- Email: [email protected]
Anodic stripping voltammetry at a new type of disposable bismuth-plated carbon paste mini-electrodes
Analytica Chimica Acta 599 (2007) 249–255
Anodic stripping voltammetry at a new type of disposable bismuth-plated carbon paste mini-electrodes L. Baldrianova a , I. Svancara a , S. Sotiropoulos b,∗ a
Department of Analytical Chemistry, Faculty of Chemical Technology, University of Pardubice, CZ-532 10 Pardubice, Czech Republic b Department of Chemistry, Aristotle University of Thessaloniki, Thessaloniki 54124, Greece Received 22 May 2007; received in revised form 19 July 2007; accepted 25 July 2007 Available online 3 August 2007
Abstract A new type of disposable carbon paste mini-electrodes (CPmEs), with dimensions in the 50–300 m range, have been fabricated by heat-shrinking the end-tip of plastic micropipette tips and filling them with carbon paste. The CPmEs have been characterized by microscopic and electrochemical means and tested as substrates for in situ plated Bi film electrodes (BiF–CPmEs), used in the determination of heavy metals by square wave anodic stripping voltammetry (SWASV). It was found that this new class of CPmEs combines the advantages of carbon paste electrodes (readily renewable surface and high surface area) with those of near-microelectrode behaviour (no stirring or electrolyte excess needed). During SWASV experiments in unstirred Pb(II) and Cd(II) solutions well-shaped stripping peaks were obtained whose height varied linearly with analyte concentration in the wide 1 × 10−8 to 10−6 M range, both in acetate buffer and unbuffered solutions. Under optimal conditions detection limits of 8 × 10−10 and 1.3 × 10−9 M were achieved for Pb(II) and Cd(II), respectively and in a trial application, these metal ions have been determined in a spiked tap water sample using a BiF–CPmE. © 2007 Elsevier B.V. All rights reserved. Keywords: Bismuth film electrodes; Microelectrodes; Carbon paste electrodes; Anodic stripping voltammetry
1. Introduction The low toxicity of Bi and its salts, when compared to Hg, as well as its satisfactory electrochemical performance (relatively wide potential window and formation of fused alloys with many metals) has led to its use as an alternative to Hg for stripping voltammetry [1–8]. Besides glassy carbon and graphite which have been the most common electrode substrates for Bi film plating [3–6], carbon paste electrodes (CPEs) are also a useful class of electrode supports for Bi stripping voltammetry [7–8] since they offer a readily renewable surface, they are easy to modify and are also low cost electrodes. Microelectrodes (i.e. electrodes with at least one of their dimensions smaller than typically 50 m [9]) have a number of properties that make them attractive in stripping analysis [10–12]. First, enhanced mass transfer rates due to non-planar
∗
Corresponding author. Tel.: +30 2310 997742; fax: +30 2310 443922. E-mail addresses: [email protected], [email protected] (S. Sotiropoulos).
0003-2670/$ – see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.aca.2007.07.067
diffusion make possible shorter accumulation periods and/or no need for stirring during that period. Second, small ohmic losses allow experiments without the need for deliberately added electrolyte (often the source of impurities). Hg-coated microelectrodes have long been used in stripping voltammetry and in particular in media of ambient or minimal electrolyte content [13–16]. However, there are only few papers on Bicoated microelectrodes [17–20], mainly of the microdisc or microfibre type. At the same time there are also rather few papers on CPE microelectrodes [21–24]; these are usually of the microdisc type and are fabricated with glass capillaries or Teflon tubing. It should be stressed that, since those electrodes had diameters larger than 200 m (with the single exception of those in [23] that were based on a glass capillary with a 75 m internal diameter), they were not strictly speaking microelectrodes. According to microelectrode terminology [9] their critical dimension should have not exceeded 50 m and hence such electrodes should better be considered as mid-size electrodes or mini-electrodes. In any case, there are no papers either on CPE mini-electrodes based on disposable plastic holders
250
L. Baldrianova et al. / Analytica Chimica Acta 599 (2007) 249–255
with a diameter smaller than 200 m or on Bi-coated such electrodes. The aim of the present work has been to introduce a new type of carbon paste mini-electrodes (CPmEs) as substrates for the preparation of in situ coated Bi film electrodes (BiF–CPmEs), to be used in the determination of metal ions by square wave anodic stripping voltammetry (SWASV). In more detail, its objectives have been (i) to fabricate CPmEs with a diameter in the 50–300 m range, using disposable micropipette tips whose end-tip is heat-shrunk to micrometer dimensions and are filled with carbon paste, (ii) to characterise the new type of CPmEs by means of microscopy and electrochemistry, (iii) to exploit the near-microelectrode behaviour of BiF–CPmEs in the SWASV determination of Pb(II) and Cd(II) from unstirred samples containing a variable acetate buffer concentration (in the 0–0.1 M range) and (iv) to use BiF–CPmEs in the determination of Pb(II) and Cd(II) in a tap water sample. 2. Experimental 2.1. Preparation and characterisation of CPmEs The holder for the CPmEs was based on the lower part (ca. 2 cm long) of disposable micropipette tips (Universal tip 500193, WPI Inc., end-tip diameter ca. 500 m). These were shrunk to an end-tip diameter in the 50–300 m range by brief exposure to a hot air stream (400–450 ◦ C) produced by a heat gun (Model 1850, Beta Utensili S.r.l.). In cases that the manual heating procedure resulted in closed end-tips the latter were cut with a razor blade under a magnifying glass, to produce a micrometer size diameter opening. The heat-shrunk holder was then packed with the carbon paste (for its composition see below) with the help of a common pin that also served as the electrical contact. The surface of the resulting CPmE was renewed between sets of measurements by further pushing the pin into the paste whereby a micrometer-sized thread of carbon paste could be squeezed out of the holder’s opening and wiped with tissue paper (A schematic diagram and a picture of a CPmE are shown in Fig. 1(A), Section 3.1 below). The diameter of the CPmEs could be checked by an optical microscope (INTRACO MICRO, s.r.o., SM 6A, Czech Republic) or by SEM (JEOL JSM-5600LV) after the electrochemical experiments. The geometric area of the CPmEs as well as their near-microelectrode behaviour was confirmed by running slow potential scan voltammograms for the reduction of ferricyanide from 5 mM potassium ferricyanide solutions in 0.1 M KCl. The carbon paste used was prepared by intimate hand-mixing of 0.5 g spectroscopic graphite powder (“RW-B”, RingsdorffWerke,Germany) with approximately 0.3 mL highly viscous silicone oil (“SO”, LUKOIL MV 12000 product; Lucebni zavody Kolin, Czech Republic). 2.2. In situ bismuth plating procedure and square wave anodic stripping voltammetry (SWASV) parameters In situ bismuth films were prepared by spiking the sample with the required Bi(III) quantity from a standard solution of
Fig. 1. Schematic diagram (A) and SEM micrograph (B) of a CPmE.
1 × 10−4 M Bi(III) containing 0.1 M HNO3 to reach always 1 × 10−6 M Bi(III) + 0.001 M HNO3 levels in the sample. The standard solution was prepared by appropriate dilution of a spectroscopic standard solution of 1000 ppm Bi(III) in 0.5 M HNO3 and consequent addition of 1 M HNO3 to reach the desired acid concentration and solution pH. The SWASV conditions were Edeposition = −1.1 V versus SCE; Efinal = + 0.5 V; tdeposition = 60 − 600 s; teq = 15 s; frequency = 25 Hz; potential step = 4 mV; amplitude = 20 mV. Before each measurement Econditioning = +1.8 V was applied on the electrode for 120 s to ensure complete Bi stripping. 2.3. Electrochemical cell and instrumentation A three-electrode cell was used, consisting of the CPmE working electrode, a SCE reference electrode (Radiometer) and a Pt coil auxiliary electrode (BAS Inc.). A laboratory-made aluminium Faraday cage housed the electrochemical cell in order to minimise electrical noise. The potentiostats used were a Model PGSTAT 100 Autolab or a AUTOLAB (Ecochimie, Utrecht, Holand), connected to a personal computer and controlled by GPES, version 4.8 software (Ecochimie, Utrecht, Holand). The pH was measured by a Radiometer Blue Line pH-meter and samples were spiked by a 20–100 L and 100–1000 L micropipetter (Eagle, from WPI Inc.).
