Anal. Chem. 2006, 78, 3745-3751
Factors Influencing Redox Magnetohydrodynamic-Induced Convection for Enhancement of Stripping Analysis Emily C. Anderson and Ingrid Fritsch*
Department of Chemistry and Biochemistry, University of Arkansas, Fayetteville, Arkansas 72701
Factors affecting the use of redox magnetohydrodynamics (MHD) to enhance the stripping analysis response to heavy metals have been investigated. The analytes were Pb2+, Cd2+, Cu2+, and Tl+ at concentrations ranging from 5 nM to 2 µM. Co-deposition of analytes with Hg2+ (to form a thin Hg film electrode) occurs along with reduction of a high concentration of Fe3+. The Fe3+ provides the high cathodic current necessary to produce a significant Lorentz force, and therefore enhanced convection and larger stripping signals and sensitivities, when the analysis is performed in the presence of an external magnetic field. The effects of varying Fe3+ concentration (1-100 mM), working electrode size (10 µm-3 mm), and magnetic field strengths (0-1.77 T) generated with electromagnets and NdFeB permanent magnets were investigated. Using 100 mM Fe3+ as the MHD-generating redox species at a 3-mm working electrode and in a magnetic field of 1.77 T, peak areas from linear sweep voltammetry were increased by as much as 159 ( 5%, compared to the signal obtained in the absence of a magnetic field. Experimental detection limits as low as 5 nM were achieved with only a 1-min preconcentration time. A field strength as low as 0.12 T offers some signal enhancement with 100 mM Fe3+. While linear scan anodic stripping voltammetry was used primarily to obtain the signals after the deposition step, potentiometric stripping analysis was also investigated. Redox MHD is an attractive alternative convection method for applications involving sample volumes too small for mechanical stirring or for in-field applications using portable devices that cannot be complicated by the instrumentation required for mechanical stirring. We have previously reported the use of high concentrations of Hg2+ to enhance anodic stripping voltammetry (ASV) by redox magnetohydrodynamics (MHD).1 The high concentrations of Hg2+ are undesirable due to the toxicity of Hg and the several heavy metal contaminants present in Hg(NO3)2 that lead to a nonzero background signal for those metals. The analyses reported here use high concentrations of Fe3+, instead, to generate a Lorentz force, and low concentrations (∼1 mM) of Hg2+ to co-deposit the Hg film electrode2 with the analytes. We have also determined * To whom correspondence should be addressed. Tel: (479) 575-6499. Fax: (479) 575-4049. E-mail:
[email protected]. (1) Clark, E. A.; Fritsch, I. Anal. Chem. 2004, 76, 2415-2418. 10.1021/ac060001v CCC: $33.50 Published on Web 05/03/2006
© 2006 American Chemical Society
the detection limits of different analytes and explored the possibility of using redox MHD in portable devices for the analysis of small volumes. Both MHD-enhanced linear scan ASV and potentiometric stripping analysis (PSA) were investigated. ASV is commonly used to analyze metals (e.g., lead, cadmium, copper, thallium, zinc, indium, etc.) in environmental samples, such as sediments3,4 and natural waters,5-9 and in biological samples, such as blood10-13 and urine.14-16 It has also been recently applied to the detection of DNA hybridization.17-20 In ASV, a deposition (preconcentration) step reduces chemical species of interest at an electrode and a second step applies a potential function that reoxidizes (or strips) the species.21 In linear scan ASV, the potential function is a simple linear sweep of the potential with time. The magnitude of the resulting anodic peak current is proportional to the amount of analyte reduced in the preconcentration step and, therefore, is proportional to the concentration of analyte in the sample.5,21,22 In potentiometric stripping analysis,23,24 the metals are oxidized out of the mercury thin film by a chemical oxidant present in solution, and the potential is measured as a function of time. The (2) Florence, T. M. J. Electroanal. Chem. 1970, 27, 273-281. (3) Olsen, K. B.; Wang, J.; Setladji, R.; Lu, J. Environ. Sci. Technol. 1994, 28 (12), 2074-2079. (4) da Silva, C. L.; Masini, J. C. Fresenius Z. Anal. Chem. 2000, 367, 284-290. (5) Petrie, L. M.; Baier, R. W. Anal. Chem. 1978, 50, 351-357. (6) Emons, H.; Baade, A.; Schoning, M. J. Electroanalysis 2000, 12, 11711176. (7) Davison, W. J. Electroanal. Chem. 1978, 87, 395-404. (8) Mart, L.; Nurnberg, H. W.; Valenta, P. Fresenius Z. Anal. Chem. 1980, 300, 350-362. (9) Richter, P.; Toral, M. I.; Abbott, B. Electroanalysis 2002, 14 (18), 12881293. (10) Prakash, R.; Srivastava, R. C.; Seth, P. K. Electroanalysis 2001, 14 (4), 303308. (11) Hardcastle, J. L.; Compton, R. G. Electroanalysis 2002, 14 (11), 753-759. (12) Liu, T. Z.; Lai, D.; Osterloh, J. D. Anal. Chem. 1997, 69, 3539-3543. (13) Kruusma, J.; Nei, L.; Hardcastle, J. L.; Compton, R. G.; Lust, E.; Keis, H. Electroanalysis 2004, 16 (5), 399-403. (14) Horng, C.-J. Analyst 1996, 121, 1511-1514. (15) Golimowski, J.; Valenta, P.; Stoeppler, M.; Nurnberg, H. W. Talanta 1979, 26, 649-656. (16) Levit, D. L. Anal. Chem. 1973, 45, 1291. (17) Wang, J.; Liu, G.; Polsky, R.; Arben, M. Electrochem. Commun. 2002, 4, 722-726. (18) Wang, J.; Liu, G.; Merkoci, A. J. Am. Chem. Soc. 2003, 125, 3214-3215. (19) Authier, L.; Grossiord, C.; Brossier, P. Anal. Chem. 2001, 73, 4450-4456. (20) Zhu, N.; Zhang, A.; He, P.; Fang, Y. Electroanalysis 2004, 16 (23), 19251930. (21) Bard, A. J.; Faulkner, L. R. Electrochemical Methods: Fundamentals and Applications. 2nd ed.; John Wiley & Sons: New York, 2001. (22) Wang, J. Analytical Electrochemistry, 2nd ed.; Wiley-VCH: New York, 2000.
