Anal. Chem. 1993, 65, 1529-1532
1529
Mercury-Free Disposable Lead Sensors Based on Potentiometric Stripping Analysis at Gold-Coated Screen-Printed Electrodes Joseph Wang’ and Baomin Tiant Department of Chemistry, New Mexico State University, Las Cruces, New Mexico 88003
Gold-coated screen-printed electrodes offer reliable quantitation of trace lead in connection with potentiometric stripping analysis (PSA). Such replacement of mercury-based sensors, with goldcoated ones, avoids environmental contamination associatedwith the disposal of mercury electrodes in connection with large-scale screening for lead poisoning. The PSA operation obviates the need for oxygen removal, offers low background contributions, and minimizes surfactant interferences. Changes in the peak intensity and position (vs mercury-coated strips) offer new selectivity dimensions. Various experimental parameters are optimized to allow convenient monitoring of micrograms per liter lead concentrations following short deposition periods. Applicability to urine and drinking water samples is illustrated. The highly stable response of these screen-printed electrodes makes them very attractive for both single-use and multiple applications.
INTRODUCTION Stripping analysis is an extremely powerful technique for measuring trace metals in environmental, clinical, or industrial samples.’ Its analytical power accrues from the unique coupling of an effective (in situ) preconcentration step with an advanced measurement scheme of the deposited metal. Mercury-based (hanging drop or thin film) electrodes have been traditionally employed for achieving high sensitivity and reprodu~ibility.~~~ However, the growing interest in applying stripping analysis for decentralized clinical screening or for on-site environmental monitoring has prompted the search for new “mercury-free”electrodes for these tasks. There is no fundamental reason to adhere to the notion that only mercury electrodes offer reliable and sensitive monitoring of trace metals. This paper describes a disposable gold-coated carbon sensor for trace measurements of lead. There are growing needs for developing a highly sensitive, reliable and portable lead analyzer to support mass screening programs aimed at preventing lead poisoning in ~ h i l d r e n . The ~ , ~ high reliability inherent to stripping measurements of lead blood in the
* Corresponding author. On leave from Sichuan University, Chengdu61064, Peoples’ Republic of China. (1) Wang, J. Stripping Analysis: Principles, Instrumentation, and Applications; VCH Publishers: Deerfield Beach, FL, 1985. ( 2 ) Peterson, W. Am. Lab. 1979, 12, 69. (3) Florence, T . M. J . Electroanal. Chem. Interfacial Electrochem. 1970, 27, 273. (4) Loeckx, R. Anal. Chem. 1986,58, 174A. ( 5 )Anal. Chem. 1990, 62, 823A. +
0003-2700/93/0365-1529$04.00/0
central laboratory: coupled with their suitability for on-site assays, has prompted increasing efforts toward the adaptation of stripping procedures for lead blood screening programs. The recent adaptation of screen-printed electrodes for stripping measurements of trace metals’ should greatly facilitate decentralized testing for lead. These extremely lowcost, mass-produced electrodes function in a manner comparable to traditional stripping electrodes and can be used as disposable sensing devices. However, due to the toxicity of mercury, it is highly desired to explore other electrode materials (particularly when single-use mass screening applications are concerned). During our evaluation of various alternative (“mercury-free”) electrodes, we have noticed the attractive potentiometric stripping behavior of lead at the gold-coated carbon strip electrode. Gold film electrodes have been traditionally employed for stripping measurements of trace metals (e.g., Se, As, Te) more electropositive than mercury.’~8~9The combination of gold-coated carbon strips with PSA yields an analytically attractive behavior, as compared to many earlier unsucessful attempts for monitoring lead without the involvement of mercury. In the following sections, we will thus describe the development of gold-coated screen-printed electrodes for decentralized testing of trace lead.