L. Baldrianova et al. / Analytica Chimica Acta 599 (2007) 249–255
2.4. Chemicals and sample preparation All chemicals used in the preparation of stock solutions were of analytical reagent grade and obtained from Merck (Damstadt, Germany) and Fluka. Concentrated CH3 COOH p.a. (Fluka) and CH3 COONa p.a. (Fluka) solutions were used for the preparation of 1 M stock acetate buffer. Stock solutions for in situ bismuth plating, were prepared from an AAS (spectroscopic grade) 1000 ppm standard solution of Bi(NO3 )3 ·5H2 O (Merck, Germany), while Pb(NO3 )2 p.a. and Cd(NO3 )2 ·4H2 O p.a. (Merck, Germany) were employed in the preparation of 0.01 M metal ion stock solutions. Where appropriate, the respective solutions were diluted as needed. All the solutions of standards and diluted standards of Pb(II) and Cd(II) were always stabilised by acidifying with 500 L of 65% HNO3 (Fluka) per 100 mL of solution (resulting in 0.07 M HNO3 concentration in the stock solutions). For the calibration of the CPmEs potassium, ferricyanide and KCl (Merck) were used. In the preparation of solutions doubly redistilled water was employed. 3. Results and discussion 3.1. Microscopic and electrochemical characterisation of carbon paste mini-electrodes (CPmEs) Fig. 1(A and B) shows the schematic diagram and a SEM micrograph of a CPmE. In the latter, the rough, undulating surface of the carbon paste material can be seen, protruding out of the plastic tip. This is surrounded by a ring of a thin, transparent plastic film covering the carbon paste material. The dimensions of the mini-electrode corresponding to the protruding carbon material shown in that micrograph are (depending on location) in the 200–300 m range. Mini-electrodes with dimensions in the 50–300 m range were fabricated by this technique (as also confirmed by the electrochemical calibration discussed below) while those of a larger diameter were dismissed. The majority of our experiments were performed on mini-electrodes with a diameter smaller than 100 m. Fig. 2 shows a slow scan voltammogram (at 5 mV s−1 ) of a CPmE in a deaerated 5 mM potassium ferricyanide + 0.1 M
251
KCl solution. A sigmoidal curve with a near-coincidence of the positive- and negative-going scans corresponding to the one-electron reduction of ferricyanide can be seen, indicating a near-steady state response, characteristic of near-microelectrode behaviour. The equation holding for the diffusion-limited current, IL , associated with the n-electron reaction of a species (with a concentration of C and a diffusion coefficient of D), at a microdisc electrode (with a radius of r) is [9] IL = 4nFCDr
(1)
applying (1) in this case for n = 1, C = 5 × 10−3 M, D = 7.63 × 10−6 cm2 s−1 [25] and considering the current at −0.3 V, the CPmE of Fig. 2 can be estimated to give a limiting current equivalent to that of a 72 m diameter disc electrode. Eq. (1) assumes pure non-planar, hemi-spherical (more precisely, hemi-cylindrical) diffusion towards a microdisc electrode and can be considered a first approximation. Strictly speaking though, electrodes with diameters larger than 50 m are not microelectrodes and both non-planar and planar diffusion effects should be taken into account. The complete equation proposed to apply for all small electrodes is due to Aoki et al. [26] IL = 4nFCDr 0.34 exp(−0.66p) + 0.66 − 0.13 −11 ×exp + 0.351p (2) p with p=
nFr 2 υ RTD
1/2 (3)
where υ is the potential scan rate. Using the values of n, C and D given above and υ = 0.005 V s−1 , Eq. (2) was solved for r with the help of Micrososft Excel Solver® and a diameter of 2r = 68 m could be estimated, i.e. within 5% of the value estimated using the simple equation of (1). Both these values are close to that of ca. 80 m determined at the end of experimentation by means of S.E.M. The slight increase in diameter may be due to tip widening caused by the carbon paste extrusion-renewal between sets of experiments in different media. All of the results presented below were obtained at this particular mini-electrode. 3.2. Effect and choice of accumulation time in the SWASV determination of Pb(II) and Cd(II) at CPmEs, from buffered and unbuffered unstirred solutions
Fig. 2. Voltammogram of a CPmE in a 5 mM ferricyanide + 0.1 M KCl deaerated solution, recorded at 5 mV s−1 potential scan rate.
Fig. 3(A) and (B) shows the SWASV picture at CPmEs from a 1 × 10−6 M Bi(III) unstirred solution containing 10−7 M Pb(II) + Cd(II) and (A) no added buffer (pH 2.8, due to 1 mM HNO3 added with the 10−6 M Bi(III) spiked from a 10−4 M Bi(III) + 0.1 M HNO3 acidified stock solution) or (B) 0.1 M acetate buffer (pH 4.7), for various accumulation times. First of all, it is seen that clear voltammetry (with Cd, Pb and Bi peaks readily visible) can be obtained both in unbuffered and buffered solutions using the mini-electrode. This would not be possible at large electrodes in unbuffered solutions, where Bi(III) hydrolysis often occurs during the accumulation period at negative potentials due to extensive solution alkalisation in the
252
L. Baldrianova et al. / Analytica Chimica Acta 599 (2007) 249–255
tion levels (this is in contrast to our recent findings with heavily activated-anodised carbon microdisc electrodes, where acetate and pH levels played an important role, presumably due to a different catalytic activity of the graphite fiber electrode material used in that work towards metal deposition and hydrogen evolution [20]). Second, as it follows from the insets to Fig. 3(A) and (B), the Pb and Cd stripping peak current increases continuously with deposition time in the 60–600 s range. At deposition times longer than 500 s, signs of signal saturation appear that may be attributed to a thickening of the Bi deposit and its transformation from a nuclei-like to a dentritic structure (as confirmed by SEM in [8]); in that case the deposited metal analyte may be partially occluded within the three-dimensional structure of the Bi matrix. In further experimentation the deposition time of 600 s was chosen since longer periods were deemed as impractical and furthermore, due to the above-mentioned saturation effect, they are not expected to enhance the stripping peak signal considerably. 3.3. SWASV at Bi–CPmEs for different Pb and Cd concentration levels
Fig. 3. Square wave anodic stripping voltammograms (SWASVs) obtained at a CPmE from an unbuffered (A) and a 0.1 M acetate buffer (B) unstirred solution of 10−6 M Bi(III) + 10−7 M Pb(II) + 10−7 M Cd(II), following accumulation–deposition at −1.1 V vs. SCE for 60, 120, 180, 240, 300, 400, 500 and 600 s (peak height increases with deposition time, tdep ). Inset: variation of Pb and Cd stripping peak current with deposition time.