Analytical Chemistry, Vol. 78, No. 11, June 1, 2006 3745
technique has been used for the analysis of a wide variety of samples, such as blood,25-27 wine,28 and seawater.29 It has also been used as the detection method for immunosensors30 and DNA hybridization assays.31 The change in potential is monitored as a function of time while the reduced forms of the analytes at the electrode are oxidized by species in the surrounding solution. In our studies, Fe3+ and Hg2+, as well as dissolved O2, can act as the chemical oxidants. Because no current is passed during the stripping step, dissolved electroactive species, which would yield overlapping background currents in voltammetry, are not a problem in PSA. PSA also has the advantage of being unaffected by charging currents and highly resistive solutions.32,33 We chose to explore redox MHD with PSA, in addition to the ASV, to demonstrate that the enhancement is general, working for other detection approaches, and because the redox species that are used to produce a Lorentz force are suitable for chemical oxidation of the analytes. Although there are several electrochemical methods to strip metals from the mercury film electrode (e.g., differential pulse voltammetry and square wave voltammetry), comparable enhancements can be expected for all of the methods because the induced convection from redox MHD occurs on the deposition step, not the stripping step. To determine the effectiveness of MHD convection on stripping analysis, we compared results from experiments performed in the absence and presence of a magnetic field. The analytes tested are Pb2+, Cd2+, Cu2+, and Tl+ at concentrations ranging from 5 nM to 2 µM. The co-deposition of analytes with Hg2+ (to form a thin Hg film electrode) occurs along with the reduction of a high concentration of Fe3+, which generates a large cathodic current, J (A/m2), in the presence of an external magnetic field, B (T). The resulting Lorentz force, FL (N/m3), with magnitude and direction governed by the right-hand rule, acts on the chargecarrying ions to induce solution convection.34-36
FL ) J × B The induced convection enhances the mass transport of analytes to the electrode. Therefore, higher quantities of analytes are preconcentrated into the thin-film electrode for a given deposition time, resulting in larger stripping signals (peak areas for linear sweep ASV and transition times for potentiometric (23) Jagner, D.; Graneli, A. Anal. Chim. Acta 1976, 83 (1), 19-26. (24) Jagner, D. Analyst 1982, 107 (1275), 593-599. (25) Jagner, D.; Josefson, M.; Westerlund, S.; Aren, K. Anal. Chem. 1981, 53, 1406-1410. (26) Almestrand, L.; Jagner, D.; Renman, L. Anal. Chim. Acta 1987, 193, 7179. (27) Tutunji, M. F. Fresenius J. Anal. Chem. 1996, 356, 90-95. (28) Anderson, L.; Jagner, D.; Josefson, M. Anal. Chem. 1982, 54, 1371-1376. (29) Riso, R. D.; Corre, P. L.; Chaumery, C. J. Anal. Chim. Acta 1997, 351, 83-89. (30) Wang, J.; Tian, B.; Rogers, K. R. Anal. Chem. 1998, 70, 1682-1685. (31) Wang, J.; Xu, D.; Kawdo, A.-N.; Polsky, R. Anal. Chem. 2001, 73, 55765581. (32) Brainina, K.; Neyman, E. In Electroanalytical Stripping Methods. Winefordner, J. D., Ed.; John Wiley & Sons: New York, 1993; Vol. 126. (33) Wang, J. Stripping Analysis: Principles, Instrumentation, and Applications; VCH Publishers: Deerfield Beach, FL, 1985. (34) Leventis, N.; Gao, X. J. Phys. Chem. B 1999, 103, 5832-5840. (35) Grant, K. M.; Hemmert, J. W.; White, H. S. J. Am. Chem. Soc. 2002, 124, 462-467. (36) Lee, J.; Gao, X.; Hardy, L. D. A.; White, H. S. J. Electrochem. Soc. 1995, 142 (6), L90-L92.
