mercury drop surface,
leading to a thick capacitive layer. Moreover, the stripping
peak corresponding to the reduction of the Ni(DMG)2
complex appears to be larger and 50 mV more cathodic
using the BiFE than the HMDE.
Finally, the preconcentration time is much longer using
the BiFE than using the HMDE, but the necessary sample

Fig. 1. SEM images for bismuth electrodeposition on the copper substrate: (a) potentiostatic electrodeposition; (b) pulsed electrodeposition. Energy of
primary electrons beam: 20 keV.

96

S. Legeai et al. / Journal of Electroanalytical Chemistry 591 (2006) 9398

20

i (A)

15
10
5

y = 1.6343 x
R2 = 0.9989

0
0

4
2+

6
7

[Ni ] x 10 mol.L

Fig. 2b. Square wave polarogram recorded at the HMDE in 0.1 mol L 1
ammonia buer supporting electrolyte. Frequency 10 Hz. Potential step
5 mV. Pulse amplitude 50 mV. Accumulation time 60 s. Accumulation
potential 400 mV vs Ag/AgCl. Deaeration time 300 s. 1: Supporting
electrolyte,
2:
[DMG] = 9.5 10 4 mol L 1,
3:
[DMG] = 9.5
10 4 mol L 1 + [Ni2+] = 1 10 6 mol L 1.

deaeration using the mercury electrode lead to similar analysis duration.

The BiFE appears then to be suitable for nickel stripping analysis. Its sensitivity and measurements reproducibility are presented below.

Fig. 3. Calibration curve of nickel in ammonia buer in the concentration

range 10 710 6 mol L 1. Square wave voltammetry. Frequency: 10 Hz.
Potential step: 5 mV. Pulse amplitude: 50 mV. Preconcentration time:
600 s. Accumulation potential: 700 mV vs Ag/AgCl.

Fig. 3 shows the calibration curve of nickel for concentrations ranging from 10 7 to 10 6 mol L 1 for a preconcentration time of 600 s. As the limit of detection for
600 s accumulation was approximately 1 10 7 mol L 1,
lower concentrations of nickel were detected for a longer
preconcentration time, namely 900 s (Fig. 4).
As can be noticed in Fig. 2, a linear calibration curve
was obtained with good correlation coecient in the 10 7
to 10 6 mol L 1 concentration range. The calibration curve
obtained in the 10 8 to 10 7 mol L 1 concentration range
(Fig. 4) is also linear, although a constant residual current
is observed for all measurements. The BiFE is characterised by a very good reproducibility: repeated sets of eight
replicate measurements were performed at nickel concentrations of 10 6 and 10 7 mol L 1 and provided values
reproducible within 5%.
It can be noticed from results shown in Figs. 3 and 4
that the calibration curve presenting the highest slope is
the one established for the shortest accumulation time.

1.5

1
i (A)

Fig. 2a. Square wave voltammogram recorded at the BiFE in 0.1 mol L 1
ammonia buer supporting electrolyte. Frequency 10 Hz. Potential step
5 mV. Pulse amplitude 50 mV. Accumulation time 300 s. Accumulation
potential
700 mV vs Ag/AgCl. 1: Supporting electrolyte,
2: [DMG] = 9.5 10 4 mol L 1, 3: [DMG] = 9.5 10 4 mol L 1 +
[Ni2+] = 1 10 6 mol L 1.

0.5
y = 0.0985x + 0.5656
R2 = 0.9995

3.2. Trace nickel analysis

In order to evaluate the analytical performances of the
BiFE for trace nickel analysis, calibration curves were
established in ammonia buer for nickel concentration
ranging from 10 8 to 10 6 mol L 1. The DMG concentration was taken equal to 10 3 mol L 1 so that the [DMG]/
[Ni(II)] ratio was always larger than 1000. The results are
presented in Figs. 3 and 4.