absence of an acidic buffer; the pH increase results from extensive hydrogen evolution at large area electrodes when exposed to the negative potentials of the accumulation step. This is because, although the rate of evolution–current density is expected to be the same irrespective of electrode size, the total amount of hydrogen evolved (and associated bulk solution alkalisation) depends on total current and hence on electrode area. Well-shaped Cd and Pb stripping peaks are located at ca. −0.80 and −0.54 V versus SCE for the unbuffered solution and at −0.82 and −0.58 V for the buffered solution (i.e. at similar values). The shape and position of Bi stripping peaks depends on deposition time and they lie in the −0.21 to −0.01 V range for the unbuffered and in the −0.20 to −0.16 V range for the buffered solution. The shift of the Bi peak to more negative potentials (easier to strip) as one passes from unbuffered (pH 2.7) to buffered (pH 4.7) solutions is in line with the expected decrease in the concentration of free Bi(III) ions which are transformed to hydroxy-complexes (e.g. BiOH+ ) [7] and which in turn have a more negative standard potential [27]. In general however, there are no striking differences regarding the background current at negative potentials as well as the shape and relative size of the Cd, Pb and Bi peaks observed in SWASVs at CPmEs in solutions of different acetate concentra-
Fig. 4 shows the SWASVs at a CPmE from an unbuffered 1 × 10−6 M Bi(III) solution of increasing Pb(II) and Cd(II) concentrations in the 3 × 10−8 to 1 × 10−6 M range (as indicated in the legend of the figure). Reasonably shaped stripping peaks are observed both for Cd and Pb. For higher analyte concentrations an additional peak deformation is seen between these two peaks which may be due to an impurity (note however that peak deformations have also been reported for experiments in unbuffered solutions at Bi-coated carbon microdiscs [20]). The inset presents plots of the Pb and Cd stripping peak current with the corresponding ion bulk concentration (linear regression equations were Y = 6.2550 × 10−2 X − 2.0190 × 10−9 with r2 = 0.9969 and Y = 2.5249 × 10−2 X − 3.6795 ×
Fig. 4. Square wave anodic stripping voltammograms (SWASVs) obtained at a CPmE from unbuffered unstirred solutions of 10−6 M Bi(III), containing equimolar concentrations of Pb(II) + Cd(II) as indicated in the graph (increasing peaks correspond to increasing metal analyte concentrations); accumulation at −1.1 V vs. SCE for 600 s. Inset: variation of Pb and Cd stripping peak current with metal ion concentration.
L. Baldrianova et al. / Analytica Chimica Acta 599 (2007) 249–255
Fig. 5. Square wave anodic stripping voltammograms (SWASVs) obtained at a CPmE from 0.1 M acetate buffer unstirred solutions of 10−6 M Bi(III), containing equimolar concentrations of Pb(II) + Cd(II) as indicated in the graph (increasing peaks correspond to increasing metal analyte concentrations); accumulation at −1.1 V vs. SCE for 600 s. Inset: variation of Pb and Cd stripping peak current with metal ion concentration.
10−10 with r2 = 0.9872, respectively). Calibration plots of reasonable linearity were obtained for Pb but of limited linearity for Cd. The rather poor response in the case of Cd (higher concentration accessible and limited linearity) may be attributed to the proximity of its stripping peak to the hydrogen evolution background current at negative potentials which is rather high in the case of the unbuffered solution of pH 2.8 (due to the quantity of HNO3 introduced to the sample by the addition of the Bi(III) stock solution, see also above). Fig. 5 shows the SWASVs at a CPmE from a buffered 1 × 10−6 M Bi(III) + 100 mM acetate buffer solution (pH 4.7) of increasing Pb(II) and Cd(II) concentrations in the
253
1 × 10−8 to 5 × 10−7 M range (as indicated in the legend of the figure). Symmetrical stripping peaks are observed both for Cd and Pb. The inset presents plots of the Pb and Cd stripping peak current with the corresponding ion bulk concentration and good linearity for both Pb and Cd was obtained in this case (linear regression equations were Y = 3.2085 × 10−1 X − 2.4891 × 10−9 with r2 = 0.9996 and Y = 2.1900 × 10−1 X − 9.4822 × 10−9 with r2 = 0.9999, respectively). Similar results were obtained at a CPmE from a weakly buffered 1 × 10−6 M Bi(III) + 10 mM acetate buffer solution of increasing Pb(II) and Cd(II) concentrations in the 1 × 10−8 to 1 × 10−7 M range. The linear regression equations of the calibration plots were Y = 3.3299 × 10−1 X − 2.8534 × 10−9 with r2 = 0.9963 for Pb and Y = 1.7920 × 10−1 X − 2.7548 × 10−9 with r2 = 0.9943 for Cd. Good linearity is achieved again, indicating that acetate buffer levels as low as 10 mM are enough to improve the linearity of the calibration plots for both Pb(II) and Cd(II) analytes. The detection limits for Pb(II) and Cd(II) in all three acetate level cases were estimated by fitting the best polynomial function (usually of a third degree) to the baseline of curves similar to those of Figs. 4 and 5, calculating the pooled standard deviation (sp ) of the experimental data, dividing that by the sensitivity (slope) of the linear approximation of the calibration plots and multiplying by 3 (3 × sp × slope−1 [15]). The sensitivity, linearity and detection limit for the three acetate levels studied are given in Tables 1 and 2 for Pb(II) and Cd(II), respectively. For comparison, the detection limits obtained at carbon microdisc electrodes (CEs, 30 m diameter) in unstirred solutions [20] and at conventional CPEs (2 mm diameter) in stirred solutions [8], are also given. It can be seen that the addition of a small (10 mM) buffer quantity significantly improves the sensitivity
Table 1 Sensitivity (expressed as Ip /CM , where Ip is the SWASV stripping peak current in A and CM the metal ion analyte concentration in M), linearity (expressed as the r2 of the fitted straight line) and detection limit (in M) for a CPmE (of a 72 m equivalent diameter), a carbon microdisc electrode (30 m diameter; CE) and a conventional CPE electrode (2 mm diameter), for the SWASV determination of Pb(II) from 10−6 M Bi(III) and variable acetate buffer content (0, 0.01, 0.1 M) solutions Pb(II)
Acetate buffer (0 M) Acetate buffer (0.01 M) Acetate buffer (0.1 M) a b
109 × detection limit (M)
CPmEa Sensitivity (AM−1 )
Linearity (r2 )
109 × detection limit (M)
CE [20]a
0.063 0.330 0.321
0.9969 0.9962 0.9996
3.19 0.77 0.87
45 11
CPE [8]b
0.50
Unstirred solutions, accumulation time of 600 s. Stirred solutions, accumulation time of 120 s.
Table 2 Sensitivity (expressed as Ip /CPb , where Ip is the SWASV stripping peak current in A and CM the metal ion analyte concentration in M), linearity (expressed as the r2 of the fitted straight line) and detection limit (in M) for a CPmE, a carbon microdisc electrode (30 m diameter; CE) and a conventional CPE electrode (2 mm diameter), for the SWASV determination of Cd(II) from 10−6 M Bi(III) and variable acetate buffer content (0, 0.01, 0.1 M) solutions Cd(II)
Acetate buffer (0 M) Acetate buffer (0.01 M) Acetate buffer (0.1 M) a b
109 × detection limit (M)
CPmEa Sensitivity (AM−1 )
Linearity (r2 )
109 × detection limit (M)
CE [20]a
0.025 0.179 0.219
0.9872 0.9943 0.9987
7.90 1.43 1.28
150
Unstirred solutions, accumulation time of 600 s. Stirred solutions, accumulation time of 120 s.