3746
Analytical Chemistry, Vol. 78, No. 11, June 1, 2006
stripping analysis) and therefore improved sensitivities compared to the analyses performed in the absence of a magnetic field. The simplicity of this method of convection makes it an attractive alternative for in-field applications using portable devices, especially in the analysis of small sample volumes where mechanical stirring is not practical. Redox-MHD convection exhibits several advantages over other convection methods. It requires no moving parts (e.g., no stirrer or rotating disk electrode). The voltage requirements remain lows the reducing potential of the current-generating species (Fe3+ in our case) is sufficient. Existing portable devices based on ASV have a number of disadvantages. One commercially available portable analysis system utilizes either a hanging mercury drop electrode (HMDE) or an on-center, glassy carbon rotating disk electrode (RDE) and requires fairly long deposition times, 10-20 min, to achieve limits of detection at the nanomolar level.37 In addition, convection is induced by either mechanical stirring (the HMDE) or rotation (the RDE), neither of which is well-suited for both small volumes and portability. The RDE has been employed on volumes as small as 20 µL (but not for the purpose of stripping voltammetry).38 However, a hydrophobic surface beneath the drop and a spin rate of e3000 rpm were required to keep the microdrop intact. A commercially available iridium microarray, which is designed to maximize diffusion of the analytes to the electrode, achieves comparable limits of detection with only a 2-min preconcentration time, but it is costly ($2000 for the array alone).39 An electrode with an integrated sound generator to enhance mass transport has been reported to achieve comparable detection limits with only a 2-min preconcentration time,40 but the system is complicated by the need for a specialized electrode. In contrast to the specialized electrodes required for most portable systems with low detection limits, the method reported in this paper uses a standard, inexpensive, commercially available or easily homemade electrode that does not need to be on-center. In addition, the time for deposition used in this paper was limited to 1 min. Fundamental studies of MHD have been performed on electrochemical systems with redox species present. Results have demonstrated that as the concentration of the redox species increases, the convection also increases.41,42 However, at very high concentrations, the increased viscosity can offset the increase in Lorentz force and the convection does not improve.42,43 In addition, an increasing magnetic field strength has been shown to improve convection, as expected from theory.36,41,42,44,45 Likewise, the (37) GAT-TEA 4000MP, Bremerhaven, Germany, http://www.gatgmbh.com/ englisch/products/polarography/TEA4000AS_eng.htm, 2004. (38) Wijayawardhana, C. A.; Halsall, H. B.; Heineman, W. R. Anal. Chim. Acta 1999, 399, 3-11. (39) TraceDetect Nano-BandTM Electrodes product specifications. Seattle, WA, http://www.tracedetect.com/specifications/instrument.htm, 2003. (40) Simm, A. O.; Banks, C. E.; Compton, R. G. Anal. Chem. 2004, 76 (17), 5051-5055. (41) Leventis, N.; Chen, M.; Gao, X.; Canalas, M.; Zhang, P. J. Phys. Chem. B 1998, 102, 3512-3522. (42) Ragsdale, S. R.; Grant, K. M.; White, H. S. J. Am. Chem. Soc. 1998, 120, 13461-13468. (43) Arumugam, P. U.; Belle, A. J.; Fritsch, I. IEEE Trans. Magn. 2004, 40 (4), 3063-3065. (44) Lee, J.; Ragsdale, S. R.; Gao, X.; White, H. S. J. Electroanal. Chem. 1997, 422, 169-177. (45) Ragsdale, S. R.; Lee, J.; Gao, X.; White, H. S. J. Phys. Chem. 1996, 100, 5913-5922.