10

-1

0
0

4
2+

6
8

[Ni ] x 10 mol.L

10

-1

Fig. 4. Calibration curve of nickel in ammonia buer in the concentration

range 10 8 to 10 7 mol L 1. Square wave voltammetry. Frequency: 10 Hz.
Potential step: 5 mV. Pulse amplitude: 50 mV. Accumulation time: 900 s.
Accumulation potential: 700 mV vs Ag/AgCl.

S. Legeai et al. / Journal of Electroanalytical Chemistry 591 (2006) 9398

Table 1
Results obtained from the analysis of Ni2+ ions in some real samples
Samples

Mineral water
Ground water
Industrial electrolytic bath
a

107 [Ni2+] (mol L 1)a

HMDE

BiFE

1.807 (2%)
2.502 (4%)
1,020,000 (3%)

1.821 (4%)
2.487 (7%)
1,030,000 (5%)

In parenthesis: SD calculated from three replicates.

This is opposite to what is usually observed using accumulation-based techniques. This might be explained by an
increase of the capacitive layer thickness at the electrode
surface with accumulation time, due to DMG molecules
adsorption. Indeed, in absence of nickel, a greater increase
of the measured current with accumulation time is
observed using the BiFE than using the HMDE. The
roughness of the bismuth deposit compared to the HMDE
surface could explain this unusual behaviour of the BiFE.
Further experiments must be undertaken in order to clear
up this point.
3.3. Application to nickel analysis in real samples
The BiFE was then applied to nickel analysis in several real samples: mineral water, characterized by a relatively simple matrix, and more complex samples like
ground water and an industrial electrolytic bath purchased from Mac Dermid S.A (France). This latter contains other metal traces such as lead or bismuth. The
results obtained with the BiFE were compared to polarographic measurements. Electrochemical measurements
were performed as previously described, using the standard addition method. Pulse polarographic measurements were carried out at the hanging mercury drop
electrode (HMDE).
Ammonia buer was used, the concentration of which
was 0.1 mol L 1. The Mac Dermid solution has been
diluted 1000 times because of its high concentration of
nickel. The analytical procedure was the same using both
electrode materials and is as follows. 10 mL of ammonia
buer were introduced in the electrochemical cell along
with 500 lL of a 0.02 mol L 1 DMG solution. The background signal was then recorded before the appropriate
volume of sample is added. The electrolytic solution was
then carefully stirred during 120 s. In the case of polarographic measurements, the solution was deaerated during
300 s by nitrogen bubbling. Preconcentration times were
120 s and 600 s for the HMDE and the BiFE, respectively.
Table 1 presents the results obtained.
As can be noticed in Table 1, similar results were
obtained with the two electrode materials, conrming the
potentiality of the BiFE to become an alternative to mercury-based electrodes for trace nickel analysis. Although
the preconcentration time is much longer using the BiFE,
the necessary sample deaeration using the mercury drop
electrode leads to similar analysis durations.