20
CPE [8]b
1.30
254
L. Baldrianova et al. / Analytica Chimica Acta 599 (2007) 249–255
and, to a less extent, the detection limit of BiF–CPmEs when compared to those in unbuffered samples. The most interesting finding however is that the BiF–CPmEs show in all cases a detection limit which is more than an order of magnitude lower than that of the bismuth film carbon microdisc electrodes (BiF–CEs) [20], under the same conditions (no stirring and 600 s accumulation time). The improved performance of carbon paste mini-electrodes when compared to the plain carbon microdisc electrodes should be attributed to the fact that CPmEs combine the advantages of near-microelectrode behaviour (high mass transfer rates) with those of carbon paste electrodes (high electroactive surface area, readily renewable). Although the mass transfer rates km , for metal ion deposition are expected to be higher at the 30 m diameter microdiscs of [20] than at the 72 m diameter mini-electrode (but still of the same order: km = 4D/πr [9], hence (km )CE /(km )CPmE = 72/30 = 2.4), the higher total deposition current at the larger diameter electrode (Eq. (1)) means that a larger amount of analyte is preconcentrated on it. Of course, the higher nominal and true surface area of the rough mini-electrode should in general result to higher capacitive currents and higher hydrogen evolution currents at the accumulation potential. However, close inspection of the SWASV curves of this work with those of [20] reveals that similarly flat baselines were recorded between the stripping peaks, presumably due to the subtraction of the reverse response from the forward one that is characteristic of the SWASV technique. Also, the relative magnitude of the hydrogen evolution current recorded close to the negative potential limit is not higher at the CPmE, presumably due to a different catalytic activity towards hydrogen evolution of the anodized graphite microdisc of [20] and the carbon paste of this work (note that CPEs are known to have an extended cathodic limit [28]). Finally, the higher surface area of CPmE may also affect the structure of the Bi deposit and of the co-deposited Pb and Cd too which in turn should determine the stripping behaviour of the latter. It has been shown by SEM studies that, unlike flat glassy carbon substrates were a dense Bi deposit is formed [2], the Bi deposits at carbon paste electrodes are thinner and comprise of sparser nuclei and dendrites [8]. Thus a higher Bi surface area for Pb and Cd is available, and their occlusion in the bulk of dense Bi deposits is avoided [8,20]. The lower detection limit at BiF–CPmEs for Pb(II) is 7.7 × 10−10 M (achieved in 0.01 M acetate buffer) and for Cd(II) is 1.28 × 10−9 M (achieved in 0.1 M acetate buffer). It should also be noted in Tables 1 and 2 that these detection limits are similar to those reported at large CPEs under usual operational conditions (for shorter accumulation times but under constant stirring [8]). 3.4. Determination of Pb and Cd in contaminated tap water samples A tap water sample from the Physical Chemistry Laboratory of the Aristotle University of Thessaloniki, was deliberately spiked with 3 × 10−8 M Pb + 3 × 10−8 M Cd + 10−6 M Bi(III). Fig. 6 presents the SWASVs obtained at a CPmE, upon standard additions of Pb(II) and Cd(II) ions in that contaminated tap water
Fig. 6. Square wave anodic stripping voltammograms (SWASVs) obtained at a CPmE in a tap water sample from the Physical Chemistry Laboratory of the Aristotle University of Thessaloniki. The sample has been deliberately spiked with 3 × 10−8 M Pb + 3 × 10−8 M Cd + 10−6 M Bi(III) and further spiked by standard additions of equimolar quantities of Pb(II) and Cd(II) (as indicated on the graph). Inset: variation of the Pb and Cd stripping peak current with quantity of Pb(II) or Cd(II) added to the sample.
sample. The total sample volume was 20 mL and no buffer addition or stirring was employed. The Inset shows the corresponding standard additions plots (Pb or Cd stripping peak current versus added quantity of Pb(II) or Cd(II)), with linear regression equations of Y = 8.444E − 02 × X + 3.015E − 09 (r2 = 0.9866) for Pb and Y = 2.193E − 02 × X + 5.974E − 10 (r2 = 0.9916) for Cd. From these the concentration of Pb(II) and Cd(II) in the spiked tap water sample was estimated as 3.5 × 10−8 M (119% recovery) and 2.7 × 10−8 M (91% recovery), respectively. 4. Conclusions (1) Carbon paste mini-electrodes (50–300 m diameter) plated in situ with Bi films (BiF–CPmEs) were successfully used in the square wave anodic stripping voltammetry (SWASV) determination of Pb(II) and Cd(II) in the wide 10−8 to 10−6 M range. Their near-microelectrode behaviour permitted measurements in unstirred solutions (due to high mass transfer rates resulting from non-planar diffusion) as well as in poorly buffered samples (due to a small extent of Bi(III) hydrolysis since at their small area limited quantities of hydrogen and hydroxyl ions are produced during the accumulation period at negative potentials). As an application of these properties, Pb(II) and Cd(II) were determined in an unstirred and unbuffered contaminated tap water sample. (2) Since CPmEs combine near-microelectrode behaviour (high mass transfer rates) and carbon paste properties (high electroactive surface area), improved detection limits are achieved in unbuffered and unstirred samples, when compared to those of carbon microdisc electrodes [20]. The detection limits for Pb(II) and Cd(II) under optimal conditions at BiF–CPmEs are 7.7 × 10−10 M (ca. 0.154 gL−1 ) and 1.28 × 10−9 M (ca. 0.152 gL−1 ), respectively, i.e. similar to those reported at large BiF–CPEs in stirred solutions [8].
L. Baldrianova et al. / Analytica Chimica Acta 599 (2007) 249–255
Acknowledgements A Greece–Czech Republic bilateral Research Grant (ET 205-␥, KONTRAKT No. 7-2006-24) is gratefully acknowledged. Czech authors would also like to acknowledge the Ministry of Education, Youth and Sport (project LC 06035) for financial support.
[10] [11] [12] [13] [14] [15] [16] [17]
References
[18] [19]
[1] [2] [3] [4] [5] [6] [7]
[8] [9]
J. Wang, Electroanalysis 17 (2005) 1341. A. Economou, Trends Anal. Chem. 24 (2005) 334. J. Wang, J. Lu, S.B. Hocevar, P.A.M. Farias, Anal. Chem. 72 (2000) 3218. J. Wang, J. Lu, S.B. Hocevar, B. Ogorevc, Electroanalysis 13 (1) (2001) 13. J. Wang, J. Lu, U.A. Kirgoz, S.B. Hocevar, B. Ogorevc, Anal. Chim. Acta 434 (2001) 29. J. Wang, Acc. Chem. Res. 35 (2002) 811. I. Svancara, L. Baldrianova, E. Tesarova, S. Hocevar, S. Elsuccary, A. Economou, S. Sotiropoulos, B. Ogorevc, K. Vytras, Electroanalysis 18 (2) (2006) 177. L. Baldrianova, I. Svancara, M. Vlcek, A. Economou, S. Sotiropoulos, Electrochim. Acta 52 (2006) 481. D. Pletcher, Microelectrodes: theory and applications, in: M.I. Montenegro, M.A. Queiros (Eds.), NATO ASI Series, Series E: Applied Science, vol. 197, Kluwer Academic Publishers, 1991, pp. 3–16.
[20] [21] [22] [23] [24] [25] [26] [27]
[28]
255
J. Osteryoung, in ibid, 1991, pp. 139–175. S.-F. Zou, H. Men, P. Wang, Chin. J. Sens. Actuat. 17 (2) (2004) 336. X. Xie, D. Stueben, Z. Berner, Anal. Lett. 38 (14) (2005) 2281. J. Wang, P. Tuzhi, Anal. Chim. Acta 197 (1987) 367. D.K.Y. Wong, A.G. Ewing, Anal. Chem. 62 (1990) 2697. M. Wojciechowski, J. Balcerzak, Anal. Chim. Acta 249 (1991) 433. L. Nyholm, G. Wikmark, Anal. Chim. Acta 257 (1992) 7. S. Hocevar, B. Ogorevc, J. Wang, B. Pihlar, Electroanalysis 14 (24) (2002) 1707. M.A. Baldo, S. Daniele, Anal. Lett. 37 (5) (2004) 995. E.A. Hutton, S.B. Hocevar, B. Ogorevc, Anal. Chim. Acta 537 (2005) 285. L. Baldrianova, I. Svancara, A. Economou, S. Sotiropoulos, Anal. Chim. Acta 580 (2006) 24. J. Wang, L.D. Hutchins, Bioelectrochem. Bioenerg. 12 (1984) 193. J. Wang, X. Zhang, M. Prakash, Anal. Chim. Acta 11 (1999) 11. Y. Zou, J. Mo, Anal. Chim. Acta 382 (1999) 145. J. Calvo Quintana, L. Idrissi, G. Palleschi, P. Albertano, A. Amine, M. El Rhazi, D. Moscone, Talanta 63 (2004) 567. M.V. Stackelberg, M. Pilgram, V. Toome, Z. Electrochem. 57 (1953) 342. K. Aoki, K. Akimoto, K. Tokuda, H. Matsuda, J. Osteryoung, J. Electroanal. Chem. 171 (1984) 219. B. Lovrecek, I. Mekjavic, M. Metikos-Hukovic, in: A.J. Bard, R. Parsons, J. Jordan (Eds.), Standard Potentials in Aqueous Solution, Marcel Dekker, Inc., NY and Basel, 1985, pp. 180–187. I. Svancara, M. Pravda, M. Hvizdalova, K. Vytras, Electroanalysis 6 (1994) 663.