influence of electrode size on the MHD effect has also been investigated.45 Consequently, we expect similar effects to occur in the deposition step of a stripping experiment. This paper reports the impact of similar parameters on redoxMHD enhanced ASV, especially as they pertain to portability. Analysis in large volumes (1 mL) was compared to small volumes (100 µL). Studies of decreasing size of working electrode (3 mm10 µm) reveals the limits of miniaturization. The effect of magnetic field strength (0-1.77 T) was investigated, because although permanent magnets are most desirable over electromagnets in simple portable devices, they have fixed and lower field strengths (∼0.6 T) than electromagnets. In addition, analytes of Pb2+, Cd2+, Cu2+, and Tl+ at varying concentrations in solutions containing 100 mM Fe3+, 1 mM Hg(NO3)2, 30 mM HNO3 and 0.1 M KNO3 were investigated to determine detection limits and sensitivity of redox-MHD enhanced ASV. Finally, a comparison of ASV to PSA was made to demonstrate the use of other stripping methods that can benefit from the use of redox-MHD-enhanced convection. EXPERIMENTAL SECTION Chemicals and Materials. All chemicals were reagent grade and used as received. Aqueous solutions were prepared with highpurity deionized (DI) water (Milli-Q, model RG). Nitric acid, potassium nitrate, and copper(II) nitrate pentahemihydrate (all certified ACS grade) were obtained from Fisher Scientific (Fair Lawn, NJ). Mercury(II) nitrate monohydrate (99.99+%), cadmium nitrate tetrahydrate, ferric nitrate nonahydrate (99.99+%), and thallium(I) nitrate (99.999%) were purchased from Aldrich Chemical Co. (St. Louis, MO). Lead(II) nitrate (99.2%) was acquired from J.T. Baker Chemical Co. (Phillipsburg, NJ). The 3-mm-diameter glassy carbon (GC) working electrode was purchased from CH Instruments, Inc. (Austin, TX). The 1-mm-diameter GC working electrode was constructed with glassy carbon rod (Alfa AESAR, Ward Hill, MA) sealed with heat shrink tubing. The 10-µm GC electrode was purchased from Cypress Systems (Lawrence, KS). For the large-volume experiments, a platinum flag counter electrode was constructed from a Pt sheet (99.99%, 0.1 mm thick, Alfa AESAR) spot welded to a Pt wire, the free end of which was sealed in a glass tube where connection was made to a copper wire. For the small-volume experiments, a Pt wire (99.95%, 0.5mm diameter, Alfa AESAR) was used as the counter electrode, and a 1-mm-diameter Ag/AgCl (saturated KCl) (Microelectrodes, Inc., Bedford, NH) was used as the reference electrode. Electrochemical Measurements. An Electrochemical Workstation model 750A (CH Instruments, Inc.) was controlled by a PC computer running CHI-750A software. A Ag/AgCl (saturated KCl) electrode served as a reference. Before each experiment, the potential at the working electrode was held at 1 V for 60 s in electrolyte (30 mM HNO3 and 0.1 M KNO3) to oxidize mercury remaining from the previous experiment. Then, the working electrode was polished sequentially (using a figure eight motion for ∼2 min for each grit size) with 1-µm diamond, 0.25-µm diamond, (Bioanalytical Systems, Inc., West Lafayette, IN), and 0.05-µm alumina (Buehler, Lake Bluff, IL) polish. The diamond polishes were used with a Nylon polishing cloth, and the alumina polish was used with a Microcloth polishing cloth (Buehler). The electrodes were sonicated (Branson 2210, Branson Ultrasonic Corp.) for 30 s in DI water between each polish grit size and after the final polish (0.05 µm). For the limit of detection studies, the
Figure 1. Experimental setup for large-volume (1 mL) experiments with an electromagnet. The pinch clamp allows for a flow-through mechanism. The Lorentz force is generated by the interaction of the current and magnetic field, which are perpendicular to each other.
Pt flag counter electrode was held at 1.5 V versus another Pt flag in electrolyte solution to oxidize out any metals that may have deposited. Large-Volume Experiments. The experimental setup for large-volume experiments is shown in Figure 1 and is the same as the setup with which our early results were obtained.1 The electrodes were placed in a round-bottom culture tube located between the poles of an electromagnet (New England TechniCoil). The electromagnet was powered by an EMS 100-50 power supply (Lambda EMI) and cooled by an Affinity F-series chiller (Lydall). The magnet poles, with 7.5-cm diameter, were set to a 2.8-cm gap. A hole was drilled in the bottom of the culture tube to which Tygon tubing was epoxied. A pinch clamp was secured on the tubing to control flow from the culture tube that results from the pull of gravity. The working electrode faced downward, so that its surface normal was perpendicular to the magnetic field to achieve a maximum Lorentz force, inducing a MHD flow parallel to the electrode surface. The tip of the reference electrode was positioned beside the working electrode with the counter electrode residing on the bottom of the culture tube, ∼1 cm below the working and reference electrodes.41 The current in the magnet power supply was adjusted to reach the desired magnetic field strength: from 0 T at 0 A to 1.77 T at 50 A. The magnetic field strength was measured with a dc magnetometer (AlphaLab Inc., Salt Lake City, UT). At the beginning of the analysis, the culture tube contained 1 mL of deposition solution consisting of analytes (Pb2+, Cd2+, Cu2+, Tl+), Hg2+, Fe3+, and electrolyte (30 mM HNO3 and 0.1 M KNO3). After 60 s of deposition at -0.8 V, 3 mL of electrolyte was added to the tube and the pinch clamp was opened to drain the solution from the bottom of the tube back to the 1-mL level. This procedure was repeated at 70 and 80 s (with the working electrode still held at -0.8 V) to lower the concentration of Fe3+ in solution and thereby lower the cathodic background current that results from reduction of the high concentration of Fe3+. The stripping step (2 V/s for linear sweep ASV and open circuit for potentiometric analysis) commenced at 90 s in the remaining electrolyte solution. Small-Volume Experiments. The experimental setup for small-volume experiments involving two permanent sintered NdFeB magnets (Magnet Sales Co.) is shown in Figure 2. The magnet dimensions were 1 in. × 1 in. × 0.5 in. with a 1.23-T residual induction and 0.55 T on the surface. The large areas of the two magnets faced each other, and they were oriented so that they attracted each other and were separated by 0.5 in., resulting Analytical Chemistry, Vol. 78, No. 11, June 1, 2006
3747
Figure 2. Experimental setup for small-volume (100 µL) experiments with permanent magnets. The Lorentz force is generated by the interaction of the current and magnetic field, which are perpendicular to each other. (Drawn approximately to scale.)
