Anal. Chem. 2004, 76, 1458-1465
Spectroelectrochemical Sensing Based on Attenuated Total Internal Reflectance Stripping Voltammetry. 2. Determination of Mercury and Lead Anne T. Maghasi,† Sean D. Conklin,† Tanya Shtoyko,† Aigars Piruska,† John N. Richardson,‡ Carl J. Seliskar,*,† and William R. Heineman*,†
Department of Chemistry, University of Cincinnati, P.O. Box 210172, Cincinnati, Ohio 45221-0172, and Department of Chemistry, Shippensburg University, 1871 Old Main Drive, Shippensburg, Pennsylvania 17257
Detection of lead and mercury by attenuated total internal reflectance spectroscopy coupled to stripping voltammetry is demonstrated. Changes in attenuation of light passing through an indium tin oxide optically transparent electrode (ITO-OTE) accompany the electrodeposition and stripping of lead and mercury on the electrode surface. The change in absorbance during stripping of electrodeposited metal constitutes the analytical response that enables detection over a range of 2.5 × 10-7-5 × 10-5 and 5 × 10-8-5 × 10-5 M for mercury and lead, respectively. The spectroelectrochemical responses of mercury and lead on the ITO surface are characterized and optimized with respect to solution conditions, the potential excitation signals used for deposition and stripping, and wavelength for detection. The deposited metals were examined by environmental scanning electron miscroscopy, and the electrodeposition pattern of lead and mercury was found to influence the optical response. Environmental monitoring of heavy metals has always relied on highly sensitive detection methods such as atomic absorption spectrometry,1 inductively coupled plasma atomic emission spectrometry,2 and anodic stripping voltammtery (ASV).3 Of these methods, ASV is widely used because it is simple, fast, and inexpensive. Detection limits as low as 10-10 M have been achieved for many trace metals using ASV.4,5 Mercury film and hanging mercury drop electrodes are routinely used as working electrodes in ASV.6 This is because * To whom correspondence should be addressed. E-mail: William.
[email protected]. Tel: (513) 556-9210. Fax: (513) 556-9239. E-mail: carl.seliskar@ uc.edu. Tel: (513) 556-9213. Fax: (513) 556-9239. † University of Cincinnati. ‡ Shippensburg University. (1) Bately, G. E.; Matousek, J. Anal. Chem. 1977, 49, 2031. (2) Habib, M.; Salin, E. Anal. Chem. 1985, 57, 2055. (3) Wang, J. In Laboratory Techniques in Electroanalytical Chemistry, 2nd ed.; Kissinger, P. T., Heineman, W. R., Eds.; Marcel Dekker: New York, 1996; pp 719-737. (4) Barendrecht, E. In Electroanalytical Chemistry; Bard, A. J., Ed.; Marcel Dekker: New York, 1967; Vol. 2. (5) Wang, J. Stripping Analysis: Principles, Instrumentation, and Applications; VCH Publishers: Deerfield Beach, FL, 1985. (6) Economou, A.; Fielden, P. R. Analyst 2003, 128, 205-212.