Anodic stripping voltammetry at a new type of disposable bismuth-plated carbon paste mini-electrodes L. Baldrianova a , I. Svancara a , S. Sotiropoulos b,∗ a
Department of Analytical Chemistry, Faculty of Chemical Technology, University of Pardubice, CZ-532 10 Pardubice, Czech Republic b Department of Chemistry, Aristotle University of Thessaloniki, Thessaloniki 54124, Greece Received 22 May 2007; received in revised form 19 July 2007; accepted 25 July 2007 Available online 3 August 2007
Abstract A new type of disposable carbon paste mini-electrodes (CPmEs), with dimensions in the 50–300 m range, have been fabricated by heat-shrinking the end-tip of plastic micropipette tips and filling them with carbon paste. The CPmEs have been characterized by microscopic and electrochemical means and tested as substrates for in situ plated Bi film electrodes (BiF–CPmEs), used in the determination of heavy metals by square wave anodic stripping voltammetry (SWASV). It was found that this new class of CPmEs combines the advantages of carbon paste electrodes (readily renewable surface and high surface area) with those of near-microelectrode behaviour (no stirring or electrolyte excess needed). During SWASV experiments in unstirred Pb(II) and Cd(II) solutions well-shaped stripping peaks were obtained whose height varied linearly with analyte concentration in the wide 1 × 10−8 to 10−6 M range, both in acetate buffer and unbuffered solutions. Under optimal conditions detection limits of 8 × 10−10 and 1.3 × 10−9 M were achieved for Pb(II) and Cd(II), respectively and in a trial application, these metal ions have been determined in a spiked tap water sample using a BiF–CPmE. © 2007 Elsevier B.V. All rights reserved. Keywords: Bismuth film electrodes; Microelectrodes; Carbon paste electrodes; Anodic stripping voltammetry
1. Introduction The low toxicity of Bi and its salts, when compared to Hg, as well as its satisfactory electrochemical performance (relatively wide potential window and formation of fused alloys with many metals) has led to its use as an alternative to Hg for stripping voltammetry [1–8]. Besides glassy carbon and graphite which have been the most common electrode substrates for Bi film plating [3–6], carbon paste electrodes (CPEs) are also a useful class of electrode supports for Bi stripping voltammetry [7–8] since they offer a readily renewable surface, they are easy to modify and are also low cost electrodes. Microelectrodes (i.e. electrodes with at least one of their dimensions smaller than typically 50 m [9]) have a number of properties that make them attractive in stripping analysis [10–12]. First, enhanced mass transfer rates due to non-planar
∗
Corresponding author. Tel.: +30 2310 997742; fax: +30 2310 443922. E-mail addresses: [email protected], [email protected] (S. Sotiropoulos).
0003-2670/$ – see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.aca.2007.07.067
diffusion make possible shorter accumulation periods and/or no need for stirring during that period. Second, small ohmic losses allow experiments without the need for deliberately added electrolyte (often the source of impurities). Hg-coated microelectrodes have long been used in stripping voltammetry and in particular in media of ambient or minimal electrolyte content [13–16]. However, there are only few papers on Bicoated microelectrodes [17–20], mainly of the microdisc or microfibre type. At the same time there are also rather few papers on CPE microelectrodes [21–24]; these are usually of the microdisc type and are fabricated with glass capillaries or Teflon tubing. It should be stressed that, since those electrodes had diameters larger than 200 m (with the single exception of those in [23] that were based on a glass capillary with a 75 m internal diameter), they were not strictly speaking microelectrodes. According to microelectrode terminology [9] their critical dimension should have not exceeded 50 m and hence such electrodes should better be considered as mid-size electrodes or mini-electrodes. In any case, there are no papers either on CPE mini-electrodes based on disposable plastic holders
250
L. Baldrianova et al. / Analytica Chimica Acta 599 (2007) 249–255
with a diameter smaller than 200 m or on Bi-coated such electrodes. The aim of the present work has been to introduce a new type of carbon paste mini-electrodes (CPmEs) as substrates for the preparation of in situ coated Bi film electrodes (BiF–CPmEs), to be used in the determination of metal ions by square wave anodic stripping voltammetry (SWASV). In more detail, its objectives have been (i) to fabricate CPmEs with a diameter in the 50–300 m range, using disposable micropipette tips whose end-tip is heat-shrunk to micrometer dimensions and are filled with carbon paste, (ii) to characterise the new type of CPmEs by means of microscopy and electrochemistry, (iii) to exploit the near-microelectrode behaviour of BiF–CPmEs in the SWASV determination of Pb(II) and Cd(II) from unstirred samples containing a variable acetate buffer concentration (in the 0–0.1 M range) and (iv) to use BiF–CPmEs in the determination of Pb(II) and Cd(II) in a tap water sample. 2. Experimental 2.1. Preparation and characterisation of CPmEs The holder for the CPmEs was based on the lower part (ca. 2 cm long) of disposable micropipette tips (Universal tip 500193, WPI Inc., end-tip diameter ca. 500 m). These were shrunk to an end-tip diameter in the 50–300 m range by brief exposure to a hot air stream (400–450 ◦ C) produced by a heat gun (Model 1850, Beta Utensili S.r.l.). In cases that the manual heating procedure resulted in closed end-tips the latter were cut with a razor blade under a magnifying glass, to produce a micrometer size diameter opening. The heat-shrunk holder was then packed with the carbon paste (for its composition see below) with the help of a common pin that also served as the electrical contact. The surface of the resulting CPmE was renewed between sets of measurements by further pushing the pin into the paste whereby a micrometer-sized thread of carbon paste could be squeezed out of the holder’s opening and wiped with tissue paper (A schematic diagram and a picture of a CPmE are shown in Fig. 1(A), Section 3.1 below). The diameter of the CPmEs could be checked by an optical microscope (INTRACO MICRO, s.r.o., SM 6A, Czech Republic) or by SEM (JEOL JSM-5600LV) after the electrochemical experiments. The geometric area of the CPmEs as well as their near-microelectrode behaviour was confirmed by running slow potential scan voltammograms for the reduction of ferricyanide from 5 mM potassium ferricyanide solutions in 0.1 M KCl. The carbon paste used was prepared by intimate hand-mixing of 0.5 g spectroscopic graphite powder (“RW-B”, RingsdorffWerke,Germany) with approximately 0.3 mL highly viscous silicone oil (“SO”, LUKOIL MV 12000 product; Lucebni zavody Kolin, Czech Republic). 2.2. In situ bismuth plating procedure and square wave anodic stripping voltammetry (SWASV) parameters In situ bismuth films were prepared by spiking the sample with the required Bi(III) quantity from a standard solution of
Fig. 1. Schematic diagram (A) and SEM micrograph (B) of a CPmE.
1 × 10−4 M Bi(III) containing 0.1 M HNO3 to reach always 1 × 10−6 M Bi(III) + 0.001 M HNO3 levels in the sample. The standard solution was prepared by appropriate dilution of a spectroscopic standard solution of 1000 ppm Bi(III) in 0.5 M HNO3 and consequent addition of 1 M HNO3 to reach the desired acid concentration and solution pH. The SWASV conditions were Edeposition = −1.1 V versus SCE; Efinal = + 0.5 V; tdeposition = 60 − 600 s; teq = 15 s; frequency = 25 Hz; potential step = 4 mV; amplitude = 20 mV. Before each measurement Econditioning = +1.8 V was applied on the electrode for 120 s to ensure complete Bi stripping. 2.3. Electrochemical cell and instrumentation A three-electrode cell was used, consisting of the CPmE working electrode, a SCE reference electrode (Radiometer) and a Pt coil auxiliary electrode (BAS Inc.). A laboratory-made aluminium Faraday cage housed the electrochemical cell in order to minimise electrical noise. The potentiostats used were a Model PGSTAT 100 Autolab or a AUTOLAB (Ecochimie, Utrecht, Holand), connected to a personal computer and controlled by GPES, version 4.8 software (Ecochimie, Utrecht, Holand). The pH was measured by a Radiometer Blue Line pH-meter and samples were spiked by a 20–100 L and 100–1000 L micropipetter (Eagle, from WPI Inc.).