97

4. Conclusion and future works

This study has shown the analytical performances of the
BiFE for trace nickel analysis. The BiFE is characterised by
a good sensitivity along with a satisfying reproducibility.
Moreover, bismuth presents the advantage of being more
environmentally-friendly than mercury, the toxicity of
which will probably lead this electrode material being progressively banned for on-site monitoring of trace pollutants.
However, further study should be conducted to determine
the suitable electrode conditioning or preparation procedure in order to use this electrode as a sensor for direct monitoring in natural waters. Indeed, the development of new
sensors, economic and renewable, is of great interest to prevent and control water pollution. Modied BiFE could then
be applied to on-site monitoring in natural waters, as its
sensitivity allows the determination of trace nickel at levels
that are lower than the one preconized by the European
Union, i.e. 20 lg L 1 (3.408 10 7 mol L 1).
Finally, further experiments will also be realised to evaluate the potentiality of the BiFE for the determination of
other pollutants of interest, such as cobalt or pesticides
containing NO2 groups, that are usually electrochemically
determined using mercury-based electrodes.
Acknowledgement
The authors thank the Rhone-Alpes region for its nancial support.
References
[1] J. Wang, Stripping Analysis, VCH, Deereld Beach, 1985.
[2] M. Taillefert, M.G.W. Luther III, D.B. Nuzzio, Electroanalysis 12
(6) (2000) 401412.
[3] M.L. Tercier, J. Bue, Electroanalysis 5 (3) (1993) 187200.
[4] J. Wang, Analytical Electrochemistry, second ed., Wiley, VCH, New
York, 2000.
[5] J. Wang, S.B. Hocevar, R.P. Deo, B. Ogorevc, Electrochem.
Commun. 3 (7) (2001) 352356.
[6] H. Eskilsson, C. Haraldsson, D. Jagner, Anal. Chim. Acta 175 (1985)
7988.
[7] B. Pihlar, P. Valenta, H.W. Nurnberg, J. Electroanal. Chem 214 (1
2) (1986) 157177.
[8] A. Bobrowski, A.M. Bond, Electroanalysis 4 (10) (1992) 975979.
[9] E.P. Gil, P. Ostapczuk, Fres. J. Anal. Chem. 346 (1011) (1993) 952
956.
[10] A. Zirino, S.H. Lieberman, C. Clavell, Environ. Sci. Technol. 12 (1)
(1978) 7379.
[11] J. Wang, M. Ariel, Anal. Chim. Acta. 99 (1) (1978) 8998.
[12] D. Jagner, Trends Anal. Chem. 2 (3) (1983) 5356.
[13] M.D. Luque de Castro, A. Izquierdo, Electroanalysis 3 (6) (1991)
457467.
[14] J. Wang, Analyst 119 (5) (1994) 763766.
[15] The Metalyzer 3000, ETG Inc., Baltimore, MD, 1995.
[16] B. Yosypchuk, L. Novotny, Crit. Rev. Anal. Chem. 32 (2002) 141.
[17] J. Wang, B. Tian, Anal. Chem. 65 (11) (1993) 15291532.
[18] M.A. Nolan, S.P. Kounaves, Anal. Chem. 71 (16) (1999) 35673573.
.A. Kirgoz, J. Liu, Electrochem. Commun. 3 (2001) 703
[19] J. Wang, U
706.
.A. Kirgoz, S.B. Hocevar, B. Ogorevc, Anal.
[20] J. Wang, J.M. Lu, U
Chem. 72 (2000) 32183222.

98

S. Legeai et al. / Journal of Electroanalytical Chemistry 591 (2006) 9398

[21] J. Wang, J. Liu, Electrochem. Commun. 2 (2000) 390393.

[22] E.A. Hutton, J.T. Van Elteren, B. Ogorevc, M.R. Smyth, Talanta 63
(2004) 849855.
[23] A. Krolicka, A. Bobrowski, Electrochem. Commun. 6 (2004) 99
104.
[24] M.A. Baldo, S. Daniele, C. Bragato, J. Phys. IV France 107 (2003)
103106.
[25] S. Legeai, K. Soropogui, M. Cretinon, O. Vittori, A. Heeren de
Oliveira, F. Barbier, M.-F. Grenier-Loustalot, Anal. and Bioanal.
Chem. 383 (2005) 839847.

[26] A.M. El-Sherik, U. Erb, J. Page, Surf. Coat. Technol. 88 (1996) 7078.
[27] K. Saber, C.C. Koch, P.S. Fedkiw, Mater. Sci. Eng. A 341 (2003)
174181.
[28] J.C. Puippe, N. Ibl, Plat. Surf. Finish. 67 (1980) 6872.
[29] B. Nguyen, E. Chainet, Actu. Chim. 4 (1990) 167173.
[30] V.P. Khlyntsev, V.M. Kozlov, T.I. Pozdeeva, Soviet Electrochem. 25
(6) (1989) 740743.
[31] F. Creusat, J. Pagetti, Galvano-Organo-Trait. Surf. 590 (1988) 875
878.
[32] W. Kim, R. Weil, Surf. Coat. Technol. 31 (1987) 143150.