1458 Analytical Chemistry, Vol. 76, No. 5, March 1, 2004
mercury has a high overvoltage for hydrogen and also forms metal amalgams that help preconcentrate the analyte metal at the electrode and give well-defined stripping peaks, hence lowering the detection limits. However, the toxicity of mercury limits its use, and this has prompted the exploration of alternative electrode materials. This is even more significant for applications where mercury itself is an analyte of interest. To address this limitation of mercury electrodes, ASV has been applied to solid electrodes such as glassy carbon.7 Preconcentrating trace metals on solid electrodes and then using spectroscopic techniques such as X-ray spectroscopy 8 and neutron activation analysis9 to measure the analytical signal have also been explored. It was found that mercury-coated platinum optically transparent electrodes could be used to optically monitor the deposition and stripping of heavy metals.10-12 Although this opened up a new approach to monitoring trace metals spectroscopically, the limitations of using mercury still persist, and it is desirable to employ an OTE without mercury. Our group has explored the use of indium tin oxide optically transparent electrodes (ITO-OTE) as substrates for electrodeposition of trace metals on the bare electrode with simultaneous monitoring of the process spectroscopically.13 This idea was based on the fact that ITO-OTEs provide an exceptionally wide potential window (-1.0 to +1.4 V) compared to many solid electrodes and high transparency in the visible wavelength range. Our initial work evaluated the use of attenuated total internal reflection with stripping voltammetry (ATR-SV) to detect lead and cadmium. In ATR-SV, the optical properties of thin metal films dictate the sensitivity of the sensor. For any given metal deposited on an electrode surface, the optical response depends on the intrinsic optical properties and morphology of deposited metal with regard (7) Heineman, W. R.; Mark, H. B. J.; Wise, J. A.; Roston, D. A. In Laboratory Techniques in Electroanalytical Chemistry; Kissinger, P. T., Heineman, W. R., Eds.; Marcel Dekker: New York, 1984; pp 499-538. (8) Vassos, B.; Hirsch, R.; Letterman, H. Anal. Chem. 1973, 45, 792. (9) Mark, H. B., Jr. J. Pharm. Belg. 1970, 25, 367. (10) Heineman, W. R.; Kuwana, T. Anal. Chem. 1971, 43, 1075-1078. (11) Heineman, W. R.; Kuwana, T. Anal. Chem. 1972, 44, 1972-1978. (12) Goelz, J. F.; Yacynych, A. M.; Mark, H., Jr.; Heineman, W. R. J. Electroanal. Chem. 1979, 103, 277-280. (13) Shtoyko, T.; Maghasi, A. T.; Richardson, J. N.; Seliskar, C. J.; Heineman, W. R. Anal. Chem. 2003, 75, 4585-4590. 10.1021/ac034830h CCC: $27.50
© 2004 American Chemical Society Published on Web 01/31/2004
to reflection, scattering, transmittance or absorbance of light.14-16 In the visible range, metals typically have high extinction coefficients17 resulting in significant attenuation of light intensity as the metal deposits. A critical factor that contributes to the optical response in ATRSV is the structure of metal films on the electrode surface. Although metal films can be formed readily on the electrode surface, their structures are not easy to control. The structure of electrodeposited films depends on the nature of the electrode surface as well as the electrolyte solution, with the former being the most critical, yet unpredictable. Many metal films form by initial nucleation at sites of surface defects and continue to grow unevenly until the surface is completely covered. The nucleation and growth of these films are also significantly influenced by the nature of interaction between the depositing metal with the electrode surface as well as the presence of other competing reactions during the electrodeposition process.18,19 A thorough understanding of the intrinsic properties of the metals as well as their electrodeposition profiles is therefore important in evaluating the response in an ATR-SV sensor. In this paper, ATR-SV has been evaluated as a detection method for sensing mercury and lead. The contributions of the electrodeposition pattern to the ATR-SV signal have been investigated by studying the nucleation and growth profiles of the analyte species on an ITO surface. We also demonstrate that ATRSV can be used in mutlianalyte sensing by simultaneously detecting both lead and mercury on the same ITO-OTE. In a companion paper, we report