L. Baldrianova et al. / Analytica Chimica Acta 599 (2007) 249–255
2.4. Chemicals and sample preparation All chemicals used in the preparation of stock solutions were of analytical reagent grade and obtained from Merck (Damstadt, Germany) and Fluka. Concentrated CH3 COOH p.a. (Fluka) and CH3 COONa p.a. (Fluka) solutions were used for the preparation of 1 M stock acetate buffer. Stock solutions for in situ bismuth plating, were prepared from an AAS (spectroscopic grade) 1000 ppm standard solution of Bi(NO3 )3 ·5H2 O (Merck, Germany), while Pb(NO3 )2 p.a. and Cd(NO3 )2 ·4H2 O p.a. (Merck, Germany) were employed in the preparation of 0.01 M metal ion stock solutions. Where appropriate, the respective solutions were diluted as needed. All the solutions of standards and diluted standards of Pb(II) and Cd(II) were always stabilised by acidifying with 500 L of 65% HNO3 (Fluka) per 100 mL of solution (resulting in 0.07 M HNO3 concentration in the stock solutions). For the calibration of the CPmEs potassium, ferricyanide and KCl (Merck) were used. In the preparation of solutions doubly redistilled water was employed. 3. Results and discussion 3.1. Microscopic and electrochemical characterisation of carbon paste mini-electrodes (CPmEs) Fig. 1(A and B) shows the schematic diagram and a SEM micrograph of a CPmE. In the latter, the rough, undulating surface of the carbon paste material can be seen, protruding out of the plastic tip. This is surrounded by a ring of a thin, transparent plastic film covering the carbon paste material. The dimensions of the mini-electrode corresponding to the protruding carbon material shown in that micrograph are (depending on location) in the 200–300 m range. Mini-electrodes with dimensions in the 50–300 m range were fabricated by this technique (as also confirmed by the electrochemical calibration discussed below) while those of a larger diameter were dismissed. The majority of our experiments were performed on mini-electrodes with a diameter smaller than 100 m. Fig. 2 shows a slow scan voltammogram (at 5 mV s−1 ) of a CPmE in a deaerated 5 mM potassium ferricyanide + 0.1 M
251
KCl solution. A sigmoidal curve with a near-coincidence of the positive- and negative-going scans corresponding to the one-electron reduction of ferricyanide can be seen, indicating a near-steady state response, characteristic of near-microelectrode behaviour. The equation holding for the diffusion-limited current, IL , associated with the n-electron reaction of a species (with a concentration of C and a diffusion coefficient of D), at a microdisc electrode (with a radius of r) is [9] IL = 4nFCDr
(1)
applying (1) in this case for n = 1, C = 5 × 10−3 M, D = 7.63 × 10−6 cm2 s−1 [25] and considering the current at −0.3 V, the CPmE of Fig. 2 can be estimated to give a limiting current equivalent to that of a 72 m diameter disc electrode. Eq. (1) assumes pure non-planar, hemi-spherical (more precisely, hemi-cylindrical) diffusion towards a microdisc electrode and can be considered a first approximation. Strictly speaking though, electrodes with diameters larger than 50 m are not microelectrodes and both non-planar and planar diffusion effects should be taken into account. The complete equation proposed to apply for all small electrodes is due to Aoki et al. [26] IL = 4nFCDr 0.34 exp(−0.66p) + 0.66 − 0.13 −11 ×exp + 0.351p (2) p with p=
nFr 2 υ RTD
1/2 (3)
where υ is the potential scan rate. Using the values of n, C and D given above and υ = 0.005 V s−1 , Eq. (2) was solved for r with the help of Micrososft Excel Solver® and a diameter of 2r = 68 m could be estimated, i.e. within 5% of the value estimated using the simple equation of (1). Both these values are close to that of ca. 80 m determined at the end of experimentation by means of S.E.M. The slight increase in diameter may be due to tip widening caused by the carbon paste extrusion-renewal between sets of experiments in different media. All of the results presented below were obtained at this particular mini-electrode. 3.2. Effect and choice of accumulation time in the SWASV determination of Pb(II) and Cd(II) at CPmEs, from buffered and unbuffered unstirred solutions
Fig. 2. Voltammogram of a CPmE in a 5 mM ferricyanide + 0.1 M KCl deaerated solution, recorded at 5 mV s−1 potential scan rate.
Fig. 3(A) and (B) shows the SWASV picture at CPmEs from a 1 × 10−6 M Bi(III) unstirred solution containing 10−7 M Pb(II) + Cd(II) and (A) no added buffer (pH 2.8, due to 1 mM HNO3 added with the 10−6 M Bi(III) spiked from a 10−4 M Bi(III) + 0.1 M HNO3 acidified stock solution) or (B) 0.1 M acetate buffer (pH 4.7), for various accumulation times. First of all, it is seen that clear voltammetry (with Cd, Pb and Bi peaks readily visible) can be obtained both in unbuffered and buffered solutions using the mini-electrode. This would not be possible at large electrodes in unbuffered solutions, where Bi(III) hydrolysis often occurs during the accumulation period at negative potentials due to extensive solution alkalisation in the
252
L. Baldrianova et al. / Analytica Chimica Acta 599 (2007) 249–255
tion levels (this is in contrast to our recent findings with heavily activated-anodised carbon microdisc electrodes, where acetate and pH levels played an important role, presumably due to a different catalytic activity of the graphite fiber electrode material used in that work towards metal deposition and hydrogen evolution [20]). Second, as it follows from the insets to Fig. 3(A) and (B), the Pb and Cd stripping peak current increases continuously with deposition time in the 60–600 s range. At deposition times longer than 500 s, signs of signal saturation appear that may be attributed to a thickening of the Bi deposit and its transformation from a nuclei-like to a dentritic structure (as confirmed by SEM in [8]); in that case the deposited metal analyte may be partially occluded within the three-dimensional structure of the Bi matrix. In further experimentation the deposition time of 600 s was chosen since longer periods were deemed as impractical and furthermore, due to the above-mentioned saturation effect, they are not expected to enhance the stripping peak signal considerably. 3.3. SWASV at Bi–CPmEs for different Pb and Cd concentration levels
Fig. 3. Square wave anodic stripping voltammograms (SWASVs) obtained at a CPmE from an unbuffered (A) and a 0.1 M acetate buffer (B) unstirred solution of 10−6 M Bi(III) + 10−7 M Pb(II) + 10−7 M Cd(II), following accumulation–deposition at −1.1 V vs. SCE for 60, 120, 180, 240, 300, 400, 500 and 600 s (peak height increases with deposition time, tdep ). Inset: variation of Pb and Cd stripping peak current with deposition time.