EXPERIMENTAL SECTION Apparatus. A TraceLab potentiometric stripping unit (PSU 20, Radiometer, Denmark) with a SAM20 sample station (Radiometer) and an IBM PS/2 55SX was used to obtain the potentiograms. The experiments employed the conventionalAg/ AgCl reference and platinum wire auxiliary electrodes of the TraceLab unit. These electrodes and the screen-printed working electrode joined the 20-mL cell by holes in its cover. The screenprinted electrodes (ExacTech Blood Glucose Strips, Medisense Inc.) were purchased from a local drugstore. The design and dimensionsof these electrodes were described in detail earlier.7J0 One printed carbon contact (2 X 8 mm) of the strip served as the substrate for the gold film. A fast-setting epoxy (Cole Parmer) was used to prevent solution creepingover areas of the conductive carbon track covered with the dielectric layer. Reagents. Deionized water was used to prepare all solutions. Stock solutions of lead and other metals (lo00 mg/L, atomic absorption standard, Aldrich) were diluted daily as required. The supporting electrolyte was 0.05 M acetate buffer solution (pH 4.5). The urine samples were obtained from a healthy volunteer. Drinking water samples were collected at the laboratory. Procedure. The gold film was preplated from a nondeaerated, stirred, 50 mg/L gold solution (in 0.25 w t % HCl) by holding the carbon strip electrode at -0.40 V for a 20-min period. The potential was then switched to +0.70 and held there for 5 min. Subsequently, the gold-coated strip was dried in air and stored overnight prior to use. (6) Boone, J.; Hearn, T.; Lewis, S. Clin. Chem. 1979, 25, 383. (7) Wang, J.; Tian, B. Anal. Chem. 1992, 64, 1706. (8) Copeland, T. R.; Skogerboe, R. K. Anal. Chem. 1974,46, 1257A. (9) Posey, R.; Andrew, R. Anal. Chim. Acta 1980, 119, 55. (10) Green, M.; Hilditch, P. Anal. Proceed. 1991, 28, 374. 0 1993 American Chemical Society
isao
ANALYTICAL
CHEMISTRY, VOL. 65, NO. ii,JUNE 1, 1993
“‘1
a
i
-0.2
I--------
P
SN
,
-0.4
,
,
r
,
-0.6
,
,
i
-0.8
POTENTIAL (V>
I
20
,
.
,
CURRENT (HA)
Flguro 2. Effect of preconcentratlon potential (A) and -ping current (8)upon the potmtbmtrlc strlpphg response for 40 pg/L b8d at the ao#coated carbon strip etlect~ode.PreconcentraM for 4 (A) and 3 (B) mln. Other ccnKMcm as in Figure 1B.
teristics are not compromised by the use of the “mercuryfree” surface. It is possible ale0to use square-wave voltammetryto follow the lead stripping process at the gold-coated screen-printed electrode. The resulting lead voltammetrk peak was broader (than the corresponding PSA one) and was &companied by a rising bac4ground current contribution associated with dissolved oxy&n and the gold surface reactions (not shown). Cmvenient measurements were thus possible only for lead concentrationshigher than 5 rcg/L (1-min deposition). Some improvements were achieved by using a background subtraction operation (utilizingdifferent preconcentration times). Ftguro 1. Stripping TSIfor 20 ked at the mercuryYet, the PSA operation yielded a sharper and more repro(A) and gold- (B) coated carbon strlp ekotrodes. Fov-mtnute ducible response and improved signal-to-noise characteristics. ~ a t - 0 . 7 V h w n a s t l r e d , ~ ~ ~ O . O I M a c e t a t e The PSA response was not compromised upon using acidic Wer (pH 4.5) 6dutbn. Strjppina clwT(Mt, +20 PA. media (e.g., HCl or “ 0 3 down to pH 1.51, aa often required in many practical applications. A more positive deposition Potentiometric stripping assays were carried out in the potential (of -0.45 V) was used in these media to minimize following manner. The 20-mL non-deaerated sample was ininterference from the hydrogen evolution reaction. The gold troduced to the cell, and the deposition potential (-0.70 V) was film can also be prepared in situ, by codeposition with the imposed on the gold-coated working electrode. Deposition periodsrangingbetween0.5and lOminwereemployed,depending target lead ions. Such operation, while yielding a well-defined on the lead concentration and the mode