results from a similar study on cadmium and copper.20 EXPERIMENTAL SECTION Reagents. Lead nitrate, sodium acetate, acetic acid, and potassium nitrate were purchased from Fisher Scientific (Fair Lawn, NJ). Mercury acetate was from J.T. Baker (Phillipsburg, NJ). All solutions were prepared with deionized water from a D2798 Nanopure water purification system (Barnstead, Boston, MA). Apparatus. ITO-coated glass (150-nm ITO over tin float or 1737F glass) was purchased from Thin Film Devices (Anaheim, CA) and cut into 1 in. × 3 in. slides. An A-2 Harrick plasma cleaner (Harrick Scientific Corp., Ossining, NY) was used to clean the ITO slides. Spectroelectrochemical measurements were performed in a plastic cell described previously.21 ITO served as the working electrode while Ag/AgCl (3 M KCl) and platinum wire mesh were used as the reference and auxiliary electrodes, respectively. Electrical contacts with the ITO were made with four copper clips attached to each corner. A Barnant 7520-35 (Barrington, IL) (14) Born, M.; Wolf, E. Principles of Optics, 6th ed.; Pergamon Press: New York, 1980. (15) Paquin, R. A. Properties of Metals; McGraw-Hill: New York, 1995. (16) Heavens, O. S. Optical Properties of Thin Films; Dover Publications: New York, 1991. (17) Bass, M., Ed. Handbook of Optics, 2nd ed.; McGraw-Hill: New York, 1995. (18) Lee, J.; Varela, H.; Uhm, S.; Tak, Y. Electrochem. Commun. 2000, 2, 646652. (19) Serruya, A.; Mostany, J.; Scharifker, B. R. J. Electroanal. Chem. 1999, 464, 39-47. (20) Shtoyko, T.; Conklin, S. D.; Maghasi, A. T.; Piruska, A.; Richardson, J. N.; Seliskar, C. J.; Heineman, W. R. Anal. Chem. 2004, 76, 1466-1473. (21) Shi, Y.; Slaterback, A. F.; Seliskar, C. J.; Heineman, W. R. Anal. Chem. 1997, 69, 3679-3686.
peristaltic pump was used to circulate analyte solution through the ATR cell. For spectroelectrochemical experiments, light from a Xe arc lamp (ILC Technology) was directed through a monochromator to obtain the selected wavelength. This light was then coupled to a 600-µm silica step-index optical fiber (Fiberguide Industries, Stirling, NJ) with a microscope objective placed at the exit slit of the monochromator. The light was transmitted to the ATR cell21 via a Schott SF6 optical prism (Karl Lambrecht, Chicago, IL) coupled to the ITO slide using a high refractive index (ηD ) 1.51) fluid (Cargille Laboratories, Cedar Grove, NJ). After propagating through the ATR cell, the light was then coupled to another SF6 optical prism and focused onto a photodiode detector (Photonic Detectors, Simi Valley, CA). A homemade potentiostat was used for electrochemical control. The current and optical responses were acquired using locally written software, and the data were exported to commercial spreadsheets for analysis. For absorbance spectrum measurements of electrodeposited metals on ITO slides, the monochromator was removed from the setup. Light exiting the ATR cell was then detected using an Ocean Optics PC2000 multichannel CCD spectrophotometer (Ocean Optics, Dunedin, FL). A Philips XL30 environmental scanning electron microscope (ESEM) equipped with an energydipersive X-ray detector (Peabody, MA) was used for all imaging measurements on the ITO slides. Procedures for Spectroelectrochemical Measurements. Prior to use, ITO slides were washed with Alconox and rinsed thoroughly with deionized water and ethanol. The slides were then dried in air before being Ar plasma cleaned for 30 min. Unless otherwise specified, aqueous 0.1 M KNO3 served as the supporting electrolyte and its spectroelectrochemical response was always recorded before each experiment. While circulating the analyte solution through the ATR cell, the potential was switched from open circuit to 0.8 V for ∼1 min to obtain the baseline absorbance. The potential was then stepped to the deposition potential (-0.6 V for Hg2+ and -0.9 V for Pb2+) for 30 min. After preconcentration, the potential was stepped to the oxidation potential (-0.1 V for Pb2+ and 0.8 V for Hg2+) to strip off the electrodeposited metal. The optical signal was continuously recorded throughout the potential changes. RESULTS AND DISCUSSION Spectroelectrochemical Characterization of Lead and Mercury at a Bare ITO-OTE. The general concept of the ATRSV sensor has been described previously.13 Before making any ATR-SV measurements, the electrochemical behavior and optical response for Hg2+ and Pb2+ solutions on a bare ITO-OTE