absence of an acidic buffer; the pH increase results from extensive hydrogen evolution at large area electrodes when exposed to the negative potentials of the accumulation step. This is because, although the rate of evolution–current density is expected to be the same irrespective of electrode size, the total amount of hydrogen evolved (and associated bulk solution alkalisation) depends on total current and hence on electrode area. Well-shaped Cd and Pb stripping peaks are located at ca. −0.80 and −0.54 V versus SCE for the unbuffered solution and at −0.82 and −0.58 V for the buffered solution (i.e. at similar values). The shape and position of Bi stripping peaks depends on deposition time and they lie in the −0.21 to −0.01 V range for the unbuffered and in the −0.20 to −0.16 V range for the buffered solution. The shift of the Bi peak to more negative potentials (easier to strip) as one passes from unbuffered (pH 2.7) to buffered (pH 4.7) solutions is in line with the expected decrease in the concentration of free Bi(III) ions which are transformed to hydroxy-complexes (e.g. BiOH+ ) [7] and which in turn have a more negative standard potential [27]. In general however, there are no striking differences regarding the background current at negative potentials as well as the shape and relative size of the Cd, Pb and Bi peaks observed in SWASVs at CPmEs in solutions of different acetate concentra-
Fig. 4 shows the SWASVs at a CPmE from an unbuffered 1 × 10−6 M Bi(III) solution of increasing Pb(II) and Cd(II) concentrations in the 3 × 10−8 to 1 × 10−6 M range (as indicated in the legend of the figure). Reasonably shaped stripping peaks are observed both for Cd and Pb. For higher analyte concentrations an additional peak deformation is seen between these two peaks which may be due to an impurity (note however that peak deformations have also been reported for experiments in unbuffered solutions at Bi-coated carbon microdiscs [20]). The inset presents plots of the Pb and Cd stripping peak current with the corresponding ion bulk concentration (linear regression equations were Y = 6.2550 × 10−2 X − 2.0190 × 10−9 with r2 = 0.9969 and Y = 2.5249 × 10−2 X − 3.6795 ×
Fig. 4. Square wave anodic stripping voltammograms (SWASVs) obtained at a CPmE from unbuffered unstirred solutions of 10−6 M Bi(III), containing equimolar concentrations of Pb(II) + Cd(II) as indicated in the graph (increasing peaks correspond to increasing metal analyte concentrations); accumulation at −1.1 V vs. SCE for 600 s. Inset: variation of Pb and Cd stripping peak current with metal ion concentration.
L. Baldrianova et al. / Analytica Chimica Acta 599 (2007) 249–255
Fig. 5. Square wave anodic stripping voltammograms (SWASVs) obtained at a CPmE from 0.1 M acetate buffer unstirred solutions of 10−6 M Bi(III), containing equimolar concentrations of Pb(II) + Cd(II) as indicated in the graph (increasing peaks correspond to increasing metal analyte concentrations); accumulation at −1.1 V vs. SCE for 600 s. Inset: variation of Pb and Cd stripping peak current with metal ion concentration.
10−10 with r2 = 0.9872, respectively). Calibration plots of reasonable linearity were obtained for Pb but of limited linearity for Cd. The rather poor response in the case of Cd (higher concentration accessible and limited linearity) may be attributed to the proximity of its stripping peak to the hydrogen evolution background current at negative potentials which is rather high in the case of the unbuffered solution of pH 2.8 (due to the quantity of HNO3 introduced to the sample by the addition of the Bi(III) stock solution, see also above). Fig. 5 shows the SWASVs at a CPmE from a buffered 1 × 10−6 M Bi(III) + 100 mM acetate buffer solution (pH 4.7) of increasing Pb(II) and Cd(II) concentrations in the
253
1 × 10−8 to 5 × 10−7 M range (as indicated in the legend of the figure). Symmetrical stripping peaks are observed both for Cd and Pb. The inset presents plots of the Pb and Cd stripping peak current with the corresponding ion bulk concentration and good linearity for both Pb and Cd was obtained in this case (linear regression equations were Y = 3.2085 × 10−1 X − 2.4891 × 10−9 with r2 = 0.9996 and Y = 2.1900 × 10−1 X − 9.4822 × 10−9 with r2 = 0.9999, respectively). Similar results were obtained at a CPmE from a weakly buffered 1 × 10−6 M Bi(III) + 10 mM acetate buffer solution of increasing Pb(II) and Cd(II) concentrations in the 1 × 10−8 to 1 × 10−7 M range. The linear regression equations of the calibration plots were Y = 3.3299 × 10−1 X − 2.8534 × 10−9 with r2 = 0.9963 for Pb and Y = 1.7920 × 10−1 X − 2.7548 × 10−9 with r2 = 0.9943 for Cd. Good linearity is achieved again, indicating that acetate buffer levels as low as 10 mM are enough to improve the linearity of the calibration plots for both Pb(II) and Cd(II) analytes. The detection limits for Pb(II) and Cd(II) in all three acetate level cases were estimated by fitting the best polynomial function (usually of a third degree) to the baseline of curves similar to those of Figs. 4 and 5, calculating the pooled standard deviation (sp ) of the experimental data, dividing that by the sensitivity (slope) of the linear approximation of the calibration plots and multiplying by 3 (3 × sp × slope−1 [15]). The sensitivity, linearity and detection limit for the three acetate levels studied are given in Tables 1 and 2 for Pb(II) and Cd(II), respectively. For comparison, the detection limits obtained at carbon microdisc electrodes (CEs, 30 m diameter) in unstirred solutions [20] and at conventional CPEs (2 mm diameter) in stirred solutions [8], are also given. It can be seen that the addition of a small (10 mM) buffer quantity significantly improves the sensitivity
Table 1 Sensitivity (expressed as Ip /CM , where Ip is the SWASV stripping peak current in A and CM the metal ion analyte concentration in M), linearity (expressed as the r2 of the fitted straight line) and detection limit (in M) for a CPmE (of a 72 m equivalent diameter), a carbon microdisc electrode (30 m diameter; CE) and a conventional CPE electrode (2 mm diameter), for the SWASV determination of Pb(II) from 10−6 M Bi(III) and variable acetate buffer content (0, 0.01, 0.1 M) solutions Pb(II)
Acetate buffer (0 M) Acetate buffer (0.01 M) Acetate buffer (0.1 M) a b
109 × detection limit (M)
CPmEa Sensitivity (AM−1 )
Linearity (r2 )
109 × detection limit (M)
CE [20]a
0.063 0.330 0.321
0.9969 0.9962 0.9996
3.19 0.77 0.87
45 11
CPE [8]b
0.50
Unstirred solutions, accumulation time of 600 s. Stirred solutions, accumulation time of 120 s.
Table 2 Sensitivity (expressed as Ip /CPb , where Ip is the SWASV stripping peak current in A and CM the metal ion analyte concentration in M), linearity (expressed as the r2 of the fitted straight line) and detection limit (in M) for a CPmE, a carbon microdisc electrode (30 m diameter; CE) and a conventional CPE electrode (2 mm diameter), for the SWASV determination of Cd(II) from 10−6 M Bi(III) and variable acetate buffer content (0, 0.01, 0.1 M) solutions Cd(II)
Acetate buffer (0 M) Acetate buffer (0.01 M) Acetate buffer (0.1 M) a b
109 × detection limit (M)
CPmEa Sensitivity (AM−1 )
Linearity (r2 )
109 × detection limit (M)
CE [20]a
0.025 0.179 0.219
0.9872 0.9943 0.9987
7.90 1.43 1.28
150
Unstirred solutions, accumulation time of 600 s. Stirred solutions, accumulation time of 120 s.