Figure 3. Effect of different concentrations of Fe(NO3)3 on ASV using a 3-mm working electrode in the large-volume stripping setup at 0 (dashed curve) and 1.77 T (solid curve). Linear sweep (2 V/s) was performed after 60-s deposition in solutions containing Fe(NO3)3 and 1 mM Hg(NO3)2, 2 µM Cu(NO3)2, 2 µM Pb(NO3)2, 2 µM Cd(NO3)2, 30 mM HNO3, and 0.1 M KNO3 and after rinsing.
in a field of 0.62 T between them. The electrodes were placed in a 2-mL centrifuge tube. The centrifuge tube was positioned between the magnets with the working electrode centered between them and its normal perpendicular to the magnetic field. For experiments involving only one permanent magnet (not illustrated), the centrifuge tube was taped to only one of the magnets with the magnetic field oriented perpendicular to the working electrode normal and the south end immediately adjacent to the tube. At the beginning of the experiment for both magnet arrangements, the tube contained 100 µL of deposition solution. After 60 s of deposition at -0.8 V, 1 mL of electrolyte was added to dilute the analyte solution, with the potential still held at -0.8 V. The stripping step commenced at 70 s. RESULTS AND DISCUSSION Redox Species Concentration. The anodic stripping peaks that result from linear scan ASV of 2 µM Cu2+, Pb2+, and Cd2+ in solutions containing 1 mM Hg2+ and various concentrations of Fe3+ are shown in Figure 3. The large-volume experimental setup was used with the 3-mm working electrode. At 1 mM Fe3+, there is only a 6 ( 1% increase in peak areas at 1.77 T, compared to 0 T. However, the peak enhancement becomes larger at higher concentrations of Fe3+: 46 ( 5% increase at 10 mM Fe3+ and 159 ( 5% increase at 100 mM Fe3+. (The enhancements reported are the average peak area increases for Cu, Pb, and Cd. Each 3748 Analytical Chemistry, Vol. 78, No. 11, June 1, 2006
experiment was performed twice, and the uncertainties reported were determined by propagation of error analysis starting with peak area uncertainties equal to half the difference of the two measurements.) The cathodic current during the preconcentration step is higher when the Fe3+ concentration is higher. Thus, the Lorentz force is increased, which results in increased solution convection and signal enhancement. Using Fe3+ as the redox species, instead of Hg2+ as we previously reported,1 allows higher concentrations of redox species to be utilized. When Hg2+ was used, the maximum enhancement was observed at 40 mM Hg2+ because higher concentrations resulted in mercury film electrodes that were thick enough to display semi-infinite behavior. Due to the use of a low concentration of Hg2+, the peaks observed in Figure 3 are consistent with thin-film behavior and exhaustive oxidation of all analyte in the film, even at the highest concentration of Fe3+, 100 mM (and therefore highest convection). As previously reported by De Vries and Van Dalen,46 thin-film behavior is predicted under conditions in which the dimensionless parameter H ) nFvl2/(DRRT) e 1.6 × 10-3, where n, F, R, and T have their usual meanings; v is scan rate; l is thickness of the mercury film; and DR is the diffusion coefficient of the analyte in the mercury film. Given the parameters used in our experiments, thin-film behavior is expected (H e 1.6 × 10-3) when l e 0.12 µm. The thickness of the mercury film, l, was determined by stripping the Hg in electrolyte solution (data not shown). The deposition solution contained 1 mM Hg2+, 100 mM Fe3+, and electrolyte. The solution was diluted with aliquots of electrolyte (as in the other large-volume experiments) before the stripping step. The peak area (Coulombs) was used to calculate the volume of Hg deposited, which was then divided by the electrode area to give l. For the 3-mm electrode, l ) 2.2 nm at 0 T and 4.7 nm at 1.77 T. These thicknesses are well below the cutoff value of 0.12 µm, indicating that thin-film behavior is dominant under these conditions. When H e 1.6 × 10-3, the peak width at half-height, b1/2, is predicted to be independent of film thickness.46 Our data at the highest concentration of Fe3+ (100 mM) is consistent with this prediction. Unlike previous experiments1 using high concentrations of Hg2+ and thicker films, no significant change in b1/2 values were observed in the presence of a magnetic field. Also, under thin-film conditions, a shift in peak potential (toward more positive potentials with increasing l) is predicted as described by the equation n(Ep - E1/2) ) -1.43 + 29.580(log H), where Ep is the peak potential.46 Given the l values determined earlier (2.2 nm for 0 T and 4.7 nm for 1.77 T), a peak potential shift of +10 mV is predicted at 1.77 T, compared to 0 T. The average of the actual peak shift is +20 mV, larger than predicted by theory. The discrepancy between theory and experimental results is possibly due to inaccuracies in our determination of mercury film thickness. In the calculation of thickness, we assumed a uniform film over the entire electrode area, an assumption that we know from visual inspection of the mercury film and from published images47,48 of mercury films is inaccurate. Magnetic Field Strength. The peak area of 2 µM Pb2+ is shown as a function of magnetic field strength in Figure 4. The large-volume experimental setup was used with the 3-mm working (46) De Vries, W. T.; Van Dalen, E. J. Electroanal. Chem. 1967, 14, 315-327. (47) Wu, H. P. Anal. Chem. 1994, 66, 3151-3157. (48) Zakharchuk, N. F.; Saraeva, S. Y.; Borisova, N. S.; Brainina, K. Z. Electroanalysis 1999, 11 (9), 614-622.