were investigated using cyclic voltammetry. Figure 1A shows the cyclic voltammogram of 1 mM Pb(NO3)2 in 0.1 M acetic acid/sodium acetate buffer (pH 5.4) (curve 1) and the corresponding optical response measured at 750 nm (curve 2). The voltammogram displays well-defined deposition and stripping peaks at -0.7 and -0.35 V, respectively. The concomitant optical absorbance increases upon initiation of deposition and peaks at the onset of the stripping process. The absorbance then returns to zero concurrently with the decay of the stripping peak. A similar experiment was conducted using 1 mM Hg(OAc)2 in 0.1 acetic acid/sodium acetate buffer (pH 5.4) and is shown in panel B. In this case, the voltammetric response (curve 1) was Analytical Chemistry, Vol. 76, No. 5, March 1, 2004
1459
Figure 1. Cyclic voltammograms and the optical responses of 1 mM Pb(NO3)2 in 0.1 M acetic acid/sodium acetate buffer (pH 5.4) (A), 1 mM Hg(OAc)2 in 0.1 M acetic acid/sodium acetate buffer (pH 5.4) (B), 1 mM Pb(NO3)2 in 0.1 M KNO3 (C), and 1 mM Hg(OAc)2 in 0.1 M KNO3 (D). The voltammograms were recorded while cycling the potential at 5 mV/s (A1, B1, C1, D1) and the optical responses measured at 750 nm for 1 mM Pb(NO3)2 (A2, C2) and at 650 nm for 1 mM Hg(OAc)2 (B2, D2) were recorded concurrently.
less than ideal when compared to that of Pb. Here, the deposition peak is broad and poorly defined and is accompanied by a pair of stripping peaks at 0.45 and 0.55 V. It is interesting to note that multiple stripping peaks are only observed at higher concentrations, thus suggesting that the peak at 0.45 V is due to stripping of Hg from Hg droplets, while the peak at 0.55 V is due to stripping of Hg from ITO. Furthermore, the measured absorbance (curve 2) increases with the onset of Hg deposition but does not decay back to zero absorbance during stripping over the time scale of the experiment. In a remote sensing application, buffering of samples might not be possible. Therefore, we chose to repeat the above experiments, but this time in the absence of buffering. Here, both Pb and Hg were examined in unbuffered 0.1 M KNO3 electrolyte solution. The reduction of Pb2+ (Figure 1, panel C) in unbuffered solution occurs stepwise with at least four discernible waves. A gradual increase in current begins at ∼-0.2 V followed by a distinct shoulder at ∼-0.5 V, another shoulder at -0.6 V, and then a sharp peak at ∼-0.7 V. Careful correlation of the optical response with the current response shows that the absorbance increases at a rate that roughly correlates with the current. Switching the direction of scan at -0.8 V resulted in a continued increase in absorbance until a small anodic stripping peak occurs at ∼-0.5 V, at which the absorbance decreases rapidly and then more slowly during the positive scan. Continuing the scan more positively than 0.4 V or stopping the scan at 0.4 V results in a continual decrease in absorbance back to the baseline of zero (not shown). This illustrates the complexity that occurs in a neutral solution in which various hydroxyl complexes of Pb2+ can exist depending on pH: Pb(OH)1+, Pb(OH)20, Pb(OH)31-, and Pb(OH)42-. All of these species have slightly different deposition potentials. Also, the current and absorbance signals during the stripping step do not correlate well, which is very different from the acidic solution experiment in Figure 1A. The absorbance continues to 1460 Analytical Chemistry, Vol. 76, No. 5, March 1, 2004
decrease even though there is no corresponding anodic current for this oxidation process. We speculate that this might be due to the presence of oxygen in the solution, which makes stripping of the Pb film by oxidation by dissolved oxygen possible. In this case, one would see the film come off spectroscopically, but without the associated anodic current. For deposition and stripping of mercury in 0.1 M KNO3, the voltammetric and optical behavior is similar to that obtained in buffered solution; an interesting difference, however, is the fact that the optical absorbance is considerably larger in the unbuffered solution. These results are summarized in Figure 1, panel D. Also, in neither case does the absorbance return to the initial baseline. We suggest for deposition and stripping of Hg this observation is due to formation of an amalgam between Hg and either indium or tin in the ITO layer of the OTE. Optical absorbances taken at long times after