of mass transport. lead response, requires 5-6 deposition/stripping cycles before Following the deposition and a 10-8rest period, the potentiogram the peak is stabilized. Subsequent work was carried out with was recorded by applying a constant oxidation current 20 PA, preplated gold films, a PSA operation, and an acetate buffer wing a quiescent solution and the BASE 5 baseline-correcting solution (of pH 4.5). command (of the TraceLab software). The electrode was Various procedure parameters, such as the deposition conditioned at +0.80 V for 50 s prior to the next measurement potential and the stripping current, have a profound effect cycle. Mercury-coated carbon strip electrodes, used for comon the potentiometric stripping response of the gold-coated parison purposes, were prepared and operated in accordance to strip electrode (Figure 2). For example, the stripping peak ref 7. All data were obtained at room temperature. for 40 c(BIL lead increases rapidly upon increasing the deposition potential between -0.2 V and -0.4 V, and then RESULTS AND DISCUSSION more slowly (Figure 2A). The tesponse decreases rapidly Characterization and Optimization. Due to various upon increasing the stripping current between 5 and 20 PA surface processes (e.g., formationof oxide or hydrogen layers) and then more slowly (Figure 2B). A potential of -0.7 V and noble metals, such as gold or platinum, have rarely been used a stripping current of 20 pA were thus selected for all for sensitive and reproducible stripping measurements of subsequent work. As expected, the response is also strongly metals more electronegative than mercury.9 However, as dependent upon the preconcentration period. Figure 3A shows stripping potentiograms for lOpg/L lead after different illustrated below, computerized potentiometric Stripping preconcentration times (2-10 min, traces a-e). Convenient analysis (PSA) successfully addresses these background processes and offers reliable trace measurements of lead at quantitation, with favorable signal-to-background characgold-coated screen-printed electrodes. Figure 1 illustrates teristics, is ohrved for all these periods. The peak increases stripping potentiograms for a non-deaerated 20 pg/L (ppb) linearly with the preconcentration time, as indicated alsofrom lead solution, obtained (under the same conditions) at the the resulting responseversus time plot (shown as inset; slope, mercury-(A) and gold-(B) carbon strip electrodes. The gold11.7 ma/min). Deposition periods required for a well-defined coated screen-printed electrode yields a well-defined lead lead peak range from 1min for 10 pg/L to 10 min for 1pg/L. peak, with a peak potential of -0.05 V and peak half width Analytical Performance. The response of the goldof 67 mV (compared to -0.49 V and 50 mV, respectively, at coated carbon strip electrodes is highly reproducible. This the mercury-coated electrode). Coupled with the flat backwas evaluated from two series of 20 repetitive measurements ground response (inherent to the computerized PSA operof 20 ~rglLlead following4- and lBmin depoeition from stirred ation), the gold-coated screen-printed electrode offers conand unatirred solutions, respectively (not shown). The lead venient measurements of microgram per liter concentrations stripping peak remained unchanged throughout these profollowing a short (4-min) depositionperiod. Overall, the data longed series (of 100 and 220 min). The relative standard of Figure 1B indicate that the signal-to-background characdeviations over the complete series were 2.5 and 3.2%. Such
POTENTIAL (V)
ANALYTICAL CHEMISTRY, VOL. 65, NO. 11, JUNE 1, 1993
A
r
I I
1591
Bi:
1
Flgu.9. (A) Strlpplngpotentlograms for 10 pg/L lead following different preconcentratbn periods: 2 (a),4 (b), 6 (c),8 (d), and 10 (e) mln. Also shown (Inset), the resulting response vs time plot. (B) Strlpplng potentiograms obtalned wlth lncreaslng lead concentration In steps of 5 Mg/L (a-e). Also shown (Inset), the resulting cailbratlon plot. Condltlons as In Figure 1B.