Figure 4. Effect of magnetic field strength on ASV of Pb using a 3-mm GC working electrode in the large-volume stripping setup. Linear sweep (2 V/s) was performed after 60-s deposition in solutions containing 100 mM Fe(NO3)3, 1 mM Hg(NO3)2, 2 µm Pb(NO3)2, 30 mM HNO3, and 0.1 M KNO3 and after rinsing.
electrode, 1 mM Hg2+, 100 mM Fe3+, and the electromagnet at various field strengths. For each magnetic field strength, three stripping voltammograms were collected, and the standard deviation is also illustrated in Figure 4. At 0.12 T, a 17 ( 9% increase in peak area is observed. In the previous report using 40 mM Hg2+ as the redox species,1 no enhancement was observed at 0.12 T. However, by using Fe3+, higher concentrations of redox species can be employed, resulting in higher current and, therefore, larger Lorentz forces and larger ASV signal enhancements. This magnetic field strength is very easily achieved by simple permanent magnets, supporting the possibility of this method of convection to be used in portable analytical devices. The peak area enhancement continues to increase with increasing magnetic field due to the larger Lorentz force and resulting solution convection. At 0.55 T, a 72 ( 12% increase is observed, and at 1.77 T, the increase is 171 ( 7%. As predicted, the peaks in the voltammograms collected at varying magnetic field strengths (not shown) display thin-film behavior, as described in the previous section. There is no change in b1/2 values with increasing magnetic field strength. The peak potentials gradually shift to more positive potentials as the magnetic field strength is increased and the mercury film becomes thicker. For example, the Pb2+ peak shifts 15 mV more positive as the magnetic field is increased from 0 to 1.77 T. This experimental shift is larger than that predicted by theory (10 mV) but in the predicted direction. Electrode Size. The effect of working electrode size on the redox MHD enhanced ASV of 2 µM Cu2+, Pb2+, and Cd2+ is shown in Figure 5. The experiments were performed with the largevolume setup using 1 mM Hg2+ and 100 mM Fe3+. There is a decrease in redox MHD with a decrease in electrode size. This finding is consistent with other reports that have shown decreasing MHD effects with decreasing electrode size.45,49 Consequently, at the 10-µm electrode, there is no peak area enhancement at 1.77 T. Thus, there is a lower limit to the electrode size that would be useful for redox MHD-enhanced ASV. Larger electrodes result in significant enhancement. A 175 ( 27% increase is observed at the 1-mm electrode, and a 159 ( 5% increase is observed at 3 mm. (The enhancements reported are the average peak area increases for Cu, Pb, and Cd. Each experiment was performed (49) Mehta, D.; White, H. S. ChemPhysChem 2003, 4, 212-214.
Figure 5. Effect of working electrode size on ASV using the largevolume stripping setup at 0 (dashed curve) and 1.77 T (solid curve). Linear sweep (2 V/s) was performed after 60-s deposition in solution containing 100 mM Fe(NO3)3, 1 mM Hg(NO3)2, 2 µM Cu(NO3)2, 2 µM Pb(NO3)2, 2 µM Cd(NO3)2, 30 mM HNO3, and 0.1 M KNO3 and after rinsing.