application of the stripping potential indicate that the Hg is eventually stripped from the ITO surface (vide infra). We emphasize that, in the absence of buffering and deoxygenation, substantial optical absorbance changes concurrent with deposition and subsequent stripping are still observed. While deoxygenation is easily accomplished at the laboratory bench, it is difficult to implement in remote sensing applications. This is especially noteworthy in the case of Pb analysis, as shown in Figure 1, panel C. Even though the anodic stripping peak is small and poorly defined compared to that obtained in a deoxygenated solution (results not shown), a large optical absorbance change is still noted. These results punctuate the superiority of the use of the optical signal for analysis of certain metals using stripping methods. To determine the optimum wavelength for absorbance measurements for mercury, spectra were recorded while depositing mercury for a period of 30 min. The ITO electrode was held at -0.6 V (vs Ag/AgCl) while a spectrum of mercury was recorded every 2 min; the results are shown in Figure 2A. From this figure it is clear that while there is a general increase in sensor absorbance at all wavelengths, the increase is greater at longer wavelengths. Experimentally obtained spectra were then compared with a theoretical spectrum based on a simple model. In this optical model, reflectance from a multilayer stack of films was calculated using matrix methods. The assumed model consisted of glass (1737F), ITO, mercury, and solution layers. Published optical constants of mercury, water, and glass were used.22 The optical constants of the ITO layer were used as determined by spectroscopic ellipsometry. Two boundary conditions were applied: one in which no mercury was deposited and another with a 5-nmthick isotropic continuous layer of mercury. Separate experiments using chronocoulometry were done to determine the thickness of the mercury layer whose spectrum is shown in Figure 2A, and these results revealed that the layer was ∼5 nm thick. To obtain theoretical absorbance, reflectance for reference (d(Hg) ) 0 nm) and sample (d(Hg) ) 5 nm) were calculated. An absorbance spectrum was then generated (Figure 2B) based on the reflectance values at the set boundary conditions. Comparison of the calculated spectrum with that obtained experimentally shows the similarity in the absorbance pattern. In (22) Palik, E. D., Ed. Handbook of Optical Constants of Solids; Academic Press: San Diego, 1998.
Figure 3. (A) ATR-SV response curves for Hg(OAc)2 in 0.1 M KNO3 deposited at -0.6 V for 30 min and stripped at 0.8 V while circulating the solutions at 50 mL/min through the ATR cell. The curves shown are for (1) 50 nM, (2) 250 nM, (3) 500 nM, (4) 2.5 µM, (5) 5 µM, (6) 25 µM, and (7) 50 µM. All absorbance signals were measured at 650 nm. (B) Calibration curve for Hg using changes in absorbance shown in (A). The inset represents the linear portion of the calibration curve for concentrations ranging from 50 nM to 2.5 µM.
Figure 2. (A) Series of absorbance spectra for 1 mM Hg(OAc)2 in 0.1 M KNO3 deposited on ITO for 2-30 min. The potential at the electrode was held at -0.6 V (vs Ag/AgCl) as Hg2+ was deposited. (B) Theoretical calculations for wavelength dependence using optical constants of Hg found in the literature. Hg layer thickness on ITO assumed to be 5 nm, and absorbance calculated at 65° relative to the plane of ITO.
both cases, there is a steady increase in absorbance with increasing wavelength. These results compare favorably despite the fact that the theoretical model is based on some very simplistic assumptions. The model assumes that the mercury layer is uniform, but microscopic images of the surface (to be discussed later) reveal that the deposition of mercury on the ITO surface is not at all uniform. The model also ignores the effects of scattering from deposited mercury droplets, which certainly plays a role in the sensor absorbance observed experimentally. However, the good agreement between the results obtained suggests that these factors do not contribute significantly to the overall absorbance of the sensor. The wavelength dependence of the absorbance of lead was characterized previously,13 and the optimum wavelength was selected at 750 nm. Calculation of the theoretical absorbance spectrum was not possible because the optical constants for lead were not available.