good precision indicates reproducible lead plating onto the gold film, as well as complete removal (stripping) of the deposited metal. Hence, in addition to single-use applications, the gold-coated carbon strip electrodes hold great promise as reusable electrodes for both centralized and decentralized stripping operations. Indeed, the same gold-coated electrode functioned in a highly stable manner over a 2-week period, performing over 60 lead (10 pg/L) runs, with no apparent loss of the response (RSD = 2.8%). Analogous measurements at the mercury-coated strip electrode exhibited a gradual decrease of the lead peak (after the second day). Apparently, the gold layer is stronglyadhered to the carbon strip substrate and is chemically stable (in the presence of oxygen). The electrode-to-electrode reproducibility is also good (RSD = 6.1% 1, as was indicated for measurements of 20 pg/L lead on six different electrodes. The analytical utility is based on the linear correlation between the stripping response and the lead concentration. Stripping potentiograms for lead solutions of increasing concentration (5-25 pglL), following a 4-min deposition, are shown in Figure 3B. The well-defined peaks allow convenient quantitation at the microgram per liter level. The data of Figure 3B yielded a linear calibration plot (shown in inset). Least-squares treatment of these data gave a slope of 4.52 ms.L/pg (correlation coefficient,0.999). A shorter deposition period (30s) was used in connection with larger concentration increments over the 50-600 pg/L range. Under these conditions, linearity prevailed up to 250 pg/L, with a curvature at higher levels (not shown). The linear range was extended to 600 pg/L when an unstirred solution was employed. However, the absence of convective transport during the deposition resulted in ca. 80% loss in the sensitivity. While quiescent solutions yield defined peaks for low micrograms per liter lead concentrations upon using longer deposition periods (e.g., 10min), more efficient deposition from unstirred solutions would be desired to address the speed requirement of large-scale screening programs. Screen-printed microelectrode arrays (e.g., multiple microbands) should be valuable to impart effective(nonlinear)mass transport conditions from quiescent solutions. Overall, detection limits of 0.1 and 0.6 pg/L lead were estimated following 20- and 4-min deposition, respectively, from stirred solutions, while a 3 pg/L value was calculated for the use of unstirred solution for 10 min. Due to changes in the position and intensity of the stripping peaks at the gold-coated electrode (versustraditional mercury surfaces), it is important to evaluate the selectivity of the lead measurement. Figure 4 displays stripping peaks for lead
Figure 4. Potentiometric Stripping response for various trace metals (at 20 pg/L), as obtained at the gold- (A) and mercury- (B) coated strip electrodes: (a) lead, (b) copper, (c)cadmlum, (d) thallium, and (e) tin. Deposition for 4 (A) and 2 (B) mln, followed by stripping wlth currents of +20 and +2 pA, respectively. Other condltbns as In Figure 18. (a), copper (b), cadmium (c), thallium (d), and tin (e) at the gold- (panel A) and mercury- (panel B) coated carbon strip electrodes. These data indicate some overlap between the lead response and the cadmium and thallium peaks at the gold-coated electrode. Note, however, that the intensity of these peaks (versus that of lead) is significantly smaller compared to those observed at the mercury-coated surface. Coupled with the fact that the “normal”blood cadmium level is significantly (300-fold)lower than that of lead,” no major interference is expected. Note alsothat the gold film electrode is not responsive to tin, as compared to the large tin signal observed at the mercury surface. Defied peaks were obtained a t the gold-coated electrode for bismuth and antimony (both at +0.18 V),withno effect upon theleadresponse (not shown). Overall, the above data indicate that other amalgam-forming trace metals can be monitored with the “mercury-free”sensors. A common problem of stripping measurements of trace metals is the interference effect caused by surface-active organic constituents of the sample matrix. However,the PSA operation is usually less sensitive to the presence of low levels of surfactants.12 Such a feature is maintained a t the goldcoated surface. For example, only 1,2, and 18% depressions of the 20 pg/L lead peak were observed in the presence of 6 mg/L sodium dodecyl sulfate, gelatin, and albumin, respectively (4-min deposition). A 12% enhancement of the peak (coupled with a distorted shape) was observed in the presence of 6 mg/L Triton X-100. Similar changes were observed for analogous experiments at the mercury-coated strip electrode. Figure 5 illustrated the suitability of the gold-coated screenprinted electrode for the determination of lead in clinical and environmental samples. Two standard additions (of 5 and (11) Ostapczuk, P. Clin. Chem. 1992, 38, 1995.