20
CPE [8]b
1.30
254
L. Baldrianova et al. / Analytica Chimica Acta 599 (2007) 249–255
and, to a less extent, the detection limit of BiF–CPmEs when compared to those in unbuffered samples. The most interesting finding however is that the BiF–CPmEs show in all cases a detection limit which is more than an order of magnitude lower than that of the bismuth film carbon microdisc electrodes (BiF–CEs) [20], under the same conditions (no stirring and 600 s accumulation time). The improved performance of carbon paste mini-electrodes when compared to the plain carbon microdisc electrodes should be attributed to the fact that CPmEs combine the advantages of near-microelectrode behaviour (high mass transfer rates) with those of carbon paste electrodes (high electroactive surface area, readily renewable). Although the mass transfer rates km , for metal ion deposition are expected to be higher at the 30 m diameter microdiscs of [20] than at the 72 m diameter mini-electrode (but still of the same order: km = 4D/πr [9], hence (km )CE /(km )CPmE = 72/30 = 2.4), the higher total deposition current at the larger diameter electrode (Eq. (1)) means that a larger amount of analyte is preconcentrated on it. Of course, the higher nominal and true surface area of the rough mini-electrode should in general result to higher capacitive currents and higher hydrogen evolution currents at the accumulation potential. However, close inspection of the SWASV curves of this work with those of [20] reveals that similarly flat baselines were recorded between the stripping peaks, presumably due to the subtraction of the reverse response from the forward one that is characteristic of the SWASV technique. Also, the relative magnitude of the hydrogen evolution current recorded close to the negative potential limit is not higher at the CPmE, presumably due to a different catalytic activity towards hydrogen evolution of the anodized graphite microdisc of [20] and the carbon paste of this work (note that CPEs are known to have an extended cathodic limit [28]). Finally, the higher surface area of CPmE may also affect the structure of the Bi deposit and of the co-deposited Pb and Cd too which in turn should determine the stripping behaviour of the latter. It has been shown by SEM studies that, unlike flat glassy carbon substrates were a dense Bi deposit is formed [2], the Bi deposits at carbon paste electrodes are thinner and comprise of sparser nuclei and dendrites [8]. Thus a higher Bi surface area for Pb and Cd is available, and their occlusion in the bulk of dense Bi deposits is avoided [8,20]. The lower detection limit at BiF–CPmEs for Pb(II) is 7.7 × 10−10 M (achieved in 0.01 M acetate buffer) and for Cd(II) is 1.28 × 10−9 M (achieved in 0.1 M acetate buffer). It should also be noted in Tables 1 and 2 that these detection limits are similar to those reported at large CPEs under usual operational conditions (for shorter accumulation times but under constant stirring [8]). 3.4. Determination of Pb and Cd in contaminated tap water samples A tap water sample from the Physical Chemistry Laboratory of the Aristotle University of Thessaloniki, was deliberately spiked with 3 × 10−8 M Pb + 3 × 10−8 M Cd + 10−6 M Bi(III). Fig. 6 presents the SWASVs obtained at a CPmE, upon standard additions of Pb(II) and Cd(II) ions in that contaminated tap water
Fig. 6. Square wave anodic stripping voltammograms (SWASVs) obtained at a CPmE in a tap water sample from the Physical Chemistry Laboratory of the Aristotle University of Thessaloniki. The sample has been deliberately spiked with 3 × 10−8 M Pb + 3 × 10−8 M Cd + 10−6 M Bi(III) and further spiked by standard additions of equimolar quantities of Pb(II) and Cd(II) (as indicated on the graph). Inset: variation of the Pb and Cd stripping peak current with quantity of Pb(II) or Cd(II) added to the sample.
sample. The total sample volume was 20 mL and no buffer addition or stirring was employed. The Inset shows the corresponding standard additions plots (Pb or Cd stripping peak current versus added quantity of Pb(II) or Cd(II)), with linear regression equations of Y = 8.444E − 02 × X + 3.015E − 09 (r2 = 0.9866) for Pb and Y = 2.193E − 02 × X + 5.974E − 10 (r2 = 0.9916) for Cd. From these the concentration of Pb(II) and Cd(II) in the spiked tap water sample was estimated as 3.5 × 10−8 M (119% recovery) and 2.7 × 10−8 M (91% recovery), respectively. 4. Conclusions (1) Carbon paste mini-electrodes (50–300 m diameter) plated in situ with Bi films (BiF–CPmEs) were successfully used in the square wave anodic stripping voltammetry (SWASV) determination of Pb(II) and Cd(II) in the wide 10−8 to 10−6 M range. Their near-microelectrode behaviour permitted measurements in unstirred solutions (due to high mass transfer rates resulting from non-planar diffusion) as well as in poorly buffered samples (due to a small extent of Bi(III) hydrolysis since at their small area limited quantities of hydrogen and hydroxyl ions are produced during the accumulation period at negative potentials). As an application of these properties, Pb(II) and Cd(II) were determined in an unstirred and unbuffered contaminated tap water sample. (2) Since CPmEs combine near-microelectrode behaviour (high mass transfer rates) and carbon paste properties (high electroactive surface area), improved detection limits are achieved in unbuffered and unstirred samples, when compared to those of carbon microdisc electrodes [20]. The detection limits for Pb(II) and Cd(II) under optimal conditions at BiF–CPmEs are 7.7 × 10−10 M (ca. 0.154 gL−1 ) and 1.28 × 10−9 M (ca. 0.152 gL−1 ), respectively, i.e. similar to those reported at large BiF–CPEs in stirred solutions [8].
L. Baldrianova et al. / Analytica Chimica Acta 599 (2007) 249–255
Acknowledgements A Greece–Czech Republic bilateral Research Grant (ET 205-␥, KONTRAKT No. 7-2006-24) is gratefully acknowledged. Czech authors would also like to acknowledge the Ministry of Education, Youth and Sport (project LC 06035) for financial support.
[10] [11] [12] [13] [14] [15] [16] [17]
References
[18] [19]
[1] [2] [3] [4] [5] [6] [7]
[8] [9]
J. Wang, Electroanalysis 17 (2005) 1341. A. Economou, Trends Anal. Chem. 24 (2005) 334. J. Wang, J. Lu, S.B. Hocevar, P.A.M. Farias, Anal. Chem. 72 (2000) 3218. J. Wang, J. Lu, S.B. Hocevar, B. Ogorevc, Electroanalysis 13 (1) (2001) 13. J. Wang, J. Lu, U.A. Kirgoz, S.B. Hocevar, B. Ogorevc, Anal. Chim. Acta 434 (2001) 29. J. Wang, Acc. Chem. Res. 35 (2002) 811. I. Svancara, L. Baldrianova, E. Tesarova, S. Hocevar, S. Elsuccary, A. Economou, S. Sotiropoulos, B. Ogorevc, K. Vytras, Electroanalysis 18 (2) (2006) 177. L. Baldrianova, I. Svancara, M. Vlcek, A. Economou, S. Sotiropoulos, Electrochim. Acta 52 (2006) 481. D. Pletcher, Microelectrodes: theory and applications, in: M.I. Montenegro, M.A. Queiros (Eds.), NATO ASI Series, Series E: Applied Science, vol. 197, Kluwer Academic Publishers, 1991, pp. 3–16.
[20] [21] [22] [23] [24] [25] [26] [27]
[28]
255
J. Osteryoung, in ibid, 1991, pp. 139–175. S.-F. Zou, H. Men, P. Wang, Chin. J. Sens. Actuat. 17 (2) (2004) 336. X. Xie, D. Stueben, Z. Berner, Anal. Lett. 38 (14) (2005) 2281. J. Wang, P. Tuzhi, Anal. Chim. Acta 197 (1987) 367. D.K.Y. Wong, A.G. Ewing, Anal. Chem. 62 (1990) 2697. M. Wojciechowski, J. Balcerzak, Anal. Chim. Acta 249 (1991) 433. L. Nyholm, G. Wikmark, Anal. Chim. Acta 257 (1992) 7. S. Hocevar, B. Ogorevc, J. Wang, B. Pihlar, Electroanalysis 14 (24) (2002) 1707. M.A. Baldo, S. Daniele, Anal. Lett. 37 (5) (2004) 995. E.A. Hutton, S.B. Hocevar, B. Ogorevc, Anal. Chim. Acta 537 (2005) 285. L. Baldrianova, I. Svancara, A. Economou, S. Sotiropoulos, Anal. Chim. Acta 580 (2006) 24. J. Wang, L.D. Hutchins, Bioelectrochem. Bioenerg. 12 (1984) 193. J. Wang, X. Zhang, M. Prakash, Anal. Chim. Acta 11 (1999) 11. Y. Zou, J. Mo, Anal. Chim. Acta 382 (1999) 145. J. Calvo Quintana, L. Idrissi, G. Palleschi, P. Albertano, A. Amine, M. El Rhazi, D. Moscone, Talanta 63 (2004) 567. M.V. Stackelberg, M. Pilgram, V. Toome, Z. Electrochem. 57 (1953) 342. K. Aoki, K. Akimoto, K. Tokuda, H. Matsuda, J. Osteryoung, J. Electroanal. Chem. 171 (1984) 219. B. Lovrecek, I. Mekjavic, M. Metikos-Hukovic, in: A.J. Bard, R. Parsons, J. Jordan (Eds.), Standard Potentials in Aqueous Solution, Marcel Dekker, Inc., NY and Basel, 1985, pp. 180–187. I. Svancara, M. Pravda, M. Hvizdalova, K. Vytras, Electroanalysis 6 (1994) 663.