Evaluation of ATR-SV Sensor for Hg2+ and Pb2+. The ATRSV response curves for Hg2+ on a bare ITO electrode are shown in Figure 3A. The absorbance versus time plots are for concentrations ranging from 5 × 10-8 to 5 × 10-5 M with the corresponding calibration curve (Figure 3B) generated with absorbance values taken after 30-min deposition. Each ATR-SV response curve was recorded by holding the electrode potential at -0.6 V for 30 min while circulating the solution through the flow cell at 50 mL/ min. The potential was then stepped to 0.8 V to strip off mercury from the electrode surface. A background response was obtained by modulating the potential for the supporting electrolyte solution, and this was subtracted from each of the responses for the Hg2+ (in 0.1 M KNO3) solutions. The sensor absorbance representing each point on the calibration curve was generated by subtracting the baseline absorbance after the stripping step from that attained after 30 min of deposition for each concentration. The stripping pattern for Hg on the ITO surface is different from that observed for Pb2+ and Cd2+ on a similar ITO electrode.13,21 After 30 min of deposition, the step to an oxidative potential did not result in an immediate return of the optical signal to baseline. Instead, there is a sharp initial decrease followed by a slow change in absorbance before eventually returning to baseline 5-20 min later depending on Hg2+ concentration. This slow removal from the ITO electrode is unique to Hg and has not been observed for stripping of other metals such as Pb, Cu, and Cd13,20 from this same electrode surface. This suggests that there is a unique interaction between mercury and the ITO Analytical Chemistry, Vol. 76, No. 5, March 1, 2004
1461
electrode surface, and this significantly affects the stripping profile of mercury from this type of electrode surface. Mercury is unique in that it can form amalgams with many metallic species, and interactions between mercury and the indium or tin on the surface could be contributing to the delay in removal of some mercury from the surface, as was described earlier in discussion of Figure 1, panels B and D. The ATR-SV calibration curve for lead was reported previously,13 and the linear range was from 5 × 10-8 to 2.5 × 10-6 M. In comparison with the data reported here for mercury, the sensor absorbance for lead is higher in magnitude for similar concentrations, and this may be attributed to the differences in microscopic electrodeposition patterns for the two metals, as these differences may contribute to the attenuation of light in ATR-SV. ESEM Surface Characterization of Hg and Pb Deposition on ITO. Sensing of heavy metals by ATR-SV relies on the attenuation of light on the ITO electrode surface due to electrodeposition of metals. The extent of attenuation depends not only on the concentration of metal ions and deposition time but also on the nature of the deposition pattern the metal assumes on the surface. It is therefore important to study the microscopic electrodeposition patterns of the various metal ions on the surface and correlate them to their contributions to the ATR-SV signal observed. A comparison of the metal film structure on the electrode may also aid in understanding the differences in the magnitude of the ATR-SV signal for similar concentrations of various metallic species. ESEM images of the ITO electrode surface as Hg is deposited are shown in Figure 4. Hg was deposited at -0.6 V from a 5 × 10-5 M mercury solution, and then ESEM images of the ITO electrode were taken at different times. Panels A-C of Figure 4 show the progression of mercury deposition on the ITO surface at 2, 10, and 30 min, respectively. It is clear from these images that the Hg features increase in size with longer deposition time without coalescence on the surface even after 30 min of deposition. The kinetics of mercury nucleation have been studied on various electrode surfaces.19,23,24 It is clear from these studies that the nucleation rate and density of nucleation sites depend on concentration and deposition potential. These studies also show that the reduction of Hg2+ occurs as a two-electron process with Hg22+ as an intermediate in the following manner:
2Hg2+ + 2e- f Hg22+ Hg22+ + 2e- f 2Hg It is also possible that, as mercury droplets form on the electrode, a disproportionation reaction that involves the deposited mercury occurs to generate Hg22+ via the following reaction:
Hg2+ + Hg h Hg22+ The generation of Hg22+ ions near the electrode surface contributes to supersaturation and increases the number density of nuclei (23) Martins, M. E.; Salvarezza, R. C.; Arvia, A. J. Electrochim. Acta 1996, 41, 2441-2449. (24) Salinas, D. R.; Cobo, E. O.; Garcia, S. G.; Bessone, J. B. J. Electroanal. Chem. 1999, 470, 120-125.