(12)Jagner, D. Analyst 1982, 107,593.
1182
ANALYTICAL CHEMISTRY, VOL. 65, NO. 11, JUNE 1, 1993
should possess a well-defined stripping behavior (with no overlap to coexisting metals). One such possibility (for blood assays) is tellurium, which yields a well-defined peak at the gold-coated electrode around +0.600 V.
E
CONCLUSIONS
CONCENTRATION lppb) C
H LLW
0.1 -01 -0 3
00
-02
POTENTIAL (V) Fi@uro1. Moaswements of lead in vine (A) and drlnking water (B). (a) Potentbprn for the sample; (b, c) successive concentratbn incrementsof 5 (A) and 2.5 (B) pg/L lead. Reconcentramfor 0 (A) and 10 (B) min. Samples: 19.86 ml of vlne 0.15 mL of 0.09 M HCI (final pH 1.8) (A); 19.00 mL of drh?klng water 0.40 mL of 0.05 M acetate buffer (pH 4.5). Deposltlonfor 0 (A) and 10 (B) min at -0.45 and -0.70 V, respecthrely. OUw conditkns as in Figure 18.
+
+
2.5 pg/L lead) to the urine (panelA) and drinkingwater (panel B) samples, respectively, resulted in well-defined stripping peaks (traces b and c). The lead peak for the original sample (trace a) can thus be quantified by means of the resulting standard addition plots (alsoshown). The reaultinglead levels in these samples correspond to 2.0 (panel A) and 1.1 pg/L (panel B). Only a simple pretreatment (pH adjustment) of these samples was required. While a standard additions scheme was employed in Figure 5, a faster quantitation (needed for decentralized testin@)could rely on the use of an internal standard (with the lead concentration being calculated from the ratio of the lead-to-standard peaks). The standard should not be commonly present in the sample and (13) Craton, D.;Jonas, C.;Williams, D.;Mum, N. Talanta 1991,38, 17.
We have demonstrated that the use of a gold coating (instead of mercury f i i ) does not affect the reliability and sensitivity of potentiometricstripping measurementsof trace lead. Such single-use "mercury-free" carbon strip electrode8 hold great promise for maas screening programs (related to childhood lead poisoning) as it eliminates environmental problems associated with the disposal of mercury sensors. Computerized PSA l m been ~ shown very attractive for such operation, as it obviate the need for oxygen removal, m i n i surfactant effeds, and offera low background contributions. Yet, reliable on-sitemonitoring of lead blood levels would require the introduction of a simple, rapid and effective sample pretreatment step. Coverage of the new electrodes with an appropriate permeelective f i a d the development of screen-printed microelectrode array@ should alsofacilitatesuch allplication. Other decentralized analytical applications (dated to environmental surveillance or industrial quality control) would benefit Rom the introduction of disposable YmercUry-freenlead sensors. The inherent stability of these gold-plated stripping electrodea indicates also great promise as reusable electrodes for routine (centralized) PSA measurements. ACKNOWLEDGMENT
Thispublication waa supported by a grantfrom the Centers for Disease Control (Contract CCR 608616-01). Its contents are solelythe responsibility of the authors and do not represent the official views of the CDC.
RECEIVEDNovember
23, 1992. Revised manuscript received January 19,1993. Accepted January 21,1993.