twice, and the uncertainties reported were determined by propagation of error analysis starting with peak area uncertainties equal to half the difference of the two measurements.) The differing enhancements at the 1- and 3-mm electrodes can be understood in terms of the differences in convection at these two sizes of electrodes and represented by differences in the mercury film thicknesses. As previously described in this report, the thickness of the mercury film, l, for each electrode size was determined by stripping the Hg in electrolyte solution. For the 3-mm electrode, l ) 2.2 nm at 0 T and 4.7 nm at 1.77 T; for the 1-mm electrode, l ) 3.6 nm at 0 T and 7.7 nm at 1.77 T; and for the 10-µm electrode, l ) 63 nm at 0 and 1.77 T. The increased film thickness at 1.77 T (compared to 0 T) for the 1- and 3-mm electrodes is expected due to the increased convection, which enhances the transport of Hg2+ as well as analytes. The thicker films observed at the 1-mm electrode (compared to the 3-mm electrode) indicate that, for the 60-s deposition used, radial diffusion makes a significant contribution at the 1-mm electrode. However, the film at the 10-µm electrode is much thicker than at the larger electrodes because radial diffusion is the dominant mode of mass transport. The magnetic field causes no change in film thickness at the 10-µm electrode because the effects of redox MHD are negligible at this electrode size under these conditions. Potentiometric Stripping Analysis. The results of PSA of 2 µM Pb2+ are shown in Figure 6. The experiments were performed with the large-volume setup using a 3-mm working electrode, 5 mM Hg2+, and 100 mM Fe3+. The rinsing steps in these experiments lower the concentration of the oxidants, Hg2+ and Analytical Chemistry, Vol. 78, No. 11, June 1, 2006
3749
Figure 6. PSA of 2 µM Pb(NO3)2 using a 3-mm GC working electrode in the large volume stripping setup at 0 (dashed curve) and 1.77 T (solid curve). PSA was performed after 60-s deposition at -0.8 V in solution containing 100 mM Fe(NO3)3, 5 mM Hg(NO3)2, 2 µM Pb(NO3)2, 30 mM HNO3, and 0.1 M KNO3 and after rinsing.
Fe3+, in solution before the stripping step. Thus, the transition times are increased, improving the sensitivity of the analysis. Using a magnetic field strength of 1.77 T, there is a 200 ( 11% increase in transition time, compared to 0 T. (Each experiment was performed twice, and the uncertainty reported was determined by propagation of error analysis starting with transition time uncertainties equal to half the difference of the two measurements.) This transition time, the time required for chemical oxidation of the Pb out of the mercury film, is a function of the amount of Pb deposited in the mercury film electrode. An increase in the transition time indicates an increase in solution convection during the preconcentration step. Therefore, redox MHD is able to provide signal enhancement of PSA as well as ASV. The larger signal enhancement observed for PSA, as compared to ASV (200% for PSA compared to 159% for ASV), can be at least partially explained by the use of a higher concentration of Hg2+. The concentration of Fe3+ used in both the PSA and ASV experiments was 100 mM. However, for the PSA experiments, the concentration of Hg2+ was 5 mM, compared to 1 mM Hg2+ for the ASV experiments. This corresponds to a 4% increase in total concentration of redox species (Hg2+ and Fe3+). Thus, it is expected that there will be a slight increase in enhancement with the magnetic field due to the higher Lorentz force caused by the higher concentration of redox species. Although the increase in enhancement is greater than what would be predicted based solely on the small increase in redox species concentration, we believe the method of convection for PSA is the same as that in ASV analyses. Thus, we anticipate similar enhancements for all stripping techniques that benefit from increased convection during the deposition step. Small-Volume Experiments. The stripping voltammograms of 2 µm Cu2+, Pb2+, and Cd2+ that were collected using the smallvolume setup are shown in Figure 7. The experiments were performed with the 1-mm GC electrode, 1 mM Hg2+, and 100 mM Fe3+. When only one permanent magnet was used, resulting in a 0.55-T magnetic field strength, the average increase in peak areas was 29 ( 4%. When two permanent magnets in an attracting arrangement were used, resulting in 0.62 T, there was a 43 ( 5% increase in the average peak areas. (Each experiment was performed twice, and the uncertainties reported were determined by propagation of error analysis starting with peak area uncertainties equal to half the difference of the two measurements.) The 3750 Analytical Chemistry, Vol. 78, No. 11, June 1, 2006
Figure 7. Effect of permanent magnet(s) on ASV using a 1-mm GC working electrode in the small-volume stripping setup at 0 T (dashed curve), using one permanent magnet at 0.55 T (dotted curve), and using two permanent magnets at 0.62 T (solid curve). Linear sweep (2 V/s) was performed after 60-s deposition in 100 mM Fe(NO3)3, 1 mM Hg(NO3)2, 2 µM Cu(NO3)2, 2 µM Pb(NO3)2, 2 µM Cd(NO3)2, 30 mM HNO3, and 0.1 M KNO3 and after diluting with 1 mL of electrolyte.