1462 Analytical Chemistry, Vol. 76, No. 5, March 1, 2004
Figure 4. Surface characterization of ITO by ESEM as Hg is deposited. Images of Hg deposited for (A) 2, (B) 10, and (C) 30 min. All images were taken at 5000× magnification. Deposition of 5 × 10-5 M Hg(OAc)2 on ITO was done at -0.6 V (vs Ag/AgCl) while circulating the solution at 50 mL/min through the ATR cell.
while inhibiting the electrodeposition of mercury in their vicinity. The contributions from Hg22+ during the nucleation of Hg therefore enhance the spatial ordering of nuclei on the electrode surface and explain why the deposition pattern for mercury is different from that observed for other metal ions. A similar study of the nucleation profile for Pb deposition on the ITO electrode surface is shown in Figure 5. Pb was deposited at -0.9 V from a 5 × 10-5 M lead solution for varying periods of time, and each electrode was then imaged by ESEM. Panels A-C of Figure 5 show the progression of lead electrodeposition on the ITO surface at 2, 10, and 30 min, respectively. These images show a general increase in the coverage of Pb on the surface with longer
Figure 6. (1) Potential excitation signal. (2) ATR-SV of Pb and Hg, co-deposited on ITO with corresponding potential steps. A mixture of 50 µM Pb2+ and 50 µM Hg2+ each in 0.1 M KNO3 were circulated at 50 mL/min through the ATR cell.
Figure 5. Surface characterization of ITO by ESEM during Pb deposition. Images of Pb deposited for (A) 2, (B) 10, and (C) 30 min. All images were taken at 5000× magnification. Deposition of 5 × 10-5 M Pb(NO3)2 on ITO was done at -0.9 V (vs Ag/AgCl) while circulating the solution at 50 mL/min through the ATR cell.
deposition time, and the Pb seems to condense into more dense structures with subsequent deposition. In comparison with the deposition pattern of Hg in Figure 4, Pb deposition seems to be more straightforward in that longer deposition times lead to a higher density of particles on the electrode. Since the attenuation of the optical signal in ATR-SV depends on the surface coverage of the electrode, among other factors, the ESEM data help to explain how the deposition pattern leads to differences in signal attenuation even for the same concentration of both metals (Pb ATR-SV data shown in ref 13). In general, the electrodeposition of each metal seems to be
affected by the metal-substrate interactions as well as the occurrence of any side reactions as in the case of mercury. Simultaneous Detection of Pb2+ and Hg2+ by ATR-SV. To evaluate the usefulness of ATR-SV for multianalyte detection, a solution containing both Pb and Hg was analyzed. The ATR-SV data obtained are shown in Figure 6 along with the potential steps applied. Initially the potential was held at 0.8 V, which was positive enough to keep both ions in the oxidized form, and then the potential was stepped to -0.6 V, which was sufficiently negative to reduce Hg2+ to Hg while Pb2+ remained in the oxidized form until the potential was switched to -0.9 V. A reversal of this process occurred as the potential was stepped to -0.1 V at which Pb2+ stripped off the electrode. However, this potential was still negative enough that Hg continued to deposit on the electrode until the potential was stepped to 0.8 V. This ATR-SV profile shows that it is possible to selectively detect two metals on a single ITO electrode by controlling the deposition and stripping potentials. In multianalyte sensing by ATR-SV, electrodeposition of all analytes on the electrode surface determines the extent of attenuation of light as it passes through the sensor. It is therefore important to determine the effect of depositing an additional metal in the presence of another. For lead and mercury, the two ions were electrodeposited on the same ITO surface and imaged by ESEM to determine the distribution of species on the surface. To discern the composition of the features on the electrode surface, X-ray mapping was done, and the results are shown along with the ESEM image for direct comparison in Figure 7. The position of Pb features on the ITO surface (Figure 7B) can be compared with that of Hg (Figure 7C) as well as In (Figure 7D) from the ITO, which is also shown. Co-deposition of the two metals on the surface seems to occur in a random manner, probably initiated at sites where defects exist on the ITO electrode. Since mercury forms amalgams with lead, the sites of lead deposition correlate with the placement of mercury on the surface. It is also worth noting that lead does not preferentially deposit where mercury is accumulated on the surface but also extends to other sites not occupied by mercury. The lead that deposits at sites without mercury seems to take on the branchlike structures observed on ITO surfaces with only lead deposited. All these Analytical Chemistry, Vol. 76, No. 5, March 1, 2004
1463
Figure 7. Surface characterization of modified ITO by ESEM and X-ray mapping of Pb and Hg co-deposited on the surface. A deposition potential of -0.9 V (vs Ag/AgCl) was applied for 10 min to the ITO while circulating a mixture of 500 µM Pb2+ and Hg2+ each at 50 mL/min through the ATR cell. All images were taken at 5000× magnification. (A) SEM image. (B) Pb MR line (2.34 keV). (C) Hg MR (2.20 keV). (D) In LR (3.28 keV).
features suggest that while the presence of mercury enhances lead deposition through amalgam formation, it does not preclude it from depositing independently on the ITO surface. CONCLUSIONS Spectroelectrochemical detection of lead and mercury by ATRSV has been demonstrated. The sensor absorbance was found to be wavelength dependent, and a comparison of the spectrum for Hg with theoretical calculations was made. At the optimum wavelength, a detection limit of 2.5 × 10-7 for mercury was achieved on a bare ITO electrode. Further improvements to this detection limit should be achievable by incorporation of chemically selective films on ITO, using longer deposition times and signal averaging. Using a waveguide as the OTE would increase the number of reflections and also lower the detection limits for the ATR-SV sensor. A possible advantage of the optical measurement versus traditional current measurement is a better signal without deoxygenation. Surface characterization of the ITO electrode with lead and mercury electrodeposited was also done by ESEM. The results 1464 Analytical Chemistry, Vol. 76, No. 5, March 1, 2004
provided insight into the nucleation and deposition profile of the metals on the surface and their contribution to the optical signal measured. The density of lead particles on the surface was significantly different from the spatial distribution of mercury on the same surface, and this was probably due to the different mechanisms involved during the reduction of each. Lead reduction occurs via a single two-electron process while reduction of Hg2+ involves the formation of Hg22+ either as an intermediate or as a disproportionation product. Branched or netted structures such as that obtained with Pb, as opposed to discrete crystallites or droplets, as those obtained with Hg, apparently contribute to the uniformity of the electrodeposited metal films and to the optical response obtained with ATR. The ability for mutliananlyte sensing was also demonstrated by detection of both species simultaneously on a bare ITO electrode. Although not shown here, the concept of ATR-SV should be amenable to sensing numerous analytes simultaneously, the only limitation being that the stripping potential of each analyte be significantly unique. This sensor is therefore potentially useful
for applications where mixtures of analytes exist such as heavy metals in environmental samples.
Phillips XL-30 ESEM was purchased with a grant from the Hayes Fund of the State of Ohio.
ACKNOWLEDGMENT Financial support was provided by the Environmental Management Science program of the U.S. Department of Energy, Office of Environmental Management (DE-FG07-99ER62311-70010). The
Received for review July 21, 2003. Accepted December 2, 2003. AC034830H
Analytical Chemistry, Vol. 76, No. 5, March 1, 2004
1465