peak potentials and b1/2 values are consistent with those described in previous sections. These results indicate that redox MHD could be used for the enhancement of on-site stripping analyses. Using portable, compact permanent magnets eliminates the need for a large electromagnet. Limits of Detection. Calibration curve and detection limit data for ASV of Tl+ and Cd2+ in the large-volume setup and Cd2+ in the small-volume setup are given in Table 1. The detection limits were not determined for Pb2+ or for Cu2+ due to the presence of nonzero background signals for these analytes, arising from heavy metal contamination in the Hg(NO3)2. The calibration plots for Cd2+ and Tl+(not shown) were linear in the concentration range of the experimental limit of detection through 2 µM (the highest concentration analyzed). The experimental limits of detection were the lowest concentrations of analytes resulting in experimental peak currents greater than 3 times the standard deviation of the background. The calculated limits of detection were obtained using the equation CL ) 3σb/m, where CL is the detection limit, σb is the standard deviation of the background current (obtained experimentally), and m is the sensitivity of the peak current.50 For the large-volume experiments, the peak area sensitivities for Cd2+ and Tl+ were increased by an average of ∼193 ( 14%, and calculated and experimental limits of detection were decreased by an average of ∼68 ( 5% using a magnetic field strength of 1.77 T. Given a sensitivity increase of 193%, it can be predicted that the detection limit with a magnet will be reduced to approximately one-third the value of that without a magnet. This predicted reduction corresponds well with the 68% decrease we observed in detection limit. For the small-volume experiments, the peak area sensitivity of Cd2+ was increased by ∼54 ( 7%, and the calculated and experimental limits of detection were decreased by an average of ∼52 ( 9%. Given the 54% increase in sensitivity, the detection limit in the presence of the magnetic field is expected to be approximately two-thirds of the detection limit in the absence of a magnetic field. This predicted detection limit reduction corresponds with the 52% decrease we observed. The voltammograms of 5 nM Cd2+ in the large-volume setup using the 3-mm electrode, 1 mM Hg2+, and 100 mM Fe3+ are (50) Rubinson, K. A.; Rubinson, J. F. Contemporary Instrumental Analysis; Prentice Hall: Upper Saddle River, NJ, 2000.
Table 1. Linear Regression Analysis (y ) mx + b) Data for Calibration Curves and Detection Limit Data for Linear Sweep (2 V/s) ASVa large volume
small volume
Tl+
sensitivity (m) based on peak area (nC/nM) y-intercept (b) based on peak area (nC) R2 based on peak area sensitivity (m) based on peak current (nA/nM) y-intercept (b) based on peak current (nA) R2 based on peak current calculated limit of detection (nM) experimental limit of detection (nM)
Cd2+
Cd2+
0T
1.77 T
0T
1.77 T
0T
0.62 T
0.25 ( 0.01 1.6 ( 3.1 0.999 5.8 ( 0.1 260 ( 92 0.999 42 ( 1 25
0.66 ( 0.02 9.9 ( 15.9 0.995 13 ( 1 1260 ( 890 0.988 18 ( 1 10
0.29 ( 0.03 -1.6 ( 9.2 0.973 8.6 ( 1.1 120 ( 320 0.967 28 ( 4 25
0.93 ( 0.02 4.4 ( 5.4 0.998 36 ( 1 -130 ( 210 0.999 6.6 ( 0.2 5.0
0.061 ( 0.003 1.5 ( 3.7 0.994 2.0 ( 0.3 320 ( 370 0.950 120 ( 20 125
0.094 ( 0.002 -0.77 ( 1.88 0.997 3.5 ( 0.2 160 ( 280 0.991 69 ( 4 50
a With 60-s deposition using a 3-mm GC for the large-volume setup and a 1-mm GC for the small-volume setup. Analyte solutions contained 100 mM Fe(NO3)3, 1 mM Hg(NO3)2, 30 mM HNO3, and 0.1 M KNO3.
Figure 8. ASV analysis of 5 nM Cd2+ using a 3-mm GC working electrode in the large-volume setup at 0 (dashed curve) and 1.77 T (solid curve). Linear sweep was performed after 60-s deposition in solution containing 100 mM Fe(NO3)3, 1 mM Hg(NO3)2, 5 nM Cd(NO3)2, 30 mM HNO3, 0.1 M KNO3 and after rinsing.
shown in Figure 8. The effectiveness of redox MHD is demonstrated by the appearance of a peak only when the experiment is performed in a magnetic field. At 0 T, the stripping peak is below the limit of detection, but at 1.77 T, the peak is observed. The ability to achieve this detection limit with only a 60-s preconcentration step and using only a simple glassy carbon disk electrode makes redox MHD-enhanced ASV competitive with commercially available analysis systems. Other systems require costly, specialized electrodes39,40 or lengthy preconcentration times.37
CONCLUSIONS The potential for redox MHD-enhanced ASV to be used in portable analytical devices has been demonstrated. The induced solution convection resulted in signal enhancements in volumes as small as 100 µL and with magnetic field strengths low enough to be achieved by small permanent magnets. MHD was successfully used to increase sensitivities and decrease limits of detection. While greater enhancements were observed in the large-volume setup than in the small-volume setup, it is because of the much larger magnetic field strength that is possible with a large electromagnet than with a small permanent magnet. Thus, ongoing work involves increasing the strength of the permanent magnets used while maintaining the small size, so that a portable instrument with a greater enhancement is possible. ACKNOWLEDGMENT Research was supported through the National Science Foundation (CHE-0096780). E.C.A. acknowledges a F.M. Beckett Summer Fellowship from the Electrochemical Society, Inc., and an American Chemical Society Division of Analytical Chemistry Graduate Fellowship, sponsored by Procter & Gamble. We thank Mr. Jerry Homesley for electronics assistance. Received for review January 1, 2006. Accepted March 24, 2006. AC060001V
Analytical Chemistry, Vol. 78, No. 11, June 1, 2006
3751
